Summary

The present paper describes two model-coupled approaches for the determination of rain rates from remote sensing data. The idea of both approaches for retrieving precipitation is to account for realistic and actual vertical pro®les of tempera-ture and relative humidity, which are taken from meso-scale model data. The original Convective Stratiform Technique (CST) method of Adler and Negri (1988) for estimating precipitation rates from infrared geostationary satellite data was improved with respect to regional adjustments to summertime mid-latitude situations and the use of the water vapour channel of Meteosat to detect mid-latitude convective clouds. In our enhanced CST (ECST) numerical model data (1D cloud model and mesoscale model) are used for a better adjustment to the actual atmospheric situation. The ECST is applied to a case with strong rain events over Germany, and it shows encouraging results in comparison with radar data, particularly for rain area detection. A second issue of the paper is to present a methodology for the improvement of rain retrievals from radar data. A spectral cloud model driven by pro®les of a meso-scale model is used to generate actual Z=R relations for the radar retrieval. Using a high-resolution raingauge network as a validation data set, the ®rst results show an improvement compared to standard Z=R relations. 1