IRRIGATION AND DRAINAGE

Irrig. and Drain. 67: 790–799 (2018)

Published online 12 September 2018 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/ird.2285

COMPARING THE USEFULNESS AND APPLICABILITY OF DIFFERENT WATER

FOOTPRINT METHODOLOGIES FOR SUSTAINABLE WATER MANAGEMENT IN
AGRICULTURE†

BETSIE LE ROUX1 , MICHAEL VAN DER LAAN1* , MARK B. GUSH2 AND KEITH L. BRISTOW1,3
1
Department of Plant and Soil Sciences, University of Pretoria, Hatfield, South Africa
2
CSIR Natural Resources and Environment Division, Stellenbosch, South Africa
3
CSIRO Agriculture and Food, Townsville, Australia

ABSTRACT
The lack of sustainability of our water resources threatens food security in many places worldwide. Different water footprint
(WF) methodologies were investigated for their ability to improve water management at various scales. Methodologies
according to the Water Footprint Network (WFN), life cycle assessment (LCA) and hydrological-based approaches were
assessed, and a working example is given for apples produced in South Africa. A fundamental viewpoint was defined and
the knowledge hierarchy applied to investigate the approaches and information generated. WFs reported simply as a volume
of water used per mass of crop produced cannot indicate the sustainability of the water use unless interpreted within the local
hydrological and environmental context. The WFN methodology appears most useful to resource managers due to its
quantitative nature and ability to compare blue and green water consumption versus water availability. The LCA approach
may be best for comparisons of the impact of different products. None of the methodologies provides a single metric that
can be used to inform wise consumer choices as it is not possible to incorporate all the complexities associated with water
use into a single number that can be used to inform sustainable water use. © 2018 John Wiley & Sons, Ltd.
key words: sustainable water use; knowledge hierarchy; Water Footprint Network; life cycle assessment; consumer awareness

Received 22 November 2017; Revised 16 July 2018; Accepted 19 July 2018

RÉSUMÉ
Le manque de durabilité de nos ressources en eau menace la sécurité alimentaire dans de nombreux endroits du monde.
Différentes méthodologies de l’empreinte d’eau ont été étudiées pour leur capacité à améliorer la gestion de l’eau à différentes
échelles. La méthodologie selon le water footprint network (WFN), l’analyse du cycle de vie (ACV) et les approches
hydrologiques ont été évaluées, et un exemple pratique est donné pour les pommes produites en Afrique du Sud. Un point de
vue fondamental a été défini et la hiérarchie des connaissances a été appliquée pour étudier les approches et les informations
générées. Les WF déclarés simplement comme un volume d’eau utilisé par masse de culture produite ne peuvent pas indiquer
la durabilité de l’utilisation de l’eau à moins d’être interprétés dans le contexte local hydrologique et environnemental. La
méthodologie WFN semble la plus utile aux gestionnaires de ressources en raison de sa nature quantitative et de sa capacité à
comparer la consommation d’eau bleue et verte par rapport à la disponibilité de l’eau. L’approche ACV peut être la meilleure pour
comparer l’impact de différents produits. Aucune des méthodologies ne fournit une seule mesure qui peut être utilisée pour
éclairer les choix judicieux des consommateurs, car il n’est pas possible d’incorporer toutes les complexités associées à
l’utilisation de l’eau dans un seul chiffre qui peut être utilisé pour une utilisation durable de l’eau. © 2018 John Wiley & Sons, Ltd.
mots clés: utilisation durable de l’eau; hiérarchie des connaissances; Water Footprint Network; évaluation du cycle de vie; sensibilisation des consommateurs

*Correspondence to: Dr Michael van der Laan, Department of Plant and Soil Sciences, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa.
+27124203665. E-mail: [email protected]
†
Comparaison de l’utilité et de l’applicabilité de différentes méthodes d’empreinte de l’eau pour une gestion durable de l’eau en agriculture.
Contract/grant sponsor: Water Research Commission; contract/grant number: WRC project No. K5/2273//4

© 2018 John Wiley & Sons, Ltd.

