Sustainable Production and Consumption 46 (2024) 442–459

Contents lists available at ScienceDirect

Sustainable Production and Consumption

journal homepage: www.elsevier.com/locate/spc

Life cycle inventory and assessment of Canadian faba bean and dry
bean production
Nicole Bamber *, Vivek Arulnathan, Lana Puddu, Alyssa Smart, Jannatul Ferdous,
Nathan Pelletier
Food Systems PRISM Lab, University of British Columbia, Okanagan Campus, 226-3247 University Way, Kelowna, BC V1V1V7, Canada

A R T I C L E I N F O A B S T R A C T

Editor: Dr. Cecile Chéron-Bessou Pulses are nutritionally beneficial and cost-effective food/feed products, and are also nitrogen-fixing crops that
can improve the environmental sustainability of annual cropping systems when included in crop rotations –
particularly with respect to GHG emissions. Canada is a major producer and exporter of pulses, therefore the
goals of this study were to (a) develop regionalized life cycle inventories (LCIs) for Canadian faba and dry beans
suitable for incorporation into public LCI databases and (b) conduct ISO 14044-compliant LCAs of Canadian faba,
navy, pinto, and kidney beans at provincial, ecozone, and national scales. LCI data were collected via survey from
~200 Canadian faba and dry bean farmers.
At the national scale of aggregation, faba beans had considerably lower impacts across all impact categories
considered than the three dry bean types. This is due to their lower input requirements (particularly fertilizer),
and higher nitrogen fixation capabilities. Among the three dry bean types considered, pinto beans had the lowest
environmental impacts across most impact categories. Fertilizer production and field-level emissions from fer­
tilizer use were the major contributors to the overall environmental impacts of the production of all bean types.
In addition to differences between bean types, there were also differences within bean types, between regions of
production and levels of regional aggregation. The results of the uncertainty analysis revealed that variability
was lowest when aggregated at the ecozone level, based on soil and climate factors, and highest when aggregated
to the national average. Taken together, these results demonstrate the utility of spatially-resolved LCI and LCA
modelling that distinguishes between different types of pulse crops in order to accurately estimate impacts, and
to support decision-making in pursuit of more sustainable food systems. Future research needs include more
primary data on field-level biological nitrogen fixation and soil carbon fluxes, since these can significantly in­
fluence estimated environmental impacts.

1. Introduction pulses with a value of $4.29 billion CAD (GTT, 2022). The Prairie
Provinces are the major pulse producers in Canada, accounting for about
Canada is one of the largest pulse producers in the world, as well as 96 % of Canadian pulse production (Statistics Canada, 2015).
the largest exporter (Pulse Canada, 2021). Pulse production in North In addition to the economic importance of pulses, there are envi­
America, particularly in Canada, has increased substantially in recent ronmental impacts and benefits associated with their production. In
decades (Statistics Canada, 2023). In 2020, pulse field crop area was just order to assess the environmental aspects of the production of pulses,
over 3.5 million hectares, representing 11.8 % of field crop area in including those associated with the supply chains of the field inputs (e.
Canada (AAFC, 2022), and Canada exported 7.46 million tonnes of g., fertilizers, fuels, etc.), and credits for biological nitrogen fixation

Abbreviations: APOS, at point of substitution; BNF, biological nitrogen fixation; C, carbon; CML, Center of Environmental Science, Leiden University; CO2, carbon
dioxide; DQI, data quality indicator; EPD, Environmental Product Declaration; GHG, greenhouse gas; ICMB, Introductory Carbon Balance Model; IPCC, International
Panel on Climate Change; ISO, International Organisation for Standardisation; LCA, life cycle assessment; LCI, life cycle inventory; LCIA, life cycle impact assessment;
N, nitrogen; N2O, nitrous oxide; NH3, ammonia; NO3, nitrate; NOx, nitrogen oxides; OMAFRA, Ontario Ministry of Agriculture, Food and Rural Affairs; P, phosphate;
SLC, Soil Landscapes of Canada; SOC, soil organic carbon.
* Corresponding author.
E-mail address: [email protected] (N. Bamber).

https://doi.org/10.1016/j.spc.2024.02.034
Received 29 November 2023; Received in revised form 23 February 2024; Accepted 27 February 2024
Available online 6 March 2024
2352-5509/© 2024 The Authors. Published by Elsevier Ltd on behalf of Institution of Chemical Engineers. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

(BNF), a systems-level environmental accounting tool called life cycle international literature (e.g., Köpke and Nemecek, 2010; Karlsson et al.,
assessment (LCA) can be used. LCA is a widely utilized, ISO standardized 2015; Vogelsang-o'Dwyer et al., 2020; Martin-gorriz et al., 2022), the
approach (ISO, 2006) to quantify resource inputs (e.g., energy, water) majority of which were in a European context, with LCI data sourced
and emissions (e.g., GHGs, air pollutants) across the entire life cycle of a from the literature, from experimental farms, or not reported. Bandekar
product or service (Guinée, 2001). LCA offers insights into the sustain­ et al. (2022) performed an LCA of pulse production and consumption in
ability of a product or service that can be used to improve product the United States, which included dry beans. They sourced their data
design, reduce resource use, inform stakeholders regarding the from a combination of literature, industry and government sources.
comparative environmental performance of goods or services, and However, in the Canadian context, there have been no LCAs of faba or
inform policy and management decisions on a multi-criteria basis. dry beans, which demonstrates a clear research gap. While regionalized,
There is substantial heterogeneity in the impacts of agricultural ISO 14044-compliant LCAs for peas and lentils produced in Canada have
production due to diverse climate, soil, and management conditions. In already been developed (Bamber et al., 2022), comparable assessments
order to understand the influences of regional parameters and local have not been reported for faba and dry bean production in Canada, nor
factors on the results of agricultural LCAs, recent studies (Nitschelm were sufficient data available to enable such studies. Thus, the goals of
et al., 2016; Yang et al., 2018) have highlighted the importance of this study were to (a) develop regionalized life cycle inventories (LCIs)
building regional life cycle inventories for agricultural products. for Canadian faba and dry beans suitable for incorporation into public
Regional resolution is essential to differentiating among agricultural LCI databases and (b) conduct ISO 14044-compliant LCAs of Canadian
products on the basis of environmental attributes, identifying appro­ faba, navy, pinto, and kidney beans at provincial, ecozone (based on soil
priate sustainability improvement opportunities, and enabling commu­ and climate zones), and national scales.
nication with respect to their sustainability benefits and impacts. In
order to enable accurate representation of Canadian pulse crops in the 3. Methods
context of life cycle-based sustainability measurement and management
activities, it is hence essential to develop high quality, regionalized data According to the ISO 14040 and 14,044 standards (ISO, 2006), an
inventories that are representative of context-specific production con­ LCA includes 4 steps: the goal and scope definition, the LCI analysis, the
ditions and practices, and to incorporate these inventories into publicly life cycle impact assessment (LCIA) and the life cycle interpretation. This
available life cycle inventory (LCI) databases. The LCI data presented in manuscript and associated supplementary file (see Electronic Supple­
this paper will enable future studies to make comparisons using mentary Material (ESM) 1 for a full LCA report) follow this format.
consistent methodology, rather than sourcing LCIA results from various
sources for a direct comparison, and hence support potential claims
regarding comparative environmental performance as well as related 3.1. Goal and scope
sustainability management decision making on a rigorous basis.
The goals of this study were:
2. Literature review
1. To develop and make publicly available ISO 14044-compliant,
The benefits of pulse cultivation, and the inclusion of pulses in crop regionalized (i.e., at ecozone, province, and national scale) LCI
rotations has been studied for some time. Pulses are considered to be datasets for navy bean, red kidney bean and pinto bean (referred to
both nutritionally beneficial (high protein and fibre, and low fat), and as dry beans), and faba bean production in the Canadian provinces of
cost- effective food products (Ojiewo et al., 2015). They are also Alberta, Saskatchewan, Manitoba, and Ontario, and terrestrial eco­
nitrogen-fixing crops that can improve the environmental sustainability zones 8 (Mixedwoods Plains), 9 (Boreal Plains), and 10 (Prairies).
of annual cropping systems when included in crop rotations, in partic­ 2. To perform ISO 14044-compliant LCAs in order to quantify the
ular with respect to reducing GHG emissions. Indeed, incorporation of resource and environmental impacts of dry bean and faba bean
pulses in cereal-based cropping systems has been found to have several production in each region, using regionally-resolved, temporally and
agronomic and environmental benefits. These include influencing soil technologically current data.
organic carbon (SOC) levels due to the amount of carbon (C) returned to
the soil in crop residue (Campbell et al., 2000), and having lower
pesticide and fertilizer requirements than cereals, which directly im­ 3.1.1. Intended applications
pacts greenhouse gas (GHG) and other emissions related to the manu­ The life cycle inventories will be made publicly available, and
facture and application of these inputs (Lemke et al., 2007). The transparently reported, to enable food industry, academia, government,
inclusion of pulse crops in rotation with grain crops also has economic and other stakeholders to conduct life cycle assessments of ingredients,
benefits due to pulses' ability to reduce the synthetic nitrogen fertilizer foods and diets that contain faba or dry beans produced in Canada using
requirements of the rotation and overall improvements in rotational high quality, regionally representative data. The presentation of the LCI
yield and quality of the grain (MacWilliam et al., 2014). Hence, there is a data will enable future comparative studies to be performed trans­
growing desire from industry, academia and governments to better parently and with methodological consistency, according to ISO stan­
quantify the environmental benefits of whole pulses and pulse in­ dards, rather than simply sourcing LCIA results.
gredients in food systems in order to support positioning pulses in the The life cycle assessments will enable establishing regionally-
context of sustainable dietary transition discourse. However, not all resolved industry benchmarks for the environmental performance of
pulses provide the same levels of BNF, and their overall, comparative faba and dry bean production in the Canadian Prairie Provinces and
environmental advantages or disadvantages are unclear. For example, Ontario. They will also support identifying regionally-resolved envi­
Hossain et al. (2016) assessed the BNF capacity of Canadian chickpeas, ronmental hotspots and potential areas for improvement for Canadian
faba beans, dry beans, lentils and field peas and found large variability faba and dry bean production. Finally, the results of the LCA studies may
between species, as well as due to regional differences in precipitation. be applied for environmental marketing communications such as Envi­
In general, they found that of all the pulse types studied, faba beans ronmental Product Declarations (EPD) for faba and dry beans produced
provided the most BNF, and dry beans provided the least. Therefore, it is in these regions. Since the LCA studies may be used to make comparative
important to assess each pulse type individually, rather than to group assertions (defined by ISO (2006) as an environmental claim asserting
them all together, which is commonly done when comparing against superiority of one product over a competing product) between different
other protein sources, as well as to assess impacts on a regional basis. products, the study methods comply with the additional requirements of
There have been several LCAs of faba and/or dry beans in the ISO 14044 in this respect.

443
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

3.1.2. Functional unit capital goods (i.e., on-farm equipment) make only minor contribu­
Both the functional unit and reference flow were defined as 1 t of tions to life cycle environmental impacts (MacWilliam et al., 2014,
dried, marketable beans at the farm gate. Dry beans are dried to 15 % 2018), and that impacts are highly sensitive to the assumed lifespan
and faba beans to 16 % (Canadian Grain Commission, 2016; Saskatch­ and use of the equipment.
ewan Pulse Growers, 2018). This is consistent with the functional units Storage (after drying) was excluded in the system boundary since a
used in previously conducted LCA and carbon footprint studies of pulses previous pulse LCA carried out in Alberta (Alberta Agriculture and
in Canada (Alberta Agriculture and Forestry, 2018; EarthShift Global, Forestry, 2018) indicated that grain drying and storage was a minor
2019; Bamber et al., 2022). contributor to the overall environmental impacts. Hence, storage is
not included, also based on the advice from expert stakeholder
3.1.3. System boundaries groups (pulse grower associations and Pulse Canada). The system
The overall scope of this attributional LCA was cradle-to-farm gate, boundary hence only includes processes up to and including the
and included the following processes (Fig. 1): production of the dried pulse product.

