Ensemble Forecast Uncertainty of the 1 December 2011 Wasatch Windstorm

Abstract

AbstractA downslope windstorm on 1 December 2011 led to considerable damage along a narrow 50-km swath at the western base of the Wasatch Mountains in northern Utah. Operational forecasts issued by the Salt Lake City National Weather Service Forecast Office provided accurate guidance for the event at 1–2-day lead times, partially based on locally generated high-resolution numerical forecasts. Forecasters highlighted the possibility of the windstorm 4 days in advance. To address the apparent reduced uncertainty for this windstorm, three 11-member three-domain ensemble forecasts were initialized at 0000 UTC 25 November, 0000 UTC 27 November, and 0000 UTC 29 November 2011 using the Weather Research and Forecasting (WRF) Model with initial and boundary conditions supplied by Global Ensemble Forecast System Reforecast, version 2 (GEFS/R2). Eight of the 11 ensemble members from the 29 November 2011 forecast (60 h before the windstorm) generated a strong, localized windstorm with outliers arising from reduced cross-barrier flow. Analysis of kinetic energy error growth suggests that the reduced uncertainty of 60-h forecasts was not primarily a result of the underdispersion of GEFS/R2 initial and boundary conditions but was related to a regional reduction in error growth in midtropospheric flow upstream of northern Utah. The ensemble initialized 2 days earlier (27 November, 108 h before the windstorm) contains fewer members that generate strong windstorms, while no members generate a windstorm in the ensemble initialized on 25 November (156 h prior). This sudden increase in uncertainty with forecast lead time results from the sensitivity of the ensemble solutions to the lateral boundary conditions imposed by the GEFS/R2 between 0000 UTC 29 November and 0000 UTC 30 November.