The Future of Food

Master of Science Thesis

The Future of Food Scenarios and the Effects on Resource Use in Agriculture

Ingrid Ym Ruth Odegard, BSc.

Institute of Environmental Sciences, Leiden University June 2011

The Future of Food Food Scenarios and the Effects on Resource Use in Agriculture Master of Science Thesis

For obtaining the degree of Master of Science in Industrial Ecology from Leiden University and Delft University of Technology

Ingrid Ym Ruth Odegard, BSc.

June 2011

Institute of Environmental Sciences - Leiden University Dr. E. van der Voet Drs. R. Huele Center for Industrial Ecology – Yale University Prof. T.E. Graedel M.Sc. Thesis I.Y.R. Odegard

2

© I.Y.R. Odegard, B.Sc. All rights reserved.

M.Sc. Thesis I.Y.R. Odegard

3

Preface This thesis presents the results of my graduation research, carried out as part of the Industrial Ecology M.Sc. program at Delft University of Technology and Leiden University. It was performed under guidance of my supervisors dr. Ester van der Voet and drs. Ruben Huele at the Institute of Environmental Sciences in Leiden and was partially conducted at the Center for Industrial Ecology at Yale University in the USA. Even though sustainability has been the focal point of my studies since 2003, when I started studying Sustainable Molecular Science and Technology at the TU Delft and Leiden University, the sustainability of our food seemed an undervalued topic in the sustainability discourse. The past years attention has been increasing, and today the ‘sustainability of food’ is a common topic on conferences, in newspapers and in the supermarket. For me, it has proven to be a deserving topic; people have always been interested in my findings. I would like to thank my first supervisor, dr. Ester van der Voet, for her support, patience and confidence. Your advice and comments were always helpful and you steered me in the right direction to make me produce the work I wanted to deliver. I would also like to thank my second supervisor, drs. Ruben Huele, for his helpful ideas and pushing me to see the broader picture. Finally, I would like to thank prof. Graedel for welcoming me at the Center for Industrial Ecology and for taking the time to meet with me and discuss my work. It had always been a dream of mine to live and study in the USA for a while, and I would like to thank my supervisors dr. Van der Voet, and prof. Graedel very much for giving me that opportunity. Without my parents I would not have been where I am now. Thank you for your love and support and for always believing that I would be able to achieve what I wanted to. I would also like to thank everyone who read part of my thesis and gave me feedback: papa Dave, Andrew, Renske, Sebastiaan, Julia and Martijn. Your feedback was very useful and I enjoyed sharing my work with you. Furthermore, I would like to thank everyone who has shown an interest in my study and were curious as to my final results; it was always motivating to talk about my study and find that you really wanted to know whether you should reduce your meat consumption or become vegetarians. Finally I would like to urge everyone who reads this to reconsider their consumption pattern, and thank you if you do. I hope I can make a contribution to making our food system more sustainable, and that we are all healthy and well-fed in 2050.

Ingrid Ym Ruth Odegard Delft, June 2011 M.Sc. Thesis I.Y.R. Odegard

4

Summary The aim of this study was to design four global food scenarios for the year 2050 and to evaluate these quantitatively with respect to their use of the natural resources land, water and fertilizers in agriculture. Resource use and food are popular topics in the current sustainability debate, because of population growth and the change seen in diet composition, related to increased welfare levels, with increased demand for animal products in developing countries. To evaluate resource use for a complete diet in different food scenarios, this study integrates three sub-studies: 1) Four scenarios were designed, which include different trends related to population, economic development, policy, technological development and diet and are specified to the year 2050. 2) A methodology was developed – Virtual Resource Content – with which the use of resources in agriculture can be calculated. 3) A model was created, with which the scenarios are quantified with respect to their resource use. The scenarios are all evaluated on a global scale, including two that are also evaluated on a regional scale. These regions are the OECD90 region (countries in the OECD in 1990), the REF region (the countries under reform such as the former Soviet Union), the ASIA region (Asia) and the ALM region (Africa, Latin America and the Middle East). The Virtual Resource Content comprises of factors for Virtual Land Content (ha kg-1), Virtual Water Content (m3 kg-1) and Virtual Fertilizer Content (for N, P2O5 and K2O, kg kg-1), on a regional and global scale. 

 



In the Affluent World (A1 – global) high, globally dispersed economic and technological development are coupled to low population growth leading to high apparent consumption of animal products. In the Full World (A2 – regional) high population growth is coupled to low regional economic and technological development, leading to lower increases in meat consumption. In the Vegetarian World (B1 – global) consumption of animal products is limited to milk and eggs. Low population growth is coupled to medium-high global economic growth and high technological development, and environmental issues are considered important. In the Low-Input World (B2 – regional) awareness concerning environmental issues results in low fertilizer inputs and is coupled to medium population growth and medium economic growth.

The results show that it is impossible to continue current trends related to meat consumption at a global level. The Affluent World shows that it may be attainable to contain land use within feasible limits, but that such a development comes at a cost: high fertilizer use. Due to intensive management of pasture areas, fertilizer use rises to a level where, in the year 2050, only 2 years of K2O use and 55 years of P2O5 use remain, based on estimates of the global resource base. Moreover, a situation such as in the Affluent World were fertilizers run out between 2050 and 2100 would never progress to such a stage because fertilizer (and thus meat) prices would have risen too much to justify such use. As shown for the Full World, less intensive management of pastures and such can cut fertilizer use; in the Full World 66 years of K2O use and 121 years of P2O5 use remain, but pasture area almost doubles. The Vegetarian World shows picture of potential natural resource use that is manageable. Water and land use stay within acceptable limits, and fertilizer reserves are large enough to continue those practices for another 100 to M.Sc. Thesis I.Y.R. Odegard

5

254 years, for respectively K2O and P2O5. It is, however, unlikely that people are willing to give up meat consumption. While the Low-Input World incorporates lower meat consumption, the low inputs of fertilizer pushes land use beyond all possible limits, in all regions. Total production increases with 40%, which is related to an increased water use of 63% (compared to 2005). On a global level this is within the moderate water stress limit, however, the ASIA region exceeds the moderate water stress level by 47%. Figure 1 shows the results for the four scenarios, compared to the situation in the year 2005.

Figure 1: Results - resource use in 2050 in the four scenarios

The results show that trade-off issues are important and need to be addressed when discussing the future of food. An assessment of resource use is only valuable when a complete picture is given. Therefore, the present study provides valuable input for assessing problem areas, but also for identifying opportunities in our agricultural system. Keywords: Food, Agriculture, Virtual Resource Content, Natural Resource Use, Water, Land, Fertilizers, Scenarios, 2050


6

Contents Preface........................................................................................................................................................... 4 Summary ....................................................................................................................................................... 5 List of Figures................................................................................................................................................. 9 List of Tables ................................................................................................................................................ 11 List of Abbreviations .................................................................................................................................... 13 1.

Introduction ......................................................................................................................................... 14 1.1 Research Goal and Research Question ....................................................................................... 16 1.2 Scientific Relevance and Contribution of Present Research ....................................................... 18 1.3 Guide to the Reader .................................................................................................................... 19

2.

Methodology ....................................................................................................................................... 20 2.1 Scope Definition .......................................................................................................................... 20 2.2 Scenario Methodology ................................................................................................................ 24 2.3 Virtual Resource Content ............................................................................................................ 28 2.4 The VRC Model ............................................................................................................................ 29

3.

Current Situation ................................................................................................................................. 31 3.1 Food Consumption ...................................................................................................................... 32 3.2 Food Production .......................................................................................................................... 38 3.3 Appropriation of Natural Resources............................................................................................ 41 3.4 Productivity ................................................................................................................................. 57 3.5 Supply and Demand .................................................................................................................... 58 3.6 Losses and Wastes ....................................................................................................................... 59 3.7 Driving Forces .............................................................................................................................. 62

4.

Driving Forces and Trends ................................................................................................................... 63 4.1 Population ................................................................................................................................... 64 4.2 Economic Development .............................................................................................................. 68 4.3 Policy ........................................................................................................................................... 71 4.4 Technological Change .................................................................................................................. 73 4.5 Diet Change ................................................................................................................................. 76

5.

The Future of Food Storylines ............................................................................................................. 82 5.1 Linkages between Driving Forces ................................................................................................ 82 5.2 A1 – The Affluent World .............................................................................................................. 85 5.3 A2 – The Full World ..................................................................................................................... 86 5.4 B1 – The Vegetarian World ......................................................................................................... 87 5.5 B2 – The Low Input World ........................................................................................................... 88


7

6.

Linkages ............................................................................................................................................... 89 6.1 Demand ....................................................................................................................................... 89 6.2 Supply .......................................................................................................................................... 92 6.3 VRC Factors.................................................................................................................................. 99 6.4 Modeling Protocol ..................................................................................................................... 102

7.

Results .............................................................................................................................................. 103 7.1 Production and Consumption ................................................................................................... 103 7.2 Land Use .................................................................................................................................... 107 7.3 Water Use .................................................................................................................................. 111 7.4 Fertilizer Use .............................................................................................................................. 113

8.

Conclusions ....................................................................................................................................... 117 8.1 Scenario Conclusions ................................................................................................................. 117 8.2 Resource Use in Agriculture ...................................................................................................... 120

9.

Discussion .......................................................................................................................................... 124

10.

Recommendations......................................................................................................................... 130

Bibliography............................................................................................................................................... 133 Appendix 1 Appendix 2 Appendix 3 Appendix 4 Appendix 5 Appendix 6 Appendix 7 Appendix 8 Appendix 9 Appendix 10 Appendix 11 Appendix 12 Appendix 13 Appendix 14 Appendix 15 Appendix 16 Appendix 17


World Regions ............................................................................................................... 138 Commodity Group Assumptions ................................................................................... 140 Waste ............................................................................................................................ 142 Feed ............................................................................................................................... 143 Seed ............................................................................................................................... 144 Other Utilitization ......................................................................................................... 145 Yield Projections ............................................................................................................ 146 Feed-mixes and Feeding Efficiency ............................................................................... 151 Production of Oil and Sugar Crops ................................................................................ 156 Land Potential................................................................................................................ 157 Virtual Water Content ................................................................................................... 158 Water Potential ............................................................................................................. 160 Meat consumption ........................................................................................................ 161 Virtual Fertilizer Content ............................................................................................... 163 Yield Projection Calculations OECD90 Region ............................................................... 166 FAO Productivity Projections ......................................................................................... 168 Results – Data ................................................................................................................ 169

8

List of Figures Figure 1: Results - resource use in 2050 in the four scenarios...................................................................... 6 Figure 2: Food production, natural resource use and associated environmental issues............................ 15 Figure 3: IPCC SRES scenario orientation .................................................................................................... 17 Figure 4: Schematic representation of the four IPCC SRES Scenarios. ........................................................ 27 Figure 5: Relationships between driving forces, supply, demand, VRC and resource use. ........................ 30 Figure 6: Apparent consumption (kcal/cap/day) in countries with different levels of development. ....... 32 Figure 7: Global average apparent food consumption (kg/cap/year)......................................................... 33 Figure 8: Regional apparent consumption (kg/cap/year). .......................................................................... 35 Figure 9: Meat consumption (kg/capita/year) for all 185 countries for the year 2004 .............................. 36 Figure 10: Regional land use per commodity group (in ha) in 2005 ........................................................... 42 Figure 11: Multiple cropping zones, rain-fed conditions ............................................................................ 43 Figure 12: Share of total water use of the global total by region ............................................................... 46 Figure 13: Fraction of global water use in agriculture per commodity group ............................................ 47 Figure 14: Share of irrigated land in arable land and permanent crops ..................................................... 48 Figure 15: Consumption of nitrogenous fertilizers, 1950-1990 .................................................................. 49 Figure 16: Regional average nitrogenous fertilizer consumption (kg/ha/year) in 2005 ............................. 51 Figure 17: Fertilizer use estimates (ton/year). ............................................................................................ 52 Figure 18: Recommended and estimated regional fertilizer use. ............................................................... 53 Figure 19: Commodity balance calculation method flowscheme. .............................................................. 59 Figure 20: World population projections until 2050 ................................................................................... 65 Figure 21: OECD90 region population projection until 2050 ...................................................................... 66 Figure 22: REF region population projections until 2050 ............................................................................ 67 Figure 23: ASIA region population projections until 2050 .......................................................................... 67 Figure 24: ALM region population projections until 2050 .......................................................................... 68 Figure 25: World PPP Projections until 2050 .............................................................................................. 70 Figure 26: Regional GDP projections (PPP). ................................................................................................ 71 Figure 27: Cereal yield projections .............................................................................................................. 76 Figure 28: Population and global food production indices, 1966-1998 ...................................................... 77 Figure 29: Meat consumption trends in different scenarios....................................................................... 80 Figure 30: Scenario Characteristics ............................................................................................................. 84 Figure 31: Population between 19050-2050 (in billions) in A1 ................................................................... 85 Figure 32: Population between 1950-2050 (in billions) in A2 ..................................................................... 86 Figure 33: Population between 1950-2050 (in billions) in B1 ..................................................................... 87 Figure 34: Population between 1950-2050 (in billions) in B2 ..................................................................... 88 Figure 35: Linkages between driving forces, supply, demand, VRC and resource use. .............................. 89 Figure 36: Example of Virtual Land Content.............................................................................................. 100 Figure 37: Example of Virtual Water Content. .......................................................................................... 101 Figure 38: Example of Virtual Fertilizer Content ....................................................................................... 101 Figure 39: Apparent consumption and intake in the four scenarios in 2050 ............................................ 103 Figure 40: Production in 2005 and 2050 ................................................................................................... 104 M.Sc. Thesis I.Y.R. Odegard

9

Figure 41: Total production, production of food and feed, apparent consumption and intake in 2005.. 105 Figure 42: Comparison of total production, production for food and feed, apparent consumption and intake (kg/cap/year) for the four scenarios .............................................................................................. 106 Figure 43: Arable land and permanent crops, harvested area in 2005 and in the four scenarios............ 108 Figure 44: Total land use in A2 and B2 ...................................................................................................... 108 Figure 45: Total land use in 2005 and for the four scenarios in 2050 ....................................................... 109 Figure 46: Total land use per capita in 2005 and for the four scenarios in 2050...................................... 110 Figure 47: Total water use in 2005 and 2050 ............................................................................................ 111 Figure 48: Regional water use in A2 and B2 .............................................................................................. 112 Figure 49: Nitrogen fertilizer use in 2005 and 2050 .................................................................................. 113 Figure 50: Phosphorous fertilizer use in 2005 and 2050 ........................................................................... 114 Figure 51: Potassium fertilizer use in 2005 and 2050 ............................................................................... 115 Figure 52: Scenario characteristics ............................................................................................................ 117 Figure 53: Scenario results - land use, water use and fertilizer use in 2050 for the four scenarios. ........ 121 Figure 54: Processing scheme oil crops ..................................................................................................... 156 Figure 55: Processing scheme sugar crops ................................................................................................ 156 Figure 56: Engel curve for meat consumption .......................................................................................... 161


10

List of Tables Table 1: Research characteristics ................................................................................................................ 18 Table 2: Data: source and specifications ..................................................................................................... 21 Table 3: Table of boundary conditions ........................................................................................................ 22 Table 4: Table of assumptions ..................................................................................................................... 23 Table 5: Scenario Methodology................................................................................................................... 25 Table 6: Regional population size and level of economic development (in PPP) in the year 2005 ............ 31 Table 7: Regional food supply (apparent consumption in kg/cap/year) in 2005 ........................................ 33 Table 8: Regional and global production of vegetable commodity groups ................................................ 38 Table 9: Regional and global production of animal commodity groups. .................................................... 39 Table 10: Import, export and net import for the four regions in the year 2005 ......................................... 40 Table 11: Area under cultivation for the years 1995, 2000 and 2005......................................................... 41 Table 12: Gross extents of land with rain-fed cultivation potential............................................................ 44 Table 13: Water use in agriculture and fraction of total water use in 2005 ............................................... 45 Table 14: Crop evapotranspiration for the four regions ............................................................................. 46 Table 15: Irrigated area ............................................................................................................................... 48 Table 16: Fertilizer production and consumption ....................................................................................... 50 Table 17: Efficiencies of animal food production, land requirements and water requirements ................ 56 Table 18: Yields for the seven commodity groups in the four regions for the year 2005 .......................... 57 Table 19: Supply and demand ..................................................................................................................... 58 Table 20: Uses other than food, % of total global production in the year 2005 ......................................... 60 Table 21: Household wastes as percentage and fraction of (edible) food supply. ..................................... 61 Table 22: Population projections ................................................................................................................ 65 Table 23: Final populations 2050 (in billions).............................................................................................. 66 Table 24: Economic Development Projections ........................................................................................... 69 Table 25: GDP Projections (in PPP) per capita for the year 2050 ................................................................ 70 Table 26: Policy Characteristics ................................................................................................................... 72 Table 27: Technological development trends ............................................................................................. 75 Table 28: Changes in the commodity composition of food by major country groups ................................ 78 Table 29: Diet Trends. ................................................................................................................................. 79 Table 30: Scenario driving forces' linkages and characteristics. ................................................................. 82 Table 31: Economic Growth Rates and Income per Capita in the A1 World .............................................. 85 Table 32: Economic Growth Rates and Income per Capita in the A2 World .............................................. 86 Table 33: Economic Growth Rates and Income per Capita in the B1 World............................................... 87 Table 34: Economic Growth Rates and Income per Capita in the B2 World............................................... 88 Table 35: Demand-side assumptions and rationale. ................................................................................... 90 Table 36: Supply-side assumptions and rationale ....................................................................................... 93 Table 37: VRC factors and their specifications. ........................................................................................... 99 Table 38: Increase in total production compared to increases in fertilizer use. ....................................... 114 M.Sc. Thesis I.Y.R. Odegard

11

Table 39: Fertilizer use and reserve base. ................................................................................................. 116 Table 40: Definition of the regions. ........................................................................................................... 138 Table 41: Commodity group assumptions................................................................................................. 140 Table 42: Items included in the commodity groups. ................................................................................. 141 Table 43: Waste (during transport, storage and processing) in 2007. ...................................................... 142 Table 44: Feed in 2007 .............................................................................................................................. 143 Table 45: Seed in 2007 .............................................................................................................................. 144 Table 46: Other utilization in 2007. ........................................................................................................... 145 Table 47: Cereal yield projection............................................................................................................... 146 Table 48: Top three yielding countries per region for fruits and vegetables ............................................ 148 Table 49: Maximum attainable yields (MAYs) and yields in 2005 and in 2050 ......................................... 148 Table 50: Feedcrop yields in 2005 and 2050 ............................................................................................. 149 Table 51: Choices for correspondence of Wirsenius' regions to IPCC regions. ......................................... 151 Table 52: Feed-mixes for the five animal products per region ................................................................. 152 Table 53: Availability of edible-type crops by-products in terms ............................................................. 155 Table 54: Division of the ‘global agro-ecological zones-study’ regions. ................................................... 157 Table 55: Global and regional gross and net extents of cultivable land ................................................... 157 Table 56: Virtual Water Content per commodity group (m3 per ton) for the four regions ...................... 159 Table 57: Global and regional renewable water resources and water stress thresholds ......................... 160 Table 58: Countries for which data was used from AQUASTAT. ............................................................... 160 Table 59: Countries for which data on renewable water resources is lacking.......................................... 160 Table 60: Parameters of the Engel curve .................................................................................................. 161 Table 61: Number of countries in income regimes - related to meat consumption - in 2050 ................. 162 Table 62: Global and regional meat consumption .................................................................................... 162 Table 63: Basis of estimates for fertilizer requirements. .......................................................................... 164 Table 64: Virtual Fertilizer Content ........................................................................................................... 165 Table 65: Future change in productivity equation .................................................................................... 166 Table 66: Parameters that represent the effect of technology on potential yield and yield gap ............. 167 Table 67: Estimated relative changes in crop productivity due to technology development .................. 167 Table 68: FAO productivity projections ..................................................................................................... 168 Table 69: Total production, losses, feed, apparent consumption and household and retail waste. ........ 169 Table 70: Total land use per scenario and total cropland use in the A2 and B2 scenarios ....................... 171 Table 71: Total water use in the scenarios and water use in the A2 and B2 scenarios. ........................... 172 Table 72: Total fertilizer use (N, P2O5 and K2O) in the four scenarios. ...................................................... 173


12

List of Abbreviations ALM ASIA AQUASTAT EF FAO FAOSTAT FBS FERTISTAT IFPRI IPCC SRES OECD OECD90 REF UN UNDP VRC WF WWF

IPCC SRES region - Africa, Latin America and Middle East region, see Appendix 1 IPCC SRES region - Asia excluding Japan, see Appendix 1 Database – FAO’s global information system on water and agriculture, developed by the Land and Water Division of the FAO. Ecological Footprint Food and Agriculture Organization of the United Nations Database – provides time-series and cross sectional data relating to food and agriculture for some 200 countries Food Balance Sheet Database – Fertilizer Use Statistics, Plant Production and Protection Division of the FAO International Food Policy Research Institute Intergovernmental Panel on Climate Change Special Report on Emissions Scenarios Organization for Economic Co-operation and Development IPCC SRES region – the countries in the OECD in 1990, see Appendix 1 IPCC SRES region – the countries under reform, see Appendix 1 United Nations United Nations Development Programme Virtual Resource Content Water Footprint World Wildlife Fund


13

1.

Introduction

Food fulfills many different purposes. We need it every day to sustain ourselves, we dine together and socialize, and it may be a comfort at times. Furthermore, it helps feed the animal we eat, and we use it to produce non-food products such as biofuel. To be able to eat strawberries in the winter, and fresh pineapple in the Netherlands, we transport food all over the world. Our relationship to food is one of deprivation on the one hand, and of waste on the other. According to various researchers, we produce enough to sustain the current global population [Rosegrant, 2001a, Penning de Vries, 1995], but approximately 1 billion people are deprived of basic nutrition [UN, 2010] while many people in the Western world, despite high caloric intakes, suffer from malnutrition. Three natural resources are of key importance in our agricultural system: land, water and fertilizers. The impact of agriculture on our land use, water use and fertilizers use is significant; agriculture is the major consumer of these resources. According to the Food and Agriculture Organization of the United Nations, the FAO, “Over the next 30 years, many of the environmental problems associated with agriculture will remain serious” [FAO, 2002, p. 7]. It is important to realize that there is no one fixed course in which the food-system will develop. Because agriculture is a human activity, we can influence the course of such development. This study will explore the factors related to agriculture and natural resource use, and will quantify the effects of four different scenarios related to the food system. Impacts of Natural Resource Use in Agriculture The impacts related to the use of natural resources in agriculture are large. As stated by Wirsenius: ‘The food and agriculture system is among the largest anthropogenic activities in terms of appropriation of land and biological primary production, as well as alteration of the grand chemical cycles of carbon, water, and nitrogen’ [Wirsenius, 2003]. Globally, agriculture – croplands and pastures – accounts for 40% of total land use [Lotze-Campen, et al. 2006; Foley et al., 2005] and 85% of the freshwater we use is used in agriculture [WWF, 2006; Foley et al., 2005]. In different parts of the world the area under cultivation is either decreasing (Europe) or expanding (South America, Asia). Fossil fuels are used extensively in the food production chain, for transport purposes but also for the production of agricultural chemicals. ‘Industrial fixation of N fertilizer increased from 100%), or whether the area equipped for irrigation is being underused (cropping intensity < 100%), or e.g. was unnecessarily equipped for irrigation. The data as reported by AQUASTAT cover respectively 92.7% and 94.9% of the irrigated area in the ASIA region and the ALM region in the year 2000 (see also Table 15), and is therefore assumed to give a fair representation. According to the AQUASTAT data, the cropping intensity in the ALM region is slightly lower than 100%: 94.9%. In the ASIA region it is well above 100%: 138.3%.

Figure 11: Multiple cropping zones, rain-fed conditions [Fischer, 2002]

Land available for further expansion of agriculture seems to be limited. According to the FAO, land suitable for rainfed agriculture, with yields above a minimum acceptable level, amount to 2.8 billion hectares in the developing world [Bruinsma, 2003, p.14]. Around 34% of this land is currently under cultivation. Most of the remaining land is concentrated in a few countries in South America and SubSaharan Africa, and is ‘very unevenly distributed’, and can therefore not be considered a ‘land reserve’ M.Sc. Thesis I.Y.R. Odegard

43

[Bruinsma, 2003, p.15]. Land availability is even more limited in parts of the REF region (Near East and North Africa) and the ASIA region (South Asia). Not only do many countries in South Asia and the Near East/North Africa have no spare land left, the land that is left is not suitable for agriculture due to environmental constraints. Furthermore, ‘a good part of the land with agricultural potential is under forest or in protected areas, in use for human settlements, or suffers from lack of infrastructure and the incidence of disease. Therefore, it should not be considered as being a reserve, readily available for agricultural expansion’ [Bruinsma, 2003, p.15]. Assuming the FAO data on agricultural land are correct, see Table 11, there is a substantial portion of agricultural land unaccounted for in the OECD90 region and the REF region, respectively 153 million hectares and 104 million hectares. The data in Table 11 show that agricultural land area has been decreasing in these regions. This could be due to various processes, e.g. urbanization or reverting land to nature. It could also be that land is left fallow, because overproduction, decreasing populations and increasing yields reduce the need for agricultural land area to remain the same. Table 12: Gross extents of land with rain-fed cultivation potential, in 1000 ha and % of total land area [Fischer, 2002]

Region

Gross extents with rain-fed cultivation potential VS + S + MS landa (1000 ha) (arable land and permanent crops as % potential)

OECD90 REF ASIA ALM

707,000 380,800 590,400 1,973,500

54.3% 64.3% 80.3% 22.9%

a

Note: VS, S and MS denote ‘very suitable’, ‘suitable’ and ‘moderately suitable’, see also Appendix 10.

Table 12 above shows the gross extent of land with rain-fed cultivation potential, according to the IIASA and the FAO [Fisher, 2002]. It is an aggregation of the land that is very suitable (VS), suitable (S) and moderately suitable (MS), with a ‘maximizing technology mix’ and measures the gross maximum available land under rainfed conditions in the four regions. Net available area is between 10% and 30% lower (see Appendix 10). 3.3.2

Water

Agriculture accounts for 85% of the global water withdrawals [Foley, 2005]. Rainfed agriculture is still predominant on a global scale. 78% of the water consumed by crops comes from green water – rainfall stored in soil moisture. These crops are grown on 71% of the world’s agricultural land, and provide 62% of the gross value of global food production. Irrigation water, provided by surface water and groundwater sources (blue water), is used on the remaining 28% of agricultural land. [De Fraiture, 2010]. Irrigation efficiency is assumed to be 60%, which means that additional water – 1050 km3 – is withdrawn from surface and groundwater sources provide the 1570 km3 needed for crop evapotranspiration [De Fraiture, 2010, p. 503]. The data shown in Table 13 and Table 14 are based on data from the Water Footprint Network [Hoekstra, 2008]. Not all countries are included in these data, but the missing countries account for only 3% of the land dedicated to arable land and permanent crops, and 3% of the global population, in the year 2005. M.Sc. Thesis I.Y.R. Odegard

44

Table 13: Water use in agriculture and fraction of total water use in 2005 [based on data from Hoekstra, 2008].

Region

Water (10 m3 year-1)

Agricultural water use (% of regional totals)

9

Renewable water

Total water use

9,536 5,728 14,773 22,926

1,730 569 3,082 1,802

OECD90 REF ASIA ALM

Water use in agriculture 1,099 481 2,828 1,781

% of regional renewable water 11,5% 8,4% 19,1% 7,8%

% of regional total water use 63,5% 84,4% 91,8% 98,9%

A higher fraction of water use is consumed in industry and in domestic settings in developed countries [Molden, 2007]. As developing countries advance, it can be expected that the water demand in sectors other than agriculture will grow. The regional averages do not provide insight into actual water scarcity. According to the FAO, countries that use more than 40% of their renewable water resources in agriculture are in a critical situation (critical water stress). Countries using more than 20% of their renewable water in agriculture pass the ‘threshold which could be used to indicate impending water scarcity’ (moderate water stress) [Bruinsma, 2003, p.15]. As can be seen in Table 13 the ASIA region as a whole is already close to that threshold. Out of the 20 countries in the ASIA region for which the Water Footprint Network provided data, 4 countries experience moderate water stress and 5 experience critical water stress. For the OECD90 region (23 countries) the same numbers apply; 9 countries in total are in a water stress situation. The REF region (20 countries) has 4 countries with moderate water stress and 3 with critical water stress. The ALM region (78 countries) has the highest percentage of countries with water stress: 11 countries have moderate water stress and 36 have critical water stress. Of the last group, 15 countries (7 in the Middle East and 5 in North Africa and 2 elsewhere) use more than 100% of their renewable sources in agriculture [based on data from Hoekstra, 2008]. Table 14 shows the annual regional consumption (the crop evapotranspiration), which is an aggregation of national consumption rates for that region. Crop evapotranspiration is ‘the combination of two separate processes whereby water is lost on the one hand from the soil surface by evaporation and on the other hand from the crop by transpiration’ [FAO, 1998, Ch. 1, p. 1]. According to Smil, evapotranspiration rates can be used to determine water needs of crops [Smil, 2001, p. 40]. The Water Footprint Network uses the term ‘evapotranspiration’ for the portion of the water footprint due to water use in agriculture. It appears that the estimation by Hoekstra may be somewhat conservative. According to the International Water Management Institute (IWMI), around 7,130 cubic kilometers are consumed annually by crops grown for the production of food [De Fraiture, 2007, p.91]. As stated above, 3% of agricultural land is not included in the Water Footprint Network (WFN) data. However, the estimate by the WFN is a significant 13% lower. The difference is probably due to the fact that in the WFN data, irrigation efficiency is not taken into account, but the water requirements are ‘defined as the total water needed for evapotranspiration, from planting to harvest for a given crop in a specific climate region, when adequate soil water is maintained by rainfall and/or irrigation so that it does not limit plant growth and crop yield’ [Chapagain, 2008]. As no country-specific data are provided by the IWMI, the WFN data will be used.


45

9

3

Table 14: Crop evapotranspiration (10 m per year) for the four regions [based on data from Hoekstra, 2008].

Crop Evapotranspiration (109 m3 per year) (average over the years 1997-2001)

Region

OECD90 REF ASIA ALM Total

3

m % 3 m % 3 m % 3 m % 3 m %

Regional consumption (% of total regional use) 720 66% 420 87% 2,655 94% 1,432 80% 5,237 85%

For export (% of total regional use) 379 33% 59 23% 174 6% 315 20% 927 15%

Total (% of total global use) 1,099 18% 481 8% 2,828 46% 1,781 28% 6,189 100%

When we take a closer look at the share for each region in the total global use, it is clear that the ASIA region is the main user of water for agriculture. This is also true per hectare. As can be seen in Figure 12, the fraction per region of water use (in m3 ha-1) is quite similar to the fractions for the regions of the global total. The main reason why the water use per hectare is much higher in the ASIA region than in the other regions, is that rice production is quite water-intensive, and that the rice production in ASIA is high. According to Hoekstra, on average on a global scale, 2291 m3 of water is needed to produce a ton of rice. Coupled to high production outputs, this yields the high water use in the ASIA region.

Water use by region Share of total water use and water use in m3 per hectare OECD90 REF ASIA ALM

Figure 12: Share of total water use of the global total by region (OECD90 18%, REF 8%, ASIA 46% and ALM 28%), and water use 3 -1 3 -1 3 -1 3 -1 3 -1 (m ha ) by region (OECD90 2767 m ha , REF 1899 m ha , ASIA 6150 m ha , ALM 4152 m ha ) [based on Hoekstra, 2008].

As can be seen in Figure 12 below, different crops have different water needs. This is also true for the same crop in different regions. The diversity in water use per crop (left pie-chart), and per kg of a certain crop (right pie-chart) is shown in Figure 14. As can be seen, different commodity groups account for quite different appropriations. Cereals take up more than half of the global water used in agriculture. M.Sc. Thesis I.Y.R. Odegard

46

When we look at how much water is needed to produce a kg of a certain commodity, it shows a completely different picture (see Figure 14, right pie-chart). Stimulants (i.e. coffee, tea, cocoa and spices) and nuts are large consumers on a per generated weight basis. So is the category ‘other’, which consists of tobacco (15% of water use in that category) and natural rubber (85% of water use in that category).

Global water use for commodity groups Fraction of total global use and use per weight unit Cereals

Roots and tubers Sugar crops Pulses Oil crops Vegetables Fruits Stimulants Nuts Fodder Fibre crops Other

Figure 13: Fraction of global water use in agriculture per commodity group (pie-chart on the left) and fractions based on use per weight unit (pie-chart on the right) [based on Hoekstra, 2008].

As stated above, around 22% of the water use in agriculture is provided by irrigation. Table 15 below shows the irrigated area for the different regions, and that area as a percentage of the regional total area dedicated to arable land and permanent crops. As we have seen that this area may be overestimated by the FAO, actual irrigated area may account for a larger share than shown in Table 15. As can be seen, the fraction of the total land base that is irrigated seems quite stable, but the absolute numbers show that the irrigated area is expanding a little in all regions except the REF region. This is in line with the shrinkage of the agricultural sector in the former Soviet Union. As the data show, the fraction of the land that is irrigated in the ALM region is close to stable, while the irrigated area is increasing. This is in line with expert opinion stating land expansion is ongoing in this region. The majority of irrigated land area expansion happened between 1950 and 2000. This area increased from 50 Mha worldwide in 1950 to over 250 Mha in the year 2000 [Smil, 2001]. According to Cassman expansion of irrigated area has slowed down recently; water supply and environmental issues are the main limiting factors for future expansion [Cassman, 1999, p.5952].


47

Table 15: Irrigated area (1000 ha) and percentage of irrigated area of total area for arable land and permanent crops in the years 1995, 2000 and 2007 for the four regions [based on FAO, 2010].

Region

OECD90 REF ASIA ALM

Irrigated Area (1000 ha) irrigated area as % of total agricultural area 1995 2000 Area % Area % Area % Area %

45,244 11.40 26,776 9.99 146,710 33.05 45,016 10.81

47,628 12.00 25,720 10.19 157,170 34.17 47,036 11.18

2007 49,372 12.86 25,136 10.27 163,107 34.41 49,111 10.89

The prevalence of irrigation within the regions is substantial, as can be seen in Figure 14. Irrigation intensity is correlated to geographical location (agro-ecological zone) and to economic development. Water requirements are not globally uniform, there are several factors that determine the water requirement for a certain crop. De Fraiture mentiones climate, mode of cultivation – rainfed or irrigated, high input or low input agriculture -, crop variety, length of growing season and crop yield as factors that determine crop water requirements [De Fraiture, 2010, p. 503]. In developing countries with unpredictable rainfall and uncertainty concerning the output price of crops, investments in agricultural technology like irrigation may carry too great a risk for (small) farmers [Molden, 2007].

