The History of Bread Production Using LC
Uploaded by
sahremmanThe History of Bread Production Using LC
Uploaded by
sahremmanArchived at http://orgprints.
org/15519
DIASreport
DIASreport•AnimalHusbandry•no.61•October2004•LifeCycleAssessmentintheAgri-foodsector
LifeCycleAssessmentintheAgri-foodsector
Proceedingsfromthe4thInternationalConference,
DIASFoulum Unitsatotherlocations
October6-8,2003,Bygholm,Denmark
ResearchCentreFoulum
P.O.Box50,DK-8830Tjele Dept.ofVarietyTesting
Tel.+4589991900.Fax+4589991919 Teglværksvej10,Tystofte
[email protected].www.agrsci.dk DK-4230Skælskør NielsHalberg(ed.)
Tel.+4558160600.Fax+4558160606
BoardofDirectors
Administration AskovExperimentalStation
Vejenvej55,DK-6600Vejen
Dept.ofFoodScience Tel.+4575360277.Fax+4575366277
Dept.ofAnimalBreedingandGenetics
Dept.ofAnimalNutritionandPhysiology ExperimentalStationforOrganicFarming
Dept.ofAnimalHealthandWelfare Rugballegaard
Dept.ofAgroecology P.O.Box536
DK-8700Horsens
Dept.ofFarmManagementandResearchFacilities Tel.+4576296000.Fax+4576296102
CommunicationDepartment
CentreManagement,Foulum Foulumgaard
P.O.Box50
DIASAarslev DK-8830Tjele
ResearchCentreAarslev Tel.89991900.Fax+4589991919
Kirstinebjergvej10,DK-5792Aarslev
Tel.+4563904343.Fax+4563904390 JyndevadExperimentalStation
DIASreport•AnimalHusbandry•no.61•October2004
Flensborgvej22,DK-6360Tinglev
Dept.ofHorticulture Tel.+4574648316.Fax+4574648489
DIASFlakkebjerg
ResearchCentreFlakkebjerg
Flakkebjerg,DK-4200Slagelse
Tel.+4558113300.Fax+4558113301
Dept.ofPlantBiology
Dept.ofCropProtection
CentreManagement,Flakkebjerg
DIASBygholm
ResearchCentreBygholm
Schüttesvej17,P.O.Box536
DK-8700Horsens
Tel.+4576296000.Fax+4576296100
Dept.ofAgriculturalEngineering
CentreManagement,Bygholm
DPILSorgenfri
DanishPestInfestationLaboratory
Skovbrynet14,DK-2800Kgs.Lyngby,
Tel:+4545878055.Fax:+4545931155,
DIASreport•AnimalHusbandry•no.61•October2004
LifeCycleAssessmentintheAgri-foodsector
Proceedingsfromthe4thInternationalConference,
October6-8,2003,Bygholm,Denmark
NielsHalberg(ed.)
DanishInstituteofAgriculturalSciencesDepartmentofAgroecology
DIAS reports primarily contain research results and
trialstatementsaimedatDanishconditions.Also,the
reports describe larger completed research projects
oractsasanappendixatmeetingsandconferences.
DIASreportsarepublishedintheseries:
Plantproduction,AnimalHusbandryandHorticulture.
Prices:
Upto50pages:55,-kr.pr.stk.
Upto75pages:85,-kr.pr.stk.
Morethan75pages:110,-kr.pr.stk.
Subscribers obtain 25% discount. Subscription can
betakenoutbycontacting:
DanishInstituteofAgriculturalSciences
P.O.Box50,DK-8830Tjele
Tlf.+4589991028
AllDIASpublicationscanbeorderedontheinternet:
www.agrsci.dk
Print:www.digisource.dk
ISSN1397-9892
Contents
Preface 7
Life cycle assessment of bread production - a comparison of eight different scenarios - 9
Braschkat, J., Patyk, A., Quirin, M., & Reinhardt, G. A. 9
When a hole matters - the story of the hole in a bread for French hotdog 17
Rosing, L. and Nielsen, A.M. 17
Possible benefits from using LCA in the agro-food chain, example from Arla Foods 23
Inger Larsson 23
Environmental Management Practices in an Italian Coffee Company using LCA
Methodology 24
Roberta Salomone 24
Slaughter offal to feed or to fuel - a comparison of systems in terms of energy balances,
land use and emissions 34
Hans B. Wittgren, Helena Elmquist and Ingrid Strid Eriksson 34
Environmental assessment of contrasting pig farming systems in France 45
Basset-Mens, C. and van der Werf, H.M.G 45
LCA of Animal Products from Different Housing Systems in Switzerland: Relevance of
Feedstuffs, Infrastructure and Energy Use 55
Erzinger, S., Dux, D., Zimmermann, A. and Badertscher Fawaz, R. 55
Qualitative Life Cycle Assessment of Thai Shrimp Product 64
Mungkung, R. and Clift, R. 64
Life Cycle Assessment of trout farming in France: a farm level approach 71
Papatryphon, E., Petit, J., Van der Werf, H. M. G. and Kaushik, S. J. 71
Environmental impacts from Danish Fish Products 78
Mikkel Thrane 78
Operation specific engine load pattern and exhaust gas emission data from vehicles used
in typical Swedish agricultural operations 89
M. Lindgren; O. Pettersson; O. Norén; P.-A. Hansson 89
An LC inventory based on representative and coherent farm types 98
Randi Dalgaard, Niels Halberg, Ib Sillebak Kristensen and Inger Larsen 98
3
The ecoinvent database: use for the agri-food sector 107
Nemecek, T., Erzinger, S. and Frischknecht, R. 107
Life Cycle Inventory of the Galician dairy sector 118
Hospido, A., Moreira, M.T. and Feijoo, G. 118
Identification of processes affected by a marginal change in demand for food products -
two examples on Danish pigs and cheese 127
Anne Merete Nielsen, Per H. Nielsen, Jørgen Dejgård Jensen, Martin Andersen and Bo P.
Weidema 127
Land occupation and transformation in the Swiss life cycle inventory database ecoinvent
2000 135
Jungbluth, N., Frischknecht, R. 135
A new land use impact assessment method for LCA: theoretical fundaments and field
validation 143
Peters, J., García Quijano, J., Content, T., Van Wyk, G., Holden, N.M., Ward, S.M. &
Muys, B. 143
Soybean Production in Argentina: New Technologies, agricultural environment and
socioeconomic impacts. The implications for the future 157
Walter A. Pengue 157
How may Quality Assurance Systems in food chains include environmental aspects
based on Life Cycle Methodology? 168
Niels Halberg 168
Sustainability in the Agrofood Sector 182
Klaas Jan Kramer & Marieke Meeusen 182
Life Cycle-Based Sustainability Indicators for Assessment of the U.S. Food System 190
Keoleian, G.A. and. Heller, M.C. 190
Most significant substances of LCA to Mediterranean Greenhouse Horticulture 199
Antón, M.A.; Castells, F.; Montero, J.I.; Muñoz, P. 199
Using LCA for the Improvement of Waste Management in Greenhouse Tomato
Production 205
Muñoz, P.; Antón, A.; Montero, J.I.; Castells, F. 205
LCA of the integrated production of oranges in the Comunidad Valenciana (Spain) 210
N. Sanjuán, G. Clemente and L. Úbeda 210
Environmental Impact Assessment at Commercial Dairy Farms 214
De Boer, I.J.M., Iepema, G., and Thomassen, M.A. 214
4
A systematic description and analysis of GHG emissions resulting from Ireland’s milk
production using LCA methodology 219
Casey, J.W, Holden, N.M. 219
Life cycle assessment results and related improvement potentials for oat and potato
products as well as for cheese 222
Katajajuuri, J.-M., Virtanen, Y., Voutilainen, P. and Tuhkanen, H.-R. 222
Effects of Intensification of Dairy Farming in New Zealand on Whole-system Resource
Use Efficiency and Environmental Emissions 226
Ledgard, S.F., Finlayson, J.D., Patterson, M.G., Carran, R.A., and Wedderburn, M.E. 226
Food purchasing processes and environmental information in the food service industry
in Sweden 230
Bergström, K., Solér, C., Adolfsson, A., Shanahan, H. 230
eLCA: website and database of IPP tools for SMEs in the agro-food sector 233
Milà i Canals, L., Fullana, P., Mantoux, F. 233
Sources of Site-Dependency and Importance of Energy Consumption in Agricultural
LCA: Apple Production in New Zealand. 236
Milà i Canals, L., Burnip, G.M., Suckling, D.M. and Cowell, S.J. 236
Procedures to improve scope definition and inventory analysis in LCAs of farming
systems 243
Hayo M. G. van der Werf and Claudine Basset-Mens 243
Predicting the future: comparing pork with Novel Protein Foods on environmental
sustainability impairment 245
Helms, M. 245
Indicators for Monitoring Environmental Relevant Trends of Food Consumption 249
Jungbluth, N., Frischknecht, R. 249
The History Of Bread Production: Using LCA In The Past 255
Mirja Van Holderbeke, Nieves Sanjuán, Theo Geerken, Daniëlle De Vooght 255
Methodological Contributions to tailor Life Cycle Assessment to the Specifics of Arable
Crop Production 261
Brentrup, F. 261
Life cycle inventory modelling in the Swiss national LCI database ecoinvent 2000 265
Jungbluth, N., Frischknecht, R. 265
Environmental profiles of foods for the Swedish market 271
Lagerberg-Fogelberg, C., Carlsson-Kanyama, A., Fogelberg, F. and Brandén, I. 271
5
Using I/O data to find hotspots in LCA - Example of a hamburger meal 273
Jacob Madsen and Suzanne Effting 273
The use of by-products from food industry as basis for livestock feed and the
consequences for the analysis of the environmental impacts of meat consumption 277
Nonhebel, S., Elferink, E.V. 277
Is manure separation a solution to environmental problems from large scale pig
production? 281
Claus Grøn Sørensen, Niels Halberg, Randi Dalgaard 281
LCA of Danish milk -system expansion in practice 285
Randi Dalgaard and Niels Halberg 285
LCC and LCA of extra-virgin olive oil: organic vs. conventional 289
Notarnicola B., Tassielli G. , Nicoletti G.M. 289
6
Preface
Focus on sustainable (agricultural) production should go hand in hand with the quest for sus-
tainable (food) consumption. This idea is in agreement with principles of Integrated Product
Policy, which is the modern chain based and product-oriented approach to environmental im-
provement and regulation adopted by national environmental authorities, the EU Commission,
UNEP and increasingly also by private companies. The regulation of agri-environmental
problems has so far focused on improvements in farm production systems in relation to local
effects. This approach does not take into consideration the developments in food processing
and consumer choices as well as more global environmental effects. In the coming decades
production and consumption of livestock products and processed food is expected to increase
significantly and the globalisation of the food systems will continue. Therefore, as much as
ever there is a need for a complementary focus on the environmental impact per unit of food
produced including the whole production chain and taking into account global effects of dif-
ferent systems for food provision and consumption patterns.
Food consumption is one of the major causes for resource use and environmental impact by
modern households. But the relative importance of these burdens in the primary production,
industrial food processing and kitchen preparation respectively differ among products. It is,
however, not easy to define and compare the environmental burden from different choices of
food. Life Cycle Assessment (LCA) is a tool for an aggregated description of emissions,
waste and the resource use from soil to kitchen per unit of different food items. The method-
ology for LCA in the food sector has been developed during the last decade and progress in
terms of methodological robustness and data availability has been demonstrated - among oth-
ers – at a series of LCA-food conferences organised in Brussels (1996, 1998) and Gothenburg
(2001). The present proceedings from the Fourth International Conference of LCA in the
Agri-Food Sector presents a number of papers based on floor- and poster presentations show-
ing both the recent advances in methodology and inventories and the wide range of applica-
tions and objectives for LCA in the food sector.
The first 10 papers and 10 of the 21 poster papers demonstrate different applications of LCA
in the primary sectors agriculture, horticulture, livestock production and aquaculture/fisheries
and in the food processing industry. Then follows a section of papers and posters addressing
methodological questions - such as system expansion and land use - and presenting new in-
ventories of life cycle data relevant to the food sector. Finally, a number of papers and posters
present other approaches to sustainability assessment of food production and consumption,
which may supplement the more “classical” LCA.
The organisers wish to thank the authors for providing revised papers for these Proceedings
and the referees for giving constructive advise for improvements of the full papers and poster
papers.
Niels Halberg
7
Acknowledgements
The organizers of the conference wish to express thanks to:
The Danish Directorate for Food, Fisheries and Agri Business (DFFE) for funding and sup-
porting the Danish project “Life Cycle Assessment of basic food products” (2000-2004),
which made it possible to host the conference.
Stiftelsen Cerealia för Forskning & Utveckling, Cerealia for giving financial support to the
conference.
Moreover, the organizers are grateful for help from the programme committee and the follow-
ing companies/organizations: 2.-0 LCA consultants, UNEP, Vito and SIK.
Birgit Sørensen, Lene Kirkegaard and Jytte Christensen at the Department of Agro-Ecology,
DIAS, are thanked for their invaluable support in organizing, web-site management and edit-
ing before, under and after the conference.
International programme committee
Dr. Bo Weidema, 2.-0 LCA consultants, DK (co-organiser)
Dr. John E. Hermansen, DIAS, DK
Dr. Sarah Cowell, Surrey, UK
Ir. Theo Geerken, Vito, BE
Dr. Niels Jungbluth, ESU-services, CH
Dr. Gregory A. Keoleian, Univ. of Michigan, USA
Env. expert Inger Larsson, Arla Foods, SE
Sen. Cons. Morten Løber, Danish Agricultural Council
Dr. Berit Mattsson, SIK, SE
Dr. Guido A. Reinhardt, IFEU, GE
Env. advisor Lone Rosing, Cerealia, DK
Env. expert Peter Schonfield, Unilever, UK
Ass. expert Anne Solgaard, UNEP
The papers in this volume and the presentations at the conference are available at the web-
site: http://www.lcafood.dk/lca_conf/
8
Life cycle assessment of bread production - a comparison of eight different
scenarios -
Braschkat, J.1), Patyk, A.1), Quirin, M.1), & Reinhardt, G. A.1)*
1)
ifeu - Institut für Energie- und Umweltforschung Heidelberg GmbH, Wilckensstr. 3, D-69120 Hei-
delberg, * Tel: +49-6221-4767-0 (ext.: -31), Fax: -19, [email protected]
Abstract
In a study the life cycle assessments of eight different ways of bread production were evalu-
ated, considering different crop production methods (conventional, organic), different milling
technologies (industrial mill, domestic mill) and different baking technologies (large bread
factory, bakery, domestic bread maker). The scenario combining organically grown wheat,
industrial milling and a large bread factory proved to be most advantageous way of producing
bread with respect to the impact categories considered in this study. The use of organically
grown wheat, however, requires more land area than the use of conventionally grown wheat.
In addition to the differences due to the crop production method, milling and baking technol-
ogy, the transport of grains, flour and bread by the consumer is of vital importance for the ul-
timate appraisal of each scenario. In fact, the transport by the end consumer easily domi-
nates the overall ecological effect depending on the distance.
Keywords: life cycle assessment, bread, agriculture.
Introduction
Bread is one of the most important foods. Like any other production process, bread produc-
tion is associated with environmental impacts due to the demand of resources and due to
emissions. The environmental impacts, however, vary depending on the way bread is pro-
duced. Disseminating information about the environmental effects of different ways of bread
production would enable the consumer to base his decision of purchasing bread or its ingredi-
ents on ecological reasons. Bread producers on the other hand might use this information to
optimise the production process. In order to provide a complete image of all relevant envi-
ronmental aspects such as resource and energy demand, greenhouse effect, ozone depletion,
acidification, eutrophication, photo smog and demand of land area, the entire life cycle from
the acquisition of raw materials, across the actual production and use, up to the final dis-
posal/recycling, needs to be considered. Life cycles within bread production may differ al-
ready during the production of the wheat, which may be grown either conventionally or or-
ganically. After crop production, milling of the grain might be done either by an industrial
mill or alternatively by a domestic mill. Finally, bread might be produced from the flour ei-
ther in a large bread factory or in a bakery or in a private household (bread maker). Each of
these single options is again associated with different transport efforts, which have to be con-
sidered as well.
This study based on Patyk (2003a) and on own calculations aimed at answering the following
questions: Which way is the most environmentally friendly way of bread production? And,
9
more specifically: Which crop production method (conventional, organic), which milling
technology (industrial or domestic mill) and which baking technology (large bread factory,
bakery, domestic bread maker) is the most advantageous one from the ecological point of
view? Which one of the process steps, including the transports, does account for the highest
or lowest environmental effects? Where, within the whole bread production chain, is it feasi-
ble to introduce ecological optimisations or to reduce environmental effects, and what are the
corresponding recommendations?
Approach and predefinitions
The combination of the above mentioned single processes result in 8 different scenarios of
bread production (Figure 1) like, for instance, bread made from organically grown wheat,
ground and baked in a private household including the associated transports. The life cycles
assessed in this comparison start off with the crop production, incorporating all steps of the
conventional as well as organic wheat production, from soil cultivation up to harvest. Data for
soil cultivation, like ploughing, seedbed preparation, sowing, application of fertilisers and
pesticides (conventional production only), harvest, transport of the grain to the farm and seed
production was taken from Borken et al. (1999). Data for the provision of diesel fuel are
based on Patyk (2003b). For fertiliser production data from Patyk & Reinhardt (1997) and for
additional operational supplements data from Gaillard et al. (1997) and Reinhardt et al. (2001)
were used. For the conventional as well as for the organic system the entire crop production
was considered within the complete crop rotation. Consequently, various effects like the value
of preliminary crops or intercropping for nitrogen fixation in the organic system, resulting in
different yields per unit area and time, were implicitly taken into account (for details we refer
to Kaltschmitt & Reinhardt (1997)).
Within flour production it is assumed that the grain is ground either in an industrial mill or in
a domestic mill. Industrial mills are operated at medium voltage; domestic mills in contrast
are run at low voltage. In this study, data for the provision of electrical energy given in
Borken et al. (1999) and data for the energy demand of industrial mills from Fritsche et al.
(2002) were used. The energy required for operating domestic mills and for the provision of
operating supplies within this study, was based on estimates.
Regarding bread production three different options were considered: a large bread factory, a
bakery and a private household using a bread maker. Ovens used in bread factories and in
bakeries are operated either by electricity or oil or natural gas while domestic bread makers
are solely driven by means of electrical power. For the commercial ovens the energy-mix ac-
cording to Fritsche et al. (2002) was applied. The energy demand of domestic bread makers
was estimated and linked with the data on energy supply published by Borken et al. (1999)
and Patyk (2003b).
10
Scenario 1 2 3 4 5 6 7 8
Crop
Conventional crop production Organic crop production
production
Retailer Retailer
Milling Industrial Domestic Domestic
Industrial mill
process mill mill mill
Retailer Retailer
Baking Bread Domestic Domestic Domestic Domestic
Bakery Bread
factory bread bread Bakery bread bread
process factory
maker maker maker maker
Retailer Retailer
End- End- End- End- End- End- End- End-
consumer consumer consumer consumer consumer consumer consumer consumer
Figure 1: Schematic representation of the 8 life cycle scenarios of bread production (solid
lines indicate standard transports, broken lines indicate transports by consumer).
Standard transports with 23 t trucks (distance 100 km, outward bound fully loaded, return
empty) were assumed for the following routes: transports of grain from the farm to the mill or
retailer, transports of flour from the mill to the bread factory or bakery or retailer and trans-
ports of bread from the bread factory to the retailer. Data on diesel fuel consumption for these
transports originate from Borken et al. (1999) and Knörr (2002). Data on the provision of die-
sel fuel and operating supplies were taken from Patyk (2003b) and Reinhardt et al. (2001) re-
spectively. For the transport of grain, flour and bread by the end-consumer it was initially as-
sumed, that the transports were done either on foot or using a bicycle and thus, the energy
demand and emissions were either zero or not significant. Finally, additional scenarios for the
transport of the bread by the consumer using a car were calculated.
Basically, the assessment follows the ISO 14040/43 standards. The functional unit is 1 kg of
bread ready for consumption at home. The environmental effects studied are listed in Table 1
including cumulated primary energy of non-renewable energy carriers, greenhouse gas emis-
sions by IPCC method, ozone depletion through N2O emissions, eutrophication and acidifica-
tion through airborne emissions and others. For details we refer to Borken et al. (1999).
11
Table 1: The environmental effects, indicators and parameters considered in this study.
Environmental effect Indicator Parameter
Energy demand Non-renewable primary energy Crude oil, natural gas, mineral
coal, lignite, Uranium
Greenhouse effect CO2-equivalents CO2, N2O, CH4
Ozone depletion N2O N2O
Acidification SO2-equivalents SO2, NOx, NH3, HCl
Eutrophication PO4-equivalents NOx, NH3
Photo smog Ethen-equivalents CH4, NMHC
Land use Land use Land use
Results
The environmental effects (Table 1) of the 8 different scenarios of bread production are pre-
sented comparatively in Figure 2. Values refer to the production of 1 kg of bread. The baking
process was the most energy-consuming step of the entire bread production process account-
ing on average for 64% of the total energy demand. The baking process using a domestic
bread maker requires 3 times more energy than in a factory and in the bakery, energy demand
is still twice as high than in a large bread factory. Due to the close correlation between energy
demand and greenhouse effect the same applies to the greenhouse effect as well. Besides, us-
ing a conventional oven for baking bread at home requires more energy on average than a
bread maker and therefore this option was not considered in this study.
Crop production, however, is much more important regarding the greenhouse effect because
of the amount of N2O released and for that reason the assessment not only depends on the
baking process but also on the way, how the crop was produced. Regarding ozone depletion,
acidification and eutrophication the situation is completely different. In these cases, all of the
scenarios based on organic crop production are most beneficial, whereas the remaining down-
stream processes did not entail any further differentiation of the results (Figure 2). Regarding
photo-smog, ethene-equivalents (NMHC) as well as the NOX-corrected ethene-equivalents
(NCPOCP) have been considered. However, this analysis did not show any significant differ-
ences between the eight scenarios.
From the environmental effects considered so far, only advantages resulted from organic pro-
duction of the wheat crop. This situation changes when considering the size of the land area
that is required for the crop production. The conventional production system requires only
65% of the area that is needed to grow the wheat organically (Figure 3). This is largely due to
the use of synthetic fertilisers and pesticides and the resulting higher yields in conventional
farming systems.
12
cultivation mill bake transport
Energy demand Greenhouse effect
10,0 800
g CO2-eqiv./kg bread
8,0
600
MJ/kg bread
6,0
400
4,0
200
2,0
0,0 0
con- con- con- Ozone
con- depletion
org- org- org- org- con- con- con- con- org-
Acidification org- org- org-
0,60 ind- ind- ind- home- ind- ind- ind- home- 3,00 ind- ind- ind- home- ind- ind- ind- home-
fac bac home home fac bac home home fac bac home home fac bac home home
g SO2-equiv./kg bread
0,50 2,50
g N2O/kg bread
0,40 2,00
0,30 1,50
0,20 1,00
0,10 0,50
0,00 0,00
con- con- con- Eutrophication
con- org- org- org- org- con- con- con- con- org- org- org- org-
0,40 ind- ind- ind- home- ind- ind- ind- home- ind- ind- ind- home- ind- ind- ind- home-
fac bac home home fac bac home home fac bac home home fac bac home home
g PO4-equiv./kg bread
0,30
0,20
0,10
0,00
con- con- con- con- org- org- org- org-
ind- ind- ind- home- ind- ind- ind- home-
fac bac home home fac bac home home
Figure 2: LCA results of the 8 scenarios of bread production regarding energy demand,
greenhouse effect, ozone depletion, acidification and eutrophication.
2,00
1,50
m2/kg bread
1,00
0,50
0,00
con org
Figure 3: Land area required to produce the grain for 1 kg of bread using a conventional (con)
and an organic (org) production system.
13
For the transport of grain, flour and bread by the consumer different transport modes (car,
public transport, bicycle/on foot), distances and additional transport purposes have to be con-
sidered. In an additional analysis, the ecological impacts were calculated for a number of dif-
ferent transport scenarios, in order to demonstrate the magnitude of the effects, owing to the
transport of 1 kg of bread by the consumer. Exemplary results are presented in Figure 4 for
energy demand and acidification. In contrast to the transport on foot or by bicycle, which in-
volves zero, respectively negligible emissions and demand of energy, the extreme case of a 4
km car transport, with the sole purpose to acquire 1 kg of bread, will entail an energy demand
of 18.6 MJ and the emission of 2.2 g SO2–equivalents. That means the energy demand due to
the bread transport is 4 times higher and acidification is 2.5 times higher than that caused by
the entire preceding bread production chain (conventional crop production assumed). In addi-
tion to this extreme transport scenario, Figure 4 indicates car transport distances, at which the
higher energy demand caused by a domestic bread maker and by a bakery, in contrast to a
large bread factory, was compensated. Similarly, the car transport distance is indicated, at
which the larger acidification due to the conventional crop production was compensated. As a
result a transport distance of about 1 km and 0.5 km may compensate for the higher energy
demand of a domestic bread maker and of a bakery compared to the bread factory. Larger
transport distances will overcompensate the higher energy demand of the bread maker and the
bakery. In contrast, the acidification due to the conventional crop production compared to the
organic crop production will be compensated at a distance not less than about 3 km and that
again only if the bread purchase is the sole purpose of the drive. Adding another 10 kg of gro-
ceries to the shopping will increase the distance, at which the higher acidification of the con-
ventional crop production is compensated, up to 29 km.
cultivation mill bake transport transport home
Energy demand Acidification
25 4
20
g SO2-equiv./kg bred
15
MJ/kg
2
10
1
5
0 0
con-home-home con-ind-bak con-ind-fac con-ind-fac con-ind-fac org-ind-fac org-ind-fac
0,53 km 0,88 km 4 km 2,9 km 4 km
by car by car by car by car by car
Figure 4: Magnitude of the energy demand and the acidification due to the transport by the
consumer in addition to the production of 1 kg of bread.
14
While the larger energy demand and the higher greenhouse gas emissions of a bread maker in
contrast to a bread factory is compensated by a 1 km car transport, the car transport will cause
a less favourable result for the bread factory in the remaining categories. However, this im-
plies that the grain or flour for a bread maker was transported either on foot or by bicycle.
Conclusions
Looking at the single options of each processing step (crop production, milling, baking) the
following conclusions may be drawn: Organically grown wheat has to be preferred over
wheat that was produced conventionally, regarding all impact categories except land use.
Flour may be produced preferably in an industrial mill rather than in a domestic mill. Ranking
the bread baking process from the most to least advantageous option results in the order bread
factory, bakery and domestic bread maker.
Looking at the entire bread production chain including the transports by the consumer the fol-
lowing conclusions may be drawn: Bread production using organically grown wheat, ground
in a industrial mill and baked by a large bread factory is the most preferable way of producing
bread. As far as possible bread producers may have to use cereals originating from organic
production. Transport of grains, flour and bread by the consumer is of vital importance for the
ultimate appraisal of each scenario. For instance, concerning the energy demand, the con-
sumer annihilates the entire ecological advantage of the bread factory, if he involves a trans-
port by car over a distance of just 1 km. As a principle consumers may have to ask for eco-
bread. Bakery products should generally be transported either on foot or with a bicycle. If
transported by car, the purchase of bread may be combined with the shopping of additional
groceries. A bakery might look for commercial energy saving measures. For instance, partici-
pating in projects like "Bäcker/Konditoren und Umwelt" organised by the bakery guild, the
BUND and the city council Heidelberg may be a first step (Bäckerinnung et al. (2003)).
Homemade bread may be baked in a domestic bread maker because the energy demand for
the production of 1 kg bread is lower than the production in a domestic oven. If using the
oven anyhow, increasing the degree of utilisation may reduce the energy demand of the do-
mestic oven.
For the decision of a specific source of bread supply not only ecological aspects may be rele-
vant but also socio-political aspects (promotion of small enterprises), economic aspects, nutri-
tional aspects, recreational aspects as well as hobby aspects play an important role. Therefore,
the selection of a specific bread supply option is based on the individual choice of all these is-
sues.
References
Borken, J., Patyk, A. and Reinhardt, G. A. 1999. Basisdaten für ökologische Bilanzierungen:
Einsatz von Nutzfahrzeugen in Gütertransport, Landwirtschaft und Bergbau. (Basic data
for life cycle assessments: Use of commercial vehicles in transportation, agriculture and
mining). Verlag Vieweg, Braunschweig/Wiesbaden
15
Bäckerinnung, Bund für Umwelt und Naturschutz (BUND) and Stadtverwaltung Heidelberg
2003. Projekt "Bäcker/Konditoren und Umwelt" (Bakery and Environment). Heidelberg,
running project
Fritsche, U. R. et al. 2002. Gesamt-Emissions-Modell Integrierter Systeme (GEMIS) (Global
Emission Model for Integrated Systems (GEMIS)) Version 4.1
Gaillard, G., Crettaz, P. and Hausheer, J. 1997. Umweltinventar der landwirtschaftlichen In-
puts. (Environmental inventory of agricultural inputs). FAT-Schriftenreihe Nr. 46, Täni-
kon/CH
Kaltschmitt, M. and Reinhardt, G. A. (ed.) 1997: Nachwachsende Energieträger: Grundlagen,
Verfahren, ökologische Bilanzierung (Biofuels: Basics, approaches, environmental as-
sessments). Verlag Vieweg, Braunschweig/Wiesbaden
Knörr, W., Höpfner, U., Lambrecht, U., Nagel, H.-J., Patyk, A. 2002. Data and Calculating
Model: Energy demand and pollutant emissions from the transportation sector in Ger-
many 1960 to 2020. Development of the software TREMOD by IFEU-Institute Heidel-
berg, Data-preparation and calculations by order of the Federal Environmental Office,
Berlin. Documentation
Patyk, A. 2003a. Inventory Analysis of Petroleum Products. By order of the German Scien-
tific Association for Crude Oil, Natural Gas and Coal, supported by the Association of
the German Petroleum Industry, Hamburg, in print
Patyk, A., 2003b. Der KEV im Bereich Landwirtschaft/ Nahrungsmittelproduktion - Ver-
gleich zur Herstellung verschiedener Brotarten. (CED in agriculture and food production
- comparison of the production of different bread types). Published in the final report to
the project "Anwendung und Kommunikaton des Kumulierten Energieverbrauchs
(KEV) als praktikabler Entscheidungsparameter für nachhaltige Produkte und Dienst-
leistungen". (Utilisation and communication of the Cumulated Energy Demand (CEV)
as a practical decision parameter for sustainable products and services)
Patyk, A. and Reinhardt, G. A. 1997. Düngemittel – Energie- und Stoffstrombilanzen. (Fertil-
izer – energy and production flow anlasysis). Verlag Vieweg, Braunschweig/Wiesbaden
Reinhardt, G. A., Jungk, N., Korsuize, G. and Patyk, A. et al. 2001. Bioenergy for Europe:
Which ones fit best? – A comparative Analysis for the Community. By IFEU/D (co-
ordinator), BLT/A, CLM/NL, CRES/GR, CTI/I, FAT/CH, INRA/F, TUD/DK. FAIR CT
98 3832.
16
When a hole matters - the story of the hole in a bread for French hotdog
Rosing, L. and Nielsen, A.M.
*) Cerealia Bakeries A/S, Birkemosevej 10, DK-8361 Hasselager. Tel.: +45 76 28 50 00. Fax: +45 76
28 50 01. [email protected]. 2.-0 LCA consultants. Borgergade 6, 1., DK-1300 Copenhagen.
Abstract
The environmental management of Cerealia Bakeries has till now focused on the bakeries’
own consumption of energy, packaging and cleaning agents. However, looking at the entire
chain of processes, which are affected when a consumer eats industrial bread products, the
major environmental impacts arise from actions by the suppliers and customers. In this pa-
per, the environmental impacts of these production technologies are assessed, and the im-
plication for future product development is discussed.
Background
During the last years, Cerealia has participated in development and application of product-
oriented environmental assessment of food. This paper is based upon an environmental as-
sessment of improvement potentials in the production of bread for French hotdog (Rosing and
Nielsen, 2003), which was made as part of the LCAfood-project. Cerealia has published part
of their own environmental data in the LCAfood-database (Nielsen et al., 2003), which in re-
turn has been the source of data on environmental data from agricultural productions for Ce-
realias work.
Identification of environmental hot-spots in the product chain of bread for French hotdog
Previous experiences show us that the hot-spots of a roll lie in agriculture’s production of
grain, including their use of fertiliser, and the consumers’ use of oven for heating up the bread
(Rosing et al., 2001). As part of the LCAfood-project, the environmental impacts from bread
for French hotdog was analysed, and a similar picture was found.
Figure 1 shows the environmentally most important processes, which are affected when a
consumer buys a bread as part of a French hotdog, and their contribution to global warming
potential, calculated with the EDIP-method (Wenzel et al., 1997). The chosen amount (17.000
pieces of bread sold from hotdog-stand) equals the product in one hour in Cerealias produc-
tion unit in Karup (18.000 pieces of bread sold from bakery minus waste in the hotdog stand).
The breads are produced of conventional wheat flour, distributed to the hotdog-stands, where
they are heated by use of a toaster. The by-products from the milling-process are used as ani-
mal feed, and thus replace alternative feed (barley). This decreased production is shown as the
green bar.
More than 50% of the total contribution to global warming comes from agriculture’s produc-
tion of grain. For the other impact types eutrophication, acidification and nature occupation
(land-use), this percentage is even higher.
17
Use of conveour toaster contributes with 25% to the total global warming. This percentage
can vary between 10 – 60% according to the size of the hotdog stand and how fast the hotdogs
are sold.
17.000
baguettes from
hotdog stand
2.3 ton CO2
18.000 frozen 360 hours
baguettes from Distribution
bakery Toaster
1.3 ton CO2 0,3 ton CO2 0,5 ton CO2
1,1 ton
conventional 14 days
wheat flour Storage in
Storage in Electricity
from mill freezer wholesale
1.1 ton CO2 0,2 ton CO2 0,2 ton CO2 1,0 ton CO2
1.4 ton 300 kg
Bread wheat, Animal feed
from farm production
1.2 ton CO2 - 0,2 ton CO2
-300 kg
Fertiliser (N) Barley,
from farm
0,4 ton CO2 -0,2 ton CO2
Figure 1. The product chain for bread for French hotdog sold at hotdog stand. Only proc-
esses, which contribute with more than 6% to total result are shown. Boxes refer to produc-
tion processes. Names of grey boxes refer to the main product of the processes. Red arrows
represent material or energy transfer between two processes; green arrows represent saved
material or energy transfer as a result of displacements and green lines represent displace-
ments. The red/green-bars in the boxes show each process’ relative, cumulated contribution to
total environmental impact.
Different production-technologies for making the hole in the bread for French hotdog and
their environmental impact
A baguette for French hot-dog must have a hole to put the sausage into. Until 15 years ago, all
hotdog stands bought standard baguettes, and made the holes themselves by sticking a thick
18
spear into them. This method was not without pitfalls. Partly because a big amount of bread
was pressed together in the bottom of the hotdog, partly because some hotdog stands had to
throw big amounts of baguettes out, when they destroyed them by e.g. sticking the spear out
of the side of the baguette.
Some hotdog-men suggested to Cerealia to find a better way of making this hole. Then the
present production technology was invented, where the baguettes after baking are chilled, fro-
zen, and drilled. In the beginning, the bread waste was sent to combustion with other normal
waste. To avoid the big amounts of waste, it was tried to sell the fresh bread-waste as animal
feed. The distribution was difficult, because it was necessary that the bread should be sold and
used within two days to avoid mould.
Cerealia then started using the bread-waste for making dried breadcrumbs. Today, all bread
waste from the drilling of breads for French hotdogs is used in this production. The bread-
crumbs are sold to household-consumers, to other food industry, or reused in-house for pro-
duction of new baguettes, where they replace flour. Excess amounts of breadcrumbs are sold
to farmers as animal feed, where they replace alternative feed (barley).
From this story we can identify three production-technologies for bread for French hotdogs,
and as a future outlook add a forth:
A. The breads are frozen and drilled. The bread waste is sold (fresh) as animal feed,
where it replaces barley.
B. The breads are frozen and drilled. The bread waste is processed into breadcrumbs, and
used in the bakery in stead of flour.
C. The breads are frozen and drilled. The bread waste is processed into breadcrumbs, and
sold as animal feed, where it replaces barley.
D. A future best-case technology, where bread-waste from drilling can be avoided. Either
the breads are baked around a stick, or a new recipe is developed where the breads can
be baked with more air in them, so the hotdog-man can make the hole with a spear,
still avoiding the excess amount of bread in the bottom of the hotdog.
In Figure 2, the contribution to environmental impacts is shown for a shift from the produc-
tion-technology 15 years ago, where the bread was sold at a weight of 90 g, and the hole was
made by the hotdog-man with a spear, to any of the four above-mentioned technologies.
Figure 2 should be read like this: if a hotdog-man changes from ordinary baguettes to ba-
guettes with a pre-drilled hole in them (C), the contributions to all studied environmental im-
pacts will decrease. The environmental impacts will decrease even more, if the bread is used
as feed without the drying of breadcrumbs, or if the dried breadcrumbs can be used as an al-
ternative to flour in another production. A hypothetical, new technology where the breads can
be produced without any drilling and/or drying of breadcrumbs, would lead to the biggest en-
vironmental improvement (D).
19
%
-10
-20
-30
-40
-50
-60
-70
-80
-90
-100
Global warming (GWP 100) Acidification Eutrophication Land use
A. Bread-waste replace grain B. Dried breadcrumbs replace flour C. Dried breadcrumbs replace grain D. Hypothetical, best-case technology,
no drilling
Figure 2. Comparison of environmental impacts (Global warming (GWP 100), Acidification,
Eutrophication, Land use) from a change in production-technology from old production,
where the hole is made (A. Bread-waste replace grain. B. Dried breadcrumbs replace flour. C.
Dried breadcrumbs replace grain. D. Hypothetical, best-case technology, no drilling). Nega-
tive bars shows that a change to the technology yields an environmental improvement.
Figure 2 shows that the last years’ shift in production-technology has not just made the hot-
dog-men more satisfied with the breads for French hotdogs, but has also been an improve-
ment for the environment. All the bars show decreased environmental impact, meaning that a
shift to any of them will be an environmental improvement. Even though the energy-
consumption at the bakery has increased, because drying of breadcrumbs is an energy-
consuming process, the total environmental impact is decreased, because even more energy-
consuming processes are avoided in the production of grain or flour.
In A and C the bread waste from the drilling replaces the same amount of alternative feed in
agriculture. When the bread-waste from the production is used as animal feed, the same
amount of grain is replaced as animal feed, and thus the same amount of land will not have to
be cultivated, whatever the bread is processed into dried breadcrumbs, or sold as fresh bread.
But in alternative A, where the drying of the bread is avoided, will off course have a lower
consumption of energy, and thereby a lower contribution to global warming, acidification and
eutrophication.
The present production-technology (C) where the hole is drilled at the bakery, and the bread-
waste is processed into dried breadcrumbs has the largest environmental impacts of the stud-
ied alternatives. Therefore a change to any of the alternative technologies could lead to envi-
ronmental improvements.
20
Discussion
Cerealias work with LCA has given a new basis for prioritizing future working efforts in the
environmental department. The hot-spot identification has pointed to agriculture and hotdog-
stands as key processes, which would be important to change to lower the total environmental
impact, but the production of bread for French hotdogs shows that environmental improve-
ments can also be found through development of the bakeries’ own productions.
Cerealias opportunity to influence agricultural production
The majority of environmental impacts from the bread come from agriculture’s production of
wheat. Based on the Danish data from the lca food database the present supply of flour-types
offers Cerealia no possibility to replace their use of conventional flour with a less environ-
mentally problematic flour-type. The alternative flour-types could maybe have lower toxic
impact, e.g. organic flour is produced with no use of pesticides or more environmental
friendly flour e.g. without straw shortener. However, the contribution to global warming,
acidification and nutrient enrichment will be similar for them, and if the flour comes from ag-
riculture with a lower productivity, the impact on nature occupation can even increase. One of
Cerealias environmental goals is a.o. to increase the use of more environmental friendly flour,
why we in the future will do some analysis of different flour types to examine the alternatives.
On the other hand, Cerealia Bakeries have a possibility to decrease environmental impacts
from agricultural production by decreasing the demand for grain through utilising of bread-
waste. The development in production-technology of bread for French hotdog is an example
of this in two respects:
1. bread-waste is used as a replacement of flour or grain
2. new types of bread are developed, where the customers needs are fulfilled with the
lowest amount of bread
When possible, Cerealia participates in research aimed at lowering the environmental impacts
from agricultural production. Knowledge from these projects will continuously be included in
the company’s decision-basis.
Cerealias opportunity to influence hotdog-men
The hotdog-stand can be the second most important process in the environmental impact of
the bread. If the hotdog is sold from small or inefficient hotdog-stands, it may even exceed
agriculture in contribution to the environmental impacts.
Cerealia has a clear possibility to influence the preparation routine of the hotdog-men, since
guidelines for preparation of the breads are part of normal marketing procedure. If Cerealia
find concrete suggestions to improve the environmental impact from the hotdog-stands, such
information could be included. However, consumers are not likely to care much about the en-
vironmental impact of the hotdog-breads, and hotdog-men are not likely to work for de-
creased environmental impacts from their products, unless it is economically and practically
21
feasible. But seeing that energy-savings and lower amounts of bread-waste mean lower cost
for the hotdog-man, this should be possible.
In contrast to the agriculture, there are no research-projects on environmental impacts from
hotdog-stands that Cerealia can join.
Cerealias opportunity to achieve environmental improvements by changes in the bakery
Organizationally it is easier for Cerealia to change the production in the bakeries, than to in-
fluence the behaviour of suppliers or customers. As showed in this analysis the absolute value
of the environmental improvements is a change in production technology of the hole. This is
not an option for now, as our customers are happy with the product. Therefore we have in-
stead looked at other possibilities within the bakery. During our environmental project “LCA
– as a tool in the learning bakery” we have created working groups which should analyse the
environmental improvements at the bakeries. Another lca could maybe show the positive de-
crease of environmental inpact for the bakery, but this work has not been done.
References
Nielsen, P.H., Nielsen A.M., Weidema B.P., Dalgaard R. and Halberg N. (2003). LCA food
database. www.lcafood.dk/.
Rosing L, Feldvoss C, Kann M, Leth K D, Mundt P, Pedersen H, Skovsby K, Christensen L
M. (2001). LCA of a roll (in Danish). Working report in the project LCAfood. Available
at www.lcafood.dk.
Rosing L and Nielsen, A.M. (2003). Bread for French hotdog – possible environmental im-
provements in the product chain. Available at www.lcafood.dk.
Wenzel H, Hauschild M and Alting L (1997). Environmental assessment of products. Volume
1. Methodology tools and case studies in product development. Chapman and Hall.
22
Possible benefits from using LCA in the agro-food chain, example from
Arla Foods
Inger Larsson
Arla Foods, Group Environmental affairs, Sweden, tel: +46 8 6773277, mobil: + 46 705 145640, fax:
+46 8 789 53 20
Arla Foods has worked with life cycle assessment since the mid 90s. We have worked with
LCA concerning packaging, transports and different milk products. The work with LCA for
packaging and transports was originally based on the public interest from consumers con-
cerning the environmental impact from packaging and transportation. The work has then de-
veloped to become complete LCA for different milk products. A life cycle assessment has, in
cooperation with the Swedish dairy association, been made for drinking milk packed in Tetra
Brik packaging. The work has been published as a scientific report and has also been pub-
lished as a brochure that can be found on www.svenskmjolk.se/english.asp Download: Milk
and the environment. The life cycle assessments have given us basically five kinds of infor-
mation to date. The production of milk at the farm including production of concentrated feed,
fertilizers, pesticides etc., is without doubt the most important impact on the life cycle of milk
and cheese. The most important environmental work for liquid milk dairies is to reduce the
loss of milk to wastewater, pig feed and landfill. For cheese dairies primary work is to pro-
duce a high cheese yield.
• Transports of milk in tankers are efficient and have only little impact on the environ-
ment when counted per liter of milk.
• The transport of milk or cheese to the home can cause substantial environmental ef-
fect.
• The choice of packaging material has an important influence on the environment. A
carton material is always better than plastic ones. Whether package material is recy-
cled or incinerated makes less difference than the choice of material.
Using LCA results, Arla Foods has now advanced environmental work in the following ways.
• We can give better answers about environmental issues to consumers, customers
and students.
• New LCA information is published in brochures and reports.
• The liquid milk dairies in Sweden have set targets to reduce all kinds of milk loss.
These dairies also have targets to avoid increasing the environmental impact of
packaging.
• The environmental work includes the Arla Foods farms in Denmark and Sweden. The
quality program, including the environmental program for Arla Foods farmers in Swe-
den and Denmark is available in Swedish and Danish, log on to www.arlafoods.dk for
more information, click download “Arlagården”.
23
Environmental Management Practices in an Italian Coffee Company using
LCA Methodology
Roberta Salomone
Dipartimento RIAM - Università di Messina - Piazza S. Pugliatti 1, 98121 Messina
Tel. 0039-90-771548; 0039-90-661499, fax 090/6764920, e-mail: [email protected]
Abstract
Coffee production has grown by nearly 200 percent since 1950; it is the most important
traded commodity in the world (after oil). Even if it is cultivated only in tropical and equatorial
areas (is the primary export of many developing countries), the most coffee is consumed in
the developed world (the United States and the European Community together import two
out of every three bags of coffee produced in the world). Considering that the coffee chain is
very wide and interspatial, with the involvement of many companies of different type and
size, each environmental decision, in whatever point of the coffee chain, should be taken in a
“life cycle thinking” perspective. It was with this intention that an environmental analysis of
impacts connected to a coffee industry located in Sicily was conducted, applying Life Cycle
Assessment methodology. System boundaries were defined to include all life cycle steps
from coffee cultivation to the consumer distribution, consumption and disposal and the func-
tional unit was defined as 1 kg of packaged coffee delivered to the final consumer. The im-
pact assessment step was performed investigating eight different impact categories (Air
acidification, Aquatic Eco-toxicity, Eutrophication, Human toxicity, Terrestrial Eco-toxicity,
Greenhouse effect, Depletion of ozone layer, Photochemical oxidant formation) and Eco-
points were used as a general weighting factor. The cultivation and the consumption stages
were identified as having the greatest environmental impacts. The aim of the study was to
recognise the “hot spots” in the product’s life cycle stages, in which environmental improve-
ments are easily achievable and to suggest options to minimise the environmental impact of
their production phases, improving process and company performance. The results showed
that the LCA methodology is suitable to assess the environmental impact associated with the
entire life cycle of an agro-food product and also to improve gate-to-gate life cycle informa-
tion that provide valuable means of understanding at company level; this means that envi-
ronmental improvements could be taken in a “life cycle thinking” perspective.
Introduction: aim of the study
Coffee production has grown by nearly 200 percent since 1950; after oil it is the most impor-
tant traded commodity in the world. Although it is grown only in tropical and equatorial areas
(it is the primary export of many developing countries) most of the coffee produced is con-
sumed in developed countries (the United States and the European Community jointly import
two out of every three bags of coffee produced in the world) (www.wri.org). Considering that
the coffee chain is very wide-ranging, involving many different types of company of varying
size, each environmental decision, at any point whatsoever of the coffee chain, should be
taken using “life cycle thinking”.
24
Environmental management has become increasingly important to productive and innovative
businesses and often involves suppliers upstream and the companies downstream. A business
that wishes to implement an effective internal environmental management system must first
of all analyse the environmental impacts of its production process and its products/services.
Inevitably, this entails identifying impact factors found at the start or end of pipeline and
therefore outside the physical confines of the business’ own productive sphere of activity
(Mirulla, 2001). Life Cycle Assessment (LCA) is making its mark as one of the most interest-
ing tools available to management for environmental assessment and control. LCA broadens
the vision of a producer giving it a more generalised view of the environmental impacts of the
production line. The business has to involve suppliers upstream and the companies down-
stream to collect inventory data in accordance with the boundaries of the system analysed.
This paper presents an environmental analysis of a coffee business adopting LCA methodol-
ogy. The analysis was carried out on a firm in Sicily (Italy) that roasts and distributes coffee.
The aim of the study is to obtain data relating to energy use, waste management and raw ma-
terial consumption in order to identify the “hot spots” in the stages of the product’s life cycle
in which environmental improvements are easily achievable then to suggest alternatives to
minimise the environmental impact of production phases, thereby improving processes and
company performance.
Methodological framework
The analysis of the environmental impacts of a coffee company located in Sicily was investi-
gated by applying Life Cycle Assessment (LCA). LCA is a methodology used for analysing
and assessing the environmental loads and potential environmental impacts of a material,
product or service throughout its entire life cycle, from raw materials extraction and process-
ing, through manufacturing, transport, use and final disposal (ISO 14040, ISO 14041, ISO
14042, ISO 14043).
The functional unit was defined as 1 kg of packaged coffee delivered to the final consumer.
The business has a wide product range but the functional unit was chosen in order to avoid al-
location (in accordance with ISO 14041), with no distinction between the various products
(e.g. different blends, different types of packaging, traditional and decaffeinated coffee, etc.).
System boundaries were defined to include all life cycle steps from coffee cultivation through
to its distribution to consumers, consumption and disposal. Production of machinery and
equipment are excluded from the system.
In figure 1 the coffee life cycle is presented: steps in the dotted boxes are not included in the
study.
25
Goal and scope definition. The goal
C offee cultivation
of the study is to examine the ways
in which the coffee roasting and dis-
Transport crude
coffee tribution company makes an impact
C om pany on the environment in order to iden-
Storing, cleaning
tify how to reduce its impacts and
W aste
and weighing
increase the environmental sustain-
ability of the product from a life cy-
R oasting W aste
cle perspective. It is important to
C ooling
ascertain the environmental aspects
of coffee processing and include the
Blending
R oasted coffee
transport
environmental impacts connected
with life cycle stages other than
Processing at
Packaging
m anufacture
G rinding external those relating solely to the company
industries
itself. The company wishes to use
Packaging Treated coffee
transport
Packaging
transport the information obtained from the
LCA study as a starting point for the
Transport Transport using development of its environmental
w ith carrier com pany vehicles
management set-up. This means col-
lecting information about the entire
D istribution
life cycle of the product and using
this information to improve the
C onsum ption W aste
company’s eco-efficiency.
W aste m anagem ent
Figure 1. Coffee life cycle.
Inventory analysis. The inventory analysis includes the following stages:
A) Cultivation. For this step literature data were collected, in particular for energy, fertilizer
and pesticide use (Diers et al., 1999). Fertilizer and pesticide production data were included
using commercially available databases (Ecobilan, 1999; Pré Consultants, 2003) while nitro-
gen and phosphorous emissions and pesticide emissions were quantified using estimation
methods (Brentrup et al., 2000; Heathwaite, 2000; Hauschild, 2000). Average coffee produc-
tion per hectare varies in relation to the type and characteristics of the land on which it is
planted, together with other ecological factors, as well as to the age of plants. The approxi-
mate yield ranges from 2 to 6.5 quintals of finished product per hectare (Barbiroli, 1970). We
assumed an average yield of 4.25 quintals/hectare.
Coffee beans can be processed in two ways: the dry method or the wet one. We assumed that
only the dry method (also called the natural one) was used to process coffee beans and that
the coffee cherries were both sun-dried and by using machines (assuming heavy fuel oil con-
sumption of 0.11 l/kg) (Diers et al., 1999) and that the whole process was done by hand. This
is very common in small or medium plantations and in regions where the temperatures are
warmer and supplies of clean, fresh water are not plentiful. The dry method produces a single
26
residue, the inner skin or outer hull, amounting to about 0.99t per 5.5t of coffee beans
(Camaggio Sancinetti and Nicoletti, 1995).
B) Processing. For this step specific site data were collected for each basic process contained
within the company box of the system flow chart. The direct material and energy inputs of the
coffee processing and packing stage are: green coffee (or dried cherries); electricity (to power
the equipment); natural gas (for the roasting step) and packing materials. The direct outputs
are: roast coffee in primary and secondary packaging; air emissions (from natural gas com-
bustion in the roaster) and waste (dust and scraps from cleaning and coffee chaff from roast-
ing).
C) Packaging. The company uses many different types of primary and secondary packaging
for roast coffee (aluminium cans, paper filters, etc). All of these have been included in the in-
ventory analysis (specific site data) whilst commercially available databases have been used
for the manufacturing of packaging materials (Ecobilan, 1999; Pré Consultants, 2003).
D) Transport. The main transportation activities take place at different life-cycle stages as fol-
lows:
1. Pesticides and fertilizers to coffee growers
2. Green coffee from growers to the coffee company premises;
3. Packaging from manufacturers to the coffee company premises;
4. Packaged coffee from the coffee company premises to local wholesalers and final
points of sale;
5. Packaged coffee from the coffee company premises to national and international
wholesalers;
6. Packaged coffee from each national and international wholesaler to final points of
sale.
Point 1 is not included in the transport calculation. Primary data for points 2 and 3 were col-
lected regarding distances travelled and quantities delivered. Primary data for point 4 were
collected on diesel oil consumption for the quantities delivered (transported using company
vehicles). In relation to point 5 it was extremely difficult to ascertain the deliveries made by
carriers to each wholesaler. Therefore estimates of the average distance between the factory
and a market town (discriminating between three market areas: regional, national and interna-
tional) were made on the basis of data provided by the company. Point 6 is not included in the
transport calculation since it was nearly impossible to collect accurate data about quantities
delivered regionally, nationally and internationally to each supermarket and shop, and from
these points of sale to each consumer. For these reasons this step is clearly underestimated.
E) Consumption. The consumption step is very difficult to measure and/or estimate because it
depends on so many different factors: consumer nationality and tastes (the amounts of coffee
and water used to make French coffee and Italian espresso differ greatly) or the type and
27
brand of coffee machine used (in particular for energy consumption) amongst others and these
differences are highly significant (+-30%). Nevertheless, in order to obtain some general in-
formation, selected data from a Pré Consultant LCA study (Pré Consultants, 2003) and spe-
cific information provided by Illycaffè Spa (Illycaffè Spa, 2003) were used. Data for the in-
ternational market refer to two different filter coffee machines used by households throughout
Europe (Pré Consultants, 2003): an electric aluminium coffee machine with a thermos jug
(machine A) and a coffee maker for use on an ordinary gas stove (machine B); we assumed
that 50% of the coffee delivered onto the international market was prepared with the first kind
of machine and 50% with the second. We further assumed the use of 7-gram mono-dose fil-
ters. Data for the Italian market refer to an electric espresso coffee machine (machine C) used
by households throughout Italy (Illycaffè Spa, 2003) assuming that 7 grams of coffee are used
for a single cup of espresso.
The use of professional coffee machines is not included because they are far more compli-
cated (they generally have other accessories that consume more energy). Water consumption
(for coffee preparation and for cleaning the machine) and sugar are also excluded as they are
assumed to be of little importance to the whole life cycle of the product and are also too diffi-
cult to model.
F) Disposal. Waste management includes packaging, coffee chaff and coffee grounds. We as-
sumed that all these materials were disposed of without any recycling.
Data quality and assumptions are expressed in the previous description of the stages included
in the inventory analysis. In general, specific on-site data were collected for the most impor-
tant aspects of the life cycle; or were obtained from scientific literature and/or commercially
available databases where on-site data were not available. The reference period for data col-
lection was the year 2001. The LCA software used was TEAM 3.0 by Ecobilan (Ecobilan,
1999(*)).
Impact assessment: main results. The impact assessment step was performed investigating
eight different impact categories (Table 1); Ecopoints were used as a general weighting factor.
An ecopoint is a measure of the overall environmental impact of a particular product or proc-
ess covering various environmental impacts (climate change, fossil fuel depletion, ozone de-
pletion, human toxicity, waste disposal, acid deposition, eutrophication, etc.) obtained by add-
ing together the score for each issue, calculated by multiplying the normalised impact with its
percentage weighting (Braunschweig and Müller-Wenk, 1993, Baldo, 2000; Ecobilan
1999(*)).
Figure 2 shows the individual contributions of the process stages (in %) to the category results
where the total of all contributions to each impact category is set at 100%.
28
Table 1. Impact categories.
Impact categories Method Unit
University of Leiden, Centre of
Air acidification g eq. hydrogen (H+)
Environmental Science (CML)
University of Leiden, Centre of
Aquatic Eco-toxicity 1e3 m3
Environmental Science (CML)
University of Leiden, Centre of
Eutrophication (water) g eq. phosphates (PO43-)
Environmental Science (CML)
University of Leiden, Centre of
Human toxicity g
Environmental Science (CML)
University of Leiden, Centre of
Terrestrial Eco-toxicity t
Environmental Science (CML)
Greenhouse effect Intergovernmental Panel on g eq. carbon dioxide
(direct, 100 y.) Climate (IPPC) (CO2)
World Meteorological Organiza- g eq. trichlorofluorome-
Depletion of ozone layer
tion (WMO) thane (CFC-11)
Photochemical oxidant forma- World Meteorological Organiza-
g eq. ethylene
tion. tion (WMO)
Cultivation Processing Packaging Transport Consumption Disposal
P hotoc hem ic al ox idant form ation
Depletion of the oz one lay er
Greenhous e effect
Terres trial Ec o-tox ic ity
Hum an Tox ic ity
E utrophic ation (water)
Aquatic Ec o-tox ic ity
A ir A c idification
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Figure 2. Impact categories studied.
From the figure it can be seen that the cultivation and the consumption stages make the great-
est impacts. The cultivation stage contributes the most to Terrestrial Eco-toxicity and Eutro-
phication (contributions greater than 97%); the consumption stage contributes the most to Air
acidification, Aquatic Eco-toxicity, Human Toxicity, Greenhouse effect, Depletion of ozone
layer and Photochemical oxidant formation (contribution exceeds 68% for all categories
cited). The disposal stage contributes to Aquatic Eco-toxicity (after consumption) and to Eu-
trophication (after cultivation). The contributions made by Transport are very limited but in-
29
fluence Photochemical oxidant formation, Greenhouse effect, Human Toxicity and Air acidi-
fication (after consumption and cultivation) and the Depletion of ozone layer and Aquatic
Eco-toxicity (after consumption but before cultivation). The contributions of the processing
and packaging stages are almost negligible (less than 1.7% for all categories).
The impact categories affected by process steps and the main emissions contributing thereto
are set out in Table 2. Only process stages with a contribution higher than 5% to each impact
categories are included in the table.
Table 2. Impact categories: main causes.
Process Impact categories %* Mainly caused by %*
Cultivation Terrestrial Eco-toxicity 100 Copper 100
Eutrophication (water) 97 Phosphates 97
Air acidification 31 Ammonia (NH3) 28
Greenhouse effect 12 Carbon Dioxide (CO2) 9
Consumption Human Toxicity 96 Sulphur Oxides (SOx) 60
Depletion ozone layer 96 Halon 1301 96
Photochemical oxidant for. 94 Hydrocarbons (except CH4) 64
Greenhouse effect 88 Carbon Dioxide (CO2) 74
Aquatic Eco-toxicity 76 Cadmium 43
Air acidification 69 Sulphur Oxides (SOx) 61
Disposal Aquatic Eco-toxicity 19 Cadmium 12
* percentage referred to the total result of a given impact category
Figure 3 shows a general weighting factor based on Ecopoints (Braunschweig and Müller-
Wenk, 1993, Baldo, 2000; Ecobilan 1999(*)) relating to energy and waste, air emissions and
water emissions for each life cycle stage considered. From the analysis of the figure it is evi-
dent that, in a general comparison among these three categories of ecopoints, air emissions are
the most relevant and they are mainly caused by SO2, NOx, CO2 emission in the consumption
stage connected to energy consumption during the use of coffee machines.
Conclusions: interpretation of main results
For a better understanding of the importance of environmental management conducted under
a life cycle perspective, figure 4 shows the Ecopoints applicable solely to the company (i.e.
from figure 1 the steps included in the company box plus the step relating to distribution made
by company vehicles). An analysis of environmental impacts made at company level alone in
order to make environmental improvements would steer management towards targeting al-
most exclusively the distribution stage (e.g. improvements to the company vehicle pool) and
the coffee roasting stage (e.g. improving energy consumption, air emissions and waste
management).
30
Figure 3. Ecopoints.
A ir W a te r E n e r g y & W a s te
120
100
80
Ecopoints
60
40
20
S t o rin g , c le a n in g a n d R o a s t in g M illin g P a c k a g in g D is t rib u t io n
w e ig h t in g
P r o c e s s in g s t e p s
Figure 4. Company level: Ecopoints.
Accordingly, at company level the main environmental improvements that could be addressed
are:
- Air emissions – principally due to fuel consumption of vehicles for local deliveries (all the
vehicles used for local deliveries run on diesel fuel) and to a lesser extent from the combus-
tion of fossil fuels in the roaster (natural gas). Improvements in fuel consumption would en-
able air emissions to be lowered .
31
- Energy consumption – electricity consumption refers to single processing steps and to fork-
lifts that are powered by electric batteries. Also improvements in energy efficiency would en-
able air emissions to be lowered.
- Waste management – waste management at company level is mainly related to coffee chaff.
At present this solid waste is disposed of alongside other urban refuse. Although the company
would be interested in seeking an alternative use, it has been discouraged from taking any ini-
tiative due to the small quantities concerned.
Under a life cycle perspective the company should approach its environmental management
decision making differently and concentrate mostly on the cultivation and consumption
stages.
- Cultivation data – although not all the data relating to cultivation have been included, it is
evident (see figure 2) that this step has a significant impact on the entire coffee life cycle,
therefore it is fundamental for the company to include the data in its environmental considera-
tions. Environmental improvements could most probably be achieved by choosing organic
and/or sustainable coffee farms as suppliers.
- Solid wastes – at company level the main solid waste is coffee chaff, but when the consump-
tion step is also taken into consideration then coffee grounds make up the largest proportion
of solid waste (apart from packing materials). Instead of being disposed of, coffee grounds
could be used as food for worms as well as for compost. The company should place bins for
composting food waste in each point of consumption to which they deliver. These can then be
collected when subsequent deliveries are made and contents used for composting in a worm
farm. The worms would process the coffee grounds into fertilizers. The amounts obtained
through collecting coffee grounds in this way together with the coffee chaff produced on the
company premises could allow it to start up a small-scale enterprise for vermiculture or com-
post production.
- Energy consumption – other methods to increase the environmental sustainability of the
company could involve setting up projects in collaboration with manufacturers of coffee ma-
chines where such projects undertake joint research aimed at improving energy efficiency.
Energy waste could also be prevented by information campaigns to heighten consumer
awareness (both at professional and household level).
- Packaging – even though this step makes no great impact, investigating more recyclable al-
ternatives to the current types of packaging used could nevertheless be very worthwhile.
From the above LCA clearly emerges as a useful tool to provide information for effective en-
vironmental management under a life cycle perspective, and as one that does not limit im-
provement opportunities to the physical confines of a company alone.
32
Acknowledgements
We would like to thank the staff of Miscela d’Oro SpA, Messina (Italy), for their cooperation.
Special thanks to Dr. Marilena Cioni for her active collaboration. www.misceladoro.com
References
Baldo G.L., LCA: Life Cycle Assessment, Ipa servizi Editore, Milano, Italia 2000;
Barbiroli G. Produzione e commercio internazionale del caffè Riccardo Patron Bologna 1970;
Braunschweig A. et Müller-Wenk R., Oekobilanzen für Unternehmungen. Eine Wegleitung
für die PraxisVerlag, Paul Haupt, Bern, pp.224. 1993. Updated version by BUWAL,
1998;
Brentrup F., Küsters J., Lammel J., Kuhlmann H. Methods to estimate on-field nitrogen emis-
sions from crop production as an input to LCA studies in the agricultural sector Interna-
tional Journal of LCA 5 (6) 200 pg. 349-357;
Camaggio Sancinetti G., Nicoletti G.M. Il ciclo di lavorazione del caffè ed i sui sottoprodotti
Nota 1: Aspetti quantitativi Industrie alimentari XXXIV 1995 pg. 1137-1146;
Diers A., Langowski H.-C., Pannkoke K., Hop R., Produkt-Ökobilanz vakuumverpackter
Röstkaffee Ecomed publishers LCA Documents vol.3 1999;
Ecobilan Group Data for Environmental Analysis and Management (DEAMTM) Version 3.0,
1999;
Ecobilan Group Tools for Environmental Analysis and Management (TEAMTM) Version 3.0,
1999(*);
Hauschild M., Estimating pesticide emissions for LCA of agricultural products in Weidema
B.P., Meeusen M.J.G. Agricultural data for Life Cycle Assessment Agricultural Eco-
nomics Research Institute (LEI), The Hague, 2000, Vol. II;
Heathwaite L., Flows of phosphourous in the environment: identifying pathways of loss from
agricultural land in Weidema B.P., Meeusen M.J.G. Agricultural data for Life Cycle As-
sessment Agricultural Economics Research Institute (LEI), The Hague, 2000, Vol. II;
Illycaffè Spa e.mail 4/4/03;
ISO 14040: 1997 Environmental management. Life cycle assessment. Principles and frame-
work;
ISO 14041: 1998 Environmental management. Life cycle assessment. Goal and scope defini-
tion and inventory analysis;
ISO 14042: 2000 Environmental management. Life cycle assessment. Life cycle impact as-
sessment;
ISO 14043: 2000 Environmental management. Life cycle assessment. Life cycle interpreta-
tion;
Mirulla R., La gestione ambientale nella supply chain: esperienze e metodi di coinvolgimento
dei fornitori in Scaramazza R. La fabbrica verde. Certificazione ambientale e imprese
sostenibili, Nuovo Studio Tecna, Roma 2001;
Pré Consultants SimaPro 5.1, Amersfoort, Netherlands, 2003; www.wri.org.
33
Slaughter offal to feed or to fuel - a comparison of systems in terms of
energy balances, land use and emissions
Hans B. Wittgren1*, Helena Elmquist2 and Ingrid Strid Eriksson2
1. JTI - Swedish Institute of Agricultural and Environmental Engineering, P.O. Box 7033, SE-750 07
Uppsala, Sweden
2. SLU - Swedish University of Agricultural Sciences, Department of Biometry and Engineering, P.O.
Box 7032, SE-750 07 Uppsala, Sweden
* Corresponding author. Current address: Linköping University, Environmental Science Programme,
Campus Norrköping, SE-601 74 Norrköping, Sweden, Phone: +46 11 36 33 48, E-mail:
[email protected]
Abstract
Large quantities of slaughter offal are being processed to meat- and bonemeal, but due to
the current legislation its use as animal feed is very restricted. One option is to use slaughter
offal directly as fuel in incineration plants. This study compared two alternative systems: (1)
Processing of slaughter offal to meat- and bonemeal (feed) and to animal fat (fuel), and (2)
Direct incineration of slaughter offal (fuel). In order to make the two alternatives comparable,
the core systems were expanded with complementary systems to produce the same
amounts of feed, generate the same amounts of energy and use the same areas of arable
land. The analysis of the two alternatives used existing life cycle data and assessments in
combination with the two simulation tools ORWARE (ORganic WAste REsearch model) and
SALSA (Systems AnaLysis for Sustainable Agriculture). The comparison showed that proc-
essing of slaughter offal to feed (alt. 1) was superior to using it as fuel (alt. 2) in terms of both
energy balances and all considered emission impact categories (global warming, acidification
and eutrophication potential). The difference between the two alternatives may be interpreted
as the price, in terms of energy and emissions, that has to be paid to lower the risk for recy-
cling of the bovine spongiform encephalopathy (BSE) agent.
Keywords: energy balances, environmental systems analysis (ESA), life cycle assessment
(LCA), slaughter offal, substance flow accounting (SFA).
Introduction
Following the discovery of bovine spongiform encephalopathy (BSE) in 1986, the countries
of Europe and the European Union have successively implemented several measures to pre-
vent recycling of the BSE agent (Heim and Kihm, 2003). Some of the most important meas-
ures have implied banning of feeding meat- and bonemeal (MBM) to farm animals. Initially,
only feeding of MBM to ruminants was banned (European Commission, 1994), but due to the
risk for cross-contamination most countries in Europe banned feeding of MBM to all farm
animals in 2001 (European Commission, 2000). Although it is possible that this total feed ban
will be relaxed in the future, it is necessary to consider alternative options for handling of
slaughter offal.
34
The most obvious alternative is to use the slaughter offal as fuel by direct incineration in
power plants. This should be a safe option in terms of managing BSE, but how does direct in-
cineration come out in comparison with production of MBM as feed in terms of resource use
and environmental impact? Such a comparison should provide part of the background for de-
cisions regarding the future handling of slaughter offal. In addition, analysis of different sys-
tems for handling slaughter offal should provide necessary input to life cycle assessments
(LCAs) of meat production.
The objective of this study was to make a quantitative comparison of two alternative systems
(Figures 1 and 2):
1. Processing of slaughter offal to MBM (feed) and to animal fat (fuel)
2. Direct incineration of slaughter offal (fuel)
The comparison was based on systems analysis in terms of energy balances, land use and
emissions.
Slaughterhouses
Emissions Emissions
1 ton slaughter offal
CORE COMPLEMENTARY
Wastewater Willow biomass
0.29 ha arable land
treatment production
Wastewater
MBM
Processing 147.4 kg feed protein Biomass
16.7 kg digestible P
Animal fat
Incineration 8 347 MJ electricity Incineration
27 289 MJ heat
Energy carriers Energy carriers
(Materials) (Materials)
Figure 1. Scheme of alternative 1: Processing of slaughter offal to MBM (feed) and to animal
fat (fuel).
Methods
Functional unit, core and complementary systems
The two alternatives were designed to treat ‘1 ton (wet weight) of slaughter offal’ - the func-
tional unit (FU) of the analysis. This treatment was performed in the ‘core’ systems (Figures 1
and 2). To make the two alternatives comparable, also ‘complementary’ systems were needed
(Figures 1 and 2). These were designed according to the following logic:
35
Slaughterhouses
Emissions Emissions
1 ton slaughter offal
CORE COMPLEMENTARY
Rapeseed cake
0.29 ha arable land
production
147.4 kg feed protein
16.7 kg digestible P
Supplementary
production of
digestible P
Energy generation
Incineration 8 347 MJ electricity
27 289 MJ heat
Energy carriers Energy carriers
(Materials) (Materials)
Figure 2. Scheme of alternative 2: Direct incineration of slaughter offal (fuel).
The core system of alt. 1 produced MBM containing certain amounts of feed protein and di-
gestible phosphorus (P). Corresponding amounts were generated in alt. 2 by including feed
production, based on rapeseed and inorganic P, in the complementary system. The production
of rapeseed required arable land, and an equivalent area of arable land was used in the com-
plementary system of alt. 1 for production of willow biomass. This choice was motivated by
the fact that it is desirable to increase the production of renewable energy carriers. The willow
biomass was incinerated to produce electricity and heat. Finally, energy generation in a natu-
ral gas powered plant was required in the complementary system of alt. 2, to balance the dif-
ference between energy generated from incineration of animal fat + willow biomass, in alt. 1,
and incineration of slaughter offal, in alt. 2. With this design, both alternatives produced the
same amounts of feed protein and digestible P, used equivalent areas of arable land and gen-
erated the same amounts of electricity and heat (Figures 1 and 2). The feed protein, however,
was not of the same quality in the two alternatives, since the amino acid composition of MBM
and rapeseed cake are quite different.
In terms of amino acid composition, soybean meal is more similar to MBM than rapeseed
cake (Rodehutscord et al., 2002). Therefore we analysed also feed production based on soy-
bean meal (from Brazil), and compared it with feed production based on rapeseed cake (from
Sweden). A complete system with soybean meal could, however, not be analysed, since we
did not have data to set up a realistic land use alternative to soybean (such as willow is in the
case of rapeseed).
36
Material flows, such as phosphorus ore or process chemicals, were not included in the analy-
sis. Furthermore, the building phases of processing, wastewater treatment, and power plants
were not considered.
Sub-systems
In the following, the different sub-systems (grey boxes in Figures 1 and 2) are briefly de-
scribed in terms of data sources, assumptions and calculations. Also the soybean meal sub-
system, and how transportation was handled, is described.
Processing (in alt. 1): In Sweden only one company, Konvex AB, processes slaughter offal to
MBM and animal fat. The data used in this study were obtained from their largest plant,
Krutmöllan, which processes approximately 60,000 tons (wet weight) annually (Konvex AB,
2001; L. Virta, personal communication, January 15, 2003). The composition of incoming
slaughter offal and the resulting products are summarised in Table 1. Data on energy use and
emissions, for the Processing sub-system as well as for the others, were recalculated to relate
to the FU (Tables 2 and 3 in the Results section).
Table 1. Composition of slaughter offal and the products meat- and bonemeal (MBM) and
animal fat at the Konvex plant in Krutmöllan, Sweden (expressed as kg per ton wet weight of
slaughter offal, e.g., the functional unit (FU)).
Component Water Dry matter Protein Fat Digestible P 2
Fraction ---------------------------------- kg FU-1 ----------------------------------
Slaughter offal 1 650 350 147 150 16
Meat- and bonemeal (MBM) 7 220 147 20 16
Animal fat 3 130 0 130 0
1. No data was available on the content of protein, fat and phosphorus in the slaughter offal. These values have been calcu-
lated from the data on product contents, assuming no losses in the process. According to Konvex AB, losses are < 1% (L.
Virta, personal communication, January 15, 2003)
2. The content of digestible P in MBM was calculated from data on total P content, assuming that 81% is digestible (Rode-
hutscord et al., 2002)
Wastewater treatment (in alt. 1): The wastewater was first treated at the slaughter offal proc-
essing plant, and this treatment was included in the Processing sub-system (above) in terms of
energy use and emissions. The pre-treated wastewater was piped to the municipal wastewater
treatment plant in Kävlinge for further treatment. Data from the wastewater treatment plant
showed that the share coming from the slaughter offal processing plant made up for 3.1%
(flow), 3.7% (BOD7), 18.8% (total nitrogen) and 2.4% (total phosphorus) of the total load in
2001 (Kävlinge kommun, 2002). Based on these shares, and somewhat arbitrary, 10 % of the
energy use in the wastewater treatment plant was allocated to the treatment of wastewater
from the slaughter offal processing plant.
37
Incineration (in alt. 1 and 2): The Swedish ORWARE (ORganic WAste REsearch) model has
a sub-model for incineration of organic waste in a combined heat and power plant (Björklund,
1998; Eriksson et al., 2002). This sub-model was used to calculate energy balances and emis-
sions for incineration of animal fat and willow biomass in alt. 1, and for incineration of
slaughter offal in alt. 2.
Willow biomass production (in alt. 1): An annual average biomass yield of 6400 kg (dry
weight) ha-1 was calculated from harvest estimates presented by the Swedish company Agro-
bränsle AB (2003). Energy use and emissions (expressed as global warming potential, acidifi-
cation potential and eutrophication potential) in a willow biomass cropping system were ob-
tained from an LCA performed for conditions in New York state by Heller et al. (2003). Their
assessment considered an estimated 23 year lifespan of a cropping system with seven harvest
rotations. The system included field preparation, planting, weed control, coppice, fertilisation,
harvest, willow stool elimination, and production of various inputs such as herbicides and fer-
tilisers. In our study we used their base case, where ammonium sulphate was used as fertiliser,
and recalculated the results to relate them to the FU.
Rapeseed cake production (in alt. 2): One part (SALSA-arable) of the Swedish SALSA (Sys-
tems AnaLysis for Sustainable Agricultural production) model was used for calculations of
energy use and emissions in rapeseed cake production (Elmquist et al., 2003). SALSA, simi-
lar to ORWARE (mentioned above), is an energy and substance flow accounting model
(SFA) complemented with life cycle assessment (LCA) methodology for evaluation of envi-
ronmental impacts (Elmquist et al., 2004; Strid Eriksson et al., 2004).
The rapeseed cake production system in SALSA-arable includes different field operations, the
soil/crop system, drying and preparation of rapeseed oil and cake, and production of inputs
such as seed and fertiliser. The yield of rapeseed was based on Swedish harvest data for win-
ter and spring rapeseed from the dominating production regions (Skåne and Östergötland, re-
spectively). It was further assumed that the rapeseed harvest was processed, with 73% ex-
change, to rapeseed cake, containing 32% protein and 1.3% P. Digestible P was assumed to
represent 30% of the P content (Rodehutscord et al., 2002).
These assumptions resulted in a protein (in rapeseed cake) harvest of 501 kg ha-1 yr-1 and a
need for 0,29 ha to equal the production of MBM protein per FU in alt. 1. Based on market
prices, the value of rapeseed cake was assumed to be 33% of the total value of the rapeseed
yield. This share of energy use and emissions for production of rapeseed was allocated to
rapeseed cake, and the rest to rapeseed oil. Since the inclusion of rapeseed cake in the com-
parison was managed via an allocation procedure, there was no need for a sub-model analys-
ing the fate of the rapeseed oil.
The amount of digestible P in rapeseed cake was only 1.8 kg FU-1 and the feed had to be
complemented with inorganic P to contain the same amount as in alt. 1. Data on energy use
38
and emissions for production of P fertiliser were used (Sundqvist et al., 2002), along with the
estimate that 80% of the P content is digestible (Rodehutscord et al., 2002).
Soybean meal production (for comparison with rapeseed cake production): The SALSA-
soybean (Strid Eriksson et al., 2004) description of the production system for soybean meal
(of Brazilian origin) include different field operations, the soil/crop system, drying of seeds,
transport from farm to extraction plant, extraction, transport from extraction plant to harbour,
ocean transport from Brazil to Germany, sea transport from Germany to Sweden, and produc-
tion of inputs such as seed and fertiliser. A soybean harvest of 2200 kg ha-1 yr-1 was assumed,
with a 70% exchange of soybean meal containing 40% protein.
These assumptions resulted in a protein (in soybean meal) harvest of 616 kg ha-1 yr-1 and a
need for 0,24 ha to equal the production of MBM protein per FU in alt. 1. The value of soy-
bean meal was assumed to be 70% of the total value of soybean, and energy use and emis-
sions allocated accordingly. The digestible P deficit was corrected for in the same way as for
rapeseed cake.
Energy generation (in alt. 2): It was assumed that the difference between electricity and heat
generated by incineration of animal fat + willow biomass (in alt. 1) and incineration of MBM
(in alt. 2) was generated in a heat and power plant fuelled with natural gas. Data on conver-
sion efficiency and emissions from a plant in Sweden (Heleneholmsverken) were obtained
from Uppenberg et al. (2001). The overall energy conversion efficiency was 94%, out of
which 25% was electricity and 75% heat.
Transportation: Transportation of the different products were not included in the analysis,
with the exeption of the transports concerning soybean meal production (mentioned above).
However, separate calculations of energy use and emissions were made for the transportation
of slaughter offal to processing and willow biomass to incineration. Data on transport dis-
tances and amounts concerning slaughter offal to processing were obtained from the process-
ing company (Konvex AB). For willow biomass, a transportation distance of 100 km between
biomass production and incineration was assumed. Data on diesel consumption and emissions
for a heavy truck with trailer (load capacity: 35 tons) were taken from Sundqvist et al. (2002).
Impact assessment
Calculations of energy use included also upstream use, for production of the energy carriers,
to yield primary energy use (Eriksson et al., 2002; Sundqvist et al., 2002). All results on en-
ergy use are thus expressed as primary energy. Upstream emissions from production of en-
ergy carriers were also calculated and are included in results on emissions.
Emissions to air and water were recalculated to the impact categories global warming poten-
tial (time horizon 100 years), acidification potential and eutrophication potential using
weighting factors according to Houghton et al. (1996) - for global warming potential - and ac-
39
cording to Lindfors et al. (1995) - for acidification (maximum scenario) and eutrophication
potential.
Results
The primary energy use was 4.3 times larger in alt. 2 as compared to alt. 1, mainly due to the
generation of electricity and heat from natural gas needed to compensate for the incineration
of willow biomass (Table 2).
The sum of primary energy use in soybean meal production was 2458 MJ FU-1 (not shown) as
compared to 1012 MJ FU-1 for rapeseed cake production. The difference was mainly due to
oil consumed for ocean transportation.
Table 2. Primary energy use per functional unit (FU) in different sub-systems of the two al-
ternatives (Figures 1 and 2).
Primary energy use
Electri- Diesel Natural Oil / Unspe- Sum
1
city gas Coal cified
---------------------------- MJ FU-1 ---------------------------------
Alternative 1
Processing of slaughter offal 963 3 232 4 195
Incineration of animal fat 72 72
Wastewater treatment 27 2 29
Willow biomass production 578 679 1 257
Incineration of willow biomass 2 052 2 052
Sum 3 115 578 3 232 2 679 7 606
Alternative 2
Incineration of slaughter offal 615 615
Rapeseed cake production 142 256 559 55 1 012
Supplementary production of digestible P 153 161 482 795
Energy generation 30 440 30 440
Sum 910 256 31 160 536 0 32 862
1 The source of data on willow biomass production (Heller et al., 2003) did not specify energy carriers beside diesel.
The global warming potential (Table 3) reflected the energy use to a large extent, since CO2-
emissions from incineration and energy generation were the dominant contributions. Emis-
sions of N2O were of the same magnitude in the two alternatives, 185 and 231 kg CO2-
equivalents FU-1, and thus a substantial contribution, 37%, in alt. 1, compared to only 9% in
alt. 2. Willow biomass and rapeseed cake production were the most important sources of N2O
in alt. 1 and 2, respectively.
