Improved Representation of Horizontal Variability in Wind Energy Forecasts Over Complex Terrain: Application to Cold Pool Events

The presence of cold-air pools (CAPs), which are characterized by low wind speeds and strong temperature gradients, can drastically affect the power output of wind farms located in regions of complex terrain. Moreover, CAPs are seen as a major challenge for wind energy forecasting because their dynamics are not well represented in mesoscale atmospheric models. In this work, we demonstrate two recent model developments that improve numerical representation of horizontal gradients, leading to more accurate CAP forecasts. First is a new diffusion scheme for potential temperature, which uses an updated finite difference stencil to calculate "truly horizontal" gradients. Second is a three-dimensional planetary boundary layer (3DPBL) scheme, which, unlike traditional one-dimensional PBL schemes, accounts for horizontal turbulent fluxes of momentum, heat, and moisture. Both model developments, implemented within the Weather Research and Forecasting (WRF) model, are validated in idealized tests. They are then applied to a realistic cold pool event observed in the Columbia River basin during the Second Wind Forecast Improvement Project (WFIP2), showing a reduction in boundary-layer wind speed bias of up to 20% compared to a standard model configuration. Beyond CAP events, the model developments presented here have the potential to improve forecasts of complex terrain flows generally, and will become increasingly useful as wind energy forecasts progress to sub-kilometer resolution.