ARTIFICIAL SKILL AND VALIDATION IN METEOROLOGICAL FORECASTING

The results of a simulation study of multiple regression prediction mo dels for meteorological forecasting are reported. The effects of sampl e size, amount, and severity of nonrepresentative data in the populati on, inclusion of noninformative predictors, and least (sum of) absolut e deviations (LAD) and least (sum of) squared deviations (LSD) regress ion models are examined on five populations constructed from meteorolo gical data. Artificial skill is shown to be a product of small sample size, LSD regression, and nonrepresentative data. Validation of sample results is examined, and LAD regression is found to be superior to LS D regression when sample size is small and nonrepresentative data are present.