 USING WATER FOOTPRINT METHODS IN AGRICULTURE 791

INTRODUCTION of journals (Hoekstra, 2016; Hoekstra et al., 2009a; Pfister

et al., 2017; Pfister and Hellweg, 2009). Several authors
Water scarcity is increasing in many places around the
criticized the WFN approach for producing information that
world. Climate change, population growth and improved
is misleading, because without the sustainability assessment
standards of living will exacerbate this even further in the
these authors claim the WFs do not reflect the local
future. Future water scarcities will present many challenges,
environmental and hydrological conditions where the water
of which global food production is of specific concern. The
is used (Perry, 2014; Wichelns, 2011a, 2011b).
water footprint (WF) concept was developed to help address
The International Commission on Irrigation and
these challenges.
Drainage (ICID) has been assisting the international irriga-
In 2009 the Water Footprint Network (WFN) (now oper-
tion community with definitions and applications of the
ating as the Water Footprint Research Alliance) published
concepts of water use efficiency and water productivity.
the first WF assessment manual (Hoekstra et al., 2009b),
Further clarity is now required to understand the definitions
which was followed by a more comprehensive manual in
and applications of WFs. The WFN, LCA and
2011 containing prescribed methodology to determine the
hydrological-based methodologies are described and evalu-
impact on water resources by individuals, communities,
ated in a simple case study for apples (Malus pumila)
businesses as well as during production processes (Hoekstra
grown in the Kouebokkeveld, South Africa. To better
et al., 2011). Hoekstra et al. (2011) distinguish between
understand the different approaches and the information
blue, green and grey WFs. Surface water and groundwater
generated, a fundamental viewpoint was defined and the
resources, which are available to multiple users, are defined
knowledge hierarchy applied. Strengths and weaknesses of
as blue water. In a crop production context, the blue WF
each approach are discussed in an attempt to identify the
consists predominantly of the irrigation water consumed.
most appropriate methodology for a specific application as
Green water is water originating from rainfall that is stored
well as the intended usefulness of the information. Through
in the soil and available for vegetation growth only.
this analysis, we hope to improve communication and un-
Hoekstra et al. (2011) proposed the concept of a grey WF,
derstanding between scientists involved in the WF debate.
which is the volume of water required to dilute pollutants
to ambient levels. Expressing water pollution impact in this
way enables the reporting of a total (blue + green + grey) MATERIALS AND METHODS
WF as a volume which includes water quality and quantity
Water Footprint Network (WFN) methodology
impacts.
Modified WF methodologies have been proposed in an at- The WFN proposes two phases, firstly a WF accounting
tempt to overcome some of the weaknesses of the WFN ap- phase and secondly a sustainability assessment of the WF.
proach, most notably the life cycle assessment (LCA) During the accounting phase, a volume of water used per
(Canals et al., 2009; Pfister et al., 2009), and the product yield is determined. During the sustainability as-
hydrological-based methodologies (Deurer et al., 2011). In sessment phase, the water use is compared to water
2014, the first International Standards Organization WF availability to reflect the local impact of the WF (Hoekstra
standard was released (ISO 14046, 2014), which, while et al., 2011).
somewhat vague in its prescription on how to calculate a The equations for blue and green WFs (Equations (1) and
WF, aligns most closely with the LCA approach. Both (2)) were taken from the WFN manual. According to
before and after this publication, there has been scientific de- Equations (1) and (2), the blue plus green WFs include water
bate between proponents supporting and opposing WF evapotranspired by the crops during cultivation, taking into
methodologies, including in the form of letters to the editors account how much of the water was provided by irrigation:

minimum ðtotal net irrigation; actual irrigation requirementÞ

Blue WF ¼ (1)
Yield

and

½Crop ET  minimum ðtotal net irrigation; actual irrigation requirementÞ

Green WF ¼ (2)
Yield

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
792 B. LE ROUX ET AL.

where Crop ET is the total crop evapotranspiration (mm). land use, and because natural vegetation would have used
The result of the blue plus green WF is therefore equal to ac- green water anyway. An alternative to the grey WF method
tual crop ET/crop yield, and is also called the volumetric is proposed by Ridoutt and Pfister (2013), which makes use
WF. Another method is described in Hoekstra et al. (2011), of advanced LCA modelling using eutrophication,
which makes use of crop ET and effective rainfall to calcu- freshwater ecotoxicity and human health impacts as impact
late the blue and green WFs. Equation (3) shows the formula indicators (Ridoutt and Pfister, 2013).
that is used to calculate the grey WF (Hoekstra et al., 2011): The International Standards Organization (ISO) published
a ‘Global WF Standard’ in August 2014 (ISO 14046, 2014),
L
Grey WF ¼ (3) which is closely related to the LCA method proposed by
C max  C nat Pfister et al. (2009). The standard gives broad and flexible
where L (kg ha-1) is the load of pollutant released into a water guidelines and includes a few important principles. Water
body, Cmax (kg m-3) the maximum concentration of pollutant footprints, according to the standard, must consider the full
at ambient water quality standards, and Cnat (kg m-3) the nat- life cycle of a product, must include an environmental im-
ural concentration of the pollutant in the receiving water. The pact assessment and must preferably be based on scientific
natural concentration of the pollutant in the receiving water evidence. The standard also has specifications on how
and the ambient water quality standards differ from one WFs are reported, in order to ensure transparency, but does
country to the next according to their individual guidelines. not include a way to report the results as product labels.
Therefore, the same pollutant load can have different grey Similar to the LCA methodology of Pfister et al. (2009), it
WFs depending on the natural background concentration is suggested that results be reported as ‘water equivalents’
and the user-selected water quality standards (Hoekstra (H2O-e). The standard also proposes the use of other mid-
et al., 2011). The blue, green and grey volumetric WF is di- point indicators firmly established in LCA methodology,
vided by crop yield (t ha-1) to give units in volume of water such as estimating eutrophication potential in ‘phosphate
per mass of crop yield (Hoekstra et al., 2011). equivalents’ in the case of nitrogen and phosphorous pollu-
tion from agriculture.
Life Cycle Assessment methodology
Hydrological-based methodology
The LCA community suggested that crop WFs according to
the WFN (a volume of water used per yield of crop) can be Deurer et al. (2011) introduced a WF method based on hy-
misleading if communicated outside the context of the local drology, aiming at considering all components of the water
circumstances and without some type of sustainability as- balance and not just water consumption. The calculation of
sessment. This community also felt they had better assess- the blue WF is based on Equation (5):
ments for the impact on water quality than the grey WF
concept. Pfister et al. (2009) suggested a WF method based Δ Blue water ¼ Dr þ Dir þ Rr þ Rir  IR (5)
on the LCA approach, in which a regional water stress index where Dr is drainage under rainfed conditions, Dir is the
(WSI) is calculated to characterize local water use impacts, difference between drainage under rainfed and irrigated con-
which follows a logistic function from 0.01 to 1, with a ditions, Rr is runoff under rainfed conditions and Rir is the
withdrawal-to-availability ratio of 0.4 (often referred to as difference between runoff under rainfed and irrigated condi-
the threshold between moderate and severe water stress) tions. Drainage and runoff collectively forms the inflow into
resulting in a WSI of 0.5. The LCA method attempts to ad- the blue water resource. IR is the amount of water abstracted
dress temporal variation by including a variation factor into from the blue water resource for irrigation and represents the
the WSI as a measure of variation in climatic conditions. outflow from the blue water resource. The calculation of the
Variation in water availability will result in a higher varia- green WF is based on Equation (6):
tion factor and a higher WSI which will in turn increase
WFs. The LCA water scarcity-weighted WF is calculated Δ Green water ¼ Dr þ ETr þ Rr  RF (6)
using Equation (4): r
where ET is the ET under rainfed conditions and RF is the
1

