Consumptive water use in livestock production – Assessment of green and blue virtual water contents of livestock products

Ylva Ran

Uppsats för avläggande av naturvetenskaplig kandidatexamen i Miljövetenskap 15 hp Institutionen för växt- och miljövetenskaper Göteborgs universitet Juni 2010

Summary Distribution and use of fresh water resources are a widely discussed matter and the competition and requirement is increasing while resources are decreasing globally. Amongst the largest consumers are households, industry and food production where food production requires by far the largest amount of water. The role and scale of the consumptive water use for food production associated with livestock production is quite unknown but definitely of a significant magnitude. In a situation with increasing world population coupled to rising food requirements the role of livestock productions is predicted to grow even faster. An assessment of the livestock sector is therefore of great importance. A thorough calculation of water productivity of livestock production enables a valid picture of the situation and what possibilities that is available for development towards a more sustainable use of water resources and increased water productivity in livestock production. This study aims to estimate the consumptive water use of livestock production in terms of green and blue water and virtual water content of livestock products. Six animal production systems have been analyzed for seven selected regions to generate results that will enable assessment for the role of livestock production in water resource requirements and what improvements can be adapted to enable increased water productivity. The calculations in this study is based on two different models where the FPD Model generated results of feed requirement for the animal systems involved in livestock production and their feed compositions in the seven selected regions. With this dataset assessments could be performed of the total feed use of the production which was further applied to the LPJmL model. In this step the amount of feed was recalculated to the consumptive water use and virtual water content of the different livestock products in selected regions. Results primarily illustrate that consumptive water use and virtual water content of livestock production is largest in Asia, in particular South and Central Asia. The division between blue and green water resources further pinpoints that withdrawal and product virtual water content of blue water resources is strongly dominated by South and Central Asia. The values are generally larger for ruminant products and beef in particular whereas milk constitutes an exception with the lowest virtual water content. Another important result from this research is the distinctly higher consumptive water use as well as virtual water contents in the developing world in relation to developed regions. The significantly large consumptive water use and virtual water content of ruminant products is a result of the fact that their feed composition largely constitutes of grazing generating high consumptive water use and virtual water content. A new approach for accounting the consumptive water use of grazing land has been used in this thesis. This has generated that results are higher in general than most assessments prior to this study. Further analysis indicates that differences in virtual water contents between ruminants and monogastric products cannot singlehandedly be explained in terms of feed use. Regional differences in particular depend on varying feed compositions and the different conditions of cultivation of the feed composites. This generates that their virtual water contents will differ depending on the location of where the feed have been cultivated and in turn have a significant impact on the livestock production in that particular region.

1

Sammanfattning Förekomst och användning av vattenresurser är något som är mycket omdiskuterat och konkurrens om samt efterfrågan på vattenresurser ökar medan tillgång av resurser minskar globalt. Den största användningen av vatten sker i hushåll, i industrin eller vid matproduktion där sistnämnda använder den betydligt största delen. Den andel av vattenanvändning för matproduktion som är kopplad till djuruppfödning och produkter associerade till denna är relativt okänd men definitivt av signifikant storlek. Med en ökande befolkning och därmed en ökande efterfrågan på mat kommer djuruppfödning spela en allt större roll vilket medför att en uppskattning av dess vattenanvändning är av stor vikt. En utförlig beräkning av vattenproduktiviteten i animaliska produkter möjliggör en sanningsenlig bild av vilka möjligheter det finns för utveckling mot ett mer hållbart nyttjande av vattenresurser och en ökad vattenproduktivitet i djuruppfödning och dess associerade animalieprodukter. Denna studie syftar till att beräkna vattenanvändningen av djursuppfödning uppdelat mellan blåa och gröna vattenresurser och det virtuella vatteninnehållet i olika animalieprodukter. Sex olika djurproduktionssystem har analyserats för sju utvalda regioner för att generera resultat som möjliggör en uppskattning av den roll djuruppfödning spelar genom sitt vattenbehov och vilka förbättringar som kan adapteras för att öka dess vattenproduktivitet. Beräkningar i denna studie är baserade på två modeller där FPD modellen genererade resultat av det totala foderbehovet för de olika boskapen i produktionen samt sammansättningen av denna föda i sju utvalda regioner. Genererade data kunde sedan en uppskattning av den totala foderanvändningen utföras och appliceras på LPJmL modellen. I detta steg beräknades sedan den totala vattenanvändningen samt vattenproduktiviteten hos de olika djurprodukterna i de utvalda regionerna. Resultaten i denna studie illustrerar framförallt att total vattenanvändning och vattenproduktivitet för djuruppfödning är störst i Asien och i synnerhet i de södra och centrala delarna. Indelningen av resurser mellan grönt och blått vatten tyder vidare på att uttag och virtuellt blåvatteninnehåll i produkter domineras av södra och centrala Asien. Värdena är generellt högre för produkter från idisslare och framförallt nötkött medan mjölk utgör ett undantag där det virtuella vatteninnehållet är lägre än för alla andra animalieprodukter. Ytterligare ett viktigt resultat från denna studie är den distinkt högre vattenanvändning som förekommer i utvecklingsregioner i relation till industrialiserade områden. Den signifikant höga vattenanvändningen och det höga virtuella vatteninnehåll i animalieprodukter associerade med idisslare beror på att deras föda till stor del består av bete vilket generar hög vattenanvändning och högt vatteninnehåll. I denna studie har en ny ansats använts för att beräkna vattenanvändningen från betesmark. Detta har genererat generellt högre resultat än de beräkningar som utförts före denna studie. Vidare analys indikerar att skillnader i virtuellt vatteninnehåll mellan djurprodukter från idisslare och övriga boskap inte enbart kan förklaras med storleken på födokonsumtion. Skillnaderna, i synnerhet mellan regioner, beror framförallt på en varierande fodersammansättning och varierande förutsättningar för att odla de grödor som ingår i djurfodret Detta genererar att grödornas virtuella vatteninnehåll kommer variera beroende på i vilken region de odlats och i sin tur ha en signifikant påverkan på boskapsuppfödningen i den specifika regionen.

2

Table of content Summary ................................................................................................................................................. 1 Sammanfattning ..................................................................................................................................... 2 Table of content ...................................................................................................................................... 3 1 Introduction .............................................................................................................................. 4 1.1 The aim of this study ..................................................................................................................... 4 2 Background .......................................................................................................................................... 5 2.1 Green and blue water resources and flows .................................................................................. 5 2.2 Regional differences in green and blue water flows ..................................................................... 5 2.3 Water use in agriculture ................................................................................................................ 7 2.4 Water use in livestock production ................................................................................................ 8 2.5 Water scarcity and agriculture .................................................................................................... 12 2.6 General ability to cope with green and blue water deficit.......................................................... 13 2.7 Virtual water ................................................................................................................................ 13 3. Methods ............................................................................................................................................ 16 3.1 FPD Model ................................................................................................................................... 16 3.2 LPJ Model..................................................................................................................................... 17 3.3 Flow chart .................................................................................................................................... 19 4. Results ............................................................................................................................................... 27 4.1 Consumptive water use ............................................................................................................... 27 4.2 Virtual water content .................................................................................................................. 30 5. Discussion.......................................................................................................................................... 35 5.1 Consumptive water use ............................................................................................................... 35 5.2 Virtual water contents................................................................................................................. 37 5.3 Comparison, new perspectives and prior studies ....................................................................... 38 5.4 Assumptions and uncertainties ................................................................................................... 40 5.5 Water scarcity.............................................................................................................................. 41 5.6 Management options .................................................................................................................. 42 6. Conclusions ....................................................................................................................................... 44 Acknowledgements .............................................................................................................................. 46 7. References......................................................................................................................................... 47 Appendixes........................................................................................................................................... 50

3

1 Introduction Freshwater is a very important, thus limited, natural resource. A growing population has increased the water requirement globally and this has created a supporting problem. Further complications are that water resources are unevenly distributed over the world. Today the trend of water deficit falls upon the developing part of the world, where population growth as well as pollution tends to be greater. The conflict of distribution of freshwater resources is rising and predicted to become a great conflict between nations, regions, industry, agriculture, livestock production and households. The food production today is by far the largest user of fresh water resources with 70 % compared with only 10 % for household use and 10 % for industry (Molden, 2007). In this study water is considered as virtual water which is water used for agricultural or industrial production and further divided into •

Blue water resources which is the liquid water resources in aquifers, rivers and lakes

•

Green water resources which is rainfall as soil moisture and the flows of such in form of evapotranspiration (Falkenmark, 1995).

All food we eat requires a green water flow as transpiration from plants (Falkenmark & Rockström, 2004). For animal products in the food production green water is indirectly connected to the product in form of pasture, fodder and grains that are eaten by the animals. This feature is considered to multiply the overall water consumption of the animal food stuffs and constitutes the majority of water used for such production (Steinfeld et al., 2006). With increasing population the overall food demand will rise. It is predicted that not only will the demand of animal food stuffs increase also but the fraction of which animal food stuffs constitute of the per capita intake (Molden, 2007). Therefore it is fundamental to estimate and understand the water used for livestock production and how it differs between livestock products, systems and regions.

1.1 The aim of this study The aim of this study was to estimate the virtual water content and consumptive green and blue water use in livestock production, and how this differ between production system, products and regions. More specifically, the following objectives are included in the thesis •

Estimate differences in consumptive water use for beef, pork meat, poultry meat, milk and egg for seven different regions

•

Relate consumptive water use to production numbers by estimating differences in virtual water contents for beef, pork meat, poultry meat, milk and egg for seven different regions

•

Calculate the regional differences in virtual water content and water use between the feed composition fed to the animals differing with livestock production systems and regions.

4

2 Background This study aims to estimate consumptive water use and virtual water contents for livestock production and livestock products. Such assessments depend on water resources and water flows as well as features that can have an impact on availability of water etc. In this chapter the resources and features associated with livestock production and issues of concern is further identified and explained.

2.1 Green and blue water resources and flows The concept of water resources can be further complemented when it comes to water requirement in agriculture and housing of livestock. The terms blue and green water flows and resources were introduced by Falkenmark (1995). The blue water resources are the liquid water found in aquifers, rivers, lakes, dams and wetlands. The blue water flows are the surface run-off, flow of ground water and base flow in rivers. Green water resources is rainfall that has infiltrated in the root zone of the soil also referred to as soil moisture (Deutsch et al., 2010). The green water flow consists of both evaporation and transpiration components, often lumped together as evapotranspiration. This inclusion into one concept may, according to Jewitt (2006), compromise the hydrological modeling and prerequisites good conceptual understanding of modeling for green water flows. Calculations of green and blue water flows in formal water resource management have proven to be extremely difficult and complex. Jewitt (2006) further suggests that the most commonly used terminology for green water flows is separated into productive and non-productive, which also means a separation between evaporation and transpiration. Evaporation is referred to as the non-productive flow and transpiration as the productive flow, which is directly linked to biomass growth (Falkenmark & Rockström, 2004). The green and blue water flows interact with each other in numerous ways and affect the flow of one another. For instance, if consumptive green water use increases upstream by improving agricultural management or by converting agricultural land to forestry for example, the blue water generation will subsequently be smaller. The calculation of green water have been highly prioritized lately and especially in semi-arid and arid regions where blue water resources are scarce. When a blue water resource is extracted for irrigation it will undergo certain redirections and reforms of its flow pattern. Some of the water will flow as evapotranspiration, a green water flow, while a large fraction will flow as run-off and return to nearby water bodies as a blue water flow. For crop cultivation and livestock production the major fresh water use will not be in terms of blue water flows but from green water flows linked to the vegetation as infiltrated rainfall returned to the atmosphere in form of water vapor (Deutsch et al., 2010).

2.2 Regional differences in green and blue water flows Green water flows will differ with hydroclimatic features between different regions and their seasonal change. In a temperate (humid) region the evaporative demand is moderate and the precipitation enough to cover this demand. Distinct seasonal differences will take place between e.g. summer and winter. The circumstances lead to a precipitation surplus which creates a moderate runoff where the green water will partition to green water flows as well as to a blue water resource and recharge rivers and aquifers. The landscape in such region will consist of recharge areas with an

5

infiltration surplus, and discharge areas where rising groundwater will meet the precipitation surplus. Shortage of water resources (both green and blue) in a temperate humid region is an unlikely concern (Falkenmark & Rockström, 2004). In the humid tropics the amount of received precipitation as well as the evaporative demand will be large. Under such conditions the green water flow will be dominating which increases the amount of water vapor in the region and the air will have a high percentage of humidity. Seasonal differences will be small in regard to temperature and tropic regions will not suffer from water deficit under normal conditions. However if such a region will endure deforestation the flow of green water will decrease since evapotranspiration from vegetation will cease and in turn lead to less vapor in the atmosphere hence a decrease in rainfall (Falkenmark & Rockström, 2004). In a semi-arid, arid or sub-humid tropical region the precipitation is generally of the same magnitude as in temperate regions. However an increase of great magnitude in demand for evaporation occurs wherefore most of the precipitation in the region will return to the atmosphere and the surplus creating a run-off will be small if existing at all. The evaporative demand simply exceeds the amount of precipitation which results in blue water resources only being recharged during certain times of the year. This generates a situation where water resources of both green and blue water are small. In areas of this nature the population growth tends to be of high rate which increases the pressure of food production and effective agricultural production and it is in these areas where the highest proportion of undernourishment is found (Figure 1).

Figure 1: Percentage of undernourished people in relation to climate zones (here savanna/step is highlighted) where red countries have the highest percentage of undernourished and white the lowest Source: Rockström et al. (2007).

6

2.3 Water use in agriculture Most of the world’s food production is produced in predominantly rainfed areas where agriculture largely depends on rainfed green water resources. An interesting aspect here is that those regions where blue water dominates food production are withdrawing about 40 percent of available blue water. Regions depending on blue water for food production are North Africa, Central Asia and the Middle East (Falkenmark & Rockström, 2004). This is however misguiding since these areas are arid or semi-arid and cannot solely depend on green water flows but have to rely on blue water resources in addition. The former accounting for water resource withdrawal does not consider the use of green water resources wherefore the water use of regions that primarily uses green water resources will not be visible. Why calculation of blue water resources and flows has been prioritized might depend on that blue water resources are the only water resource considered as an economic good by economists and can be controlled by engineers. Falkenmark & Rockström (2004) further identifies that the importance and calculations of green water flows have been well known by other branches of science (apart from economics). Problems have occurred since green water flows and resources are not properly recognized in conventional water resource assessments and water policy decisionmaking. For agriculture to be able to meet the estimated increased demand of food from the predicted population growth in the future the increasing rate of agricultural production is expected to have to reach 3 percent per year. This rate exceeds the growth rate of the green revolution. This situation has induced that agricultural regions will suffer great environmental problems which are directly linked to water flows. Changes in use of land surface and increased biomass production will negatively affect the ability of rainfall to increase the soil moisture and production of runoff (Falkenmark & Rockström, 2004). In table 1 some of the calculations of global consumptive blue and green water use in agriculture are listed. The results differ widely with water demand varying between 7 000 and 15 000 km3 depending on which model has been used for the assessment. The models used for each calculation are listed in the left hand column of the table.

7

Table 1: Total global consumptive blue and green water use in agriculture calculated with different models Source: Hoff et al. (2010); de Fraiture & Wichelns (2007). Model

Rainfed cropland 3 (km )

Irrigated Cropland 3 (km )

Grazing land 3 (km )

Total 3 (km )

Green water

Green water

Blue water

Green water

919-959

1 180-1 448

-

Blue & Green water 11 003-11 271

GCWM

9 823

GEPIC

6 371

-

927

-

7 298

H08

9 540

850

1 530

4 320

15 390

IMPACT

-

1 703

1 425

-

-

LPJmL

-

381

a

1 364

3 269

WaterGap

8 290

-

1 300

3 260

WBM

9 406

-

1 301

4 413

Watersim

4 910

1 570

650

840

b

a

a a

c

d

c

-

c

12 850

c

15 120 e

7 970

a. Consumptive green water use for marked irrigated croplands are calculated only for cropping periods, not per year as the value derived from Watersim model. b. Consumptive blue water use for irrigated crop land calculated per year. c. Consumptive green water use for grazing lands are calculated on the assumption that 1/3 of the total consumptive green water use of grazing lands constitutes the consumptive water use for agriculture related to quantity of grass eaten by grazing animals. d. Consumptive green water use for grazing lands calculated with the Watersim model only calculates for the amount of evaporation actually consumed not for the total area wherefore numbers are lower than for other models. e. The Watersim value is excluding consumptive water use for feed crops, with that amount included the total consumptive 3 water use would be 9282 km .

2.4 Water use in livestock production The overall demand for water in livestock production is influenced by several factors such as type of animal, its activity, feed intake and diet, quality of available water, temperature of water and temperature of the ambient environment (Lardy et al., 2008). Conditions that will influence the water requirement of livestock is the physiological condition of the animal and the availability of water. Cattle with constant availability to water compared with cattle only allowed water access twice a day will produce more milk and more butterfat. In the same way a gestational or lactating animal will have larger water consumption than a non-gestational or non-lactating one. The diet has an impact on drinking water consumption of poultry. An increased level of fat, protein, salt, potassium and high level of crude fiber in the diet will increase the drinking water consumption (Lardy et al., 2008). Different products will require different amounts of water in their production. This requirement is as mentioned also depending on production site, water use efficiency of feed baskets and specific time as well as other production conditions. To produce a certain amount of meat, milk or egg it is

8

necessary to take into consideration a number of parameters. What kind of animal is used for production, where the animal is kept, what the animal diet constitutes of, where the feed is produced etc. will have an impact on the water requirement of the product (Steinfeld et al., 2006). Livestock production requires high amounts of water. One kilogram of grain used in livestock feed requires about 1000 to 2000 kg of water if the feed is grown in the Netherlands or Canada. The same grain will, however, require approximately 3000 to 5000 kg of water if grown in an arid region like Egypt or Israel. That difference in water requirement will have an impact on the total water use for a specific product relying on the grain from a particular region. Livestock in itself contains between 5 and 20 times more virtual water per kg product than crop products (Chapagain & Hoekstra, 2003). Consumptive water use in livestock are generally divided into two categories •

Drinking and process water – direct blue water use

•

Water use for production of feed, fodder and grazing – blue and green water use

Up to this point most calculations have been performed on the first category of consumptive water use but recently numerous scientists have highlighted the importance of including consumptive green water use in calculations regarding water use in livestock (Stenifeld et al., 2006; Falkenmark & Rockström, 2004; Ridoutt & Pfister, 2009; don Peden, 2007). The second type of consumptive water use is significantly larger than the first one referring only to blue water requirements and the relationship between the two and how they differ between regions are listed in table 2. Table 2: Estimated water needs for livestock. Source: Pedro (2007) adapted from Thornton et al. (2002) Estimated water needs for maintenance of 3 livestock (Mm )

SubSaharan Africa

South and Central Asia

West Asia and North Africa

East Asia

South- South East Asia Asia

Central a Asia

Total

Drinking water

2,2

3,7

0,5

1

0,8

2,5

0,2

0,9

Feed production

111

176

26

50

38

124

11

536

a. Newly independent states in Central Asia

2.4.1 Drinking and process water

Water is consumed by the animal itself in terms of drinking needs to support the physiological functions of the animal. Approximately 60-70 % of an animal body is water and animals need water for services that maintain vital physiological functions. The water intake of livestock is maintained by drinking and eating. Water leaves the body through respiration, evaporation, defecation and urination. Vital physiological functions will suffer from water deficit, which will result in reduced milk, meat and egg production (Steinfeld et al., 2006). The water requirement of animals increase with increasing temperature and it is also shown that water losses tend to increase under such conditions (Lardy et al., 2008). Animals suffering from water deficiency can experience a depression of vital physiological functions more quickly and drastically than any other nutrient deficiency and domestic animals can only survive without water for seven days. Signs of dehydration are tightening of skin, weight loss and drying of mucous membranes and eyes (Lardy et al., 2008).

9

Processing of animal products, including slaughtering and tanning of skins, demand a high level of hygiene and quality when processed. This in turn results in a large consumption of water and a large amount of waste water generated from these activities. Generally poultry processing facilities and dairy products have a higher demand for water per unit of weight than processing facilities for carcasses of cattle and pork. Tanning consumes large amounts of water as well. However, the environmental impact of the emitted pollutants from such processing is considered to be of greater concern (Steinfeld et al., 2006). There is also a water requirement for services such as cooling and washing of the animal and feed production. Production facilities need to be kept clean and animal product production need processing, which also requires water (Deutsch et al., 2010). The water use per animal differs with different production systems. In extensive systems the water requirement per animal will be greater than for intensive (industrialized) production systems. However, the intensive system will have a larger service water demand for cooling and cleaning of the facilities. Extensive systems tend to use more water for the animals´ feed supply. In some regions, where extensive livestock systems are dominating, the fresh water withdrawal for the drinking and servicing of livestock will be a large fraction of the total withdrawal. In most regions, however, the water use for drinking and servicing livestock constitutes only a small fraction of the total water requirement for the livestock production. Globally the water consumption of these processes is estimated to be only 0.6 percent of all freshwater use (Steinfeld et al., 2006).

2.4.2 Water for production of feed, fodder and grazing

Undoubtedly the largest amount of water used in livestock production is the amount used for feed production and the amount of feed produced is growing globally (Deutsch et al., 2010). Livestock stand for the largest anthropogenic land use in the world where the majority of land and water are dedicated to feed production (Steinfeld et al., 2006). Agriculture is also the economic sector that uses the largest portion of fresh water resources. Approximately 70 % of blue water resources from rivers, lakes and groundwater are used, and as much as 86 % of total blue and green water resources altogether (Deutsch et al., 2006; Chapagain & Hoekstra, 2007, 2008). Evapotranspiration is the main process for depletion of water for agricultural land used to grow crops and grass for grazing. The amount of green and blue water evaporated through feed crop land are enormous and estimated to be as much as 45 % of the water used for food production globally (Zimmer & Renault, 2003). Feed production is argued to affect green and blue water flows in three ways - through withdrawal of blue water for irrigation, through land cover change (e.g. when rain forest is converted to agricultural land) and through alterations in water division due to changes in land use management (Deutsch et al., 2010). Steinfeld et al. (2006) argued that a large fraction of this water use is environmentally insignificant, since a large amount is through evapotranspiration of land used for grazing such as non cultivated fodder land and grasslands. This land is not arable and therefore the opportunity cost will be small. Furthermore, in those situations where land for grazing is actually of agricultural potential, the land is generally situated in tropical areas where water is not limiting. Therefore, the opportunity cost of the land use will be that of lost agricultural soil and not the water resources (Steinfeld et.al, 2006). However, the water use from such areas must still be considered as environmentally significant, since freshwater is a scarce natural resource on a global scale.

10

Scientists differ in their opinions on however the water use for evaporation of grazing lands should be considered as a water use and be included in assessments. Pimentel (1997; 2004) argues in contrast to Steinfeld that this evaporation definitely should be considered as a water use and be accounted for in calculation of water flows and water consumption of livestock and agriculture. Others, e.g. Falkenmark & Rockström (2004), further contribute to the disagreement with defending the approach of accounting the evaporation from pasture lands only from the fraction of vegetation actually consumed by livestock while the rest is not accounted for in terms of water requirement for livestock. Deutsch et al. (2010) further argues that when allocating water resource use in livestock production there are three major issues that should be reviewed i.e. the multifunctionality of ecosystems, the degradation of ecosystems and the hydrological effects associated with livestock and livestock production. When all these issues are taken into consideration the different ways of approaching the evaporative water use of grazing land can all be partly or fully correct depending on prevailed conditions. When e.g. land is dedicated to grazing in an ecosystem in total balance where such activity would not generate any degradation of the land but merely be using another service that the healthy ecosystem can deliver evaporation should not be accounted for. However if the activity of grazing would contribute to the degradation of ecosystems and the deliverance of other ecosystem services for other purposes the evapotranspiration of pasture land should be considered a water use. Approximately 20 % of pastures are considered as degraded due to e.g. increased erosion, bush encroachment and sedimentation in runoff waters. When the third issue of the hydrological effects of grazing are included as well there can also be opposite effects than if only degradation of ecosystems are taken into consideration. As mentioned grazing are often carried out in deforested areas. This is clearly viewed as a degradation of an ecosystem but forests generally have a higher amount of green water flows than grazing land. A shift from forest ecosystem towards grazing land would therefore generate a decreased water use instead of increased. Costs would then increase instead of decrease which would not be the case if other resources than water would be considered (Deutsch et al., 2010) The different approaches in allocation of water use will generate widely different results when it comes to livestock production. If, as in the first approach, a grazing ecosystem is healthy and the extra cost of water use is not included in the water requirement is estimated to 12 000 liters of water per kg meat. If, instead conditions follows the second approach, the water requirements will be as high as 30 300 liters of water per kg of meat. The apparent complexity of situations favors the approach of Falkenmark & Rockström (2004) where the evapotranspiration water use of grazing land should be evaluated for each case and no general allocation for all grasslands should be used. This will allow the best possible appreciation of the significance of the water use of grazing lands in allocation of water consumption in livestock production. Lately the assessment of water use in livestock production have improved by including green water in calculations but are still considered to be quite far from reality. This has various reasons. Primarily no consideration is taken to what production system is used (intensive or extensive systems) which has proven to have enormous impact on the amount of water required. Secondly the entire hydrological cycle need to be regarded since there can be different effects on different parts of the

11

cycle. Thirdly the water use has to be directly linked to changes in ecosystems (Deutsch et al., 2010). The predicted change towards more intensive livestock production systems is an example that is predicted to deliver such ecosystem changes. Land area, today dedicated to pasture, are predicted to lose area to crop growth e.g. Steinfeld et al. (2006) and Rockström et al. (2007) estimated that 1000 km3 more green water will be used for grain feed production for livestock production.

2.5 Water scarcity and agriculture The concept and problems of water scarcity have been given many meanings referring to blue or green water scarcity. The former is the direct scarcity of water in rivers and aquifers, the latter the scarcity in soil affecting agriculture (Falkenmark et al., 2007). Household water delivery problems (and to that the availability of clean water), problems related to mobilization of freshwater resources and infrastructural development and economics and the size of the population competing for the resource are all different kinds of water scarcity issues. Freshwater resources depend on rainfall which constitutes the ultimate source of such resources. This implies that water scarcity is coupled with climate and temperature patterns where areas of low latitudes with higher evaporative demand suffer from water scarcity in a greater extent (Falkenmark et al., 2007). Since agriculture requires the largest amount of global water resources water scarcity is a major concern. Water scarce regions also tend to be large agricultural producers why the global picture of water scarcity (Figure 2) is even more important when considering consumptive water use of agriculture. Predictions of an increasing world population require that more food is produced and the pressure on agriculture to generate a much larger yield increases significantly. There is also a trend of increasing consumption of livestock products and the fraction of calorie intake per capita from animal products are thought to increase all over the world and in the developing world in particular (Falkenmark & Rockström, 2004; Steinfeld et al., 2006). Consumption and production rates are expected to grow with 2,5 % -4 % annually in developing countries but only with 0,5 % a year in developed regions (don Peden, 2007).

Figurer 2: Water scarcity expressed as withdrawals in percentage of total availability. Source: Rekacewicz (2006)

12

2.6 General ability to cope with green and blue water deficit Water management is only one of many solutions to deal with water deficit. To increase and develop irrigation with blue water resources is one solution dealing with water scarcity and have been the major solution of the past. However this solution is argued to be inadequate in meeting the water requirements in the future (Falkenmark & Rockström, 2004). The importance of dealing with losses of green water flows and capture of such resources have been brought forward. In rainfed agriculture e.g. Sub-Saharan Africa different methods of rainwater harvesting have been developed to enable coping with water scarcity (Falkenmark & Rockström, 2004; Molden, 2007). However to enable production improvements in areas suffering from water deficit a combination of water, crop and soil management must be adapted to generate sustainable development (Falkenmark & Rockström, 2004). As mentioned the regions that will most likely suffer from water deficit are the ones in most need of available water for increasing food and agricultural production. Their ability to cope with water scarce conditions will thus be crucial for their future development and their possibility of meeting a predicted rapid population growth hand-in-hand with increasing demand for food. This capacity of coping will definitely depend on accessibility of resources such as reservoirs and energy. It will however undoubtedly be affected by financial and political structure in the region as well as available natural resources (Falkenmark & Rockström, 2004). This further brings complications to regions that already are put under great pressure. These regions are the one with the largest expected population growth rate. They are also dealing with socio-economic issues such as HIV/AIDS and other diseases, high maternal mortality, gender inequity, high percentage of unemployed, education etc. (UNSTAT, 2010)

2.7 Virtual water The amount of water that is required for producing a certain industrial or agricultural product or service is defined as “virtual water” (Allan, 1993; 1994). The virtual water is simply water embodied in the product and not real water (Hoekstra, 2003). The volume of virtual water differs depending with sector and its production of products and services. Hoekstra (2003) further describes virtual water from two approaches namely the production point of view and the consumption point of view. The second approach identifies the water use as consumption site specific. Which implies the specific amount of water that would have been required to produce the specific product or service in the same place where it is consumed (Chapagain & Hoekstra, 2003). The assumption is made from the consumer point of view mentioned above and is important when trying to estimate the benefits or costs of e.g. importing a good or a service instead of producing it locally. It aims to answer the question if the product will consume more water when produced at a specific site compared to another site. In this study the first approach is used since the assessment of water use of that approach is site specific (Chapagain & Hoekstra, 2003). This will allow for consideration of the production conditions in the calculation of virtual water, e.g. time and place of the production as well as local and time specific water use efficiency (Chapagain & Hoekstra, 2003; Hoekstra, 2003). In the assessment of water requirements in agricultural production such as housing of livestock the virtual water flow between nations has to be considered. If a country exports a product that required

13

a certain amount of water for production, then water is exported in a virtual form. In a water-scarce nation there might be an attempt to produce and export products with low virtual water content but to import products with a high content of virtual water (Chapagain & Hoekstra, 2003; Hoekstra, 2003). That behavior will imply that the flow of virtual water out of a water-scarce nation will be as small as possible, and that the burden of production with high water requirement will fall on a waterrich nation. Figure 3 and 4 illustrate global flows of virtual water between regions pinpointing net exporting and importing areas.

Figure 3: Virtual water trade balances between nations (1995-1999). The scale starts from red color illustrating countries with a net import of virtual water. Green colored countries have a net export. Source: Hoekstra (2003).

Figure 4: Virtual water trade between regions (1995-1999). The largest net flows of virtual water are illustrated 3 as white arrows with a flow larger than 100 Gm . Source: Hoekstra (2003).

2.6.1 Virtual water content and water footprints

The virtual water content of a product is simply the freshwater embodied in that particular item (WFN, 2010). It refers to the volume of water required for the complete production chain of the product and is related to the term water footprint commonly used in the literature. Water footprint is the virtual water content calculated for e.g. a product, nation or region but is a more geographically explicit multi-dimensional indicator expressing not only volumes but what kind of water is being used as well as when and where (WFN, 2010). 14

Another term widely used in literature is the water productivity of products, nations or regions which simply is the inverse virtual water content (WFN, 2010). The input is rather consumptive water use than livestock production size why this thesis has focused on virtual water content instead of water productivity. To enable comparison with former estimates of consumptive water use associated with livestock however the invert of generated virtual water content results will be used in the discussion.

15

3. Methods The analysis has been performed with input data of regional livestock production numbers from the Food and Agricultural Organization statistics department (FAO STAT 2010, year 2005). Input data have been derived for seven geographical regions and the year of 2005. To be able to match the present regional classifications by FAO (FAOSTAT, 2010) and Wirsenius (2000), based on the older categorization, a certain regrouping was done prior to the analyze. Based on this data the feed and food amounts in terms of biomass to support this production were calculated. The consumptive water use for growing this biomass was estimated from data and finally the virtual water content per kilogram of animal product. Two models were used to recalculate input data in forms applicable for further analyze •

The Food Phytomass Demand (FPD) Model – calculating for phytomass feed demand for livestock animal production systems

•

The Lund-Postdam-Jena managed Land (LPJmL) model – calculating the virtual water content in regard to agricultural crops in a pixel-area based system applicable on both national, regional and global level

These models are described in further detail below. In this research the consumptive water use of livestock production is assumed to be the consumptive water used to produce the feed composition of the different animals producing livestock food items. This assumption will generate a small bias, however the water for drinking and caretaking of animals is considered an insignificant part of the total consumptive water use for livestock production (Steinfeld et al., 2006; don Peden, 2007) wherefore the assumption can be justified. The influence of trade in livestock production has also been excluded in this research and all items of the animals feed composition is assumed to be produced within each region.

3.1 FPD Model The FPD Model can be referred to as a bottom-up model and calculates necessary above-ground phytomass production divided between numerous crops and pasture species. Processes are described on both energy and mass balance basis and all major phytomass types connected to the food system are covered. All use of by-products and residues of biomass are also included in the FPD Model (Wirsenius, 2000). Further the model calculates energy requirements and in addition to them feed requirement of animal systems based on the nature of the flowchart and with input data from production numbers and feed use data from FAO. Since input data in the FPD model are of the biomass side and not the production side as in this research the model also delivers a general picture of the composition of demanded feed for animals in livestock production. The results and data derived from the FPD model and used in this research are on a regional basis that takes regional differences in consideration and delivers results on a regional and global scale.

16

3.2 LPJ Model The LPJ model (Lund-Postdam-Jena model) couples biogeochemistry and biogeography. The model represents land and atmosphere exchange of water and carbon and terrestrial vegetation dynamics on a 0,5 degreed pixel scale (Gerten et al., 2004) where each pixel are functioning as in figure 5.

Figure 5: Water balance for a pixel in the LPJ model. Thick arrows indicate water fluxes. Source: Gerten et al. (2004)

The model has been successfully applied to compare observations such as monitored seasonal cycle and the interannual variability of atmospheric carbon dioxide concentration (Sitch et al., 2003; Zaehle et al., 2005; Prentice et al., 2000; Peylin et al., 2005), water requirement for irrigation (Rost et al., 2008), river runoff and evapotranspiration (Gerten et al., 2004), soil moisture (Wagner et al., 2003), validation of phenology and yields (Bondeau et al., 2007) and virtual water content of certain crops (Fader et al., 2009). The premature LPJ model were process-based and simulated the production, phenology and growth of nine plant functional types, PFTs (Sitch et al., 2003). Bondeau et al. (2007) added 13 crop functional types (CFTs) to the model to ensure coverage of arable crops and managed grasslands. The CFTs correspond to particular species or groups of species of crops and/or grasses functioning similarly. They share fundamental physiological and biophysical characteristics with the premature PFTs but include agro-ecosystem-oriented functions and parameters to the model listed in table 3. This further enhances estimation of changes in fluxes and basic stocks of water and carbon. Such flux changes are caused by land use due to net primary production, biome productivity, heterotrophic respiration and evapotranspiration. When CFTs are used in the LPJ model it is referred to as LPJ managed Land (-LPJmL).

17

Table 3: Features/parameters of the LPJml model for phenology and growth of CFTs. Source: Bondeau et al. (2007) Crop Functional Types (CFTs

Local adaptation

Base Maximal tempera Leaf Area ture (°C) Index, LAI

Irrigation priority

Optimum harvest index (HI) at harvest (0-1)

Minimum HI at harvest (01)

Temperate Cereals (wheat, barley, rye, oats

Sowing date (sdate), phenological heat units (phu), vernalization requirements (pvd)

0

3-7

8

0,4

0,2

Rice

Double cropping in tropical asia, sdate, base temperature (tb)

10

5

1

0,5

0,25

Maize

Sdate, tb

(5-15)

3-7

2

0,5

0,3

Tropical cereals (millet, sorghum

Sdate

10

2,5

7

0,25

0,01

Pulses (lentils)

No

3

4

6

0,6

0,01

Temperate roots (sugar beet)

No

3

5

5

2

1,1

1

Tropical roots (cassava)

Sdate

20

5

11

2

1,1

2

Sunflower

Sdate, phu

6

3

4

0,3

0,2

Soybean

No

10

3

3

0,3

0,01

Groundnuts

Sdate

14

4

10

0,4

0,3

Rapeseed

Sdate, phu, pvd

0

4,5

9

0,3

0,15

Modelling with LPJmL also depend on additional input data. Monthly averages of e.g. temperature, precipitation, days with precipitation, hours of sunshine are used (Bondeau et al., 2007) and soil texture and concentration of CO2 (Sitch et al., 2003). Monthly precipitation data is diverted over each day determined by a generator (Gerten et al., 2004). LPJmL further includes a dataset of land use connected to each CFT. This implies that the cover of each CFT in percentage is represented for each pixel both irrigated and rain-fed on a yearly basis.

1 ,2

HI is larger than one for belowground storage organs

18

3.3 Flow chart The flow chart can be viewed as a simplified version of the Food Phytomass Demand (FPD) Model developed in the doctoral thesis by Wirsenius (2000) but with an additional step linking the phytomass demand with water productivity. Five different animal systems, each with several subsystems have been identified suited for analysis following the system of Wirsenius (2000) and the nature of input data listed in table 4. Results are presented as consumptive water use and virtual water content of the products generated from these animal systems also listed in table 4 as animal food items. Table 4: Animal production systems in livestock production and their associated livestock food items Animal production system

Livestock food item

Beef cattle

Beef meat

-Buffalo subsystem Cattle milk producing system

Milk

-Buffalo subsystem Sheep

Sheep and goat meat

Goat Pig

Pork meat

Poultry

Poultry meat

Egg producing system

Egg

Livestock production in the selected regions for the sample year of 2005 has been calculated to analyze consumptive water use and virtual water content per kilo in relation to the production size and illustrated in figure 6. Beef and milk cattle dominate production where milk production is largest in terms of weight for the majority of studied regions. Europe dominates milk production with over 200 million tonnes (MT). Sub-Saharan Africa seems to be the region with the smallest livestock production in general. Asia is the dominating producer for livestock products other than milk (dominated by Europe) and beef which is largest in Latin America and Caribbean.

19

Figure 6: Livestock production divided between animal systems in selected regions for 2005. Source: FAO STAT (2010).

The flow chart is presented in figure 7 and data is put into the flow chart in the right hand box in the figure (figure 7) consisting of livestock production data from FAO. Following the flow chart calculations with the input data have been made primarily of feed use linked to production size based on feed requirement conversion factors and feed mix compositions for each animal system and subsystem. These are based on data withdrawn from the FPD model. Input data from FAO are in terms of tonnes of production. When converted using feed requirement factors from the FPD model (Wirsenius, 2000;2010) the production will be represented as total amount of feed use in tonnes of dry mass.

