Further Applications of Empirical Orthogonal Functions of Wind Fields for Tropical Cyclone Motion Studies

Thomas B. Schott Department of Meteorology, Naval Postgraduate School, Monterey, CA 93943

Search for other papers by Thomas B. Schott in
Current site
Google Scholar
PubMed
Close
,
Johnny C-L. Chan Department of Meteorology, Naval Postgraduate School, Monterey, CA 93943

Search for other papers by Johnny C-L. Chan in
Current site
Google Scholar
PubMed
Close
, and
Russell L. Elsberry Department of Meteorology, Naval Postgraduate School, Monterey, CA 93943

Search for other papers by Russell L. Elsberry in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A physical basis is provided for representing the large-scale operationally analyzed wind fields around western North Pacific tropical cyclones by use of empirical orthogonal functions (EOFs). The synoptic differences in the environmental flow are demonstrated for cyclones having different initial directions and translation speeds. It is also shown that the wind-based EOF coefficients may be used to differentiate between future tracks of cyclones that have a similar initial direction of motion. Thus, the small sets of EOF coefficients that are used in statistical regression techniques for track prediction by Peak et al. have physical meaning and are not statistical artifacts.

A new application of the EOF coefficients is to post-process the track predictions from the One-way influence Tropical Cyclone Model (OTCM), which is presently the best objective aid at the Joint Typhoon Warning Center (Guam). A 30% reduction in the forecast error at 72 h is achieved in the dependent sample. Thus, the synoptic influences represented by the EOF coefficients can differentiate situations in which the dynamical model (OTCM) is likely to provide poor guidance.

Abstract

A physical basis is provided for representing the large-scale operationally analyzed wind fields around western North Pacific tropical cyclones by use of empirical orthogonal functions (EOFs). The synoptic differences in the environmental flow are demonstrated for cyclones having different initial directions and translation speeds. It is also shown that the wind-based EOF coefficients may be used to differentiate between future tracks of cyclones that have a similar initial direction of motion. Thus, the small sets of EOF coefficients that are used in statistical regression techniques for track prediction by Peak et al. have physical meaning and are not statistical artifacts.

A new application of the EOF coefficients is to post-process the track predictions from the One-way influence Tropical Cyclone Model (OTCM), which is presently the best objective aid at the Joint Typhoon Warning Center (Guam). A 30% reduction in the forecast error at 72 h is achieved in the dependent sample. Thus, the synoptic influences represented by the EOF coefficients can differentiate situations in which the dynamical model (OTCM) is likely to provide poor guidance.

Save