Performance of the Revised SFMR Algorithm: 2015 Season in Review

The previous operational version of the Stepped Frequency Microwave Radiometer (SFMR) algorithm performed well at hurricane strength but suffered from a significant high wind speed bias at tropical depression and storm strength wind speeds, especially within moderate-to-heavy precipitation. A revised algorithm was developed to reduce this bias by comparing the SFMR wind speeds to those of the GPS dropsondes, and a reduction of the high bias was produced. Additionally, a correction to the rain/wind separation within the microwave signal improved the range of the attainable rain rates of the SFMR. Therefore, the revised algorithm should provide more accurate wind speeds in all conditions and should provide a more realistic representation of the coincident rain rate. This revised algorithm became operational on all NOAA and Air Force Reserve hurricane hunter aircraft during the 2015 hurricane season, which flew into several tropical storms (Danny and Erika) as well as several intense hurricanes (Joaquin and Patricia).

In this work, an evaluation of the performance of the revised version is provided. Comparison of the weak wind speed regime in heavy rain is specifically analyzed to provide a quantitative estimate of the improvement. As in the algorithm development, these data are compared to GPS dropsonde surface-adjusted wind speeds. The prior version of the algorithm slightly underestimated very strong hurricane winds (> 65 m/s), and with the available data in several major hurricanes during 2015, evaluation of the upper end of the wind speed spectrum is quantifiable. Preliminary results reveal that the SFMR wind speeds compare well to the dropsonde wind speeds in all conditions and that a slight increase in the wind speeds at the upper end of the regime is apparent. Comparisons between the revised algorithm rain rate and estimates from the NOAA Tail Doppler radar (TDR) indicate that the revised version provides improved estimates of the rain intensity. The results from the 2015 hurricane season are promising in that improved wind speeds and rain rates were provided using the revised algorithm. Future work regarding the 2015 SFMR data include an examination of the extreme case of Patricia against other extreme cases (i.e. Megi, Felix, and Wilma) for the purpose of verifying the SFMR's capability to capture these extreme events with accuracy.