Assessing boundary layer heights with ERA5, WRF, and scanning lidar measurements

Understanding the atmospheric phenomena of the offshore wind sites helps inform the potential risks from wake propagation, an issue becoming more prominent as wind farms grow larger and more densely concentrated. The atmospheric boundary layer (ABL) represents a crucial component for evaluating wake recovery as it defines the limit of the atmosphere where all convective activity occurs, and where the wind speed is affected by phenomena at the surface (heating, friction, etc.), before reaching the free atmosphere. This parameter is directly linked to stability conditions and determines the “space” for momentum flux recovery in wakes.

Currently the ABL is a parameterized variable in wake models, either defined by a theoretical model or adapted from reanalysis data such as ERA5. This becomes an issue for offshore applications as conditions are mostly stably stratified, thus highlighting the importance of correctly representing the ABL for accurate model wake propagation impacts on energy production.

Lidars present a viable solution to provide in-situ measurements that could help validate offshore ABL outputs from these models, which are highly relied upon in wind resource assessment and monitoring. This study presents a bias and sensitivity analysis of offshore lidar ABL measurements compared to ERA5 and WRF ABL model outputs. This analysis builds off Santos et al work from the GLOBE campaign of using scanning lidar measurements to validate ERA5 and WRF boundary layer heights by further investigating the sensitivity of this bias and adding a site with coastal conditions.  Two separate campaigns lasting 6+ months in the North Sea were assessed: one located completely offshore from the GLOBE campaign, and another coastal from the FLOW project. The method used for lidar ABL measurements consists of an image binarization method to retrieve mixing layer heights from vertical velocity profiles and residual layers heights from return signal profiles. Several averaging methods were assessed to combine both residual and mixing layer heights to be compatible with the singular ABL heights given by the models

The results demonstrate a gross overestimation of the ABL from both ERA5 and WRF, though WRF seems to perform slightly better than ERA5. The analysis is still ongoing, but a clear diurnal cycle sensitivity is prevalent indicating potential misrepresentation in offshore heat flux processes in the model parameterization. The lidar ABL distinction between mixing and residual layer provides an interesting perspective for convective and stable ABL conditions, both of which are represented differently depending on model parameterization. The purpose of presenting these results is to provide insight into how models could be corrected by on-site measurements by understanding their sensitivities and ultimately improving ABL representation in wake modelling.