The Water Footprint of an Animal
We follow the water footprint definitions and methodology as set out in Hoekstra and others (2011). The blue water footprint refers to consumption of blue water resources (surface and groundwater) along the supply chain of a product. ‘Consumption’ refers to loss of water from the available ground-surface water body in a catchment area. Losses occur when water evaporates, returns to another catchment area or the sea or is incorporated into a product. The green water footprint refers to consumption of green water resources (rainwater in so far as it does not become runoff). The grey water footprint refers to pollution and is defined as the volume of freshwater that is required to assimilate the load of pollutants given natural background concentrations and existing ambient water quality standards.
We consider eight farm animal categories: beef and dairy cattle, pigs, sheep, goats, broiler and layer chickens, and horses. When estimating total feed amounts and total water footprints per category, we include ‘buffaloes’ in the category of ‘beef cattle’ and ‘asses and mules’ in the category of ‘horses’.
The water footprint of a live animal consists of different components: the indirect water footprint of the feed and the direct water footprint related to the drinking water and service water consumed (Chapagain and Hoekstra 2003, 2004). The water footprint of an animal is expressed as:
$$ {\text{WF}}[a,c,s] = {\text{WF}}_{\text{feed}} [a,c,s] + {\text{WF}}_{\text{drink}} [a,c,s] + {\text{WF}}_{\text{serv}} [a,c,s] $$
(1)
where WFfeed[a,c,s], WFdrink[a,c,s] and WFserv[a,c,s] represent the water footprint of an animal for animal category a in country c in production systems s related to feed, drinking water and service-water consumption, respectively. Service water refers to the water used to clean the farmyard, wash the animal and carry out other services necessary to maintain the environment. The water footprint of an animal and its three components can be expressed in terms of m3/y/animal, or, when summed over the lifetime of the animal, in terms of m3/animal. For beef cattle, pigs, sheep, goats and broiler chickens—animals that provide their products after they have been slaughtered—it is most useful to look at the water footprint of the animal at the end of its lifetime, because it is this total that will be allocated to the various products (for example, meat, leather). For dairy cattle and layer chickens, it is most straightforward to look at the water footprint of the animal per year (averaged over its lifetime), because one can easily relate this annual animal water footprint to its average annual production (milk, eggs).
The water footprint of an animal related to the feed consumed consists of two parts: the water footprint of the various feed ingredients and the water that is used to mix the feed:
$$ {\text{WF}}_{\text{feed}} [a,c,s] = \frac{{\sum\limits_{p = 1}^{n} {\left( {{\text{Feed}}[a,c,s,p] \times {\text{WF}}_{\text{prod}}^{ * } [p]} \right) + \text{WF}_{\text{mixing}} [a,c,s]} }}{{{\text{Pop}}^{ * } [a,c,s]}} $$
(2)
Feed[a,c,s,p] represents the annual amount of feed ingredient p consumed by animal category a in country c and production system s (ton/y), WF
*prod
[p] the water footprint of feed ingredient p (m3/ton), WFmixing[a,c,s] the volume of water consumed for mixing the feed for animal category a in country c and production system s (m3/y/animal) and Pop*[a,c,s] the number of slaughtered animals per year or the number of milk or egg producing animals in a year for animal category a in country c and production system s.
