Modelling spatial dynamics of cropping systems to assess agricultural practices at the catchment scale
Résumé
Assessing the environmental impacts of agricultural practices at the catchment scale increasingly involves the use of spatially distributed models that include cropping systems as input. Use of these models for diagnosis and evaluation requires large datasets at large spatial and temporal scales. The description of spatial dynamics of cropping systems at a fine resolution (i.e. field level) is particularly needed. As these data are scarce or not available, our objective was to propose a model that first reconstructs spatial dynamics of past cropping systems from available data, and second constructs alternative spatial cropping systems, corresponding to agricultural practice scenarios. Classification systems for farms and fertilisation practices were defined to classify farm-level organisation for crop-succession and crop management strategies, in particular nitrogen fertilisation. Winter cover crops and multiple strategies per crop at the farm level can be represented with this model. Cropping systems are modelled in three steps: we modelled crop cover succession in summer with Markov chains based on empirical data and in winter with rules based on expert agronomic knowledge: lastly a Knapsack-based algorithm was used to allocate a crop management system to fields with constraints on several crop management strategies per crop type. Results of a case study performed on the Fremeur catchment (western France) show that this model respects the main constraints for cropping system modelling to assess agricultural practices at the catchment scale. From a scenario discussed with stakeholders and regional authorities characterised by stable areas of summer crops, the decrease in bare soil area in winter and overall reduction in total nitrogen fertilisation for different types of crops and farms we demonstrate the model's ability to simulate coherent cropping systems. Thus, if faced with a lack of data or knowledge about the local factors that drive changes in cropping systems, the model can construct the spatial dynamics of cropping systems, which can be used as input to distributed nitrate transfer models to assess agricultural practices at the catchment scale.