Field operations included crop rotation, tillage, harrowing, seeding, 3.1.4. Geographical, temporal, and technological boundaries
land rolling, fertilizer application, pesticide and other agri-chemical This study was designed to represent contemporary faba bean, navy
applications, swathing, irrigation (when relevant), and harvesting. bean, red kidney bean and pinto bean production in the Canadian
Farm inputs/outputs included inputs of seed, fertilizer, plant pro­ provinces of Alberta, Saskatchewan, Manitoba, and Ontario, aggregated
tection products, desiccants, inoculants, irrigation, energy carriers (i. at provincial, ecozone, and national scales. This aggregation was per­
e., fuels and electricity for farm operations), and outputs of har­ formed because ecozones share macro climatic factors such as flora and
vested/dried pulses and fixed nitrogen (N). fauna, geographical formations, precipitation, temperature, and soil
Transportation included the transport of seed, fertilizer, plant conditions. Provinces share regulations and recommendations regarding
protection products, desiccants, inoculants and fuel inputs to the farm inputs and farming methods from provincial growers' associations.
farm. Different levels of regional specificity are important to different con­
Field-level emissions of carbon dioxide (CO2), nitrous oxide (N2O) sumers – for example, within Canada the province of origin may be
(direct and indirect), ammonia (NH3), nitrate (NO3), nitrogen oxides important while, internationally, only the Canadian average impacts are
(NOx), phosphate (P), and losses of crop protection products (fun­ relevant since beans are intermingled and marketed as Canadian only.
gicides, herbicides, and insecticides) to air, soil and water. The temporal bounds for the foreground system spanned 2018 to 2020
Soil organic carbon sequestration, taking into account climate, soil, for crop yield data, but otherwise encompassed the 2020 field season for
and management practices. all other foreground system (i.e., field-level) data that were collected.
Nitrogen credit from BNF and crop residues was considered as an
additional function of pulse production in the foreground system 3.1.5. Allocation procedure
models, included using system expansion. The production of pulses includes no co-products, but does include
Post-harvest activities (not including storage) including the drying the additional function of BNF, which was included using a system
and aeration of the pulse crop. expansion approach. The service provided by BNF is the provision of N
to the next crop in rotation, meaning that that crop would have to use
The following processes were excluded from the system boundary: less N fertilizer than if the pulses had not been included in the rotation.
This was modelled as an N fertilizer credit. Specifically, a negative input
Capital goods such as farm infrastructure and machinery used on of ammonia was included in the inventory model for the amount of N
the farm. Previous LCAs of pulses in Western Canada have indicated fixed. Ammonia was chosen since it is the simplest form of N fertilizer,
that the production, distribution, storage, and disposal of on-farm from which all other N fertilizers are derived. For any allocation in the

Fig. 1. Flow diagram of faba and dry bean production system boundaries, from the inputs of resources from the ecosphere to the production of materials and
chemicals as inputs to agricultural processes producing crop yield and N fixation (modelled as a credit of N fertilizer), including all life cycle emissions to
the ecosphere.

444
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

background LCI datasets, pre-allocated processes in the ecoinvent “At Table 1

the point of substitution” (APOS) dataset were used, and when running LCI data sources for ICBM model in Holos.
impact assessment results, physical allocation was chosen to best Holos data requirement Source/method of calculation
represent the biophysical relationships between any upstream co-
Region (SLC polygon) SLC polygons with most occurrence of
products and associated flows (while working within the constraints of farms (from latitude and longitude
the pre-allocated ecoinvent background processes). coordinates given by farmers),
representing the most common soil types
3.1.6. Life cycle inventory in the ecozone/province
Climate data (precipitation, potential NIR data by ecodistrict averaged to
The majority of the foreground data came directly from a represen­ evapotranspiration, temperature) provincial and ecozone scales
tative sample of Canadian faba and dry bean growers. Data on field Total area of field Area per 1 t of beans produced (by
operations, irrigation, fertilizer and crop protection product application, province/ecozone)
yield, seeding rate, seeded area, electricity use and fuel use were Crop rotation Most common crop rotation(s) from
survey responses
collected from ~200 survey responses from growers in Alberta, Sas­
Yield Survey responses for past 3 years (by
katchewan, Manitoba, and Ontario (see Appendix A of ESM 1 for a copy province/ecozone)
of the survey questions). Farm-level emissions of carbon dioxide (CO2), Moisture content Saskatchewan Pulse Growers (2018) (used
direct and indirect nitrous oxide (N2O), and other reactive N species for all regions)
emissions (i.e. NH3, NOx, NO3) from fertilizer to air, leaching to ground Tillage (no till, reduced, or intensive) Survey responses for whether tillage was
performed, and machinery type classified
water and run off to surface water were calculated using the Interna­ as either reduced or intensive based on
tional Panel on Climate Change (IPCC) Tier II approach (Hergoualc'h expert opinion that moldboard ploughing
et al., 2019), with regionally specific emission factors and biomass to 30 cm (plus discing, harrowing etc.) is
partitioning factors from Thiagarajan et al. (2018), Rochette et al. considered intensive, and to only 15 cm is
reduced (by province/ecozone)
(2018), and Government of Canada (2021) (equations detailed in ESM1,
Irrigation Survey responses for whether irrigation
Section 2.1.12.9.1.1). Emissions of pesticides were modelled as per was performed, and volume of water
Margni et al. (2002), which estimates that 76.5 % of the active ingre­ applied when applicable (by province/
dient of each agricultural plant protection product is emitted to soil, 8.5 ecozone)
% to water, and 10 % to air. This means a total of 95 % of the applied Number of passes of pesticide Survey responses for number of passes of
application each type of pesticide application (by
active ingredient is emitted back to the environment. The last 5 % is
province/ecozone)
taken up by the plant and was therefore not considered as an emission in Fertilizer Survey responses for types of fertilizer
this model. Nitrogen credit values for nitrogen fixed by the plants were applied and amount (by province/
estimated according to the data provided by Barker (2007) for Top Crop ecozone), timing is spring based on expert
opinion (for all regions)
Manager.
% straw returned to soil Expert opinion (for all regions)
The ICBM model in Holos v.4 was chosen to model soil carbon dy­
namics since at the time of performing the calculations. This was the
model used in the Canadian National Inventory Report, and is more analysis. All secondary data used were modified, as described above, to
representative of individual farm dynamics than what can be achieved be as representative as possible of the production conditions in the
using default factors (Kröbel et al., 2016). The net soil carbon seques­ respective regions of Canada. Data quality scores were assigned to all life
tration or emissions were calculated per functional unit of 1 t of beans cycle inventory model data points by comparing them to the data quality
for the current crop rotations including faba and dry beans, under cur­ descriptors in the standard ecoinvent pedigree matrix developed by
rent management practices, in each region of production. In cases where Ciroth et al. (2016). The pedigree matrix comprises 5 data quality in­
there were multiple different data choices possible for a given bean and dicators (DQIs) respectively addressing: data reliability, completeness,
region (e.g., no tillage, reduced tillage, and intensive tillage all present temporal correlation, geographical correlation, further technological
in different proportions), the simulation was run multiple times and correlation; with 5 possible data quality scores ranging from high
averaged based on the proportions of production for each option (see quality (1) to low quality (5). Each of these scores correspond to a
Table 1 for data inputs to the Holos model and their sources). The ICBM specific uncertainty value that is represented as the geometric standard
simulation uses location-specific soil and climate data to model soil deviation of a lognormal distribution (Frischknecht and Rebitzer, 2005)
carbon dynamics from 1985 to 2021. The output data from ICBM gives (see ESM 1 Section 2.1.12.11 for further details).
the change in soil carbon per year per crop in the rotation. These data In addition to parameter uncertainty from data quality, stochastic
were used to calculate the average change in soil carbon sequestration parameter uncertainty was also considered. Stochastic uncertainty re­
over this period, assuming the same management practices were fers to inherent variability of data representing a flow or process, as
employed the entire time. This was modelled as a negative output of represented by a characteristic distribution. To quantify stochastic un­
carbon dioxide emission to air. certainty, wherever possible, a lognormal uncertainty distribution was
Background system data in this study included life cycle inventories calculated directly from the raw data (multiple replications of the same
for the production and transportation of inputs used in faba and dry bean measurement for the same treatment). When the stochastic uncertainty
production, including seeds, fertilizers, crop protection products, and distribution of the data was not known, the generic basic sectoral un­
energy carriers. These data were sourced from life cycle inventory da­ certainty factors from Frischknecht et al. (2005) were used. These basic
tabases such as Ecoinvent and Agri-footprint and were modified where uncertainty factors are meant to represent the amount of stochastic
necessary to best approximate Canadian conditions. All sources of uncertainty for LCI data that is commonly observed in data from the
background system LCIs, and any modifications applied, are described sector represented.
in ESM 1 Section 2.1.12. Transportation distances for LCI inputs were
estimated based on transportation distances indicated in Alberta Agri­ 3.1.7. Life cycle impact assessment
culture and Forestry (2018).
3.1.7.1. Impact assessment categories. Based on the ISO requirements,
3.1.6.1. Data quality and uncertainty. The farm-specific primary data all EPD requirements, and previous LCA studies of beans, the impact cat­
met the desired data quality standards since they came from farms in egories (and associated characterization models/factors) from IMPACT
bean-growing regions of Canada within 5 years of the time of the World+ selected for this study were: climate change, mineral resources

