Experiments with a straightforward model for the spatial forecast of fog/low stratus clearance based on multi-source data

Abstract

A straightforward model for the spatial calculation of the time of fog clearance is presented which is based on thermodynamic equations, spatial data sets such as NOAA-AVHRR satellite data, Digital Elevation Model, and horizontal and vertical meteorological observations. The model has been tested on two days with extended fog layers within the study area. The local validation of the model reveals an accuracy in fog clearance of 4 minutes by comparing the model result with meteorological observations. A spatial validation by means of a reference NOAA overpass indicates an under-estimation of the fog-covered area by the model of 2.9% at the time of the reference image. Five minutes before reference time, the spatial correspondence of the modelled and the reference fog coverage increases to 98.9%. The temporal deviation of uncoinciding pixels between reference and modelled fog image at reference time is less than ±30 minutes in 68% of the total number of fog-covered pixels and therefore matches the accuracy of local tephigram methods. However, for individual pixels a time error of up to ±60 minutes occurs for the time of fog clearance. Time errors are probably mainly due to an inaccurate estimation of fog thickness.