The Water Footprint of Feed Ingredients
The water footprints of the different crops, roughages and crop by-products (WF
*prod
[p], m3/ton) that are eaten by the various farm animals have been calculated following the methodology developed by Hoekstra and Chapagain (2008) and Hoekstra and others (2011). The water footprints of feed crops were estimated using a crop water use model that estimates crop water footprints at a 5 × 5 arc minute spatial resolution globally (Mekonnen and Hoekstra 2010, 2011). Grey water footprints were estimated by looking at leaching and runoff of nitrogen-fertilizers only, following Mekonnen and Hoekstra (2010, 2011). As animal feed in a country originates from domestic production and imported products, for the calculation of the water footprint of animal feed in a country, we have taken a weighted average water footprint according to the relative volumes of domestic production and import:
$$ {\text{WF}}_{\text{prod}}^{ *} [p] = \frac{{P[p] \times {\text{WF}}_{\text{prod}} [p] + \sum\limits_{{n_{e} }} {\left( {T_{i} [n_{e} ,p] \times {\text{WF}}_{\text{prod}} [n_{e} ,p]} \right)} }}{{P[p] + \sum\limits_{{n_{e} }} {T_{i} [n_{e} ,p]} }} $$
(3)
in which P[p] is the production quantity of feed product p in a country (ton/y), T
i
[n
e
,p] the imported quantity of feed product p from exporting nation n
e
(ton/y), WFprod[p] the water footprint of feed product p when produced in the nation considered (m3/ton) and WFprod[n
e
,p] the water footprint of feed product p as in the exporting nation n
e
(m3/ton). The water footprint of crop residues such as bran, straw, chaff and leaves and tops from sugar beet have already been accounted for in the main product, therefore their water footprint was set equal to zero.
Volume and Composition of Feed
The volume and composition of the feed consumed vary depending on the type of animal, the production system and the country. The amount of feed consumed is estimated following the approach of Hendy and others (1995), in which the total annual feed consumption (including both concentrates and roughages) is calculated based on annual production of animal products and feed conversion efficiencies. Only for horses we have used the approach as in Chapagain and Hoekstra (2003), which means that we multiplied the estimated feed consumption per animal by the number of animals, thus arriving at an estimate of the total feed consumed by horses.
The total feed per production system for both ruminants and non-ruminants animals is calculated as follows:
$$ {\text{Feed}}[a,c,s] = {\text{FCE}}[a,c,s] \times P[a,c,s] $$
(4)
where Feed[a,c,s] is the total amount of feed consumed by animal category a (ton/y) in country c and production system s, FCE[a,c,s] the feed conversion efficiency (kg dry mass of feed/kg of product) for animal category a in country c and production system s, and P[a,c,s] the total amount of product (meat, milk or egg) produced by animal category a (ton/y) in country c and production system s. The estimated global amount of feed consumption per animal category and world region is presented in Appendix II (Supplementary material). Feed consumption per production system is shown in Appendix III (Supplementary material).
Estimating Feed Conversion Efficiencies
Feed conversion efficiency is defined as the amount of feed consumed per unit of produced animal product (for example, meat, milk, egg). Feed conversion efficiencies were estimated separately for each animal category (beef cattle, dairy cattle, sheep, goats, pigs, broiler chickens and egg-layer chickens), for each animal production system and per country. Although the term used may suggest precisely the opposite, animals that have a low ‘feed conversion efficiency’ are efficient users of feed. We use the term here as generally used in livestock studies. The feed conversion efficiencies (FCE, kg dry mass/kg product) for non-ruminants (pigs and chickens) were adopted from Hendy and others (1995). For ruminants (cattle, goats, sheep), feed conversion efficiencies were estimated through dividing feed intake per capita by annual production (of beef, milk, sheep and goat meat) per capita:
$$ {\text{FCE}}[a,c,s] = \frac{{{\text{FI}}[a,c,s]}}{{{\text{PO}}[a,c,s]}} $$
(5)
where FI[a,c,s] is the feed intake per head by ruminant animal category a in country c and production system s (kg dry mass/y/animal), and PO[a,c,s] the product output per head for ruminant animal category a in country c and production system s (kg product/y/animal). The product output (beef, milk, sheep and goat meat) per animal for ruminants is calculated as:
$$ {\text{PO}}[a,c,s] = \frac{P[a,c,s]}{{{\text{Pop}}[a,c,s]}} $$
(6)
in which P[a,c,s] is the total annual production of beef, milk, sheep meat or goat meat in country c in production system s (kg/y) and Pop[a,c,s] the total population of beef cattle, dairy cattle, sheep or goats in that country and production system. Region-specific feed conversion efficiencies are presented in Appendix I (Supplementary material).