445
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

use, fossil and nuclear energy use, terrestrial acidification, freshwater harvest energy use, and nitrogen credits, and had lower inputs of N
eutrophication, freshwater ecotoxicity, particulate matter formation, fertilizers, total pesticides, and fuel use for field activities, compared to
water scarcity, land occupation, photochemical oxidant formation, dry beans (Tables 2–4 and ESM 1 Appendix B). There was also variability
ionizing radiation, and ozone layer depletion (Bulle et al., 2019). Impact within the three dry bean types studied, but with no clear trends. In
assessment results were reported at the mid-point for all LCIA addition to variability between bean types, there was also variability
categories. between regions within the same bean type. For example, compared to
the other provinces, faba beans had the highest yields in Alberta, and the
3.1.7.2. Uncertainty and sensitivity analysis. An uncertainty analysis highest seeding rates in Saskatchewan. The observed variations within
using Monte Carlo simulations was performed to determine the overall bean types may be due to climatic differences. For example, according to
uncertainty of the impact assessment results based on both stochastic expert opinion, pulse farmers in the north seed more to get the same
uncertainty and data quality. Monte Carlo simulation was selected since number of viable crop plants.
it is the most commonly used method by LCA practitioners (Bamber The most commonly applied N fertilizers across all four bean types
et al., 2020). We used a Monte Carlo simulation for uncertainty propa­ were urea (46–0-0) and urea ammonium nitrate (UAN) (28–0-0)
gation for each regional aggregated data set (1000 runs, as is typical in (Table 2). The commonly applied P fertilizer across bean types was
agricultural LCAs). Sensitivity analyses were performed to determine monoammonium phosphate (MAP). Potash was the commonly applied K
the sensitivity of the carbon sequestration data and the overall results to fertilizer across all bean types. Sulphur fertilizer inputs were realized
the choice of crop rotation and Soil Landscapes of Canada (SLC) polygon through the application of elemental sulphur and ammonium sulphate
used in the ICBM model in Holos, and the calculation of the N credit. (AMS). Biological nitrogen fixation and the associated N credit co-
The crop rotations included in the ICBM model were chosen as the product for faba beans was almost twice as much as that of dry beans
most common crop rotation(s) indicated by farmers. A sensitivity (ESM 1 Table 34). Among pesticides, herbicides were the most
analysis was performed by instead using the second most commonly frequently applied (Table 3). Some of the commonly used active in­
indicated crop rotation(s) for each bean and region based on the ICBM gredients in herbicide inputs were glyphosate, bentazon, imazamox, and
carbon modelling in Holos. The SLC polygons used to represent the soil clethodim. Among desiccants, diquat dibromide and saflufenacil were
conditions in the ICBM model were chosen based on the polygons with both used in three of the four bean types analyzed (Table 3). The three
the most occurrence of farms (from the latitude and longitude co­ commonly applied fungicide active ingredients for all four bean types
ordinates given by farmers). For the sensitivity analysis, the second most were boscalid, picoxystrobin, and prothioconazole. Insecticides were
common SLC polygons were chosen instead. In the cases in which there applied sparingly for faba and dry bean production, with lambda-
were no polygons that occurred more than once, additional polygons cyhalothrin being the only common insecticide across all four bean
were chosen based on their location in the predominant soil type for the types.
ecozone or province modelled (see ESM 1 Section 3.1.13.4.3 for details The most fuel intensive operation for all bean types was harvesting
on the crop rotations and SLC polygons used). (Table 4). The other two fuel intensive field activities were tillage and
The N credit values were estimated rather than measured or calcu­ planting/seeding. For faba beans, no irrigation was used in Manitoba
lated based on farm-specific data (other than yield). Therefore, a and ecozone 9, while Alberta, Saskatchewan, and ecozone 10 had 27.37
sensitivity analysis was also performed to test the results of alternative %, 17.32 %, and 40.13 % of seeded area irrigated, respectively. Based on
methods of N credit estimation. For the first alternative, the N fertilizer expert opinion, these proportions of irrigated area may be slightly
mix applied to the beans was used in calculating the N credit estimate higher than the true proportions, potentially due to the inclusion of seed
instead of ammonia fertilizer. For the second alternative N credit growers in the survey respondents. No irrigation was used for navy
calculation, the N fertilizer mix credit was considered along with beans in Ontario/ecozone 8, while Manitoba and ecozone 10 had 7.14 %
different estimates of nitrogen fixed. For faba beans, a newsletter from and 14.73 % of seeded area irrigated, respectively. For pinto beans, no
Saskatchewan Pulse Growers was used to model nitrogen fixation rates irrigation was used in Manitoba, while Alberta and ecozone 10 had 100
(pers. comm. Pulse Canada). The newsletter reported that faba beans % and 18.59 % of seeded area irrigated, respectively. One hundred
fixed 8–17 g N per kg of grain for the benefit of the following crop in the percent of pinto bean seeded area in Alberta was considered irrigated
rotation. Assuming a mean value of 12.5 g of N fixed per kg, N credits based on expert opinion, despite only 35.17 % of the seeded area
were calculated for the reported yields of faba beans. For dry beans, data captured through the surveys being irrigated. Manitoba had 42.89 % of
from North Platte Natural Resources District in the United States was its red kidney bean seeded area irrigated and Ontario red kidney beans
used (pers. comm. Pulse Canada). This study reported that dry beans fix were 4.28 % irrigated. Based on expert opinion, these values are likely
45 kg of N/ha for use by the following crops in rotation, which was also high, due to the low sample sizes in these areas.
scaled by the yield to estimate the amount of N fixed per kg beans. There were large differences in the electricity used for post-harvest
activities in different regions and for different bean types (Table 4).
4. Results and discussion This was primarily due to the differences in the share of yield that was
reported to be aerated. Two production regions – navy beans in Man­
4.1. Life cycle inventory itoba and pinto beans in Alberta – did not report any aeration. Navy and
pinto bean survey respondents did not report that any of the beans
The full life cycle inventory results at the ecozone, province and produced were dried using electricity. No drying using propane/natural
national aggregated scales are provided in Appendix B of ESM 1, and the gas was done for pinto beans. The average transportation distance for all
complete openLCA model for this study is publicly available on the Open inputs combined was 11.80 t-kilometers/t for faba beans, 10.23 t-kilo­
Science Framework (https://osf.io/3y8r4/?view_only=84d95d278a674 meters/t for navy beans, 9.86 t-kilometers/t for pinto beans, and 10.57 t-
e8883ea95fedb71768d), in the Canadian Agri-food Life-cycle Data kilometers/t for red kidney beans (ESM 1 Table 40). Transportation of
Centre at https://caldc.ca, and available for purchase in ecoinvent seed and fertilizers were the major contributors to transportation.
versions 3.9 and later. The availability of these transparently reported
data will enable methodological consistency for future ISO-compliant 4.2. Life cycle impact assessment
comparative studies.
In general, the three types of dry beans studied (navy, pinto, and red Overall, at the national scale of aggregation, faba beans had
kidney beans) had fairly similar LCI data, but faba beans were quite considerably lower impacts across all impact categories considered than
different. On average, faba beans had higher yields, seeding rates, post- the three dry beans (Fig. 2). This was primarily due to lower levels of
fertilizer use and the higher N fertilizer credits for BNF. Complete LCIA

446
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

Table 2
Seeding rate (kg seed/tonne harvested crop), yield (kg harvested/ha), and major fertilizer inputs of urea, UAN, ammonia, MAP, potash, ammonium sulphate, and
sulphur (kg fertilizer product/tonne harvested crop).
Region Seed Yield Fertilizer inputs

Urea UAN Ammonia MAP Potash Ammonium sulphate Sulphur

Faba beans
Alberta 55.35 4598.52 0.263 10.233 0.473 0.600
Manitoba 73.23 3040.27 5.958 5.958 5.213
Saskatchewan 82.76 3244.39 15.279 1.607
Ecozone 09 67.93 3777.10 9.487 1.331 0.451 0.695
Ecozone 10 66.33 3508.77 0.281 11.688 0.992 0.094

Navy beans
Alberta 29.31 2748.52 37.394 12.121
Manitoba 26.89 2230.52 11.216 18.887 3.031 15.554 3.086 2.132
Ontario 22.34 2739.36 11.206 44.909 0.572 34.187 3.469
Ecozone 08 22.34 2739.36 11.206 44.909 0.572 34.187 3.469
Ecozone 10 27.66 2240.73 10.687 17.996 5.461 17.147 2.941 2.032

Pinto beans
Alberta 24.08 3296.28 28.525 21.508 2.934 2.789
Manitoba 24.51 2616.70 18.344 23.127 3.562 10.327 0.503 0.397
Ecozone 10 22.95 2763.22 18.887 20.631 3.178 11.163 0.387 0.451 0.722

Red kidney beans

Manitoba 45.22 2132.21 25.888 13.310 11.108 6.025 4.972
Ontario 33.40 2696.50 11.827 22.735 35.373 35.373 7.975
Ecozone 08 33.40 2696.50 11.827 22.735 35.373 35.373 7.975
Ecozone 10 45.22 2132.21 25.888 13.310 11.108 6.025 4.972

Table 3
Application rates of common pesticides, including glyphosate, bentazon, imazamox, clethodim, lambda-cyhalothrin, diquat dibromide and prothioconazole (kg active
ingredient/t harvested crop).
Region Glyphosate Bentazon Imazamox Clethodim Lambda-cyhalothrin Diquat dibromide Prothioconazole

Faba beans
Alberta 0.066 0.064 0.003 0.001 0.0002 0.048 0.003
Manitoba 0.057 0.100 0.005 0.001 0.084
Saskatchewan 0.060 0.079 0.005 0.003 0.0001 0.021
Ecozone 9 0.073 0.088 0.004 0.002 0.0000 0.039
Ecozone 10 0.056 0.065 0.003 0.001 0.0002 0.046 0.003

Navy beans
Manitoba 0.0486 0.4328 0.006 0.0092 0.022 0.008
Ontario 0.0783 0.1103 0.0002 0.0003 0.002 0.047
Ecozone 8 0.0783 0.1103 0.0002 0.0003 0.002 0.047
Ecozone 10 0.0444 0.4186 0.006 0.0084 0.024 0.010

Pinto beans
Alberta 0.075 0.363 0.051 0.034
Manitoba 0.050 0.254 0.005 0.0027 0.003 0.004
Ecozone 10 0.051 0.262 0.005 0.0022 0.002 0.009

Red kidney beans

Manitoba 0.084 0.45 0.004 0.0097 0.017
Ontario 0.018 0.13 0.0005 0.0005 0.018
Ecozone 8 0.018 0.13 0.0005 0.0005 0.018
Ecozone 10 0.084 0.45 0.004 0.0097 0.017

results for faba bean production can be found in Appendix C of ESM 1. impacts) and highly significant (42–85 % reduction in water scarcity
Fertilizer production, field-level emissions from fertilizer use, and impacts). Since the N credit was modelled as an avoided fertilizer input,
fuel use for field activities were the three largest contributors to the these impacts varied in proportion to the impacts of N fertilizer pro­
overall environmental impacts of faba bean production (Fig. 3). The duction for each impact category.
other two major sources of impacts were seed inputs and pesticide use. Except for land occupation, photochemical oxidation, and ozone
Soil carbon only influenced climate change impacts, producing a 27–44 layer depletion, field-level and upstream fertilizer emissions were the
% net reduction in estimated GHG emissions. The impacts of the N credit highest contributor to all of the impact categories analyzed for all three
for the co-product of N fixed by the crop, included using system dry bean types (Figs. 4–6). Overall, the impacts for red kidney bean
expansion, varied depending on the impact category, with its influence production were higher than impacts of pinto bean and navy bean
ranging between trivial (<0.1 % reduction in freshwater eutrophication production for climate change, fossil and nuclear energy use, freshwater

447
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

Table 4
Fuel use (L diesel used/tonne harvest crop) for field activities including first tillage, fertilizer application, seeding/planting, fungicide application and harvesting;
electricity use (kWh/tonne harvested crop) for aeration and drying; fuel use (kWh/tonne harvested crop) for drying.
Region Litres/tonne kWh/tonne

First Fertilizer Seeding/ Fungicide Harvesting Electricity use for Electricity use for Fuel use for
tillage application Planting application aeration drying drying

Faba beans
Alberta 0.296 0 1.770 0.095 4.833 107.39 80.79 175
Manitoba 2.227 0.108 4.231 0.142 10.376 80.79
Saskatchewan 0.666 0.071 2.574 0.213 6.905 174.99
Ecozone 9 0.603 0.102 2.284 0.192 5.733 235.89 0.071
Ecozone 10 0.727 0.056 2.434 0.098 6.734 109.38 0.086 5.75

Navy beans
Manitoba 2.531 0.980 2.746 0.218 10.887
Ontario 3.156 0.310 1.869 0.262 7.038 9.61 16.00
Ecozone 8 3.156 0.310 1.869 0.262 7.038 9.61 16.00
Ecozone 10 2.540 0.367 2.703 0.235 10.909 15.65

Pinto beans
Alberta 3.361 0.260 1.658 0.230 8.133
Manitoba 2.359 0.350 2.181 0.211 8.624 0.07
Ecozone 10 2.457 0.392 2.093 0.232 8.519 11.67

