The development and application of land use/land use intensity data from SPOT/Vegetation and Census of Agriculture data

The objective of this research is to develop the first Canada-wide, integrated land use/land cover database with supplemental land use intensity data within agricultural regions. We combined the 1998 growing season composite of Canada from processed SPOT4/Vegetation (VGT) data with thematically detailed, coarse resolution Census of Agriculture data from 1996. An enhanced, unsupervised classification procedure was applied to the VGT data to create a map with a small number of unlabelled clusters. Using known patterns of land cover, the spatial resolution of the Census data was increased in some regions to the degree that it provided clear guidance on the spectral characteristics that defined, and could therefore be used to label, particular agricultural land uses in the clustered VGT image. The results of the classification process remain consistent with the Census data. The data developed here will be useful for environmental monitoring and impact assessment, such as for endangered species modelling, an application which we demonstrate.