Since 1998, the Advanced Microwave Sounding Unit (AMSU) flown on board the NOAA KLM series of polar-orbiting satellites has been used to estimate Tropical Cyclone (TC) intensity. The AMSU-A (sounder) instrument provides thermal information in the upper portion of the troposphere using channels in the ~55 GHz band. Warm temperature anomalies in this portion of the troposphere associated with subsidence in TC cores are directly related to the TC minimum sea level pressure (MSLP). An algorithm developed by the UW-CIMSS TC team has been used in real-time since 2000 to estimate TC MSLP using a statistical regression of AMSU-derived brightness temperature anomalies.

A number of improvements have been incorporated into the original algorithm to account for instrument sub-sampling and these have led to improved performance. The AMSU instrument has a variable resolution dependant on scan geometry that ranges from 48 km at nadir to 100 km at the limb. As such, the instrument will sub-sample the “real” upper level warm anomaly associated with TCs when the diameter of the storm’s warm core, as constrained by the eyewall convection, is significantly smaller than the instrument field of view (FOV) resolution. To address this, a statistical correction for sub-sampled TC’s (small warm cores) was developed. Conversely, another correction has been added for storms with well-resolved (large) eyes. Other upgrades to the algorithm logic have resulted in reducing the estimation error. Subsequent to these improvements, a statistical analysis versus recon MSLP measurements in the Atlantic basin shows the algorithm performance is superior in skill to the Dvorak technique used at the various TC forecast centers. A discussion of the recent algorithm upgrades and performance metrics will be presented.