CROP GROWTH MODELS: THEIR USEFULNESS AND LIMITATIONS
Management asks for predictive models, whereas research demands explanatory models. It is argued that for predictive purposes, well-parameterized regression or simple physiological models are the most promising way. To assist in research, however, the physiological models are usually superior as they combine explanation and integration of the underlying principles. The model should fit the particular objectives of the experiment. A promising strategy is, therefore, to develop a series of submodels of varying complexity for the different processes, while putting emphasis on simple approaches.