An Evaluation of Landsat Thematoc Mapper (TM) and Spot- HRV Data, in Grasslands Inventory in the UK

Abstract

The principal aim of this project is to qualitatively and quantitatively evaluate satellite imagery for the mapping of unimproved semi-natural chalk grassland in lowland Britain and to develop a methodology that permits the classification of chalk grassland. The rationale behind the project is to provide various governmental and independent agencies with an operational information system capable of creating quickly and inexpensively natural resource maps, specifically chalk semi-natural grassland maps, as an aid to environmental, and ecological planning and for monitoring change. Within the study areas, major land cover classifications were evaluated with the assistance of ground data and existing topographical maps. The use of temporal and spatial features for classification of the multisensor data sets were investigated and the spectral characteristics of different chalk grassland types analysed. Detailed chalk grass sub-community categories, as defined by English Nature (EN) for the Salisbury Plain Range, were poorly represented by the spectral classes. However, broader groupings interactively chosen with ecological significance were mapped successfully with TM and SPOT. Off the main range areas isolated Special Sites of Scientific Interest (SSSI) ‘status’ chalk grass fields were also successfully mapped using July TM data. SSSI classification using a earlier spring TM image was moderately successful and a June SPOT scene produced results which were of limited value.