Objective Identification of Environmental Patterns Related to Tropical Cyclone Track Forecast Errors
Abstract
The increase in skill of numerical model guidance and the use of consensus forecast techniques have led to significant improvements in the accuracy of tropical cyclone track forecasts at ranges beyond 72 hours. Identification of instances when the forecast track from an individual numerical model may be in error could lead to additional improvement in the accuracy of tropical cyclone track forecasts. An objective methodology is tested to characterize the spread among the three primary global numerical model forecast tracks used as guidance by the Joint Typhoon Warning Center. Statistically significant principal components derived from empirical orthogonal functions of mid-tropospheric height and vorticity forecast fields identify cases of large spread among model forecasts. Cases in which the three-model average forecast track resulted in a large error were characterized by a distribution of principal components such that one component was significantly different from the other two. Removal of the forecast track associated with the outlying principal component resulted in a reduced forecast error. Therefore, the objective methodology may be utilized to define a selective consensus by removing forecast tracks from consideration based on the projection of forecast fields onto empirical orthogonal functions and inspecting the distribution of the resulting principal components.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2006
- Accession Number
- ADA457263
Entities
People
- Elizabeth R. Sanabia
Organizations
- Naval Postgraduate School