The sum of emissions contributing to acidification potential, particularly NOx and to some ex-
tent SO2, were similar in alt. 1 and alt. 2 (Table 3). It was, however, noted that the acidifica-
tion potential was a factor 3.4 larger for willow biomass production than for rapeseed cake
40
production, and a factor 5.6 larger for soybean meal production than for rapeseed cake pro-
duction (not shown). In the later case, larger energy use and less efficient NOx reduction dur-
ing combustion in the soybean meal production were the reasons. Concerning willow biomass
production, it was assumed that the larger emissions, in comparison with rapeseed cake pro-
duction, were at least partly a consequence of the different sources of data for calculations of
emissions from, e.g., field operations and fertiliser manufacturing.
Eutrophication potential was more than twice as a high in alt. 2 as compared to alt. 1 (Table
3). The difference was a result of very different assumptions regarding leaching of nitrate and
phosphorus in willow biomass production and rapeseed cake production, respectively. The as-
sessment of willow biomass production assumed no leaching (Heller et al., 2003), while the
SALSA-arable model assumed leaching of 94 kg NO3- ha-1 yr-1 and 0.9 kg P ha-1 yr-1 from
rapeseed production. SALSA-soybean assumed similar leaching from soybean production,
but since 70% was allocated to soybean meal, as compared to only 33% to rapeseed cake, the
leaching per FU was higher in the case of soybean meal. In addition, NOx-emissions were
higher for soybean meal production, making the total eutrophication potential 111 kg O2-
consumption FU-1 (not shown) as compared to 55.6 kg in rapeseed cake production (Table 3).
In alt. 1, transportation of slaughter offal to processing and of willow biomass to incineration
used 509 and 327 MJ FU-1, respectively. The reason for higher use in transportation of
slaughter offal, although quantities were smaller, is that average distances are much longer
than was assumed for willow biomass transportation (100 km). Taken together, these two
transports contributed 63 kg CO2-equivalents FU-1 to global warming potential, 0.5 kg SO2-
equivalents FU-1 to acidification potential, and 3.0 kg O2-consumption FU-1 to eutrophication
potential. If these were the only transports, they would increase primary energy use in alt. 1
by 11%, global warming potential by 13%, acidification potential by 9% and eutrophication
potential by 10%.
Discussion
The comparison showed that processing of slaughter offal to feed (alt. 1) was superior to us-
ing it as fuel (alt. 2) in terms of both energy balances and all three emission impact categories.
The difference between the two alternatives may be interpreted as the price, in terms of en-
ergy and emissions, that has to be paid to lower the risk for recycling of the BSE agent.
The comparison did not take into account the fact that protein from MBM and rapeseed cake
have different amino acid composition. It is therefore not realistic to fully exchange one for
the other in feed preparation (Rodehutscord et al., 2002). In a more stringent comparison, alt.
2 would thus have to be complemented with further sub-systems for production of supple-
mentary feed ingredients, probably leading to an even larger difference in terms of energy use
and emissions.
41
Table 3. Emissions to air and water expressed as the three impact categories global warming
(kg CO2-equivalents), acidification (kg SO2-equivalents) and eutrophication (kg O2-
consumption) per functional unit (FU) for different sub-systems of the two alternatives (Fig-
ures 1 and 2).
Global warming Acidification Eutrophication
kg CO2-eq. FU-1 kg SO2-eq. FU-1 kg O2-cnsmp. FU-1
Alternative 1
Processing of slaughter offal 211 1.4 5.7
Incineration of animal fat 1 0.3 2.2
Wastewater treatment 0 0.0 1.7
Willow biomass production 228 1.6 7.6
Incineration of willow biomass 59 1.9 14.6
Sum 499 5.2 31.8
Alternative 2
Incineration of slaughter offal 69 0.4 3.6
Rapeseed cake production 238 0.5 55.6
Supplementary production of digestible P 54 0.6 1.1
Energy generation 2 091 4.2 30.8
Sum 2 451 5.7 91.1
Soybean meal is more similar to MBM and its use as a feed ingredient has increased follow-
ing the ban on feeding of MBM to farm animals. In comparison with rapeseed cake produc-
tion, soybean meal production was calculated to have higher energy use and larger emissions.
The major reason is that soybean meal was assumed to be produced in Brazil, with ocean
transportation to Sweden included in the calculations, while rapeseed cake was assumed to be
produced locally in Sweden, with no transportation needed.
It is likely that road transportation (not included in Tables 2 and 3) would add more to energy
use and emissions in alt. 1 than in alt. 2. Transport distances for slaughter offal would proba-
bly be shorter for alt. 2 since it is reasonable to assume that incineration can be implemented
at several places, while in alt. 1 all slaughter offal is taken to one single processing plant. A
further reason is that the weight of willow biomass (3.8 tons FU-1) is much larger than the
weights of any of the other products involved in the comparison.
Since energy balances were one focus of the study, willow biomass production was chosen as
alternative to rapeseed production concerning land use. Another option could have been to
consider permanent fallow as the alternative land use. This would have made the comparison
turn around since complementary energy generation would have been needed in alt. 1 instead
of alt. 2. Permanent fallow was, however, considered as too different economically from rape-
seed production, when the farm economics dimension was not included in the comparison.
This is not to say that alt. 1 and alt. 2 in the present study were economically equal. To serve
as true decision support, most comparisons of systems need to include economic and social
dimensions in addition to the resource and environmental dimensions.
42
The present study took short-cuts in the sense that already existing tools and assessments, not
designed to analyse systems for management of slaughter offal, where used. This is some-
times necessary, since the choice is between a simplified analysis or no analysis at all. The al-
ternative of a seam-less assessment, with tailor-made tools, is presently not economically pos-
sible in most applications. This state-of- the-art points to the need for development of agri-
food sector assessment tools that are highly compatible with each other.
References
Agrobränsle AB (2003). Kalkyl [online]. Agrobränsle AB, Örebro, Sweden. Available from:
http://www.agrobransle.se [Accessed 7 July 2003]. (In Swedish).
Björklund, A., 1998. Environmental systems analysis of waste management with emphasis on
substance flows and environmental impact. Thesis (licentiate). TRITA-KET-IM
1998:16, Department of Chemical Engineering and Technology, KTH - Royal Institute
of Technology, Stockholm, Sweden.
Elmquist, H., Strid Eriksson, I., Öborn, I. and Nybrant, T., 2003. An environmental systems
analysis of winter wheat, barley and rapeseed production on an arable farm in Sweden -
A study on effects of N fertilisation rates. Journal of Cleaner Production, submitted
manuscript.
Elmquist, H., Lindgren, U. and Mäkilä, K., 2004. Decision-making and environmental im-
pacts - A dynamic simulation model of a farm business. Food 21 Report, Department of
Biometry and Engineering, SLU - Swedish University of Agricultural Sciences, Upp-
sala, Sweden, in press.
Eriksson, O., Frostell, B., Björklund, A., Assefa, G., Sundqvist, J.-O., Granath, J., Carlsson,
M., Baky, A., and Thyselius, L., 2002. ORWARE - A simulation tool for waste man-
agement. Resources, Conservation and Recycling 36 (4), 287-307.
European Commission (EC), 1994. Commission Decision 94/381/EC of 27 June 1994 con-
cerning certain protection measures with regard to bovine spongiform encephalopathy
and the feeding of mammalian derived protein (Text with EEA relevance). Off. J. Eur.
Communities L 172 (07/07/1994), 23-24.
European Commission (EC), 2000. Council Decision 2000/766/EC of 4 December 2000 con-
cerning certain protection measures with regard to transmissible spongiform encephalo-
pathies and the feeding of animal protein. Off. J. Eur. Communities L 306 (07/12/2000),
32-33.
Heim, D. and Kihm, U., 2003. Risk management of transmissible spongiform encephalo-
pathies in Europe. Rev. sci. tech. Off. int. Epiz. 22 (1), 179-199.
Heller, M. C., Keoleian, G. A. and Volk, T. A., 2003. Life cycle assessment of a willow bio-
energy cropping system. Biomass and Bioenergy 25, 147-165.
Houghton, J. T., Meiro Filho, L. G., Callander, B. A., Harris, N., Kattenburg, A. and Maskell,
K. (eds), 1996. Climate Change 1995: The Science of Climate Change: Contribution of
Working Group I to the Second Assessment Report of the Intergovernmental Panel on
Climate Change. Cambridge University Press, Cambridge, UK.
43
Konvex AB, 2001. Konvex AB Krutmöllan - Ansökan om tillstånd enligt 9 kap. miljöbalken.
Konvex AB, Lidköping, Sweden. (In Swedish).
Kävlinge kommun, 2002. Kävlinge avloppsreningsverk år 2001 - Miljörapport enligt mil-
jöskyddslagen. Kävlinge kommun, Kävlinge, Sweden. (In Swedish).
Lindfors, L.-G., Christiansen, K., Hoffman, L., Virtanen, Y., Juntilla, V., Hanssen, O.-J.,
Rønning, A., Ekvall, T. and Finnveden, G., 1995. Nordic Guidelines on Life-Cycle As-
sessment. Nord 1995:20, Nordic Council of Ministers, Copenhagen, Denmark.
Rodehutscord, M., Abel, H. J., Friedt, W., Wenk, C., Flachowsky, G., Ahlgrimm, H.-J., Joh-
nke, B., Kühl, R. and Breves, G., 2002. Consequences of the ban of by-products from
terrestrial animals in livestock feeding in Germany and the European Union: alterna-
tives, nutrient and energy cycles, plant production, and economic aspects. Arch. Anim.
Nutr. 56, 67-91.
Strid Eriksson, I., Elmquist, H., Stern, S., Nybrant, T., 2004. Environmental systems analysis
of pig production - the impact of feed choice. Int. J. LCA., submitted manuscript.
Sundqvist, J.-O., Baky, A., Carlsson, M., Eriksson, O. and Granath, J., 2002. Hur skall
hushållsavfallet tas om hand? Utvärdering av olika behandlingsmetoder. IVL Report B
1462, IVL - Swedish Environmental Research Institute, Stockholm, Sweden. (In Swed-
ish with English summary).
Uppenberg, S., Brandel, M., Lindfors, L.-G., Marcus, H.-O., Wachtmeister, A., Zetterberg, L.,
1999. Miljöfaktabok för bränslen - Del 2. Bakgrundsinformation och teknisk bilaga.
IVL Report B 1334B-2, IVL - Swedish Environmental Research Institute, Stockholm,
Sweden. (In Swedish).
44
Environmental assessment of contrasting pig farming systems in France
Basset-Mens, C.* and van der Werf, H.M.G*
*: INRA, UMR Sol Agronomie Spatialisation de Rennes-Quimper, 65, rue de Saint Brieuc CS 84215,
35 042 Rennes Cedex, France. Tel : +33 02 23 48 52 22. Fax : +33 02 23 48 54 30. Email : bas-
[email protected]
Abstract
The aim of this study was to produce a multicriteria environmental assessment of three con-
trasting pig farming systems : conventional Good Agricultural Practice (GAP), “Label Rouge”
(LR) and “Agriculture Biologique” (AB), using the Life Cycle Assessment (LCA) method. Av-
erage, favourable and unfavourable scenarios were defined and evaluated for each produc-
tion mode for seven impact categories. Expressed per hectare, LR and AB had lower im-
pacts than GAP for eutrophication and acidification, a higher (LR) or similar (AB) impact for
climate change and was less productive (respectively - 14% and – 45%). We identified “hot-
spots” for each system. However, for the three scenarios important margins of improvement
were highlighted.
Keywords: LCA; pig; uncertainty; hot-spots.
Introduction
In Brittany (France), the successful development of intensive pig production since Word War
II has led to a severe degradation of the environment. Hence, the current pig production
model is in crisis. Although alternative farming systems are favourably considered by society,
none of these have yet been assessed for their environmental impact as far as pig production is
concerned. The global diagnosis of the environmental performance of current and more pro-
spective farming systems appears as an emerging need and this evaluation should include an
estimation of the uncertainty of the main results (Huijbregts et al., 2001; Guinée et al., 2002).
Among environmental assessment approaches, Life Cycle Assessment (LCA) has been identi-
fied as a valuable tool for the environmental evaluation of farming systems (van der Werf and
Petit, 2002). The objectives of this study were to explore the diversity of pig farming systems
using LCA with a prospective view and to assess the robustness of the results. We were fac-
ing the following methodological issues: (i). How should contrasting farming systems with
different degrees of optimisation be described and how should an environmental inventory be
carried out? (ii). How should the uncertainty of the results be assessed? This paper proposes a
scenario-based approach for the environmental assessment of contrasting farming systems in-
cluding an estimation of the uncertainty of the results.
Materials and methods
Evaluation methodology
Environmental impacts associated with pig farming systems were evaluated using Life Cycle
Assessment (LCA). In this approach, the potential environmental impacts of a product are as-
sessed by quantifying and evaluating the resources consumed and the emissions to the envi-
45
ronment at all stages of its life cycle – from the extraction of resources, through the produc-
tion of materials, product parts and the product itself, and the use of the product to its reuse,
recycling or final disposal (Guinée et al., 2002). The present study only deals with the proc-
esses up to and including the production on the farm. We compared three scenarios for pig
production. The Good Agricultural Practice (GAP) scenario corresponds to conventional pro-
duction, optimised in particular with respect to fertilisation practices, as specified in the
French “Agriculture Raisonnée” standards (Rosenberg and Gallot, 2002). The “Agriculture
Biologique” (AB) scenario corresponds to organic agriculture according to the French version
of European rules for organic animal production (Ministère de l’Agriculture et de la Pêche,
2000) and the European rules for organic crop production (CEE, 1991). The Label Rouge
(LR) scenario corresponds to the “Porc Fermier Label Rouge” quality label, as specified in its
rules of production (Groupements des fermiers d’Argoat, 2000). Impacts were referenced to
two functional units referring to two farming system functions: units of pig produced (in kg
live weight at slaughter) and units of land used (in ha).
Farming system scenarios
Data on crop production, transport distances, feed composition and system performance were
based on statistics, estimates from experts and data from growers’ associations. The feed
compositions respected the specific rules of each production mode. According to the function
and development stage of the animal, six diets were used for GAP (van der Werf et al., in
press) and LR and five diets for AB.
For all crops, production corresponded to good agricultural practice, i.e. fertilisation accord-
ing to anticipated crop needs and integrated pest management for GAP and LR. For the three
scenarios, we assumed that pig manure (liquid manure for GAP, solid manure for LR, com-
posted solid manure for AB) was used to fertilise Brittany-grown crops used as feed ingredi-
ents. The overall amount of manure or compost used for crop-based feed ingredients was ad-
justed, so as to correspond to the amount of manure that feeding the feed these ingredients
were part of would yield. Additional N, P and K fertiliser was applied according to crop
needs. For LR and GAP, yield levels were averages for 1996 – 2000 (AGRESTE, 2001; FAO,
2002). The yield levels of AB crops were defined according to the judgement of experts from
the region the crops were produced in. Yields are lower than conventional yields: from -15%
for maize to -40% for wheat and barley. For the processes concerning the transformation of
crop products into feed ingredients and the production of feed, the inventory of resources used
and emissions to the environment was limited to resources and emissions associated with the
use of non-renewable energy. For ingredients resulting from processes yielding more than one
product (e.g. soy cake, wheat gluten), resource use and emissions were allocated according to
the economic value. Data for feed production (involving, amongst others: grinding, heating,
mixing, pelleting) were from Sanders (2000).
We distinguished two stages in pig production: piglet production (PP) and weaning to slaugh-
tering (WS). For GAP, data on technical performance for both PP and WS (Table 1) were ac-
46
cording to published statistics (ITP, 2001). For LR, data concerning PP were from ITP (2001),
data concerning WS were averages supplied by the LR producers’ association. For AB, data
on technical performance were based on an optimised model of organic pig production (Ber-
ger, 2000) adjusted according to expert judgement. For GAP and LR, manure was stored,
while for AB, manure was composted, involving one or two turnings and taking 4 to 5
months.
Table 1. Characteristics of GAP, LR and AB pig farming systems.
GAP* LR* AB*
Piglet production
Housing Slatted floor Outdoor Outdoor
Weaned piglet/productive sow/year 25.5 22.6 20.3
Weaning age, days 25.7 28 42
Surface per sow, m2 <4 1000 1000
Feed per sow (boar included), kg/year 1313 1490 1695
Weaning to slaughtering
Housing Slatted floor Straw litter Straw litter
Surface per pig, m2 0.85 2.6 2.3
Feed : gain ratio 2.7 2.9 3.2
Slaughter age, days 175 190 195
Slaughter weight, kg 113 115 120
*: GAP = Conventional production with good agricultural practice; LR = respects the “Porc Fermier Label Rouge” quality
label; AB = organic agriculture
Climate Land Kg of
Eutroph. Acidif. Energy Pesticide
change use pig
kg ha kg ha kg ha kg ha kg kg ha
ha
350
300
250
200
150
100
50
0
G L A G L A G L A G L A G L A G L A G L A G L A G L A G L A G L A G L A
G= Good Agricultural Practice; L= "Label Rouge"; A="Agriculture Biologique"
Figure 1. LCA results and estimated uncertainty for the GAP, LR and AB scenarios, ex-
pressed per kg of pig and per hectare as a percentage of the average result for GAP.
47
Eutrophi- Climate Acidifica- Terrestrial Energy
cation change tion toxicity use
140
Litter production
120
Building construction
100
Crop and feed
80 production
Compost production
60
40 Weaning to slaughtering
20 Piglet production
0
G L A G L A G L A G L A G L A
G=Good Agricultural Practice; L= Label Rouge; A=Agriculture Biologique
Figure 2. Contribution of six life cycle stages to five impact categories for the GAP, LR and
AB scenarios, expressed per hectare and as a percentage of GAP.
Inventory data
Inventory data (resource use and emissions to the environment) were based on input-output
data collected from different sources. Data associated with the production and delivery of in-
puts for crop production (fertilisers, pesticides, tractor fuel and machines) were derived ac-
cording to Nemecek and Heil (2001). Data for energy carriers for road and sea transport were
from the BUWAL 250 database (BUWAL, 1996). Data concerning resource use and emis-
sions associated with buildings (production and delivery of materials, construction) were from
Kanyarushoki (2001). Ammonia emissions due to the application of ammonium nitrate fertil-
iser were estimated according to ECETOC (1994) and ammonia emissions following applica-
tion of slurry were according to Morvan et Leterme (2001). Ammonia and nitrous oxide emis-
sions from slurry in pig buildings were from IPCC (1996) and UNECE (1999). Methane
emissions due to enteric fermentation and housing type were from IPCC (1996). Data on the
production of excreta, emissions from buildings, during storage, during composting and from
crops and paddocks, were chiefly obtained with the support of an expert panel from the Insti-
tut National de la Recherche Agronomique. The panel comprised: J. Y. Dourmad, Th. Mor-
van, J.M. Paillat, P. Robin and F. Vertès. The panel based its expertise on their experiments,
simulation models and on their interpretation of the available literature. The environmental
impacts due to the slurry or the manure applied on the crops were associated with the crop
production.
48
Uncertainty analysis
In order to explore the robustness of our results, an uncertainty analysis was conducted. Crop
yields, WS feed to gain ratio, field emissions (NH3, N2O and NO3) and emissions of NH3 and
N2O from buildings, manure storage and composting were identified as important issues for
the variability of results. For each emission and parameter concerning these issues, specific
high and low values reflecting what we coined “realistic” rather than overall variability were
defined in addition to the default reference value. The intervals thus defined were much larger
for emissions at field level (on average –40% and +83%) than for emissions in buildings (on
average ±18%) and for technical parameters (±6% for feed to gain ratio and around ±10-15%
for crop yield). The “realistic” uncertainty interval thus defined contains about two thirds of
the overall variability for the parameter concerned. We combined all favourable values (better
performance, lower emissions) for key-parameters to obtain a favourable variant of each sce-
nario and similarly we combined all unfavourable values (poorer performance, higher emis-
sions) to obtain an unfavourable variant of each scenario. In addition to the default scenario
based on the reference values, these two variants are proposed as indicators of overall uncer-
tainty for each scenario.
Impact assessment
The environmental impact categories considered in this study are: eutrophication (in kg PO4-
eq), climate change (in kg CO2-eq), acidification (in kg SO2-eq), terrestrial toxicity due to
heavy metal accumulation (in kg 1.4-dichlorobenzene-eq), energy use (in MJ Lower Heating
Value (LHV)-eq), land use (in m2.year) and pesticide use (in kg of active ingredient). Eutro-
phication, acidification, terrestrial ecotoxicity potentials were calculated using characterisa-
tion factors by Guinée et al. (2002). Global warming Potential for a 100 year time horizon
(GWP100) was calculated according to the GWP100 factors by IPCC (Houghton et. al.,
1996). Energy use was calculated using the LHV proposed in the SIMAPRO 1.1 method (PRé
Consultants, 1997). Pesticide use (in kg active matter used) refers to the global quantity of
pesticide used for crop production.
Results
Per kg of pig produced, eutrophication was smallest for LR (0.0166 kg PO4-eq) followed by
GAP (0.0208) and AB (0.0216). Per ha, eutrophication was largest for GAP (38.3 kg PO4-eq)
intermediate for LR (26.4) and smallest for AB (22.9). Both per kg and per ha, overall uncer-
tainty was large and asymmetrical, particularly for AB (Fig. 1). From 64% (AB) to 71% (LR)
of eutrophication was due to crop and feed production. 10% (AB) to 27% (GAP) was due to
weaning to slaughtering and 6% (GAP) to 10% (AB) was due to piglet production because of
emissions of NH3 and N2O in buildings. For AB, 15% was due to compost production. Litter
production accounted for 1.3% (AB) and 2.8% (LR), building construction contributed 0.1%
(GAP, LR, AB) (Fig. 2).
Per kg of pig, climate change was 2.3 kg CO2-eq for GAP, 3.46 for LR and 3.97 for AB. Per
ha, climate change was larger for LR (5510 kg CO2-eq) than for AB (4022) and GAP (4236).
49
Both per kg and per ha uncertainty intervals were very large and asymmetrical for the three
scenarios (Fig. 1). From 54% (AB) to 73% (GAP) of climate change was due to crop and feed
production, 20% (GAP) to 35% (LR, AB) was due to weaning to slaughtering, 4% (LR, AB)
to 6% (GAP) to piglet production and for AB 5% was due to compost production. Litter pro-
duction accounted for 1.2% (AB) and 2.1% (LR) and building construction for 0.7% (AB) to
1.4% (GAP) (Fig. 2).
Acidification per kg of pig was 0.0226 kg SO2-eq for LR, 0.0372 for AB and 0.0435 for GAP.
Acidification per ha was very close for LR (36.0) and AB (37.7) and much larger (80.1) for
GAP. Both per kg and per ha, uncertainty intervals for acidification were smaller than for eu-
trophication and climate change, in particular when expressed per ha for LR and AB (Fig. 1).
Pig production (PP and WS) was the main contributor to acidification for GAP (71%) and LR
(64%), while for AB, compost production contributed more (40%) than pig production (35%).
Crop and feed production accounted for 24% (AB) to 34% (LR), litter production accounted
for 0.5% (AB) and 1.4% (LR), building construction accounted for 0.3% (GAP, AB) to 0.6%
(LR) (Fig. 2).
Per kg of pig, terrestrial toxicity was 0.0165 kg 1.4-dichlorobenzene-eq for GAP, 0.0184 for
LR and 0.0304 for AB. Per ha the three scenarios had similar impacts. Per kg, the uncertainty
intervals were relatively small, expressed per ha, the intervals were very small or non-existent
(not shown). Crop and feed production was the main source of terrestrial toxicity for LR
(81%) and AB (92%) and the only one for GAP. For LR, (outdoor) piglet production contrib-
uted 14% and 2.6% for AB. Litter production contributed 5% for LR and AB (Fig. 2).
Energy use per kg of pig was close for GAP (15.9 MJ) and for LR (17.9), for AB, energy use
was larger (22.2). Conversely, energy use per ha was close for GAP (29282 MJ) and for LR
(28503) but smaller for AB (22492). The uncertainty intervals expressed per kg were small,
expressed per ha they were very small (Fig. 1).
Crop and feed production was the main contributor to energy use, ranging from 74% for GAP
to 96% for AB. For GAP, PP plus WS contributed for 23% while no energy was used for
these stages for LR and AB. Building construction accounted for 2.0% (AB) to 2.7% (GAP)
of energy use. Straw litter production accounted for 1.3% (AB) to 1.9% (LR) (Fig. 2).
Land use per kg of pig was larger for AB (9.87 m2.year) than for GAP (5.43) and LR (6.28).
The uncertainty intervals were relatively small (Fig. 1). Crop and feed production was the
most important contributor to land use, ranging from 89% (LR) to 100% (GAP) (not shown).
Outdoor piglet production accounted for 6% (AB) and 8% (LR) of land use. Litter production
accounted for 3.1% (AB) and 3.4% (LR) of land use (not shown). With respect to the amount
of pig per ha, GAP was the most productive scenario (1840 kg/ha) followed by LR (1590) and
AB (1010) (Fig. 1).
50
Pesticide use per kg of pig was close for GAP (1.37 g) and LR (1.44) and much smaller for
AB (0.239 g). Per ha it was close for GAP (2.52 kg/ha) and LR (2.29) and much smaller for
AB (0.24 kg/ha). The uncertainty intervals expressed per kg were relatively small, expressed
per ha, the intervals were very small or non-existent. For GAP and AB, crop and feed produc-
tion was the sole contributor to pesticide use. For LR, litter production contributed 4% (not
shown).
Comparison of the three scenarios for impacts and yield
Current intensive pig production systems have a poor image with the general public because
they are associated with environmental pollution (in particular poor water quality) on a re-
gional scale (Petit and van der Werf, 2003). Given this perspective, we based our examination
of the three scenarios primarily on their impacts and productivity per ha. Expressed per ha,
LR had lower impacts than GAP for eutrophication and acidification, a higher impact for cli-
mate change, the other impacts were similar for GAP and LR; LR was less productive (- 14%)
(Fig. 1). Expressed per ha, AB had lower impacts than GAP for eutrophication, acidification,
energy use and pesticide use, for climate change and terrestrial toxicity impacts were similar;
pig production was 45% less for AB than for GAP (Fig. 1).
The lower impacts per ha of LR and AB relative to GAP for eutrophication and acidification
resulted from lower emissions of eutrophying and acidifying substances from the weaning to
slaughtering stage in straw-litter buildings (despite compost-related emissions for AB) and
from the crop and feed production stage (Fig. 2). Pesticide use per ha was much lower for AB
but not zero as the feed contained 10% of non-organic ingredients.
Hot-spots and margin for improvement
The GAP scenario did worse than LR and AB for eutrophication and acidification. NO3 (crop
and feed production) and NH3 (crop and feed production and pig production: PP and WS)
were the predominant substances contributing to these two regional impacts. As fertilisation
was optimised in the GAP scenario, the systematic introduction of catch crops (rather than
50%) seems the most promising measure to reduce nitrate emissions. Although the field ap-
plication of slurry was relatively optimised for GAP, ammonia losses can be further reduced
by adopting slurry injection technology. Decreasing dietary protein can lead to significant re-
duction of ammonia emission along the building, storage and application stages (Portejoie et
al., 2002). Similarly, the coverage of slurry stores can significantly reduce ammonia emis-
sions during storage and application (Portejoie et al., 2003). Pesticide use was another major
drawback of the GAP scenario.
The LR scenario contributed more to climate change than GAP and AB, due to the WS stage,
when pigs are in a straw litter building. We identified N2O, a very potent greenhouse gas, as
the predominant substance contributing to climate change for this stage. Litter-based housing
systems suffer from a low level of standardisation. So far, the type of litter (Robin et al.,
1999), litter management (Kermarrec, 1999) and animal density (Robin, 2002) have been
51
identified as influent parameters affecting gaseous emissions in litter systems. As for GAP, a
low-protein diet could help reduce emissions for LR. The LR scenario thus seems to have a
large margin of improvement. As for GAP, pesticide use is a drawback of LR.
Expressed per ha, AB had lower acidification and eutrophication impacts than GAP, ex-
pressed per kg, AB and GAP were similar for these impacts. For AB, 15% of eutrophication
and 40% of acidification were due to compost production, which thus constitutes a major dis-
advantage for AB. However, the emissions associated with compost production can be re-
duced by adapting the heap composition (Ekinci et al., 2000; Sommer and Møller, 2000) and
by optimising composting techniques such as reducing the turning frequency or covering the
heap (Robin et al., 2001). As for LR, the AB straw litter system presents a major contribution
to climate change. Like GAP and LR, AB seems to have an important margin of improve-
ment. Finally, in certain contexts, the low productivity per ha of AB might be a limitation to
its development.
Conclusion
Our comparison of three contrasting and relatively optimised scenarios of pig production pro-
duces useful knowledge for decision makers at different levels. For each scenario, hot-spots
as well as important margins of improvement have been identified. The use of favourable and
unfavourable scenarios allowed an estimation of the uncertainty of the results.
References
AGRESTE, 2001. Tableaux de l’Agriculture Bretonne 2001. Direction Régionale de
l’Agriculture et de la Forêt, Service Régional de Statistique Agricole, Rennes, France.
Berger, F., 2000. Les défis du porc biologique. Chambre régionale d’agriculture des Pays de
Loire. France.
BUWAL, 1996. Ökoinventare für Verpackungen. Schriftenreihe Umwelt Nr. 250/1+2, Bun-
desamt für Umwelt, Wald und Landschaft, Bern, Switzerland.
CEE, 1991. Règlement (CEE) n°2092/91 du conseil du 24 juin 1991 concernant le mode de
production biologique de produits agricoles et sa présentation sur les produits agricoles
et les denrées alimentaires (JO L 198 du 22.7.1991).
ECETOC, 1994. Ammonia emissions to air in western Europe. Technical report no. 62. Euro-
pean Chemical Industry Ecology & Toxicology Centre, Brussels, Belgium.
Ekinci, K., Keener, H.M., Elwell, D.L., 2000. Composting short paper fiber with broiler litter
and additives. Part 1: effects of initial pH and carbon/nitrogen ratio on ammonia emis-
sion. Compost Sciences and Utilization 82: 160 – 172.
FAO, 2002. FAOSTAT Agriculture Data. Available at : http://apps.fao.org/.
Groupement des fermiers d’Argoat, 2000. Cahier des charges éleveur – porcs fermiers élevés
en plein air. Homologation n°20-88. Saint-Brieuc, France.
Guinée, J.B., Gorrée, M., Heijungs, R., Huppes, G., Kleijn, R., de Koning, A., van Oers, L.,
Wegener Sleeswijk, A., Suh, S., Udo de Haes, H.A., de Bruijn, H., van Duin, R.,
52
Huijbregts, M.A.J., 2002. Life cycle assessment. An operational guide to the ISO stan-
dards. Centre of Environmental Science, Leiden University, Leiden, The Netherlands.
Houghton, J.T., Meira Filho, L.G., Callander, B.A., Harris, N., Kattenberg, A., Maskell, K.,
1996. Climate change 1995: the science of climate change. Cambridge University Press,
Cambridge, United Kingdom.
Huijbregts M.A.J., Norris, G., Bretz, R., Ciroth, A., Maurice, B., von Bahr, B., Weidema, B.,
de Beaufort A.S.H., 2001. Framework for Modelling Data Uncertainty in Life Cycle In-
ventories. Int. J. LCA 6 (3): 127 – 132.
IPCC, 1996. Intergovernmental Panel on Climate Change. Revised 1996 IPCC Guidelines for
National Greenhouse Gas Inventories : Reference Manual (3). Available at :
http://www.ipcc-nggip.iges.or.jp/public/gl/invs6.htm.
ITP, 2001. Résultats Gestion Technico-économique. ITP, Paris, France.
Kanyarushoki, C., 2001. Evaluation de l’impact environnemental des bâtiments porcins – Ap-
plication de la méthode d’analyse du cycle de vie. Mémoire de fin d’études DESS. Uni-
versité Rennes 1, Rennes, France.
Kermarrec, C., 1999. Bilan et transformations de l’azote en élevage intensif de porcs sur litiè-
re. PhD thesis. Université de Rennes I, Rennes, France.
Ministère de l’Agriculture et de la Pêche, 2000. Cahier des charges concernant le mode de
prodution et de préparation biologique des animaux et des produits animaux définissant
les modalités d’application du règlement CEE n° 2092/91 modifié du conseil et/ou
complétant les dispositions du règlement CEE n° 2092/91. Homologué par l’arrêté in-
terministériel du 28 août 2000. J.O.R.F. du 30 août 2000. Direction des politiques éco-
nomique et internationale, Paris, France.
Morvan, Th., Leterme, Ph., 2001.Vers une prévision opérationnelle des flux d’azote résultant
de l’épandage de lisier : paramétrage d’un modèle dynamique de Simulation des trans-
formations de l’azote des lisiers (STAL). Ingénieries 26: 17 – 26.
Nemecek, T., Heil, A., 2001. SALCA – Swiss Agricultural Life Cycle Assessment Database.
Version 012, December 2001. FAL, Swiss Federal Research Station for Agroecology
and Agriculture, Zuerich, Switserland.
Petit, J., van der Werf, H M G, 2003. Perception of the environmental impacts of current and
alternative modes of pig production by stakeholder groups. Journal of Environmental
Management 68: 377-386.
Portejoie, S., Dourmad, J-Y., Martinez, J., Lebreton, Y., 2002. Effet de la réduction du taux
protéique de l’aliment sur la volatilisation ammoniacale des effluents porcins. Journées
Rech. Porcine en France 34: 167-174.
Portejoie, S., Martinez, J., Guiziou, F., Coste, C.M., 2003. Effect of covering pig slurry stores
on the ammonia emission processes. Bioresource Technology 87: 199-207.
PRé Consultants, 1997. SimaPro 2 method. In: Database Manual. Pré Consultants B.V., Am-
ersfoort, The Netherlands.
Robin, P., 2002. Maîtrise des émissions gazeuses en bâtiments sur litière (validation en éle-
vages des résultats acquis en station expérimentale). Rapport final première année
(2002). Convention DERF-INRA n°A 01592-2001-2002. Rennes, France.
53
Robin, P., De Oliveira, P.A., Kermarrec, C., 1999. Productions d’ammoniac, de protoxyde
d’azote et d’eau par différentes litières de porcs durant la phase de croissance. Journées
Rech. Porcine en France 31: 111-115.
Robin, P., Aubert, C., Bline, D., Tricot, G., Vacher, V., 2001. Maîtrise du compostage de fu-
mier de volailles à la ferme : mélange initial, retournements, couvertures. Sciences et
Techniques Avicoles 35: 9-12.
Rosenberg, P.E., Gallot, J., 2002. Référentiel de l’agriculture raisonnée. Arrêté du 30 avril
2002 relatif au référentiel de l’agriculture raisonnée. Journal Officiel de la République
Française 104 (4 mai 2002): 8519.
Sanders, J., 2000. L’evaluation des systèmes de production animale: l’Analyse de Cycle de
Vie peut-elle servir à construire des indicateurs environnementaux pour l’alimentation
porcine ? INRA, UMR SAS, Rennes, France.
Sommer, S.G., Møller, H.B., 2000. Emission of greenhouse gases during composting of deep
litter from pig production – effect of straw content. Journal of Agricultural Science,
Cambridge 134: 327-335.
UNECE, 1999. Control options/techniques for preventing and abating emissions of reduced
nitrogen compounds. United Nations Economic and Social Council, Economic Commis-
sion for Europe.
Van der Werf, H.M.G., Petit, J., 2002. Evaluation of the environmental impact of agriculture
at the farm level: a comparison of twelve indicator-based methods. Agriculture, Ecosys-
tems and Environment 93 (1): 131 – 145.
Van der Werf, H.M.G., Petit, J., Sanders, J. The environmental impacts of the production of
concentrated feed : the case of pig feed in Bretagne. Agricultural Systems, in press.
54
LCA of Animal Products from Different Housing Systems in Switzerland:
Relevance of Feedstuffs, Infrastructure and Energy Use
Erzinger, S.*, Dux, D., Zimmermann, A. and Badertscher Fawaz, R.
*Agroscope FAT Taenikon, Swiss Federal Research Station for Agricultural Economics and Engineer-
ing, CH-8356 Ettenhausen, Switzerland. Phone +41 52 3683239, fax +41 52 3651190,
[email protected]
Abstract
For many foodstuffs, especially animal products, agricultural production is an environmentally
relevant production step. Milk and meat are produced in many different housing systems.
The differences in the environmental impacts of these production systems are as yet little
known. The goal of our study was to assess the environmental impacts of conventional and
animal-friendly housing systems in Switzerland. We compared case studies of two housing
systems for dairy cows and for fattening pigs. The most important factor proved to be the
feeding regime. The building infrastructure was also relevant, especially for energy consump-
tion and human toxicity. Improvements in the choice and production of feedstuffs should be
introduced. Possible measures are low-emission fertilisation, efficient mechanisation, short
transport distances and little drying.
Keywords: animal farming, housing system, life cycle assessment, milk, pork
Introduction
Animal-friendly housing is finding increasing practical application in Switzerland, partly as a
result of animal welfare legislation, partly within the framework of direct payments for espe-
cially animal-friendly husbandry or in the context of various label schemes. Not much infor-
mation is available on the environmental impacts of different animal housing systems, data
gaps being particularly common in relation to new types of housing. The goal of the study
was to compare the environmental impacts of various animal housing systems for dairy cows
and fattening pigs using life cycle assessments.
Methods and systems investigated
Each of the following systems were compared for different herd sizes and feed variants:
• Milk from cows in "cubicle housing" and "tied housing" and
• Fattening pigs from "multi-surface systems with littered lying area and exercise yard"
and from "pens with fully slatted floors".
Special emphasis was placed on an accurate study of the buildings. This includes construction
(building materials, construction processes), operation, maintenance and dismantling of the
buildings. Service life of the buildings is 50 years. Building parts and equipment with a
shorter life time are replaced during this 50 years (Nemecek et al., 2003)
55
The comparison of these different animal housing systems was carried out using production
processes which were determined on the basis of empirical values and expert know-how. Ad-
ditionally, the pork case study is based on a survey of around 90 fattening pig farms. All the
influencing variables unconnected with housing type (e.g. mechanisation) were identically de-
fined where possible. However there were also differences between the processes resulting
not from housing type but from labour organisation and farm management. The processes
were not defined and optimised with reference to ecological criteria, but illustrated conditions
normally found on the farm. The study should therefore be regarded as a case study.
The farm gate was the system boundary for assessment. All preliminary stages and infrastruc-
ture (buildings, machinery) were included. The processing and distribution of the milk and
fattened pigs did not form part of the study.
Direct emissions of ammonia, methane and nitrous oxide from animal metabolism, housing
areas and slurry storage were difficult to assess. No reliable data on emissions were available,
particularly for the animal-friendly systems. Four scenarios were therefore defined for dairy
cattle and for fattening pigs (Tab. 1). The scenarios “Standard1” and Standard2” base on cur-
rent emission factors published in Switzerland and Germany respectively. The scenario
“High” uses very high emission factors, whereas the scenario “Low” uses very low factors.
These two scenarios do not distinguish between housing systems. The intention of the scenar-
ios “High” and “Low” is to get a range for the relevance of direct emissions from the hous-
ings compared to the total emissions from the entire systems.
Table 1. Scenarios for direct emissions from animal housing (incl. slurry pit) for dairy cows
and fattening pigs [kg gas/(place year)]. TH = tied housing, CH = cubicle housing; FSF =
fully slatted floors, MSS = multi-surface system.
Dairy cows
NH3 CH4 N2O
Scenario TH CH TH CH TH CH
Standard1 1 11.2 22.9 186 183 0.5 1.0
Standard2 2 6.3 16.7 126 126 0.7 1.3
High 3 41.5 42.1 240 240 1.4 1.8
Low 3 3.5 4.1 50 50 0.4 0.8
Fattening pigs
NH3 CH4 N2O
Scenario FSF MSS FSF MSS FSF MSS
Standard1 1 2.3 2.3 4.7 5.1 0.05 0.05
Standard2 2 3.8 5.0 4.0 2.5 0.12 0.12
High 3 7.1 7.1 7.0 7.0 0.52 0.52
Low 3 1.1 1.1 0.5 0.5 0.02 0.02
1: Ammonia (NH3): Menzi et al. (1997); Methane (CH4): Minonzio et al. (1998); Nitrous oxide (N2O): Schmid et al. (2000);
all in Nemecek (2002)
2: UBA (2002), adapted
3: own assumptions based on analysis of the literature
56
Twelve environmental impacts were calculated. The impacts were assessed according to
Rossier and Gaillard (2001) and Nemecek et al. (2003a). Neither noise, odour nor impact on
biodiversity, soil fertility or landscape were studied. Life cycle assessment calculations were
carried out using TEAM software (Version 3.0) and the SALCA life cycle inventory database
of the Swiss Agricultural Research Stations (Nemecek, 2002). Only three of the twelve envi-
ronmental impacts investigated will be discussed here, namely energy consumption, eutrophi-
cation and ecotoxicity. These indicators have proved particularly meaningful in previous stud-
ies (Rossier and Gaillard, 2001).
Milk case study
Milk is the most important raw product of Swiss farming and is frequently produced in inten-
sive grassland-based systems. Production is organised on a small scale, many farms owning
around 20 cows. Although tied housing systems are widespread (approx. 80%), new buildings
are almost exclusively loose housing. The systems studied are described in Table 2.
Table 2. Description of the types of dairy cattle housing studied.
Tied housing Cubicle housing
Herd size 20 and 40 dairy cattle places
Functional unit 1 litre cooled milk from farm tank
By-products [kg live
• Starter calves
weight/(cow y)]
40.5 44.3
• rearing calves 58.8 54.3
• culled cattle 205.0 190.3
Milk yield 1 [kg 6987 7103
ECM/(cow y)]
Feed without silage 2 hay, grass, fodder beet, maize cubes, concentrate, pasture
with silage grass silage, maize silage, grass, hay, concentrate, pasture
Pasture 3 60 half days pasture 198 half days pasture
• Ventilation
Building with exercise yard
gravity flues free ventilation
• Milking pipeline milking system herringbone milking parlour
1: Difference due to higher fertility and longer useful life in loose housing (Badertscher, 2003)
2: Prerequisite for the production of cheese made from raw milk
3: Tied housing: minimum animal welfare legislation; cubicle housing: full growing season
Environmental impact was split between milk and by-products (Tab. 2) using a financial allo-
cation. Under this method, the milk accounted for 88% of the environmental impact.
Results
Table 3 and Figure 1 show the results of the housing systems, herd sizes and feed variants
studied.
57
Table 3. Environmental impacts of milk from different dairy cow housing systems and feed
variants (per kg cooled milk at farm gate). TH = tied housing, CH = cubicle housing.
Herd size 20 cow places 40 cow places
Feed without silage with silage without silage with silage
Housing type TH CH TH CH TH CH TH CH
Energy [MJeq/kg] 6.0 5.8 4.6 4.4 5.7 5.4 4.4 4.1
Eutrophication [gPO4eq/kg] 4.2 4.5 3.7 4.1 4.2 4.5 3.7 4.1
Ecotoxicity [gZneq/kg] 0.21 0.21 0.21 0.20 0.20 0.20 0.21 0.20
The environmental impact of milk production is determined mainly by the feedstuffs (Fig. 1),
which contribute between 50 and 85% to the three impacts analysed. This includes the agri-
cultural production, processing (e.g. hay aeration) and transports. The buildings (construction,
operation and dismantling) play a significant role in energy consumption with a share of ap-
proximately 25%.
120%
Direct Emissions
(Standard1 scenario)
100%
Various
80%
Feedstuff (for cows)
60%
Heifers (incl. buildings
40% and feedstuff)
Buildings (for cows)
20%
TH = Tied housing
0% CH = Cubicle housing
TH- CH- TH- CH- TH- CH- TH- CH- TH- CH- TH- CH- ns = non silage feeding
ns ns s s ns ns s s ns ns s s s = silage feeding
Energy consumption Eutrophication Ecotoxicity TH-ns = 100%
Figure 1. Comparison of the environmental impact of milk from tied housing and cubicle
housing with 40 cow places each.
The variants with silage feed needed 24% less energy than those without silage. This differ-
ence is above all due to hay ventilation and drying of maize cubes in the non-silage feeding
variant. Average energy consumption was 5 MJeq per kilo of milk. Compared with the me-
tabolisable energy of 2.7 MJ/kg milk, primary energy input was thus almost twice as high as
energy output for human nutrition.
Maize cubes (chopped, dried and pressed maize for non-silage feeding) and concentrate (cere-
als) were the feedstuffs with a high contribution to the three environmental impacts analysed.
Hay aeration used around 10% of total energy consumption in the variants without silage
58
feed. The impact of silage, grass, pasture and fodder beet was low. The environmental impact
of feedstuffs was determined mainly by agricultural production and drying. Transport and
processing were of less importance.
Herd size barely affected environmental impact of the entire system as well as of the build-
ings used. Both herd sizes studied showed a similar impact per cow place. The housing type
only affected direct emissions, other environmental impacts were similar. Direct emissions
from housing and slurry pit were caused mainly by ammonia emissions, accounting for be-
tween 3 and 40% of total eutrophication, depending on the emission scenario (Tab. 4). In the
standard scenarios, the direct emissions from housing were around twice as high for cubicle
housing as for tied housing.
Table 4. Influence of scenarios for direct emissions on „eutrophication“ environmental im-
pact for the „40 places without silage feed“ variant. TH = tied housing, CH = cubicle housing.
For definition of scenarios see Tab. 1.
Scenario Standard1 Standard2 High Low
Housing type TH CH TH CH TH CH TH CH
Eutrophication [gPO4eq/kg milk] 4.2 4.7 4.0 4.4 6.1 6.0 3.8 3.8
- of which direct emission from housing 0.5 1.0 0.3 0.8 2.4 2.3 0.1 0.1
Share direct emission from housing [%] 12 22 8 17 40 39 3 3
Share other system [%] 88 78 92 83 60 61 97 97
Discussion
In particular the type of feedstuff (including cultivation, conservation, processing and trans-
ports) determines the potential environmental impact of milk production. The way to reduce
environmental impact is to optimise roughage production (Nemecek and Huguenin, 2002).
Possible measures are:
• extensive roughage production with low fertiliser use
• more pasture, less grass harvesting
• silage or field dried hay instead of hay aeration and maize drying. Use of renewable
energy carriers in hay aeration.
There was hardly any difference in the environmental impact of the building infrastructure
and usage in the two dairy housing systems studied. The environmental soundness of dairy
housing may be optimised by ensuring minimum possible energy consumption during milk
production and cooling. Direct emissions from the cowsheds could be expected to differ be-
tween the systems, but these cannot be quantified using the data available. Research is needed
here.
Other studies often take no account of the infrastructure, and the basic data and calculation
methods are seldom accurately described. Comparisons are therefore difficult. Compared with
59
the mean values of 35 dairy farms from Rossier and Gaillard (2001), the energy and eutrophi-
cation values calculated here are roughly one third lower, but twice as high for ecotoxicity –
although they fall within the variability noted there. The figures for primary energy consump-
tion are higher than in Cederberg (1998), Haas et al. (2001) or Hogaas Eide (2002).
Pork case study
Two herd sizes and feed variants were studied for pork (Tab. 5). Pig fattening gives rise to no
by-products, so no allocation was required. The total environmental impacts were down to the
animals produced.
Table 5. Description of the fattening pig housing systems studied.
Fully slatted floors Multi-surface system
Herd size 300 and 1000 fattening pig places
Functional unit 1 kg pig (live weight at farm gate)
• Daily gain [kg/day] 0.748 0.757
• Days to reach slaughter 111 107
weight
• feed conversion [MJ DE 36 39
per kg live weight gain]
Feed Complete diet Complete diet
Whey Whey, energy and protein sup- –
plementary feed, fat
Building No exercise yard With exercise yard and littered ly-
ing area.
• Ventilation Forced ventilation Free ventilation
• Heating Electric radiators No heating
Results
The results of the housing systems, herd sizes and feed variants studied are shown in Table 6
and Figure 2.
Table 6. Environmental impact of different housing systems and feeding variants for fatten-
ing pigs (per kg live weight at farm gate). FSF = fully slatted floors, MSS = multi-surface sys-
tem.
Herd size 300 pig places 1000 pig places
Feed Whey Complete diet Whey Complete diet
Housing system FSF FSF MSS FSF FSF MSS
Energy [MJeq/kg] 30.7 33.3 27.8 27.7 30.2 27.6
Eutrophication [gPO4eq/kg] 28.8 36.2 37.5 29.9 37.3 38.5
Ecotoxicity [gZneq/kg] 0.80 0.75 0.78 0.81 0.76 0.81
60
Feedstuff provision is by far the greatest factor affecting the environmental impact of pig fat-
tening (Fig. 2). Depending on the type of environmental impact, the feedstuffs used in fatten-
ing and piglet production account for up to 90% of the environmental impact.
140%
Direct Emissions
(Standard1 scenario)
120%
Various
100% Feedstuffs (for
fattening pigs)
80% Piglets (incl. buildings
and feedstuffs)
60% Buildings (for fattening
pigs)
40% FSF = Fully slatted floor
MSS = Multi surface
20% system
w = whey
0% cf = complete feed
FSF- FSF- MSS- FSF- FSF- MSS- FSF- FSF- MSS-
w cf cf w cf cf w cf cf FSF-w = 100%
Energy consumption Eutrophication Ecotoxicity
Figure 2. Comparison of the environmental impact of pork from fully slatted and multi-
surface housing with two feed variants and herd size of 300 fattening pigs.
The buildings have a relevant influence on energy consumption, in housing with fully slatted
floors the buildings account for almost 30% of energy consumption.
Direct emissions from the fattening pigs housing contributed roughly 5% to eutrophication in
the scenario “Low” and around 35% in the scenario “High”.
Herd size had a limited effect on energy consumption. Although the building-related energy
consumption per kilogram of meat produced did fall as herd size increased, the energy con-
sumption for the supply of feedstuffs was unchanged. Herd size also barely affected any of
the other environmental impacts studied.
Discussion
The feedstuffs used have a crucial effect on the potential environmental impact of pork pro-
duction. Good feed utilisation and the use of feedstuffs produced by environmentally sound
methods are therefore of primary importance in reducing environmental pollution. The key
factors here are:
• Agricultural production using low-emission fertiliser as well as efficient mechani-
sation,
• Little transportation and drying and
61
• Use of by-products from milk processing, milling, sugar and oil production or other
industrial processes – provided that these products are not contaminated with pollut-
ants or competing with other fields of application.
Conclusions
The results shown here provide pointers to the environmental impact of different housing sys-
tems for fattening pigs and dairy cattle.
The differences found were caused not only by the actual housing system, but in particular by
differences in the feeding regime. The present results should therefore always be interpreted
as the combined result of feeding regime and housing system.
Feedstuff provision was central – both to environmental impact and production cost (Bad-
ertscher, 2003).
Infrastructure (buildings and installations) had a significant effect on the overall environ-
mental impact of animal production. It should therefore definitely be taken into account by
LCAs. Herd size, on the other hand, had little effect on the results. There are large data gaps
in the assessment of direct emissions from housing, an influence on eutrophication.
LCAs are well suited to analysing the environmental impact of entire animal production sys-
tems, assuming the availability of high quality LCI databases. As shown above, the most im-
portant factor in optimising animal production is to provide environmentally sound feedstuffs.
Here we should also be dealing with biodiversity and landscape issues beyond the LCA cur-
rently used.
Acknowledgment
The project was funded by the Swiss Federal Veterinary Office. Issues of economic viability,
workload, animal health and product quality were studied in addition to the environmental life
cycle assessment discussed here (Badertscher, 2003).
Literature
Badertscher, R., 2003. Tierwohl und Wirtschaftlichkeit: Ein Widerspruch? Paper to be pre-
sented at the Gewisola Tagung Hohenheim 2003 Hohenheim (Germany), Sept 29 - Oct
01, 2003.
Cederberg, C., 1998. Life Cycle Assessment of Milk Production. A Comparison of Conven-
tional and Organic Farming. SIK-Document 143. Swedish Institute for Food and Bio-
technology, Gothenburg.
Erzinger, S. and Badertscher Fawaz, R., 2001. Life Cycle Assessment of animal housing sys-
tems as part of an overall assessment. Proceedings International Conference on LCA in
Foods, p. 118-122. SIK-Report Nr. 643. Swedish Institute for Food and Biotechnology,
Gothenburg.
62
Haas, G., Wetterich, F. and Köpke, U., 2001. Comparing intensive, extensified and organic
grassland farming in southern Germany by process life cycle assessment. Agriculture,
ecosystems and environment 83, 43-53.
Hogaas Eide, M., 2002. Life cycle assessment (LCA) of industrial milk production. Int J LCA
7 (2), 115-126.
Menzi, H., Frick, R. and Kaufmann, R., 1997. Ammoniak-Emissionen in der Schweiz: Aus-
mass und technische Beurteilung des Reduktionspotentials. Schriftenreihe der FAL 26.
Swiss Federal Research Station for Agroecology and Agriculture, Zurich-Reckenholz.
Minonzio, G., Grub, A. and Fuhrer, J., 1998. Methan-Emissionen der schweizerischen Land-
wirtschaft. Schriftenreihe Umwelt Nr. 298. Swiss Agency for the Environment, Forests
and Landscape (SAEFL), Berne.
Nemecek, T., Heil, A., Huguenin, O., Meier, S., Erzinger, S., Blaser, S., Dux, D. and Zim-
mermann, A., 2003. Life Cycle Inventories of Agricultural Production Systems. Final
report ecoinvent 2000 No. 15. FAL Reckenholz, FAT Tänikon, Swiss Centre for Life
Cycle Inventories, Dübendorf, CH, retrieved from: www.ecoinvent.ch.
Nemecek, T., Heil, A., Erzinger, S. and Zimmermann, A., 2003a. SALCA Swiss Agricultural
Life Cycle Assessment Database. Version 031. Swiss Federal Research Station for
Agroecology and Agriculture, Zurich-Reckenholz.
Nemecek, T., 2002. SALCA-Templates. Swiss Agricultural Life Cycle Assessment Database.
Beschreibung der Mustersysteme „Betrieb“ und „Kultur“, Version 1.20. Swiss Federal
Research Station for Agroecology and Agriculture, Zurich-Reckenholz.
Nemecek, T. and Huguenin, O., 2002. Beurteilung von Graslandsystemen. Schriftenreihe der
FAL 38, 16-18. Swiss Federal Research Station for Agroecology and Agriculture, Zu-
rich-Reckenholz.
Rossier, D. and Gaillard, G., 2001. Bilan écologique de l'exploitation agricole. Méthode et
application à 50 entreprises. Swiss Federal Research Station for Agroecology and Agri-
culture, Zurich-Reckenholz (available from [email protected]).
Schmid, M., Neftel, A. and Fuhrer, J., 2000. Lachgasemissionen aus der Schweizer Landwirt-
schaft. Schriftenreihe der FAL 33. Swiss Federal Research Station for Agroecology and
Agriculture, Zurich-Reckenholz.
UBA (eds.), 2002. BMVEL/UBA-Ammoniak-Emissionsinventar der deutschen Landwirt-
schaft und Minderungsszenarien bis zum Jahre 2010. German Federal Environmental
Agency, Berlin.
63
Qualitative Life Cycle Assessment of Thai Shrimp Product
Mungkung, R. and Clift, R.
Corresponding address: Centre for Environmental Strategy (CES), School of Engineering, University
of Surrey, Guildford, Surrey GU2 7XH, UK Phone: +44 (0)1483 689559 Fax: +44 (0)1483 686671 E-
mail: [email protected]
Abstract
Economic and social importance of shrimp farming is evident in terms of export values, income
generation, and employment opportunity. However, the continual rise of shrimp consumption and
proliferation of shrimp farms has become controversial because of their associated potential envi-
ronmental impacts. The impacts of shrimp farming have been highlighted and their sustainability
has been under criticism. Much attention has been paid to Thailand as the world’s leading pro-
ducer of farmed shrimps. In view of the contentious issues surrounding shrimp aquaculture, quali-
tative Life Cycle Assessment (LCA) has been applied to evaluate the principal environmental in-
terventions over the life cycle shrimp production chain. Key issues have been identified and
potential impacts from different farming systems were assessed and compared based on the
qualitative LCA results and knowledge of systems. However, practical approaches permitting the
quantification of environmental impacts associated with shrimp aquaculture such as depletion of
wild broodstock, depletion of marine fish stock, loss of biodiversity, impacts on land use and im-
pacts of chemical use need to be further investigated.
Keywords: shrimp aquaculture; Life Cycle Assessment (LCA); Thailand
Introduction
Thailand has been a major world shrimp producer since the 1990s. Export of farm-raised
black tiger prawn (Penaeus monodon) products has benefited the whole country. Shrimp pro-
duction has also generated substantial social welfare, particularly to people living in coastal
zones. In addition, upstream and downstream industries associated with shrimp production
have generated numerous job opportunities particularly for women. However, the shrimp in-
dustry in Thailand has been controversial because of its associated potential environmental
impacts. Several innovative culturing technologies has been developed and applied to mini-
mise the negative environmental consequences especially at the farm stage which is believed
to be the most important stage driving the whole industry. However, the wider environmental
sustainability implications of these culturing practices have not been thoroughly demon-
strated. In this study, qualitative Life Cycle Assessment (LCA) is applied to identify and as-
sess the environmental interventions at different stages in the shrimp production chain to give
a comprehensive view of their environmental consequences.
Shrimp Aquaculture in Thailand
Shrimp farming in Thailand originally developed as a by-product from the salt fields along
the seacoast of the Inner Gulf of Thailand. The conversion of salt fields into shrimp ponds
was fortuitous as the salt fields were close to the sea, thus offering ideal conditions to obtain
the wild shrimp larvae and facilitating shrimp farming. Several factors have also promoted the
64
rapid growth of shrimp farming: the regional decrease of shrimp from wild capture, increasing
demand for farmed shrimp, and a massive collapse of shrimp farming due to disease and pol-
lution in Taiwan. Subsequently, the growth of the shrimp market has been the main driving
force promoting rapid expansion of the shrimp farming in Thailand. At the beginning, shrimp
ponds were constructed close to, or within, the mangrove area to obtain wild juvenile shrimps
and to facilitate pumping of seawater into and out of the farms. However, the unfavourable
conditions of acidic water due to high pyrite content and pond bottoms made uneven by roots
led to the abandonment of numerous farms. The failure of shrimp ponds in mangrove areas
initiated shifts in development of new ponds to more suitable and productive sites in coastal
areas. This included the development of inland shrimp farming at low salinity in the Central
Plain (Suphanburi, Nakhon Pathom, Ayuthaya, Angthong, Prachinburi, Nakhon Nayok and
Chacheongsao). The emergence of inland shrimp farming in the most intensive and produc-
tive area of rice cultivation produced a heated debate over the environmental, social and eco-
nomic implications of the activity. Nevertheless, it has finally been banned due to its potential
impacts.
The main hubs of shrimp production presently are in the eastern and southern parts of the
country. However, with uncontrolled shrimp pond expansion and inappropriate culture prac-
tices, shrimp farming in Thailand has been through ‘boom and burst’ cycles with very high
production followed by a sudden collapse caused by outbreak of disease associated with poor
water quality management (Nissapa and Boromthanarat, 2002). This has raised concern for
the sustainability of shrimp aquaculture.
In addition to the concern over unsustainable farming practices, the social and economic sus-
tainabilities of shrimp farming industry have also received much attention. As shrimp farming
utilises common property resources, this tends to affect social equity distribution leading to
competition for access to resources. From an economic perspective, even though shrimp farm-
ing is considered as a ‘quick-in-return’ investment, the financial performance of shrimp farm-
ing is unpredictable. Shrimp farmers have to invest a considerable amount of money for crop
production due to the high capital investment of the shrimp ponds as well as the high cost of
input factors. Moreover, production is subjected to climate variability, the possibility of dis-
ease outbreak and other adverse factors resulting in crop failure. The profitability of the crop
is intimately associated with current shrimp price which are subject to the supply and demand
of the global market. In light of these factors and the limited control that a farmer can have
over them, shrimp farming, to a certain extent, is an inherently risky investment.
In view of the shrimp industry outlook, the intensive farming system is still a common prac-
tice. However, the farm expansion has been limited because of the less land availability and
the previous experiences of boom-and-burst consequences. Fishmeal shortage has also be-
come more serious and this is linked to the production of trash fish that are not being sustain-
ably managed. Much is still unknown regarding disease prevention and control. Vaccination
against diseases has not been successfully developed, and attempts to prevent or eradicate dis-
65
eases have failed until now. Besides, Thailand has been faced with the export problems due to
the increased tariff rate by the European Union (EU) to revoke the Generalized System of
Preferences (GSP) privileges since 1999. This has caused a decrease in Thai exports to the EU
(Sitthipongpanich, 2001).
Non-tariff barriers, such as food safety requirements and certification and eco-labelling
schemes, have also been proposed. The shrimp market is still uncertain, and the price fluctu-
ates depending on the global demand.
Research Methodology
The study has focused on the shrimp aquaculture production in Thailand. Following a cradle-
to-grave approach, the life cycle stage of shrimp production includes trawling (capture of wild
broodstock), hatchery (post-larvae rearing), farming, processing, storage, transportation to
importing countries, consumption and waste management. Qualitative LCA methodology has
been used as the analytical tool to identify and assess the environmental impacts associated
with shrimp production. Due mainly to concerns over the environmental sustainability at the
farm level, five farming types with diverse culturing techniques and farming management
systems were selected from the shrimp farms in the Eastern region. These five case studies,
which have been analysed, are representative of current shrimp farming practices in Thailand.
Features specific to the individual farming system are respectively: “biological and Code of
Conduct (CoC)” farming applying the CoC (Code of Conduct for Responsible Marine Aqua-
culture, the environmental management programme of shrimp farming developed by Depart-
ment of Fisheries, Thailand) as well as biological management strategies such as using man-
grove to filter organic matter to minimise water pollution and minimising of chemical usage;
“conventional and CoC” farming applying the concept of environmental management system
to minimise the impacts and only approved drugs can be used when necessary; “probiotic”
farming utilising probiotic microbes to digest waste in order to maintain the water quality dur-
ing culturing; “ecological” farming aspiring to raise shrimps naturally by optimizing input
factors in order to sustain the pond productivity with no use of chemicals; and “ongoing-to-be
organic” farming operating at lower stock density (not more than 312,500 of post-larvae per
hectare) with best available organic inputs and chemical usage is completely eliminated.
Environmental interventions in every stage along the block-frozen shrimp production chains
were qualitatively identified. Site visits to hatcheries, farms, and processing plants were con-
ducted to understand fully the nature of their activities. Hatchery operators and shrimp farm
managers were interviewed to obtain more information on hatchery operations and farming
managements. Processing lines at the shrimp processing factory were “walked through” to en-
sure familiarity with the processing method used. Information regarding the transportation of
product to destinations by refrigerated container ship was obtained by interviewing the ship-
ping company.
66
Life Cycle Stages of Shrimp Production
Shrimp culturing starts with the capture of wild broodstock so that the hatchery cycle can be
established for producing post-larvae. At the hatchery, the broodstock is subsequently induced
to spawn through photoperiod manipulation. After hatching, a combination of phytoplankton,
artemia cyst and artificial diet is fed until the hatchlings transform to post-larvae stage 15,
where they are transferred to culturing ponds. At the farm, the ponds are constructed with a
suitable depth and bank slope and their bottoms are compacted to provide an appropriate habi-
tat for shrimps. Water is introduced into the pond and the water quality is prepared into an op-
timal regime for shrimps. After that, the stocking is carried out with a prior acclimatization of
post-larvae. High-protein artificial feed is used for growing shrimps. During culturing, water
exchange is carried out in order to dilute the waste accumulated in the pond. Aeration systems
are used to maintain a sufficient oxygen level in the ponds and to allow organic decomposi-
tion to take place. The crop production takes about 120 days or more. At harvesting, the water
is drained out of the pond through a net. Shrimps are collected and immediately put into icy
waters. Thereafter the sizing is performed and they are subsequently transported by refriger-
ated truck to Central Auction Shrimp Market, where the shrimps are further distributed to
processing plants. The shrimps are processed into block-frozen shrimps and transported to
overseas by refrigerated container ship. At post-harvest, accumulated sludge is taken out from
the pond bottom for further treatment and disposal. The soil surface is turned over and the
ponds are allowed to dry out before a new crop is introduced.
Results and Discussions
The principal life cycle stages of shrimp production are shown in Figure 1. Through the quali-
tative LCA analysis, the environmental impacts associated with shrimp aquaculture produc-
tion chains have been identified, as follows. Trawling for wild broodstock has a potential to
cause a stress on marine ecosystems such as destruction of seabed, changes of benthic com-
munities, and loss of biodiversity. Introduction of the wild broodstock and post- larvae can
also be a source of pathogens or diseases. At the hatchery, artemia cyst and artificial feed are
the most crucial inputs in addition to the introduction of broodstock; of these the utilization of
resources for producing feed and use of energy for harvesting and processing artemia cysts
are likely to have the most significant impacts. Intensive farming system required higher
stocking rate and this might lead to the deletion of wild shrimp populations.
At the farm site, the change of land use as a result of the construction of shrimp ponds could
affect the soil quality and biodiversity. Land requirement of shrimp ponds may come into
competition with other possible land uses when the scale of activities increases. Pond man-
agement, during and after the culturing period, highly influences the degree of water and soil
quality deterioration. The higher demand of trash fish to supply fishmeal as well as fish oils
required for formulated shrimp feed production may possibly increase pressure on the capture
fisheries leading to the depletion of wild catch as well as marine fish stock. The over-
harvesting of juvenile fish and non-target species might affect marine food chains and conse-
quently marine biodiversity and ecosystems. Uneaten feed as a result of poor feeding man-
67
agement is the main source of nutrients dissolved in wastewater and sludge accumulated in
shrimp ponds. However, using microbes to digest protein wastes together with sufficient aera-
tion is likely to reduce air emissions such as hydrogen sulphide and ammonia which are
highly odorous and detrimental to shrimps’ health.
Another aspect at the farm stage is the ecological toxicity of chemicals and antibiotics used.
Excessive use of chemicals and antibiotics possibly create the problem of antibiotic resistance
development and residue accumulation, and toxic effects on non-target species in the ponds
and surrounding ecosystems may change the ecological structure. Accumulation of the
chemical used for improving soil and water quality as well as for treating diseases during the
culture is a potential problem. Nevertheless, the human activities, such as removing sediment
from the bottom pond, adding fertilizer or lime to improving the soil quality, and drying the
pond between crops, influence the transport and fate of pollutant chemicals. At post-harvest,
improper discharge of wastewater and sludge could affect the water supply quality apart from
the external environmental quality. Moreover, some diseases could directly be triggered or
spread more effectively by poor environmental conditions. Human toxicity associated with
antibiotic contaminated shrimps is probably not an issue as potentially hazard antibiotics and
chemicals have been banned, and thereby eliminated from shrimp farming in Thailand. Water
and electricity consumption are the key issues at a processing plant. Air emissions from the
fuel used by shipping vessel as well as the energy consumed by refrigerated containers are the
most significant environmental impact from the transportation stage. Food waste generated
i.e. shrimp shells should be disposed properly.
Focusing on the farm stage, the environmental impacts of different shrimp farming systems
were compared based on the results from the qualitative LCA and the knowledge of farming
systems. Conventional and CoC farming types used a considerable amount of seawater due to
the higher rate of water exchange during the culturing period. The energy used for aeration
systems in this farming type was also fairly high. The wastewater treatment using mangrove
wetland systems at Biological and CoC farm seemed to be an effective mean to reduce the or-
ganic loads of discharge. The probiotic manufacturer has claimed that the probiotic products
have been proved that they are environmentally safe and sound. However, the mode of action
of probitoic substances associated with their remediation effects was not clear. Inputs of pro-
biotic farming were not significantly different from conventional intensive farming, except
the probitoic substances. Ecological farming consumed fewer resources because of no water
exchange during the culturing period and the under-rate feeding management. Organic farm-
ing used less energy due to the lower stocking rate. Considering the key factors affecting the
environmental performance of shrimp farms, two main issues identified are farm location and
choice of culturing practice. The comparison of the impacts from different shrimp farming
systems is not straightforward if the farms are located in the different geographical areas.
68
Wild broodstock Hatchery Wastewater
Seawater Shrimp post larvae
Land
Feed
Chemicals Transportation
Electricity
Land Landoccupation/transformation impacts
Pond construction
Shrimp pond
Seawater
Feed
Electricity Wastewater (water exchanging)
Shrimp Farming
Fertilizers Air emissions
Chemicals
Antibiotics
Fuel
Shrimp adults
Electricity Harvesting Wastewater (water drainage)
Ice Air emissions
Sludge
Transportation
Shrimp products
Water
Shrimp wastes
Chemicals Processing Wastewater
Electricity
Solid wastes
Papers/plastic
Transportation
Fuel Distribution* Air emissions
Solid wastes
*Distribution means the distribution from storage to wholesalers, retailers and consumers
Figure 1. Principal flow chart of the life cycle of shrimp production.
This is due to the intrinsic properties of geographical location as well as climate conditions at-
tached to a particular site. In searching for the most environmental preferable farming system,
it is rather sensible to apply a qualitative LCA to assess the potential environmental impacts
associated with the choices of site location and culturing technique during the planning phase
of site selection and for pond management during the farming operation. Another issue of
69
concern related to the comparison at national level, between countries, is that there are sig-
nificant differences between and within countries regarding the levels of production intensity,
types of resource utilisation, farm numbers and their sizes (Kongkeo, 1997). Most shrimp
farms in developing countries are operated by a family. How to aggregate the assessment of
individual farms to assess cumulative effects into an “integrated” local, or even national, im-
pact assessment are still problematic. Some effects and impacts are still unknown, such as fate
and transport of chemicals used previously, and impacts on biodiversity. Other barriers to
LCA studies, in Thailand as well as other developing countries, are a lack of LCA expertise
and database systems. From the qualitative LCA results, nevertheless, careful site selection
can help minimise unnecessary impacts such as selecting areas without flooding, avoiding ar-
eas with potential sources of contamination and to make sure that the selected area have
enough available sources of water supply. The choice of farming system, especially the inten-
sity of inputs and management strategies, should be selected based on the particular environ-
mental condition of the site. For instance, the stocking density should not exceed the pond’s
carrying capacity that is associated with the pond soil and water properties to accommodate
the shrimps’ growth. Even permitted chemicals should be used as little as possible and only as
necessary. Primary treatment of wastewater drained from the shrimp pond after crop produc-
tion is needed, for example by sedimentation. Growing fish species in alternate periods, for
instance Tilapia, could reduce the nutrient level in the wastewater. Accumulated sludge in the
pond bottom soil has to be managed properly, for example by collection in a sludge storage
pond for further disposal. Pond drying between crops should be long enough to allow the de-
composition process to come to completion. Transportation time from farm to table should be
minimised and efficient use of water and energy in processing should be applied. Storage time
at the processing plant before transporting to overseas should be as short as possible.
Conclusion
The environmental impacts associated with shrimp aquaculture are dependent on the charac-
teristics of the site and the choice of culturing practice. The qualitative LCA results provided
a better view of the environmental interventions over the life cycle of shrimp production.
However, practical approaches permitting the quantification of environmental impacts associ-
ated with aquaculture need to be further investigated. Application of LCA studies of different
farming systems and sites will ascertain the practices with the least environmental impacts to
make the shrimp industry more sustainable.
References
Kongkeo, H. 1997. Comparison of intensive shrimp farming systems in Indonesia, Philippines,
Taiwan and Thailand. Aquaucture Research 28: 789-796.
Nissapa, A. and Boromthanarat, S. 2002. A Case Study on Institutional Aspects of Shrimp Aqua-
culture in Thailand. Report prepared under the World Bank, NACA, WWF and FAO Con-
sortium Program on Shrimp Farming and the Environment. Work in Progress for Public
Discussion. Published by the Consortium. 72 pp.
Sitthipongpanich, T., 2001. The Thai frozen shrimp industry. Industry Note: Food & Agribusiness
Research. Robobank International, the Global Food & Agribusiness Knowledge Bank: 1-5.
70
Life Cycle Assessment of trout farming in France: a farm level approach
Papatryphon, E.∗1, Petit, J.1, Van der Werf, H. M. G.1 and Kaushik, S. J.2
1
Environmental Systems Analysis Research Group, UMR Sol Agronomie Spatialisation, Institute Na-
tional de la Recherche Agronomique, 65 rue de Saint Brieuc-CS 84215, 35042 Rennes Cedex, France.
2
Fish Nutrition Research Laboratory, Unité Mixte INRA-IFREMER, Unité d’Hydrobiologie, 64310
Saint-Pée-sur-Nivelle, France.
Abstract
The aim of this study was to evaluate the applicability of the Life Cycle Assessment (LCA)
methodology for the comparison of the environmental impacts of individual farms and as a
tool towards understanding and, therefore, assessing the effects of management
choices/decisions at the farm level. Eight trout farms were used to obtain data for the inven-
tory regarding the process of fish production. Results indicate that there is a significant vari-
ability in the environmental impacts among the farms and that most of the variability can be
explained by a detailed analysis of farm parameters. Our results point towards the particular
importance of two parameters-metrics, which could be used as indicators of overall environ-
mental performance: 1) the feed efficiency and 2) the production intensity during the dry sea-
son, expressed as production per volume of fresh water use. We believe that LCA has good
potential for the comparison of farming systems at the farm level.
Keywords: Aquaculture, rainbow trout farming, LCA methodology, farm systems analysis.
Introduction
The environmental impacts of aquaculture activities have received considerable attention, es-
pecially in terms of the production of carnivorous fish, like the salmonids (Naylor et al., 1998;
2000). Rainbow trout comprises approximately 68% by weight of farmed fish production in
France, with an annual production of approximately 40000 tonnes, which sets it as the leading
economic activity of the sector (FAO, 2001). Under the umbrella of the theme “Sustainable
Aquaculture”, the French National Institute of Agronomic Research and the French Inter-
Professional Committee of Aquaculture Products have been engaged in the development and
application of LCA for the evaluation of the environmental impacts of the various aquaculture
sectors, but also as a tool towards optimum farm management and strategic planning.
LCAs in agriculture have been mainly conducted as a means to photograph the environmental
impacts of production processes, and to further position them in a global context or to com-
pare production processes widely different from each other. There have only been a few at-
tempts, to our knowledge, aiming at using LCA on an individual farm basis, and which could
serve as a tool to optimise farm management decisions and strategic planning (Haas et al.,
2000).
∗
Present address: European Commission, Institute for Prospective Technological Studies, Edificio Expo, C/Inca
Garcilaso s/n, E-41092 Seville – Spain, tel: +34 95 4488 402, fax: +34 95 4488 326, [email protected]
71
Goal and Scope
Aims and functional unit
The overall aim of the research project is to develop and apply the LCA methodology for the
evaluation of the environmental impacts of fish farming in France, as a means to support the
communication and decision making of the stakeholders involved, namely the fish farming
industry, the research community, as well as the administrators responsible for putting for-
ward the directives concerning the evaluation and evolution of the industry. The specific goal
of the study described herein was to assess the potential of using LCA as a tool for the identi-
fication and demonstration of the potential variability in the environmental impacts due to dif-
ferent choices in farm management, such as in technical sophistication, product type, produc-
tion planning, and geographical origin, to name a few. Since the trout production industry
comprises the most significant aquaculture activity in France, it was selected as the first can-
didate for evaluation. At first stage, the production process is followed to the farm gate, the
functional unit being one ton of trout live weight.
System boundaries
The system to be studied covers the farm production of trout, as well as the following up-
stream processes as related to non-renewable energy: a) the production and use of primary in-
puts to the farm (oxygen, veterinary treatments and feed), b) the production and transforma-
tion of feed ingredients, including the agricultural and fishery phase of the various
ingredients, c) the production of equipment used on the farm, d) the construction and produc-
tion of necessary infrastructure and e) transportation in all stages. For the assessment at the
farm level, the production of the fish inputs as eggs or juveniles was not taken into account,
so as to calculate and compare the efficiencies of each farm. Fish inputs were included how-
ever in the calculations of the final scenarios.
System description
Rainbow trout aquaculture in France is carried out in freshwater, flow-through raceway-type
installations, mainly using river water, and following the intensive aquaculture model, as they
rely completely on external provisions of feed. The most recent survey on aquaculture pro-
duction was published in 2000 (Agreste, 2000), indicating the existence of three main product
types (as defined by their size and not by other qualitative traits), namely: portion trout (250-
300 g mean fish weight), large trout (900-1500 g mean fish weight) and very large trout
(2000-3000 g mean fish weight). Approximately 50% of the trout produced is of the portion
size while the other 50% is trout of larger sizes. From a geographical point of view, 50% of
the production is carried out in only two regions, Aquitaine and Bretagne, on an approxi-
mately equal basis. Eighty percent of the production by weight is produced by only 20% of
the production facilities, indicating the importance of large production capacity farms, also
being the ones with the highest investments towards intensification, mainly related to water
treatment (aeration, oxygenation, recycling etc). Data, concerning the year 2002, from eight
individual farms that were located in Aquitaine and Bretagne, were collected in an effort to
cover a large part of the variability regarding on-farm production practices, namely final
72
product size and production intensity (expressed as fish production per unit volume of fresh
water use during the driest month). The reason for choosing these two parameters as our basis
of farm categorisation arises from our belief that on-farm production practises vary widely as
these parameters change. The individual farms were also chosen based on their potential to be
included in the different production scenarios (to complete the production cycle, from the
broodstock to the final product), as well as the availability of data, and their willingness to
collaborate and share data.
Production scenarios were constructed based on data obtained from the above mentioned
farms in order to simulate as best as possible the complete production process for each indi-
vidual farm. To this end, three separate production processes (corresponding to separate farm
units) were used: A) The production of broodstock and eggs, B) the production of juveniles
and portion trout and C) the production of large and very large trout. Two scenario types,
which are also the most important production scenarios of the French trout farming, were
identified in the farms we assessed: Type I, from the egg to portion size trout and Type II,
from the egg to large and very large trout.
Allocation
Economic allocation was used for all processes yielding by- and co- products during the pro-
duction of feed ingredients. On the fish farm, trout was considered as the only product and,
therefore, allocation was not necessary. In farms where trout of various market sizes is pro-
duced, the product was considered to be identical, and a weighted average was used for cate-
gorisation. The production of manure from trout farms was considered as a waste product and
as a result no allocation was used either.
Selection of impact categories
The impact categories that were included in this study are Eutrophication, Acidification, Cli-
mate change, Energy Use and Biotic Resource Use (expressed as Net Primary Production
Use). The first four were chosen as they are among those that are considered critical for inclu-
sion in all LCA studies (Guinée, 2002). Furthermore, Eutrophication, Energy Use and Biotic
Resource Use were judged to be the most relevant categories for the system we studied, and
extra care was taken towards their assessment. Biotic Resource Use was calculated as the di-
rect use of net primary production, while indirect use, such as that due to land use or other ac-
tivities, was not taken into account (Papatryphon et al., in press).
Inventory Analysis
Data regarding inputs and outputs to, and from, the farms, were directly obtained from pro-
ducers records. Information regarding the data and assumptions for the production of feed can
be found elsewhere (Papatryphon et al., in press). Data on oxygen production and transport, as
well as the production of equipment used on the farms were obtained from experts. Data re-
garding the infrastructures were calculated based on data obtained from each farm. Emissions
regarding all agricultural phases were calculated according to Papatryphon et al. (in press).
73
Results
Environmental Impact Assessment – farm level
Table 1 shows the range as well as the relative variation for the impacts associated with the
five impact categories included in this analysis. The assessment at this stage does not include
the upstream processes regarding the production of fish that supply the farm, since we are in-
terested at obtaining the impacts at the farm level (and not the production scenarios). A varia-
tion of 40-50% is found for most impact categories except for energy use, in which case the
variation reaches approximately 90% and for biotic resource use, for which it is lower at 24%.
In an effort to identify the factors responsible for the variability in the calculated impacts, the
farms were categorised in three groups relative to the final product type (Table 1), indicating
that there is an increasing trend for all impact categories as the size of the final product in-
creases.
Table 1. Variation in the total calculated impacts among seven individual farms categorised
per final product type (as main product) for the production of 1 ton of rainbow trout live
weight.