Figure 14: Share of irrigated land in arable land and permanent crops [FAO, 2009]

Competition over rainwater resources is usually not an issue; precipitation – the water used in rain-fed agriculture –– can be thought of as a characteristic of the land. For irrigation, however, water needs to be extracted from groundwater or surface water resources and thus competes with other sectors [WWAP, 2009]. According to the International Water Management Institute, ‘Irrigation water was M.Sc. Thesis I.Y.R. Odegard

48

essential to achieve the gains from high-yielding fertilizer responsive crops’ [Molden, 2007, p. 59]. This is collaborated by the FAO, stating that ‘Irrigated crops require higher levels of fertilization for optimal productivity, and there is often synergy between the irrigation and the fertilizers’ [FAO, 2006, p.20]. Combining the use of fertilizers to irrigation can raise yields significantly. An experiment with wheat production in Morocco showed that yields were increased by 18% and 35% (for different types of wheat) by adding irrigation, and by 68% and 71% by combining irrigation and fertilization [FAO, 2006]. In the developing countries, irrigation will play an important role in agricultural development. The FAO projects a 40 million hectare expansion, totaling 242 million ha equipped for irrigation in the developing countries in the year 2030. This figure represent 60% of the total potential in these countries. Cropping intensity will be raised, because of this trend [Bruinsma, 2003]. These FAO projections include raising irrigation efficiency, from 38% to 42%. Such efficiency improvements are especially necessary in countries where water withdrawals account for high portions of the renewable sources (see section on water stress above). 3.3.3

Fertilizer

Use of synthetic fertilizers has increased steadily over the past half century [Pimentel, 1990; Vitousek, 1997]. The production of synthetic nitrogen fertilizers on an industrial scale started in 1913 with the commercialization of what became known as the Haber-Bosch process. Figure 15 shows the increased consumption of nitrogenous fertilizers on a global scale, for the USA and for China. Some of the obvious reasons of this increased use are the expansion of agricultural land and the increase in yields due to the use of higher yielding varieties, which also have higher harvest indexes. Another, not so obvious, reason is pointed out by Pimentel. Agricultural businesses have become more and more specialized (in the USA partially due to tax incentives), a process which separated the livestock industry from feed grain production. This discouraged the recycling of animal waste and the use of crop rotation schemes to introduce nitrogen through growth of leguminous crops, leading to the increased need for synthetic fertilizers [Pimentel, 1990].

Figure 15: Consumption of nitrogenous fertilizers, 1950-1990 [Smil, 2002].


49

Nitrogen fertilizer provides one of the three macronutrients, while P2O5 and K2O fertilizer provides phosphorous and potassium. Phosphorous containing fertilizers are produced from phosphate rock, which is a non-renewable resource. Current global production (in 2005) is estimated at 1.2% of the current reserve, and at 0.4% of the potential reserve base [based on USGS, 2009 and IFA, 2011]. Potassium shows similar figures; production in 2005 was at 0.7% of the current reserve base, and at 0.3% of the potential reserve base [based on USGS, 2009 and IFA, 2011]. Three factors determine the fertilizer requirements by a crop on a specific site: (1) the nutrient demand of the crop, (2) the nutrients made available by other sources (for example nitrogen by leguminous crops grown off-season), and (3) the efficiency of fertilization, which determines the levels of nutrients available for the plant [Appel, 1994]. While optimum fertilization rates for a specific crop may vary considerably between sites, one may reasonably presume that regional averages should be estimable. Policy has, however, had a significant impact on fertilizer consumption, unrelated to actual requirements; ‘In countries where a centrally planned system, with its heavy support to agriculture and the allocation of fertilizers according to plans, was replaced around 1990 by a market-oriented system, fertilizer consumption fell abruptly’ [FAO, 2006, p.43]. Currently, government support is available for the use of synthetic fertilizers in certain developing countries, causing their use to increase [FAO, 2006, p. 43]. Whether this raises fertilizer use to appropriate levels is uncertain. Increased application of fertilizer is not necessarily better. The ‘yield response curve’ usually shows a decreasing additional yield with further increasing fertilizer input, beyond a certain level of fertilizer input [FAO, 1984; Howarth, 2002]. Furthermore, the FAO states that excessive application of fertilizer can suppress yields [FAO, 1984]. The link between fertilizer use and policy is corroborated by Smil, who states that heavy subsidies in the former Communist countries and in the European Union led to excessive use of nitrogen fertilizers [Smil, 2001, p.108]. Table 16 shows the annual production and consumption of fertilizers (the sum of N, P2O5 and K2O fertilizers) in 2007 as well as the net import for the four regions. When positive, ‘net import’ is the amount of fertilizer imported, while a negative ‘net import’ means that amount will be exported. Table 16: Fertilizer production and consumption (sum of N, P2O5 and K2O) in 2007, and estimated average use in kg per hectare per year [based on FAO, 2010].

Region OECD90 REF ASIA ALM

Production of Fertilizers

Consumption of Fertilizer

Net Import

(N, P2O5, K2O) in tons



48,338,121

55,152,447

6,814,326

29,271,173

10,084,498

-19,186,675

75,147,624

85,864,873

10,717,249

22,804,196

28,370,527

5,566,331

a

Use in kg per ha was estimated by dividing consumption by the area under cultivation on arable land and permanent crops as given by the FAO. As stated above, this area may be overrated, which would result in underestimation. On the other hand, multiple cropping is not taken into account, which results in overestimation.


50

According to the FAO, around 10 million ton is applied to pastures [FAO, 2006, p.60], a little less than 6% of the total fertilizer production. Figure 16 gives insight into regional differences, and differences between crops. Figure 16 shows the N-fertilizer consumption per hectare per year for the various crops, per region [based on FERTISTAT, 2011]. Average regional consumption per crop for K2O and P2O5 fertilizers are given in Appendix 14.

Figure 16: Regional average nitrogenous fertilizer consumption (kg/ha/year) in 2005, [based on FERTISTAT, 2011].

In Figure 17 a comparison is made between the actual use of N, P2O5 and K2O fertilizer on a global scale (FAOSTAT data, only available as aggregated figure), the use recommended by the FAO (see Appendix 14) and an estimate of global fertilizer consumption based on an estimate of regional consumptions based on data from the FAO [FERTISTAT, 2011], as was also used for Figure 16 above. It shows that the estimated use is a lower than the actual use in 2005. This may be due to the fact that not all foodstuff production is included in the present analysis (e.g. treenuts are excluded), furthermore, data is only reported for 155 countries. The figure also shows that the estimated recommended use is significantly higher than the actual use. When comparing recommended and actual use on a regional scale it shows that the recommended use of potassium fertilizer (K2O) is much higher in all regions than the actual use. Recommended use of phosphorous fertilizer (P2O5) is higher than the actual use in all regions except the ASIA region. Figure 18 shows the recommended and estimated fertilizer use in the OECD90 region. Figures for the other three regions are given in Appendix 14.


51

Figure 17: Fertilizer use estimates (ton/year): actual, recommended and estimated fertilizer use based on regional data [based on FAO, 2010; FAO, 1984; FERTISTAT, 2011].

The ‘recommended use rate’ given in Figure 17 may be higher for several reasons. First, fertilizer efficiency is likely to have improved between the 1980s and now. Furthermore, how and whether fertilizer efficiency is included in the FAOs recommendation is unclear. Use of the word ‘requirements’ implies a baseline value, but due to the much higher than average values, it is more likely that ‘application rate’ is meant. This leaves room for further optimization of application rate through better management practices. Another explanation could be that currently incorrect NPK ratios are applied. Smil states that fertilizer applications in China specifically have been deficient in P and K. He gives a worldwide average NPK ratio of 100:18:22 (probably based on data from the mid-nineties, according to the data in Figure 17 the NPK ratio was 100:44:30 in the year 2005). For China this was 100:14:8, whereas in the United States it was 100:16:35 [Smil, 2001, p.311]. According to the data from Figure 17, the ratio should be 100:38:62.


52

Figure 18: Recommended and estimated regional fertilizer use (kg fertilizer per ton generated product), [based on FAO, 2984; FERTISTAT, 2011].

The FAO predicts that fertilizer use will increase in the developing countries, although at a slower rate than in the past, and not in all countries. According to the FAO, East Asia would still have the highest fertilizer consumption rate in 2030, of 266 kg per hectare, while Sub-Saharan Africa would still have the lowest, of under 10 kg per hectare [Bruinsma, 2003, p.17]. There are several reasons the FAO predicts a slowdown in the increase in fertilizer use. The current levels of fertilizer applied are deemed quite high, agricultural production growth is predicted to decelerate and the FAO predicts an increase in fertilizer use efficiency [Bruinsma, 2003]. Fertilizer uptake rates vary considerably, Appel reports uptake rates on different sites of between 33% and 96% [Appel, 1994]. There are many ways in which such fertilizer losses can be reduced. Direct measures, pointed out by Smil, include ‘soil testing, choice of appropriate fertilizer compounds, maintenance of proper nutrient ratios, and attention to the timing and placement of fertilizers’ [Smil, 2000, p.114]. Furthermore, the need for fertilization can be reduced, and the efficiency improved, by the planting of leguminous crops, or by optimizing conditions for other diazotrophs (bacteria that fix nitrogen) and by ‘good agronomic practices embracing crop rotations, conservation tillage and weed control’ [Smil, 2000, p.114]. The FAO states that fertilizer input to wheat can be reduced by 30-40 kg N/ha in the case of rotation with a leguminous crop. For potatoes this value is even higher: 40-50 kg N/ha [FAO, 1984]. According to Smil, significant gains can be made during the next two generations (assumed to be 50 years from the year 2000). Fertilizer efficiency could be raised by at least 25-30%. For ‘modernizing countries’ this would mean average uptakes rates of around 50-55%, for ‘affluent nations’ such uptake rates would be around 65-70% [Smil, 2000]. Crop yield response to fertilizer and irrigation go hand in hand, as investing in either technology improves overall yields when the other is implemented as well. M.Sc. Thesis I.Y.R. Odegard

53

3.3.4

Organic Agriculture

The term organic agriculture is described by Badgley as ‘farming practices that may be called agroecological, sustainable, or ecological; utilize natural (non-synthetic) nutrient-cycling processes; exclude or rarely use synthetic pesticides; and sustain or regenerate soil quality’ [Badgely, 2006] and will be used as such here. There are advantages to raising crops organically, such as better water retention, lower fossil energy inputs, higher soil organic matter and higher soil nitrogen [Pimentel, 2005]. According to the FAO, the agricultural area managed organically in 2008 was a little over 14 million hectares. This represents 0.3% of the current agricultural land base. Of the organic agricultural area, 66% is located in South-America, 12% in North America, respectively 7% and 9% in Southern Europe and Northern Europe. Western Europe, Eastern Europe and Western Asia account for the remaining 6%, of which two-thirds is located in Western Asia [FAOSTAT, 2010]. Critics claim that organic agriculture would not be able to feed the current population. Lower yields would mean expansion of agricultural area, thereby offsetting the environmental advantages [Badgely, 2007; Smil, 2001]. According to Smil, ‘The only way to support 10 billion people by traditional cropping dependent solely on recycling of organic matter and rotations with legumes would be to double, or even to triple the extent of currently cultivated land’ [Smil, 2001, pp. 4647]. Others, however, claim that with suitable management practices, such as rotation cropping and growing leguminous crops off-season, reasonable yields can be achieved. Badgley et al, in a literature research consisting of a global dataset of 293 examples, compared organic (or low input) agricultural yields to conventional yields. Their analysis shows that organic yields vary from 0.816 to 1.005 times conventional yields in the developed world, and between 1.573 to 3.995 times conventional yields in the developing world [Badgley, 2007]. Mäder et al, state that in their 21-year study done in Central Europe, comparing organic crop yields to conventional crop yields, mean crop yields were found to be 20% lower in the organic system. Depending on the crop, yields in the organic system ranged from 58% to 90% of those in the conventional system. In the organic system, inputs of fertilizer were reduced 34% to 53% and inputs of pesticides were reduced by 97%. They state that currently, organic yields in Europe are typically 60% to 70% of current conventional yields [Mäder, 2002]. Pimentel et al also compare organic and conventional yields, from data gathered from a 21-year study in the US comparing two different organic systems to a conventional system. Corn yields were significantly lower in the first five years of the study (20% and 30%), but after this transition period, yields were similar (differing only 2% and 3%). Furthermore, they found that corn yields in dry years were significantly higher in the organic systems (28% and 34% higher). During a drought, soybean yields also responded favorably in the organic systems, with yields being 35% to 50% higher than in the conventional system. It was found that improved water retention in the organic systems accounted for these higher yields [Pimentel, 2005]. Poudel et al also report organic crop yields (for tomato and corn), which are comparable to conventional yields [Poudel, 2002]. The summary above, of some studies comparing organic yields to conventional yields, shows that researchers can achieve similar-to-conventional yields with organic or low-input agriculture. The question is whether these results can be obtained on a large scale. As stated by Smil ‘even a complete recycling of all organic wastes from the current harvested land and from all confined domestic animals M.Sc. Thesis I.Y.R. Odegard

54

would not be able to supply all macronutrients removed from soils by modern high-yield cropping’ [Smil, 2001, p. 46]. This does, however, exclude the use of growing cover crops and leguminous crops that increase nitrogen availability by biofixation. Badgley et al. estimate that growing cover crops off-season would make enough nitrogen available to replace synthetic fertilizer [Badgley, 2007]. The amount of nitrogen available after growing leguminous crops depends on the harvesting index – the ratio of edible or usable biomass to total biomass. Harvesting indexes have gone up substantially with the green revolution, and differ widely for different crops; between 30% for beans and 80% for soybeans [Smil, 2001]. Soybeans may not even be able to fix all the nitrogen they need, and in the US around 20% of the planted area for soybeans receive additional fertilizer. Grown as green manure (crops grown to fixate nitrogen that are plowed back completely into the soil) soybeans, like clover and alfalfa, provide good enrichment. One of the advantages of green manures is that nitrogen recovery is much higher than it is for synthetic fertilizers. Recovery for green manures is generally around 70%, but can be up to 90%. Uptake rates can be as low as 18% for synthetic fertilizer, and are in general a little over half of the applied nitrogen [Smil, 2001]. Growing cover crops off-season is a good way to provide natural fertilization, but economic viability may hamper implementation. The same argument applies to the growing of green manures. According to Smil, in 1975 green manure cultivation was at its peak, at 9.9 Mha, which dropped to only 4 Mha in 1989 [Smil, 2001, p. 120]. Smil states that the reason this practice lost its desirability is the pressure to produce more food on limited land resources. Another issue related to economic viability is the use of animal manure and specifically the transportation costs involved. As stated above, animal manure does not provide enough nutrients to supply the global demand for micronutrients. This does not mean that animal manure cannot be used to provide fertilization on a local scale. The costs of transportation, however, are in many cases too high to be financially more attractive than the use of synthetic fertilizers. Furthermore, there would be a need to supplement with green manures, off-season cover crops and/or low inputs of synthetic fertilizer. In addition, in the developed world, agriculture and animal husbandry have developed as separate industries (see also Section 3.3.3), which makes it difficult to close cycles. In regions where agriculture is less developed and in general already more integrated with animal husbandry, possibilities for joint development incorporating the closing of cycles (thus the use of animal manures) may be promising. 3.3.5

Animal Husbandry

We have seen that the production of animal products has increased significantly over the past decades. As will be explained in Chapter 4, this increased demand for animal products is correlated to the level of economic development. An assessment of the linkages between the use of natural resources and animal production has to start with an analysis of the current feeding methods. While pasturing, at first glance, may seem intensive from a land use perspective, feeding animals feed cereals and other food crops is much more land intensive because of the production of these crops. According to Smil, actual space needed to keep animals is much lower than the land area needed to grow feed crops; ‘the optimum allotment of space for growing and finishing pigs is about one m2/head; the two animals that occupy sequentially that area during one year will consume 600 kg of feed, which assuming that the pigs are raised on a mixture of concentrate feed, will need on the order of 1,000 m2 of arable land to grow’ [Smil, M.Sc. Thesis I.Y.R. Odegard

55

2001]. The land needed for the production of feed accounts for most of the land use during the whole life cycle of the animal food production industry. Pastures – also providing feed – are included. Table 17 shows the feeding efficiencies for different animals. These differences stem from differences in size and thus metabolic rate, and differences in gestation and lactation periods. As can be seen, cattle are least efficient in metabolizing feed to meat, while poultry is most efficient. Table 17: Efficiencies of animal food production, land requirements and water requirements [Smil, 2001].

Efficiency factor a

Feed (kg/kg EW ) b 2 Land requirements (m /Mcal) c Water requirements (g/kcal) d Food energy (kcal/kg )

Beef

Pork

Commodity Poultry

20.0 6-10 25-35 1,200

7.3 2-2.5 5 3,100

4.5 2.5-3 6 1,800

Eggs

Milk

2.8 1.5-2 1.5 1,600

1.1 1-1.5 10-15 650

a

EW = edible weight. Requirements are based on a common share of 20% of the total coming from by-products, with a minimum 15% share of ruminant roughage. Average feed crop yield was assumed to be 6 t/ha. c Requirements are based on animal food production in temperate climates. d Slaughter-weight – weight as reported by FAO b

There are differences between animal production in the developing world and in the developed world. The feed/meat ratio – the feed input divided by the meat output - is much lower in developing countries for several reasons: (1) Animals are more often pastured, (2) animals receive household waste as food input, and (3) harvest by-products are more often fed to animals in developing countries [SOW-VU, 2005; Wirsenius, 2000]. The FAO estimates that 666 million tonnes of cereals, 35% of the total global cereal consumption, are currently used as feed [Alexandratos, 2006, p.51]. The data in Appendix 4 show that this figure was 31.7% of global cereal production in 2007. Regional data show that in the OECD90, the REF and the ALM region cereal feed production is higher than average, with respectively 48.5%, 42.3% and 36.8% of regional cereal production being used as feed, while in the ASIA region it is almost half of the global average: 15.9%. A quick estimate showing the difference in feeding methods can be given by dividing the tonnage of animal products over the available acreage of pastures. While this does not yield useable information on the yield of animal products per hectare because animals are also fed food crops, it does show the significant difference between the regions, which reflects the increased use of such food crops as feed in the OECD90 region. In this region the ‘yield’ of animal products is around 504 kg/hectare. The ASIA region comes close to that figure, with 494 kg/hectare. Production of pork is high in the ASIA region, and pigs need less feed to gain weight (half of what is needed to produce beef) and the edible weight is also higher (15% higher) for pork than for beef *Smil, 2001, p.157+. In the REF region the ‘yield’ comes to 322 kg/hectare, and the ALM region has the lowest ‘yield’ of 106 kg/hectare. Feeding efficiencies are higher in the industrial world [Wirsenius, 2000; Wirsenius, 2010] and the numbers reflect that pasturing animals is more common in the regions with lower ‘yields’ of animal products per hectare. Furthermore, slaughter weight is lower in the developing world [FAO, 2000]. Appendix 8 elaborates on feed-mixes and feeding efficiencies. M.Sc. Thesis I.Y.R. Odegard

56

3.4

Productivity

Table 18 shows the global and regional yields (in tons per hectare) for the seven commodity groups, in ton of generated product per harvested hectare, for the year 2005. Yield increases, as documented by the FAO between the early sixties and the late nineties, varied between an 11% increase (sunflower) and a 106% increase (wheat). The three main cereal crops maize, rice and wheat showed increases of respectively 99.6%, 84.4% and 106% [FAO, 2000]. Several factors played a role in the ‘Green Revolution’ in the mid-twentieth century. As summarized by Cassman, there were three ‘production factors’ that increased yields so substantially that food production kept up with population growth. Technological development played a large role in the development and implementation of these factors. The three factors are (1) the introduction of new varieties of cereal crops which have a higher harvesting index and ‘better’ plant characteristics such as increased stalk strength, (2) the introduction of synthetic nitrogen fertilizer, and (3) increased use of irrigation. These ‘production factors’ – driven by technological development – were accompanied by government policies and economic and social development [Cassman, 1999]. Analysis of FAOSTAT data has shown that yields have been increasing for all commodity groups, but that the yield gap – the difference between the maximum attainable yield and the current yield – is much smaller in the industrialized world than in the developing world. Thus even though yields are lower in the developing for most commodity groups, room for improvement is larger. Table 18: Yields (tons per hectare harvested) for the seven commodity groups in the four regions for the year 2005 [based on FAOSTAT, 2011].

Commodity Group

Yield in 2005 (tons per hectare harvested) OECD90 REF ASIA ALM

Cereals Fruits Oil Crops Pulses Roots and Tubers Sugar Crops Vegetables

WORLD

4.92

2.43

3.61

1.98

3.33

13.06

5.10

9.81

10.06

10.04

1.56

1.34

2.02

1.12

1.58

1.86

1.73

0.75

0.62

0.86

36.00

13.55

17.04

9.74

13.64

64.90

31.88

60.21

69.08

62.23

27.84

16.15

16.81

13.64

17.28

Note: Yields are given in tons per hectare harvested.

In the future it can be expected that a combination of policies, economic and social development and technological development will define agricultural practices. As explained, the potential yield depends on the agro-ecological zone, and therefore cannot be changed. Agricultural management practices, socioeconomic conditions and policy also have an influence, and, according to the FAO adoption of better varieties and fertilization will be able to increase yields where there is the agroecological potential for it to happen [Bruinsma, 2003, p.15]. More information about the maximum attainable yields and potential future yields can be found in Appendix 7.


57

3.5

Supply and Demand

To assess whether current food production is sufficient to nourish the present global population, and whether the regional populations can be self-sufficient, supply data will be compared to demand data. The ‘current supply’, given in Table 19, is based on the FAO estimate of food available in the four regions in 2005 [FAOSTAT, 2011]. However, this includes import and export. To assess the amounts that regions would be able to feed their inhabitants, allowance needs to be made for imports and export. Food may be exported for economic reasons while inhabitants of that country are insufficiently nourished. Therefore, the food fraction of the net import (as given in Section 3.2) was subtracted. Because it is unknown for which purpose foodstuff is traded, the food fraction of net import was assumed equal to the food fraction of total production; i.e. the fraction of total production which is destined to be food. The demand data are based on the FAO’s threshold of sufficient nourishment of between an average of 2800 and 2900 kcal per person per day [FAO, 2006]. A category ‘other’, which includes nuts and alcoholic beverages provides on average on a global around 77 kcal per capita per day. The example demand proposed in Table 19 provides a reference for comparison to the regional supplies. Table 19: Supply and demand [Based on data from Bruinsma, 2003; FAOSTAT, 2011]

Commodity group

Cereals, food Roots and tubers Sugar and sweeteners Vegetable oils Oil crops Pulses Meats Milk Eggs c Vegetables a Fruit Other Total kcal/capita/day)

Current Supplyb (kcal/cap/day) (2005)

Example Demanda 1503 156 229 263 56 196 119

277 2800

OECD90 1482 139 471 422 49 55 499 359 52 80 99

REF 1693 242 270 251 16 22 225 291 45 76 51

ASIA 1328 107 118 240 57 45 142 60 32 92 64

ALM 905 204 303 228 57 79 339 11 45 12 111

World 1251 146 217 253 53 54 248 104 39 69 80

3709

3183

2291

2297

2519

Note: values may not add up due to round-off errors. a Demand is based on the global commodity composition in 1997-1999, which adds up to 2804 kcal/cap/day [Bruinsma, 2003, p. 53]. Other food includes fruits, vegetables, eggs, oil crops, nuts and alcoholic beverages, these are not specified by Bruinsma. b Supply data is based on ‘food’ minus ‘food fraction of net import’, thus neglecting imports and exports and is given for the year 2005 [FAOSTAT, 2011].

As explained in Box 3.1, the 2800-kcal threshold takes inequitable distribution into account. Many sources state that there is enough food to feed the global population, but the data given in Table 19 do not corroborate this. When allowing for unfair distribution – setting a required average of 2800 kcal per capita per day – the global supply is not enough to meet global demand in kcal. As can be seen, the global average comes to 2519 kcal per capita per day. M.Sc. Thesis I.Y.R. Odegard

58

When taking a closer look at the regional supply averages, it becomes clear that the differences are vast. In the OECD90 region, supply exceeds demand in all categories except for ‘roots and tubers’, and totals at 3709 kcal per capita per day. The REF region comes in second with an, also high, value of 3183 kcal per capita per day. The ASIA region and the ALM region both do not produce adequate amounts of food the ASIA region has an average of 2291 kcal per capita per day, while the ALM region had 2297 kcal per capita per day. Because import and export are not included, actual availability is in reality a little more favorable for the developing regions. The ASIA region ‘imports’ 39 kcal/cap/day and the ALM region 118 kcal/cap/day. This still does not raise supply enough to meet actual demand. Furthermore, the fact remains, however, that on a global scale the threshold for sufficient nourishment is not met.

3.6

Losses and Wastes

Losses and wastes occur during all stages of the food life cycle. National food supplies are calculated by the FAO by subtracting the export quantity and adding the import quantity and the stock changes to the production quantity. This yields the domestic supply quantity. From this other uses, i.e. feed, seed, other utilities, food that is processed and cannot be converted to primary equivalents, and waste (i.e. wastes that occur during transport, processing and storage) are subtracted to yield the food supply. This is shown in Figure Pre-harvest and harvest losses are not included, but are not included in production data either, and neither are household and retail losses. Production Quantity

+

Import Quantity

+

Stock Variation

-

Export Quantity

=

Domestic Supply Quantity

-

Feed

-

Seed

- Processed -

Other Utilities

-

Waste

=

Food

Figure 19: Commodity balance calculation method flowscheme [based on FAOSTAT, 2011].

Here, the term losses is used in the broader sense, including all losses, both unpreventable and preventable, between harvest and retail. Food used for the following purposes (categories as defined by the FAO), are thus considered losses: feed, seed, waste (during transport, storage and processing), other utilities and processing. The term waste is only used for those losses that are preventable, i.e. household and retail waste. Household and retail waster includes both edible losses, which can be considered waste, and in-edible losses, such as peels, which are called refuse and are unpreventable. A note that should be made is that the category ‘food’ still includes refuse, i.e. for unprocessed fresh vegetables and fruits, and meat (bone-in weight) and thus the edible supply is lower still. The ‘food’ fraction of production in Table 20 therefore underestimates the amount of edible food for certain commodity groups. Loss of foodstuff towards processed food is most significant for oil crops and sugar crops, and in these cases the production of the secondary foodstuffs (i.e. vegetable oils and sugar and sweeteners) are included in the present study. Processing of other commodity groups ranges between 0.1% (vegetables) and 10% (fruits) and are considered losses because of lack of information. Wine is, according to the FAO, not included in the ‘fruit-data’, although it is the only commodity mentioned under ‘crops processed’ with a fruit-origin. Fruit that is processed is processed into juice, and the pulp is removed and used as animal feed. Some sort of allocation would be necessary to account for the use of fruit for this purpose, M.Sc. Thesis I.Y.R. Odegard

59

which the FAO does not do. ‘Processed’ (as well as feed, seed, waste and other utilities) is subtracted from the domestic supply, and thus is not included in what the FAO estimates is available as ‘food’. An estimation of the fraction of production that is not available as either food or feed can be made by subtracting the global data for seed, waste, processed food and other utilities from the global production data and comparing these result to what the FAO estimates as ‘food’. Such a method can only be applied on a global scale, thereby avoiding confusion because of trade. Table 20: Uses other than food, % of total global production in the year 2005 [based on FAOSTAT, 2010].

Commodity

Cereals Fruits Oil Crops Vegetable oils Pulses Roots and tubers Sugar crops Sugar and sweeteners Vegetables Eggs Meat Milk a

Purposes of Food Production (% of total global production in 2005) Food 46 80.8 9.7 56.5 62.3 58.1 1.7 84.6 86.6 88.5 99.1 83.4

Feed 36.7 0.9 5.8 0.7 24.4 22.3 1.4 0.3 4.5 0.1 0 11.6

Seed 3.2 0 2.3 0 6.1 4.8 1.3 0 0 6 0 0

a

Waste 4 8.9 2.7 1 4.2 8.2 0.9 0.1 8.4 4.5 0.4 2.3

Processing 6.6 9.1 77 4.4 1.9 1.1 93.9 4.8 0.5 0 0.1 0.2

Other Utilities 3.5 0.3 2.5 37.4 1.1 5.5 0.8 10.2 0 0.9 0.4 2.5

Waste during storage, transport and processing.

Appendix 3 shows the transport, processing and storage wastes as recorded by the FAO for the different commodity groups in the four regions, for the year 2007. Notable are the high wastes in the ALM region; this region has higher than average wastes in all but one category (eggs) and the highest wastes in 6 out of 10 categories. This is explained by the fact that wastes are often estimated as a fixed percentage of the available supply, dependent on the region and that distribution wastes are higher is countries with more humid and hotter climates. It is important to realized that these wastes are quantified based on climate rather than economic development. While this latter factor may play a role, it is uncertain whether it would cut wastes due to higher efficiency, or raise wastes due to higher quality standards and higher quantities being processed. Edible Food Waste It is difficult (and outside the scope of this thesis) to assess whether the waste during storage, transport and processing as reported by the FAO are preventable. Therefore it is assumed here that they are unpreventable. Losses that occur in the household stage, however, can be significant and are for a large part preventable. When considering losses on the household level, one can distinguish between refuse – unpreventable losses – and wastes – preventable losses. Losses of processed products can be considered a 100% waste, while unprocessed products are still partially in-edible. The portion of the primary commodity groups that is processed varies significantly with wealth level, and thus varies between regions. According to the FAO, most of the calories we consume come from processed food. On a global level 75-80% of the calories consumed come from processed products, thus only 20-25% are consumed M.Sc. Thesis I.Y.R. Odegard

60

in their primary product form. The difference between developing and developed countries in their processing rates are large. In the developed countries 55% of the production is processed, while this figure is 22% in the developing world. Economic development is a major factor in the creation of food waste, especially in households. According to Smil, for the years 1992-1994 the difference between the FAO’s FBS and consumption measured by intake surveys in the USA was between 40 and 45 percent. This figure was a little lower for Japan in the mid-1990s: 30 percent. Compared to countries in transition and developing countries this gap is big; around 2000 it was under 15 percent in China, under 10 percent in India and seemed to be non-existent in African countries. According to Smil around 10-15% of food losses may be unpreventable, but the losses in developed countries are inexcusable [Smil, 2001, p. 210]. From this quote it becomes clear that Smil includes both refuse and wastes in his estimation of household losses. The figures are, however, corroborated by the data in Table 19. Average available supplies in the OECD90 and in the REF region are well above the 2900-kcal threshold, respectively 28% and 10%. At the same time, supplies in the ASIA region and the ALM region are well below the threshold; both 21%. Table 21 shows an estimate of the waste that occurs at the retail level and at the ‘foodservice and consumer’ level and is defined as losses from the edible food supply. Refuse losses are thus already subtracted. The data applies to the US, for the year 1995 [Kantor, 1997]. Similar data were used by Cuéllar in 2010, based on data from 1995 [Cuéllar, 2010]. In that year GDP in the USA, in PPP, was a little over 27,000 US$. In comparison, PPP in Japan in that year was around 22,800, but in China and India it was only respectively 3,300 US$ and 1,460 US$. Table 21: Household wastes as percentage and fraction of (edible) food supply [Kantor, 1997].

‘Losses from edible food supply’ Data for the USA in 1995 (Kantor, 1997) (% of edible food supply) Group as defined by Kantor Cereals Fruits Vegetables Pulses Vegetable oils Roots and tubers Sugar and sweeteners Milk Eggs Meats

‘Grain products’ ‘Fruit’ ‘Vegetables’ ‘Dry beans, peas and lentils’ ‘Fats and oils’ ‘Caloric sweeteners’ ‘Dairy products’ ‘Eggs’ ‘Meat, poultry, and fish’

Retail food loss (%) 2 2 2 1 1 1 2 2 1

Foodservice and consumer food loss (%) 30 23 24 15 32 30 30 29 15

No specific data on the inedible fraction of food supply is available. The data in Table 21 can be used to estimate the difference between apparent consumption (supply, which still includes household and retail waste and refuse) and intake. It needs to be kept in mind, however, that this will yield an overestimation of availability for some commodity groups, specifically fruits, vegetables and meat. M.Sc. Thesis I.Y.R. Odegard

61

3.7

Driving Forces

Factors that are important to determine the future potential appropriation of land, water and fertilizers, by agriculture can now be identified. The previous sections have shown that there are substantial differences in the kind and the quantity of the commodity groups that are consumed and produced in the four regions. It was shown that there is a correlation between economic development and the type of food consumed; increased welfare increases the consumption of animal protein in people’s diet. The amount of food wasted also increases with increasing welfare. It can also be assumed that such waste depends on policy and that losses can be reduced by policy measures. Furthermore, and obviously, there is a link between population size and the total amount of food consumed. Section 3.5 showed that supply is sufficient in the OECD90 and the REF region to fulfill demand, while in the ASIA and ALM regions supply does not yet equal demand. In Section 3.4 the differences in productivity in the four regions were shown. Yields (productivity per ha) vary substantially and were shown to be related to the level of technological development. The use of irrigation and fertilizer are part of the agro-technology development in a country, and their use is correlated to higher yields. Furthermore, policy in a country was shown to have a significant influence on the way natural resources are used. The driving forces identified here – population, economic development, policy, technological development and diet – will be discussed in Chapter 4.


62

4.

Driving Forces and Trends

Chapter 2 describes the scenario methodology used in this research and elaborates on the importance of driving forces. These driving forces – those that are important when studying potential food futures – will be elaborated on in this chapter. As was pointed out in Chapter 2, the IPCC (Intergovernmental Panel on Climate Change) design regarding differences and similarities in possible futures is taken as a guideline for building food scenarios in this study. Five driving forces were identified in Chapter 3, and were selected to assess the potential impact on land use, water use and energy use in the food system. These include some basic driving forces defined by the IPCC, but do not include the driving forces used by the IPCC SRES (Special Report on Emission Scenarios) that relate to energy and resource availability. Furthermore, the IPCC driving force ‘land-use changes’ is not considered, as this is what will be modeled in this study. As the IPCC does not consider feedback driving forces – e.g. land quality degradation as a result of climate change – such forces will not be considered here. The driving force ‘diet change’ was added, as is it an important driver for changes in the environmental impact of the food system. The driving forces considered in this study are: 

Population



Economic and Social Development



Policy



Technological Change



Diet Change

The importance of these five forces will be clarified in the following sections. A qualitative description of the matter will be given, as well as possible trends. Trends are quantitative descriptions of the topic resulting from different assumptions as to what the future will look like. For example, population growth will be very different in a low-fertility world than in a high-fertility world. These basic assumptions will link the different trends of the driving forces in significantly different ways. This process will lead to four different ‘internally consistent views of what the future might turn out to be’ [Ringland, 1998]. The IPCC SRES scenarios will be used as a guideline and basis for these four different ‘worlds’ throughout this section. The driving forces’ trends and their importance in the context of agriculture and our regional and global food system are explored in detail in the following sections. As the IPCC has already determined which trends fit which scenarios for certain driving forces, for example for population, the SRES scenario name (e.g. A1) will be given for those trends in the following sections. For driving forces which are more open to interpretation, e.g. diet change, trends are projected consistent with other driving forces that are already linked to a certain scenario. The linkages between driving forces and trends in driving forces will be described in Chapter 5, where all trends will be coupled to a specific scenario and the scenario storylines will be given. This will create the basis the quantification of trends given in Chapter 6. M.Sc. Thesis I.Y.R. Odegard

63

4.1

Population

Population is one of the key driving forces used by the IPCC. Quantification of the emission scenarios depends significantly on the population change in the studied period. This means the variety in outcomes of the scenarios is very much correlated to the large variety in population sizes for the different scenarios. The same is true for food scenarios; a large population will need much more food than a small population. Not only the magnitude of the global population is important to evaluate the extent of our environmental impact, the locations where the population is increasing or decreasing will also influence land use, water use and fossil fuel use. Population growth will have a different impact on the food system in different regions. When regional self-sufficiency is an issue, it is important to know the population size in the region, and the amount of food which will be demanded. Zero population growth does not necessarily translate into zero growth in demand for food. In regions with inadequate food consumption levels, demand will continue to rise until the demand is met. Increasing consumption levels and population growth is linked to economic and social development, which will be elaborated on in Section 4.5. In this section population projections for the world and the 4 regions are presented. These were compiled using 2008 data from the UN (United Nations) and the IIASA (International Institute for Applied Systems Analysis). In Section 5.1 the relationship between the population growth trends and the four alternative scenario paradigms will be elaborated on. Population Trends To fully explore the complete spectrum of potential futures, the four IPCC scenarios use three different population projections; the A1 and the B1 scenarios use the same projection. The projections and the characteristics of these three projections that will be used for the definition of the scenarios are shown in Table 22. Because the IPCC Emissions Scenarios are nine years old, adjustments have been made to these earlier predictions. As is shown in Table 22 the population projections were updated from the ones used by the IPCC. The sources remain the same; IIASA data was used for the low and high projections, UN data was used for the medium projection. The IIASA low and high projections represent the outer limits to an 80 percent uncertainty range. This means there is a 10% chance the population will be higher and a 10% chance it will be lower in 2050, than the projections that are given.