Estimating the Total Annual Production of Animal Products
The meat production (P
meat, ton/y) per animal category a (beef cattle, pigs, sheep and goats) in country c and production system s is estimated by multiplying the carcass yield per slaughtered animal by the annual number of animals slaughtered:
$$ P_{\text{meat}} [a,c,s] = {\text{CY}}[a,c,s] \times {\text{SA}}[a,c,s] $$
(7)
The carcass yield (CY, kg/animal) for each animal category per production system was estimated by combining country average carcass yield data from FAO (2009) with data on animal live weight per production system per economic region (Hendy and others 1995) and data on carcass weight as percentage of live weight (FAO 2003). The obtained carcass yields were scaled such that the total meat production per animal category equals the value provided by FAO (2009). The number of slaughtered animals per production system (SA, number of animal/y) was calculated by multiplying the total animal number by the animal off-take rate per production system:
$$ {\text{SA}}[a,c,s] = {\text{Pop}}[a,c,s] \times {\text{OR}}[a,c,s] $$
(8)
where Pop[a,c,s] is the population of animal category a in country c for production system s and OR[a,c,s] the off-take rate, which is the fraction of the animal population that is taken out in a given year for slaughter (dimensionless).
Milk and egg production per production system and country were calculated as:
$$ P_{\text{milk}} [a,c,s] = {\text{MY}}[a,c,s] \times {\text{DC}}[a,c,s] $$
(9)
$$ P_{\text{egg}} [a,c,s] = f_{\text{egg}} [a,c,s] \times P_{\text{egg}} [a,c] $$
(10)
where P
milk[a,c,s] and P
egg[a,c,s] represent production of milk and egg in country c and production system s, respectively (ton/y), MY[a,c,s] milk yield per dairy cow in country c and production system s (ton/dairy cow), DC[a,c,s] the number of dairy cows in country c and production system s, f
egg[a,c,s] the fraction of egg produced in country c and production system s and P
egg[a,c] the total amount of egg produced in country c (ton/y).
Estimating the Feed Composition
Animal feeds are generally divided into ‘concentrates’ and ‘roughages’. The volume of concentrate feed has been estimated per animal category and per production system as:
$$ {\text{Concentrate}}[a,c,s] = {\text{Feed}}[a,c,s] \times f_{c} [a,c,s] $$
(11)
where Concentrate[a,c,s] is the volume of concentrate feed consumed by animal category a in country c and production system s (ton/y) and f
c
[a,c,s] the fraction of concentrate in the total feed for animal category a in country c and production system s. For the latter variable, data have been obtained from Bouwman and others (2005) and Hendy and others (1995).
The composition of concentrate feeds varies across animal species and regions of the world. To our knowledge, there are no datasets with global coverage on the composition of feed for the different animals per country. Therefore, we have made a number of assumptions concerning the concentrate feed composition of the different animal species. According to Hendy and others (1995), the diets of pigs and poultry include, on average, 50–60% cereals, 10–20% oil meals and 15–25% ‘other concentrates’ (grain substitutes, milling by-products, non-conventional concentrates). Wheeler and others (1981) provide the feed composition in terms of major crop categories for the different animal categories. We have used these and other sources in combination with FAOSTAT country average concentrate feed values for the period 1996–2003 (FAO 2009) to estimate the diet composition of the different animal species. To estimate the feed in terms of specific crops per animal, we first estimated the feed in terms of major crop categories following Wheeler and others (1981). The feed in terms of major crop categories is further distributed to each crop proportional to the crop’s share in its crops category as obtained from FAO (2009). The roughage feed is divided into fodder, grass and crop residues using the data obtained from Bouwman and others (2005). The estimated fraction of concentrate feed in total feed dry matter, per animal category, production system and world region is presented in Appendix IV (Supplementary material).