Red kidney beans

Manitoba 1.998 0.392 2.491 0.391 10.598 22.38
Ontario 5.276 0.376 2.054 0.274 8.023 0.38 0.009 13.63
Ecozone 8 5.276 0.376 2.054 0.274 8.023 0.38 0.009 13.63
Ecozone 10 1.998 0.391 2.491 0.391 10.598 22.38

eutrophication, ozone layer depletion, and freshwater ecotoxicity. For slightly higher in Alberta than Manitoba, primarily due to higher
the other impact categories, navy beans had the highest impacts. Pinto amounts of C sequestration in Manitoba (Fig. 4A). Kidney bean pro­
beans had the lowest environmental impacts overall among the three duction in Ontario/ecozone 8 had higher GHG emissions than Man­
dry bean types. N fertilizer credits had a much lower influence on the itoba/ecozone 10 due to higher fertilizer and fuel inputs (Fig. 5A). For
impacts associated with dry bean production, often producing a 1–3 % navy beans, climate change impacts were higher in Ontario/ecozone 8,
reduction in the overall impacts for most impact categories. This is due with the main difference being higher GHG emissions associated with
to the correspondingly lower N fixation rates for dry beans compared to fertilizer use (Fig. 6A).
faba beans. As with faba beans, soil carbon sequestered reduced climate
change impacts significantly, though its influence was much more var­ 4.2.1.2. Mineral resource use. Fertilizer inputs were the highest con­
ied, ranging between − 1 % and − 28 %. tributors to mineral resource use impacts for all four bean types, though
Detailed analyses of each impact category are presented below. their share for faba beans (47 % at the national level) was lower than for
dry beans (82–87 %) due to the lower fertilizer application levels in faba
bean production (Fig. 2B). These fertilizer-related impacts were pri­
marily due to upstream fertilizer production – specifically the produc­
4.2.1.1. Climate change. On average, fertilizers (48 %) and fuel for field tion of urea ammonium nitrate (UAN) and monoammonium phosphate
operations (37 %) were the highest contributors to climate change im­ (MAP). The N credit from BNF accounted for 10–15 % reductions in
pacts for faba beans, while fertilizers (74–80 %) contributed most to impacts for faba beans, and 1–2 % reductions for dry beans.
climate change impacts for all three dry bean types (Fig. 2A). Field Overall, impacts of faba bean production in Saskatchewan were the
operations emissions came primarily from CO2 emitted in the combus­ highest at the provincial level (Fig. 3B). In contrast, the mineral resource
tion of diesel to run agricultural machinery and the upstream emissions use impacts at the ecozone level were almost identical. Kidney beans and
associated with producing fuel. Most of the fertilizer-related impacts navy beans had considerably higher mineral resource use impacts (over
came from field-level emissions of applied fertilizer, as well as emissions 60 % higher) than pinto beans, which can be attributed to higher inputs
associated with upstream phosphorus fertilizer production. Soil carbon of MAP for kidney bean production in Ontario/Ecozone 8, and lower
sequestration resulted in between 3 % (red kidney beans) and 25 % yields (Fig. 2B). Overall, red kidney bean production had the highest
(pinto beans) reductions in climate change impacts. Nitrogen fixation mineral resource use impacts among the three dry bean types at both the
had similarly significant impacts on faba bean production – producing a provincial and ecozone levels. Red kidney and navy beans produced in
13 % reduction in the climate change impacts. The impacts of nitrogen Ontario/ecozone 8 had higher impacts than Manitoba/ecozone 10 due
fixation were considerably lower (2 %) for dry beans due to lower N to higher levels of fertilizer application (Figs. 5B–6B).
fixation and higher levels of fertilizer use.
Overall, faba bean production in Alberta had the lowest GHG emis­ 4.2.1.3. Fossil and nuclear energy use. For fossil and nuclear energy use,
sions at the provincial level – primarily due to the lower fuel use, higher fertilizers contributed a higher share of the impacts for dry beans
soil carbon sequestration and higher N fixation (Fig. 3A). At the ecozone (52–58 %) than faba beans (25 % of the national average) with field
level, the overall GHG emissions of faba bean production in ecozone 9 operations being the other major contributor for both faba beans (50 %)
were significantly lower than ecozone 10, due to differences in BNF and and dry beans (29–34 %) (Fig. 2C). The impacts due to fertilizers can be
SOC sequestration. At the national scale, kidney beans (368.8 kg CO2- attributed to the production of urea and MAP, and the impacts due to
eq.) had higher GHG emissions than navy beans (341.5 kg CO2-eq.) and field activities can be attributed to the production of diesel used in farm
pinto beans (214.3 kg CO2-eq.) (Fig. 2A). For pinto beans, impacts were machinery. Overall, the N credit contributed a 13–26 % reduction in

448
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

Inoculant A) Climate change B) Mineral resources use

N credit 500
b d b b
Post-harvest acvies 400 2.3
Soil carbon sequestraon c
c

kg deprived
300 1.8

kg CO2e
Irrigaon
200 1.3
Pescides a
100 0.8 a
Ferlizer
Field operaons 0 0.3
Seed -100 -0.2
Agricultural land Faba Navy Pinto Red Faba Navy Pinto Red
Total
kidney kidney

C) Fossil and nuclear D) Freshwater E) Terrestrial acidificaon

Eutrophicaon 2.5E-05 b d
3700 energy use d 0.014
b d 2.0E-05 c
c b c
kg PO4-lim eq.
MJ deprived

2700

kg SO2-eq
1.5E-05
0.009 a
1700 1.0E-05
a a
0.004 5.0E-06
700
0.0E+00
-300 -0.001 -5.0E-06
Faba Navy Pinto Red Faba Navy Pinto Red Faba Navy Pinto Red
kidney kidney kidney
F) Parculate maer H) Land occupaon,
G) Water scarcity
formaon biodiversity c
b c
d b

m2 arable land-eq. yr
0.35 b d b
m3 world-eq.

c 180
kg PM2.5-eq.

3900 a
0.25
0.15 80
a a 1900
0.05
-0.05 -20 -100
Faba Navy Pinto Red Faba Navy Pinto Red Faba Navy Pinto Red
kidney kidney kidney
I) Photochemical oxidant J) Ionizing radiaon K) Ozone layer depleon
formaon b d 5.5E-05 d
1.9 1400 b
b
kg CFC-11 eq.

d c
Bq C-14 eq.

c
kg NMVOC-eq.

1.4 3.5E-05
c 900
a a
0.9 a
400 1.5E-05
0.4

-0.1 -100 -5.0E-06

Faba Navy Pinto Red Faba Navy Pinto Red Faba Navy Pinto Red
kidney kidney kidney

L) Freshwater ecotoxicity
c
4.8E+06
3.8E+06
d
b
2.8E+06
CTUe

1.8E+06 a
8.0E+05
-2.0E+05
Faba Navy Pinto Red
kidney

Fig. 2. Impact assessment results (IMPACT World+) per tonne of faba and dry beans produced at the national scale of aggregation. Results presented for A) climate
change, B) mineral resources use, C) fossil and nuclear energy use, D) freshwater eutrophication, E) terrestrial acidification, F) particulate matter formation, G) water
scarcity, H) land occupation, biodiversity, I) photochemical oxidant formation, J) ionizing radiation, K) ozone layer depletion, and L) freshwater ecotoxicity. Small
letters above bars (a, b, c, d) refer bean types that have significantly different impacts (p<0.05).

449
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

Inoculant
A) Climate change B) Mineral resources use
N credit
150 0.7 c
Post-harvest acvies b b
Soil carbon sequestraon b a a
100 0.5

kg deprived
Irrigaon c b
a

kg CO2e
Pescides a
50 0.3
Ferlizer
Field operaons 0 0.1
Seed
Agricultural land -50 -0.1
Total AB MB SK 9 10 AB MB SK 9 10

C) Fossil and nuclear 0.014 D) Freshwater E) Terrestrial acidificaon

b c
1700 b energy use Eutrophicaon
c 3.8E-06
b

kg PO4-lim eq.
c 0.009 b a b a
MJ deprived

kg SO2-eq
a b a a
a
700 1.8E-06
0.004

-300 -0.001 -2.0E-07

AB MB SK 9 10 AB MB SK 9 10 AB MB SK 9 10

F) Parculate maer G) Water scarcity H) Land occupaon,

formaon 40
biodiversity
b b b b c

m2 arable land-eq. yr
a 3900 b
0.055 a a
m3 world-eq.

a a
kg PM2.5-eq.

a 20 c
0.035 b a
1900
0
0.015

-0.005 -20 -100

AB MB SK 9 10 AB MB SK 9 10 AB MB SK 9 10

I) Photochemical oxidant J) Ionizing radiaon K) Ozone layer depleon

formaon 750 b c 2.5E-05 b
b
a c b
kg NMVOC-eq.

0.9 b
kg CFC-11 eq.

550 a
Bq C-14 eq.

c b a 1.5E-05 a
a
a 350
0.4 5.0E-06
150

-0.1 -50 -5.0E-06

AB MB SK 9 10 AB MB SK 9 10 AB MB SK 9 10

L) Freshwater ecotoxicity
c
1.4E+06
a
CTUe

9.0E+05 b
a b

4.0E+05

-1.0E+05
AB MB SK 9 10

Fig. 3. Impact assessment results (IMPACT World+) per tonne of faba beans produced at the provincial (Alberta, Manitoba, Saskatchewan) and ecozone (Ecozone 9
Boreal Plains, Ecozone 10 Prairies) scales of aggregation. Results presented for A) climate change, B) mineral resources use, C) fossil and nuclear energy use, D)
freshwater eutrophication, E) terrestrial acidification, F) particulate matter formation, G) water scarcity, H) land occupation, biodiversity, I) photochemical oxidant
formation, J) ionizing radiation, K) ozone layer depletion, and L) freshwater ecotoxicity. Small letters above bars (a, b, c) refer to regions within one scale (provincial
or ecozone) that have significantly different impacts (p < 0.05). Provinces and ecozones were not compared to each other for statistical purposes.

450
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

N credit A) Climate change B) Mineral resources use

Post-harvest acvies 400 1.9
b a
Soil carbon sequestraon a a
300 b a 1.4

kg deprived
Irrigaon

kg CO2e
Pescides 200
0.9
Ferlizer 100
Field operaons 0.4
Seed 0
Agricultural land -100 -0.1
Total AB MB 10 AB MB 10
C) Fossil and nuclear D) Freshwater E) Terrestrial acidificaon
energy use eutrophicaon a
2800 a 1.9E-05 b
b a b a a
0.009 a
kg PO4-lim eq
1.4E-05
MJ deprived

kg SO2 eq
1800
9.0E-06
0.004
800
4.0E-06

-200 -0.001 -1.0E-06

AB MB 10 AB MB 10 AB MB 10
F) Parculate maer G) Water scarcity H) Land occupaon,
0.29
a formaon a biodiversity

m2 arable land eq .yr

b a 280 b a
a
kg PM2.5 eq

m3 world-eq

3900
0.19 a
180 b

0.09 1900
80

-0.01 -20 -100

AB MB 10 AB MB 10 AB MB 10

I) Photochemical oxidant J) Ionizing radiaon K) Ozone layer depleon

a b a
1.4 formaon a 3.5E-05
b
a a 950 a
b
kg NMVOC eq

2.5E-05
kg CFC-11 eq
Bq C-14 eq

0.9
1.5E-05
450
0.4
5.0E-06

-0.1 -50 -5.0E-06

AB MB 10 AB MB 10 AB MB 10

L) Freshwater ecotoxicity
a
1.9E+06 b a

1.4E+06
CTUe

9.0E+05

4.0E+05

-1.0E+05
AB MB 10

Fig. 4. Sensitivity analysis of nitrogen fixed, modelled as an N fertilizer credit: Original N credit calculated from Barker (2007) for Top Crop Manage and modelled as
a credit of ammonia fertilizer, Alternate N% N credit calculated as a credit of the fertilizer mix applied to the beans instead of ammonia, and Alternate N Fixed N
credit calculated according to a newsletter from Saskatchewan Pulse Growers for faba beans, and data from North Platte Natural Resources District in the United
States for dry beans (pers. comm. Pulse Canada).

