Predicting Wet Snow Icing Risks on the Grid Edge

This presentation highlights a successful research-to-operations effort improving the prediction of power outages caused by wet snow icing. Wet snow icing is poorly understood, with no existing electric distribution engineering standards for wet snow loading, and no widely accepted meteorological standards to identify conditions when wet snow icing occurs. This work presents a method to identify wet snow icing potential using surface wet bulb temperature, in addition to an outage prediction method. GIS analytics are used to derive grid-edge impacts to sample wet snow risks.

Testing of NWP performance demonstrates the value of using a mesoscale ensemble to predict wet snow icing conditions with up to three days lead time prior to the start of a wet snow icing event. A 21-member WRF-based mesoscale ensemble was compared to a 44-member global model ensemble consisting of the North American Ensemble Forecast System, Global Forecast System, and Global Deterministic Prediction System for several major northeast US coastal cyclones. The mesoscale system generally showed greater skill for predicting precipitation phase (e.g., wet vs. dry snow) than the global system, whereas the mesoscale and global system performed similarly for the total amount of forecast precipitation through the first three forecast days.