Verification of numerical weather prediction and avalanche forecasting

This thesis is an attempt to help weather and avalanche forecasters in their decisionmaking process based on meteorological predictions by (1) quantifying some of the many uncertainties of meteorological forecasts and (2) combining numerical weather and avalanche prediction. In particular, output from two numerical weather prediction models was verified with surface data from two study sites: the ski area Whistler/Blackcomb in the Coast Mountains and the highway operation Kootenay Pass in the southern Selkirk Mountains. They represent two different mountain climates in British Columbia. The two highresolution, real-time, numerical weather forecast models that were running at U B C make daily forecasts on multiple nested grids out to 48 hours into the future. Domains with grid-point spacings of 2 km, 10 km, and 30 km were used for Kootenay Pass. At Whistler/Blackcomb, the grid-point spacings were 3.3 km and 10 km. Both weather models perform well. Temperature and wind predictions are very accurate with the post-processing Kalman-predictor correction method. Precipitation amount is under-forecast, which should be considered when used in avalanche prediction. The model with the 30 km grid spacing has comparable results to the higher-resolution model. The 2km-grid of this model performs slightly better than the corresponding lOkm-grid at the Kootenay Pass stations. 24-hour forecasts were generally more accurate than 48-hour forecasts. The forecast output for Kootenay Pass was used as input for a local avalanche forecasting model. This 24-hour avalanche forecast was verified against observed avalanche occurrence and against the 12-hour avalanche forecast with current weather observations. The combination of numerical weather prediction and numerical avalanche forecasting looks very promising. The avalanche model output for the test run with numerically predicted weather data is very similar to the run with observed weather data. This indicates that avalanches may be predicted statistically out to 24 hours into the future with high-resolution numerical weather prediction as input. However, the weather and avalanche forecast errors should be taken into account during operational use.Science, Faculty ofEarth, Ocean and Atmospheric Sciences, Department ofGraduat