7C.2 Update on SATellite-based CONsensus (SATCON) Approach to TC Intensity Estimation

Tuesday, 17 April 2012: 1:45 PM
Champions AB (Sawgrass Marriott)
Derrick Herndon, CIMSS, Madison, WI; and C. S. Velden and J. D. Hawkins
Manuscript (174.9 kB)

The SATellite CONsensus (SATCON) algorithm developed at CIMSS objectively combines Tropical Cyclone (TC) intensity estimates analyzed from satellite infrared and microwave-based methods to produce a weighted consensus estimate which is more skillful than the individual members. SATCON provides the TC forecaster with the ability to quickly reconcile differences in objective intensity methods, and as a comparative guidance tool for evaluating subjective Dvorak estimates. Real-time SATCON estimates have been provided to NHC, BOM and JTWC along with other TC forecast agencies as part of a demonstration phase since the 2008 TC season.

Current members of SATCON include the CIMSS ADT along with the CIMSS and CIRA AMSU algorithms. Each member of SATCON is weighted based on that member's performance in a given situation (eye size, intensity, infrared scene type, etc.). Separate weights are used for minimum sea level pressure (MSLP) and maximum sustained wind estimates (MSW). The SATCON algorithm logic and weighting structures were developed using coincident reconnaissance aircraft data from storms in the Atlantic, Eastern Pacific and Western Pacific. To date, we have accrued 562 cases when all three SATCON members were available and matched up with recon data. This dataset is divided into a dependent development sample, and an independent validation sample using a “leave every other case out” approach.

In addition to updating the consensus weight structure as new cases become available, a number of recent changes have been made to further improve the MSW estimates. These changes include using the TC eye size objectively measured by either infrared from geostationary satellites or 85-91 Ghz microwave frequencies from polar orbiter satellites to adjust the MSW in cases when the eye size deviates from climatology. A fourth pseudo-member is added to the MSW estimates which uses the statistically superior SATCON MSLP estimate along with TC size, latitude, storm translation speed and eyewall intensity/robustness to produce an estimate of the MSW. This pressure-wind derived estimate is then averaged with the weighted SATCON MSW estimate to produce a final MSW estimate.

Additional members to SATCON are currently being evaluated for inclusion in the method. An SSMIS sounder technique developed at CIMSS using logic similar to the CIMSS AMSU intensity method shows skill comparable to the CIMSS AMSU and ADT methods and thus should add skill to the consensus (see poster session for further information on this method). Another objective passive microwave-based method developed at NRL which identifies important features in the 85-91 Ghz frequencies from SSMI and SSMIS also shows promise as a member of SATCON.

These upgrades, as well as the resultant performance statistics, will be presented. Further information on SATCON can be found at:


- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner