1. MODIS Atmospheric Profiles Suzanne Wetzel Seemann, CIMSS MOD07 Developer Retrievals are performed in 5x5 FOV (approximately 5km resolution) clear-sky radiances over land and ocean for both day and night. Algorithm is a statistical regression and has the option for a subsequent nonlinear physical retrieval. Regression predictors include MODIS infrared radiances from bands 25, 27-36 (4.4 - 14.2mm). Clear sky determined by MODIS cloud mask (MOD35).

2. MODIS IR Bands Spectral Position MODIS

3. MODIS IR Bands Profile Sensitivity - Temperature

4. MODIS IR Bands Profile Sensitivity – Water Vapor

5. Atmospheric Profile Output Atmospheric precipitable water vapor (total, high - 250 hPa to 700 hPa, and low- 920 hPa to the surface ) Profiles of temperature and moisture (20 levels) Total column ozone Stability indices (lifted index, total totals) Surface Skin Temperature

6. Algorithm Discussion R  =   s B  s (T s )  s (p s ) -  0 ps B  (T(p)) d   (p) + r  s   (p s )  0 ps B  (T(p)) d  *  (p) + R  sun   1+sec  (p s ) r  s sun RTE (no scattering) in LTE R…radiance,  …wavenumber, s…surface, p…pressure, sun…solar, T…temperature, B…Planck function,  …emissivity,r…reflectivity,  …level to space transmittance, ...local solar zenith angle  *…level to surface transmittance [  *=   (p s )/   (p)]

7. Algorithm Discussion - continue R is measured by MODIS for = 4.4 - 14.2  m (R 25, R 27, … R 36 ) R can be considered a nonlinear function of the atmospheric properties including T, q, ozone, surface pressure, skin temperature, and emissivity. We can infer a statistical regression relationship using calculated radiances from a global set of radiosonde profiles and surface data. Relationship is inverted to retrieve atmospheric properties from observed MODIS radiances.

8. Global radiosondes: data set drawn from NOAA-88, TIGR-3, ozonesondes, ECMWF analyses, desert radiosondes containing 15000+ global radiosonde profiles of temperature, moisture, and ozone used for training data set. RT model: Radiance calculations for each training profile are made using a 101 pressure layer transmittance model. MODIS instrument noise is added to calculated spectral band radiances. Radiosonde temperature-moisture-ozone profile / calculated MODIS radiance pairs are used to create the statistical regression relationship. Bias corrections are applied to the observed MODIS radiances to account for forward model error, spectral response uncertainty, and calibration error. Algorithm Discussion - continue

9. New BT zones: Land Zone 1: < 272, 1978 profiles (< 275) Zone 2: 272-287, 2538 profiles (269-290) Zone 3: 287-296, 2807 profiles (284- 299) Zone 4: 296-350, 2226 profiles (293-353) Ocean Zone 1: < 283.5, 2214 profiles (< 286.5) Zone 2: 283.5-293, 2900 profiles (280.5-296) Zone 3: 293-350, 2437 profiles (290-353) MODIS Land – Sea Classified Retrievals OLD BT 11  m ZONES Zone 1: < 245 K Zone 2: 245-269 K Zone 3: 269-285 K Zone 4: 285-294 K Zone 5:294-300 K Zone 6: 300-310 K Zone 7: > 310 K

10. 10 AIRS Clear-Sky Regression Retrieval Single FOV Eigenvector Regression Retrieval of T, q, T s, TPW, O 3, and  s under clear conditions Preparation of representative trainingsets Forward Model Calculations using SARTA Application of BT/scanang-classification scheme Retrieval Validation/Comparison: ECMWF analysis, global RAOBs, MODIS and GOES Retrievals, L2 Standard Product X = C Y T C = X Y (Y T Y) -1 Regression Model Least squares regression solution  TIGR3 & Noaa88 & ECMWF & special desert and polar cases  Ecosystem assigned to each point to get realistic surface pressure, surface skin temperature and surface emissivity. 295  BT290  BT 6 285  BT  295280  BT  300 5 275  BT  285270  BT  290 4 265  BT  275260  BT  280 3 255  BT  265250  BT  270 2 BT  255BT  260 1 BT@1000 cm -1 observations BT@1000 cm -1 training Class X…Atmospheric State, C…Coefficients, Y…Measurements 11 scanning angles  with sec(  )=1 +  *i i=0,1,2,…10 and  =0.0524

11. Global gridded emissivity (0.05 degree resolution) at each of the 8 inflection points, derived from the laboratory “baseline” emissivity and measured Aqua MOD11 emissivity for August 2003. Improved physical based emissivity spectra is assigned to each training profile for better characterization of the surface

12. Solar Zenith >= 62 and < 95 Solar Zenith >= 95 Solar Zenith < 62 Sonde Surface Air T IRT Skin T Land Skin T relationship derived from SGP-CART site, as a function of solar zenith (3 categories) & azimuth (8 categories) Improved Surface Skin Temperature Assignment

13. Histogram of BT11 from 4 MODIS desert granules in June 1. Separated training into seven BT regression “zones” to limit the retrievals to training data with physical relevance to the observed conditions. There was little training data in the warmest “zones” for 11  m BT > 300 o K: 2. Added 900 new radiosondes to the original NOAA-88 training data set. Radiosondes were from the north African desert from all months in 2001. Recent improvements to the algorithm : Histogram of BT11 from NOAA-88 training data set

