The 10-year data set includes all storms with intensities greater than or equal to 120 kts as denoted in the best track files created by the National Hurricane Center (Atlantic and East Pacific), the Central Pacific Hurricane Center (central pacific) and the Joint Typhoon Warning Center (for the rest of the TC basins). Storms that approached 120 kts, ~110 kt were also included. The resampling method of Poe (1990) was used for processing storm-centered SSM/I 85 GHz that enabled enhanced eyewall structure analysis.
The western Pacific basin clearly leads the way with the percentage (80%) of intense systems (> 120 kts) that contain concentric eyewalls. This is likely due in part to the fact these storms have more uninterrupted time prior to encountering unfavorable environmental conditions (shear, dry air intrusion, cold SSTs, etc). The Atlantic comes in second with 70%, but this number is deceiving since the number of intense storms is significantly smaller than the western Pacific and the 2005 season dramatically changed the 10-year statistics. The eastern Pacific and southern hemisphere trail with percent of concentric eyewalls at 50% and 40% respectively.
The data clearly outlines preferred eyewall evolutions: a) outer eyewall completely encircles the original inner eyewall which collapses as the outer eyewall becomes the primary eyewall, b) the eyewall cycle repeats multiple times while environmental conditions are favorable for development, c) eyewall cycles are interrupted and multiple eyewalls exist in tandem for days at a time, d) annual eyewall configuration evolves, e) outer eyewall is created at large radius and remains for an extended time while inner eye decays and fades away or barely exists. In addition, formation of secondary eyewalls typically occurs via asymmetric convection and the wrapping of one dominant rainband.
The passive microwave data set suffers from obvious limitations of temporal coverage from a polar orbiter sensor, but this is partially offset by combining a suite of near real-time operational and research microwave imagers and sounders (SSM/I, TMI, AMSR-E, WindSat, SSMIS, AMSU-B and MHS). Temporal sampling has improved during the later portions of this data set as more new sensors joined the constellation. Thus, it is possible that early time frames missed some storms with double eyewalls due to sampling issues alone and represent minimum values.