Precipitation has significant social and economic impacts on highly populated areas such as southern Ontario through weather events associated with severe rainfall. Although steady improvement is made in quantitative precipitation forecasts (QPF), the skills in QPF are still very low. In particular, during the warm season/early warm season the skill levels are the lowest at the time of year when the area of heavy precipitation associated with small-scale weather systems is largest. The high social and economic impact of heavy/extreme precipitation, together with the low forecast skill, motivates this research to evaluate mesoscale numerical weather prediction model performance in QPF. The objectives of this study are to assess the mesoscale model capability in prediction of severe and/or high-impact precipitation events in southern Ontario through a case study.

A series of the rainfall events that occurred during the middle of May 2000 over southern Ontario are examined, with an emphasis on the period of May 10-12, 2000. This early warm season rainfall event is characteristic of positive temperature anomalies of 10-15oC, typical of temperatures in the warm season. In this study, the severe rainfall event has been simulated at very high resolutions using a mesoscale model MC2 with a nesting technique. The MC2 simulated mean-sea-level pressure field showed that on 06 UTC May 10, 2000, there was a low-pressure system moving into southern Ontario. The strong low-level south and southwesterly flows associated with the cyclone were dominant for moisture transport into southern Ontario, which contributed to the development of precipitation. After a high-pressure ridge moved out of the region, a low-pressure system was again dominant on May 12, 2000, which resulted in heavy rainfall over southern Ontario.

The model performance is evaluated in terms of its accuracy and skills in QPF. Quantitative analyses of model prediction accuracy and skills in QPF have been carried out through comparisons of model predicted precipitation with observations archived from the raingauge network in southern Ontario. An alternative evaluation is also performed through comparisons between simulated streamflows using a coupled-atmospheric-hydrologic model and observed streamflows.