No abstract

The Pan and Parapan American Games (PA15) are the third largest sporting event in the world and were held in Toronto in the summer of 2015 (10–26 July and 7–15 August). This was used as an opportunity to coordinate and showcase existing innovative research and development activities related to weather, air quality (AQ), and health at Environment and Climate Change Canada. New observational technologies included weather stations based on compact sensors that were augmented with black globe thermometers, two Doppler lidars, two wave buoys, a 3D lightning mapping array, two new AQ stations, and low-cost AQ and ultraviolet sensors. These were supplemented by observations from other agencies, four mobile vehicles, two mobile AQ laboratories, and two supersites with enhanced vertical profiling. High-resolution modeling for weather (250 m and 1 km), AQ (2.5 km), lake circulation (2 km), and wave models (250-m, 1-km, and 2.5-km ensembles) were run. The focus of the science, which guided the design of the observation network, was to characterize and investigate the lake breeze, which affects thunderstorm initiation, air pollutant transport, and heat stress. Experimental forecasts and nowcasts were provided by research support desks. Web portals provided access to the experimental products for other government departments, public health authorities, and PA15 decision-makers. The data have been released through the government of Canada’s Open Data Portal and as a World Meteorological Organization’s Global Atmospheric Watch Urban Research Meteorology and Environment dataset.

The field phase of the Second Alliance Icing Research Study (AIRS II) was conducted from November 2003 to February 2004, with the main center of interest being near Mirabel, Quebec. The AIRS II project operational objectives are to: a) develop techniques/systems to remotely detect, diagnose and forecast hazardous winter conditions at airports, b) improve weather forecasts of aircraft icing conditions, c) better characterize the aircraft-icing environment and d) improve our understanding of the icing process and its effect on aircraft. In order to support the operational objectives, the following science objectives are being addressed to: a) investigate the conditions associated with supercooled large drop formation, b) determine conditions governing cloud glaciation, c) document the spatial distribution of ice crystals and supercooled water and the conditions under which they co-exist, and d) verify the response of remote sensors to various cloud particles, and determine how this can be exploited to remotely determine cloud composition. Five research aircraft were involved in the field project. These aircraft flew special flight operations over a network of ground in-situ and remote-sensing meteorological measurement systems, located at Mirabel, Quebec. Data were collected to evaluate some prototype airport weather forecasting systems, which use satellite and surface-based remote sensors, PIREPS, and numerical forecast models. The project will also be used in North America and Europe to further develop numerical forecast models, and forecast systems, which predict aircraft icing over large areas. AIRS II is an exciting collaborative effort involving approximately 26 government and university groups from Canada, the United States and Europe. It will assist in providing the aviation community better tools to avoid aircraft icing, and to improve the efficiency of airport operations.

No abstract

An object-based forecasting, nowcasting, and alerting system prototype was demonstrated during the summer 2015 Environment Canada Pan Am Science Showcase (ECPASS) in Toronto. Part of this demonstration involved the generation of experimental thunderstorm threat areas by both automated NWP postprocessing algorithms and by a pair of human forecasters. This paper first develops a rigorous verification methodology for the intercomparison of continuous as well as categorical probabilistic thunderstorm forecasts. The methodology is then applied to the intercomparison of thunderstorm forecasts made during ECPASS. Statistical postprocessing of forecasts by smoothing with optimal bandwidth followed by recalibration is found to improve the skill scores of all thunderstorm forecasts studied at all lead times between 6 and 48 h. In addition, the calibrated ensemble mean forecasts are found to be better than the calibrated deterministic thunderstorm forecasts for all lead times considered, though postprocessing of the convective rain-rate forecast gives results that are statistically comparable with the ensemble mean forecast. Thunderstorm threat areas that were automatically generated by thresholding the output of NWP-based postprocessed algorithms have better scores than those generated by human forecasters for most lead times beyond 9 h, indicating that they could be integrated as an automated tool for providing high-quality “first-guess” thunderstorm threat areas in an object-based forecasting, nowcasting, and alerting system. A unique contribution of this paper is a novel verification methodology for the fair comparison between continuous and categorical probabilistic forecasts, a methodology that could be used for other experiments involving human- and automatically generated object-based forecasts derived from probabilistic forecasts.