Ship-based lidar measurements for validating ASCAT-derived and ERA5 offshore wind profiles

Abstract

The accurate characterisation of offshore wind resources is crucial for the efficient planning and design of wind energy projects. However, the scarcity of in situ observations in marine environments requires the exploration of alternative and reliable data sources. In response to this challenge, this study presents a comprehensive comparison between wind profiles derived from the Advanced Scatterometer (ASCAT) satellite observations and the ECMWF Reanalysis fifth-generation (ERA5) dataset against ship-based lidar measurements in the northern Baltic Sea. The aim is to investigate the applicability of ship-based lidar measurements for validating these datasets and to better understand the reliability, accuracy, and limitations of ASCAT- and ERA5-derived wind statistics for offshore wind characterisation at wind turbine operating heights. To extrapolate ASCAT observations from sea level to turbine rotating heights, a mean correction of atmospheric stability effects based on ERA5 and a probabilistic adaptation of the Monin–Obukhov similarity theory were implemented. The comparison between the two gridded datasets, extrapolated ASCAT and ERA5, reveals an overall good agreement in average wind speeds at 100 m height, with ASCAT exhibiting overall mean wind speeds approximately 0.6 m s−1 higher than ERA5 across the entire study region. However, excluding regions within 40 km of the coastline reduces this bias to around 0.4 m s−1, highlighting the negative impact of coastal contamination in ASCAT measurements and the difficulties ERA5 faces in accurately capturing wind conditions in complex coastal areas due to its coarse resolution. The validation against the ship-based lidar measurements shows a comparable performance of both datasets, with bias below ±0.2 m s−1 at heights between 90–170 m, with an overestimation by ASCAT and underestimation by ERA5. Both datasets show deteriorating performance with height, which is particularly notable in ASCAT profiles, with rapidly increasing biases above 170 m, peaking at around 0.5 m s−1 at 270 m.