Spatial biases revealed by LiDAR in a multiphysics WRF ensemble designed for offshore wind

Abstract

Numerical weather predictions (NWPs) have become essential in offshore wind energy planning and operations. Thus, rigorous assessments of NWP model performance are critical to integrating offshore wind power into existing power systems. Taking advantage of two LiDAR buoys launched off the coast of New York in 2019, we assess the performance of a multiphysics Weather Research and Forecast (WRF) model ensemble with a 1.33-km spatial resolution for estimating the power system impacts associated with New York’s offshore wind target. Our work is the first to report WRF horizontal wind speed biases not only at multiple heights above sea level but also at two locations while still considering all seasons. WRF tends to overpredict wind speeds during spring and summer and underpredict wind speeds during winter. However, the patterns in wind speed biases differ substantially between the two buoys offering compelling evidence against spatially uniform biases, which impacts the performance of numerous bias correction methods frequently used to post-process WRF data. Therefore, additional measurements of wind speeds throughout the lower atmosphere are necessary to fully characterize bias patterns. With the recent goal set by the U.S. to install 30 GW of offshore wind by 2030 — largely along the East Coast, mispredictions carry important policy implications. Absent accurate offshore wind uncertainty forecasts, power system operators throughout the Eastern Interconnection will be forced to dispatch their most expensive and likely high emitting power plants to compensate for periods of underperformance.