High Quality Wind Retrievals for Hurricanes Using the SeaWinds Scatterometer W. Linwood Jones and Ian Adams Central Florida Remote Sensing Lab Univ. of Central FL Presentation Outline Hurricane Wind Vector Retrieval Issues CFRSL Hurricane Retrieval Algorithm QRad Rain Rate Algorithm Isabel and Fabian Results
Conclusion Issues with L2B Wind Vector Retrieval for Extreme Wind Events Wind retrievals not tuned for high wind speeds Geophysical Model Function developed for wind speeds < 20 m/s
CFRSL Approach Tropical Cyclone GMF tuned for retrieving high wind speeds Blended UMASS aircraft scat measurements and NSCAT GMF
Issues with L2B Wind Vector Retrieval for Extreme Wind Events Egg spatial resolution @ 25 km sampling is too coarse CFRSL Approach Uses
gridded slice sigma-o @ 12.5 Km sampling Latitude 25 km 50 km
25 km resolution using eggs Longitude Range gated slices from egg 30 km 25 km 6 km
12.5 km Issues with L2B Wind Vector Retrieval for Extreme Wind Events Rain has negative effects on wind measurements Rain attenuation lowers retrieved wind speeds Volume scattering increases retrieved wind
speeds and alters directional signatures Rain splash increases retrieved wind speeds and changes anisotropic ocean surface features MUDDH performance in hurricanes is poor Issues with L2B Wind Vector Retrieval for Extreme Wind Events Rain has negative effects on wind measurements
CFRSL Approach Simultaneous rain measurement using QRad Rain attenuation and volume backscatter correction provided Rain splash effect included in GMF Rain Quality Flags Provided
CFRSL Hurricane Retrieval Algorithm Slice sigma-0s for improved spatial resolution - 12.5 km Sigma-0 minimum eye location Wind speed retrieved from individual sigma-0 (assumed spiral wind directions) Rain inferred using QRad rain rate algorithm Sigma-0 corrections from atmospheric transmissivity and rain volume backscatter Flag pixels that correspond to high rain attenuation
QRad Algorithm Block Diagram QScat L1B sig-0 L2A Tbs Bin Slice Sig-0 Locate Center Atmos. Atten
and Backscatter Correction Retrieve Hurricane Wind Speed Wind Retrieval Hurricane Wind Speed QRad
Rain Rate Spiral Wind Direction Based upon Average of 4-flavor Sigma-0s 12.5 km sampling
Latitude Hurricane Eye Location Longitude QRad/SRad Rain Algorithm Look at noise channel of scatterometer Subtract
echo channel Calibrate radiometric temperature vs. TMI Train brightness temperature/rain rate relationship via TMI Coincident with scatterometer measurement QRad Integ. Rain Rate Product Improved CFRSL rain retrieval algorithm provides earth-gridded 50 km data
product Instantaneous km*mm/hr Resampled integrated rain rates > 2.4 to 12.5 km wind vector grid Latitude
QRad TRMM 2A12 Instantaneous Rain Rate Longitude Land QRad TMI
QRad TRMM 3B42RT HQ Instantaneous Collocated Low Rain event TMI QRad QRad TRMM 3B42RT HQ Instantaneous Collocated Moderate Rain event TMI
QRad QRad TRMM 3B42RT HQ Instantaneous Collocated High Rain Event TMI QRad CFRSL Integ. Rain Rate, km*mm/hr QRad (CFRSL) & TMI Integrated Rain
Rates for Simultaneous Events TMI Integ. Rain Rate, km*mm/hr AMSR vs. SRad Path Attenuation, Horizontal Polarization AMSR Attenuation SRad Attenuation 102
2003 Hurricane Results QuikSCAT Rev. 22055, Sept 13, 2003 22:04 Z (m/sec) HRD Wind Field, Time Interpolated CFRSL Wind Retrieval (mm/hr)
SSMI Rain Rate QRad Rain Rate SeaWinds Rev. 03927, Sept 15, 2003 15:40 Z (m/sec) HRD Wind Field, Time Interpolated CFRSL Wind Retrieval
(mm/hr) SSMI Rain Rate SRad Rain Rate QuikSCAT Rev. 22084, Sept 15, 2003 22:52 Z (m/sec) HRD Wind Field, Time Interpolated
CFRSL Wind Retrieval (mm/hr) SSMI Rain Rate SRad Rain Rate SeaWinds Rev. 03934, Sept 16, 2003 02:52 Z (m/sec)
HRD Wind Field 01:30 Z CFRSL Wind Retrieval (mm/hr) SSMI Rain Rate SRad Rain Rate SeaWinds Rev. 03934, Sept 16, 2003 02:52 Z
CFRSL Wind Retrieval JPL Wind Speed (m/s) CFRSL Wind Speed (m/s) JPL Wind Field HRD Modeled Wind Speed (m/s) HRD Modeled Wind Speed (m/s)
QuikSCAT vs. H* Wind Analysis SeaWinds vs. HRD Model CFRSL Algorithm Composite Wind Speeds Conclusion Ready to test current algorithm in operational environment
Acceptable estimates of high rain rates inside storm conditions Reasonable wind speed retrievals up to 40 - 45 m/s Work must be done to improve model function and rain backscatter and attenuation estimates Will
utilize UMass IWRAP and SFMR data set Have simultaneous AMSR and SRad for the 2003 hurricane season