Portion trout Large trout Very large trout Relative variation (%)1
Number of farms 2 3 2
Eutrophication kg PO4-eq 46.3-63.8 58-72 65.9-74.8 47
Climate change kg CO2-eq 1760-1850 1960-2290 2430-2760 44
Acidification kg SO2-eq 12.1-13.7 14.4-17 16.5-19.1 45
Energy use MJ 30000-42300 41000-57900 51900-78200 89
Biotic resource use kg C 48700-50700 53000-57400 60500-62200 24
1
Relative variation = (maximum value - minimum value)/((minimum value + maximum value)/2)
Contribution analysis – farm level
A process contribution analysis was conducted for all relevant impact categories. The analy-
sis indicates that feed production has the highest contribution to the impact categories biotic
resource use (100%), climate change (mean 83%, range 77-88%), acidification (mean 82%,
range 75-87%) and energy use (mean 52%, range 36-73%), while it is responsible for only
6% (range 5.5-6.6%) of the impact towards eutrophication. Fish production is the major con-
tributor to eutrophication (mean 94%, range 93-94%). Electricity and liquid oxygen use are
the second and third most important contributors to energy use, respectively, followed by in-
frastructure, equipment, diesel and formol production. In order to understand the variation
within each product category, the farms were also categorised based on the production inten-
sity. A comparison of the energy use among the different farms indicated, that there is a clear
trend towards an increase in energy use with increasing production intensity and with increas-
ing average product size and, therefore, these two parameters may be determinant factors for
this category (Table 2). The variation and larger overlap in the impact to eutrophication be-
tween the different product types is mainly related to the variation in the feed efficiency en-
countered between the different farms and to the difference in the degree of solid waste re-
moval (from 0 to approximately 30%).
74
Table 2. The variation in energy use for the production of 1 ton of rainbow trout live weight
in 7 farms categorised by production intensity and final product size (as main product).
Category Production intensity Final Number Energy
(g production/m3 fresh water use) product size of farms Intensity (MJ)
1 low Portion 1 30000
2 medium Portion 1 42300
3 medium Large 1 41000
4 medium Very large 1 51900
5 high Large 2 57600
6 high Very large 1 78200
Environmental Impact Assessment - production scenarios
Table 3 shows the range as well as the relative variation for the impacts associated with the
five impact categories included in the analysis of the two production scenarios. The relative
variation is similar from what was found in Table 1 in the comparison at the farm level, since
in the production scenarios the final impact is higher but it increases more or less equally for
both scenarios.
Table 3. Variation in the total calculated impacts among two farm production scenarios (I:
portion trout and II: large and very large trout) for the production of 1 ton of rainbow trout
live weight.
Scenarios Relative variation (%)1
I II
Number of farms 2 5
Eutrophication kg PO4-eq 48.4-65.4 68.8-79.6 49
Climate change kg CO2-eq 1800-1902 2120-2877 46
Acidification kg SO2-eq 12.4-18.7 16-20.3 48
Energy use MJ 31517-44166 44664-80884 88
Biotic resource use kg C 49662-51962 59797-65753 28
1
Relative variation = (maximum value - minimum value)/((minimum value + maximum value)/2)
Contribution analysis – production scenarios
Table 4 shows the contribution of each production process to the impacts of the complete
production scenarios. For simplicity and since the relative contributions of each production
process (A, B or C) to the final impacts were similar for all impact categories, a range is pre-
sented. For Scenario type I, the process concerning the production of the final product has, by
far, the largest contribution (95-98%). Similarly, the production process of the final product is
also the most important contributor for the Scenario type II (78-97%), although the production
of juveniles/ portion sized fish can become an important contributor to the overall production
process as well (3-21%).
75
Table 4. Contribution of individual processes to the final impacts for the two production sce-
nario types studied (range of % contribution for the different impact categories and different
farms used).
Processes included Scenario Type I Scenario Type II
Range Range
A. Production of large and very large trout 78-96.7
B. Production of juveniles and portion trout 95.2-98 3.2-21
C. Production of broodstock and eggs 2-4.8 0.07-1
Discussion
Farm impact assessment
A comparison among the different farms indicates that there is significant variability in all
environmental impacts for the same product species, namely the rainbow trout. As the rain-
bow trout industry is the most mature and competitive aquaculture industry in France, it has
evolved in a way that it has diversified in terms of the production practises. The most impor-
tant difference regarding the production orientation among trout farms is the final product size
which, by default, has significant implications in the zootechnical, environmental and eco-
nomic performance of the farms. Results from this analysis, which is based however on a lim-
ited set of farms, indicate that as final product size increases so do the environmental impacts
for all impact categories assessed, when expressed on a unit mass production basis. For the
impact categories Biotic resource use, Climate change and Acidification, feed is the major
contributor to the environmental impacts. Furthermore, feed efficiency is directly related to
the size of the fish being produced, since as fish size increases, feed efficiency decreases. It is,
therefore, not surprising to find that there is a correlation between the calculated environ-
mental impacts and the final product type, which is due, in turn, to the underlying correlation
of fish size and feed efficiency. A more detailed analysis regarding the environmental impacts
of feeds using LCA has previously been conducted (Papatryphon et al., in press). For the im-
pact category Eutrophication, the process of fish production is by far the greatest contributor.
However, since the vast majority of eutrophying potential on the farm is of feed origin, it is
clear that the variation in feed efficiency will also be the major parameter affecting the score
to this category. Solid waste removal in French farms is not widely practiced, and from the
farms assessed, only four had in place some type of removal. Two types of waste removal are
mainly used in French fish farms, either filters specially designed for this purpose or the use
of sedimentation basins at the outlet of the farms. When the removal of solid waste is not
taken into account, there is a clear trend towards an increase in the impact to eutrophication as
final product size increases. However, due to the existence of the removal technologies on
some of the farms, the reported overlap in eutrophication was observed. For the impact cate-
gory of energy use, the situation seems to be a little different, since a large part of the impact
is due to the use of electricity on the farms, but also to the use of other high energy carrying
inputs, such as liquid oxygen. Since electricity and oxygen are mostly required when the pro-
duction intensity is high, which in turn may be due to farm management practices (elevated
stock size, high feed distribution) and to environmental parameters (low river flow, high tem-
76
perature) a good correlation was observed between the production intensity during the dry
month and energy use. A correlation was also observed between final product size and energy
use, as feed is also an important contributor to energy use and as feed efficiency varies with
fish size. It seems, therefore, that feed efficiency and production intensity are the deciding
factors affecting energy use. Furthermore, as high production intensity is related to high level
of fuel and electricity use (direct or indirect, i.e. through oxygen use), it follows that feed effi-
ciency and production intensity are also the deciding factors affecting the impact categories
climate change and acidification.
Farm Scenario impact assessment
An assessment of complete production scenarios is necessary from an LCA point of view, in
order to obtain a complete picture of the life cycle impacts of a product. Our assessment indi-
cated that the large majority of the potential impacts is associated with the production process
found at the end of the production chain, as would be expected, since very small quantities of
the processes located upstream are required for the production of the final product. Moreover,
the effect of the fish hatchery on the overall impact is very small, which is due to the high fe-
cundity of fish, as compared to other farm animals. Nevertheless, it is of interest that large
variations can be found in the contribution of the production of juveniles and portion trout to
the production of large and very large trout, which indicates that the overall environmental
performance of a product may be more or less different from the environmental performance
of the last stage of production. We believe that the variation calculated herein may provide a
good indication of the expected variability of the potential environmental impacts of French
trout farming under the impact assessment methodology used.
References
Agreste 2000. Recensement de la Pisciculture Française en 1998 : Salmoniculture et Pisciculture
Marine. Agreste Données Chiffrées Agriculture, No 124. Ministère de l'agriculture et de la
pêche, Service Central des Enquêtes et Etudes Statistiques, Paris, 128 pp. (In French).
Guinée, J.B. 2002. Handbook on Life Cycle Assessment. Operational Guide to the ISO Stan-
dards, Kluwer Academic Publishers, Dordrecht, 708 pp.
FAO, 2001. FAO yearbook, Fishery statistics, Aquaculture production 2001. Vol 92/2.
Haas, G., Wetterich, F. and Geier, U. 2000. Life Cycle Assessment framework in agriculture
on the farm level. Int. J. LCA 5, 345-348.
Naylor, R.L., Goldburg, R.J., Mooney, H., Beveridge, M., Clay, J., Folke, C., Kautsky, N.,
Lubchenco, J., Primavera, J. and Williams, M. 1998. Nature’s subsidies to shrimp and
salmon farming. Science 282, 883-884.
Naylor, R.L., Goldburg, R.J., Primavera, J.H., Kautsky, N., Beveridge, M.C.M., Clay, J., Folke,
C., Lubchenco, J., Mooney, H. and Troell, M. 2000. Effect of aquaculture on world fish
supplies. Nature 405, 1017-1024.
Papatryphon, E., Petit, J., Kaushik, S. J. and Van der Werf, H. M. G. Environmental impact as-
sessment of salmonid feeds using Life Cycle Assessment. Ambio, in press.
77
Environmental impacts from Danish Fish Products
Mikkel Thrane
Department of Development and Planning. Aalborg University,
Fibigerstraede 13, 9220 Aalborg, Denmark. Phone +45 96 35 83 16, e-mail [email protected].
This article presents the results from an environmental assessment of Danish fish products
carried out as part of a PhD project at Aalborg University. The article includes data for energy
and water consumption for a wide range of fish products, but also presents the results of a
life cycle assessment (LCA) of 1 kg frozen flatfish filet. It is concluded that the fishing stage is
one of the most important stages, not only for flatfish – but Danish fish products en general. It
is argued that the fuel consumption represents an important impact potential and that im-
provements in fuel efficiency are consistent with other objectives, such as a reduction of dis-
card, overexploitation and seabed impact. In this regard, it is shown that that the fuel con-
sumption can be reduced with up till a factor 15 in some cases. It is suggested that more
attention should be given to the development of cleaner production methods in the fishing
stage, and that the previous focus on the processing stage in terms of environmental regula-
tion and cleaner production can be characterized as sub-optimization from a life cycle per-
spective.
Key words: Danish fishery, Energy consumption, LCA, Flatfish, Environmental policies
Introduction
As environmental policy becomes more product-oriented, an increasing need for seeing envi-
ronmental impacts in a life cycle perspective (from sea to table) is occurring. Traditional envi-
ronmental regulation mainly focuses on the companies and their (on site) emissions, but fre-
quently larger environmental impacts are found elsewhere in the life cycle.
So, far few studies have been made of the environmental impacts of fish products from a
“cradle to grave” perspective, but recently LCA studies from Sweden and Island have been
published suggesting that the fishing stage is the environmental hot-spot for Atlantic cod (Ey-
jólfsdóttir et al., 2003; Ziegler et al. 2003). Previously, Danish LCA screenings of pickled
herring, canned mackerel, and frozen peeled blue mussels have come to the same conclusion
(Christensen et al. 2001).
The studies mentioned above only include a limited number of fish species and have some
methodological shortcomings in relation to co-product allocation (typically handled by mass
or socio-economic criteria) and representativeness – especially in previous Danish LCA
screenings.
The dissertation ”Environmental Impacts from Danish Fish Products” (Thrane, 2004a) in-
cludes a detailed analysis of the environmental impacts from Danish fish- and shellfish prod-
ucts. The analysis represent state-of-the-art LCA methodology (that is a consequential ap-
78
proach), which is characterized by market based modelling of the product system, use of mar-
ginal data and the application of system expansion for handling of co-product allocation (Ek-
vall and Weidema, 2004). The methodological aspects are further described in the following.
Methodological aspects
The environmental assessment in Thrane (2004a) is carried out in three separate steps: a
MECO analysis, a quantitative LCA, and finally a qualitative LCA:
• The MECO analysis provides information about the inputs and outputs of materials,
energy, chemicals, and other aspects - for all major species groups, and can be per-
ceived as an LCA without the impact assessment phase.
• The LCA includes a detailed LCA of flatfish but LCA screenings are also made of
other fish products for the purpose of generalization.
• Finally, the qualitative LCA provides an overall assessment of impact categories,
which have not been included in the previous LCA, addressing all major species
groups.
The present article presents key findings from the MECO analysis, as well as the results from
the detailed LCA of flatfish - but results for the LCA screenings and the qualitative LCA are
also discussed briefly.
Data collection
Data for the fishing stage has been relatively detailed. The data include a sample of 330 ves-
sels representing a population of 1528 vessel making up 99% of the total Danish catches,
measured in standard-catch-value. (Danish Research Institute of Food Economics, 2001).
These data have been supplemented with data for specific fishing gear and vessels sizes (Niel-
sen, 2002) and personal interviews with fishermen, slip-ways and producers of antifouling
agents.
The data for fish processing are based on literature studies supplemented with empirical stud-
ies. The data for the processing stage represent 10 different plants and the main references are
Andersen et al (1996), Matcon A/S, and Dansk Energi Analyse A/S (1995). In the LCA of
flatfish, detailed empirical data from one large plant have been applied as the main source.
For the wholesale- and transport stage, I have used a combination of empirical data (one com-
pany in each case) and theoretical calculations based on literature. For the consumer stage, I
have only used literature data, while the data for the retail stage are based on empirical data
from three supermarket chains combined with literature data for cooling processes etc. For
more details, see Thrane (2004a)
79
Co-product allocation
Both for the detailed LCA of flatfish and the LCA screenings of other fish products, co-
production allocation has been avoided through system expansion in the fishing stage, where
several species typically are caught at the same time (target fish and by-catch). System expan-
sion has also been applied at the processing stage, where a number of by-products occur after
the filleting process. In both cases, the applied methodology follows the principles for co-
product allocation described in Weidema (2001). Allocation by physical causality (weight or
volume) has been applied for most other stages / processes, while economical allocations have
been used for exchanges related to shopping at the use stage.
Data presented for energy consumption in this article are slightly overestimated, especially for
the processing stage, as avoided exchanges haven’t been considered in the MECO analysis.
System delimitation
Material flows representing more than 0,1% of one kg of fish product dispatched from the
processing stage are included, but for certain chemicals such as anti fouling agents, I have in-
cluded smaller flows. The cut-off criteria for energy consumption have been 0,1 MJ per kg of
frozen flatfish, dispatched form the processing stage.
As mentioned, market based system delimitation has been applied. As the purpose of the
study has been to elucidate the exchanges and impacts from Danish fish products in a regula-
tion perspective, the product chain has been fixed to Danish producers and European super-
markets and consumers. The latter is because more than 90% of the Danish fish most are ex-
ported – mainly to the European market. Hence, compared to an attributional LCA, the
consequential approach has not affected the system delimitation with respect to the immediate
product chain (foreground data).
Other methodological aspects of the LCA
In the LCA case study of flatfish the functional unit is one kg consumed frozen flatfish filet
(IQF) in consumer packaging made of cardboard in boxes of 300 gram each. The results pre-
sented here only include characterization (the first step of impact assessment). An updated
version of the method known as EDIP 96 described in Wenzel et al. (1997) has been used.
The PC tool “SimaPro version 5.1” has been used for the calculations and the updated EDIP
method is entered by 2.-0 Consultants as part of the LCAfood project. Several databases have
been used for related processes, such as materials, chemicals, and energy. In this regard, I
have used the ETH-ESU database, which includes capital goods for all energy processes. The
database is around 10 years old and large uncertainties exist. Specifically for the direct emis-
sions at the fishing stage, I have used data from the European Environmental Agency (Euro-
pean Environmental Agency, 2001). Data from the Danish LCAfood project have been used
for different types of food ancillaries and avoided emissions such as soy protein and minced
80
pork (LCAfood, 2003). This project also uses the ETH-ESU 96 database and most of the data
are also based on system expansion.
MECO analysis - exchanges in a life cycle perspective
The MECO study includes an analysis of exchanges for eight product chains. All exchanges
are measured per kg consumed fish product and the exchanges have obviously been adjusted
for product spillage along the product chain. The results for energy and water consumption
are illustrated in the following.
Energy consumption (Heat and combustion)
The studies of the direct heat and combustion related energy consumption show that the fish-
ery is the most important life cycle stages for demersal fish and shellfish (see table 1). Nor-
way lobster represents the largest energy consumption (roughly 1000 MJ per kg consumed
lobster meat) of which the fishing stage makes up 95%. For flatfish and other shellfish the en-
ergy consumption is well over 100 MJ per kg consumed fish. Other important life cycle stages
are the use stage, but also processing for products that involve boiling such as shrimps.
Table 1. Energy consumption for heat and combustion in MJ per kg consumed fish product -
for eight product types. All products are frozen except the pelagic species.
Demersal fish Shell fish Pelagic
(frozen filets) (boiled, peeled & frozen) (pickled/canned)
Codfish Flatfish Prawn Shrimp N. lobster Muss. Herring Mack.
(block) (IQF) (IQF) (IQF) (IQF) (IQF) (in jar) (canned)
Fishery 41 110 101 136 941 5 16 5
Landing 1 1 1 1 0 4 0 0
Processing 2 3 12 12 12 6 3 3
Wholesale 0 0 0 0 0 0 0 0
Transport 1 1 1 1 1 1 4 2
(if whole) 4 5 5 5 8 9 2 2
Retail 0 0 0 0 0 0 0 0
Consumer 14 11 16 16 44 12 9 9
Total 59 126 130 166 998 28 32 18
The energy analysis shows great differences between the energy consumption in the fishing
stage, depending on the fish species that are caught. For instance, the energy consumption for
Norway lobster is 6 litre per kg caught whole lobster, while it is 1 litre per kg caught whole
flatfish, 0,36 litre per kg caught whole codfish and 0,06 litre pr kg caught whole mackerel
(Thrane, 2004b) Energy consumption (electricity).
I have not included a separate table with information about the electricity consumption, but
roughly 50% of the total electricity consumption is related to the use stage for all products.
The total electricity consumption varies between 8 MJ and 28 MJ per kg. The consumption
81
for the processing stage only represent 1-5 MJ per kg consumed fish products - depending on
the species. Even though electricity consumption only represents around 40% of the electric-
ity produced, it must be concluded that the heat and combustion related energy consumption
is more important.
Water consumption
The most important stages with respect to ground water consumption are the processing- and
use stage. The water consumption in the processing stage can be more than 100 litre per kg
consumed shrimp and prawn (partly because of the boiling process), while it is around 10-30
litre for other species. For the use stage the water consumption is assessed to be around 15-20
litre per kg consumed fish - slightly more if dishwashing is done by hand and slightly less
when dishwashing machine is used.
Considering the differences between different species, the picture is somewhat similar to en-
ergy consumption. Shellfish represents the largest exchanges (Norway lobster, prawn, shrimp
and mussel), followed by demersal (flat- and codfish) and pelagic fish (herring and mackerel)
that have relatively small exchanges at all stages.
LCA case study of flatfish
As mentioned, the MECO analysis does not include impact assessment. That is why the LCA
is carried out. The characterization results for one kg consumed frozen flatfish filet (IQF)
caught by the average fishing method, are illustrated in figure 1. Each impact category is
separately discussed in the following. As it appears the dominating stages are fishing, use, and
retail, but the various impact categories are further discussed in the following.
100%
Use (consumption)
80% Retail
Transport
60%
Wholesale
40% Processing
20% Auction
Fishery
0%
ETWC
ETWS
ETWA
Nutrient en.
depletion
formation
warming
Acidification
Ozone
Global
Ozone
-20%
-40%
-60%
-80%
Figure 1. Characterization results for one kg consumed flatfish filets (IQF). Based on the
Danish EDIP method.
82
GWP – gram CO2 equivalents
As it appear in figure 1, the it is roughly half of the global warming potential (GWP) that is
related to fishing, while the remaining mainly originates from the use and retail stage. Com-
bustion of diesel is the dominating process for fishing while electricity consumption is domi-
nating for the retail stage. For the use stage, it is partly electricity for cooling and food prepa-
ration, and partly combustion of fuel for transport, which are the important factors. The Dutch
method ECOindicator 99 method shows similar results.
ODP - CFC11 equivalents
The potential contribution to stratospheric ozone depletion (ODP) is also mainly caused by
fishing (~90%) due to the production of diesel, which involves Halon 1301, and the emissions
of HCFC-22 cooling agents that are used on the fishing vessels. Again, the EcoIndicator 99
shows similar results. It is estimated that the uncertainty is somewhat larger than for global
warming, as several of the ozone depleting substances occurs only because the ETH database
is relatively old.
Acidification potential – gram SO2 equivalents
For the acidification potential, the largest contribution also originates from fishing, mainly
due to NOx emissions. As for the GWP and ODP, the use and retail stage are also somewhat
important, mainly due to SOx emissions. When the EcoIndicator (99) method is applied, the
fishery dominates even more.
Methodological aspects: According to Wenzel et al. (1997) contributions to acidification in
robust areas such as the sea are less significant. As most of the Danish fishing activity takes
place in the North Sea, it must be assumed that the actual acidification impact related to the
fishing is significantly smaller than suggested above. Thus, it is probably more reasonable to
argue that there are three potential hot spots: fishing, use and retail.
Eutrophication potential – gram NO3 equivalents
As for previous impact categories, the fishing stage completely dominates for eutrophication,
mainly due to NOx emissions caused by the combustion of diesel. In contrast, the impact po-
tential from processing is large but negative. The reason is that the by-products from process-
ing are assumed to substitute soy protein, which is the marginal protein source on the global
market according to Weidema (2003). The substitution of soy protein lead to avoided water
borne emissions of nitrate in South America the overall contribution therefore becomes nega-
tive. The large avoided impact potential from South America hides that some emissions actu-
ally occur from the processing stage. Still the impact potential remains insignificant, as the
emissions of wastewater are treated in modern biological wastewater treatment plants. When
the EcoIndicator 99 method is applied the contribution from fishery becomes even more sig-
nificant, but in this method N and P are not included in the characterization factors, and is
therefore difficult to compare.
83
Methodological aspects: The eutrophication potential analyzed here does not distinguish be-
tween terrestrial and aquatic etutrophication, which is a weakness. Furthermore, site specific
aspects are not included in the modelling, and the results are therefore quite uncertain.
Photochemical ozone formation potential – C2H4 equivalents
For photochemical ozone formation potential, the main contribution also comes from fishing
followed by the use stage. In both cases, this is due to the emission of non methane VOC
(hydrocarbons except methane), which is related to the production of diesel, but also the com-
bustion process itself.
Methodology: This impact category shown in this paper does not distinguish between ozone
leading to lower crop yields in agriculture, and ozone causing respiratory problems in urban
areas. Nor are site-specific aspects included in the assessment. This is obviously a weakness.
Substances that contribute to ozone formation may drift over long distances, but according to
Wenzel et al. (1997) a site-specific factor between 0 and 1 should be applied for emissions in
sparsely populated areas (low background NOx levels) such as desserts or the sea. Hence, it
would probably be more reasonable to argue that fishing and use both are potential hot-spots.
Eco-toxicity potential – measured in m3 water or soil
The EDIP method includes three types of eco-toxicity: Eco-toxicity water (acute and chronic)
as well as eco-toxicity soil (chronic). The unit is m3 that express the water or soil volume re-
quired to dilute the given exchange to a concentration below the Predicted No Effect
Concentration (PNEC). For further details see – Wenzel et al. (1997).
As illustrated in figure 1, both categories of eco-toxicity water are completely dominated by
the fishing stages. This is caused by the emissions of TBT from the anti fouling paint. These
results are plausible, as sex changes have been registered for 10 different kinds of Sea snails
in Denmark, mainly because of biocides from antifouling agents such as TBT and copper
compounds.
For eco-toxicity soil (chronic), the main contribution is the processing stage followed by the
use-, auction- and retail stages. The most important substances are acetone followed by for-
maldehyde and hexane. This is also related to energy production. Normalized and weighted
results are not presented in this paper, but it should mentioning that eco-toxicity soil (chronic)
becomes completely insignificant after the normalization step. This shows that the processing
stage remains relatively insignificant even tough the contribution to ETSC is large.
Future developments
Future scenarios suggest that the fishing stage will remain important even though TBT is sub-
stituted. The alternative antifouling pains are also somewhat problematic, and future scenarios
suggest that fishery will remain dominating for eco-toxicity water (acute and chronic) even
though the toxicity potential is greatly reduced from antifouling agents. Furthermore, the fuel
84
consumption is expected to grow in the fishing stage, while the energy efficiency probably
will increase at other life cycle stages. For further details, see Thrane (2004a).
Other impact categories
There are a large number of other impact categories, which have been separately analyzed in a
qualitative LCA – that is seabed impacts, land use, waste, use of non-renewable abiotic re-
sources, use of groundwater, exploitation of fish, discard and by-catch, human toxicity poten-
tial, occupational health and safety, noise and accidents, and animal welfare.
The analysis shows that different conclusion can be obtained by focusing separately on other
impact categories, but overall the qualitative LCA appear to strengthen the previous conclu-
sions – thus pointing towards fishing as the overall hot-spot followed by use and retail. Fish-
ing is characterized by overexploitation of certain fish stocks, a high frequency of injuries and
accidents among fishermen (including fatal accidents), seabed impact inflicted by bottom
dragged fishing gear (e.g. bottom- and beam trawl), by-catch of sea-mammals, and discard of
fish.
Other fish products
Apart from the detailed LCA on flatfish, the dissertation includes a number of LCA screen-
ings of other typical Danish fish products.
The screenings also suggest that the fishing-, use- and retail stages are the hot-spots (in this
order) for products based on frozen shrimp, prawn and Norway lobster (all in cardboard pack-
aging). Glass, aluminium or steel packaging would further increase the impact potential for
processing, but these packaging materials have only been analysed for mackerel and herring.
For pelagic fish exemplified by pickled herring in glass jars and canned mackerel, the proc-
essing and transport stages are relatively more important, while the opposite is the case for the
fishing- and retail stages. Still, the qualitative LCA points towards the fishing stage as the
most important stage for all fish products. Thus, while the processing stage, can be important
for some products due to packaging, the overall conclusion is that the hot-spots are indeed to
be found in other stages of the life cycle – mainly the fishing and use stage.
Improvement potentials in the fishing stage
The analysis suggests that the largest potential for improvements is in the fishing stage. As it
appear in figure 2, it is possible to reduce the fuel consumption per kg caught flatfish with a
factor 15 by switching from beam trawl to Danish seine. If all flattish were caught by Danish
seine or passive fishing methods such as gillnet, it would theoretically be possible to save
around 30.000 m3 fuel per year in the Danish fishery alone. This is 15% of the total fuel con-
sumption in one year. This figure is rather theoretical and probably overestimated, but im-
provement potential also exists for other species groups.
85
Improvements can also be obtained by promoting the use of purse seine in the fisheries target-
ing herring and mackerel (and maybe other species), and by promoting the use of Danish
seine, gillnet and long line in the fishery after codfish. The latter is also illustrated in figure 2.
The negative fuel consumption for codfish caught by Danish seine is a result of system ex-
pansion – where the by catch substitutes relatively a more fuel consuming fishery.
Liters diesel per kg caught target fish Liters diesel per kroner landed mixed fish
3,00 Flatfish 0,20
2,50 0,16
2,00
0,12
1,50 Codfish
0,08
1,00
0,50 0,04
0,00 0,00
Beam trawl Bottom trawl Danish seine Trawl Gill net Danish seine
(+200 GT) (38 GT) (34 GT) (19GT) (19 GT) (47 GT)
2,61 0,98 0,18 0,46 0,24 -0,20
0,17 0,06 0,02 0,04 0,02 0,03
Figure 2. Fuel consumption per kg caught flat- and codfish in year 20001. (Danish Research
Institute of Food Economics, 2001; Nielsen, 2002; Beam trawler, 2001).
Figure 2 also shows the fuel consumption per landing value (white columns) for both flatfish
and codfish. As it appear Danish seine is roughly 10 times more effective in this respects. For
further details about potentials for improvements - see Thrane (2004a)
Conclusion
The MECO analysis as well as the LCA and the qualitative LCA points towards the fishing
stage as the overall environmental hotspot followed by the use and retail stage for flatfish
products. Still, LCA screenings of other fish products suggest that the hot-spot distribution is
much similar for most other fish products, except for canned mackerel and pickled herring,
where the processing and use stages are the hot spots. Still, if we include the results from the
qualitative LCA, it is suggested that the fishing and use stage remains the overall spots for
Danish fish products as such.
1
The data are based on fishing vessel accounts. Flatfish (bottom trawl) is based on 16 accounts, flatfish (Danish Seine)
comprises 9 accounts and flatfish (beam trawl) is based on interview with the owner of three trawlers. For codfish the number
of accounts is 15 for bottom trawl, 22 for gill net/long line and finally 8 for Danish Seine.
86
There are significant differences (up till a factor 600) in the fuel consumption in the fishing
stage depending on the target fish. Still, what appear to be even more interesting the fuel con-
sumption varies considerably as a function of the fishing gear - even considering the same
target species. In this article it is argued that a difference of a factor 15 exist between the fuel
consumption per kg caught flatfish – depending on the type of fishing gear applied. It is ar-
gued that we need to address these differences in the search for a more fuel-efficient and sus-
tainable fishery. In this regard, improvements in fuel efficiency appear to be consistent with
other objectives, such as reduced impacts on sea floor habitats and reduced overexploitation
of fish stocks
So far, the processing stage has been the focus of attention concerning the impacts on the ex-
ternal environment. They have been regulated through environmental approvals and a number
of cleaner technology projects have been initiated through the 1990s. The focus has been wa-
ter consumption and wastewater emissions. However, the LCA study suggests that the direct
emissions of nutrients from the fish industry are insignificant due to wastewater treatment.
This is a paradox and suggest that we need to draw more attention to the fishing stage – not
only concerning overexploitation, but also in a wider environmental context that include con-
siderations of how the fish are caught - not only how many.
References
Andersen, Erik et al. (1996): Overview of cleaner production practices in the fish industry,
Miljøprojekt fra Miljøstyrelsen nr. 317, København K
Beam Trawler, 2001: Personal communication with the owner of three beam trawlers around
200 GT+ targeting European Plaice in the North Sea. Detailed catch composition for
year 2000 have been available.
Christensen P. et al. (2001): LCA of Danish Fish products, Proceedings, International Confer-
ence on LCA in Foods, Gothenburg, Sweden 26-27 April, 2001, SIK-Dokument 143,
Gothenburg
Danish Research Institute of Food Economics (2001): Account statistics for fishery 2000,
Ministry of Food, Agriculture and Fisheries, Copenhagen.
ETH-ESU 96 (1996): Inventory data for Swiss and the Western European energy supply
situation. Data include emissions from primary energy extraction and construction of in-
frastructures and energy conversion and transmission, The report and its database has
been compiled, edited and revised by the ESU-group at ETH Zürich and the section of
Comprehensive Systems Analysis at the PSI in Villigen.
European Environmental Agency (2001): EMEP/CORINAIR Emission Inventory Guidebook
- 3rd edition. Technical report No 30. (Group 8: Other mobile sources).
Ekvall, T., and B. P. Weidema (2004): System boundaries and Input Data in Consequential
Life Cycle Inventory Analaysis. International Journal of LCA. OnlineFirst.
Eyjólfsdóttir, H. R., E. Yngvadóttir, H. Jónsdóttir, and B. Skúladóttir. (2003): Environmental
effects of fish on the onsumer´s dish - Life cycle assessment of Icelandic frozen cod
products, Icelandic Fisheries Laboratories (IFL), Reykjavik, Island.
87
LCAfood (2003): LCA data on basic food products produced and consumed in Denmark, The
present database has been created by Per H. Nielsen, Ph.D., Anne Merete Nielsen,
M.Sc., Bo P. Weidema, Ph.D. from 2.-0 LCA consultants and Randi Dalgaard, M.Sc.,
Niels Halberg, Ph.D., Danish Institute of Agriculture Sciences, see www.lcafood.dk
Matcon A/S and Dansk Energi Analyse A/S, 1995: Branch energy analysis, Udarbejdet af
Matcon A/S and Dansk Energi Analyse A/S for Energistyrelsen, Fiskeindustriens Ar-
bejdsgiverforening og Dansk Industri, København.
Nielsen, R. (2002): Average fuel consumption and catch composition for different vessels,
gear types and target species - year 2000, Danish Research Institute of Food Economics,
Copenhagen.
Thrane, M. 2004b. Energy Consumption in the Danish Fishery - Identification of Key Factors.
Journal of Industrial Ecology, 8(1): 223-239
<http://mitpress.mit.edu/catalog/item/default.asp?tid=12724&ttype=6>.
Thrane, M. (2004a): Environmental impacts forms Danish fish products, Ph. D dissertation,
Department of Development and Planning, Aalborg University, Denmark. DRAFT
Thrane, M. (2004b): Energy Consumption in the Danish fishery – Identification of key factors
Journal of Industrial Ecology.
Weidema, B. P. (2001): Avoiding co-product allocation in life-cycle assessment, Journal of
Industrial Ecology 4(3):11-33.
Weidema, B. P. 2003: Market information in life cycle assessments. Environmental project
no. 863. Danish Environmental Protection Agency, Copenhagen.
Wenzel, H. et al. (1997): Environmental Assessment of Products, Volume 1: Methodology,
tools and case studies in product development, Chapman & Hall, London
Ziegler F et al. (2003): Life cycle assessment of frozen cod filets including fishery-specific
environmental impacts, International Journal of life cycle assessment 8 (1) 39–47,
ecomed publishers, Germany.
88
Operation specific engine load pattern and exhaust gas emission data from
vehicles used in typical Swedish agricultural operations
M. Lindgren; O. Pettersson; O. Norén; P.-A. Hansson
* Department of Biometry and Engineering. P.O. Box 7032, SE 750 07 Uppsala, Sweden, tel: +46 18
67 12 39, fax: +46 18 67 35 29, [email protected];
Swedish Institute of Agricultural and Environmental Engineering. P.O. Box 7033, SE 750 07 Uppsala,
Sweden;
Abstract
The purpose of the work was to present operation specific engine load data, fuel consump-
tion and exhaust gas emission data from vehicles used in typical agricultural operations.
Three agricultural tractors and one combined harvester were equipped with instruments for
measuring engine speed and loading torque. Recorded data were combined with emission
data measured in a test bench in order to calculate operation specific fuel consumption and
emission data. The result showed that the engine load patterns were rather independent of
vehicle used but that the variations between different operations were big, both in terms of
engine load pattern and resulting engine exhaust gas emissions.
Keywords: emission, fuel consumption, operation specific, agricultural
Introduction
High quality data on fuel consumption and engine exhaust gas emission amounts from agri-
cultural tractors and other agricultural vehicles are needed in calculations of environmental
loads caused from different food production and agricultural strategies (Hansson and Matts-
son, 1999). Fuel consumption and engine exhaust gas emission data are normally produced
through measurements according to different standards. In Europe, USA and Japan there is an
obligation for manufactures to certify new engine models to emissions performance standards.
The ISO 8178 (ISO, 1996) and ECE R49 (EEC, 2000) standards currently used for agricul-
tural tractors are based on emission measurements under steady-state conditions. Emissions
are measured individually at different modes (combination of engine speed and torque) and
the resulting average emission values (one for each emission) are obtained by weighting the
values of the different modes together. These standards are used for all types of nonroad mo-
bile machinery (EU, 2000) and the weighting factors are not adapted to agricultural conditions
(Treiber and Sauerteig, 1991: Rinaldi and Näf, 1992).
Due to the varying use of agricultural vehicles it is not possible to design one set of weighting
factors valid for the average use of agricultural vehicles (Renius, 1994: Hansson et al., 2001).
Earlier studies (Hansson et al., 2001) have shown that operational specific fuel consumption
and emission amounts cannot be accurately calculated without account being taken to the ac-
tual engine load in the operation performed. Lindgren et al. (2003) found that the engine load
strongly varied between different typical operations with agricultural tractors, from low load
and engine speed operations like spreading fertiliser to high load and engine speed operations
89
like a stubb puller operation. However, the amount of detailed operations specific fuel con-
sumption and emission data available for these vehicles is very limited.
Most fuel consumption and emission data used today are obtained from measurements under
steady-state conditions with no consideration to the engine‘s emission characteristics at tran-
sient conditions i.e. variations in engine speed and torque. Hansson et al. (2003) found that
the fuel efficiency decreased with increasing proportion of fast variations in engine load while
Lindgren et al. (2003) found that the emission amounts during a front end loading operation
including fast load variations increased compared to steady-state conditions. Lindgren et al.
(2003) also found that the variations in engine load for most operations with agricultural vehi-
cles were rather slow. For most operations, except for high transient operations like front end
loading, fuel consumption and emission amounts can be analysed with the semi-static ap-
proach as described by Hansson et al. (1998). This has also been confirmed by Cornetti et al.
(1988).
The purpose of the work was to present operation specific fuel consumption and emission
data from vehicles used in typical agricultural operations. The purpose was also to obtain op-
eration specific engine load data for the vehicles mentioned.
Material and methods
Vehicles and instrumentation
Two medium-sized tractors, a Valtra 6600 and a Valtra 6650 with a rated engine power of 75
kW and 81 kW respectively, and one large tractor, a Case IH MX 270 with a rated engine
power of 240 kW were studied. Furthermore, one combined harvester, a Massey Ferguson
7254 with a rated engine power of 160 kW, was included in the study. The Valtra 6600 tractor
was manufactured in 1996, before engine exhaust gas emissions got regulated through emis-
sion performance standards. The Valtra 6650 and the Massey Ferguson were delivered in
2000 and 2002 and both complied with the European stage I regulation (EU, 2000). The Case
IH MX 270 was delivered in 2002 and passed the stage II regulation. All vehicles were
equipped with instrumentation for measuring engine speed, vehicle speed and fuel consump-
tion. All data were sampled with 1 Hz and stored on a data logger. The measurement system
is described in more detail in Pettersson et al. (2002).
Selection of operations
Six different farms, both with and without livestock handling, were studied. The main crite-
rion was that the farm should have existing mechanical equipment suitable for the vehicle in
question. The vehicles were delivered to the farm with the intention to replace the existing
vehicles. Moreover, the drivers were instructed to use the vehicles in a normal manner. All the
farms were situated in the middle of Sweden and the predominant textural class of soil was
silty clay loams to clayey soils according to the Swedish soil materials e.g. more than 30 %
clay (Hillel, 1982).
90
One individual operation for each type of operations was chosen as representative. The selec-
tion was based on statistical methods and calculated the fit of the individual operation to the
sum of all operations of that specific operation. The operation with the best fit was chosen for
further analysis and was considered to represent an as ‘normal’ operation as possible.
Emission measurements
The engines were tested at the Swedish Machinery Testing Institute located in Umeå, Swe-
den. All tests were conducted with a low sulphur (maximum 10-ppm sulphur) diesel fuel with
low aromatic content, classified as Swedish environmental class 1 diesel fuel. Emissions of
CO2, CO, hydrocarbons (HC) and nitrogen oxides (NOx) were measured as well as fuel con-
sumption. Particulate matter emissions was only measured as an average value for the whole
cycle and thus not possible to include in this study although particulate matter emissions
could be substantial. The engine dynamometer used in the transient tests was a fast response
Schenck eddy-current dynamometer with 400 kW maximum power. Engine output speed and
torque was controlled through an electronic fuel pump rack control in combination with the
dynamometer brake power. The engine dynamometer, control-system and analysis instru-
ments are described in more detail in Wetterberg et al. (2002).
Fuel consumption and emissions of CO2, CO, HC and NOx were measured according to a 20
mode steady-state cycle. The 20 mode steady state cycle used was based on the 8 mode inter-
national ISO 8178 test cycle (ISO, 1996) and extended with 12 additional modes in order to
increase the resolution, se Figure 1. The fuel consumption and emissions were measured in
accordance with the ISO 8178 regulation (ISO, 1996).
Torque [% of maximum torque]
100
80
60
40
20
0
0 20 40 60 80 100
Engine speed [% of rated speed]
Figure 1. 20-mode steady-state test cycle based on ISO 8178 (∆) and extended with 12 addi-
tional modes (Χ).
91
Data analysis
The data analysis was based on the following parts:
(1) The development of an expression deciding the engine torque in Nm (moment) as a func-
tion of engine speed and fuel consumption. The proposed model was similar to that ex-
τ = c1 × f + c2 × f 2 + c3 × f 3 + (c4 × f + c5 × f 2 + c6 × f 3 )× n + (c7 × f + c8 × f 2 + c9 × f 3 )× n 2
pressed by Jahns (1983) and consisted of nine coefficients specific to the engine:
where τ was engine torque, n was engine speed, f was steady state fuel consumption and
ci were engine specific coefficients. The coefficients c1 to c9 were found for each engine
by least-squares regression (R2=1.00) using values of engine speed, engine torque and
steady-state fuel consumption obtained from the emission measurements;
(2) The development of a matrix over emissions of CO2, CO, HC and NOx for all possible
steady-state combinations of engine speed and torque data for each engine, within the
operational range of the engine. Engine speed and torque data were rounded towards the
nearest integer. The data for the matrix were obtained from the 20-mode steady-state
emission measurement by two-dimensional interpolation;
(3) Registration of engine speed, fuel consumption etc when performing the work
operation to be analysed;
(4) Calculation of the engine torque using the recorded engine speed and fuel con-
sumption data from part (3) as inputs to the expression from part (1);
(5) Calculation of operation specific weighting factors by dividing the operational range of
the engine into eight smaller areas, one for each mode in the ISO 8178 standard (Hansson
et al., 1999). The weighting factors were calculated as the relative frequency of recorded
combinations of engine speed and torque data in each of the eight areas.
(6) Calculation of operation specific emission data using the time series of engine speed and
torque data from part (4) as inputs to the matrix in part (2);
Result
Operation specific engine load pattern
Engine load patterns were recorded for several different operations and varying conditions. In
Table 1 and 2 is the engine load pattern presented as weighting factors according to the ISO
8178 standard for recorded operations with the agricultural tractors and the combined har-
vester, respectively. The variations in engine load between different operations was substan-
tial e.g. from fairly steady-state, low engine speed and low load operations like urine manure
spreading and the PTO driven tedder operation to operations with highly varying engine load
like the transport operations or the power demanding precision chopper operation.
Operation specific emissions
The operation specific fuel consumption and emission data for the operations recorded are
presented in Table 3 and 4. Data of average engine power and cultivated area per hour, where
required, is also shown in Table 3 and 4. Table 3 includes multiples of several operations de-
pending on the different vehicles used. Each vehicle, e.g. engine, produces different amounts
92
of emissions due to different technical designs etc. Examples of the variation in fuel con-
sumption and emission amounts between different operations are presented in Figure 2. The
data are presented as relative values compared to the ISO 8178 amounts.
The result presented in table 3 and 4 can easily be recalculated to g l-1 or g MJ-1 fuel con-
sumed by use of the density (814 g l-1) and energy content of diesel (35.15 MJ l-1).
Table 1. Operation specific engine load data for typical operations with agricultural tractors.
Operation Distribution of engine operations by ISO 8178 modes in %
Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6 Mode 7 Mode 8
ISO 8178 C1 15 15 15 10 10 10 10 15
Bale wrappers 0 0 0 0 1 7 40 52
Baler 0 32 10 0 0 1 56 1
Fertiliser spreader 1 6 4 1 1 1 43 43
Forest trailer 4 4 2 0 2 2 6 80
Front end loading 0 0 2 8 0 0 73 17
Harrowing, heavy 56 1 0 0 23 16 2 2
Harrowing, normal 27 47 7 0 3 11 5 0
Mower conditioner, heavy 38 24 2 0 1 35 0 0
Mower conditioner, light 0 0 0 1 0 1 95 3
Ploughing 61 8 6 9 6 2 5 3
Ploughing 38 13 8 6 8 9 18 0
Precision chopper, heavy 89 10 1 0 0 0 0 0
Rolling 0 0 1 0 0 2 83 14
Semi-liquid manure spreader 0 36 63 1 0 0 0 0
Solid manure spreader 13 26 26 1 0 18 16 0
Sowing, high engine speed 1 25 67 4 0 0 3 0
Sowing, low engine speed 0 3 3 0 0 7 79 8
Stubb puller 73 26 1 0 0 0 0 0
Tedder 0 0 0 0 0 1 99 0
Transport of gravel 13 7 12 5 4 6 12 41
Transport of manure 44 13 17 4 16 1 4 1
Transport on country road 37 27 14 8 7 3 2 2
Urine manure filling 0 0 10 40 0 0 40 2
Urine manure spreader 0 1 2 1 0 0 95 1
Table 2. Operation specific engine load data for typical operations with a combined harvester.
Operation Distribution of engine operations by ISO 8178 modes in %
Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6 Mode 7 Mode 8
ISO 8178 C1 15 15 15 10 10 10 10 15
Barley harvesting 6 26 50 18 0 0 0 0
Oats harvesting 4 23 49 24 0 0 0 0
Rapeseed harvesting 5 22 54 19 0 0 0 0
Winter wheat harvesting 18 44 28 10 0 0 0 0
93
Table 3. Operation specific fuel consumption and emission data for typical operations with
agricultural tractors.
Operation Power Work rate Fuel Emissions (g/h) Vehicle
% ha/h kg/h CO NOx HC
Bale wrappers 27 5.0 22.3 236 5.9 Valtra 6600
Baler 56 9.5 22.9 369 9.3 Valtra 6600
Fertiliser spreader 21 11.0 4.4 32.3 181 7.2 Valtra 6600
Forest trailer 2 3.0 36.1 148 6.2 Valtra 6600
Front end loading 16 3.8 22.6 77 6.6 Valtra 6650
Harrowing, heavy 90 4.9 14.8 99.8 716 10.3 Valtra 6650
Harrowing, normal 76 8.0 12.8 27.0 469 10.2 Valtra 6600
Mower conditioner, heavy 79 2.5 12.9 25.8 512 9.4 Valtra 6600
Mower conditioner, light 33 1.6 6.8 24.3 172 9.1 Valtra 6650
Ploughing 68 0.7 11.6 27.5 449 9.4 Valtra 6600
Ploughing 80 0.6 13.5 88.9 624 10.2 Valtra 6650
Ploughing 70 1.9 34.4 107.8 1202 11.7 Case IH
Precision chopper, heavy 99 0.8 17.3 30.8 618 10.6 Valtra 6600
Rolling 37 5.8 6.9 17.2 216 7.5 Valtra 6650
Semi-liquid manure spreader 53 6.9 10.3 29.5 310 11.4 Valtra 6650
Solid manure spreader 59 2.4 10.5 33.6 385 10.4 Valtra 6650
Sowing 48 6.6 24.3 79.0 679 11.2 Case IH
Sowing, high engine speed 52 2.3 9.5 29.0 311 10.7 Valtra 6600
Sowing, low engine speed 41 2 7.6 21.2 243 8.4 Valtra 6650
Stubb puller 92 1.3 15.9 30.0 564 10.7 Valtra 6600
Tedder 32 2.2 6.2 25.8 241 7.9 Valtra 6600
Transport of gravel 35 6.4 33.3 257 7.9 Valtra 6600
Transport of manure 79 13.3 87.5 596 10.8 Valtra 6650
Transport on country road 72 12.3 62.5 496 10.9 Valtra 6650
Urine manure filling 15 4.1 40.4 140 8.8 Valtra 6600
Urine manure filling 22 5.9 28.0 102 9.5 Valtra 6650
Urine manure spreader 37 4.4 7.0 25.0 267 8.6 Valtra 6600
Urine manure spreader 35 6.7 6.7 23.5 181 8.5 Valtra 6650
Table 4. Operation specific fuel consumption and emission data for typical operations with a
combined harvester.
Operation Power Work rate Fuel Emissions (g/h)
% ha/h kg/h CO NOx HC Vehicle
Barley harvesting 48 1.6 22.9 211 739 12.6 MF 7254
Oats harvesting 44 1.8 21.7 191 675 12.1 MF 7254
Rapeseed harvesting 44 2.4 21.2 175 680 11.5 MF 7254
Winter wheat harvesting 63 2.1 27.8 337 1045 14.7 MF 7254
Discussion
Standardised test cycles like ISO 8178 shows a fairly uniform engine load with slightly em-
phasis on the high speed and load area. Only a minority of the typical agricultural operations
studied had an engine load pattern that resembled the standardised one, i.e. on-road transport
with universal trailer. The results clearly shows that the standardised weighting factors and
thus the resulting fuel consumption and emission data is not adapted to the real use of the
tractor. Therefore, the use of one single set of engine load data, i.e. weighting factors, and fuel
consumption and emission data valid for all operations with typical agricultural vehicles, as
for example in life cycle assessment analyses (LCA), may give misleading results.
94
300
250
200
%
150
100
50
0
Transport of manure
ISO 8178
Urine manure
Harrowing
Ploughing
manure spreader
Rolling
Front end loading
Mower conditioner
spreader
Semi-liquid
Figure 2. Fuel consumption and emission data for different operations with the Valtra 6650
agricultural tractor relative to the ISO 8178 standard measurement: , fuel consumption; ,
CO; , NOx; , HC.
The standardised measurement method used today for nonroad mobile machinery, ISO 8178,
is based on measurements at steady-state load conditions. When measure at steady-state con-
ditions no respect is take to the engine’s characteristics at fast variations in engine speed and
torque. Resent research work has shown that transients in engine speed and torque strongly
influence the fuel consumption and emission amounts (Lindgren et al. 2003). However, most
operations studied in this work were of less transient nature. For operations frequently includ-
ing fast variations in engine speed and load, such as front end loading, the effects of transients
are considerable and have to be included. Furthermore, a transient test cycle, ‘nonroad tran-
sient cycle’, is under consideration for use in forthcoming emission regulations, stage III B in
Europe (EU, 2002) and tier 4 in USA (US EPA, 2003). A transient test cycle includes sections
of acceleration as well as steady-state conditions and emissions are measured continuously
over the whole transient test cycle.
Emission values published for different diesel engines show rather wide variations, especially
when comparing a non-regulated stage 0 engine with a new engine that comply with the stage
2 emission regulation. However, the engine load patterns presented within this study showed
that the engine load pattern were rather independent of vehicle used, a harrowing operation
with the Valtra 6650 and the Case IH MX tractor were almost the same in terms of engine
load pattern. The engine load for a specific operation varied slightly depending on the size of
the implement compared to the engine power available.
The engine load patterns presented in this study can be used to calculate operations specific
fuel consumption and emission data also for vehicles with engines and fuels other than these
95
in the original measurements. In order to obtain operations specific fuel consumption and
emission data for a new vehicle just multiply the individual mode of the ISO 8178 test for the
new vehicle with the corresponding weighting factors for the operation in question.
Table 5 shows an example of how the fuel consumption data measured according to the stan-
dardised ISO 8178 regulation can be adjusted to better represent the operation in question,
harrowing. Operation specific emission data for other vehicles or fuels than these in the origi-
nal measurements can be obtained with the same technique.
Table 5. Calculation of operation specific fuel consumption and emission data from an ISO
8178 test and operation specific engine load data.
Mode Fuel consumption Weighting factor (%) Fuel consumption (kg/h)
kg/h ISO 8178 Harrowing ISO 8178 Harrowing
1 17.7 15 56 2.7 9.9
2 13.6 15 1 2.0 0.1
3 10.0 15 0 1.5 0
4 4.6 10 0 0.5 0
5 14.1 10 23 1.4 3.2
6 10.5 10 16 1.1 1.7
7 7.4 10 2 0.7 0.2
8 0.8 15 2 0.1 0
Sum 100 100 10.0 15.1
References
Cornetti, G., Klein, K., Fränkle, G. J. and Stein, H. J., 1988. US transient cycle versus ECE R49
13-mode cycle. Society of Automotive Engineers (SAE), Paper 880715.
EEC, 2000. Uniform provisions concerning the approval of compression-ignition (C.I.) and natu-
ral gas (NG) engines as well as positive-ignition (P.I.) engines fuelled with liquefied petro-
leum gas (LPG) and vehicles equipped with C.I. and NG engines and P.I. engines fuelled
with LPG, with regard to the emissions of pollutants by the engine. Regulation NO. 49, Re-
vision 3, E/ECE/324, The European Union.
EU, 2000. Directive 2000/25/EC of the European Parliament and of the Council of 22 May 2000
on action to be taken against the emission of gaseous and particulate pollutants by engines
intended to power agricultural or forestry tractors and amending Council Directive
74/150/EEC. European Union Official Journal L 173, pp 1-34.
EU, 2002. COM(2002) 765 final 27.12.2002, Proposal for a Directive of the European Parliament
and of the Council amending Directive 97/68/EC on the approximation of the Laws of the
Member States relating to measures against the Emission of Gaseous and Particulate Pollut-
ants from Internal Combustion Engines to be Installed in Non-Road Mobile Machinery. The
European Union.
Hansson, P-A., Norén, O. and Bohm, M., 1999. Effects of specific operational weighting factors
on standardized measurements of tractor engine performance. Journal of Agricultural
Engineering Research 74, 347-353.
96
Hansson, P-A., Burström, A., Norén, O. and Bohm, M., 1998. Bestämning av motoremissioner
från arbetsmaskiner inom jord och skogsbruk [Engine emissions from agricultural tractors
and forestry machines]. Department of Agricultural Engineering, Uppsala, Sweden, Report
232.
Hansson, P-A., Lindgren, M., Nordin, M. and Pettersson, O., 2003. A methodology for measuring
the effects of transient loads on the fuel efficiency of agricultural tractors. Applied Engi-
neering in Agriculture, 19(3), 94-100.
Hansson, P-A., Lindgren, M. and Norén, O., 2001 A Comparison between different methods of
calculating average engine emissions for agricultural tractors. Journal of Agricultural Engi-
neering Research, 80(1), 37-43, doi: 10.1006/jaer.2001.0710.
Hansson, P-A., and Mattsson, B., 1999. Influence of derived operation-specific tractor emission
data on results from an LCI on wheat production. The International Journal of Life Cycle
Assessment, 1999. 4(4), 202-206.
Hillel, D., 1982. Introduction to soil physics. Academic press limited, London ,UK.
ISO, 1996. ISO 8178-4; Reciprocating internal combustion engines- Exhaust emission measure-
ments- Part 4: Test cycles for different engine applications. International Organisation of
Standardisation.
Jahns, G., 1983. A method of describing diesel engine performance maps. ASAE Paper No 83-
103, St Joseph, Mich.: ASAE.
Lindgren, M., Pettersson, O., Hansson, P. and Norén, O., 2003. Jordbruks- och
anläggningsmaskiners motorbelastning och avgasemissioner samt metoder att minska
bränsleförbrukning och avgasemissioner [Engine load pattern and engine exhaust gas
emissions from off-road vehicles and methods to reduce fuel consumption and engine
exhaust gas emissions]. Report Agriculture and Industry 308, Swedish Institute of
Agricultural and Environmental Engineering.
Pettersson, O., Norén, O., Hansson, P-A. and Lindgren, M., 2002. A system for onboard determi-
nation of engine powerby measuring fuel consumption at 1 Hz. EurAgEng Budape2002.
Paper Number 02-PM-024.
Renius, K. T., 1994. Trends in Tractor Design with Particular Reference to Europe. Journal of
Agricultural Engineering Research, 57(1), 3-22.
Rinaldi, M. and Näf, E., 1992. Motorauslastung von Landwirtschaftstraktoren (Engine loads on
agricultural tractors. FAT-berichte No. 426. Tänikon, Switzerland.
Treiber, P. and Sauerteig, J., 1991. Present and future European exhaust emission regulations for
off-road diesel engines. Society of Automotive Engineers (SAE), Paper 911808.
US EPA, 2003. Notice of proposed Rulemaking. Code of Federal Regulation 40 Part 69, 80, 89,
1039, 1065, and 1068. Control of Emissions of Air Pollution from Nonroad Diesel Engines
and Fuel. USA Environmental Protection Agency.
Wetterberg, C., Holmgren, K., Rosell, L., Johansson, L. and Magnusson, R., 2002. The influence
of the fuel on emissions from diesel engines in large off-road machines. Report AL90B
2002:16295, Swedish National Road Administration, Borlänge, Sweden.
97
An LC inventory based on representative and coherent farm types
Randi Dalgaard1, Niels Halberg1, Ib Sillebak Kristensen1 and Inger Larsen2
1
Danish Institute of Agricultural Sciences, Department of Agroecology, Research Centre Foulum, P.O.
Box 50, DK-8830 Tjele. Tel: +45 8999 1254 - Fax: +45 8999 1200. E-mail:
[email protected]. 2Danish Research Institute of Food Economics
Abstract
There is a need for valid and representative data regarding the production, resource use and
emissions from typical farming systems in Denmark for analysis of the environmental impact
of different systems and as input to product oriented analyses such as Life Cycle Assess-
ments of basic food items. An inventory of 31 farm types was constructed on the basis of
2138 farm accounts from 1999 selected and weighted to be representative for the Danish
farming sector. The farm accounts were grouped according to the major soil types, the num-
ber of standard working hours, the most important enterprise (dairy, pig, different cash crops)
and the stocking rate (livestock units per hectare). For each group the account data on the
average inputs and outputs, land use and herd structure was used to establish a farm type
model with coherency between livestock production, total feed use, land use, yields, im-
ported feed, home-grown feed, manure production, fertiliser use and crop production. The
set of farm types were scaled up to national level thus representing the whole Danish agricul-
tural sector for the included products. The sum of area and yield by crop, number and pro-
duction by livestock type and the use of fertiliser, energy and concentrated feed was checked
against national level statistics and corrected accordingly across all farm types. Resource
use and emissions in each farm type was established using standard nutrient concentrations
and models for nutrient cycling, energy use and emissions of e.g. ammonia, nitrous oxides
and methane. For LCA the product oriented inventory was established using system expan-
sion rather than allocations to account for the secondary enterprises in the livestock farm
types. Data are made available on a web-based database and may be used for analyses of
the primary production systems or as input for LCA across the whole production chain.
Background
For most products the primary agricultural food production is an important determinant of the
total resource use and environmental impact, which is why life cycle assessment (LCA) of
food products must carefully address the question of data quality for agricultural production.
Many existing Inventories for LCA of agricultural products have used case studies, based on
recordings on a limited number of farms. However, there is a large variation in the resource
use and environmental impact between farms with the same main enterprise (Halberg, 1999;
Weidema et al., 2002). Thus, an LCA that aims at a more general validity must be based on a
larger sample of farm data being representative for the systems in question (average or mar-
ginal depending on the purpose of the LCA) and preferably be checked against statistical in-
formation from the level the sample will represent (e.g. regional or national).
98
This paper present an LCI which is based on representative farm accounts and is used to
model the input and production of typical farms using a method that allows to check that the
models are consistent with higher level statistical information following ideas described by
Halberg et al. (2000).
Objective
The objectives of this paper is:
• To present a method for establishing LCI for important farm types based on represen-
tative data for the Danish agricultural sector and farm models.
• To give examples of LCI data and discuss problems and advantages in using represen-
tative statistical farm data for LCI.
Methods
All Danish farms are obliged to keep detailed records of purchases and sales for tax purposes
and the yearly accounts are made with professional help. A representative data set of these ac-
counts, 2138, are reported by the advisors to the Danish Research Institute of Food Econom-
ics (DRIFE) and constitute the basic empirical data input to the model of representative farm
types presented here. The accounts include besides economical data, technical data on the
land use, livestock numbers in different groups and cash crop yields including cereals. The
representative data set was based on farm accounts from 1999, sampled as to present the total
Danish agricultural sector of the main livestock and crop production. Thus, each farm account
is given a weight to allow for division into sub-populations/groups and for scaling up the
sample to national level (Larsen, 2003).
The accounts in the data set were divided into 31 groups. Each group contained from 5 to 185
accounts and represented one of the 31 farm types according to soil type (loamy vs. sandy),
main enterprise (dairy, beef, pig, poultry and different cash crop types), organic vs. conven-
tional and animal density (e.g. livestock units per ha). For each farm type a detailed model
was established partly based directly on the average accounts data within each group and
partly on general knowledge as explained in the following: Step 1: Modelling coherent farm
types which have a realistic balance between crop and livestock production, use of inputs and
sale of products. Step 2: Modelling the emissions (CH4, CO2, NO3, HPO3, NH3 and N2O)
from the individual farm types.
Step 1. Modelling farm types
The average partition of land with different crops and the number of livestock in each group
was used to establish the production of each farm type. The accounts also gave information on
crop yields and amounts of cash crops (e.g. cereals, rape seed, potatoes, grass seeds) and live-
stock products sold (milk, meat, live animals). This information was thus used to establish
the level of production within each type and the general crop-livestock interaction (e.g. how
much grassland was used for cattle). However, because the use of external inputs like pur-
chased feed and fertiliser was only available in monetary units the exact feed and fertiliser use
99
was modelled using standards. Due to the public regulation of manure and fertiliser use in
Denmark representative average values for feed efficiency in livestock production (e.g. feed
use per kg live weight pig) and the production of Nitrogen (N) and Phosphorus (P) in manure
by livestock types is well established (Poulsen et al., 2001). Moreover, each farm has a fertil-
iser quota based on official crop-N norms deducted the plant available manure-N produced or
imported. Thus, the fertiliser use on the farm types was calculated using these norms. As part
of Danish compliance with the Nitrates Directive the use of manure-N is limited (e.g. 140 kg
manure N per ha on pig farms) why some farms are obliged to export manure to cash crop
farms. This was modelled as transfer of manure from farm types with high stocking rate to o-
ther types, which then reduced the fertiliser input accordingly.
This way a coherent model of crop-livestock interactions was established for each farm type
with a consistent relation between livestock production, use of home-grown vs. imported feed
and export of cash crops. Energy use for traction was modelled following Dalgaard et al.
(2001) where each crop is assigned a number of field operations multiplied by diesel use per
ha. Electricity use was estimated directly from the accounts. The total land use and yields of
each crop, the number of livestock, imported feed and fertiliser etc. across all farm types were
then checked against national level statistical information (Agricultural Statistics, 2000) to
make sure that the typology as whole was consistent and representative for Danish agricul-
tural sector. As shown in table 1 the data set based on farm accounts is in good agreement
with the Danish national statistics (Agricultural Statistics, 2000) for land use and for pig and
milk production. The total area and yield of major cash crops (not shown) also fits to national
statistics.
Table 1. Selected data from the typology of farm models scaled up to national level and com-
pared with the Danish national statistics (Agricultural Statistics, 2000)
Typology of farm types Danish national statistics Deviation
from nat. stat.
Slaughtering pigs 1produced, 1000 20639 20801 -1%
Sows (yearly basis) 1000 1083 1052 3%
Milking cows2, 1000 633 661 -4%
Milk production, 1000 tons 4624 4455 4%
Agricultural area, 1000 ha 2585 2644 -2%
Area with cereals, 1000 ha 1395 1448 -4%
Area with roughage, 1000 ha 567 570 -1%
Fertiliser N, 1000 tons N 226 252 -10%
Soybean meal, 1000 tons N 142 156 -9%
Grain, 1000 tons 6571 6728 -6%
Diesel and fuel, PJ 13 14 -18%
1
Living weight = 100 kg
2
Milking cows
100
The typology of farm models did, however, not account satisfactorily for the total use of fer-
tiliser. Therefore, the farm models were adjusted using some of the slack in the determination
of individual fertiliser quotas per farm and finally the still unexplained difference was cor-
rected using an overall factor on the input to all farm types. The model also underestimated
the total use of diesel and fuel by 20%, and therefore the farm models were adjusted accord-
ingly.
Use of medicine is not considered and pesticide use was not included in the first version. Re-
source use and emissions related to the construction and maintenance of buildings and ma-
chinery used on the farm was not included.
Step 2. Modelling emissions
The emissions of gasses and other substances relevant for LCA impact categories were calcu-
lated based on the established resource use and production including land use and herd struc-
ture. The emissions of green house gasses were calculated using standard IPCC methodology
for methane production from livestock and nitrous oxide production from soils and all rele-
vant manure and fertiliser compartments (IPCC, 1997; 2000). Following the TIER 2 of the
IPCC principles specific data for Danish crops and manure handling were used. The CO2
emission was calculated from the use of fossil fuel for traction and stables. Emissions related
to the production of farm inputs like fertiliser and soybean meal, which happen outside the
farm may be included in a second step and have been established as separate processes in the
LCI database (Nielsen et al., 2003).
Emissions of nitrate for the eutrophication/nutrient enrichment impact category was assumed
to be equal to the farm gate balance minus ammonia losses, denitrification (Kristensen et al.,
2003) and net change in soil N status. The ammonia emission from stables, manure storage
and handling was calculated using standard values from Andersen et al. (1999). Denitrifica-
tion was estimated using the method of Winter (2003), and net change in soil N status was
modelled using the method of Petersen et al. (2002).
Table 2 shows the aggregated emissions over all farm types compared with national statistics
for emissions of green house gasses (Gyldenkærne et al., 2004) and ammonia (Andersen et
al., 2001). The difference in nitrous oxide emission was expected since we used more detailed
information regarding crop residues than in the national nitrous oxide budget. The methane
emission was 10% lower and the ammonia emission was 1% lower than national statistics.
Table 2. Selected emissions from the typology of farm models scaled up to national level and
compared with the Danish national statistics (Andersen et al., 2001; Gyldenkærne et al., 2004,)
Typology of farm types Danish national statistics Deviation from nat. stat.
N2O (1000 tons) 22 20 9%
CH4 (1000 tons) 160 177 -10%
NH3 (1000 tons N) 76 77 -1%
101
Results
The resulting 31 farm type models after correction for national level consistency shows inputs
and outputs used to produce specific amounts of livestock and cash crop products with differ-
ent land use according to major enterprise and livestock density. Detailed results are presented
at an open database (Nielsen et al., 2003). Table 3 shows a part of the inputs and outputs as-
sociated with production at the different dairy farm types.
Table 3. Main characteristic, inputs and outputs associated with agricultural production at
eight different dairy farm types. Data are provided per farm per year.
Farm type 4 5 6 7 16 17 18 19
Characteristics
Soil type Loamy (clay) Sandy
Stocking rate <1.4 1.4-2.3 >2.3 Organic <1.4 1.4-2.3 >2.3 Organic
(Livestock Units/ha) farms farms
Pct. of Danish 0.9 1.7 5.3 0.2 3.8 7.9 0.7 1.4
3
farms
Number cows 55 55 82 62 48 67 76 85
Land area (ha) 99 50 44 88 81 65 48 102
Milk yield per cow
7227 7288 7053 6811 7431 7429 7125 6866
per year
Pct. of total Danish
4 7 3 1 15 43 4 9
milk production
Pct. of cows' feed
83 64 36 74 85 66 42 71
produced on farm
Inputs
Soybean meal,
tons 59 70 168 15 49 77 125 24
Spring barley, tons 0 65 177 104 0 92 211 154
Fertiliser, kg N 10689 4486 2096 0 8806 6602 3580 0
Fertiliser, kg P 1016 554 0 0 872 909 758 0
Diesel, MJ 515111 292549 326952 384807 409783 376043 336181 439502
Electricity, kWh 46190 30003 44258 39399 34929 42162 45563 55127
Outputs
Milk, tons 399 398 576 424 355 499 538 583
Bread wheat, tons 76 17 34 27 37 12 8 8
Beef meat, tons 25 15 20 16 20 21 24 18
Rape seed, tons 8 1 0 0 6 0 0 0
More than 50% of the total Danish milk was produced on the sandy soil types with low and
medium stocking rate. There are differences in farm size and the percentage of feed produced
on farm between the types. Farm types with high stocking rate produce a smaller part of the
feed on the farm and import more soybean meal compared to farm types with lower stocking
3
Percentage of Danish farms represented by the farm type
102
rate. The average organic farm is larger than the conventional farm types, has lower milk
yield per cow and crop yields per ha and produces more feed on the farm, especially based on
grass-clover leys in crop rotation with cereals. In the model the organic farm import around
20 kg N per ha in manure from conventional farms.
The resulting environmental impact per kg milk produced at farm-gate after system expansion
and displacement of cash crops is shown in table 4. Milk produced at farm types with low
stocking rates (farm type 4 and 16) shows a tendency to lower environmental impact than
milk produced at farm types with medium stocking rates (farm type 7 and 17). The farms with
high stocking rate have to export manure according to public regulation, which decreases
emissions from the farm. Land use per kg milk increase with higher stocking rate, because the
land used for imported feeds are involved.
Similar results for pig meat and major cash crops on farm level and per kg product ex farm
are presented by Nielsen et al. (2003) on the open database: www.lcafoods.dk.
Table 4. Environmental impact from production of 1 kg of milk from six different conven-
tional dairy farm types
Units Farm type Farm type Farm type Farm type Farm type Farm type
(eqv.) 4 5 6 16 17 18
Global warming g CO2 754 910 726 943 1030 998
Eutrophication g NO3 14.3 36.2 22.7 46.9 52.3 50.6
Acidification g SO2 7.6 9.6 10.1 9.0 10.0 10.9
Photochemical
smog g ethene 0.25 0.26 0.24 0.27 0.28 0.30
Land use m2year 1.18 1.36 1.48 1.31 1.38 1.57
Discussion and conclusion
The present LCI is based on realistic levels of resources used per unit of produced product
and reflects average production levels and efficiency within different farm types. The types
are all consistent in terms of crop-livestock interactions. The typology accounts for most input
and output of the Danish agricultural production including the exchange of manure between
farm types. The factors soil types and livestock density were assumed to be the primary sys-
tematic determinants of the level of resource use and emissions from farms. The farm models
allow for the calculation of emissions per kg of product using system expansion and dis-
placement as demonstrated by Nielsen et al. (2003). The process of system expansion is,
however, not straightforward and involves critical assumptions regarding marginal producers
of the avoided products. In the case of exported manure the methodological choice of using
system expansion gave a different allocation of ammonia losses than is often used when com-
paring nutrient balances and losses from farming systems (e.g. Kristensen et al., 2003). Be-
cause the exported manure only displaces an amount of N fertiliser equal to the plant avail-
103
able N content (i.e. the part of total manure-N taken up by the crop when compared to fertil-
izer in trials) the ammonia losses from spreading the manure on the importing farm is still in-
cluded in the emissions of the manure producing farm.
The basis for the established typology of farm models is a set of representative farm accounts
on the form that is used for statistical purpose including the Danish reporting to the Farm Ac-
count Data Network (FADN), which again forms part of EU agricultural statistics (Poppe et
al., 2000). Thus, this type of data will be available for most European countries, which again
could facilitate the development of more uniform methods for LCI establishment across dif-
ferent countries. Another advantage of this method is that it may be updated relatively easily
with data for the subsequent years when accounts data are available.
The major drawback of the method from the authors’ point of view is that the large variation
between farms in e.g. feed or fertiliser use efficiency due to differences in farmers’ manage-
ment skills and strategic choices regarding crop rotation and feed planning is not reflected in
differences between the farm types. This was, however, a necessary choice based on the pri-
mary purpose: To get representative and statistically valid data for an LCI to be used for com-
parison of different products and securing a valid baseline for LCA on processed food prod-
ucts. The amounts of feed and fertiliser purchased could have been modelled based on the
monetary information using standard prices per unit but that might have introduced another
bias because of differences in the actual price per unit paid (e.g. large farms that get discount
prices would in reality have used more feed or fertiliser than estimated from average prices).
One hypothesis could be that farmers in the marginal types would be more efficient than the
average farmers and thus have a lower resource use and emissions per kg product delivered.
The results show differences in resource use and emissions per kg product between farm
types, but more sensitivity analyses are needed in order to determine if these differences are
significant.
Another drawback is the relatively large number of small co-enterprises in the farm types re-
sulting from combining a large number of farm accounts with different co-enterprises (e.g.
two dairy farms growing five hectares with cash crops, one bread wheat, the other sugar beets
will result in a type growing 2.5 hectares of each). This results in a number of co-enterprises
that have to be compensated for through system expansion. A solution to this would be to
eliminate some of these co-enterprises in the modelled farm types, which however further
would detach the model from the empirical data.
The typology did not initially account for the total use of fertiliser in Danish agriculture why a
correction factor was used. While this secures consistency with national level statistics it is
not a totally satisfactory solution because the error may in fact belong to underestimation in
specific rather than all types. Fertiliser use in Danish farms is strongly regulated presently and
it was considered most realistic to adjust all farm types equally in order to fit the national sta-
tistic.
104
It can be concluded that the resulting LCI demonstrates successfully a method to establish co-
herent and representative inventories of agricultural production based on generally available
data.
Reference
Agricultural Statistics. 2000. Statistik om landbrug, gartneri og skovbrug. Danmarks Statistik.
ISBN 87-501-1195-7. 287 pp.
Andersen, J.M., Sommer, S.G., Hutchings, N.J., Kristensen, V.F. and Poulsen, H.D., 1999.
Emission af ammoniak fra landbruget - status og kilder. Ammoniakfordampning - rede-
gørelse nr. 1, Danmarks JordbrugsForskning, 63 pp.
Andersen, J.M., Poulsen, H.D., Børsting, C.F., Rom, H.B., Sommer, S.G. and Hutchings, N.J.
2001. Ammoniakemission fra landbruget siden midten af 80’erne. Fagligt rapport fra
DMU nr. 353. 47 pp.
Dalgaard, T., Halberg, N. and Porter, J.R., 2001. A model for fossil energy use in Danish ag-
riculture used to compare organic and conventional farming. Agriculture, Ecosystems &
Environment 87(1): 51-65.
Gyldenkærne, S. and Mikkelsen, M.H. 2004. Projection of Greenhouse Gas Emission from
the Agricultural sector. Research Notes from NERI No. 194, National Environmental
Research Institute. Denmark. 50 pp.
Halberg, N., 1999. Indicators of resource use and environmental impact for use in a decision
aid for Danish livestock farmers. Agriculture, Ecosystems & Environment 76 17-30.
Halberg, N., Kristensen, I.S. and Dalgaard, T., 2000. Linking data sources and models at the
levels of processes, farmtypes and regions. Agricultural data for Life Cycle Assess-
ments. Series 2.00.01, vol.1. Agricultural economics research Institute (LEI), The Ha-
gue, the Netherlands. p. 16-30. ISBN 90-5242-563-9.
IPCC, 1997. Greenhouse gas inventories. Revised 1996 IPCC Guidelines for National Green-
house Gas Inventories. Available on-line: http://www.ipcc.ch/pub/guide.htm
IPCC, 2000. IPCC Good practice Guidance and Uncertainty Management in National Green-
house Gas Inventories. Chapter 4. Agriculture. 4.1-4.83. IPCC. Available on-line:
http://www.ipcc.ch/pub/guide.htm
Kristensen, I. S., Halberg, N., Nielsen, A. H., Dalgaard, R., and Hutchings, N., 2003. N-
turnover on danish mixed dairy farms, 1-21. Paper for workshop: “Nutrient manage-
ment on farm scale: how to attain European and national policy objectives in regions
with intensive dairy farming?”. 23-25 June 2003, Quimper, France. Available on-line:
http://www.agrsci.dk/jbs/isk/DK_country_report_partII.pdf.
Larsen, I., 2003. LCA project. Method description. Available online, 1/10 2003 at
www.lcafood.dk/database.
Nielsen, P.H., Nielsen, A.M., Weidema , B.P., Dalgaard, R. and Halberg N., 2003. LCA food
database. Available online, 31/8 2003 at www.lcafood.dk/database.
105
Petersen, B.M. & Berntsen, J., 2002. Omsætning i jordpuljen på forskellige bedriftstyper.
Temadag arrangeret af Afd. for Jordbrugssystemer 24. april 2002. Forskningscenter
Foulum. Intern rapport nr. 157: 13-24.
Poppe, J.K. and Meeusen, J.G., 2000. Using a Farm Accountancy Data Network in data man-
agement for LCA. Agricultural data for Life Cycle Assessments. Series 2.00.01, vol.2.
Agricultural economics research Institute (LEI), The Hague, the Netherlands. p. 115-
127. ISBN 90-5242-563-9.
Poulsen, H.D., Børsting, C.F., Rom, H.B. and Sommer, S.G. 2001., Kvælstof, fosfor og kali-
um i husdyrgødning -normtal 2000. DJF rapport nr. 36. Husdyrbrug.
Winter, F., 2003. Empirisk model beregning af N2O-emission og denitrifikation som funktion
af jordtype og gødningsmængde. Unpublished note from Danmarks JordbrugsForsk-
ning. 10pp. (Empirical model of N2O-emission and denitrifikation as a function of soil
type and fertiliser application)
Weidema, B.P., Thodberg, L., Nielsen, A.H., Kristensen, I.S., Hermansen, J. & Hvid, S.K.,
2002. Produktorienteret miljøindsats i landbrugets primærproduktion. Vurdering af vir-
kemidler. Arbejdsrapport fra Miljøstyrelsen 19, 74 pp.
http://www.mst.dk/udgiv/publikationer/2002/87-7972-131-1/html.
106
The ecoinvent database: use for the agri-food sector
Nemecek, T.*, Erzinger, S. and Frischknecht, R.
*Swiss Federal Research Station for Agroecology and Agriculture, FAL-Reckenholz, Reckenholzstr.
191, CH-8046 Zurich, Switzerland. Phone +41 1 3777254, fax +41 1 3777201, tho-
[email protected].
Swiss Federal Research Station for Agricultural Economics and Engineering, FAT-Taenikon, CH-
8356 Ettenhausen, Switzerland.
ecoinvent Centre, c/o EMPA, Überlandstrasse 129, CH-8600 Dübendorf, Switzerland.
Abstract
Life cycle inventory (LCI) data are the basis of every LCA study and are very important for its
quality. The ecoinvent database provides over 2500 LCI datasets from the areas energy sys-
tems, transportation, waste disposal, construction, chemicals, detergents, papers and agri-
culture with reference to Switzerland or Europe. It is useful to the agri-food sector by provid-
ing highly detailed data on agricultural plant products, infrastructure, means of production
and processes and datasets from various other economic sectors having close relationships
with the food industry. The data exchange format EcoSpold allows recording, documentation
and exchange of LCI data in a standardised form.
Keywords: life cycle inventory, LCI, database, agricultural systems, agriculture, food indus-
try
Introduction
Life cycle assessment (LCA) has proved to be a powerful tool for the environmental im-
provement of production processes in the agri-food sector (e.g. Anderson 2000). However, the
increased use of the LCA method to analyse systems is hindered by the lack of agreement on
the use of methods and by the poor availability of life cycle inventory (LCI) data.
The problem of availability of LCI data has various aspects:
• the data collection process is very time-consuming,
• production data are often confidential and not available to the public,
• LCI data are available in different, often incompatible formats (standardised formats
like SPOLD are not widely used),
• data are collected for different regions and time periods, which leads to inconsisten-
cies,
• the underlying standards and quality levels of the datasets are different.
To circumvent these difficulties, databases are needed that offer consistent data for many eco-
nomic branches. Such a database has been built in the frame of the Swiss project ecoinvent
2000.
107
The ecoinvent database
The project ecoinvent 2000 was initiated by the ETH domain and Swiss Federal Offices to
promote the life cycle approach in various economic sectors and to provide a basis for the “In-
tegrated Product Policy”. The goals of the project are:
• to harmonise the LCIs of the participating institutes on the basis of common quality
guidelines and by using the common data exchange format EcoSpold (derived from
the SPOLD99-standard),
• to update the data for the year 2000 and
• to allow access to the database via the internet and subsequently through different
LCA software tools (see Jungbluth and Frischknecht 2003).
Twelve institutes have participated in the process of data acquisition and harmonisation. The
database contains LCIs from the areas energy systems, transportation, waste disposal, con-
struction, chemicals, detergents, papers and agriculture, which refer to the geographic context
of Switzerland and/or Western Europe.
The main characteristics of the ecoinvent database are:
• The documentation is assured by meta-data in the database and by extended
reports, provided on a CD-ROM.
• Geographic reference: each dataset refers to a specified country or region and to a
defined process stage (e.g. “at plant”, “at farm” or “at regional storehouse”).
• Common quality standards were used to define and to document the datasets.
• The datasets were submitted to an internal review process prior to inclusion in the
database.
• Detailed recording of resources and emissions: over 1000 resources and emissions
belonging to 20 categories were distinguished (Table 1).
• Land use was accounted for by a newly developed scheme (see Jungbluth 2003).
• Matrix calculation: a mechanism of matrix inversion allows to solve the problem of
mutual interdependencies of the datasets.
• The data exchange format ECOSPOLD is used to import and export the data and the
meta-data (see Jungbluth and Frischknecht, 2003, see also Tables 3 and 4).
• The uncertainty of each figure is described by quantitative and qualitative indicators.
The ecoinvent database is accessible on the internet since September 2003. Further informa-
tion is available on www.ecoinvent.ch.
LCI of agricultural systems in ecoinvent
Categories of emissions
The distinction of several subcategories in the ecoinvent database (Table 1) is useful for agri-
cultural applications. The pollution of an industrial soil has certainly not the same impact and
significance as the pollution of an agricultural soil which serves for the production of food. In
108
the latter case, harmful substances can enter the human food chain and be ingested with con-
sequent impacts on human health.
Table 1. Categories and subcategories of emissions and resources considered in ecoinvent.
Only subcategories that contain data in ecoinvent data v1.0 are presented.
Category SubCategory
low population density
low population density, long-term
air lower stratosphere + upper troposphere
high population density
unspecified
agricultural
forestry
soil
industrial
unspecified
ground-
ground-, long-term
lake
water
ocean
river
unspecified
in air
biotic
resource in ground
land
in water
The substances released during agricultural production are considered to be emitted to agri-
cultural soil. Emissions in air from agricultural production are considered to go into “air, low
population density”. On the other hand, the production processes for agricultural means of
production like fertilisers and pesticides are assumed to take place in urban areas and conse-
quently related emissions are counted as emissions into “air, high population density”. Low
resp. high population density refer to the conditions in Central Europe. The distinction of
these subcategories is relevant for impact categories like ozone formation or human toxicity.
