Student Seminar Series
Department of Atmospheric & Oceanic Sciences
a talk by
Ensemble predictability of a major freezing rain event in Montreal
Accurate forecasts of precipitation in winter storms are required for suitable preparations to be made by society. However, these forecasts are prone to large errors due to model physics parameterizations, uncertainty in the initial conditions and numerical error growth. One aspect of precipitation forecasts that has been less studied is the uncertainty in diagnosing the type of precipitation the forecasts predict. To further our understanding of sources of model uncertainty, we produce an ensemble forecast for a significant freezing rain event in Montreal (24th January 2017). The case chosen is notable as it represents a major forecast bust, with a Snowfall Warning having been issued by Environment Canada the day before the event. Regional soundings during the event reveal an elevated frontal inversion characteristic of freezing precipitation, with near-zero temperatures near the 850 hPa level. This layer was responsible for melting snowfall generated aloft into rain, which froze upon impact with the subzero ground surface. The marginal nature of the warm layer aloft with respect to the freezing point (0oC) promoted a large degree of uncertainty in the precipitation-type forecast. This uncertainty, and the resulting lack of predictability in precipitation type, is reflected by a large degree of variation in the dominant precipitation type over the six-hour period of heaviest precipitation in different members of the ensemble. We use ensemble sensitivity analysis to relate the variability with concurrent synoptic-scale fields to highlight atmospheric conditions for which errors may exert significant influence on the precipitation type.