Twelve years (1966–78) of precipitation forecasts made at National Weather Service forecast offices have been evaluated with a view to determining accuracy trends. Relative precipitation frequency, which proved to be negatively correlated to forecast accuracy, accounted for generally more than 80% of the variance. Winter forecasts were better than summer forecasts, and accuracy differed insignificantly among the eastern, southern, central, and western regions. After the contribution of relative precipitation frequency was removed, accuracy improvement, amounting to about 1% in seven years, could be detected only in the eastern and central region winter forecasts.

This content is only available as a PDF.