Abstract

The extratropical transition (ET) of tropical cyclones in the North Atlantic basin is explored using the framework of the Cyclone Phase Space (CPS). Analyses from the U.S. Navy Operational Global Atmospheric Prediction System (NOGAPS) and National Centers for Environmental Prediction Aviation Model (AVN) are used to diagnose CPS locations of 19 Atlantic tropical cyclones that underwent ET in the years 1998-2002. A nonhierarchical cluster analysis technique, the k-means method, is employed to determine physically meaningful clusters of cyclone structure in this CPS dataset. The cluster analysis successfully partitions observations into subsets representing distinct cyclone structures. It is also found that, while current operational cyclone classification schemes are successful at defining some storm structures, cluster analysis assists in defining a "gray area" of transitioning tropical cyclones, an area where operational schemes overlap. A mean path through the CPS of tropical cyclones undergoing ET is proposed. This CPS path is explored using a synoptic life cycle developed from NOGAPS-derived clusters. The resulting climatology highlights distinct stages of ET, beginning with a tropical cyclone isolated from midlatitude influences to its north and ending with an extratropical cyclone, deeply embedded in the midlatitude westerlies. Similar cluster solutions, derived using data from both the NOGAPS and AVN models, suggest that the clusters found in this analysis are not model specific and indicate natural stages of the ET process. However, there are differences between the model CPS solutions, suggesting the use of an observational approach to validating cyclone phase and thus to defining ET.

Original languageEnglish (US)
Pages (from-to)2916-2937
Number of pages22
JournalMonthly Weather Review
Volume132
Issue number12
DOIs
StatePublished - Dec 1 2004

All Science Journal Classification (ASJC) codes

  • Atmospheric Science

Fingerprint Dive into the research topics of 'Characterization of extratropical transition using cluster analysis'. Together they form a unique fingerprint.

  • Cite this