Abstract
Farming activities are major drivers of the landscape-related ecological patterning because of their multiple influences on both non-arable and arable landscape elements and mosaics. Uncertainties still remain about the way individual farmer decisions and the aggregation of their activities in space contribute to these mosaics at local landscape scales, therefore about possible levers of action in farms for ensuring sustainable landscapes. The general objective of the present study was to give an assessment of the way farms contribute to crop-mosaic patterning at local landscape scales. We developed a method for comparing five contrasted local study cases of agriculture (five farm sets), by combining a statistical modeling of crop-rotation organization in farms, from empirical data, and a crop-mosaic simulation approach, applied in each case. Statistical rules of crop rotation allocation to fields in farms were defined from tests involving the following potential allocation factors: (1) farm household-enterprise, (2) field physical environment and (3) farm territory configuration. The statistical rules setup from each set of factors were implemented in rule-based simulations separately for each of the five local study cases. We used a raster representation of fields (one pixel by field), to concentrate on crop-rotation management-related effects, independently from the shape of fields and of diverse perennial landscape elements. An index of connectedness was measured for each mosaic, and was used as an indicator of crop-clustering effects due to crop-rotation organization in farms. The results showed that crop-clustering effects at local landscape scales differed according to the following factors of crop-mosaic patterning: (a) the spatial distribution of crop-rotation allocation factors over the field mosaic, (b) the specialization of crop rotations to their allocation factors, and (c) the development of crop rotations over time. These results also demonstrated that crop rotations, even if run as regular crop sequences over time, lead to significant year to year crop-mosaic variability. The results finally made explicit the differences between local study cases. We conclude from these results by pinpointing trade-offs to find between farm and landscape management.