Presentation is loading. Please wait.

Presentation is loading. Please wait.

AIRS (Atmospheric Infrared Sounder) Regression Retrieval (Level 2)

Similar presentations


Presentation on theme: "AIRS (Atmospheric Infrared Sounder) Regression Retrieval (Level 2)"— Presentation transcript:

1 AIRS (Atmospheric Infrared Sounder) Regression Retrieval (Level 2)

2 Level 0 to Level 2 Level 0: raw data Level 1A: geolocated radiance in counts Level 1B: calibrated radiance in physical units Level 2: retrieved physical variables (temperature, humidity and ozone profiles, surface skin temperature, total precipitable water, total ozone content, cloud top height...)

3 Regression Model X = C Y T C = X Y (Y T Y) -1 1. Regression Model 2. Least squares regression solution Y…measurements [nprofs x nchannels] C...Regression coefficients [nlevels x nchannels] X... Atmospheric variables [nlevels x nprofs]

4 Regression Retrieval (1) C = X tr Y tr (Y tr T Y tr ) -1 X = C Y T 1. Calculate Regression Coefficients 2. Perform Retrieval (RTV) Y… Measurements [nprofs x nchannels] C... Regression coefficients [nlevels x nchannels] X... Atmospheric variables [nlevels x nprofs] Subscript tr refers to trainingset

5 Regression Retrieval (2) C =  X tr  Y tr (  Y tr T  Y tr ) -1 with  X tr = X tr – mean(X tr )  Y tr = Y tr – mean(Y tr )  X = C  Y T or X = mean(X tr ) + C  Y T with  X = X – mean(X tr )  Y = Y – mean(Y tr ) 1. Calculate Regression Coefficients 2. Perform Retrieval (RTV)

6 Principal Components (PC) Regression Retrieval 1. Calculate Regression Coefficients M = Cov(Y tr ) U = eig(M) A tr =  Y tr U C =  X tr A tr (A tr T A tr ) -1 X = mean(X tr ) + C A T with A=  Y U,  Y = Y – mean(Y tr ) 2. Perform Retrieval (RTV) M… covariance matrix [nchannels x nchannels] U... First few eigenvectors of M [nchannels x npc] A... Projection Coefficients (or amplitudes) [nsamples x npc]

7 The Trainingset X tr... Representative set of atmospheric variables including temperature, moisture, ozone, surface pressure, surface skin temperature and surface skin emissivities Y tr... Corresponding set of simulated radiances, calculated by a fast RT (radiative transfer) forward model

8 Radiance received by AIRS  Upwelling IR radiation from surface  Upwelling IR radiation from atm. layers  Reflected downwelling IR radiation  Reflected solar radiation R…radiance,  …wavenumber, s…surface, p…pressure, sun…solar, T…temperature, B…Planck function,  …emissivity,  …level to space transmittance, ...local solar zenith angle r…reflectivity, with r = (1-  )/ ,  *…level to surface (downwelling) transmittance [  *=   2 (p s )/   (p)]

9 Fast Radiative Transfer Forward Model Fast Model Regression : - Computation of line-by-line Transmittance  for FM training data set - Convolve with AIRS SRF (spectral response function) - Solve regression scheme  = AC for coefficients C using predictors A (predictors are functions of T, p, absorber amount, scanang …) Calculate transmittance  for any other profile Solve RTE to get radiance R 

10 IMAPP AIRS Regression Retrieval Results: Comparison with co-located radiosonde observations (RAOBs)

11 Co-located RAOB / AIRS single profile retrieval (RTV) Temperature (pixel 1259) RAOB Residual (Obs – Calc BT) Retrieval Residual (Obs – Calc BT ) residual = observed minus calculated spectrum RAOB – RTV Brightness Temperature (BT) residual

12 Co-located RAOB / AIRS single profile retrieval (RTV) RAOB Residual (Obs – Calc BT) Retrieval Residual (Obs – Calc BT ) Humidity (pixel 1516) Brightness Temperature (BT) residual RAOB – RTV residual = observed minus calculated spectrum

13 HumidityTemperature RMS of RAOB – RTV STDEV of RAOB – RTV Co-located RAOB / AIRS retrieval statistics (1899 profiles)

14 RMS of Residual (Obs – Calc BT) Stdev of Residual (Obs – Calc BT) Mean of Residual (Obs – Calc BT) Radiosonde ObservationsAIRS Retrieval

15 IMAPP AIRS Regression Retrieval Results: Global (240 granules) retrievals and retrievals over the CIMSS direct broadcast area