effective rainfall, excluding any water that is intercepted by
Blue WF ¼ Water Consumption m Functional Unit
3
WSI
the plant cover. Collectively Dr, ETr and Rr forms the out-
(4)
flows from the green water resource and RF is the inflow
The results are a stress-weighted index reported as ‘water into the green water resource.
equivalents’ (H2O-e) which gives an indication of a product According to this method, a negative WF is possible if the
or activities’ impact on water resources (Ridoutt and Pfister, recharge of the blue water resource through return flows and
2010). For this approach, green water use is not considered precipitation exceeds the volumes abstracted. A negative
because it is accounted for in other LCA metrics, such as blue WF is therefore required for a catchment or aquifer to

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
 USING WATER FOOTPRINT METHODS IN AGRICULTURE 793

sustain ecosystems downstream. A positive WF indicates area under irrigation in this WMA at 730 km2. Water foot-
water abstraction exceeds recharge through return flows print methodologies were applied to apple production in
and precipitation (Deurer et al., 2011). A zero WF is possi- the Olifants-Doorn WMA to assess their ability to improve
ble if return flows and precipitation are equal to abstraction water resource management of agriculture in this water-
volumes. Data used to calculate hydrological-based WFs are scarce region (Van der Laan, 2017).
obtained on a local scale and over an annual water cycle
(Herath et al., 2014). Blue and green water use are divided Fundamental viewpoint
by yield to obtain the WFs. Grey WFs are calculated in
the same way as proposed by Hoekstra et al. (2011). The fundamental viewpoint is that WF assessments or
metrics must primarily promote sustainable water use
(Figure 1: Level A). Each WF method was evaluated ac-
Case study—the water footprint of apples according to
cording to this viewpoint. Sustainable water use is deter-
different methodologies
mined by several variables (Figure 1: Level B), including:
The Olifants-Doorn water management area (WMA),
located in the Western Cape Province of South Africa, is a re- • variables in the hydrological system that determine wa-
gion experiencing extreme water scarcity. Climatic condi- ter availability. Such variables include climatic/weather
tions vary considerably across the Olifants-Doorn WMA as conditions, soil types, topography, landscape character-
a result of the variation in topography. The mean annual pre- istics and land use;
cipitation (MAP) ranges from approximately 1500 mm in the • variables that define the systems that determine water
south-west to less than 100 mm in the far north (DEA&DP, demands and receive impacts from water use. Such
2011). With the MAP over much of the WMA being less variables include the ecological, social and economic
than 200 mm, the result is that, except in the wetter south- systems (including agriculture);
west, the climate is not suitable for dryland farming on a • variables related to water use, including water use
large scale. Irrigation in the WMA depends heavily on management (catchment and national level), water use
surface water (76%) and groundwater (16%) as sources of efficiency and water productivity (field and consumer
supply. The deciduous and citrus fruit industries are particu- level).
larly at risk as they utilize large volumes of water per unit
weight of fruit produced (Dzikiti and Schachtschneider, Most of these variables are difficult or impossible to ma-
2015). Consequently, more than 90% of the land in the nipulate, but more efficient water use management can be
Olifants-Doorn WMA is used as grazing for livestock, enforced through policies and regulation, and water use
predominantly for sheep and goats. However, the principal efficiency can be achieved through increasing public and
economic activity in the WMA is irrigated agriculture, and commercial enterprise awareness (Figure 1: Level C). In or-
87% of total water use is for irrigation (DEA&DP, 2011). der to manage and increase the efficiency of water use, the
A recent estimate (Bailey and Pitman, 2015) puts the total volumes of water consumed and polluted must be measured

