Sensitivity of the WRF model wind simulation and wind energy production estimates to planetary boundary layer parameterizations for onshore and offshore areas in the Iberian Peninsula

This work aims to assess the Weather and Research Forecasting (WRF) model wind simulation and wind energy production estimates sensitivity to different planetary boundary layer parameterization schemes. Five WRF simulations considering different sets of planetary boundary layer (PBL) and surface layer (SL) parameterization schemes were performed, and their results compared to measured wind data collected at five offshore buoys and thirteen onshore wind measuring stations located in the Iberian Peninsula. The objective is to determine which of these model configurations produces wind simulations and wind energy productions estimates closest to measured wind data and wind energy production estimates derived from measurements, aiming to provide guidelines for onshore and offshore wind energy assessment studies focused on areas where measured wind data is not available and numerical modelling is necessary. This work focuses on the Iberian Peninsula, an area with intensive wind energy penetration due to its favourable wind conditions, which combined with its large coastline makes this area a promising one for the future installation of offshore wind farms. The results presented in this work show that, although no major differences are seen among the simulations in terms of wind speed and direction simulation accuracy, in terms of wind energy production estimates the differences are not negligible due to the high sensitivity of the wind energy production to the wind simulation accuracy. The PBL-SL parameterization set composed by the schemes ACM2-PX is the one with the lowest errors when compared to observed wind data, when considering all onshore and offshore sites together. The ACM2 PBL scheme combines features of local and non-local closure schemes and the PX LSM scheme provides a better parameterization of the surface meteorology, which proved to be important in the model performance. However, for offshore sites the PBL-SL parameterizations QNSE-QNSE produced the best wind energy production estimates. Due to the close dependence of each PBL and SL scheme performance on the surrounding synoptic conditions and atmospheric stability, it is expected that for different geographical areas and/or temporal periods these schemes may show different results. However, the fact that this study includes one complete year of simulation for a considerably wide geographical area, including the different synoptic conditions that typically occur in a annual cycle, provides a solid base of confidence that the conclusions drawn from this work may be applied to other periods and/or geographical areas. (C) 2014 Elsevier Ltd. All rights reserved.