Forecasting hail size by combining a NWP model with a hail growth model

We investigate the feasibility of producing maps of forecast maximum hail size by using prognostic soundings from an NWP model as input for a hail growth model. Specifically, the GEM model of the Canadian Meteorological Service is used to provide intitial data for HAILCAST. Maps depicting hail size are generated for each summer day. The forecast hailsize maps are compared with radar observations and surface hail reports for Alberta. Verification statistics indicate that that the forecast technique is skillful at predicting the main treat areas. However, the technique displays limited skill when forecasting the distribution of hail on a small spatial scale.