Figure 1. Schematic representation of the role of water footprint assessments towards the goal of sustainable water use. Blue ovals are variables that impact on
sustainable water use, but are difficult or impossible to manipulate. Red rectangular boxes indicate variables that impact sustainable water use. [Colour figure
can be viewed at wileyonlinelibrary.com]

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
794 B. LE ROUX ET AL.

and characterized according to local water resource avail- developed to quantify these needs. Soil type is captured by
ability and sensitivity (Figure 1: Levels E and F). Impact the green WF, more green water will be available for soils
characterization should be informed by both the hydrologi- with a higher water-holding capacity. If all relevant informa-
cal system that influences water availability, as well as the tion, such as landscape and land-use effects on runoff, is in-
environmental setting where water is required. Current and cluded in the calculation of water availability, all aspects of
future water demands and management practices should also the hydrological cycle will be included. The method pro-
be considered as part of the environmental assessment and poses the concept of virtual water, which can inform policies
sustainability (Figure 1: Level B). by providing a simple universal way to estimate consump-
By means of a literature review the fundamental tive use per yield of product, and can assist catchment man-
viewpoints of the three WF methods were compared with agement practices. The WFN has been successful in raising
the fundamental viewpoint defined for this review as illus- awareness of water use impacts as a result of water con-
trated in Figure 1 and discussed below. It must be noted that sumed to make and distribute products. Whether the WFs
the similarities investigated here are only a reflection of the that are communicated to consumers will enable them to
aspects that are considered by the various methods and not make wise decisions that will lead to more sustainable water
an indication of how successfully these aspects are measured. use is uncertain, especially if they are used outside the local
context or without the sustainability assessment (Figure 1).
Compared to the fundamental viewpoint in Figure 1, the
RESULTS LCA approach includes some aspects of the hydrology. Wa-
ter availability is determined using monthly and annual rain-
Case study—the water footprint of apples according to
fall data. Landscape characteristics are considered in terms
different methodologies
of stream flow regulation in the particular catchment. Green
For the 12-year-old apple orchard-scale WF, taking into ac- water is excluded, because it can only be accessed through
count all water uses and a fruit yield of 61.5 t ha-1, according occupation of land. Therefore, the effects of soil types and
to the WFN approach we estimated a blue WF of 133 m3 t-1 topography are not addressed. This indicator includes many
and a green WF of 32 m3 t-1. To estimate the blue WF ac- aspects of the environment and people, by determining im-
cording to the LCA method, the blue WF was multiplied pacts on human health (social need), ecosystem quality
by the WSI for the region (0.78) to get a value of 104 (ecological sustainability) and resource depletion (economic
H2O-e t-1. For the hydrological-based approach, the blue requirements). However, ecological, social and economic
WF was estimated to be 41 m3 t-1. It was, however, not systems are extremely complex and measurements of these
possible to calculate green WFs according to the hydrologi- endpoint indicators are mostly calculated with many uncer-
cal approach in the same way as the published methodology tainties (Goedkoop et al., 2013). The WSI generated by this
prescribes, because modelling under rainfed-only conditions method can theoretically be used on product labels for
is required, and as this crop grows in summer in a region awareness raising, but it cannot really contain all the infor-
with a winter rainfall, a marketable yield will not be mation needed by consumers to make wise decisions that
achieved with which to calculate the green WF. A similar will lead to sustainable water use with all the complexities
situation developed when trying to estimate green WFs of this involves (Figure 2).
vegetable crops grown in winter in a summer rainfall region Compared to the fundamental viewpoint in Figure 1, the
(Van der Laan, 2017). As pointed out by a reviewer, it is hydrological approach provides information on the local cli-
difficult to learn from numbers that are completely different; mate and geographical features that determine water inputs,
not even the signs are the same. outputs and changes in water storage to produce a sustain-
ability indicator that includes all components of the hydro-
logical cycle. The sustainability indicator of this approach
Analysis in terms of the fundamental viewpoint
does not address social needs and economic requirements.
The WFN proposed a useful way to measure water con- Ecological impacts due to pollution are taken into account
sumption and pollution. The method differs from the funda- through the grey WF, but the impacts on ecosystem sustain-
mental viewpoint in that the WF is calculated before the ability due to a reduction in water availability and changes
sustainability assessment, and is therefore not a sustaina- in river flows are not yet considered. It also does not
bility indicator in itself (note the inverse in Levels D and consider current and future water use and environmental
E for Figure 1). Ecological impacts are accounted for by management practices, and the effectiveness of the approach
considering water quality impacts through grey WFs and to inform consumers to make wise decisions on their water
impacts due to consumption by including ecological flow re- uses is not clear (Figure 3).
quirements. The method mentions the need to reserve flow Although the volumetric WF according to the WFN,
for basic human and ecological needs, but has not yet been which is a volume of water used by the crop (ET)/mass of

