Context and objectives
Preservation of wetlands as winter habitats for water birds is one of the priorities of the Belgian Nature Conservation Institute. Terrain observations indicate a fast and continuous decline in areal extent of the old florally rich grasslands in the Flemish Polders. These grasslands form an important part of the winter range habitat of large populations of migratory geese. The strong decrease in area is due to changes in agricultural practices, including a shift from permanent grasslands to croplands. The effect is twofold: part of the goose population became endangered, and damage is caused to agricultural crops as the geese shift to non-permanent grasslands and croplands.
Purpose of the study : evaluation of the technical feasibility and the cost efficiency of the use of satellite remote sensing for (a) the mapping of the present habitat condition, (2) the localisation and quantification of the changes which took place in the past, and (c) monitoring of the habitat in the future.
Expected scientific results
Classification of grasslands requires a multitemporal image set. The best results are obtained using two scenes, the first one taken at the beginning of the growing season, the second one taken at the end of the growing season. The grasslands can be subdivided into 3 classes (meadows, hay-land and grass cropland) by applying the decision tree classification method.
Since a second multitemporal image set taken at a similar period of the year was not available, the localisation of the changes was performed on the basis of a satellite image and historical maps. The legends of the map and the image classification had to be aggregated in order to obtain comparable source material, thus reducing the level of thematic detail. About 40% of the areal extent of the old grasslands has been converted to cropland between 1959 and 1993, which is unfavourable for the population of migratory geese.
Satellite remote sensing seems interesting for the mapping of a large area (the Netherlands-Belgium-France) in the context of the RAMSAR-convention. The costs related to this methodology are rather high in comparison with the results obtained. Multiple use of the information sources and infrastructure (GIS) for various applications increases the cost efficiency.