The new-high resolution HYRAS global radiation dataset for Germany and its river catchments

Within the research program BMDV Network of Experts of the Federal Ministry for Digital and Transport, the Hydrometeorological raster datasets (HYRAS) of the Deutscher Wetterdienst (DWD) have been extended by one variable: the global radiation. Previously, precipitation, temperature (daily mean, maximum, minimum), and relative humidity data were available, covering Germany and its river catchments from 1951- 2020 with a spatiotemporal resolution of 5km x 5km and 1 day.

The HYRAS datasets are frequently used in various applications, such as analyzing current and historical climate conditions, as well as for the adjustment of the bias in climate model data which results in more accurate climate change impact assessments. Another application of the HYRAS datasets is the usage as input data for a water balance model. Since global radiation is an important parameter to correctly model evapotranspiration, together with the other HYRAS variables, the global radiation leads to more realistic results of the long-term development of the modeled runoff as well as high-flow and low-flow situations in river basins. Therefore, a gridded dataset of global radiation should extent as far back into the past as possible to include trends and decadal variability.

The interpolation methodology and the availability of suitable measurement data have a significant impact on data quality and thus the suitability of the gridded dataset for the aforementioned applications. Station measurements of global radiation are available only very sporadically for the period 1951-1980 for Germany and its river catchments, especially at the very beginning of the time series. Thus, station measurements of sunshine duration were also considered to interpolate the HYRAS global radiation dataset. These measurements are widely available and cover the entire area, making them a valuable resource for the interpolation. To convert the sunshine duration measurements into global radiation, a simple linear regression method according to Angstrom and Prescott can be used, which shows good results for monthly but not for daily values. Especially a large bias for low global radiation values is caused by using the simple linear regression. To improve the quality of this method the total column cloud liquid water content from ERA5 reanalysis data has been implemented as additional predictor. 

After a short overview about dataset generation, a comparison between the new HYRAS global radiation dataset, station measurements of global radiation and further reference datasets like the ERA5 reanalysis and the CM SAF SARAHv2.1 satellite dataset will be presented.