Evaluating Atlantic Tropical Cyclone Track Error Distributions for Use in Probabilistic Forecasts of Wind Distribution

Abstract

This thesis investigates whether the National Hurricane Center (NHC) operational product for producing probabilistic forecasts of tropical cyclone (TC) wind distributions could be further improved by examining the distributions of track errors it draws upon to calculate probabilities. The track spread/skill relationship for several global ensemble prediction system forecasts is examined as a condition for a description of a full probability distribution function. The 2007, 2008, and 2009 NHC official track forecasts are compared to the ensemble prediction system model along-, cross-, and forecast-track errors. Significant differences in statistical properties were then identified among the groups to determine whether conditioning based on geographic location was warranted. Examination of each regional distribution interval suggests that differences in distributions existed for along-track and cross-track errors. Because errors for ensemble mean and deterministic forecasts typically have larger mean errors and larger variance than official forecast errors, it is unlikely that independent error distributions based on these models would refine the PDFs used in the probabilistic model. However, this should be tested with a sensitivity analysis and verified with the probability swath. Overall, conditional formatting suggests that the NHC probability product may be improved if the Monte Carlo (MC) model would draw from refined distributions of track errors based on TC location.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2010
Accession Number
ADA531556

Entities

People

  • Jay M. Neese

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Air Force
  • Chi Square Test
  • Cyclones
  • Data Science
  • Databases
  • Distribution Functions
  • Geographic Regions
  • Information Science
  • Models
  • Probabilistic Models
  • Probability
  • Probability Distribution Functions
  • Probability Distributions
  • Statistical Analysis
  • Storm Surges
  • Tropical Cyclones
  • United States

Fields of Study

  • Environmental science

Readers

  • Atmospheric Science/Meteorology
  • Computational Modeling and Simulation
  • Statistical inference.