The limitations of spatial land use data in environmental analysis

Analysis of detailed agricultural parcel and farm location data in central Belgium is used to reveal landscape structure and to identify the processes that are responsible for this structure. Agricultural land use patterns are determined, at least partly, by the relative location of fields to farms. In particular, grassland is located closer to farms than other land uses. General-purpose land cover maps (CORINE and PELCOM) are compared to raster maps generated from the detailed parcel data. Both the CORINE and PELCOM strongly overestimate arable land use in the study area, and neither is able to correctly represent minor land use classes. The farm location–land use relationship as well as land use quantities and other spatial metrics are found to be sensitive to the format, source and resolution of the spatial data for the study area. This suggests that the use of PELCOM is not appropriate when landscape scale processes affect the composition and spatial patterns of land use. In this case, it is always better to aggregate fine resolution spatial data.