20

Feed composition of animal system in selected regions from FPD

Virtual water content per CFT and region from LPJ

Green water use & productivity of livestock production in selected regions

Temperate cereals

Temperate cereals

Rice

Rice

Maize

Maize

Tropical cereals

Tropical cereals

Pulses

Pulses

Temperate roots Tropical roots

Blue water use & productivity of livestock production in selected regions

Sunflower Soybean

Total feed use divided between 13 Crop Tropical roots Functionl Types, CFTs Sunflower (tonnes of DM)

Total feed use of livestock production system in selected regions (tonnes of DM)

Temperate roots

Soybean

Groundnuts

Groundnuts

Rapeseed

Rapeseed

Pasture & grass

Pasture & grass

Other

Other

Feed requirement of animal system from FPD

Figure 7: Flow chart model: Complete process used for the data analysis of livestock production of the seven different regions

Feed use differs between animal system, CFT and region. Figure 8 illustrates the total feed use of different animal systems over the selected regions. Individual feed use of all animal production systems are presented in Appendix C. Largest numbers are found throughout Latin America, Asia and Africa where quantities of used feed are larger than those of North America and Europe. The overall feed use of cattle tend to be higher for all regions with a few exceptions such as feed use of pig in e.g. East Asia and feed use of where some regions does not have livestock production of buffalo. Feed use of milk cattle is e.g. of significant quantity in Europe and hence the production of the sector (figure 8). The same correlation of large production numbers and feed use of high quantity is found for all animal sectors although a trend can be seen that the feed use in developing regions tend to be higher in relation to production than that of developed regions. E.g. the production of milk cattle in Europe is several times larger than in Latin America although the quantities of used feed are of the same numbers. Milk cattle tend to dominate production in most of the selected regions (figure 6) whereas beef cattle use the largest amount of feed following the results of feed requirements (Appendix B). These figures clearly imply that the production quantity as well as feed use of different animal systems varies between regions. Asia is the region where most buffalo production takes place and the only region where feed dedicated to buffalos is significant.

21

Livestock production numbers. Input data from FAO (tonnes)

Figure 8: Total feed use of animal systems in selected regions for 2005.

In the next step of the flow chart the total amount of feed categorized according to the FPD feed composition for the different animal systems is reclassified to match the crop functional types (CFTs) applicable to the LPJ model. The results are summarized in figure 9 A presented as feed use of each CFT as a percentage of the total amount of feed used in each region. The individual feed compositions of pig production system and beef production system are illustrated in figure 9 B and C. All feed compositions are further presented in Appendix A.

22

A)

Other Non eaten food

100%

Fish meal

90%

Chicken meal Goat meal

80%

Sheep meal Beef buffalo meal

70% Percentage

Dairy buffalo meal 60%

Beef cattle meal Dairy cattle meal

50%

Pasture & Grass Rapeseed

40%

Groundnuts

30%

Soybean

20%

Sunflower

10%

Temp roots

Trop roots Trop Cereals

0% East Asia

Latin North North Africa South & Sub-Saharan America & & West Asia America & Central Asia Africa Caribbean Oceania

Maize

Europe

Rice Temp Cereals

B) Other

100%

Non eaten food

90%

Fish meal Meat-type chicken meal

Percentage

80%

Goat meal

70%

Sheep meal

60%

Beef buffalo meal

50%

Beef cattle meal

Dairy buffalo meal Dairy cattle meal

40%

Rapeseed

30%

Groundnuts Soybean

20%

Sunflower

10%

Tropical roots Temperate roots

0% East Asia

Latin N. Africa North South & C. Sub-Sah. America & W. Asia America Asia Africa

Europe

Tropical Cereals Maize Rice Temperate Cereals

C)

Percentage

100%

Other

90%

Pasture & Grass

80%

Rapeseed

70%

Groundnuts

60%

Soybean Sunflower

50%

Tropical roots

40%

Temperate roots

30%

Tropical Cereals

20%

Maize

10%

Rice

0% East Asia

Latin America

N. Africa & W. Asia

North America

23

South & C. Asia

Sub-Sah. Africa

Europe

Temperate Cereals

Figure 9: Regional feed use compositions presented as the total feed uses expressed as CFT percentages of total feed use (A), the feed composition of the pig livestock system in selected regions (B) and the feed compostionf of beef cattlelivestock system in selected regions (C).

The matching of input between the models generated some problems. Feed composites such as “other” and “non eaten food” does not match any of the CFTs generating LPJmL results. An average LPJmL for all CFTs were used for the category of other (consisting of only plant items), however, the LPJmL model calculates values only for phytomass wherefore the feed composite non eaten food constituting of other items than phytomass cannot be given the same average. Data from food balance sheets (FAO STAT 2010, year 2005) of the relation between calorie intake from animal and vegetable food in concerned regions were used to calculate a second average LPJmL of this feed composition category using the plant LPJmL average and an average of the results from the other animal food productions. The final step of the flow chart is when the feed amount in tonnes of dry mass of CFTs is converted to water use with virtual water contents, divided into green and blue water, from the selected regions derived from the LPJ model. This delivers results of virtual water content of livestock production systems and their produced products as well as their total water footprint in form of their total water use in selected regions. The results of such calculations as well as the development and establishment of the used feed requirement factors and feed compositions are as mentioned presented in Appendix A and B. The calculation of actual amounts of feed use for each livestock production system in selected regions is presented in Appendix C. Since the FPD model is based on input data from FAO from 1993 (FAO STAT, 2000) a publication of Wirsenius, Azar & Berndes (2010) has been used to enable an interpolation of inputs from FPD to the sample year of this research. The consulted study contributes with data based on the FPD model (Wirsenius, pers. comm., 2010) from different prediction scenarios of future differences in parameters such as land use, efficiency of animal systems and change of production system influencing feed requirement and feed mixes. These data have then been interpolated with the early data to fit the dataset for the year of analysis namely 2005. FPD input in this study has therefore been from the years of 1993 and 2030. Presented feed requirements in Appendix B are the interpolated values based on data and calculation between the two data sets.

24

Feed requirement of animal system from FPD

Feed compositiont of animal system from FPD

Total feed use of dairy milk cows in region (tonnes of DM)

Milk production. Input data from FAO

Fraction of meat production from milk production system from FPD

Meat production. Input data from FAO

Total feed use of dairy bulls & replacement heifers (tonnes of DM)

Total feed use of beef cattle (tonnes of DM)

Temperat

Feed use e cereals for milk Rice production divided Maize between 13 Crop Tropical cereals Functionl Types, Pulses CFTs (tonnes of Temperat e roots DM) Feed use for meat production divided between 13 Crop Functionl Types, CFTs (tonnes of DM)

Virtual water content per CFT and region from LPJ

Rice Maize Tropical cereals Pulses

Tropical roots

Temperat e roots Tropical roots

Sunflower

Sunflower

Rapeseed

Blue water use & productivity of milk Groundnut production in s selected regions Rapeseed

Pasture & grass

Pasture & grass

Other

Other

Temperat e cereals

Temperat e cereals

Rice

Rice

Groundnut s

Maize Feed use for meat Tropical cereals production divided Pulses between 13 Temperat Crop e roots Functionl Tropical Types, roots CFTs (tonnes of Sunflower DM)

Maize Tropical cereals Pulses Temperat e roots Tropical roots

Rapeseed Pasture & grass Other

Green water use & productivity of meat production in selected regions

Sunflower

Green water use & productivity of meat Groundnut production in s selected regions Rapeseed Soybean

Groundnut s Feed composition of animal system from FPD

Green water use & productivity of milk production in selected regions

Soybean

Soybean

Soybean

Feed requirement of animal system from FPD

Temperat e cereals

Virtual water content per CFT and region from LPJ

Pasture & grass Other

Figure 10: Flow chart model: Cattle milk & Cattle meat production system. Red arrows illustrate feed used for meat production and blue arrows feed use for milk production.

Figure 10 illustrates the second flow chart developed for this research. The chart describes the path production of dairy cattle cows and beef cattle will follow and the same path applies for products from the livestock systems of buffalo. Production systems producing egg and milk products have a more complex nature than other production systems. This is due to their contribution to the meat production in terms of carcass bodies of those animals whose primary purpose is producing egg and milk products but that will generate meat products as they are taken out of production and replaced. Dairy cattle will contribute further with carcass bodies of calves and replacement heifers that will not serve as dairy cows and will go directly to slaughter. The feed requirement for these animals with secondary production is assumed to differ throughout their lifetime. The carcass body of a dairy cow is in this research assumed to be produced mainly before her first calve and with that her actual conversion to a dairy cow. Before the first calf she should be considered as a replacement heifer and they have a higher feed demand than dairy cows. By assuming that the body, later constituting meat products, is produced before the first calf the feed required for the meat production of dairy cows can be separated from the total amount of feed used for milk production. Since this feed, although the direct cost will fall upon the dairy cattle production system, should be considered together with the rest of the feed required for meat production. When presenting results of livestock products in this manner a complete water use

25

footprint for different livestock products will be generated and not only the footprint of different livestock systems. The feed that the dairy cow eats in the remaining life is assumed to be invested in milk production in its full amount. This assumption generates a small bias since some of the feed is still required for the functioning of the animal and therefore invested in the body maintenance of the dairy cow hence her meat production and not her milk production.

26

4. Results Results are presented as green and blue water divided between consumptive water use and virtual water contents of the produced livestock products: beef, sheep and goat meat, poultry meat, pork meat, egg and milk for the selected regions.

4.1 Consumptive water use Consumptive blue and green water use of the different livestock production systems have been calculated with the LPJmL model which delivers results in form of production quantity of CFTs and virtual water content of each CFT in m3/ton of drymass (DM) on a country level. The total amounts of consumptive blue and green water used for each production system are then calculated with the summary of the quantity of each CFT used as feed for each region. Generated results presents total quantity of green and blue water required for the production of each livestock animal system in the selected regions. Figure 11 illustrates the total consumptive water use in selected regions for all livestock products. The bars present the consumptive blue and green water use as fractions of the total amount. Results vary between around 500-2800 Gm3 between regions with South, Central and East Asia being the regions where consumptive water use is largest. The consumptive water use seems to be of larger magnitude in developing regions with the exception of North Africa and West Asia with the smallest consumptive water use followed by North America, Oceania and Europe, all regions using less than 1000 Gm3 for livestock production. Fractions of consumptive blue water use are significantly smaller than green water and highest in South and Central Asia with close to 20% of the total consumptive water use constituted of blue water resources. Industrialized regions and Sub-Saharan Africa have the smallest fraction of blue water of their total consumptive water use.

Figure 11: Consumptive water use of all livestock products in selected regions

27

Consumptive blue water use for different livestock products illustrated in Figure 12 A are much less than consumptive green water use for all livestock production and regions ranging between 0-350 Gm3 as contrasted with 0-1450 Gm3 for green water (Figure 12 B).

Figure 12: Consumptive blue water use (A) and green water use (B) in livestock production divided between products and selected regions.

The amount of blue water is significantly highest for production of beef in amounts up to 350 Gm3 in South and Central Asia and seems to be largest in developing regions and Asia in particular. Correspondingly the consumptive green water use for beef is dominated by South and Central Asia with amounts close to 1500 Gm3 followed by Latin America & Caribbean with amounts up to 1000 Gm3 respectively. South and Central Asia dominates consumptive green water use in general, not only for beef production e.g. the region demands about 550 Gm3 for milk production. Milk 28

production uses between 100-550 Gm3 of water and beef products in a range of 250-2700 Gm3. The variation between regions are of great nature and for beef products the average global usage is less than 1000 Gm3 and milk production normally demands around 100 Gm3. Europe, North Africa and West Asia are the regions where water requirement of beef production are smallest. However, for milk production Europe stand for one of the largest requirements although the amount of virtual water used for milk is still of smaller nature than that for beef production. The consumptive water use of sheep and goat meat is much more evenly distributed than for other ruminants and their livestock products. The consumptive water use is dominated by the developing world. Europe, North America and Oceania require much less water resources than other regions. Largest quantity of virtual water diverted to sheep and goat production is used in South and Central Asia (375 Gm3) where the blue water requirement constitutes a rather large fraction of about 50 Gm3. The second largest consumptive water use for production takes place in Sub-Saharan Africa (225Gm3) however the fraction of blue water is evidently smaller for this region in comparison to other regions with high consumptive water use for sheep and goat meat. Other developing regions have consumptive water use in amounts of 175-225 Gm3, where the blue water fraction constitutes approximately 25 Gm3. The only monogastric production system that singlehandedly shows large consumptive water use for a specific region is pork meat production. Consumptive water use is distinctly largest in East Asia with quantities of approximately 900 Gm3 for both green and blue water followed by Europe (400 Gm3). East Asia also requires the largest amount of blue water for pork production with almost 100 Gm3 where the second largest consumptive blue water use only constitutes 25 Gm3 (Europe). Poultry livestock systems have the smallest consumptive virtual water usage when livestock products from sheep and goats are considered together. Consumptive use from egg production is distinctly dominated by East Asia with a quanity of 120 Gm3 and the significant blue water fraction close to 20 Gm3. The demand is evenly distributed between other regions and in the range of 20-40 Gm3, with Latin America and Caribbean as the second lhighest demanding region and North America and Oceania the lowest. Poultry production Latin America and Caribbean has the largest consumptive water use (about 120 Gm3) followed by East Asia (about 100 Gm3). However, the largest consumptive blue water use takes place in East Asia with approximately 20 Gm3. The fraction of blue water is large in several regions for poultry and egg production. Asia has a large consumptive blue water use in general with over 10 Gm3 for all production which should be considered a significant amount in relation to the total consumptive water use of the production.

