Search Results
Abstract
Canadian Global Environmental Multiscale (GEM) numerical model output was compared with the meteorological data from an enhanced observational network to investigate the model’s ability to predict Lake Ontario lake breezes and their characteristics for two cases in the Greater Toronto Area—one in which the large-scale wind opposed the lake breeze and one in which it was in the same direction as the lake breeze. An enhanced observational network of surface meteorological stations, a C-band radar, and two Doppler wind lidars were deployed among other sensors during the 2015 Pan and Parapan American Games in Toronto. The GEM model was run for three nested domains with grid spacings of 2.5, 1, and 0.25 km. Comparisons between the model predictions and ground-based observations showed that the model successfully predicted lake breezes for the two events. The results indicated that using GEM 1 and 0.25 km increased the forecast accuracy of the lake-breeze location, updraft intensity, and depth. The accuracy of the modeled lake breeze timing was approximately ±135 min. The model underpredicted the surface cooling caused by the lake breeze. The GEM 0.25-km model significantly improved the temperature forecast accuracy during the lake-breeze circulations, reducing the bias by up to 72%, but it mainly underpredicted the moisture and overpredicted the surface wind speed. Root-mean-square errors of wind direction forecasts were generally high because of large biases and high variability of errors.
Abstract
Canadian Global Environmental Multiscale (GEM) numerical model output was compared with the meteorological data from an enhanced observational network to investigate the model’s ability to predict Lake Ontario lake breezes and their characteristics for two cases in the Greater Toronto Area—one in which the large-scale wind opposed the lake breeze and one in which it was in the same direction as the lake breeze. An enhanced observational network of surface meteorological stations, a C-band radar, and two Doppler wind lidars were deployed among other sensors during the 2015 Pan and Parapan American Games in Toronto. The GEM model was run for three nested domains with grid spacings of 2.5, 1, and 0.25 km. Comparisons between the model predictions and ground-based observations showed that the model successfully predicted lake breezes for the two events. The results indicated that using GEM 1 and 0.25 km increased the forecast accuracy of the lake-breeze location, updraft intensity, and depth. The accuracy of the modeled lake breeze timing was approximately ±135 min. The model underpredicted the surface cooling caused by the lake breeze. The GEM 0.25-km model significantly improved the temperature forecast accuracy during the lake-breeze circulations, reducing the bias by up to 72%, but it mainly underpredicted the moisture and overpredicted the surface wind speed. Root-mean-square errors of wind direction forecasts were generally high because of large biases and high variability of errors.