451
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

N credit A) Climate change B) Mineral resources use

Post-harvest acvies 450 b a
Soil carbon sequestraon a b a
350 b 2.8
Irrigaon

kg deprived
a b

kg CO2e
Pescides 250
1.8
Ferlizer 150
Field operaons 0.8
Seed 50
Agricultural land -50 -0.2
Total MB ON 8 10 MB ON 8 10
C) Fossil and nuclear D) Freshwater E) Terrestrial acidificaon
energy use eutrophicaon 2.9E-05 a b
3800 b a
0.014 a b
a b a a b a
kg PO4-lim eq

kg SO2 eq
MJ deprived

2800 1.9E-05
0.009
1800
0.004 9.0E-06
800

-200 -0.001 -1.0E-06

MB ON 8 10 MB ON 8 10 MB ON 8 10
F) Parculate maer G) Water scarcity H) Land occupaon,
0.38 formaon 190 a b biodiversity
a b 5900 a b
b a

m2 arable land eq .yr

b a b a
0.28 140
m3 world-eq
kg PM2.5 eq

3900
0.18 90
1900
0.08 40

-0.02 -10 -100

MB ON 8 10 MB ON 8 10 MB ON 8 10
I) Photochemical oxidant J) Ionizing radiaon K) Ozone layer depleon
formaon 1900 5.5E-05 b a
b a b a
1.4 a b 4.5E-05 a b
1400
kg CFC-11 eq

a b
kg NMVOC eq

Bq C-14 eq

3.5E-05
0.9
900 2.5E-05

0.4 1.5E-05
400
5.0E-06
-0.1 -100 -5.0E-06
MB ON 8 10 MB ON 8 10 MB ON 8 10

L) Freshwater ecotoxicity
3.9E+06 a b a b

2.9E+06
CTUe

1.9E+06

9.0E+05

-1.0E+05
MB ON 8 10

Fig. 5. LCIA results (IMPACT World+ Climate Change) for A) faba (provincial) and B) red kidney (ecozone) beans for sensitivity analysis of N credit estimates.
Original N credit calculated from Barker (2007) for Top Crop Manage and modelled as a credit of ammonia fertilizer, Alternate N% N credit calculated as a credit the
fertilizer mix applied to the beans instead of ammonia, and Alternate N Fixed N credit calculated according to a newsletter from Saskatchewan Pulse Growers for faba
beans, and data from North Platte Natural Resources District in the United States for dry beans (pers. comm. Pulse Canada).

452
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

N credit
A) Climate change B) Mineral resources use
Post-harvest acvies 600 3.8 b a
b a
Soil carbon sequestraon
Irrigaon 400 b 2.8

kg deprived
a

kg CO2e
Pescides a b
Ferlizer 200 1.8
Field operaons 0 0.8
Seed
Agricultural land -200 -0.2
Total MB ON 8 10 MB ON 8 10
C) Fossil and nuclear D) Freshwater E) Terrestrial acidificaon
energy use eutrophicaon 2.9E-05 b a
0.014 b
b a b a a
3600 b
a b

kg PO4-lim eq

kg SO2 eq
1.9E-05
MJ deprived

2600 0.009
a
1600
0.004 9.0E-06
600
-400 -0.001 -1.0E-06
MB ON 8 10 MB ON 8 10 MB ON 8 10

F) Parculate maer G) Water scarcity H) Land occupaon,

formaon b a biodiversity
b
5900 a b
b a 180

m2 arable land eq .yr

0.39 a
a b b a
m3 world-eq

130
kg PM2.5 eq

0.29 3900
0.19 80
1900
0.09 30

-0.01 -20 -100

MB ON 8 10 MB ON 8 10 MB ON 8 10
I) Photochemical oxidant J) Ionizing radiaon K) Ozone layer depleon
formaon b a 5.5E-05 b a
1.9 b a 1950
a a a b
kg NMVOC eq

kg CFC-11 eq

1450 a b
Bq C-14 eq

3.5E-05
0.9 950
1.5E-05
450

-0.1 -50 -5.0E-06

MB ON 8 10 MB ON 8 10 MB ON 8 10

L) Freshwater ecotoxicity
b a
3.9E+06

2.9E+06
a b
CTUe

1.9E+06

9.0E+05

-1.0E+05
MB ON 8 10

Fig. 6. Impact assessment results (IMPACT World+) per tonne of navy beans produced at the provincial (Manitoba, Ontario) and ecozone (Ecozone 8 Mixedwood
Plains, Ecozone 10 Prairies) scales of aggregation. Results presented for A) climate change, B) mineral resources use, C) fossil and nuclear energy use, D) freshwater
eutrophication, E) terrestrial acidification, F) particulate matter formation, G) water scarcity, H) land occupation, biodiversity, I) photochemical oxidant formation,
J) ionizing radiation, K) ozone layer depletion, and L) freshwater ecotoxicity. Small letters above bars (a, b) refer to regions within one scale (provincial or ecozone)
that have significantly different impacts (p<0.05). Provinces and ecozones were not compared to each other for statistical purposes.

453
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

fossil and nuclear energy use impacts of faba beans, and only a 3–4 % Overall, the particulate matter formation impacts for faba bean
reduction for dry beans. production were highest in Manitoba and ecozone 10 (Fig. 3F). The
Overall, fossil and nuclear energy use impacts of faba bean produc­ impacts of pinto bean production were lower than the impacts of kidney
tion were highest in Manitoba at the provincial level – about 30 % more bean production and navy bean production (Fig. 2F). Navy beans in
than the next highest (Saskatchewan) (Fig. 3C). At the ecozone level, the Ontario/ecozone 8 had higher impacts than Manitoba/ecozone 10, due
overall impacts in ecozone 10 were slightly higher than in ecozone 9, to higher fertilizer inputs (Fig. 6F). For red kidney beans, the regional
with most of this difference resulting from different fuel use levels. trend was the opposite, with Ontario/ecozone 8 having lower impacts,
Kidney beans had higher fossil and nuclear energy use impacts overall again due to differences in fertilizer application (Fig. 5F). Pinto beans
compared to pinto and navy beans, due to higher MAP application and produced in Alberta had higher impacts than Manitoba (Fig. 4F).
fuel use for field activities (Fig. 2C). Pinto bean production in Alberta
had higher fossil and nuclear energy use than in Manitoba due to higher 4.2.1.7. Water scarcity. Fertilizer use was again the major contributor
application rates of MAP and urea, and higher fuel use (Fig. 4C). For to water scarcity impacts of all dry bean production (71–79 % of the
kidney and navy beans, production in Ontario/ecozone 8 had higher national average), with irrigation accounting for <5 % of the impacts in
impacts than Manitoba/ecozone 10 due to significantly higher MAP the case of navy beans and red kidney beans (Fig. 2G). Irrigation
application rates and higher fuel use (Figs. 5C–6C). accounted for 17 % of water scarcity impacts of pinto beans due to the
higher levels of irrigation in Alberta. Water scarcity impacts attributed
4.2.1.4. Freshwater eutrophication. For freshwater eutrophication, there to fertilizers were due to the production of urea and MAP fertilizers. For
was less variability in the contribution analysis, with fertilizers (pre­ faba beans, irrigation was the biggest contributor to water scarcity im­
dominantly on-field phosphate emissions resulting from fertilizer pacts (46 %), followed by fertilizer use (30 %). Faba beans in Manitoba
application) accounting for most of the impacts (81–84 %) across all four did not report any irrigation, while the reported average irrigation water
bean types (Fig. 2D). The N credit included via system expansion had usage in Alberta (0.50 m3/t) was about five times higher than in Sas­
limited influence on freshwater eutrophication, causing only a < 1 % katchewan (0.08 m3/t). Hence, irrigation was the highest contributor to
reduction in the impacts of all bean types. water scarcity impacts in Alberta (64 %) but only the fourth largest
On average, faba bean production in Saskatchewan had higher contributor in Saskatchewan (14 %). The differences in contributions of
freshwater eutrophication impacts compared to Alberta and Manitoba, irrigation were larger at the ecozone level. Irrigation contributed 57 %
due to higher fertilizer application and associated emissions (Fig. 3D). At of the water scarcity impacts in ecozone 10 and 0 % in ecozone 9. The
the ecozone level, impacts were similar for ecozone 9 and ecozone 10. overall net water scarcity impacts of dry bean production were reduced
On average, kidney beans had higher freshwater eutrophication impacts by 42–87 %, and 2–5 % due to nitrogen fixation, for faba beans and dry
compared to pinto and navy beans (Fig. 2D). There were no major dif­ beans, respectively.
ferences between regions for both pinto and red kidney beans For faba beans, water scarcity impacts were highest in Alberta and
(Figs. 4D–5D), but impacts of navy bean production in Ontario were ecozone 10, due to higher levels of irrigation in those regions (Fig. 3G).
almost twice as high as in Manitoba, again due to differences in fertilizer The water scarcity impact of pinto bean production was higher than
application and associated emissions (Fig. 6D). kidney beans and navy beans (Fig. 2G). Pinto beans had higher impacts
in Alberta due to higher levels of irrigation (Fig. 4G). For red kidney
4.2.1.5. Terrestrial acidification. At the national scale, fertilizers beans, impacts were higher in Manitoba/ecozone 10 due to slightly
accounted for a higher share of terrestrial acidification impacts in dry higher irrigation, as well as upstream impacts from pesticide and fer­
beans (84–89 %) compared to faba beans (52 %), with fuel use being the tilizer production (Fig. 5G). Navy beans had higher impacts in Ontario/
second highest contributor for both faba (34 %) and dry beans (7–11 %) ecozone 8, despite lower levels of irrigation, due to higher upstream
(Fig. 2E). Fertilizer use impacts were mainly the result of field level impacts from higher levels of fertilizer application (Fig. 6G).
emissions of ammonia and nitric oxide from fertilizer application, and
the emissions associated with the production of MAP fertilizer. The 4.2.1.8. Land occupation. Agricultural land use accounted for 93–97 %
contribution of field activities was due to the nitrogen oxides and of land occupation for all bean types at the national scale (Fig. 2H).
sulphur dioxide emissions caused by burning fuel for field operations. These contributions had minimal variability, ranging between 92 % and
The N credit from BNF contributed a 3–5 % reduction in impacts for faba 95 % (faba beans), and between 96 % and 98 % (dry beans), of the
beans, and < 1 % for dry beans. overall land use impacts across all production regions. N credits for ni­
For faba beans, at the provincial level, terrestrial acidification im­ trogen fixation did not contribute to significant reductions in land
pacts were highest in Manitoba (Fig. 3E). At the ecozone level, the im­ occupation.
pacts of faba bean production in ecozone 10 were higher than ecozone 9. The land occupation impacts of faba beans were ~ 50 % and 40 %
Overall, the production-weighted average terrestrial acidification im­ higher in Manitoba and Saskatchewan, respectively, compared to the
pacts of navy bean production were higher than for the production of impacts in Alberta – primarily due to the differences in reported yields
pinto and red kidney beans (Fig. 2E). Pinto beans produced in Alberta (Fig. 3H). At the ecozone level, the impacts were ~ 10 % higher in
had higher impacts than Manitoba (Fig. 4E). Navy beans had higher ecozone 10. Overall, land occupation impacts were the highest in
acidification impacts in Ontario/ecozone 8 (Fig. 6E), and red kidney Manitoba (3719.90 m2 arable land eq.). For pinto beans, production in
beans had higher impacts in Manitoba/ecozone 10 Fig. 5E). All regional Manitoba had ~25 % higher land occupation impacts due to the dif­
differences were driven by differences in fertilizer application and ferences in reported yields for the respective regions (Fig. 4H). Similarly,
associated impacts. navy bean yields were ~ 25 % higher in Manitoba compared to Ontario
(Fig. 6H). For red kidney beans, the land occupation impacts for pro­
4.2.1.6. Particulate matter formation. For particulate matter formation duction in Manitoba were ~ 30 % higher than for Ontario (Fig. 5H).
impacts at the national scale, fertilizers (82–87 % for dry beans, 51 % for Comparing the different dry bean types, land occupation impacts for
faba beans) and fuel use (8–12 % for dry beans, 35 % for faba beans) pinto beans were the lowest and navy beans the highest (Fig. 2H).
were the major contributing factors (Fig. 2F). The impacts associated
with fertilizer use can be attributed to field-level emissions and the 4.2.1.9. Photochemical oxidant formation. Fuel use was the main
production of MAP fertilizer. N credits produced a 3–4 % decrease in contributor to photochemical oxidant formation impacts of faba, pinto
overall particulate matter formation impacts for faba beans, and < 1 % and kidney bean production (71–75 %) (Fig. 2I). For navy beans, both
for dry beans. fuel use (48 %) and fertilizers (47 %) contributed equally to

