During the first quarter of 2015 the United States experienced a widespread and extended episode of low surface wind speeds. This episode had a strong impact on wind power generation. Some wind farms did not generate enough cash for their steady payments, and the value of some assets decreased. Although the wind industry expressed their concerns, the episode has not received much attention from the scientific community and remains weakly understood. In this paper we aim to fill this gap and advance understanding of the underlying processes at seasonal time scales. Using retrospective climate predictions, we find that high sea surface temperatures in the western tropical Pacific ocean associated with a strongly positive phase of the North Pacific Mode played a central role to establish and maintain those wind anomalies. In a more general way it has also been shown that interannual variability of wind speed over North America is not only dominated by El Niño/Southern Oscillation but also by other sea surface temperature variations in the tropical Pacific. This new knowledge can be useful for industry stakeholders to anticipate future periods of low wind speed.

This work was funded by the EU projects S2S4E (GA 776787), EUCP (GA 776613), CLIM4ENERGY (C3S_441_Lot2_CEA), EUCLEIA (GA 607085), INDECIS (GA 690462), and MEDSCOPE (GA 690462). Omar Bellprat has been funded by the European Space Agency (ESA) Living Planet Fellowship Programme under the project VERITAS-CCI.We thank Daniel Cabezón for helping and informing on the wind drought episode and Javier García-Serrano for the insights on the interpretation of the physical processes. All the analyses have been performed using the statistical software R (R Core Team, 2015). The startR and s2dverification R packages have been used to read data sets, compute EOFs, and plot maps. Model output from the numerical experiments can be accessed through EUDAT facilities at the link http://doi.org/10.23728/b2share.71078 a8618414e9b906cfa6bb7d2cdab. The reanalysis data and models used are listed in the references. We also acknowledge Javier Vegas and Nicolau Manubens for technical support, as well as the EC-Earth consortium for the model development.

Peer Reviewed