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
 USING WATER FOOTPRINT METHODS IN AGRICULTURE 795

Figure 2. Similarities between the fundamental viewpoints of the WFN (2.1), LCA (2.2) and hydrological (2.3) methodologies and the fundamental viewpoint
proposed in Figure 1. Hexagon shapes (green) indicate similarities in the viewpoints, pentagon shapes (red) indicate aspects partly included, and rectangular
shapes (orange) indicate aspects lacking in each of the methodologies. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 3. The knowledge hierarchy (Rowley, 2007). [Colour figure can be viewed at wileyonlinelibrary.com]

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
796 B. LE ROUX ET AL.

produce, does not indicate the sustainability of the water use • according to the hydrological-based methodology, blue
unless it is interpreted within the local hydrological and en- WFs are the difference between volumes abstracted
vironmental context, for example by simply stating that it through irrigation and volumes recharged due to deep
takes 162 m3 blue plus green water to produce 1 t of carrots drainage and runoff. In the original method by Deurer
in winter, it was considered to be the most useful methodol- et al. (2011), runoff was considered to re-enter the blue
ogy to be used for catchment management and decision- water source, which was groundwater in their case,
making. Reasons for this include: because of the flat topography of their study area. How-
ever, in different circumstances runoff may flow out
• the methodology is well developed, WFs are relatively from a catchment and not replenish the local aquifer.
simple to calculate and understand and can be included In this case the original method will overestimate
in a water balance for further interpretation on the sus- aquifer replenishment and underestimate blue WFs. A
tainability of water use; carefully calculated water balance is important to un-
• the quantitative nature of these WFs can potentially be derstand the sustainability of a water use, but the issue
used in different information systems, such as water al- on including or excluding runoff illustrates how diffi-
location and water use licensing services, given that it cult it is to standardize a water balance calculation.
is first interpreted in terms of the local environmental We therefore recommend that water use data, as calcu-
and hydrological environment; lated by the WFN method, be used in a site-specific
• by altering the functional units, these metrics can be water balance, as illustrated by Le Roux et al. (2017);
used for applications such as understanding WFs per • green WF calculations are complex to make and may
nutritional unit produced or economic gain; require sophisticated modelling to estimate what the
• these WFs can reveal impacts on water resources in yield would be under rainfed conditions;
different seasons of a hydrological or calendar year; • WFs according to the hydrological-based approach are
• it can indicate high WFs of certain crop species, or calculated over a year. To determine WFs for short-
certain growing regions, such as those which experi- season vegetable crops, for example, crop sequences
ence relatively high vapour pressure deficits or have need to be used for a year. This may conceal the high
poor soils; WFs of certain crops in the sequence, and the impact
• it allows for local contextualization if there is suitable on water resources in dry seasons. The method is there-
information to conduct the sustainability assessment. fore more suitable for perennial crops;
• the WF according to the hydrological-based approach
The LCA methodology has some important strengths, most may provide valuable information to a water resources
notably the more advanced calculation of water quality im- manager, but care is needed as the result may be coun-
pacts in terms of eutrophication, freshwater ecotoxicity and terintuitive. For example, if water used by agriculture
human health (Pfister et al., 2009). The method takes in a catchment was exactly the same as the volume of
multiple environmental impacts into account simulta- water received, the collective WF will be zero, but if
neously, including water consumption and carbon footprints. all the available water is used there will be no discharge
Considering the unique geohydrological characteristics and from the catchment, which will affect the downstream
water issues that often exist, for example in the Olifants- users and aquatic ecosystems. If a product is labelled
Doorn WMA case study, more detailed local WS indices with a zero WF, it would seem that its production
are required. Although spatial variations may impact the was carried out sustainably. However, in this case a
WSI, it most likely will also be sensitive to temporal varia- zero WF does not indicate sustainable water use during
tions within a year and over longer periods. A WSI may need production. One instead needs a negative WF to be sus-
modification when, for example, commercial agriculture ex- tainable, but how negative should your WF be? This
pands. The variation factor is lower for catchments with will be different for each catchment, and the method
dams or aquifers that regulate flows and reduce variations does not include guidelines on how to include the water
in water availability (Pfister et al., 2009). Groundwater requirements of downstream users or specify the vol-
availability will reduce variations in overall water umes of water that are required to flow from a particu-
availability, which will reduce the WSI. lar catchment. You could argue that an activity that
To calculate blue and green WFs, the hydrological-based supplies fresh water to a catchment, like a desalination
method takes all water flows into account, as opposed to the plant or waste water treatment plant, will have a nega-
WFN, which considers crop ET only. Although the tive WF. But, if agriculture in a catchment uses less
hydrological-based method seems more biophysically com- than the available rainfall, they are not actually supply-
prehensive than the WFN method, the following issues were ing water to the aquifer, and should therefore not have
encountered in the assessment of the method: a negative WF;