29

4.2 Virtual water content The Virtual Water Content, VWC, of each grown crop functional type have been calculated in relation to where they are grown and the amount of produced CFT in that particular region in figure 13.

Figure 13: Virtual water content of the crop functional types (except the CFT of non eaten food) grown in selected regions.

The figure clearly illustrates that VWC differs widely between regions as well as type of CFT. Peaking values of e.g. groundnuts in Europe have VWC of up to 7000 m3/ ton of dry matter (DM). SubSaharan Africa has a significantly high value of VWC for rapeseed (6000 m3/ ton of DM), Latin America and Caribbean has an almost as high VWC for pasture and grass and tropical roots has a VWC of close to 5000 m3/ ton of DM for tropical roots and tropical cereals. If single peak values are disregarded tropical cereals in general have high values of VWC in all regions. The lowest VWC are those of temperate roots with a VWC under 500 m3/ ton of DM in all regions except for South and Central Asia. Virtual water contents for each animal product in all regions have been calculated and are presented in Figure 14. South and Central Asia has the highest virtual water contents (VWC) where ruminants generates the peak values (300 and 250 m3/kg) in contrast to Europe with the smallest VWCs of approximately 10-20 m3/kg. The virutal water content of industrialized regions are distinctly lower than the developing regions and Asia in particular. In general, ruminant livestock products dominate virtual water contents as well as consumtive water use. Regional peak values of pork and poultry products can be found in e.g. Europe and Sub-Saharan Africa.

30

Figure 14: Virtual water content (VWC) for each animal product in selected regions

VWC of each product has also been calculated relating consumptive water use to production numbers for each livestock product in the selected regions.

31

Figure 15 illustrates the results in form of blue virtual water content (A) and green virtual water content (B) per livestock product and between selected regions.

Figure 15: Virtual water content of different livestock products in selected regions illustrated as blue virtual water content in figure (A) and green virtual water content in figure (B) divided between livestock products.

The figure clearly illustrates that both blue and green VWC of beef products is distinctly higher than the VWC of other livestock products in general. The virtual blue water content for beef ranges between less than 5 m3/kg in Europe and 60 m3/kg in South and Central Asia and the green VWC from about 20 m3/kg in Europe to 250 m3/kg in South and Central Asia which are the regions with the highest and lowest VWCs. Sub-Saharan Africa has the second highest value although the fraction of blue water is very small in relation where North Africa and Asia have a higher VWC of blue water. Sheep and goat meat have the second highest virtual green water content ranging from 0-200 m3/kg (from zero since not all regions produce sheep and goat meat). South and Central Asia has a 32

remarkably high VWC (200m3/kg) in comparison with other regions and Asia seems to have higher VWC for sheep and goat meat in general. The blue water VWC for sheep and goat ranges between 050 m3/kg with the significantly highest content in South and Central Asia followed by North Africa and remaining Asia. Poultry and egg products as well as pork and milk all have very low values of both blue and green virtual water content. The variation between regions for these products also seems smaller than that of other livestock products. Milk production has the lowest virtual water contents (between 1-10 m3/kg) with low values of both blue and green virtual water contents. The highest total VWC is found in Sub-Saharan Africa followed by North Africa and Asia (excluding East Asia) (9,5 and 4 m3/kg) where North Africa and Asia has higher VWC of blue water. Developed regions has distinclty lower VWC than the developing world. In poultry and egg prodcution the blue water fraction of VWC is significantly larger than for milk production and constitutes a large fraction of the total VWC in all developing regions except SubSaharan Africa. However last-named region has the significantly highest total VWC for both products of about 16-18 m3/kg . In South and Central Asia the blue water VWC constitute one third of the total VWC in the region. Pork production represents the highest VWC of the monogastric production systems with SubSaharan Africa in top with around 40 m3/kg of DM. Lowest VWC is found in North America and Caribbean (4 m3/kg of DM). VWC fraction of blue water resembles that of poultry and egg production with significant values in all developing regions except Sub-Saharan Africa. Europe, Latin America and Caribbean have significant green virtual water contents for pork production but the blue virtual water content is insignificant. North Africa, West Asia and East Asia has high virtual water contents of both blue and green water. Evident is that South and Central Asia have the highest values of blue virtual water content for all livestock products with the only exception of pork production where North Africa and West Asia have a higher content. Blue virtual water content is clearly higher for developing regions than developed regions where the values for Europe, North Amercia and Oceania are insignificant for all products except beef production although values are still low in relation to other regions. Virtual green water content varies more between regions and products than blue water content but the results corresponds fairly well with the highest values in South and Central Asia for ruminant livestock products. For monogastric livestock production and milk however, Sub-Saharan Africa have the highest green virtual water content in general dominating both pork, milk, egg and poultry and following South and Central Asia with the second highest green virtual water content of beef production in comparison with having only the fourth largest blue water VWC for beef production. The results of total consumptive green-, blue and total water use and virtual water content for each livestock products and regions are summarized and presented in table 5 below. The total consumtive water use estimated in this research constitutes of 9680 km3 of water where 8600 km3 is constituted by green and 1085 km3 is constituted by blue water resources.

33

Table 5: Summary of results calculated with FPD and LPJmL models in this study. Consumptive water use (km3)

Blue water Green water Total

Virtual water content green + blue water (m3/kg)

East Asia

Latin America & Caribbean

North Africa & West Asia

North South & SubAmerica Central Saharan & Asia Africa Oceania

237

61

112

28

566

22

59

1 085

1 834

1 380

488

709

2 218

1 086

882

8 597

737

2 784

1 108

941

9 682

2 072

East Asia

1 441

Latin America & Caribbean

600

North Africa & West Asia

North South & SubAmerica Central Saharan & Asia Africa Oceania

Beef

83,0

61,9

115,7

27,1

308,9

Sheep & Goat

87,9

0,0

64,3

36,1

5,8

7,3

1,9

3,2

Pork

16,3

12,8

21,0

Milk

1,8

2,2

4,1

Poultry

Egg

Virtual water content green + blue water (litres/kg)

Europe Total

3,9

East Asia

6,3

Latin America & Caribbean

6,2

North Africa & West Asia

Europe Average

186,6

20,1

114,7

243,5

0,0

14,0

63,7

10,2

16,9

3,4

7,0

4,1

12,1

40,7

15,9

17,6

1,0

3,7

9,5

1,0

3,3

2,3

7,4

14,7

2,4

6,2

North South & SubAmerica Central Saharan & Asia Africa Oceania

Europe Average

Beef

83 000

61 900

115 700

27 100

308 900

186 600

20 100

114 700

Sheep & Goat

87 900

0

64 300

36 100

243 500

0

14 000

63 700

5 800

7 300

1 900

3 200

10 200

16 900

3 400

7 000

Pork

16 300

12 800

21 000

4 100

12 100

40 700

15 900

17 600

Milk

1 800

2 200

4 100

1 000

3 700

9 500

1 000

3 300

Egg

3 900

6 300

6 200

2 300

7 400

14 700

2 400

6 200

Poultry

34

5. Discussion This study was performed to assess the regional variation and differences between livestock food items in terms of their consumptive blue and green water use and blue and green virtual water content. Results are presented as total numbers as well as divided between blue and green water resources to clearly illustrate the purpose of this study.

5.1 Consumptive water use Results show that consumptive water use for beef production in terms of both blue and green water resources is larger than for all other livestock food items. This can depend on the distinctly large beef production numbers in Latin America in relation to the consumptive water use which is significantly higher than for Asia and Africa. Milk production has the second largest consumptive water use and result patterns comply with beef production trends where consumptive water use is largest in South and Central Asia followed by Latin America. Total consumptive water use for sheep and goat meat production are low in relation to other ruminant livestock food items but is in general higher than for monogastric food items with the exception of pork production in some regions. In general the consumptive blue water use of livestock products is highest for Asia and South and Central Asia in particular. Other regions with a high consumptive blue water use and virtual blue water content are North Africa, West Asia, Latin America and Caribbean. All mentioned regions are developing regions with a high temperature climate and suffering from water scarcity of some extent. This results in a higher demand for irrigated agriculture in relation to regions richer in blue and in particular green water resources. Since mentioned regions are all largely depending on agriculture they are forced to manage water scarcity and make sure that enough water resources are dedicated to support agriculture. Manage solutions have in the past been dominated by irrigated cultivation which is directly linked to large withdrawals of blue water resources (Falkenmark & Rockström, 2004). This further explains their dominance of consumptive blue water use in relation to other regions. In pork meat production the largest requirement of green and blue water resources is dominated by East Asia although Europe requires a noticeable amount of green water resources. Production numbers agree with consumptive water use results where South and Central Asia produces most pork meat followed by Europe. The same pattern applies for feed use, however, differences are larger. Feed use is six times larger for South and Central Asia than and twice as large for Europe as remaining regions whereas differences in production numbers are not as significant. Consumptive blue water use for poultry meat and egg production is both dominated by East Asia where the most differentiating result is that egg production requires a slightly higher amount of blue water resources than poultry meat. Green water requirement of poultry meat production and the consumptive water use of such in total are highest in Latin America and Caribbean and where East Asia requires the second largest amount whereas for egg production opposite order applies. East Asia has an egg production more than twice as big as for all other regions which could explain the large consumptive water use in the region. For poultry meat the production is highest in North Africa and West Asia but there is no large difference between developing regions in magnitude of production numbers. Feed use for egg production is generally low for all regions and the same

35

applies for poultry production although feed use of the latter is slightly higher. Feed use follows production numbers and is dominated by South and Central Asia followed by Europe.

5.1.2 Virtual water content in relation to consumptive water use

It is evident by the results from this study is that feed use of animals has a large impact on their associated consumptive water use. Results of consumptive water use are generally higher for Asia and especially South and Central Asia. Beef and milk production greatly depend on grass as feed for animals. Grass or grasslike CFTs in form of fodder and pasture from managed grassland or as pasture from non managed grasslands requires large land areas for growing. Land areas of such magnitude require large amounts of water resources, in particular green water resources (mainly in form of precipitation). Managed grasslands might also require irrigation to some extent and therefore have additional consumptive blue water use. The dominance of grazing as the largest feed composite for ruminants explains the significantly higher consumptive water use for the livestock products they produce. The distinctly higher requirement of feed in terms of dry mass (expressed in weight) is also explained by features of grazing as the dominant feed composite. To receive the same amount of calorie intake from grazing and fodder in relation to intake of high energy containing feed such as grains, a much larger amount of feed have to be consumed (Wirsenius, 2000; Steinfeld et al., 2006). Results primarily seem to depend on that feed use is significantly larger for ruminants than for monogastrics where ruminants eat more than twice as much feed and fodder in general. This feature, however, only explains generated results to some extent. If feed use would indeed be the only factor influencing consumptive water use of ruminant livestock products their water requirement would have to be higher for all situations where feed use was higher. When comparing results of feed use and consumptive water use this appears not to be true in some cases. In the case of milk production in Europe the feed use (in tonnes) is about four times greater than that of pork meat production in the same region (and production of milk is about five times larger in Europe in relation to pork production). However, the total consumptive water use of green and blue resources for milk production is about 220 km3 in relation to close to 400 km3 for pork meat production in Europe. Results of consumptive water use therefore have to depend on the composition of feed for ruminants and monogastrics rather than the amount of feed they have consumed. Feed compositions of monogastrics are quite similar in all regions and the same applies for the different ruminants. This indicates that in the case of milk and pork production in Europe the internal feed composites of a pig must require such large amounts of water for cultivation in the region (hence the differences of virtual water contents of the CFTs depending on where they are cultivated illustrated in figure 13) to generate such a high consumptive water use. It can also be argued that if ruminants would have a feed composition similar to that of monogastrics their consumptive water use would be largely affected. Primarily the effect would be that the smaller amount of required feed would result in a smaller amount of required water resources. The consumptive water use of grazing lands would decrease if livestock production changes towards more intensive farming. However, the fraction of the ruminants feed, today consisting of grasslike CFTs would be substituted with for example maize, grains or other high energy containing

36

concentrated feed which evidently have a significantly higher requirement of water in some regions as in the example of pork and milk production in Europe. This could result in that even if the feed requirement of ruminants would decrease largely there is a possibility that they would require more water resources than expected.

