Modelling nutrient responses in the field

Models of the yield responses of crops to applied nutrients are a recent addition to the methods available for making fertilizer recommendations. They have a place in integrating nutrient information with information on other factors which affect yield and its response to added nutrients. This review deals with nitrogen models classified into three groups: those which predict yield-response curves based on empirical factors; those which simulate the yield response from complex simulation models of many processes regulating crop growth and the soil environment; and those which aim to simulate yield and selected processes based on simplified functional relationships which apply to a target region or industry. Three case studies representing the three classes of model are drawn from research on dryland wheat in different parts of Australia. They show examples in which models provide information which is unobtainable from experimental procedures and which provide information useful to farmers in making decisions about fertilizers.Suggestions are made for future developments in crop-nutrient modelling including further comparisons of models, linkage of models with tissue tests, modelling co-limiting nutrients, deciding on the appropriate level of detail within a model and the need for methods for calibrating and testing models on attributes other than yield alone.