Combination of satellite-derived and ground-based measurement data of solar irradiance and sunshine duration for Germany

Since the 1980s, satellite imagery in the visible radiation spectrum has been used to calculate gridded data of cloud information and, subsequently, of solar radiation at the Earth's surface. These satellite-based radiation data are progressively improving in quality and will experience a significant increase in spatial resolution with the current transition to the new third generation of Meteosat satellites (MTG). The German Meteorological Service (DWD) is currently expanding the use of these data to supplement conventional solar radiation measurements at ground stations and to increase the spatial coverage of solar radiation information over Germany. Within DUETT solar radiation data from 42 pyranometer stations are combined with data based on measurements from METEOSAT-SEVIRI (with a spatial resolution of about 5 km). As products, hourly data of the global horizontal irradiance (GHI) and the sunshine duration (SDU) are provided on a 2 x 2 km grid for Germany in near real-time.

Merging is performed in four main steps: 1. Calculation of hourly gridded data based on 15-minute satellite observations with 5 km resolution and subsequent interpolation to a 2 km grid. 2. Correction of errors in the satellite data related with snow cover and the use of climatological data of water vapour column. 3. Determination of systematic deviations of the corrected satellite data from the ground measurement data to achieve a ‘global’ bias correction. 4. Determination of the remaining residual deviations at the individual locations of the ground measurement stations and interpolation of these ‘local’ biases to a 2 km target grid using ‘Ordinary Kriging’ to receive a ‘regional’ bias correction.

Based on the combined grid data, additional point data are determined at the coordinates of 576 measurement sites of DWD. These pseudo station data are optimized by a subsequent correction regarding the influence of surrounding topography. For both, grid and point data, uncertainties are estimated based on three known error sources. We present the latest version of the merging procedure for the two parameters GHI and SDU. The steps of the data combination are illustrated and validation results based on cross validation are presented. Furthermore, an outlook is given on the upcoming utilisation of the high-resolution data from MTG satellites within DUETT.