454
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

photochemical oxidant formation, due to higher fertilizer use for navy pesticide use (Fig. 4L). For red kidney beans, impacts were similar be­
beans. The photochemical oxidant formation impacts due to fuel use can tween Manitoba/ecozone 10 and Ontario/ecozone 8 due to both higher
be attributed to the emissions of nitrogen oxides and non-methane pesticide use and lower fertilizer use in Manitoba (Fig. 5L). For navy
volatile organic compounds (NMVOC) caused by burning diesel in beans, impacts were lower in Manitoba/ecozone 8 due to lower fertilizer
agricultural machinery. Impacts from fertilizers were due to the nitrogen use (Fig. 6L). Overall, freshwater ecotoxicity impacts of pinto produc­
oxides emissions in the production of urea and MAP fertilizers. Nitrogen tion were considerably higher than both navy beans and red kidney
credits from BNF reduced impacts by 2–4 % for faba beans, and 1 % for beans, due to higher upstream fertilizer impacts (Fig. 2L).
dry beans.
For faba beans, the impacts in Manitoba and ecozone 10 were the
highest due to higher fuel use for field activities (Fig. 3I). The photo­ 4.3. Life cycle interpretation
chemical oxidant formation impacts of navy bean production were
higher than for both kidney and pinto beans, due to the higher fertilizer 4.3.1. Uncertainty and data quality analysis
use (Fig. 2I). Pinto beans had higher impacts in Alberta, due to higher The results of the Monte Carlo simulation with 1000 runs for the
fuel use for field operations and fertilizer use (Fig. 4I). Red kidney beans LCIA results of faba and dry beans are represented as standard error bars
and navy beans had higher impacts in Ontario/ecozone 8, mainly due to in the LCIA results presented in Figs. 2–6. The differences in uncertainty
differences in fertilizer use (Fig. 5I). In the case of navy beans, the higher came largely from the foreground data variability, calculated as the
fertilizer use outweighed lower fuel use in this region (Fig. 6I). geometric standard deviation of the LCI input and output data.
For faba beans at the provincial level, uncertainty was lowest in
4.2.1.10. Ionizing radiation. There was considerable variability in the Alberta for all impact categories except water scarcity. Uncertainty was
contributing factors for ionizing radiation impacts across the bean types. generally higher in Manitoba, likely as a result of fewer survey responses
Fertilizer use (32–49 %), fuel for field operations (28–44 %), and pes­ (5) compared to the sample sizes for Alberta (22) and Saskatchewan
ticides (13–23 %) all contributed significantly to ionizing radiation (23). At the ecozone level, standard errors in faba bean results were
impacts across all bean types at the national level (Fig. 2J). The ionizing higher for ecozone 9, except for freshwater eutrophication, ionizing
radiation impacts from fuel use are primarily due to the production of radiation, photochemical oxidant formation, and water scarcity. Over­
the diesel used for agricultural machinery. Production of urea and MAP all, uncertainties were higher at the provincial level than at the ecozone
fertilizers was the primary reason for the impacts from fertilizer use. level. This is consistent with the findings for peas and lentils (Bamber
Production of herbicides – especially Bentazon and glyphosate – et al., 2022), where the differences between regions were more often
contributed most to the impacts from pesticide use. The N credit reduced statistically significant for ecozones than for provinces. This is likely
impacts around 1–2 % for faba beans and < 1 % for dry beans. because ecozones are based on soil and climate differences (which are
The ionizing radiation impacts of faba bean production were highest influential with respect to life cycle inventory and impact assessment
in Saskatchewan and Manitoba (Fig. 3J). For dry beans, the impacts of results), rather than political borders as is the case for provinces. This
pinto bean production were considerably lower than those for navy indicates that there are more similarities between farms based on soil
beans and red kidney beans, mainly due to lower levels of fertilizer use and climate and related managements practices and emissions, than
(Fig. 2J). The ionizing radiation impacts of pinto bean production were based on provincial guidelines and regulations.
slightly higher in Alberta than Manitoba, due to slightly higher impacts For dry beans, the uncertainties are more difficult to interpret due to
from fuel use, fertilizer, and pesticide use (Fig. 4J). For red kidney and small sample sizes and data for ecozones and provinces being the same
navy kidney beans, impacts were much higher in Ontario/ecozone 8 due in cases where all survey responses from certain provinces are in the
to higher fertilizer use (Figs. 5J–6J). same ecozone. For navy beans, results for Ontario had higher uncer­
tainty than those for Manitoba, except with respect to land occupation.
4.2.1.11. Ozone layer depletion. Significant variability in the contribu­ At the ecozone level, higher uncertainties were distributed largely
tions of fertilizers (11–41 %), fuel use (31–51 %), and pesticides (15–22 evenly between ecozone 8 and 10. While the standard errors for Ontario
%) can be seen with respect to ozone depletion impacts (Fig. 2K). These and ecozone 8 are similar, ecozone 10 also includes data from Alberta,
impacts are due to emissions resulting from upstream emissions of which weren't modelled at the provincial scale due to a sample size of 3
methane (bromotrifluoro-, Halon 1301 and tetrachloro-, R-10). N credits or fewer survey responses. This showcases the unreliability of data from
contributed an 8–13 % reduction in overall ozone layer depletion im­ small sample sizes. The differences in standard errors between the
pacts for faba beans, and 2–4 % for dry beans. Manitoba and ecozone 10 (Manitoba and Alberta) production regions
For faba beans, ozone layer depletion impacts were highest for for navy beans were between 6 % and 78 % percent, depending on the
Manitoba and ecozone 10 (Fig. 3K). The ozone layer depletion impacts impact category. For kidney beans, Ontario had higher uncertainties
did not vary much between Alberta and Manitoba for pinto bean pro­ than Manitoba for a majority of the impact categories. Since these re­
duction (Fig. 4K). For kidney beans, the ozone depletion impacts were ~ gions at the provincial scale directly correspond to the two production
45 % higher in Ontario/ecozone 8 compared to Manitoba/ecozone 10 regions at the ecozone scale (unlike navy beans), ecozone 8 had higher
due to higher inputs, particularly of fertilizers (Fig. 5K). Similar differ­ uncertainties than ecozone 10. For pinto beans, uncertainties had less
ences were also observed for navy bean production in Manitoba and variability – sometimes higher in Alberta and other times in Manitoba.
Ontario (Fig. 6K). Overall, ozone layer depletion impacts were higher for Uncertainties were generally higher at the national level than both
kidney beans compared to pinto beans and navy beans (Fig. 2K). provincial and ecozone scales across the four bean types considered.
This is indicative of the variability that exists between regions, meaning
4.2.1.12. Freshwater ecotoxicity. At the national level, fertilizers were that farms within ecozones or provinces have similarities and are
the biggest contributor to freshwater ecotoxicity impacts for both faba different from farms in other ecozones or provinces.
beans (64 %) and dry beans (70–91 %) (Fig. 2L). These impacts can be
attributed to upstream emissions of aluminium, copper, and iron in urea 4.3.2. Sensitivity analysis
and MAP production. N fixation offset about 8 % of the freshwater
ecotoxicity impacts of faba bean production. Comparatively, the 4.3.2.1. Soil carbon. While modelling the results for soil carbon
reduction in impacts was smaller for dry beans (1–2 %). sequestration using the ICBM model in Holos, it was found that the re­
Faba bean impacts were highest in Saskatchewan and ecozone 9 sults were highly dependent on the location (SLC polygon) and the crop
(Fig. 3L). Pinto bean impacts were higher in Alberta due to higher rotation. Since most farms were in different SLC polygons and used
different crop rotations, the most common ones were chosen for

455
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

inclusion in the models. Therefore, a sensitivity analysis was performed fertilizer mix applied to beans has a much lower percentage of N than
to assess how the results would change if, instead, the second most ammonia (Fig. 7 – alternate N%). Using the alternate sources to calcu­
common SLC polygons and crop rotations were used to calculate the soil late the amount of N fixed (specified in Section 3.1.7.2) resulted in the
carbon sequestration for the faba and dry bean fields (Table 5). estimation of much higher N credits for both faba and dry beans (Fig. 7 –
Alternate crop rotations did not produce any major changes in the alternate N fixed).
estimates of soil carbon sequestered for both faba bean and pinto bean The impacts of alternate N credit estimates on the overall impacts of
production. Comparatively, navy beans and red kidney beans had faba and dry bean production are considerable. When the nitrogen fer­
considerable changes in carbon sequestered due to alternate crop rota­ tilizer mix is used as the N credit, rather than ammonia, it produced a 10-
tions. For example, estimated soil carbon sequestered was reduced by fold increase in the N credit estimate for faba beans, and a 2- to 5-fold
68 % for navy bean production in Manitoba when the crop rotation increase in the N credit estimates for dry beans. When this reported N
considered changed from “wheat-corn-navy beans” to “corn-navy beans- content is used along with the alternate (higher) estimates of N fixed (as
canola-wheat-soy”. For Ontario/ecozone 8, soil carbon sequestered mentioned in Section 3.1.7.2), the N credit estimate increased by 15- to
increased by over 5 times to 21.5 kg CO2 when the crop rotation 30-times across all bean types. This clearly identifies the modelling of
modelled changed from “corn-kidney beans-winter wheat” to “wheat- the N credit as a key focus for future research efforts. Field-level mea­
corn-soy-corn-kidney beans”. When the region modelled (SLC polygon) surements and model validation are required to ensure accurate
in HOLOS is changed, a 40–70 % change in soil carbon sequestration was modelling of BNF for use in LCA and other sustainability assessments.
observed in all production regions except Ontario/ecozone 8. For navy The changes in estimated climate change impacts resulting from the
and red kidney bean production in Ontario/ecozone 8, changing the SLC alternate N-credit estimates for faba and red kidney beans are presented
polygon resulted in a 10-fold increase in soil carbon sequestered. For all as examples in Fig. 8. The climate change impacts of faba bean pro­
production regions except navy bean production in Manitoba/ecozone duction at the provincial level were up to 2 times lower when the N
10, soil carbon sequestered increased due to the change in the SLC content (%) and background production of the reported fertilizer mix
polygon. ESM 1 Section 3.3.2.1 shows the overall changes in climate were used for N credit calculation. When the alternated N fixation values
change impacts (soil carbon sequestration is modelled as CO2 taken out were used, the climate change impacts were reduced further, by 3–5
of the atmosphere, thus only contributes to this impact category), due to times compared to the impacts originally reported. These reductions
the changes in crop rotation and SLC polygon. In the context of the entire were so significant that the beneficial impacts of estimated N credit
life cycles of the beans, change in crop rotation did not greatly alter the values outweighed all the positive (detrimental) impacts of faba bean
results, however the differences in overall impacts were non-trivial for production, resulting in net negative (beneficial) climate change im­
the SLC polygon sensitivity analysis. pacts. This trend was observed across most impact categories for faba
beans.
4.3.2.2. N credit for nitrogen fixation. The N credit contributed signifi­ The changes were not as influential to the climate change impacts of
cantly to the LCIA results, particularly for faba beans where it reduced dry beans, but still very significant when both reported N content in
the life cycle impacts in many categories by over 10 %, leading to more fertilizer and alternate N fixed values are considered in N credit esti­
beneficial outcomes overall. These benefits would be greater if using the mation. For example, climate change impacts reduced by 34 % for red
fertilizer mix applied to the beans as the N fertilizer credit, since the N kidney bean production in Ontario/ecozone 8 and by 23 % in Manitoba/
ecozone 10. Overall, N credits applied to account for nitrogen fixed by
the plant can have a significant impact on the overall environmental
Table 5
impacts reported. This shows that our original estimates were conser­
Results of sensitivity analysis showing the changes in soil carbon sequestered (kg
vative, which is a good approach in this case since they are not based on
CO2 sequestered per tonne of crop) due to modelling alternate crop rotations and
selecting alternate SLC polygons, using the second most common crop rotations
any farm-specific data other than the yield.
or SLC polygon locations from the survey responses, instead of the most common
in the original results. 4.3.3. Comparison to previous studies
The results of the current study were compared to the only other LCA
Region Carbon Carbon sequestered Carbon sequestered
sequestered (kg (kg CO2t− 1) – (kg CO2t− 1) –
CO2t− 1) – alternate crop alternate SLC
baseline rotation polygon
N credit sensivity analysis
Faba beans
200
Alberta 18.1 25.2 31.0
kg N ferlizer credit/tonne beans