14. PW High 700-300 hPa PW Low 920 hPa - sfc PW (mm): 0 3 6 9 12 15 18

17. Surface Skin Temperature

18. Dobson Total Column Ozone, 2 October 2002 Terra MODIS direct broadcast

19. All cases: bias = 2.65 mm, rms = 4.3 mm, n = 314 Emis = 0.98, PFAAST MODIS, GOES, and radiosonde TPW (mm) Microwave radiometer TPW (mm) All cases: bias = 0.33mm, rms = 2.58mm, n = 313 Dry cases: bias = -0.6mm, rms = 1.88, n = 201 Wet cases: bias = 2.02mm, rms = 3.49mm, n = 112 Gridded emis, PFAAST All cases: bias = -0.04mm, rms = 2.49mm, n = 313 Dry cases: bias = -0.68, rms = 1.97, n = 201 Wet cases: bias = 1.12mm, rms = 3.22mm, n = 112 Gridded emis, pCRTM Comparison of TPW from Terra MODIS (red dots), GOES-8 (blue diamonds), and radiosonde (black crosses) with the SGP ARM- CART microwave water radiometer 313 clear sky cases from April 2001 to August 2005

20. Global Total Ozone (Dobson) for December 1, 2004 MODIS MOD07 TOMS

21. Global Total Precipitable Water Comparison 22 May 2002 MODIS TPW SSM/I f-14 TPW Ascending and descending passes were averaged 0 6 12 18 24 30 36 42 48 54 60 66 72 TPW (mm):

22. TPW (mm) for 2 June 2001 over North America MODIS Statistical RetrievalMODIS Physical RetrievalGOES-8 and GOES-10 Day Night

23. Mean difference MODIS - TOMS = 4.4 dob RMS = 27.4 dob mean abs % error: abs(M-T)/T = 5.9% N = 10,614 Total Ozone from MODIS (top) and TOMS (bottom) May 22, 2002 MODIS (dob) TOMS (dob) 240 260 280 300 320 340 360 380 400 420 440 Latitude ( o ) - 60 - 40 - 20 0 20 40 60 80 - 40 - 20 0 20 40 60 80 % Error (M-T)/T

24. MODIS profiles agree well with radiosondes and NCEP-GDAS when the atmospheric temperature and moisture is fairly smooth and monotonic: But not so well with smaller-scale features, such as isolated dry or moist layers: MODIS GDAS Sonde Temperature Pressure (hPa) Mixing RatioTemperature Pressure (hPa) Mixing Ratio Temperature Pressure (hPa) Mixing Ratio Temperature Pressure (hPa) Mixing Ratio

25. Isobaric Surfaces/Profiles of Temperature 13 October 2002 Terra MODIS direct broadcast oKoK Temperature Pressure (hPa) 300 hPa 500 hPa 850 hPa

26. Mixing Ratio (g/kg) Pressure (hPa) 300 hPa 500 hPa 850 hPa Isobaric Surfaces/Profiles of Moisture 13 October 2002 Terra MODIS direct broadcast

27. Access to EOS Data SPoRT (GHCC) receives MODIS data from Univ. of Wisconsin DB system Aqua and Terra MODIS, testing AIRS near real time, full resolution to 250m MCIDAS ADDE server – subset by region and channel – reduces 700mb down to 10-20mb FOG PRODUCT Near real-time products EOS atmospheric science team from UW GHCC generated products from real time data stream (composites, LST/SST, TPW) SST from USF Earth Science Enterprise National Aeronautics and Space Administration

28. EOS Products AIRS Profile – Valparaiso, Fl – 15 Oct 2003 AIRS 1845 UTC VPS 1800 UTC good portrayal of dry layer on top of low level moisture AIRS assesses changes in stability AIRS profiles will map temperature and moisture gradients and help diagnose asynoptic changes in atmospheric stability. MODIS land surface temperatures are used in forecasting morning low temperatures and in IFPS validation. Earth Science Enterprise National Aeronautics and Space Administration

29. Earth Science Enterprise National Aeronautics and Space Administration EOS Products MODIS 250m visible and color composite imagery can detect tornado damage tracks and help in storm intensity assessment. Color composite imagery and aerosol optical depth derived from MODIS can identify regions of restricted visibility with significant impact on aviation. Smoke from Montana forest fires – summer 2003 High concentrations of aerosols MODIS Aerosol Optical Depth Van Buren F4 - 300yds 21 mi 3:13-3:45p Fredericktown F2 - 800yds 9 mi 3:45-4:00p Hendrickson F4 - 650yds 21 mi 3:45-4:18p

30. May-July 2002 trends inferred from daily MODIS TPW * Continuous pulsing motion of moisture is evident * Global circulations are obvious esp around subtropical highs (e.g. clockwise around Bermuda high in Jun, counter clockwise around southern Pacific high in Dec) * Indian monsoon evident Jun-Jul-Aug * Gulf of Mexico moisture moving into central US appears May - Jun * Indonesian region has year round high moisture (often global max) * TPW follows the Sun – latitudinal moisture bands connecting continents drift N & S with seasons

31. Future Work Global radiance bias adjustment improvements Combined retrievals from Aqua MODIS and AIRS radiances Applications to climate and weather studies

32. Some aspects of AIRS Sounding Retrieval and their impact on IMAPP Products Dr Pradeep Kumar Thapliyal, ASD/MOG/RESIPA Space Applications Centre (ISRO), INDIA ABSTRACT As a part of International MODIS/AIRS Processing Package (IMAPP) an algorithm has been developed at Cooperative Institute for Meteorological Satellite Studies (CIMSS) to retrieve atmospheric and surface parameters from AIRS-L1B radiance measurements. In this presentation some aspects of the AIRS sounding retrieval, based on principal component regression (PCR), will be discussed. Presentation will mainly focus on retrieval sensitivity to infrared (IR) spectral surface emissivity, training data classification (global versus regional), sunglint/solar-reflection effect, etc. Some interesting features in AIRS observed radiance spectra that might help in detecting boundary-layer temperature inversion, will also be presented.

34. Global Vs. Regional IMAPP Profile Performance