Transforming the spatial basis of agricultural census cover data

Land cover data from the June agricultural census of England and Wales have historically been published annually at the parish level. This level of spatial resolution masks heterogeneity of land use within parishes, limiting the value of the data for detailed spatial analyses. However, variation in land characteristics across parishes can be used to model variation in land cover, thereby allowing the spatial basis of the census data to be transformed. This procedure, referred to as areal interpolation, is illustrated using two land classification systems based on soils for a study area in northern England. The results indicate that the spatially transformed data offer a better description of the distribution of land covers than that provided by the raw census data. They also demonstrate the extra value that can be obtained by combining extant data sources within a GIS framework. Potential roles for areal interpolation in environmental modelling and policy formulation are discussed briefly.