All emissions to surface water from agricultural production are counted as emissions to “wa-
ter, river” (e.g. P-emissions by run-off or erosion). As only Swiss agriculture was considered,
emissions to sea water are irrelevant. Furthermore, since only a small share of the agricultural
area is located in direct proximity of lakes, it was assumed that all emissions to surface water
go into rivers. Leaching of nitrate and phosphorus are considered as emissions into ground
water.
109
Field emissions
Direct field emissions are considered by means of models that partly use situation-specific pa-
rameters:
• Ammonia emissions are calculated with models described my Menzi et al. (1997).
Constant release factors are applied in the case of mineral fertilisers. For slurry, liquid
manure and solid manure the content of NH4+, the average monthly temperature and
humidity and the quantity of manure spread per hectare are taken into consideration.
• Potential nitrate leaching is estimated on a monthly basis by accounting for N-
mineralisation in the soil and the N-uptake by the vegetation (specific to each crop). If
the mineralisation exceeds the uptake, nitrate leaching can potentially occur. In addi-
tion, a risk of nitrate leaching from fertiliser application during unfavourable periods
is calculated, by taking the crop, the month of the application and the potential rooting
depth into account (Oberholzer & Walther, 2001).
• Three paths of P-emissions to water are considered: run-off (as phosphate) and erosion
(as phosphorus) to rivers and leaching to ground water (as phosphate). The category of
land use, type of fertiliser, quantity of P spread and the characteristics and duration of
the soil cover (for erosion) are considered. The model used is derived from Prasuhn &
Grünig (2001).
• N2O-emissions are estimated by using an adapted IPCC-method (Schmid et al., 2000).
-
Indirect emissions from the conversion of NH3 and NO3 into N2O are considered in
the inventory as well.
• Heavy metal emissions are assessed by a simple input-output balance, by taking all in-
puts (fertilisers, seed, pesticides) and outputs (products and straw) to resp. from the
field caused by the farmer.
• Pesticide application are accounted for as emissions of the active ingredients in agri-
cultural soil.
Resource use
CO2 uptake by the vegetation is taken into account as well as the potential heat energy con-
tained in the biomass. The biotic flows of CO2 are clearly separated from the release of CO2
during the combustion of fossil fuels. This lets the database user the opportunity either to con-
sider biotic CO2-flows (including also the CO2-release during the usage of the agricultural
product) or not. The content of organic C in the soil is assumed as constant.
Land occupation and transformation is accounted for agricultural land as well as for non-
agricultural areas (Jungbluth 2003).
110
Available datasets for agricultural systems and their use for the agri-food sector
The ecoinvent database provides three types of datasets useful for the agri-food sector:
1. Datasets on agricultural products. Data on plant products are included like the ar-
able crop products wheat, rye, barley, silage and grain maize, potatoes, sugar and fod-
der beets, fava beans, soybeans, peas, sunflower and rape seed. For most crops data-
sets for integrated (IP, a production respecting a set of rules for environmental
protection defined by the Swiss government) and organic production are present. For
the cereals and rape seed an additional integrated extensive variant is calculated, cor-
responding to a crop without treatment by fungicides, insecticides and growth regula-
tors, receiving extra subsidies in Switzerland. Three types of hay (intensive IP, inten-
sive organic and extensive) are also included. The datasets refer to model crops, which
have been defined on the basis of statistics (like FADN, import statistics, a.o.), rec-
ommendations, pilot farm networks and surveys of seed suppliers. The datasets were
validated by a panel of experts. Table 2 shows an overview of the datasets of agricul-
tural production in ecoinvent.
2. Datasets on agricultural means of production. The database contains a number of
modules that allow the calculation of most systems in arable and fodder crop produc-
tion and animal husbandry. LCAs of special productions like vineyards, vegetable and
fruit production are partly possible. ecoinvent contains datasets on agricultural build-
ings, machinery, work and drying processes, fertilisers, pesticides, seed and animal
feed (see Table 2).
3. Background datasets for the food processing industry. Furthermore, the database
contains various modules that are required for LCA studies in the food sector: energy
systems, transports, detergents, packaging materials, construction materials and proc-
esses as well as waste management.
111
Table 2. Overview of the datasets for agricultural systems available in the ecoinvent database.
Number of
Subcategory modules Example of inventories for the subcategories
Name Location Unit
potatoes organic, at farm CH kg
Agricultural
products
Plant production 59 rape seed extensive, at farm CH kg
hay intensive IP, at farm CH kg
Animal production 1 tallow, at plant CH kg
lime, from carbonation, at regional storehouse CH kg
Mineral fertilisers 24 ammonium nitrate, as N, at regional storehouse RER kg
urea, as N, at regional storehouse RER kg
Agricultural inputs
Organic fertilisers 6 horn meal, at regional storehouse CH kg
cyclic N-compounds, at regional storehouse RER kg
Pesticides 68 [Sulfonyl]urea-compounds, at regional storehouse CH kg
pesticide unspecified, at regional storehouse CH kg
Seed 22 sugar beet seed IP, at regional storehouse CH kg
wheat organic, at fodder mill CH kg
Feed 10
wheat IP, at fodder mill CH kg
dried roughage store, air dried, solar CH kg
structure
Buildings 13
Infra-
label housing system, pig CH pig place
Machinery 6 agricultural machinery, tillage, production CH kg
loose housing system, cattle, operation CH LU
Building usage 8
Operation of in-
dried roughage store, air dried, solar, operation CH kg
frastructure
haying, by rotary tedder CH ha
Machinery and
33 tillage, ploughing CH ha
equipment usage
milking CH kg
Drying 4 grain drying, high temperature CH kg
Total 254
Examples from crop production
Table 3 shows an example of unit process inventory data and Table 4 of meta-data for the
dataset “wheat, extensive CH”. The details of each dataset together with its description are
available from the database.
112
Table 3. Example of a unit process inventory (multi-output process) in the ecoinvent database (wheat from extensive integrated production). CH
= Switzerland, RER = Europe, MA = Morocco, na = not considered.
Unit process inventory for: 1 ha wheat extensive, CH
Uncertainty information Allocation
Wheat grains Wheat straw
Input/ Location/ Uncert SD extensive, at extensive, at
output Exchanges Category Unit Value Type 95% Uncert Scores farm CH (kg) farm CH (kg)
Input ammonium nitrate, as N, at regional storehouse RER kg 5.71E+01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input urea, as N, at regional storehouse RER kg 2.01E+01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input diammonium phosphate, as N, at regional storehouse RER kg 5.10+E00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input calcium ammonium nitrate, as N, at regional storehouse RER kg 2.87E+01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input ammonium sulphate, as N, at regional storehouse RER kg 4.32E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input triple super phosphate, as P205, at regional storehouse RER kg 1.91E+01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input single super phosphate, as P205, at regional storehouse RER kg 7.71E-01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input diammonium phosphate, as P205, at regional storehouse RER kg 1.30E+01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input phosphate rock, as P205, beneficiated, dry, at plant MA kg 1.12E+01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input thomas meal, as P205, at regional storehouse RER kg 2.42E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input potassium chloride, as K20, at regional storehouse RER kg 1.56E+01 1 1.07 (2,1,1,1,1,na) 92% 8%
Input potassium sulphate, as K20, at regional storehouse RER kg 1.03E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input tillage, ploughing CH ha 1.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input tillage, harrowing, by spring tine harrow CH ha 2.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input tillage, currying, by weeder CH ha 1.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input sowing CH ha 1.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input fertilising, by broadcaster CH ha 4.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
solid manure loading and spreading, by hydraulic loader
Input CH kg 9.62E+02 1 1.07 (2,1,1,1,1,na) 92% 8%
and spreader
Input slurry spreading, by vacuum tanker CH m3 8.25E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input application of plant protection products, by field sprayer CH ha 1.00E+00 1 1.13 (2,2,3,1,1,na) 92% 8%
Input combine harvesting CH ha 1.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input transport, tractor and trailer CH tkm 1.39E-02 1 1.07 (2,1,1,1,1,na) 100%
Input grain drying, low temperature CH kg 6.39E+01 1 1.07 (2,1,1,1,1,na) 100%
Input baling CH unit 4.70E+00 1 1.07 (2,1,1,1,1,na) 100%
Input loading bales CH unit 2.05E+01 1 1.07 (2,1,1,1,1,na) 100%
Input tillage, cultivating, chiselling CH ha 1.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8%
Input nitrile-compounds, at regional storehouse CH kg 2.00E-01 1 1.13 (2,2,3,1,1,na) 92% 8%
Input phenoxy-compounds, at regional storehouse CH kg 3.50E-01 1 1.13 (2,2,3,1,1,na) 92% 8%
113
Input (sulfonyl)lurea-compounds, at regional storehouse CH kg 7.20E-01 1 1.13 (2,2,3,1,1,na) 92% 8%
Input wheat seed IP, at regional storehouse CH kg 1.80E+02 1 1.07 (2,1,1,1,1,na) 92% 8%
Input transport, freight, rail CH tkm 5.87E+01 1 2.71 (4,5,na,na,na,na) 92% 8%
Input transport, lorry 28t CH tkm 4.68E+01 1 2.71 (4,5,na,na,na,na) 92% 8%
Input transport, van <3,5t CH tkm 2.74E+00 1 2.71 (4,5,na,na,na,na) 92% 8%
Input transport, lorry 40t CH tkm 7.00E+00 1 2.71 (4,5,na,na,na,na) 92% 8%
Input transport, barge RER tkm 4.66E+02 1 2.71 (4,5,na,na,na,na) 92% 8%
Input Occupation, arable, non-irigated resource/land m2a 7.94E+03 1 1.77 (2,1,1,1,1,na) 92% 8%
Input Transformation, from pasture and meadow, intensive resource/land m2 2.90E+03 1 2.67 (2,1,1,1,1,na) 92% 8%
Input Transformation, from arable, non-irigated resource/land m2 7.10E+03 1 2.67 (2,1,1,1,1,na) 92% 8%
Input Transformation, to arable, non-irigated resource/land m2 1.00E+04 1 2.67 (2,1,1,1,1,na) 92% 8%
Input Carbon dioxide, in air resource/in air kg 1.16E+04 1 1.07 (2,2,1,1,1,na) 61% 39%
Input Energy, gross calorific value, in biomass resource/biotic MJ 1.40E+05 1 1.07 (2,2,1,1,1,na) 59% 41%
air/low population
Output Dinitrogen monoxide ykg 4.30E+00 1 1.61 (2,2,1,1,1,na) 92% 8%
density
air/low population
Output Nitrogen oxides ykg 9.03E-01 1 1.61 (2,2,1,1,1,na) 92% 8%
density
air/low population
Output Ammonia ykg 1.06E+01 1 1.30 (2,2,1,1,1,na) 92% 8%
density
Output Nitrate water/ground- kg 1.73E+02 1 1.77 (2,2,1,1,1,na) 92% 8%
Output Phosphate water/ground- kg 1.85E-01 1 1.77 (2,2,1,1,1,na) 92% 8%
Output Phosporus water/river kg 2.58E-01 1 1.77 (2,2,1,1,1,na) 92% 8%
Output Phosphate water/river kg 5.64E-01 1 1.77 (2,2,1,1,1,na) 92% 8%
Output Cadmium soil/agricultural kg 2.72E-03 1 1.77 (2,2,1,1,1,na) 42% 58%
Output Copper soil/agricultural kg -1.82E-02 1 1.77 (2,2,1,1,1,na) 55% 45%
Output Lead soil/agricultural kg -9.56E-03 1 1.77 (2,2,1,1,1,na) 5% 95%
Output Zinc soil/agricultural kg -1.39E-01 1 1.77 (2,2,1,1,1,na) 79% 21%
Output Nickel soil/agricultural kg 6.77E-03 1 1.77 (2,2,1,1,1,na) 47% 53%
Output Chromium soil/agricultural kg 5.52E-02 1 1.77 (2,2,1,1,1,na) 46% 54%
Output Mercury soil/agricultural kg -1.39E-04 1 1.77 (2,2,1,1,1,na) 18% 82%
Output Difenoconazole soil/agricultural kg 1.80E-02 1 1.32 (2,2,3,1,1,na) 92% 8%
Output Ioxynil soil/agricultural kg 2.00E-01 1 1.32 (2,2,3,1,1,na) 92% 8%
Output Isoproturon soil/agricultural kg 7.20E-01 1 1.32 (2,2,3,1,1,na) 92% 8%
Output Mecoprop-P soil/agricultural kg 3.50E-01 1 1.32 (2,2,3,1,1,na) 92% 8%
Output wheat grains extensive, at farm CH kg 5.37E+03 100%
Output wheat straw extensive, at farm CH kg 3.27E+03 100%
114
Table 4. Example of meta-data for “wheat extensive, CH” (extract).
Type Field name Field contents
ReferenceFunction Name Wheat extensive
Geography Location CH
InfrastructureProcess 0 (=no)
Unit Ha
Amount 1
The inventory includes the processes of soil cultiva-
tion, sowing, weed control, fertilisation, pest and
pathogen control, harvest and drying of the grains.
Machine infrastructure and a seed for machine shel-
tering is included. Inputs of fertilisers, pesticides and
ReferenceFunction IncludedProcesses seed as well as their transports to the farm are con-
sidered. The direct emissions on the field are also in-
cluded. The system boundary is at the farm gate.
LocalName Weizen Extenso
Inventory refers to the production of 1 kg wheat ex-
tensive grains respectively straw, both with a moisture
content of 15%. The multioutput-process ‘wheat ex-
tensive’ delivers the co-products ‘wheat grains exten-
GeneralComment sive, at farm’ and ‘wheat straw extensive, at farm’.
Economic allocation with allocation factor of 92.5% to
grains (exceptions see report).
Category agricultural production
SubCategory plant production
StartDate 1996
EndDate 1999
TimePeriod DataValidFor
EntirePeriod 1 (=yes)
OtherPeriodText The yield data have been collected for the years
1996-1999.
Geography Text Refers to an average production in the Swiss low-
lands.
Integrated production with extensive plant protection
Technology Text (no fungicides, insecticides and growth regulators)
Percent 86
ProductionVolume CH prod of wheat: 561000t in 2000. % refers to fract
of tot area in lowlands
Representativeness Data were compiled from statistics, pilot network, fer-
tilising recommendations, documents from extension
services, information provided by retailers and expert
SamplingProcedure knowledge. The production data were verified and ad-
justed by a group of experts.
115
Figure 1 shows an example of calculated results for several arable crops. The cumulative energy
demand (CED) is determined by the use of machines (mechanisation), the application of mineral
fertilisers and by grain drying. For mechanisation the processes of soil cultivation/seedbed prepa-
ration and harvest are most relevant. The production of mineral fertilisers – especially nitrogen
fertilisers – is highly demanding on energy resources. Mineral fertilisers have the highest share
in cereals and oil crops. Drying is very important in grain maize, followed by oil crops, grain
legumes and cereals. This shows the importance of harvesting in dry conditions. Pesticide pro-
duction has only little relevance for energy use.
Arable crop products IP, at farm: cumulative energy demand, non-renewable energy resources,
total [MJ-Eq/kg DM]
10
9 Seed
8 Green manure
7 Pesticides production
6
K-fertilisers
5
P-fertilisers
4
3 N-fertilisers
2 Transports of inputs
1
Drying
0
Mechanisation and fuel
rye grains
soy beans
barley grains
sunflower
wheat grains
fava beans
rape seed
potatoes
sugar beets
fodder beets
protein peas
grain maize
silage maize
consumption
Figure 1. Cumulative energy demand of non-renewable energy resources for agricultural crops
from integrated production per kg dry matter (DM).
Discussion and conclusion
The ecoinvent database provides datasets that allow the calculation of LCA for agricultural pro-
duction and food processing. The database offers not only data on the agricultural sector, but also
on the most important economic sectors. It is detailed and accompanied by an extensive docu-
mentation.
The database contains also minima and maxima (confidence limits) for each figure. This allows
to estimate the uncertainty of the information. However, these indications are rather rough and
116
would profit from methodical developments in the future. In addition, the parameters are varied
independently in the Monte-Carlo simulations, although they are not independent in reality.
Further developments of the database can be the extension to other economic sectors not yet
covered and to countries outside Europe. For agriculture in particular, the database would profit
from including animal products and datasets from other European countries.
The ecoinvent contains a large number of inventory datasets containing detailed data on re-
sources and emissions for 20 environmental subcategories. Most current LCIA methods do not
make use of this information. We encourage scientists who are developing LCIA methods to
consider the structure now available and to refine their characterisations factors for the different
subcategories. This will contribute to an increased reliability and credibility of LCA.
The EcoSpold format allows to exchange unit-process inventory data, LCI results and also LCIA
methods and results in the same format. Since an EcoSpold-interface is implemented in the most
important LCA-software tools, this is a great progress towards better data exchange and more ef-
ficient LCA studies. It is expected that ecoinvent should boost the application of the LCA
method in the agri-food sector.
References
Andersson, K., 2000. LCA of food products and production systems. Int. J. LCA, 5 (4): 239-248.
Jungbluth, N., 2003. Land occupation and transformation in life cycle inventories. Paper presented at
the 4th International Conference on: Life Cycle Assessment in the Agri-food sector, Horsens,
Denmark, October 6-8, 2003.
Jungbluth, N. and Frischknecht R., 2003. Life cycle inventory modelling in the Swiss ecoinvent da-
tabase. Poster presented at the 4th International Conference on: Life Cycle Assessment in the
Agri-food sector, Horsens, Denmark, October 6-8, 2003.
Menzi, H., Frick, R. & Kaufmann, R., 1997. Ammoniak-Emissionen in der Schweiz: Ausmass und
technische Beurteilung des Reduktionspotentials. Schriftenreihe der FAL 26, Zürich, 124 p.
Nemecek, T., Heil, A., Huguenin, O., Meier, S., Erzinger, S., Blaser, S., Dux, D. and Zimmermann
A., 2003. Life Cycle Inventories of Agricultural Production Systems. Final report ecoinvent
2000 No. 15. FAL Reckenholz, FAT Tänikon, Swiss Centre for Life Cycle Inventories,
Duebendorf, CH, Online-Version under: www.ecoinvent.ch.
Oberholzer, H.R. & Walther, U. (2001) Modell zur Abschätzung des Nitratauswaschungsspotenzials.
In: Neue Erkenntnisse zu Stickstoffflüssen im Ackerbau, FAL-Tagung 6.4.2001, 6p.
Prasuhn V. & Grünig K. (2001) Evaluation der Ökomassnahmen Phosphorbelastung der Oberflä-
chengewässer durch Bodenerosion. Schriftenreihe der FAL 37, 151p.
Schmid, M., Neftel, A. & Fuhrer, J., 2000. Lachgasemissionen aus der Schweizer Landwirtschaft.
FAL, Schriftenreihe der FAL 33, 131 p.
117
Life Cycle Inventory of the Galician dairy sector
Hospido, A.*, Moreira, M.T. and Feijoo, G.
Department of Chemical Engineering. University of Santiago de Compostela. 15782 – Santiago de Com-
postela (Spain). T: +34981563100, ext.16020 // F: +34981547168 // e-mail: [email protected]
Abstract
Dairy industry has been extensively studied from Life Cycle Assessment (LCA) in many Euro-
pean countries, nevertheless, concerning LCA in Spain, little work has been reported, and a
global and reliable inventory is still lacking. In this work particular attention has been paid to the
Life Cycle Inventory (LCI) of this sector by performing an exhausted fieldwork considering the
contribution of several subsystems: farms, fodder factories, dairies and the manufacture of
tetrabrik containers. For each subsystem, average data as well as associated standard devia-
tions are presented. The majority of these data suffer from a high variability, which means that
representative production and processing scheme can be difficult to establish.
Key words: life cycle inventory, dairy, farm, milk, and fodder
Background
In April 2000, the European Commission published an extensive report concerning the environ-
mental impact of the dairy production in the European Union (CEAS Consultants Ltd, 2000),
where a classification of countries according to herd size was presented. Spain was included on
the third group characterized by a huge number of many small farms. Most of the LCA studies
on milk production were carried out for countries belonging to the second one (Berlin, 2002;
Cederberg and Mattson, 2000; Haas et al., 2001; Høgaas, 2002), but in the case of the third
group, a reliable inventory for Life Cycle Analysis was still lacking.
In Galicia, a Spanish region with an important milk production, statistical data indicate that 75%
of its farms have less than 10 cows, while the remaining 25% have encountered an industrialisa-
tion and modernisation process with a superior surface and herd size, which correspond to a
renovated picture of the rural area.
Moreover, since mid-80´s dairy facilities in Galicia have suffered an important recession, a high
number of farms have been obliged to close: 1,159 on year 2,001 and 1,435 on year 2,002, this
rising tendency being observed during last years. In fact, the number of farms in Galicia has de-
creased from 109,284 farms in 1,982 to 24,910 at the end of 2,002. The earliest farms to undergo
these consequences are the smaller ones because they cannot be competitive so the farms that
have a significant contribution in the productive framework are the bigger ones. For this reason,
three of them were selected for this study.
118
Milk processing is a more uniform process as the technology and facilities are common for the
majority of the dairies. Three dairies, which an important market quota, have been chosen to be
inventoried.
Method
Goal and Scope Definition
The objective of this study is to examine the total life cycle of production and processing of milk
in order to quantify the potential environmental impact. Three dairies and three farms have been
selected as representative of Galician milk industry to define both production and processing
scenarios. Additionally, other relevant subsystems have been identified and studied separately:
fodder production (also called concentrate mixture) and packaging, the former by considering
two important fodder factories and the latter by public report data. Our main focus of attention is
LCI, the second phase at LCA methodology. This stage requires the collection of extensive data
on the physical inputs and outputs of the processes and related procedures under evaluation. As it
was mentioned before, these data were compiled on a case-by-case basis bearing in mind the im-
portance of data with high quality and reliability.
Functional unit
The functional unit (FU) selected is 1 litre of packaged liquid milk, ready to be delivered, which
corresponds to the standard and more widespread final presentation of milk at Galician and
Spanish homes.
System boundaries
The life cycle of milk production included in this analysis is shown in Figure 1, where the fol-
lowing hierarchy has been established:
• First level: Main stages of global process.
• Second level: Inputs that have an industrial process associated and have been analysed in
detail.
• Third level: Inputs taken from LCA databases.
Apart from the main product, other outputs of the system such as co-products, waste and emis-
sions to water, air or soil, are also included in the inventory. The most often omitted subsystem is
the consumer phase and it has not been considered in the present study.
Allocation rules
During the performance of LCA, allocation problems arise when the life cycle of different prod-
ucts are connected. In other works (Høgaas, 2002), a cow is defined as a multifunctional produc-
tion system, supplying several products: milk, meat, skin and manure. The allocation rule applied
by Høgaas for the two main products (milk and meat) was based on the biological demand of
119
fodder while the other two products were not considered: the manure was used as a fertiliser at
the farm and it was not looked upon as a product, and the skin of the cows was omitted due to its
much lower importance (economic and mass). Although the farms evaluated were specialised in
milk production, the study of economic revenue per cow based on historical Galician market
prices from years 2000-2002 for milk and meat has entailed the following distribution of the as-
sociated economical benefits: 87% for milk and 13% for meat. This figure is important and
therefore economic allocation was used for raw milk production in our study.
FARM SUBSYSTEM
SILAGE MAIZE
ALFALFA CLEANING
ELEMENTS
MILK PRODUCTION
FODDER ELECTRICITY
(FARM)
DIESEL IMPACTS
ASSOCIATED
TRANSPORT
FUEL ELECTRICITY
CLEANING OTHER
ELEMENTS PACKAGING
MILK PROCESSING ELEMENTS
(DAIRY)
STERILISATION
ELEMENTS TETRA BRIK
IMPACTS
ASSOCIATED
DAIRY SUBSYSTEM
Figure 1. Schematic flow chart of the life cycle of milk. The main stages of the process are rep-
resented in blocks; Inputs associated with industrial processes in triangles; Inputs from LCA da-
tabases in continuous circles; impacts associated in discontinuous circles.
At the milk processing stage, the dairies chosen are basically mono-functional and the cream ob-
tained as a co-product represents less than 2.5% of the total annual production, consequently, al-
location rules considering cream were not applied. During fodder production and tetrabrik manu-
facture process, mass criteria have been chosen depending on the distribution of final products.
120
Data Collection
To assess the most accurate environmental impacts associated to Galician milk production, we
mainly considered data from Galician industries. Real data from farms, fodder factories and dair-
ies were collected in consecutive periods during the last three years. In particular, elements com-
ing from the agricultural subsystem (maize, alfalfa and silage) have been quantified and data cor-
responding to their harvest belong to a comprehensive study made in Catalonia (Milá et al.,
1998).
There are other types of data whose production systems are not present in Galicia and they have
been obtained from companies from other regions. For instance, tetrabrik containers are only
manufactured in a factory located in Arganda del Rey (Madrid) and its inventory is obtained
from several reports (Tetrapack Spain, 1999, 2000 and 2001). In relation to the electricity pro-
duction profile, an electrical percentage distribution according to data from the Institute for Di-
versification and Energy Saving (IDAE, 2002) has been used: 35.8% of the electricity is pro-
duced from coal, 27.6% is nuclear, 13.9% is hydroelectric, 9.9% is obtained from oil power
plants, 9.7% from gas power plants, 2.2% from wind power plants, 0.6% from waste use and
0.3% from biomass use. However, due to the non-availability of data quantifying the environ-
mental burdens associated to the different ways to produce electricity in Spain, we chose data
from the database BUWAL 250 (SAEFL, 1998).
Results
LCI at Farms
Three well-managed farms were inventoried (Table 1). Their renewed facilities, consisting of an
automatic milking system with recollection pipes and a storage tank, are a good example of the
modernisation philosophy that is in practice nowadays.
Table 1. General data about the farms inventoried.
Farm A Farm B Farm C
Portomarin Carballedo Sarria
Location
(Lugo) (Lugo) (Lugo)
Operative since 1997 1996 1994
228,286 353,725 427,050
Annual Production (L)
(year 2000) (year 2001) (year 2002)
Average Size (animals) 50 60 67
Dairy cows 23 38 50
Suckler cows 27 22 17
Average yield (L/cow/year) 9,925 9,309 8,541
To obtain the annual milk production for each farm, a daily measurement took place. Then, this
yearly production is handled to obtain the inventory data on the basis of the chosen functional
unit considering annual consumption. Figure 2 shows bimonthly electricity spending form Farms
B and C that were measured to get the annual consumption.
121
4000
ELECTRICITY CONSUMPTION (kWh)
3750
3500
3250
3000
2750
2500
Jan-Feb Mar-Apr May-Jun Jul-Aug Sep-Oct Nov-Dec
FARM B FARM C
Figure 2. Bimonthly variation of electricity consumption at Farms B and C.
Table 2 shows the results of the fieldwork developed there; the average values as well as the
standard deviations are presented. Data have been mainly quantified according to the annual con-
sumption of each element and the opinion of the farmers.
Table 2. Inventory Data of the Farm (FU = 1 L of raw milk).
INPUTS
From the TECHNOSPHERE From the NATURE
Materials and fuels Raw materials
1. Food 1. Water 4.20 ± 0.25 L
1a. Maize 482.9 ± 191.1 g
1b. Fodder 424.0 ± 36.7 g
1c. Silage 551.6 ± 208.1 g
1d. Alfalfa 145.0 ± 88.7 g
2. Alkaline detergent 0.78 ± 0.06 mL
3. Acid solution 0.11 ± 0.06 mL
4. Disinfectant 0.83 ± 0.62 mL
5. Paper 0.418 g
6. Linchpin 9.00·10-5 units
7. Diesel 4.25 ± 0.10 mL
50.25 ± 12.96 Wh
Electricity
1. Electricity
OUTPUTS
To the TECHNOSPHERE To the NATURE
13.01 ± 0.99 g
Products and co-products Emissions to air
1. Raw milk 1L 1. Methane *
Emissions to water
Waste for treatment 1. Wastewater ** 1.30 ± 0.10 L
1. Urban Solid Waste 0.28 g COD 8.22 g L-1
TSS 2.70 g L-1
* According to available data (EPA, 2002), an adult cow emits 120 kg of methane per year due to enteric fermentation.
** COD = Chemical Organic Demand // TSS = Total Suspended Solids
122
Regarding food components, maize and silage are produced on each farm (internal elements)
whereas fodder and alfalfa are considered external elements that have to be exported from out-
side.
LCI at Fodder Factories
Maize, fodder, silage and alfalfa are the main components in animal food, and their final mixture
has to attain acceptable levels of protein and energy supplement (between 15.5-18.5% protein
and 1.70-1.72 Mcal kg-1 of dry material).
Fodder plays an important role in animal food and this subsystem has been studied in detail and a
life cycle inventory of this manufacture process has been carried out. The inventory data pro-
ceeded from two factories sited in Galicia (Table 3). The first one has an annual production of
over 100,000 tonnes and an animal distribution of 60% for cattle, 35% for pig and 5% for other
animals. The second one has an annual production of 90,000 tonnes, which is distributed in the
following percentages: 90% for cattle destined to milk production and 10% for cattle belonging
to rural families destined to their own milk consumption. The distribution of the production in
terms of areas of influence is the following: 98% of total production is delivered within 30-40
km and 2% goes to longer distances (100 km).
LCI at Dairies
All the Galician dairies have a similar technology and a comparable size, so the dairy inventory
data were calculated on the basis of their annual production for the years 2001 and 2002: around
200 millions litres. The products produced are packaged liquid milk: 71%, whole; 18%, semi-
skimmed and 11%, skimmed milk. Cream is obtained as a co-product and it is sold for further
processing in other factories.
Transport by isothermal trucks from farms to dairies has been included and studied in detail,
identifying all the routes associated to each factory as well as their distance and amount of milk
collected.
Table 4 displays the inventory data for Dairy Subsystem.
123
Table 3. Inventory Data of the Fodder Factory (FU = 1 kg of fodder).
INPUTS
From the TECHNOSPHERE From the NATURE
Materials and fuels Raw materials
1. Raw materials 1. Water 66.62 mL
1a. Maize 165.8 ± 34.2 g
1b. Barley 189.2 ± 15.9 g
1c. Wheat 47.7 ± 7.7 g
1d. Rye 82.22 g
1e. Soy bean 143.7 ± 10.4 g
1f. Soy shell 19.4 ± 0.6 g
1g. Gluten 104.8 ± 15.2 g
1h. Cotton seed 34.7 ± 5.3 g
1i. Molasses 22.1 ± 1.3 g
1j. Calcium carbonate. 13.2 ± 3.9 g
1k. Phosphate carbonate. 4.6 ± 2.1 g
1l. Alfalfa 3.6 ± 1.7 g
2. Paper bags 1.39 ± 0.03 g
Electricity
1. Electricity 49.1 ± 3.3 Wh
OUTPUTS
To the TECHNOSPHERE To the NATURE
Products and co-products
1. Fodder 1 kg
Waste for treatment
1. Urban Solid Waste 0.11 kg
2. Oil 0.08 g
Discussion
Average data as well as the associated standard deviations have been presented separately for
each subsystem. As it can be observed, the majority of these data suffer from a high variability
(defined as the ratio between the standard deviation and the mean), which means that representa-
tive production and processing scheme can be difficult to establish. For instance, all the elements
included on the food ration, with the exception of fodder or concentrate feed, undergo variability
around 50%.
The next step will be the evaluation of this inventory in order to identify the most pollutant sub-
systems along the process. In this study, focus has been paid on the LCI so this will be beyond
the scope of the study.
124
Table 4. Inventory Data of the Dairy (FU = 1 L of packaged milk).
INPUTS
From the TECHNOSPHERE From the NATURE
Materials and fuels Raw materials
1. Raw milk 1.11 ± 0.09 L 1. Water 3.18 ± 1.74 L
2. Tetra-brik 1.01 ± 0.01 u
3. Cardboard 12.26 ± 6.42 g
4. Film 1.12 ± 1.33 g
6. Hydrogen peroxide 0.69 ± 0.34 g
7. Nitric acid 1.12 ± 1.04 g
8. Sodium hydroxide 1.95 ± 0.45 g
9. Fuel 8.06 ± 2.32 g
Transport
1. Truck 40 ton 281 ± 104 kgkm
Electricity
1. Electricity 46.12 ± 10.90 Wh
OUTPUTS
To the TECHNOSPHERE To the NATURE
Products and co-products Solid emissions
1. Packaged milk 1L 2.155 g
46.42 ± 34.16 g
1. Combustion waste
2. Cream
0.20 ± 0.01 g
Emissions to air
Waste for treatment 1. SO2
1. Cardboard 0.34 ± 0.05 g 2. NOx 6.83 ± 4.75 g
2. Film 0.785 g 3. CO 2.14 ± 2.39 g
3. Used oil 0.04 g Emissions to water
4. Used oil filters 10-7 units 1. Wastewater 1.06 ± 1.25 L
5. Used Tetra-brik 0.008 ± 0.006 u Emissions to soil
1. Sludge 40.46 ± 17.74 g
Future Outlook
The research planned for the future will continue in several directions:
• In order to achieve a higher quality of data at farm level, we are trying to proceed with a
more complete inventory including new production systems for each stage. For instance,
we are carrying out the inventory of two additional farms, which are examples of com-
pletely different herd sizes (one farm of 20 cows and the other of 170 heads) in order to
identify the influence of extreme cases on the environmental impact associated. At the
moment those inventories are not finished so the conclusions about this aspect are not
still available. Regarding dairy level and bearing in mind the average technology and
specialising degree at Galician factories, this influence is supposed to be not significant
and this is the reason why we are concentrated on farm level.
125
• We would like to get a more comprehensive knowledge about the influence of cereals on
the fodder factories. We have used data from international databases on the basis of the
assumed country of origin for each one. The next step to check the reliability of this as-
sumption is the comparison with the data handled by the major farm cooperative that
manages imports and supplies raw cereals to an important number of the fodder factories
in Galicia.
• Another point of attention will be the consideration of the named Ecological Milk, which
is commercialised in our region since 2002. Nowadays, only one company is processing
this ecological product and its market share is still small, but bearing in mind the social
growing environment awareness, this product is supposed to have a relevant contribution
in a near future.
References
Berlin, J., 2002. Environmental Life Cycle Assessment (LCA) of Semi-Hard Cheese. Interna-
tional Dairy Journal 12, 939-953.
CEAS Consultants (Wye) Ltd and The European Forum on Nature Conservation and
Pastoralism, 2000. The environmental impact of dairy production in the EU, CEAS
1779/BDB, Brussels, Belgium. 176 pp. Available at http://europa.eu.int/
comm/environment/agriculture/pdf/dairy.pdf
Cederberg, C. and Mattson, B., 2000. Life cycle assessment of milk production - a comparison of
conventional and organic farming. Journal of Cleaner Production 8, 49-60.
EPA, 2002. U.S. Environmental Protection Agency, www.epa.gov.
Haas, G., Wetterich, F. and Köpke, U., 2001. Comparing intensive, extensified and organic
grassland farming in southern Germany by process life cycle assessment. Agriculture, Eco-
systems & Environment 83, 43-53.
Høgaas, M.E., 2002. Life Cycle Assessment of Industrial Milk Production, Chalmers University
of Technology, Gothenburg, Sweden. 53 pp.
IDAE, 2002. Instituto para la Diversificación y Ahorro de la Energía (in Spanish), www.idae.es.
Milá, L., Rieradevall, J., Domènech, X., Fullana, P. and Puig, R., 1998. Anàlisi del Cicle de Vida
de la Pell (in Catalonian), Bellaterra, Spain.191 pp.
SAEFL, 1998. Life Cycle Inventories for Packagings, Environmental Serie No. 250/I and 250/II,
Berne, Switzerland. 552 pp.
Tetrapack Spain, 1999. Declaración Ambiental 1998 (in Spanish). 10 pp.
Tetrapack Spain, 2000. Declaración Ambiental 1999 (in Spanish). 10 pp.
Tetrapack Spain, 2001. Informe de Sostenibilidad 2000 (in Spanish). 52 pp. Available at
http://www.tetrapak.es/sostenibilidad/index.htm
126
Identification of processes affected by a marginal change in demand for food
products - two examples on Danish pigs and cheese
Anne Merete Nielsen*, Per H. Nielsen1, Jørgen Dejgård Jensen2, Martin Andersen2 and Bo P. Weidema1.
*) 2.-0 LCA consultants, Borgergade 6, 1., DK-1300 Copenhagen. Tel. +45 333 22 8 22. Fax: +45
339****. [email protected]. 1) 2.-0 LCA consultants, Borgergade 6, 1., DK-1300 Copenhagen. 2) Danish
Research Institute of Food Economics, Rolighedsvej 25, DK-1958 Frederiksberg C.
Abstract
In environmental assessments of products, co-products should be dealt with by use of system
expansion. This theoretical consensus has existed for some time. However, till now still many
environmental assessments are based on economically or mass allocated data. The LCAfood
database is the first consistent model of Danish food-production, with a widespread use of sys-
tem expansion. Using quantified as well as more qualitative knowledge on market structures
and production economics, the affected processes are identified for a range of basic food prod-
ucts, in agriculture as well as in food processing industry. It is crucial to identify which technolo-
gies are affected by a product demand prior to data collection, as the work can be focused on
the most important processes, and the explanatory power of the environmental assessment can
be maximised.
Keywords: environmental assessment, milk production, system expansion, milk-quotas
Background
Agriculture has significant contribution to many negative environmental impacts (see eg.
Kramer, 2000), and product-oriented environmental assessment is a promising approach to
achieve new insights on how to decrease environmental impacts from food consumption. Be-
cause of agriculture’s high degree of co-productions due to crop rotations and linked production
of plants and animals, it has not been straightforward to find out how to assess the impacts of the
individual products. The theoretical consensus has existed for some time that co-products should
be dealt with by use of system expansion (Guinée et al., 2001), and some basic principles for the
implementation are available (Weidema et al. 1999, and Weidema, 2001). However, till now
most environmental assessments of agricultural products are based on economically or mass al-
located data.
In system expansion, the main-product for which the production is optimised should be identi-
fied, and the credit of the by-products should be given by subtracting the processes, which they
affect, e.g. an alternative production of a similar product. A key-point in the application of sys-
tem expansion is therefore the identification of the affected producer/technology.
127
To identify affected processes, it is necessary to model consequences in society, and treat the dy-
namics of human behaviour stringent and exact, even though they lie outside the field of natural
science, which is the base of most practitioners of environmental assessments. Luckily other sci-
ences offer the models for these dynamics, such as economy, marketing and politology. A limit-
ing factor for inclusion of knowledge from these sources can be the perception of “objective-
ness” in natural science. Logics and models from these other sciences are often perceived as
subjective.
For years, the use of average versus marginal data has been discussed in the field of life-cycle as-
sessment (LCA). However, in economic analysis it is a given that analysis made to support deci-
sions on the most efficient use of resources should be based upon marginal data, i.e. knowledge
on the consequences of the change under study, all other things being equal (see e.g. Hardwick et
al., 1990). Now the theoretical consensus seems to be formed among LCA-practitioners as well:
if the assessment shall predict the consequences of an action, data must come from the processes,
which are actually affected.
The LCAfood database (Nielsen et al, 2003) is the first consistent model of Danish food-
production, with a widespread use of system expansion and application of marginal data. In this
paper two examples of system expansion with affected processes are presented.
Framework for identification of affected processes
Weidema et al. (1999) provided the first applicable guide to identification of affected processes
in environmental assessments. Rephrasing, but following the string of logic, our identification of
affected producers has gone through two kinds of questions.
1. Is the supplying market increasing or decreasing, and are any of the supplying technolo-
gies constrained?
This question addresses the nature of the market, which delivers the products, and is necessary to
identify which of the possible suppliers are most likely to be affected by a small change in de-
mand. If the market is increasing, the general trend will be that new productions are being estab-
lished, with the most competitive technologies. A little change in demand will therefore affect
these technologies. If the market is decreasing on the other hand, the general trend will be, that
the least competitive producers are stopping their production. A little extra demand will therefore
make them stay longer or shorter on the market.
Some producers can be constrained by technical limits (e.g. availability of land), legislation (e.g.
limits to application of fertiliser) or other limits.
128
2. Which supplier is most likely to be affected by a small change in demand?
To answer this, one must list the technologies, which are technically capable of delivering the
relevant product (as identified in question 1) and identify which of them is a) unconstrained and
b) most respectively least competitive (for increasing respectively decreasing market).
To answer this, the total production-economical costs and benefits of the technologies should be
estimated more or less quantitatively.
Processes affected by demand for pork
The nature of the market for agricultural products
The majority of Danish pigs are produced at highly specialised farms, but productions also exist
at more diverse farms, such as sugar beet or potato-producing farms. Pork is traded on a global,
increasing market (FAOSTAT).
Possibly affected technologies and their competitiveness
Pork can be produced in all countries worldwide. Like most industrial productions, pork-
production has a clear tendency that larger production units can be run more efficiently than
smaller units (economy of scale). Therefore the most competitive production will most likely be
an intensive production.
The Danish production of pigs is to some extent limited by public regulation of manure distribu-
tion on farmland. However, since this is increasingly overcome by longer transport of the ma-
nure, the production can be considered unconstrained.
Danish pig producers have been divided into 9 main categories, and the reaction of the pig pro-
ducers on a slight increase in market price (i.e. an increased demand) is estimated by modelling
in the econometric model ESMERALDA (Jensen et al., 2001). The model is based on market re-
lated experiences including effects from public regulation of e.g. manure distribution, and pre-
dicts that an increased demand for pork will be met by a varying increases of pig production in
pig farm categories (see Table 1). The production, which is affected by a small change in de-
mand, is therefore a weighted average of production in these farms.
According to modelling in ESMERALDA, the pig-producing farm of low intensity, situated on
sandy soil would react stronger to a change in pork demand than any other farms (see Table 1).
This single farm-type was selected as a reasonable representative for the marginal producers of
pigs, because estimates of environmental data for this farm type were more accurate than for
other farm types due to a better modelling of the environmentally significant production of
grower pigs.
129
Table 1. The classification of pig-producing farms their contribution to total Danish pig produc-
tion in 1999, and their predicted contribution to a small increase of production. Dejgaard and
Andersen, (2003).
Soil- Intensitivity of Contribution to total Danish pig Contribution to affected
Main-product type production production production
of the farm (animal density) (%) (%)
Sugar beet Clay 3.9 6.5
Seeds Clay 3.4 4.6
Pigs Clay Low 3.7 10.3
Pigs Clay Medium 4.4 4.2
Pigs Clay High 23.1 3.7
Potatoes Sand 1.3 6.7
Pigs Sand Low 11.1 32.7
Pigs Sand Medium 4.7 5.3
Pigs Sand High 34.1 10.3
Residual group 10.3 15.7
All farmtypes 100.0 100.0
Processes affected by the demand for cheese
The nature of the market for cheese
Cheese is produced on specialised cheese-dairies with input of milk and output of cheese, whey
and cream. Production is increasing in Denmark and globally (FAOSTAT). The market for
cheese is strongly influenced by the public regulations (milk quotas) and is therefore not affected
directly by a small change in demand at least in a short time perspective.
A screening showed that the cheese production itself had only a small contribution to the envi-
ronmental impact from cheese. Instead the most important process to identify correctly appeared
to be the production of milk, or - to phrase it more precisely - the processes affected by the use of
milk. Therefore focus turned from the affected cheese producing technology, to the affected milk
supplier.
Possibly affected technologies and their competitiveness
A change in the cheese-production does not affect milk production in agriculture, because quotas
limit the Danish production of milk. Demand for milk for cheese production must therefore in-
fluence other processes using milk.
Milk from agriculture can be used to produce drinking milk, cheese, yoghurt, butter and dry
milk. Following the dynamics of market economy, the production with the lowest alternative
costs (i.e. lowest revenue) will be decreasing its production.
130
Milk powder is produced in excess and donated to developing countries (Nielsen et al., 2003).
Personal communication with the production planners at the dairies indicates that the dry milk
production is increased or decreased as demand for other milk products changes (Weidema,
2003). Hence a change in demand for cheese will induce a change in dry milk production and
donation. Since dry milk is donated to developing countries it is assumed that it does not displace
any other products at these markets and that only dry milk production is influenced by a slight
change in demand for cheese. It has not been possible to justify the later assumptions and this
point is still open for discussion and development.
Example of the importance of identification of affected processes
The processes that are affected by the demand for cheese as identified above are shown in Figure
1. Figure 2 shows an alternative system delimitation where market regulation of milk production
is neglected. This could be relevant for a scenario analysis to assess environmental impacts from
cheese when and if the quota system is abolished.
The figures should be read like this: Boxes refer to production processes. Names of grey boxes
refer to the main product of the processes. Red arrows represent material or energy transfer be-
tween two processes; green arrows represent saved material or energy transfer as a result of dis-
placements and green lines represent displacements. The numbers and the red/green-bars in the
boxes show each process’ cumulated contribution to total environmental impact measured as a
single score indicator. Numbers (Pt) indicate amount of “personal equivalents”, which is the en-
vironmental impact from Danish production per inhabitant. Please note that the environmental
indicator of one process cannot directly be calculated as the sum of the environmental indicators
of the processes at the below level. This is due to the looped structure of the diagram, where e.g.
the cumulated indicator for electricity in Figure 1 is partly passed on to the cheese-production,
partly to the production of whey powder. The processes and figures can be found with further
explanation in the SimaPro-database available in Nielsen et al. (2003). Figure 1 shows that the
total environmental impact from consumption of 10 tons cheese is equivalent to 1 personal
equivalent. This impact is the sum of increased environmental impacts in the cheese-production
itself as well as in the processing of by-products (whey and cream). Some reduced environmental
impacts are subtracted, because cheese production leads to decreased production of milk powder,
and because the whey powder is used as animal feed, and thus replaces barley.
Figure 2 shows that if the quota regime were abolished, the main environmental impact from
cheese would lie in the agriculture’s production of milk and their inputs. Two productions would
be reduced (green bars): the meat from the milking cows would replace alternative production of
meat, and the production of rape seed would decreased, because of the milk-farm’s by-
production of these, and because the soy oil produced as by-product to the soy-meal would re-
place rape seed oil.
131
Cheese
1 Pt
Electricity Whey powder Milk powder Butter
(natural gas) production production production
1 Pt 2 Pt - 3 Pt 0,2 Pt
Spring Barley, Heat (dairies)
from farm
- 1 Pt 1 Pt
Figure 1. The cumulated environmental impact in the product chain for 10 ton of cheese. All
shown processes will be affected by a change in demand for cheese under the present market
structure. Only processes contributing with more than 20% of total score are shown.
Cheese
20 Pt
Milk from farm Heat (dairies)
16 Pt 3 Pt
Live stock feed Live stock feed Electricity Cattle, from
(spring barley) (soy) farm
3 Pt 7 Pt 3 Pt 5 Pt
Soy Meal Rape seed,
from farm
6 Pt
-4 Pt
Figure 2. The cumulated environmental impact in the product chain for 10 ton of cheese under
the assumption that a milk farm will be affected by an extra production of cheese. Only proc-
esses contributing with more than 15% of total score are shown.
132
It can be seen that the two different identifications of affected technologies yield highly different
results. In a comparison with e.g. other products for sandwich filling, the difference between the
results in figure 1 and 2 is enough to significantly change the conclusions. If data are used to
identify the hotspot-processes in the product chain of cheese, the conclusions will also be irrele-
vant and misleading.
Discussion and outlook
In the LCAfood database, environmental data for the Danish production of basic food is mod-
elled using system expansion. The affected processes are identified based on quantified as well
as more qualitative knowledge on market structures and production economics. This work led us
to some key-findings.
Necessary cooperation between production engineers and market-economists
Lifecycle assessment used to be a matter of inputs and outputs from processes linked to each
other by a physical flow of materials. However, other processes than those directly involved in
the physical chain are influenced by a product demand and these processes influence the final re-
sult of an assessment, as shown in the cheese-example above. Hence, lifecycle assessment is no
longer just a natural science accomplished by environmental departments and process engineers
in co-operation, but a combined natural and social science, which also requires inputs from mar-
keting and economic departments.
The level of details in identification of affected technologies
The level of details in the identification of the affected producer can vary. In the pork-example,
the affected producer was identified according to some key factors in the production technology
(namely the production intensity, soil type and main production of the farm). In the case of
cheese, it was found that the limiting factor to decrease uncertainty of the assessment was not to
identify the cheese-producing technology, but to identify which processes was affected by the
cheese-production’s consumption of milk. This can be seen from the big differences between
Figure 1 and 2. Therefore the level of details with which the affected producer is identified
should depend upon the factors, which affect the result of the assessment the most.
The workload of market-based LCA
Market-based identification of affected technologies adds a new task to the life cycle assessment.
Since this will be on the expense of other tasks of the assessment, it should only be added, if it
actually increases the explanatory power of the study. The workload for the identification of sys-
tem delimitation will increase, especially if the economical information must be found outside
the group performing the LCA. Workload for data collection may be increased, because it is less
possible to use standard databases, which are based on average data. On the other hand, identifi-
cation of affected processes prior to the data-collection can help focus the data-collection on the
133
processes, which are most important to the result, and can thus avoid collection of unnecessary
data. The advantage of the approach comes in the assessment. Conclusions based on allocation
can be very unsatisfactory, if different allocations yield different results. Conclusions based on
system-expanded data where affected technologies are identified are easier to interpret, even if
there is doubt about the right identification of affected technology, because there is a clear bond
between the market-assumption and the different results.
References
FAOSTAT. FAO Statistical Databases. Available at http://apps.fao.org/
Guinée, J. B., Gorrée, M., Heijungs, R., Huppes, G., Kleijn, R., Koning, A. de, Oers, L. van,
Sleeswijk, A. W., Suh, S., Udo de Haes, H. A., Bruijn, H. de, Duin, R. van, Huijbregts, M.
A. J., Lindeijer, E., Roorda, A. A. H., Ven, B. L. van der, Weidema, B. P. (2001). LCA –
An operational guide to the ISO standard. Centre of Environmental Science, Leiden Uni-
versity.
Hardwick, P., Khan, B., Langmead, J. (1990). An introduction to modern economics. 3rd edition.
Longman Group UK Limited.
Jensen, J.D., Andersen, M. & Kristensen K. (2001). A Regional Econometric Sector Model for
Danish Agriculture – A Documentation of the Regionalized ESMERALDA Model. Danish
Research Institute of Food Economics, report no. 129.
Jensen, J.D. and Andersen, M. (2003). Marginal producers of a selection of agricultural products.
Working paper. Danish Research Institute of Food Economics. In press.
Kramer, K.J. (2000). Food Matters – On reducing energy use and greenhouse gas emissions from
household food consumption. Rijkuniversiteit Groningen
Nielsen P.H., Nielsen A.M., Weidema B.P., Dalgaard R. and Halberg N. (2003). LCA food data-
base. www.lcafood.dk/database.
Weidema, B.P., Frees, N. and Nielsen, A.M. (1999). Marginal Production Technologies for Life
Cycle Inventories. International Journal of LCA 4 (1), 48 – 56.
Weidema, B. P. (2001). Avoiding Co-Product Allocation in Life-Cycle Assessment. Journal of
Industrial Ecology 4 (3), 11 - 33.
Weidema B.P. (2003). Market information in life cycle assessments. Environmental Project no.
863. Danish Environmental Protection Agency.
134
Land occupation and transformation in the Swiss life cycle inventory data-
base ecoinvent 2000
Jungbluth, N., Frischknecht, R.
*ESU-services, Kanzleistrasse 4, CH-8610 Uster, T. +41 1 940 61 91 or +41 1 940 6132, F. +41 1
940 61 94, [email protected], [email protected]
Centre for life cycle inventories in the ETH domain, c/o EMPA, CH - 8600 Dübendorf
1. Abstract
Land occupation and land transformation gets more and more attention in life cycle inventory
analyses and life cycle impact assessment methods. It is especially important for agricultural
and forestry products. However, consistent land occupation and transformation figures for
unit processes of an economic sector or even complete process networks including agricul-
ture, forestry, energy supply, transport and waste treatment services, materials production,
etc. are still rare. For the Swiss ecoinvent 2000 project, one emphasis is put on a systematic
registration and quantification of land occupation and transformation. For that purpose the
ecoinvent 2000 project group developed a simplified methodology that allows for a relatively
efficient data compilation and data handling while at the same time minimising information
loss in view of future developments in LCIA (life cycle impact assessment). The approach
uses the CORINE land cover typology which has punctually been enlarged (e.g., land occu-
pation and transformation on the sea bottom, the so-called benthos) to cover all LCI specific
land types. It considers as much as possible the experiences and recommendations of the
SETAC LCIA working group. Occupation on the one hand and the land types just before and
after a land transformation on the other hand are reported separately. Hence, the three types
of elementary flows for land use are: “occupation, …” (in m2.a), “transformation, from …” and
“transformation, to …” (both in m2). On the level of cumulative LCI results, the balance of the
total surfaces transformed, for instance "transformation, to forest" minus "transformation,
from forest", indicates whether the surface of forests de- or increased (negative and positive
balance, respectively) due to the supply of the functional unit at issue. The presentation ex-
plains and substantiates the method applied with the help of a case study. The developed
methodology can be used to record land use patterns in life cycle inventories for all types of
products.
Keywords: land use, occupation, transformation, life cycle inventory
2. Introduction
Land occupation and land transformation gets more and more attention in life cycle inventory
analyses and life cycle impact assessment methods (Lindeijer et al. 2001). It is especially im-
portant for agricultural and forestry products. However, consistent land occupation and trans-
formation figures for unit processes of an economic sector or even complete process networks
including agriculture, forestry, energy supply, transport and waste treatment services, materi-
als production, etc. are still rare and the documentation is not sufficient for an in depth impact
assessment. For the Swiss ecoinvent 2000 project, one emphasis is put on a systematic regis-
tration and quantification of land occupation and transformation (ecoinvent Centre 2004). For
that purpose the ecoinvent 2000 project group developed a simplified methodology that al-
135
lows for a relatively efficient data compilation and data handling while at the same time
minimising information loss in view of future developments in LCIA.
3. Land use and its impacts
The impact category “land use” in LCIA covers a range of consequences of human land use.
Land use mainly has impacts on the following area’s of protection (Guinèe et al. 2001):
• Natural Resources
• Natural Environment
• Man-made Environment
According to Lindeijer et al. (2001) the impacts of land use are classified in four groups. Cer-
tain land use types might also have a positive impact:
1. Increase of land competition
2. Degradation of biodiversity
3. Degradation of life support functions
4. Degradation of cultural values
The methodology used in the ecoinvent database concentrates on the second and - as far as
possible - the third impact (Frischknecht et al. 2004).
4. Distinction between land occupation and land transformation
A distinction is made between
• land occupation (i.e., the operation of a power station hinders the occupied land from
changing to state it would have under uninfluenced conditions), and
• land transformation (i.e., a new assembly plant for airbus airplanes require the con-
version of a former natural resort to industrial land; a gravel-pit is converted to a natu-
ral resort by active re-cultivation).
For land occupation the surface as well as the duration required for the production of a certain
amount of products and services are important. That is why land occupation is recorded in
square metres times time (m2a).
Clearly defined and relatively short changes in the land use type are recorded as land occupa-
tion (e.g., the construction of underground natural gas pipelines, which converts agricultural
land to an excavation site). For these construction processes as well as for active restoration
activities after decommissioning, the land use category "land occupation, construction site" is
applied.
Land transformation links a state during an economic activity with a state before and a state
after that activity (road construction, power plant erection, active mine restoration, etc.). But it
also may occur during the economic activity itself (open pit lignite extraction).
136
For particular processes the land use type before starting the activity may well be known.
However, it is difficult to assess in detail all the land use types which have been converted by
the production processes recorded within the ecoinvent project. If not known, the land use
type "transformation, from unknown" is applied. Continental or regional statistics about land
transformation over time may then be used later on to attribute specific land use type to this
land use type "unknown".
Land transformation consists of an entry
1. land transformation, from land use type X, and
2. land transformation, to land use type X.
For land transformation at the beginning of an economic activity the land use type encoun-
tered at that point in time is recorded. This starting state, such as "transformation, from for-
est", is recorded in m2. The transformation to the land use type valid during the economic ac-
tivity is recorded as well. For gravel extraction, for instance, the m2 "transformation, to
mineral extraction site" are recorded.
This land transformation needs to be attributed to the total amount of products and services
delivered (the life time production of a power plant, one production cycle of a forest, time pe-
riod until the depletion of a mine, etc.).
Tabel 1 shows the time periods applied in the ecoinvent project if no specific information is
available (Frischknecht et al. 2004).
Active restoration at the end of an extraction or production process is modelled as a separate
unit process (restoration, gravel-pit", "restoration, copper mine"). This process includes tech-
nical requirements such as diesel for construction machines, seeds, etc.. Additionally, land
transformation to the final land use type (and from the land use type during the operation of
the economic process) are recorded (e.g., "transformation, from mineral extraction site", and
"transformation, to pasture and meadow"). Such restoration processes may be required by an
infrastructure process (such as a power plant: "restoration, power plant") or by the production
process itself (in case land transformation and restoration takes place during the economic ac-
tivity (e.g., lignite extraction).
Land transformation caused by the use of the land for new purposes is attributed to this future
new uses. No land transformation after the assumed life time is recorded for actual land uses
that are likely not change in the future (such as transport infrastructure, agricultural land) as
well as for land abandoned and subjected to natural succession.
5. Regional differentiation
Inventory data are collected on the level of national averages. Hence, no regional differentia-
tion can be made. Unit processes are described by a geographic code, be it a country or a con-
137
tinent or an international organisation. This geographic code provides information about
where the land occupation and transformation of the process at issue takes place. However,
the ecoinvent database does not yet allow for an automatic evaluation of this information.
Table 1. Land occupation types used in ecoinvent based on the CORINE land cover types
classification. The same types are used for land transformation, according to the naming rules
as described in the text.
english name CORINE class Use period
Occupation, arable CORINE 21 1
Occupation, arable, non-irrigated CORINE 211 1
Occupation, arable, non-irrigated, diverse-intensive CORINE 211b 1
Occupation, arable, non-irrigated, fallow CORINE 211c 1
Occupation, arable, non-irrigated, monotone-intensive CORINE 211a 1
Occupation, construction site CORINE 133 not considered
Occupation, dump site CORINE 132 10
Occupation, dump site, benthos CORINE 132a 1
Occupation, forest CORINE 31 80
Occupation, forest, extensive CORINE 31a 100
Occupation, forest, intensive CORINE 31b 80
Occupation, forest, intensive, clear-cutting CORINE 31b2 80
Occupation, forest, intensive, normal CORINE 31b1 60
Occupation, forest, intensive, short-cycle CORINE 31b3 30
Occupation, heterogeneous, agricultural CORINE 243a 100
Occupation, industrial area CORINE 121 50
Occupation, industrial area, benthos CORINE 121c ??
Occupation, industrial area, built up CORINE 121a 50
Occupation, industrial area, vegetation CORINE 121b 50
Occupation, mineral extraction site CORINE 131 20
Occupation, pasture and meadow CORINE 231 30
Occupation, pasture and meadow, extensive CORINE 231b 20
Occupation, pasture and meadow, intensive CORINE 231a 20
Occupation, permanent crop CORINE 22 20
Occupation, permanent crop, fruit CORINE 222a 15
Occupation, permanent crop, fruit, extensive CORINE 222b 15
Occupation, permanent crop, fruit, intensive CORINE 222a 15
Occupation, permanent crop, vine CORINE 221 25
Occupation, permanent crop, vine, extensive CORINE 221b 25
Occupation, permanent crop, vine, intensive CORINE 221a 25
Occupation, sea and ocean CORINE 523 ??
Occupation, shrub land, sclerophyllous CORINE 323 100
Occupation, traffic area, rail embankment CORINE 122d 50
Occupation, traffic area, rail network CORINE 122c 100
Occupation, traffic area, road embankment CORINE 122b 50
Occupation, traffic area, road network CORINE 122a 100
Occupation, unknown CORINE x 1
Occupation, urban, continuously built CORINE 111 80
Occupation, urban, discontinuously built CORINE 112 80
Occupation, water bodies, artificial CORINE 512a 100
Occupation, water courses, artificial CORINE 511a 100
138
6. Naming rules
The differentiation between transformation and occupation is reflected in the naming of land
use elementary flows. It takes pattern from the naming proposals of a Dutch project (Lindeijer
& Alfers 2001) and deviates from the provisional proposals of the SETAC Europe working
group (de Beaufort-Langeveld et al. 2003; Hischier et al. 2001):
• Occupation, type, subtype
• Transformation, from type of occupation
• Transformation, to type of occupation
The different levels of details in describing the land use type are separated by commas:
• Occupation, arable
• Occupation, arable, non-irrigated
• Occupation, arable, non-irrigated, monotone-intensive
The highest level of information detail is always used and recorded in the inventories.
7. Land use types in ecoinvent
The definition of land use types takes pattern from the CORINE land cover types (Bossard et
al. 2000). New, and partly more detailed land use types have been added to the CORINE table
using the same systematic. Only land cover types required in the ecoinvent project have been
added. The list may however easily be extended if required.
Land use types do not include national or even regional differentiation. For instance, the land
use type "pasture and meadow, extensive" covers land occupation (and transformation) by
Alpine pastures as well as South American cattle pastures.
If the land use types before and after the operation phase of the economic activity are not
known, land transformation is recorded with "transformation, from unknown" and "transfor-
mation, to unknown", respectively. With that, the sum of all "transformation, to ..." equals the
sum of all "transformation, from ...". From these figures the net transformation of land use ty-
pes can be calculated.
If the total amount of m2 "transformation, from forest" is larger than the amount of m2 "trans-
formation, to forest", calculated for the production of 1kWh of Swiss low voltage electricity,
the production of this kWh reduced the total amount of forest.
8. Allocation issues
Land transformation and occupation may be allocated among several products or services.
Roads for instance are built for personal and freight traffic. Allocation is done with regard to
the influence of each of the products / services on the impact categories of land occupation
and transformation listed above.
139
9. Attribution of land use to operation or construction/dismantling (infrastructure)
Land occupation and transformation may in some cases either be attributed to the infrastruc-
ture or the operation of a process. The surface of a greenhouse may rather be recorded in the
infrastructure part whereas the farm land would rather be recorded in the operation of an agri-
cultural process. As a rule, agricultural and forest property surfaces are attributed to the opera-
tion as long as they do not include buildings and roads. Land use by buildings, forest roads,
greenhouses and the like are attributed to the infrastructure.
Ecoinvent allows for calculation of results excluding infrastructure requirements. Hence, land
used by infrastructure is neglected and may influence impact assessment results substantially.
Care must be taken when comparing LCIA results computed without infrastructure contribu-
tions.
10. Example
The ecoinvent approach to land use is illustrated with the following example of gravel extrac-
tion:
• total area (1000 m*1000 m = 1'000'000 m2),
• gravel-pit used during 20 years,
• 1'000'000 tons of gravel extracted per year,
• 2 years of restoration activities,
• diesel consumption of 500 t (21.3 TJ) per year during extraction and of 100 t (4.3 TJ)
during restoration.
The inventories for the two processes "gravel, at extraction/kg/RER/0" and "restoration,
gravel-pit/m2/RER/0" are shown in Tabel 2. Land occupation is calculated by dividing the to-
tal surface by the total amount of gravel extracted per year. Land transformation is calculated
by dividing the total surface by the total life time production (1'000'000m2 / 20a *
1'000'000'000 kg/a = 5.0 * 10-5 m2/kg). The requirements for restoration are also evenly at-
tributed to the life time production of the gravel-pit. Restoration leads to a forest (assump-
tion). This restoration activity may not be included in the inventory of forestry products (e.g.,
timber). For timber production, the correct land transformation before starting that activity is
"transformation, from forest". In the column results one can see that the total surface trans-
formed is 5.0 * 10-5m2 per kg gravel extracted and that the net transformation is "from un-
known" to "forest". The two transformations "to mineral extraction site" and "from mineral
extraction site" cancel each other out.
140
Table 2. Example of unit process raw data including land transformation and land occupation.
raw data results
gravel, at ex- restoration, gravel, at ex-
traction gravel-pit traction
kg m2 kg
2
Resource Occupation, mineral extraction site m a 0.001 0.001
use Occupation, construction site ma 2
2 0.0001
Transformation, from unknown m2 5.00E-05 5.00E-05
2
Transformation, to mineral extraction site m 5.00E-05 5.00E-05
Transformation, from mineral extraction site m2 1 5.00E-05
Transformation, to forest m2 1 5.00E-05
Techno- restoration, gravel-pit m2 5.00E-05 5.00E-05
sphere diesel, burned in building machine MJ 0.021 4.3 0.0212
Inputs
... ...
Output gravel, at extraction kg 1 1
restoration, gravel-pit m2 1
11. Conclusions
The methodology developed within the ecoinvent 2000 project can be used to record land use
patterns in life cycle inventories for all types of products. It fits for the implementation of dif-
ferent LCIA methodologies. The experiences within the ecoinvent project showed that the
method is quite easy to apply for the recording of land occupation. But, it is still time consum-
ing to investigate the land transformation patterns for a given process because the necessary
data are quite often not available. Therefore it might be useful to elaborate clear guidelines
how this transformation can be investigated.
12. References
Bossard M., Feranec J. and Otahel J. (2000) CORINE land cover technical guide - Addendum
2000. 40. European Environment Agency (EEA), Copenhagen (DK), retrieved from:
www.eea.eu.int, http://etc.satellus.se/I&CLC2000/download.htm.
de Beaufort-Langeveld A. S. H., Bretz R., van Hoof G., Hischier R., Jean P., Tanner T. and
Huijbregts M. (2003) Code of Life-Cycle Inventory Practice (includes CD-ROM). SE-
TAC, retrieved from: www.setac.org.
ecoinvent Centre 2004. ecoinvent Centre (2004) ecoinvent data v1.1. ISBN 3-905594-38-2.
Swiss Centre for Life Cycle Inventories, Dübendorf, CH, retrieved from:
www.ecoinvent.ch.
Frischknecht R., Jungbluth N., Althaus H.-J., Doka G., Dones R., Hischier R., Hellweg S.,
Nemecek T., Rebitzer G. and Spielmann M. (2004) Overview and Methodology. Final
report ecoinvent 2000 No. 1. Swiss Centre for Life Cycle Inventories, Dübendorf, CH,
retrieved from: www.ecoinvent.ch.
Guinèe J. B., Gorrée M., Heijungs R., Huppes G., Kleijn R., de Koning A., van Oers L.,
Wegener Sleeswijk A., Suh S., Udo de Haes H. A., de Bruijn H., van Duin R., Hui-
jbregts M. A. J., Lindeijer E., Roorda A. A. H. and Weidema B. P. (2001) Life cycle as-
141
sessment; An operational guide to the ISO standards; Parts 1 and 2. Ministry of Hous-
ing, Spatial Planning and Environment (VROM) and Centre of Environmental Science
(CML), Den Haag and Leiden, The Netherlands.
Hischier R., Baitz M., Bretz R., Frischknecht R., Jungbluth N., Marheineke T., McKeown P.,
Oele M., Osset P., Renner I., Skone T., Wessman H. and de Beaufort A. S. H. (2001)
Guidelines for Consistent Reporting of Exchanges from/to Nature within Life Cycle In-
ventories (LCI). In: Int. J. LCA, 6(4), pp. 192-198.
Lindeijer E., Müller-Wenk R., Steen B., with written contributions from, Baitz M., Broers J.,
Finnveden G., ten Houten M., Köllner T., May J., Mila i Canals L., Renner I. and Wei-
dema B. (2001) Impact Assessment of resources and land use. SETAC WIA-2 taskforce
on resources and land.
Lindeijer E. and Alfers A. (2001) Summary of Step A of the Delfts Cluster Research Pro-
gramme on Land Use in LCA. In: Int. J. LCA, 6(3), pp. 186.
142
A new land use impact assessment method for LCA: theoretical fundaments
and field validation
Peters, J. (1), García Quijano, J. (2), Content, T. (3), Van Wyk, G. (4), Holden, N.M. (1), Ward, S.M.
(1) and Muys, B (3,*).
(1) Department of Agricultural and Food Engineering, University College Dublin, Ireland
(2) Division of Engineering and Applied Sciences, Harvard University, USA
(3) Laboratory for Forest, Nature and Landscape Research, Katholieke Universiteit Leuven, Vital De-
costerstraat 102, B-3000 Leuven, Belgium, Tel: +32 16329726, Fax: +32 16329760,
[email protected]
(4) Faculty of Agricultural and Forestry Sciences, University of Stellenbosch, South Africa
Abstract
For LCA studies with a major part of the life cycle in agriculture and forestry, incorporation of
a land use impact category is essential. Different method proposals have been formulated in
the last 6-7 years, but their common weakness is an arbitrary choice of indicators due to a
lack of solid theoretical basis. Other frequently reported problems of the existing proposals
are related to the choice of a reference system, the lack of universal applicability and
time/space issues such as the incompatibility between land use change and permanent land
occupation. The new method proposal addresses most of these problems. A theoretical
background based on ecosystem thermodynamics uses the hypothesis that in absence of
human land use impact, all ecosystems tend to maximize the internal exergy level and con-
trol over incoming and outgoing exergy fluxes. In order to measure land use impact, the de-
viation from the site specific maximum ecosystem performance in exergy terms is estimated
using 17 quantitative indicators and aggregated into four thematic scores. The indicator
scores for ecosystem biomass and structure and for biodiversity are quantifying the land use
impact on the ecosystem exergy level, while the indicator scores for water and for soil and
nutrients are quantifying the land use impact on ecosystem buffering capacity for exergy
fluxes. Thematic scores are multiplied by the area x time needed for the production of the
functional unit. Test results from different land uses in several countries under cold temper-
ate, mediterranean, subtropical and tropical climates indicate that the method is workable
and universally applicable. Results are discussed and recommendations for further im-
provements are formulated.
Keywords: Life Cycle Assessment, land use impact assessment
1. Introduction
Many human activities have important spatial needs (e.g. for the extraction of resources, for
production processes, for landfill). The use of land for a given production process will often
make the land unavailable for other uses. Land transformation and occupation may also
change the “quality” of the land. The land use impact category is especially relevant in LCA
studies of products with a major part of their life cycle in the agriculture and forestry sector.
Several methods have been proposed to assess the land use impact (Sturm and Westphal,
1996, Baitz et al., 1998, Bradley et al., 1998, Giegrich and Sturm, 1998, Lindeijer et al., 1998
Schweinle, 1999, Køllner, 1999). The strengths and weaknesses of these proposals and the
differences between them have been tested in some comparative studies (Giegrich et al., 1999,
Leplae, 2000, Peters, 2002). Two expert groups published recommendations for developing
better land use impact assessment methods: the COST E9 action (LCA for forestry and forest
143
products) working group on land use (Schweinle et al., 2002) and the SETAC working group
on resources and land use (Lindeijer, 2002).
In this paper a newly developed method based on these recommendations is presented. It
starts from a scientific concept for the choice of indicators, as recommended by COST E9
(Schweinle et al., 2002). For the impact assessment it combines the strong points from the
methods of Sturm and Westphal (1996), Baitz et al. (1998), Giegrich and Sturm (1998)
Lindeijer et al. (1998), and Schweinle (1999), and adds some new ideas, while remaining
simple and flexible.
2. Methods
2.1. Basic concept and reference system
COST E9 working group 2 proposed to use ecosystem exergy as a basic concept (Schweinle
et al., 2002). Ecosystems are systems that are open to energy and/or material flows and reside
in states of thermodynamic non-equilibrium. When in thermodynamic non-equilibrium, the
second law of thermodynamics cause ecosystems to counter applied gradients and oppose
movement towards equilibrium (Schneider and Kay, 1994). Ecosystems attempt to develop
away from the thermodynamic equilibrium by gaining exergy and losing entropy. The goal
function of an ecosystem can be defined as the continuous tendency to go as far as possible
from the thermodynamic equilibrium by: (i) exergy storage maximisation (Bendoricchio and
Jørgensen, 1997, Fath et al., 2001, Scott, 2003) and (ii) maximising external exergy flow dis-
sipation (Schneider and Kay, 1994, Fath et al., 2001). Maximisation of exergy storage, makes
ecosystems build up biomass and embedded information (genetic heritage) (Bendoricchio and
Jørgensen, 1997) in complex structures with greater biodiversity and more hierarchical levels.
The ecosystems tendency to maximize exergy dissipation results in more exergy capture and
flow in the system, more energy and nutrient cycling within the ecosystem, higher trophic
structures, more respiration and transpiration, more ecosystem biomass and a greater organ-
ism diversity (Schneider and Kay, 1994). It is hypothesised that for any site (i.e. a combina-
tion of abiotic circumstances such as climate, soil, aspect) the potential natural vegetation (i.e.
the climax system) is the ecosystem with greatest possible exergy storage and dissipation
level under natural circumstances (Muys et al., 2001). The potential natural vegetation is cho-
sen as the reference system in this method. Any human activity altering physical, chemical or
biological components of an ecosystem will have a direct impact on the exergy storage and
dissipation (Schweinle et al., 2002). Measuring exergy storage and dissipation could provide a
measure of land use impact on ecosystems. Direct exergy storage calculation has not been
possible (Scott, 2003), but indirect top-down approaches using indicators have been proposed
(Luvall et al., 2001, Schweinle et al., 2002).
The exergy concept may be linked with environmental issues of concern which are grouped in
three areas of protection (or safeguard) (International Organisation for Standardisation, 1997,
Udo de Haes, 1999): (i) natural environment (e.g. human interventions causing change in ex-
ergy storage or dissipation results in ecosystem transition from climax to a secondary state);
144
(ii) natural resources (e.g. desertification reduces exergy storage and dissipation and adversely
impacts on the economic value of land) and (iii) human health (e.g. biomass/biodiversity de-
cline reducing supply of plant materials or changes to soil processes causing health problems
(e.g. Oliver, 1997)).
2.2. Land use impact calculation
The considered land must be classified as site types with homogenous abiotic variables. For
every site type, the reference system, which is the potential natural vegetation, must be identi-
fied. The method describes the land use impact by 17 quantitative impact indicators divided
over 4 themes: soil, water, vegetation and biodiversity (Table 1). A score (∆Q) is attributed to
the chosen indicators that reflects the difference in quality between the actual situation and the
reference. In exergy terms, human land use can affect the ecosystem quality in two ways: it
can directly change the exergy storage level and the ability of the ecosystem to control exergy
flows (e.g. removal of living biomass), and it can change the site quality (e.g. by irrigation,
fertilisation, etc.). Changing the site will lead to another climax system with another maxi-
mum exergy level. The indicators are chosen in such a way that the reference has a value
equal to or close to zero and that impacts cause a positive value with a maximum of 100. A
natural ecosystem will get an impact score (close to) zero, while anthropogenically modified
ecosystems (with lower exergy levels) will get a positive impact score proportional with the
impact magnitude. It is however possible that human interventions (e.g. introduction of new
species) could increase the buffering capacity of the ecosystem without changing the site qua-
lity. Therefore, a negative impact score up to –25 is provided for in the method.
Table 1. Impact indicators grouped per theme: soil, water, vegetation and biodiversity.
( ) ⎞⎟ *100
Code Indicator Formula Units
⎛ area aff * permref − permact ⎛ cm ⎞
⎜ ⎜ ha d⎟
S1 Soil compaction
⎜ ⎟ ×
⎜⎜ ha cm ⎟⎟
⎝ ⎠
⎝ d⎠
areatot * permref
where areaaff = area affected; areatot = total area;
permref = permeability at the reference state; permact =
⎛ areaaff * depth timesS 2 ⎞
permeability at the actual state
⎛ ha * m yr ⎞
⎜ ⎟ * 100 ⎜⎜ ×
S2 Soil structure dis-
⎜ rot ⎟⎠
⎝ ⎝ ha yr ⎠
turbance by *
ploughing, etc. areatot
where timesS2 = number of soil works per rotation period;
rot = length of rotation period (in years)
⎛ 100 * USLE
⎜⎜
⎞
⎟⎟ * 100 ⎛ t ⎞
⎜ ha ⋅ yr ⎟
S3 Soil erosion
⎝ Soil depth ⎠ ⎜ t ⎟
⎜ ⎟
⎝ ⎠
where USLE = soil loss in t ha-1 yr-1; soil depth = total ha
rootable soil depth in t ha-1
⎛ CEC act ⎞ ⎛ Meq ⎞
⎜1 − ⎟ * 100 ⎜ ⎟
S4 Cation Exchange
⎜ CEC ⎟ ⎜ ⎟
⎝ ⎠
Capacity (CEC) 100 g
⎜⎜ ⎟⎟
⎝ ⎠
ref Meq
⎛ ⎞
100 g
⎛%⎞
⎜1 − BS act ⎟ * 100 ⎜ ⎟
S5 Base Saturation
⎜ BS ⎟ ⎝%⎠
⎝ ⎠
(BS)
ref
145
⎛ ETact ⎞ ⎛ mm ⎞
⎜1 − ⎟ *100 ⎜ yr ⎟
W1 Evapotrans-
⎜ ET ⎟ ⎜ mm ⎟
⎝ ⎠
piration (ET)
⎜ ⎟
⎝ ⎠
ref
yr
⎛ mm ⎞
⎜ yr ⎟
W2 Surface runoff SR
P − ET ⎜ mm ⎟
(SR) * 100
⎜ yr ⎟⎠
⎝
⎛ TABact ⎞ ⎛ t ⎞
⎜1 − ⎟ *100 ⎜ ha ⎟
V1 Total above-
⎜ TAB ⎟ ⎜⎜ t ⎟⎟
⎝ ⎠
ground living bio-
mass (TAB) ⎝ ha ⎠
⎛ LAI act ⎞ ⎛ m² ⎞
ref
⎜1 − ⎟ *100 ⎜ m² ⎟
V2 Leaf area index
⎜ LAI ⎟ ⎜⎜ m² ⎟⎟
⎝ ⎠
(LAI)
⎝ m² ⎠
ref
⎛ H act ⎞ ⎛m⎞
⎜1 − ⎟ *100 ⎜ ⎟
V3 Vegetation height
⎜ H ⎟ ⎝m⎠
⎝ ⎠
(H)
⎡ ⎛ NPPact − AH ⎞⎤ ⎛t ⎞
ref
⎢1 − ⎜⎜ ⎟⎥ *100 ⎜ ha⋅yr− ha⋅yr
V4 Free Net Primary t
⎢⎣ ⎝ NPPref ⎟⎥ ⎜ t
⎠⎦
Production (fNPP)
⎜ ha⋅yr ⎠
⎝
⎛ crop biomass ⎞
where AH = annual harvest
⎛ t ⎞
⎜⎜ ⎟⎟ * 100 ⎜ ha ⋅ yr ⎟
V5 Crop biomass
⎜ t ⎟
⎝ total biomass ⎠ ⎜ ha ⋅ yr ⎟
⎝ ⎠
⎡⎛ area irr + area drain ⎞⎤ ⎛ ha ⎞
⎢⎜⎜ ⎟⎟⎥ *100 ⎜ ⎟
B1 Artificial change of
⎝ ha ⎠
⎣⎝ ⎠⎦
water balance
total area
where areairr = irrigated area;
⎡⎛ area aff times B 2 ⎞⎤
areadrain = drained area
⎛ ha yr ⎞
⎢⎜⎜ ⎟⎥ * 100 ⎜⎜ × ⎟⎟
B2 Liming, fertilisa-
⎢⎣⎝ areatot rot ⎟⎠⎥⎦ ⎝ ha yr ⎠
tion, empoverish- *
ment
⎡⎛ area aff times B 3 ⎞⎤
where timesB2 = number of applications per rotation period
⎛ ha yr ⎞
⎢⎜⎜ ⎟⎥ * 100 ⎜⎜ × ⎟⎟
B3 Biocides
⎣⎢⎝ areatot rot ⎟⎠⎦⎥ ⎝ ha yr ⎠
*
where timesB3 = number of applications per rotation period
⎛ ⎞ ⎛ number ⎞
⎜⎜ ⎟⎟ *100 ⎜ ⎟
B4 Canopy cover of Ex
⎝ total cov er ⎠ ⎝ number ⎠
exotic plant spe-
cies (Ex)
⎛ ⎞ ⎛ number ⎞
⎜1 − Spact ⎟ *100 ⎜ ⎟
B5 Number of plant
⎜ Sp ⎟ ⎝ number ⎠
⎝ ref ⎠
species (Sp)
The 17 different indicator scores are aggregated in 4 thematic scores for soil (∆QS), water
(∆QW), vegetation (∆QV) and biodiversity (∆QB). The aggregation is done as follows:
∑ ∆Q
n
∆Q S = i =1
N =5
Si
where Eq.(1).
N
∑ ∆Q
m
∆ QW = M =2
j =1
Wj
where Eq.(2).
M
146
∑ ∆Q
x
∆ QV = X =5
p =1
Vp
where Eq.(3).
X
∑ ∆Q
y
∆Q B = Y =5
q =1
Bq
where Eq.(4).
Y
The soil and water themes reflect the buffering capacity of the ecosystem for exergy flows.
The vegetation and biodiversity themes reflect the exergy level of the ecosystem itself, in
terms of biomass, complex trophical networks and genetic information.
