PREDICTING THE SPATIAL-DISTRIBUTION OF BUZZARD BUTEO-BUTEO NESTING AREAS USING A GEOGRAPHICAL INFORMATION-SYSTEM AND REMOTE-SENSING

1. Predictive models of animal distributions based on habitat can be u sed to assess the likely effects of changes in landuse on a species. I n this study we developed a model of the distribution of buzzard nests in part of Argyll, Scotland. The model was tested on a further study site. 2. Habitat was described in terms of vegetation cover, derived f rom satellite imagery, and topography, using a digital terrain model t o classify altitude, slope, aspect and ruggedness. This data base was incorporated into a Geographical Information System. 3. Environmental data, in the form of areas and boundary lengths of vegetation types an d landscape classifications, were extracted from the data base for cir cular areas of various radii from the centre of 500 m grid cells cover ing each study area. We also included counts of buildings and lengths of roadways. 4. Both logistic regression analysis and discriminant fun ction analysis were used to produce classification models, which assig ned each grid cell a probability that it contained a buzzard nesting a rea. The best predictive model was based on median altitude, total bou ndary length between all vegetation categories, the amount of moorland and the length of boundary between pre-thicket forestry and open grou nd. 5. This model successfully reclassified 96 . 88% of grid cells in the areas from which it was developed and 82 . 35% in a test area. Pre vious studies have frequently predicted the distribution of species wi thin the environment, but here we were able to predict the distributio n of nesting areas within the distribution of a species.