64

Table 22: Population projections [IPCC, 2000; Lutz, 2008; UN, 2009]

Growth

Characteristics

Projection in SRES

Projection Update

Scenario

Low

Lowest population trajectory. Low fertility, low mortality and central migration rates.

IIASA “low” (1996)

IIASA “low” (2008)

A1 and B1

Medium

Medium population trajectory. Approaching stabilization of population in Asia between 20502100 and in the rest of the developing world towards 2100.

UN Long Range medium projection (1998)

UN Long Range medium projection (2009)

B2

High

Highest population trajectory. Declining fertility in most regions and stabilization above replacement levels.

IIASA “high” (1996)

IIASA “high” (2008)

A2

The following figures show the low, medium and high population projections for the world and the four regions. A list of the countries in each region is provided in Appendix 1. Comparing initial and final population sizes, the global population shows an increase in population size for all three projections over the given time period, which is shown in Figure 20. The IIASA-low projection is the only projection which shows a decline before 2050; the population size in 2045 is slightly higher than in 2050. The final populations for respectively the IIASA-low, the UN-medium and the IIASA-high projections are 7.78 billion, 9.15 billion and 9.9 billion.

Figure 20: World population projections until 2050 [based on UN, 2009; Lutz, 2008], lines appear in same order as in the legend.

Figures 2 to 5 show population projections until 2050 for the four regions. The final populations for the world and the four separate regions for the three different projections are given in Table 23. As is shown, the final global population will most likely be somewhere between 7.78 billion and 9.9 billion, with the M.Sc. Thesis I.Y.R. Odegard

65

UN medium projection being 9.15 billion. In all projections the ASIA region is close to half of the total global population. Table 23: Final populations 2050 (in billions) [based on Lutz, 2008; UN, 2009]

Region OECD90 REF ALM ASIA World

IIASA-low

UN-Medium

IIASA-high

0.88 0.31 2.39 3.78 7.78

1.06 0.37 3.09 4.60 9.15

1.15 0.41 3.57 4.91 9.9

Figure 21 below shows the three population projections for the OECD90 region. In the IIASA low projection, the population starts to decline starting in the year 2025, while growth stagnates in the UN medium projection. In the IIASA high projection the population continues to increase, mostly due to growth in North America; 76% of the population growth between 2010 and 2050 happens there.

Figure 21: OECD90 region population projection until 2050 [based on UN, 2009; Lutz, 2008], lines appear in same order as in the legend.

In Figure 22 below, the population projections are given for the countries in the REF region. From 2025 on all three projections show a decline, with the decline starting in respectively 1995 and 2010 for the UN medium and IIASA low projections. The dip shown around 1995 could be due to the fact that for data until 2005 the UN medium projection was used.


66

Figure 22: REF region population projections until 2050 [based on UN, 2009; Lutz, 2008], lines appear in same order as in the legend.

As stated above, around half of the global population in 2050 can be found in the ASIA region. According to the IIASA low projection population starts to decline in 2035 in this region, while the population continues to increase over the given period in the other two projections, even though growth rates decline, as can be seen in Figure 23.

Figure 23: ASIA region population projections until 2050 [based on UN, 2009; Lutz, 2008], lines appear in same order as in the legend.

Figure 24 below shows the population projections for the ALM region: Africa, Latin America and the Middle East. As can be seen, this region has the highest growth rates, and none of the projections show any significant slowing down of these growth rates. Between 2005 and 2050 the population of the ALM region increases by 40% (for the IIASA low projection) to 112% (for the IIASA high projection). M.Sc. Thesis I.Y.R. Odegard

67

Figure 24: ALM region population projections until 2050 [based on UN, 2009; Lutz, 2008], lines appear in same order as in the legend.

The IIASA projections, unlike the UN projection, do not project on a per country basis, but have defined 13 regions for which population projections are given: Japan/Oceania, Western Europe, North America (together forming the OECD90 region), Central Asia, Eastern Europe, European Soviet Union (together forming the REF region), South Asia, China Region, Pacific Asia (together forming the ASIA region), and North Africa, Sub-Saharan Africa and the Middle East (together forming the ALM region). Projections per country were estimated by calculating the fractions of the populations in the year 2005 of the 13 regions named above, and applying these same fractions to the population projections for those regions for the year 2050.

4.2

Economic Development

Economic development is an important driving force when studying any environmental impact on a global scale and it is therefore one of the key driving forces in the IPCC SRES. Such development is linked to increases in resource use and in consumption and is therefore essential when studying the future impact of our food system. The environmental effects of economic development depend on the level of development of the country. With increasing GDP (Gross Domestic Product), the poor will increase their spending on food first. Furthermore, as will be further elaborated in Section 4.5, with increasing wealth, diets tend to change toward increased meat consumption which is directly linked to an increase in foodcrop production for feed and pasture requirements. The IPCC uses GDP as a measure of economic development in the SRES scenarios. The IPCC has received some criticism for using market exchange rate (MER) as opposed to purchasing power parity (PPP), because the latter more accurately reflects the increase in purchasing power. In this study PPP will be used because it gives a more accurate view of the ability people have of spending money on certain items. As is described in numerous articles, people change their diet both quantitatively and qualitatively M.Sc. Thesis I.Y.R. Odegard

68

with increasing wealth; ‘As their purchasing power grows, countries in relatively early stages of modernization will see appreciable increases in average per capita intakes of both food energy and protein’ [Smil, 2001, p. 207]. In the following section PPP projections for the four regions up to 2050 for four different scenarios of economic development will be presented. Economic Development Trends In the SRES scenarios 4 different levels of economic development are explored. These are shown in Table 24 together with the data and the source which was used for respectively the SRES scenarios and this study. The data used by the IPCC in the SRES scenarios was developed specifically for the SRES by the IPCC, because long-term economic development projections were not available. For this study, PPP data was obtained from the PBL (Planbureau voor de Leefomgeving or Netherlands Environmental Assessment Agency), which defined PPP for regions compatible with this study. The PBL based the PPP data on the World Development Indicators (WDI) from the World Bank. Table 24: Economic Development Projections [van Vliet, 2010; IPCC, 2000; PBL, 2009]

Growth

Characteristics

Projection in SRES

Projection Update

Low

Business-as-usual economic growth in industrialized countries, slow growth in other countries. Large gaps in economic prosperity and between regions. Medium economic growth and slow convergence between countries. High economic growth with convergence between industrialized and less-industrialized countries. Highest economic growth. Development converges and income gaps between regions decrease.

IPCC SRES “A2” (2001)

PBL “A2” (2010)

IPCC SRES “B2” (2001)

PBL “B2” (2010)

IPCC SRES “B1” (2001)

PBL “B1” (2010)

IPCC SRES “A1” (2001)

PBL “A1” (2010)

Low-Medium Medium-High

High

As can be seen in Table 24, the PBL names their PPP projections for each of the four SRES scenarios: A1, A2, B1 and B2. As such, these data will be used for quantification of each scenario in Chapter 6. The figures below show the PPP projections for the incomes per capita for the world and the four separate regions, given in US$-1995 [Van Vliet, 2010]. The data is summarized in Table 25 below.


69

Table 25: GDP Projections (in PPP) per capita for the year 2050 (1000 US$-1995) [van Vliet, 2010].

Region

OECD90 REF ALM ASIA World

2000

Low 2050

26.5 4.3 4.4 3.1 6.8

41.5 11.5 10.6 7.1 11.9

GDP in PPP (1000 US$-1995) Low-Medium Medium-High 2050 2050 50.4 19.1 11.7 15.4 17.7

54.5 22.3 19.1 16.6 22.2

High 2050 64.9 32.2 21.8 23.9 28.4

As can be seen in Table 25 and Figure 25 the GDP projections show a marked increase over the coming 40 years. The A1 projection shows the highest growth, following by the B1, B2 and A2 projections, respectively. Differences between regions are, even in the medium-high and high projections, however, still significant. This is shown in Figure 26.

Figure 25: World PPP Projections until 2050 [based on Van Vliet, 2010], lines appear in same order as in the legend.

As can be seen in Figure 26, the difference between the projections for the OECD90 region and the other three regions is still substantial in 2050. Even so, the income gap between the OECD90 region and the other regions declines. In the year 2000, PPP in the OECD90 region was respectively 6.2, 8.5 and 6.0 as high as in, respectively, the REF region, the ASIA region and the ALM region. Depending on the projection, these numbers have declined to respectively 2.0-3.0, 3.0-5.8 and 2.7-3.9. Thus, while the income gap is still significant, progress has been made towards closing this gap.


70

Figure 26: Regional GDP projections (PPP) [based on Van Vliet, 2010], lines appear in same order as in the legend.

The PPP projections given above will serve as a measure for economic and social development. Such development is linked to the food system in various ways. For example, countries with a higher GDP generally have higher levels of technological development, making it possible to produce agricultural products using less labor. In this study, the PPP projections will provide the basis for the determination of changes in diet and of technological change.

4.3

Policy

‘Government policies are among the dynamics that influence population growth, economic and social development, technological change, resource exploitation, and pollution management’ [IPCC, 2000, p.155]. Policy should be seen as a broad concept; here it will help explore how the perceived importance of various issues will impact the use of natural resources, and thereby it will provide the backbone to the scenarios. The ‘scenario world’ is a simplified version – a model – of the real world, and so it is assumed that policy will be strictly enforced. This is in line with IPCC reasoning, stating that policies cannot be quantified, and that the scenario storylines ‘give a broad characterization of the areas of policy emphasis thought to be associated with particular economic, technological and environmental outcomes, as reflected in alternative scenario assumptions in the models’ [IPCC, 2000, Section 3.7.3].


71

Policy Trends The IPCC mentions two kinds of policies that are important in this context: agriculture policies and environmental policies [IPCC, 2000]. Agriculture policies are focused on government support, which steers agriculture into a certain direction through the use of subsidies, tariffs or price controls. Environmental policy influences consumer choices, e.g. by taxing meat, and poses restraints on the use of natural resources, e.g. prohibiting the expansion of agricultural land in certain areas. Table 26 below gives a description of the policy characteristics related to the four different aspects that are the main characteristics of the IPCC scenarios; a focus on either economic or environmental aspects combined with a focus on either globalization or regionalization. There are four specific sub-topics whose interpretation is determined or influenced by policy: trade, use of resources, diets and losses. Trade is be restricted in the regional scenarios; regions will have to be self-sufficient. In the global scenarios there are no trade barriers. Expansion of the use of natural resources is restricted in the ‘environment’ scenarios, whereas the ‘economy’ place no restrictions on the use of resources. Furthermore, efficiency of the use of resources – water and fertilizer - is a specific policy objective in the ‘environment’ scenarios. Issues related to diet will be discussed elaborately in Section 4.5. Policy influences diets by either maintaining the status quo (a Western diet in the A scenarios) or directing people and industry towards more sustainable choices (a vegetarian or organic diet in the B scenarios). It seems as though the ‘Western diet trend’ is the way we are headed, however, in a scenario study it is important to explore a range of options. Table 26: Policy Characteristics [based on IPCC, 2000; PBL, 2009; own interpretation]

Policy Characteristics

Consequences

Globalization

A policy focus on globalization will promote free interaction between industrialized countries and lessindustrialized countries.

Policy will allow free trade of food between regions; no trade barriers are in place and food will be distributed equitably.

1

Regionalization

A policy focus on regionalization will limit interaction between regions by e.g. trade barriers.

Trade between regions will be difficult and regions will try to be self-sufficient. Food will have to be produced and distributed within regions.

2

Economy

A focus on economy will mainly not be a focus on environment. Governments will not try to influence people’s food choices or subsidize sustainable practices significantly. Combined with globalization, a focus on economic development will stimulate technological development.

Policy will not pose restraints on the use of natural resources to produce food. People’s diet will become more and more ‘Western’. There is no (governmental) pressure to reduce household and retail wastes.

A


Scenario Paradigm

72

Environment

4.4

Measures are taken to ensure environmental sustainability and resource efficiency. People are directed toward sustainable food choices and policy will favor technological innovation focused on sustainable practices.

Policy will control resource use; land expansion for agriculture will be restricted, especially in high-biodiversity areas like rainforests. Irrigation and fertilization will be used sustainably. People’s diet will shift toward either vegetarian or organic. There is a focus on resource efficiency and reduction household and retail waste.

B


In the IPCC scenarios the driving force ‘technological change’ focuses on changes in energy supply. As this is not relevant in this context, the basic dynamics – e.g. rate of change and direction – will be taken as guidelines to describe technological change in agriculture. The PBL (The “Netherlands Environmental Assessment Agency”) gives an indication of the direction of change related to the food system. These sources, together with a personal interpretation, provide the basis for what ‘technological change’ will look like in different scenarios. Technological change or development in agriculture is seen here as all measures that are related to crop management. The main parameter influenced by technological change is yield. In Chapter 3 it was shown that yields have increased substantially during the green revolution, and that they are still increasing in all regions for all commodity groups. Three aspects of technological development important to yield projections with clear linkages to natural resource use were identified in Chapter 3: irrigation, use of synthetic fertilizer and different agricultural management practice. This section will elaborate on the linkages between these three factors and yield increase projections and will lay the foundation for quantification of yields and yield increases for 2050, for the four scenarios in the four different regions. There is more to technological development than irrigation and fertilization. Research in biotechnology and in plant and pest ecology should be integrated with plant and animal production to optimize soil, water and nutrient use efficiencies [Bruinsma, 2003]. Information and communication technologies could also play a major role here; such technologies can be used to e.g. optimize irrigation and fertilization [Smil, 2001]. Another matter of importance is the geographical range in research. As many people continue to consume locally produced food and depend on rainfed agriculture, it is important for research to have a broad focus because demand continues to grow in many regions. Technological development will only occur given certain investments, policy implementation and knowledge transfer. According to the FAO, ‘The need for further increases in production in the future while conserving the resource base of agriculture and minimizing adverse effects on the wider environment, calls for ever greater contributions from agricultural research’ [Bruinsma, 2003, p.17]. This scenario-study will not go into those issues, but will consider levels of technology development that are consistent with the given M.Sc. Thesis I.Y.R. Odegard

73

scenarios. These issues will not be elaborated on, but what is taken into account in yield projections is whether it is plausible that such development takes place. Ewert et al. give two factors which result in an increase in yields that are driven by technological development; increasing the potential yield (or the ‘maximum attainable yield’) and closing of the ‘yield gap’ *Ewert, 2005+. The yield gap is the difference between the potential yield and the actual yield. The potential yield depends on the agro-ecological zone (which cannot be changed for a certain location and is related to e.g. solar radiation and temperature) and is defined as the yield in that environment where nutrients and water are not limiting factors, and pests and diseases are controlled [Cassman, 2003]. External factors, however, limit the potential yield to the “attainable yield”, for which the available water resources are taken into account. The actual yield is limited further by other factors related to technological development, policy and the socioeconomic environment (e.g. external inputs like fertilizer or pest control). In practice it turns out that yields stagnate at the 80% of the potential yield threshold because of a lack of economic viability [Cassman, 2003]. Irrigation and Fertilization As was elaborated on in Chapter 3, both the use of irrigation and the use of fertilization increase crop yields. Their simultaneous use even increases the yield more than the sum of the individual benefits. The question here is what projections related to fertilization and irrigation, and their respective efficiencies, consistent with the technology development trends above, would look like. In Table 27 below the rate of change and associated characteristics of technological development in the four scenarios are described. As was shown in Chapter 3, between 10% (OECD90 region) and 34% (ASIA region) of the arable land and permanent crops was irrigated in the year 2007. Whether irrigation is profitable depends on annual rainfall, the agro-ecological zone and the crop. In general, yields are higher on irrigated lands than they are with rainfed agriculture. Current wheat yields (global average) are 42% higher on irrigated lands, and rice yields show an even more pronounced difference; yields are 112.5% higher on irrigated lands [De Fraiture, 2007]. The current yield gap is roughly the same for irrigated agriculture as for rainfed agriculture because not only the current yields are higher for irrigated agriculture, the maximum potential is also higher *De Fraiture, 2007+. Irrigation competes with other sectors for the use of “blue water” (extracted from groundwater and surface water sources), especially in areas where water scarcity is an issue. Irrigation efficiency (or water requirement ratio) is defined as the ratio between the irrigation water requirements and the amount of water withdrawn for irrigation [WWAP, 2009]. Implementation of innovative irrigation measures can improve this ratio, thereby decreasing the “blue water” water that is extracted. As was also shown in Chapter 3, fertilizer use differs per region and per crop. Higher inputs are correlated to higher yields, although this relation does not continue indefinitely; yields approach a certain value after which additional application does not increase the yield any further. According to the International Institute for Applied Systems Analysis (IIASA), ‘On average, long-term achievable yields are 10%, 20%, and 55% lower than maximum attainable yields, at high, intermediate, and low levels of inputs, respectively’ [Fisher, 2002]. Such values are corroborated by Ewert et al. Their projections for the year M.Sc. Thesis I.Y.R. Odegard

74

2050 in the EU are that in the A scenarios (high level of inputs) the yield is roughly 30% lower than the maximum projected yield in 2080, while the yields in the B1 scenario (intermediate level of inputs) are projected to be 38% lower. In the B2 scenario (low level of inputs) the yields are projected to be 52% lower [Ewert, 2005]. Fertilizer application rates seem to be higher than absolutely necessary. Like water use for irrigation, fertilizer use can be decreased without loss of with minor loss in productivity with implementation of better application methods and timing, and different – less volatile – sources. According to Smil, application rates can be dropped significantly with minor loss in productivity [Smil, 2001]. Productivity Trends In Table 27 the rate of technological change and the characteristics related to this change are summarized. These characteristics were based on ‘scenarios of European agricultural land use’ developed by Ewert et al [Ewert, 2005], the IPCC SRES, interpretation of the IPCC scenarios by the PBL, the personal interpretation of the information given in these, and other, sources. Table 27: Technological development trends [IPCC, 2000; Ewert, 2005; Ewert, 2006; PBL, 2009; personal interpretation]

Rate of Characteristics Change

Consequences

Scenario

Slow

Technological change is slow and more fragmented than in other scenarios. There is no global dispersion of technological innovation and the focus is on local solutions.

Potential yield does not increase much and the yield gap is not closed. Irrigation is implemented in all suitable areas. Fertilization and irrigation efficiencies do not change.

A2

SlowMedium

Technological change is less rapid and focused on a diverse set of solutions, on a local scale, with an emphasis on the quality of life.

Potential yield is increased a little, the yield gap is not closed. Use of synthetic fertilizer is minimized, which reduces yields. Irrigated area may increase regionally to feed the regional population and irrigation and fertilizer efficiencies are raised.

B2

Medium

Technological change is directed by societal concerns. There is a focus on the global introduction of clean and resource-efficient technologies.

Potential yield is increased, and yield gap is partially closed. Irrigation efficiency is raised, and fertilizer efficiency is optimized.

B1

Rapid

Technological change is focused on a rapid and global introduction of new and innovative technologies.

Increasing the food supply is done by focusing on potential yield increases and efforts are made toward closing the yield gap. Irrigation is implemented in all suitable areas. Fertilization and irrigation efficiencies are not considered important from and environmental perspective, but improve due to implementation of new and innovative technologies.

A1


75

Yield Projections The most important conclusion which needs to be drawn from the assessment in Table 27 are projections related to yields. Various studies have been made related to cereal yields, as cereal crops are the main staple in the global diet. Furthermore, as meat consumption is projected to increase, cereal crops will further gain in importance because of their contribution as feedstuff. Figure 27 shows the historic global average cereal yield between 1960 and 2005, extrapolated with 6 different yield projections. The yield up to 2005 includes both rainfed and irrigated yield, which is why it is possible for one of the projections (for rainfed agriculture) for the year 2050 to dip below the yield in 2005.

Figure 27: Cereal yield projections, ‘I’ indicates irrigated yield, while ‘R’ indicates rainfed yield [based on De Fraiture, 2010], lines appear in the same order as in the legend.

4.5

Diet Change

It was shown in Chapter 3 that different foodstuffs have different land, water and fertilizer requirements. Economic growth will ensure an adequate food supply for a growing number of people, but it will also change the composition of people’s diets. Such changes in consumption patterns will act as a driving force of potential changes in our food system over the coming years. For example, Vinnari points out that ‘meat consumption has been identified as problematic from at least three perspectives: an environmental perspective, an animal perspective and a human perspective’ [Vinnari, 2009, p.269]. As was described in Section 4.3, policy can encourage or discourage certain trends. This makes diet change a different type of driving force in this study than for example population growth; it is a derivative of economic development and policy. Trends that relate to changes in diet composition are examined in this section, and will be linked directly to trends in economic development and policy.


76

Traditional diets evolved over hundreds of years, yielding nutritionally adequate diets of complementary and regionally available foodstuffs [Pollan, 2008; Gerbens-Leenes, 2005]. Fueled by the rapid globalization, modern transportation and conservation methods of the past century, a much wider variety of foodstuffs and beverages has become available to many people in large parts of the world. Furthermore, welfare has been rising steadily, which not only influences the quantity of food people eat, but also changes the composition of people’s diet qualitatively. According to Lotze-Campen, economic factors, mainly food prices and income, are the main determinants of people’s diet [Lotze-Campen, 2006, p.112]. Gerbens-Leenes adds several important factors: ‘personal preference, habit, convenience, social relations, ethnic heritage, religion, tradition, culture and nutritional requirements’ [Gerbens-Leenes, 2005]. The relationship between different food choices and the environmental impact related to those choices has steadily gained attention. Specifically the increase in meat consumption, the associated increase in cereal production for feed and the relation of meat consumption to welfare levels have been studied [Vinnari, 2009; Gerbens-Leenes, 2004; Keyzer, 2005; Grigg, 1995; Smil, 2001].

b

Figure 28: Population and global food production indices, 1966-1998 [Rosegrant, 2001 ]

Diet Trends As mentioned above, per capita income is the main determining factor in food consumption and dietary transition. Several researchers have described the diet transitions and the changes in commodity distribution [Smil, 2001; Keyzer, 2005; Alexandratos, 2006], which are partially due to economic development. Not only meat intake increases with increasing wealth; the complete composition of people’s diet changes. This is shown in Table 28 below, which presents the diet projections by the FAO for the industrial countries and the developing countries for the year 2050.


77

Table 28: Changes in the commodity composition of food by major country groups [Alexandratos, 2006, p.25].

Kg/cap/year

1969/71

1979/81

1989/91

1999/01

2030

2050

Developing Countries Cereals, food Cereals, all uses Roots and tubers Sugar and sweeteners Pulses, dry Vegetable oils, oilseeds and products Meat (carcass weight) Milk and dairy (excl. butter) Other food (kcal/cap/day) Total food (kcal/cap/day)

146.3 191.8 78.8 14.7 9.2 4.9

161.7 219.1 69.6 17.5 7.8 6.5

173.7 238.6 60.1 19.2 7.3 8.6

165.7 238.0 67.0 20.7 6.7 10.4

166 268 75 25 7 14

163 279 77 26 7 16

10.7 28.6 123 2111

13.7 34 140 2308

18.2 38.1 171 2520

26.7 45.2 242 2654

38 67 285 2960

44 78 300 3070

Industrial Countries Cereals, food Cereals, all uses Roots and tubers Sugar and sweeteners Pulses, dry Vegetable oils, oilseeds and products Meat (carcass weight) Milk and dairy (excl. butter) Other food (kcal/cap/day) Total food (kcal/cap/day)

132.3 531.1 74.2 40.5 3.4 13.2

139.4 542.0 67.1 36.7 2.8 15.7

154.4 543.7 69.4 32.6 3.2 18.5

162.4 591.8 66.7 33.1 3.6 21.5

159 641 61 32 4 24

156 665 57 32 4 24

69.7 189.1 486 3046

78.5 201.0 500 3133

84.3 211.2 521 3292

90.2 214.0 525 3446

99 223 565 3520

103 227 580 3540

Note: Cereals food consumption includes the grain equivalent of beer consumption and of corn sweeteners.

Both industrial countries and developing countries have seen remarkable increases in the consumption of several commodity groups, but there are various interesting trends. Meat consumption in the developing countries (the ASIA region and the ALM region) in the year 2000 was about two and a half times as high as in 1970, while, in the same period, meat consumption rose with 30% in the industrial countries (the OECD90 region). Even so, a person in the developing world only consumed approximately 30% of the amount of meat consumed per capita in the industrial world. Not only does the amount of meat consumed increase with increasing purchasing power, the type of meat – beef, pork or poultry also changes. It is, however, beyond the scope of this research to incorporate this. Trends for milk consumption are similar; consumption increased by 60% in the developing countries and by 13 % in the industrial countries, while milk consumption is still 4.7 times as high in the industrial countries. Cereal consumption provides other interesting information. While consumption of cereals as a food source was 2% higher in the developing world than in the industrial world in the year 2000, cereal consumption including feed was 2.5 times as high in the industrial world. This is, of course, due to the much higher intake of meat products in the industrial world. M.Sc. Thesis I.Y.R. Odegard

78

As can be seen in Table 28, consumption of sugar has been decreasing in the industrialized world, and increasing in the developing world. The gap is, however, still significant; consumption is about 30% lower per capita in the developing world. Consumption of vegetable oils has been increasing steadily in both regions, but the consumption in the industrialized world is twice as high as in the developing world. Pulses constitute a larger part of the diet of poorer people, as do roots and tubers. The consumption of the latter is decreasing in the industrialized world, while the FAO projects their consumption will be increasing in the time period until 2050. The consumption of pulses seems to be converging; consumption in the industrialized world is increasing again (since 1980) and is lower than the consumption in the developing world, where consumption seems to be decreasing. According to the FAO, fruit consumption in Africa has risen 6% between 1970 and 2000; from 51.4 kg/cap/year to 54.6 kg/cap/year. During that same period, fruit consumption in Western Europe rose 8.5%, but was close to double the consumption in Africa; 99.1 kg/cap/year in 1970 and 107.4 kg/cap/year in 2000. Vegetables consumption shows a similar trend, even though consumption rose faster. Consumption rose 24% between 1970 and 2000 in Africa, while it rose 12% in Western Europe. However, total consumption was almost twice as high in Western Europe in 2000: 98.7 kg/cap/year as opposed to only 54.15 kg/cap/year in Africa. Most authors agree that there are no indications of a transition away from a per capita increase in meat consumption. Scenario studies, however, become more interesting when a diverse set of futures is presented. In the IPCC SRES there are two scenarios in which sustainability is incorporated, while two take a business-as-usual approach to energy use. Here, likewise, two trends are based on the ‘Western Diet’ diet transition – a trend based on the increase in meat consumption related to economic growth, without taking environmental issues into account (scenarios A1 and A2). Two alternatives are proposed: the ‘Vegetarian Diet’ (a diet based on vegetable products, dairy products and eggs, based on business-asusual caloric intake) and the ‘Low-input Diet’ (agriculture with low-inputs of fertilizers, and reduced meat consumption). Meat consumption projections made by the PBL and a method developed by the Centre for World Food Studies to calculate meat consumption based on PPP are taken as a guideline for the A scenarios, while food consumption in the B scenarios is interpreted more freely. Table 29 elaborates on the diet trends. Table 29: Diet Trends [PBL, 2010; own interpretation].

SRES Projection [PBL, 2010]

SRES Scenario

Current Interpretation

Trend

‘Fast increase in per capita consumption of livestock products as a result of GDP increase’ ‘Slow increase in per capita consumption of livestock products as a result of GDP increase’

A1

Business-as-usual increase in meat consumption, based on PPP, calculation with ‘Keyzer equation’ *Keyzer, 2005]. Average apparent consumption needs to exceed the FAO threshold of 2900 kcal per capita per day. Business-as-usual increase in meat consumption, based on PPP, calculated with ‘Keyzer equation’ *Keyzer, 2005]. Consumption of vegetable sources is based on historic trends. Average apparent consumption needs to exceed the FAO threshold of 2900 kcal per capita per day.

Western Diet


A2

Western Diet

79

‘Per capita consumption of livestock products is 10% lower than in A1 scenario in 2050 and 20% lower than in A1 in 2100’ ‘Moderate increase in per capita consumption of livestock products as a result of GDP increase’

B1

Business-as-usual increase in caloric intake, based on PPP. Diet based on vegetable sources, dairy and eggs. Consumption of sugar and sweeteners is reduced and of vegetables and fruits increased due to health concerns. Average apparent consumption conforms to a lowered FAO threshold of 2800 kcal per capita per day.

Vegetarian Diet

B2

Meat consumption conforms to half of that determined by the ‘Keyzer equation’. Consumption of sugar and sweeteners is reduced and of vegetables and fruits increased due to health concerns. Average apparent consumption conforms to a lowered FAO threshold of 2800 kcal per capita per day. Agriculture uses low-inputs of fertilizers.

Low-input Diet

As presented in Table 29, meat consumption in the A scenarios follows a business-as-usual trend based on purchasing power. A mathematical relationship between PPP and meat demand was established by Keyzer et al, based on data for 125 countries over a 26 year period [Keyzer, 2005]. Figure 29 shows meat consumption projections.

Figure 29: Meat consumption trends in different scenarios [based on Keyzer, 2005; Van Vliet, 2010; Alexandratos 2006].

For the A scenarios, these were based on the equation derived by Keyzer, combined to data on PPP. The FAO projections for different subregions in the ASIA region and the ALM region were averaged relative to population size, to fit the ASIA region and the ALM region. The projection for the B2 scenario is, like the A scenarios, based on the Keyzer equation, but adjusted to fit the scenario paradigm; projected meat consumption is cut in half.


80

While in both A scenarios the ASIA region and the ALM region consume the least meat, consumption in these regions goes up from respectively 27 kg and 32 kg per capita per year in the year 2005, to respectively 53 kg and 57 kg per capita per year in the year 2050 for the A2 scenario and respectively 91 kg and 85 kg per capita per year in the year 2050 for the A1 scenario. The global average meat consumption increases significantly in both scenarios; it increases 60% in the A2 scenario and 141% in the A1 scenario. Increase in consumption is much higher in the developing regions than in the developed regions. The increase in the developing regions is between 80% (ALM, A2) and 239 (ASIA A1), while the increase in the developed regions is between 20% (OECD90, A2) and 94% (REF A1).


81

5.

The Future of Food Storylines

In Section 5.1 the 5 driving forces identified in Chapter 3 and discussed in Chapter 4 are grouped to create four consistent scenarios. These scenarios are subsequently described in Section 5.2 to 5.5 in, mostly qualitative, scenario storylines.

5.1

Linkages between Driving Forces

To understand the importance of the driving forces in scenarios, it is necessary to explore the linkages between these forces. Certain driving forces may have a positive feedback on others, while some may be completely unrelated. For example, a low population growth is positively related to high economic development. The linkages reflect IPCC reasoning in compatibility of trends and consistency between them. Some pre-defined linkages, e.g. the relation between population growth to economic and social development, will not be changed here. Table 30 shows how the main driving forces and their subthemes are fitted into consistent scenarios. The plusses and minuses indicate the development of the trend related to the topic named. A plus or minus either indicates whether something is true (+) or not (), e.g. for ‘globalization’, or whether the trend is projected to increase at a higher (+) or lower (-) than average rate, e.g. for population growth. The source and rationale for these choices is given. Where no source is mentioned, the trend projection is a personal interpretation. Table 30: Scenario driving forces' linkages and characteristics.

Driving Forces

A1 A2 B1 B2

Source and rationale

Population growth

-

+

-

+/-

Economic and social development

+

-

+

+/-

+ + -

+ +

+ + -

+ +

+

-

-

-

[IPCC, 2000; UN, 2009; Lutz, 2008] Low population growth is correlated to (high) economic development and (rapid) technological development. [IPCC, 2000; UN, 2009; Lutz, 2008] High economic development is correlated to rapid spread of technological knowledge and thus to globalization. [IPCC, 2000] The main scenario directions are linked in different ways to create 4 significantly different scenarios. Globalization allows for free trade and global dispersion of knowledge. Self-sufficiency is considered important in the ‘regional’ scenarios. The focus on ‘economy’ or ‘environment’ determines aspects related to technological change. [Kantor, 1997; Cuéllar, 2010; Bender, 1994; own interpretation] High levels of household and retail waste are related to high levels of economic development, while low levels of waste are related to environmental consciousness

Policy

Themes

Globalization Regionalization Economy Environment

a

Waste


82

Technological change

Diet change

Rate of change

++

-

+

+/-

Fertilizer use

+

+

+/-

-

Increase irrigated area

+

+

-

-

Fertilizer and irrigation efficiency

+

-

+

+

Productivity

++

-

+

+/-

Rise in meat consumption

+

+

-

-

and to low levels of economic development. [IPCC, 2000] Higher economic development levels are correlated to higher rates of change. Policy stimulates (‘economy scenarios’) or discourages (‘environment scenarios’) the use of fertilizer. Increase of the area under irrigation is related to levels of economic development, and policy restrictions on water use in water-scarce areas. High levels of efficiency are related to a focus on environmental issues, and also to economic development and dispersion of technological innovations. Productivity (yield) is correlated to economic growth, use of fertilizer and irrigation, and technological development. Globalization leads to dispersion of technological innovations. Meat consumption is correlated to levels of economic development. A policy focus on the environment and a greater environmental consciousness can lead to a decrease in meat consumption.

Note: a plus or a minus (or a double plus or plus-minus) indicates the growth rate of the topic in the specific scenario, relative to the other scenarios; thus while population growth is positive in the A1 scenario compared to now, its relative growth rate is low compared to population growth in e.g. A2. The four IPCC paradigm indicators – globalization, regionalization, economy and environment - are either ‘true’ or ‘false’ for a specific scenario, although this can be viewed a s a growth rate as well; a world is never truly solely globalized or regionalized. a Household and retail waste

Figure 28 summarizes the information from Table 30 above. The labeling used by the IPCC, i.e. A1, A2, B1, 2, is used in this figure. Additionally, scenario names are added to increase recognizability. Table 30 and Figure 28 provide the basis for the scenario storylines in the following sections.


