Crop models are increasingly used in food security assessments studies to quantify possible effects of climate and other changes on food production. Despite the opportunities that crop models provide for such studies there are clear research demands for improving crop models and allowing their use in combination with climate and other models. These include among others issues related to model uncertainty, data quality, and lack of knowledge about model integration and scaling. As crop models are typically developed for field scale application for one or few seasons methods of up-scaling crop models from the field to larger spatial and temporal scales as relevant for food security studies are of particular concern. Among the different scaling methods available the aspect of required modeling detail when changing the scale of model application is unresolved.

This paper describes results from studies to understand the implication of differences in modeling detail on simulation results. It further presents selected results on the effect of input data aggregation on simulations for models with different detail. The considered results suggest that modeling detail should depend on input data resolution. However, more systematic analyses are required to enable formalization of changing modeling detail with the scale of model application. As crop models in food security assessment studies are typically used in conjunction with other models, progress in large scale crop modeling will also depend on the data demands of other models (e.g. economic models) to which crop models are linked to.