The scope for improvement is tremendous (Molden et al., 2007): rainfed farming covers most of the world croplands (80%), and produces most of the world's food (60-70%). Poverty is particularly concentrated in tropical developing countries in rural areas where rainfed farming is practiced (Castillo et al., 2007). Half of the currently malnourished are concentrated in the arid, semiarid and dry subhumid areas where agriculture is very risky due to extreme vari­ability of rainfall, long dry seasons, and recurrent droughts, floods and dry spells (Rockstrom et al., 2007). Current pro­ductivity is generally very low (yields generally less than half of irrigated systems and in temperate regions where water risks are much lower). Even in these regions, there is gener­ally enough water to double or often quadruple yields in rainfed farming systems. In these areas the challenge is to reduce water related risks rather than coping with absolute scarcity of water. With small investments large relative im­provements in agricultural and water productivity can be achieved in rainfed agriculture. Small investments providing 1000 m3 ha-1 (100 mm ha-1) of extra water for supplemental irrigation can unlock the potential and more than double water and agricultural productivity in small-scale rainfed agriculture, which is a very small investment compared to the  10000-15000 m3 ha-1  storage infrastructure required to enable full surface irrigation (Rockstrom et al., 2007). Provided that there are sufficient other factor inputs (e.g., N), the major hurdle for rain water harvesting and supple­mental irrigation systems is cost effectiveness. Investment in R&D for low cost small scale technologies is therefore im­portant to realize gains. This approach can address seasonal variability in rainfall (expected to increase with climate change) but have little impact in conditions of more severe interannual variability (very low rainfall), which can only be addressed by systems with storage (dams and ground-water)  or buffering  (lag in hydrologic response to that river flows are substantially maintained through drought periods).
         Climate change will require a new look at water stor­age, to mitigate the impact of more extreme weather, cope with changes in total amounts of precipitation, and cope with changing distribution of precipitation, including shifts in ratios between snowfall and rainfall. Developing more storage (reservoirs and groundwater storage) and hydraulic infrastructure provides stakeholders with more influence in determining the precise allocation to desired activities in­cluding agriculture and hydropower production.
         In the process of adapting to climate change multiple interests at the basin scale can be incorporated and man­aged, and tradeoffs with other livelihood and environmen­tal interests included in the planning (Faurés et al., 2007). Storage will itself be more vulnerable to climatic extremes resulting from climate change, and therefore be less reli­able.  Furthermore,  it will  have  proportionately  greater impacts on wetland and riverine ecosystems, which are already under stress. The arguments on the relative merits of further storage will become sharper and more pressing (Molden et al., 2007). The role of groundwater as a stra­tegic reserve will increase (Shah et al., 2007) How to plan appropriate and sustainable storage systems that address climate change is a pressing need for future AKST develop­ment.

6.8.2 Sustainable use of bioenergy

6.8.2.1 Liquid biofuels for transport
Current trends indicate that a large-scale expansion of pro­duction of 1st generation biofuels for transport will create huge demands on agricultural land and water—causing potentially large negative social and environmental effects, e.g., rising food prices, deforestation, depletion of water resources (see Chapter 4) that may outweigh positive ef­fects. The following options are currently being discussed as means to alleviating these problems.

Reducing land and water requirements through increasing yields of agricultural feedstocks. Efforts are currently fo­cused on increasing biofuel yields per hectare while reduc­ing agricultural input requirements by optimizing cropping methods or breeding higher yielding crops. For example, Brazil has been able to increase yields and reduce crop vul­nerability to drought and pests by developing more than 550 different varieties of sugar cane, each adapted to dif­ferent local climates, rainfall patterns and diseases (GTZ, 2005). Both conventional breeding and genetic engineering are being employed to further enhance crop characteristics such as starch or oil content to increase their value as energy crops. There is a great variety of crops in developing coun­tries that are believed to hold large yield potential but more research is needed to develop this potential (Cassman et al., 2006; Ortiz et al., 2006; Woods, 2006). However, even if yields can successfully be increased, several problems will persist for the production of liquid biofuels on a large scale.
          Total land area under cultivation will still need to ex­pand considerably in order to meet large-scale demand for biofuels and food production (Table 6-5).
           Land availability and quality as well as social and envi­ronmental value and vulnerability of this land differ widely by country and region and needs to be carefully assessed at the local level (FAO, 2000; WBGU, 2003; European Envi­ronment Agency, 2006). Moreover, various studies predict that water will be a considerable limiting factor for which feedstock production and other land uses (e.g., food produc­tion, ecosystems) would increasingly compete (Giampietro et al., 1997; Berndes, 2002; De Fraiture et al., 2007). In addition to these environmental problems, special care must be taken to avoid displacement and marginalization of poor people who often have weakly enforceable or informal property and land-use rights and are thus particularly vulnerable (Fritsche et al., 2005; FBOMS, 2006; The Guardian, 2007).
          Economic competitiveness will continue to be an issue. Even in Brazil, the world leader in efficient ethanol produc­tion, biofuels are competitive only under particularly favor­able market conditions. To increase total land area under production, less productive areas would have to be brought into production, either for bioenergy feedstocks directly or for other agricultural crops which may be displaced on the most productive lands. This depends on economic incentives for farmers and investments in productivity enhancements and could have strong effects on agricultural systems and further accentuate food price effects.
           Environmental concerns, associated with issues such as high-input feedstock production, the conversion of pristine land for agricultural production, the employment of trans-