Multi-decadal offshore wind power variability can be mitigated through optimized European allocation

• Main Authors: Neubacher, Charlotte, Witthout, Dirk, Wohland, Jan
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus 2021
• Subjects: North Atlantic; North Atlantic oscillation
• Online Access: https://hdl.handle.net/20.500.11850/461585; https://doi.org/10.3929/ethz-b-000461585

Wind power is a vital ingredient for energy system transformation in line with the Paris Agreement. Limited land availability for onshore wind parks and higher wind speeds over sea make offshore wind energy increasingly attractive. While wind variability on different timescales poses challenges for planning and system integration, little focus has been given to multi-decadal variability. Our research therefore focuses on the characteristics of wind power on timescales exceeding ten years. Based on detrended wind data from the coupled centennial reanalysis CERA-20C, we calculate European long-term offshore wind power potential and analyze its variability focusing on three locations with distinct climatic conditions: the German North Sea, the Greek Mediterranean and the Portuguese Atlantic coast. We find strong indications for two significant multi-decadal modes that are identified consistently using two independent spectral analysis methods and in the 20-year running mean time series. In winter, the long-term evolution of wind power and the North Atlantic Oscillation (NAO) are directly linked in Germany and Portugal. While German North Sea wind power is positively correlated with the NAO (r=0.82), Portuguese Atlantic coast generation is anti-correlated with the NAO (r=−0.91). We evaluate the corresponding potential for spatial balancing in Europe and report substantial benefits from European cooperation. In particular, optimized allocations off the Portuguese Atlantic coast and in the German North Sea allow to reduce multi-decadal generation variance by a factor of 3–10 compared with country-level approaches. ISSN:1680-7340 ISSN:1680-7359