Advanced offshore renewable energy resource assessment, combining re-analysis datasets with bias correction techniques and satellite data, analysing the sensitivity of these techniques to the provided data

The present work focuses on the necessity of a deep and accurate knowledge of metocean conditions for Offshore Renewable Energy devices. In-situ measurements are accurate, but often not-available (due to malfunctions) or expensive to obtain on a large time-scale, and reanalysis datasets are characterized by an intrinsic uncertainty. The possibility to exploit wave direction in bias correction methods was therefore analyzed here. Three directional strategies were tried; their results (centred on the significant wave height’s correction) were computed and compared across four locations around the Iberian peninsula (each offering different wave and wind characteristics for an optimal cross-validation) according to apposite statistical parameters. A comparison was made with the most performing non- directional technique, and improvements in the data quality were actually observed in most methods, with the Gumbel-based techniques generally overperforming the linearly-spaced ones. The study also addressed the selection of the most appropriate time window for the data analysis: the duration (=number of years) and the instant (=exact time period) of the time window for correction factor calculation can in fact affect the result of the bias correction technique. The need to be able to obtain correction factors from an identification period, to later deploy them in a successive application window was to be addressed as well. A temporal sensitivity analysis was therefore carried out, considering two different application periods (2015-2019 and 2019-2023) to, again, cross-validate the results obtained in each case. The results shown that only few ones are actually necessary; corrections performed basing on only 1 year already provide satisfactory results, even though the dispersion among the different selected periods is still large. It significantly reduces, as more years of data are used, independently from the selected period, to the point when further enlarging the identification window brings little to no significant improvement to the bias correction. One more check was made, aimed at assessing whether the identification window could be shortened down to half a year or to a single season, in order to make the observation process cheaper and more efficient; each season was analyzed and compared to the other. It was observed that winter months correction factors overall better encompass the variety of wave heights (and relative distribution) and provide satisfactory correction factors for metocean data calibration.