16 Globally Retrieved Temperature at 850 mbar With CloudmaskWithout Cloudmask Ascending Grans Descending Grans

17 Globally Retrieved Moisture at 850 mbar Ascending Grans Descending Grans With CloudmaskWithout Cloudmask

18 CIMSS Direct Broadcast area: AIRS measurements (10-23-2003) Brightness Temperature [K] at 1000 cm -1 (no cloudmask)

19 Temperature [K] at 700 mbar (no cloudmask) Humidity [g/kg] at 700 mbar (no cloudmask) CIMSS Direct Broadcast area: IMAPP AIRS Retrieval (10-23-2003)

20 Surface Skin Temperature [K] (no cloudmask) Total Precipitable Water [cm] (no cloudmask) CIMSS Direct Broadcast area: IMAPP AIRS Retrieval (10-23-2003)

21 Ozone [ppmv] at 9.5 mbar (no cloudmask) Total Ozone [Dobson Units] (no cloudmask) CIMSS Direct Broadcast area: IMAPP AIRS Retrieval (10-23-2003)

22 Surface Reflectivity at 2641 cm -1 (no cloudmask) Surface Emissivity at 908 cm -1 (no cloudmask) CIMSS Direct Broadcast area: IMAPP AIRS Retrieval (10-23-2003)

23 IMAPP AIRS Regression Retrieval Results: Comparison with ECMWF analysis fields

24 AIRS RTV vs. ECMWF Analysis: Temperature at 700 mbar (G192, 09-02-2003) With CloudmaskWithout Cloudmask AIRS RTV ECMWF ANL

25 AIRS RTV vs. ECMWF Analysis: Temperature at 500 mbar and at selected pixel AIRS RTV ECMWF ANL IMAPP RTV ECMWF – IMAPP RTV Pixel 28/49

26 AIRS RTV vs. ECMWF Analysis: Humidity at 750 mbar and at selected pixel ECMWF ANL IMAPP RTV ECMWF – IMAPP RTV Pixel 78/53 AIRS RTV ECMWF ANL

27 AIRS RTV vs. ECMWF Analysis: Humidity along scanline 65 (without cloudmask) Scanline 65 IMAPP AIRS RetrievalECMWF Analysis

28 AIRS RTV vs. ECMWF Analysis: Humidity along scanline 65 (with cloudmask) Scanline 65 IMAPP AIRS RetrievalECMWF Analysis

29 RMS and STDEV of ECMWF minus AIRS RTV (G192, 09-02-2003, 4758 clear pixels) HumidityTemperature Root Mean Square (RMS) Error Standard Deviation Surface Skin Temperature [K] Total Precipitable Water [cm]

30 AIRS RTV vs. ECMWF Analysis: Spatial mean of Brightness Temperature (BT) residual RMS of Residual (Obs – Calc BT) Stdev of Residual (Obs – Calc BT) Mean of Residual (Obs – Calc BT) IMAPP AIRS RetrievalECMWF Analysis

31 AIRS RTV vs. ECMWF Analysis: Spectral mean of Brightness Temperature (BT) residual AIRS RTV ECMWF Analysis Mean Long WaveMean Short Wave Mean Mid Wave

32 IMAPP AIRS Regression Retrieval Results: Comparison with ECMWF analysis fields and L2 operational product

33 AIRS RTV vs. ECMWF Analysis vs. Operational Product: Temperature at 500 mbar (G192, 04-04-2004) IMAPP AIRS RTV ECMWF ANL Operational AIRS RTV available on AMSU footprint (=3x3 AIRS FOVs) only missing areas  retrieval not successful or not validated yet

34 AIRS RTV vs. ECMWF Analysis vs. Operational Product: Temperature at selected pixel ECMWF ANL IMAPP RTV OP RTV ECMWF – IMAPP RTV ECMWF – OP RTV Pixel 22/6

35 AIRS RTV vs. ECMWF Analysis vs. Operational Product: Humidity at 600 mbar (G192, 04-04-2004) IMAPP AIRS RTV ECMWF ANL Operational AIRS RTV available on AMSU footprint (=3x3 AIRS FOVs) only missing areas  retrieval not successful or not validated yet

36 AIRS RTV vs. ECMWF Analysis vs. Operational Product: Humidity at selected pixel ECMWF ANL IMAPP RTV OP RTV ECMWF – IMAPP RTV ECMWF – OP RTV Pixel 14/15

37 IMAPP AIRS Regression Retrieval Results: Comparison with MODIS and GOES retrievals

38 MODIS RTV vs. AIRS RTV vs. GOES RTV: Temperature and Humidity at 620 mbar (G192, 09-02-2003) AIRS RTVGOES RTVMODIS RTV Temperature @620 Humidity @620

