Development, validation and applications of simulation models for agroecosystems: Problems and perspectives

Abstract

Several key issues relevant to the development, validation, and applications of simulation models for soil-crop systems are discussed. Development of models with modular components providing multiple options for conceptual representation of the system components is recommended. Combining simulation models with the concepts of artificial intelligence will facilitate the development of user-interactive interfaces which permit the user to customize the model, based on expert guidance, for a specific application. Most models do not account for the spatial and temporal variabilities in input parameters. Uncertainty in model predictions resulting from such variations in the input parameters needs to be accounted for. Minimum data sets required for model development and validation as well as objective criteria for assessing model performance need to be identified. The application of crop-soil simulation models to estimate the probable success of a specific crop production management recommendation (i.e., risk analysis) and evaluating the regional variations in crop performance using spatial modelling techniques are discussed.