Day-Ahead Predictability of Complex Terrain Flows for Wind Resource Production: A Case Study of the Washoe Zephyr

Abstract

detailed description of the meteorological conditions of the Washoe Valley (Nevada) and simulations that examine the predictability of the westerly high wind event known as the Washoe Zephyr are presented. Numerical weather model prediction skill is computed for day-ahead (24?48 h) forecasts of wind speed at a meteorological tower on the Virginia Hills range relative to a persistence forecast based on a seasonal climatology constructed of hourly mean observations. The model predictions are shown to be more skillful than a climatology based on seasonal and hourly means during winter and less skillful than the seasonal-hourly climatology (SHC) during summer. Overall skill of the forecasted winds tends to increase with finer horizontal grid spacing. Phase errors compose the largest component of the error decomposition and large phase errors are associated with the onset and decay of the diurnally forced Washoe Zephyr during summer and synoptically forced high wind events and valley rotors during winter. The correlation coefficient between forecasts and observations for all forecast horizontal grid spacings considered is shown to depend roughly linearly on the ratio of the integrated power spectral density in the synoptic band to the integrated power spectral density in the combined diurnal and subdiurnal band.