5.2 Virtual water contents The difference in virtual water content and consumptive water use of ruminant livestock food items in relation to monogastric products are mostly due to the differences in consumed feed amounts. However, results show regional differences that can only be explained with the composition of their feed in the same way they explained differences between consumptive water use of pork and milk production in Europe. Such results are even more distinct among monogastrics since their feed compositions are much more diverse than that of ruminants mostly containing the CFT of grass and pasture. The blue and green virtual water content of beef products agrees with the results of water requirement, however regional differences appear when comparing consumptive use and VWC. South and Central Asia has the largest consumptive water use followed by Latin America and Caribbean. Virtual water contents are also highest for South and Central Asia however followed by Sub-Saharan Africa and not Latin America which has lower virtual water content of beef products than several regions of lower consumptive water use. Both regions have relatively low production numbers in relation to the amount of consumptive water use generating high virtual water contents for beef products. Sheep and goat meat have a rather low consumptive water use in general. Despite this the products have high virtual water contents primarily in Asia but also to some extent in North America and Oceania following the trend of ruminant livestock products. Milk products have significantly lower virtual water content than all other livestock food items. This is due to that global milk production is more than four times greater than other livestock products expressed in weight. Virtual water content of pork food items does for example not comply with consumptive water use patterns to any great extent. Green virtual water content is distinctly highest in Sub-Saharan Africa followed by North Africa and West Asia whereas blue VWC is highest in North Africa and West Asia closely followed by South and Central Asia. Since results of consumptive water use, feed use and production does not conform to virtual water contents the high VWC for example in Africa therefore depends on the diet of pigs not amount of consumed feed. For all regions with significant virtual water contents of pork meat the feed to some extent constitute of non eaten food which has significantly higher virtual water content than other CFTs. The virtual water content of non eaten food is not given by any model but calculated through assumptions of regional consuming patterns. It can be argued that this virtual water content should be set to zero since it is embedded in other VWCs of cultivated crops. If the animals would eat another feed composite instead of non eaten food however that composite would have a virtual water content of its own. This resulted in that non eaten food in this study was concluded to

37

generate most vital results if treated separately but appointed with a virtual water content calculated according to its regionally assumed content. Feed compositions in regions with the highest virtual water contents largely constitutes of tropical roots, temperate roots, maize and soybean of varying percentage. This further explains the higher virtual water contents where for example maize have a significantly high VWC in Sub-Saharan Africa and tropical roots in North Africa and West Asia (Figure 13). The green virtual water contents regarding poultry meat and egg production once again peaks in Sub-Saharan Africa for both livestock food items, and second largest contents are found in Latin America and Caribbean. Blue virtual water content however is significantly higher for South and Central Asia for both poultry meat and eggs although the content in North Africa and West Asia is almost as large. The diet of both meat- and egg producing chickens in Sub-Saharan Africa largely constitutes of maize which as mentioned above has a significantly high VWC for the particular region. Latin America and Caribbean has the second highest VWC for cultivation of maize explaining the results of the second largest virtual water content for all animal products since maize is a significant feed composite in this region as well. Furthermore the diet of poultry for both products constitutes of several other CFTs such as groundnuts, soybeans and cereals, all with high VWC for concerned regions. In general the feed composition for regions with high VWC and consumptive water use tend to have much more diverse feed composition and constitute of temperate cereals and maize to a larger extent than regions where results show lower consumptive use and virtual water contents.

5.3 Comparison, new perspectives and prior studies In this thesis the virtual water content rather than the water footprint of a product have been calculated since the assumption that all feed incorporated into livestock production was produced locally. To be able to express received results as actual water footprints of livestock products this assumption would not be possible and trade of feed composites would have to be included in the calculations to enable a geographically valid indicator. Assessments of virtual water contents and water productivity associated with livestock production vary greatly between different studies prior to this analysis. The virtual water content of livestock products expressed as water productivity by Chapagain & Hoekstra (2003) are 0,082 kg/m3 for beef, 0,12 kg/m3 for sheep and goat meat, 0,29 kg/m3 for pork, 0,22-0,51 kg/m3 for poultry meat and 0,79 kg/m3 for milk. Results from this study are in relation from an average virtual water content for beef 0,0087 kg/m3, sheep and goat meat 0,016 kg/m3, pork 0,057 kg/m3, poultry 0,14 kg/m3 and milk 0,30 kg/m3. Water productivities of livestock food items from Chapagain & Hoekstra (2003) are up to ten times higher than calculations from this thesis. However the numbers from Chapagain & Hoekstra (2003) have been argued to be lower in relation to other studies prior to this analysis (don Peden, 2007). When results are put in relation to total consumptive water use of agriculture calculated from different models (table 1) the great differences between assessments are even clearer hence the importance of which model is used for calculation. Zimmer & Renault (2003) roughly estimated that 45 % of the total consumptive water use of food production is dedicated to feed for live animals. When comparing the total consumptive water use of

38

about 9700 km3 calculated in this research with those numbers of consumptive water use of agriculture listed in table 1 such assumption does not seem far from the truth if one of the higher results of agriculture’s consumptive water use should be reviewed. The assessment of the worlds consumptive water use in Livestock´s long shadow (Steinfeld et al., 2006) for barley, maize soybean and wheat is in the size range of 1300 km3 annually. In relation to the average feed use composition those four crops constitutes about 30 % of feed given to animals indicating lower overall consumptive water use than in this study. Other assessments of water footprints for livestock products range between 16 000-30 000 liters per kg beef (WFN, 2010) about ten times less than the largest virtual water content of beef estimated in this research. If the assumption of the fraction of consumptive water use associated with livestock production by Zimmer & Renault (2003) would be accepted these estimates seem to low in relation to the consumptive water use of agriculture calculated by the majority of models in table 1 and livestock production size relationships similar to 2005. Results from this study however remain high in relation to those numbers but resemble former assessments of a water requirement of 100 000 liters of water per kg beef (however grainfed) (don Peden, 2007). The difference between results can be due to many reasons. Variation in assumptions of consumptive water use regarding forage can be as large as seventyfold (don Pedro, 2007). The greatest impact on the results however seems to be in which way grazing lands are regarded and accounted for. The CFT that represents grazing, pasture and grass, constitute the largest feed use in all regions alone accounting for more than 45 % and up to 80 %. In former studies there have been assumptions about grazing animals eating approximately half or even one third of the above ground biomass (don Peden, 2007; de Fraiture, 2007). There is a general disagreement on how grazing lands should be treated and accounted for with different ways of treating the consumptive water use (Steinfeld et al., 2006; Falkenmark & Rockström, 2004; Deutsch et al., 2010; Chapagain & Hoekstra, 2003) In this research grazing lands have been treated with a new approach. Since input in this study have been livestock production and the feed requirement of such the calculations have been performed from the actual requirement of grass expressed in dry mass and not starting from the land areas consisting of grass lands and their consumptive water use as a total. This approach induces that no assumptions of how much grass the animals eat of the total amount of grown grass on grasslands is necessary since such features are already known. The numbers derived from the LPJmL model have been of virtual water content on a national level and calculations have been made of the actual consumptive water use for growing a certain amount of grass in a particular region. The new approach also generates that the results do not necessarily accord with any other results found in literature. Reviewing the results from this new approach can lead to conclusions that former assumptions of how much of consumptive water use of grazing lands should be considered consumptive water use of livestock are understated. If input is amount of used grass (in weight) as in this research the consumptive water use tend to be significantly higher than what has been assumed in past studies. This conclusion should be considered even more valid since the LPJmL model used for calculation in this thesis generates amongst the lowest results of consumptive green water use in general (table 1) and for grasslands in particular indicating that other models using the approach from this study would receive even higher results.

39

The results of virtual water contents and consumptive water use from this thesis also indicates that it is first when results of the LPJmL are connected with the actual feed use of livestock (where grass like crops are the largest feed composite globally) the results of higher water productivity and consumptive water use are received. This implies that it is merely the amount of grass consumed by livestock that have been understated and not the virtual water content of the CFT in particular regions.

5.4 Assumptions and uncertainties To carry out this study (as well as for other studies accounting for consumptive water use of livestock production) assumptions on how to deal with ruminants and grazing lands had to be adopted. The assumption used in this study is likely to generate higher results than in reality. It is likely that some of the grazing lands accounted for does not generate any land degradation. Such grazing lands would not require water resources that could be used elsewhere, wherefore it is not certain that their consumptive water use should be included. To enable a correct exclusion however, each grazing land area should be considered separately to identify their proper impact, which in the frame of this study has been impossible. Without detailed knowledge about all grazing lands the inclusion of all their water demand will generate results of highest scientific validity. The uncertainty generated from the assumption that feed consumed by dairy cattle before their first calve is dedicated to meat production should also be mentioned. However, the uncertainty that would appear if all feed was instead diverted solely to milk production or meat production would generate a much larger bias. The part of meat production dairy cattle constitutes is of significant nature in some regions and in particular where milk production is rather small. Besides the great uncertainty when calculating for grazing and ruminants there are several other reasons for such fluctuations between different assessments. A contributing factor is that livestock production involves several animal systems contributing to a high biophysical diversity (don Peden, 2007). When investigating in global perspective additional uncertainties reside. There is great socioeconomic diversity between regions and the complex nature of describing all trade flows between nations in a trustworthy way. The demand for proper and applicable reporting of agricultural production, livestock production and imports and exports of all associated goods on a national or regional level is extremely high if results are to be valid. However such knowledge as well as knowledge of the water productivity of livestock production is greatly unknown and the demand for increasing knowledge worldwide grows larger. The gap of knowledge also applies for trade flows of feed, fodder and livestock food items which had to be completely excluded in this thesis. To generate a true picture trade flows can not only be described in terms of export and import numbers. The trade flow of a good have to be marked from where the export starts to where the good ends up since the LPJmL model is pixel based and calculates for production locally. The import in one country in terms of feed have to be appointed to the specific country where the crops have been grown to give the model the ability of correct calculation of virtual water content and consumptive water use of the good. There is also a great need of updating the feed requirements and feed compositions of the different animal systems. In this study feed requirements were calculated based on data from 2000 and the

40

generated uncertainties would be smaller if they were based on present numbers. The different feed compositions also have to be kept up to date since they differ with differing production systems etc.

5.5 Water scarcity Calculations in this study indicates that consumptive use of water resources are largest in Asia (with the exception of West Asia) followed by developing regions. Consumptive use of water resources is significantly smaller in developed regions as well as the virtual water content of livestock products produced there. Consumptive use of blue water resources follows the same pattern as overall use where semi-arid regions in Asia require the largest amount of blue water resources. Results of regional consumptive water use strongly agree with the global situation of water scarcity illustrated in figure 2. They also conform to the global picture of the situation of undernourishment (Figure 1). Future predictions strongly indicate that withdrawals will continue to increase in already water scarce areas in terms of both blue and green water resources (Figure 2). Predictions of increasing demand for livestock products also indicate that associated resource requirement will rise. Consumptive blue water use have to decrease and not increase since resource withdrawals are already pushed to an alarming level resulting in water deficit, the pressure on rainfed agriculture increases significantly. Calculations that approximately 15 % of water requirement from food production can be met by blue water resources (Falkenmark et al., 2007) indicates that with a demand for larger future agricultural yields the dependence of irrigated agriculture must cease. Development towards more effective rainfed agriculture will become crucial for meeting the rising demand for food (Falkenmark & Rockström, 2004). If green water losses can be effectively reduced in areas where consumptive green water use is large the result could greatly benefit food and livestock production. Since consumptive green water use constitutes the significantly larger fraction of livestock and agricultures water demand, and especially in areas most sensitive to green water scarcity due to climatic reasons, this development must be given the main focus in the nearby future to achieve the goal of food production and poverty reduction (Falkenmark et al., 2007; Falkenmark & Rockström, 2004). When viewing the regional differences between results of virtual water contents and consumptive water use of different livestock products in this study it is clear that the differences between regions depend on their regional virtual water contents and the composition of feed given to animals. The trend of the more water consumptive beef production due to grazing apply for all regions resulting in the larger virtual water content of such products. However the regional differences regarding this production (and all other livestock production not relying on grazing) depend on other differences in the feed composition of live animals. The increasing rate of consumptive water use associated with grazing lands are thought to decrease or stagnate over the next couple of years since livestock production tend to shift towards more mixed or intensive farming (Steinfeld et al., 2006; Kemp-Benedict, 2006). This could result in a reduction of consumptive water use associated with livestock production but could also generate the opposite development. Mixed and especially intensive farming depend on feed given to animals consisting of grains and other crops belonging to CFTs that in general have high virtual water content in relation to other crops (Wirsenius, 2000; Bondeau et al., 2007). The fraction of such concentrate 41

feed as well as other feed than grazing is expected to increase significantly in the near future and especially in the developing world. The results in this research indicate that these CFTs are the main contributing reason to high regional virtual water contents of livestock products. A shift toward mixed or intensive farming would demand that more agricultural land would be dedicated to production of such feed. Land areas that are today non managed or managed grass lands would have to be converted to high yielding agricultural land and require much more resources than the former grasslands (Steinfeld et al., 2006). This indicates increased pressure on regional agriculture to manage and decrease consumptive water use associated with agriculture and develop a less resource consuming agriculture production to prevent that consumptive water use of livestock increases even more in the nearby future.