Land use denotes the use of a piece of land for a certain purpose during a certain period of
time. So space and time are essential parameters. The dimension of land use is therefore area
x time per functional unit (FU), i.e. the area needed to produce 1 functional unit in 1 rotation
period. The final thematic impact scores are:
S S = ∆ Q S * ( area × time ) FU * FU −1
Eq.(5).
S W = ∆ QW * ( area × time ) FU * FU −1
Eq.(6).
S V = ∆ Q V * ( area × time ) FU * FU −1
Eq.(7).
S B = ∆ Q B * ( area × time ) FU * FU −1
Eq.(8).
Eq.(5-8) can be applied to cases of land use occupation and land use change. This distinction
is proposed by the COST E 9 working group on land use (Schweinle et al., 2002): land use
occupation is the continuous use of a certain piece of land for a certain period of time, land
use change denotes a more or less abrupt change from one land use to another. A poplar plan-
tation with a rotation period of 25 years is considered as an example of land use occupation.
To calculate the mean land use impact over 100 years for the different themes Eq.(9) has to be
used. The factor (area x time)FU is assumed to be constant over the rotation periods.
∑ ∆Q
n
Si = i=0
* (area × time )FU −1
n
* FU Eq.(9).
N
for i = soil, water, vegetation and biodiversity
As can be seen in Eq.(10) land use change compares the difference in quality of the last rota-
tion of the former land use with the first rotation of the new land use. A degradation of land
quality will lead to positive values, land quality improvement to negative values. The (area x
time)FU factor is for the new land use. A visual representation is given in Figure 1.
S i = [∆Q2 − ∆Q1 ]* (area × time )FU * FU −1 Eq.(10).
for i = soil, water, vegetation and biodiversity
147
Q Q reference system
∆Q1
∆Q 2
Q land use 1
Q land use 2
n = rotation period
t
Figure 1. Land use quality (Q) as a function of time (t) during land use change.
2.3. Test scenario definition
The proposed land use impact assessment method has to fit into the LCA framework. Impor-
tant criteria are that should be met include: universal applicability; the reference system must
be definable; the assessment must be quantitative;, a meaningful functional unit must be used;
the dimension must be appropriate; linearity; sufficient sensitivity to land use change; and in-
dicator efficiency. To test the method against these criteria different test scenarios were as-
sessed in different climatic and geographical zones.
Multifunctional forest in Flanders (TEMP multi for)
A forest where wood production is combined with high ecological and recreational value,
characterized by long rotations (150 years), managed with a thinning frequency of once every
10 years and regenerated with a group selection system. For the simulation we used inventory
data of Meerdaal Forest near Leuven, a 1200 ha FSC certified ancient woodland. (Muys et al.,
2002).
Short rotation coppice in Flanders (TEMP en for)
Afforested agricultural lands with willow and poplar clones. Fully mechanised harvesting oc-
curs in 3-year rotation cycles for the above-ground biomass and 25-year rotation cycles for
the below-ground biomass (Muys et al., 2002).
Eucalypt plantation in Spain (MED eucalypt)
Different Eucalyptus globulus plantations are analysed, mainly in Galicia and Huelva. Rota-
tion periods of 12 years are used to produce pulp (Madueño, 2003).
Tropical rainforest conservation (TROP land use)
Tropical moist forest in South-western Cameroon that is part of the Tropenbos Cameroon
Programme. Primary forest, selective logging, shifting cultivation and agricultural land uses
are assessed. Logging activities focus on Azobé (Lophira alata Banks ex Gaertn.), Tali
(Erythrophleum ivorense A. Chev) and Padouk (Pterocarpus soyauxii Taub). The shifting cul-
tivation system involves the clearing and burning of primary and old secondary forest just be-
148
fore the rainy season. Cleared land is cultivated and left after a maximum of five years. Agri-
cultural land use consists of cassava cultivation. (Leplea, 2000).
Subtropical agroforestry plantation (SUBTROP land use)
The plantation combines avocado (Persea americana Mills) cultivation with Eucalypt (Euca-
lyptus grandis Hill ex Maiden) forestry. A clearcut system is used and leftovers are burned.
Eucalypts have rotation cycles of 25 years and are thinned twice and pruned three times. The
rotation cycle for avocado varies between 25 and 30 years. Avocado trees are pruned to keep
a constant height of 7.5 m. (Content, 2003).
Mediterranean pine plantation (MED land use)
Afforestation of semi -natural scrubland with Pinus radiata based on data from the FSC certi-
fied Jonkershoek plantation near Stellenbosch, Western Cape, South Africa. The rotation
length is 30 years and includes two thinnings and a final clearcut. (Peters, 2002). To test the
suitability of the functional unit, the linearity of the method and the dimensions of the end re-
sult, results for the pine plantation are expressed per functional unit Eq.(5-8), where one tonne
of wood produced is the functional unit. The methods output (thematic scores) for the test
scenarios were analysed by Monte Carlo – Chi Square analysis. It is assumed that independ-
ent and dependent variables are discrete. The Chi Square test statistic is:
χ = ∑∑
2 (O jk − E jk ) 2 Eq.(11).
j k E jk
with Ojk = Observed value row j, column k;
Ejk = Expected value row j, column k
The Monte Carlo test was used because it can handle small expected values, small sample
sizes and degrees of freedom do not have to be verified. Monte Carlo also allows for null hy-
pothesis (H0) variation, depending on the randomisation. H0 states that distribution is uniform
over the whole matrix, rows and columns when respectively the whole matrix, columns and
rows are randomised.
2.4. Principal Component Analysis (PCA)
For testing on indicator efficiency, PCA was used. A data matrix of land uses versus indicator
scores was constructed, and from the data matrix, a correlation matrix was calculated. Princi-
ple Components – orthogonal axes explaining maximum data variation - were defined and
presented graphically, with data presented as vectors. Distance between data is proportional to
data similarity; angles between vectors are proportional to correlation. Orthogonal projection
of vectors on the principal axis are eigenvalues and are proportional to correlation between
vector and principle axis.
149
3. Results
3.1. Land use impact assessment of different test scenarios
All scenarios were divided into homogenous site classes, and for each site class the reference
vegetation was identified. Indicator scores were determined and aggregated into thematic
scores according Eq.(1-4) and summarised in Table 2a. Thematic scores of different land uses
were tested one by one using Monte Carlo – Chi Square analysis with H0: there is no signifi-
cant difference between the two land uses analysed. Some important results are:
• Land use impact decreases with decreasing land use intensity. Infrastructure has the
greatest impact followed by agricultural land use and forestry. Utilisation of natural
resources (TROP primary forest, MED fynbos, MED indigenous forest and SUB-
TROP kloof forest) had the lowest impact.
• Impacts of intensive land uses (TROP infrastructure, MED infrastructure, SUBTROP
agricultural land use) do not differ significantly.
• Reference vegetations suggest that there is no significant difference (p >0.05) between
TROP primary forest, MED indigenous forest, and SUBTROP kloof forest, but there
are significant difference (p <0.05) between MED fynbos and TROP primary forest,
MED indigenous forest and SUBTROP kloof forest.
• There is no significant difference (p >0.05) between the two Eucalypt plantations as-
sessed (MED eucalypt and SUBTROP eucalypt).
• In Flanders, multifunctional forestry (TEMP multi for) has lower impacts on vegeta-
tion and biodiversity than energy forestry (TEMP en for). Energy forestry has lower
impacts on soil and water. There is no overall significant difference (p >0.05).
• There is no significant difference (p >0.05) between selective logging and shifting cul-
tivation in tropical rainforest (TROP). Though we can conclude that shifting cultiva-
tion has a slightly higher impact, because impacts are greater at 0.05 significance level
compared with primary forest, and no significance difference (p >0.05) can be seen
between primary forest and selective logging.
• Fynbos (MED fynbos) and indigenous forest (MED indigenous forest) are both refer-
ence vegetations in the Western Cape. As a consequence impacts are close to zero.
Fynbos has a higher water impact than indigenous forest becasue natural burning cy-
cles lead to poor water flux buffering. There is a significant difference (p <0.05) be-
tween both land uses.
• Mediterranean pine plantation (MED pine plantation) has a fairly low impact on all
themes except for biodiversity. Impact on water is negative, indicating a higher water
flux buffering capacity than MED fynbos and MED indigenous forest.
• MED Fire belt has a high impact on vegetation, because it is burned every 4 years,
keeping vegetation shorter and less structured than the reference vegetation.
• In the subtropical agroforestry scenario, Avocado has significantly higher impacts
than Eucalypt (p <0.05).
150
• There is no significant difference (p >0.05) between SUBTROP secondary forest and
SUBTROP kloof forest impacts, except for vegetation. Secondary forest is not as
structured as the reference kloof forest vegetation.
Thematic scores for the different land uses in the Mediterranean pine plantation scenario were
scaled spatially to get thematic scores for the whole management unit. Results of Eq.(1-4). are
0.91, 0.47, 6.87 and 10.00 respectively. The (area x time)FU * FU-1 is 1,675 m2 yr t-1, i.e. 1,675
m² of pine plantation are needed to harvest 1 t wood after 1 rotation period. Using Eq.(5-8) re-
sults in SS 1524 m2 yr t-1, SW 782 m2 yr t-1, SV 11507 m2 yr t-1 and SB 16750 m2 yr t-1 (Table
2b).
3.2. Principal Component Analysis
The complexity of natural systems, and our limited understanding of cause-effect relation-
ships within them means that it is not possible to include all relevant energy and material
flows in the impact assessment. It is possible however, to characterise a limited number of
impacts using indicators. The method defines 17 quantitative indicators within 4 themes. Indi-
cator choice is based on the exergy concept. To test the indicator efficiency a PCA was used.
For all indicators and land uses of the mediterranean pine plantation and subtropical agrofor-
estry plantation scenario a correlation matrix was made and principle components were calcu-
lated. Results are presented graphically in Figure 2, and both axes explain 73% of the varia-
tion. Distance between points is proportional to similarity, angles between vectors are
proportional to correlation. 3 clusters can be distinguished. Soil erosion (S3) is not correlated
with other indicators, probably because of the low erosion value for all land uses. Soil struc-
ture disturbance (S2), liming and fertilisation (B2) and biocides (B3) are strongly correlated
because silvicultural practices always have a related soil disturbance. Furthermore these indi-
cators have a very similar build-up through time. A second cluster consisted of indicators S4,
S5, W1, V2 and V3. All these indicators have a similar build –up through time, but correla-
tion also results because of ecological variables (which are difficult to clearly define). It is not
surprising that V3 and W1 are correlated, LAI is often used to model evapotranspiration. The
third cluster consists of S1, W2, V1, V4, V5, B1, B4, and B5. A possible explanation of cor-
relations between these indicators for the land uses assessed in PCA is that a higher crop per-
centage per land use results in greater annual harvest (V4), greater crop biomass (V5), greater
cover of exotic plants (B4), lower number of species (B5). Soil compaction (S1) and surface
runoff (W2) are positively correlated in general.
151
Table 2. (a) Thematic impact scores for the different land uses; (b) Thematic impact scores
per FU for the Mediterranean pine plantation scenario
(a) THEMATIC IMPACT SCORES [/]: ∆QS ∆QW ∆QV ∆QB
SCENARIO
CLIMATIC ZONE LAND USE
TEMP multi for 2.6 5.3 25.1 10.3
TEMP en for -1 -3.8 44.5 35
MED eucalypt 5.8 4 34.4 38.8
TROP primary forest 0 1.5 0 0
selective logging 1.4 2.6 9.2 2.4
shifting cultivation 1.8 4 13.1 17
agricultural land use 20.5 21 81 52
SUBTROP eucalyptus 8.2 -4.4 33.3 39.5
avocado 9.7 24.9 38.7 87.5
wetland 0.2 -8.9 46.6 17.3
secondary forest 0 8.25 25.4 1.3
infrastructure 60 79.2 100 60
kloofbos 0 7.3 0 0
MED pine plantation 0.5 -3.7 4.6 15.9
fynbos 0.5 5 0 0
indigenous forest 0 0 0 0
infrastructure 20 47.3 100 40
fire belt 0.5 5 39.1 0
(b) THEMATIC IMPACT SCORES PER FU [m² yr t-1]: SS SW SV SB
FU = 1 ton of wood leaving the plantation
subtropical plantation scenario 1524 782 11507 16750
4. Discussion and Conclusion
The method allowed for comparison of different land uses in different climatic or geographi-
cal regions in a fully quantitative matter. The potential natural vegetation varied over all sce-
narios assessed and was a useful reference. Land use impact can be expressed per FU and the
dimensions [m2 yr FU-1] are suitable for LCA application. The linearity requirement is satis-
fied, doubling the FU results in a doubled impact score. The methods sensitivity is moderate.
All indicators and the methodology are fully quantitative, consequently every change in input
results in a changes of one or more thematic impact scores. A weakness is the subjective
minimum threshold of –25% for land uses, which increase the buffering capacity without
changing the site quality. The minimum threshold eliminates variation. The various indicators
deal with different components of the ecosystem, but they do not describe the ecosystem in-
dependently, so double counting is possible with the system, as is indicated by the PCA re-
sults. The PCA identifies which indicators are not necessary to assess and compare land use
impact of subtropical agroforestry and mediterranean plantations with equal accuracy, but
generalisation about indicator correlations are not possible. Assessing other land uses might
lead to different indicator correlations. Conclusions of the method for criteria defined in sec-
tion 2.2. are presented in Table 3.
152
B3 S2
0,40 B2
0,20 B4
S1
B5
pca2
V5 V4
S3 B1
0,00
V1 W2
-0,20 V3
W1
V2
S5 S4
-0,10 0,00 0,10 0,20 0,30
pca1
Figure 2. PCA of the MED pine plantation and SUBTROP agroforestry plantation scenario.
Data labels are indicator codes (see Table 1)
The method does not aggregate theme scores into an overall score. This is interesting when
LCA is used for decision-making. Production cycle 1 scores better under soil and water
themes, while production scenario 2 scores better on vegetation and biodiversity for the land
use category. To decide which is better, regional interests can be included. For example in
South Africa where water is an important issue, production cycle 1 is preferred, while in Ire-
land where water availability is currently of little concern, production cycle 2 might be pre-
ferred because of biodiversity reasons.
Table 3. Results of the method for different criteria.
Legend: + = good, +/- = moderate, - = poor
criteria
use of functional unit
universal applicable
indicator efficiency
workable reference
system definition
quantitative
dimension
sensitivity
linearity
+ + + + + + +/- +/-
153
5. Recommendations
• Further testing of the method is needed. Special attention should be given to the indi-
cator choices. Until now, many indicators seem to correlate, suggesting indicators
could be eliminating to avoid double counting. Distinguishing between related land
uses one might require all indicators.
• Indicators are designed as ratios between the actual -and reference states. All indica-
tors are dimensionless, except for S2 and S3, which should be addressed.
• It is possible to add new indicators. A feasible indicator is based on red list species:
⎛ ⎞
⎜ ⎟
⎜1 − ⎟ ∗ 100
number of red list species
⎜ ⎟
act
⎝ ⎠
number of red list species ref
This indicator represents nature value and, in exergy terms, it is important to conserve biodi-
versity as genetic information is a way to store exergy.
• Indicator minimum threshold needs revision. All indicators have a maximum of 100,
but indicators do not have an intrinsic minimum. The –25 threshold was chosen to
limit positive land use quality changes of man made land uses, but clearly affects sen-
sitivity. Lowering the minimum threshold to –100 might solve sensitivity problems.
• The potential natural vegetation seems to be an excellent reference system, but prob-
lems can rise where no natural ecosystem is left for measurements. Literature data can
be used but may affect overall reliability.
• Reference vegetation has the greatest possible exergy content for the site so impact
should be 0, which is the case for all indicators except for soil erosion (S3) and surface
runoff (W2). S2 and W2 should be adapted, so that reference vegetations score 0 on
all themes.
• A general methodology to deal with uncertainty should be developed. The method as
presented allows for comparison of many land use impacts, but reliability of assess-
ment should be compared as well.
• Within themes the calculation (Eq. 1-4) of the impact scores gives equal weight to
each indicator (e.g. in the soil theme soil erosion is as important as base saturation).
This subjective weighting system needs reconsideration.
• Different themes cannot be compared within the same land use. All land uses have
low soil and water scores, suggesting that soil and water indicators are less sensitive.
Weighing factors can be attributed to indicators in order to make themes comparable,
or themes themselves can be weighted. An important point is that different weighing
systems should be used for different ecosystems but that objectivity and comparability
of different land use assessments would be affected.
6. Acknowledgements
The authors would like to thank the Flemish Government, Belgium for funding through the
project PBO 98/48/16 and the Environmental Protection Agency Ireland for funding through
ERTDI project LF 5.1.2b supported by the National Development Plan (2000-2006).
154
7. References
Baitz, M., Kreissig, J. and Wolf, M., 2000. Methode zur Integration der Naturraum-
Inanspruchnahme in Ökobilanzen. Forstw Cbl 119: 128-149.
Bendoricchio, G. and Jørgensen, S.E., 1997. Exergy as a goal function of ecosystem dynamic.
Ecological modelling 102: 5-15.
Bradley, M., Krokowski, K., Bresky, J., Pettersson, B., Woodtli, H., Nehm, F. and Daniel, P.,
1998. Proposal for a new forestry assessment method in LCA. LCA graphic paper and
print products (part 1). Axel Springer Verlag, Stora, Canfor study. 40 pp.
Content, T., 2003. Land use impact of a subtropical forestry plantation in Westfalia, South
Africa: evaluation of a newly developed LCA method for impact category land use and
comparison with land uses in different climatic regions. Diplom work, Laboratory for
Forest, Nature and Landscape Research, KULeuven. 65pp.
Fath, B.D., Patten, B.C. and Choi, J.S., 2001. Complementary of ecological goal functions. J.
theor. Biol. 208: 493-506.
Giegrich, J. and Sturm, K., 1998. Operationalisierung der Wirkungskategorie
Naturraumbeanspruchung. IFEU, Heidelberg. 19pp.
Giegrich, J., Schweinle, J., and Baitz, M., 1999. Verifizierung verschiedener Methoden zur
Wirkungsabschätzung des Wirkkriteriums Naturrauminanspruchnahme, Landnutzung,
Landverbrauch. Deutsche Gesellschaft fuer Holzforschung, Muenchen. 94 pp.
International Organisation for Standardisation, 1998. ISO/DIS 14042: Environmental man-
agement – Life cycle assessment – Life cycle impact assessment. Geneva.
Köllner, T., 2000. Species-pool effect potentials (SPEP) as a yardstick to evaluate land-use
impacts on biodiversity. Journal of cleaner production 8: 293-311.
Leplae, H., 2000. Comparison of different LCA methods for the impact category land use in a
tropical moist evergreen forest (Ebom, Cameroon). Diplom work, Laboratory for Forest,
Nature and Landscape Research, KULeuven. 89 pp.
Lindeijer, E.W., Van Kampen, M., Fraanje, P.J., Van Dobben, H.F., Nabuurs, G.J.,
Schouwenberg, E., Prins, A.H., Dankers, N. and Leopold, M.F., 1998. Biodiversity and
life support indicators for land use impacts in LCA. Rijkswaterstaat, Delft. report W-
DWW-98-059.
Lindeijer, E., Muller-Wenk, R. and Steen, B. (eds.), Baitz, M., Broers, J., Cederberg, C.,
Finnvelden, G., ten Houten, M., Köllner, T., Mattsson, B., May, J., Milà e Cannals, L.,
Renner, I. and Wiedema, B. (contributors), 2002. Impact assessment of resources and
land use. Chapter in Udo de Haes, H.A. et al., 2002.
Luvall, J.C., Kay, J.J., Fraser, R.F., 2001. Thermal remote sensing and the thermodynamics of
ecosystem development. In: Ulgiati, S., Brown, M.T., Giampietro, M., Herendeen, R.A.,
Mayumi, K. (Eds.), 2001. Advances in energy studies: exploring supplies, constraints,
and strategies. SGE dictoriale, Padova. 147-158 pp.
Madueño Murillo, M., 2003. Land use impact assessment of Eucalyptus globulus in Spain.
Diplom work, Laboratory of Forest, Nature and Landscape Research, KULeuven. 61 pp.
155
Muys, B. , Wagendorp, T. and Coppin, P., 2001. Ecosystem exergy as an indicator of land use
impact in LCA. In: Ulgiati et al. Advances in energy studies: exploring supplies, con-
straints and strategies. Portovenere, Italy. pp 275-284.
Muys, B., Deckmyn, G., Moons, E., García Quijano, J., Proost, S. and Ceulemans, R., 2003.
An integrated decision support tool for the prediction and evaluation of efficiency, envi-
ronmental impact and total social cost of forestry projects in the framework of the Kyoto
Protocol. International conference: Decision support for multiple purpose forestry, April
23-25, Vienna, Austria. 17 pp.
Oliver, M.A., 1997. Soil and human health: a review. European Journal of Soil Science 48:
573-592.
Peters, J., 2002. Comparative study of different LCA methods for the impact category land
use in a Radiata pine plantation, Jonkershoek, South Africa. Diplom work Laboratory
for Forest, Nature and Landscape Research, KULeuven. 106 pp.
Schneider, E.D. and Kay, J.J., 1994. Life as a manifestation of the Second law of Thermody-
namics. Mathematical and Computer Modelling 19: 25-48.
Schweinle, J., 1999. Methode zur Integration des Aspektes der Flächennutzung in die Ökobi-
lanzierung. In: Giegrich, J., Schweinle, J. and Baitz, M.. Verifizierung verschiedener
Methoden zur Wirkungsabschätzung des Wirkkriteriums Naturrauminanspruchnahme,
Landnutzung, Landverbrauch. Deutsche Gesellschaft fuer Holzforschung, Muenchen. 94
pp.
Schweinle, J., Doka, G., Hillier, W., Kaila, S., Köllner, T., Kreißig, J.,
Muys, B., Garcia Quijano, J., Salpakivi-Salomaa, P., Swan, G., and Wessman, H., 2002.
The Assessment of Environmental Impacts caused by Land Use in the Life Cycle As-
sessment of Forestry and Forest Products - Guidelines, Hints and Recommendations. Fi-
nal Report of Working Group 2 “Land use” of COST Action E9. Mitteilungen der Bun-
desforschungsanstalt für Forst-und Holzwirtschaft, Hamburg. 96 pp.
Scott, D.S., 2003. Exergy. International journal of hydrogen energy 28: 369-375.
Sturm, K. and Westphal, C., 1996. Entwicklung einer Methode zur Bestimmung der
Flächenqualitäten für Waldöksysteme für die Zwecke der Produkt-Ökobilanz. Büro für
angewandte Waldökologie, Duvensee. 40 pp.
Udo de Haes, H.A., Jolliet, O., Finnvelden, G., Hauschild, M., Krewitt, W. and Müller-Wenk,
R., 1999. Best available practice regarding impact categories and category indicators in
life cycle impact assessment. Int J LCA 4(2): 66-74.
156
Soybean Production in Argentina: New Technologies, agricultural envi-
ronment and socioeconomic impacts. The implications for the future
Walter A. Pengue
GEPAMA – www.gepama.com.ar, Grupo de Ecología del Paisaje y Medio Ambiente (Landscape
Ecology and Environment Group), Universidad de Buenos Aires – Argentina, wapen-
[email protected]
Introduction
Soybean has transformed itself into the most important crop of Argentina. In the same way,
the country is at the global top ranking in relation with the adoption of the transgenic soybean
technology. This year, planted surface with this crop rise to around 11,000,000 ha with a pro-
duction of 34,000,000 metric tons (95% is transgenic).
For farmers, Round-up Ready (RR) soybean came to solve one of the main problems for the
farm management: weed control, obtaining a cost reduction in the herbicide price, less fossil
energy consumption and simple application that made the technical package offered irresista-
ble. For the private pesticides and seed production sector, it opened the unique possibility to
concentrate and rearrange the business of production and commercialization of insecticides
and weed killers to the new biotechnological alternative.
The main area where soybean were produced was The Pampas, one of the most productive
places in the world. But currently, due to the need for larger scale production, farmers are go-
ing out, increasing the pressure on more environmental sensitive areas. This situation is re-
produced in Bolivia, Brazil and Paraguay.
The Pampas prairie is a vast, flat region of Argentina that comprises more that 50 million hec-
tares of arable lands for crop and cattle production. Agriculture in the pampas has a short his-
tory (a little more than 100 years), and shared several common features with the agricultural
history of the North American Great Plains. Both ecoregions were mostly native rangelands
until the end of the 19th century and the beginning of the 20th, and both of them were later in-
troduced into crop (cereal crops and oil seeds) and cattle production on dryland conditions.
The Pampas prairie is not homogeneous in soils (Morello, J and Matteucci, S, 1997). Using
soils and rainfall patterns, the Pampas can be divided (Viglizzo, 2002) into five homogeneous
areas: 1) Rolling Pampas, 2), Central Pampas (which could be subdivided in Suhumid on the
East and Semiarid on the West, 3) Southern Pampas, 4), Flooding Pampas and 5) Mesopota-
mian Pampas.
In Argentina, specially in the Pampas, and now in areas out of it, from the north to the west,
soybean production has, during last five years, displaced 4,600,000 ha dedicated earlier to
other production, like dairy, fruticulture, horticulture, cattle or other agricultural sectors. More
157
than fifty percent of the whole agrifood sector in Argentina (73,000,000 metric tons) will
come this year from soybean sector. This unusual situation may endanger the stability of the
Argentina economy or at least several sectors of it, as well as the food sovereignty of the
country itself.
The increase of the soybean sector, which responds specifically to a global demand – Argen-
tina consumes a very little of its own production – has produced important impacts on the en-
vironment, the economy and the society.
New technologies imported, success in economic terms of the No Tillage model, Transgenic
Soybean and an explosive consumption of very specific pesticides have produced a particular
combination that doubled the argentine production during the last decade: the “Input Dec-
ade”.
Landscape transformation in the rural sector is evident, the homogenization of the rural land-
scape and the transformation of virgin areas, could produce consequences that we are now
evaluating, because this type of process has not a prior history in Argentina, neither in South
American or global agriculture in general, and its impacts must be evaluated thoroughly.
Soybean Production in Argentina
Since 1997, the private companies, in many cases with the support and expectations of the
government sector, has established that the transgenic package (Round- up Ready soybean +
glyphosate) could offer to the country a real competitive advantage.
These advantages added to the already existing comparative advantages of the country could
make Argentina one of the most efficient countries for producing and trading agricultural
commodities.
In this way, since the 1996/1997 season, there has been a strong campaign for the commer-
cialization of RRsoybean, that grew from 20% to 95% of the surface planted with GMO-soya
in 2002/2003. In seven years, there was rapid adoptions of the new technology by growers, so
that in the current season the whole production of argentine soybean is transgenic. Argentina
did not generate the new technology, which has been imported by an international company’s
branch from USA.
Few years ago, traditional cultivation of grains alternated with fallow seasons for cattle pas-
ture. This rotation system allowed maintaining the agronomic and environment system in the
long-term. But, in the 1980s, world market prices for grains and oilseeds increased, while at
the same time productivity of raising cattle declined. Agriculture became more lucrative,
since the production of soybean in rotation with wheat or sunflower allows for three harvests
in two year. Furthermore, the opening of the economy to the global market, the end of hyper-
inflation due to the fixation of the argentine peso against the US dollar, and abolition of ex-
158
port levies on agricultural products, triggered an investment in new technologies. This new
framework favored the import of machinery and agricultural inputs as pesticides, fertilizers
and royalties on seeds at low prices and their use in oilseed production under No Tillage sys-
tem for export markets (Picture Nº 1).
Agriculture intensification under No Tillage model
Hectares Equivalent litres
12000000 120000000
No Tillage Glyphosate
10000000 100000000
8000000 80000000
6000000 60000000
4000000 40000000
2000000 20000000
0 0
2
0
/9
/9
/9
/9
/9
/9
/9
/9
/0
/9
/9
/9
/9
/9
/9
/9
/9
/0
91
92
93
94
95
96
97
98
99
91
92
93
94
95
96
97
98
99
Hectares Sowing Equipment
10000000 3500
9000000
RR Soybean No Tillage Seeder
3000
8000000
7000000 2500
6000000
5000000 2000
4000000 1500
3000000
2000000 1000
1000000
500
0
0
2
0
/9
/9
/9
/9
/9
/9
/9
/9
/0
91/92 92/93 93/94 94/95 95/96 96/97 97/98 98/99
91
92
93
94
95
96
97
98
99
Source: SAGPyA, Aapresid, Pengue, 2000.
Picture Nº 1. Agriculture Intensification under the No Tillage Model in Argentina.
(RR soybean means GMO-soybean resistant to Glyphosate).
The intensification of the production system was followed by a decline in soil fertility and in-
crease of soil erosion (Prego, E, 1996) Consequently, fertilizer consumption stepped up from
0.3 million tons in 1990 to 2.5 million tons in 1999. Another step was the continuos increase
of No Tillage system that is directly associated with the high consumption of herbicides –
such as glyphosate – reinforced with the release of transgenic soybean that are tolerant to this
herbicide (package glyphosate + Roundup Ready soybean).
The main factors that produce the rapid adoption of transgenic soybean are:
a) Lower herbicide prices. In Argentina, from a price of $ 28/litre goes down now to $
3/l, much less expensive that in the USA. Four companies (Monsanto, Atanor, Nidera
and Dow) dominate more than 80% of glyphosate market in Argentina, mainly im-
ported from USA, EU and China.
159
b) Fewer expenses on labor, fuel and machinery. No Tillage and more effective herbicide
application allow for crop cultivation with less labor and fewer machinery cycles.
c) Complete knowledge of the technological package associated to No Tillage + Soy-
bean.
d) Seed prices and self-reproduction. In Argentina, farmers don’t pay technological fee
for seeds and they reproduce the new seeds in theirs fields. This year the “white bag”
(seed with no certificate and fiscalization) is around 300.000 ton.
Risks and profits under conditions of technological changing
Biotechnology is emerging at a period of worsening inequalities between the developing
countries and the industrialized world. The income gap between the fifth of the world’s peo-
ple living in the richest countries and the fifth in the poorest was 74 to 1 in 1997, up from 60
to 1 in 1990 and 30 to 1 in 1960 (UNDP, 1999). And of the US$ 460 billion spent in R&D
worldwide, only one tenth was spent in developing world where 80% of the world’s popula-
tion resides (UNESCO, 1999). These figures imply that much of developing countries are
unlikely to benefit from biotechnology.
Although many of the developing countries are interested in the role of biotechnology for im-
proving nutrition and reducing hunger, the majority of current agricultural biotechnology ef-
forts are driven by the markets in the developed world: thus much of the research focus is on
crops that are staple varieties for animal foods, attributes that minimize labor and comfort for
farmers (unique herbicides, insecticides) or improve the quality of foods. Many of the crop
varieties, traits and environmental or health conditions that could be important for large parts
of the developing world are still largely ignored.
The intensification of agriculture implies for South American countries like Argentina and
Brazil (the two main crop growers) two important transformations of land use:
An intensive production, under high input technology on common agricultural lands in the
whole Pampas.
An extensive production, on new lands, gaining and advancing on marginal areas (agricultural
border) with new varieties of soybean (transgenic and not transgenic), bred specifically to be
adapted to these new virginal lands (Photo Nº 1).
160
Photo Nº 1. Clearfield of lands in the Yungas forest prepared for soybean production (Salta,
Argentina, 2002).
Three decades before, soybean was a botanical curiosity. Nowadays, it is the engine of
MERCOSUR. It is the third exportation good (after coffee and sugar) and the first of Argen-
tina. But, both countries have followed different goals and different views of markets. While
Argentine followed the United States and continue with the intensification of OGM produc-
tion, the production and release of engineered crops is under discussion in Brazil, where the
government only allowed planted transgenic soybean this season, with an open end for the
next future (Table Nº 1).
Table Nº 1. Growing of surface implanted with soybean in Argentina and Brazil.
Country decade 1970 1980 1990 2000
Argentina Hectares 50,000 2,000,000 5,000,000 8,000,000
Brazil Hectares 1,000,000 6,000,000 12,000,000 13,000,000
Source: Pengue, Walter. Seminar Sustainable Agriculture in the Third World, pp.71-87 (2002), Brussels.
The current situation seems to be a bifurcation of the world market. By one side, those coun-
tries that accept engineered crops and those that do not accept engineered crops or insist that
those crops and foods have to be labeled.
161
Current situation and trends for agriculture production in Argentina.
Argentina can be considered as a “natural” country, free till the first years of this decade of
high inputs of chemicals as fertilizers, insecticides or herbicides for its crops (Table Nº 2).
This is a “market value”. But, in hands of globalization and facing an important soybean de-
mand, the country is changing its system of production, intensifying agriculture, with high
consumption of imported chemicals, new varieties of crops and a class of agricultural bio-
technology that implies more consumption of herbicides, with active principles [ingredients?]
imported too.
Table Nº 2. Some agriculture indicators in selection agriculture economies.
Argentina USA France
Insecticides (gr./ha) 250 1000 3000
Fertilizer (kg. /ha) 25 100 300
Herbicides (gr./ha) 250 900 2000
Changes in farm area (%) 18 5 - 2.5
Native mammals under danger of extinction (%) 10 11 50
Native birds under danger of extinction (%) 2 8 40
Native reptiles under danger of extinction (%) 0 6 38
Sources: INTA (1995), INRA(1995), USDA (1996), Pengue (1996)
Historically, Argentina has been characterized for its natural conditions, that even following
the intensification of the “green revolution”, the country did not consume much chemicals.
Only the erosion and nutrient exportation have been important as a consequence of wrong
management and the incorporation of the package for soybean, without the right evaluation of
the environmental context. But nowadays, adding to the problems with the soil resource, the
entire ecosystem will be involved. The “new biorevolution”, in the way that is being promoted
in Argentina, will allow increasing the agricultural cycles, diminishing the length of fallow
fields and restoration, increasing the impacts and pressure over natural resources, the social
system and the economy.
Agriculture intensification has produced environmental, economical and social consequences
that have not been evaluated conspicuity in the country. Probably, the new biorevolution
could exacerbate the weak conditions of the system: Intensification of agriculture, globaliza-
tion, large farm concentration, low levels of credit for small farmers, dependence of imported
supplies, dependence of technology, apropiation of large farms by outside owners, concentra-
tion of seeds and chemicals on a very few agricultural firms.
This simplification of agriculture will produce effects that will affect the commercial position
of Argentina in the meantime: degradation of soils and biodiversity, rural migration, concen-
tration in large farms only producing high yielding crops in place of more natural foods.
162
There are social and economic consequences related with the important changes and trans-
formation of national economy. Since 1991, starting the period of dollar convertibility and
opening of the argentine market, changes in the mode of production have led to a number of
social transformations for the agricultural sector:
a) Dependence on imports. Grains and soybean have become the main goods for foreign
markets, boosting the dependence on import of the inputs. Local production of pesti-
cides rose 16.6%, while 43.6% are imported and the other 39.8% are produced in Ar-
gentina with imported drugs. Glyphosate consumed in 2003 is around 160,000,000 lt.
b) Concentration of holdings. New technological package offered in a context of profit
margins falling down by half between 1992 and 1999, makes it very difficult to sur-
vive for many farmers indebted with bank loans of high interest rates to pay back for
these investments in machinery, chemical inputs and seeds. This situation favors the
concentration of holdings and many farmers (especially small and medium size grow-
ers which were the train of the argentine economy) disappeared. Between 1992 and
1999, the number of farms in Las Pampas declined from 170,000 to 116,000, while the
average size of a producer’s farm increased from 243 to 357 hectares. In 2003: 532
hectares.
c) Dumping prices. Argentina as many developing countries subsidize neither its farmers
nor the goods they produce, but are being affected by those governments that subsi-
dize the production of commodities in developed countries. In this way, these activi-
ties promote an intensification of agriculture production in developing countries, over-
exploitation of resources and subutilization of goods (that excluded the valuation of
externalities).
d) Exclusion of small farmers, who cannot get financial support, for the acquisition of the
technological package.
e) Adverse consequences for organic farming by contamination or gene flow.
It is a real consideration that short-term economic and social objectives that ignore mid and
long-term environmental effects put the future sustainability of the society at risk. However,
although indicators to measure social or economic changes are abundant, indicators for as-
sessing environmental changes are scarce. The generation and development of proper indica-
tors for an agro-environmental information system are essential to get a permanent quality as-
sessment of rural environments.
About the current situation and the exploitation of the environment under a typical situation of
pressure and technological change we can question if our Pampas are sustainable at this time?
Where were we and where are we in environmental terms?, What are the tendencies?, Which
are the most worrying ones?, and which are the most appropriate indicators for an encompass-
ing evaluation of the environment of the Pampas today. Another question is how these indica-
tors are related to the social and economical ones mentioned before.
The first results are available from studies that evaluated twelve indicators: Land use, con-
sumption of fossil energy, fossil energy use efficiency, nitrogen and phosphorus balances, ni-
163
trogen and phosphorus risk, pesticide contamination risk, relative levels of habitat interven-
tion, changes in Carbon stock and greenhouse gases balance.
Land use is the most important factor that drives the environmental behavior of the region
(Table Nº 3). All indicators, from fossil energy consumption to contamination risk, from ero-
sion risk to greenhouse gases emission, are particularly sensitive to land use. Technology is
the next factor (Viglizzo, op. cit).
Table Nº 3. Changes in the area allocated to predominant annual crops (soybean, maize,
wheat and sunflower) in the Argentina Pampas during the period 1960-2000.
Percentage of the Total Area
Pampas Area / Year 1960 1988 1996
Regional Average 23.70 30.30 40.00
Rolling Pampas 28.90 47.60 63.40
Central Subhumid 31.30 38.30 53.60
Central Semiarid 21.70 38.40 39.10
Southern Pampas 23.40 32.40 36.80
Flooding Pampas 12.20 8.20 13.20
Mesopotamian Pampas 10.40 7.60 10.40
Source: National Program of Agro-Environmental Management, Argentina, 2002.
Trends in fossil energy consumption in The Pampas indicate that intensification is increasing
at high rates. Land productivity and fossil energy consumption have almost doubled in less
than ten years. This shows a direction of Argentinean agriculture towards a more intensive
model, departing from the traditional semi-intensive one. Rolling Pampas has an energy
budget that highly exceeds the whole Pampas average (Table Nº 4).
Table Nº 4. Energy productivity and fossil energy consumption in the Argentine Pampas dur-
ing the period 1960-2000. Comparison of trends among ecologically homogeneous areas.
Energy Productivity (Gj/ha/year) Fossil energy consumption (Gj/ha/year)
1960 1988 1996 1960 1988 1996
Regional Area 6.40 13.45 22.16 1.30 1.68 3.31
Rolling Pampas 9.03 24.11 31.92 1.27 1.95 3.79
Central Subhumid 6.39 14.40 25.59 1.88 2.00 3.81
Central Semiarid 2.75 4.33 8.43 1.19 1.88 2.68
Southern 5.44 11.23 19.48 1.15 1.78 3.12
Flooding Pampas 3.48 4.19 10.91 0.56 0.50 1.43
Mesopotamian 3.21 3.42 13.92 0.56 0.51 1.86
Source: National Program of Agro-Environmental Management, Argentina, 2002.
164
In a general context, the Pampas has not been fertilized till the beginning of the nineties. The
nutrient budget of the Pampas had some stabilization before this time, by the rotation of crops
and cattle, the most common production system in the area. But it was in the nineties when
the land use transformations and an increase in fertilizer use drove the Argentinean Pampas
into more intensive models that are typical of the northern hemisphere.
Soybean has had and will have an emblematic role in relation with nutrient balance, loss of
quality and richness of ours soils.
Each year the country exports with its grains a considerable amount of nutrients – especially
nitrogen, phosphorus and potassium – that in the process of intensification, are not replen-
ished. Argentina exports yearly around 3,500,000 metric tons of nutrients – with no recogni-
tion in the market prices, increasing the “ecological debt” (Martinez Alier and Oliveras,
2003). Soybean, the engine of this transformation, represents around fifty percent of this aver-
age. If we compensate the natural depletion with mineral fertilizers, Argentina will need
around 3,326,786 metric tons of nitrogen and phosphorous fertilizers and an amount of
900,000,000 American dollars to buy it in the market (Table Nº 5) (Pengue, W, 2003).
Table Nº 5. Stimation of nutrients (N, P) exportation and the cost for soybean harvest
2002/2003 (34,000,000 metric tons).
Nitrogen Phosphorous Total
Nutrient Extraction in metric tons 1,020,000 227,800 1,247,800
Equivalent in Mineral Fertilizers in metric tons 2,217,400 1,109,386 3,326,786
Cost Stimation reposition (US$) 576,524,000 332,816,000 909,340,000
Source: Pengue, W. La economía y los subsidios ambientales: Una Deuda Ecológica en la Pampa Argentina. Fronteras Nº 2:
7-8. Año 2. Number 2. GEPAMA.FADU.UBA. Buenos Aires. 2003.
Stimation for next season (2003/2004) considered that around 30% of the whole soybean area
(4,500,000 hectares) will be fertilized with mineral fertilizers. In 2002/2003 surface implanted
with soybean rose to 12,900,000 ha and next season the estimation is around 13,600.000 hec-
tares. The scenery shows a trend in important depletion of nutrients in ours soils that will be
consumed completely in 50 years (Ventimiglia, 2003).
Under No Tillage system, indicators show that soil erosion risk tend to decline. Although the
area cultivated with annual crops expanded, minimum and No Tillage practices compensated
the more intensive use of land.
The use of land and the expansion of the agricultural border is the most relevant impact factor
after losing biodiversity. Aggressive agronomic practices, global demand for soybean and at-
tractive prices without institutional national regulations and economic instruments, reinforce
the negative impact of intensive land use on habitats and biodiversity. Argentina is one of the
165
countries in South America that posses less territory as protected area (4.8% of 2,777,815
km2; others countries: Brazil 16.8%, Bolivia 22.4%, Peru 9.9%) (Burkart, 1999)
Final Comments
The intensification of Argentine agriculture, represented by the use of No-tillage practice and
glyphosate, has allowed for the homogenization of production based on transgenic soybean as
the dominant crop.
The present export-oriented commodity production system is most likely to drive more
smaller farmers out of business. For them, a diversification beyond global commodity mar-
kets, be they non-transgenic for export or other crops for internal purposes, might render an
alternative development trajectory. However, this would require a drastic turn in Argentine’s
agricultural policy, namely to play a more active role and to subsidize small-scale farmers.
The overwhelming domination of transgenic soybean makes farmers especially vulnerable to
changes in the global commodity markets due to the preoccupation regarding the safety of
GMOs.
About environment, first indicators show interesting impacts, some of them directly related to
soybean production and intensification practices. The transition towards a more intensive
model of agricultural production, both in terms of land use and technology application, has
characterized the nineties. Many farming systems in Argentina resemble some intensive mod-
els that are very common in the United States or Europe.
Nutrients depletion is a new complex discussion that must be solved with holistic agroproduc-
tive policies, not with current decisions as increasing of application mineral fertilizers.
Deforestation and expansion of agricultural borders must be analyzed fast and need policy de-
cisions to avoid an important loss of biodiversity and habitat.
Indicators showing a negative trend are important keys to identify critical problems that will
require a more specific attention of researchers and government. An increase in contamination
risks, fertilizer supplies, lost of biodiversity, and monoproduction must to be discussed to as-
sure the future sustainability of the Pampas.
References
Brodnig, G. Biotechnology in international trade. Weatherhead Center for International Af-
fairs. Harvard University, Set. 1999.
Burkart, R. Conservación de la biodiversidad en bosques naturales productivos del subtrópico
argentino in Matteucci, S and Solbrig, O. Comp. Biodiversidad y Uso de la tierra. Eu-
deba. Unesco. Buenos Aires, 1999.
166
CID – Center for International Development. Agricultural Research in Africa: Technological
opportunities and institutional challenges: Report of a Seminar. Center for International
Development. Harvard University. 1999.
James, C. The global review of commercialized transgenic crops. ISAAA, Brief paper – 2000,
Ithaca, N Y, USA. 2000.
Lehman, V and Pengue, W. Herbicide tolerant soybean: Just another step in a technology
treadmill? Biotechnology and Development Monitor Nº 43, September 2000.
Martinez Alier, J and Oliveras, A. Deuda Ecológica y Deuda Externa. Quién debe a quién?.
Icaria. Barcelona, 2003.
Morello, J et al. Argentina, granero del mundo ¿Hasta cuando?. Orientación gráfica Editora.
Bs. As. 1997.
Morello, J y Matteucci, S. Estado actual del subsistema ecológico del nucleo maicero de la
Pampa Húmeda. Capitulo Nº 4. En Argentina, Granero del Mundo, ¿Hasta cuándo?.
Orientación Gráfica Editora. Buenos Aires. 1997.
Pengue, W. The agriculture’s sustainability in Argentina. The fourth biennial meeting of
ISEE, Boston University. 1996.
Pengue, W. Sojas Transgénicas: Tecnología y Mercados. Realidad Económica. IADE. Buenos
Aires, Mayo, 1999.
Pengue, W. Cultivos Transgénicos ¿Hacia dónde vamos? Lugar Editorial. UNESCO, 208
páginas. Buenos Aires. 2000.
Pengue, W. Commoditación o diversificación de la producción agricola argentina. Transgéni-
cos. Jornada de Biotecnología en el agro. Universidad Nacional de La Plata. 2000.
Pengue, W. Impactos tecnológicos y ambientales de la liberación de OGMs. Conferencia In-
ternacional sobre Comercio, Ambiente y Desarrollo Sustentable. Programa de las Na-
ciones Unidas para el Medio Ambiente. PNUMA/UNEP Mexico, Febrero, 2001.
Prego, J et. al. El deterioro del ambiente en la Argentina (suelo-agua-vegetación-fauna). 3º
Edición. Fundación para la Educación, la ciencia y la cultura. Buenos Aires, 1997.
Sachs, J. The Economist. 1999.
UNDP. United Nations Development Programme. Human development reports 1999: Global-
ization with a human face. UNDP, New York. 1999.
UNESCO – United Nations Scientific, Educational, Scientific and Cultural Organization.
World Science Report. Paris. 1998.
Ventimiglia, L. El suelo, una caja de ahorros que puede quedar sin fondos. La Nación. Su-
plemento Campo. Pag. 7. Buenos Aires, October 18th. 2003
Viglizzo, E. F et al. La sustentabilidad ambiental del agro pampeano. Ediciones INTA. Bue-
nos Aires, 2002.
167
How may Quality Assurance Systems in food chains include environmental
aspects based on Life Cycle Methodology?
Niels Halberg
Danish Institute of Agricultural Sciences, Department of Agroecology, Research Centre Foulum, P.O.
Box 50, DK-8830 Tjele. Tel: +45 8999 1206 - Fax: +45 8999 1200. E-mail: [email protected]
Abstract:
The number of Quality Assurance Systems (QAS) for food products is increasing and the so is
the topics they cover, from traditional intrinsic product characteristics such as percent meat in
slaughtered pigs and protein content in milk to food safety issues such as zoonoses and pesti-
cide residues and in some cases aspects of animal welfare. This development is linked to de-
mands for risk controlling systems such as HACCP and traceability systems that would allow
food safety problems to be traced to a small number of producers or farms. The large retail
companies (supermarkets) are an important driving force for this development because of their
efforts to build consumer trust in food products and loyalty to the companies own brands. Envi-
ronmental characteristics of food products and information on their production methods are be-
coming part of some QAS but not mostly in the form of qualitative information e.g. certification
that the farmers have used Good Agricultural Practice (GAP). The paper gives examples of this
and then discuss this development in relation to LCA based environmental appraisal of food
products. The development of quantitative (tools for) environmental appraisal of agriculture and
food production is becoming more productoriented improving the possibilities of assessing the
regional and global impacts of food production chains and consumption. But these systems
building on LCA does not so far seem to be linked with the development of QAS for food. The
paper finally discuss the possibilities for linking the food safety related traceability systems and
gives an example of on-going work to establish LCA based QAS in a meat processing system.
Introduction
In the wake of the great European food scares of the 1990’ties a number of Quality Assurance
Systemes (QAS) have been introduced or improved by food processing and retail companies in
order to increase and regain consumers’ confidence in food products (Schiefer, 2004) and to se-
cure against liability from unforeseen food hazards. The more elaborate QAS build on a number
of safety and control measures at critical points in the production process to avoid contamination
and spread of food hazards. To secure transparency there is often an intensive information flow
and assurance of tracking and tracing of product components between different steps in the prod-
uct chain. Some slaughterhouses, for example, keep track of the meat quality from each primary
producer thus keeping records that allow to track batches of meat from supermaket back to a
very small number of farms. This increased interest in quality control and traceability has only to
a limited extent included externalities of the production such as the animal welfare and environ-
168
mental impacts through the production chain even though these may be considered as equally
important attributes of food products from a societal point of view.
Consumers appreciate a number of different quality aspects of food products among which are
both intrinsic and extrinsic. Following Steenkamp (1990) the intrinsic characteristics include the
organo-leptic or sensoric quality such as colour, taste, visible fat (called “quality cues” if they are
observable before purchase and “experience attributes” if they may first be ascertained while
consuming the product). Intrinsic characteristics that relate to food safety may often not be re-
vealed immediately (“credence attributes”) and it is off course no surprise that many consumers
have become interested in information regarding the risk of zoonoses, bacteria or other food haz-
ards when choosing food products. Extrinsic characteristics – besides price - relate to the condi-
tions of livestock in the production chain (animal welfare), to the resource use and environmental
impact from the production and to other aspects not observable from the product itself (e.g.
GMO free, organic production, regional product)(Brom, 2000; Verbeke and Viane, 2000).
It has been proposed that the environmental information from e.g. Life Cycle Assessment (LCA)
may be used to guide consumer choices (Nilsson et al., 2004) at least in public procurement
(Anonymous, 2002). The EU Integrated Product Policy (IPP) considers LCA as one of the cor-
nerstones (Anonymous, 2003) for improvement of the knowledge and transperancy concerning
the environmental impacts related to production and consumption. However, it is not clear how
significant environmental labelling is for consumers’ preferences and other aspects of food qual-
ity are probably more important (Brom, 2000; Verbeke and Viane, 2000; Nilsson et al., 2004). It
seems as if presently environmental aspects has less priority compared with food safety issues in
the development of traceability and documentation in QAS (maybe because food contamination
can have direct impact on specific consumers while most environmental issues have a less direct
impact on particular consumers using a specific product). But in a larger context the environ-
mental characteristics of food production chains are important because the food production and
consumption is one of the larger contributors to a family’s environmental impact (Anonymous,
1996; Wilting et al., 1999; Spangenberg & Lorek, 2002). Livestock products are particularly im-
portant for the emissions of nutrients and greenhouse gasses (see several papers in this volume)
and for land use (Gerbens-Leenes et al., 2002) and projections of global food demands and pro-
duction foresee a significant global rise in consumption of meat and milk (Delgado et al., 1999).
Therefore, it seems relevant to seek ways to improve environmental appraisal of agriculture and
food chains with the aim of reducing the environmental load per kg product produced and con-
sumed.
There are examples of larger food companies - such as Arla, Unilever and Cerelia - performing
LCA on specific projects (Larsson, 2004; McKeown, 2001; Rosing et al., 2004) either as part of
their product development or in order to be prepared against criticism from environmentally con-
169
scious consumer groups. Also, a number of food processing plants like slaughterhouses and dair-
ies have used energy accounting tools or other environmental management tools and some have
become ISO 14001 certified such as the Danish slaughterhouse “Tican”. However, often the
most significant environmental impacts from food production are happening in the primary pro-
duction rather than the processing stages. Therefore, attempts to describe, appraise and document
environmental characteristics of food products should include the whole production chain. A
number of European tools for farm level environmental appraisal and reporting exist for volun-
tary use, some of which are linked to advisory tools for farm planning (Halberg et al., 2004). But
these tools are seldom linked with the rest of the food chain and existing labels for environmen-
tally friendly production are usually not based on quantitative information of potential environ-
mental impact from the specific producers (Nilsson et al., 2004).
The aim of this paper is therefore
• to discuss trends in Quality Assurance Systems (QAS) used in European food chains and
the different methods used for appraisal of environmental characteristics of food products
• to propose that environmental appraisal as part of QAS should be based on quantified in-
formation through the product chain and be linked to traceability principles.
Quality Assurance Systems in European food sector
A number of QAS for food products exist in Europe and other countries differing in both their
organisational set-up and in the degree to which they build on international certification stan-
dards (Schiefer, 2004). Meat and animal products have been at the core of food scandals in the
last decade and not surprisingly some of the strongest QAS are to be found within pork produc-
tion, such as the Dutch IKB (Trijp et al., 1997), a relatively new German system called QS
(Quality and Safety for Food Products, from Producer to Consumer" (Nienhoff, 2004; www.q-
s.info) and the “Danish” brand (Anonymous, 2003). These management schemes aim at securing
a high degree of traceability through a strong information flow where results from tests and
measurements along the production chain are continuously fed backwards with the aim of im-
proving performance (e.g. feeding back information to farmers on carcass quality or bacterial
counts in milk). Most often this relates to analyses of ”classical” quality parameters such as fat
content, freezing point and bacterial counts and for medicine residues and other more recent tests
such as Aflatoxines. Thus, most often QAS builds on two key concepts: Traceability and
HACCP. HACCP (Hazard Analysis and Critical Control Points (CFSAN, 2004; Danske Slag-
terier, 2003) is a systematic approach to monitoring quality and risks of contamination of food
products during a production process and is standardised among others in Denmark and Holland
and internationally in ISO 22000.
Moreover, increasingly the QAS seek to include whole product chains in order to improve cross-
border trade and share information not only backwards/upstream but also downstream in the
170
chain as documentation coming with the product. However, when it comes to information re-
garding the extrinsic characteristics such as animal welfare and environmental performance such
information is presently not accumulated and exchanged through the value chain in a quantitative
form. The QS system for example includes demand for compliance with other German regulation
of manure use but no quantification of nutrient surplus or losses. Likewise, the French QAS
“Agri Confiance” for certification of a variety of livestock products and processed crop products
has only recently started to include environmental aspects in a separately managed ISO 14001
scheme called Agri Confidence® Quality-Environment (www.cooperation-agricole.asso.fr cit.
Cederberg, 2004). The intention of labels such as IKB are to add value to a product (pork) in the
eye of the consumer by guaranteeing that it is produced according to some criteria, which are be-
lieved to be important (Trijp et al., 1997). This added value should then give an advantage in the
form of either higher prices or increased consumer loyalty to products labelled with the IKB
brand.
In addition to the brands and QAS of the food processing industry large retail companies also
begin to focus on traceability and documentation of the origin and quality of the products they
sell. There seems to be a trend towards non-price competition (competing on the products’ qual-
ity attributes) and building consumer loyalty towards the retail companies’ own brands. An in-
creasing proportion of the retail sale of food and household commodities happens under the retail
companies’ own brands (Arfini and Mancini, 2004). The British Tesco and Sainsbury for exam-
ple both have their own brands that account for over 50% of their total sales, the French Carre-
four and Intermarche sell 20 and 29% of their turnover in their own brands and this proportion is
increasing also in the Nordic COOP chain. This has economical advantages for the retail busi-
ness and gives them stronger control and flexibility (e.g. they may change their suppliers without
the consumers noticing). To minimise the risk of loosing consumer confidence in these trade-
marks due to food scandals the retail business is now very active in quality assurance and there-
fore demand quality control measures and traceability backwards in the food chain.
Arfini and Manicini (2004) studied the British Retail Consortium (BRC) as an example of this
involvement. The BRC is an association of major retail chains and distribution companies in the
UK and has as its major function to translate consumer demands and interests into demands for
the products’ characteristics and performance through the supply chain. Thus, the BRC’s so-
called “Technical standard and protocol for companies supplying retailer branded food products”
include demands that companies establish hygiene and safety control systems based on the
HACCP method. Suppliers should also adopt a documented quality management system includ-
ing requirements of minimum levels and recommendations on good practice, following a stan-
dard such as EN45011, which is basically in compliance with ISO 9001. This way (expected)
consumer demands concerning food safety and product quality have been translated into contrac-
171
tual requirements that suppliers of food products (also foreign) have to comply with when deal-
ing with the members of BRC, which is the majority of the large retail chains in the UK.
According to Krieger and Schiefer (2004) the primary agricultural production (e.g. the pig fatten-
ing facility) will become integrated in HACCP systems in the future and HACCP systems al-
ready exist for fruits and vegetables (Hernandez-Souchez et al., 2004). However, the authors do
not find it likely that these quality assurance schemes will be used to claim higher prices for the
certified products. Rather these concepts will be considered the standard or basic quality for a
number of food items. Contrary to this Broom (2000) argues that food safety may be assured by
labelling schemes if they are backed by government control systems. Other issues that are not
necessarily relevant to all consumers in their role as consumers - such as animal welfare or envi-
ronmental issues – may also be assured by labelling. The European Commissions study on “envi-
ronmental product declaration schemes” (Anonymous, 2002) advise that information on the envi-
ronmental performance of food products be considered in e.g. public procurement. Nilsson et al.
(2004) find that existing labelling schemes often lack credibility in the sense that “they are per-
ceived interesting and trusted by consumers” and should therefore be backed by a more factual
appraisal of the actual production methods and their environmental impact. Whether the QAS
will be used to give information to consumers or will remain primarily a process between agents
in the food processing chain is not clear, but in both cases it seems relevant to discuss the poten-
tial role of LCA or other forms of environmental assessment to supplement such systems.
Good Agricultural Practice and environmental Quality management of food products
Some initiatives aim at securing minimum standards for the environmental performance of agri-
cultural products, especially through the establishment of certification schemes1. A major objec-
tive among retail companies is to avoid pesticide scandals caused by either too high amounts of
residues in the products or hazardous use by the farmers. This is primarily secured by either ban-
ning certain pesticides in specific products or demanding rules for pesticide use and storage, see
below. Other aspects are included such as the prudent use of water and fertiliser but most often in
a non-quantitative form, which seem less rigorous than for the above mentioned meat quality and
food safety issues. Therefore, it is relevant to discuss the possibility of using quantified informa-
tion based on actual use of inputs and/or estimated emissions in the environmental assessment of
food products.
A number of food products are produced and sold under labels claiming some form of environ-
mental consideration. One example is the bread wheat and rye sold under the NATUR+ label
1
Certified organic farming is off course also a certification that certain practices have been followed and that no
pesticides have been used but it does not quantify the environmental impact as such and will not be considered here.
172
(Cederberg, 2004; Swedish Seal, 2004) owned by the Swedish farmers and used for most bread
and flour sold in Scandinavian supermarkets such as COOP in Denmark and Sweden. The
NATUR+ label guarantees that no chemical “plant growth regulators” have been used (following
specific worries for food safety of this otherwise legal crop treatment) and the rules also ban pre-
harvest Round-up use and use of sewage sludge. It may be discussed whether these rules in real-
ity address a consumer concern for chemical residues in the bread rather than care for the envi-
ronment, but recently also rules for fertiliser planning and minimum requirements for “green
zones” on the farm have been included.
Under the British Farm Standard logo of a “Little Red Tractor” exists a number of guidelines for
assured production with rules of how farmers should take environmental considerations in their
planning and management. One example is the Assured Produce Scheme (APS), which promotes
safe and environmentally responsible production of fruit, salads and vegetables through the use
of integrated crop management (ICM). According to the home page (Anonymous, 1993) APS “is
designed to maintain consumers' confidence in the safety and integrity of the produce they eat”.
Growers must follow the best production advice contained in the crop specific protocols that
form the basis of the scheme. For example, the use of fertiliser should be based on crop norms
and soil analyses and pesticides should only be used after observation of a critical level of a pest
in the crops. APS thus follows a logic of Good Agricultural Practice (GAP) and is crop/field
level based, not product oriented. APS is an independently assessed assurance scheme and farm-
ers have to be certified and inspected to sell products labelled with the Little Red Tractor. Other
QAS under the little red tractor cover e.g. pigs (Assured British Pigs) and chicken (Assured
Chicken Production).
The APS is one of many examples of a labelling scheme certified under the umbrella EurepGAP,
which is an initiative owned by a consortium of European retail companies (supermarket chains).
The EurepGAP is based on the socalled FoodPLUS / STATUTES which have the objectives
to:“Encourage adoption of commercially viable Farm Assurance Schemes, which promotes the
minimisation of agrochemical inputs, within Europe and world wide.
Develop a Good Agricultural Practice (GAP) Framework for benchmarking existing Farm As-
surance Schemes and Standards including traceability.” (EurepGAP, 2003).
The Danish IP label (Integrated Production) for vegetables (outdoor as well as greenhouse crops)
is owned by an independent group of horticulturalists organised under producer organisations
(GAU/DEG/GASA) (Anonymous, 2004a). Danish IP is based on the idea of promoting the use
of good crop rotations and other preventive measures to reduce the need for pesticides as much
as possible. As an example the producers of IP tomatoes and cucumbers have to record and
document that they purchase and use biological control of pests. It was originally the hope
among the initiating producers that the IP label would qualify for a price premium but this has
173
not been realized. However, the Danish IP label is credited for the relatively high proportion of
Danish produced vegetables sold in supermarkets. The Danish IP is currently undergoing ad-
justments to comply with the EurepGAP standards. This implies some changes in the level of
documentation and in specific rules for e.g. storage of pesticides but not in the actual environ-
mental performance of the farms. This is because the EurepGAP standards do not include spe-
cific quantified limits for e.g. fertiliser use or environmental impact.
Environmental assessment as part of quality assurance schemes
Why should the food business use precise documentation of environmental impacts of food
products? There are a least two reason why documentation of environmental characteristics
should be included in QAS. The first reason is related to the interests of the brands and food
companies and the second is related to the societal interests in environmental improvements in
the food chain.
1. The advantages for the food companies of branding builds partly on the ideas that the
perceived better quality associated with the brand and other brand associations increases
the consumers experience and satisfaction with a product (van Trijp et al., 1997). As dis-
cussed in the introduction “quality” is more than the intrinsic characteristics of a product
and includes external characteristics such as environment and animal welfare aspects of
the production process. Therefore, it may be an advantage towards at least some parts of
consumers to have documentation that environmental care is part of the brand policy. Or,
at least the retail companies and the brand owners should try to avoid critical stories con-
cerning the environmental impact from their suppliers, e.g. damaging losses of critical
pesticides.
2. From a broader environmental perspective there could be an advantage of including envi-
ronmental assessment in QAS because important environmental impacts from food con-
sumption are regional and global such as nutrient losses and greenhouse gas emissions.
The larger part of these environmental impacts happen in the primary production and
there seems to be a potential for improvement as demonstrated by both LCA studies (de
Boer, 2003; Haas et al., 2000; Halberg, 1999; Erzinger, 2004) and the tests of farm level
green accounts (Halberg et al. in press). To record and report environmental information
would facilitate better control, regulation and improvements based on incentives from
e.g. the retail companies with their own brands. Following the ideas of the IPP informa-
tion regarding the environmental impacts accumulated through the food processing chain
would also facilitate better choices among retail companies and professionals in kitchens
and restaurants.
As mentioned above, quality parameters like carcass quality is communicated up- and down-
stream in the food chain partly as quantitative information but the environmental information - if
174
used at all - is based mostly on GAP, adherence to decision rules etc., not on the actual result,
e.g. resources used per kg product or LCA type of information. It should however be possible to
use quantified environmental assessments because this type of information is the baseline of
many types of green accounts for farms in Europe (Halberg et al., 2004), but these are most often
not integrated with the product chain QAS. Nilsson et al. (2004) analysed the credibility of 58
eco-labelling schemes and conclude that presently ”There is no labelling system that covers the
entire food production chain which could install ecoefficiency in the production chain”. The au-
thors call for an “alternative approach [that] could measure appointed quality aspects in indica-
tors for the whole food product chain and report them to interested parties and consumers” (Nils-
son et al., 2004).
Some of the objectives for including environmental characteristics in QAS may be fulfilled by
GAP type rules to be followed by the farmers, e.g. certification that they only use legal pesticides
and only after pest-infections above a certain threshold have been observed by the farmer in his
fields (such as is the generic rules in EurepGAP standards). From the point of view of the retail
sector this approach seems to limit the risk of food scandals caused by misuse or overuse of pes-
ticides by the primary producers. They may claim that they have done their best to limit this and
thus avoid liability in case of such a case becoming public.
However, this rule based method for environmental QA does not significantly document any im-
provements in environmental performance compared with standard practice, especially not in
countries with a high standard for public regulation. Rules for GAP that simply demand the
farmer to make a fertiliser plan or to use pesticides only after inspection in the crops do not dis-
tinguish between farmers who use only small amount of pesticides or fertilisers (e.g. because
they have a better crop rotation) and those who in reality rely on standard dosages. Halberg et al.
(2004) compared different European concepts for farm level environmental appraisal (Input out-
put accounting, Green Accounts etc.) and found a similar distinction between management (rule)
based indicators and quantitative indicators based on results of environmental performance on
farms (e.g. nutrient surplus per ha or energy use per kg product). It was concluded that the results
based indicators were more suitable to link with advisory tools for improving farm performance
based on e.g. benchmarking. Benchmarking is here understood as “the process of improving per-
formance by continuously identifying, understanding and adapting outstanding practices and
processes found inside and outside the organisation” (Amer. Prod & Quality Center, 1999, cit.
EEA, 2001). In other words, benchmarking is to compare one’s own results with other produc-
ers’ performance and thereby identifying ”best practices” among comparable producers. The
process also involves the tasks of understanding these differences, thus learning from others and
using this to set goals for one self and the engage in activities to improve one’s own practices. To
perform benchmarking and facilitate improvements there is a need for quantitative assessments
of environmental characteristics of food products based on the actual processes and resource use.
175
The quantitative, results based environmental information also has the advantage that they may
describe the environmental impacts accumulated through the production chain, such as energy
use and nutrient losses per kg product using LCA methods (as demonstrated in several papers in
this volume). This would facilitate a product based environmental appraisal of food products in
line with ideas of IPP (Anonymous, 2003). Therefore, it should be recommended to use quantita-
tive environmental information based on the actual results from the production processes as part
of QAS in the future. The existing LCA methodologies could be used as a starting point for this
merge between IPP and QAS and supplemented with processes to include environmental aspects
in product development (Nielsen & Wenzel, 2002).
To our knowledge the Dutch system “MPS” (Anonymous, 2004b) for green house production of
flowers is the only environmental QAS for agricultural products which is both accredited by the
retail sector in Europe, and builds on quantified information in a way that would fit in a product
oriented environmental appraisal building on IPP/LCA principles. The MPS is accredited under
EurepGAP but unlike most other GAP approaches this system is based on a quantification of the
actual use of energy, pesticides and fertiliser per batch of flowers finished by the certified grow-
ers. Thus MPS is an example of how environmental QAS may be used for both documentation
towards retail chains and benchmarking between growers. Some of the major retail chains have
recently demanded MPS certification of flowerpots for their stores. The growers under MPS
have to report their use of energy, pesticides and fertiliser every four weeks to the MPS organisa-
tion, where quarterly reports are compiled and sent out to growers with comparisons of their per-
formance against standard limits and results of other growers. If a grower uses more of one input
factor than the standard different measures are taken – depending on the severity of the exceed-
ing and ultimately a grower may loose the certification for a period of time. The MPS approach
demonstrates the feasibility of benchmarking between producers but the pot flower enterprises
have a relatively short production chain from grower to retail compared with e.g. livestock prod-
ucts. However when looking at the existing dataflow on product quality in the pork sector includ-
ing environmental data does not seem impossible from a practical viewpoint.
Fig. 1 gives an overview of the idea of creating an information flow on environmental profiles of
livestock products along the physical flow of agricultural products from primary producers to re-
tail. This idea is presently being tested in a case study involving a private Danish slaughterhouse
and a number of its major suppliers of fattening pigs. The involved farms will establish Life cy-
cle based green accounts of their production of fattening pigs with the help of local production
advisors. These will be collected at the slaughter house for two purposes: 1.The accounts from
different farms will in anonymous form be compared and fed back to farmers in a benchmarking
exercise, where each farmer may assess his environmental performance in comparison with other
producers delivering to this specific company. 2. The information on resource use and emissions
176
in the primary production will be supplemented with environmental information from the slaugh-
tering and other processes including transport to give an environmental profile per kg product de-
livered to the retailers and professional kitchens. It will be part of the project to explore which
type of information the professional buyers will be interested in and how to convert LCA type of
information into a format that is understandable for these stakeholders. It is not the idea to pre-
sent this information to ordinary consumers because the LCA based information itself is assumed
to be too complicated for laymen to relate to in the purchasing situation.
Conlusions
Rule based environmental quality assurance based on GAP is becoming part of the overall QAS
in the food sector. The GAP approach may give some quality assurance for the food retail sector
helping to reduce risks of food scandals from e.g. pesticide misuse analogously to HACCP sys-
tems reducing other food hazards. But the GAP approach found in most environmental QAS
presently does not satisfactorily quantify the actual environmental performance on the farms nor
does it allow benchmarking between farms, supplier cooperatives and products. Moreover, the
GAP approach is not suitable for a product-oriented appraisal of environmental impacts from
food products in line with the increasing interest in Integrated Product Policy. There is a need for
development of an environmental quality assurance scheme that records and exchange informa-
tion up- and downstream in the food chain and allows both primary producers and the food in-
dustry to continuously benchmark their performance and the retail sector to assure their custom-
ers that products are environmentally sustainable. Life Cycle Assessment and Life Cycle
Management seems obvious tools for this and will be used in an attempt to develop such a sys-
tem within a Danish slaughterhouse company.
177
Suppliers of
feed and fertilisers
Farm 1 Farm 2 Farm 3
Processing plant
(slaughterhouse, dairy)
Cooperative or private
Products and
resources Processing
prepared meals
Information
2 2
Transport and traders
Professionel
Retail
kitchens
Consumers
1 Input/output accounts on farm level transformed into environmental profile per kg product
from farms. Data for comparisons between farms re. ressource use and emissions per kg
product, (benchmarking).
2 Documentation of environmental profiles of products accumulated through the chain in
life cycle terminology
Figure 1. Simplified description of a food chain with physical flow of products and
(potential) flow of information regarding quality and enviromental characteristics.
178
References
Anonymous, 1993. Assured produce scheme, available on-line (sept. 2004) at:
http://www.assuredproduce.co.uk/aproduce/
Anonymous, 1996. Miljøbelastningen ved familiens aktiviteter (Environmental impact from the
family’s activities), Danish Consumer Board, report 1996. 1, 78 pp.
Anonymous, 2002. Evaluation of Environmental Product Declaration Schemes. Final Report.
European Commission, DG Environment, 120 pp.
Anonymous, 2003. Integrated Product Policy – Building on Environmental Life-Cycle Thinking.
Communication from the commission to the council and the European Parliament. Com-
mision of the European Communities, Brussels, 18.06.2003,. COM (2003) 302 final.
Anonymous, 2004a: Danish IP: Gartneribrugets Afsætningsudvalg (GAU) Available on-line
(sept. 2004) http://www.dansk-ip.dk.
Anonymous, 2004b: Miljø Programma Sierteelt. Available on-line (sept. 2004): www.my-
mps.com
Arfini, F., Mancini, M.C, 2003. British Retail Consortium (BRC) Standard: a New Challenge for
Firms Involved in the Food Chain. Analysis of Economic and Managerial Aspects. (ed.) G.
Schiefer, U. Rickert. Proceedings of the 82nd Seminar of the European Association of Agri-
cultural Economists (EAAE), May 14-16, Bonn, Germany. Vol. A, 23-31.
Brom, F.W.A., 2000. Food, consumer concerns, and trust: Food ethics for a globalizing market.
Journal of Agricultural and Environmental Ethics 12, 127-139.
Cederberg, C., 2004. Systems for assurance of farm production in Western Europe. In. Thodberg
et al. Produktorienteret miljøindsats i landbrugssektoren - forudsætninger og fremsyn. Re-
port no. 900. App. E, p. 185-204 (in English). Available on-line (Sept. 2004) at
www.mst.dk/hompage (publications)
Cederberg, C., Mattsson, B., 2000. Life Cycle Assessment of Milk Production - A Comparison
of Conventional and Organic Farming. J. Cleaner Prod. 8, 49-60.
CFSAN, 2004. Center for Food Safety & Applied Nutrition. Hazard Analysis and Critical Con-
trol Point. HACCP overview. Availble on-line (Sept. 2004):
http://vm.cfsan.fda.gov/~lrd/haccp.html
Dalgaard, T., Halberg, N. and Porter, J.R., 2001. A model for fossil energy use in Danish agricul-
ture used to compare organic and conventional farming. Agriculture, Ecosystems & Envi-
ronment 87(1): 51-65.
Danske Slagterier, 2003: Description of the 'Danish Quality Guarantee', HACCP and traceability.
Available on-line (Sept. 2004):
http://www.danskeslagterier.dk/smcms/Danish_English/Danish_Quality/Index.htm?ID=1968
De Boer, I.J.M. 2003. Environmental impact assessment of conventional and organic milk pro-
duction. Livestock Production Science 80: 69-77.
179
Delgado, C., M. Rosegrant, H. Steinfeld, S. Ehui, and C. Courbois, 1999. Livestock to 2020: The
next food revolution. IFPRI. Discussion paper 28, 70 pp. Available on-line (August 2004):
http://www.ifpri.org/2020/dp/dp28.pdf
EEA, 2001. Environmental benchmarking for local authorities: From concept to practice. Envi-
ronmental issue report no. 20. European Environmental Agency.
Erzinger, S. and Badertscher Fawaz, R., 2001. Life Cycle Assessment of animal housing systems
as part of an overall assessment. Proceedings International Conference on LCA in Foods, p.
118-122. SIK-Report Nr. 643. Swedish Institute for Food and Biotechnology, Gothenburg.
EurepGAP, 2004: FoodPLUS principles and generic rules for GAP, available on-line (sept.
2004) at: http://www.eurep.org/sites/index_e.html
Gerbens-Leenes, P.W., Nonhebel, S., 2002. Analysis – Consumption patterns and their effects on
land required for food. Ecological Economics 42, 185-199.
Grønborg Eskesen, 1999. Danske gartnerier og MPS. Gartneritidende uge 29-30, 2003. Tilgæn-
gelig online: http://www.gartneritidende.dk/artikler.
Haas, G., Wetterich, F., Geier, U., 2000. Life Cycle Assessment Framework in Agriculture on
the Farm level. Int. J. LCA 5, 345-348.
Halberg, N., 1999. Indicators of ressource use and environmental impact for use in a decision aid
for Danish Livestock Farmers. Agriculture, Ecosystems and Environment 76, P. 17-30.
Halberg, N., Verschuur, G. & Goodlass, G. 2004. Farm level environmental indicators; are they
useful? An overview of green accounting systems for European farms. Agriculture, Ecosys-
tems and Environment. In Press.
Krieger, S., Schiefer, G., 2003. Quality Management Schemes in Europe and Beyond. In: G.
Schiefer, U. Rickert (eds.). Proceedings of the 82nd Seminar of the European Association of
Agricultural Economists (EAAE), May 14-16, Bonn, Germany. Vol. A, 35-47.
Larsson, I., 2004. Possible of the benefits from using LCA in the agro-food chain, example from
Arla Foods. In: Halberg, N. (ed.) Life Cycle Assessment in the Agri-food sector. Proceed-
ings of the 4th International Conference, October 6-8, 2003, Horsens, Denmark.
McKeown, P.J., 2001. The application of LCA methodology to the life of frozen vegetables. In:
Proceedings of International Conference on LCA in Foods, Gothenburg, Sweden, 26-27
April 2001. SIK-document 143, 83-87.
Nienhoff, H.J., 2004. QS Quality and Safety: A Netchain Quality management Aggroach. In: G.
Schiefer, U. Rickert (eds.). Proceedings of the 82nd Seminar of the European Association of
Agricultural Economists (EAAE), May 14-16, Bonn, Germany. Vol. A, 627-629.
Nilsson, H., Tuncer, B, Thidell, Å., 2004. The use of eco-labeling like initiatives on food prod-
ucts to promote quality assurance – is there enough credibility? Journal of Cleaner produc-
tion 12, 517-526.
Rosing, L. and Nielsen, A.M., 2004. When a hole matters - the story of the hole in a bread for
French hotdog. In: Halberg, N. (ed.) Life Cycle Assessment in the Agri-food sector. Pro-
ceedings of the 4th International Conference, October 6-8, 2003, Horsens, Denmark.
180
Schiefer, G., 2004. From Enterprise Activity ‘Quality Management’ to Sector Initiative ‘Quality
Assurance’: Development, Situation and Perspectives. In: G. Schiefer, U. Rickert (eds.).
Proceedings of the 82nd Seminar of the European Association of Agricultural Economists
(EAAE), May 14-16, Bonn, Germany. Vol. A, 3-20.
Spangenberg, J.H. & Lorek, S., 2002. Environmentally sustainable household consumption:
From aggregate environmental pressures to priority fields of action. Ecological Economics,
43, 127-140.
Steenkamp, J-B. E.M., 1990. Conceptual Model of the Quality Perception Process. Journal of
Business Research 21, 309-333.
Steenkamp, J-B. E.M., 1997. Dynamics in consumer behavior with respect to agricultural and
food products. In: B. Wierenga, A. Van Tilburg, K. Grunert, J-B. E.M. Steenkamp, M.
Wedel. (eds.). Agricultural Marketing and Consumer Behavior in a Changing World. Chap-
ter 8, 143-188.
Swedish Seal, 2004. About Swedish Seal of Quality. Available on-line (Sept. 2004):
www.svensktsigill.com.
Van Trijp, H.C.M., Steenkamp, J-B. E.M., Candel, M.J.J.M., 1997. Quality labelling as instru-
ment to create product equity: The case of IKB in the Netherlands. In: B. Wierenga, A. Van
Tilburg, K. Grunert, J-B. E.M. Steenkamp, M. Wedel (eds.). Agricultural Marketing and
Consumer Behavior in a Changing World, 201-216.
Verbeke, W.A.J., Viaene, J., 2000. Ethical challenges for livestock production: Meeting con-
sumer concerns about meat safety and animal welfare. Journal of Agricultural and Envi-
ronmental Ethics 12, 141-151.
Wilting, H.C., Moll, H.C., Nonhebel, S., 1999. Direct and indirect energy use of European
households. Groningen: IVEM, University of Groningen. IVEM report.
181
Sustainability in the Agrofood Sector
Klaas Jan Kramer & Marieke Meeusen
Agricultural Economic Research Institute (LEI), PO Box 29703, 2502 LS The Hague, the Netherlands
Phone: +31-70-3358245, Fax: +31-70-3615624, E-mail: [email protected];
[email protected]
Abstract
Sustainability is a hot issue. Society is demanding products and services that are produced and
delivered in a sustainable manner. However, the concept of sustainability is very broad and
needs concretising. To determine sustainability in agricultural product chains, LEI has devel-
oped a sustainability checklist. Elements of the checklist were used in the development of an
annual social & environmental report on the Dutch horticultural sector. The availability of infor-
mation/data and their application as indicators are a main factor in the determination of sustain-
ability aspects. Analyses of chain information systems showed that these systems provide sev-
eral linkages which are helpful in determining sustainability
Keywords: sustainability, checklist, indicators, information.
Introduction
Climate change, biodiversity, economic growth, human rights, poverty, labour conditions, etc. –
all these factors are considered important regarding the sustainable development of our society,
where such development means fulfilling the needs of the present generation without endanger-
ing the fulfilment of the needs of future generations. Companies are entering a new phase in
which integrated responsibilities for mankind, the economy and the environment are becoming a
prerequisite for good entrepreneurship. In other words, they are approaching sustainability from
the perspective of people, planet and profit (the three P’s).
The sustainability issue is increasingly under scrutiny at the chain level. The Dutch government
is encouraging the agrofood business sector to take innovative steps towards sustainable devel-
opment from the chain point of view. To gauge the sustainability of chains, the Agricultural
Economics Research Institute (LEI) is involved in several research projects to determine and
measure the progress of sustainability in agrofood chains. This paper presents some of the latest
results of our sustainability chain research.
Motive
Indicators are a useful instrument to determine the level of a product’s or a company’s sustain-
ability. Although there is still no precise definition of sustainable development, it is possible to
begin working on indicators. The first question to ask is which issues and topics are involved in
sustainable agrofood chains. The next question is how the sustainability of chains can be deter-
182
mined in a relatively straightforward manner. A set of simple instruments would be a helpful tool
in this. These tools should preferably contain information that is already available. As a result,
insight will be obtained into the sustainability of agrofood chains relatively quickly and without
incurring excessive costs.
LEI was asked by the Dutch Ministry of Agriculture to draw up a checklist of issues or indicators
to determine the sustainability of agrofood chains. LEI was also asked to provide an insight into
the availability of the data used to determine sustainability. In this article we summarise the find-
ings of several LEI projects aimed at determining sustainability with the help of chain informa-
tion systems.
Aim
The aim of this paper is to provide greater insight into the possible ways of measuring the sus-
tainability of agrofood chains. There are two distinct steps in this, viz. to establish:
• which indicators, topics and items are involved in sustainable agrofood chains, and
• what information is already available in order to make a relatively quick and cheap
measurement.
