Nowcasting Thunderstorms at Cape Canaveral, Florida, Using an Improved Neumann-Pfeffer Thunderstorm Index

摘要

The Neumann-Pfeffer Thunderstorm Index (NPTI) is used daily by the 45th WeatherSquadron during the convective season to estimate the probability of afternoon thunderstorms. The current I , developed by Charles J. Neumann in the 1960s, i based on only 13 years of data taken over 30 years ago. The index was in desperate need of an upgrade. Following the multiple regression techniques outlined by Neumann, this thesis examines whether or not including additional data would improve the performance of the NPTI. After performing the multiple regressions and retuning the regression coefficients, both NPTs were validated using a 2-year independent data set. Then, several measures of accuracy were computed to compare the current NPTI, the upgraded NOPTI, and 24-hour persistence. At lower cutoff percentages, the current NPTI and the upgraded version performed quite similarly; persistence was the worst of the three methods. However, at higher cutoff percentages, persistence out-performed both versions of the NPTI. Both NPTs still performed equally well. It is recommended that the current NPTI should continue to be used operationally since the upgraded NPTI did not offer any significant improvement. Furthermore, because persistence out-performed either NPTI at highter cutoff levels, a new forecast method must be developed and implemented immediately.