An Operational Experiment in the Statistical-Dynamical Prediction of Tropical Cyclone Motion

Charles J. Neumann National Hurricane Center, NOAA, Miami, Fla. 33124

Search for other papers by Charles J. Neumann in
Current site
Google Scholar
PubMed
Close
and
Miles B. Lawrence National Hurricane Center, NOAA, Miami, Fla. 33124

Search for other papers by Miles B. Lawrence in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

Current statistical models for the prediction of tropical cyclone motion use predictors derived from climatology, persistence, and observed geopotential height data. This paper describes an operational experiment conducted during the 1973 and 1974 Atlantic hurricane seasons whereby prognostic 500 mb height data from the National Meteorological Center's primitive equation model were also included as statistical predictors.

Both the “perfect-prog” and a “simulated-model-output-statistics” (SMOS) approach were utilized to introduce the prognostic height data into the statistical prediction equations. Compared to the current “state-of-the-art” of tropical cyclone forecasting, the perfect-prog technique gave relatively poor displacement forecasts for the first half of the 72 h forecast period but excellent forecasts for the latter half. The SMOS method performed well over the entire period but the 72.h displacement error was somewhat greater than that of the perfect-prog equations.

The results of the test are extremely encouraging and suggest that independent predictive information obtained from the numerical prognoses can be objectively used to improve the performance of current statistical tropical cyclone prediction models.

Abstract

Current statistical models for the prediction of tropical cyclone motion use predictors derived from climatology, persistence, and observed geopotential height data. This paper describes an operational experiment conducted during the 1973 and 1974 Atlantic hurricane seasons whereby prognostic 500 mb height data from the National Meteorological Center's primitive equation model were also included as statistical predictors.

Both the “perfect-prog” and a “simulated-model-output-statistics” (SMOS) approach were utilized to introduce the prognostic height data into the statistical prediction equations. Compared to the current “state-of-the-art” of tropical cyclone forecasting, the perfect-prog technique gave relatively poor displacement forecasts for the first half of the 72 h forecast period but excellent forecasts for the latter half. The SMOS method performed well over the entire period but the 72.h displacement error was somewhat greater than that of the perfect-prog equations.

The results of the test are extremely encouraging and suggest that independent predictive information obtained from the numerical prognoses can be objectively used to improve the performance of current statistical tropical cyclone prediction models.

Save