Checklist for the sustainability of agrofood chains
Instruments that address the comprehensive scope of the definition of sustainability were ex-
pressly requested. These instruments were to elaborate on the three P’s, viz. people (the socio-
ethical component), planet (the environmental component) and profit (the economic component).
In addition, they were to focus on the aggregate nature of the industry with respect to sustainabil-
ity. They also had to take account of the consumer stage and other production chains. A fourth
request was to respond to the challenge of harmonising potential within the business community.
The instruments were to provide added value to current practice and to stimulate enterprises to
become more aware, but they were not to set unreachable goals. These requests were the starting
point for the checklist that was developed. This checklist provides an overview of all the catego-
ries and items that jointly determine the sustainability of agrofood chains.
An inventory of existing initiatives showed that although there are a great many of them they (a)
have very little connection with each other, (b) generally have no bearing on the analysis level of
chains, (c) have a specific bearing on the agricultural sector in only a limited number of cases,
(d) pay little attention to the mutual trade-off between the various sustainability aspects, and (e)
lack any conceptual or theoretical basis. An exception to this is the Sustainability Score Card, in
which there is also an important connection with the private sector. However, the Score Card in-
strument is not made available to third parties. The Sustainable Corporate Performance model
developed by the University of Groningen is also worthy of mention, but in 2002 no significant
efforts had been made within this model to define indicators. The Life Cycle Analysis (LCA)
183
was selected as one of the few instruments considered suitable for measuring environmental as-
pects along the whole chain.
In short, no system of measurable, scientifically-based indicators suitable for determining sus-
tainability – either in the broadest sense (i.e. covering the three P's) or more narrowly – had been
developed for agrofood chains. It was also apparent that the private sector was not ready for the
far-reaching operationalisation of the concept. Although many companies are devoting attention
to the concept of sustainability, they do not have a detailed operational plan for collecting peri-
odic and standardised data. Therefore a checklist format was chosen, to be based on an inventory
of many initiatives. The following aspects were examined: (a) the chosen system and set-up of
measuring instruments, and (b) the chosen topics and the underlying issues.
The first version of the checklist for the measurement of sustainability in agrofood chains con-
tained not only the outcomes and all topics but also a section on ‘vision/mission/strategy’ and
one on ‘measures’. This was intended to enable the measurement of the efforts and intentions
behind sustainability. The third section comprised the ‘outcomes’. Table 1 – which is a part of
the AKK/LEI sustainability checklist (Meeusen & ten Pierick, 2002) – presents some of the re-
sults of the top-down approach.
In 2003, LEI began work on the second version of the checklist. A good basis was sought to
serve as a conceptual framework for its development. This basis was found in Wood's Corporate
Social Performance model (1991). LEI is now using this to draw up the new checklist. Here,
links are made chiefly with the work of the Global Reporting Initiative (GRI).
Application: annual social & environmental report on the Dutch horticultural sector
LEI developed an annual social & environmental report covering 2002 for the Department of
Greenhouse Horticulture of the Dutch Agricultural and Horticultural Organisation. To decrease
the environmental impact of the Dutch greenhouse horticulture sector, the government and the
sector drew up the voluntary agreement ‘Greenhouse horticulture and the environment’. This
agreement contains ambitious objectives for 2010 concerning increasing energy efficiency and
decreasing the use of minerals and pesticides.
In 2000, employers’ organisations and trade unions signed a declaration concerning the devel-
opment of a voluntary agreement in the field of labour conditions. This agreement aims at de-
creasing the physical impacts (the major cause of absence due to illness and disablement) and
improving the process of reintegration.
184
Table 1. Part of the LEI/AKK sustainability checklist (Meeusen & ten Pierick, 2002).
Dimension Category Aspect Explanation
Planet Transportation Reducing freight trans-
port
Energy Reducing energy use
Renewable energy Promoting its use
Materials Reuse of materials
Water Water quality Reducing emissions
Air Air quality Reducing emissions
Fauna Biodiversity Preventing the reduction in diver-
sity of sorts and types of animals
Profit Costs and effi- Price/quality ratio Increasing the price/quality ratio of
ciency products and services
Ethics in business- Control and certification Checking whether demands have
to-business context been met
Employment Quantity of employment Increasing the number of jobs
People
Working conditions Workplace Improving the location, interior (er-
gonomic) and safety
Food safety Reducing food-pollution compo-
nents
Norm and values Emancipation Stimulating integration of the eld-
erly, handicapped, immigrants,
women, etc.
Social responsibility Welfare Contributing to the health, housing,
safety, education, etc. of the com-
munity
The Dutch greenhouse horticultural sector has made some significant improvements in the field
of sustainable production. The annual social & environmental report presents the steps taken by
the sector to reduce its environmental impact and to become more socially sustainable. The 2002
social & environmental report partly covers some elements of the checklist and concerns the fol-
lowing topics: energy use, pesticides use, mineral use, environmental labelling, lighting, waste,
labour, education, labour conditions and labour circumstances. These topics consist of various
indicators. Figures 1 and 2 show the development of two different indicators: energy efficiency
(energy per unit of production) and labour supply.
185
100%
90%
80%
70%
60%
50%
40%
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
year
Energy efficiency (1980=100% )
Figure 1. Development of the energy efficiency in the Dutch horticultural sector (van der Knijf
& Benninga, 2003).
6
5
4
% absence
3
2
1
0
1998 1999 2000 2001 2002
Netherlands Agriculture Greenhouse horticulture
Figure 2. Development of absence due to illness (Netherlands Ministry of Social Affairs and
Employment, 2002).
Information available for determining sustainability
This section 1) presents the results of research into the possibilities to use data from the Farm
Accountancy Data Network (FADN) in determining sustainability, and 2) describes research into
the possibilities to combine sustainability and chain information systems. In the latter project, the
focus was on how chain information systems can be used to determine sustainability.
186
The FADN
The next step in developing the checklist was to formulate indicators. Based on several selection
criteria, the appropriate indicators for sustainable agricultural chains were formulated. Selection
criteria for good indicators were, for example, scientific validity, communicational quality,
availability of data, representativeness, reproducibility, etc. In a LEI project carried out after in-
dicators for sustainability had been developed, several possible indicators were judged with the
help of several selection criteria. For the aspects of data availability, the possible indicators were
judged on the information contained in the Farm Accountancy Data Network (FADN). Various
data from a random sample of agricultural and horticultural holdings are stored in the network.
Most aspects related to the economic dimension of sustainability can be determined, sometimes
with another calculation step, with the help of the information contained in the FADN.
Possible indicators in the economic dimension are, for instance, number of jobs, financial results,
investments in capital, R&D, human capital, and certification. Possible indicators in the social
dimension are, for instance, number of registered complaints, number of days of illness (%), em-
ployee education/courses, contribution to local economies, and nature conservation. The FADN
also contains information that can be used to determine some of the environmental aspects of
sustainability, such as the total use of energy, mineral leaching, toxic emissions, waste, and land
use.
In short, the FADN provides a sound basis for linking data with indicators. However, the empha-
sis here lies chiefly on the primary production phase; the FADN contains less information with
which to determine the social aspects of the sustainability of agricultural holdings or chains. A
second point is that for most of these environmental aspects, the information in the FADN first
has to be transformed in order to create information about the environmental sustainability (this
topic is dealt with later in this paper). This transformation also takes place in an environmental
Life Cycle Assessment (LCA). In an LCA study of a product or service, flows within the eco-
nomic system and flows out of the system into the environment are determined (i.e. all extrac-
tions of resources from the environment and emissions to the environment), when possible in a
quantitative way. Based on this data, the potential impacts on natural resources, the environment
and human health are assessed. The FADN serves as an input to carry out LCAs (Vrolijk et al.,
2002).
Chain information
The availability of data is one of the impediments in the determination of sustainability in agro-
food chains. In developing the annual social & environmental report, many data sources had to
be consulted in order to obtain the appropriate information with which to determine the current
level of sustainability. The collection of the necessary data is also the main focus of attention
when carrying out an LCA. This factor led LEI to look at ways in which data can be supplied and
187
collated in a different relationship. Together with ATO, LEI investigated the extent to which
LCA and chain information systems (CISs) can be used to make a report assessing sustainability
in the chain.
A CIS is defined as the way the chain actors share information. Two different systems can be
distinguished: a system in which information is transferred by the various actors in the chain, and
one which is placed outside the chain and acts like a type of database. In the latter system, all the
actors in the chain have access to the system. Today, most CISs are developed to trace and/or
track products. Tracking provides information about the current position in a chain, while tracing
provides information about where the products have been and under what circumstances. Food
safety and legislation are important motivations behind the growing attention to tracing and
tracking and to CISs. Currently, the most important reason for a company to pay attention to
tracing and tracking is to comply with laws and to limit damage claims. It is expected that com-
panies will start to use tracing and tracking in a more positive way, viz. to better control the pro-
duction processes (i.e. to shift from tracing to tracking).
In the research mentioned above, the possibilities for combining sustainability and CISs was in-
vestigated. This was carried out with the help of previous studies. Existing CISs were investi-
gated in more detail: it was examined what kind of information is gathered in these systems and
whether there are some links with sustainability.
Case study: the Groeinet information system (vegetables)
Groeinet develops registration programmes for environmental registrations, product registrations,
logistics, etc. Agricultural companies can register their use of pesticides and fertilisers as well as
the yield per crop. While the Groeinet information system was not developed to provide an in-
sight into sustainability, the system does provide information useful in the determination of as-
pects of sustainability. For example, information about pesticides and fertilisers can be used to
determine contributions to several environmental impacts. The production of these inputs is ac-
companied by, for example, the use of energy, which can be very high; for example, the fertiliser
calcium ammonium nitrate (CAN) is produced from natural gas in the Netherlands, and it takes
0.7 m3 of natural gas to produce 1 kg of CAN. However, databases with information about the
energy contents of inputs are necessary to determine the total energy use. The use of pesticides
leads to emissions of toxic compounds to water, soil and air. With information about the eco-
toxic effects of pesticides use, the recorded information regarding pesticides can be used to de-
termine sustainability as regards the impact on the quality of water, soil and air.
The case studies showed that the Groeinet information system contains information relevant to
sustainability. However, additional information is required to translate the used inputs into useful
information about sustainability. For some aspects of sustainability this can be done easily. The
188
case studies showed that this is the case for environmental sustainability aspects. However, the
determination of aspects of social sustainability is more difficult. This is because CISs focus on
products. Environmental aspects could also be related to products, as in an LCA. Aspects of so-
cial and economic sustainability are more related to processes than to products. However, there
are possibilities to relate social aspects to a product, for example, the number of labour hours or
the labour happiness level required to produce a certain number of products. Additional research
is required to work this out in further detail (Kramer et al., 2002).
Summary
This article focused on the development of issues and themes to determine the progress in mak-
ing agricultural chains sustainable. For some sustainability aspects, no indicators are available.
However, for the Dutch horticultural sector, some environmental and social sustainability indica-
tors have been developed. To determine environmental sustainability in chains, elements of LCA
can be used. The development of an annual social & environmental report on the Dutch horticul-
tural sector showed that the availability of information is still a major impediment to the deter-
mination of sustainability. This problem occurs mainly at the level of the production chain. On
the farm level, several systems (mainly the FADN) contain the necessary information. A possible
solution to this would be to combine chain information systems with sustainability.
References
Netherlands Ministry of Social Affairs and Employment (2002). Nulmeting en onderzoek stand
der techniek fysieke en psychische belasting arboconvenant agrarische sectoren, The
Hague, the Netherlands.
Meeusen, M.J.G. and E. ten Pierick (2002), Meten van duurzaamheid
Naar een instrument voor agroketens, Landbouw-Economisch Instituut (LEI), The Hague, the
Netherlands
Vrolijk, H.C.J., G. Cotteleer, K.J. Kramer, T.C. van Leeuwen and H.H. Luesink (2003). Per-
formance indicators. Landbouw-Economisch Instituut (LEI), The Hague, the Netherlands
Kramer, K.J, Thors, M. and J. Wolfert. (2003). Duurzaamheid in agrofood ketens, Landbouw-
Economisch Instituut (LEI), The Hague, the Netherlands
Knijff, van der A. and J. Benninga (2003). Energie in de glastuinbouw van Nederland; ontwikke-
lingen in de sector en op de bedrijven t/m 2002. Landbouw-Economisch Instituut (LEI),
The Hague, the Netherlands.
Wood, D.J. (1991). ‘Corporate Social Performance Revisited’. Academy of Management Re-
view, vol. 16, no. 4, 691-718.
189
Life Cycle-Based Sustainability Indicators for Assessment of the U.S. Food
System
Keoleian, G.A.* and. Heller, M.C.
*Center for Sustainable Systems, School of Natural Resources and Environment, 430 East University,
University of Michigan, Ann Arbor, MI 48109-1115 U.S.A.;
phone: 1 734 764 3194; fax: 1 734 647 5841; [email protected]
Abstract
The overall sustainability of any food system is influenced not only by agricultural practices but also
by food consumption behaviors, food processing and distribution activities. A product life cycle ap-
proach provides a useful framework for studying the links between production and consumption ac-
tivities. The United States has a highly productive agricultural system as measured by the total
amount of food output. Over the last century dramatic changes to this system have been imple-
mented to enhance this productivity metric but the environmental, social, and economic conse-
quences have also been significant. This paper highlights a select set of indicators covering the life
cycle management of the entire food system. Indicators address economic, social, and environ-
mental aspects of each life cycle stage: origin of (genetic) resource; agricultural growing and produc-
tion; food processing, packaging and distribution; preparation and consumption; and end of life.
Introduction
A recent assessment of the sustainability of the U.S. food system was conducted by the Center
for Sustainable Systems. A life cycle framework was used to identify and organize a set of so-
cial, economic and environmental indicators for evaluating the current U.S. food system. Even
though a comprehensive life cycle assessment (LCA) of the food system is not currently possi-
ble, using a life cycle framework does provide a systematic basis for developing indicators. A
matrix defined by life cycle stages and the three dimensions of sustainability served to group the
indicators. A total of 57 indicators were proposed and evaluated. The full set of indicators are
presented in two publications by the authors (Heller and Keoleian, 2003; Heller and Keoleian,
2000). The purpose of this paper is to highlight key findings of our investigation. In particular,
current trends for a select set of ten indicators that threaten the long-term economic, social, and
environmental sustainability of the US food system are discussed.
Method
Table 1 presents the full matrix of sustainability indicators developed in Heller and Keoleian
2003. The rows represent major stages of the food system; indicators for each stage are catego-
rized into the “triad” of sustainability: economic, social, and environmental. In many incidences,
this division is somewhat arbitrary since particular indicators often address more than one aspect
of sustainability. Also identified in Table 1 are the primary stakeholders involved or influential
in each stage of the food system.
190
Table 1. Life Cycle Sustainability Indicators for the Food System (Heller and Keoleian 2003).
Life cycle stage Stakeholders INDICATORS
Economic Social Environmental
Origin of (genetic) Farmers -degree of -diversity in seed purchasing and seed col- -ratio of naturally pollinated plants to geneti-
resource – seed Breeders farmer/operator control of lecting options cally modified/ hybrid plants per acre
production, animal Seed Companies seed production/breeding -degree of cross-species manipulation -reproductive ability of plant or animal
breeding -% of disease resistant organisms
Agricultural grow- Farm operators -rates of agricultural land -average age of farmers -rate of soil loss vs. regeneration
ing and production Farm workers conversion -diversity and structure of industry, size of -soil microbial activity, balance of nutrients/acre
Ag. Industry -output/input productivity farms, # farms per capita -quantity of chemical inputs/ unit of production
Ag. Schools -% return on investment -hours of labor/ yield and / income -air pollutants/ unit of production
Government -cost of entry to business -avg. farm wages vs. other professions -number of species/acre
Animals -farmer savings and in- -# of legal laborers on farms, ratio of migrant -water withdrawal vs. recharge rates
surance plans workers to local laborers, -# of comtaminated or eutrophic bodies of sur-
-flexibility in bank loan -% workers with health benefits. face water or groundwater
requirements to foster -# of active agrarian community organizations -% waste utilized as a resource
environmentally sustain- -% of ag. Schools that offer sustainable ag. -veterinary costs
able practices programs, encourage sustainable practices -energy input/ unit of production
-level of gov’t support -# animals/unit, time animals spend outdoors -ratio of renewable to non-renewable energy
(animal welfare) -portion of harvest lost due to pests, diseases
Food processing, Food processors -relative profits received -quality of life and worker satisfaction in food -energy requirement for processing, packaging
packaging and dis- Packaging pro- by farmer vs. processor processing industry and transportation
tribution viders vs. retailer -nutritional value of food product -waste produced/ unit of food
Wholesalers -geographic proximity of -food safety -% of waste and byproducts utilized in food
Retailers grower, processor, pack- processing industry
ager, retailer -% of food lost due to spoilage/mishandling
Preparation and Consumers -portion of consumer dis- -rates of malnutrition -energy use in preparation, storage, refrigera-
consumption Food service posable income spent on -rates of obesity tion
Nutritionists/ food -health costs from diet related dis- -packaging waste/ calories consumed
Health pro- -% of food dollar spent ease/conditions -ratio of local vs. non-local and seasonal vs.
fessionals outside the home -balance of average diet non-seasonal consumption
-% of products with consumer labels
-degree of consumer
literacy regarding food system consequences,
product quality vs. appearance, etc.
-time for food preparation
End of life Consumers -ratio of food wasted to -ratio of (edible) food wasted vs. donated to -amount of food waste composted vs. sent to
Waste managers food consumed in the US food gatherers landfill/incinerator/ waste water treatment
Food recovery & -$ spent on food disposal
gleaning orgs
191
Table 2 summarizes the scope and boundary conditions for the assessment. The current
agricultural system is highly productive, yielding an estimated 161 million metric tons of
edible food in 1995 for a U.S. population of 263 million. As a result of affordable food and
food assistance, more than 90% of U.S. households were food secure, meaning they had
assured access at all times to enough food for an active healthy life. Still, 9.7% of U.S.
households – about 10 million – were food insecure over the 1996-1998 period. The food
energy available for consumption, based mainly on national disappearance of food was 15900
kJ (3800 kilocalories) per capita per day in 1994 (NASS 1999). However, this value includes
food that is wasted at the retail and consumer level. According to USDA’s Continuing Survey
of Food Intakes by Individuals (CSFII), the surveyed caloric intake was 8382 kJ based on
1994-1996 surveys (Tippett and Cleveland 1999).
Table 2. Scope and boundary condition for this sustainability assessment.
• System boundary encompass: origin of (genetic) resource; agricultural growing and produc-
tion; food processing, packaging and distribution; preparation and consumption; and end of
•
life management
•
Disaggregation of food from other agricultural production is not completely possible
U.S. exports and imports of agricultural products are both significant but indicators are not
•
corrected for agricultural trade ratios
Analysis is conducted at a national scale recognizing, however, that there is significant re-
gional variation in indicators
Key Findings
A select set of indicators from Table 1 is presented here. These were chosen by the authors
based on their current importance in influencing the long-term sustainability of the US food
system.
1. Rapid conversion of prime farmland
The land available for agriculture is an primary indicator of its stability. According to the Na-
tional Resources Conservation Service, cropland and pastureland (including CRP land) to-
taled 213.4 million ha in 1997, a 5% decrease from 1982 (NRCS 1999). 5.7 million ha of
cropland and pastureland were converted to urban or built-up land over this 15 year period,
amounting to over 50% of the decrease. As prime farmland is being developed, less stable
non-prime farmland in arid regions is being added to the base, leading to increased erosion
rates and irrigation demands (Harlin 1995).
2. Depletion of topsoil exceeds regeneration
The National Resources Inventory of the USDA (1999b) reports that 1700 megatonnes (Mt, =
million metric tonnes) of soil eroded from U.S. land in 1997, 760 Mt from wind and 960 Mt
from sheet and rill (caused by water) (NRCS 1999). Improvements have been made in recent
years: the number reported for 1982 is 2780 Mt (NRCS 1999). 45.3 million ha (30% of crop-
land) were determined to be excessively eroding in 1997 (erosion rates greater than that
192
deemed tolerable), totaling to 1200 Mt of eroded soil per year (NRCS 1999). If the 1700 Mt
of topsoil lost in 1997 were evenly distributed over all of the U.S. cropland, the average rate
of erosion would be 9.9 tonnes per ha per year. This translates into 2.5 cm of topsoil lost from
all U.S. cropland every 34 years. It should be quickly recognized that this practice is not sus-
tainable.
3. Rate of groundwater withdrawal exceeds recharge in major agricultural regions
Agriculture greatly affects the quantity of water consumed in the U.S., primarily through
irrigation of crops and through livestock production. In 1995, 507 million m3 per day
(134,000 million gallons per day) of freshwater were withdrawn for irrigation purposes (39%
of total freshwater withdrawal), 185 million m3 per day of this from groundwater sources.
Water consumption for livestock totaled 20.8 million m3 per day in 1995, 41% of which was
from groundwater (Solley et al. 1998). The concern is that, in certain regions of the country,
withdrawal from groundwater sources is exceeding the natural recharge rate of aquifers. An
excellent case-in-point is the Ogallala aquifer in the High Plains states. The Ogallala aquifer
is largely a nonrenewable resource that has been mined at rates that greatly exceed recharge.
Pumping from the aquifer is measured in feet per year while replacement, trickling in from
the surface only, occurs at less than an inch a year (Opie 1993). Irrigation rates have de-
creased slowly over recent years, but current withdrawal rates still cannot be sustained long
into the future.
4. Losses to pests increasing
According to the Census of Agriculture, $7600 million were spent on agricultural chemicals
in 1997, up from $4700 million in 1987 (USDA 1999a). Average application rates appear to
have decreased, however, from a two decade high of 2.0 kg ha-1 in 1987 to 1.3 kg ha-1 in 1996
(Kellogg et al. 1999). Interestingly, while pesticide use is generally seen as profitable in terms
of direct crop returns, it has not necessarily led to decreases in crop loss. Even with a tenfold
increase in insecticide use from 1945 to 1989, total crop losses from insect damage have
nearly doubled from 7% to 13% (Pimentel et al. 1991). This rise in crop losses is partly
caused by changes in agricultural practices such as abandoning crop rotations and increased
crop homogeneity.
5. Reduction in genetic diversity
Modern agricultural systems have also become increasingly genetically uniform. Only 10-20
crops provide 80-90% of the world’s calories (Brown 1981). In the U.S., 42% of the soy-
beans, 43% of the corn, and 38% of the wheat grown in 1980 were dominated by the top 6 va-
rieties (Duvick 1984). Often, these varieties originate from an even smaller genetic base. For
example, the hundreds of corn hybrids grown in the U.S. are largely based on about 12 inbred
lines that originated from a few open-pollinated varieties of a single race of the some 200
known races of corn. (Soule et al. 1990). Repeated warnings have been sounded from the re-
search community about the extreme vulnerability associated with this limited genetic diver-
sity (CAST 1999; Wood and Lenne 1999). Lack of biodiversity in crops leads to pest and dis-
193
ease susceptibility (National Research Council 1972; Brown 1983; Altieri 1994). Insufficient
crop genetic diversity led to the corn leaf blight in southern U.S. in 1970 (CAST 1999) and is
partly responsible for outbreaks of Fusarium head blight in wheat and barley in Minnesota
and the Dakotas (McMullan et al. 1997). The forces driving genetic uniformity also lead to
abandonment and loss of locally adapted varieties that are the necessary resources to meeting
future plant breeding challenges.
6. Poor economic conditions for farmers
Industrialization of agriculture has made production extremely capital intensive, leading to
high cost of entry. An estimated $500,000 in assets is needed to support a farm household
(ERS 2000). This, combined with return on investment considerably lower than can be re-
ceived in other business ventures, is contributing to declining numbers of young farmers.
Rates of return on farm business equity in 1998 were reported at 2.15%, and have averaged –
0.3% since 1980 (3.8% since 1990) (ERS 2000). Nearly half (48%) of all farms reported a
negative net cash return (net loss) in 1997. Ninety-two percent of the farms reporting losses in
1997 were relatively small, with sales worth less than $50,000 (USDA 1999a). Still, farm
households in the U.S. receive incomes on par with average U.S. households. The 1997 aver-
age household income for farm operator households was $52,300. However, 89% of this
household income comes from off-farm sources (USDA 1999b).
Providing inexpensive food for Americans has long been a central tenet in U.S. agricultural
policy. Yet, as food processing, handling, and marketing have increased, the farmer has re-
ceived smaller and smaller portions of the American food bill. USDA estimates that the
farmer’s gross return on a consumer’s dollar spent on food in 1998 was 20 cents (Elitzak
1999a) (in 1975 it was 40 cents (Elitzak 1999b)). The remaining 80% of the food bill is dis-
tributed among marketing labor, packaging, advertising and other categories.
7. Declining entry of young farmers into the profession
Another prominent trend on America’s farms is the advancing age of farm operators. Accord-
ing to the Census of Agriculture, the average age of farm operators in 1997 was 54.3 years,
and 61% of the operators were 55 and over. In 1954, only 37% of farmers were 55 and over.
Comparatively, 11.7% of the civilian labor force was age 55 or above in 1997 (18% in 1954)
(ERS 2000). The percentage of farmers under 35 years of age dropped from 15% in 1954 to
7.8% in 1997 (ERS 2000). Furthermore, farm labor has dropped significantly over the past 50
years, from 9.9 million workers in 1950 to 2.8 million in 1998 (USDA 1999b).
8. Obesity rates are rising along with the costs of diet related diseases
Data from the National Health and Nutrition Examination Surveys of 1977-80 and 1988-1994
demonstrate that the prevalence of obesity is on the rise throughout the American population.
The number of overweight individuals rose over the time between surveys from 25.4% to
34.9% among American adults, from 7.6% to 13.7% among children ages 6-11 years, and
from 5.7% to 11.5% among adolescents (Nestle and Jacobson 2000). Under an updated defi-
194
nition presented in the 2000 Dietary Guidelines (USDA 2000), 60% of males and 46% of fe-
males 20 years and over were overweight in 1994-1996 (Tippett and Cleveland 1999). In-
creasingly, scientific studies confirm that America’s diet of high fat intakes and low intakes of
whole, fiber-containing foods such as whole grains, vegetables, and fruits has a significant
impact on our health, quality of life, and longevity. Diet is a significant factor in the risk of
coronary heart disease, certain types of cancer, and stroke – the three leading causes of death
in the United States (National Research Council 1989). Estimates of diet-related medical
costs, loss of productivity, and value of premature deaths reach $71,000 million per year
(Frazão 1999). Estimates of the direct health care costs of obesity alone range from $39,000
million (Allison 1999) to $52,000 million (Wolf and Colditz 1998) annually. The prevalence
of overweight and obese Americans was highlighted as a major agenda issue at the National
Nutrition Summit in May of 2000 (Scannell et al., 2000).
9. A large fraction of edible food is wasted
41.2 Mt of food is wasted at home and at food service establishments, amounting to 26% of
the edible food available for human consumption in the U.S. Fresh fruits and vegetables ac-
counted for 19% of these losses, and an additional 18% was fluid milk (Kantor et al., 1997).
Examinations of household garbage by researchers at the University of Arizona concluded
that large quantities of single food items – entire heads of lettuce, half-eaten boxes of crackers
– accounted for a larger share of household food loss than did plate scraps. They also found
that specialty products such as sour cream, hot dog buns, or impulse items had a higher fre-
quency in household garbage than did frequently purchased staples like bread, milk and cereal
(Kantor et al., 1997).
10. Heavy reliance on fossil fuel
In total, providing the 15900 kJ of food energy available per capita per day in the United
States is estimated to consume 10.8 million TJ annually. This represents about 10% of the to-
tal energy consumed in the United States (EIA 2001). By our estimates, therefore, it takes
about 7.3 units of (primarily) fossil energy to produce one unit of food energy in the U.S. food
system. This value is somewhat lower than estimates reported by others. Pimentel and Pimen-
tel 1996 and Hall et al. 1986 both estimate 10 units of input energy per unit of output food en-
ergy.
Conclusions
Land, sufficient topsoil, water, and human capital are all essential inputs for a sustainable
food system. A sustainable food system must also be founded on a sustainable diet. In the
most general sense, this would be a diet that matched energy intake with energy expenditure
while supplying necessary nutrients for a healthy lifestyle. The greatest leverage point for en-
hancing the sustainability of the U.S. food system lies with the level of consumption and
amount of food waste. Significant improvements in diet not only have direct health benefits
and reduced costs of diet related diseases but also more than proportionally reduce environ-
mental impacts from agricultural production. The opportunity to reduce food production is
195
tremendous by limiting excess consumption estimated at 8382 kJ per capita per day and edi-
ble food waste estimated at 26%. A reduction by one third is not unrealistic.
The economics of the U.S. food system also needs some fundamental adjustments to reverse
unsustainable social and environmental impacts. Entry of young farmers into the profession is
declining and production is shifting to larger scale farms, which are less ecologically sustain-
able. A systems-based solution would combine a reduction in food consumption and waste
while maintaining revenues to farmers for less food output. The disposable income spent on
food could be held constant in this scenario and costs of diet related diseases would be dra-
matically reduced. Until society places a higher value on food, the reported unsustainable pat-
terns will continue. It is clear that governmental policies that address both production and
consumption are necessary to advance the sustainability of our food system.
References
Allison, D. B., 1999. “The Direct Health Care Costs of Obesity in the United States.” Ameri-
can Journal of Public Health 89(8): 1194-1199.
Altieri, M. A., 1994. Biodiversity and Pest Management in Agroecosystems. New York, Ha-
worth Press.
Brown, N. J., 1981. Biological Diversity: The Global Challenge. The U.S. Strategy Confer-
ence on Biological Diversity, U.S. Department of State Publication.
Brown, W. L., 1983. Economic Botany 37: 1-12.
CAST, 1999. Benefits of Biodiversity, Council for Agricultural Science and Technology.
Duvick, D. N., 1984. “Genetic Deversity in Major Farm Crops on the Farm and in Reserve.”
Economic Botany 38(2): 161-178.
EIA, 2001. Annual Energy Review 2000. Washington, DC, Energy Information Administra-
tion, U.S. Dept. of Energy.
Elitzak, H., 1999a. “Desire for Convenience Drives Marketing Costs.” Food Review 22(3):
23-25.
Elitzak, H., 1999b. Food Cost Review, 1950-97. Washington, DC, Economic Research Ser-
vice, USDA.
ERS, 2000. The Aging of Farm Operators and Participation of Beginning Producers in Farm-
ing, Economic Research Service, USDA.
ERS, 2000. Farm Business Balance Sheet and Financial Ratios, Economic Research Service,
USDA.
Frazão, E., 1999. High Costs of Poor Eating Patterns in the United States. America's Eating
Habits: Changes and Consequences. E. Frazão. Washington, DC, Economic Research
Service, USDA. Agriculture Information Bulletin No. 750.
Furst, T., M. Connors, Bisogni, C.A., Sobal, J., Falk, L.W., 1996. “Food Choice: A Concep-
tual Model of the Process.” Appetite 26(3): 247-265.
Hall, C. A. S., C. J. Cleveland, Kaufmann, R., 1986. Energy and Resource Quality. New
York, Wiley Interscience.
196
Harlin, J. M., 1995. Rangeland to Cropland Conversions in the Western States: Implications
for Soil Loss and Sustainability. Forum of the Association for Arid Land Studies XI,
Lubbock, TX, International Center for Arid and Semiarid Land Studies.
Heller, M. C., and Keoleian, G. A., 2000. Life Cycle-based Sustainability Indicators for
Assessment of the U.S. Food System. Center for Sustainable Systems, Ann Arbor, MI.
CSS00-04, available: http://css.snre.umich.edu/css_doc/CSS_2000_4.pdf.
Heller, M.C., G.A. Keoleian (2003) Assessing the sustainability of the U.S. food system: a
life cycle perspective. Agricultural Systems 76: 1007-1041.
Kantor, L. S., K. Lipton, Manchester, A., Oliveira, V., 1997. “Estimating and Addressing
America's Food Losses.” Food Review 20(1).
Kellogg, R. L., Nehring, R., Grube, A., Plotkin, S., Goss, D. W., and Wallace, S., 1999.
Trends in the Potential for Environmental Risk from Pesticide Loss from Farm Fields.
Proceedings of The State of North America's Private Land Conference (Chicago, IL,
January 19-21, 1999), National Resources Conservation Service, USDA.
McMullan, M., R. Jones, Gallenberg, D., 1997. “Scab of wheat and barley: A re-emerging
disease of devastating impact.” Plant Disease 81: 1340-1348.
NASS, 1999. Chapter XIII: Consumption and Family Living. Agricultural Statistics 1999.
Washington, DC, National Agricultural Statistics Service, USDA.
National Research Council, 1972. Genetic Vulnerability of Major Crops. Washington, DC,
National Academy Press.
National Research Council, 1989. Diet and Health: Implications for Reducing Chronic Dis-
ease Risk. Washington, DC, National Academy Press.
Nestle, M. and M. F. Jacobson, 2000. “Halting the Obesity Epidemic: A Public Health Policy
Approach.” Public Health Reports 115(Jan/Feb.): 12-24.
NRCS, 1999. 1997 National Resources Inventory Summary Report. Washington, DC, Natural
Resources Conservation Service, USDA.
Opie, J., 1993. Ogallala: Water for a dry land. Lincoln, University of Nebraska Press.
Pimentel, D., L. McLaughin, Zepp, A., Lakitan, B., Kraus, T., Kleinman, P., Vancini, F.,
Roach, W. J., Graap, E., Keeton, W. S., and Selig, G., 1991. Environmental and eco-
nomic impacts of reducing U.S. agricultural pesticide use. Handbook on Pest Manage-
ment in Agriculture. D. Pimentel. Boca Raton, FL, CRC Press: 679-718.
Pimentel, D. and M. Pimentel, 1996. Energy Use in Fruit, Vegetable, and Forage Production.
Food, Energy, and Society. D. Pimentel and M. Pimentel. Niwot, CO, University Press
of Colorado: 131-147.
Scannell, K. M., M. E. Conway, Gordner, R. L., Klein, L. J., and Zorn, M, 2000. National Nu-
trition Summit: information resources, National Library of Medicine (U.S.).
Solley, W. B., R. R. Pierce, Perlman, H. A, 1998. Estimated Use of Water in the United States
in 1995. Denver, CO, U.S. Department of the Interior.
Soule, J., D. Carre, Jackson, W., 1990. Ecological Impact of Modern Agriculture. Agroecol-
ogy. C. R. Carroll, J. H. Vandermeer and P. M. Rosset. New York, McGraw-Hill Pub-
lishing Co.: 175.
197
Tippett, K. S. and L. E. Cleveland, 1999. How Current Diets Stack Up: Comparison with Die-
tary Guidelines. America's Eating Habits: Changes and Consequences. E. Frazão.
Washington, DC, Economic Research Service, Food and Rural Economics Division,
USDA. Agriculture Information Bulletin No. 750.
USDA, 1999a. 1997 Census of Agriculture, U.S. Dept. of Agriculture.
USDA, 1999b. Agriculture Fact Book 1999. Washington, DC, U.S. Dept. of Agriculture, Of-
fice of Communications.
USDA, 2000. Dietary Guidelines for Americans. Washington, DC, U.S. Dept. of Agriculture.
Wolf, A. M. and G. A. Colditz, 1998. “Current estimates of the economic cost of obesity in
the United States.” Obesity Research 6: 97-106.
Wood, D. and J. M. Lenne, Eds., 1999. Agrobiodiversity: Characterization, Utilization, and
Management. New York, CABI Publishing.
198
Most significant substances of LCA to Mediterranean Greenhouse Horti-
culture
Antón, M.A.*(a); Castells, F.(b); Montero, J.I. (a); Muñoz, P.(a)
a
Institut de Recerca i Tecnologia Agroalimentària. Centre de Cabrils. Ctra. de Cabrils s.n. 08348,
Cabrils (Barcelona). Spain. Tel: 34937507511 Fax: 34937533954 [email protected]
b
Department of Chemical Enginering. ESTEQ. Universitat Rovira Virgili 43007 Tarragona. Spain
Abstract
LCA has been used as a tool to identify the environmental burdens associated to protected
cultivation in the Mediterranean area. A tomato crop was cultivated in a steel-framed green-
house; plants were transplanted directly in the soil. With the exception of toxicity indicators,
the main sources of environmental impact were production and use of the fertilizer needed,
and the manufacture of the greenhouse structure. From the 470 substances considered in
this study, only 21 contributed more than 5% of the total impact of one category. Among the
21 substances, 16 contributed to air or water emissions, being 6 of them pesticides, and 5
were non-renewable resources. Results from this study also suggest that there is a potential
for simplifying LCA methodology by considering a restricted number of substances.
Keywords: simplified LCA, protected horticulture
Introduction
Previous works on the application of LCA in protected horticulture were focused in North
European Regions. They showed that reducing the heating requirements must be a priority in
order to limit the environmental load. (Jolliet, 1993, Bucher y col., 1996, Nienhuis y col.,
1996, Jungbluth y col., 2000, Van Woerden, 2001). Regarding protected horticulture in south-
ern areas, where there is no need of heating or these needs are much lower, there was a strong
lack of information on which greenhouse production and handling processes have bigger ef-
fects on the environmental impact of greenhouse production. Antón (2004) studied the main
environmental burdens associated to protected horticulture in Mediterranean countries com-
paring different production means. In this work we have evaluated the traditional soil crop-
ping system and try to identify the main contributor substances in order to promote simplified
LCA at more local level.
Material and methods
A tomato crop was cultivated in a traditional greenhouse with steel frame and PE film as the
cover. Plants were transplanted directly in the soil. This study takes into consideration only
the means of production. Therefore the analysis finishes at the moment that tomatoes are har-
vested, “farm’s gate”. The kg of tomato produced was chosen as the functional unit. Agro-
nomic data were collected from different tomato growers of Maresme, an area located 30 km
North from Barcelona.
199
In order to know the relative importance of the different processes, greenhouse structure, pes-
ticides and fertiliser production, greenhouse, pest, fertirrigation and waste management were
assessed as subsystems.
1. Tomato production
Due to its complexity and in order to facilitate study, the tomato production was di-
vided into five sub-systems:
1.a) Greenhouse management during tomato production, GM
1.b) Fertilizer production, F
1.c) Fertilization and irrigation, FR
1.d) Pesticide production, P
1.e) Pest management, PM
2. Manufacture
Two further subsystems were added to take into consideration the different process
and materials used in the greenhouse structure and for the auxiliary equipment, the fer-
tilization and irrigation system:
2.a) Greenhouse structure, G
2.b) Auxiliary equipment, R
3. Waste
The final system analysed included the management of waste generated during and at
the end of greenhouse crop cultivation, W
Results
Table 1 shows the overall impact of the global process of tomato greenhouse cropping for the
different impact categories and their contribution to the different subsystems considered. With
relation to climate change, it is important to point out the negative impact due to CO2 fixation
by the crop.
Fertiliser production was the main stage in the cycle that contributed to climate change (83%)
and depletion of non-renewable resources (65%). During crop production, pesticides were
mainly responsible for the toxicity indicator scores, while the fertirrigation was the main con-
tributor to eutrophication (60%) mainly due to nitrogen compounds emissions. The green-
house structure manufacturation (steel frame and cladding) mainly contributed to the catego-
ries of photochemical oxidant formation (45%) caused by hydrocarbons emissions to the air,
climate change (34%), mostly CO2 emissions, and air acidification (28%), principally due to
release of NOx and SOx.
From the 470 substances considered in this study, only 21 contributed more than 5% of the to-
tal impact of one category. Among the 21 substances, 16 contributed to air or water emis-
sions, being 6 of them pesticides, and 5 were non renewable resources (table 2). From those
10 emitted substances non pesticides, 8 were cited also as main contributor substances in
Cowell (1998). These substances and their contribution of each category are shown in figure
200
1. Toxicity potentials are mainly due to the use of pesticides. Nevertheless it has to be men-
tioned that pesticide types and application may change from one exploitation to another as a
function of the pests affecting the crop every season.
Conclusions
Further research must be orientated towards reducing the environmental impact of the materi-
als used in the structure and looking forward a more rational management criteria in the sup-
ply of nutrients to the crop in order to reduce fertiliser use and avoid the leachates
Since only a reduced number of substances are the main contributors for the environmental
impact, it may be possible to simplify the assessment by focusing on a restricted number of
substances.
Acknowledgements
The authors thank Montse Martí and Jordi Ariño for providing useful information related to
tomato crop farmers. This research was partially supported by INIA nº SC00-080-C2 and nº
RTA03-096-C5-2.
Literature cited
Antón, A. (2004). Utilización del Análisis del Ciclo de Vida en la Evaluación del Impacto
Ambiental del cultivo bajo Invernadero Mediterráneo. Tesis doctoral. Programa d'En-
ginyeria Ambiental. Universitat Politècnica de Catalunya. Barcelona.
Bucher, J., Chelkova, N., van der Ende, J.P., van der Hulst, P., Fernández Soler, M., Szarvas,
G., Vrbensky, R. (1996). Tomato North-South: Recommendations for the improvement
of environmental quality of tomato production using different environmental labels.
Centre of Environmental Science, Leiden University. Leiden. xxxi+96+31
Cowell, S.J. (1998). Environmental Life Cycle Assessment of Agricultural Systems: Integra-
tion Into Decision-Making. Ph. D. thesis. Centre for Environmental Strategy. University
of Surrey. Guildford.
Jolliet, O. (1993). Bilan écologique de la production de tomates en serre. Revue S. Vitic. Ar-
boric. Hortic. 25 (4): 261-267.
Jungbluth, N., Tietje, O., Scholz, R.W. (2000). Food Purchases: Impacts from the Consum-
ers' Point of View Investigated with a Modular LCA. International Journal of Life Cycle
Assessment. 5: 134-142.
Nienhuis, J.K., Vreede, P.J.A.d., De Vreede, P.J.A., Brumfield, R.G. (1996). Utility of the
environmental life cycle assessment method in horticulture. En: Proceedings of the
XIIIth International Symposium on Horticultural Economics, Rutgers. (Ed). New
Brunswick, New Jersey, USA. No. 429: 531-538.
Van Woerden, S. (2001). The application of Life Cycle Analysis in glasshouse horticulture.
En: International Conference LCA in Foods. (Ed). Gothenburg. 136-140.
201
Antón, A. (2004). Utilización del Análisis del Ciclo de Vida en la Evaluación del Impacto
Ambiental del cultivo bajo Invernadero Mediterráneo. Tesis doctoral. Programa d'En-
ginyeria Ambiental. Universitat Politècnica de Catalunya. Barcelona.
Bucher, J., Chelkova, N., van der Ende, J.P., van der Hulst, P., Fernández Soler, M., Szarvas,
G., Vrbensky, R. (1996). Tomato North-South: Recommendations for the improvement
of environmental quality of tomato production using different environmental labels.
Centre of Environmental Science, Leiden University. Leiden. xxxi+96+31
Cowell, S.J. (1998). Environmental Life Cycle Assessment of Agricultural Systems: Integra-
tion Into Decision-Making. Ph. D. thesis. Centre for Environmental Strategy. University
of Surrey. Guildford.
Jolliet, O. (1993). Bilan écologique de la production de tomates en serre. Revue S. Vitic. Ar-
boric. Hortic. 25 (4): 261-267.
Jungbluth, N., Tietje, O., Scholz, R.W. (2000). Food Purchases: Impacts from the Consum-
ers' Point of View Investigated with a Modular LCA. International Journal of Life Cycle
Assessment. 5: 134-142.
Nienhuis, J.K., Vreede, P.J.A.d., De Vreede, P.J.A., Brumfield, R.G. (1996). Utility of the
environmental life cycle assessment method in horticulture. En: Proceedings of the
XIIIth International Symposium on Horticultural Economics, Rutgers. (Ed). New
Brunswick, New Jersey, USA. No. 429: 531-538.
Van Woerden, S. (2001). The application of Life Cycle Analysis in glasshouse horticulture.
En: International Conference LCA in Foods. (Ed). Gothenburg. 136-140.
202
% % %
100 100 100
80 (a) NH3 80 (r) BaSO4 80
(a) SOx
(a) NOx (r) Nat. Gas
60 60 60
(a) N2O (r) Oil (a) NOx
(w) NO3- 40 (r) KCl
40 40 (a) NH3
(w) N (r) Phosphate
20 20 20
0 0 0
Eutrophication (CML, g eq. PO4) Dep. Non ren. resources (yr-1) Acidification (ETH, g eq. H+)
% % %
100 100 100
80 80 80
60 (a) N2O 60 60
(a) CH4 (a) CxHy
40 (a) CO2 40 40 (a) Halon
20 20 20
0 0 0
Climate change (g eq. CO2) Phot. oxidant (WMO, g eq. ethylene) Dep. Ozone layer (WMO, g eq. CFC-11)
% %
100 %
100 100
(f) Cyromazine
80 80 80
(w) Cyromazine (f) Iprodione (s) Azoxystrobin
60 60 60
(f) Captan
(w) Benomyl
40 40 (f) 40
(w) Azoxystrobin Bromopropylate (s) Cyromazine
20 20 (f) Azoxystrobin 20
0 0 0
Aquatic Tox. (CST, eq. Zn w) Human Tox. (CST, eq. Pb a) Terrestrial Tox. (CST, eq. Zn a)
Figure 1. Main contributors’ substances to the total impact categories
203
Table 1. Values for each one of the subsystems of the tomato crop production for the different
environmental categories.
FAB PRODUCTION WASTE
TOTAL G R GM F FR P PM W
Eutrophica- EI g eq. PO4 2.7E-1 2.4E-2 7.3E-3 7.3E-3 4.2E-2 1.6E-1 9.0E-5 2.3E-4 2.6E-2
tion
Dep. non BR year-1 8.1E-3 1.6E-3 6.1E-4 8.2E-4 5.3E-3 4.2E-5 1.1E-5 1.1E-5 -1.9E-4
renov. re-
sources
Acidification AI g eq. H+ 3.0E-2 8.6E-3 3.3E-3 4.3E-3 7.1E-3 6.5E-3 7.3E-5 1.9E-4 3.5E-5
Climate CCI g eq. CO2 1.2E+2 4.0E+1 1.2E+1 - 9.7E+1 3.1E+1 4.2E-1 1.1E+0 -7.5E+0
change 5.7E+1
Dep. ozone ODI g eq. 2.3E-5 2.3E-6 8.0E-7 1.6E-5 2.5E-6 4.7E-7 1.5E-7 1.2E-7 6.4E-7
layer CFC11
Phot. oxi- POI g eq. 1.7E-1 7.7E-2 3.2E-2 4.1E-2 1.9E-2 1.4E-3 2.8E-4 3.7E-4 1.2E-3
dant ethylene
Aquatic Tox ATI eq. Zn 6.6E+0 2.0E-3 2.6E-4 1.7E-3 7.5E-4 9.2E-5 1.8E-5 6.3E+0 2.3E-1
water
Human Tox HTI eq. Pb air 4.9E+4 4.3E-1 1.2E-1 3.0E-1 3.4E-1 6.9E-2 6.5E-3 4.9E+4 1.3E-1
Terrestrial TTI eq. Zn air 1.7E+2 4.6E-5 8.5E-6 2.3E-5 3.9E-5 5.0E-6 5.2E-7 1.7E+2 6.1E-4
Tox
Table 2. Main substances contributing to impact assessment.
Type Substance Category Abbreviation
air Ammonia Acidification AI
air Carbon Dioxide Climate change CCI
air Halon Deplet ozone layer DOI
air Hydrocarbons (except methane) Phot. oxidant POI
air Methane Climate change, Phot. oxidant CCI, POI
air Nitrous Oxide Climate change CCI
air Nitrogen Oxides Eutrophication, Air Acidification EI, AI
air Sulphur Oxides Air Acidification AI
water Nitrates Eutrophication EI
water Nitrogen Eutrophication EI
resource Barium Sulphate Depletion non renovable re- BR
sources
resource Oil (in ground) Depletion non renovable re- BR
sources
resource Natural Gas (in ground) Depletion non renovable re- BR
sources
resource Potassium Chloride Depletion non renovable re- BR
sources
resource Phosphate Depletion non renovable re- BR
sources
Pesticides
food, soil, water Azoxystrobin Human, Terrestrial and Aquatic HTI, TTI, ATI
Tox
water Benomyl Aquatic Tox ATI
food Bromopropylate Human Tox HTI
food Captan Human Tox HTI
food, soil, water Cyromazine Human, Terrestrial and Aquatic HTI, TTI, ATI
Tox
food Iprodione Human Tox HTI
204
Using LCA for the Improvement of Waste Management in Greenhouse
Tomato Production
Muñoz, P.(a); Antón, A.(a); Montero, J.I. (a); Castells, F.(b)
a
Institut de Recerca i Tecnologia Agroalimentària. Centre de Cabrils. Ctra de Cabrils s.n. 08348,
Cabrils (Barcelona). Spain
b
Department of Chemical Enginering. ESTEQ. Universitat Rovira Virgili 43007 Tarragona. Spain,
[email protected]
Abstract
Protected cultivation is aimed to obtain higher yields by modifying and improving natural cli-
matic conditions. In spite of the fact that Mediterranean horticulture is mainly based on low-
technology cold greenhouses, and therefore input resources are less than those used by
more complex greenhouses, protected cultivation has a certain environmental impact. One of
the bottlenecks associated to this system of production is the large amount of solid waste
production. Waste can be biologic such as non-yield biomass and organic substrates, plas-
tics (cover film, mulching etc…) or minerals (such as steel and mineral substrates). Life Cy-
cle Assessment, LCA, has been used to assess different scenarios of waste treatment gen-
erated in a greenhouse tomato crop. The production of plastic waste from the different
material used was estimated in 1750 kg ha-1 year-1. Soilless closed systems, which reduce
contamination from fertilisers and use of water, generated an estimated waste of 1150 kg ha-1
year-1 of polyethylene used in substrate bags and soil cover. Finally there is a non-yield bio-
mass of 20000 kg of dry matter ha-1 year-1. Different scenarios such as landfill, incineration
and compost of the biomass have been evaluated. Results shows that for most indicators
categories the waste management is important in the life cycle of greenhouse tomato
production.
Keywords: compost, incineration and landfill
Introduction
European Union Landfill Directive (EC, 1999) requires their members to take measures to re-
duce biodegradable wastes going to landfills to 75% by 2004 and 35% by 2014. It is clear that
the current recovery and recycling practices in Europe are not able to reach the targets set by
this directive unless composting as a way of organic recycling would be performed at a com-
mercial scale (Ren, 2003).
Meanwhile in recent years, this practice is being promoted for municipal solid waste and also
for waste from livestock and gardens. For the other agricultural sectors composting activities
are not well planned. Horticulture and fruticulture are important contributors to biomass
waste, which is caused by non-harvested part of the plants when the crop is finished in horti-
culture, or by the pruning waste in fruticulture. For instance, for a tomato crop there could be
a non-yield biomass next to 20000 kg of dry matter ha-1 year-1 (Stanghellini et al., 2003)
205
Also, greenhouse industry generates and important quantity of plastics waste. The production
of plastic waste for renewing the polyethylene covers was estimated in 1000 kg ha-1 year-1.
Plastic waste from irrigation system is another 500 kg ha-1 year-1, and in addition 250 kg ha-1
year-1 generated from other materials such as film fastening bars, tutoring pincers, plastic
strings, …(Antón, 2004). Closed system is a growing system where the water drained from
the root zone is recollected and reused for irrigation of the same crop. Soilless closed systems,
which reduce contamination from fertilisers and use of water, generate an estimated waste of
1150 kg ha-1 year-1 of polyethylene used in substrate bags and soil cover.
This study uses LCA to compare different waste management for biodegradable matter and
plastic waste from the horticultural sector. The aim is to promote and give diffusion to the
importance of waste management in the global cycle of crop production to advance in
the sustainable use of natural resources and pollution prevention following the waste
policy of the European Union.
Material and Methods
Figure 1 shows the schematic process of the system under analysis, which was divided in
three systems:
1. INF that includes production of greenhouse structure and components of the fertirriga-
tion system
2. PROD production of the tomato crop, which take into account production of agro-
chemicals, fertilizers and pesticides, fertirrigation and integrated pest management
3. WASTE
Tomato was grown in a greenhouse with steel frame structure and LDPE cover film. Plants
were grown in a closed system where the substrate was perlite bags.
Three scenarios of waste management were considered: A) biomass compost taking into ac-
count avoided environmental loads and plastic landfill., B) biomass and compost landfill and
C) biomass and compost incineration. Data from landfill an incineration were obtained from
DEAM database and corrected for non-yield biomass. Non-yield biomass production was
evaluated as 0,044 g dry weight FU-1. The composition of biomass was considered as C27H38O16N
(Haug, 1993). It was accepted that the proportion of anaerobic decomposition of this biomass
was the same as that of the decomposition of glucose (Soliva, personnel communication). Fol-
lowing this approach the anaerobic decomposition of one mol of non-yield biomass produced
594 g of CO2 and 189 g of CH4 (Antón, 2004). Data from compost plant were obtained from
(AGA, 2002). To calculate the avoided environmental loads by using compost, data from
Soliva (1998)and Rovira (1997) were used.
The following impacts categories, typically used in LCA, were assessed: CCI-Climate
Change, (g eq. CO2); WMO-Depletion of the ozone layer, (g eq. CFC-11), WMO-
Photochemical oxidant formation, (g eq. ethylene); ETH-Air Acidification, (g eq. H+), BR-
206
Depletion of abiotic resources, (yr-1), CML-Eutrophication, (g eq. PO4). In greenhouse pro-
duction the main contributions to toxicity categories are pesticides. In the waste subsystem
toxicity categories are not so important in comparison with the total production system and
therefore are not presented.
Results
Comparing scenario B) landfill to A) compost of the biomass, the subsystem WASTE in
landfill disposal presents a 60 times higher score for the greenhouse effect, IPCC, and
Photochemical oxidant formation, mainly caused by the methane emissions that comes from
the decomposition of the biomass. This scenario is also 6.5 and 3.7 times higher for
acidification and eutrophication respectively. Composting the biomass also represents a
reduction on the depletion of non renewable resources due to the reductionon the use of
chemical fertilizers (figure 2).
Comparation of scenario C) Incineration of waste, plastics and biomass to A) presented seven
times higher impact score in greenhouse effect (climate change) and acidification. Depletion
of the ozone layer and Photochemical oxidant formation showed ratios of 1.3. Besides,
eutrophication had smaller effect in C) than in A). Incineration produced human and
ecosystems toxicities due to heavy metal emissions that there are not shown here because they
are small compared to toxicity produced by pesticides (figure 2).
Impacts in Scenario A) are clearly smaller than B) or C) and they are caused mainly by the
plastic landfill. Recycling plastic and the use of biodegradable plastics are subjects for future
research.
Conclusions
• The main impact of the greenhouse tomato production is due to the waste of biomass
and plastics. Especially categories such as climate change, eutrophication and
photochemical oxidant formation are strongly influenced by the different treatments.
Therefore suitable waste management is the best practicable environmental option.
• Compost of biodegradable matter is the best way of managing the waste to improve
the impact assessment for most of the considered impacts categories.
Acknowledgements The authors thank Julio Rodrigo, Rafi Caceres and Montse Soliva for
providing useful information related to compost process. This research was partially sup-
ported by INIA nº SC00-080-C2 and nº RTA03-096-C5-2.
References
AGA. 2002. Avaluació ambiental de diferents estratègies per a la gestió dels residus munici-
pals ordinaris (RMO). Grup AGA-Centre d'Innovació SIMPPLE-STQ-URV i Junta de
Residus (Departament de Medi Ambient).
207
Antón, A. 2004. Utilización del Análisis del Ciclo de Vida en la Evaluación del Impacto Am-
biental del cultivo bajo Invernadero Mediterráneo. Tesis doctoral. Programa d'Engin-
yeria Ambiental. Universitat Politècnica de Catalunya. Barcelona.
EC. 1999. Landfill Directive. European Parliament and Council Directive 1999/31/EC of 26
April.
Haug, R. 1993. The practical handbook of compost engineering. . Boca Raton, Florida. Lewis
Publishers.
Ren, X. 2003. Biodegradable plastics: a solution or a challege? Journal of Cleaner Production.
11: 27-40.
Rovira, S. 1997. Composició de l'extracte aquos d'un sòl adobat amb residus orgànics. TFC.
Escola Superior d'Agricultura de Barcelona.
Soliva, M. 1998. Residus orgànics per a l'agricultura: un tema de recerca a l'escola superior
d'agricultura de Barcelona (ESAB). Arxius de l'Escola Superior d'Agricultura, Sèrie cin-
quena, Nº 1. Barcelona.
Stanghellini, C., Kempkes, F.L.K., Knies, P. 2003. Enhancing Environmental Quality in Ag-
ricultural Systems. Proc. Proceedings of the International Symposium on Managing
Greenhouse Crops in Saline Environment. International Society Horticultural Science.
Acta Horticulturae. Pisa. Italy. 609: 277-283.
Greenhouse Irrgitation
structure system
Fertilisers Pesticides
Greenhouse
management Fertirrigation Pest management
FOREGROUND
WASTE
BACKGROUND
Figure 1. Schematic processes diagram of the system under study.
208
a) Non-yield biomass compost and plastic landfill
100%
80%
60%
40%
20%
0%
-20%
g eq. yr-1 g eq. H+ g eq. g eq. g eq.
PO4 CO2 CFC-11 ethylene
b) Non-yield biomass and plastic landfill
100%
80%
60%
40%
20%
0%
-20%
g eq. yr-1 g eq. H+ g eq. g eq. g eq.
PO4 CO2 CFC-11 ethylene
c) Non-yield biomass and plastic incineration
100%
80%
60%
40%
20%
0%
-20%
g eq. yr-1 g eq. H+ g eq. g eq. g eq.
PO4 CO2 CFC-11 ethylene
WASTE PROD INF
Figure 2. Influence of different waste management, WASTE, in six impact categories related
to subsystems PROD, crop cultivation, and INF, production of estructure.
209
LCA of the integrated production of oranges in the Comunidad Valenciana
(Spain)
N. Sanjuán*, G. Clemente and L. Úbeda
*Department of Food Technology. Universidad Politécnica de Valencia. Camino de Vera s/n.
46021Valencia. Spain. Tel. (+34) 963879366. Fax (+34) 963879839. [email protected]
Abstract
An LCA of integrated orange production in the Comunidad Valenciana was performed. The
functional unit was 1kg of oranges. The production of agrochemicals, the production and use
of energy in agriculture (for watering and machinery), and agricultural practices were studied.
Hot spots were detected. The lack of environmental information for agricultural LCA in Spain
and the need to adapt certain methodological aspects to Spanish soil and climate character-
istics have also been pointed out.
Key words: oranges, integrated farming, LCA
Introduction
Spain is the fourth largest orange producer in the world, with 5.4 million t of oranges har-
vested in 2000, of which ca. 3.6 million t are produced in the Comunidad Valenciana. Around
70% of the citric fruits of this region are exported, mainly to the European Union. Since the
1980s, both the state and regional governments have encouraged integrated farming to de-
velop more sustainable agricultural practices. In the year 2000, more than 60% of the regis-
tered orange cultivated areas and around 65% of the total production were a result of inte-
grated production (IP).
The main characteristic of IP is the strict control of agricultural practices that is carried out in
order to optimize available resources and technologies. Nevertheless, it is important to have a
holistic view of the agricultural processes to avoid the displacement of environmental prob-
lems from one part of the cycle to another. In this sense, LCA has proved to be a useful tool.
Goals and scope definition
The objective of this study is to carry out an LCA of the integrated production of oranges in
the Comunidad Valenciana and the purpose is twofold: to evaluate the environmental impact
of the agricultural practices carried out in IP of oranges in the Comunidad Valenciana, and to
contribute to developing the application of LCA methodology to agricultural practices in
Spain. The functional unit was 1kg of oranges leaving the farm gate.
Description of the system under study. System boundaries
In this study the agricultural production of oranges (variety navelina) according to IP system
(DOGV, 2001) in the Comunidad Valenciana corresponding to the year 2000 was consid-
210
ered, taking into account a farm in full production (adult trees) with 500 trees/ha, a
plantation frame of 4x5 m and an average yield of 30,000 kg oranges/ha.
From among the different systems of watering, drip irrigation using well water was
studied. An average annual water volume of 5000 m3/ha and 85% irrigation efficiency
was taken into account. No tillage was done. Four herbicide treatments were carried
out each year: the first in February with a commercial mix of glyphosate (18% w/v ac-
tive ingredient) and MCPA (18% w/v a.i.); the second in March with glyphosate (36%
w/v a.i.); the third in May, again using a commercial mix of glyphosate (18% w/v a.i.)
and MCPA (18% w/v a.i.); and the last in August with glyphosate (36% w/v a.i.).
Trees were pruned by hand and according to the IP normative, the ground pruning was
left on the soil surface. The following solid soluble fertilizers were applied together
with water: ammonium nitrate (33.5% de N), phosphoric acid (54% P2O5) and potas-
sium sulphate (13% N y 46% K2O). The applied rates were 782, 120 and 293 kg/ha re-
spectively. Once a year sheep manure was applied (3600 kg/ha).
The stages considered were agrochemical production (fertilizers and pesticides), pro-
duction of energy directly used in agriculture (for machinery and watering system) and
agricultural practices. The inventory was based on data for average farm practices pro-
vided by FECOAV (Federación de Cooperativas Agrícolas Valencianas). Data
regarding ammonium nitrate and phosphoric acid production have been obtained from
DEAM database, and data for potassium nitrate were based on Davis and Haglund
(1999). Emissions that originate in manure production were not included because they
were allocated to the manure producer. The emissions from the input of fertilisers to
the agricultural soil were obtained from a nutrient balance. Data on energy consump-
tion for pesticide production were from Green (1987) and, when the active ingredient
was not available, the extrapolation method proposed by Audsley et al. (1997) was
carried out. Transport of agrochemicals to the farm was not included. Although these
products are mostly formulated in Spain, their active ingredients are produced in many
different countries.
The production of capital goods (agricultural machinery, watering pumps and build-
ings) was not included as they have a long life. Data regarding energy consumption of
agricultural machinery were obtained from a study carried out by Gracia et al. (1986)
into orange cultivation practices in Valencia. Although these data are not current, it
can be considered that there has been little change in the practices over the years given
the large number of existing small farms. The energy for watering was computed from
the pressure and the volume of water needed. Data on energy production was obtained
from the aforementioned DEAM database.
211
Impact categories
The impact categories selected were: acidification, non-renewable resources depletion and eu-
trophication, according to the CML method; global warming and photochemical oxidant for-
mation using the WMO method; ozone depletion following the POPC method; and terrestrial
and human toxicity, with the USES method.
Results
From the impact assessment results (Figure 1) it is clear that the fertilizer production greatly
contributes to acidification (86% of total impact), mainly due to ammonia emissions in am-
monium nitrate production, and to non-renewable resources depletion (84%), attributable to
rock phosphate, potassium chloride and natural gas consumption. Fertilizer production also
contributes to the greenhouse effect (52%) and photochemical oxidant formation (42%) given
the emissions from combustion processes, and ozone depletion (25%) mainly due to Halon
301 formation during the combustion processes in the presence of F and Br. The production
and use of energy for watering and agricultural machinery also plays a decisive role in photo-
chemical oxidant formation and ozone depletion (33% and 41% of total impact, respectively).
The agricultural practices are what mainly lead to eutrophication (99.9%), due to nitrate
leaching. These practices mostly responsible for human and terrestrial toxicity (96 and 85%,
respectively) mainly due to the use of copper as fungicide and also to 31% of the greenhouse
effect. The production of pesticides causes 15% of the depletion of non-renewable resources
(mainly because of the extraction of Cu), 34 % of ozone layer depletion and 25% of photo-
chemical oxidant formation, the latter two impacts being the result of fuel combustion.
Agriculture Machinery and watering Fertilizer production Pesticide production
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
CML-resource
photochemical
depletion [Halon
Eutrophication
USES 2.0-human
USES 2.0-
WMO - Global
acidification
ecotoxicity
terrestrial
1301 (CF3Br)]
POCP-ozone
CML-air
depletion
oxidation
warming
WMO-
CML-
toxicity
Figure 1. The environmental profile of IP oranges in Valencia.
212
Discussion
Hot spots and improvements
Nowadays, the eutrophication caused by agricultural practices is one of the main problems in
Valencia. In order to prevent this, nutrient balances should be performed. Regarding the im-
pact of fertilizer production, a comparative study of the manufacturing process of fertilizers is
needed to determine if there are significant differences in the efficiency of use of resources
and the emissions caused in the production process. In order to reduce the depletion of rock
phosphate alternative sources of P should be applied, taking also into account that in Valen-
cian soils, P is mostly immobilized.
Methodology problems
One of the critical problems when performing this LCA was the lack of data, mainly with re-
spect to fertilizers, pesticides and machinery production. Other important aspects are the
emissions derived from manure. This lack of data has obliged us to use those agrochemicals
available in the databases instead of selecting, in specific cases, the most representative ones.
As for the methodology, it is also important to point out that the evaluation methods for toxic-
ity consider only specific soil and climate characteristics. For example, in this LCA, copper
contributed greatly to toxicity impact categories, yet we must bear in mind that given the
characteristics of Valencian soils (basic pH), copper is retained and its toxic effect decreases.
References
Audsley, E. (coordinator) Alber, S., Clift, R., Cowell, S., Gaillard, G., Hausheer, J., Jolliett,
O., Kleijn, R., Mortensen, B., Pearce, D., Roger, E., Teulon, H., Weidema, B. and van
Zeirs H. 1997. Harmonisation of Environmental Life Cycle Assessment for Agriculture.
Final Report. Concerted Action AIR3-CT94-2028. Silsoe Research Institute. Bedford,
United Kingdom.
Davis, J. and Haglund, C. 1999. Life Cycle Inventory (LCI) of Fertiliser Production. Fertiliser
Products Used in Sweden and Western Europe. SIK report nº654. Chalmers University
of Technology. The Swedish Institute for Food and Biotechnology (SIK). Gothenburg,
Sweden.
DOGV (Diari Oficial de la Comunitat Valenciana). 2001. 3 de gener de 2001. RESOLUCIÒ
de 23 de novembre de 2000, del director general d’Innovació Agrària i Ramaderia, per
la qual s’estableixen les normes per a la Producció Integrada en cítrics en l’àmbit de la
Comunitat Valenciana. [2000/9741].
Gracia, C., Val, L. and Juste, F. 1986. Consumo Energético de las Operaciones Mecanizadas
en el Cultivo de los Agrios. Investigación Agraria. Producción y Protección de
Vegetales. 1 (2), 185-207.
Green, M. 1987. Energy in Pesticide Manufacture. Distribution and Use. In: Helsel, Z. (ed.)
Energy in Plant Nutrition and Pest Control. Elsevier. Amsterdam, The Netherlands.
213
Environmental Impact Assessment at Commercial Dairy Farms
Comparison of LCA, ecological footprint analysis, and input-output accounting
De Boer, I.J.M.*, Iepema, G., and Thomassen, M.A.
Animal Production Systems Group, Wageningen Institute of Animal Sciences
PO Box 338, 6700 AH Wageningen, Wageningen University, Wageningen, The Netherlands
Abstract
The aim of this article was to determine effectiveness of environmental indicators obtained
from input-output accounting (IO), Ecological Footprint Analysis (EFP) and Life Cycle As-
sessment (LCA) to show differences among commercial dairy farms. In total, 11 environ-
mental indicators were quantified and correlated. Results show that input-output accounting
of nutrients yields effective indicators with respect to eutrophication and acidification. Re-
garding land and energy use, however, EFP and LCA yield similar indicators. EFP, subse-
quently, sums land and energy use into one final unit, i.e., biologically productive area, which
is shown to be ineffective. LCA appears a useful technique to deduce effective, environ-
mental indicators. LCA results with respect to acidification potential, for example, showed
that on-farm NH3 emission per ha or per kg milk is an effective indicator to show differences
among production systems and to improve performance with a system.
Key words: organic dairy production, environmental impact assessment, input-output ac-
counting, ecological footprint analysis, life cycle assessment.
Introduction
To show differences in environmental impact among production systems, such as for example
organic and conventional milk production, environmental impact should be assessed at a large
number of commercial farms for each production system of interest. In principle, two differ-
ent approaches can be used to assess the environmental impact at commercial agricultural
farms. First, the input-output accounting approach (IO), which computes the difference in nu-
trients entering and leaving the farm gate, while the farm itself is considered a black box.
Second, a “cradle to grave or farm-gate” approach, which computes the integrated environ-
mental impact of an agricultural activity throughout its entire life cycle, such as Ecological
Footprint analysis (EFP) and Life Cycle Assessment (LCA). The aim of this article is to de-
termine the effectiveness of environmental indicators deduced from IO, EFP and LCA, to
show differences among commercial dairy farms.
Material and methods
Data. The environmental impact of eight commercial organic dairy farms was assessed, using
IO, EFP, and LCA. These eight farms participated in a demonstration project of Dutch or-
ganic dairy farmers, the so-called BIOVEEM project (Smolders and Wagenaar, 2004). In this
project, a large number of on-farm data are gathered regularly in order to improve farm per-
formance. Additional on-farm and off-farm data were required, which were obtained by ques-
tionnaires and expert consultation.
214
Input-output accounting. For each dairy farm, we computed the annual surplus of nitrogen (N)
and phosphate (P2O5) using input-output accounting at farm level. This surplus is derived
from the difference between farm inputs and farm outputs, with respect to N and P2O5, while
the farm itself is considered a black box (Ondersteijn et al., 2002). Farm inputs of N and P2O5
considered were: clover N-fixation, deposition, and import through roughage, concentrates,
animals and manure. Farm outputs of N and P2O5 considered were: animal products, i.e.,
milk, meat, manure or living animals, and plant products, i.e., roughage or crops. The differ-
ence between inputs and outputs is called the farm surplus, and is assumed to be lost to the
environment. This farm surplus is expressed per ha farm area. To further diversify the N sur-
plus, we first corrected this surplus for annual NH3 emission. The corrected N surplus per
farm area, therefore, includes environmental losses of NOx, N2O and NO3- only. This ap-
proach, therefore, yields three indicators, i.e., NH3 emission (kg N) per ha, surplus of N (kg)
per ha, and surplus of P2O5 (kg) per ha.
Ecological Footprint Analysis. For each dairy farm, an ecological footprint (EFP) was com-
puted. A farms EFP is the biologically productive area (BPA) needed to produce all resources
and to absorb waste (i.e., CO2 from fossil fuel combustion) generated by that farm (Wacker-
nagel and Rees, 1996). To compute EFP of a dairy farm we, therefore, keep track of land and
energy requirements of all resources imported into the farm or used on the farm. Subse-
quently, land use and energy use is summed into BPA, assuming that 1 ha of wood land ab-
sorbs all CO2 released during combustion of 100 GJ of energy. To determine BPA of co-
products, such as feed ingredients, economic allocation is used. Finally, a farms BPA is allo-
cated to milk production, based on economic allocation, and expressed per kg of Fat and Pro-
tein Corrected Milk (i.e., kg FPCM = (0.337 + 0.116×%fat + 0.06×%protein)×kg milk pro-
duction (CVB, 2000). EFP, therefore, yields one indicator, i.e., BPA (m2) per kg of FPCM.
Life Cycle Assessment. For each dairy farm, a “cradle to farm-gate” LCA of one kg of FPCM
was performed (de Boer, 2003). The following impact categories were assessed: land use, en-
ergy use, global warming potential (GWP; CO2, CH4, N2O), eutrophication potential (EP;
NOx, PO43-, NO3-, NH3, NH4+) and acidification potential (AP; SO2, NOx, NH3). Economic al-
location was used. LCA, therefore, yields the following environmental indicators: ha land
use/kg FPCM, MJ energy use/kg FPCM, GWP in CO2-eq/kg FPCM, EP in NO3- eq/ha or in
NO3-eq/kg FPCM, AP in SO2 eq/ha or in SO2 eq/kg FPCM.
Results and discussion
Input-output accounting. For each farm, results of environmental indicators computed are
given in Table 1. The average N surplus was 82.5 kg per ha, with a standard deviation of 61.6.
Variation in N surplus mainly was due to variation in clover N-fixation and import of rough-
age. The average P2O5 surplus was 5.4 kg per ha, with a standard deviation of 16. This ex-
treme variation was due to the fact that halve of the farms had a negative and half of the farms
a positive P2O5 surplus.
215
Ecological footprint analysis. BPA of an organic dairy farm on average was 1.85 m2 per kg of
FPCM, with a coefficient of variation of 18%. EFP sums land use and energy use in one final
unit (i.e., BPA) by assuming that 1 ha of woodland fixes all CO2 released during combustion
of 100 GJ of energy. For farm no. 1, for example, BPA/kg FPCM is 1.39, which is the sum of
1.21 ha due to actual land use and 0.18 ha required for CO2 absorption due to energy combus-
tion (see Table 1). No correlation, however, was found between BPA from land use and BPA
from energy combustion for eight farms studied. Summation of BPA from land use and BPA
from energy combustion, therefore, implies loss of information regarding the environmental
performance of commercial farms, and, therefore, seems inappropriate.
Life cycle assessment. An organic dairy farm used an average of 1.6 m2 land per kg FPCM.
From this, 69% is on-farm land, such as grassland and arable land, whereas 31% is off-farm
land required mainly for cultivation and transport of feed. An organic dairy farm used an av-
erage of 2.48 MJ energy per kg FPCM. From this energy use, 40% is direct, on-farm energy
consumption, whereas 60% is required for the production of farm inputs like concentrates,
purchased roughage and extern labour. GWP of an average farm was 1.81 kg CO2-eq/kg
FPCM, of which 78% is due to on-farm emission of CH4 and N2O mainly, each explaining
around 50%. EP of an average farm was 82.1 g NO3- eq/kg FPCM or 720.3 kg NO3- eq/ha, of
which around 50% is due to on-farm emission of mainly NO3-, PO4-and NH3. Off-farm EP is
explained mainly by cultivation and transport of feed. AP of an average farm was 104 kg SO2
eq/ha or 11.8 g SO2 eq/kg FPCM. AP is for around 70% due to on-farm emission of mainly
NH3.
Correlations between environmental indicators. As described previously, EP per ha was for
47% due to on-farm leaching of NO3-, PO4- and emission of NH3. A correlation was found be-
tween the LCA indicator EP per ha (i.e., g NO3- eq per ha) and the input-output accounting in-
dicator N surplus per ha (correlation 0.8; p=0.01) and between EP per ha and P2O5 surplus per
ha (correlation 0.75; p=0.03). Similarly, a correlation was found between AP and NH3 emis-
sion. Expressed per ha farm area, AP and NH3 show a correlation of 0.92 (p=0.001). AP per
ha farm area also correlated to Dutch Livestock Units (LU) per ha (0.92; p=0.001), which is
due to the fact that computation of NH3 emission used largely depends on animal numbers.
No correlation, however, was found between fossil fuel use and GWP per kg FPCM (correla-
tion 0.2; p=0.6), which is due to the fact that CH4 and N2O and not CO2 are main contributors
to GWP.
Conclusion
Unlike input-output accounting, ecological footprint analysis and especially LCA of milk
production at commercial dairy farms appears time-consuming. LCA results, however, show
insight into the environmental impact of various processes in the chain of milk production.
This insight is highly relevant to deduce environmental indicators necessary to show differ-
ences in environmental impact among production systems. Regarding land and energy use,
e.g., ecological footprint analysis and LCA yield similar indicators. Regarding AP, for exam-
216
ple, a comparison of dairy production systems could be based on the indicator on-farm NH3
emission per kg FPCM or per ha farm area. Regarding EP for current dairy production sys-
tems, N and P surplus per ha are indicators that detect variation among and within production
systems.
References
De Boer, I.J.M. 2003. Environmental impact assessment of conventional and organic milk
production. Livestock Production Science 80:69-77.
CVB, 2000. Table book animal nutrition; Livestock norms and feeding values. Central Buri
for Animal, Lelystad, The Netherlands.
Ondersteijn, C.J.M, A.C.G. Beldman, C.H.G. Daatselaar, G.W.J. Giesen, Huirne, R.B.M.
2002. The Dutch Mineral Accounting Systems and the European Nitrate Directive: Im-
plications for N and P management and farm performance. Agriculture, Ecosystems and
Environment 92:283-296.
Wackernagel, M. W. Rees. 1996. Our ecological Footprint. Reducing Human Impact on the
Earth. New Society Publishers, Canada.
Smolders, G., Wagenaar, J.P., 2004. BIOVEEM. Practice and research on 10 organic dairy
farms 1997-2001. Applied Research, Animal Sciences Group, Wageningen University
and Research Center, Lelystad (in press).
Acknowledgements
The authors thank Trudy Straetemans for her contribution to this research during her M.Sc.
thesis work.
217
Table 1. Results of environmental indicators obtained from input-output accounting, ecological footprint analysis and life cycle assessment for
eight organic dairy farms studied.
Farm 1 to 8
Environmental indicators 1 2 3 4 5 6 7 8 Mean (SD)
Input-output accounting
kg N surplus/ha 47.3 197.3 49.1 87.3 152.8 25.0 31.8 69.7 82.5 (61.6)
kg NH3/ha 35.1 43.8 20.4 32.8 30.4 40.6 25.4 42.2 33.8 (8.3)
kg P2O5 surplus/ha -4.3 21.5 12.3 -0.5 29.7 -15.3 -12.0 11.8 5.4 (16.0)
Ecological Footprint
m2 BPA/kg FPCM 1.39 2.38 2.02 1.79 2.07 1.37 1.85 1.91 1.85 (0.33)
• land 1.21 2.11 1.71 1.53 1.64 1.19 1.72 1.68 1.60 (0.30)
• energy combustion 0.18 0.27 0.31 0.26 0.43 0.19 0.13 0.23 0.25 (0.09)
Life Cycle Assessment
land use (m2/kg FPCM) 1.21 2.11 1.71 1.53 1.64 1.19 1.72 1.68 1.60 (0.30)
energy use (MJ/kg FPCM) 1.78 2.65 3.11 2.55 4.26 1.88 1.31 2.31 2.48 (0.91)
GWP (kg CO2 eq/kg FPCM) 1.17 1.77 1.64 3.85 1.76 1.16 1.56 1.60 1.81 (0.86)
EP (g NO3- eq/kg FPCM) 65.0 101.4 74.4 97.6 153.3 34.3 37.7 93.4 82.1 (38.6)
EP (kg NO3- eq/ha) 756.1 1068.5 449.1 977.7 1132.8 357.8 187.2 833.2 720.3 (350.7)
AP (g SO2 eq/kg FPCM) 10.1 15.0 9.8 11.0 12.4 10.3 10.9 15.0 11.8 (2.14)
AP (kg SO2 eq/ha) 117.0 158.4 59.0 110.2 91.7 107.8 53.9 134.1 104.0 (35.4)
GWP = Global Warming Potential, EP = Eutrophication Potential, AP = Acidification Potential.
218
A systematic description and analysis of GHG emissions resulting from Ire-
land’s milk production using LCA methodology
Casey, J.W*, Holden, N.M
Department of Agricultural and Food Engineering, University College Dublin, Earlsfort Terrace,
Dublin 2, Ireland. Tel. +353(1)7165557 Fax: +353(1)4752119, [email protected]
Abstract
Life Cycle Assessment (LCA) methodology was used to assess the greenhouse gas (GHG)
emissions resulting from rotational grazing for dairy production in Ireland. A system defini-
tion was formulated with average herd size (47 cows), total inorganic nitrogen usage (175 kg
N ha-1), and concentrate consumption (818 kg animal yr-1) estimated from national data.
Mechanisation inputs were also included. System processes that contribute to GHG emissions
are identified. Emission factors were ascribed to the identified GHG contributors to assess the
effect of 1litre of milk production (FU) on climate change.
Keywords: GHG emissions, LCA methodology, dairy production, Ireland
Introduction
Greenhouse gas emissions attributable to dairy farms arise primarily from three main sources
- methane (CH4) from livestock, nitrous oxide (N2O) from pasture improvement and carbon
dioxide (CO2) from energy generated by the combustion of fossil fuels. The objectives of this
study were to: (i) produce a system inventory for all input and outputs; (ii) identify GHG
contributers to the system; and (iii) assess GHG emissions per functional unit (FU) using part
of the LCA methodology.
System description
The systems boundary is defined by the environmental burden of Irish milk production with
respect to GHG emissions. Political boundaries are not considered as limits to the system.
Distribution and milk processing are beyond the scope of the boundary. In Ireland over 1.2 m
dairy cows on c. 27,000 farms consume grass in the field or as conserved forage, with addi-
tional feed as concentrates (40% home grown and 60% imported). Weather permits 220-240
days grazing and housed feed is either grass silage and/or forage maize apart from concen-
trates. The system can operate over a range of farm sizes to produce 5.55 x 109 L of milk
quota (European Community, 2003).