83

Economic Unrestrained Use of Natural Resources

A1 The Affluent World

A2 The Full World

Low Population Growth High Economic Growth Rapid Spread of Agro-Technology Medium Irrigation Efficiency High Fertilizer Efficiency High Productivity Western Diet

High Population Growth Low Economic Growth Slow Spread of Agro-Technology Medium Fertilizer Efficiency Low Irrigation Efficiency Medium Productivity Western Diet

Globalization

Regionalization

Global Food Distribution

Regional Food Distribution

B1 The Vegetarian World

B2 The Low-Input World

Low Population Growth Medium-High Economic Growth Spread of Sustainable Agro-technologiy High Irrigation Efficiency High Fertilizer Efficiency High Productivity Vegetarian Diet

Medium Population Growth Low-Medium Economic Growth Spread of Sustainable Agro-technology High Irrigation Efficiency High Fertilizer Efficiency Low Productivity Organic Diet

Environment Efficient Use of Natural Resources

Figure 30: Scenario Characteristics


84

5.2

A1 – The Affluent World

In the A1 world, there is a focus on globalization and economic development. Regional and global population growth is shown in Figure 28. The world population will have grown to 7.78 billion in 2050. The combination of a focus on globalization and economic development results in a decrease in inequity between regions and a convergence of welfare on a global scale. For this to happen, developing countries will undergo a remarkable transition in economic development. Dispersion of technological knowledge is both a driver and a outcome of this development. The relatively low population growth is coupled to very high GDP growth.

Figure 31: Population between 19050-2050 (in billions) in A1 [based on Lutz, 2008].

Economic development will happen rapidly on a global scale. Table 31 shows the economic growth rates (in % per year) and the income per capita in the world and its 4 regions. By 2050, GDP in the poorest region – the ALM region – will have increased to the current level of OECD90 development. Table 31: Economic Growth Rates and Income per Capita in the A1 World [IPCC, 2001; Van Vliet, 2010]

Region

World OECD90 REF ASIA ALM

Economic Growth Rates (% per year) 1950-1990 1990-2050 4.0 3.9 4.8 6.4 4.0

3.6 2.0 4.1 6.2 5.5

Income per Capita in PPP (103 US$-1995 per capita) 2000 2050 6.8 26.5 4.3 3.1 4.4

28.4 64.9 32.2 23.9 21.8

Policy will be focused on stimulating the economy, rather than protecting the environment, and will be focused on achieving a global food market. Important for the future of the food system are policies that create an increasing interaction between regions. Trade barriers between countries will be lifted, resulting in the absence of restrictions on trade between the four regions. In all four regions, diets will continue to converge towards a diet which is high in meat consumption, the ‘Western’ diet, as a result of the high economic and social development. This will create incentives for technological development in agriculture, resulting in higher feeding efficiencies and relatively high yields; yield gaps are closed by 80% of the gap in the year 2005. The incentives to create a larger supply continue to be focused mostly on the supply side, and household and retail waste are high. The rapid technological change and the interactions between regions will push the use of irrigation in agriculture as well as the high use of agricultural chemicals. M.Sc. Thesis I.Y.R. Odegard

85

5.3

A2 – The Full World

While the main focus in the A2 world is on regionalization and economy, economic growth rates are low and as can be seen in Table 32 development is distributed very unequally. Table 32: Economic Growth Rates and Income per Capita in the A2 World [IPCC, 2001; Van Vliet, 2010]

Region



2.3 1.6 2.3 3.9 3.8

Income per Capita in PPP (103 US$-1995 per capita) 2000 2050 6.8 26.5 4.3 3.1 4.4

11.9 41.5 11.5 7.1 10.6

Interactions between regions are limited, which displays itself in high inequities between regions, and only very slowly converging demographic trends. Population growth is high, and is still increasing in 2050 in all regions except the REF region, as can be seen in Figure 28. The ALM region in particular shows no sign of a decreasing population growth. The world population will have grown to 9.9 billion in 2050. Policy will be formulated and implemented on a regional scale. This will result in protection of local markets and limits to global trade of food products. The environment is not much of a policy issue, and no limitations will be placed on acreage expansion and irrigated area expansion in an effort to feed a world Figure 32: Population between 1950-2050 (in billions) in A2 population of close to 10 billion people. No measures [based on Lutz, 2008]. are taken to decrease animal product consumption or have consumers make healthier food choices, which increases the use of natural resources. Because technological change is slow, and is focused on a regional scale, specific technologies that would raise yields or lower the environmental impact of agriculture are implemented on a limited scale. This results in only a partial closure of the yield gap and no improvement in irrigation efficiency and fertilizer efficiency. Because economic development is low, household and retail waste is not as high as in the A1 world.


86

5.4

B1 – The Vegetarian World

The B1 world is a world in which globalization and taking care of the environment are core values. Similar to the A1 world, the B1 world experiences the same low population growth, shown in Figure 30, and a high growth in GDP, shown in Table 33. The world population will have grown to 7.78 billion in 2050. The B1 world distinguishes itself from the A1 world by the explicit focus on the environment. This will influence the impact agriculture has on the environment through policy, the types of technological change and the diet of the population. Policy in the B1 world will have a strong focus on conservation of the environment. For agriculture this means that there will be no conversion of nature to Figure 33: Population between 1950-2050 (in billions) in B1 cropland, if possible, cropland will be given back to [based on Lutz, 2008]. nature. Table 33: Economic Growth Rates and Income per Capita in the B1 World [IPCC, 2001; Van Vliet, 2010]

Region



3.1 1.8 3.1 5.5 5.0

Income per Capita in PPP (103 1995US$ per capita) 2000 2050 6.8 26.5 4.3 3.1 4.4

22.2 54.5 22.3 16.6 19.1

Technological change, linked to the fairly high economic growth, will increase yields and feeding efficiencies. Yield gaps will be close to 80% of the year-2005 gap, and because of improvements in fertilizer and irrigation efficiency, resource input is more effective in terms of generated output. Irrigated area does not expand as much as in the A worlds. While the focus in the A1 worlds to increase supply is on the supply-side, in the B worlds the demand side gets attention and while household and retail waste increase due to economic development it is only half of that in the A1 world. Furthermore, healthy food choices are important in the B1 world. While meat consumption is cut entirely, protein consumption is considered in diet choices, leading to a higher intake of milk, eggs, oil crops and pulses. Moreover, sugar and sweetener intake decreases, while intake of fruits and vegetables increases. Because globalization is a driver in this world, there will be no trade restrictions, and food will be distributed fairly over the world. Furthermore, because of greater equity, apparent consumption can be lower than the under-nutrition threshold of 2900 kcal set by the FAO, by a 100 kcal. M.Sc. Thesis I.Y.R. Odegard

87

5.5

B2 – The Low Input World

In the B2 world the environment and regionalization play a large role. The B2 world will experience a medium growth in GDP, but as can be seen in Table 34 the differences between the regions in income per capita will still be quite substantial in 2050. Table 34: Economic Growth Rates and Income per Capita in the B2 World [IPCC, 2001; Van Vliet, 2010]

Region



2.8 1.4 3.0 5.5 4.1

Income per Capita in PPP (103 1995US$ per capita) 2000 2050 6.8 26.5 4.3 3.1 4.4

17.7 50.4 19.1 15.4 11.7

As shown in Figure 31, the B2 world experiences a medium population growth, hitting a maximum population a little later than 2050. As can be seen in the figure on the right, population in the OECD90 and REF regions experience very little growth and negative growth respectively, and the ASIA and ALM region will still be growing steadily in 2050. The world population will have grown to 9.15 billion in 2050. This is a world where attention to sustainability is focused on a local scale. For agriculture this will mean a shift towards food grown with low inputs of fertilizers. This is supported by local policy which has a strong focus on local environment and environmental solutions on a local scale. This does, however, lead to Figure 34: Population between 1950-2050 (in billions) in low yields; half of those in the A2 world. While B2 [based on UN, 2009] fertilizer input is low, fertilizer efficiency does increase. Similar to the B1 world, policy has a tendency to protect nature, but because solutions are implemented on a local scale there is no trend towards a globally efficient food system. Technological change is less rapid than in the A1 and B1 scenarios. Irrigated area and irrigation efficiency increase at the same rate as in the B1 world. Diets shift to being their diet towards either a ‘western’ or a ‘vegetarian’ lifestyle –meat consumption is still half that predicted by the economic growth rates–, but instead purchase their food locally, in an effort to be more environmentally friendly. Similar to the B1 world, people make healthier choices concerning their diet and increase their fruits and vegetables consumption. Because of the combination of moderate economic growth and environmental consciousness, household and retail waste levels are half of those in the A1 world. M.Sc. Thesis I.Y.R. Odegard

88

6.

Linkages

The following section will discuss the quantification of the driving forces elaborated on in previous chapters. Table 35 and Table 36 give detailed information about the assumptions made, respectively those related to demand and supply. Figure 35 below shows the relationships between the driving forces and the demand and supply of food. As can be seen some driving forces have an influence on demand while others have an influence on supply. Where the quantification is not straightforward, the method of quantification is explained below.

Policy

Population

Economic Development

Diet

Demand

Losses

Supply

Production

Resource Use

VRC-factors


Waste Policy

Figure 35: Linkages between driving forces, supply, demand, Virtual Resource Content and resource use.

6.1

Demand

As shown in Figure 35 above, economic development and policy influence diets, while population of course influences the total consumption. Furthermore, economic development and policy also have an impact on the amount of retail and household waste. The basis for the diets in the four scenarios are the projections made by the FAO. These projection are made for the world, for developing countries, for industrial countries and for transition countries, for the years 2030 and 2050. These projections have been adapted to be consistent with the scenarios in this study. Changes have been made that are related to the level of economic development and its influence on meat consumption. Meat consumption in the A scenarios was calculated using a relationship between meat consumption and PPP, defined by Keyzer et al [Keyzer, 2005]. In the B scenarios, people make diet choices related to greater health consciousness, which for example increases fruits and vegetables consumption. Furthermore, in the B2 scenario meat consumption as defined by the Keyzer equation is cut in half, while in the B1 scenario meat consumption is cut altogether. Also, the level of apparent consumption is based on the undernutrition threshold, which is reduced in the B scenarios because of higher equity and lower household waste levels. Meat consumption, consumption of other commodity groups and levels of household and retail waste are quantified in Table 35 for the year 2005 and for the four scenarios. In the A1, B1 and B2 scenarios the total food quantity in kcal/cap/day is slightly lower than the level of the related FAO projection. In the A2 scenario this quantity is slightly higher in the OECD90 region and the REF region, while slightly lower in the ASIA region and the REF region, which corresponds to the difference in definition of regions. M.Sc. Thesis I.Y.R. Odegard

89

Table 35: Demand-side assumptions and rationale.

Scenario Topic

Quantification

Source and rationale

All

In 2050, total meat consumption is divided between the different types as such: of total consumption: 26% = beef, 33% = poultry, 41% = pork. Waste fraction for oil crops, sugar crops and roots and tubers is set equal to the waste fraction for pulses.

[FAOSTAT, 2010] Proportions in 2005, global level.

Eradication of undernutrition

Caloric apparent consumption exceeds the FAO undernutrition threshold of an average of 2900 kcal per capita per day in the A scenarios and of a reduced level of 2800 kcal per capita per day in the B scenarios.

[Alexandratos, 2006] Greater equity and lower waste levels allow for a lower undernutrition threshold level.

A1, A2

Meat consumption

Meat consumption is based on the ‘Keyzer equation’, inputs: PPP projections and population projections.

*Keyzer, 2005+; ‘Keyzer equation’, based on PPP. [Van Vliet, 2010]; PPP projections for A1, A2, B2 scenarios. [IIASA, 2008] high and low population projections (IIASA divides the world into 13 regions; the proportion of populations on a country level in 2050 is assumed equal to the proportion in 2000). [UN, 2009]

2005

Meat consumption Other commodity groups Household and retail waste

Global average consumption of meat in 2005 according to the FAO. Global average consumption of commodity groups in 2005 according to the FAO. Global household and retail waste is half of the 1995 USA levels defined by Kantor.

[FAOSTAT, 2010]

Meat consumption

Average meat consumption: OECD90 = 130 kg/cap/year REF = 99 kg/cap/year ASIA = 91kg/cap/year ALM = 85 kg/cap/year World = 93 kg/cap/year Basis: diet in ‘industrial countries in 2050’, according to FAO.

Based on [Keyzer, 2005]

Meat type

Household and retail waste

A1

Other commodity groups


[Kantor, 1997] No values are given for the commodity groups mentioned; a low waste fraction is chosen: pulses.

[FAOSTAT, 2010]

[Kantor, 1997] Household and retail waste is quite high in the industrialized regions, but close to non-existent in the developing world.

[Alexandratos, 2006, p.25] Because of high economic development, diet transition occurs rapidly on a global scale.

90


A2

Meat consumption


Household and retail waste B1

Meat consumption Other commodity groups

Eggs, oil crops, sugar crops, vegetables and fruit are set to the apparent consumption level of the year 2005. Level of food waste in the USA in 1995.

[FAOSTAT, 2010]

Average meat consumption: OECD90 = 107 kg/cap/year REF = 79 kg/cap/year ASIA = 53 kg/cap/year ALM =57 kg/cap/year World = 62 kg/cap/year Basis: diets in 2030 according to FAO. OECD90 = ‘industrial countries’ REF = ‘transition countries’ ASIA = ‘developing countries’ ALM = ‘developing countries’ Eggs, oil crops, sugar crops, vegetables and fruit are set to the apparent consumption level of the year 2005. Household and retail waste is half of the 1995 USA levels defined by Kantor.

Based on [Keyzer, 2005]

[Alexandratos, 2006, p.25] Because of low economic development, diets undergo a slower transition.

[FAOSTAT, 2010]

[Kantor, 1997] Lower waste levels because of lower economic development levels.

Meat consumption is zero. Basis for the diet is the global average diet in 2050 according to the FAO.

Consumption of eggs, oil crops, pulses and milk is increased to increase the diet’s protein content (between brackets former figure): Oil crops = 10 kg/cap/year (7) Pulses = 10 kg/cap/year (6) Eggs = 10 kg/cap/year (8) Milk = 130 kg/cap/year (100) b Sugar consumption is cut in half . Apparent consumption of fruit and vegetables consumption is set to 148 kg/cap/year, based on a healthy intake b of 300 gr/cap/day , excluding 25%


[Kantor, 1997] Household and retail food waste is positively correlated to purchasing power and negatively to environmental consciousness; waste levels in A1 are the highest.

[Alexandratos, 2006, p.25] Low waste levels and greater equity allow for lower apparent consumption rates. Healthier choices are made in the B scenarios; apparent meat consumption is cut in half, and so is apparent sugar consumption. Apparent consumption of fruit and vegetables is high at 300 gr/cap/day.

91

refuse and 10% household waste.

B2


Household and retail waste is half of the 1995 USA levels defined by a Kantor .

[Kantor, 1997] Lower waste levels because of higher environmental consciousness.

Meat consumption

Diet of reduced meat consumption. The average meat consumption is half of that given by the ‘Keyzer equation’: OECD90 = 57.5 kg/cap/year REF = 43 kg/cap/year ASIA = 41 kg/cap/year ALM =30.5 kg/cap/year World =37 kg/cap/year Basis: diets in 2050 according to FAO. OECD90 = ‘industrial countries’ REF = ‘transition countries’ ASIA = ‘developing countries’ ALM = ‘developing countries’ Fruit and vegetable consumption is set to the levels in the B1 scenario. Sugar consumption is cut in half. Household and retail waste is half of the 1995 USA levels defined by Kantor.

Environmental consciousness influences peoples diet choices to a certain degree, but not as much as in B1.



6.2

[Alexandratos, 2006, p.25] Similar to FAO projections, population growth rates and economic development are intermediate, thus diets follow the FAO projections for 2050. Some healthier choices are made related to consumption of fruits, vegetables and sugar and sweeteners. [Kantor, 1997] Lower waste levels because of higher environmental consciousness.

Supply

As shown in Figure 35 above, policy and technological change influence supply. Furthermore, policy also has an impact on the amount of ‘losses’ of foodstuffs – the fraction of the total production which is used for other purposes. This section discusses such losses, irrigated area, irrigation efficiency, fertilizer efficiency and productivity. These topics are quantified and summarized in Table 36. A portion of the production of potential foodstuffs is used towards other purposes than food. A significant quantity is used as animal feed. Furthermore, part of the production quantity is lost during storage, processing and transportation, is used as seed, and is used to produce other utilities (e.g. biofuel). Relative values (the fraction of total production) of seed, waste (storage, processing and transport), processed food and other utilities are kept constant over all scenarios and all regions to increase transparency. Foodstuff used as feed depends on total meat consumption, the type of meat consumed and on the global or regional feeding efficiency, and thus varies between regions and scenarios. Feeding efficiencies are based on data from Wirsenius [Wirsenius, 2000]. For the global scenarios a relatively high feeding efficiency, equal to the feeding efficiency of the ‘North America and Oceania’ region (as defined by Wirsenius), was chosen, because of the rapid global technology transfer in these scenarios. For the regional scenarios regional feeding efficiencies were chosen, based on data


92

availability and prospects for growth; regions with the higher feeding efficiency were chosen because of potential improvements in the next decades. The feeding efficiencies can be found in Appendix 8. For the definition of the linkage between productivity and irrigation, and definition of irrigated area, the scenarios by the International Water Management Institute were taken as a guideline [De Fraiture, 2007; De Fraiture, 2010]. For the countries for which irrigation cropping patterns are available (mostly developing countries), 75% of the area under irrigation is cultivated for cereals. Another 8.5% is cultivated for oil crops, while the other commodity groups each account for only a couple percent. De Fraiture, however, states that 9% of the irrigated area is accounted for by cotton (which was included in the estimate for oil crops above, which may thus be underestimated). For the four scenarios it was assumed that 75% of the irrigated area is used for cereal production. Because of lack of data, irrigated area and irrigation efficiency are not accounted for in the 2005 situation. Separate yields for rainfed and irrigated cultivation are only given for cereals, as most irrigation takes place on cereal crops. Some data about the maximum attainable yield for some crops, for both irrigated agriculture and rainfed agriculture, is available. However, the data on current and historic yields are an average of rainfed and irrigated yields, and data on irrigated area and rainfed area per crop, on a country level, is not available. Potential yields in the year 2050 were estimated using the methodology based on closure of the yield gap, also used by De Fraiture to estimate potential cereal yields [De Fraiture, 2010]. The maximum attainable yield for the different commodity groups in the different regions was estimated based on data from the FAO and IIASA [Fisher, 2002]. In the scenarios with a high or rapid economic and technological development (A1 and B1), 80% of the yield gap was closed. In the low development scenario (A2) 20% of the yield gap was closed. In the B2 scenario regional yields were defined as being half the regional yields in the A2 scenarios, which results in low-input agriculture as fertilizer input is defined on a per generated output basis. Appendix 7 gives an elaborate explanation of the methodology and the results related to yields. Table 36: Supply-side assumptions and rationale

Scenario Topic All

Losses


Quantification

Source and rationale a

Use of foodstuff for seed, waste , processing and other utilities is set at a fixed rate of production in all scenarios in all regions. The fraction of the total production destined for food and feed: Cereals 0.827 Fruit 0.817 Pulses 0.867 Roots and tubers 0.804 Vegetables 0.911 Oil crops 0.595 Sugar crops 0.828 Eggs 0.886 Meat 0.991 Milk 0.834

Absolute use of foodstuff for these purposes will increase with population, it is reasonable to assume a consistent relative relation. While the FAO estimates that waste during storage, processing and transportation is higher in developing countries due to warmer and more humid climates, waste is also reported to increase with increasing level of development. Fractions are based on global averages in the year 2005 [FAOSTAT, 2010].

93

2005

Processing of oil crops

Vegetable oils are extracted from oil crops. The extraction rate is fixed at the same value for 2005 and for all regions and all scenarios.

Processing of sugar crops

Sugar and sweeteners are extracted from sugar crops. The extraction rate is fixed at the same value for 2005 and for all regions and all scenarios.

Irrigation and multiple cropping

Of the area under irrigation, 75% is cultivated for cereals.

Irrigated and rain-fed yields.

Only for the commodity group ‘cereals’ a distinction is made between rain-fed and irrigated yields.

Productivity

Global average yields for 2005, calculated by dividing total production over total harvested area. Yields for oil crops and sugar crops were converted using extraction rates given in Appendix 9.

Irrigated area and irrigation efficiency Water use

Irrigated area and irrigation efficiency was not taken into account in the modeling of the 2005 situation. 3 Average water use (m /ton):

Fertilizer use and efficiency

A1

Productivity


Cereals 1,571 Fruits 844 Oil crops 2,209 Pulses 3,790 Roots and tubers 375 Sugar crops 165 Vegetables 264 Fertilizer use for 2005 was modeled using global average baseline c requirements .

Commodity group yields (ton/ha):

The extraction rate is based on the production of vegetable oils and the processed amount of oil crops, in the year b 2005 . [FAOSTAT, 2010] The extractions rate is based on the production of sugar and sweeteners and b the processed amount of sugar crops . [FAOSTAT, 2010] [Aquastat, 2011], [De Fraiture, 2010] Figure is based on data on countries for which cropping patterns are available (mostly developing countries) [AQUASTAT, 2011]. Cereal yields are well documented, and are divided into irrigated and rain-fed yields. Such a division is usually not made for the other commodity groups. [FAOSTAT, 2010] The FAO reports specific oil crop yields (e.g. soybeans) and sugar crop yields (e.g. sugar cane) per country in unconverted values, but the average in converted values, which thus need to be reconverted.

[based on Hoekstra, 2008] Global average crop evapotranspiration, elaborated on in Appendix 11.

Fertilizer requirements based on [FERTISTAT, 2011] and [FAO, 1984]. Estimated requirements are compared to aggregations of regional average, and to the total global use as estimated by the FAO, see Appendix 14 and Section 3.3.3. Yields are based on bridging the yield gap by 80% [De Fraiture, 2010], methodology in Appendix 7.

94

Fruit 19.01 Oil crops 4.48 Pulses 3.13 Sugar crops 69.18 Roots 37.37 Vegetables 27.06 Cereal yields based on “Optimistic scenario (2050)”. Global average cereal yield: Rainfed 3.88 ton/ha Irrigated 5.74 ton/ha Feeding efficiency based on the ‘North America and Oceania’ region as defined by Wirsenius.

Irrigated area and irrigation efficiency

Water use


A2

Productivity


Permanent pasture are managed intensively; yields are equal to those of the pasture yield in Western Europe (3.2 ton DM /ha). Harvested-conserved grass-legume is also assumed to be harvested from high-intensively managed land, giving the same yield as mentioned for permanent pasture. With an emphasis on area expansion, irrigated area will increase to 454 million hectare worldwide; the area as projected by the AREA-scenario. Current irrigation efficiency of 60% is maintained. Equal to ‘water use 2005’ – global 3 average use in m /ton of generated product. Fertilizer use is assumed linearly correlated to yields. Requirements are set 10% lower than baseline c requirements (kg NPK/kg foodstuff) . Commodity group yields (ton/ha), range indicates difference between regions (regionally specified data can be found in Appendix 7): Fruit 6.18-16.07 Oil crops 1.86-2.83

[De Fraiture, 2010]

[Wirsenius, 2000] Rapid technology diffusion and economic growth leads to adoption of feeding efficiencies as they were defined by Wirsenius for the ‘North America and Oceania’ region. Based on [Wirsenius, 2000] and [FAO 1984].

[De Fraiture, 2010]

[De Fraiture, 2010] Emphasis on area expansion rather than efficiency. Global average crop evapotranspiration [based on Hoekstra, 2008]. Fertilizer requirements based on [FERTISTAT, 2011] and [FAO, 1984]. Higher fertilizer efficiency due to rapid global technological development. Yields are based on bridging the yield gap by 20% [De Fraiture, 2010], methodology in Appendix 7.

95

Pulses 1.27-2.23 Sugar crops 35.52-70.62 Roots 17.51-37.45 Vegetables 18.15-30.64 Cereal yields based on “Pessimistic scenario (2050)”. Regional average e cereal yields (respectively rainfed and irrigated): 5.2 ton/ha 7.2 ton/ha REF 2.5 ton/ha 4.2 ton/ha ASIA 2.6 ton/ha 4.9 ton/ha ALM 2.25 ton/ha 4.3 ton/ha Feeding efficiency based on the following regions as defined by Wirsenius: OECD90 ‘North America and Oceania’ ‘East Europe’ REF ‘East Asia’ ‘Latin America and ASIA Caribbean ALM Permanent pasture yields are equal to the global average pasture yield (1.6 ton DM/ha). Cropland pasture is managed intensively and has a higher yield (3.2 ton DM/ha), as does harvested-conserved grass-legume.

[De Fraiture, 2010]

OECD90

Irrigation area and irrigation efficiency

Water use


Current irrigation efficiency of 60% is maintained. With an emphasis on area expansion, irrigated area will increase to 453 million hectare worldwide; the area as projected by the AREA-scenario. Regional irrigated area: OECD90 50 million ha REF 37 million ha ASIA 304 million ha ALM 62 million ha 3 Water use (m /ton), range indicates difference between regions (regionally

[Wirsenius, 2000]

Feeding efficiencies lag behind compared to the global scenarios. Regional efficiencies as defined by Wirsenius were chosen based on data availability (i.e. no pork production takes place in the North Africa & West Asia region) and growth prospects. Based on [Wirsenius, 2000] and [FAO, 1984] and [FERTISTAT 2011]

[De Fraiture, 2010] Emphasis on area expansion rather than efficiency. [De Fraiture, 2010]

Region specific crop evapotranspiration [based on Hoekstra, 2008]. Elaborated on

96

B1

specified data can be found in Appendix 11): Cereals 1,099-2,069 Fruits 844 Oil crops 2,002-2,226 Pulses 3,790 Roots and tubers 375 Sugar crops 122-165 Vegetables 264

in Appendix 11.


Fertilizer use is assumed linearly correlated to yields. Requirements are equal to baseline requirements (kg c NPK/kg foodstuff) .

Fertilizer requirements based [FERTISTAT, 2011] and [FAO, 1984].

Productivity

Commodity group yields (ton/ha): Fruit 19.01 Oil crops 4.48 Pulses 3.13 Sugar crops 69.18 Roots 37.37 Vegetables 27.06 Cereal yields based on “Trade scenario (2050)”. Global average cereal yield:

Yields are based on bridging the yield gap by 80% [De Fraiture, 2010], methodology in Appendix 7.

Rainfed 3.9 ton/ha Irrigated 4.94 ton/ha Feeding efficiency based on the ‘North America and Oceania’ region as defined by Wirsenius.

Pasture yields are equal to the global average pasture yield (1.6 ton DM/ha), as are yields for harvested-conserved grass-legume. There is no intensive management of cropland pasture, for which the yields are thus chosen to correspond to those of permanent pasture. Irrigated area and irrigation efficiency

Water use


on

[De Fraiture, 2010]

[Wirsenius, 2000] Rapid technology diffusion and economic growth leads to adoption of feeding efficiencies as they were defined for the North America and Oceania region. Based on [Wirsenius, 2000].

Irrigated will expand in line with the YIELD-scenario; irrigated area will increase to 363 million hectare worldwide in 2050. Irrigation efficiency increases to 65%.

[De Fraiture, 2010]

Equal to ‘water use 2005’ – global

Global average crop evapotranspiration

[De Fraiture, 2010] Emphasis on optimal strategies and efficient use of resources.

97

3

B2

average use in m /ton of generated product.

[based on Hoekstra, 2008].


Fertilizer use is assumed linearly correlated to yields. Fertilizer use is c 15% lower than the baseline .

Fertilizer requirements are based on [FERTISTAT, 2011] and [FAO, 1984]. Higher fertilizer efficiency is due to increased environmental consciousness resulting in better agricultural practices.

Productivity

Commodity group yields (ton/ha):

Fertilizer use rates are related to the yields. Reducing fertilizer input means reducing yields; regional B2 yields are half of those in A2.

Cereals, rainfed Cereals, irrigated Fruit Oil crops Pulses Sugar crops Roots Vegetables Feeding efficiency following regions Wirsenius: OECD90

Irrigated area and irrigation efficiency

1.125-2.6 2.1-3.6 3.09-8.035 0.93-1.06 0.635-1.115 17.76-35.31 8.755-18.725 9.045-15.32 based on the as defined by

‘North America and Oceania’ REF ‘East Europe’ ASIA ‘East Asia’ ALM ‘Latin America and Caribbean Pasture yields are equal to the global average pasture yield (1.6 ton DM/ha), as are yields for harvested-conserved grass-legume. There is no intensive management of cropland pasture, for which the yields are thus chosen to correspond to those of permanent pasture. Irrigated area will increase to 363 million hectare worldwide in 2050. Regional irrigated area (million hectares): OECD90 REF ASIA ALM

Based on [Wirsenius, 2000].

[De Fraiture, 2010]

45 34 237 47

Irrigation efficiency increases to 65%.


[Wirsenius, 2000] Feeding efficiencies lag behind compared to the global scenarios. Regional efficiencies as defined by Wirsenius were chosen based on data availability (i.e. no pork production takes place in the North Africa & West Asia region) and growth prospects.

[De Fraiture, 2010] Emphasis on optimal strategies and efficient use of resources.

98

Water use

Equal to regional average water use in A2.

Regional average crop evapotranspiration [based on Hoekstra, 2008].


Yields are assumed linearly correlated to fertilizer use. Use levels are set to 85% of the baseline requirements (kg N/kg foodstuff), yields are half of those c in the year 2005 .

Fertilizer requirements are based on [FERTISTAT, 2011] and [FAO, 1984]. Fertilizer efficiency increases with technological development and environmental consciousness.

a

Waste includes waste during storage, processing and transport [FAO, 2010]. Appendix 9 shows flowschemes for oil crops and sugar crops . c Appendix 14 discusses fertilizer requirements. b

6.3

VRC Factors

The previous sections have shown all assumptions made related to supply and demand. It is, however, important to distinguish between the scenario study and the Virtual Resource Content methodology. The Virtual Resource Content factors and the basic structure of the model, which were developed for evaluating the scenarios, can be used for purposes other than the quantification of the scenarios defined in this study. The characteristics of the Virtual Resource Content factors are elaborated on in Table 37. They are specified per region, and variations are made per scenario. For the Virtual Land Content factors these variations are incorporated in the factors – these represent the yield improvements. For the Virtual Water Content factors and the Virtual Fertilizer Content factors these are defined separately to increase transparency in the model; they represent the irrigation efficiency and the fertilizer efficiency. Table 37: VRC factors and their specifications.

VRC factor

Unit

Virtual Land Content

m /kg (or yield in ton/ha)

Virtual Water Content

m /ton


Specified for

Spatial and temporal specifications and linkage to scenario characteristics

2

Commodity groups: irrigated cereals, rainfed cereals, fruit, oil crops, pulses, roots and tubers, sugar crops, vegetables. Feed crops: whole maize, pasture, cropland pasture, harvested-conserved grass legume.

3

Commodity groups: irrigated cereals, rainfed cereals, fruit, oil crops, pulses, roots and tubers, sugar crops, vegetables. Feed crops: whole maize.

Global  2005  2050-A1/B1: 80% closure of yield gap Regional  2050-A2: 20% closure of yield gap  2050-B2: Half of A2 yields – between 51% and 102% of yields in 2005. Regionally specified  2005: basis  2050: basis plus incorporation of irrigation efficiency (60%-65%) for cereals.

99

Virtual Fertilizer Content

kg N/kg kg K2O/kg kg P2O5/kg

Commodity groups: cereals, fruit, oil crops, pulses, roots and tubers, sugar crops, vegetables. Feed crops: whole maize, fodder (other feedstuffs).

Regionally specified  2005: basis  2050: basis plus incorporation of fertilizer efficiency (reduction of requirements of between 0% and 15%) for all commodity groups.

The following figures show examples of the three different Virtual Resource Contents. Figure 36 shows the Virtual Land Contents (in m2/kg of generated product) for rainfed and irrigated cereals. These Virtual Land Content-factors are specified to scenario characteristics; the partial closure of the yield gap differs per scenario and is incorporated in the Virtual Land Content factors, per commodity group. Appendix 7 gives more information concerning Virtual Land Content. Yield projections (inverse of Virtual Land Content factors) and how these were derived is elaborated on for all commodity groups and scenarios there. The Virtual Water Content factors are defined per commodity and per region, and are averaged to obtain a global factor per commodity. The factors are given in m3/ton of generated product. Figure 37 shows the different values for cereals as an example: one global factor and four regionalized factors. These represent requirements: irrigation efficiency is not incorporated in these factors. Appendix 11 gives more information on the Virtual Water Content factors. Values are given per region for all commodity groups.

Figure 36: Example of Virtual Land Content - cereals in the A1 scenario and the A2 scenario.


100

Figure 37: Example of Virtual Water Content - cereals in the four regions.

Virtual Fertilizer Content factors are defined for the three macronutrients and given per commodity (e.g. kg N/kg generated product). Figure 38 shows the global average Virtual Fertilizer Contents for all commodity groups and for the three different macronutrients. Such values were calculated for all four regions. Appendix 14 elaborates on the methodology behind the Virtual Fertilizer Content, and the way in which such values were derived for all commodity groups and regions.

Figure 38: Example of Virtual Fertilizer Content - all commodity groups and all three fertilizer, global figures.


101

6.4

Modeling Protocol

The way in which the assumptions, elaborated on in the previous sections, are put together to provide the VRC model is explained below, by explaining the key concepts. Regional and Global Scenarios While A1 and B1 scenarios are global, the A2 and B2 scenarios are regional and thus all assumptions are quantified on a regional level. This provides the opportunity to compare regions, evaluate self-sufficiency and gain deeper insight into the importance of driving forces. Sugar and Oil Demand for sugar and sweeteners and vegetable oils is converted to demand for sugar crops and oil crops and added to those commodity groups. It is assumed sugar crops and oil crops are primarily produced for their sugar and oil content. The by-products may be used as feed, and thus do have economic value. They are, however, not allocated to; production is solely allocated to food production. Feed Requirements Feed requirements are calculated for the 5 animal products separately, and subsequently aggregated per feed type. Those feed types that are also food crops are added to the total demand for food, so that fertilizer requirements and water requirements are taken into account for such feed production of foodcrops. Demand for by-products (edible-type, conversion-type, and non-eaten food) is checked against total by-product production; 50% of by-product production is considered suitable for feed. If availability is too low, the missing portion is substituted by the feed-type most appropriate for the specific animal. Calculation of by-product production is done over the total production, i.e. food, feed and production for other purposes. Irrigated Area and Irrigation Efficiency For all scenarios and regions it is assumed that 75% of the irrigated area is under cultivation for cereals. Total production, given the higher irrigated cereal yield, on these areas is subtracted from the total demand, yielding the production on non-irrigated areas. Irrigation is only considered for cereals, thus irrigation efficiency is only relevant for irrigated cereal production. Water requirements (in m 3/ton cereals) were divided by the respective irrigation efficiencies: 0.6 for the A scenarios and 0.65 for the B scenarios. Fraction of Production for Other Purposes The fraction of food production for other purposes – i.e. seed, processing, other utilities and waste – was assumed equal to that fraction in the year 2005, on a global scale. After aggregation of food demand and feed demand, every commodity is multiplied by a set value, which is the same for the global scenarios, and for each region in the regional scenarios.


102

7.

