Cumulonimbus and Thunderstorm Stepwise Multiple Regression Objective Forecast Study for Salt Lake City

摘要

The University of California BMDP2R Stepwise Multiple Regression program was used to obtain prediction equations for cumulonimbus or precipitation and for thunder at the Salt Lake City airport weather station. Fourteen years of data for the five months May through September were analyzed. The data were broken into two time periods; the spring and fall months May-June 14 and September and the summer months June 15 to August 31. The Observed 12Z Salt Lake City radiosonde (RAOB) was used to forecast the probability of occurrence of cumulonimbus or thunder from RAOB time to 2400 MST (O700Z). The five year independent data 12Z RAOB skill scores for cumulonimbus were 0.62 for spring and fall and 0.50 for the summer months. The 12Z RAOB thunder equation independent data skill scores were 0.26 for spring and fall and 0.36 for the summer months. Tables are presented to assist the forecaster in determining the probability of occurrence of cumulonimbus or thunder for the various regression equations and for single predictors. High and low probability forecasts can be used to improve the forecast confidence. Single predictors such as the average relative humidity from the surface to 400 millibars for cumulonimbus or lifting index or precipitable water for thunder can be used to obtain skill scores slightly smaller than those of the best equations. (Author)