Objective Evaluation of a Large Number of 250-m Forecasts with Urban-GEM in the Context of the 2015 Pan American Games in Toronto, Canada

Impressive simulations have been recently presented with numerical weather prediction (NWP) models having sub-km grid spacing. In urban environments, these models have been shown to realistically produce small-scale features such as urban heat islands, breezes, and convergence zones. Often, they are even able to represent the impact of urban areas on clouds and precipitation. These achievements have mostly been demonstrated in the context of case studies. An effort is presented here to objectively substantiate these positive results with a series of 250-m runs with an urban configuration of the Global Environmental Multiscale (GEM) model. A total of 150 summertime integrations performed over Toronto, Canada, as part of Environment and Climate Change Canada (ECCC) science showcase project for the 2015 Pan American Games, have been used for the evaluation. The results, to be presented at the conference, indicate significant improvements for air temperature and humidity, as well as precipitation compared with ECCC's 2.5-km GEM currently run at Operations.