Modelling of Radiation-Based Thermal Stress Indicators for Urban Numerical Weather Prediction

Heat-health meteorological products would benefit from high-resolution meteorological forecasts of advanced heat-stress indices in urban areas. The goal of this study is to implement two widely used radiation-based thermal stress indices, The Wet-Bulb Globe Temperature (WBGT) and the Universal and Thermal Climate Index (UTCI), into the Canadian urbanized version of the GEM numerical weather prediction model. Intermediate important variables representing thermal load on a body, the mean radiant temperature (TMR), and a synthetic variable equivalent to the globe temperature (TG) as measured with a globe thermometer sensor, are also made available. Results from sub-kilometer (grid spacing of 250-m) meteorological forecasts over the Greater Toronto Area (GTA), Ontario, Canada, are evaluated with measurements during two cases. 15 July 2015 represents a clear-sky day whereas 19 July 2015 experiences cloudy conditions with heat warnings issued by environment and Climate Change Canada (ECCC) and the city of Toronto. Spatial distribution reveals that both WBGT and UTCI are sensitive to mesoscale features such as the lake-breeze flow, UTCI having a larger range of values. Accurate prediction of WBGT and its indicators is found in particular in clear-sky conditions. Index of scattering solar radiation from GEM is then found to be more realistic than fixed values. Results for the cloudy day are still relatively accurate although more sensitive to the cloud representation in the atmospheric model that will require improvements. The importance of using a distinct definition of TMR-[WBGT] and TMR-[UTCI] is highlighted, as differences up to 50 % are found due to a different energy partitioning. TMR and TG are linked through an hysteresis cycle and TG forecasting is sensitive to the wind speed and to the parametrization of the convective coefficient at the globe sensor’s surface, with the current ISO formulation being the most accurate.