systems, crop diversity is still being created and preserved locally, and the importance of local in situ conservation efforts has more recently been acknowledged under Article 8 of the CBD. In situ conservation of crops and seeds on the farm or community level operates under a number of constraints, partly organizational, partly economic. These constraints can more easily be overcome if biodiversity management is part of an integrated approach-such as sustainable land management.

     It is notable that plant varieties and animal breeds -very much like farming systems-are intricately linked to languages, environmental knowledge, farming systems, and the evolution of human societies. They embody history, both in their form which is a result of selection and adaptation to human needs, and through the knowledge that is associated with them. In participatory research and selection, such knowledge has increasingly been validated and valued.

     In the contemporary context of rapid land use change, the complex coevolution of agrobiodiversity, ecosystems and human societies needs to be documented, analyzed and validated. An appropriate level for this task is the landscape. Cultural landscapes are complex but spatially bounded expressions of ecosystems that have evolved under the influence of biophysical factors as well as of human societies. They provide the context to understand how management practices have shaped the productive and characteristic landscapes of cultivated systems, and how crop knowledge fits into these patterns (Brookfield et al., 2003).

Agriculture and climate change

Agriculture contributes to climate change through the release of greenhouse gases in its production processes. It is a significant emitter of CH4 (50% of global emissions) and N2O (70%) (Bathia et al., 2004). The levels of its emissions are determined by various aspects of agricultural production: frequency of cultivation, presence of irrigation, the size of livestock production, the burning of crop residues and cleared areas. In many cases, emissions are difficult to mitigate because they are linked to the very nature of production; in a number of cases, however, technical measures can be adopted to mitigate emissions from specific sources.

     Agricultural activities account for 15% of global greenhouse gas (methane, nitrous oxide and carbon dioxide) emissions (Baumert et al., 2005). Two-fifths of these emissions are a result of land use or soil management practices. Methane emissions from cattle and other livestock account for just over a quarter of the emissions. Wetland rice production and manure management also contribute a substantial amount of methane. Land clearing and burning of biomass also contributes to carbon dioxide production.

     Changes in land use, especially those associated with agriculture, have negatively affected the net ability of ecosystems to sequester carbon. For instance the carbon rich grasslands and forests in temperate zones have been replaced by crops with much lower capacity to sequester carbon. By storing up to 40% of terrestrial carbon, forests play a key role, and despite a slow increase in forests in the northern hemisphere, the benefits are lost due to increased deforestation in the tropics (Matthews et al., 2000).

There is considerable potential in agriculture for mitigating climate change impacts. Changing crop regimes and

modifying crop rotations, reducing tillage, returning crop residues into the soil and increasing the production of renewable energy are just a few options for reducing emissions (Wassmann and Vlek, 2004).

     Climate change poses the question of risks for food security both globally and for marginal or vulnerable agroecological zones. People's livelihoods are threatened, as we know, if they lack resilience and the purchasing power to bridge production losses on their farms. The magnitude of the threat to the agricultural sector, and to small-scale farmers in particular, is thus also dependent on the performance of the non-agricultural sectors of developing economies, and on the opportunities they provide. Adaptation to climate change is therefore an important topic for AKST. The need and the capacity to adapt vary considerably from region to region, and from farmer to farmer (Smit, 1993; McCarthy et al., 2001).

     Change in water runoff by 2050 is expected to be considerable (Figure 1-19). Some regions will have up to 20% less runoff, while others will experience increases of the same order, and only few countries will have similar conditions as at present (HDR, 2006). Improving water use efficiency, adapting to the risks related to topography, and changing the timing of farming operations are some examples of adaptation that will be required.

     Adaptation has a cost and often requires investments in infrastructure. Therefore, where resource endowments are already thin, adverse impacts may be multiplied by the lack of resources to respond. Farmers are masters in adapting to changing environmental conditions because this has been their business for thousands of years. This is a knowledge base farmers will need to maintain and improve, even if climate change may pose challenges that go beyond problems tackled in the past.

Sustainability implications of AKST

A key objective of agricultural policies since the 1950s, both in industrialized and in developing countries, has been to increase crop production. In its production focus, these policies have often failed to recognize the links between agricultural production and the ecosystems in which it is embedded. By maximizing provisioning services, crop production has often affected the functioning of the supporting ecosystem services.

     In the 1960s and 1970s, for instance, irrigated agriculture was intensified in Asia and elsewhere to boost production of one major food crop: rice. The effort was underpinned with massive public investments in crop research, infrastructure and extension systems. While successful in terms of production and low commodity prices, this Green Revolution led in some cases to environmentally harmful practices such as excessive use of fertilizers or pesticides. As evidence of negative impacts on the environment-particularly on soil and water-emerged, a number of corrective measures were envisaged.

     In Indonesia, for example, a major effort was undertaken in the 1980s to introduce integrated pest management (IPM) in intensive rice production (Röling and van de Vliert, 1994). This required that farmers have better knowledge of pests and their predators-knowledge that could be used to reestablish pest-predator balances in rice agroecosystems,