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
 USING WATER FOOTPRINT METHODS IN AGRICULTURE 797

• using the hydrological-based method for crops for wisdom implying that a volume used is ‘good’ or ‘bad’,
which drainage plus runoff is higher than irrigation, ‘high’ or ‘low’. Using hypothetical examples, Wichelns
will likely produce a negative blue WF. It is confusing (2017) took WFs according to the WFN approach (i.e. data)
to obtain a negative blue WF for a crop that is heavily and rightfully interpreted them together with socio-
reliant on irrigation, even if deep drainage plus runoff economic impacts, such as employment opportunities and
is greater than the total amount irrigated. This was the use of energy (i.e. produced information). Similarly, Ridoutt
case for our study in the Olifants-Doorn WMA because and Pfister (2010) interpreted data on water use in terms of
irrigation is applied during the dry summer season, water availability, which provided important information.
while rainfall during the wet winter months (when the The above examples demonstrate how a water resource
orchard is dormant) results in recharge and runoff. manager should apply water use data in the decision-making
process. It is, however, also important to note that these
kinds of assessments would not be possible without the data
Analysis in terms of the knowledge hierarchy
on water use (such as the WFs calculated according to the
The aim of any method to improve water management prac- WFN approach). For this reason, the LCA and
tices should be to encourage sustainable water use. This hydrological-based communities developed modified meth-
must be done at national, regional and local levels to address odologies seeking to interpret the data to obtain information
water resources management and/or to change the behaviour (a better understanding of the water use in terms of water
of consumers. The so-called knowledge hierarchy (Ackoff, availability and the hydrology) and wisdom (the LCA meth-
1989) provides a useful way to better understand the differ- odology potentially providing consumers with a label that
ence between WF methodologies and the complexities in- will indicate the degree of impact).
volved in developing and using them (Figure 3). Although it is very important to move from data to
According to Chaffey and Wood (2005) as cited by wisdom, there are many complexities involved in standard-
Rowley (2007), higher orders of the hierarchy have more izing a method on these higher levels of the knowledge hier-
meaning than lower ones. As indicated in Figure 3, data archy. Figure 3 also indicates this, by showing that the
which are meaningless in themselves are at the bottom of higher levels of the knowledge hierarchy cannot be pro-
the knowledge hierarchy. Data that are interpreted become grammed and calculated by computers. For example, the
information, knowledge is the know-how or experience of WSI calculated for the LCA methodology considers the
what to do with information, and wisdom is the judgement availability of water within a certain area. The WSI is based
about whether our actions are right or wrong. In a water on the withdrawal to availability ratio (WTA), of which 0.4
management context, the volume of water that is used to is assumed to be the threshold indicating severe water stress.
produce a product is data. These data only become informa- However, this WTA of 0.4 is a fixed number, which does
tive when interpreted in a local context of water availability not apply to all catchments at all times. For example, the
and environmental sustainability. Somehow the information Ganges River is characterized by severe flood and drought
should be communicated to consumers, producers and water cycles (Sharma et al., 2010), and a different WTA ratio
resource managers so they make wise decisions that will would be needed to represent the threshold indicating severe
lead to the sustainability of the water used to produce a water stress in these different cycles. Although it is impor-
product. Awad and Ghaziri (2004) as cited by Rowley tant to consider water availability in relation to water use,
(2007) said that data can be programmed, while wisdom it is not the only consideration in terms of sustainability.
cannot be programmed or generated by a computer (Fig- Water requirements of the ecosystem, people and economy
ure 3). For these reasons it is not possible to develop a WF must all be considered. This can become very complex, tak-
method of which the outcome is an undisputed number that ing ecological water requirements as an example. It is com-
can be used on product labels and will indicate ‘right’ or monly recognized that flow reductions in rivers are not
‘wrong’ to a consumer. There is probably no ‘right’ or desirable (Lake, 2003) and that floods are important ecolog-
‘wrong’ answer and it might be more important to indicate ical events that flush the river free of alien vegetation and
if one is moving towards sustainability or away from it. sediments (Rountree, 2014). However, changes in the sea-
Volumetric WFs according to the WFN methodology are sonality of flows such as increasing dry season flows and de-
at the level of data, defining the WF as a volume of water creasing wet season flows, which is common in irrigation
used to produce a product or provide a service. Data are of- schemes, also have an impact on river ecosystems (Pattie
ten most valuable, because a water resource manager can in- et al., 1985; Lake, 2003; Rountree, 2014). Aquatic species
terpret them within his/her specific location to obtain the are adapted to certain flow regimes, which support connec-
necessary information for decision-making. However, care tivity in the aquatic ecosystem and habitats (Bunn and
must be taken not to communicate a WF defined as a vol- Arthington, 2002). Changes to geohydrological characteris-
ume of water used, which is mere data, as information or tics, such as groundwater converted to surface water, are