180
Manitoba 21.7 22.6 46.5 160
Saskatchewan 23.1 24.4 47.9 140
Ecozone 9 35.7 36.2 67.8 120
Ecozone 10 24.4 20.8 44.7
100
80
60
40
Navy beans 20
Manitoba 90.6 29.5 47.2 0
Ontario 4.7 21.5 44.8
Ecozone 8 4.7 21.5 44.8
Ecozone 10 25.8 29.5 47.3

N credit N credit (alternate N%) N credit (alternate N fixed)

Pinto beans
Alberta 24.1 25.5 51.8
Manitoba 24.5 22.2 71.1 Fig. 7. Sensitivity analysis of nitrogen fixed, modelled as an N fertilizer credit:
Ecozone 10 33.0 30.0 68.3
Original N credit calculated from Barker (2007) for Top Crop Manage and
modelled as a credit of ammonia fertilizer, Alternate N% N credit calculated as
Red kidney beans a credit of the fertilizer mix applied to the beans instead of ammonia, and
Manitoba 27.6 8.8 51.3 Alternate N Fixed N credit calculated according to a newsletter from Sas­
Ontario 3.4 20.6 44.7 katchewan Pulse Growers for faba beans, and data from North Platte Natural
Ecozone 8 3.4 26.9 46.0
Resources District in the United States for dry beans (pers. comm.
Ecozone 10 27.6 6.8 49.2
Pulse Canada).

456
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

are best compared against the results of dry bean production in Canada.
It should be noted that the Greek study was done over 10 years before
Inoculant A) Climate change - Faba beans

the current study, analyzed different bean types, and its system
N credit 200

Post-harvest acvies 100

boundaries were significantly different (ESM 1 Section 2.1.5). Hence the
Soil carbon sequestraon 0 value of comparison to the current study is limited. Impact assessment
Irrigaon kg CO2e
-100 was done using the CML baseline methodology to enable comparison
Pescides
-200 with the Greek study. Table 6 presents the impact assessment results of
Ferlizer
-300
the Greek study, the current study with carbon sequestered and N fix­
Field operaons ation included, and the current study without the positive impacts of
carbon sequestration and N fixation – since the Greek study did not
-400
Seed

Agricultural land
-500
AB AB- AB-N MB MB- MB - SK SK-N% SK-N
account for either of these credits.
Total N% fixed N% N fixed Overall, the impacts reported for dry bean production in Greece were
lower than the impacts reported in this study. However, the Greek study
fixed

B) Climate change - Red kidney beans did not account for seed inputs and post harvest activities. Also, the
N credit study modelled inputs of sheep manure (considered a free input without
any impacts associated with its production) as a substitute for synthetic
500
Post-harvest acvies

Soil carbon sequestraon 400

nitrogen fertilizers, which likely reduced the overall impacts reported.
Irrigaon 300 Taking global warming potential (GWP) as an example, conventional
Pescides 200 plake bean production in Greece had an overall impact of 302 kg CO2-eq.
kg CO2e

Ferlizer
100
compared to 399 kg CO2-eq. (navy beans), 302 kg CO2-eq. (pinto beans),
Field operaons and 397 kg CO2-eq. (red kidney beans) without inclusion of carbon
sequestration and N credits. The climate change impact of gigantes
0
Seed

Agricultural land
-100 beans (247 kg CO2-eq.) was comparable to pinto beans. The results of
Total -200 other impact categories were varied, with the Greek study reporting
higher impacts than the current study for some categories (e.g., acidi­
EZ8 EZ8-N% EZ8-N EZ10 EZ10-N% EZ-N
fixed fixed
fication), and considerably lower impacts in other impact categories (e.
Fig. 8. LCIA results (IMPACT World+ Climate Change) for A) faba (provincial) g., freshwater ecotoxicity). Given the differences in system boundaries
and B) red kidney (ecozone) beans for sensitivity analysis of N credit estimates. and the different version of the impact assessment method used in this
Original N credit calculated from Barker (2007) for Top Crop Manage and study (CML 2015) compared to the Greek study (CML 2000), these
modelled as a credit of ammonia fertilizer, Alternate N% N credit calculated as comparisons should not be considered definitive.
a credit the fertilizer mix applied to the beans instead of ammonia, and Alter­ When considering the contributions to environmental impacts, the
nate N Fixed N credit calculated according to a newsletter from Saskatchewan
Greek study found that irrigation, sheep manure, and fuel inputs were
Pulse Growers for faba beans, and data from North Platte Natural Resources
the biggest contributors. The Greek study had considerably more irri­
District in the United States for dry beans (pers. comm. Pulse Canada).
gation and hence it is possible that irrigation was more influential in
Greek bean production than in Canada. The impacts associated with
of beans available in literature - Abeliotis et al. (2013) – which analyzed
sheep manure are likely to be considerably lower than the production
the production of gigantes and plake beans in Greece. Both gigantes and
and emissions associated with synthetic manure for dry bean production
plake (plaki in Greek) are Greek dishes that are commonly made with
in Canada. Since the study did not report any allocation procedure for
dry beans such as white runner beans. So, the results of the Greek study
sheep manure, it appears that manure was modelled as an input without

Table 6
Comparison of CML baseline 2000 impact assessment results of current study (with and without N credit and C sequestration) to the impacts of bean production in
Greece (Abeliotis et al., 2013).
All results Impact Abiotic Global Ozone Human Freshwater Marine Terrestrial Photochemical Acidification Eutrophication
per tonne category depletion warming layer toxicity aquatic aquatic ecotoxicity oxidation
of bean depletion ecotoxicity ecotoxicity
produced
Indicator kg Sb eq. kg CO2 kg CFC- kg 1,4- kg 1,4-DB kg 1,4-DB kg 1,4-DB kg C2H2 eq. kg SO2 eq. kg PO4 eq.
(CML
unit eq. 11 eq. DB eq. eq. eq. eq.
2000)
2
Abeliotis Gigantes 0.732 247 4.74 × 9.94 − 0.1 40,000 1.28 × 1.25 × 10− 13.2 2.2
et al. convent. 10− 6 10–1
2
(2013) Plake 0.866 302 5.74 × 12 − 0.0117 48,400 0.155 1.46 × 10− 16.2 2.7
convent. 10− 6
This study Faba 1.09 × 57.30 1.08 × 41.21 89.81 47,016 1.01 0.01 0.59 1.48
10− 5 10− 5
Navy 3.48 × 350.14 3.31 × 150.28 433.93 133,294 5.42 0.03 4.63 5.74
10− 5 10− 5
Pinto 3.59 × 221.49 2.51 × 177.59 227.69 151,915 2.60 0.02 3.16 4.63
10− 5 10− 5
Red 3.53 × 337.42 3.72 × 203.08 438.87 162,828 5.79 0.03 4.01 5.21
Kidney 10− 5 10− 5
This study Faba 1.26 × 89.10 1.20 × 45.27 92.06 50,257 1.05 0.01 0.61 1.49
(no N or 10− 5 10− 5
C Navy 3.57 × 399.10 3.43 × 152.60 435.12 135,084 5.44 0.03 4.65 5.74
credits) 10− 5 10− 5
Pinto 3.68 × 302.09 2.57 × 179.67 228.84 153,569 2.62 0.03 3.17 4.63
10− 5 10− 5
Red 3.62 × 396.96 3.85 × 205.42 440.07 164,637 5.81 0.04 4.02 5.21
Kidney 10− 5 10− 5

457
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

any upstream impacts associated with its production i.e., it was 5. Conclusions
considered a waste stream from sheep farming. Since the paper did not
report on the source of the sheep manure, it is difficult to ascertain if this The cradle-to-farm gate life cycle environmental impacts of Cana­
modelling choice is acceptable but, nevertheless, it is likely the major dian faba and dry bean production were assessed using field-level data
source of differences in impacts reported in the two studies. On the other supplied by Canadian pulse farmers aggregated at the ecozone, pro­
hand, both studies report that the fuel used in farm machinery as a major vincial, and national scales, along with well-established emissions
source of impacts. modelling methodologies. The main contributors to the impacts of both
faba and dry bean production were fertilizer and fuel use for field ac­
tivities. Fertilizer production required large amounts of water, land, and
4.4. Limitations mineral resources, and was responsible for many of the resulting emis­
sions during the fertilizer production processes. This contributed sub­
While crop protection does not feature prominently in results for stantially to acidification, ionizing radiation and ecotoxicity emissions.
many of the impact categories assessed, it is important to note that these Fertilizer application on the farm also resulted in field-level emissions
results are conservative estimates since several herbicide, fungicide and such as N2O, CO2, P and NH3, which contributed to climate change,
insecticide emissions were unspecified in the available ImpactWorld+ eutrophication, and acidification impacts. Fuel extraction and produc­
LCIA methods. Hence insufficient data availability could be a limiting tion also contributed substantially to the demand for water, land, and
factor with respect to evaluating the potential impacts of crop protection fossil energy resources. The combustion of the fuel on the farm emitted
products. e.g., CO2, NOx and particulate emissions that contributed to climate
The study was also limited with respect to access to accurate farm- change, photochemical oxidation, ozone layer depletion and particulate
level energy use data for both field operations and post-harvest activ­ matter formation.
ities. As an estimation, activity data were collected from farmers, and Although the trends were similar, there were variations in results
fuel efficiency data from the Ontario Ministry of Agriculture, Food and between different crop types, regions, and levels of spatial aggregation.
Rural Affairs (OMAFRA 2020) were used to calculate total fuel use. The Overall, the impacts for red kidney bean production were higher than
single source from OMAFRA was used to estimate differences between impacts of pinto and navy bean production, except for navy bean pro­
bean types and regions based on differences in field activities and ma­ duction in Ontario, which generally had the highest impacts. Impacts for
chinery used, rather than also including the different methodologies faba beans were generally lower than dry beans due to their lower input
employed by each province to estimate fuel use efficiency. To better requirements, particularly for fertilizer, as well as the considerable N
assess the impacts of field activities, a dedicated effort is required to credit associated with faba bean cultivation. With faba beans, a much
collect energy use data directly from a sufficient sample of farms pro­ higher proportion of farms reported performing post-harvest aeration
ducing each bean type in each region. For post-harvest activities, data and drying compared to all other beans, but this did not contribute
used for aerating and drying wheat were used as substitutes (Team significantly to the overall impacts.
Alberta, 2021) for data for bean drying. To improve the accuracy of While there were some variations between crops, regions and levels
energy use estimates for aerating and drying beans, dedicated mea­ of regional aggregation, similar trends were seen in terms of the highest
surement of energy consumption from bean aeration and drying facil­ contributors to the estimated impacts. In general, fuel and fertilizer were
ities is required. Experts highlighted that often these post-harvest identified as hotspots for potential reductions in emissions and resource
activities are done outside the farm and hence farmers have limited use. Adopting best management practices may enable reducing impacts.
capacity to provide accurate data – thus necessitating the use of sec­ This could include, for example, implementation of soil N testing, re­
ondary data sources. ductions in fertilizer application, precision agriculture, crop rotation
The ICBM model in the Holos tool, used to calculate soil carbon decisions, adoptions of reduced tillage or no tillage strategies, or other
sequestration, uses soil data at a SLC polygon scale, whereas the LCI data reductions in mechanized field operations. In addition, changes to land
in this report were aggregated at provincial or ecozone scale. Therefore, management practices could be used in combination with lower fertil­
the most common SLC polygons on which farms were located for each izer inputs to reduce the field-level emissions associated with fertilizer
province and ecozone were used in the ICBM soil carbon calculations. application and land management. Further research is required to
This methodology precludes adequately accounting for some of the identify the most regionally applicable strategies and assess the life cycle
variability in soil conditions at province and ecozone scale. Similarly, impacts of their implementation. Finally, considering the significant N
the ICBM model in Holos requires crop rotation data. However, there credit associated with cultivating faba beans as part of crop rotations
were a large variety of crop rotations reported by farmers. Therefore, the (which the sensitivity analysis showed could be even higher than re­
most common responses were included in the ICBM model and averaged ported), it would appear that expanded cultivation of faba beans (pro­
to estimate the potential soil carbon sequestration levels in each region. vided sufficient markets and regions providing conducive growing
This may underestimate the actual variability in crop rotations in each conditions are available) could provide substantial environmental ben­
region for each bean type. The results of the sensitivity analysis showed efits in the context of Canadian field crop production.
that these factors, particularly the choice of SLC polygon, had a non- The results of the sensitivity analyses show the importance of as­
trivial impact on the estimates of soil carbon sequestration, and over­ sumptions made with regard to modelling soil dynamics such as soil
all life cycle GHG emissions, further underscoring the essential nature of carbon sequestration or emissions, and biological nitrogen fixation
highly regionalized life cycle inventory modelling to accurately account leading to N credits for the next crops in rotation. Altering assumptions
for the context-specific environmental performance profile of agricul­ around both parameters resulted in significant variability in results. The
tural field crop products. Therefore, it may be more accurate to auto­ results of agricultural LCAs, such as this one, are often highly dependent
mate the calculation of soil carbon sequestration to get an estimate from on such field-level dynamics. Therefore, soil measurements and model
each farm before aggregating data to the ecozone or provincial scale. validation should be a priority for future agricultural sustainability
Finally, assumed nitrogen fixation levels and associated fertilizer research. Additionally, LCIA methods do not fully account for all system
credits can clearly have a large influence on the overall estimated im­ dynamics associated with field-level processes such as biological nitro­
pacts of bean production, especially faba beans. In the current study, this gen fixation. Therefore, consideration should also be given to a wider
was calculated using generic estimates of the amount of N fixed by crop range of ecosystem services and functions when interpreting the results
type, reported yield, and the nitrogen content of ammonia fertilizer. of agricultural LCAs in the context of informing policy and agricultural
Actual measurements, or more regionalized calculations, are required to decision-making.
increase the reliability of the results in this regard.