5.6 Management options To decrease the pressure of livestock production on blue and green water resources there are some available management solutions dealing with situations of water scarcity. Livestock production can be associated with scarcity of both blue and green water resources (Falkenmark et al., 2007) and they have different mitigation methods. Options are usually of three types namely, reduced water use, reduced depletion process or improved replenishment of water resources (Steinfeld et al., 2006). Blue water scarcity induced by a large demand for resources is managed by more sustainable use of resources, reduction of unnecessary leakages and spills (e.g. by changing irrigation system to drip irrigation), reduction of household use and by preventing pollution of resources (Falkenmark et al., 2007). Other blue water scarcity situations are the result of a large population competing for a scarce resource. Such deficit is mitigated by reallocating resources which can be done in several ways e.g. by desalination of sea water, water storage or rainwater harvesting (Falkenmark et al., 2007). An important mitigating measure when dealing with green water scarcity is soil conservation and this can also be performed in numerous ways however where green water resources are scarce due to lack of rainfall the most efficient mitigation measure is through irrigation (Falkenmark et al., 2007). The main feature for managing water scarcity associated with agriculture is to develop a sustainable agriculture where the soil quality is maintained and where green water losses as well as blue water losses are minimized to increase water productivity. This study primarily highlights the importance of reviewing the feed given to animals in the livestock production and the selection of which type of livestock production is consumed or carried out. A shift toward more water productive composites in the feed could decrease consumptive water use significantly on a regional level. An increase in use of by-products such as crop residues, other wastes and non-eaten food would induce a reduction of water requirement (Steinfeld et al., 2006; don Peden, 2007) as well as using crops with as low virtual water contents as possible for production in a specific region. To shift production from ruminant livestock production would also generate a decrease of consumptive water use and lower virtual water content of products. Such a shift can be greatly influenced by the choice of consumers which relies on their knowledge about the differences in water productivity between both products and where they originate from.

42

Since trade is not accounted for in this research any conclusions of such impact cannot be made but evident is that if there is a necessity to use a certain crop as feed the consumptive water use might be significantly reduced if such composite is imported from a region where the virtual water content and consumptive water use is low. Management solutions to reduce consumptive water use falls on agriculture but also on policy makers and consumers. For consumers to understand the consequences of their consumption and to enable for them to make choices of for example more water conservative goods an easily comprehensive indicator is required. Water footprinting could be argued to be such a solution similar to carbon footprinting that have enabled a comprehensive consumer adapted life-cycle analysis regarding emission of greenhouse gases. It is argued however, that water footprinting does not generate the same reliable results since they are a summation of different kinds of consumptive water uses and from regions that differs in availability of water resources (Pfister & Ridoutt, 2009). What is missing in water footprinting is the relation to socio-economic and environmental impacts. To choose a good with lower water footprint does not necessary mean that the product is less harmful. If for example the water footprint of one product is lower than another but the former have a large fraction of consumptive blue water use it might be more environmentally beneficial to choose the product with the higher green water based water footprint (Pfister & Ridoutt, 2009). In a recent attempt of life-cycle analysis of two products in terms of water footprint Pfister & Ridoutt (2009) introduces water stress to the footprint bringing features of availability of water resources and water scarcity to the agenda which adds to results of virtual water contents a more broaden perspective bringing water footprinting closer to becoming as successful as carbon footprints for the consumers.

43

6. Conclusions Results were analyzed for six different livestock food items and seven regions to enable a global perspective and coverage of the complete livestock production. Calculations in this study have shown that: •

Total global consumptive water use of livestock production in this research is calculated to about 9700 km3 of water out of which 89 % is consumptive green water use. The result indicates higher volumes than most former assessments and the water requirement of the sector constitutes a significant part of the total consumptive water use of agriculture and the requirement of total global food production.

•

Consumptive water use is dominated by ruminants. Consumptive blue water use is significantly higher for South and Central Asia constituting approximately 20 % of the total consumptive water use associated with livestock production in the region. Other regions requiring a significant amount of blue water are other regions in Asia, Latin America, Caribbean and Europe. The green consumptive water use comply with blue water patterns and is of greatest magnitude in South and Central Asia with East Asia as the second largest consumptive green water user.

•

Virtual water contents (m3/kg) of ruminant livestock products (except milk) are higher in comparison to monogastric products. Virtual water content is (as total consumptive water use) dominated by South and Central Asia with the same distinct fraction of 20 % blue water content. The blue virtual water contents are however only significant in Asia. Regions with large green virtual water contents of livestock products are, despite South and Central Asia, Sub-Saharan Africa followed by North Africa and West Asia.

•

In general, both virtual water content and consumptive water use is of more significant nature in the developing regions with a few exceptions of specific products but when the regional production also tends to be distinctly large.

The methodology in this study has enabled an analysis that has identified peak values and coupled them with feed compositions as specific CFTs, amount of feed required, livestock food item and region. The data analysis and comparison of generated results further implies that: •

The higher virtual water contents and water requirement of ruminants compared with monogastrics depend on their feed composition. Their feed consumption largely consists of grazing which generates a significantly higher requirement of primarily green water resources in turn resulting in high virtual water contents. The amount of grazing for feed is of such distinctly larger amount than other feed for livestock production wherefore having such a significant imprint on results.

•

Variation patterns in virtual water contents and consumptive water use does not collaborate. Virtual water contents collaborate with neither amount of feed used nor production numbers which to some extent explain fluctuation patterns for consumptive water use. This indicate that it is primarily composition of feed constitutes that matter in terms of virtual water contents of livestock products and not the amount of feed they required for production. 44

•

The regional differences in virtual water contents can further be explained by the composition of feed. Regional differences in virtual water contents of feed composites that are of particular significance for a certain product in a certain region collaborates with regional peaking virtual water contents. For example Sub-Saharan Africa tend to have high virtual water contents for products where feed composition constitutes of maize by an important magnitude since this CFT has a distinctly high virtual water content when grown in the region.

•

The results of water requirement and therefore consumptive water use of different regions comply well with the global picture of water scarcity. With the predicted population growth resulting in an increased demand for food this indicates that the already large consumptive water use of livestock production can be expected to grow further.

This study should be considered as a pre-study and future research is of great importance to enable improved results and an understanding of the scale and impact of the livestock sector on future water resource handling. This indicates that: •

Studies of this kind are of great importance to enable livestock production to develop towards further efficiency. It is crucial for the sector to decrease consumptive water use, minimize water losses and develop towards lower virtual water content of production.

•

To support such development towards a more effective water usage further research has to be performed to fill the gap of knowledge of the consumptive water use of livestock production.

•

Calculations of grazing and other features of the production have to be further improved generating better estimations of the consumptive water use of livestock production.

•

Guide consumers to wise decisions when it comes to purchasing of livestock food items -

Less ruminant meat and more food items from pig and poultry production systems

-

Less animal foods altogether

45

Acknowledgments I sincerely would like to thank thank Mats Lannerstad and Louise Karlberg at Stockholm Environmental Institute and Stockholm Resilience Center for making this thesis possible. Not only by accepting me as an intern at SEI but with most valuable thoughts, discussions and comments throughout the thesis. In addition I would like to thank Stefan Wirsenius at Chalmers University for sharing his model data and making it possible to bring calculations forward by discussions and guiding through data analysis. I would also like to thank Jens Heinke and Holger Hoff at PIK research center and SEI for managing of the LPJmL model and valuable discussions and comments about the content of the thesis, analyze of results and hopefully continued research. Further I would like to thank Lisa Deutsch at the Resilience center for her assistance. Finally I would also like to thank Stockholm Environmental Institute for accepting me as an intern to write this thesis and Gothenburg University and Göran Dave for allowing me to write my thesis at SEI and comments throughout the work.

46

7. References Allan, J. A. (1993) Fortunately there are substitutes for water otherwise our hydro-political futures would be impossible. Priorities for water resources allocation and management, ODA. London, pp. 13-26. Allan, J.A. (1994) Overall perspectives on countries and regions. In: Rogers, P. and Lydon, P. Water in the Arab World: perspectives and prognoses, Harvard University Press, Cambridge, Massachusetts, pp. 65-100. Chapagain, A. K. & Hoekstra, A. Y. (2003) Virtual water flows between nations in relation to trade in livestock and livestock products. Value of Water Research Report Series No 13. Institute for Water Education, UNESCO-IHE. Delft, Netherlands. De Fraiture & C. Wichelns, D (2007) Looking ahead to 2050: scenarios of alternative investment approaches. In Molden, D (ed)(2007) Water for food ,Water for life: a comprehensive assessment of water management in agriculture. International water management institute. Earthscan. London, UK. Deutsch, L. Falkenmark, M. Gordon, L. Rockström, J. Folke, C. (2010) Water-mediated ecological consequences of intensification and expansion of livestock production. In Livestock in a changing landscape ed. Steinfeld, H. Mooney, H.A. Schneider, F and Neville, L.E. Washington DC, USA, pp 97111. Don Peden (2007) Looking Water and livestock for human development. In Molden, D (ed)(2007) Water for food ,Water for life: a comprehensive assessment of water management in agriculture. International water management institute. Earthscan. London, UK. Fader, M. Rost, S. Müller, C. Bondeau, A. Gerten, D (2009) Virtual water content of temperate cereals and maize: Present and potential future patterns. Journal of Hydrology, doi: 10.1016/j.jhydrol.2009.12.011 Falkenmark, M. (1995) Land-water linkages: A synopsis, land and water integration and river basin management. Land and Water Bulletin, 15-16. FAO (Food and Agriculture Organiszation). Rome, Italy. Falkenmark, M. & Rockström, J. (2004) Balancing water for humans and nature: The new approach in ecohydrology. Earthscan. London, UK. Falkenmark, M. Berntell, A. Jägerskog, A. Lundqvist, J. Matz, M & Tropp, H. (2007) On the verge of a new water scarcity: A call for good governance and human ingenuity. SIWI Policy brief, SIWI. Stockholm, Sweden. FAOSTAT (Food and agriculture Organization Statistics) (2000). Statistical database. Accessed in 2000 for sample year 1993. Available at: http://faostat.fao.org/ FAOSTAT (Food and agriculture Organization Statistics) (2010). Statistical database. Accessed in may 2010 for sample year 2005. Available at: http://faostat.fao.org/ Gerten, D. Schaphoff, S. Haberlandt, U. Lucht, W. Sitch, S (2004)Terrestrial vegetation and water balance - hydrological evaluation of a dynamic global vegetation model. Journal of hydrology, 286, 249-270. Hoekstra, A.Y. (ed) (2003) Virtual water trade: Proceedings of the international expert meeting on virtual water trade. Value of water research report series No 12. Institute for Water Education, UNESCO-IHE. Delft, Netherlands. Available at: http://www.waterfootprint.org/Reports/Report12.pdf

47

Hoff, H. Falkenmark, M. Gerten, D. Gordon, L. Karlberg, L. Rockström, J (2009) Greening the global water system. Journal of Hydrology 384, 177-186. Jewitt, G. (2006) Integrating blue and green water flows for water resources management and planning. School of Bioresources and Envirnomental Hydrology. University of KwaZulu-Natal, South Africa. Kemp-Benedict, E (2006) Land for Livestock Scenario Notes: Comprehensive assessment for water in agriculture. Stockholm Environmental Institute. Stockholm, Sweden. Lardy, G. Stoltenhow, C. Johnson, R. (2008) Livestock and Water. North Dakota State University, Fargo, North Dakota. Molden, D. (ed) (2007) Water for food ,Water for life: a comprehensive assessment of water management in agriculture. International water management institute. Earthscan. London, UK. Peylin, P. Bousquet, P. Le Quéré C. Sitch, S. Friedlingstein, P. McKinley, G.A. Gruber, N. Rayner, P.J Ciais, P (2005) Multiple constraints on regional CO2 flux variations over land and oceans. Global Biogeochemical Cycles,19. GB1011, doi: 10.1029/2003GB002214 Pfister, S & Ridoutt, B.G. (2009) A revised approach to water footprinting to make transparent impacts of consumption and production on global freshwater scarcity. Global Environmental Change, 20 (Issue 1), 113-120. Pimentel, D. Housner, J. Preiss, E. White, O. Fang, H. Mesnik, L. Barsky, T. Tariche, S. Schreck, J and Alpert, S (1997) Water resources: Agriculture, the environment, and society. BioScience 47 (2), 97106. Pimentel, D. Berger, B. Filiberto, D. Newton, M. Wolfe, B. Karabinakis, E. Clark, S. Poon, E. Abbett, E. Nandagopal, S (2004) Water resources, agriculture and the environment. College of Agriculture and Life Sciences. Cornell University, Ithaca. New York, USA. Prentice, I C. Heimann, M. Sitch, S. (2000) The carbon balance of the terrestrial biosphere: ecosystem models and atmospheric observations. Ecological Applications 10, 1553-1573. Rekacewicz, P (2006) Increased global water stress. Maps and Graphics Library, UNEP/GRID Arendal . Available at: http://maps.grida.no/go/graphic/increased-global-water-stress Rockström, J. Lannerstad, M. Falkenmark, M. (2007) Assessing the water challenge of a new green revolution in developing countries. In Proceedings of the National Academy of Sciences of the United States of America, 104 (No. 15), 6253-6260. Rost, S. Gerten, D. Bondeau, A. Lucht, W. Rohwer , J. and Schaphoff, S. (2008) Agricultural green and blue water consumption and its influence on the global water system. Water Resources Research 44, p. W09405. Sitch, S. Smith, B. Prentice, C. Arneth, A. Bondeau, A. Cramer, W. Kaplan, J.O. Levis, S. Lucht, W. Sykes, M.T. Thonike, K. Venevsky, S. (2003) Evaluation of ecosystem dynamics, plant geography and terrestrial carbon cycling in the LPJ dynamic global vegetation model. Global Change Biology 9, 161– 185. Steinfeld, H. Gerber, P. Wassenaar, T. Castel, V. Rosales, M. de Haan, C. (2006) Livestock´s long shadow. Environmental issues and options. FAO. Rome, Italy.