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
798 B. LE ROUX ET AL.

undesirable (Rountree, 2014). Maintaining natural flows of catchment to basin water management (Le Roux et al.,
rivers is more important if the aquatic and riparian biodiver- 2017). Engineers may prefer the LCA approach because of
sity is sensitive containing, for example, red data species its compatibility with determining other potential impacts si-
(Rountree, 2014). Often a water resource manager has to de- multaneously, and being able to link it to a functional unit
cide whether to allocate water to people or ecosystems, allowing product cross-comparisons. From this study, it
which involves trade-offs of various impacts. These are only can be concluded that the choice of method depends on
some of the complexities associated with the water require- the objective (s) that it is used for, and it is hoped that this
ments of an ecosystem, and the LCA methodology does not review will contribute to further discussion and evolution to-
address this. wards a practical and useful WF methodology.

CONCLUSIONS
ACKNOWLEDGMENTS
One of the drawbacks of water becoming a global resource
is that water users become disconnected from, and unaware This research originated from a project initiated, managed
of, the impacts of their water uses. It is therefore important and funded by the Water Research Commission (WRC pro-
ject No. K5/2273//4: Water footprint of selected vegetable
to consider ways of providing quantitative information on
water use and/or impact and to influence consumer behav- and fruit crops produced in South Africa), now published
iour towards sustainability. How this should be done has as WRC Report No. TT 722/17 (see Van der Laan, 2017).
The first author, Betsie le Roux, received financial support
been debated by scientists involved in WF assessments.
The volumetric WF of the WFN, even with the sustainabil- for research from the WRC and a bursary from the National
ity assessment, is not a suitable metric for communication to Research Foundation (NRF) of South Africa (NRF Grant
number: 88572).
consumers and for product labelling, because it does not
fully describe the context of the water use, including
socio-economic factors. The other methods have attempted
REFERENCES
to interpret and modify the WFN metric, most notably the
LCA method that aimed to produce product labels. As the Ackoff RL. 1989. From data to wisdom. Journal of Applied Systems
ISO Standard (ISO 14046, 2014) does not specify ways of Analysis 16: 3–9.
reporting WFs to consumers for awareness raising, this Awad EM, Ghaziri HM. 2004. Knowledge Management. Upper Saddle
indicates that they too struggled with the complexity of stan- River, NJ, USA: Pearson Education International.
Bailey A, Pitman W. 2015. Water Resources of South Africa: 2012 Study
dardizing such a method.
(WR2012). Water Research Commission Report No. 2143/1. Pretoria,
This study on WFs has indicated that calculating WF la- South Africa.
bels still requires much refinement and debate, and at best Bunn ES, Arthington HA. 2002. Basic principles and ecological conse-
may result in a symbol indicating responsible water use or quences of altered flow regimes for aquatic biodiversity. Environmental
stewardship, as opposed to a quantitative or even water-scar- Management 30: 492–507.
city-weighted volumetric WF label. Consumers need all Canals LM, Chenoweth J, Chapagain A, Orr S, Antón A, Clift R. 2009.
Assessing freshwater use impacts in LCA: Part I—inventory modelling
levels of the knowledge hierarchy (data, information, and characterisation factors for the main impact pathways. The Interna-
knowledge and wisdom) to make wise decisions about the tional Journal of Life Cycle Assessment 14: 28–42.
products they buy. However, influencing consumers through Chaffey D, Wood S. 2005. Business Information Management: Improving
education may have unpredictable outcomes. Some con- Performance Using Information Systems. Harlow, UK: FT Prentice Hall.
Department of Environmental Affairs and Development Planning
sumers may choose products based on potential impacts on
(DEA&DP). 2011. Western Cape IWRM Action Plan: Status Quo Report
people, while others could make decisions based on ecolog- Final Draft. Available online. URL: http://www.westerncape.gov.za/
ical sustainability. Advertisements and marketing also influ- eng/pubs/public_info/P/205596 (accessed 28 January 2013).
ence market demands and the interpretation of information. Deurer M, Green SR, Clothier BE, Mowat A. 2011. Can product water foot-
Future studies must pay attention to the various ways in prints indicate the hydrological impact of primary production? A case
study of New Zealand kiwifruit. Journal of Hydrology 408: 246–256.
which consumer behaviour can be influenced to change
Dzikiti S, Schachtschneider K. 2015. Water Stewardship for Stone Fruit
market demand. Personal experience has shown that Farmers: Technical Report to World Wide Fund for Nature, South
second-year biological sciences undergraduate students at Africa. Available online. URL: http://awsassets.wwf.org.za/downloads/
the University of Pretoria relate better to the concept of a wwf_water_stewardshiptechreport_2015.pdf
WF than to water use efficiency or water productivity. Goedkoop M, Heijungs R, Huijbregts M, De Schryver A, Struijs J, Van
Ultimately it may be that natural scientists prefer the Zelm R. 2013. ReCiPe 2008 A life cycle impact assessment method
which comprises harmonised category indicators at the midpoint and
WFN approach due to its volumetric nature, which means the endpoint level. First edition (revised) Report I: Characterisation.
it is quantitative and can be applied to identify hotspots at Ministry of Housing, Spatial Planning and Environment (VROM), Den
larger spatial scales (Multsch et al., 2016), or be used in Haag, the Netherlands, 133 pp.

