A Model for Decision Making Based on NWS Monthly Temperature Outlooks

Abstract

A Gaussian model for evaluating the probability of occurrence of forecast-contingent monthly average temperature and degree day outcomes is developed by use of forecast-verification data, and proposed for use in decision making. The model 1) guides current National Weather Service forecasting to ensure consistency with demonstrated skill, and 2) interprets the forecast statements i as projected perturbations of the mean and variance of a standard temperature variable t. A forecast thus specifies at each map point a local perturbed variable (t|i) that should be the basis for local decision making. A figure showing t and t|i curves on normal probability paper makes clear how significantly the probabilities of a given outcome of interest can differ for different forecast statements. Use of model results in decision making and in setting priorities for forecast development is discussed.