food can be met by improving land and water productivity in irrigated areas (de Fraiture et al., 2007). The scope for productivity improvement in rainfed areas is equally prom­ising. In the high yield variant all additional land and water for food can be met by improving water productivity (de Fraiture et al., 2007). But there is considerable risk associ­ated with this strategy. Yield improvements in rainfed areas are more uncertain than in irrigated areas because of high risk for individual farmers. If yield improvement targets are not achieved (i.e., adoption of water harvesting techniques is low or fluctuations in production due to climate variability are too high), the shortfall has to be met mainly by imports, because the scope of area expansion is limited. The scope for irrigated area expansion is limited, though groundwater expansion by private well owners will continue.

5.4.5.3 MENA
In the MENA region the scope to expand irrigated areas is very limited due to severe water shortages. Rainfed agricul­ture is risky due to unreliable rainfall. With climate change, variation in rainfall within the year and between years will further increase, particularly in semiarid areas. Trade will play an increasingly important role in food supply.
     Table 5-21 shows the outcomes of a variant in which all high potential options are successfully implemented. The results from the WATERSIM model show that a major part of additional water use to meet future food demand can be met by increasing the output per unit of water, through appropriate investments in both irrigated and rainfed ag­riculture, thus relieving pressure on water resources. The output per unit water in rainfed areas increases by 31%. The potential in sub-Saharan Africa is highest (75%), while in OECD countries where productivity already is high the output per unit water increases by 20%. Overall the scope for enhancing water productivity in irrigated areas is higher than in rainfed areas (48% and 31%, respectively). In South Asia the output per unit of water can be improved by 62%. When multiple uses of water are encouraged and fisheries and livestock production are integrated, the output per unit of water in value terms may be even higher. Improvement of water productivity is often associated with higher fertil­izer use, which may result in increased polluted return flows from agricultural areas. A challenge for AKST is to develop ways in which the tradeoff between enhanced water produc­tivity and polluted return flows is minimized.
     While a major part of additional water demand in agri­culture can be met by improvements in water productivity on existing areas, further development of water resources is essential, particularly in sub-Saharan Africa where infra­structure is scarce. In total, irrigated areas expand by 50 mil­lion hectares (16%). In sub-Saharan Africa the expansion is largest (78%), in the MENA region the expansion is negli­gible because of severe water constraints. Agricultural water diversions will increase by 15% globally. A major challenge is to manage this water with minimal adverse impacts on environmental services, while providing the necessary gains in food production and poverty alleviation.
     In the realization of optimistic water productivity AKST plays an essential role. Challenges for AKST are listed in Table 5-22.

5.4.6 Changing preferences for meat and certified organic products
Consumer preferences are evolving for both meat-focused diets and foods that are produced using integrated nutrient management. These two trends could (both individually and collectively) lead to several important differences from the reference case presented here. Rising interest in the health and environmental impacts, among other concerns, of con­ventional agriculture has pointed many consumers towards changing dietary habits away from meat and towards prod­ucts that are certified in their use of better nutrient manage­ment practices (Knudsen et al., 2006, Steinfeld et al., 2006). As a result of the slowdown in meat demand, there is the potential for a shift in consumer preferences that would de­crease the share of meat products in the typical person's diet and emphasize nonmeat foods. The main consequence of growing consumer demand for certified products that come from integrated nutrient management, which includes both meat and nonmeat commodities, will be the shift in produc­tion toward certified practices that would impact produc­tivity. The impact on productivity depends on the region in which it is practiced, however. In industrial country regions, which already practice high-input intensity, conventional agriculture, the adoption of integrated nutrient manage­ment techniques would likely lower productivity and cause higher unit costs of production, while still providing greater satisfaction to those consumers who value such products. In regions like sub-Saharan Africa, on the other hand—where fairly low-intensity agriculture is still widely practiced—the adoption of integrated nutrient management techniques will likely cause an increase in yields, over and above the refer­ence levels.
     The IMPACT modeling framework, which was described earlier, was used to simulate these trends for comparing and contrasting with the reference case. Though the shift toward a less meat-intensive diet has the potential to be a global phe­nomenon, introduction of production techniques that practice integrated nutrient management is more practical in indus­trial country regions due to infrastructure and institutional requirements that are more readily available and applicable (see Halberg et al., 2006 for further discussion).

5.4.6.1 Specification of the low growth in meat demand policy issue
The global slowdown in the growth of meat demand is im­plemented via adjustments to the income demand elastici­ties for meat and vegetarian foods. Income demand elastici­ties for meat products (beef, pork, poultry, and sheep/goat) decline at a faster pace than in the reference case. At the same time income demand elasticities decline at a slower pace for vegetarian foods (fruits and vegetables, legumes, roots, tubers,  and cereal grains).  Elasticities for  animal products such as dairy and eggs are left the same as in the reference case. This happens globally using a differentiated set of multipliers for developed versus developing regions, and assumes that the slowdown in meat demand is stronger in the industrialized regions, compared to that in developing regions. Regional average income demand elasticities for meat and nonmeat foods for IAASTD regions are presented in Table 5-23. The effect, in general, is that the meat income