Optimizing boundary layer physics perturbations in the COSMO-DE Ensemble towards renewable energies

During the last years, the proportion of renewables in the German power supply has significantly increased. The weather-dependent nature of renewable energies, particularly of electricity generated via wind and solar power plants creates a direct link between weather and power production. The high variability of weather results in fluctuations in power output, which pose great challenges to the power grid. Therefore, reliable and accurate weather forecasts are required in order to ensure energy security and to minimize costs for grid development and maintenance.

The COSMO-DE Ensemble Prediction System (EPS), with its high resolution and ability to simulate deep convection, is a well-suited tool for producing reliable weather forecasts and quantifying their associated uncertainties. The future development of the model will be concentrated to a large degree on a more accurate simulation of weather variables that are relevant for renewable power production. In order to address these challenges, the German Weather Service has started two projects to optimize its forecasts for power production variables. Both the EWeLiNE and the ORKA research projects are funded by the Federal Ministry for the Environment and collaborate with German transmission system operators. In addition, the ORKA project works together with energy & meteo systems while the EWeLiNE project is a collaboration with the Fraunhofer Institute (IWES).

ORKA aims to optimize ensemble forecasts for renewable energies for the very short time range, which is especially relevant for net security issues. In the course of the project, ensemble power forecasts are analyzed to identify critical weather situations. The performances of COSMO-DE-EPS in these situations are improved by varying the physical parameterizations and changing initialization schemes. Particular attention is paid to the realistic simulation of atmospheric conditions at hub height, e.g. by improving physics parameterizations for the boundary layer. Initial studies have shown that during the transition from stable stratification at night to well mixed conditions during the day, it is difficult to accurately model boundary layer wind speeds. Regarding solar radiation, it becomes apparent that the simulation of fog as well as low stratus clouds needs to be revised. Additionally, the physics perturbation scheme will be improved to produce a more realistic probability density for variables that are relevant for renewable energies.

This presentation will show initial results of numerical experiments with the COSMO-DE deterministic and ensemble models, as well as outline the plans for further research in the course of the ORKA project.