Linking crop models with a geographic information system to assist decisionmaking : a prototype for the Indian semiarid tropics

Where adequate climatic, soil, and crop data exist, simulation models allow some extrapolation into these less developed areas and thus provide some insights into fertilizer behavior in different environments. Even where adequate data exist, simulation models are generally run in a mode that renders the results specific to one site. Input data can be changed to investigate crop responses to different management regimes at other sites, but decisionmakers often need information that relates to a whole region, where there may be a wide mix of soil types and weather conditions. In addition, this information needs to be presented in such a way that nonspecialists can readily pick out the important conclusions. One way in which such information can be produced is through coupling crop simulation models with a Geographic Information System (GIS), a data base and analysis system that contains the spatial data needed to run the models and that maps model outputs of interest to decisionmakers. This document outlines the need for crop model geographic information systems, the possible uses of such systems, and the methods required in their construction. It also describes, as an example, a study investigating nitrogen fertilizer efficiency in Maharashtra State in the Indian semiarid tropics, using a simulation model of the development, growth, and yield of sorghum coupled with a GIS. The spatial data bases of the GIS contain information on soils, weather, and other inputs needed by the sorghum model, and the system allows regional analyses of model output to be performed and maps to be drawn. There are substantial limitations in the example, and the results should be treated with caution, but the main objective of this document is to demonstrate the power and utility of the approach. A crop model GIS could be used in a number of ways: (1) to act as a store of information on farming systems that can be updated to produce timely statistics of current land use patterns and production levels and how they are changing over time; (2) as a short-term policy tool, in relation to forecasting for the coming year or the current season, where regional simulations of yield and resource use would give an estimate of requirements for imports of agricultural inputs and exports of commodities, aid requirements in response to a bad season, or likely international loan requirements; and (3) as a long-term policy tool, to investigate the effects on regional production and resource requirements of economic, technological, or climatic change, or of substantial policy or trade changes to the economic environment within which farmers operate. (Author abstract)