Resource use dynamics and interactions in the tropics: Scaling up in space and time

We discuss the temporal and spatial dynamics of nutrient resources and water within cropping and livestock systems, their interactions and those with other resources such as labour. Shortterm dynamics (within season) revolve around nutrient availability and losses as a function of soil moisture dynamics. Longer-term effects (multiple seasons and years) are related to residual effects of crop management in successive seasons and to changes in soil organic matter contents. Spatial patterns of resource use are consistent across different tropical farming systems. Farmers preferentially allocate manure, mineral fertilizers and labour to fields close to the homestead, resulting in strong negative soil fertility gradients away from the homestead. Livestock are the central means of concentration of nutrients within farming systems, resulting in their inequitable redistribution from common lands and poorer households to richer households. Productivity gains achieved by concentration on home plots are at the expense of long-term declining productivity on remote fields. Restricted availability of inputs leads to a form of self-organization resulting in repeating patterns of farm organization that are recognisable across sub-Saharan Africa. Principles for enhancing allocation efficiency of scarce resources are required that address the dynamics of interacting temporal and spatial scales. Managed variability that creates gradients of soil fertility can have major effects on resource use efficiency of both nutrients and water, necessitating analysis of trade-offs at farm scale. Investment decisions of farming families are shaped through complex interactions among competing demands for investment of cash and labour within and beyond farm boundaries. Combinations of socio-economic and agro-ecological conditions can provide windows of opportunity in both time and space that favour investment in particular forms of management. Past research provides a vast array of technologies to improve agricultural production, and understanding of the underlying processes. A research framework is proposed representing farm systems as sets of interacting components. This framework can be used ex-ante, to assist in targeting technologies to specific types of farmers, and for identification of more appropriate technologies. It can be used to explore short and long-term trade-offs of management strategies and to evaluate effects of policy on farms varying in resource endowment. (c) 2005 Elsevier Ltd. All rights reserved.