Results

In this chapter the results regarding the resource use in the four scenarios are presented. Section 7.1 will elaborate on the scenario results concerning production and consumption in the year 2050. Sections 7.2, 7.3 and 7.4 respectively discuss land use, water use and fertilizer use. The results for the four scenarios are compared to the situation in the year 2005. These results were modeled in the same way as the resource uses for the scenarios, using global average data. These data were fitted so that the result matched the data given by the FAO.

7.1

Production and Consumption

Figure 1 shows the apparent consumption and the intake for the year 2005 for the four scenarios in kcal per person per day. As can be seen, apparent consumption exceeds the FAO ‘sufficient nutrition threshold’ of 2900 kcal in the A scenarios and in the B2 scenario. Because of the assumption of higher equity in the B1 scenario, the threshold was set to a lower 2800-kcal limit, which it exceeds.

Figure 39: Apparent consumption (AC) and intake in the four scenarios in 2050, components in the legend are shown bottomup in the bar chart. Lines show level of apparent consumption (solid line) and intake (dashed line) in 2005.

It is clear that animal sources provide a greater part of the total caloric value in the A scenarios than in the B scenarios. In A1 meat consumption accounts for 25% of apparent consumption (and 27%) of the intake in caloric value. In A2, B1 and B2 this is respectively 17%, 8% and 14%. The total production of the 10 main commodity groups is shown in Figure 40. This includes feed, seed, processing, other utilities, waste during storage, transportation and processing. In these charts feed is included only where it concerns food crops that are grown for feed; cereals, pulses, roots and tubers and vegetables are all fed M.Sc. Thesis I.Y.R. Odegard

103

to animals. While the population is lowest in the A1 and B1 scenarios, the total production is highest in the A1 scenario. Total production of animal products is a little over 60% higher in A1 than in A2. This is reflected by the higher cereal production in the A1 scenario. The feed mix in the A2 and B2 scenarios includes non-eaten food for pork production. Because availability is too low, it is substituted by vegetables, which accounts for the comparatively high vegetables production in the A2 and B2 scenarios. Sugar crops are not fed directly to animals, but as can be seen, production is much higher in the A scenarios; people in the B scenarios make healthier food choices. This also explains the higher fruits and vegetables production in the B scenarios.

Figure 40: Production in 2005 and 2050, components in the legend are shown bottom-up in the bar chart.

Figure 41 shows the total global production of foodcrops in kilograms per commodity per capita in 2005. This includes other uses, i.e. seed, processing, other utilities, waste during storage, transportation and processing. Furthermore, in the same bar diagram, the production of food and feed, the apparent consumption and the intake (food minus household and retail waste), are shown. Production and consumption of vegetable oils and sugars and sweeteners are included in respectively the categories ‘oil crops’ and ‘sugar crops’. Of the total production, 15.5% is used for other purposes than food and feed. Of the production of foodcrops, 30% is used as feed. Between total production (including animal products) and intake, 44% is lost.


104

Figure 41: Total production, production of food and feed, apparent consumption and intake in 2005 (kg/cap/year), components in the legend are shown bottom-up in the bar chart.

The figures below show the same, for the four scenarios. Production of food and feed is proportionate to the total production. Therefore, relative losses between total production and the portion of the total production destined for food and feed is similar in all scenarios: between 17.9% and 15.5%. Because animal products are included in this calculation, and their losses are a little lower relatively, losses in the B scenarios are at the higher end of that spectrum. As can be seen, the difference between apparent consumption and intake (thus household and retail waste) is much larger in the A1 scenario than in the other scenarios; 28% of caloric supply is wasted, while this is only 14-15% in the other scenarios.


105

Figure 42: Comparison of total production, production for food and feed, apparent consumption and intake (kg/cap/year) for the four scenarios, components in the legend are shown bottom-up in the bar chart.

Production and Consumption – Analysis The figures presented in the previous section represent all assumptions made related to supply and demand in 2050 for the four scenarios. A few issues deserve attention. First of all, total production (aggregated weight) increases in all scenarios. In 2005, 8.3 billion tons of food-crops were produced. This increases by respectively 80%, 73%, 6% and 40% in the A1, A2, B1 and B2 scenarios. Second, incidence of undernourishment is low, because apparent consumption levels exceed the 2800 or 2900 kcal/cap/day threshold, depending on the scenario (elaborated on in Chapter 3). The FAO estimates that in 2050, there will still be 15 countries with average apparent consumption levels of under 2700 kcal, totaling 746 million people. With a low coefficient of variation (CV) of 0.2, this means it should be assumed that 2.5% of these people are undernourished, totaling 18.65 million people. Compared to the estimated 1 billion in 2009, this is a reduction of 98%. The FAO estimates all developing countries (102) to have an average apparent consumption level of 3070 kcal/cap/day in 2050. This is higher than the modeled apparent consumption levels for 2050 in the ASIA region (2909 kcal/cap/day in A2, 2895 kcal/cap/day in B2) and the ALM region (2930 kcal/cap/day in A2, 2838 kcal/cap/day in B2). However, because a small group of food products is not taken into account in this scenario study (e.g. nuts, for full list of boundary conditions, see Chapter 2) and because food consumption may be underestimated for developing countries, it can be assumed that apparent consumption is a little higher (around 100 kcal/cap/day). Incidence of undernourishment can thus be expected to be similarly low as in the FAO projection. What stands out when taking a closer look at apparent consumption and intake, is that global average apparent consumption is highest in A1 – almost 500 kcal per capita per day higher – but average intake is M.Sc. Thesis I.Y.R. Odegard

106

a little lower – a little over 100 kcal/capita/day lower. Especially wastage animal of animal products indirectly leads to higher production, because the associated feed can also be seen as waste. The wasted animal products in A1 are associated with 602 million tons of cereals, 4.9 million tons of pulses, 2.8 million tons of roots and tubers, and 6.8 million tons of oil crops, respectively 11%, 6.3%, 0.5% and 6.5% of the total production. Finally, the most obvious difference between the scenarios are the much higher total production levels in the A1 scenario. Feed is included in the production figures, and without a doubt the higher consumption of animal products causes the high per capita production of foodcrops. In A1 37% of the crop production that is left after losses are taken into account (‘food and feed’), is for feeding purposes. In A2 this is 29%, which is comparable to the situation in 2005, when it was 30%. In the B scenarios these figures are substantially lower; in B2 it is 19% and in B1 only 9%. As was pointed out above, vegetables production is comparatively high in the A2 and B2 scenarios because there is insufficient non-eaten food to feed pigs. Vegetables provide the substitute. Per capita apparent consumption of vegetables in A2 is comparable that in A1, while in B2 it is comparable to that in B1.

7.2

Land Use

Figure 43 shows the total global land use (in harvested area) for the year 2005 and the four scenarios. This is the land needed to produce food, feed and all other uses. The horizontal lines indicate the area suitable for agriculture. As explained in Section 3.3 and Appendix 10, land can be suitable but not available; between 70 and 90% of the land suitable may be available. Furthermore, there are different classes of land; very suitable (VS), suitable (S), moderately suitable (MS) and marginally suitable (mS). The lines in Figure 43 indicate, from bottom to top: 70% of all VS+S land, 70% of all VS+S+MS land, 90% of all VS+S land and 90% of all VS+S+MS land. As can be seen land use in B2 exceeds all those categories. The bars show the harvested area, while the lines shows actual land. This means that if the cropping intensity is higher than 100%, the required land area is lower than the harvested area. For the B2 scenario, however, this is not an option. Low-input agriculture, as is practiced in B2, should be combined with longer fallow periods; lowering the cropping intensity below 100%. Land use in A2 comes very close to the lowest limit. Figure 44 below shows the total land use in the A2 and B2 world, specified for the four regions. The grey line indicates the harvested area in the year 2005. Only for the ASIA region, this value exceeds the lower boundary of all VS+S+MS land. Cropping intensity, is, however quite high in the ASIA region. Assuming an average cropping intensity of 138% on irrigated lands (current estimated average [based on FAO/NRL, 2011]), the actual area under cultivation in 2050 could be 116 million hectares lower. This would still lead to an exceedance of available VS+S land. In the B2 scenario, only VS+S land is considered appropriate for agriculture, however, as can be seen below, all regions lack enough of this resource to fulfill the demand.


107

Figure 43: Arable land and permanent crops, harvested area in 2005 and in the four scenarios. Solid lines show lower and upper limit of VS+S land (relevant for the B scenarios), dotted lines show lower and upper limits of VS+S+MS land (relevant for the A scenarios), components in the legend are shown bottom-up in the bar chart.

Figure 44: Total land use in A2 and B2. Grey line indicates current harvested area. For A2, dotted lines indicate the lower and upper limit of total available area (VS+S+MS land), for B2 dotted lines indicate the lower limit of VS+S land area, components in the legend are shown bottom-up in the bar chart.

The main factors which create the difference between the scenarios are meat consumption, feeding efficiency and yield projections. For example, average apparent consumption in kcal per capita is around 15% higher in A1 than in A2. Even though yields are much higher in A1 than in A2, harvested cropland area per capita is around 17% higher in A1 than in A2. If land used for pastures and non-foodcrop feed M.Sc. Thesis I.Y.R. Odegard

108

(harvested grass legumes and whole cereals) is added, it becomes clear that feeding efficiency is much higher in A1. This is shown in Figure 45 below. While aggregated total production is higher in A1 than in A2, total land use is higher in A2. The black lines between ‘2005’and ‘A12050’ indicate what the FAO estimates as ‘arable land’ (lower line) and ‘permanent meadows and pastures’ in 2005 *FAOSTAT, 2011+. As can be seen, this does not correspond to the area given for 2005. As discussed in Chapter 3, there is a discrepancy between ‘arable land’ and the aggregate of all harvested areas for all commodity groups. This is at least partially due to the fact that cropping intensity is not taken into account. The discrepancy between the modeled area for pasture and the other non-food feedcrops and what the FAO claims is land devoted to permanent pasture could be due to the fact that not all permanent pastures are used as intensively as they could be. Another way of looking at this is that the pasture area in the model is used to the max, while in the real world they are not. Even so, Figure 45 shows that pasture area increase significantly in A2 and B2.

Figure 45: Total land use in 2005 and for the four scenarios in 2050. The black lines indicate 'arable land' (1.534 billion hectares) and 'permanent meadows and pastures' (3.382 billion hectares) in 2005 according to the FAO [FAOSTAT, 2011], components in the legend are shown bottom-up in the bar chart.

Land use for food and feed crops in A1, A2 and B2 is higher than in 2005. Total land use, including land use for animal feed such as pasture, is, however, lower in A1 and B1. This also means land use per capita in A1 and B1 is lower than in 2005, shown in Figure 46 below.


109

Figure 46: Total land use per capita in 2005 and for the four scenarios in 2050, components in the legend are shown bottomup in the bar chart.

This shows the effect of higher yields and higher feeding efficiencies. Even though the populations in A1 and B1 are higher than in 2005, the land use per capita is low, as is the total land use. As can be seen, total land use in B1 is about half that of A1, which is due to the lower consumption animal products. Land Use – Analysis Land use for foodcrops (including foodcrop-feed) increases in all scenarios but the B1 scenario. Total land use – including pastures and other non-food feeds – show that a combination of high yields, high feeding efficiencies and intensive management of pastures decreases total land use in A1. Due to a much lower consumption of animal products and higher feeding efficiencies and yields, total land use – including pastures – in B1 is lower than cropland use in 2005 (land use excluding non-food feed). In 2005, the harvested cropland area was 1.28 billion hectares, this increases by 31% in A1, to 1.68 billion hectares, by 50% in A2 to 1.92 billion hectares and by 165% in B2 to 3.39 billion hectares. In B1 harvested cropland area actually decreases by 33%, to 0.86 billion hectares. Land use for pastures and such decreases in the A1 and B1 scenarios, by respectively 41% and 85%, from 2.3 billion hectares in 2005 to respectively 1.3 billion hectares and 0.3 billion hectares. In the A2 and B2 scenarios land use for pastures and such increase by respectively 93% and 132%, to 4.4 billion hectares and 5.3 billion hectares. When focusing on the four regions in the A2 and B2 scenarios, it shows that land use increases in all regions for both scenarios. In A2, the ASIA region will not be able to be self-sufficient, and the REF region comes very close reaching the lower limit of available suitable land. In B2, because of policy restraints, agriculture is restricted to VS+S land. However, the demand exceeds the available land area in all regions.


110

7.3

Water Use

Figure 47 shows the total global water use for the year 2005 and the four scenarios. Direct water use by animals (i.e. drinking water) is insignificant compared to the water needed to grow crops, it adds up to around 0.1% of total water use. Water needed to grow food-crop feed (including whole maize) and other uses is included. The lines in Figure 47 indicate the global threshold for moderate water stress (lower line) and critical stress (upper line), based on FAO criteria (see Chapter 3 and Appendix 11). The fact that water use in the A1 and A2 scenarios already exceed the moderate water stress threshold, by respectively 43% and 18%, and that the B2 scenarios comes quite close, indicates that many countries will have water problems in all scenarios. From an estimated value of 6,010 km3 in 2005, the water use increases to 15,091 km1 in A1, an increase of 151%. In A2 it also increases substantially, to 12,497 km3, a rise of 108%. Water use in the B scenarios are significantly lower, 8,682 km3 in B1, an increase of 44%, and 9,783 km3 in B2, an increase of 63%.

Figure 47: Total water use in 2005 and 2050. Lines indicate global thresholds for moderate water stress (lower line) and critical water stress (upper line), components in the legend are shown bottom-up in the bar chart.

Furthermore, comparing global water use to global average water stress thresholds may underestimate regional issues. Figure 48 below shows the regional situations for the A2 and B2 scenario (note that the values on the y axis differ). As can be seen, the ASIA region exceeds the moderate water stress threshold in the A2 scenario, by 84%, and is only 8% short of the critical water stress threshold. The OECD90 region, the REF region and the ALM region come quite close to the moderate water stress threshold; they are respectively 6%, 0.5% and 5% short of this threshold.


111

Figure 48: Regional water use in A2 and B2, components in the legend are shown bottom-up in the bar chart. Lines indicate the moderate water stress and critical water stress levels (the latter is not shown for B2).

In B2, water use in the ASIA region also exceeds the moderate water stress level, by 47%, and it is 26% short of the critical water stress limit. Water use in the other regions are below the moderate water stress thresholds, respectively 30%, 38% and 25% for the OECD90 region, the REF region and the ALM region. Water Use – Analysis The water use was modeled at 6,019 km3 per year in 2005, which is somewhat lower than the estimate made by the Water Footprint Network: 6,189 km3 per year [Hoekstra, 2008]. The International Water Management Institute estimates water use in agriculture to be even higher: 7,130 km3 per year [De Fraiture, 2007]. Irrigation was not taken into account in modeling the 2005 situation, which means that water use is slightly underestimated because irrigation efficiency is around 60%. As can be seen in Figure 47, in all scenarios the bulk of water use is required for cereals production. The slightly higher irrigation efficiency, a raise from 60% to 65%, saves respectively 3% and 1.5% of total water use in B1 and B2. Furthermore, because certain commodity groups, i.e. stimulants, nuts and spices (totaling only 5.5% of global water use in agriculture in 2000 [Hoekstra, 2008]), are not taken into account, a lower 2005 estimate is reasonable. As Figure 47 shows, high feeding efficiencies come at a price: high water use in agriculture. As it is, A1 and A2 both already exceed the moderate water stress threshold, by respectively 43% and 18%. Their water uses are respectively 15,091 km1 in A1 and 12,497 km3 in A2, which represent increases of respectively 151% and 108%. Water uses in the B scenarios are significantly lower, 8,682 km3 in B1, an increase of 44%, and 9,783 km3 in B2, an increase of 63%.


112

7.4

Fertilizer Use

As was shown in Chapter 3, actual fertilizer use as estimated by the FAO, requirements based on recommendations by the FAO and an aggregation of average regional use differ. The fertilizer use was modeled using the use levels based on estimated requirements, combined to higher fertilizer efficiencies in the A1, B1 and B2 scenarios. Figure 49 shows nitrogenous fertilizer use levels for the year 2005 and for the four scenarios in 2050. Two assessments of fertilizer use in 2005 are given. ‘2005-FAO’ is the total amount of N-fertilizer used in 2005, as estimated by the FAO. This is given as an aggregated value, and is not available per commodity group (it is shown in dark blue, but includes all commodity groups, not just irrigated cereals). It includes fertilizer used on intensively managed pastures. ‘2005’ is the modeled use, and as can be seen is a little higher: 13%. Fertilizer use for the four scenarios was modeled using the estimated recommended use rates based on requirements.

Figure 49: Nitrogen fertilizer use in 2005 and 2050. '2005-FAO' indicates actual use as estimated by the FAO, while '2005' indicates the use rate modeled with requirement levels.

Figure 49, Figure 50 and Figure 51 respectively show N, P2O5 and K2O fertilizer use for the total production, including feed and other uses such as seed and other utilities. Fertilizer use follows production; it is defined as fertilizer input per generated output. The reason the increases in fertilizer use are not equal to the increase in total production (in tons produced) is because the composition of the total production changes. Table 38 shows the increase in the total production, and the increases in fertilizer use compared to the situation in 2005. The comparison is made based on the recommended use rates as shown in the figures under ‘2005’.


113

Table 38: Increase in total production compared to increases in fertilizer use.

Scenario

A1 A2 B1 B2

Increase in total production of crops (increase in % since 2005)

Increase in total fertilizer use (increase in %, compared to ‘2005’ – estimated requirements) N P K

80

408.5

277.3

455.4

73

127.0

103.3

46.6

6

-2.5

4.8

4.7

40

19.2

24.4

-18.8

While fertilizer use does follow production, increased fertilizer efficiency can lower fertilizer use while maintaining yields. Fertilizer use on irrigated cereals in A1 and A2 seem equal, and indeed their use rates are very close. However, production of irrigated cereals is 15% higher in A1 than in A2, and fertilizer efficiency is 10% higher. B1 also provides an interesting example. While total production in B1 is 6% higher than it was in 2005, N-fertilizer consumption was 2.5% lower, while use of K2O and P2O5 fertilizers is respectively 4.7% and 4.8% higher.

Figure 50: Phosphorous fertilizer use in 2005 and 2050. ‘2005--FAO' indicates actual use as estimated by the FAO, while '2005' indicates the use rate modeled with requirement levels.


114

Figure 51: Potassium fertilizer use in 2005 and 2050. ‘2005--FAO' indicates actual use as estimated by the FAO, while '2005' indicates the use rate modeled with requirement levels.

Fertilizer Use – Analysis The most striking result concerning fertilizer use are the high use rates in A1. Use of fertilizers to produce fodder (animal feed) is much higher in the A1 scenario than in the other scenarios. The reason for this is that permanent pastures, as well as cropland pastures and harvested-conserved grass-legumes, are intensively managed. These three categories are included in the ‘fodder’ category in the fertilizer use figures. It illustrates a common topic in sustainability issues; a trade-off is made land use and water and fertilizer use. Land use is relatively low in A1, but water use and fertilizer use exceed the limits of reason. Fertilizer use for fodder is much lower in A2, even though total meat is also high, however, pastures are not intensively managed in A2, which is reflected in the much higher land use for pastures in A2. There are a number of other interesting aspects. Fertilizer use on irrigated area is similar in the A scenarios; this is a coincidence. While irrigated area is the same in both regions, yields are higher in A1 and thus is the associated fertilizer use. Furthermore, there is reason to believe, as described in Section 3.3, that the current NPK ratio of application is too much in favor of N and P fertilizer, while K2O fertilizer is underused. This explains the gap between what the FAO estimates as use (‘2005-FAO’) and what was estimated as recommended use (‘2005’). Because K2O and P2O5 fertilizers are non-renewable resources, thresholds of maximum use cannot be given, as they were for land and water. However, the use rates can be checked against reserve bases to give an indication whether the use rates exceed reasonable limits, which is shown in Table 39.


115

Table 39: Fertilizer use and reserve base [based on IFA, 2011; USGA, 2009].

A1 A2 B1 B2

K2O P2O5 K2O P2O5 K2O P2O5 K2O P2O5

Average Use (between 2005 and 2050 - modeled) (‘000 ton)

Use in 2050

Reserves left

(‘000 ton)

(in years at 2050 use-level)

238873

449580

2

113622

177340

55

73413

118660

66

72733

95563

121

56457

84748

100

49572

49241

254

46988

65811

135

54238

58572

211

Note: Average use rates between 2050 and 2005 were based on the average between use in 2005 and modeled use in 2050. Use in 2005 was taken as the average of production in the years 1999-2009. Assessment of the reserve base includes reserves that are currently uneconomical to mine, but that may be in the future.

Table 39 may give an explanation for the lower than optimal application rates of K2O. The K2O reserves will be depleted by the year 2051 at the levels applied in A1. In A2 too resources are depleting fast, there will only be 67 years’ worth of K2O resources left. The B scenarios show more promising figures; in 2050 there will be over 100 years left for K2O and over 200 years for P2O5.


116

8.

Conclusions

The aim of this study was to design four food scenarios for the year 2050 and to evaluate these quantitatively with respect to their use of the natural resources land, water and fertilizers. To achieve this, a model was designed which would enable the evaluation of natural resource use with the Virtual Resource Content concept, for the different scenarios and for the seven main vegetal food commodity groups and 5 main animal food commodity groups. In the previous chapter the land use, the water use and the fertilizer use have been presented and analyzed. In this chapter an answer will be given to the main research question: What are the regional and global consequences with respect to the use of natural resources – concerning land, water and fertilizers – for four food scenarios evaluated for the year 2050?

8.1

Scenario Conclusions

Figure 52 shows the main characteristics of the four scenarios. For the A2 and B2 scenarios the results were regionally and globally specified, while results were given on a global scale for the A1 and B1 scenarios. For the regional scenarios self-sufficiency for the four regions was evaluated, while for the global scenarios a global average diet was modeled and evaluated. Each scenario has its own story to tell with respect to its use of the natural resources land, water and fertilizers. These stories will be given in the next sections.

Economic Unrestrained Use of Natural Resources

A1 The Affluent World

A2 The Full World

Low Population Growth High Economic Growth Rapid Spread of Agro-Technology Medium Irrigation Efficiency High Fertilizer Efficiency High Productivity Western Diet

High Population Growth Low Economic Growth Slow Spread of Agro-Technology Medium Fertilizer Efficiency Low Irrigation Efficiency Medium Productivity Western Diet

Globalization

Regionalization

Global Food Distribution

Regional Food Distribution

B1 The Vegetarian World

B2 The Low-Input World

Low Population Growth Medium-High Economic Growth Spread of Sustainable Agro-technologiy High Irrigation Efficiency High Fertilizer Efficiency High Productivity Vegetarian Diet

Medium Population Growth Low-Medium Economic Growth Spread of Sustainable Agro-technology High Irrigation Efficiency High Fertilizer Efficiency Low Productivity Organic Diet

Environment Efficient Use of Natural Resources

Figure 52: Scenario characteristics


117

8.1.1

A1 – The Affluent World

The question central to the A1 scenario – The Affluent World – is: What will the effects be given a worldwide shift to Western agricultural management practices and a Western Diet? Production of foodcrops increases by 80% (weight basis) in the Affluent World. This is due to the high per capita apparent consumption of meat: 93 kg per capita per year. This is higher than the current apparent consumption in the Netherlands ~73.8 kg in 2005. It is, however, much lower than the level in the USA 123.6 kg. Another factor is the relatively high losses and wastes in A1; 28% of total caloric supply is wasted in retail and households. The losses for other purposes than food and feed are proportionate to the total production, and thus high in the Affluent World because production of feed is relatively high. The FAO undernutrition threshold of 2900 kcal/cap/day for apparent consumption is passed. However, because household and retail waste is high, intake is estimated at a lower level than for the Full World, even though levels of apparent consumption in the Full World are lower. Not only meat consumption follows a ‘western’ trend; feeding efficiency is high and so are yields. The Affluent World clearly illustrates the trade-off that can be made between the different resource uses. Land use for crop production increases 31%, even though corresponding production increases 80%, while land use for pastures and such decreases 41%. This has to be compensated elsewhere, and water use increases by 151%, and exceeds the global average threshold for moderate water stress by 43%. Furthermore, fertilizer use – including pastures and such - increases between 277% and 455%. Especially these last figures show the effects of intensive management of pastures, for example, of the total increase in use of N fertilizer, 59% is used for intensive management of pasture lands. In light of the reserve base, this is clearly an undesirable and uneconomical situation, as only 1 year (K2O) and 52 years (P2O5) of resources remain in 2050. The high fertilizer use and high water use in the Affluent World illustrate a common topic in sustainability issues; a solution may seem sustainable from one angle – land use is low in the Affluent World – but unachievable and impractical from another – fertilizer use and water use are extremely high. 8.1.2

A2 – The Full World

The main question in the A2 scenario – The Full World – is: What will the effects be when regions have to be self-sufficient in a world where population growth was high and economic development low between now and the year 2050? Because of the low economic and technological development, yields increase at a much slower rate in the Full World scenario; only 20% of the yield gap is closed. Meat demand increases along with PPP, and thus increases significantly; 60% on a global level relative to 2005, increasing from 39 kg/cap/year in 2005 to 62 kg/cap/year in 2050. This is still much lower than the increase in meat consumption in the Full World, which is 140%. With a population that is 27% higher in the Full World than in the Affluent World, total production is only 4% lower in A2, corresponding to an increase of 73% relative to 2005.


118

Because yields and feeding efficiencies only improve a little, land use is high. Land area under cultivation for foodcrops (including those used for feed and other purposes) expands by 50%. Land for pastures and such almost doubles; it increases by 193%. The fact that pastures are less intensively managed is reflected in the lower fertilizer uses; they increase by between 47% and 127%. Still substantial increases, but significantly lower than those in the Affluent World. Because the Full World is regionally specified, regional results deserve attention. The land area potentially suitable for agriculture in the Full World is defined as between 70% and 90% of the total available area of very suitable, suitable and moderately suitable land. The OECD90 region and the ALM region both use little over half of the lower limit, respectively 55% and 56%. The REF region comes quite close to the lower limit, it is only 6% short. The ASIA region exceeds both limits, and thus will not be able to supply its demand, even with higher cropping intensities. Water use increases by 108% in the Full World, globally leading to an exceedance of the moderate water stress threshold by 18% of water a year. Similar to land, use of water in the ASIA region is problematic. This region exceeds its moderate water stress level by 84%, and is only 8% short of the critical water stress level. The other regions come quite close to their respective moderate water stress levels; the OECD90, REF and ALM region are respectively 6%, 0.5% and 5% short of this level. Use of fertilizers is high, but can be sustained for another 67 years for K2O and 115 years for P2O5 if levels are retained at their 2050 value. 8.1.3

B1 – The Vegetarian World

The question central to the B1 scenario – The Vegetarian World – is: What will the effects be of a global shift to a vegetarian diet? Because of higher equity in the B scenarios, the undernutrition threshold was lowered to 2800 kcal/cap/day, and even though total production only increases by 6% this threshold is exceeded. Compared to 2005, the share of crop production for food and feed used as feed drops from 37% to 9%. High yields and feeding efficiencies, combined with low consumption of animal products make the Vegetarian World the scenario with the lowest use of natural resources. The harvested cropland area decreases relative to 2005; it drops to 0.86 billion hectares, a reduction of 33%. Water use increases by 44%, to 8,682 km3. Furthermore, production of fruits and oil crops increases, while production of vegetables and sugar crops is lower than in 2005. The former two have higher water needs per generated ton; respectively an average of 844 m3/ton and 2209 m3/ton, to 264m3/ton and 165 m3/ton for vegetables and sugar crops. Even though water use increases by 44%, it is still 1911 km3 or 18% short of the moderate water threshold. Fertilizer use does not change much relative to 2005; use of nitrogen fertilizer decreases 2.5% relative to 2005, while use of phosphorous and potassium fertilizers increase respectively 4.8% and 4.7%. The Vegetarian World shows a much better situation regarding the remaining reserve base of these fertilizers; use can be sustained for another 103 years for K2O and 246 years for P2O5 if levels are retained at their 2050 value.


119

8.1.4

B2 – The Low-Input World

The B2 scenario – The Low-Input World – revolves around the following question: What will the effects be when regions have to be self-sufficient given low-input agricultural practices? The Low-Input World is characterized by medium population growth, medium economic development and a meat consumption which is similar to the global level in 2005, but more equitable distribution between the regions. The main characteristic, and the most significant and influential, are the lower yields as a results of reduced inputs of fertilizers. Thus while total production of crops increases by 40%, and only 19% of the production for food and feed is dedicated to feed, the global land use increases by 165%. Furthermore, regional land use exceed the defined limits for the Low-Input World; the 70% limit of very suitable and suitable land. Because pasture lands are not intensively managed, this area increases from an estimated 2.3 billion ha in 2005 to 5.3 billion hectares in 2050. Water use increases by 63%, to 9,783 km3, only 809.6 km3 or 7.6% short of the moderate water stress threshold. A closer look at regional values shows that the ASIA region exceeds the threshold, by 47%. The other regions – the OECD90 region, the REF region and the ALM region – are respectively 30%, 38% and 25% short of their moderate water stress thresholds. As expected, fertilizer use increases at much lower rates than it does in the A scenarios. Use nitrogen and phosphorous fertilizer increase with respectively 19.2% and 24.4%, while use of potassium fertilizer decreases by 18.8%. The Low-Input World aims to reduce external inputs in agriculture. It shows a much better situation regarding the remaining reserve base of these fertilizers than the A scenarios; use can be sustained for another 140 years for K2O and 203 years for P2O5 if levels are retained at their 2050 value.

8.2

Resource Use in Agriculture

The previous sections elaborated on conclusions per scenario. Figure 53 below shows the main results regarding land use, water use and fertilizer use in the different scenarios in 2050. The figures can all be found, with full explanation, in Chapter 7. Here they are presented to explain the main conclusions. As can be seen in the figures, land use exceeds availability in the B2 scenario, the Low-Input World. Furthermore, in both A scenarios, the Affluent World and the Full World, the moderate water stress threshold is exceeded. Relatively low land use in the A1 scenarios, the Affluent World, is offset by very high fertilizer use. The only scenario in which the limits to the natural resources are respected is the B1 scenario, the Vegetarian World.


120

Figure 53: Scenario results - land use, water use and fertilizer use in 2050 for the four scenarios.

In any study aiming to give result on a global level, data availability and reliability is an issue. Therefore, when evaluating the results of the scenarios it is important to keep in mind that they are ‘what if?’ scenarios, based on a list of assumptions. If the quantifications to these assumptions would change, naturally, so would the results. It is important to remember that scenarios of this kind are not made to assess what will happen, but what could happen, given such a set of consistent assumptions. The results should also be evaluated in this light. Because economic variables, such as prices, were not taken into account, the effects on resource use were modeled given the assumption that a given demand is fulfilled. It is not reasonable, however, to assume that if this creates an unsustainable situation, such a situation would not occur; such situations occur at present, e.g. countries in the Middle East using more than M.Sc. Thesis I.Y.R. Odegard

121

100% of their renewable water in agriculture. Assuming that situations that seem untenable will not occur is thus too simple a proposition. The results show that trade-off issues are important and need to be addressed when discussing the future of food; it is important to show the impacts of all three major inputs when presenting scenarios. This is substantiated by the results shown in Figure 53, which clearly show that for example low land use in the Affluent World/ A1 scenario is offset by high fertilizer us. The opposite is true for the Low-Input World/ B2 scenario; land use exceeds the limits of availability, while fertilizer use is relatively low. An assessment of resource use is only valuable when a complete picture is given. Therefore, the present study provides valuable input for assessing problem areas, but also for identifying opportunities in our agricultural system. Resource use differs widely between the scenarios. While land use in the Affluent World is low, water use and fertilizer use exceed the limits, which makes it an unsustainable and even improbable situation. This is also true for the Full World, which shows that all regions are close to, or by far exceed (i.e. the ASIA region) the moderate water stress level. As total production in the Affluent World is similar to that of the Full World, it is certain that similar regional figures for water use apply to the Affluent World. The Low-Input World shows the opposite case; while fertilizer inputs may be low, land use exceeds all limits. Only in the Vegetarian World all the resource use is within the limits these resources pose. From a demand-side perspective it can be concluded that current trends in meat consumption cannot continue indefinitely due to lack of natural resources to support such a system. Moreover, a situation such as in the Affluent World were fertilizers run out between 2050 and 2100 would never progress to such a stage because fertilizer prices would have risen too much to justify such use. Or, put differently, such use rates would raise meat prices to a level where they would only be affordable to the very rich, and the global average consumption would not reach the modeled levels. This study clearly illustrates the inequity in the food system: the lifestyles of many in the industrialized world cannot be supported on a global scale. One potential development could be that as food prices increase due to strains on natural resources, people’s motivation to decrease their food waste rises. However, food wastes in households and retail are highest in the industrialized world – exactly where people can afford such wastage. What this scenario study has shown is that it is possible to feed a growing world population a balanced and adequate diet, i.e. a vegetarian diet, while respecting the limitations our natural resources pose. It is, however, unlikely that people are willing to give up meat consumption. This study also points out opportunities in our agricultural system. From a supply-side perspective it can be concluded that technological development is of vital importance. Yield increases reduce land requirements and technological development can improve the efficiencies of irrigation and fertilization. The potential for growth in the developing countries is large, and also quite necessary in view of the population increases that are to be expected. The current low yields in many developing countries reflect the potential for development. Such development is not something which will happen naturally; it will take long-term dedication of such stakeholders as governments, industry, development agencies and of course farmers. Furthermore, the Low-Input World scenario has shown that it is absolutely necessary to increase yields, while the Affluent World scenario shows that even maximizing yields and feeding efficiencies is not enough if a western diet is adopted on a global scale. These results give a clear message to decision makers on all levels of society – government, companies, retail and households. M.Sc. Thesis I.Y.R. Odegard

122

Because the current diet in the industrialized world cannot be sustained on a global scale, people in the industrialized world will have to change their lifestyles if they want to have a more equitable world, or accept the fact that their food choices have large scale consequences. This may be a tenable situation for individuals, as long as they can afford what they consume, however, governments will have to make farreaching decisions if they are serious about reaching global equity and eliminating undernutrition.


123

9.