39 Accurate atmospheric sounding and cloud variability observations provide an understanding of atmospheric dynamic and cloud-radiation processes that need to be included in climate models. Furthermore they are critical for initializing storm scale models for the prediction of severe weather events. This requires an algorithm which fully utilizes the information about the atmospheric state inherent in single field-of-view hyperspectral radiances (e.g. AIRS, IASI, CrIS) is fast to be useful for real-time applications provides T, H 2 O, O 3 profiles, as well as surface temperature, surface emissivity, cloud top pressure, cloud optical thickness, CO 2 concentration simultaneously under all weather conditions at single- FOV resolution accounts for the non-linearity between radiances and cloud parameters and atmospheric moisture accounts for the annual change of CO 2 Improved Profile and Cloud Top Height Retrieval by Using Dual Regression Elisabeth Weisz & others

40 Dual Regression Retrieval Algorithm – Part 1 Clear trained EOF regression results CLEARCLOUDY EOF regression Clear radiance calculations SARTA v1.7 (UMBC) Regression RTV T, H 2 O, O 3, T skin, Emissivity, CO 2 at single FOV BT, Scanning angle and CO 2 Classification Cloudy radiance calculations Fast RT cloud model (Wei, Yang) Scanning angle, Cloud height and CO 2 Classification T, H 2 O, O 3, T skin, CO 2, CTOP, COT at single FOV Additional Predictors (spres, solzen) Global training data set (SeeBor V5) 15704 profiles Global cloudy training data set 19948 profiles Clear Regression CoefficientsCloudy Regression Coefficients Cloud trained EOF regression results X…atmospheric state, C… regression coefficients, A…compressed radiances (projection coefficients)

41 The Training Sets Clear TrainingSet (SeeBorV5.0 clear-sky global training database) 15704 globally distributed profiles of T, H 2 O, O 3 at 101 levels taken from NOAA-88, ECMWF and TIGR-3 data sets, including radiosondes and ozonesondes. Quality and saturation checks have been applied. T skin has been assigned using a relationship between T skin and T airs derived for the SGP ARM site. Surface Emissivity was assigned using a global land surface emissivity data base and the standard sea water emissivity spectrum. A few eigenvector coefficients are used in the regression, which are reconstructed to full spectrum after the retrieval. CO 2 concentration is assigned as a random normally distributed value ranging from 360 and 400 ppmv Cloudy TrainingSet 19948 globally distributed profiles, with 7969 from the clear set. CTP pressure was determined from the relative humidity structure of the profile Randomly distributed COTH ranging from 0.01 and 10 Randomly distributed CPS ranging from 5 to 35 for water droplets and 10 to 50 for ice crystals T skin has been assigned from T air by adding a random number Surface Emissivities as in the clear set CO 2 concentration as in the clear set

42 The Classifications Class T b @920 cm -1 T b @920 cm -1 (training) (observations) 1 T b  260 K T b  255 K 2 250  T b  270 K 255  T b  265 K 3 260  T b  280 K 265  T b  275 K 4 270  T b  290 K 275  T b  285 K 5 280  T b  300 K 285  T b  295 K 6 290  T b 295  T b 1.Scanning angle classification: 11 sets ranging from 0° to 49°. 2.BT classification (for the clear set only) 1.Cloud Height classification (for the cloudy set only): 8 overlapping classes between 100 and 1000 hPa (200 hPa range each). CTOP is determined iteratively until the optimal cloud category is selected. 2.CO 2 classification: 5 sets defined by overlapping periods from 2002 to 2013. Annual mean values from ftp:://ftp.cmdl.noaa.gov/ccg/co2/trends/co2_annmean_gl.txt or via 370+2(year-2000). yearsAnn mean CO 2 [ppmv] 12002-2005375.7 22004-2007379.8 32006-2009383.7 42008-2011387.9 52010-2013392.6

43 Cloud Height classification (for the cloudy set only): 8 overlapping classes between 100 and 1000 hPa (200 hPa range each) i.e. 100-300 hPa, 200-400 hPa, ….800-1000 hPa. CTOP is determined iteratively until the optimal cloud category is selected. CO 2 classification: 5 sets defined by overlapping periods from 2002 to 2013. Annual mean values from ftp:://ftp.cmdl.noaa.gov/ccg/co2/trends/co2_annmean_gl.txt or via 370+2*(year-2000). yearsAnn mean CO 2 [ppmv] 12002-2005375.7 22004-2007379.8 32006-2009383.7 42008-2011387.9 52010-2013392.6

44 Dual Regression Retrieval Algorithm – Part 1 Clear trained EOF regression results CLEARCLOUDY EOF regression Clear radiance calculations SARTA v1.7 (UMBC) Regression RTV T, H 2 O, O 3, T skin, Emissivity, CO 2 at single FOV BT, Scanning angle and CO 2 Classification Cloudy radiance calculations Fast RT cloud model (Wei, Yang) Scanning angle, Cloud height and CO 2 Classification T, H 2 O, O 3, T skin, CO 2, CTOP, COT at single FOV Additional Predictors (spres, solzen) Global training data set (SeeBor V5) 15704 profiles Global cloudy training data set 19948 profiles Clear Regression CoefficientsCloudy Regression Coefficients Cloud trained EOF regression results X…atmospheric state, C… regression coefficients, A…compressed radiances (projection coefficients)