48

United Nations Statistics Division, UN STAT (2010) Millenium development goal indicators. United Nations. Accessed: 19-03-2010. Available at: http://mdgs.un.org/unsd/mdg/Default.aspx Wagner, W. Scipal, K. Pathe, C. Gerten, D. Lucht, W. Rudolf, B (2003) Evaluation of the agreement between the first global remotely sensed soil moisture data with model and precipitation data. Journal of Geophysical Research, 108, 4611, doi: 10.1029/2003JD003663 WFN (Water Footprint Network), 2010. Glossary, Water footprint, Water productivity, Virtual water content. Accessed: 21-05-2010. Available at: http://www.waterfootprint.org/?page=files/Glossary Wirsenius, S (2000) Human use of land and organic materials modeling the turnover of biomass in the global food system. PhD Thesis. Department of physical resource theory, Chalmers University of Technology and University of Gothenburg. Gothenburg, Sweden. Wirsenius, S. Azar, C. Berndes, G (2010) Preserving natural ecosystems and global biodiversity: How much land can be spared for nature by dietary changes and livestock productivity increases? Department of energy and environment, Chalmers University of Technology. Gothenburg, Sweden. Wirsenius, S (2010). Researcher, Dep. of Energy and Environment, Chalmers University of Thechnology, Gothenburg, Sweden, personal communication. Zaele, S. Sitch, S. Smith, B. Hatterman, F. (2005) Effects of parameter uncertainties on the modeling of the terrestrial biosphere dynamics. Global Biogeochemical Cycles, 19, GB3020, doi: 10.1029/2004GB002395 Zimmer, D. Renault, D (2003) Virtual water in food production and global trade: review of methodological issues and preliminary results. In: Virtual water trade: Proceedings of the international expert meeting on virtual water trade (2003) Hoekstra, A.Y. (ed). Value of water research report series No 12. Institute for Water Education, UNESCO-IHE. Delft, Netherlands. Available at: http://www.waterfootprint.org/Reports/Report12.pdf

49

Appendix A. Feed compositions Feed compositions have been calculated through the FPD model regionally for each animal system and have been reclassified into 13 CFTs applicable to the LPJml model. Below the different feed compositions are presented (Figure 16). The composition of feed varies greatly between ruminants in grazing animal systems (cattle milk and dairy cows, replacement heifers, beef carcass, buffalo milk and dairy cows, replacement heifers, buffalo carcass, sheep and goat) and monogastric animals in non-grazing systems (pig carcass, meattype chicken and egg producing chicken). Feed composition of ruminants is dominated by CFTs consisting of grasslike plants and here lumped together as Pasture & Grass. For all ruminants grazing is greatest in Latin America constituting about 90 % of feed composition in relation to Europe, North America & Oceania where grazing for milk cattle is approximately 40 % and for other ruminants around 70 %. In Europe, North America & Oceania and East Asia the grazing in general seems to be a little less for most of the animal systems with a few exceptions (e.g. buffalo replacement heifers where approximately 98% of the feed composition is of permanent & cropland pasture). Other than pasture ruminants feed composition mainly constitutes of maize (between 10-30 %), temperate cereals (in an average of about 10%), rice (in warmer climate-regions of about 10-20%), soybean and sunflower (differing widely between 1-15 %) with magnitudes depending on animal type as well as region. Maize seems to be the largest section in most animal systems and regions other than pasture followed by temperate cereals. Monogastrics have a feed composition of more varying kind. Cereals are a significant section of the composition. In a majority of regions of temperate sort which for example in North America and Europe composes up to 40% of the feed composition. Tropical cereals are fed to all animals in tropical regions in amounts up to 10 % of the total feed composition. Animals also feed largely upon maize throughout all regions of varying amounts differing from approximately 10 % in East Asia to 50 % in North America & Oceania. Soybeans are also fed, although in small amounts, to all monogastric animals but constitute an important part of the feed composition of both chicken types. In East & West Asia, Africa and Europe the feed composition of pigs greatly constitutes of non eaten food which is basically domestic waste used as feed. This sector is not significant for any other livestock system or in any other regions. East Asia also deviates in the pig livestock sector with a severe part of feed consisting of the category “other”. This is due to a large amount of forage-vegetables fed to pigs in that particular region.

50

51

52

53

Figure 16: Feed composition of different animal systems in selected regions

54

Appendix B. Feed requirement Feed requirement for all animal systems and subsystems have been calculated for the production of meat, milk and egg on a regional level presented in table 6. Feed requirement of ruminants such as the animal systems of cattle, buffalo, sheep and goat tend to be significantly higher than for monogastric livestock systems where feed and fodder are of different type. There also tend to be higher feed demand for the animal systems in the developing world than the developed regions. Differences between the feed requirements of systems within regions are also of much larger magnitude in the less developed regions. Requirement for grazing animals can be more than five times greater in e.g. South and Central Asia and Sub-Saharan Africa compared to Europe, North America & Oceania. Even differences in demand of non-grazing animal systems between regions are of much less significant nature in the developed world. The evident exception which applies for all selected regions is that of the cattle milk and buffalo milk systems. Dairy cows have different feed requirement for producing milk and the meat where feed eaten in their early life stadiums is invested in the animal body that at slaughter will contribute to the meat production. The feed requirement for meat production is evidently higher than that of milk production. The feed demand for meat production of the dairy cattle system will however throughout all system and regions be lower than that of beef cattle. This is due to that the meat produced is not the primary production the invested feed is aimed for. The quantity of meat is not of great significance however, less feed resources will be diverted to this type of production compared to beef production. The feed requirement of beef animals tend to be at least twice as big as the requirement for bulls and replacement heifers that are the parts of the milk producing systems that produces meat. The egg producing animal system follow the same pattern of different feed requirements for different products where the production of meat has a slightly higher requirement than egg production. However the difference is much less and the requirement of the egg-producing system is remarkably lower than for all other animals except the total requirement of milk cattle calculated for their total production. Milk cattle consistently have the lowest feed demand which depend on their large production in terms of dry mass and their long lifetime which further contributes to large production numbers per animal. The differences of poultry livestock production is generally small within as well as outside regions compared to all other livestock production.

55

Table 6: Feed requirement factors calculated for all animal system in selected regions interpolated for the year of 2005 by input data from 1993 and 2030 (tonnes of feed/tonnes of product). Animal system and region

Feed requirement factor total production

Feed requirement factor meat production

Feed requirement factor milk/egg production

East Asia 2,7

Cattle milk & dairy cow carcass Dairy cattle bulls & heifer carcass

43,8

Beef cattle carcass

88,9

Buffalo milk & dairy cow carcass

7,4

Dairy buffalo bulls & heifer carcass

43,8

2,1

79,7

5,5

79,7

Beef buffalo carcass

183,3

Sheep carcass

72,4

Goat carcass

71,7

Pig carcass

6,6

Meat-type chicken carcass

4,1

Chicken eggs

2,9

4,1

2,8

1,6

10,2

1,5

16,9

3,8

Europé Cattle milk & dairy cow carcass Dairy cattle bulls & heifer carcass

10,2

Beef cattle carcass

32,2

Buffalo milk & dairy cow carcass

4,1

Dairy buffalo bulls & heifer carcass

16,9

Beef buffalo carcass

38,9

Sheep carcass

48,0

Goat carcass

72,5

Pig carcass

4,9

Meat-type chicken carcass

3,2

Chicken eggs

2,6

3,2

2,5

3,6

42,1

3,1

21,2

4,6

Latin America & Caribbean Cattle milk & dairy cow carcass Dairy cattle bulls & heifer carcass

42,1

Beef cattle carcass

77,9

Buffalo milk & dairy cow carcass

5,0

Dairy buffalo bulls & heifer carcass

21,2

Beef buffalo carcass

23,4

Sheep carcass

123,8

Goat carcass

124,4

Pig carcass

9,2

Meat-type chicken carcass

3,4

Chicken eggs

2,8

3,4

2,7

3,3

25,1

3,0

16,3

3,7

North Africa & West Asia Cattle milk & dairy cow carcass Dairy cattle bulls & heifer carcass

25,1

Beef cattle carcass

83,7

Buffalo milk & dairy cow carcass

4,0

56

Dairy buffalo bulls & heifer carcass

16,3

Beef buffalo carcass

55,6

Sheep carcass

60,5

Goat carcass

89,3

Pig carcass

6,4

Meat-type chicken carcass

4,1

Chicken eggs

3,1

4,1

3,0

1,2

11,1

1,0

16,9

4,0

North America & Oceania Cattle milk & dairy cow carcass Dairy cattle bulls & heifer carcass

11,0

Beef cattle carcass

25,8

Buffalo milk & dairy cow carcass

4,3

Dairy buffalo bulls & heifer carcass

16,9

Beef buffalo carcass

59,8

Sheep carcass

59,0

Goat carcass

68,9

Pig carcass

4,3

Meat-type chicken carcass

2,9

Chicken eggs

2,3

2,9

2,3

4,1

99,7

2,7

78,1

2,6

South & Central Asia Cattle milk & dairy cow carcass Dairy cattle bulls & heifer carcass

99,7

Beef cattle carcass

210,1

Buffalo milk & dairy cow carcass

3,6

Dairy buffalo bulls & heifer carcass

78,1

Beef buffalo carcass

163,0

Sheep carcass

77,2

Goat carcass

91,6

Pig carcass

8,9

Meat-type chicken carcass

4,7

Chicken eggs

3,5

4,7

3,4

8,4

56,8

6,9

19,9

4,6

4,9

4,2

Sub-Saharan Africa Cattle milk & dairy cow carcass Dairy cattle bulls & heifer carcass

56,8

Beef cattle carcass

121,0

Buffalo milk & dairy cow carcass

5,0

Dairy buffalo bulls & heifer carcass

19,9

Beef buffalo carcass

72,0

Sheep carcass

109,4

Goat carcass

101,3

Pig carcass

10,8

Meat-type chicken carcass

4,9

Chicken eggs

4,3

57

Appendix C. Feed use Calculations have been performed to be able to illustrate how the feed use among the different animal systems varies in terms of amount of each CFT they eat and how that use varies globally. Calculations are illustrated in figure 17. Feed use for the different ruminants are of similar composition which also applies for all monogastric animals. Feed quantities illustrates the results presented above under feed composition in terms of megatonnes of dry mass. Ruminants demand large quantities of different types of grazing which constitutes most of their feed use. Numbers differ between regions and are of largest nature for milk and beef cattle where beef cattle uses feed in the size of 1000 MT of DM in e.g. Latin America in comparison with sheep and goat which uses about 30 MT of DM respectively 10 MT of DM in the same region. Amount of feed used for goat and sheep is dominated by East Asia where the production demand approximately 90 MT of DM only for grass and pasture. Additional, sheep and goat uses temperate and tropical cereals, maize and rice (primarily in Asia) of varying amounts all in the scale of 10 MT of DM. Cattle seem to eat the same groups of CFTs but for cattle milk, beef and buffalo beef in amounts of about 50-100 MT of DM instead. Feed use of chickens varies around 10-20 MT of DM for all components where only maize deviates for all regions, in some such as East Asia with up to 80 MT of DM for egg producing chickens and 50 MT of DM for meat producing chickens. Other than maize which is the single dominating CFT of feed for chickens both temperate and tropical cereals, sunflower and soybean occur in mentioned magnitude. Quantities of feed used for livestock production of pig varies greatly between selected regions. Largest amount of feed is dedicated in East Asia and Europe. The regions where largest amounts of feed is used feed non eaten food to a great extent in Europe about 50 MT of DM and in East Asia up to 70 MT. In East Asia the greatest feed used for livestock production of pig is forage vegetables and this generates that the CFT “other” in the region have a magnitute of 150 MT of DM. Furtheremore temperate cereals in both regions are used in amounts of 70 MT, the usage of maize is at an average of between 10-20 MT of DM. Roots, sunflower, soybean etc. are fed in amounts about 10 MT of DM in all regions.

58

59

60

61

Figure 17: Feed use sorted in CFTs of animal systems in selected regions.

62