The Dvorak tropical cyclone intensity estimation technique using visible satellite data was introduced in 1975, with an additional technique using infrared satellite data introduced in 1984. The techniques have performed well throughout ocean basins across the world for nearly two decades of operational use. There are, however, cases where the intensity of a tropical cyclone using the Dvorak rules has been inaccurate, particularly with the infrared techniques. This often occurs when the exact location and/or structure of the low level circulation center is not known. Instances include the central dense overcast feature, the central cold cover, or the embedded center pattern (infrared imagery only). The Special Sensor Microwave/ Imager (SSM/I) has the advantage of sensing though thick cloud cover which can show the structure of a tropical cyclone that would otherwise be unseen in infrared and visible satellite imagery. SSM/I imagery can indicate the presence and location of low level circulation centers, banding features, and detect whether or not an eye is developing or already exists. These particular structural features of the tropical cyclone are not always discernible in conventional satellite imagery. The purpose of this study is not to replace Dvorak’s techniques, but to develop similar rules for estimating the intensity of tropical cyclones using SSM/I imagery. Further, techniques are developed to incorporate the SSM/I imagery into Dvorak’s conventional techniques in order to improve the intensity estimates.
Using time-matched visible, infrared and SSM/I imagery in cases where the intensity estimate given by the Dvorak infrared or visible satellite rules is clear cut, features in the SSM/I imagery are found that are indicative of tropical cyclone intensity. The techniques derived from the matched imagery are then applied to several tropical cyclones and are evaluated against the best-track data to determine their utility. Examples are shown where there exists discrepancies between the best-track intensity (as derived from real time conventional Dvorak analysis as well as synoptic data) and the intensity as determined from the SSM/I imagery. An attempt is made to resolve these discrepancies.