Rainwater harvesting to retain water has    been seen in some semiarid areas in ESAP. Rainwater harvesting has shown    considerable potential in semiarid areas because it could supply limited    irrigation at the key stages of crop growth by using stored rainwater. A    number of cases in China    and India    have shown significant increase in crop productivity through rainwater    harvesting. In Gansu Province in China, for example, yields of    maize and wheat on the experimental sites increased over 50% (Liu et al.,    2005).  
           Rainwater harvesting technology is    simple for local people to install and operate. It is convenient because it    provides water at the point of consumption and family members have full    control of their own systems, reducing operating and maintenance problems.    The disadvantage is the limited supply and uncertainty of rainfall. In    addition, numerous small-scale water-harvesting and storage systems in a    basin could have similar effects on river flows and aquatic ecosystems as a    large dam and canal irrigation. For example, along the Yellow     River, bunds and plugging gullies were effective in encouraging    agriculture and in reducing erosion, but evidence showed these practices    reduced river discharge (Zhuetal., 2003).  
           Improved land management techniques and    agricultural  production  systems     have  received  growing     attention for improving water productivity of rainfed systems. Such    technology has been referred to as "green [soil] water management."    In some areas, minimum or zero tillage proved effective in improving soil    moisture and crop yields in rainfed land (Hatibu and Rockström, 2005).    Mulching, terracing, contouring and microbasins are also important in    maximizing rainfall infiltration into the soil to increase yields. No-till    and conservation agriculture maintains and improves crop yields and    resilience against drought and other hazards, while protecting and    stimulating the soil. The essential features of conservation agriculture are minimal    soil disturbance and maintenance of a permanent cover of live or dead    vegetative material. The cover protects the soil against erosion and water    loss from runoff or evaporation. A major impediment to successfully    introducing conservation agriculture is that management skills are complex.    In many ESAP countries, any production system that includes crop rotation is    complex because it calls for coherent management over more than one or two    crop seasons. Farmers who have adopted these systems need to understand them    and the reasons for the various procedures to be able to adapt them to their    needs and conditions to balance crop rotation with market requirements (Box 2-1).  
           Recently, increasing emphasis has been    on integrated rainwater and irrigation water management. Because obtaining    additional water for irrigation is difficult and water in rainfed systems is    unreliable, agricultural water management has shifted from pure rainfed or    fully irrigated systems to emphasizing intricately connected soil    conservation and supplemental, drip, ground and surface irrigation. 
        2.2.2     Development and application of modern    technology and inputs 
      2.2.2.1   High-yielding varieties—the Green Revolution 
        The    historical focus by international and national research institutes has been    food crop production technology, em-  | 
       | 
    Box 2-1. Potential of rainfed agriculture.  
      There are    different views on the potential of rainfed systems. Evidence exists for    great potential for poverty reduction from new approaches to enhancing    rainfed agricultural systems. New pro-poor small-scale, low-cost approaches    such as treadle pumps, water bags and water harvesting are key to unlocking    rainfed potential and reducing poverty on marginal rainfed lands. Although    crop yields seem low considering the amounts of land, water, labor and capita    required, new technologies are available to help farmers predict uncertain    variables such as rainfall. This improved predictability can help increase    the contribution of rainfed agriculture.  
           Others state that although rainfed    agriculture has been the focus of research for many years, gains are not forthcoming.    Dependence on approaches to enhancing rainfed agriculture involves high risk    due to climate variability, particularly affecting small scale and poor    farmers. As poor people often live in semiarid agricultural environments    where the ability to cope with weather variation is very low, the failure of    crop often means starvation or even death. A study in three semiarid    watersheds in India    showed that large scale investments in soil and water conservation did not    have a significant impact on dryland yields, at least not under prolonged    conditions of drought (Bouma and Scott, 2006).  
        phasizing    improved yield varieties—the Green Revolution. Modern plant breeding and    improved agronomy, including the use of inorganic fertilizer and pesticides,    have been components of the strategy to increase production (Friedman,    1990). Nearly three-quarters, 71%, of production growth since 1961 has been    from yield increases. Increased yields have contributed to greater food    security within developing regions and contributed to declining real prices    for food grains.  
               In the 1960s, when the International    Rice Research Institute (IRRI) was formed, breeders found the main    constraint to rice yield was the architecture of traditional tropical rice    cultivars (Khush et al., 2001). Although tall cultivars responded positively    to nitrogen fertilizers, competed well with weeds, and provided much straw    for fodder, fuel and construction, they lodged and lost yield. The Japanese    had realized the value of short-straw cultivars in the quest for high yield    and introduced the trait into rice around 1900. By the 1950s, semidwarf rice    could be found among the landraces in many Asian countries, including in    subtropical China.    Taichung Native 1 (TN1), a semidwarf cultivar from Taiwan    (China),    was first planted in the tropics in the late 1950s, but it was highly    susceptible to major diseases and insects in the tropics (Peng and Khush,    2003). In 1962, IRRI introduced dwarfness into tropical rice by crossing the    dwarf Taiwanese cultivar Dee-geo-woo-gen into the tall Indonesian cultivar    Peta. The result was IR8 (the 8th cross), the     first  lodging-resistant  and     fertilizer-responsive  cultivar.    Farmers rapidly adopted it and it became the symbol of the Green Revolution    in Asia. After the release of IR8,  |