Verification of the Weather Research and Forecasting Model for Alberta

Abstract

We simulated three heavy summer precipitation events in Alberta, Canada, using the Weather Research and Forecasting (WRF) model and compared the output precipitation data with observations to determine model accuracy. Each storm was simulated multiple times, using five different cumulus parameterization schemes and three grid resolutions. The explicit scheme, when simulated using 6 km grid resolution, was found to be the most accurate scheme when simulating these heavy precipitation events. We also used the WRF model to perform daily forecasts during 2011-2012. The forecasts include 2-meter maximum and minimum temperatures, 10-meter wind speed, sea-level surface pressure, and daily accumulated precipitation. When compared against observations, the WRF forecasts showed seasonal differences and tendencies such as forecasting a smaller range of diurnal temperatures than observed. WRF was found to have high skill scores when forecasting maximum temperature against climatology and persistence forecasts, but was less skillful when forecasting minimum temperatures.