Analysis of model forecasts of maximum temperature before and after cold front passages in the southern plains

Forecasting maximum temperatures before and after the passage of significant cold fronts can be a challenge. The speed at which the numerical weather prediction models move the cold air is one important factor. Another is the magnitude of the warming predicted ahead of the front and the cooling behind the front. An analysis of the temperature forecasts produced by the National Centers for Environmental Prediction's Global Forecast System (GSF) and the North American Mesoscale (NAM) model, and that of the European Centre for Medium Range Weather Forecasts (ECMWF) will be presented, using Model Output Statistics (MOS) and direct model output from each model to address the latter issue. For the purpose of this investigation, a significant cold front was defined as one in which the maximum surface temperature change from one day to the next was 15 or more Fahrenheit degrees.

Data from more than thirty frontal passages at Tulsa International Airport were compiled for significant frontal passages through the southern plains from October 2007 through January 2009. Maximum temperature forecasts from the GFS-MOS, NAM-MOS and ECMWF guidance were analyzed to identify errors and biases in the guidance and, by extension, in the underlying numerical weather prediction models.

The analysis indicated that, while all the models had comparable mean absolute errors, the GFS-based guidance was notable in consistently underforecast both the warming ahead of the fronts and the cooling behind them. An awareness of these errors and biases should allow forecasters to add value to the numerical guidance of significant frontal passages.