Bias correction of Copernicus Marine reanalysis for western Mediterranean wave energy applications

Reliable wave energy assessment requires reanalysis datasets that accurately represent wave-height and wave-period statistics, their joint occurrence, and the sea states that control device energy production. This study evaluates bias correction of the Copernicus Marine Mediterranean wave reanalysis for western Mediterranean wave energy applications using collocated in-situ observations from seven offshore sites. Biases in significant wave height ( H s ) and mean wave period ( T m 02 ) are quantified and corrected using mean-based, distribution-based, and conditional quantile-based methods. In addition to standard statistical and distributional diagnostics, the preservation of the joint H s – T m 02 structure is assessed, and the propagation of metocean uncertainty into wave energy flux and annual energy production (AEP) is evaluated. The original reanalysis reproduces the temporal variability of H s well, but systematically underestimates both H s and T m 02 . Distribution-based corrections improve the marginal distributions of both variables, while the conditional quantile delta mapping approach (BC-CQDM) provides the most consistent improvement in the joint H s – T m 02 occurrence structure. To assess bias-propagation sensitivity, AEP is computed for six representative WEC concepts using different published or scaled power matrices. The same metocean bias produces markedly different AEP errors across devices, ranging from overestimation for short-period devices to underestimation for longer-period or narrower-band devices. A Shapley-style signed decomposition shows that wave-period bias is the dominant contributor to AEP error for most concepts. The results demonstrate that reanalysis suitability for wave energy applications cannot be assessed from wave-resource statistics alone. Bias-correction method selection should account for the target metric, the preservation of joint wave-height–period dependence, and device-specific power-response characteristics. An application-specific decision procedure is therefore proposed to support bias-correction selection for western Mediterranean wave energy assessments.