WIND ENERGY FORECAST ENSEMBLES USING A FULLY-COUPLED GROUNDWATER TO ATMOSPHERE MODEL

As wind energy becomes an increasingly important component in the renewable energy portfolio, accurate wind forecasts become critical. We apply the PF.WRF model, a fully-coupled groundwater-to-atmosphere model incorporating the parallel three-dimensional variably-saturated hydrologic model ParFlow and the Weather Research and Forecasting (WRF) atmospheric model to simulate the components of the hydrologic cycle from bedrock to the top of the atmosphere. Model components are coupled via moisture and energy fluxes in the Noah land surface model. The fully-coupled model dynamically simulates important meteorological effects of interactions between the land surface and the atmosphere. In this study, we complete an ensemble of simulations for a wind energy production site on the west coast of North America using varying stochastic random fields of subsurface hydraulic conductivity and forced with meteorological data from the North American Regional Reanalysis dataset for a series of wind ramp events, and compare the simulated results with observational data. We attribute error between the modeled and observed data to uncertainty in the statistical representation of subsurface heterogeneity. We reduce subsurface uncertainty by conditioning the stochastic random field representation of subsurface hydraulic conductivity to values drawn from a “truth” simulation. This reduction of uncertainty propagates through subsurface and surface variables, and into atmospheric variables like wind speed.