Materials & Methods
Data were collected from National Farm Surveys and the Fertiliser Survey (all produced by
Teagasc the Irish Agricultural and Food Development authority) and by personal communica-
tions. It was assumed that all manure produced on the farms is used as fertiliser and nitrogen
excretion data were taken from Smith and Frost (2000). A system inventory was constructed
(Table 1) and for each element a GHG emission (scaled to CO2 as a reference) is being allo-
219
cated (IPCC, 1996). The defined FU is: the production of national milk quota in litres, scaled
to the output of one litre over a time frame of one year (total greenhouse gas emissions per
functional unit – TGE/FU). The global warming potential (GWP) index will be used to assess
the system for total GHG emissions per FU.
Results
The Irish system can be summarized as follows: average herd size is 47 cows, mean fertilizer
applied is 175 kg N ha-1, the mean concentrate used is 819 kg cow yr-1 and the average milk
output is 4822 L cow yr-1. The sources of GHG emissions are outlined in Table 1. The calcu-
lated TGE/FU is 1.36.
Discussion and Conclusions
To assess Irish milk production for GHG emissions it was thought that the LCA methodology
could be partly used. However there are a number of limitations associated with applying the
LCA methodology to agricultural production. For example using the LCA framework for the
study suggested that elimination of processes with small emissions would make a more robust
and reliable inventory. However the project set out to assess the entire environmental burden
of milk production in terms of GHG emissions irrespective of the magnitude of the emission.
Another issue is that most of the emission data are not from Ireland. The data are mainly UK,
German, Danish and Dutch. For example dung emissions were taken from UK data and ma-
nipulated to suit the Irish system. A figure of 1.36 TGE/FU is presented from preliminary ex-
amination of the system not including the contributions of concentrate feed emissions and
diesel usage associated with silage making.
Acknowledgements
The authors would like acknowledge the assistance of Sanderine Nonhebel, Jose Potting,
Emiel Elferink and their colleagues at the IVEM University of Groningen, Netherlands for
their assistance. This work was funded by the Environmental Protection Agency under the
National Development Plan (2000-2006)
References
European Community, 2003. COMMISSION REGULATION (EC) No 572/2003.
IPCC, Intergovernmental Panel on Climate Change, 1996. Climate Change 1995. The Science
of Climate Change. Contribution of Working Group I to the Second Assessment Report
of the Intergovernmental Panel on Climate Change. Cambridge University Press, UK.
Smith, K.A. & Frost, J.P., 2000. Nitrogen excretion by farm livestock with respect to land
spreading requirements and controlling nitrogen losses to ground and surface waters.
Part 1: Cattle and sheep. Bioresource Technology 71, 173-181.
220
Table 1. Identified GHG contributors (Preliminary), kgs CO2 equivalents.
Category Input Output Methane Carbon dioxide Nitrous Oxide
Size in ha 40
Cows in milk 47.2 4720
Other stock 69.7 3485
-1
Stocking rate (LU ha ) 1.8
-1
Milk Output (L cow yr ) 4822
Total litres per farm 227598
Farm management
Fertiliser production (kg N) 7000 12 18620 94
Importation (fert. km) 1040 0.36 234 0.05
Merchant to farm (fert. km) 40 0.32 208 0.05
-1
Fertilizer applied (kg N ha ) 175 85
Diesel used (kg) 3204 2 11406 2
Electricity (kwh) 8338 6504
Manure management
Slurry indoors Storage 257 4
Dung in field 0.001 48
Collecting yard 0.002 0.00004
Spreading 7.7 20
Other cattle
Slurry indoors Storage 194 2.5
Dung in field 0.04 20
Spreading 5 13
Total kg CO2 eqv. 182362 36971 89684
Total 309017
TGE/FU 1.36
221
Life cycle assessment results and related improvement potentials for oat
and potato products as well as for cheese
Katajajuuri, J.-M.a,*, Virtanen, Y.a, Voutilainen, P.a and Tuhkanen, H.-R.b
a
MTT Agrifood Research Finland, Environmental Management, FIN-31600 Jokioinen, FINLAND
*
[email protected], Phone: +358-(0)3-41881, Fax: +358-(0)3-4188 2222
b
MTT Agrifood Research Finland, Karilantie 2A, 50600 Mikkeli, Finland
Abstract
LCA studies were performed for oat flakes, potato flour, hard cheese and IQF (Individually
Quick-Freeze) potato gratin with cheese sauce produced in Finland. The main objective of
the study was to compile reliable environmental impact data for all stages of the current pro-
duction and supply systems and to identify and evaluate improvement potentials of them. No
comparisons between products were made. Data for the system models were acquired from
the field thus providing a reliable basis to analyse the sources of environmental impacts (“hot
spots”) and to consider respective improvement possibilities. In the cheese production sys-
tem, special attention was paid on the management of complicated nutrient flows and wash-
outs at the farms. As one part of improvement assessment, the effect of animal husbandry
practices, like different feeding strategies, on environmental impacts in the entire cheese sys-
tem were studied.
Keywords: Life Cycle Assessment, food, agriculture, field oriented data acquisition, im-
provement assessment.
Background
Finnish agricultural and food industry and trade, in co-operation with Finnish research insti-
tutes MTT and VTT, started in early 2000 a process to produce reliable environmental per-
formance data on relevant Finnish food production and supply systems. This joint national ef-
fort was expected to support R&D and innovation activities in food production and had a
further aim to improve the environmental performance of products according to the novel so-
cietal responsibility principle and integrated product policy (IPP).
So far the studies have covered feed barley and oat, silage, pasture, dry hay, feed concen-
trates, potatoes, oat, milk and cheese. The studies were a part of Environmental Cluster Re-
search Programme and national quality strategy for food in Finland, which covers the entire
food chain from farm inputs to the consumer. The studies were financed by the Ministry of
Environment and Ministry of Agriculture and Forestry, as well as by the participating compa-
nies and the research organisations MTT and VTT. The main objectives of the studies have
been to compile reliable environmental impact data for all process stages involved in the sys-
tems considered, to identify and assess the significance of the impact sources (“hot spots”),
and to draw and evaluate possibilities to improve the environmental performance of the sys-
tems. The central data on the current supply webs were based on empirical investigation of
the real processes. One important aim behind this was to get the different parties involved in
222
the supply webs to learn more about product oriented environmental management, respective
assessment of environmental impacts, and related benefits, i.e. learning by doing. This gives a
real possibility to seek continuous improvements in the supply chains. Principles and benefits
of supply web management based LCA are more widely presented and discussed by Virtanen
and Poikkimäki (2003) and Poikkimäki and Virtanen (2003).
Methods and field oriented data acquisition
The central data was based on empirical investigation of the real processes. Mainly the parties
of the supply chains did fieldwork for the data acquisition. Hence, data from industrial proc-
esses came from Finnish plants producing and processing raw materials and products needed
in the systems. Average Finnish grid electricity data was used for electricity. Local energy
production, like steam and utilisation of biogas from landfills etc, was taken as it currently
appeared in the systems. Consumer products were assumed to be delivered over Finland ac-
cording to the current regional market shares. Emissions of delivery webs were modelled us-
ing realistic delivery routes with initial loading, retail stops, and final discharge of return load
each. Logistics was modelled in collaboration with Finnish logistics companies, including re-
tail product losses. The data for retail refrigeration and freezing was drawn using nominal
electricity consumptions of the refrigeration devices and assessed average product through-
puts of the cold stores. Household cooking (oat meal and IQF gratin) and freezing (IQF
gratin) were assessed based on appliance consumptions and respective times.
Milk production data and detailed feeding values were drawn from about 700 farms. Data ac-
quisition for farm cultivation was based on individual farm level cropping plans and realiza-
tion of these, and on interviews of farmers. Washouts (N and P) and air emissions (NH3, CH4,
N2O) were calculated on the basis of nutrient balances, nationally applied P washout models,
IPCC reference manual (1997) for greenhouse gases and Finnish agricultural knowledge at
MTT. Data acquisition methods and sources are described in detail in project reports (Vouti-
lainen et al. 2003a, Katajajuuri et al. 2003, Voutilainen et al. 2003b.) Tools to handle input-
output data collection electronically and generation of LCI data for farm level are under con-
struction in Finland. Technique to be used of that might be a part of Finnish quality databank
(see e.g. Katajajuuri & Loikkanen, 2000 or
http://www.mmm.fi/english/agriculture/food_quality.htm). National IO tables were consid-
ered too generic for the assessments, and for seeking real improvements in particular chains
and farms. The principal starting point was to model systems as they currently were. Because
the systems studied were quite complicated with several by-products and various secondary
outputs, allocation couldn’t be completely avoided. Allocation principles utilised have been
discussed in Katajajuuri and Voutilainen (2002). However, it was found that allocation prin-
ciples have a big effect on the results, and in the future avoiding allocation through system
expansion e.g. for milk and meat production should be carried out as proposed by Cederberg
and Stadig (2001) and Weidema (2001). This will be especially important when making com-
parisons with different kinds of foodstuffs. The management of the data uncertainties and the
ranges of data variation is an important research area, especially in agricultural systems. For
223
this reason the development of quantitative uncertainty assessment was also included in the
study using the gratin case for testing. Stochastic modelling, performed by Monte Carlo and
Latin Hypercube simulation methods, was chosen as techniques for uncertainty assessment
(see Voutilainen et al. 2003b).
LCA results in general terms
Nitrogen and phosphorus emissions to water and the respective eutrophication potentials of
the studied systems were nearly all due to the nitrogen and phosphorous emissions from culti-
vation, and in milk production also due to washing of milk storage tanks and milking equip-
ment on dairy farms. Other processes of the systems had very little influence on them. In-
stead, energy-related emissions, like carbon dioxides, varied considerably depending on the
characteristics of many processes of systems, such as the fuels assumed for energy produc-
tion, degrees of industrial refining, yields of processing, and need and efficiencies of freezing.
The contributions of the different production stages to acidification and global warming po-
tentials (GWP) differ substantially between the product systems. In cheese system, when dai-
ry cattle involved, acidification and GWP were dominated by the milk farm: crop and milk
production. The case studies clearly showed that the environmental impacts of the systems are
e.g. case-, allocation principle-, and production system dependent. Therefore, generalisation
of the contributions of the life cycle stages to other food chains should be avoided. When
household cooking was included, its contribution especially to global warming potential of
the systems was found to be enormous, with microwave as a positive exception.
Improvement assessment of the dairy farms in cheese production system
Data sets, including feeding inputs, from 700 milk production farms were divided into three
categories: farms producing milk at overall average level (7155 kg milk/cow/year), at lowest
level (average 6409 kg from 143 farms) and at highest level (average 7906 kg from 144
farms). Increase in milk output is influenced by many inter-related farm parameters. The pur-
pose of the assessment was to form an overview of the importance of the milk output as such.
When assessing environmental impacts of different feeding strategies, with different levels of
concentrates and crude proteins of concentrates, farm data could not be used because there are
many factors influencing milk output and nutrients excretions. Instead, results from 14 milk
production trials with 112 different diets were used to study the N and P excretion and wash-
outs with different level of concentrates of the total feeding. Furthermore, results from 27
milk production trials with 186 different diets were used to study the N and P excretion and
washouts with different crude protein contents in concentrates (Nousiainen et al. 2003 and Yr-
jänen et al. 2003). In addition to these contributions of crop yields, nitrogen inputs, lifetime of
dairy cattle and density of dairy cattle per ha to environmental impacts of the system were as-
sessed. In studied region, average figures of dairy cows were 3,1 calving times and 2.5 lacta-
tion periods. As an example of results, increased number of lactation period decreased rapidly
N emissions per cheese ton. This was a result of a dilution of feeds needed to raise a two-year-
old heifer and a greater production capacity in later lactations.
224
Acknowledgements
We wish to acknowledge The Finnish Environmental Cluster Research Programme and the
Ministry of Agriculture and Forestry and the Ministry of Environment and companies (Ruo-
kaKesko, Kemira Agro, Lännen Tehtaat, Raisio Group, Suomen Rehu & Valio) for the finan-
cial support, and thank their several representatives that worked in this research, and our re-
search colleagues.
References
Cederberg, C. & Stadig, M. 2003. System Expansion and Allocation in Life Cycle Assessment of
Milk and Beef Production. Int J LCA 8 (6) 350-356 (2003).
Katajajuuri, J.-M. & Loikkanen, T. Enhancing competitiveness through green innovation - Finnish
food industry committed to environmental data production - LCA pilot case for barley. 9th
International Conference of the Greening of Industry Network, Sustainability at the Millen-
nium: Globalization, Competitiveness, and the Public Trust, January 21-25, 2001, Bangkok,
Thailand. 9 p.
Katajajuuri & Voutilainen (2002). LCA case studies on Finnish food production chains - man-
agement of secondary inflows and outflows. SETAC Europe 10th LCA Case Studies Sym-
posium "Recycling, close-loop economy, secondary resources", 2-3 December, Barcelona,
Spain. 6 p.
Katajajuuri, J.-M., Voutilainen, P., Tuhkanen, H.-R. & Honkasalo, N. 2003. Environmental im-
pacts of oat flakes. Maa- ja elintarviketalous 33: 47 p. In Finnish.
http://www.mtt.fi/met/pdf/met33.pdf
Nousiainen, J., Kytölä, K., Khalili, H. & Huhtanen, P. 2003. Methods of improving N utilization
through feeding strategies on dairy farms. In Uusi-Kämppä, Markku Yli-Halla & Kaarina
Grék: Reducing environmental loading from dairy farming. Maa- ja elintarviketalous 25: p.
26-39. In Finnish. http://www.mtt.fi/met/pdf/met25.pdf
Poikkimäki, S & Virtanen, Y. 2003. Supply web integrated life cycle assessment. PTR report NO.
49. ISBN 951-8988-30-7.
Virtanen, Y. & Poikkimäki, S. 2003. Applicability of the Product Life-Cycle Analysis. European
Brewery Convention, 29th International Congress, Dublin Ireland 2003.
Voutilainen, P., Tuhkanen, H.-R., Katajajuuri, J.-M., Nousiainen & Honkasalo, N. 2003a. Envi-
ronmental impacts and improvement possibilities of Emmental blue-label cheese. Maa- ja
elintarviketalous 35: 91 p. In Finnish. http://www.mtt.fi/met/pdf/met35.pdf
Voutilainen, P., Katajajuuri, J.-M., Tuhkanen, H.-R. & Honkasalo 2003b. Environmental impacts
of Kesäpöytä cheese-cream potato gratin and Pirkka potato flour. Maa- ja elintarviketalous
34: 54 p. In Finnish. http://www.mtt.fi/met/pdf/met34.pdf
Yrjänen, S., Nousiainen, J., Kytölä, K., Khalili, H. & Huhtanen, P. The feasibility of minimising
faecal excretion on dairy farms. In Uusi-Kämppä, Markku Yli-Halla & Kaarina Grék: Re-
ducing environmental loading from dairy farming. Maa- ja elintarviketalous 25: p. 13-25. In
Finnish. http://www.mtt.fi/met/pdf/met25.pdf
Weidema, B. 2001. Avoiding co-product allocation in life-cycle assessment. Journal of Industrial
Ecology 4(3): 11-33.
225
Effects of Intensification of Dairy Farming in New Zealand on Whole-
system Resource Use Efficiency and Environmental Emissions
Ledgard, S.F.1*, Finlayson, J.D.1, Patterson, M.G. 2, Carran, R.A.3, and Wedderburn, M.E.1
1
AgResearch Ruakura, Private Bag 3123, Hamilton, NZ
2
Massey University, Private Bag 11222, Palmerston North, NZ
3
AgResearch Grasslands, Private Bag 11008, Palmerston North, NZ
*AgResearch Ruakura, Private Bag 3123, Hamilton, New Zealand,
phone ++6478385133, fax ++6478385155, [email protected]
Abstract
A life cycle assessment (LCA) approach was used to estimate whole system resource use
and environmental emissions for the dairy industry in the Waikato region of New Zealand.
Application of a “top-down” LCA method using input/output analyses revealed marked differ-
ences in the relative contribution between dairy farms, dairy factories and indirect contribu-
tors. For example, the relative use of energy by farms : factories : indirect contributors was 1
: 3.0 : 1.8, whereas the corresponding relative greenhouse gas emissions were 1 : 0.2 : 0.3.
A “bottom-up” LCA method was used to evaluate the whole farm-system (dairy farm + graz-
ing and forage land) effects of intensification using nitrogen fertiliser or forage crop integra-
tion. Fertiliser nitrogen increased production and economic efficiency but decreased envi-
ronmental efficiency. In contrast, increased use of forage produced off-farm increased the
total use of land and production efficiency, with no loss in environmental efficiency (per litre
milk).
Introduction
Milk production on dairy farms in New Zealand has been steadily increasing over time. This
has occurred through a number of factors including increased feed supply through greater use
of nitrogen (N) fertiliser and increased use of supplementary feeds. The effects of this on effi-
ciency of use of resources such as energy, nutrients and land are uncertain.
Evaluation of dairy farm efficiency should go beyond consideration of only the dairy farm
unit to incorporate total land and use of resources. Ideally, a whole-system evaluation should
account for other indirect contributors (e.g. energy used for fertiliser production) and the dairy
processing system. As well as resource use efficiency, the whole system evaluation should
account for the impacts of intensification on air and water quality. In this paper, LCA meth-
odologies are applied to examine resource use efficiency and environmental emissions in the
dairy production chain, and to determine the impacts of intensification practices.
Keywords. Dairy, emissions, intensification, resources, whole-system.
Methodology
Waikato Dairy Industry
A “top-down” LCA approach used Input/Output matrices (Patterson and McDonald, 1996) for
23 or 48 sector models of the Waikato and NZ economies for a range of resource uses and
226
emissions. These were used to calculate life-cycle multipliers and define the indirect contribu-
tion of all industries or sectors to the Waikato dairy industry. This evaluation was based on
1997/98, a recent period when detailed national data was available from StatsNZ. There are
about 6000 dairy farms in the Waikato region (about 40% of New Zealand’s total) and nine
dairy processing factories.
Farm economic, productivity and resource use data for Waikato dairy farms was obtained
from dairy industry statistics and from a database (Dexcel ProfitWatch) of 128 farms. The
OVERSEER® nutrient budget model (Ledgard et al., 1999) was used to estimate N leaching
and total N emissions to waterways from the “average” Waikato dairy farm. It was also used
to estimate IPCC-based emissions of the greenhouse gases (GHG), methane and nitrous ox-
ide. Farm emissions of CO2 were estimated using fuel and electricity data, and emission fac-
tors from Wells (2001). Milk input, fuel use and nutrient discharge information for the nine
Waikato dairy factories was obtained from the Dairy companies. Energy use data was derived
from NZ statistics.
Dairy Farm Intensification
Data for 2000/2001 for the average 83 ha Waikato dairy farm was derived from the Dexcel
ProfitWatch database. This farm was estimated to use 15 ha of an average intensive Waikato
beef farm for grazing non-lactating animals (grazing replacement animals off farm is common
practice in New Zealand). Similarly, the farm purchased forage which would be produced on
2 ha of a ‘typical’ double-cropping block (yielding 26 t DM/ha/yr of maize and oats silage).
Methodology outlined in the previous section was applied in the evaluation of:
1. Base farm system, with 83 ha dairy farm producing 10 m3 milk/ha,
2. Base farm system plus extra 200 kg fertiliser-N/ha on the dairy farm producing 12 m3
milk/ha, and
3. Base farm system plus extra 2 t DM/ha of silage (from an extra 6.4 ha of forage
block); assumed to produce 12 m3 milk/ha on the dairy farm.
All options were assumed to carry 2.8 cows/ha on the dairy farm. Milk production responses
were based on average research data. A “bottom-up” LCA method (e.g. Cederberg, 1998) was
used for whole-system analysis to the point of milk in the farm vat ready for collection and
processing.
Results and Discussion
Resource use and emissions
Relative differences between farms and dairy factories in the use of resources and level of
emissions varied greatly (Table 1). For example, the relative use of energy by farms : facto-
ries : indirect contributors was 1 : 3.0 : 1.8, whereas the corresponding relative greenhouse
gas emissions were 1 : 0.2 : 0.3. There was a large difference in the source of the GHG emis-
sions. Factory emissions were all associated with energy consumption. In contrast, on dairy
227
farms the use of energy only generated about 2% of GHG emissions and the main sources
were nitrous oxide and methane from grazing animals. Calculated direct emissions of N to
waterways from farms were about 200 times that from factories. Most of the farm N emis-
sions were estimated to come from N leaching losses to groundwater. This evaluation high-
lighted the benefits of using LCA to determine the relative contributions from different direct
and indirect contributors to total resource use and emissions. Thus, it was a valuable tool for
identifying inefficiencies in the production system.
Dairy farm intensification
The two intensification options involved increased use of N fertiliser or forage. In terms of
land use efficiency, N fertiliser was the most efficient at increasing productivity per unit of
land area, particularly when the whole-system land use was accounted for (Table 2). How-
ever, for N leaching, the farm emissions per m3 milk increased by about 70% for the +200N
system, and there was a similar increase when estimated on a whole-system basis. In contrast,
the forage treatment decreased N leaching per m3 milk by 10% on the dairy farm, but this ef-
ficiency gain was reduced when total land use was considered. Greenhouse gas emissions per
m3 milk were similar for the base farm and +forage system, but increased by about 15% for
the +200N system. The latter was mainly due to increased N2O emissions. This evaluation
highlighted that the choice of intensification method influences the potential for gain in dairy
farm system efficiency.
Benchmarking
The LCA methodologies were useful for determining “hot-spots” in the dairy production
chain, for evaluating the total impacts of land intensification or improved management prac-
tices, and for benchmarking farms or whole industries within and between countries. Data for
GHG emissions per m3 milk for the average Waikato dairy farm were similar to those for the
Swedish dairy farm of Cederberg (1998). While the energy-related CO2 emissions were grea-
ter for the Swedish farm, this was countered by lower methane emissions per unit of milk
from high-producing Swedish cows. Total farm energy use per unit of milk production on the
Swedish farm was over 5-fold higher than that of the Waikato farm on a whole-system basis.
This was mainly due to high fuel use in the Swedish farm system for crop production, feeding
and heating the farm dairy. The NZ farm system with all-year-round grazing of long-term
permanent legume-based pastures is a low energy requiring system, but this advantage may
diminish with intensification. Further research is required to determine whether the energy
advantage of the NZ farm system is sufficient to compensate for “food-miles” or the energy
cost associated with shipping dairy produce from NZ to Europe.
References
Cederberg, C., 1998. Life cycle assessment of milk production – a comparison of conven-
tional and organic farming. SIK report. The Swedish Institute for Food and Biotechnol-
ogy. SIK, Gothenburg, Sweden.
228
Ledgard, S.F., Williams, P.H., Broom, F.D., Thorrold, B.S., Wheeler, D.M. and Willis, V.J.,
1999. OVERSEERTM – a nutrient budgeting model for pastoral farming, wheat, pota-
toes, apples and kiwifruit. In: Currie, L.D., Hedley, M.J., Horne, D.J. & P Loganathan,
P. (eds.). Best soil management practices for production. Massey University, NZ. pp
143-152.
Patterson, M. G. and McDonald, G., 1996. Regional level environmental accounting systems
in New Zealand, using Input Output methodologies. In: Tracking progress: Linking en-
vironmental economy through indicators and accounting systems. Australian Academy
of Science, Fenner Conference, The University of New South Wales, Sydney.
Wells, C., 2001. Total energy indicators of agricultural sustainability: Dairy farming case
study. Report to MAF; Technical Paper 2001/3. 79pp.
Table 1. Some examples of direct resource use and emissions from Waikato dairy farms and
factories, the indirect contributions and total embodied values for the whole Waikato dairy in-
dustry.
Farms Factories Indirect Total
(direct) (direct) contributors
Energy use (TJ) 997 2970 1843 5810
GHG (Gg CO2-equiv.) 3074 616 1030 4720
N emission to water (Gg) 19.4 0.1 7.0 26.5
Table 2. Effect of intensification from 10 (average) to 12 m3 milk/ha, using N fertiliser (+200
kg N/ha/yr) or extra bought in forage (+2 t DM/ha/yr as maize + oats silage), on environ-
mental emissions. Data for the dairy farm only and the whole-system (dairy farm + grazing +
forage land) are compared.
Dairy Farm Whole-system
Av. +200N +Forage Av. +200N +Forage
3
Milk (m /ha/yr) 10.0 12.0 12.0 8.3 10.0 9.4
N leached (kg/ha/yr) 36 74 38 32 64 35
GHG (kg CO2- 8590 11970 10780* 7790 10590 9940*
equiv/ha/yr)
Efficiency indices:
kg N leached/m3 milk 3.6 6.2 3.2 3.9 6.4 3.7
kg CO2-equiv/m3 milk 859 998 898* 939 1059 1058*
* does not account for CO2 emissions from cultivated soil
229
Food purchasing processes and environmental information in the food ser-
vice industry in Sweden
Bergström, K., Solér, C., Adolfsson, A., Shanahan, H.
* Kerstin Bergström, PhD student, Göteborg University, Department of Home Economics, Box 300,
405 30 Göteborg, Sweden, Phone +46 (0)31-773 4204. Fax +46 (0)31-773 4209,
[email protected]
Cecilia Solér, PhD, Gothenburg School of Economics and commercial Law, Department of Business
Administration, Box 600, 405 30 Göteborg, Sweden, [email protected]
Anna Adolfsson, PhD student, Göteborg University, Department of Home Economics, Box 300, 405 30
Göteborg, Sweden, [email protected]
Helena Shanahan, ass. professor, Göteborg University, Department of Home Economics, Box 300,
405 30 Göteborg, Sweden, [email protected]
Abstract
Food consumption affects the environment in several ways. In Sweden one fifth of all food is
consumed outside homes. It is therefore highly justified to study food purchasing processes
and needs for environmental information in the food service industry. Purchasing managers
have been interviewed in food production companies, wholesalers, local and regional public
authorities, restaurant catering and retailing. They actively apply quality assurance in order to
reduce environmental impact of cooling media, transport and packaging. However purchas-
ing managers lack knowledge about environmental impact of other stages in the food chain.
This makes it difficult to value impacts of foods and to make environmentally sound purchase
decisions.
Keywords: environmentally friendly foods, procurement, institutional and restaurant cater-
ing.
Background
Today’s food consumption affects the environment in numerous ways. Throughout the life
cycle of food, which includes agricultural production, storage, transportation, processing,
preparation and waste disposal, resources are used and emissions are released to the environ-
ment. This project has a focus with substantial potential for broadening the understanding of
how purchasing managers could contribute to a more environmentally friendly food industry
as well as suggesting templates for tailoring the information required by users.
The research programme is titled: Designing and evaluating the impacts of environmental in-
formation in food service institutions and the food wholesale sector. The project is supported
by the Foundation for Strategic Environmental Research, MISTRA, 2002-2004. It focuses on
how different factors interact in food purchasing, especially how environmental information
related to foods affects food purchase in institutional catering and the wholesalers sector. The
project is a co-operation between Göteborg University, the Mid Sweden University, Swedish
University of Agricultural Sciences, the Swedish Defence Research Agency and the Univer-
sity of Trollhättan/Uddevalla. In the initial phase of the project, implemented in the autumn
230
2002, purchasing managers in the commercial and public companies were interviewed. In the
phase to follow the environmental impact of selected foods will be calculated using a life-
cycle inventory approach. Based on these calculations different types of environmental infor-
mation labeling will be developed and tested on purchasing managers in the commercial and
public sectors. The effects of different environmental purchase decisions based on labeling
will be studied and illustrated in a number of scenarios describing changes in resource use in
the food service industry. The long-term objective of the project is to contribute to patterns of
food production and consumption with substantially lower resource use and emission levels
than today. This will contribute to societal goals to safeguard ecosystems for sustained gen-
eration of ecological services, which is a mutual prerequisite for economic and societal sus-
tainability.
Purpose of the poster
The purpose of the poster is to report findings from the initial phase which studied practices
and needs in relation to environmental information of producers, suppliers and corporate cus-
tomers within the food service industry.
Method used
Informal interviews were held with purchasing managers in national, regional and local au-
thorities in food service institutions, and with restaurant and retailing managers as well as
managers as suppliers of food. The interviews focused on decision situations as well as organ-
isational specific factors that affect the use of and perceived need for environmental informa-
tion. The interviews were phenomenological. This method focuses on agents personal views
on certain subjects and results in descriptions of the agents experience. Here the description of
experiences focused on the work of using and communicating environmental information.
Results
The results point out that purchasing is a complex information situation. The procurement is
an important and time consuming work for purchasing managers. Communication between
participating companies, departments and suppliers in the study is illustrated in table 1. The
information about inquiries, procurement, agreements, purchasing, deliveries and other food
items flows in many directions.
The phenomenological analysis shows that the respondents represent four different perspec-
tives when using environmental information in the purchasing process. The first perspective
can be called “to regard financial facts”, the second is “to work according to the law”, the
third is “to adjust according to demands” and lastly “to be in control”.
All the respondents said that environmental information about food is needed, when making
inquiries about food for procurement and when negotiating written agreements. The informa-
tion is also needed in training programmes for employees, in production, and when labeling
and marketing. It is also needed when making quality revisions.
231
Conclusion
It can be concluded that the purchasing process is very complex and that information and food
items, at different degree of processing, flow in many directions. Moreover, purchasers have
different perspectives on environmental purchase decisions. In both the commercial and pub-
lic food service industry, work is actively carried out through quality assurance to reduce the
environmental impacts of foods. Areas included especially are cooling media, transport and
packaging. However, purchasing managers lack knowledge about environmental impacts
from other stages in the food chain. This makes it difficult to value the total impacts of foods
on the environment and consequently to make environmentally sound purchase decisions.
Table 1. The table is illustrating the participating companies, departments and suppliers in the
study and the connections between them.
Food Food suppliers/ Public food Commercial Retailing
producers Whole service in local and public food
sales and regional service through
authorities catering chains
Food Food purchasing Food purchasing Food purchasing Food purchasing Food purchasing
purchase process for pro- process for process for pub- process of food process for re-
ducers, also be- wholesalers, be- lic catering in lo- for global cater- tailers
ing suppliers ing suppliers cal government ing chains
agencies and
county councils
Food In-house food Food production Food sales in re-
production production in in catering for tailing stores.
canteens with canteens with Food prepara-
public manage- contract man- tion in private
ment agement households
Food Meal Meal consump- Meal consump-
consump- consumption in tion in canteens tion in private
tion canteens and and dining households
dining rooms rooms
232
eLCA: website and database of IPP tools for SMEs in the agro-food sector
Milà i Canals, L., Fullana, P., Mantoux, F.
* Escola Superior de Comerç Internacional, Pg. Pujades 1, ES-08003, Barcelona (Catalonia), Spain.
+34 93 415 11 46. +34 93 295 47 20. [email protected]
Abstract
The new approach of environmental policy in Europe, the so-called Integrated Product Policy
(IPP), seeks to improve environmental performance of products and services in a life cycle
perspective by integrating different tools of reduction of environmental impacts. Examples of
such tools are LCA (Life Cycle Assessment) and eco-design, eco-labelling, green purchasing
and ecotaxes. The major barriers for its incorporation by Small- and Medium-sized Enter-
prises (SMEs) concern the data availability as well as the knowledge and resources to man-
age this information. The European project eLCA aims to support SMEs in the implementa-
tion of IPP tools by providing through a web portal all the technical and managerial
information (pre-elaborated LCA data, guidelines for the integration of LCA and eco-design,
legislation, description of market tools, etc.), training resources and case studies, specialised
software (LCA, eco-design) and on-line consulting services. Thus, the SMEs of the agro-food
sector will find Good Agricultural Practices (GAP), simplified LCA data for the conduction of
LCA, sources of economic incentives from the CAP reform, etc.
Keywords: IPP, tools, LCA extension, Internet.
Introduction: the eLCA project
Integrated Product Policy (IPP) is a new approach for environmental policy in Europe. It aims
to improve the overall environmental performance of products and services by integrating dif-
ferent environmental tools and stakeholders. European policy in this area includes LCA, and
its application to eco-design, eco-labelling and “green” purchasing, as evidenced in the new
EC communication of June 2003. Small- and Medium-sized Enterprises (SMEs) provide the
majority of economic activity across Europe but are unlikely to implement IPP because of the
high cost of gathering data as well as the knowledge and resources to manage this informa-
tion.
eLCA is a pan-European project funded by the Commission under its eContent programme. It
provides a web portal (ecosmes.net) containing pre-compiled life cycle data linked to a sim-
plified life cycle analysis tool in order to overcome the barriers that prevent SMEs from using
LCA. The web site also provides information on eco-design, legislation, market tools, as well
as training packages and guidelines derived from case studies. The database will be pre-
populated with the results of several product chain studies. Stakeholders will use the portal to
carry out studies on their product, supplying data at the same time as using the information.
This use of Internet will benefit the environment by facilitating rapid access to environmental
information and tools for SMEs allowing them to respond to “green” procurement and other
supply chain pressures to reduce the environmental impacts of their products.
233
The overall objective of the web portal is to develop the market for IPP services, creating a
platform for consulting services on training and LCA-based studies and tools. Partnership
agreements are envisaged for the distribution of consulting services in many sectors. The
partners of the project include public agencies, research centres, Universities, trade associa-
tions and service centres, and consultancies.
Demand for environmental information in the agro-industrial sector
SMEs represent also the majority of industries in the agro-industrial sector, and are generally
run by small farmers who have not the knowledge to incorporate all the environmental issues
at stake in their decision-making. It has been suggested that the farmer’s decisions on tech-
nique implementation may have a greater effect on the environmental performance of the ag-
ricultural system than the choice of technology (Milà i Canals, 2003). Other factors such as
the physical site conditions (soil type, weather, etc. see Cowell and Clift, 1998, and Milà i
Canals, 2003), and the socio-economical context of the farm affect the environmental per-
formance of agricultural systems (see Figure 1).
The socio-economical context in the agro-industrial sector is formed by a wide range of ac-
tors, from the inputs suppliers (fertilisers, pesticides, seeds…) to the final consumers. The IPP
approach actually pretends to integrate all the information and communication tools that work
between these actors, in order to foster a reduction of the environmental impacts of the whole
supply chain, instead of focusing on the production stage. This integration includes legislative
pressure but also communication and information exchange between the actors in the supply
chain, e.g.: retailers asking for environmental performance to their suppliers, consumers de-
manding eco-labelling, etc.
In this context, agro-industrial SMEs are joining efforts with research centres and public bod-
ies in several research projects aiming at facilitating adoption of environmental tools. These
projects have focused on process-based tools such as Environmental Management Systems
(EMS) and recommendations on Good Agricultural Practices. The IPP approach is an oppor-
tunity to expand the research topics to product-focused tools such as LCA.
Wholesale retailers have been active for a long time in response to consumers’ demand of in-
formation, and have set their own programmes for product information (type-I-like eco-
labelling schemes). These programmes, together with the increasing presence of EMS within
the sector, are increasing the demand for information upstream in the product chain. The in-
creased use of LCA will facilitate the provision of this information.
Finally, the consumers’ pressure and demand for environmentally friendly agrarian products
is another driver for the multiplication of environmental product information systems requir-
ing for life cycle data.
234
eLCA and the agro-industrial sector
The eLCA project is presenting the first results for several industrial sectors in September
2003, but the agro-industrial sector is not yet included. So far, the activities of the project in
this sector are focused on the dissemination of LCA. It is essential that this dissemination be
done so the farmers perceive LCA as a way to increase their competitiveness, to facilitate leg-
islation compliance or the communication with the rest of the supply chain, in order to get
positive reactions.
In Spain, where the extension of the eLCA services to the agro-industrial sector is first
planned, the activities will begin with LCA training to relevant stakeholders: research and
training centres, Ministry for agriculture… This training is aiming at fostering the demand for
LCA studies and LCA data. Then, product panels will be organised in strategic sectors, such
as fresh fruit and pork. Product panels serve as a start for supply chain studies and definition
of Good Agricultural Practices (GAP), which in turn provide the basic information for the
pre-population of the website. Apart from simplified LCA data for the conduction of LCA,
the SMEs of the agro-food sector will find training and consultancy on GAP, sources of eco-
nomic incentives from the CAP reform, tools to communicate with the other actors of the
supply chain, etc. Partnership agreements with stakeholders of the agro-industrial supply
chain seek the promotion of the use of the website, and include the integration of already ex-
isting services and contents.
References
Cowell, S.J. and Clift, R. 1998. Site-dependency in LCAs involving agricultural production.
8th Annual Meeting of SETAC-Europe. Abstracts. SETAC. 14-18 April 1998. Bordeaux,
France.
Milà i Canals, L. 2003. Contributions to LCA Methodology for Agricultural Systems, Site-
dependency and soil degradation impact assessment. PhD Thesis. Universitat Autònoma
de Barcelona, Spain.
Environmental interventions
Physical site Producer’s
conditions technique
Technology
type
predicted specific e.g.: time type of
emissions e.g.: soil’s buffer substance devoted to inputs
capacity, and operations, policy, subsidies, retail
permeability, amounts election of a.i. value of outputs, etc.
weather, etc. and fertiliser,
etc. Socio-economical
cost of la
b context (of the farm)
cost of fu our/
site-dependency el, etc.
Figure 1. Aspects influencing the environmental impacts of the farm (Milà i Canals, 2003).
235
Sources of Site-Dependency and Importance of Energy Consumption in
Agricultural LCA: Apple Production in New Zealand.
Milà i Canals, L., Burnip, G.M., Suckling, D.M. and Cowell, S.J.
* Escola Superior de Comerç Internacional, Pg. Pujades 1, ES-08003, Barcelona (Catalonia), Spain.
+34 93 415 11 46. +34 93 295 47 20. [email protected]
The Horticulture and Food Research Institute of New Zealand Ltd, PO Box 51, Lincoln (Canterbury),
New Zealand
The Horticulture and Food Research Institute of New Zealand Ltd, PO Box 51, Lincoln (Canterbury),
New Zealand
Centre for Environmental Strategy, The University of Surrey, Guildford (Surrey), GU2 5XH United
Kingdom
Abstract
Research conducted in New Zealand apple orchards shows that farmers’ technique exerts a
considerable effect on the LCA results, introducing variances of e.g. 30-50% in energy con-
sumption when different farmers perform the same field operation. Energy consumption is
found to be significantly higher in organic farming than in integrated farming in apple produc-
tion in New Zealand, and it contributes above 50% to most impact categories considered in
the study. Therefore, holistic approaches such as LCA covering the different environmental
impacts from agriculture should be promoted when designing certification schemes or as-
sessing the environmental soundness of agricultural technologies.
Keywords: apple production, site-dependency, farmer technique, New Zealand, comparative
LCA
Introduction: Goals and scope of the study
The LCA study aims at detecting the environmental hotspots of two different systems for ap-
ple production, Integrated Fruit Production (IFP) and Organic Fruit Production (OFP), in two
New Zealand regions: Central Otago (CO) and Hawke’s Bay (HB) (Milà i Canals et al.,
2001). This presentation analyses the effect of farmer’s technique on the results using data
from specific sites (Milà i Canals, 2003).
The functional unit has been set to the production of 1 ton of Braeburn apples of export or lo-
cal market quality in New Zealand. The system’s physical boundaries are set in the whole or-
chard, including a tree wind shelter that is usually found in New Zealand apple orchards. On
the vertical axis, soil is considered as part of the system (and thus of technosphere) down to a
depth of 1 m. Substances crossing these boundaries will be considered as emissions to the en-
vironment. In the case of ancillaries, only machinery use has been analysed. Soil quality deg-
radation has not been assessed due to lack of methodology. Farming infrastructure (buildings,
irrigation infrastructures, etc.) and its maintenance has neither been included. As for the time
boundaries, only one year of the orchard’s high yield period has been considered in the study.
Even though soil quality has not been assessed in the apple LCA, the substances emitted to
236
soil remaining in soil after the time boundaries are crossed (at harvest) are also considered as
an emission to soil. Their inclusion is consistent with the need of leaving the soil in the same
conditions as it is found in the beginning of the system (Audsley et al., 1997, p. 85). Finally,
from a life cycle perspective, only the phases from cradle to gate are analysed, as the transpor-
tation of the finished product, consumption, and final waste disposal are not relevant for the
purposes of the study.
Data for agricultural inputs consumption and agricultural practices were obtained directly
from four individual producers (one for each technology and region), who filled a question-
naire for the season 1999-2000. Some further checking had to be done by telephone inter-
views in most cases from June to August 2000.
The impact assessment phase includes the impact categories usually considered in LCA:
global warming, photochemical oxidants formation, acidification, nutrification, human toxic-
ity (air, water and soil), ecological toxicity (acute and chronic for aquatic ecosystems and
chronic for terrestrial ones) (Milà i Canals et al., 2001). Characterisation factors were ob-
tained from Hauschild and Wenzel (1998), and new toxicity factors have been calculated for
the pesticides used in IFP using the method described in Hauschild and Wenzel (1998) (see
Milà i Canals, 2003). Besides, an indicator on the competition aspect of land use has been in-
cluded as a measure of land use efficiency, expressed in ha year. Finally, a simple indicator
expressing the amount of non-renewable energy consumed in each site is included.
The inventory parameters mostly determining the life cycle impact assessment (LCIA) results
(dominance analysis) are checked for their confidence by estimating error margins. This
qualitative uncertainty analysis assures the consistency of LCA results.
System’s description and LCI
The production operations have been deeply analysed as part of the foreground system for the
apple LCA, including detailed calculations for field emissions and energy consumption for
mechanisation. Data for the processes in the “background system” (agro-chemicals produc-
tion, fertilisers production, machinery production and delivery of energy carriers and trans-
portation) was gathered from the literature, with the exception of the substances for the bio-
logical pest control used in OFP, which have been described in this study (Milà i Canals,
2003).
Field processes
Figure 1 is a graphical representation of the operations taking place during apple production,
with the direct inputs to the apple orchard and a rough graphical representation of the timing
of different operations for New Zealand, from May to July (when the trees are pruned) to
April (harvest ends). Field emissions have been considered mainly for understorey manage-
ment (herbicide emissions in IFP), fertiliser use, thinning (only in IFP as well), and pest and
disease management. Energy consumption, on the other hand, has been studied for all field
237
operations. Table 1 depicts the operations in which differences between regions and technolo-
gies exist.
Field emissions
Emissions of ammonia, nitrate, nitrous and nitrogen oxides, methane and heavy metals are es-
timated for the fertilisers used from references in the literature. In the case of synthetic pesti-
cides, a detailed partition analysis is done based on Hauschild (2000) to estimate emissions to
air, surface water, groundwater and soil. Heavy metals emissions are also considered for min-
eral pesticides.
Results
Figure 2 gives an overview of the relative contribution to each impact category by the sites
under study. It must be noted that IFP systems have lower values for land competition than
OFP ones, i.e.: they produce more apples in less surface, even though OFP_HB has a similar
value for land competition than IFP orchards. The land competition factor has an obvious ef-
fect on the LCA results, as systems with lower productivity and pack-out will be charged
more for their impacts. Energy consumption determines the impacts on photochemical oxi-
dants formation and ecological toxicity, and to a lesser extent the human toxicity through air
emissions, acidification and global warming (the two latter also affected by fertiliser emis-
sions in IFP), and this is why a further analysis of energy consumption is given in Figure 3.
In the case of human toxicity through water and soil emissions, only integrated sites show
relevant impacts (see Figure 2). These impact categories are dominated by emissions of syn-
thetic pesticides, which are only used in IFP systems. The huge differences (above one order
of magnitude) between the IFP sites are due to local conditions (mainly the soil type) and to
farmer’s practices (choice of active ingredient, method of application etc.). Finally, ecological
toxicity in aquatic ecosystems is dominated by emissions related to direct energy consumption
and to inherent energy used in the production of inputs (machinery, pesticides and fertilisers),
and no clear differences may be observed between organic and integrated systems.
Sources of energy consumption
Figure 3 shows the systems consuming more energy, and the sources of energy consumption.
What first comes into sight from the figure is that organic systems have higher energy con-
sumption than IFP systems. Of all the producers participating in the study, OFP_HB is the
one with the highest energy consumption, due to the extraordinary use of hydra-ladders for
pruning, thinning, and harvesting. The energy consumption for understorey management in
OFP_CO is also noticeable, and can be explained by the intensive mowing of the orchard;
also the fact that the engine of the tractor is of a higher cc rating than usual partly explains
this higher contribution.
From Figure 3 it is obvious that direct energy consumption by field operations is the main
cause of energy consumption (70-75% in IFP and 83-90% in OFP). Pruning and thinning
238
have a greater share in organic systems than in integrated ones, due to the higher mechanisa-
tion needed to perform these operations manually. The contribution of pesticides production
to energy consumption is noteworthy in integrated systems, where it represents from 11% to
18%. Also energy consumption related to machinery production is relevant, and contributes
7% to 15% to total consumption.
Discussion
The first conclusion arising from these results is that the results are highly dependent on the
characteristics of the site. Firstly, most impacts are directly dominated by producer’s practices
in some way or other: selection of fertiliser or pesticide active ingredients, efficiency in the
use of machinery, etc. For instance, the same field operation (e.g.: mowing, thinning, pruning,
harvesting…) performed by different farmers results in variances of 30-50% in energy con-
sumption. Physical site conditions (particularly soil type) also have a significant effect on the
impact categories, mainly through their effect on field emissions; they act as “filters” reducing
or increasing emissions to the environment from the amounts used by the farmer.
Besides, it can be stated that integrated production presents a wider variety of impact sources
than OFP, and these are both related to energy emissions and field emissions from pesticides
and fertilisers. In the case of organic apple production, energy consumption is a clear focus of
impact generation. This is because inputs used in organic fruit production are in principle less
problematic than those used in IFP. Apart from this overall distribution of impact sources,
ample variations appear in the relative contributions of each producer’s item to the impact
categories.
In summary, site characteristics have been found to affect the LCA results to a bigger extent
than the choice of technology (organic or integrated) in many impact categories. Actually,
only the impact categories that are clearly affected by substances only used in IFP (synthetic
pesticides) show clear differences between IFP and OFP (Human Toxicity through water and
soil). Also those impact categories chiefly dominated by energy consumption (Eco-Toxicity
through soil and Photochemical Oxidants Formation) present clear differences, because in the
New Zealand apple LCA consistently higher energy consumption has been found for organic
orchards.
Energy consumption (mainly related to the intensive mechanisation of field operations) seri-
ously hampers the environmental preference of OFP over IFP in New Zealand, where it is
significantly higher in organic farming than in integrated farming. Above 50% of most impact
categories considered in the study is due to energy-related emissions. The degree of mechani-
sation should thus be considered when designing certification schemes or assessing the envi-
ronmental soundness of agriculture.
239
References
Audsley, E. (coord.), Alber, S., Clift, R., Cowell, S., Crettaz, P., Gaillard, G., Hausheer, J., Jolliet,
O., Kleijn, R., Mortensen, B., Pearce, D., Roger, E., Teulon, H., Weidema, B., Van Zeijts,
H. 1997. Harmonisation of Environmental Life Cycle Assessment for Agriculture. Final
Report. Concerted Action AIR3-CT94-2028. European Commission. DG VI Agriculture.
Hauschild, M. and Wenzel, H. 1998. Environmental Assessment of Products. Volume 2 Scien-
tific background. Chapman & Hall, London, UK.
Hauschild, M.Z. 2000. Estimating pesticide emissions for LCA of agricultural products. B.P.
Weidema and M.J.G. Meeusen (eds.). Agricultural data for life cycle assessments. Volume
2, LCANet Food, pp. 64-79. The Hague, The Netherlands.
Milà i Canals, L., Burnip, G.M., Suckling, D.M., Cowell, S.J. 2001. Comparative LCA of apple
production in New Zealand under two production schemes: Integrated and Organic Fruit
Production. International Conference on LCA in Foods. Proceedings. Gothenburg, Sweden,
27th April 2001.
Milà i Canals, L. 2003. Contributions to LCA Methodology for Agricultural Systems, Site-
dependency and soil degradation impact assessment. PhD Thesis. Universitat Autònoma de
Barcelona, Spain.
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May
agro- Understorey Management
chemicals
emissions
Fertiliser use Fertil.
Field
Emissions (from
fertilisers direct energy
Energy consumption
Thinning consumption +
field emissions)
Pest & Disease Management
machinery
Prun. Prun.
energy
Frost Fighting
Irrigation APPLES
water
Harvesting
SOI L SOIL’
Figure 1. Field operations in the production stage of the apple life cycle in New Zealand
(Milà i Canals, 2003).
240
pet.
Com
Land
n
mptio
gy c onsu
Ener
on
ificati
Acid
tion
orma
xid. F
Phot
oche
m . O
OFP_CO
ronic
il Ch
x. So OFP_HB
Ecoto
ute
t. Ac
Ecot
ox . Wa IFP_CO
ronic
ox. W
a t. Ch IFP_HB
Ecot
oil
an To x S
Hum
r
Wate
an Tox
Hum
ir
ox A
an T
H um
g
armin
al W
Glob
0 20 40 60 80 100
Figure 2. Main results of the New Zealand Apple LCA.
Mechanization
Machinery P.
1400
Fertilisers P.
1200 Pesticides P.
Herbicides P.
1000
800
MJ
600
400
200
0
IFP_HB IFP_CO OFP_HB OFP_CO
Figure 3. Energy consumption by different input items. Results refer to 1 ton of apples of ex-
port or local quality.
241
Table 1. Main differences in field operations as depending from technology and region (Milà
i Canals, 2003).
Differences in technology
Operation type (IFP / OFP) Differences in regions
YES (use of herbicides in IFP;
Understorey Management NO
mulching more usual in OFP)
Fertilising YES (type of fertilisers) NO
YES (approach to pest man-
YES (intensity of pests related
Pest and Disease Management agement and type of sub-
to climate)
stances)
YES (more time-intensive in
Pruning NO
OFP; fate of prunings)
YES (chemical thinning in IFP;
Thinning NO
type of substances)
YES (slightly higher water con- YES (source of water; irrigation
Irrigation
sumption in OFP expected) system)
YES (greater need in CO than
Frost Fighting NO
in HB)
Harvesting NO NO
242
Procedures to improve scope definition and inventory analysis in LCAs of
farming systems
Hayo M. G. van der Werf and Claudine Basset-Mens
INRA, UMR Sol Agronomie Spatialisation, 65, rue de Saint Brieuc CS 84215, 35 042 Rennes Cedex,
France. Email :[email protected]
We identify three methodological weaknesses in Life Cycle Assessments of farming systems.
• Definition of the technology used is often based on data for a single farm. Compared
with other economic activities agriculture involves many production units (farms) with
major variability in production processes.
• Data on economic and environmental outputs often poorly represent the processes
concerned with respect to specific farmer practices and the effect of climate.
• No assessment of the uncertainty of results is made.
We propose procedures to better address these topics.
This study compared three scenarios for pig production: a) Good Agricultural Practice (GAP),
b) quality label “Label Rouge” (LR) and c) “Agriculture Biologique” (AB).
“Technology coverage”, i.e. the determination of the technology used, is part of the scope
definition. The technology used was defined for crop production, feed production, housing
type, access to pasture, storage, treatment and field application of manure. It was preferably
based on the official requirements and the production rules for GAP, AB and LR. Other
sources used were, in order of preference: statistical data, literature references, data from in-
dustry and expert opinions. If these were not sufficient, farms were visited to collect the data.
Generally data sources were of better quality for GAP than for AB, with LR intermediate.
This approach allowed us to determine technology coverage for each scenario with a satisfac-
tory degree of confidence.
Data quality has a major influence on results. In agriculture both economic and environmental
outputs of processes are strongly affected by farmer production practices (e.g. timing of fer-
tilisation) and climate factors. Output data should be as specific and representative as possi-
ble, taking into account farmer practices and climate. Output data were preferably based on
simulation models taking into account technology, practices and climate. If an appropriate
model was not available, we used measured data collected under conditions corresponding
closely to those of our scenarios. If fully representative models or data were not available, we
used less specific models or data, as published in the international literature. As a last resort,
we relied on expert opinion.
In order to assess uncertainty of results we first identified key parameters of economic and
environmental output. For each of these a high and low value was defined in addition to the
243
default reference value. High and low values were chosen to reflect realistic rather than ex-
treme values, so that the uncertainty interval defined by these values would contain 60% to
70% of the variability for the parameter concerned. Values corresponding to improved eco-
nomic output or lower emissions were labelled “favourable”, conversely, values reflecting
worse economic output and higher emissions were labelled “unfavourable”. By combining on
the one hand all “favourable” values for key-parameters in a “favourable” scenario and on the
other hand all “unfavourable” values in an “unfavourable” scenario we obtained two sub-
scenarios (“favourable” and “unfavourable”) for each farming system, reflecting an assess-
ment of overall uncertainty.
244
Predicting the future: comparing pork with Novel Protein Foods on envi-
ronmental sustainability impairment
Helms, M.*
*Institute for Environmental Studies, Vrije Universiteit, De Boelelaan 1087, 1081 HV Amsterdam, The
Netherlands, Tel: +31 (0)20 444 9557, Fax: +31 (0)20 444 9553, Email: [email protected]
Abstract
A method is being developed to assess the environmental sustainability of protein production
chains and facilitate a deliberate consumer choice. Pork is compared to a virtual plant protein
product based on green peas. Application of the method to the latter product requires an ab-
stract approach in which sustainability is assessed on the basis of impaired ecosystem func-
tioning by a representative selection of production processes. The present focus is on pri-
mary production and the regulatory parameters selected are biodiversity, carbon, nitrogen
and water.
Introduction
The present manner in which food is produced and consumed has large impacts on the envi-
ronment. These impacts are expected to increase due to a growing world population and in-
creasing consumption of animal products. As a result, the sustainability of future food produc-
tion systems is questioned (e.g. see (Tilman et al., 2002). One option to change the system is
on the consumer side in the choice of the preferred diet and the selection of foodstuffs.
PROFETAS (PROtein Foods, Environment, Technology And Society) is a research program
concerned with the development of more sustainable food production systems (see
(PROFETAS, 2002). The focus is on protein production and consumption and the prospects
to replace meat in the Western diet with new plant protein products – the so-called Novel Pro-
tein Foods (NPFs). The primary hypothesis is that a substantial shift from animal to plant pro-
tein foods is a) environmentally more sustainable than present trends, b) technologically fea-
sible, and c) socially desirable.
The present study focuses on the first part of the hypothesis and examines the development of
a method to assess the environmental sustainability of protein production chains. This should
facilitate the consumers’ choice between the more sustainable of alternative foodstuffs. A
program-wide case study has been adopted in which the conventional pork production system
is compared to the production of a Novel Protein Food based on dry green peas.
Environmental sustainability assessment
An interesting complication is the virtual nature of the NPF-chain, which precludes the use of
conventional LCA: the product has yet to be developed, hence the details of the production
system are largely unknown. The potential applications of the method to products with un-
available data, plus a general desire to reduce LCA-like data requirements, argue for a less de-
tailed approach. Therefore, a method is proposed in which the “less detailed approach” is
245
sought in reduction of the complicated and elaborate production chains to a small number of
essential processes. These are selected on a macrolevel to make them representative of agri-
cultural based food production in general and are than applied to the pork-NPF example (table
1).
The environmental performance of the main processes is analysed as a measure of impaired
environmental sustainability. The latter is considered an anthropocentric concept in which
human welfare and permanence of human society is a central issue. This means that the value
and importance of the environment is strongly tied to the generation of natural capital that can
be used by humans and the permanence of ecosystem services, which provide natural capital
and essential living conditions. De Groot et al. (2002) provide a long list of ecosystem ser-
vices classified in four groups: regulation functions such as climate regulation, soil formation
and pollination, habitat functions – the provision of refugia and nurseries -, production func-
tions such as primary production and genetic resources, and information functions for spiri-
tual, educational and recreational uses. A long-lasting generation of these ecosystem services
requires that a minimum level of ecosystem health and functioning be retained.
The proces of parameter selection
Crucial to the method of assessment is the selection of parameters describing the impairment
of ecosystem health and functioning. Focus is on the translation of resource use and environ-
mental problems to the mechanisms behind the loss of regulation, habitat and production ser-
vices.
The main environmental problems of, here for example, agricultural activities in the primary
production phase, are listed in table 2. The importance of specific resource uses or environ-
mental impacts are determined by the contribution of the process(es) to total human-driven
impact or human use. In general, human dominance of the global ecosystem is most promi-
nent in land transformation, marine ecosystems, the C-cycle, the hydrological cycle, the N-
cycle, the synthesis of persistent organic chemicals and in changes in biodiversity (Vitousek
et al., 1997). In turn, the latter is mainly determined by changes in land use, CO2-
concentration in the atmosphere, nitrogen deposition and acid rain, climate change and the in-
troduction of exotic species (Sala et al., 2002). Returning to the example of agriculture; it
plays a prominent role in land conversion and domination of the C-cycle, the N-cycle and the
hydrological cycle (Helms and Aiking, 2003).
Currently the translation to sustained ecosystem functioning must be made and it is these cy-
cles (carbon, nitrogen, water) that may play an important role. Alexander et al. (1997) tell us
“Managing and finding solutions to many of the important environmental problems facing
humanity begin with understanding and integrating biogeochemical cycles and the scales at
which they operate”. Tilman (1997) confirms the importance of closing cycles – “Ecosystems
attain a sustainable level of functioning when (…) rates of loss and gain of organic matter and
nutrients are in balance” – and emphasises the role of species richness in this matter.
246
Conclusion
Although presently finer details still lay hidden, the importance of the complex interactions
between biogeochemical cycles, biodiversity and ecosystem functioning as well as the degen-
erating effect of food production is clear. Therefore, the pork and NPF chains should be com-
pared primarily in relation to impaired regulatory function of biodiversity, the C-cycle, the N-
cycle and the hydrological cycle.
Aknowledgement
The author gratefully acknowledges The Netherlands Organization for Scientific Research
(NWO) for generous financial support of the PROFETAS research programme (grant number
455.10.300) and would like to thank Harry Aiking for helpful comments.
Literature Cited
De Groot, R. S., M. A. Wilson, and R. M. J. Boumans. 2002. A typology for the classifica-
tion, description and valuation of ecosystem functions, goods and services. Ecological
Economics 41:393-408.
Helms, M. and H. Aiking. 2003. Food and the environment: towards sustainability indicators
for protein production. Proceedings of ECOSUD 2003, Fourth International Conference
on Ecosystems and Sustainable Development, June 4-6, Siena.
PROFETAS. 2002. "PROFETAS - Protein Foods, Environment, Technology and Society."
Available from http://www.profetas.nl.
Sala, O. E., F. S. Chapin III, J. J. Armesto, E. Berlow, J. Bloomfield, R. Dirzo, E. Huber-
Sanwald, L. F. Huenneke, R. B. Jackson, A. Kinzig, R. Leemans, D. M. Lodge, H. A.
Mooney, M. Oesterheld, N. LeRoy Poff, M. T. Sykes, B. H. Walker, M. Walker, and D.
H. Wall. 2002. Global biodiversity scenarios for the year 2100. Science 287, no. 10
March 2000:1770-1774.
Tilman, D. 1997. Biodiversity and ecosystem functioning. In Nature's services. Societal de-
pendence on natural ecosystems, edited by Daily, G. C. (Washington DC: Island Press).
Tilman, D., K. G. Cassman, P. A. Matson, R. Naylor, and S. Polasky. 2002. Agricultural sus-
tainability and intensive production practices. Nature 418, no. 8 august 2002:671-677.
Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of
Earth's ecosystems. Science 277, no. 25 July 1997:494-499.
247
Table 1. Main processes in pork and NPF-production.
Pork chain Novel Protein Food chain
Primary production Primary production
Crop processing Crop processing
Feed fabrication
Animal production NPF fabrication
Animal processing
Food fabrication Food fabrication
Table 2. Environmental impacts of agriculture.
Agricultural activities Environmental impacts Impaired regulatory ecosys-
tem components
Land conversion & Habitat loss Biodiversity
land use
Energy use Global warming C-cycle
N-fertilization Acidification N-cycle
Ozone destruction
P-fertilization Eutrophication
Tillage Erosion
Irrigation Water depletion Hydrological cycle
Pest control Water pollution
Based on Helms and Aiking (2003)
248
Indicators for Monitoring Environmental Relevant Trends of Food Con-
sumption
Jungbluth, N., Frischknecht, R.
*ESU-services, Kanzleistrasse 4, CH-8610 Uster, T. +41 1 940 61 91 or +41 1 940 6132, F. +41 1
940 61 94, [email protected], [email protected]
1. Abstract
Indicators are necessary to compare the state of sustainable development in different coun-
tries and to compare the relevance of consumption patterns between different countries. In-
dicators are also used for monitoring the success of political decisions in the time. This paper
develops indicators for monitoring the environmental impacts of food consumption patterns.
The research work has been initiated by the OECD. It starts with a review of indicators al-
ready proposed by the OECD. Environmental impacts of food consumption can be analysed
with different environmental assessment methodologies. Results from the review of different
case studies are used to propose a set of indicators that covers all types of environmental
impacts related to food consumption patterns in different countries. The indicators are easy
to calculate based on statistical data. They cover important environmental aspects like agri-
cultural production methods applied, transportation patterns, consumption levels of product
categories, household behaviour and other non-quantifiable key issues.
Keywords: indicator, food consumption, trends, environmental impacts, sustainable devel-
opment
2. Introduction
Indicators to measure the progress towards sustainable household consumption have been
proposed by the Working Group on the State of the Environment (1999). Only general food
consumption trends were included in the indicator set (consumption in kg/capita/year and per-
centage change 1970-1995). In an actual study (Jungbluth & Frischknecht 2000a, b) addi-
tional indicators are suggested that could be used to improve the monitoring of environmental
impacts from household food choice, preparation and disposal. Furthermore related data
sources are identified.
3. Critical review of indicators proposed by OECD
This chapter discusses indicators proposed by OECD for household food consumption. Table
1 shows these indicators (Working Group on the State of the Environment 1999). The two
proposals, type of food and growing method, cover some of the important problems.
The type of food is especially important with respect to the level of meat consumption and the
comparison of the total consumed food products. The consumption in kg per capita for differ-
ent product groups can be linked with environmental indicators, e.g. MIPS (Loske &
Bleischwitz 1996), EF (Wackernagel et al. 2000), energy use and CO2 (Kramer & Moll
1995), or information from an LCA valued with an impact assessment method (Jungbluth
249
2000) to compare the environmental impacts over a period of time or for different countries.
The examples show the differences in the relevance of different product groups.
Table 1. Indicators for household food consumption proposed by OECD.
• Reflects a) consumer choices related to food
Proposed indicators Comments
Food consumption intensities and pat-
terns: categories and to growing methods, b) shifts in
1. by type of food (fish, meat, etc.) in kg
• To
demand towards organically grown agricultural
per capita, as % of total products
be read in connection with information on the
2. by growing method and level of proc- environmental effects of the various growing and
ess (share of processed food, share production methods (including effects on e.g. fish
of organically grown produce over to-
tal agricultural produce consumed) as • Needs further investigation on policy relevance, and
stocks)
% of total on actual implications for the environment
The indicator for the growing method describes developments in agriculture and food proc-
essing that influence the level of environmental impacts caused by fertilizers, pesticides and
soil degradation. This indicator needs some clarification on the production methods distin-
guished (e.g. what is processed food and how should different types of processed food be dis-
tinguished). The calculation of environmental impacts would be possible if these indicators
would be multiplied with results of one of the methods described. Thus, it is for example pos-
sible to estimate the average energy use for different types of food and to combine this with
indicator 1. But, it would probably need some research work to estimate the indicator values
per kg.
4. Proposal of indicators for monitoring environmentally relevant trends of food con-
sumption
A proposal for household consumption indicators with special focus on food consumption
patterns in different countries is now elaborated. These indicators should help to quantify and
compare the environmental impacts due to food consumption patterns in different countries
over different years. The indicators are discussed in theme blocks for e.g. production meth-
ods, transports, etc. Main focus while determining the indicators is laid on (see also Working
Group on the State of the Environment 1999:8):
• Policy relevance for monitoring and priority setting.
• Analytical soundness of the indicator with regard to the environmental impacts.
• Measurability and proposal for appropriate methods to quantify the environmental im-
pacts (roughly ordered according to their appropriateness and the data availability).
4.1 Production methods
1. Share and per capita availability of products from e.g. organic, integrated, conven-
tional and greenhouse production
This indicator aims to measure shifts in agricultural production practice. It should cover in-
digenous products as well as imports. The higher the share of products from intensive produc-
250
tion like greenhouses the higher is the environmental impact. The indicator serves to control
the success of political measures like labeling schemes or environmental subventions for the
agricultural sector. Agricultural and foreign trade statistics are the basis for this indicator.
2. Share and per capita consumption of food products with different degrees of process-
ing (fresh, chilled, conserved, deep-frozen, pre-prepared, ready made, self-service and
restaurant)
This indicator aims to measure the shifts in consumption patterns towards more processed
food products. A shift from fresh to conserved and pre-prepared products leads to a rise in en-
ergy use and environmental impacts. Statistics for per capita food availability and sales of
food retailers and restaurants should serve as a basis for the calculation of this indicator.
3. Total energy use per capita and the share of different economic sectors (chemical in-
dustry, agriculture, food industry, retailers, restaurant, freight carriers, households)
for meeting the food demand
In countries with a good database for input-output statistics this indicator might monitor more
directly the impact due to developments in different stages of processing. And it helps to iden-
tify the overall importance of different stages in the life cycle of food products.
4. Percentage of actors in the food chain that have implemented an environmental audit-
ing or management scheme for their company
An environmental audit helps to identify possibilities for the reduction of environmental im-
pacts. This indicator might also be considered in an indicator set for different industries, as it
is not directly related to food consumption patterns.
5. Food products produced with genetically modified organisms
The use of genetically modified organisms in agriculture is a theme of critical debate. Today
it is difficult to assess the environmental impacts due to an increased use of these organisms
and it is unclear how to weight negative or positive aspects.
4.2 Transportation
6. Per capita average distance and mode of transportation for domestic food transports
This indicator should measure the domestic transports of food products. A shift from train to
road and air based transports and a rise in total transports for food products per capita indicate
a rise in environmental impacts. The success of policy measures like energy taxes or reduc-
tion of subsidies (e.g. for road transports) can be controlled directly with this indicator. Aver-
age transport distances can be investigated with national transport statistics using WASD
(Carlsson 1997).
7. Per capita average transport distance and transport modes of imported food products
This indicator should measure the impacts due to transports of imported food products. A
shift from train to road and air based transports and a rise in total transports for food products
per capita indicate a rise in environmental impacts. The indicator helps to identify environ-
mental impacts due to globalization and diversification of consumption patterns.
251
4.3 Purchasing
8. Share of different Eco-labels for food products sold in a country
Eco-labels help consumers to buy less environmentally harmful products. Information about
the sales share of different labels helps to measure the acceptance of these labels. Only labels
with widely accepted guidelines that show a considerable improvement in comparison to con-
ventional production.
9. Types of food distribution (direct on farm, market, small shop, supermarket, fast-food,
restaurants, etc.)
The share and frequency of visiting different types of food dealers indicates environmental
impacts due to e.g. home transports, land use and construction of buildings. Large supermar-
kets as well as farms that sell their products directly are often accessible only by private cars.
A high share of supermarkets indicates a dissipation of areas for living, shopping and working
in big cities. This leads to increased environmental impacts due to transports and land use.
4.4 Consumption level
10. Per capita food availability (kg or MJ nutrition value per head) and share of different
product categories (meat, vegetables, grains, fats, beverages, etc.) in food consumption
The availability (production + imports - exports) of food differs from country to country. If
the availability is higher than the actual demand this leads to food wastes, to over consump-
tion or to long time storage with negative health effects and/or unnecessary environmental
impacts. This indicator serves to compare the level of food availability and consumption in
different countries. As environmental impacts vary among different product groups this indi-
cator can also analyze environmental impacts due to changing consumption patterns, e.g. a
rising share of meat products. Statistical data might be available from foreign trade, agricul-
tural and consumption statistics.
11. Food availability against food consumption as an indicator for wastage or per capita
food waste from waste statistics
It is possible to assess the share of non-consumed products directly if data for food availabil-
ity (from agricultural and foreign trade statistics) as well as for food consumption (nutrition
studies) are available. The amount of food waste might also be investigated from national
waste statistics. The edible parts should be differentiated according to the way of treatment
(incineration, deposition, composting, etc.).
4.5 Household behaviour
12. Per capita packaging wastes, recycling quotas and means of waste treatment for dif-
ferent materials like glass, paper, metals or plastics
The amount and type of packaging wastes from food products and the way to take care for
them is one indicator to assess the environmental impacts due to packaging. A rise in packag-
ing wastes per capita indicates higher environmental impacts while an increased share of re-
cycled wastes indicates an environmental improvement. Data can be gained from national
waste and recycling statistics.
252
13. Mobility for home transport
The share of different transport modes (foot, bike, public transport, car), while buying (food)
products, indicates the environmental impact due to home transports. A rising share of private
cars indicates rising impacts due to fuel use, noise and land occupation for streets, etc.. Data
can be found in transport statistics for the share and distances of different transport modes
while purchasing goods.
14. Distribution and energy use of household appliances for food storage and preparation
The number and size of household appliances for food storage and preparation (stoves,
freezer, deep-freezer, and small appliances for food preparation) indicates the direct energy
use in the household. Data for kitchen equipment can be found in household statistics. The
average energy use due to the use of white goods can be assessed.
4.6 Discussion
As the indicators aim to describe the environmental impacts of food consumption over the
whole life cycle it is important to reflect not only indigenous food production but to account
also for im- and exports.
Some of the indicators proposed are also linked to other fields of household consumption or
they might interfere with indicator sets for sustainable development in other economic sec-
tors. The indicator for eco-labelling (0.) might for example also be dealt with in indicator sets
for agricultural production practice. Or the share of different modes of transportation (0.) to
bring food products to the household might also be covered by an indicator describing the
general impacts of private mobility. Prior to a final decision about the indicators it should be
crosschecked with proposals for other indicators of sustainable consumption patterns if there
are any unnecessary double counting.
5. Conclusions and Outlook
In order to monitor the success of policy measures over a time period and to compare the en-
vironmental impacts related to household food consumption in different countries it is neces-
sary to use certain indicators. Different indicators are proposed. In our view out of this the fol-
lowing six indicators are preferred with regard to the data available and the importance of the
measured impact:
1. Share and per capita availability of products from e.g. organic, integrated, conven-
tional and greenhouse production
2. Share and per capita consumption of food products with different degrees of process-
ing (fresh, chilled, conserved, deep-frozen, pre-prepared, ready made, self-service and
restaurant)
6.+7.Per capita average distance and mode of transportation for domestic and imported food
transports
9. Types of food distribution (direct on farm, market, small shop, supermarket, fast-food,
restaurants, etc.)
253
10. Per capita food availability (kg or MJ nutrition value per head) and share of different
product categories (meat, vegetables, grains, fats, beverages, etc.) in food consumption
14. Distribution and energy use of household appliances for food storage and preparation
These indicators cover important environmental impacts due to household food consumption
and the calculation is based on available statistical databases. The indicators should be cross-
checked with indicators from other fields of household consumption or for industries to en-
sure that there is no double counting of the same effects.
Further research work is necessary after an agreement on the indicators and methods to be
used by OECD decision-makers in order to generate generic quantitative factors that can be
used to calculate the environmental impacts based on statistical data for different countries.
References
CARLSSON, A. 1997: “Weighted Average Source Points and Distances for Consumption Origin -
Tools for Environmental Impact Assessment.” In Ecological Economics (accepted 1/97).
JUNGBLUTH, N. 2000: Umweltfolgen des Nahrungsmittelkonsums: Beurteilung von Produktmerk-
malen auf Grundlage einer modularen Ökobilanz. Dissertation Nr. 13499, Eidgenössische
Technische Hochschule Zürich, Umweltnatur- und Umweltsozialwissenschaften, dissertati-
on.de, 317 Seiten,www.jungbluth.de.vu, Berlin, D.
JUNGBLUTH, N. & FRISCHKNECHT, R. 2000a: Environmental Impacts from Household Food Con-
sumption: Methodologies for Evaluation and Indicators for Monitoring. ESU-services for
OECD Environment Directorate, Project on Sustainable Consumption, www.esu-
services.ch, Uster.
JUNGBLUTH, N. & FRISCHKNECHT, R. 2000b: Household Food Consumption Patterns: Part B -
Evaluation methodologies and indicators. Sustainable consumption: Sector case study series
No. ENV/EPOC/GEEI(2000)9/ANN1, ESU-services for Organisation for Economic Co-
operation and Development (OECD), Environment Directorate, Project on Sustainable Con-
sumption, 49 Seiten, Paris.
KRAMER, K. J. & MOLL, H. C. 1995: Energie voedt: nadere analyses van het indirecte energie-
verbruik van voeding. Final report to the NRP global Air Pollution and Global Change,
IVEM research report No. 77, Center for Energy and Environmental Studies of the Univer-
sity of Groningen (IVEM RUG), The Netherlands.
LOSKE, R. & BLEISCHWITZ, R. 1996: Zukunftsfähiges Deutschland. Ein Beitrag zu einer global
nachhaltigen Entwicklung. BUND UND MISEREOR, Wuppertal Institut für Klima, Umwelt,
Energie GmbH, ISBN/ISSN 3-7643-5278-7, 453 Seiten, Birkhäuser Verlag, Basel, Boston,
Berlin.
WACKERNAGEL, M., DHOLAKIA, R., DEUMLING, D., RICHARDSON, D. 2000: Assess your House-
hold's Ecological Footprint 2.0. www.rprogress.org, San Francisco and Centre for Sustain-
ability Studies, Xalapa.
WORKING GROUP ON THE STATE OF THE ENVIRONMENT 1999: Towards more sustainable house-
hold consumption patterns: Indicators to measure progress. Environment Directorate, Envi-
ronment Policy Committee, Organisation for Economic Co-operation and Development
(OECD), www.oecd.org/env/consumption/scp26.htm, Paris.
254
The History Of Bread Production: Using LCA In The Past
Mirja Van Holderbeke1, Nieves Sanjuán2, Theo Geerken1, Daniëlle De Vooght3
1
Institute for Technological Research (Vito), Integral Environmental Studies, Boeretang 200, B-2400
Mol, Belgium, [email protected], 2Universidad Politécnica de Valencia, 3Free University
of Brussels
Abstract
To get a clear understanding of the process of (un)sustainable development during the past
centuries, the environmental, social and economic developments of four basic needs (drink-
ing water, bread, travelling over land and heated accommodation) are investigated both on
the production- as well as the consumption side, starting from the pre-industrial period (1800)
until the year 2000. For the quantitative environmental assessment, the LCA methodology is
used. The project is sponsored by the Belgian Federal Public Planning Service Science Pol-
icy. This paper shows the first LCA results and thoughts ensuing from the bread case study.
Keywords: bread, history, LCA, sustainability.
LCA goal and functional unit
The goal of the LCA is to compare the environmental effects of producing and distributing
bread in the period 1800 till 2000. The functional unit is “1 kg representative bread”.
Life cycle description of bread for the key years 1800, 1900 and 2000
The specifications given below attempt to represent an average of the way in which the basic
need bread was fulfilled for all people at that time.
Description 1800
Rye bread was usually consumed in Belgium. The grain was locally produced and only hu-
man or animal energy were used in agriculture. Manure and other organic wastes were applied
in low rates to improve the soils. Plant protection was based on cultural measures, e.g. three-
course rotation. Transportation to the mill, bakery and consumer was done by push-carts and
carts drawn by dogs or horses. Mills operated on natural energies (wind and water). People
baked at home in ovens using brushwood. Summarising, only data about the emissions for
baking have to be taken into account. Data about the rye yield/ha are extrapolated from fig-
ures for other years. The consumption of brushwood and subsequent emissions were deter-
mined by emission measurements in a 19th century baker’s oven at open air Museum Bokrijk,
Belgium. The assimilation of CO2 in wood and grain was not taken into account. The amount
of flour/kg bread was founded on Eiselen (1995).
Description 1900
White bread made of wheat was habitually eaten in Belgium. About 70% of the wheat was
imported from the USA by steam trains and steamships, the remaining 30% was produced in
Belgium. The use of chemical fertilizers and pesticides became only important after World
255
War II. Until then, producers depended on the native fertility of the soil to provide the nutri-
ents needed for growth of wheat and on cultural practices and resistant varieties to control
weeds, diseases and insects (Paulsen, 2002). Manure was seldom applied (Kansas State Board
of Agriculture, 1903) and relating to machinery, no fuel was consumed because all equipment
was propelled by farmers or animals. Regarding the energy used in the mills (Bauters, 1998
and own calculations), 41% of the flour was produced by water and wind energy and 59% by
steam engines using coal. Transportation of flour to bakeries was done by horse carts. At the
beginning of the 20th century, dough was prepared manually and the ovens operated mainly
on coal or wood, only a few ovens used gas. The distribution of bread to the consumers was
done by carrier-tricycle. Resulting, only data concerning the fuel consumption and subsequent
emissions to transport wheat, milling in steam mills and baking have to be taken into account.
The yields of wheat crops in Belgium and the USA were founded on NIS (1962) and USDA
(2002). The flour yield resulting from and the power needed during the milling process are
based on figures from Ammann (1914). The coal consumption in the steam mills was founded
on Vierendeel (1921). Emissions (SO2, CO and CO2) due to coal combustion were obtained
from measurements in Bokrijk. We assume an equal division between coal and wood ovens.
The fuel consumption was founded on Ammann (1914). The coal consumption to cross the
Atlantic by steamship per kg cargo was calculated using data from SEWSS (2003). Regarding
the steam trains, coal consumption was estimated using data from Sinclair (1898). Emissions
(SO2, CO and CO2) due to coal combustion were based on measurements in Bokrijk.
Description 2000
Nowadays, wheat bread is still the most consumed bread in Belgium. The wheat is mainly lo-
cally produced or from European origin. Agricultural processes are mechanised and chemical
fertilisers and plant protection products are used to enlarge the yields. Transportation of wheat
to Belgian mills is generally done by trains, trucks and ships. Mills operate on electricity and
flour is distributed by trucks. Baking processes usually use electricity and heating gas. Trans-
portation of bread to households is done by car, bicycle or by foot. For the LCI, all life cycle
steps, from agricultural production until distribution of bread to the consumer, have to be con-
sidered.
The wheat used originated in Belgium (30 %), France (43.4%), Germany (19,6%) and The
Netherlands (7%) (CLEA—CEA, 2000). The yields of wheat crops in these countries have
been taken from FAOstat (FAO, 2003). Data about fuel consumption involved with growing
and harvesting of wheat were obtained from Nielsen and Luoma (1999). Sowing rates were
taken from Moerschener and Gerowit (1999) and L'Institut du Genech (2002). The amount
and type of fertilizers applied to wheat in France, Germany and the Netherlands are based on
Ekboir (2002), ITCF (2002) and Moerschner and Gerowitt (2000). Concerning the fertilizers
production, data are taken from the SimaPro data base or based on Davis and Haglund (1999).
Nutrient balances were founded on L’Institut de Genech (2002), Lopez Bellido (1991), Aud-
sley et al. (1997), Bentrup et al. (2000), Castillon (2003) and Hansen (2000). The use of plant
protection products was taken from Eurostat (2002). Regarding the energy requirements for
256
pesticide production, different sources have been consulted (Audsley et al., 1997; Weidema et
al., 1995; Ceuterick and Spirinckx, 1997; Van den Broek et al., 2002).
Data about the milling yield, energy and water consumption and prices of flour and by-
products were obtained from CERES, the most important mill in Belgium. An allocation
based on the weight was made between flour and by-products. Data about water and energy
use for baking have been obtained from PRESTI (1996). The amount of flour/kg bread was
founded on Andersson (1999). Transportation of wheat to the mill was founded on own as-
sumptions, statistical data and information from CERES. Trucks transport the flour to baker-
ies (CERES), the average distance taken into account is based on own assumptions. Distribu-
tion of the bread to the consumer was established on own assumptions and on statistical data
(MOBEL, 2000).
First LCI/LCA results for bread in the years 1800, 1900 and 2000
Figure 1 shows the relative contributions of the life cycle phases in 2000. The agriculture sub-
system is a hot spot for most of the impact categories studied. The potential contributions to
global warming are associated with the use of fossil fuels for baking and the fertiliser produc-
tion for the agricultural phase.
Figure shows the relative environmental profiles for the years 1800 and 2000. Surprisingly,
the impact categories global warming and photochemical oxidation are much higher for the
year 1800, due to the emissions resulting from the brushwood combustion. The baking proc-
ess nowadays uses four times less energy. Acidification and eutrophication are much higher in
2000, both mainly caused by the production and use of fertilisers.
For the year 1900 not all necessary emission data were available. Therefore we can only com-
pare the emissions of CO, CO2 and SO2 for the years 1800, 1900 and 2000 (Figure 3). We can
see a large efficiency improvement in the last century. The highest level of these emissions
was observed for 1900, followed by the ones for 1800. This is due to the combustion of coal
during the transport of grain by steamships and steam trains from the USA in 1900, and to a
lesser extent to the combustion processes in the milling and baking phases. At the beginning
of the 19th century, all emissions during the bread life cycle were caused by the combustion of
brushwood during the baking process. In 2000, CO and CO2 emissions are attributed to the
energy use for fertilizers production, the baking process and distribution to the consumer.
