Verification of Mesoscale Forecasts Made during MAP'88 and MAP'89

Abstract

Two experimental forecasting projects, each called Mesoscale Applications Project (MAP), were conducted jointly by the National Severe Storms Laboratory and the National Weather Service Forecast Office at Norman, Oklahoma, during 1988 and 1989. This paper focuses primarily on the verification of the MAP'88 and MAP'89 experimental forecasts, and combines the results with those from a similar experiment run in 1987, to examine the evolution of forecast skill over that three-year period. Results suggest that the severe-weather outlooks issued on a given experiment day exhibited good skill, with the skill being fairly stable over the three-year period studied. For outlooks issued the day before, the skill was notably higher in 1987 than in the subsequent years. Convective-mode forecasts ranged from poor to moderate skill levels, and did not change significantly from results obtained in 1987. Areal lightning forecasts were attempted in 1988 and 1989, with skill increasing more or less as the valid area increased, that area being defined as a circle ranging from 10 to 40 km of Norman. Advance outlooks for lightning, issued the day before the anticipated event, showed little or no skill. Some discussion of the possible reasons for the observed forecasting skill and its trends is presented. Several aspects of forecasting experiments in general are discussed also, based on experience during the MAP experiments.