Assessing wave climate trends in the Bay of Biscay through an intercomparison of wave hindcasts and reanalyses

The Bay of Biscay, located in the Northeast Atlantic Ocean, is exposed to energetic waves coming from the open ocean that have crucial effects on the coast. Knowledge of the wave climate and trends in this region are critical to better understand the last decade’s evolution of coastal hazards and morphology and to anticipate their potential future changes. This study aims to characterize the long-term trends of the present wave climate over the second half of the twentieth century in the Bay of Biscay through a robust and homogeneous intercomparison of five-wave datasets (Corrected ERA-40 (C-ERA-40), ECMWF Reanalysis Interim (ERA-Interim), Bay Of Biscay Wave Atlas (BOBWA-10kH), ANEMOC, and Bertin and Dodet 2010)). The comparison of the quality of the datasets against offshore and nearshore measurements reveals that at offshore locations, global reanalyses slightly underestimate wave heights, while regional hindcasts overestimate wave heights, especially for the highest quantiles. At coastal locations, BOBWA-10kH is the dataset that compares the best with observations. Concerning long time-scale features, the comparison highlights that the main significant trends are similarly present in the five datasets, especially during summer for which there is an increase of significant wave heights and mean wave periods (up to +15 cm and +0.6 s over the period 1970–2001) as well as a southerly shift of wave directions (around ?0.4° year^?1). Over the same period, an increase of high quantiles of wave heights during the autumn season (around 3 cm year^?1 for 90th quantile of significant wave heights (SWH₉₀)) is also apparent. During winter, significant trends are much lower than during summer and autumn despite a slight increase of wave heights and periods during 1958–2001. These trends can be related to modifications in the wave-type occurrence. Finally, the trends common to the five datasets are discussed by analyzing the similarities with centennial trends issued from longer time-scale studies and exploring the various factors that could explain them.