A Statistical Comparison of the Forecasts Produced by the NGM and LFM for the 1987/88 Cool Season

Abstract

Forecasts from the National Meteorological Center's (NMC) nested grid model (NGM) and limited-area fine-mesh model (LFM) were compared objectively for the 1987/88 cool season. Mean values of various predicted variables were computed for each model for each 6-h projection during the 0- to 48-h forecast period; the mean values from the two models were then compared to see how the models’ forecasts changed with increasing forecast projection. The results showed a tendency for the NGM to warm (cool) with increasing projection in the eastern (western) United States. The LFM warmed (cooled) with time in the central (western) United States. The relative humidity in the NGM tended to increase with increasing projection, especially at 500 mb. Both models generally overpredicted the frequencies of small amounts of precipitation; the NGM underpredicted the frequency of the larger amounts of precipitation. In addition to the mean value comparisons, the models’ forecasts were evaluated on the basis of correlation with observed temperature, cloud cover, and precipitation. In general, the forecasts produced by the NGM were slightly better correlated with the observed weather than the forecasts produced by the LFM.