Estimating maximum surface winds from hurricane reconnaissance measurements

Scientists have long studied the best means to estimate hurricane surface winds from measurements collected far above (at typical 10,000 ft altitudes). In 2003 results from GPS dropsondes suggested that a simple 90% adjustment could do the job. Now results from a new instrument with superior sampling abilities (the Stepped-Frequency Microwave Radiometer or SFMR),indicate that the 90% rule only works on the most intense storms and is biased high on the others. Powell, Uhlhorn, and Kepert compared radial profiles of surface winds measured by the SFMR to radial profiles of flight-level winds to determine the slant ratio of the maximum surface wind speed to the maximum flight-level wind speed. The SFMR's advantage over the sonde is relatively fine radial resolution compared to only one or two spot surface sonde measurements that might be available along the same flight leg. Since the eyewall radius of maximum wind tilts outward with height, applying a slant ratio provides a more representative estimate of the maximum surface wind than ratios based on the top and bottom of GPS sonde profiles. Larger slant surface wind factors are found in small storms with large values of inertial stability and small values of relative angular momentum at the flight-level radius of maximum wind, which is consistent with Kepert's boundary layer theory. The GPS sonde-based “90% rule” surface winds assessed using this new data set have a high (~ 5 m/s) bias and substantially larger RMS errors than the new technique. A new regression model for the slant surface wind factor based upon SFMR data is presented, and used to make retrospective estimates of maximum surface wind speeds for significant Atlantic basin storms, including Hurricanes Allen (1980), Gilbert (1988), Hugo (1989), Andrew (1992), and Mitch (1998). The new slant surface wind factor model has been implemented in the AOML Hurricane Wind Analysis System (H*Wind) and is applicable for “best track” reanalysis of historical storms and for training satellite pattern recognition intensity estimate techniques during time periods with concurrent satellite and flight-level observations, to assess changes in global tropical cyclone activity.