On the development of a spatially transferable bias-correction framework: Assessing Spatial

All stages of marine studies and operations base their success on the accurate knowledge of metocean conditions at the site of interest. In fact, wind and wave phenomena highly affect facility design as well as operation planning in offshore environments. The scarce distribution of offshore measurement stations represents a considerable limit for this purpose. Therefore, reanalysis models, such as widely used Copernicus’ ERA5, are very helpful thanks to their spatial and temporal coverage. Nonetheless, reanalysis data introduce non-negligible uncertainties in all main geophysical variables, affecting the computation of design parameters and wave power estimates.
Several studies have already tested statistical correction methods, in order to reduce the bias between model data and on-site observations in specific locations. Following preliminary assessments previously carried out by the same authors, this study aims at the correction of spatially broad areas beyond the common correction of specific locations, combining re-analysis data with a limited number of observation points within the selected area via a novel a spatially transferable bias-correction framework.
Using the North Sea as a case study, two sub-areas of study are described within the North Sea, containing 6 and 18 measurements points respectively. The buoys provide accurate measurements of wind and wave parameters for a period covering the years 2018 and 2019. In this study, the focus is set on significant wave height (Hs), peak wave period (Tp) and 10-m wind speed (Uw).
For each area, a leave-one-out and leave-two-out (according to measurement availability) spatial cross validation approach is adopted. An already tested bias correction method, based on Quantile Mapping, is used. The technique is calibrated on a subset of the buoys, in order to interpolate the correction factors all over the area. These are used to calibrate ERA5 points contained in the area, with remaining buoys serving for result validation.
The study aims at exploring the applicability of such spatially transferable correction framework within each area, accounting for the different conditions across the ocean and the potential offered by multiple observation points in a relatively restricted area. To assess the impact of this novel framework, the variations in the uncertainty of wave power estimation (using raw and corrected ERA5 data) are analysed.