The effects of climate change on vector-borne diseases are currently widely discussed. Special attention has been focused on the unprecedented emergence of Bluetongue (BT) in Europe. In previous studies, low-resolution (1km2) satellite-derived data have been used to model the distribution of Culicoides imicola, the major vector in south-western Europe. To understand local differences in distribution of BT outbreaks in southern Corsica, the environment of BT-free and BT-infected sheep farms was described at a finer scale using high resolution remote sensing data (SPOT satellite, pixel of 10x10 m) and a digital elevation model (DEM). From the DEM, topographic variables such as mean altitude, slope, sunshine and aspect were obtained. Land-cover information was produced by classifying the satellite image with a supervised object-oriented nearest-neighbour method. Finally, landscape metrics were calculated to evaluate the number, diversity, length of edge and connectance of vegetation patches. A geographical information system was developed to extract these data in the neighbouring of sheep farms. As little is known on the flight range and bionomics of C. imicola, the environment was described at three scales: within a 500 m, 1 km or 2 km radius buffer around the sheep farms. Models highlighted the role of environmental variables such as latitude, sunshine, and some vegetation types such as prairies. The model developed at 1 km had the highest AUC receiver operating characteristic (ROC AUC = 0,90) which represents the best trade-off between specificity and sensitivity. Validation of this model was carried out on the same data set (sensitivity and specificity = 85%) as well as on a new data set originating from the region of Ajaccio (Corsica) situated 40 km north of the study area. Data, methods, results, possible applications in disease-free areas and limits of this approach are discussed. (Texte intégral)