45 Cloud Top Altitude Level where T cloudy -T clear > 3 K at all levels below that level Final Profile (clear-trained above, cloud-trained below cloud level) Clear trained EOF regression resultsCloud trained EOF regression results Dual Regression Retrieval Algorithm – Part 2 General Idea

46 1.A retrieval (incl. CO 2 amount) is generated by using the regression set associated with current year. The CO 2 classification is refined if necessary and the retrieval is run again. 2.A first estimate of the CTOP from the clear-trained and cloudy-trained retrieved T profile is obtained. The cloud classification is refined if necessary (i.e. if T derived cloud class differs from cloud-trained solution). 3.If CTOP is close to the surface and the maximum difference between clear and cloudy retrieved Tskin<2.5K the clear-trained solution is used as the as the final at all levels, otherwise the cloud-trained profiles is used. 4.If CTOP≤300mbar the cloud-trained profile is used as the final at all levels. 5.A second (upper level cirrus) cloud top pressure is found if relative humidity exceeds 70% at a level above the cloud altitude derived by the T profile. 6.If maximum difference between clear-trained and cloud-trained T retrieval exceeds 25K the final retrieval below the cloud top is set to missing. 7.Effective cloud emissivity, i.e. (T skin -T skin_gdas )/(T skin_ctop -T skin_gdas ), the cloud condition, and the vertical mean and standard deviation of differences between final T retrieval and GDAS T above and below the cloud top are used to assign quality flags to the final retrieval product. Dual Regression Retrieval Algorithm – Part 2 Some details

47 SurfSkinTemp [K]CTOP [mbar]COTH #/Bias/Stdev/RMSE (orig) 16359/0.45/2.21/2.2616009/-25.9/177.3/79.216009/1.6/2.9/3.3 #/Bias/Stdev/RMSE (DR) 16359/-0.07/1.57/1.5716009/1.1/156.1/156.116009/-0.1/1.7/1.7 Retrieval error statistics (simulated) Unstratified (orig) vs. stratified (DR) Retrieval error statistics (simulated) Unstratified (orig) vs. stratified (DR)

48 CTH comparison with CloudSat and Calipso tropical polar

49 T and H 2 O Error Statistics (comparison with radiosondes over CONUS, Dec 2009) T and H 2 O Error Statistics (comparison with radiosondes over CONUS, Dec 2009)

50 MODIS 1km images (1705, 1710) High Dense Cirrus Cirrus Outflow Eye Hurricane Isabel (Sept-13, 2003)

51 .. AIRS Temperature (at 850 hPa) and Humidity (at 700 hPa)

52 AIRS Temperature and Humidity (at 300 hPa)

53 Isabel Eye Sounding Location: 22.6N, 62W Relative Humidity [%]Temperature [K] Pressure [mbar]

54 MODIS – 19:46 UTC 22 May 2011  Joplin GOES Visible 22:30 UTC  Joplin Joplin Tornado  Joplin Joplin, MO Tornado (May 22, 2011)

55 RH 850 hPa ☐ Joplin RH 700 hPaTemp 500 hPa ☐ Joplin 1.5-km RH Increased 3.0-km RH Decreased 5-km Temp. Decreased IASI (at 15:40 UTC) and AIRS (at 19:45 UTC) H 2 O and Temperature Change IASI (153855Z) May 22, 2011 1540 UTC AIRS (G197) May 22, 2011 1945 UTC

56 IASI (1540 UTC)AIRS (1945 UTC) IASI and AIRS Temperature, Dew Point Temperature and Humidity at Joplin, MO. Dew Point Solid lines: temperature Dashed lines: dew point temperature Temp Lat/Lon: 37.0/-94.6Lat/Lon: 37.1/-94.5 Temperature [K]

57 AIRS at 19:40 vs. RAOB at 18:00 for Springfield (75 mi east of Joplin, MO) Solid lines: temperature Dashed lines: dew point temperature Temperature [K] Pressure [mbar]

58 IASI/AIRS Stability Change (Total Totals)

59 The cloud height stratified dual (EOF) regression retrieval provides all-sky condition vertical temperature and moisture profiles with profiles being retrieved down to cloud level and below thin and/or scattered to broken clouds. The single FOV algorithm is capable to capture the spatial detail of the atmospheric state which will benefit various applications, in particular severe weather forecasting. Algorithm will be integrated in the UW/CIMSS IMAPP AIRS software package. Retrieval products will include TPW, TOC, Lifted Index and CAPE. Improved Profile and Cloud Top Height Retrieval by Using Dual Regression Summary


Download ppt "AIRS (Atmospheric Infrared Sounder) Regression Retrieval (Level 2)"

Similar presentations


Ads by Google