Fog detection and fog mapping using low cost Meteosat‐WEFAX transmission

Nowcasting the extent of fog and diurnal fog dynamics is important for traffic management as well as for air‐quality studies. Imagery from operational weather satellites provide a spatial database several times a day which can be used for fog monitoring given a classification scheme which is nearly operational. Unfortunately, the use of high‐resolution imagery is often expensive and time consuming, so that it is not suitable for nowcasting at local weather offices. For that reason, a method for detecting fog based on easily accessible Meteosat‐WEFAX imagery has been developed. The algorithm uses the Meteosat VIS‐channel (C‐Format) and includes illumination correction and calibration of the WEFAX signal which provides a synthetic reflectance factor. The procedure is improved by making use of an additional image of surface texture. The discrimination between fog and other surfaces has been performed by the application of a bi‐dimensional threshold test based on both images of synthetic reflectance and texture and a final snow/ice/edge correction. The resulting binary fog image has proved to be of encouraging accuracy by comparison with a classification based on NOAA‐AVHRR data as well as visibility observations. Especially in cases of fog optical depths less than 4 or a patchy fog structure, classification results should be interpreted carefully. Binary fog masks are combined into daily fog maps which clearly show diurnal fog dynamics on a European scale.