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)
 USING WATER FOOTPRINT METHODS IN AGRICULTURE 799

Herath I, Green S, Horne D, Singh R, Clothier B. 2014. Quantifying and LCA and ISO water footprint: a response to Hoekstra (2016) ‘A critique
reducing the water footprint of rain-fed potato production, part I: measuring on the water-scarcity weighted water footprint in LCA’. Ecological
the net use of blue and green water. Journal of Cleaner Production 81: Indicators 72: 352–359.
111–119. Pfister S, Hellweg S. 2009. The water ‘shoesize’ vs. footprint of bioenergy.
Hoekstra AY. 2016. A critique on the water-scarcity weighted water Proceedings of the National Academy of Sciences 106: E93–E94.
footprint in LCA. Ecological Indicators 66: 564–573. Pfister S, Koehler A, Hellweg S. 2009. Assessing the environmental im-
Hoekstra AY, Aldaya MM, Chapagain AK, Mekonnen MM. 2011. The Wa- pacts of freshwater consumption in LCA. Environmental Science &
ter Footprint Assessment Manual: Setting the Global Standard. Technology 43: 4098–4104.
Routledge: London, UK. Ridoutt BG, Pfister S. 2010. A revised approach to water footprinting to
Hoekstra AY, Chapagain AK, Aldaya MM, Mekonnen MM. 2009b. Water make transparent the impacts of consumption and production on
Footprint Manual: State of the Art 2009. Enschede, the Netherlands: Wa- global freshwater scarcity. Global Environmental Change 20(1):
ter Footprint Network. 113–120.
Hoekstra AY, Gerbens-Leenes W, Van der Meer TH. 2009a. Reply to Ridoutt BG, Pfister S. 2013. A new water footprint calculation method
Pfister and Hellweg: water footprint accounting, impact assessment, integrating consumptive and degradative water use into a single stand-
and life-cycle assessment. Proceedings of the National Academy of alone weighted indicator. The International Journal of Life Cycle
Sciences 106: E114–E114. Assessment 18: 204–207.
International Standard Organization (ISO). 2014. ISO 14046. Environmen- Rountree M. 2014. Information on aquatic impacts due to changes in river
tal Management Water Footprint—Principles, Requirements and Guide- flows. Personal communication, South Africa.
lines. Geneva, Switzerland. Rowley JE. 2007. The wisdom hierarchy: representations of the DIKW hi-
Lake PS. 2003. Ecological effects of perturbation by drought in flowing wa- erarchy. Journal of Information Science 33: 163–180.
ters. Freshwater Biology 48: 1161–1172. Sharma B, Amarasinghe U, Xueliang C, De Condappa D, Shah T, Mukherji
Le Roux B, Van der Laan M, Vahrmeijer T, Bristow KL, Annandale JG. A, Bharati L, Ambili G, Qureshi A, Pant D, Xenarios S, Singh R,
2017. Establishing and testing a catchment water footprint framework Smaktin V. 2010. The Indus and the Ganges: river basins under extreme
to inform sustainable irrigation water use for an aquifer under stress. pressure. Water International 35(5): 493–521.
Science of the Total Environment 599: 1119–1129. Van der Laan M (ed.). 2017. Application of Water Footprint Accounting
Multsch S, Pahlow M, Ellensohn J, Michalik T, Frede H-G, Breuer L. 2016. for Selected Fruit and Vegetable Crops in South Africa. Water Research
A hotspot analysis of water footprints and groundwater decline in the Commission Report No. TT 722/17. Pretoria, South Africa.
High Plains aquifer region, USA. Regional Environmental Change 16: Wichelns D. 2011a. Assessing water footprints will not be helpful in
2419–2428. improving water management or ensuring food security. International
Pattie F, Ward WB, Patton DR. 1985. Effects of regulated water flows on Journal of Water Resources Development 27(3): 607–619.
regeneration of Fremont cottonwood. Journal of Range Management Wichelns D. 2011b. Virtual water and water footprints. Compelling no-
38: 135–138. tions, but notably flawed. GAIA-Ecological Perspectives for Science
Perry C. 2014. Water footprints: path to enlightenment, or false trail? and Society 20(3): 171–175.
Agricultural Water Management 134: 119–125. Wichelns D. 2017. Volumetric water footprints, applied in a global context,
Pfister S, Boulay A-M, Berger M, Hadjikakou M, Motoshita AM, Hess T, do not provide insight regarding water scarcity or water quality
Ridoutt B, Weinzettel J, Scherer L, Döll P. 2017. Understanding the degradation. Ecological Indicators 74: 420–426.

© 2018 John Wiley & Sons, Ltd. Irrig. and Drain. 67: 790–799 (2018)