458
N. Bamber et al. Sustainable Production and Consumption 46 (2024) 442–459

Declaration of competing interest Hergoualc’h, Kristell, Akiyama, Hiroko, Bernoux, Martial, Chirinda, Ngonidzashe, del
Prado, Agustin, Kasimir, Åsa, MacDonald, James Douglas, Ogle, Stephen Michael,
Kristiina Regina, T.J. van der W., 2019. 2019 Refinement to the 2006 IPCC
The authors declare the following financial interests/personal re­ Guidelines for National Greenhouse Gas Inventories.
lationships which may be considered as potential competing interests: Hossain, Z., Wang, X., Hamel, C., et al., 2016. Biological nitrogen fixation by pulse crops
Nicole Bamber reports financial support was provided by Pulse Canada. on semiarid Canadian prairies. Can. J. Plant Sci. 97, 119–131. https://doi.org/
10.1139/cjps-2016-0185.
If there are other authors, they declare that they have no known ISO, 2006. ISO 14044: Environmental Management - Life Cycle Assessment -
competing financial interests or personal relationships that could have Requirements and Guilelines. Int. Organ. Stand. (ISO), Geneva, Switz.
appeared to influence the work reported in this paper. Karlsson, H., Ahlgren, S., Strid, I., Hansson, P.A., 2015. Faba beans for biorefinery
feedstock or feed? Greenhouse gas and energy balances of different applications.
Agric. Syst. 141, 138–148. https://doi.org/10.1016/j.agsy.2015.10.004.
Appendix A. Supplementary data Köpke, U., Nemecek, T., 2010. Ecological services of faba bean. F Crop Res 115,
217–233. https://doi.org/10.1016/j.fcr.2009.10.012.
Kröbel, R., Bolinder, M.A., Janzen, H.H., et al., 2016. Canadian farm-level soil carbon
Supplementary data to this article can be found online at https://doi. change assessment by merging the greenhouse gas model Holos with the
org/10.1016/j.spc.2024.02.034. Introductory Carbon Balance Model (ICBM). AGSY 143, 76–85. https://doi.org/
10.1016/j.agsy.2015.12.010.
Lemke, R.L., Zhong, Z., Campbell, C.A., Zentner, R., 2007. Can pulse crops play a role in
References mitigating greenhouse gases from North American agriculture? Agron. J. 99 (6),
1719–1725. https://doi.org/10.2134/agronj2006.0327s.
AAFC, 2022. Canada: Outlook for Principal Field Crops, 03-18-2022. MacWilliam, S., Wismer, M., Kulshreshtha, S., 2014. Life cycle and economic assessment
Abeliotis, K., Detsis, V., Pappia, C., 2013. Life cycle assessment of bean production in the of Western Canadian pulse systems: the inclusion of pulses in crop rotations. Agric.
Prespa National Park, Greece. J. Clean. Prod. 41, 89–96. https://doi.org/10.1016/j. Syst. 123, 43–53. https://doi.org/10.1016/j.agsy.2013.08.009.
jclepro.2012.09.032. MacWilliam, S., Parker, D., Marinangeli, C.P.F., Trémorin, D., 2018. A meta-analysis
Alberta Agriculture and Forestry, 2018. Life Cycle Assessment of Alberta Pea Production. approach to examining the greenhouse gas implications of including dry peas (Pisum
Bamber, N., Turner, I., Arulnathan, V., et al., 2020. Comparing sources and analysis of sativum L.) and lentils (Lens culinaris M.) in crop rotations in western Canada. Agric.
uncertainty in consequential and attributional life cycle assessment: review of Syst. 166, 101–110. https://doi.org/10.1016/j.agsy.2018.07.016.
current practice and recommendations. Int. J. Life Cycle Assess. 25, 168–180. Margni, M., Rossier, D., Crettaz, P., Jolliet, O., 2002. Life cycle impact assessment of
https://doi.org/10.1007/s11367-019-01663-1. pesticides on human health and ecosystems. Agric. Ecosyst. Environ. 93, 379–392.
Bamber, N., Dutta, B., Heidari, M.D., et al., 2022. Spatially resolved inventory and https://doi.org/10.1016/S0167-8809(01)00336-X.
emissions modelling for pea and lentil life cycle assessment. Sustain Prod Consum Martin-gorriz, B., Zabala, J.A., Sánchez-navarro, V., et al., 2022. Intercropping practices
33, 738–755. https://doi.org/10.1016/j.spc.2022.07.030. in Mediterranean mandarin orchards from an environmental and economic
Bandekar, P.A., Putman, B., Thoma, G., Matlock, M., 2022. Cradle-to-grave life cycle perspective. Agric 12, 1–17. https://doi.org/10.3390/agriculture12050574.
assessment of production and consumption of pulses in the United States. J. Environ. Nitschelm, L., Aubin, J., Corson, M.S., et al., 2016. Spatial differentiation in Life Cycle
Manag. 302, 114062 https://doi.org/10.1016/j.jenvman.2021.114062. Assessment LCA applied to an agricultural territory: current practices and method
Barker, B., 2007. Nitrogen credit. https://www.topcropmanager.com/nitrogen-credit development. J. Clean. Prod. 112, 2472–2484. https://doi.org/10.1016/j.
-906/. jclepro.2015.09.138.
Bulle, C., Margni, M., Patouillard, L., et al., 2019. IMPACT World+: a globally Ojiewo, C., Keatinge, D.J.D.H., Hughes, J., et al., 2015. The role of vegetables and
regionalized life cycle impact assessment method. Int. J. Life Cycle Assess. 24, legumes in assuring food, nutrition, and income security for vulnerable groups in
1653–1674. https://doi.org/10.1007/s11367-019-01583-0. Sub-Saharan Africa. World Med. Health Policy 7, 187–210. https://doi.org/
Campbell, C.A., Zentner, R.P., Selles, F., Biederbeck, V.O., McConkey, B.G., Blomert, B., 10.1002/wmh3.148.
Jefferson, P.G., 2000. Quantifying short-term effects of crop rotations on soil organic Pulse Canada, 2021. Growing regions. http://bouncebeta.com/stage/pulses/producers
carbon in southwestern Saskatchewan. Can. J. Soil Sci. 80, 193–202. https://doi. -industry/about-pulses/growing-regions/.
org/10.4141/S99-045. Rochette, P., Liang, C., Pelster, D., et al., 2018. Soil nitrous oxide emissions from
Canadian Grain Commission, 2016. Tough and damp moisture ranges for Canadian agricultural soils in Canada: exploring relationships with soil, crop and climatic
grains. https://www.grainscanada.gc.ca/en/grain-quality/grain-grading/gradin variables. Agric. Ecosyst. Environ. 254, 69–81. https://doi.org/10.1016/j.
g-factors/moisture-content/tough-damp-ranges.html. agee.2017.10.021.
Ciroth, A., Muller, S., Weidema, B., Lesage, P., 2016. Empirically based uncertainty Saskatchewan Pulse Growers, 2018. Post-harvest Storage of Pulses.
factors for the pedigree matrix in ecoinvent. Int. J. Life Cycle Assess. 21, 1338–1348. Statistics Canada, 2015. Pulses in Canada. https://www150.statcan.gc.ca/n1/pub/
https://doi.org/10.1007/S11367-013-0670-5/TABLES/6. 96-325-x/2014001/article/14041-eng.htm. (Accessed November 2023).
EarthShift Global, 2019. Milestone 3: Methodology for Low Carbon Fuel Pathways and Statistics Canada, 2023. Canadian Agriculture at a Glance.
Default Carbon Intensities. Team Alberta, 2021. Evaluating Energy Efficiency of On-farm Grain Conditioning
Frischknecht, R., Rebitzer, G., 2005. The ecoinvent database system: a comprehensive Systems.
web-based LCA database. J. Clean. Prod. https://doi.org/10.1016/j. Thiagarajan, A., Fan, J., McConkey, B.G., et al., 2018. Dry matter partitioning and
jclepro.2005.05.002. residue N content for 11 major field crops in Canada adjusted for rooting depth and
Frischknecht, R., Jungbluth, N., Althaus, H.J., Doka, G., Dones, R., Heck, T., Hellweg, S., yield. Can. J. Soil Sci. 98, 574–579. https://doi.org/10.1139/cjss-2017-0144.
Hischier, R., Nemecek, T., Rebitzer, G., Spielmann, M., 2005. The ecoinvent Vogelsang-o’Dwyer, M., Petersen, I.L., Joehnke, M.S., et al., 2020. Comparison of faba
database: overview and methodological framework. Int. J. Life Cycle Assess. 10 (1), bean protein ingredients environmental performance. Foods 9, 322.
3–9. https://doi.org/10.1065/lca2004.10.181.1. Yang, Y., Tao, M., Suh, S., 2018. Geographic variability of agriculture requires sector-
Government of Canada, 2021. GHG Inventory and Emission Factors. specific uncertainty characterization. Int. J. Life Cycle Assess. 23, 1581–1589.
GTT, 2022. Global Trade Tracker. https://doi.org/10.1007/s11367-017-1388-6.
Guinée, J., 2001. Handbook on Life Cyle Assessment — Operational Guide to the ISO
Standards. Editorial in Int J LCA 6 (5), 255.

459