An evaluation and post-processing of COSMO-reanalyses for renewable energy applications

Christopher Frank, Susanne Crewell, Jan Keller, Christian Ohlwein, Andreas Hense and Sabrina Wahl

In recent years the volatility of renewable energies grew to one of the most interesting topics in the energy branch. In particular wind- and solar energy production are strongly influenced by local weather conditions. Thus, the energy transition from fossil energy sources to renewable energy sources leads to enormous variability in energy feeds. The problem with volatile energy feeds is that energy demand and supply must be balanced at any time. To ensure these equilibrium it is of great relevance to know where, when and to what extent renewable energy can be generated. Atmospheric reanalyses represent a suitable dataset to investigate the spatiotemporal variability, potential and availability of renewable energy sources.

Reanalyses combine various observations and model simulations using a physical numerical weather prediction (NWP) model and a data assimilation scheme. Thus, reanalyses provide consistent and realistic state estimates. In this study we use the high resolution regional reanalyses.

COSMO-REA developed within the Hans-Ertel Centre for Weather Research. The reanalyses are based on the NWP model COSMO of the German weather service. A 6 km horizontal resoulted reanalysis called COSMO-REA6 covers the EURO-CORDEX region with 40 vertical layers. A second reanalysis COSMO-REA2 covers an extended COSMO-DE domain with a horizontal resolution of 2 km and 50 vertical layers. The coarser reanalysis is available from 1995 to 2014 and  the finer from 2007 to 2013.

As reanalyses always provide estimates of the real atmospheric state, this study starts with an analysis of the representativity of the relevant quantities wind speed and global radiation. Especially the global radiation shows deficits caused by insufficiently known aerosols in the atmosphere and optically too thin clouds. A post-processing is applied to account for these kind of deficits in the simulated global radiation. This post-processing method will be introduced and discussed. In future, the resulting data set will be used for investigations of the variability and availability of solar and wind energy yields. Compared to observations, the benefit by using reanalyses for these kind of evaluations is the physical, spatiotemporal consistency of wind and solar simulated power yields.