Discussion

It would be interesting to compare the results presented in this study with other studies. Because, by the authors knowledge, there are no studies that assess the future use of all three resources, the scope for comparison is limited. Other studies do not assess a complete diet, have a limited time frame or only assess a single scenario. For example, an interesting scenario study by Wirsenius was published in August 2010, which explores the effects of increased feed-to-food efficiency, dietary changes, and decreased wastage. Clearly, the issues identified as important correspond to the ones in this study. However, because the scenarios are made for 2030, and include only one population growth trend, comparison could only be done on a very basic level, and is not within the scope of this study. It does show the relevance and popularity of the topic, and – because of the different scope – the importance of further research of food scenarios as done in this study. A general issue encountered in processing data was the fact that different organizations use different definitions of regions. While the IPCC definition is useful from the perspective of economic development, and therefore of diet change, a lot of work went into processing data to fit the IPCC regions. A solution would be to model on a country level, which is outside the scope of this study. The need for processing data to fit the current regions did help in getting acquainted with the data, which helped in understanding results. The PPP data which was used for the calculation of meat consumption projections was defined per sub-region and not per country. Using country specified data will give more detailed results. Other data issues are related to the fact that the FAO was the prime source of agriculture data. First of all, when this study was started, data was not available online. Halfway through, data concerning production and consumption was put online, in Food Balance Sheets, for 184 countries, although data was not available for every one of those countries. Furthermore, even though the FAO provides the most complete data, their data can be very intransparant and the FAO did not respond to numerous emails asking for clarification of different issues. Cropping intensity deserves clarification in the FAO databases, as does fertilizer use per crop (although both may be difficult to monitor/assess/calculate/balance), total agricultural area and fallow land, and the way in which calories are calculated from the food supply. Furthermore, as data is still being added, the different parts of the database do not always match. For instance, even though Singapore has a very small agricultural output, which is reported under Production, it is not even listed as a country under Commodity Balances – which is where data is given on e.g. production, but also import, export, domestic supply quantity and food. Other such instances have been encountered. For example, data for a number of countries is not given at all. Data for these countries seems to have been set at zero before (in the FBSs), however, this could be confusing as it was not clear whether the data entry actually was zero or unknown. Now these countries seem to be eliminated from the list, which seems odd too, as they are quite large; e.g. Afghanistan and Iraq. Also, concerning oil crops and sugar crops, the FAO states that ‘conversion factors are applied to values when calculating totals’, which total and which conversion factors they mean is not specified. The item ‘metadata’ to which is referred for further information does not contain any information on these issues. This results in issues like the following. Production of oil crops (Oilcrops + total) under Commodity Balances, for Brazil in 2005 gives a result of ~57*106 tonnes. Under Production, Oilcrops Primary + (Total) yields a result of M.Sc. Thesis I.Y.R. Odegard

124

~10*106 tonnes. The sum of the produced quantities (also under Production) of individual oil crops is ~62*106 tonnes. In short, the data should be handled carefully, and this study, and studies such as these would benefit from a sensitivity analysis, were the strength of the results are tested by varying the input parameters. For this study, such an analysis carried too far, but it can be kept in mind that a scenario study already gives an idea about the importance of different inputs by incorporating different trends. Furthermore, FAO data was mainly used to provide insight into the current situation and to be able to compare future situations to the current situation. The scenario diets were modeled for the total global population, and compared to total global availability of resources. This way gaps in FAO data were bypassed. The value of the results and the limitations of the model and the data which was used will be elaborated on by discussing a number of relevant issues; assumptions related to fertilizer use, losses, wastes, sugar and oil, cropping intensity and yield projections will be discussed in the following sections. Virtual Fertilizer Content Fertilizer use was modeled using estimates of requirements recommended by the FAO [FAO, 1984]. Aggregated (over the different commodity groups) estimated requirements for N and P2O5 came quite close to the FAO estimate. For K2O, however, estimates of requirements were much higher than the FAO estimate of current use. Because there is reason to believe that the NPK ratio is currently not optimal, with a preference for N and P, Virtual Fertilizer Content was modeled using the requirement estimates. There are several issues regarding the required use rates, which will be discussed here. First, the relation between fertilizer input and generated product was assumed linear; fertilizer use was defined as input of N, P2O5 or K2O per ton of generated product. However, the fertilizer response curve usually approaches a maximum value; yield response to additional input is marginal. Thus, for high yielding crops, total fertilizer needs may be overestimated. To counter this issue to some extent, the use rates were based on recommendations for high-yielding crops; also a valid assumption since the FAO source which was used to estimate requirements was written in 1984, and improvements have been made since. This brings us to the second issue; recommendations may have changed since 1984. Since the fertilizer requirements were established and used to model the Virtual Fertilizer Content, another, more up-to-date FAO document was found in which fertilizer requirements are discussed. It was beyond the scope of this thesis to check whether more recent data would change the applied use rates. A third issue, also connected to the recommended requirements, is the basis on which the rates were established for the commodity groups sugar crops and vegetables. These were based on one of the commodities in each group; sugar cane for sugar crops and tomatoes for vegetables. Sugar cane accounts for 84% of total sugar crop production, and since no recommendations are given for sugar beet, on a global level this is a reasonable assumption. For the OECD90 and REF region, however, with temperate climates and thus a preference of sugar beet over sugar cane, this may not be the case. As requirements for sugar beet are unknown, it is uncertain whether this leads to overestimating or underestimating fertilizer use. However, fertilizer consumption for sugar crops is relatively low, and thus total fertilizer use will not change substantially. For the vegetable commodity group, requirements are only given for tomatoes, cucumbers and onions. Cucumbers and onions account for respectively 4.9% and 7.7% of total vegetable production. Tomatoes are the single largest vegetables product; they account for 14.7% of total production. M.Sc. Thesis I.Y.R. Odegard

125

Recommendations for onions and tomatoes are similar. Because those for tomatoes are a little lower, and tomatoes account for a larger share of total production, those rates provide the most reasonable estimate. Of course, soil characteristics are important for establishing local fertilizer needs. It is assumed that local conditions which would lead to overestimates or underestimates counter each other out. Since assessment of the scenario results, it was found that nitrogenous fertilizer needs were by mistake undervalued by a factor 10. Of course, in future assessments this should be adjusted. Total nitrogen fertilizer use would increase respectively 5%, 11%, 17% and 16% in the A1, A2, B1 and B2 scenarios. Losses There are several issues related to losses and wastes in the food system that deserve to be discussed. First, there is the assumption that the fraction which is used for purposes other than food or feed – seed, waste, other utilities and processing, calculated per commodity group – is a constant relative to the total production. For seed this is a reasonable assumption. The FAO acknowledges that waste – during storage, transport and processing – is estimated as being a fraction of the total production; it is not measured. While some argue [Bender, 1994] that wastes are lower in industrialized countries, it is actually the case that wastes are estimated as being higher in developing countries because these have in general different climates – more humid and hotter. For the use of foodstuffs for processing and other utilities it may be more reasonable to relate production levels to either population or economic development. Processing of food products increases with economic development. This can be reasonably assumed to be the case for other utilities too. The current assumption does lead to the result that for scenarios in which meat consumption, and thus subsequently feed production, is high (A1), but for which population growth is equal and economic development similar to another (B1), food production for other purposes is higher. For the commodity groups oil crops, sugar crops and for sugar and sweeteners and vegetable oils, pretty much all (>94%) of the production for other purposes goes to processing and other utilities. For fruits and vegetables, and inherently for vegetable oils and sugar and sweeteners, as they are derived products, the FAO explicitly does not report seed. Processing accounts for little over half of the total fruit production for other purposes (52%), for which wine production is excluded. The figure for cereals is similar: 56%. For the commodity groups roots and tubers, pulses and vegetables, these figures are much lower, respectively 35%, 10% and 1%. Because of the increased demand for animal products in the A scenarios, use for other purposes per capita goes up. In the B scenarios, however, it goes down, and because of the link to economic development, use for other utilities and processing may be underestimated in those calculations. Also, because foodstuff use for processing and other utilities is correlated to the level of economic development, and global average values are used, the estimates for other uses may be overestimated for developing regions, and underestimated for industrialized regions. The previous issue links to the production of alcoholic beverages. These are the product of the processing of either cereal crops or fruit crops. Because of conversion issues, they are not included in the FAO food supply and commodity balance data for cereals and fruit. It cannot be reported under ‘processing’ because this category is reported after ‘import’ and ‘export’. Production as reported by the FAO for the year 2005 of cereals excluding alcoholic beverages is 10% lower than the reported total production. For fruit this is only 1%, which makes inclusion less relevant. M.Sc. Thesis I.Y.R. Odegard

126

Wastes Occurrence of household and retail waste has been studied by various researchers. However, a study resulting in a clear, and for such purposes as this study useable, correlation between such wastes and economic development and/or other factors, is lacking. Estimates for household and retail wastes in this study were based on estimates made for the year 1995 for the USA [Kantor, 1997], as this source provided the most detailed and clear numbers. These levels of household waste by Kantor were defined as portion of the edible food supply. This implies that specifically for the vegetable commodity groups fruits, vegetables and roots and tubers the household waste is underestimated, as refuse is included in the apparent consumption. Furthermore, animal meat products are reported in slaughter-weight, which is bone-in. Whether inclusion of refuse is also true on a caloric level, is unclear. The FAO did not answer questions about the methodology used to calculate the caloric supply. For example, the FAO estimates the caloric value of 100 grams of apple in the Netherlands to be 50 kcal, whereas the Dutch Nutrition Centre estimates 100 grams of apple to contain either 58 kcal (without peel) or 60 kcal (with peel). For bananas, the FAO estimates 100 gram to contain 70 kcal, while the Nutrition Centre gives a value of 95 kcal. Similar figures apply to roots and tubers. While the FAO estimates 100 grams of potato in the Netherlands to contain 67 kcal, the Nutrition Centre estimates 100 grams of boiled potatoes to be 83 kcal. For the two vegetables for which both organizations give data, tomato and onion, the estimates by the Nutrition Centre are – only slightly – lower than those by the FAO. Thus while the estimates of intake may be overestimated because wastes are given as portion of the edible food supply, the supply or apparent consumption may be underestimated, given that the caloric values for fruits and roots and tubers given by the FAO are relatively low. For determining the resource use in the scenarios, wastes play an additional role because in the A2 and B2 scenarios, in all regions except for the OECD90 region, non-eaten food is fed to pigs. In case of lack of availability, true for both scenarios for both the ALM region and the REF region, it is substituted by vegetables. Sugar and Oil Production of sugar crops and of oil crops is characterized by the fact that their primary form is relatively unimportant in people’s diet, but that their processed products are. This makes it necessary to make assumptions about extraction rates. Because both commodity groups incorporate various different crops, such extraction rates may differ between regions because of a different commodity group composition. For example, oil content of soybean is 18% (weight basis), while it is 43% for groundnuts. Yields are reported by the FAO in converted values for countries and regions, but in unconverted values per crop. There are several reasons why it is necessary to know demand in unconverted values. Oil crops in their primary form are included in the feed-mixes and by-products from vegetable oil production and sugar production are also used in the feed-mixes. Furthermore, sugar crop and oil crop yield projections concerning maximum attainable yields are given in unconverted values. Finally, oil crops and sugar crops are part, although not substantial, of people’s diet in their unconverted form. Current yields of these commodity groups were estimated by dividing total production, in unconverted values, over the harvested area. Maximum attainable yields (MAYs) were based on the MAYs for different crops, M.Sc. Thesis I.Y.R. Odegard

127

proportionate to current regional and global production levels. The underlying assumption is that shares of different crops in the total, both globally and regionally, will not change. Conversion factors were used to convert the amount of vegetable oils and sugar and sweeteners to respectively oil crops and sugar crops. These conversion factors are based on global data for 2005. It was assumed that all processing of oil crops and sugar crops is done to yield vegetable oils and sugar and sweeteners. On a weight basis, only 36% (oil) and 12% (sugar) is extracted. Because it is beyond the scope of this thesis to make regional projections for conversion factor, and furthermore to increase transparency and bypass the trade issues involved in regional calculations, global averages were used. An interesting note about the area under cultivation for oil crops worldwide; two months ago this area was reported as 2.52*108 hectare for the year 2005. Now it is reported as 0.54*108 hectare. This has a significant influence on yields and warrants further investigation. Cropping Intensity The FAO reports yields in tons per harvested area, not in tons per area per year. This makes comparison of yields per harvest possible and easy, however, in the case of multiple cropping, yields per year per hectare may be twice or three times as high. Some data are given relating to cropping intensity. These data are not, however, given for all countries, e.g. the only OECD90 country listed is Turkey. Furthermore, just how the multiple cropping occurs is unclear; which crops are alternated for example. Furthermore, multiple cropping is reported in terms of how much of the irrigated area is cultivated per year and per crop, resulting in a higher than 100% intensity if the aggregate is higher than the area equipped for irrigation. For some countries, the cropping intensity is lower than 100%. It is unclear whether such land equipped for irrigation is not used for cultivation at all, or whether the irrigation systems on those lands are unused. To retain transparency, results relating to land use are given in harvested area, and not in land under cultivation. Irrigation and Yield Projections Irrigated area accounts for around 20% of arable land and permanent crops – between 10% and 12% in the OECD90, REF and ALM region, and around 34% in the ASIA region. Because the difference between current and historic irrigated yields and rainfed yields of crops other than cereals is unclear, a distinction was only made for cereals. However, high yields in the OECD90 region indicate that irrigation may be common for other crops as well, e.g. for sugar crops and roots and tubers. Maximum attainable yields are important for establishing the yield projections, and are higher for irrigated crops than for rainfed. As no historic data were found on the difference between irrigated and rainfed yields for the different commodity groups in the different regions, yield projections were based on the current yields given by the FAO, which include both rainfed and irrigated areas. The maximum attainable yields used in these projections were based on the MAYs for rainfed agriculture. MAYs for irrigated agriculture are higher, but, as was shown, water use already exceeds the moderate water stress limit in the A scenarios, while the B scenarios come quite close to this limit.


128

The assumption that 80% of the yield gap is closed in the A1 and B1 scenarios is a considerable challenge. It is put somewhat in perspective by the use of the lower MAYs for rainfed agriculture. Still, it is generally assumed that a closure of the yield gap by 80% is the highest which can possibly achieved. Furthermore, such high yields may only be reasonably achieved on ‘very suitable’ land, while they were assumed to be attainable across all land in the present scenarios. This results in scenario results which are on the optimistic side. Diet The total caloric intake as determined for the diets in the different scenarios and the different regions approximate the FA projections and therefore give a complete picture with regard to caloric intake. In the A1, B1 and B2 scenarios the total food quantity in kcal/cap/day is slightly lower than the level of the related FAO projection. In the A2 scenario this quantity is slightly higher in the OECD90 region and the REF region, while slightly lower in the ASIA region and the REF region, which corresponds to the difference in definition of regions between this study and the FAO definition. Overall this means that the results as presented in Chapter 7 present a complete picture of resource use. Caloric values of foodstuffs provide a way to add foodstuffs in a uniform way and the ability to evaluate whether it is quantitatively sufficient. Quality cannot, however, be solely evaluated by checking the caloric intake. While health effects and malnutrition of food intake are beyond the scope of this study, some comments will be made here. Variety is important to receive adequate nutrition, which is incorporated by the variety of commodity groups included in the diets. However, such variety does not ensure that people make healthy choices, which is reflected in the fact that over half (50.1%) of the adult population in 15 out of 27 EU countries was overweight and the average obesity rate in the EU was 15.5% in 2008 [OECD, 2010]. Fruit and vegetable consumption is not specified explicitly in the FAO projections concerning future apparent consumption. Here, in the A scenarios, consumption of fruits and vegetables was set at the global average consumption in 2005. This may, however, be too low an estimate. In 2005 the global average consumption was 66 kg/cap/year for fruits and 117 kg/cap/year for vegetables. In comparison, in the EU, these rates were on average respectively 105 kg/cap/year and 116 kg/cap/year. For vegetables, with a slightly lower than global average supply in the EU, 10 countries exceed this level, ranging from 118 kg/cap/year for Belgium to 241 kg/cap/year for Greece [OECD, 2010]. It may therefore be reasonable to assume higher levels of fruits and vegetables consumption. The level of consumption of foodstuff in the miscellaneous category (undefined, but includes fruit and vegetables) increases with 51 kcal/cap/day between the year 2000 and 2050, according to the average global FAO projection for food supply [Alexandratos, 2006]. In the B scenarios, consumption of fruits and vegetables was raised, in line with recommendations from the European Commission to increase the consumption of fruits and vegetables [OECD, 2010]. An important question related to meat consumption and resource use is the type of meat consumed. The proportion between beef, pork and poultry meat was kept constant at the 2005 level. Scenario results would differ if the proportion changes. Cultural and religious beliefs play a role in this; beef and poultry will be preferred in Muslim countries and pork and poultry in India. The FAO projects higher growth rates for poultry meat than for beef and pork. M.Sc. Thesis I.Y.R. Odegard

129

10.

Recommendations

In the previous chapter some methodological issues related to Virtual Fertilizer Content, losses, wastes, sugar and oil, cropping intensity, irrigation and yield projections, and diet were addressed. Some recommendations for further improvement of the model and its assumptions, and suggestions for further research are given here. Virtual Fertilizer Content

The assumptions concerning fertilizer use and the recommended requirements could be checked against more recent values. Furthermore, as fertilizer use on pastures can account for a significant part of total fertilizer use in a country, the relationship between fertilizer use and pasture yields could be more thoroughly studied. The Virtual Fertilizer Content factor of N for oil crops was miscalculated, and should be ten times larger. This would increase the estimates further, by around 5%-17% of total use, for the different scenarios. The Low-Input World scenario (B2) could be extended to present an ‘Organic Agriculture’ scenario. Synthetic input use can be lowered further by applying other, organic, inputs, or by changing agricultural management practices. This topic is too large to fit into the scope of the present research, but deserves attention because organic agriculture has its advantages (as elaborated on in Section 3.3.4). In fact, research is currently being done at Wageningen University concerning yields in organic agriculture at a global scale [De Ponti, 2011].

Losses

Currently alcoholic beverages are not included in the model or its results. When this study was started, only information on cereal production excluding alcohol was available. As around 10% of the total production of cereals seems to be used to produce such drinks, it is worthwhile to incorporate them. Other utilities are currently included in the category losses, and are thus assumed to increase linearly with increasing demand and thus production. As other utilities include biofuels and other biobased materials, this category may increase in importance in the future. It would be interesting to include projections concerning e.g. government policy related to biofuels into the food scenarios, or to use the basic structure of the model to quantify scenarios related to biobased products.

Wastes


There seems to be confusion in the literature about what constitutes ‘wastes’ and ‘losses’, and this issue should be handled with care. The terms are used loosely and they are not reported consistently. Wastes, 130

thus household and retail wastes, but also wastes during storage, processing and transport, are important factors related to food production, and deserve the attention they are currently getting. However, it is important that issues are clear. When the FAO estimates ‘storage, processing and transport wastes’ as a proportion of total production in a country, based on its climate, it is nonsensical to try to find correlations between these wastes and economic development to make projections. It is similarly nonsensical to call losses such as seed and other utilities ‘wastes’ as they clearly serve a valuable purpose. Furthermore, all too often, it is unclear or dismissed that the food supply values given by the FAO still include refuse. Sugar and oil

There are a couple of questions related to sugar crops and oil crops production that deserve further attention (all elaborated on in the discussion section):  

Why do the sugar yields in the OECD90 region exceed the MAY? Will the assumed conversion factors change in case of a change in composition of the commodity group, e.g. increased demand for feed will favor soy production?  What is the correct area under cultivation for oil crops? The FAO data concerning oil crops is intransparant and deserves further elaboration and clarification in FAOSTAT. Cropping intensity

Cropping intensity should be reported along with area under cultivation, for all countries and for all crops and commodity groups. Double cropping can reduce land use, but is currently reported in a format which is intransparant. Furthermore, it is unclear whether significant areas in the OECD90 and REF are left fallow; there is an inexplicable gap between the total area under cultivation for the main commodity groups and the total agricultural area.

Yield projections


It is reasonable to assume that different land qualities have different maximum attainable yields, and these could be integrated in the model. This could also be done for rainfed and irrigated agriculture for crops other than cereals. Some of the strengths of the model and the scenarios are their simplicity and transparency, which would be compromised in case of much further elaboration of these matters. However, because these matters have a great influence on future supplies, they deserve attention, especially in combined to water use and fertilizer use.

131

Diet

Fruit and vegetable consumption are probably underestimated in the A scenarios. Resource use in the A scenarios already exceeds our biocapacity, adaption would further increase resource use. Effects of changes in the composition of meat consumption pose an interesting topic for further research. It may yield interesting results as inputs are much low for poultry meat and pork than for beef, and different types of meat are preferred in different cultural and religious systems.

Recommendations regarding further research related to the Virtual Resource Content model can focus on further elaboration of the VRC methodology, for which suggestions are pointed out above. Furthermore, the current scenarios can be elaborated on, or other scenarios can be modeled. Finally, it might be possible to link the methodology to other tools that assess resource use and its impacts. Regional results were modeled for the A2 and B2 scenarios. This could also be done for the A1 and B1 scenarios. The regional assessment can be further elaborated on by estimating the resources available per capita in the four regions. This, together with supply and demand data can be used to optimize trade schemes. Furthermore, as pastures are an important part of the animal diet, and fertilizer inputs are very important in the A1 scenario to achieve the high yields necessary to maintain low land use, it is worthwhile to explore different scenarios concerning pasture utilization. Additionally, a lot of interesting data can be extracted from the model, e.g. the inputs needed to produce a kg of protein of a kcal, the efficiency of meat production, the influence on resource use of different types of meat. Also interesting would be the creation of other scenarios, and quantifying their effects on resource use. For example, bio-based materials or fuels have been mentioned, which could become more important in the future. Other scenarios can be used for different purposes, for example increasing consumer awareness by evaluating current consumption patterns. For instance, the resource use in agriculture of an average Dutch diet could be calculated and compared to the global average, or the resource requirements, would the global population consume a Dutch diet. Linking this model to a calorie check, such as the one at Voedingscentrum.nl, may be interesting and give further incentive to change unhealthy habits. However, it is important to be cautious, as food is essential and one would not want to give some people further reason to reduce their intake. Addition of such a model to footprint calculators may thus be more appropriate. In fact, one was recently put online by the Water Footprint Network. Linked to this are the options for policymakers. The results can be useful to stimulate or discourage certain consumer behavior, for example through higher taxes on meat, or for industrialized countries to support foreign aid projects.


132

Bibliography

Alexandratos, N., Bruinsma, J., Bödeker, G., Schmidhuber, J., Broca, S., Shetty, P., Grazia Ottaviani, M., World agriculture: towards 2030/2050, Prospects for food, nutrition, agriculture and major commodity groups, Interim report, Global Perspective Studies Unit, Food and Agriculture Organization of the United Nations, Rome, June 2006 Badgley, C., Moghtader, J., Quintero, E., Zakem, E., Chappell, M.J., Avile´s-Va´zquez, K., Samulon, A., Perfecto, I., Organic agriculture and the global food supply, Renewable Agriculture and Food Systems, 22(2), 86-108, 2006 Bender, W.H., An end use analysis of global food requirements, Food Policy, Vol. 19 (4), pp. 381-395, 1994 Börjeson, L., Höjer, M., Dreborg, K-H., Ekvall, T., Finnveden, G., Scenario types and techniques: Towards a user’s guide, Futures 38, 723-739, 2006 Bruinsma, J. (ed.), World Agriculture: towards 2015/2030, An FAO perspective, FAO, Earthscan publications Ltd, 2003 Cassman, K.G., Ecological intensification of cereal production systems: Yield potential, soil quality, and precision agriculture, Proceedings of the National Academy of Sciences, Vol. 96, pp. 5952-5959, May 1999 Chapagain, A.K., Hoekstra, A.Y., The global component of freshwater demand and supply: an assessment of virtual water flows between nations as a result of trade in agricultural and industrial products, Water International, Vol. 33, No., 1, pp.19-32, 2008 Cuéllar, A.D., Webber, A.E., Wasted Food, Wasted Energy: The Embedded Energy in Food Waste in the United States, Environ. Sci. Technol. Vol. 44, pp. 6464-6469, 2010 De Fraiture, C., Wichelns, D., Rockström, J., Kemp-Benedict, E., Looking ahead to 2050: scenarios of alternative investment approaches, in: Molden, D. (ed), Water for Food, Water for Life, Earthscan, London, and International Water Management Institute, Colombo, 2007, available at http://www.iwmi.cgiar.org/assessment/Publications/books.htm De Fraiture, C., Wichelns, D., Satisfying future water demand for agriculture, Agricultural Water Management, Vol. 97, p. 502-511, 2010 De Ponti, T., personal communication, June 2011, Duarte, C. M., Holmer, M., Olsen, Y., Soto, D., Marbà, N., Guiu, J., Black, K., Karakassis, I., Will the Oceans Help Feed Humanity? BioScience, Vol. 59, No. 11, p. 967-976, December 2009


133

Ewert, F., Rounsevell, M.D.A., Reginster, I., Metzger, M.J., Leemans, R., Future scenarios of European agricultural land use, I. Estimating changes in crop productivity, Agriculture, Ecosystems and Environment, Vol. 107, p. 101-116, 2005 Ewert, F., Rounsevell, M., Reginster, I., Metzger, M., Leemans, R., Technology Development and Climate Change as Drivers of Future Agricultural Land Use, in: Brouwer, F., McCarl, B.A., Agriculture and Climate Beyond 2015, p. 33-51, 2006 FAO, Food and Agriculture Organization of the United Nations, Fertilizer and Plant Nutrition Service, Land and Water Development Division, Fertilizer and plant nutrition guide, FAO Fertilizer and Plant Nutrition Bulletin 9, Rome, 1984 FAO, Food and Agriculture Organization of the United Nations, Allen, R.G., Pereira, L.,S., Raes, D., Smith, M., Crop evapotranspiration - Guidelines for computing crop water requirements, FAO Irrigation and drainage paper 56, Rome, 1998, http://www.fao.org/docrep/x0490e/x0490e00.htm FAO, Food and Agriculture Organization of the United Nations, Inter-temporal Changes of Conversion Factors, Extraction Rates, Productivity of Crops and Livestock and Related Matters: 1963-1967 to 19931997, August 2000, accessed December 2010, http://www.fao.org/fileadmin/templates/ess/documents/methodology/totdoc.pdf FAO, Food and Agriculture Organization of the United Nations, World Agriculture: towards 2015/2030, Summary Report, Rome, 2002, ftp://ftp.fao.org/docrep/fao/004/y3557e/y3557e.pdf FAO, Food and Agriculture Organization of the United Nations, Fertilizer use by crop, FAO Fertilizer and Plan Nutrition Bulletin, Rome, 2006 FAO4, Food and Agriculture Organization of the United Nations, FAOSTAT, Glossary, last updated: 2010, accessed: December 2010, http://faostat.fao.org/site/375/default.aspx Fischer, G., Van Velthuizen, H., Medow, S., Nachtergaele, F., at IIASA, International Institute for Applied Systems Analysis, and FAO, Food and Agriculture Organization of the United Nations, Global Agroecological Assessment for Agriculture in the 21st Century CD-ROM, 2002, available at http://www.iiasa.ac.at/Research/LUC/SAEZ/ accessed November 2010 Foley, J.A., DeFries, R., Asner, G.P., Barford, C., Bonan, G., Carpenter, S.R., Chapin, F.S., Coe, M.T., Daily, G.C., Gibbs, H.K., Helkowski, J.H., Holloway, T., Howard, E.A., Kucharik, C.J., Monfreda, C., Patz, J.A., Prentice, I.C., Ramankutty, N., Snyder, P.K., Global Consequences of Land Use, Science, Vol. 309, July 2005 Gerbens-Leenes, W., Nonhebel, S., Food and land use. The Influence of consumption patterns on the use of agricultural resources, Appetite, Vol. 45, p. 24-31, 2005 Grigg, D., The Pattern of World Protein Consumption,Geoforum, Vol. 26, No. 1, pp. 1-17, 1995 Hoekstra, A.Y., Chapagain, A.K., Globalization of Water: Sharing the Planet’s Freshwater Resources, Wiley-Blackwell, Oxford, UK, 2008 M.Sc. Thesis I.Y.R. Odegard

134

Hoekstra, A.Y., Human appropriation of natural capital: A comparison of ecological footprint and water footprint analysis, Ecological Economics, 68, p. 1963-1974, 2009 Howarth, R.W., Boyer, E.W., Pabich, W.J., Galloway, J.N., Nitrogen Use in the United States from 19612000 and Potential Future Trends, Ambio, Vol. 31, No. 2, March 2002 IPCC, Intergovernmental Panel on Climate Change, Nakicenovik, N., Alcamo, J., Davis, G., de Vries, B., Fenman, J., et al., Special Report on Emission Scenarios: A Special Report of Working Group III of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, 2000, available at: http://www.ipcc.ch/ipccreports/sres/emission/index.php?idp=87 IFA, International Fertilizer Association, Production and Trade Statistics, last updated: February 22nd, 2011, accessed April 2011, available at: http://www.fertilizer.org/ifa/HomePage/STATISTICS/Productionand-trade , 2011 Kantor, L.S., Lipton, K., Manchester, A., Oliveira, V., Estimating and Addressing America’s Food Losses, FoodReview, pp. 2-12, January-April, 1997 Keyzer, M.A., Merbis, M.D., Pavel, I.F.P.W, Wesenbeeck, C.F.A., Diet shifts towards meat and the effects on cereal use: can we feed the animals in 2030? Ecological Economics, Vol. 55, p. 187-202, 2005 Lotze-Campen, H., Müller, C., Bondeau, A., Smith, P., Lucht, W., Rising Food Demand, Climate Change and the Use of Land and Water, in: Brouwer, F., McCarl, B.A., Agriculture and Climate Beyond 2015, Springer, The Netherlands, pp. 109-129, 2006 Lutz, W., Sanderson, W., Scherbov, S., IIASA's 2007 Probabilistic World Population Projections, IIASA World Population Program Online Database of Results, 2008, accessed October 2009, available at: http://www.iiasa.ac.at/Research/POP/proj07/index.html?sb=5, http://www.iiasa.ac.at/Research/POP/proj07/ptotr_all.xls, Nijdam, D.S., Wilting, H.C., Environmental load due to private consumption, RIVM rapport 771404004, in Dutch, 2003 Mäder, P., Flieβbach, A., Dubois, D., Gunst, L., Fried, P., Niggli, U., Soil Fertility and Biodiversity in Organic Farming, Science, Vol. 296, May 2002 Molden, D., Frenken, K., Barker, R., De Fraiture, C., Mati, B., Svendsen, M., Sadoff, C., Finlayson, C.M., Trends in water and agricultural development, in: Molden, D. (ed), Water for Food, Water for Life, Earthscan, London, and International Water Management Institute, Colombo, 2007, available at http://www.iwmi.cgiar.org/assessment/Publications/books.htm Penning de Vries, F.W.T., Van Keulen, H., Rabbinge, R., Natural resources and limits of food production in 2040, in: Bouma, J. et al. (eds), Eco-regional approaches for sustainable land use and food production, pp. 65-87, Kluwer Academic Publishers, the Netherlands, 1995


135

OECD, Health at a Glance: Europe 2010, OECD Publishing, accessed May 2011, available at http://dx.doi.org/10.1787/health_glance-2010-en, 2010 PBL, Planbureau voor de leefomgeving, SRES Scenarios, last updated: December 23rd 2009, accessed June 7th 2010, http://www.pbl.nl/en/themasites/image/scenarios/sres_scenarios/index.html Pimentel, D., Environmental and Social Implications of Waste in the U.S. Agriculture and Food Sectors, Journal of Agricultural Ethics, Vol. 3, No. 1, 1990 Pimentel, D., hepperly, P., Hanson, J., Douds, D., Seidel, R., Environmental, Energetic, and Economic Comparisons of Organic and Conventional Farming Systems, BioScience, Vol. 55, No. 7, July 2005 Pollan, M., In Defense of Food: An Eater's Manifesto, Penguin Press, New York, 2008 Pollan, M., The Omnivore’s Dilemma, A Natural History of Four Meals, Penguin Press, New York, 2006 Poudel, D.D., Horwarth, W.R., Lanini, L.T., Temple, S.R., Van Bruggen, A.H.C., Comparison of soil availability and leaching potential, crop yields and weeds in organic, low-input and conventional farming systems in Northern California, Agriculture, Ecosystems and Environment, Vol. 90, pp. 125-137, 2002 Ringland, G., Scenario Planning, Managing for the Future, John Wiley & Sons Ltd, England, 1998 Rosegrant, M.W., Paisner, M.S., Meijer, S., Witcover, J., Global Food Projections to 2020, Emerging Trends and Alternative Futures, International Food Policy Research Institute, Washington, D.C., August 2001a Rosegrant, M.W., Paisner, M.S., Meijer, S., Witcover, J., 2020 Global Food Outlook, Trends, Alternatives and Choices, A 2020 Vision for Food, Agriculture, and the Environment Initiative, International Food Policy Research Institute, Washington, D.C., August 2001b Schoemaker, P.J.H., Scenario Planning: A Tool for Strategic Thinking, Sloan Management Review, Winter, 1995 Schneider F., Consideration of food losses in life cycle approach of food supply chain, Institute of Waste Management, BOKU-University of Natural Resources and Applied Sciences Vienna, Switzerland, 2007 Shearer, A.W., Approaching scenario-based studies: three perceptions about the future and considerations for landscape planning, Environment and Planning B: Planning and Design 2005, Vol. 32, pp. 67-87 Smil, V., Feeding the World; A Challenge for the Twenty-First Century, The MIT Press, Cambridge, USA, 2001 SOW-VU, Can we feed the animals? Centre for World Food Studies, Amsterdam, 2005. Tilman, D., Fargione, J., Wolff, B., D’Antiono, C., Dobson, A., Howarth, R., Schindler, D., Schlesinger, W.H., Simberloff, D., Swackhamer, D., Forecasting Agriculturally Driven Global Environmental Change, Science, Vol. 292, April 2001 M.Sc. Thesis I.Y.R. Odegard

136

UN, United Nations, UN Population Division, retrieved October 2009, from World Population Prospects: The 2008 Revision Population Database: http://esa.un.org/unpp, 2009 UN, United Nations, The Millennium Development Goals Report 2010, New York, 2010 http://www.un.org/millenniumgoals/pdf/MDG%20Report%202010%20En%20r15%20low%20res%2020100615%20-.pdf#page=13 USGS, U.S. Geological Survey, Mineral Commodity Summaries 2009, 2009 Van Vliet, J., PBL, SRES PPP data, Personal Communication, 2010 Vinnari, M., Tapio, P., Future images of meat consumption in 2030, Futures, Vol. 41, 269-278, 2009 Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., Human Domination of Earth’s Ecosystems, Science 277, 494, 1997 Wirsenius, S., Human Use of Land and Organix Materials, Modeling the Turnover of Biomass in the Global Food System, Department of Physical Resource, Theory, Chalmers University of Technology and Göteborg University, Göteborg, Sweden, 2000 Wirsenius, S., The Biomass Metabolism of the Food System, A Model-Based Survey of the Global and Regional Turnover of Food Biomass, Journal of Industrial Ecology, Vol. 7, No. 1, p. 47-80, 2003 Wirsenius, S., Azar, C., Berndes, G., How much land is needed for global food production under scenarios of dietary changes and livestock productivity increases in 2030?, Agricultural Systems, Vol. 103, p. 621638, 2010 WWAP, World Water Assessment Programme, The United Nations World Water Development Report 3: Water in a Changing World, Paris: UNESCO, London: Earthscan, 2009 WWF, Living Planet Report 2006, edited by Halls, C., in corporation with the Zoological Society of London and the Global Footprint Network, 2006 FAO Databases: FAOSTAT, FAO, Time-series and cross sectional data relating to food and agriculture, last updated: 2011, accessed: 2010-2011, http://www.faostat.fao.org AQUASTAT, FAO/NRL, FAO’s Information System on Water and Agriculture, last updated 2011, accessed 2010-2011, http://www.fao.org/nr/water/aquastat/main/index.stm FERTISTAT, Food and Agriculture Organization of the United Nations, Plant Production and Protection Division, Fertilizer Use Statistics, updated 2007, accessed: February 2011, http://www.fao.org/ag/agl/fertistat/index_en.htm


137

Appendix 1

World Regions

The definition of the four regions as used in this study [based on IPCC, 2000 and FAO, 2010]. Table 40: Definition of the regions [IPCC, 2000; FAO, 2010].