However, a significant data uncertainty on this distribution phase exists. SO2 emissions in
2000 are mainly caused by the energy used in the baking process.
Using LCA in the past: potential, problems and additional assessments
The first LCI and LCA results show already a surprising picture of the environmental impacts
accompanying the bread production during the last two centuries. Combining this knowledge
with social and economic data will give surely new insights to the sustainability development
in the 19th and 20th century. In the economic assessment we will express the number of man
257
hours needed to produce one bread yourself in the past related to the number of man-hours we
work today to buy a bread.
However, some practical problems arose during our project sofar. One of the main problems
found when applying LCA to the past is the lack and uncertainty of historical data, such as
emissions from steam machines or data as manure composition necessary to perform nutrient
balances. The emissions from steam machines can probably still be measured.
Another important aspect to take into account when assessing processes in the past are the dif-
ferent kinds of energy sources. In 1800 and 1900 mainly all the energy was of natural, human
or animal character. Typically animal and human labor is not included in an LCA, but in this
case they are the motor of many stages of the life cycle. The inclusion of an energy indicator
(summing fossil, human and animal energy) could be evaluated in order to get an idea of the
general energy efficiency improvement.
Another problem is the assessment of land use in the past. We probably will include a simple
land use indicator defined as the area of agricultural land needed to produce the amount of
grain used to make 1 kg of bread. Its value in1800 is lower (8.1 m2/ha) than in 1900 (10.5
m2/ha). This is caused by the more efficient rye milling process in 1800, accounting for a
flour yield of nearly 100%. Nevertheless the value obtained for 2000 is the lowest (1.3 m2/ha),
due to the higher yields of the crops and milling process, and also to the protein content of
flour that allows to decrease the ratio kg of flour per kg bread.
References
Ammann, L. 1914. Meunerie et boulangerie. Librairie J.-B. Baillière et Fils. Paris France.
Andersson, K. 1999. Life Cycle Assessment (LCA) of Bread produced on different Scales,
Case study, AFR-Report 214 CIT Ekologik, Stiftelsen Chalmers Industriteknik, Goth-
enburg.
Audsley, E. 1997. Harmonisation of environmental life cycle assessment for agriculture. Final
report concerted action, AIR3-CT94-2028. Silsoe Research Institute, Silsoe, UK.
Bauters, P. 1998. Van zadelsteen tot zetelkruier. Tweeduizend jaar molens in Vlaanderen, 1.
Geschiedenis van het malen met natuurlijke drijfkracht, Gent.
Bentrup, F.; Küsters, J.; Lammel, J.. Kuhlmann, H. 2000. Methods to estimate on-field emis-
sions from crop production as an input to LCA studies in the agricultural sector. Interna-
tional Journal of Life Cycle Assessment, 5(6), 349-357.
Castillon, P. 2003. Personal communication. ARVALIS-Institut du végétal. Baziège, France.
Ceuterick, D. and Spirinckx, C. 1997. Comparative LCA of biodiesel and fossil diesel fuel.
VITO report 1997/PPE/R/026.
CLE-CEA. 2000. Landbouwstatistisch jaarboek.
Davis, J. and Haglund, C. 1999. Life Cycle Inventory (LCI) of Fertiliser Production. Fertilis-
ers products in Sweden and Western Europe. SIK-report N° 654. Göteborg, Sweden.
Eiselen Hermann (edt.). 1995. Brotkultur. DuMont Verlag, Köln.
258
Ekboir, J. (ed.). 2002. CIMMYT 2000-2001 World Wheat Overview and outlook: Developing
No till Packages for Small-Scale Farmers. Mexico, DF, CMMYT.
Eurostat. 2002. The use of plant protection products in the European Union. Eurostat, Euro-
pean Comission, Luxemburg.
FAO. 2003. FAOstat, FAO statistical data bases. http://apps.fao.org/cgi-bin/nph-
db.pl?subset=agriculture
Hansen, J. 2000. Nitrogen balances in agriculture.
http://europa.eu.int/comm/eurostat/datashop/print-
catalogue/EN?catalogue=Eurostat&product=KS-NQ-00-016-__-I-EN
Institut de Genech. 2002.
http://home.nordnet.fr/~jlnieuviarts/menu/fiches_techniques/productions_vegetales/blete
ndreraisonne/bletendreraisonne.htm
ITCF. 2002. Blé tendre. Marchés, debouches, techniques culturales, récolte et conservation.
Institute Technique des Cereals et des Fourrages. France.
Kansas State Board of Agriculture. 1903. Thirteenth Biennial Report of the Kansas State
Board of Agriculture to the Legislature of the State for the years 1901 and 1902. Part IV.
Reports of Kansas wheat –growing experience and practice, by counties. Pp. 573-632.
Mobel http://www.mobel.be/
Moerschner, J.; Gerowitt, B. 2000. Direct and indirect energy use in arable farming - an ex-
ample on winter wheat in Northern Germany. In: Agricultural data for Life Cycle As-
sessments. Vol 1. Ed: B. P. Weidema and M.J.G. Meeusen. Agricultural Economics Re-
search Institute, The Hague.
NIS. 1962. Nationaal Instituut voor Statistiek , Landbouwstatistieken, 1900-1961, Brussel.
Nielsen, V.; Luoma, T. 2000. Energy consumption: overview of data foundation and extract
of results. In: Agricultural data for Life Cycle Assessments. Vol 1. Ed: B. P. Weidema
and M.J.G. Meeusen. Agricultural Economics Research Institute, The Hague.
Paulsen, G.M. 2002. International contributions to improvement and marketing of Kansas
wheat. http://www.kswheat.com/education/facts/history.htm
PRESTI. 1996. Sectoriële studie over de brood- en banketbakkers, ijsberijders en
chocoladebewerkers. PRESTI Programma OVAM.
SEWSS, Sucursal Española de World Ship Society, personal communication.
Sinclair, A. 1890. The LOCOMOTIVE ENGINE. Running and Management, Seventeenth
Edition -. http://www.railroadextra.com/chapt24.Html
USDA. 2002. Historical track records. United States Department of Agriculture. National
Agriculture Statistics Service. http://www.usda.gov/nass/pubs/histdata.htm
van den Broek, R.; Treffers, D.J.; Meeusen, M.; van Wijk, A.; Nieuwlaar, E.; Turkenburg, W.
2002. Green energy or organic food? A life cycle assessment comparing two uses of set-
aside land. Journal of Industrial Ecology 5(3), 65-87.
Vierendeel, A. 1921. Esquisse d'une histoire de la technique. Vol 1. Vromant & Co, Im-
primeurs-Editeurs. Brussels, Belgium.
259
Weidema, B.P.; Pederson, R.L. and Drtivsholm, 1995. Life cycle screening of food products.
Two examples and some methodological proposals. Danish Academy of Technical Sci-
ences.
agriculture milling baking distribution to consumer
0,62 kg 1,6 e-04 kg 2,9 e-03 kg 2,6 e-03 kg
3-
CO2 eq. C2H2 eq. SO2 eq. PO4 eq.
100%
75%
50%
25%
0%
global warming photochemical acidification eutrophication
(GWP100) oxidation
Figure 1. The environmental profile of Belgian bread in the year 2000.
bread 1800 bread 2000 CO2 SO2 CO
4,5 80
Relative contribution (%)
100%
4 70
80%
3,5
60
3
kg CO2, g SO2
60% 50
g CO
2,5
40
40%
2
30
1,5
20%
20
1
0% 0,5 10
global warming photochemical acidification eutrophication
(GWP100) oxidation
0 0
1800 1900 2000
Figure 2. Comparison of environmental pro-
Figure 3. Comparison of CO2, SO2 and CO
files of 1 kg bread in years 1800 and 2000.
emissions for the years 1800, 1900 and 2000.
260
Methodological Contributions to tailor Life Cycle Assessment to the Specif-
ics of Arable Crop Production
Brentrup, F.
Hydro Agri, Centre for Plant Nutrition Hanninghof, Hanninghof 35, D-48249 Duelmen, Germany,
Tel.: +49-2594-798 137, Fax.: +49-2594-7455, e-mail: [email protected]
Abstract
This paper summarizes the main results of a PhD study entitled “Life Cycle Assessment to
evaluate the environmental impact of arable crop production” (Brentrup, 2003a). The study
proposes new methodological contributions for 1) the estimation of diffuse, in-field nitrogen
emission as an input to the Life Cycle Inventory, 2) the impact assessment of the consump-
tion of abiotic resources like fossil fuels or minerals, 3) the impact assessment of land use 4)
and the aggregation of the different environmental impacts into summarizing environmental
indices via normalization and weighting. The single modules were integrated into a compre-
hensive LCA approach, which was then tested in a case study. In this case study the envi-
ronmental impact of different nitrogen fertilizing intensities in wheat production was analysed.
Keywords: nitrogen emissions, resource consumption, land use, weighting, wheat production,
fertilizer.
Introduction
The Life Cycle Assessment (LCA) methodology provides the theoretical framework for the
evaluation of the environmental impact of products or production systems, and thus, can also
be applied on arable crop production. This is particularly important since agriculture is ex-
pected to comply with the principles of sustainability, which include the three core elements
economy, society and environment. In order to evaluate the sustainability of different agricul-
tural production systems, it is necessary to have appropriate indicators for all of these ele-
ments in place. Principally, LCA is able to provide those indicators for the environmental as-
pect.
However, since most of the available ready-to-use LCA models were not specifically de-
signed for agricultural applications they show some difficulties when for instance applied to
arable production. Therefore, this study presents contributions to the LCA methodology in or-
der to tailor LCA more to the specifics of arable crop production. These contributions concern
the inventory as well as the impact assessment phase and are briefly described in the follow-
ing (see Fig. 1).
Results
Methods to estimate diffuse, in-field nitrogen emissions as an input to LCA studies
Based on a literature study, structured methods for the estimation of diffuse, in-field nitrogen
emissions (ammonia, nitrous oxide, nitrate) were selected. For LCA studies including arable
farming it is particularly important to derive reasonable estimates of these highly variable
261
emissions. The selected methods consider important soil, climate and management parameters
(Brentrup et al., 2000).
Conceptual considerations on the impact assessment of abiotic resource consumption
In contrast to other approaches, this study suggests to treat the consumption of resources,
which are not substitutable by each other, as separate environmental problems. A final aggre-
gation of non-equivalent resources like phosphate rock and fossil fuels into a summarizing re-
source depletion indicator is found to be only possible after an explicit weighting procedure,
which has been developed in this study (Brentrup et al., 2002a).
Impact assessment of land use based on the ‘Hemeroby’ concept
Arable farming uses huge quantities of land for crop production. An assessment of the envi-
ronmental impacts of land use in LCA has to include two dimensions: (a) the size of an area
used for a certain period of time and (b) the potential of a specific land use type to degrade the
naturalness of the area under use. Whereas the first aspect can be directly expressed as a
physical quantity, the latter aspect needs an appropriate indicator. The Hemeroby concept
provides such an indicator, since this concept was specifically developed in order to evaluate
the level of naturalness of land area. Hemeroby is a measure for the human influence on eco-
systems, which defines the level of naturalness of different land use types (e.g. urban area or
extensive pasture) according to their deviation from a natural reference situation. This study
employs the Hemeroby concept in order to assess the impacts of different land use types
within LCA (Brentrup et al., 2002b).
Aggregation into environmental indices
An evaluation of the different environmental effects that are relevant to arable production re-
garding their potential to harm the environment is performed in order to enable an aggregation
of the separate indicators per effect into two summarizing environmental indicators: (a) for
abiotic resources and (b) for impacts on ecosystems and human health. This weighting was
realized by a comparison of the current status of each effect with defined target values for the
respective effects (“distance-to-target principle”). This study suggests internationally agreed
environmental targets to be employed in this procedure because they represent a consensus of
science, economy and society (Brentrup et al., 2003b).
Case study on wheat production at different fertilizing intensities
After these methodological developments the method was tested in a case study. In this case
study the environmental impact of different N fertilizer rates in winter wheat production was
analyzed (Brentrup et al., 2003c). The case study revealed that the aggregated environmental
impact per ton of wheat grain increases clearly at N rates exceeding the crop demand and at
zero N fertilization (see Fig. 2). In the first case aquatic eutrophication was the major prob-
lem, whereas in the latter case this is land use. From reduced to economic optimum N rates
the environmental indicator values increased rather slightly. At economic optimum N fertili-
zation (192 kg N/ha) aquatic eutrophication contributed most to the aggregated indicator; ter-
262
restrial eutrophication, acidification, climate change and land use show similar contributions
to the aggregated value.
References
Brentrup, F., Küsters, J., Lammel, J., Kuhlmann, H., 2000. Methods to estimate on-field ni-
trogen emissions from crop production as input to LCA studies in the agricultural sector.
Int. J. LCA 5, 349-357.
Brentrup, F., Küsters, J., Lammel, J., Kuhlmann, H., 2002a. Impact Assessment of Abiotic
Resource Consumption - Conceptual Considerations. Int. J. LCA 7, 301-307.
Brentrup, F., Küsters, J., Lammel, J., Kuhlmann, H., 2002b. Life Cycle Impact Assessment of
Land Use Based on the Hemeroby Concept. Int. J. LCA 7, 339-348.
Brentrup, F., 2003. Life Cycle Assessment to Evaluate the Environmental Impact of Arable
Crop Production. PhD thesis, Univ. of Hanover, Germany. 187 pp.
Brentrup, F., Küsters, J., Lammel, J., Kuhlmann, H., 2003b. Environmental impact assess-
ment of agricultural production systems using the life cycle assessment methodology. I.
Theoretical concept of a LCA method tailored to crop production. Europ. J. Agronomy,
in press.
Brentrup, F., Küsters, J., Lammel, J., Kuhlmann, H., 2003c. Environmental impact assessment
of agricultural production systems using the life cycle assessment methodology. II. The
application to N fertilizer use in winter wheat production systems. Europ. J. Agronomy,
in press.
General LCA concept Methodological contribution
Life Cycle Inventory Estimation methods for diffuse,
on-field N-emissions
Characterization methods for
Aggregation to impact category indicators • consumption of abiotic resources
Li
fe
(Characterization) • land use
Cy
cle
Im
Normalization Development of a consistent set of
pa
ct
• normalization values
As
• weighting factors
se
ss
Weighting for environmental impacts relevant
m
en
to arable crop production
t
Environmental Development of
indicator • Environmental index (EcoX)
• Resource depletion index (RDI)
Figure 1. LCA concept and methodological contributions proposed in this study.
263
0,8 10
Aqua. eutroph.
Terr. eutroph.
yield 9
0,7 Acidification
Climate change
8
Land use
0,6
7
Grain yield (t/ha)
0,5
Environmental indicator
6
(EcoX/t grain)
0,4 5
4
0,3
3
0,2
2
0,1
1
0,0 0
N0 (0) N1 (48) N2 (96) N3 (144) N4 (192) N5 (240) N6 (288)
Plot no. (N fertilizer rate in kg N/ha)
Figure 2. Aggregated environmental indicator values (EcoX) per ton of grain (stacked bars)
and yields (t/ha, dots) at increasing N fertilizer rates.
264
Life cycle inventory modelling in the Swiss national LCI database ecoinvent
2000
Jungbluth, N., Frischknecht, R.
*ESU-services, Kanzleistrasse 4, CH-8610 Uster, T. +41 1 940 61 91 or +41 1 940 6132, F. +41 1
940 61 94, [email protected], [email protected] Centre for life cycle inventories in
the ETH domain, c/o EMPA, CH - 8600 Dübendorf
1. Abstract
In late 2000 the project ecoinvent 2000 has officially been launched. Several Swiss federal
agencies and nine institutes of the ETH domain agreed on a joint effort to harmonise and up-
date life cycle inventory (LCI) data for its use in life cycle assessment (LCA). The goal is a
unified and generic set of LCI data of high quality which is valid for Swiss and Western Euro-
pean conditions. The Centre for life cycle inventories in the ETH domain has developed a
central database building on past experiences with a large network-based LCI database de-
veloped at ETH Zurich. The database comprises in a first version LCI data from the energy,
transport, building materials, chemicals, detergents, paper and pulp, waste treatment and ag-
ricultural sector (ecoinvent data v1.0). The content of the database will be made publicly
available via the web page www.ecoinvent.ch. A data exchange format (ECOSPOLD) has
been developed for the data im- and export to different software tools. This format uses the
extended mark-up language (XML) for data exchange. The consistent and coherent LCI
datasets for basic processes make it easier to perform LCA studies, and increase the credi-
bility and acceptance of the life cycle results.
Keywords: ecoinvent, life cycle inventory, database, background data.
2. Motivation and problem setting
Up to now, several public Life Cycle Assessment (LCA) databases exist in Switzerland,
partly covering the same economic sectors (Frischknecht et al. 1996, Gaillard et al. 1997,
Habersatter et al. 1998). However, life cycle inventory data for a particular material or proc-
ess available from the databases often do not coincide and therefore the outcome of an LCA is
(also) dependent on the institute working on it. Furthermore the efforts required to maintain
and update comprehensive and high quality LCA-databases are beyond the capacity of any
individual institute.
At the same time, LCA gets more and more attention by industry and authorities as one im-
portant tool for e.g., Integrated Product Policy, Technology Assessment or Design for the En-
vironment. In parallel with this increasing trend in LCA applications the demand for high
quality, reliable, transparent and consistent LCA data increased as well. Only a few publicly
available LCI databases fulfil these criteria and most of them were published in the nineties.
265
3. Goal of ecoinvent 2000
That is why LCA-institutes in the ETH domain (Swiss federal Institutes of Technology (ETH)
Zürich and Lausanne, Swiss Federal Laboratories for Materials Testing and Research
(EMPA) St. Gallen and Dübendorf, Paul Scherrer Institute (PSI) Villigen, and the Swiss Fed-
eral Institute for Environmental Science and Technology (EAWAG)) as well as the LCA-
department of the Swiss Federal Research Station for Agroecology and agriculture (FAL) in
Zurich agreed on a close co-operation. Together with the Swiss Agency for the Environment,
Forests and Landscape (SAEFL or BUWAL), the Swiss Agency for Energy (BFE) and other
agencies they founded the Centre for Life Cycle Inventories in the ETH domain. The database
comprises LCI data from the energy, transport, building materials, chemicals, paper and pulp,
waste treatment and agricultural sectors (see Table 1).
A large, network-based database and efficient calculation routines are required for handling,
storage, calculation and presentation of data and are developed in the course of the project.
These components partly take pattern from preceding work performed at ETH Zurich
(Frischknecht & Kolm 1995).
4. Basic structure of the ecoinvent database system
The ecoinvent 2000 database system consists of the following components (see Figure 1):
1. The central database,
2. Calculation routines,
3. The local databases,
4. The administration tool,
5. The query tool,
6. The data (exchange) format,
7. The editor.
Ad 1. The central database contains LCI data and Life Cycle Impact Assessment methods
such as the Swiss Ecological Scarcity 1997, Eco-indicator 99 or the CML characterisation
scheme 2001. The database is located on a server and accessible via Internet.
Ad 2. Data will be supplied by the partner institutes as non-terminated unit processes. The
computation of cumulative inventory results is performed with powerful calculation routines
related to the central database. Input data as well as calculations will include (cumulative) un-
certainty ranges.
Ad 3. Commercially available LCA-software such as Gabi, SimaPro, Team and Umberto are
used as local databases. These local databases are suited for an implementation and use of
ecoinvent data. The ecoinvent data (exchange) format is recommended for that purpose.
266
Ad 4. The administration tool supports the integration of datasets delivered by the co-
operating institutes into the central database. It helps to verify the completeness of datasets,
calculates inventories and (normalised and weighted) category indicator results and ensures
the accessibility for clients respecting the users’ rights.
Ad 5. The Query tool is used to interrogate the database and to download datasets from the
central database. It enables the search for individual processes, for processes of a certain eco-
nomic sector (e.g., transport or energy sector) or for data from a certain institute. General in-
formation (so-called meta information) about the processes (technology, age, geographic cov-
erage, et cetera) is accessible to everybody whereas the quantitative LCI data is only
accessible for registered members (clients) of ecoinvent database.
Ad 6. The data exchange format lists all data fields that need to be completed when data is
imported into the central database for the first time. It has evolved from the international
SPOLD data exchange format (Weidema 1999) and takes pattern from the committee draft of
the international technical specification ISO 14048 (International Organization for Standardi-
zation (Iso) 2001). Some of the data fields are mandatory, i.e. information must be provided.
Among other features, the data exchange format allows for specifying upper and lower esti-
mates (or the coefficient of variance) as well as the probability distribution (e.g., lognormal).
Ad 7. The local administrators use the editor and EXCEL software to create new datasets and to
change, enlarge or delete existing datasets. The editor administrates the module names (via a
direct link to the central database, where the index of module names is placed). The editor
acts as the interface between the local administrator and the central database and generates
files in the ecoinvent 2000 data format.
5. Quality guidelines ecoinvent 2000
In the next section selected aspects of the quality guidelines for life cycle inventory analyses
performed within the ecoinvent 2000 project are described.
5.1. Introduction
The creation of one central life cycle assessment database requires a high degree of co-
ordination and harmonisation. Besides structural aspects and naming conventions, content-
related aspects have been discussed and unified. This guarantees a maximum degree of con-
sistency of process data available in the database. Here we focus on content-related aspects.
5.2. Content-related aspects
While structural aspects and naming rules are in many cases controllable by the software, a
consistent application of content-related rules is less straightforward. Nevertheless, clear rules
are required in order to minimise differences caused by individual, unsystematic choices of
the LCI practitioners involved.
267
System boundaries are drawn based on expert knowledge and not based on fixed rules such as
mass or energy shares. If the emission of a pollutant must be expected but no data are avail-
able, estimates are used in order to identify whether or not this pollutant may be environmen-
tally relevant.
Electricity is supplied on high, medium and low voltage with increasing losses and investment
requirements. Hence electricity demand of processes must be linked to the correct (or most
likely) voltage. The supply mix (as well as the export mix) is calculated based on the domes-
tic production plus the imports. In cases where the electricity mix actually purchased deviates
from the average supply mix of a nation (or region) such specific mixes (or particular power
plant technologies) are used in the model.
Standard transport distances are applied for materials such as steel, cement, basic chemicals et
cetera, in case the exact distances are unknown. A similar approach is chosen for waste treat-
ment processes. If no particular information is available, standard waste treatment processes
defined per material are applied. It is supposed that inert materials go to landfill, plastics are
incinerated and metals are recycled.
Allocation is an ubiquitous issue that calls for a harmonised approach. A cut-off approach is
used for recycled materials and for by-products (outputs with no economic value that are not
sent to waste treatment but are used in other processes). No burdens and no requirements of a
preceding process chain and of a process are allocated to the recycled materials and by-
products, respectively. On the other hand no benefits are granted for any subsequent use of
recycled material or by-product. No fixed prescriptions are made for joint product allocation
(co-products) except that system expansion (especially the "avoided burden"-concept) is not
recommended.
Fossil and renewable carbon are distinguished for CO2-, CH4- and CO-emissions. For renew-
able energy sources and materials an equal amount of CO2 is registered as a resource con-
sumption according to the binding capacity of the corresponding crops. Carbon that is emitted
as CO is considered when calculating CO2-emissions. On the other hand, CO will get a global
warming potential assuming its subsequent conversion to CO2.
Uncertainty of flow data is quantified on the level of unit processes. If uncertainty is not
known (because not stated in the sources used or because not known by the company provid-
ing the data) a standardised procedure is used for estimations. A data quality matrix has been
developed which takes pattern from the pedigree matrix published by (Pedersen Weidema &
Wesnaes 1996). Scores from 1 to 5 are given for reliability, completeness, temporal correla-
tion, geographical correlation, further technological correlation and sample size. Fixed uncer-
tainty factors are attributed to each of the scores and an additional basic uncertainty is attrib-
uted to categories of exchanges (such as electricity and thermal energy consumption, groups
of combustion emissions, waste treatment requirements and the like). In most cases a log-
268
normal distribution is assumed. With the help of this standardised uncertainty factors, the ge-
ovariance is determined for each individual exchange in the unit processes.
6. Outlook
The software system presented in this paper is in its final development phase. The database is
fed with LCI datasets and life cycle impact assessment methods. The size of the economic
part of the matrix (up to 3'000 unit processes in its first version) certainly poses a real chal-
lenge for the project team in terms of database response and computation time.
The use of XML technology for the exchange of data between the participating institutes and
between them and the central database is a challenge not only for the project team but also for
the LCA software suppliers who are encouraged to implement the ecoinvent data exchange
format. Variations of the data exchange format are possible thanks to the flexibility of the
XML technology. This should further enhance the acceptability of this format in the LCA
community.
In fall 2003 the database will go online. By then, LCI data with the reference year 2000 will
be available via www.ecoinvent.ch for many basic products and services (such as energy sup-
ply, transportation and waste treatment services, building materials, chemicals and agricul-
tural products) that make part of most LCI process networks.
References
Frischknecht, R., Bollens, U., Bosshart, S., Ciot, M., Ciseri, L., Doka, G., Dones, R., Gantner, U.,
Hischier, R., Martin, A. 1996: Ökoinventare von Energiesystemen: Grundlagen für den öko-
logischen Vergleich von Energiesystemen und den Einbezug von Energiesystemen in Ökobi-
lanzen für die Schweiz. Auflage No. 3, Gruppe Energie - Stoffe - Umwelt (ESU), Eidgenös-
sische Technische Hochschule Zürich und Sektion Ganzheitliche Systemanalysen, Paul
Scherrer Institut, Villigen, www.energieforschung.ch, Bundesamt für Energie (Hrsg.), Bern,
CH.
Frischknecht, R. & Kolm, P. 1995: "Modellansatz und Algorithmus zur Berechnung von Ökobi-
lanzen im Rahmen der Datenbank ecoinvent." In Schmidt M., S. A., Stoffstromanalysen in
Ökobilanzen und Öko-Audits. Pages: 79-95, Springer-Verlag, Berlin Heidelberg.
Gaillard, G., Crettaz, P., Hausheer, J. 1997: Umweltinventar der landwirtschaftlichen Inputs im
Pflanzenbau. FAT-Schriftenreihe No. 46, Eidg. Forschungsanstalt für Agrarwirtschaft und
Landtechnik (FAT), Tänikon.
Habersatter, K., Fecker, I., Dall`Acqua, S., Fawer, M., Fallscheer, F., Förster, R., Maillefer, C.,
Ménard, M., Reusser, L., Som, C. 1998: Ökoinventare für Verpackungen. Schriftenreihe
Umwelt No. 250, 2. korrigierte und aktualisierte Auflage, Bundesamt für Umwelt, Wald und
Landschaft, Bern, Schweiz.
International Organization for Standardization (ISO) 2001: Environmental Management - Life Cy-
cle Assessment - Data documentation format. Technical Specification ISO/DTS 14048 Ge-
neva.
Pedersen Weidema, B. & Wesnaes, M. S. 1996: “Data quality management for life cycle invento-
ries - an example of using data quality indicators.” In Journal of Cleaner Production Vol. 4
(3-4): 167-174.
Weidema, B. K. 1999: SPOLD '99 format - an electronic data format for exchange of LCI data
(1999.06.24). SPOLD, www.spold.org.
269
Table 1. Database content, responsible institutes and their partners in LCI data compilation
(see www.ecoinvent.ch for addresses and responsible persons).
Database content Responsible Institute Partners
Energy supply Paul Scherrer Institute (PSI) ESU-services
Fuels
Heat production
Electricity production
Plastics Swiss Federal Laboratories for Doka Ökobilanzen, Chuda-
Paper and Board Materials Testing and Research coff Öko-science
Basic Chemicals (EMPA)
Detergents
Waste treatment services
Metals Swiss Federal Laboratories for Chudacoff Öko-science
Wood Materials Testing and Research
Building materials (EMPA)
Basic chemicals
Transport services Swiss Federal Institute of Tech-
nology Zurich, (ETHZ UNS)
Basic chemicals Swiss Federal Institute of Tech-
nology Zurich, (ETHZ LTC)
Agricultural products and Federal Research Station for Research Station for Agricul-
processes Agroecology and Agriculture, tural Economics and Engi-
(FAL) neering (FAT)
Local Administrator Local
database
Editor
Central
Administrator
Administration Central Calculation
tool Database routines
Query tool
Data consumer XY
Figure 1. The basic structure of ecoinvent database system.
270
Environmental profiles of foods for the Swedish market
Lagerberg-Fogelberg, C.a)*), Carlsson-Kanyama, A.b), Fogelberg, F.c) and Brandén, I.d)
a)
Swedish University of Agricultural Sciences and Mid Sweden University, Centre for Sustainable Ag-
riculture, P. O. Box 7047, SE-750 07 Uppsala, Sweden, telephone: +46 18 67 16 48, Fax: +46 18 67
35 71, e-mail: [email protected]
b)
Environmental Strategies Research Group, FOI, SE-172 90 Stockholm, Sweden, e-mail:
[email protected]
c)
Swedish University of Agricultural Sciences, Dept. of Plant Science,P.O. Box 44, SE-230 53 Alnarp,
Sweden, e-mail: [email protected]
d)
Mid Sweden University, Dept. of Natural and Environmental Sciences, SE-831 25 Östersund, Swe-
den, e-mail: [email protected]
Keywords: agriculture, energy, GHG, wholesale, LCA, processing, resource, retail.
The wholesale and food service sectors control food delivered to large numbers of consumers.
About 5.6 million meals are served each day in Swedish institutional households, for example
in private and public restaurants in schools, military defence and public health care. The pur-
chasers of food service institutions and the food wholesale sector have a key position in the
food chain (Figure 1) in that they control large flows of foods, thereby allowing them to prac-
tice potential influence on the food supply chain. Consequently these sectors have a great im-
pact on the overall environmental effects of food consumption patterns.
Environmental profiles of a number of foods are estimated from a life cycle perspective. Sup-
ply chains are currently being mapped for the analyzed products. We focus on fresh carrots,
tomatoes, apples and meat, frozen broccoli, onion and chicken and dried beans/peas.
Preliminary findings suggest complex patterns of delivery of foods for the wholesale and food
service institutions sector. The companies participating in the study buy foods from a wide
variety of companies. Companies supplying goods and services are both large and small and
several are specialized in trading and logistics. As illustrated by Figure 2, the complexity is
not only displayed by the many actors and choices at each level but also by the cross flows of
products, i.e. companies in several cases purchase a certain product from more than one
source while the upward flows of the product passes through several companies. Conse-
quently the product may eventually be of the same origin but delivered through different sup-
ply chains.
The above project is a part of the ongoing three-year project “Designing and evaluating the
impacts of environmental information in food service institutions and the food wholesale sec-
tor” which is funded by the Foundation for Strategic Environmental Research, MISTRA. The
project focuses on how different factors interact in food purchase and how environmental in-
formation of foods affect the decisions of food purchase of institutional households and
wholesalers. The project investigates how well-designed environmental information can im-
prove environmental performance in food service institutions and the wholesale sector. Needs
271
and practices of suppliers, producers and corporate consumers in relation to environmental in-
formation about food are investigated. Simplified descriptions of life-cycle related environ-
mental impacts from selected food products are developed. The focus is on product groups
where substantial differences between product choices may be expected. Different types of
environmental information and how they interact with factors such as norms and directives
will be tested. Recommendations for design of information for specific key users, such as
purchasing managers in food service institutions and the wholesale sector will be addressed.
The overall potential impact of such information packages on the environmental performance
of the studied sector will also be evaluated.
For more and updated information please visit the project homepage at www.e-info.se.
Food service
institutions
Primary Production, Processing, Waste
Wholesale Consumption
production incl storage incl storage management
Distribution
Direct control
Indirect control
Figure 1. Generic food supply chain. Transportation occur between all compartments.
Primary producers, Primary producers, Primary producers, Primary producers,
Spain1 Spain2 Sweden The Netherlands
E E E E E E
D D D D
C C C C C
B B B
Figure 2. Preliminary supply chain for fresh table tomatoes for one purchasing manager (A).
272
Using I/O data to find hotspots in LCA - Example of a hamburger meal
Jacob Madsen and Suzanne Effting
PRé Consultants, Plotterweg 12, Amersfoort, The Netherlands, 3821 BB Tel +31 33 4555022,
fax. +31 33 4555024 email: [email protected], [email protected]
Introduction
Each country has an economic input output table that specifies the value of the purchases be-
tween sectors within the country and abroad (the imports), as well as the supplies to other sec-
tors and the exports. Furthermore all other major costs and revenues are specified. These ta-
bles have been used by several LCA experts to compile input output (IO) databases [1].
In the present study we compare the environmental impact of a hamburger meal based on data
from traditional LCA process databases and data from the Dutch Input Output database con-
taining sector IO data both from the Netherlands and the rest of the world. Since LCA is an it-
erative process it is essential to have data available of your full life cycle. When performing
an LCA you first identify the hotspots in the life cycle and the results you use in the refine-
ment of your study. The advantage of IO data is that it covers an entire economy thus includ-
ing all products. However IO data lack process specificity since all products are aggregated
into sectors. With this in mind the objective of this study is to validate the data from the IO
database in order to fill out the gaps that normally occur in the data collection process.
Databases
In our work we aim to use existing databases to provide a preliminary picture of the environ-
mental impact of the product under study. A simplified model of the life cycle of a hamburger
meal was build using process data based on Buwal [2] for the packaging and IVAM [3] for
the food products that go into the hamburger meal i.e. Cola, French fries and a hamburger
sandwich. At the same time an identical model was built using I/O data from the Dutch I/O
database [4].
The Dutch I/O database
The Dutch input output database consists of a matrix with 105 producing sectors and their
imports. Regarding emissions the Dutch “emission registry” system maintains a detailed data
inventory for industrial activities. This database has been used to run queries that produce
datasheets per sector. In order to calculate the impact of imported products, the “rest of the
world” was split up into three regions:
• OECD countries in Europe
• Other OECD countries
• Non OECD countries
For each of these regions, thirty sectors were defined that were taken from the DIMITRI [5]
and EDGAR [6] database. The EDGAR database already has data on Energy use, CO2, NOx
273
and SOx per country and per sector. To cover the other stressors a wide range of sources has
been consulted. In order to focus the efforts, an analysis was made using the GTAP [7] data-
base to identify which countries or regions contribute most to an industrial activity. The focus
was to find data for these countries and regions first, and extrapolate this data over the whole
region. Of course the data collection was not complete, and often extrapolations have had an
important influence.
The use of different sector definitions for the Netherlands (105 sectors) and the three regions
that cover the rest of the world (30 sectors each) requires a conversion routine.
An aggregation table has been constructed that specifies which of the 105 Dutch sectors can
be aggregated to one of the 30 international sectors.
Each Dutch sector has 105 domestic purchases and 105 imports. The 105 imports were con-
verted into 30 imports using the aggregation table.
Dutch trade statistics have been used to determine which share of each import comes from the
3 regions which are used to model the world outside the Netherlands. Competing and non
competing imports have been treated in the same way. Also here a 30 sector aggregation has
been used.
Inventory
An average hamburger meal was estimated to consist of:
• 83 gram beef • 225 gram potatoes • 20 gram plastic
• 88 gram bread • 376 gram cola • 100 km truck transport
• 40 gram tomatoes, salad etc • 20 gram paper • 7,5 MJ heat
• 0,7 kWh Electricity
The life cycle was modelled in SimaPro 5.1 and calculated with the Eco-indicator 99 method.
Results
When comparing the results of the process LCA with the IO LCA we observe the same pat-
tern. The categories land use, fossil fuel use and respiratory inorganic substances dominate the
overall environmental impacts, see figure 1.
However when looking at the process contribution we noticed that the I/O data for potatoes,
tomatoes and salad have a significant environmental impact (figure 2). The agricultural sector
is very heterogeneous and the average product from this sector is probably not representative
for products like potatoes, tomatoes and salad. Additionally, the average product from the
‘animal based food sector’ underestimates the impact from cow meat when comparing the two
274
data sources since cow meat typically has a higher impact in comparison with for example
chicken and pork.
Furthermore we notice a significant environmental impact from the services related to the
production of a hamburger meal. In this context services are marketing, advertisement etc.
The impacts from services are something that seldom is taken into account in traditional LCA.
Conclusions
The results showed that the I/O database provides representative data for identification of the
hotspots in the life cycle of a hamburger meal. However, when examining the process contri-
bution we discovered that the environmental impacts from potatoes, tomatoes and salad were
high. Additionally the environmental load from cow meat was low compared to process LCA
data. The agricultural sector is a heterogeneous sector and the average product from this sec-
tor is not representing products like tomatoes, salad etc. Therefore, additional data collection
regarding these inputs would be needed. This also supports the general assumption that IO
LCA is complete in system boundaries but lack process and product specificity. At the same
time, LCA's based on process databases often are specific and detailed but incomplete due to
cut-offs. One solution is the hybrid method where you combine the two methods. We recom-
mend to use available process LCA data and I/O to fill in the data gaps and complete the in-
ventory of the whole life cycle.
References
1. Joshi, S.; Product Environmental Life-Cycle Assessment Using Input-output Techniques,
Journal of Industrial Ecology, Volume 3, nr 2&3, 2000
2. BUWAL 250, Life Cycle Inventories for Packagings, Volume II. Environmental Series No.
250, Bern, 1998
3. IVAM LCA Data 4. Ewijk, H.A.L., Krutwagen, B.T.J.M., Saft, R.J., Otto, A. IVAM 2002
4. Goedkoop M., Database Manual Dutch Input Output Database 95, PRé Consultants, Am-
ersfoort, 2003
5. Wilting, H.C. et. Al. Dimitri 1.0 : Beschrijving en toepassing van een dynamisch input-
output model (Dimitri 1.0 : Description and application of a Dynamic input-output
model, in Dutch), RIVM report 778001 005, Bilthoven, the Netherlands, 2001
6. Olivier, J.G.J et al. Description of EDGAR 2.0. RIVM report 771060002, Bilthoven the
Netherlands 1996; a recent version 3.0, can be accessed on
www.rivm.nl/ieweb/ieweb/index.html?databases/edgar.html.
7. GTAP database, Center for Global Trade analysis, Purdue University,
www.gtap.agecon.purdue.edu/
275
0,50
0,45
0,40 Fossil f uels
Minerals
0,35 Land use
Eco-indicator [Pt]
Acidification/ Eutrophication
0,30
Ecotoxicity
0,25 Ozone layer
Radiation
0,20 Climate change
Resp. inorganics
0,15
Resp. organics
Carcinogens
0,10
0,05
0,00
I/O Process
Fig. 1. Comparison of the environmental impact from the life cycle of a hamburger meal
when using process LCI data and I/O LCI data.
0,50
0,50 Input/Output Process LCA
Input/Output Process LCA
0,45
Services -
0,45
Services -
0,40
Energy Energy
0,40
0,35
Transport over Energy Energy
Eco-indicator [Pt]
Truck
land
0,30
0,35 Paper, PaperTransport
+ EPS over
[Pt]
Chemical&plastics Truck
0,25 land
Beverages
0,30 Cola
Paper,
Eco-indicator
0,20 Paper + EPS
Agriculture Chemical&plastics
Potatoes
0,25
0,15 Beverages
Agriculture
Tomato+salad
Cola
0,20
0,10
Animal based food Agricult ure
Cow meat Potat oes
0,05
0,15 Other food
BreadAgricult ure
0,00
Tomato+salad
0,10 I/O Process
Animal based f ood
Cow meat
0,05
Other food
Bread
0,00
I/O Process
Fig. 2. The process contributions to the environmental impact of the life cycle of a hamburger
meal when using process LCI data and I/O LCI data
276
The use of by-products from food industry as basis for livestock feed and
the consequences for the analysis of the environmental impacts of meat con-
sumption
Nonhebel, S., Elferink, E.V.*
* Center for Energy and Environmental Studies, Nijenborgh 4, 9747AG Groningen, The Netherlands,
Tel: (++)31-50-3637177, fax (++)31-50-3637168
Abstract
The production of meat puts a large claim on resources. The allocation used, however, is of
influence on the environmental impact of meat. These allocations can change suddenly due
to economic or food safety regulations. Meat production is, however, closely linked with the
production of other food products. Focussing on pork only instead of the entire food system
makes the environmental impact of pork opaque. Therefore, determination of the environ-
mental impact of the combined production of pork and by-products gives far more consistent
results than an analysis of individual products.
Keywords: allocation, meat, environmental impact, by-products, system analysis.
Introduction
Analysis of the environmental impact of food packages has shown that especially the con-
sumption of meat goes together with large environmental impacts. A more detailed analysis of
the meat production system, however, shows that a large fraction, over 50%, of the livestock
concentrated feed is generated from by-products from the food industry e.g. sugar pulp and
molasses from the sugar industry, oil-seed cakes from the vegetable oil industry, but also
slaughter wastes (Dutch statistical bureau of agriculture and horticulture, 2000). This large
fraction of by-products in the feed complicates the analysis of the environmental impacts of
meat. When these by-products are considered as unwanted ‘waste-streams’ of the food indus-
try the environmental impacts allocated to meat are very low, and the production of meat can
be considered as an environmental friendly method to convert a by-product in a highly valued
food item (Nonhebel, 2003). In that case meat production even prevents dumping wastes.
Presently by-products from the food industry can be considered as co-products, since the food
industry obtains a considerable income on selling their ‘waste-streams’ to the livestock fodder
industry. However, new insights in food safety and economical trends have led and will lead
to changes in permitted by-products in livestock fodder. BSE in cows, for instance, led to a
ban on the use of slaughter waste as an ingredient for livestock fodder. This implies that pres-
ently available by-products can change in value from highly valued basis of livestock fodder,
to very unwanted by-products. Such changed regulations may have large consequences for the
derived environmental impacts of meat.
277
This paper focuses on the environmental impacts allocated to the by-products, various alloca-
tion procedures are recognised and the consequences for the results are evaluated.
Method
System description
Production of food products of animal origin as meat milk and eggs, requires a complex sys-
tem with interaction at different levels of scale and between systems, e.g. local vs. global and
livestock vs. crop. Analysis at a higher level of scale than currently done in environmental as-
sessments, leads to improved insight in the system, the environmental impacts and its major
players (Nonhebel, 2003).
Dutch pork production and consumption is studied in this paper as a model sector for western
meat production. In the Netherlands pigs are mainly kept in an intensive manner, meaning
that they are kept in barns and are fed concentrated feed. Dutch pig production can roughly be
divided in three sectors, a pedigree sector, a multiply sector and a fattener sector. After fatten-
ing pigs are being slaughtered, processed and consumed. Transportation occurs between all
sectors. The fodder industry produces the feed, produced from two streams of raw materials.
The first are imported raw materials primarily grown for concentrated feed, e.g. grains, pulses
and tapioca. The second are different kinds of by-products from the Dutch food industry. The
largest streams of by-products, representing 66% of total available, are; soybean cake, by-
products from the sugar industry, beet pulp and molasses, and potato peels and residues
(Dutch statistical bureau of agriculture and horticulture, 2000). The by-products are relatively
cheap, homogeneous and have a good nutritional value. Therefore these streams are very suit-
able ingredients in concentrated feed but also for feeding directly to livestock. The nutritional
value of fodder for pigs is expressed in Energy value pig (Evp). This nutritional value is de-
termined from literature for each type of raw material (Centraal Veevoederbureau, 1997).
Taking loses due to premature dead and cutting into account, 4 Evp is needed to produce 1
kilogram of pork (Elferink, 2001).
The indicator used in this paper is energy input. The energy input is allocated to the by-
products until fodder production. It is assumed that energy input in the remainder of the pork
production and consumption chain is independent of the raw materials used for fodder, and is
therefore left out in this analysis.
Allocations used
There are several options for allocating environmental effects (Proce, 1986). In this study by-
products are allocated in three different manners:
1. On a by-product basis, meaning that no environmental impact was attributed to the by-
product but only the main product.
2. On the present economical value as a raw material for fodder.
3. On mass ratio
278
Table 1 shows the allocation ratio for the different crops and concentrate ingredients as used
in this study as well as the energy required for yielding the crop and the Evp. The table is de-
rived from data from; the Dutch statistical bureau of agriculture and horticulture (CBS), Agri-
cultural Economics Research Institute (LEI) and the Dutch Central Bureau for Livestock
Feeding (CVB).
Results
Table 2 shows the energy input to produce the raw materials for 1 kilogram vegetable oil,
sugar and potato product. The energy input for fodder is not on a kilogram base but depends
on the amount of by-product released from the food industries to produce 1 kilogram of main
product. Results are compared with feeding on wheat only. This is comparable with a scenario
in which food regulations forbid the use of by-products in fodder. The amount of grain is
equal to the total Evp, 4.8 Evp, of the by-products generated.
If allocated on by-product, fodder has no impact, while the ‘main’ products have a high im-
pact. This scenario is not realistic because the food industry gains a lot on selling by-products.
When allocated on mass the energy input of by-products is slightly lower than for wheat. An
allocation on economical value, however, shows an energy impact that is substantial lower for
fodder on the different by-products than for fodder on wheat.
The allocation used has consequences for the calculated environmental impact of a product.
However, total energy use of al products together, e.g. meat, sugar, etc., is constant as long as
the food system doesn’t change. When the system changes, however, energy input changes
also. In this case total energy input increases when by-products are not being fed anymore.
Discussion and conclusion
The allocation used has a large influence on the energy input required to produce fodder for
pork. These allocations can change suddenly due to new (food) regulations or economic
trends. Making the environmental impact of pork very opaque. It is further shown that there is
a strong correlation between the food production industry and the livestock fodder industry.
These industries affect each other in the choice of agriculture products used as basis for their
production process. The determination of the energy input of the combined production of pork
and by-products will give far more consistent results than an analysis of individual products.
This recycling of by-products as analysed in here for pork also occurs in other parts of the
food production system. More attention to these product linkages would improve the insight
in the environmental impacts of food production system as a whole.
References
Centraal Veevoederbureau, 1997. Voedernormen landbouwhuisdieren en voederwaarde
veevoeders; Verkorte tabel. Lelystad.
Dutch statistical bureau of agriculture and horticulture, 2000. Land en Tuinbouwcijfers. 's-
Gravenhage, LEI-DLO/ CBS.
279
Elferink, E. V., 2001. Vlees een duurzame eiwitbron? Een modelstudie naar het inzetten van
organische reststromen in de veehouderij. Groningen, RuG, Center for Energy and Envi-
ronmental Studies (IVEM).
Nonhebel, S., 2003 On resource use in food production systems: The value of livestock as
'rest-stream upgrading system'. Ecological Economics. Accepted.
Proce, C., 1986. Energieverbruik in de nederlandse akkerbouw en veehouderij. IVEM-report
nr. 17. Groningen, IVEM.
Table 1. Allocation ratios, yields and inputs for the crops and their ingredients used for con-
centrated pig feed.
Crop Products Yield Energy input Evp Allocation (%)
ton/ha) (Mj/ kg) mass price by-product
Wheat Grain 8.2 3.10 1.01 100 100 100
Soybean oil, cakes 2.5 2.95 (1.39) 0.96 20, 80 50, 50 100, 0
Sugar beet sugar, beet pulp, 57.9 0.55 (0.26) 0.98 59, 24, 91, 5, 100, 0,
Molasses 0.80 17 4 0
Potato Potato, peels 47 0.94 (0.98) 1.21 77, 23 99, 1 100, 0
Table 2. Energy input of agricultural products allocated in four different ways.
Allocation Oil Sugar Potato Fodder Total
(MJ/kg) (MJ/kg) (MJ/kg) (MJ) (MJ)
by-product 26.3 9.1 3.5 0.0 38.9
price 13.2 8.2 3.5 14.0 38.9
mass 5.3 5.3 2.7 25.6 38.9
grain 26.3 9.1 3.5 19.7 58.6
280
Is manure separation a solution to environmental problems from large scale
pig production?
Claus Grøn Sørensen1, Niels Halberg2, Randi Dalgaard2
1
Danish Institute of Agricultural Sciences, Department of Agricultural Engineering, Research Centre
Bygholm, DK-8700 Horsens, Tel: 7629 6023. E-mail: [email protected]
2
Danish Institute of Agricultural Sciences, Department of Agroecology, Research Centre Foulum, P.O.
Box 50, DK-8830 Tjele.
There is a rapid structural development of the agricultural holdings in Denmark. It is esti-
mated that the number of dairy farms and the number of pig farms will be reduced by 50%
within the next 10 years. However, milk production will be maintained and pork production
increased at the same time. This development will result in an increased specialization and a
marked enlargement of each holding with respect to land use, substance turnover, and eco-
nomic turnover. Hereby, the economic and environmental impact of the actions taken by each
farmer is increased considerably.
In relation to the structural development, new technologies are introduced and new production
systems emerge. These include other ways of handling slurry including biogas production and
separation, and other ways of carrying out tillage of the arable land. These technologies influ-
ences resource use (energy etc.), nutrient efficiency in the system, use of toxic elements, and
eventually environmental impact, (greenhouse effect, acidification, eutrophication, and
ecotoxicity) of the agricultural production. Often an introduced technique can have a positive
influence in one aspect but a negative influence in another aspect and it may be difficult to
make an easy assessment of it.
In order to comply with national priorities and international obligations, there is a need to
know how the new production methods may affect the range of environmental impact in ques-
tion and to identify where in the production process the critical events occur. Also, it is impor-
tant to make such information available for decision makers at different levels including poli-
ticians, farmer organizations, the advisory service, and companies involved in development of
the new technologies. This is the objective of the project: Evaluation of innovative agricul-
tural production systems through a life cycle assessment (LCA) methodology (2003-2005).
Overall, the aim of the project is to investigate and evaluate the impact of introducing new
production technologies within, pork production, and plant production in resource use, and
the environmental impact using an LCA methodology. Moreover, it is the aim to identify the
most 'critical' production process or production phases in relation to specified resource use
and environmental impact parameters (greenhouse effect, acidification, eutrophication and
eco-toxicity) resulting from the production of a specific amount of food-products.
281
Specifically, it is the aim, within the above context, to evaluate:
1. Different ways and methods of handling liquid manure on pig farms including estab-
lishment of on-farm biogas systems and handling of separation fractions.
2. The impact of introducing new precision farming technologies (e.g. new tillage tech-
niques as e.g. no-ploughing) in arable farming on energy use, nutrient efficiency, and
pesticide use.
New innovative systems for slurry handling on pig farms, and minimal tillage techniques are
identified and formulated in a collaborative process including the advisory service, research-
ers, and farmer organizations. Each system is 'optimized' in land use as well as in machinery
and equipment. The resource use, the technical efficiencies, the emissions, and the changes in
carbon soil sink in relation to the agricultural production process will then be measured in
private farms or estimated. The systems are evaluated through a life cycle assessment meth-
odology as well as through traditional economic tools. By using the LCA methodology the
environmental impact is evaluated per unit of production in agreement with the idea of the ef-
forts in the EU to improve the environmental policy through a product-oriented approach. In
addition, estimates of the environmental impact on a per-hectare basis and of farm economy
are obtained.
Scenario identification
A number of models of large scale pig production were derived from a participatory work-
shop with pig producers and representatives from companies involved in manure separation
and biogas techniques. Figure 1 shows an example of a future farm with a total production of
83850 pigs on 3 units including sow/piglet production. Danish regulation limits the applica-
tion of manure to 140 kg total N in manure applied per ha. Therefore, there is an increasing
interest in methods for reducing the amount of slurry to be handled and transported. Separa-
tion of manure or slurry into liquid and fibre phases may reduce the amount of water trans-
ported when exporting nutrients from the pig farm to agricultural land on other farm. How-
ever, it is not clear how this effect will reduce the total environmental impact from intensive
pig production. Different systems using manure separation and biogas will be studied in pri-
vate farms and modelled within an LCA framework in order to compare the environmental
impacts. The need for additional knowledge as regards methods and operational data for the
selected model farms will be determined through a detailed identification of the existing data
on systems for separation, biogas, etc. This evaluation will determine the amount of additional
data to be collected from private farms and separation installations.
Transport and logistics
As part of the modelling framework, specific models on, for example, the logistics of different
manure distribution strategies will be used to evaluate the transport efforts in terms of labour
input and transport distances. For example, the average transport distance is estimated for
various different scenarios involving amounts of manure, harmonization demands, and sepa-
ration.
282
Farm unit 1
Supplmemtal field
3733 ha
Mobile
Sows: 3225 sows = 750 LU 9430 t manure
decanter
16100 t manure 85 t N(18 kg N/ha)
centrifuge
72 t N 75 t P(20 kg P/ha)
24 t P
Farm limit
Field: 388 ha
13868 t manure
54 t N(140kg N/ha)
6 t P(16kg P/ha) Solid fraction, store
T manure TN TP Hygienic
Farm1 2258 18 138
Farm2 3586 33 28 treatment
Farm3 3586 33 28
Sum 9430 85 75
Farm unit 2 Farm unit 3
41925 fattening pigs = 1400 LU Mobile 41925 fattening pigs = 1400 LU Mobile
26000 t manure decanter 26000 t manure decanter
134 tN centrifuge 134 tN centrifuge
38 tP 38 tP
Field: 716 ha Field: 716 ha
22000 t manure 22000 t manure
100 t N(140kg N/ha) 100 t N(140kg N/ha)
9 t P(13kg P/ha) 9 t P(13kg P/ha)
Figure 1. Scenario: low-technology slurry separation.
283
Separation of slurry entails a number of advantages, depending on the type of separation tech-
nology, including lower costs of storage, transport, and application of slurry, as well as better
utilization of the nutrients in the field. A typical transport scenario involves: (1) that the liquid
fraction from the decanting separation is applied following the N-norm, (2) that a certain
amount of the separated solid fraction is applied following the P-norm within the harmoniza-
tion area supplemental to the liquid fraction, and (3) that the surplus solid fraction is trans-
ported out from the farm to a secondary recipient location for application. Figure 2 shows the
transport distances for decanting fractions as a function of the distance to “manure-free” areas
receptive to manure application and as a function of varying harmonization demands in terms
of LU ha−1.
For smaller distances to secondary localities, the average transport distance is only dependent
to a small degree on the harmonization demands. The dependency grows when this distance is
increased. The transport distance/work is reduced by 61–78% as compared with the baseline
scenario with no separation. The modeling suite has the advantage of coping with numerous
different farm scenarios and application scenarios. In addition, the results from using the
modeling suite may be further elaborated and serve as input for the estimation of economic
indicators.
18 80
16 78
Weighed average transport distance, km
76
Reduction in transport efforts, %
14
74
12
72
10
70
8
68
6
66
4
64
2 62
0 60
10 50 100
Distance to secondar recipient locality, km
Figure 2. Distance for total transport of manure for application on primary farm land, trans-
port to secondary recipient land, and subsequently application. Default values: 1000 LU,
equaling 20 000 t of slurry per year. The dosage for the liquid fraction and the solid fraction
from the separation is 180 kg N per ha (utilized 145 kg N per ha) equaling 37 t ha−1 and 21 kg
P ha−1 equaling 3 t ha−1. The reduction of the transport distance is related to a baseline sce-
nario involving no separation and application of raw slurry with a dosage of 21 kg P ha−1
equaling 14 t ha−1. The shape of the adjoining land is assumed to be a quarter circle with an
, 2.5 LU/ha; ●, 1.4 LU/ha; ○, 1.7 LU/ha; ▼, 2.0 LU/ha; ∇, 2.5 LU/ha.
area corresponding to the respective harmonization area: , 1.4 LU/ha; , 1.7 LU/ha;
, 2.0 LU/ha;
284
LCA of Danish milk -system expansion in practice
Randi Dalgaard and Niels Halberg
Danish Institute of Agricultural Sciences, Department of Agroecology, Research Centre Foulum, P.O. Box
50, DK-8830 Tjele. Tel: +45 8999 1254 - Fax: +45 8999 1200. E-mail: [email protected]
Background
Danish mixed dairy farms do not only produce milk but also several co-products (e.g. meat, bread
wheat, manure, rape seed). When estimating the emissions related to milk production it is difficult
to decide to what extent the emissions are related to milk and to what extent they are related to the
co-products. For example the methane emission from the cows enteric fermentation is both related
to the milk and beef production. In most LCA-studies on milk, co-products allocation has been used
as a solution to the problem. Co-products allocation implies partitioning of the emissions between
the products and the partitioning factor has mostly been based on economic values or mass. Co-
products allocation is not recommended as best practice by International Standard Organization
(ISO 14041,1998), and can be avoided by use of system expansion, as shown in this study.
Objectives
To make LCA of milk produced on a Danish farm and accounting for co-products by the use of sys-
tem expansion
Methods
A typology of 31 farm models representing the Danish agricultural sector was established. Inputs
(e.g. electricity, soybean meal, fertiliser) to each farm type and emissions were estimated from ac-
count data and modelling as described by Dalgaard et al. (2004). Co-products were accounted for
by use of system expansion. Milk was considered as the main product on farm types producing milk
and the co-products (e.g. beef, sugar beet, bread wheat) were assumed to substitute products on
other farms, and thereby contributing with a negative amount of emissions. Data describing re-
source use and emissions per functional unit of fertiliser, soybean meal etc. was derived from Niel-
sen et al. (2003).
Results
The processes affected by the production of 1 kg of Danish milk from farm gate are shown in figure
1. These processes can be divided into three main categories: 1) Processes before the farm (e.g.
soybean cultivation, grain feed production) 2) Processes on farm (e.g. fodder production, manure
handling) 3) Processes after the farm (e.g. avoided production of meat, sugar beet, fertiliser).
The dairy farm type shown in figure 1 has an agricultural area of 48 hectares with grass-clover in
crop rotation on 19% of the land. The farm type produces 538 tons milk per year and the average
milk yield per cow is 7100 kg milk. It has a high stocking density (2.7 LSU per hectare.)
285
Production of soybean meal imported to farm
When producing soybean meal there is a co-production of soy oil. Soy oil substitutes rapeseed oil
on the world market and it can be assumed less rapeseed will be produced.
Thus a demand for soybean meal increases soybean production and decreases rapeseed production
(Weidema, 1999). Soybeans are capable of N-fixation and rapeseeds are not, and therefore increas-
ing soybean meal production lead to a decrease in fertiliser use.
Production of milk on farm
In the system the dairy farm is the major contributor to global warming (1061 g CO2 equivalents per
kg milk). Methane from cows, nitrous oxides from crop residues and manure handling are the most
important emissions of greenhouse gasses.
Avoided production of meat
For each kg of milk 44 g of cattle meat is produced. The more expensive parts (tenderloin etc.) re-
place beef and the cheaper parts replace pork. The avoided production of meat contributes with –
533 g of CO2 eqv. per kg of milk.
Avoided production of fertiliser
According to Danish legislation a dairy farm with a stocking rate higher than 2.3 LSU per hectare is
obliged to sell part of the manure. A part of the Nitrogen and Phosphorus in the manure will replace
artificial fertiliser used on cash crop farms. This will again reduce greenhouse gas (GHG) emissions
from fertiliser production, which is then deducted from the GHG of the Dairy farm. But the saved
GHG emissions from the avoided production of fertiliser is very small, and only contribute with -2
g CO2 per kg milk.
The farm data used above represents one of 8 farm types divided by soil groups and stocking den-
sity to be found in the LCA-food database (Nielsen et al., 2003). Figure 2 shows GHG emissions
from four conventional dairy farm types and the part of emissions related to the use of concentrate
feed in the form of soybean meal.
Discussion and conclusion
Manure sold from a farm will replace artificial fertiliser as shown in figure 1. Thereby a farm that
sells manure gives rise to a decrease in GHG emission from fertiliser production, because it is as-
sumed that the manure-receiving farm will import less fertiliser. But the ammonia loss from manure
is much higher than ammonia loss from fertiliser, which means that the manure-producing farm
contributes to an increase in ammonia emission on the manure-receiving farm. This emission is the-
refore included in the environmental impact from milk production in our study due to the use of sy-
stems expansion rather than allocation. Two of the dairy types (farm type 6 and 18) in this study sell
manure and emissions connected to sold manure is counterbalanced in the LCA database (Nielsen et
286
sen et al., 2003) used for the analysis. However the counterbalancing has very little impact at the re-
sults.
The study has showed that system expansion in agricultural production systems is possible in prac-
tice. LCA of milk also showed that the dairy farm is the major contributer to global warming. The
results of this method for different livestock products in a number of impact categories may be
found on the open LCA database (Nielsen et al., 2003).
A more detailed analysis of the contributions of on-farm processes and use of external inputs is
needed to explain the differences between the farming systems. Moreover, the present data structure
does not allow to estimate the uncertainty or variation on the emission estimates. Sensitivity analy-
sis should be carried out in order to quantify how important the differences between the farm types
are compared to differences in farm practise within the types.
References
Dalgaard, R., Halberg, N., Kristensen, I.S. and Larsen, I. 2004. An LC inventory based on represen-
tative and coherent farm types. In: Halberg , N. (ed.) Life Cycle Assessment in the Agri-food
sector. Proceedings of 4th International Conference. DIAS report. Danish Institute of Agricul-
tural Science. Denmark (this volume).
ISO 14041. 1998. Environmental management -Life cycle assessment –Goal and scope definition
and inventory analysis. International standard. Reference number: ISO 14041: 1998(E).
Nielsen, P.H., Nielsen, A.M., Weidema , B.P., Dalgaard, R. and Halberg N., 2003. LCA food data-
base. Available online, 31/8 2003 at www.lcafood.dk .
Weidema, B. 1999. System expansions to handle co-products of renewable materials. Pp. 45-48 in
Presentation Summaries of the 7th LCA Case Studies Symposium. Brussels: SETAC-Europe.
287
1 kg m ilk
farm gate
M eat
M eat
44 g
44 g
S oy bean
m eal
220 g
S ugar
S ugar
beet
beet
83 g
83 g
F eed
grain
390 g
B read
B read
wheat
wheat
16 g
16 g
E lectricity
and
D iesel
F ertilizer
M anure
6 g
4 g N
(N +P +K )
P roc esses before farm P roc esses on farm C o-products from farm A voided production
(470 g C O2 eqv.) (1061 g C O2 eqv.) ( - 533 g C O2 eqv.)
-
Figure 1. Processes affected when producing 1 kg of Danish milk on sandy soil. Data derived from
the LCA-food database (farm type 18).
1,2
0,8
0,6
0,4
0,2
0
Farm type 4 Farm type 5 Farm type 16 Farm type 17
from other processes from soybean meal
Figure 2. GHG emissions (kg CO2 eqv.) per kg milk produced at six different farm types. Farm ty-
pe 4: Loamy soil, 0.9 LSU per hectare. Farm type 5: Loamy soil, 1.7 LSU per hectare. Farm type
16: Sandy soil, 1.1 LSU per hectare. Farm type 17: Sandy soil, 1.8 LSU per hectare.
288
LCC and LCA of extra-virgin olive oil: organic vs. conventional
Notarnicola B.*, Tassielli G. *, Nicoletti G.M.**
*Dipartimento di Scienze Geografiche e Merceologiche, Facoltà di Economia, Università di Bari, via C. Ro-
salba, 53, 70124 Bari, Italy.
Tel: 0039 080 5049084; fax: 0039 080 5049019; e-mail: [email protected]
**
Facoltà di Economia, Università di Foggia, via IV Novembre 1, 71100, Foggia, Italy.
Abstract
Olive oil represents a relevant productive sector in Puglia, a region of the South of Italy, since it
covers more than the 50% of the whole Italian output and about the 18% of the EU output. In the
last years, the production of organic extra virgin olive oil is highly increasing due to a new con-
sumer behaviour and to the high organoleptic, nutritional and healthiness quality of this product.
But organic extra-virgin olive oil still remains a niche product because of its market price remarka-
bly higher than the other oil and fats. In this paper the production systems of the conventional and
organic extra-virgin olive oil have been compared, in order to assess their environmental and cost
profiles, and to verify if the two dimensions – environmental and economical – converge in the
same direction.
Keywords: LCC, LCA, extra-virgin olive oil, organic agriculture
1 Introduction
In the last ten years the Italian agricultural area cultivated by following organic practices is re-
markably increased, going from a 0.5% of the total area in the 1993 to the 8% in the 2001. The total
Italian “organic” area is about 1,200,000 ha, about the 15% more than the previous year, featuring
about 60,000 farms. The “organic” area is shared among fodder (56%), wheat (18%), arboreal culti-
vations (19%) and vegetables (6%) (Compagnoni et al., 2001). Among the area cultivated by arbo-
reals (about 228,000 ha) olive tree is particularly important, since it covers 44,175 ha. The olive oil
production of the Southern Italian region Apulia covers about the 50% of the total Italian produc-
tion and about the 18% of the EU production. In the light of these data and of the importance of the
Apulian oliviculture and olive oil production (1182 olive oil factories on the Italian total of 5514), a
high rate of growth for the organic oliviculture can be forecasted in this region. Apart the producer
countries, olive oil price remains remarkably higher than other oils and fats even if it is character-
ised by a clear better environmental performance due to the lack of chemical treatments (Nicoletti et
al., 2001a). This situation, which is exactly the contrary of what an environmentally friendly policy
should do – to support on the market place environmentally friendly products –, becomes much mo-
re marked in the case of the organic oil in which, together with the typical olive oil high price,
basically due to the cost of labour in the extremely delicate operation of olives harvesting, one has
to add the additional costs due to the minor yields (about 30%) of the organic soil. Therefore, even
more than the conventional one, organic extra-virgin olive oil is destined to be a niche product. But
we think that a superior product from the health, nutritional and environmental point of view should
289
be affordable by everyone. In this study the results of a Life Cycle Costing (LCC) and the first re-
sults of a more comprehensive comparative Life Cycle Assessment (LCA) of organic and conven-
tional extra-virgin olive oil are shown in order to identify the relative economical and environ-
mental scores and to verify if the dimensions converge or not in the same direction.
2 Methods
The methodologies which have been used are: the LCC as suggested by the guidelines stated by
White (White et al., 1996) which divide the costs into three categories, conventional company costs
(typical costs which appear in the company accounts), less tangible, hidden and indirect costs (less
measurable and quantifiable, often obscured by placement in an overheads account) and external
costs (the costs which are not paid by the polluter, but by the polluted); the LCA as stated by the
ISO 14040 rules (ISO, 1996). The functional unit is 1 kg extra-virgin olive oil and the analysis is
from cradle-to-gate. The physical and economical data have been taken directly from farms, olive
oil factories and by databases. The internal and external costs are respectively shown in Table 1 and
2. The external costs relative to the energy have been taken from the ExternE National Implementa-
tion Italian Report (ExternE, 1997), while those relative to the use of pesticides and fertilisers from
a study of the Bocconi, Milan, Italy in which the production and social costs of the organic and con-
ventional agriculture have been compared. The study has taken into account the impact of the
agricultural activities on the water and has monetarised these impact showing that the damage
caused by the conventional agriculture due to fertilisers and pesticides in terms of reclamation and
decontamination costs is 33 times superior to that of the organic agriculture.
3 Results and conclusions
As other studies on the organic systems (Nicoletti et al., 2001 b), the results of the LCA (Fig. 1)
show that the organic system scores worse than the conventional one in all the impact categories
with the exception of the NP, HTP, TETP and FAETP (the impact assessment method used is the
CML 2000 as stated in Guinèe et al., 2002, with the exception of the ADP category which is substi-
tuted by the EC, Energy Content, and with the addition of LU, Land Use). The minor yield (about
the 30%) of the organic system is the reason of this result. By going through the evaluation step
(Fig. 2) it can be found out that the organic system results more eco-compatible than the conven-
tional of about 5 times due to the relevant difference in the TETP and FAETP impact categories.
Figure 2 shows the differences in the results if one accounts for the external costs. If one does not
consider the external costs, the organic oil has a superior cost profile which is basically due to its
minor agricultural yields (about the 30% less). On the contrary, by adding the external costs, which
are not actually paid by the farmer and by the olive oil companies, to the conventional company and
to the less tangible, hidden and indirect company costs, it can be found out that the organic oil has a
minor total cost compared to the conventional oil. This result enlightens the need to account for ex-
ternal costs as the European Commission has started to do (Labouze E. et al., 2003). The options for
environmental improvement in the conventional system are mainly related to a more reasonable use
of pesticides while, in the case of the organic, a reuse of the brushwood as fuel rather than their un-
290
controlled burning on the fields, could lead to a better environmental profile both in the HTP and in
the POCP. Moreover, in the organic system the “traditional” extraction method has been used in the
inventory set-up; on the contrary, the AIAB guidelines (AIAB, 2001) permits to the organic oil pro-
ducer to use the “continuous-extraction method” which is characterised by an energy consumption
double than the traditional process. It would be desirable a major attention of these guidelines to the
energy consumption, since the consumer who is interested in organic foods would like to buy a
more eco-compatible product, which is characterised not only by the absence of the use of fertilisers
and pesticides but also by the least energy consumption. On the cost side, Fig. 2 has shown the im-
portance in taking into account the external costs. Since a minor cost of the organic olive oil com-
pared to the conventional one is not obtainable on the market place just with the “market laws”, it is
necessary to promote tools of public intervention which could end up in reducing the gap between
the cost of the conventional oil calculated by the traditional cost accounting methods and those cal-
culated by following the LCC approach. The aim should be that on the basis of the same quality
standards, products with a better environmental profile should have a minor market price compared
to the concurrent; exactly the contrary of the present situation in which the most eco-compatible
products have higher market prices.
References
AIAB (Organic Agriculture Italian Association), 2001. Guidelines for the production of extra-virgin
olive oil, (in Italian). www.aiab.it
Compagnoni A., Pinton R., Zanoli R., 2001. Organic farming in Italy 2001. www.organic-
europe.net
ExternE National Implementation Italy. Final Report. October 1997. http://externe.jrc.es/it.pdf
Guinèe J.B., Gorèe M., Heijungs R., Huppes G., Kleijn R., de Koning A., van Oers L., Wegener
Sleeswijk A., Suh S., Udo de Haes H.A., de Bruijn H., van Duin R., Huijbregts M.A.J., 2002.
Handbook on Life Cycle Assessment, Operational Guide to ISO Standards, Dordrecht, Klu-
wer Academic Publishers.
International Organization for Standardization ISO/DIS 14040, 1996. Environmental management -
Life cycle assessment - Principles and framework, ISO/TC207/SC5.
Labouze E., Monier Veronique, Le guern Y., 2003. Study on external environmental effects related
to the life cycle of products and services. European Commission, Directorate General Envi-
ronment.
Nicoletti G.M., Notarnicola B., Tassielli G., 2001a. Comparative LCA of virgin olive oil vs. seed
oil, Proceedings of the International Conference LCA in Foods, 26-27 April, 2001, Gothen-
burg, Sweden.
Nicoletti G.M., Notarnicola B., Tassielli G., 2001b. Comparison of conventional and organic wine,
Proceedings of the International Conference LCA in Foods, 26-27 April, 2001, Gothenburg,
Sweden.
White A.L., Savage D., Shapiro K., 1996. Life Cycle Costing: concept and applications. In Curran
M.A. (ed.). Environmental Life Cycle Assessment, Mc Graw Hill.
291
Table 1. Internal costs of the two systems for functional unit (€).
Agricultural phase Organic Conventional
Pesticides 0.171 0.117
Fertilisers 0.268 0.181
Lube oil 0.023 0.011
Electric energy 0.143 0.085
Water 0.077 0.046
Diesel 0.084 0.048
Labour 4.344 2.864
Organic certification costs 0.064 -
Total 5.174 3.352
Transports 0.078 0.039
Industrial phase
Electric energy 0.014 0.024
Labour 0.089 0.045
Water 0.002 0.022
Packaging 0.298 0.298
Waste authority 0.015 0.015
Organic certification costs 0.009 -
HACCP certification costs 0.0009 0.0009
Total 0.428 0.405
Total 5.680 3.796
Table 2. External costs of the two systems for functional unit (€).
Organic Conventional
External costs of energy 0.664 0.533
External costs of fertilisers and pesticides 0.439 9.870
292
100%
90%
80%
70%
60%
organic olive oil
50%
conventional olive oil
40%
30%
20%
10%
0%
EC GWP ODP AP POCP HTP FAETP MAETP TETP NP LU
Figure 1. LCA results.
100%
90%
80%
70% Organic
60%
Conventional
50%
40%
30%
20%
10%
0%
LCA LCC w ithout external costs LCC w ith external costs
Figure 2. LCA-LCC without external costs and with external costs.
293
You might also like
- Ruviaro - Life Cycle Assessment in Brazilian Agriculture Facing Worldwide Trends (Journal-of-Cleaner-Production)No ratings yetRuviaro - Life Cycle Assessment in Brazilian Agriculture Facing Worldwide Trends (Journal-of-Cleaner-Production)16 pages
- Download Full (Ebook) Sustainable Food Systems from Agriculture to Industry: Improving Production and Processing by Charis Michel Galanakis ISBN 9780128119358, 9783213303311, 0128119357, 3213303313 PDF All Chapters100% (12)Download Full (Ebook) Sustainable Food Systems from Agriculture to Industry: Improving Production and Processing by Charis Michel Galanakis ISBN 9780128119358, 9783213303311, 0128119357, 3213303313 PDF All Chapters65 pages
- Simplified Life Cycle Assessment of Galician Milk Production: A. Hospido, M.T. Moreira, G. FeijooNo ratings yetSimplified Life Cycle Assessment of Galician Milk Production: A. Hospido, M.T. Moreira, G. Feijoo14 pages
- Comparative Analysis of Environmental Impacts of ANo ratings yetComparative Analysis of Environmental Impacts of A12 pages
- Life Cycle Assessment of Food: International Agricultural Engineering Journal November 2013No ratings yetLife Cycle Assessment of Food: International Agricultural Engineering Journal November 20137 pages
- Clark 2017 Environ. Res. Lett. 12 064016No ratings yetClark 2017 Environ. Res. Lett. 12 06401612 pages
- Systematic review of greenhouse gas emissions for different fresh food categoriesNo ratings yetSystematic review of greenhouse gas emissions for different fresh food categories32 pages
- PDF Sustainable Food Systems from Agriculture to Industry Improving Production and Processing Galanakis download100% (4)PDF Sustainable Food Systems from Agriculture to Industry Improving Production and Processing Galanakis download41 pages
- Sustainable Food Systems from Agriculture to Industry Improving Production and Processing Galanakis - The latest ebook version is now available for instant access100% (1)Sustainable Food Systems from Agriculture to Industry Improving Production and Processing Galanakis - The latest ebook version is now available for instant access60 pages
- ROY ET AL 2009 A Review of Life Cycle Assessment (LCA) On Some Food ProductsNo ratings yetROY ET AL 2009 A Review of Life Cycle Assessment (LCA) On Some Food Products10 pages
- Reducing Food 'S Environmental Impacts Through Producers and ConsumersNo ratings yetReducing Food 'S Environmental Impacts Through Producers and Consumers7 pages
- 1 - Reducing Food's Environmental Impacts Through Producers and ConsumersNo ratings yet1 - Reducing Food's Environmental Impacts Through Producers and Consumers7 pages
- Environmental Impact Assessment of Agricultural Production Sy 2004 EuropeanNo ratings yetEnvironmental Impact Assessment of Agricultural Production Sy 2004 European18 pages
- The Environmental Impacts of Food Production and ConsumptionNo ratings yetThe Environmental Impacts of Food Production and Consumption9 pages
- Clune_2017_Systematic review of greenhouse gas emissions for different fresh food categoriesNo ratings yetClune_2017_Systematic review of greenhouse gas emissions for different fresh food categories18 pages
- Instant Download Sustainable Food Systems From Agriculture To Industry Improving Production and Processing 1st Edition Charis Michel Galanakis PDF All Chapters100% (3)Instant Download Sustainable Food Systems From Agriculture To Industry Improving Production and Processing 1st Edition Charis Michel Galanakis PDF All Chapters39 pages
- Journal of Cleaner Production: E. Crenna, T. Sinkko, S. SalaNo ratings yetJournal of Cleaner Production: E. Crenna, T. Sinkko, S. Sala14 pages
- An Evaluation of Environmental SustainabNo ratings yetAn Evaluation of Environmental Sustainab10 pages
- Potential of Life Cycle Assessment To Support Environmental Decision Making at Commercial Dairy FarmsNo ratings yetPotential of Life Cycle Assessment To Support Environmental Decision Making at Commercial Dairy Farms11 pages
- Submission 6.1. Mr Andrew Benskin (14.7.2021)No ratings yetSubmission 6.1. Mr Andrew Benskin (14.7.2021)159 pages
- life-cycle-assessment-on-agricultural-production-a-mini-1908rjaiNo ratings yetlife-cycle-assessment-on-agricultural-production-a-mini-1908rjai16 pages
- Sustainable Food Production and ConsumptionNo ratings yetSustainable Food Production and Consumption270 pages
- Clark 2017 Environ. Res. Lett. 12 064016No ratings yetClark 2017 Environ. Res. Lett. 12 06401612 pages
- Sustainable Food Systems from Agriculture to Industry Improving Production and Processing 1st Edition Charis Michel Galanakis - The ebook is ready for instant download and accessNo ratings yetSustainable Food Systems from Agriculture to Industry Improving Production and Processing 1st Edition Charis Michel Galanakis - The ebook is ready for instant download and access65 pages
- Download Full Food Systems Modelling: Tools for Assessing Sustainability in Food and Agriculture Christian J. Peters PDF All Chapters100% (3)Download Full Food Systems Modelling: Tools for Assessing Sustainability in Food and Agriculture Christian J. Peters PDF All Chapters50 pages
- Carbon Footprint of Food e Approaches From National Inputeoutput Statistics and A LCA of A Food PortionNo ratings yetCarbon Footprint of Food e Approaches From National Inputeoutput Statistics and A LCA of A Food Portion8 pages
- Immediate download (eBook PDF) Encyclopedia of Food Security and Sustainability ebooks 2024100% (1)Immediate download (eBook PDF) Encyclopedia of Food Security and Sustainability ebooks 202450 pages
- Full (Ebook PDF) Encyclopedia of Food Security and Sustainability PDF All Chapters100% (5)Full (Ebook PDF) Encyclopedia of Food Security and Sustainability PDF All Chapters40 pages
- Life Cycle Assessment of Indoor Vertical Farms: Burleigh Dodds Series in Agricultural ScienceNo ratings yetLife Cycle Assessment of Indoor Vertical Farms: Burleigh Dodds Series in Agricultural Science23 pages
- Inorganic and Organic Components of SoilNo ratings yetInorganic and Organic Components of Soil33 pages
- Life Cycle Assessment Research ProposalNo ratings yetLife Cycle Assessment Research Proposal14 pages
- Water Research & Technology: Environmental ScienceNo ratings yetWater Research & Technology: Environmental Science17 pages
- Agroforestry Production For Energy Production in RwandaNo ratings yetAgroforestry Production For Energy Production in Rwanda11 pages
- Peanut Production: Bima, West Nusa Tenggara IndonesiaNo ratings yetPeanut Production: Bima, West Nusa Tenggara Indonesia10 pages
- Aphids as Crop Pests 2nd Edition Harrington download100% (1)Aphids as Crop Pests 2nd Edition Harrington download63 pages
- The War On Nutrition - Rationing in America and Britain During World War II0% (1)The War On Nutrition - Rationing in America and Britain During World War II44 pages
- Post-harvest technology and food processing innovations: Prospects and challenges in BangladeshNo ratings yetPost-harvest technology and food processing innovations: Prospects and challenges in Bangladesh11 pages
- A Review On The Growth, Yield and Oil Contents of Brassica Under Rainfed ConditionsNo ratings yetA Review On The Growth, Yield and Oil Contents of Brassica Under Rainfed Conditions7 pages
- Analisis Kadar Mineral Dalam Abu Buah Nipa (Nypa Fructicans) Kaliwanggu Teluk Kendari Sulawesi TenggaraNo ratings yetAnalisis Kadar Mineral Dalam Abu Buah Nipa (Nypa Fructicans) Kaliwanggu Teluk Kendari Sulawesi Tenggara8 pages
- Ge Elec 1 Prelim Reviewer: Module 1: Introduction To Ecosystem ECOSYSTEM First Appeared in ANo ratings yetGe Elec 1 Prelim Reviewer: Module 1: Introduction To Ecosystem ECOSYSTEM First Appeared in A11 pages
- Procedia Cirp: Niken Kusumowardani, Benny Tjahjono, ConceptualisationNo ratings yetProcedia Cirp: Niken Kusumowardani, Benny Tjahjono, Conceptualisation6 pages
- 4 - Siri Institutional Area, Hauz Khas, New Delhi 110016, Tel - No. 26510314 Fax: 011-26962370 E.mailNo ratings yet4 - Siri Institutional Area, Hauz Khas, New Delhi 110016, Tel - No. 26510314 Fax: 011-26962370 E.mail24 pages
- Maths STD 4 Annual Revisio Worksheet 2023 PDFNo ratings yetMaths STD 4 Annual Revisio Worksheet 2023 PDF8 pages
- Project Report On Production of Menthol Crystal (Flakes), Menthol Crystal (Bold) Dmo & Mint Oil StorageNo ratings yetProject Report On Production of Menthol Crystal (Flakes), Menthol Crystal (Bold) Dmo & Mint Oil Storage6 pages
- Introduction to Animal Production Course Information_ Revised_January 2025No ratings yetIntroduction to Animal Production Course Information_ Revised_January 202512 pages
