Geostationary Satellite-based Estimation Method for Typhoon Radius of Maximum Wind

Estimating the typhoon sizes, including the radius of maximum wind (RMW) and the wind radii, is a challenging aspect of typhoon monitoring and forecasting. Currently, methods for estimating typhoon sizes include ground-based pressure and wind measurements, airborne Stepped Frequency Microwave Radiometer (SFMR) measurements, satellite-based infrared (IR) and microwave instrument retrievals, etc. Retrievals of sea surface winds based on the microwave scatterometers/radiometers suffer from coarse spatial resolution (40-50 km) and susceptibility to heavy rainfall; meanwhile, IR brightness temperatures lack a direct physical correlation with sea surface winds at the pixel level.

Due to the lack of regular aircraft reconnaissance, the determination of typhoon sizes in the western North Pacific relies solely on IR and microwave retrievals. Our assessment based on synthetic aperture radar (SAR) wind products indicates that the Joint Typhoon Warning Center's (JTWC) best track dataset has a better estimation of typhoon intensities than inner sizes, with an uncertainty within 15% for maximum sustained wind (Vmax) estimation, but as high as 30-60% for RMW estimation, which is above the global average of 25-40%.

This study establishes an RMW estimation algorithm for eyed typhoons based on geostationary satellite IR observations, with key steps including: (a) determining the typhoon center; (b) distinguishing clear-eye cases from unclear-eye cases; (c) estimating the eyewall radius (Reye) separately for clear- and unclear-eye cases; (d) estimating the RMW.

A TC-red-green-blue (TC-RGB) composite was designed by using satellite multichannel observations (reflectance, brightness temperature, and brightness temperature differences), which can effectively differentiate convective clouds, cirrus clouds, and low clouds, proving very effective in identifying exposed low-level circulation centers. Combining the TC-RGB composite image with 10-min atmospheric motion vectors products, a precise typhoon center location can be obtained through interactive human-computer methods. Using SAR positioning results as a reference, the root mean square difference (RMSD) for eyed typhoon positioning results was calculated, with the JTWC dataset direct interpolation result being ~8.6 km, and the interactive method being ~6.6 km, reducing by over 20%.

Subsequently, an objective method to differentiate between clear- and unclear-eye typhoons was established, along with an IR-based method for measuring Reye. For clear-eye typhoons, the calculated Reye has a correlation coefficient as high as 0.89 with the SAR observed RMW (RMW_SAR); for unclear-eye typhoons, the correlation coefficient between Reye and RMW_SAR also reaches 0.82.

Ultimately, an RMW regression equation for eyewall typhoons was established based on Reye and RMW_SAR, with a mean absolute error (MAE) of 5.46 km and a root mean square error (RMSE) of 7.35 km, nearly 40% less than the JTWC dataset.