Satellite Retrieval of Atmospheric Water Budget Over Gulf of Mexico-Caribbean Sea Basin

1. Satellite Retrieval ofAtmospheric Water Budget OverGulf of Mexico-Caribbean Sea Basin Pablo Santos 1 & Eric A. Smith 2 1 National Weather Service, Miami, FL [Florida State Univ. Dept. of Meteorology Graduate Student Program] 2 NASA/Goddard Space Flight Center, Greenbelt, MD Presented at 81st AMS Annual Meeting Conference on Global Change & Climate Variations [January 14-18, 2001; Albuquerque, NM]

2. interest in understanding all terms of atmospheric branch of hydrological cycle within enclosed ocean basin interest in quantifying water budget components from satellite measurements only interest in extending past research which has been restricted largely to requiring balance between divergence of WV transport & E - P develop remote sensing-based & portable retrieval methodology, using measurements from SSM/I & GOES satellites for calculating atmospheric water budget, verifying divergence residue with radiosonde-based estimates quantify mean & variance properties of Gulf of Mexico-Caribbean Sea Basin atmospheric water budget in time-space understand drawbacks within convectively active region under assumption that local rate of change of PW and cloud LWP are negligible in context of regionally & seasonally based atmospheric water balance Motivation and Objectives

3. Study Area

4. Data • Research utilizes data from three sources: • DMSP (SSM/I) • used to retrieve precipitable water (W), cloud liquid water path (Wc), precipitation (P), surface air temperature (Ta), surface specific humidity (qa), & surface wind speed (U) • GOES (Imager) • used to retrieve sea surface temperature (SST) & surface air temperatue (Ta) -- plus W, Wc, P, & qa with adjustments by SSM/I • Upper Air T-q-V profiles

5. Methodology • development of algorithms for water budget variables • calculation & validation of water budget • analysis of water budget • atmospheric water budget equation for vertically integrated total water in atmosphere where: W = precipitable water; Wc = cloud liquid water path; Q = vertically integrated water vapor transport; E = evaporation; & P = precipitation:

6. Design of Retrieval System Evaporation Clayson & Curry (1996) Four Inputs Precipitable Water Cloud Liquid Water Path Precipitation SSM/I Greenwald et al. (1995), Lojou (1994), Petty (1994), Wentz (1995) GOES Split Window Technique Chesters (1980), Crosson et al. (1993) Combined SSM/I-GOES GOES W estimate adjusted by SSM/I SSM/I Alishouse (1990), Greenwald et al. (1995), Lojou (1994), Weng and Grody (1994) Combined SSM/I-GOES ∑ GOES detects cloud area ∑ GOES LWP estimate equals closest SSM/I retrieved mean cloud LWP in time ∑ data from two SSM/Is are used to increase data sampling SSM/I ∑ NESDIS SSM/I alg (Ferraro et al. , 1998) cross-validated against 4 other SSM/I algs ∑ SSM/I retrievals within expected uncertainty of ground truth (Smith et al. , 1998) Sea Surface Temp (GOES) Legeckis (1997) Surface Air Temperature (SSM/I-GOES) Clayson and Curry (1996) SSM/I - cld classification GOES - cld top temp & SST Ta func of cld type & SST Rain Screening Ferraro et al. (1998) Surface Specific Humidity calculated from GOES as function of SST & adjusted by SSM/I qa via Schlüssel (1996) Combined SSM/I-GOES ∑ GOES retrievals based on probability matching (PM) between SSM/I retrieved rainrates & GOES EBBTs ∑ PM technique based on work of Turk et al. (2000) & Grose et al. (2000) Cloud detection based on NESDIS & NAVY operational screening techniques Legeckis (1997) & May (1998) Surface Wind Speed (SSM/I) Bates (1991), Clayson & Curry (1996), Schlüssel (1996) GOES E estimate uses closest in time SSM/I mean U retrieval from two SSM/Is

7. Water Budget Calculation Once precipitable water (W), cloud liquid water path (Wc), precipitation (P), & evaporation (E) are retrieved at each grid point, divergence of vertically integrated water vapor transport (—∑Q) for time t at grid coordinate (i,j) is calculated using centered difference form of water balance equation, as follows:

8. Validation of Water Budget • Divergence of water vapor transport obtained as residue in water balance equation. • Residue term compared to direct measure of divergence obtained from ring of operational radiosonde stations surrounding basins in study area. • Direct measure obtained from line-integral technique (application of Greens’ theorem) in conjunction with sounding ring:

9. Analysis • Algorithm sensitivity analyses and multiple-algorithm intercomparisons are being used to define retrieval system uncertainties. • Budget calculations are being performed on 2 x 4 km grid mesh with hourly mean composites generated on 0.25 x 0.25 degree grid mesh for 6 one-month periods (Oct 1997; Jan, Apr, Jul, Oct 1998; Jan 1999). • Analyses are aimed at quantifying regional water budget variability of Gulf of Mexico-Caribbean Sea Basin to enable pursuit of key scientific objectives. • Final conclusions are to be based on analyses of filtered daily time series of sub-regional averages & gridded monthly/diurnal-composites with respect to individual terms of water budget equation.

10. Algorithm Intercomparison Results - Jan 98

11. Precipitable Water (W)intercomparison results(Jan/Jul 98)SSM/I-GOES W algorithm(Santos Adj-GOES)used for final budget calculations

12. Cloud Liquid Water Path (Wc)intercomparison results(Jan 98)Alishouse SSM/I algorithm used as control for GOES-8 retrievalsSSM/I-GOES Wc algorithm used for final budget calculations

13. Precipitation (P) intercomparison results(Jan 98)NESDIS SSM/I algorithm used as control for developing probability matching lookup tables between SSM/I rainfall & GOES-8 EBBTs (these tables used for GOES-8 rain retrievals)SSM/I-GOES P algorithm used for final budget calculations

14. Sea Surface Temperature (SST)intercomparison results(Jan/Jul 98) SST is input parameter for Clayson & Curry (1996) Evaporation algorithmLegeckis GOES algorithm used for final SST retrievals

15. Surface Air Temperature (Ta) intercomparison results(Jan/Jul 98)Ta is input parameter for Clayson & Curry (1996) Evaporation algorithmSSM/I-GOES algorithm(modified Clayson & Curry where GOES used in deriving cloud type) used for final Ta retrievals

16. Surface Specific Humidity (qa)intercomparison results(Jan/Jul 98)qa is input parameter for Clayson & Curry (1996) Evaporation algorithmSSM/I-GOES algorithm used for final qa retrievals

17. Surface Wind Speed (U)intercomparison results(Jan/Jul 98)U is input parameter for Clayson & Curry (1996) Evaporation algorithmSchlüssel SSM/I algorithm used for final U retrievals

18. Ocean surface latent heat flux from Clayson & Curry (1996) Evaporation algorithm(based on surface renewal theory)Comparisons to vertical q gradient over surface layer [qs(SST) - qa]& to surface wind speed (U)[Jan 98]

19. Ocean surface latent heat flux from Clayson & Curry (1996) Evaporation algorithm(based on surface renewal theory)Comparisons to vertical q gradient over surface layer [qs(SST) - qa]& to surface wind speed (U)[Jul 98]

20. Jan 98 Hourly Composites of Rate of Change of Precipitable Water (W)

27. Jan 98 Hourly Composites of Rate of Change of Cloud Liquid Water Path (Wc)

34. Jan 98 Hourly Composites of Input Parameters forClayson (1996)Evaporation Algorithm:Sea Surface Temperature (SST), Surface Air Temperature (Ta), Surface Specific Humidity (qa)