OECD90 REGION North America (NAM) Canada

United States of America

Western Europe (WEU) Austria Belgium Cyprus Denmark Finland France

Germany Greece Iceland Ireland Italy Luxembourg

Madeira Malta Netherlands Norway Portugal Spain

Sweden Switzerland Turkey United Kingdom

Pacific OECD (PAO) Australia Japan New Zealand Note: The following countries present in the IPCC list were excluded because they are not listed in the FAO country list. These are part of other countries (overseas) territories: Guam, Puerto Rico, Virgin Islands, Azores, Canary Islands, Faeroe Islands, Gibraltar, Greenland, and Isle of Man. Also missing are: Liechtenstein, Monaco and Andorra.

REF REGION (countries undergoing economic reform) Central and Eastern Europe (EEU) Albania Bosnia and Herzegovina Bulgaria Croatia

Czech Republic Hungary Poland Republic of Macedonia

Romania Slovakia Slovenia Serbia and Montenegro

Yugoslavia

Newly independent states (NIS) of the former Soviet Union (FSU) Armenia Georgia Lithuania Turkmenistan Azerbaijan Kazakhstan Republic of Moldova Ukraine Belarus Kyrgyzstan Russian Federation Uzbekistan Estonia Latvia Tajikistan Note: ‘Serbia and Montenegro’ was substituted for ‘The Former Yugoslav’ (as it is named in the IPCC list), as was ‘Slovakia’ for the ‘Slovak Republic’.

ASIA REGION Centrally planned Asia and China (CPA) Cambodia China

Korea (DPR) Laos (DPR)

Mongolia Vietnam

Bhutan India

Nepal Pakistan

South Asia (SAS) Afghanistan Bangladesh

Sri Lanka

Other Pacific Asia (PAS) American Samoa Malaysia Republic of Korea Tonga Brunei Darussalam Myanmar Singapore Vanuatu Fiji New Caledonia Solomon Islands French Polynesia Papua New Guinea Timor-Leste Indonesia Philippines Thailand Note: ‘Hong Kong’ is not listed separately in the FAO country-list, as well as the ‘Maldives’, ‘Gilbert-Kiribati’, ‘Taiwan, province of China’ and Western Samoa. ‘Timor-Leste’ was added, this country became independent in 2002.


138

ALM REGION (Africa and Latin America) Middle East and North Africa (MEA) Algeria Bahrain Egypt (Arab Republic) Iran (Islamic Republic) Iraq

Israel Jordan Kuwait Lebanon Libya (SPLAJ)

Morocco Oman Qatar Saudi Arabia Sudan

Syria (Arab Republic) Tunisia United Arab Emirates Yemen Occupied Palestinian Territory Saint Vincent and the Grenadines Santa Lucia Suriname Trinidad and Tobago Uruguay Venezuela

Latin America and the Caribbean (LAM) Antigua and Barbuda

Costa Rica

Honduras

Argentina Bahamas Barbados Belize Bolivia Brazil Chile Colombia

Cuba Dominica Dominican Republic Ecuador El Salvador Guatemala Guyana Haiti

Jamaica Martinique Mexico Nicaragua Panama Paraguay Peru Saint Kitts and Nevis

Sub-Saharan Africa (AFR) Angola

Democratic Republic of Madagascar Seychelles Congo Benin Equatorial Guinea Malawi Sierra Leone Botswana Eritrea Mali Somalia Burkina Faso Ethiopia Mauritania South Africa Burundi Gabon Mauritius Swaziland Cameroon Gambia Mozambique Tanzania Cape Verde Ghana Namibia Togo Central African Republic Guinea Niger Uganda Chad Guinea-Bissau Nigeria Zambia Comoro Kenya Rwanda Zimbabwe Congo Lesotho Sao Tome and Principe Cote d’Ivoire Liberia Senegal Note: The following countries are not included in the FAO factsheets and are part of other countries (overseas) territories: Bermuda, French Guyana, Guadeloupe, Netherlands Antilles, British Indian Ocean Territory, Reunion and Saint Helena. The Democratic Republic of Congo was added. Also missing are Grenada and Djibouti.

The ‘missing’ countries together accounted for 0.1% of the total world population in the year 2005.


139

Appendix 2

Commodity Group Assumptions

Table 41: Commodity group assumptions

Commodity Group

Caloric Value (kcal per kg)

Protein Content f (g per kg)

Cereals Roots and tubers Sugar crops Sugar and sweeteners Oil crops Vegetable oils Pulses Meat Milk Vegetables Fruit Eggs

3208.9 824.4 289 3484.6 2871.7 8735.1 3418.8 1982.6 558.7 242.5 458.2 1426.7

79.7 12.7 1.7 0.6 141.8 1 215.4 122.7 33.3 d 12.3 c 5 111.0

Dry weight fraction (% dry weight per harvested a weight) 88 30 26 90 100 10 20 -

Note: All caloric values and protein contents were based on world averages reported by FAOSTAT under Food Supply – Crops Primary Equivalent and Livestock and Fish Primary Equivalent. Caloric values of commodity groups may vary between regions because of different compositions. a [Wirsenius, 2000]


140

Table 42: Items included in the commodity groups.

Items Included in Commodity Groups Cereals Wheat Barley Rye Millet Cereals, other

Rice, paddy (milled equivalent) Maize Oats Sorghum

Roots and tubers Potatoes Cassava Roots, other

Sweet potatoes Yams

Sugar crops and sweeteners Sugar cane Sugar non-centrifugal Sweeteners, nes

Sugar beet Sugar (raw equivalent) Honey

Oil crops and vegetable oils Soybeans and soybean oil Sunflower seed and sunflower seed oil Cottonseed and cottonseed oil Sesame seed and sesame seed oil Olives Maize germ oil

Groundnuts (shelled equivalent) and groundnut oil Rape and mustard seed and rape and mustard oil Coconuts (including copra and copra oil) and coconut oil Palm oil, palm kernels and palm kernel oil Rice bran oil Oilcrops, other and oilcrops oil, other

Pulses Beans Pulses, other

Peas

Vegetables Tomatoes Vegetables, other

Onions

Fruit Oranges and mandarins Grapefruit Bananas Apples Dates Fruit, other


Lemons and limes Citrus, other Plantains Pineapples Grapes

141

Appendix 3

Waste

Table 43: Waste (during transport, storage and processing) in 2007 [based on FAOSTAT, 2010].

Commodity

Cereals % of total Roots and tubers % of total Sugar crops % of total Oil crops % of total Pulses % of total Vegetables % of total Fruits % of total Eggs % of total Meat % of total Milk % of total


Waste during transportation and storage  tons in 2007  % of total production World OECD90

REF

ASIA

ALM

85,738,291 3.6%

8,240,231 1.1%

6,494,882 2.8%

39,480,964 3.8%

31,522,211 9.0%

58,613,974 8.4%

5,239,392 6.3%

3,897,906 4.2%

19,069,122 7.3%

30,407,553 11.7%

59,891,852 3.3%

173,920 0.08%

364,710 0.5%

8,540,126 1.3%

50,813,096 5.8%

10,438,441 7.0%

996,481 3.3%

576,728 6.3%

6,269,687 8.5%

2,595,545 7.3%

2,644,214 4.7%

257,643 2.2%

133,308 4.7%

991,193 4.2%

1,262,070 7.1%

7,572,981 8.3%

14,499,732 10.7%

2,468,639 4.4%

45,921,942 7.8%

12,838,967 10.0%

50,176,340 6.7%

7,197,473 7.2%

1,099,082 0.5%

20,964,282 9.2%

20,915,502 10.3%

2,891,421 4.2%

271,233 2.6%

46,183 0.3%

1,869,300 5.6%

704,705 2.1%

1,142,940 0.4%

273,935 0.3%

85,410 0.5%

197,20 0.02%

763,875 1.4%

15,767,789 2.2%

1,137,084 0.4%

819,936 0.8%

9,048,976 4.8%

4,761,794 4.0%

142

Appendix 4

Feed

Table 44: Feed in 2007 [based on FAOSTAT, 2010].

Commodity



Feed  tons in 2007  % of total production World OECD90

REF

ASIA

ALM

745,878,305 31.7%

352,962,020 48.5%

97,664,273 42.3%

166,096,091 15.9%

129,155,921 36.8%

1,61E+08 23.1%

7,431,405 9.0%

27,543,302 29.5%

64,167,500 24.5%

62,207,062 24.0%

36,903,992 2.0%

193,309 0.09%

10,000,373 13.2%

6,761,432 1.0%

12,497,560 1.4%

26,827,777 18.0%

10,002,061 32.8%

2,536,372 27.6%

10,924,250 14.8%

3,365,092 9.5%

10,685,440 19.1%

3,492,036 29.9%

1,784,584 63.4%

3,848,975 16.2%

1,559,844 8.8%

40,221,451 4.4%

5,191,192 3.8%

6,378,496 11.3%

27,704,983 4.7%

946,780 0.7%

5,216,606 0.7%

78,743 0.07%

503,298 0.2%

752,088 0.3%

3,882,477 1.9%

55,465 0.09%

165 0.00001%

55,300 1.1%

0 0.0%

0 0.0%

8,829,595 3.6%

8,815,756 10.0%

6,700 0.04%

770 0.000009%

6,369 0.01%

71,450,531 10.5%

11,335,223 4.2%

21,667,066 21.4%

25,945,623 13.6%

1,250,219 10.4%

143

Appendix 5

Seed

Table 45: Seed in 2007 [based on FAOSTAT, 2010].

Commodity

Cereals % of total Roots and tubers % of total Sugar crops % of total Oil crops % of total Pulses % of total a Vegetables % of total Fruits % of total Eggs % of total

Seed  tons in 2007  % of total production World OECD90

REF

ASIA

ALM

67,421,309 2.9%

15,066,469 2.1%

22,830,215 10.0%

21,930,682 2.1%

7,593,943 2.2%

34,477,598 4.9%

4,598,908 5.6%

17,618,798 18.9%

6,687,945 2.6%

5,571,947 2.1%

27,481,119 1.5%

1,703,000 0.8%

0 0.0%

24,915,916 3.7%

862,203 0.09%

10,917,741 7.3%

3,164,796 10.4%

950,543 10.3%

3,914,465 5.3%

2,887,936 8.2%

3,873,360 6.9%

790,797 6.8%

369,234 13.1%

1,388,926 5.9%

1,324,403 7.5%

111,537 0.01%

26,882 0.02%

30,242 0.05%

19,501 0.003%

34,912 0.03%

0 0.0%

0 0.0%

0 0.0%

0 0.0%

0 0.0%

3,939,412 1,324,409 251,666 1,238,984 3,939,412 6.2% 9.2% 4.8% 3.7% 11.0% a ‘Only those vegetables which are cultivated principally for human consumption belong to this group. Consequently, vegetables 5 grown principally for animal feed should be excluded, as should vegetables cultivated for seed’ *FAO , 2010].


144

Appendix 6

Other Utilitization

Table 46: Other utilization in 2007 [based on FAOSTAT, 2010].

Commodity



Global ‘other utilization’  tons in 2007  % of total production 79,266,054 3.4% 41,096,189 5.9% 18,467,045 1.0% 13,805,594 9.3% 574,688 1.0% 462,375 0.05% 1,728,398 0.2% 613,743 1.0% 953,594 0.4% 17,228,369 2.5%

145

Appendix 7

Yield Projections

Most yield projections are made solely for cereals. Ewert et a. give a metholodogy for making yield projections for different commodity groups, which can only reasonably be applied to the OECD90 region (see also Appendix 15), and can thus only be used as for comparison. The FAO gives productivity projections (see Appendix 16) for the commodity groups cereals, oil crops and sugar crops, but it is unclear to what extent cropping intensity and land expansion play a role. Because this study gives an holistic picture, and includes the 7 commodity groups in four different regions, different yield projections have to be made. First, the cereal yield projections will be introduced. The yield projections made for the other commodity groups are based on the same methodology. Cereal yield projections were extracted from [De Fraiture, 2010]. De Fraiture assumes that in an “optimistic scenario” 80% of the yield gap is bridged, while in a “pessimistic” scenario 20% of the yield gap is bridged. Economic and technological development is high in the A1 and B1 scenarios, thus for these scenarios a bridging of 80% of the yield gap was chosen. Development in the A2 scenario is low, which makes bridging the yield gap with 20% reasonable. Table 47 shows the cereal yield projection for the four scenarios, and for the four regions in the A2 and B2 scenarios, for both rainfed and irrigated cereal yields. Table 47: Cereal yield projection [based on De Fraiture, 2010].

Scenario

Region

2005 A1 A2

World World OECD90 REF ASIA ALM World OECD90 REF ASIA ALM

B1 B2

Cereal Yield (ton/ha) Rain-fed Irrigated 3.33 3.88 5.20 2.50 2.60 2.25 3.90 2.60 1.25 1.30 1.13

5.74 7.20 4.20 4.90 4.30 4.94 3.60 2.10 2.45 2.15

In line with the reasoning behind the cereal yield projections [De Fraiture, 2010], yield projections for the other commodity groups were based on the exploitable yield gap and the maximum attainable yield (MAY). Data from the FAO and the IIASA were used to estimate maximum attainable yields [Fisher, 2010]. The potential yield (or maximum attainable yield) depends on the ecological zone, the crop and the management practice. This means that within a region and within a commodity group yields can differ. One of the assumptions underlying the yield projections made here for the commodity groups is that the composition of the production of crops within a commodity group does not change. This is important M.Sc. Thesis I.Y.R. Odegard

146

because crop yields within a commodity group can differ widely. For example, the yield of ‘beans, dry’ in 2005 in The Netherlands was 3.6 ton/ha, while the yield of ‘broad beans, horse beans, dry’ was 6.4 ton/ha. If production was shifted to only this latter type, yields would necessarily be higher. Maximum attainable yields for the OECD90 region and the REF region were based on the data for the ‘temperate’ zone. For the ASIA region and the ALM region the average between ‘tropical’ and ‘temperate’ was used. The global MAYs were set at the same value as those for the ASIA region and the ALM region, as these are slightly higher than the OECD90 and REF region. Another issue which need to be kept in mind in the definition of global average yields is that the potential to increase yields is much higher in those regions where the yield gap is still high: the less developed regions. Thus the yield increases in these regions accounts for a larger part of the total increase than the increase in the industrialized regions. Current regional average yields were calculated using FAOSTAT data. Total production in the region was divided by the total harvested area. No data concerning maximum attainable yields for fruits and vegetables was available. Therefore, an estimate of the maximum attainable yield was made using historic data. On a regional level, the fruits and vegetables yields of the top three yielding countries was averaged. This was taken as the maximum attainable yield for these commodity groups for the regions. On a global level, the yields of those 12 countries was averaged, and that value was taken as an estimate of the global MAY. Some adjustments were made as in some cases the countries in the top three yield range do not offer much variety in terms of production, which greatly influences the average yield. For instance, in the OECD90 region, Iceland has the highest vegetable yield, but it produces only 5 out of 25 crops in the vegetable commodity group, while the Netherlands produces 20 out of 25. Because Iceland produces the high yielding crops ‘cucumbers and gherkins’ and ‘tomatoes’, the average yield in Iceland is 33% higher than in the second leading country for vegetable yields: the Netherlands. It was checked that the countries from which a maximum attainable yield was calculated either produce a significant number of crops or have a yield comparable to those of the other countries in the top. MAYs for sugar beet were significantly underestimated by the FAO and the IIASA [Fisher, 2002], as current yields in the OECD90 region are much higher that the estimated MAY. The same is true for the roots and tubers yield in the OECD90 region, which was higher than the estimated MAY in the year 2005. Part of the explanation for the difference can be that sugar crops may be irrigated in the OECD90 region, but this does not account for the complete difference. The MAYs for these commodity groups were adjusted, using the same methodology which was used to estimate the MAYs for fruits and vegetables. The oil crop production data given by the FAO is given in ‘converted’ values, to account for differences in oil fractions of different oil crops. To estimate unconverted values yields were converted using the extraction rate for oil crops as established in Appendix 9.


147

Table 48: Top three yielding countries per region for fruits and vegetables [based on FAOSTAT, 2011].

Top three yielding countries per region for fruit and vegetables (number of crops grown in commodity group) st

Fruit OECD90 REF ASIA ALM

nd

rd

1

2

3

Belgium (10) Slovenia (13) Republic of Korea (12) Costa Rica (12)

Netherlands (10) Armenia (11) Indonesia (8) Israel (27)

USA (27) Macedonia (15) Malaysia (10) Honduras (13)

Netherlands (20) Armenia (10) Republic of Korea (16) Kuwait (18)

Austria (19) Poland (15) China (24) United Arab Emirates (9)

Luxembourg (6) Slovenia (13) Singapore (6) Jordan (20)

Vegetables OECD90 REF ASIA ALM

Virtual Land Content Yield projections (ton/ha) can be seen as an inverse of Virtual Land Content (m2/kg). Because higher inputs, better management practices and technological development (different strains) lead to higher yields, such projections are adapted to the scenarios. The methodology of closure of the yield gap, as explained above, can be used for other scenarios, for different ‘closures’. Table 49 shows the maximum attainable yields, the yields in 2005, and the projected yields in 2050 for the four scenarios as used in this study. Table 49: Maximum attainable yields (MAYs) and yields in 2005 and in 2050 [based on Fisher, 2002; FAOSTAT, 2011; De Fraiture, 2010]. Yield projections for 2050 take scenario characteristic into account.

Maximum Attainable Yields and Yields in 2005 and 2050 OECD90 Scenario Fruit MAY 28.12 Yield – 2005 13.06 Yield – 2050 (A2) 16.07 Yield – 2050 (B2) 8.03 Yield – 2050 (A1/B1) Oil crops MAY 4.40 Yield – 2005 1.56 Yield – 2050 (A2) 2.13 Yield – 2050 (B2) 1.06 Yield – 2050 (A1/B1) Pulses MAY 3.7 Yield – 2005 1.86 Yield – 2050 (A2) 2.23 Yield – 2050 (B2) 1.11 Yield – 2050 (A1/B1) -


REF

ASIA

ALM

World

10.5 5.10 6.18 3.09 4.64 1.34 2.00 1.00 3.7 1.73 2.12 1.06 -

14.88 9.81 10.83 5.41 6.03 2.02 2.83 1.41 3.5 0.75 1.30 0.65 -

25.37 10.06 13.12 6.56 4.81 1.12 1.86 0.93 3.87 0.62 1.27 0.64 -

21.25 10.04 19.01 5.21 1.58

4.48 3.70 0.86 3.13

148

Sugar crops

Roots tubers

and

Vegetables

MAY Yield – 2005 Yield – 2050 (A2) Yield – 2050 (B2) Yield – 2050 (A1/B1) MAY Yield – 2005 Yield – 2050 (A2) Yield – 2050 (B2) Yield – 2050 (A1/B1) MAY Yield – 2005 Yield – 2050 (A2) Yield – 2050 (B2) Yield – 2050 (A1/B1)

71.06 64.90 66.13 33.06 43.26 36.00 37.45 18.73 41.84 27.84 30.64 15.32 -

50.09 31.88 35.52 17.76 34.33 13.55 17.70 8.85 26.14 16.15 18.15 9.08 -

65.27 60.21 61.22 30.61 41.30 17.04 21.89 10.95 23.24 16.81 18.09 9.05 -

76.77 69.08 70.62 35.31 48.60 9.74 17.51 8.76 45.45 13.64 20.00 10.00 -

70.92 62.23 69.18 43.30 13.64

37.37 29.50 17.28

27.06

Notes: 1) Regional yields are based on assumption that 20% of the yield gap is closed (A2 scenario), while global yields are based on the assumption that 80% of the yield gap is closed (A1 and B1 scenarios). 2) MAYs are based on maximum attainable crop yields, with high input levels under rainfed conditions as defined by the FAO and the IIASA [Fisher, 2002]. They are averaged according to the proportion of production of main crops in 2005 in the regions. As these crops were assumed to be grown under rainfed conditions, the upper boundary (representing the most productive cultivar) was chosen to compensate for the lower maximum yields under such conditions. Values were converted from values in dry weight, which can be found in Appendix 3. No information was given for fruits and vegetables; maximum attainable yields were based on the average of the highest three yields (by three countries, for the whole commodity group) achieved in a region. 3) B2 yields are set to half the A2 yields. This corresponds to a level of between 51% and 102% of current yields; OECD90: between 51% and 68% of current, REF: between 56% and 75% of current, ASIA: between 51% and 86% of current, ALM: between 65% and 102% of current. The higher values in the ALM region indicate that that region has a higher potential for improvement; the yield gaps are higher than in the other regions.

Feed Yields The projected feed crop yields shown in Table 50 – for pasture, harvested-conserved grass-legumes, cropland pasture – are based on data from Wirsenius [Wirsenius, 2003, p.245], and on FAOSTAT and the global agro-ecological zones study by the FAO and IIASA for whole cereals [Fisher, 2002; FAOSTAT 2011]. Table 50: Feedcrop yields in 2005 and 2050 [based on Wirsenius, 2003; Fisher, 2002, FAOSTAT, 2011].

Feedcrop yields in 2005 and 2050 (ton/ha) Pasture

Harvested-conserved grasslegume


Yield – 2005 Yield – 2050 (A2) Yield – 2050 (B2) Yield – 2050 (A1) Yield – 2050 (B1) Yield – 2005 Yield – 2050 (A2) Yield – 2050 (B2) Yield – 2050 (A1) Yield – 2050 (B1)

OECD90

REF

ASIA

ALM

World

6.4 3.2 12.8 6.4 -

6.4 3.2 12.8 6.4 -

6.4 3.2 12.8 6.4 -

6.4 3.2 12.8 6.4 -

6.4 12.8 6.4 6.4 12.8 6.4

149

Cropland pasture

Whole cereals

Yield – 2005 Yield – 2050 (A2) Yield – 2050 (B2) Yield – 2050 (A1) Yield – 2050 (B1) MAY Yield – 2005 Yield – 2050 (A2) Yield – 2050 (B2) Yield – 2050 (A1) Yield – 2050 (B1)

12.8 6.4 62.67 13.92 21.3 10.7 -

12.8 6.4 62.67 11.40 19.5 9.7 -

12.8 6.4 63.33 4.51 14.7 7.3 -

12.8 6.4 63.33 7.75 17.0 8.5 -

6.4 12.8 6.4 64.0 8.1 47.7 47.7

Notes: Pasture, cropland pasture and harvested-conserved grass-legume are based on the world average ‘permanent pasture phytomass per total area of permanent grassland’ of 1.6 Mg DM/ha (linked to a pasture DM of 0.25, thus yielding 6.4 ton/ha) when not intensively managed, and the ‘West Europe’ average of 3.2 Mg DM/ha ((linked to a pasture DM of 0.25, thus yielding 12.8 ton/ha) when intensively managed [Wirsenius, 2003, p. 245]. B2 yields are assumed half of the A2 yields. Whole cereals yields are based on 90% availability of whole maize yields [Wirsenius, 2003], MAY includes the non-available 10%.


150

Appendix 8

Feed-mixes and Feeding Efficiency

Five animal foodstuffs are included in this study: beef, pork, poultry meat, eggs and milk. These categories account for 79% of the global animal caloric supply in the year 2005 [based on FAOSTAT, 2011]. The remaining 21% constitutes of by-products such as butter and offals, and are thus implicitly included. ‘Feed’ is included in FAOSTAT, but this only refers to the part of the feed-mix which consists of edible-type foodstuffs; it is assumed that ‘feed’ as reported by the FAO can also be used as food. Feedmixes, however, contain a number of different components other than foodstuffs. The feed-mixes for the 5 animal product systems in this study were obtained from Wirsenius, as well as the feeding efficiencies [Wirsenius, 2003]. The feeding-efficiency is defined as the feed intake per fresh weight (carcass-weight) produced. The data are quite detailed and therefore provide a good basis for the calculations done in this study. All phases of animal husbandry are included, e.g. reproduction, replacement, gestation and lactation. Milk used as feed (lactation) is viewed as a system internal input, and is therefore subtracted from the overall feeding requirements. A little over 11% of milk is considered ‘feed’ by the FAO, this should therefore be considered a loss. In the research done by Wirsenius, data is given for the year 1993, which is also the starting point in his article written in 2010 [Wirsenius, 2010]. The feed-mixes are composed of 14 different ingredients. Not all of these are included in the feed-mix for each of the five animal systems; e.g. cattle have a preference for pasture, poultry have a preference for cereals, while pigs prefer forage crops. To include a measure of development in the scenarios, four of the regions defined by Wirsenius were chosen to represent the regions in the A2 and B2 scenarios, and 1 region was chosen to represent the A1 and B1 scenario. In the A1 and B1 scenario, the feed mixes defined for the ‘North America and Oceania’ region were chosen to represent the global feeding efficiency in 2050 because it is the most efficient in terms of the weight of the feed input in dry matter. Four of the Wirsenius regions were chosen to represent the four regions in the A2 and B2 scenarios, based on geographical correspondence combined with highest efficiency. Table 51: Choices for correspondence of Wirsenius' regions to IPCC regions.

Region

Wirsenius Region

OECD90 REF ASIA ALM

North America and Oceania Eastern Europe East Asia Latin America and Caribbean

Buffalo, sheep and goat meat are grouped under ‘cattle, beef’, as these animal systems are similar in their requirements and efficiency. For the year 2005, feed-mixes and efficiencies were represented by the global average in 1993. Some progress and development has taken place since then, but it was found that these data are a reasonable approximation to the situation in the year 2005. Overall, in DM, the calculated total feed was only 0.1% higher than the FAO estimate. Cereals provide the bulk of foodcropfeed: 86.7% on DM-basis. The modeled value was only 1.73% lower than the FAO estimate. The other values differ more widely: pulses were modeled 13.6% higher, roots 2.3% lower and oil crops 45.7% M.Sc. Thesis I.Y.R. Odegard

151

higher. This may be due to the fact that feed-mixes have changed. Because the total is so close, as well as the cereals value, and the more efficient regions were chosen to represent the 2050 situation, the values presented below provide a good approximation. Table 52: Feed-mixes for the five animal products per region [based on Wirsenius, 2003]. a

Feed type

Cattle – Milk, Feed input (kg DM) per kg product b c d e OECD90 REF ASIA ALM

Crops, cereals Crops, oil Crops, pulses Crops, roots Pasture Forage- vegetables Edible-type crops by-products Conversion by-products Non-eaten food Whole-cereals - forage Harvested-conserved grass-legume Cropland pasture - forage Meat and bone meal System external inputs

0.361 0 0 0 0.240 0 0.013 0.034 0 0.196 0.147 0.208 0 0

Total Feed type Crops cereals Crops, oil Crops, pulses Crops, roots Pasture Forage-vegetables Edible-type crops by-products Conversion by-products Non-eaten food Whole-cereals - forage Harvested-conserved grass-legume Cropland pasture - forage Meat and bone meal System external inputs

Total Feed type Crops, cereals Crops, oil Crops, pulses Crops, roots Pasture Forage vegetables Edible-type crops by-products Conversion by-products


0.093 0 0 0 1.213 0 0.521 0.034 0 0.147 0.417 0.074 0 0

1.2 2.1 2.9 3.9 Cattle – Beef, Feed input (kg DM) per kg product OECD90 REF ASIA ALM

2.5

0 0 0 0 14.691 0 2.390 0.263 0 4.882 15.022 0.751 0 0

0 0 0 0 1.816 0 0.832 0 0 0 0.252 0 0 0

WORLD

0.090 0 0 0 2.971 0 0.685 0.040 0 0 0.114 0 0 0

4.620 0 0 0 12.778 0 0.859 0.378 0 2.699 2.159 3.508 0 0

0 0 0 0 1.066 0 0.088 0.014 0 0.248 0.663 0.020 0 0

0 0 0 0 70.359 0 31.227 2.208 0 0 4.206 0 0 0

27 38 108 Pork, Feed input (kg DM) per kg product OECD90 REF ASIA 2.838 0.016 0 0.021 0 0 0 0.823

1.791 0.111 0.046 0.251 0 1.071 0.368 0.590

0.617 0.146 0 0.328 0 2.406 0.826 0.591

WORLD

1.573 0 0 0 62.981 0 14.907 0.579 0 0 4.961 0 0 0

2.195 0 0 0 34.453 0 13.465 0.808 0 1.154 5.771 1.154 0 0

85

59

ALM

WORLD

2.475 0.360 0.089 0.876 0 0 0.613 1.552

1.249 0.213 0.053 0.462 0 1.234 0.545 0.680

152

Non-eaten food Whole-cereals - forage Harvested-conserved grass-legume Cropland pasture - forage Meat and bone meal System external inputs

0 0 0 0 0 0

Total Feed type Crops, cereals Crops, oil Crops, pulses Crops, roots Pasture Forage-vegetables Edible-type crops by-products Conversion by-products Non-eaten food Whole-cereals - forage Harvested-conserved grass-legume Cropland pasture - forage Meat and bone meal System external inputs

Feed type Crops, cereals Crops, oil Crops, pulses Crops, roots Pasture Forage-vegetables Edible-type crops by-products Conversion by-products Non-eaten food Whole-cereals - forage Harvested-conserved grass-legume Cropland pasture - forage Meat and bone meal System external inputs

Total

1.289 0 0 0 0 0

3.7 4.6 6.2 Poultry, Feed input (kg DM) per kg product OECD90 REF ASIA 1.890 0.173 0.152 0 0 0 0 0.525 0 0 0 0 0.049 0.081

Total

0.375 0 0 0 0 0

3.278 0.038 0.031 0 0 0 0 0.417 0 0 0 0 0.042 0.085

3.213 0.161 0.132 0 0 0 0 0.342 0 0 0 0 0.044 0.178

2.9 3.9 4.1 Eggs, Feed input (kg DM) per kg product OECD90 REF ASIA

2.553 0 0 0 0 0.096

0.810 0 0 0 0 0.057

8.6

5.3

ALM

WORLD

2.601 0.106 0.093 0 0 0 0 0.619 0 0 0 0 0 0.161

2.423 0.178 0.146 0 0 0 0 0.604 0 0 0 0 0.039 0.078

3.6

3.5

ALM

WORLD

1.776 0 0 0 0 0 0 0.472 0 0 0 0 0.052 0

2.646 0 0 0 0 0 0 0.321 0 0 0 0 0.033 0

2.562 0.01 0.008 0.059 0 0 0 0.250 0 0 0 0 0.032 0.097

2.206 0 0 0 0 0 0 0.464 0 0 0 0 0.030 0

2.300 0.005 0.004 0.029 0 0 0 0.474 0 0 0 0 0.032 0.064

2.3

3

3

2.7

2.9

a

Units in kg of feed-intake in dry weight per kg of animal product produced. Based on the ‘North America and Oceania’ region *Wirsenius, 2003+. c Based on the ‘East Europe’ region *Wirsenius, 2003+. d Based on the ‘East Asia’ region *Wirsenius, 2003+. e Based on the ‘Latin America and Caribbean’ region *Wirsenius, 2003+. b


153

The Feed-Mix As stated above, the FAO gives data on feed. The use of primary commodity groups as feed is only significant for cereals, oil crops, pulses and roots and tubers. As can be seen in Table 52 these are the first four mentioned in the feed-mix table above and correspond to the primary commodity groups under the same names mentioned throughout the report. Permanent pasture has four subgroups, grouping native or oversown grass-legume systems to either temperate or tropical systems. While native systems receive no additional management, oversown systems have a relatively high management intensity; e.g. they are fertilized and irrigated. Three different inputs are grouped under the heading animal forage crops: forage vegetables, whole cereals, and cropland pasture. Wirsenius refers to forage crops as ‘grasses and legumes cultivated for harvest (that is, not for grazing), whole cereals and other fodder crops’ [Wirsenius, 2003, p.6]. Forage vegetables are only used in pork production in the REF region and in the ASIA region. As can be seen in Appendix 11, vegetables are used as feed input, even though they are not mentioned separately in the table above. Forage vegetables are assumed equal to vegetables (in terms of yield and inputs) as mentioned throughout this report. Whole-cereals refers to ‘whole-maize’, eaten as silage. The recovery rate for silage produced from whole-maize is 90%, in all regions. Whole-maize is assumed to be equal to ‘maize, green’ *FAOSTAT, 2011+ and ‘maize (silage)’ *Fisher, 2002+, in terms of production characteristics, i.e. yield. Water input data was provided by the Water Footprint Network [Hoekstra, 2008], and fertilizer requirements are based on green maize yield [FAOTSTAT, 2011] in USA and fertilizer proportion per hectare of 150:70:90 NPK in US in ’98 *FERTISTAT, 2007+. Cropland pasture consists of grass and legumes grown on land that is suitable for growing food-crops. Cropland pasture resembles oversown permanent pasture, and will be considered as such regarding inputs and management practices. Depending on the scenario yields are equal to intensively managed pastures Harvested-conserved grass-legume consists of ‘Grass-legume hay, grass-legume silage and whole-cereals silage’ [Wirsenius, 2003, p.88]. Its yield per hectare in dry matter is assumed equal to that of cropland pasture or permanent pasture, depending on the scenario. Data on yields of feedcrops is given in Appendix 7. Meat and bone meal and system external inputs constitute an insignificant part of the feed-mix. Meat and bone meal are animal-type conversion by-products. Since meat consumption will increase availability will not pose problems, as only poultry in general and pork in the ALM region have this input as part of their feed-mix. System external inputs are composed of fish, cotton oil and cotton meal. Since these flows are not considered in this study, and availability is not an issue in the calculations made by Wirsenius, their availability is not considered to be a problem here. Non-eaten food is only part of the pork feed-mix and only for the REF region, the ASIA region and the ALM region. Still, it constitutes between 8% (REF region) and 30% (ALM region) of the pork feed-mix. Of course, this input can only be an input if enough non-eaten food is available. If not, forage crops will balance the difference, as this is the preferred feed-input for pork.


154

Edible-type crops by-products are the ‘left-overs’ after harvest has taken place and/or crops have gone through a first cleaning process. Cereals, roots and tubers, sugar crops and oil crops are included in Wirsenius’ calculations: ‘cereals straw and stover’, ‘starchy roots tops’, ’sugar crops tops and leaves’, ‘oil crops by-products’. The amount available of such edible-type crops by-products depends on the harvest index; this factor determines which part of the harvest is useful as food, and which part is not. The harvest index was assumed the same as calculated by Wirsenius. The total amount of edible-type crops generated was decreased by the amount not recovered (which varies between 0% and 34%) and the amount of food crops generated was subsequently divided by the former value. This calculation yields factors indicating how much edible-type crops by-products are generated per unit of edible-type crops. The results are shown in Table 53. Table 53: Availability of edible-type crops by-products in terms of fraction of production [based on Wirsenius, 2003].

Edible-type crops by-products Cereals straw and stover Starchy roots tops Sugar crops tops and leaves Oil crops by-products

Availability 1.30 * cereals production 0.90 * roots and tubers production 0.63 * sugar crops production 1.08 * oil crops production

Conversion by-products are limited to the oil crops and sugar crops. When the oil and the sugar is extracted from the oil crops and sugar crops, these commodities are converted to vegetable oils and to sugar and sweeteners. This process yields significant amounts of by-products that are very suitable as animal feed because all protein ends up in the oil cake, while hardly any remains in the vegetable oils. Appendix 9 shows the process flows for oil crops and sugar crops. The availability of oil-type and sugartype conversion by-products is determined by the fractions given in these diagrams, and were determined on the basis of oil crop and sugar crop production and processing, and vegetable oil and sugar and sweetener production in the year 2005. Physical and Economic Availability Even though, on a regional or global scale, inputs of the feed-mixes may be available, it may not be physically possible to ensure the availability where it is necessary, or it may not be economical to make certain inputs available in certain places. Therefore, only a part of certain flows, where availability is an issue, will be taken into account. This method is also used by Wirsenius [Wirsenius, 2010]. The flows in question are: edible-type crops by-products, conversion by-products and non-eaten food. If the feed requirements are less than half of the actually available amount, either globally or regionally depending on the scenario, the feed-mix is considered ‘ok’. If not, the missing flow will be substituted with the ‘balancing flow’; i.e. the flow that is considered most appropriate for the specific animal: pasture for cattle, forage for pigs and cereals for poultry.


155

Appendix 9

Production of Oil and Sugar Crops

Production oil crops (ton/year)

Other utilities, 3% of Production Food, 10% of Production Feed, 6% of Production Seed, 2.3% of Production Waste, 2.7% of Production

Food 57 w-% of vegetable oils

Processing (76w-% of production)

Vegetable oils

Other Utilities

36 w-% of processing

37 w-% of vegetable oils

Losses 6 w-% of vegetable oils

Conversion by-products Potential feed

Figure 54: Processing scheme oil crops [based on FAO, 2010, on data for the year 2005].

Production sugar crops (ton/year)

Other utilities, 0.7% of Production Food, 1.7% of Production Feed, 1.4% of Production Seed, 1.3% of Production Waste, 0.9% of Production

Food 85 w-% of sugar and sweeteners

Processing Sugar and sweeteners

(94w-% of production)

12 w-% of processing

Other Utilities 10 w-% of sugar and sweeteners

Losses 5 w-% of sugar and sweeteners

Conversion by-products Potential feed

Figure 55: Processing scheme sugar crops [based on FAO, 2010, on data for the year 2005].


156

Appendix 10

Land Potential

The extent of land available in the four regions was estimated using data from Global Agro-Ecological Zones study by the FAO and the IIASA [Fisher, 2002]. Gross extents of land with rain-fed cultivation potential are given for 22 regions or countries. These were grouped to correspond most closely to the four regions in this study, as shown in Table 54. Table 54: Division of the ‘global agro-ecological zones-study’ regions.

Region

FAO/IIASA regions

OECD90 REF ASIA ALM

North America, Northern Europe, Southern Europe, Western Europe, Oceania, Japan Eastern Europe, Russian Federation, Central Asia Polynesia, Western Asia, Southeast Asia, South Asia, East Asia Caribbean, Central America, South America, Eastern Africa, Middle Africa, Northern Africa, Southern Africa, Western Africa

The extents of cultivable land suitable are shown in Table 55. These values give a ‘gross extent’, which does not mean this is actually available. According to the Fisher et al, between 10 and 30% of gross suitable areas may not be available for agriculture *Fisher, 2002+. The ‘net’ estimate is also shown between brackets in Table 55. Table 55: Global and regional gross and net extents of cultivable land [based on Fisher, 2002].

Region

OECD90 REF ASIA ALM WORLD

Land with cultivation potential – gross extent (106 ha) (net extent range – 70-90% of gross extent) VS+S MS 482.7 (337.9-434.4) 226.3 (158.2-203.4) 490.2 (343.1-441.2) 1658.2 (1160.7-1492.4) 2,857.4 (2,000.2-2,571.7)

224.3 (157.0-201.9) 154.5 (108.2-139.1) 100.2 (70.1-90.2) 315.3 (220.7-283.8) 764.3 (535.0-687.9)

VS+S+MS 707 (494.9-636.3) 380.8 (266.6-342.7) 590.4 (413.3-531.4) 1973.5 (1381.4-1776.2) 3,621.7 (2,535.2-3,259.5)

Notes: It is assumed that similar yields can be attained on very suitable, suitable and moderately suitable areas which, however, leads to overestimation of attainable production quantities, because yields are most likely lower on less suitable areas. The extent of land with cultivation potential which is deemed suitable for agricultural purposes differs per scenario; the VS+S+MS areas fit the A scenarios, while only the VS+S areas fit the B scenarios.


157

Appendix 11

Virtual Water Content

Regional averages are needed to calculate future water use. These are, however, not available as such. The Water Footprint Network provides data on the global average water requirements (in terms of crop evapotranspiration) of primary crops. Furthermore, they provide data on the water needs of foodstuffs and primary crops for different countries (see also Section 3.3.2 in Chapter 2). When the regional water appropriation of agriculture is calculated using the global average water requirements of commodity groups, the results deviate substantially from the data given in Table 13 and Table 14. Because certain crops are not taken into account, lower estimates of appropriations are acceptable. The OECD90 region, however, shows an appropriation of over 25% more than what is shown in Table 13 and Table 14. As shown in Table 14 the total global water use in agriculture is 6,189 billion m3 per year; while the water use of the seven main vegetable commodity groups comes to 5,706 billion m3 per year. To see if more appropriate (regionally specified) values could be obtained, the water needs of primary crops in different countries were used to obtain regional averages for five crops with high global total use: rice (21.3% of total global water use in agriculture), wheat (12.4%), maize (8.6%), soybeans (4.5%) and sugar cane (3.4%). Together they account for a little over half of the total water use in agriculture. These crops are part of the commodity groups cereals, oil crops and sugar crops. This method can only be reasonably applied to staple crops that account for high global uses and are grown in all regions. This means that for the commodity groups roots and tubers, pulses, vegetables and fruits, the global average water appropriations (m3/ton) will be assumed to be reasonable estimates for regional water use. With the regionally specified data (the global averages) the OECD90 region has a water use which is 5% lower than the value given in Table 13, clearly a better choice than the result with the original calculations. The REF region, however, rose to being 43% higher than the value given in Table 13 with the regionally specified data, instead of being 20% short with the original data. This could be due to the fact that only 10 out of 27 countries in the REF region are accounted for. Values for the ASIA region did not change much; from being 3% short with the original data to being 5% short with the regionally specified data. In the ALM region, the gap between the calculated appropriation and the appropriation given in Table 13 decreased from 38% to 28%. If the regionally specified data are chosen for the OECD90 region and the ALM region, while the original global averages are maintained for the REF region and the ASIA region, 95% of the water appropriation given in Table 13 is accounted for. Crops that were not taken into account in this study (stimulants, nuts, fibre crops and some fodder crops), account for about 10% of global water use according to Hoekstra. Because irrigation efficiency is not taken into account for crops other than cereals, the values below provide a realistic approximation.


158

3

Table 56: Virtual Water Content per commodity group (m per ton) for the four regions [based on Hoekstra, 2008]

Commodity Group

Cereals Fruits Oil Crops Pulses Roots and tubers Sugar crops Vegetables 3 Total (in billion m ) 3 Actual total (see Table 13, in billion m ) Difference in % (total to actual total)

Virtual Water Content Per region per commodity (in m3 per ton) OECD90 1,099 844 2,226 3,790 375 122 264 1,043 1,099 -5%

REF 1,571 844 2,002 3,790 375 165 264 383 481 -20%

ASIA 1,571 844 2,002 3,790 375 165 264 2,732 2,828 -3%

ALM 2,069 844 2,002 3,790 375 151 264 1,291 1,781 -28%

World 1,571 844 2,209 3,790 375 165 264 5,669 5,707 -1%

The deviations in the REF region and the ALM region seem big, but because some commodity groups are left out (e.g. stimulants and nuts) gaps are to be expected. The global ‘actual total’ in Table 56 shows the total water requirements for the commodity groups produced in 2005. The difference between the total (calculated with the regional adjustments) and the original total water use shows a deviation of only 5%. The Virtual Water Content of ‘whole maize’ was set to 143.4 m3/ton, as defined by Hoekstra for ‘maize for forage and silage’ *Hoekstra, 2008+. These regional averages can be used in further calculations.


159

Appendix 12

Water Potential

Table 57 shows the global and regional renewable water resources, and corresponding water thresholds. Moderate water stress occurs at a use of 20% of the total renewable resources, while critical water stress occurs at a use of 40% of the total. Of course, if the global or regional water use does not reach either threshold, it does not mean that water stress is not experienced locally. Table 57: Global and regional renewable water resources and water stress thresholds [based on Hoekstra, 2008].

Region Total OECD90 REF ASIA ALM WORLD

Renewable water resources (km3) 40% threshold – 20% threshold – critical water stress moderate water stress

9,535.58 5,728.18 14,773.37 22,926.14 52,963.27

3,814.2 2,291.3 5,909.3 9,170.5 21,185.30

1,907.1 1,145.6 2,954.7 4,585.2 10,592.60

Data from the Water Footprint Network was supplemented by data from the FAO [AQUASTAT, 2011]. Together, they account for 170 out of 182 countries. Table 58: Countries for which data was used from AQUASTAT.

Region

Countries for which AQUASTAT data was used

OECD90 REF ASIA ALM

Ireland, New Zealand Bosnia and Herzegovina, Croatia, Estonia, Slovakia, Slovenia, Tajikistan Brunei Darussalam, Timor-Leste Bahamas, Comoros, Congo, Equatorial Guinea, Eritrea, Guinea, Guinea-Bissau, Lesotho, Niger, Occupied, Palestinian, Territories, Saint Kitts and Nevis, Sao Tome and Principe, United Republic of Tanzania, Togo, Uganda, United Arab Emirates, Uruguay, Venezuela.

The countries for which no data are given by either AQUASTAT or WFN are listed in Table 59. Their contribution is insignificant on a global and regional scale. Table 59: Countries for which data on renewable water resources is lacking.

Region

Countries for which no data on renewable water resources is given

OECD90 REF ASIA ALM

(0) Serbia and Montenegro (1) American Samoa, French Polynesia, New Caledonia, Samoa, Tonga, Vanuatu (6) Antigua and Barbuda, Dominica, Saint Lucia, Saint Vincent and the Grenadines, Seychelles (5)


160

Appendix 13

Meat consumption

Keyzer et al. [Keyzer, 2005] determined the effect of income growth (GDP measured in PPP) on meat consumption, based on the Engel curve. In general, the Engel curve describes the relationship between individual income (on the x-axis) and consumption of a good (on the y-axis). While for some commodity groups this relation may be simply positively or negatively linear, the relation could also be S-shaped, indicating that beyond a first threshold, consumption increases more steeply with increasing income, and beyond a second threshold this relative consumption increase is reduced again. This is shown in Figure 56 [Keyzer, 2005]. This model is a simplification, based on a nonlinear Engel curve which was based on data for 125 countries, for the years 1975 to 1997 [Keyzer, 2005]. Figure 56 shows that there are three different regimes related to two income thresholds and two consumption thresholds.

Figure 56: Engel curve for meat consumption. On the x-axis,

y and y represent the income thresholds. On the y-axis, c1 and

c2 represent the consumption thresholds [Keyzer, 2005].

The model which fits the curve above and the data on which it was based is represented by the following system of linear equations: a2  b2  yi   1  ( yi  y ),  ci ( yi )  a2  b2  yi ,  a2  b2  yi   3 ( yi  y ),

if yi  y ,

  if y  yi  y ,  if yi  y 

The parameters that correspond to the parameters in the equation are shown in Table 60. Table 60: Parameters of the Engel curve [Keyzer, 2005].

Parameter

Estimate

a2

-1.182

b2

8.070

1

-4.820

3

-7.090

Income thresholds:

In per capita GDP in PPP, corresponding to US$-1992


161

y

2,200 US$

y yi

9,700 US$ In ‘000 US$

Consumption thresholds:

Meat consumption per capita

c1

16.6 kg/year

c2

77.1 kg/year

Table 61 shows the number of countries in the different regions in the different income regimes in 2050. Table 61: Number of countries in income regimes - related to meat consumption - in 2050 [based on Keyzer, 2005; Van Vliet, 2010].

Regions (number of countries)

Scenarios

Number of countries in income regimes in 2050 st

OECD90 (26) REF (27) ASIA (31) ALM (98)

A1/A2/B2 A1/A2/B2 A1/B2 A2 A1 A2/B2

nd

1 regime < 2,200 US$ -

2 regime 2,200-9,700 US$ 22 30 46

rd

3 regime > 9,700 US$ 26 27 31 9 68 52

Table 62 shows the average global and regional meat consumptions in for the 3 scenarios in which meat is consumed. The meat consumption levels in B2 are halved relative to the projection made by the ‘Keyzer equation’, as explained in Chapter 6. Table 62: Global and regional meat consumption (kg/cap/year) [based on Keyzer, 2005; Van Vliet, 2010].

Scenario

Meat Consumption (kg per capita per year, margin between brackets for regions) OECD90 REF ASIA ALM

World

89.34

51.06

26.74

31.92

38.81

129.68

99.07

90.66

84.76

93.47

A2

(99.18-141.38) 106.94

(95.79-107.91) 79.02

(84.76-110.26) 53.32

(62.57-104.57) 57.41

61.95

B2

(82.30-116.19) 57.34

(77.27-83.76) 43.19

(36.59-93.21) 41.25

(25.97-84.97) 30.52

39.67

2005 A1


162

Appendix 14

Virtual Fertilizer Content

As was shown in Section 3.3.3, recommended fertilizer use varies with crop, region, and local soil characteristics and management practices. It was also shown that actual fertilizer use may depend on factors that have nothing to do with proper agricultural management, e.g. subsidies raising fertilizer use well above recommended values. Another issue important to fertilizer use rates, as was also discussed in Chapter 3, is fertilizer efficiency. Application timing and methods can significantly reduce use rates, without reducing yields. Fertilizer use is reported in FAOSTAT, separately for N, P2O5 and K2O fertilizers, but not specified per commodity group, and thus only valuable as a measure for comparison. Fertilizer use per country per crop is reported by the FAO in their FERTISTAT database. Data for 37 crops or commodity groups is given for countries in all regions for a year (or period) between 1995 and 2004. A total of 1022 instances of fertilizer use for a specific crop for a specific country are given (including for each of those 1022 instances, use of N, P2O5 and K2O fertilizer) [FERTISTAT, 2011]. This gives insight into regional differences in current fertilizer application rates. It does not, however, give insight into requirements and differences in requirements. Data on ‘nutrient removal by crops’, reported by the FAO in the Fertilizer and plant nutrition guide, are given in ‘kg/ha’ with values corresponding to high yields and to low yields, although not for all crops. Furthermore, data is given for specific crops, in terms of output per input of fertilizer, for specific crops. Two different levels of nutrient removal by crops (for high yield and low yield) are given for wheat, maize, rice, potato, sweet potato, cassava and sugar cane, other data is given for onions, tomatoes, cucumber, soybeans, beans and groundnuts. Here, recommended fertilizer use rates were based on estimates of fertilizer requirements per generated product. The method is similar to that used to estimate feed requirements and water requirements. This means that requirements are expressed in input per output, on a weight basis. This implicates a linear relationship between fertilizer input and generated output, which cannot be used for individual crops in individual countries, but does provide a reasonable approximation on a regional and global scale. As the guide, on which the projections were based, was written in 1984, and the scenarios are designed for 2050, requirements are calculated using the requirements given for the high yield variety. These are lower per generated output than the requirements for low yields, and may thus underestimate fertilizer use. On the other hand, fertilizer use in the REF region and in the ALM region seems to be much lower than the rates recommended by the FAO. Furthermore, in over 60 years (compared to the ‘80s), a lot of progress can be made in terms of agricultural management, e.g. in application methods and proper timing, reducing the required application rates. The assumptions on which the estimates for the requirements of the commodity-groups are based are given in Table 63. The data were extracted from the ‘Fertilizer and Plant Nutrition Guide’ by the FAO *FAO, 1984]. Differences between regions in fertilizer requirements were based solely on differences in crops grown, e.g. it was assumed that no rice is grown in the OECD90 and in the REF region. The levels of nutrient removal by crops give a baseline indication of fertilizer requirements, although some adjustments need to be made. Pulses and soybeans (oil crops) are both leguminous, and thus the extraction of nitrogen is much higher than the actual fertilizer requirements. Regional production of different cereals crops varies, for which M.Sc. Thesis I.Y.R. Odegard

163

adjustments were made. No adjustments were made for potential regional differences in soil conditions and subsequent variances in fertilizer requirements, because it is outside the scope of this thesis. Table 63: Basis of estimates for fertilizer requirements [based on FAO, 1984; Fertistat, 2007].

Commodity group

Basis of estimate

Cereals

Requirements are weighted per region, based on the production proportions of maize, rice and wheat. Requirements are based on those for ‘fruit trees’, which are given in kg/ha. Yield assumption: fruit yield in USA in 2005 Requirements equal to ‘soybeans’ for N, as pulses are leguminous crops. Requirements for ‘beans’, for P2O5 and K2O. N requirements based on average oil crop yield and average N use in the OECD90 in 2005. P2O5 and K2O requirements based on ‘soybeans’. Requirements are based on those for ‘sugar cane’. Requirements are weighted per region, based on the production proportions of cassava, potato and sweet potato. In the OECD90 and REF region only production of potatoes was assumed. Requirements are based on those for ‘tomatoes’. Requirements given for intensive management of grassland, yield based on Wirsenius’ * yield on average world yield of cropland phytomass (5.1 kg DM/ha) . Requirements based on green maize yield [FAOTSTAT, 2011] in USA in and fertilizer proportion per hectare of 150:70:90 NPK in US in ’98 *FERTISTAT, 2007].

Fruit Pulses Oilcrops

Sugar crops Roots and tubers

Vegetables Fodder Whole Maize

*Note: Fertilizer requirements on intensively managed grasslands are based on the world average yield of cropland phytomass. This value is higher than the yields in all regions related to permanent grassland phytomass. Such a value corresponds to the values given for Puerto Rico in Table 22 [FAO, 1984], for the cutting interval of 40 days. Further research concerning fertilizer use requirements for growth of fodder crops is recommended.

Because fertilizer efficiency and agricultural management also play a role in fertilizer uptake and therefore in fertilizer requirements, rates were adjusted in scenarios were efficient use of natural resources is high on the agenda, i.e. scenario B1, or technological development increases efficiencies, i.e. A1. These choices are shown in Table 35 in Chapter 5, along with their rationale. No allowances were made for the fact that the combination of irrigation and fertilization raises yields further. However, this is implicitly incorporated in the cereal yield projections. Virtual Fertilizer Content Table 64 shows the Virtual Fertilizer Content values, based on the methodology described above, for all commodity groups and all regions. These values represent a baseline, and improvements in efficiency will result in lower applications.


164

Table 64: Virtual Fertilizer Content (kg fertilizer/kg generated product)

OECD90

REF

ASIA

ALM

0.036533 0.009302 0.0027 (0.027) 0.0125 0.004375 0.0011 0.00275 0.019608 0.0088

0.0384 0.009302 0.0027 (0.027) 0.0125 0.004375 0.0011 0.00275 0.019608 0.0088

0.0446 0.009302 0.0027 (0.027) 0.0125 0.00347 0.0011 0.00275 0.019608 0.0088

0.038 0.009302 0.0027 (0.027) 0.0125 0.0032 0.0011 0.00275 0.019608 0.0088

0.00944 0.003721 0.018333 0.020833 0.002 0.0009 0.00075 0.004902 0.0041

0.00952 0.003721 0.018333 0.020833 0.002 0.0009 0.00075 0.004902 0.0041

0.0184 0.003721 0.018333 0.020833 0.0016 0.0009 0.00075 0.004902 0.0041

0.0118 0.003721 0.018333 0.020833 0.00174 0.0009 0.00075 0.004902 0.0041

0.014275 0.006977 0.040417 0.05 0.00775 0.0034 0.00375 0.019608 0.0053

0.0176 0.006977 0.040417 0.05 0.00775 0.0034 0.00375 0.019608 0.0053

0.0191 0.006977 0.040417 0.05 0.00712 0.0034 0.00375 0.019608 0.0053

0.0144 0.006977 0.040417 0.05 0.00833 0.0034 0.00375 0.019608 0.0053

N Cereals Fruit * Oil Pulses Roots Sugar Veggies Fodder Whole Maize

P Cereals Fruit Oil Pulses Roots Sugar Veggies Fodder Whole Maize

K Cereals Fruit Oil Pulses Roots Sugar Veggies Fodder Whole Maize

* Erratum: Since processing of the results with the values given in this table, it was discovered that the Virtual Fertilizer Content value for nitrogenous fertilizer for oil crops was calculated as being a factor 10 too small, and should actual be set to 0.027 kg/kg. This means the results concerning fertilizer use in 2050 are estimated too low. The mistake was discovered too late to make adjustments. Since these values were calculated, the FAO seems to have adjusted their data concerning oil crops; the area under cultivation seems to have doubled. This is further elaborated on in the discussion. Note: Pasture, cropland pasture and harvested-conserved grass-legume are all grouped under ‘fodder’.


165

Appendix 15

Yield Projection Calculations OECD90 Region

Ewert et al. have shown that even though yield may differ between crops and countries, the difference between relative yield changes (% change from one year to the next) tend to be small, and converge over time [Ewert, 2005]. In the future, these relative yield changes are expected to merge. Productivity estimations based on the relative yield changes thus has an important advantage: ‘yield changes can be compared and averaged across crops and countries to avoid unnecessary complexity’ [Ewert, 2005, p.106]. Technology was identified as the most important driver of productivity change. Technology development, as used by Ewert et al., refers to all measures related to crop management. Ewert et al. solely focus on food production in Europe (EU15 member countries, Norway and Switzerland). Their interpretation of the IPCC SRES is somewhat different than the one made in this study, and therefore the yield projections can only be used as a Here, the equations established by Ewert et al. were used to calculate potential productivity increases for the OECD90. It is assumed that the technological development parameters generated by Ewert are also representative for the whole of the OECD90 region. Relative yield changes in the baseline year t0 (the year 2007) were calculated as the average relative yield change for the years 2000-2007. For pulses this period was increased to include the years 1995-2007, because of fluctuations. The future change in productivity can be calculated with the equation shown in Table 65. Table 65: Future change in productivity equation [Ewert, 2005, p.105]

( )

∫

(

(

( )

( ))

)

future change in productivity ( )

( )

( )

relative yield change at t0 yearly increment in the relative yield change with reference to the baseline year t0 (calculated as ( ) ) represents the potential yield as a relative fraction of the current yield represents the actual yield as a relative fraction of potential yield in the future

As becomes clear from the equation and parameters in Table 65, future yields can be increased by increasing the potential yield and reducing the yield gap. Table 66 below show the values for the parameters which are used in the equation above. These values represent the effect that technology will have on the potential yield and the yield gap. Both are dependent on time and the scenario. Table 67 shows the relative changes in crop productivity as a result of technology development.


166

Table 66: Values of parameters that represent the effect of technology on potential yield ( different scenarios and time slices [Ewert, 2005].

Parameter

Year

( )

2020 2050 2020 2050

( )

A1 0.90 0.80 0.85 0.90

Scenario A2 0.80 0.60 0.85 0.90

( ))

and yield gap (

B1 0.60 0.40 0.85 0.90

( ))

for

B2 0.20 0.00 0.60 0.60

Table 67: Estimated relative changes in crop productivity due to technology development for the OECD90 region, the four scenarios and for the year 2050 [based on Ewert, 2005; FAO, 2010]

OECD90 Region Cereals (incl. feed) Oil Crops Roots and Tubers Sugar Crops Vegetables b Fruits Pulses

A1 1.7076 1.6388 1.3679 1.6712 1.4488 1.4488 1.0283

Relative yield change A2 1.5653 1.5103 1.2939 1.5362 1.3585 1.3585 1.0226

a

B1 1.3988 1.3600 1.2074 1.3783 1.2529 1.2529 1.0160

B2 1.0516 1.0466 1.0268 1.0490 1.0327 1.0327 1.0021

a

Based on relative yield changes averaged over the years 2000-2007, except for pulses for which the relative yield changes were averaged over the years 1995-2007 (due to extreme variations). Extrapolated for OECD90, based on the equation in Table 65 [from Ewert, 2005, p.105] and data from the FAO [FAO, 2010]. b Relative yield changes for fruit were assumed equal to those of vegetables.

The interpretation made by Ewert et al. of the IPCC SRES is somewhat different than the one made in this study, and therefore the yield projections can only be used as a reference for comparison.


167

Appendix 16

FAO Productivity Projections

Table 68: FAO productivity projections

Region

OECD90

Yield

Cereals (incl. feed) c

Oil Crops

REF

Yield

Sugar Crops Cereals (incl. feed) c

Oil Crops

ASIA

Yield

Sugar Crops Cereals (incl. feed) c

Oil Crops

Sugar Crops ALM

Yield

Cereals (incl. feed) c

Oil Crops

Sugar Crops c

Growth rate (percent p.a.) (relative yield change) 2000-2030 0.5 (1.1614) 1.5 (1.5631) -0.1 0.8 (1.27) 1.6 (1.6099) 0.0 1.0-1.3 (1.3478-1.4733) 1.8-2.1 (1.7078-1.8654) 2.2 (1.9210) 1.3-2.5 (1.4733-2.0976) 1.7-3.7 (1.6582-2.9741) 2.2 (1.9210)

2030-2050 0.7 (1.1497) 1.3 (1.2948) -0.4 0.4 (1.0831) 1.4 (1.3206) -1.5 0.7-0.2 (1.2328-1.0408) 1.4 (1.3206) 1.6 (1.3736) 0.7-1.8 (1.1497-1.4287) 1.1-2.0 (1.2446-1.4859) 1.6 (1.3736)

Includes productivity increases due to non-food uses such as biofuel.


168

Appendix 17

Results – Data

Table 69: Total production (in million tons per year and kg/cap/year), losses (kg/cap/year), feed (kg/cap/year), apparent consumption (kg/cap/year) and household and retail waste (kg/cap/year), per commodity group per scenario.

Commodity group

Total production (million tons per year) 2005 A1

Cereals Fruit Pulses Roots and tubers Vegetables Oil crops Sugar crops Eggs Meat Milk Total

Commodity group

5,475.4 628.5 77.7 577.7 999.2 1,052.1 2,453.7 70.2 730.1 2,117.6 14,182.2

Total production (kg per capita per year) 2005 A1


Commodity group

314.51 80.65 9.81 111.84 265.41 73.06 230.98 9.47 39.69 98.82 1,234.25

703.78 80.78 9.99 74.26 128.43 135.23 315.39 9.03 93.84 272.18 1,822.92

Total losses (kg per capita per year) 2005 A1



2,048.2 525.2 63.9 728.3 1,728.4 475.8 1,504.2 61.7 258.5 643.6 8,037.8

54.41 14.76 1.31 21.92 23.62 29.72 32.75 1.08 0.36 16.40 196.33

121.75 14.78 1.33 14.55 11.43 55.01 44.72 1.03 0.84 45.18 310.64

A2

B1

B2

3,697.4 810.6 101.8 1,235.9 2,567.7 928.1 2,615.7 90.6 624.2 1,075.1 13,747.1

2,216.1 1,409.4 89.7 725.7 1,263.9 750.5 1,056.1 87.8 0.0 1,212.7 8,811.9

2,744.6 1,651.4 81.4 1,007.0 2,114.7 881.5 1,250.6 82.3 362.7 1,091.8 11,268.0

A2

B1

B2

373.47 81.88 10.28 124.84 259.36 93.74 264.21 9.15 63.05 108.60 1,388.58

284.85 181.15 11.53 93.28 162.46 96.46 135.75 11.29 155.88 1,132.65

299.96 180.48 8.89 110.05 231.11 96.34 136.68 9.00 39.64 119.32 1,231.48

A2

B1

B2

64.61 14.98 1.37 24.47 23.08 38.13 37.47 1.04 0.57 18.03 223.75

49.28 33.15 1.53 18.28 14.46 39.24 19.25 1.29 25.88 202.36

51.89 33.03 1.18 21.57 20.57 39.19 19.38 1.03 0.36 19.81 208.01

169

Commodity group

Total feed per capita (kg per capita per year) 2005 A1


Commodity group

426.03 4.66 2.70 6.55 439.94

A2

B1

B2

141.51 2.37 24.99 117.70 5.77 292.34

73.57 73.57

86.67 1.28 13.40 63.09 3.11 167.55

Total apparent consumption (kg per capita per year) 2005 A1 A2


Commodity group

146.62 65.89 5.91 65.67 116.58 37.31 198.23 8.39 39.33 82.42 766.35

156.00 66.00 4.00 57.00 117.00 73.67 270.67 8.00 93.00 227.00 1,072.33

167.35 66.89 6.54 75.38 118.58 49.84 226.75 8.11 62.48 90.57 872.49

Total household and retail waste (kg per capita per year) 2005 A1 A2



113.48 2.60 24.25 125.21 6.03 271.57

23.46 8.24 0.47 5.25 15.16 2.98 15.86 1.30 3.15 13.19 89.05

49.92 16.50 0.64 9.12 30.42 11.79 43.31 2.48 14.88 72.64 251.69

26.78 8.36 0.52 6.03 15.42 3.99 18.14 1.26 5.00 14.49 99.98

B1

B2

162.00 148.00 10.00 75.00 148.00 57.22 116.50 10.00 130.00 856.72

161.39 147.45 6.43 75.08 147.45 54.04 117.30 7.97 39.28 99.52 855.92

B1

B2

25.92 18.50 0.80 6.00 19.24 4.58 9.32 1.55 20.80 106.71

25.82 18.43 0.51 6.01 19.17 4.32 9.38 1.24 3.14 15.92 103.95

170

Table 70: Total land use per scenario and total cropland use in the A2 and B2 scenarios (million hectares).

Commodity group

Total Land Use (million hectare) 2005 A1

Cereals – irrigated Cereals – rainfed Fruit Oil crops Pulses Roots and tubers Sugar crops Vegetables Pasture Harvested-conserved grass-legume Cropland pasture Whole cereals (Whole Maize) Total

Commodity group

340.5 907.5 33.1 234.8 24.8 15.5 35.5 36.9 883.5 86.7 321.2 59.7 2,979.7

A2

B1

B2

339.8 712.7 69.7 411.7 73.5 61.7 42.2 151.2 4,179.2 148.6 60.8 58.8 6,309.9

272.2 223.4 74.1 167.5 28.7 19.4 15.3 46.7 151.5 38.7 131.7 13.9 1,183.1

272.2 1,502.8 284.6 774.0 119.2 99.4 40.4 227.3 5,003.7 181.9 76.4 74.9 8,656.8

Cropland Use in A2 (million hectare) OECD90 REF

Cereals – irrigated Cereals – rainfed Fruit Oil crops Pulses Roots and tubers Sugar crops Vegetables Whole Maize Total

Commodity group

37.5 116.32 5.76 73.68 5.55 2.44 5.47 4.81 21.31 272.84

Cropland Use in B2 (million hectare) OECD90



0 615.1 52.3 301.1 74.3 53.4 24.2 100.0 1,985.9 168.5 77.7 67.5 3,520.0

33.75 179.95 23.83 130.03 7.52 4.12 4.12 11.21 28.07 422.60

ASIA

ALM

27.75 121.52 5.31 23.35 2.38 5.78 4.39 26.66 37.51 254.65

228.00 47.24 36.60 143.46 35.41 24.04 19.83 83.75 0 618.33

46.50 427.65 22.01 171.19 30.20 29.47 12.52 35.93 0 775.47

REF

ASIA

ALM

25.50 159.04 214.2 417.3 2.72 7.75 4.19 28.97 46.79 906.46

177.75 470.11 153.85 291.26 62.18 43.34 19.67 131.07 0 1,349.23

35.25 693.73 85.46 310.94 46.75 44.19 11.47 56.00 0.00 1,283.79

171

3

Table 71: Total water use in the scenarios and water use in the A2 and B2 scenarios (billion m ).

Commodity group

Total Water Use (billion m3) 2005 A1


Commodity group

5,117.5 5,531.4 530.4 2,324.1 294.5 216.6 404.9 263.8 408.2 15,091.4

A2

B1

B2

4,414.4 3,325.8 684.1 1,893.1 385.9 463.5 403.7 756.3 169.9 12,496.7

3,250.6 1,368.7 1,189.5 1,657.8 340.1 272.2 174.3 333.7 94.8 8,681.7

1,628.6 3,401.4 1,393.8 1,795.8 308.4 377.6 193.4 575.6 108.2 9,782.8

Total Water Use in A2 (billion m3) OECD90 REF


Commodity group

494.5 664.7 78.2 349.4 46.9 34.3 44.1 38.9 65.1 1,816.1

305.2 477.3 27.7 93.5 19.1 38.4 25.7 127.7 104.8 1,219.4

Total Water Use in B2 (billion m3) OECD90 REF



0 3,217.7 443.3 1,051.1 242.2 273.1 248.2 456.3 78.2 6,010.1

205.4 514.2 161.6 308.2 31.8 28.9 20.6 45.3 42.9 1,358.9

129.4 312.3 55.9 83.5 10.9 25.7 12.3 69.4 65.4 764.8

ASIA

ALM

2,925.2 193.9 334.5 812.8 174.5 197.3 200.3 400.0 0.0 5,238.5

689.5 1,990.8 243.7 637.5 145.3 193.5 133.5 189.7 0.0 4,223.5

ASIA

ALM

1,052.5 960.1 703.1 825.1 153.2 177.9 99.3 313.0 0.0 4,284.2

241.2 1,614.7 473.2 578.9 112.5 145.1 61.2 147.8 0.0 3,374.6

172

Table 72: Total fertilizer use (N, P2O5 and K2O) in the four scenarios (million tons per year).

Commodity group

Total Fertilizer – N (million tons per year) 2005 A1

Cereals – irrigated Cereals – rainfed Fruit Oil crops Pulses Roots and tubers Sugar crops Vegetables Fodder Whole Maize Total

Commodity group

71.78 129.30 5.26 2.56 0.87 1.88 2.43 2.47 291.72 22.55 530.82

Total Fertilizer – P2O5 (million tons per year) 2005 A1


Commodity group

0.00 28.87 1.94 8.71 1.33 1.27 1.35 1.30 0 2.24 47.01

24.80 44.67 2.09 17.33 1.46 0.91 1.99 0.67 72.93 10.50 177.34

Total Fertilizer Use – K2O (million tons per year) 2005 A1



0.00 83.57 4.88 1.28 0.80 2.64 1.65 4.75 0.00 4.80 104.37

0.00 34.80 3.68 19.22 3.20 5.57 5.11 6.48 0.00 2.89 80.95

29.88 53.84 3.96 38.25 3.5 3.97 7.51 3.37 291.72 13.58 449.58

A2

B1

B2

71.76 75.80 7.54 2.51 1.27 4.32 2.88 7.88 52.53 10.43 236.92

46.65 30.22 11.14 1.72 0.95 2.23 0.99 2.95 0.00 4.95 101.8

24.48 69.40 13.06 2.02 0.86 2.99 1.17 5.10 0.00 5.65 124.73

A2

B1

B2

26.57 22.22 3.02 17.01 2.12 2.13 2.35 2.15 13.13 4.86 95.56

16.12 10.44 4.43 11.67 1.59 1.08 0.81 0.81 0.00 2.30 49.24

8.98 22.75 5.22 13.74 1.44 1.47 0.96 1.39 0.00 2.63 58.57

A2

B1

B2

27.88 25.14 3.32 20.48 2.48 2.65 3.26 2.59 25.46 5.40 118.66

19.42 12.58 8.39 25.77 3.81 4.71 3.05 4.03 0.00 2.98 84.74

9.48 24.67 5.75 16.34 1.65 1.84 1.32 1.83 0.00 2.94 65.82

173