Probabilistic precipitation type forecasting based on GEFS ensemble forecasts of vertical temperature profiles

Freezing precipitation types can have a substantial impact on air and ground transportation, and reliable predictions of them can help limit associated safety hazards and disruptions of travel and commerce. The type of precipitation (rain, snow, freezing rain, etc.) observed during a precipitation event depends, among other factors, on temperature at various elevations in the atmosphere. Forecasts of these temperature profiles several days ahead are highly uncertain, and entail large uncertainty about the precipitation type. Probability forecasts communicate that uncertainty, and are therefore more appropriate for these types of predictions than deterministic forecasts, which only provide a single "best guess" of precipitation type.

In a new study published in the April issue of the Monthly Weather Review, CIRES and NOAA researchers at the Physical Sciences Division propose a new statistical method for producing reliable probability forecasts of precipitation type based on temperature profile forecasts from NOAA’s Global Ensemble Forecast System. By using information from the entire temperature profile (rather than just a few summary statistics) this method generates probability forecasts that are more skillful than those obtained with the method currently used by forecast offices, especially when it comes to predicting the challenging freezing precipitation types.

Freezing rain (Credit: Barb DeLuisi, NOAA)
Freezing rain (Credit: Barb DeLuisi, NOAA)
Freezing rain (Credit: Barb DeLuisi, NOAA)

Authors of Probabilistic precipitation type forecasting based on GEFS ensemble forecasts of vertical temperature profiles are Michael Scheuerer, Scott Gregory, Tom Hamill of the Physical Sciences Laboratory, and Phillip Shafer of NOAA's Meteorological Development Laboratory.