Left Hand Side Wind Maxima in North Atlantic and Western Pacific Tropical Cyclones

Most parametric tropical cyclone risk models assume a circularly symmetric wind field that is modified by forward motion that yields a wind speed maximum on the right hand side of the system (northern hemisphere). Extratropical transitioning, interaction with adjacent weather systems, upper level shear, and other processes can result in different wind distributions for tropical cyclones that are important to account for when forecasting landfall wind speeds and when estimating property loss. Developing a parametric model that properly includes such asymmetries really requires several decades of detailed surface wind information over the basin for which the model is being developed. In this study the Climate Forecast System Reanalysis (CFSR) and Best Track Data Sets are used for the period 1979-2008 to determine the climatology of wind asymmetries associated with tropical cyclones in both the North Atlantic (NATL) and Northwestern Pacific (WPAC) basins. For each 6 hourly track point and for each cyclone in each basin, a scan is made of the CFSR surface wind speeds surrounding the center. If the maximum wind speed is found on the Left Hand Side the Track Point is tagged as a Left Hand Side Maximum (LHSM) point. A comparison of NATL and WPAC results shows significant differences in the relative frequencies of LHSM Track Points. Specifically, nearly 30% of the track points exhibit LHSMs in the WPAC while only about 20% of the track points in the NATL exhibit LHSMs. Additionally, the biggest differences between the WPAC and NATL LHSM distributions come from the lack of seasonality for such occurrences in the WPAC and the relatively high percentage of such occurrences south of where extratropical transitioning would be expected to play a role. This presentation will provide a more thorough comparison of the climatologies, more detailed explanations for the differences, and illustrations of how such information can be used in parametric risk models.