Status of the LBLRTM Forward Model for
Passive Remote Sensing
Mark Shephard, Vivienne Payne, Karen Cady-Pereira and
Tony Clough
Workshop on Advanced High Spectral Resolution Infrared Observations
University of Wisconsin
26-28 April 2006
1
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
Topics
• Status of LBLRTM Analytic Jacobian
• Temperature and Emissivity of the Surface
• Temperature Profiles
• Species Profiles
• Upwelling/Downwelling
• Status of LBLRTM for Passive Remote Sensing
• Temperature
• Retrieval of Spectroscopic Parameters from
High Resolution Spectral Observations
• SHIS,TES, AIRS, AERI
2
Temperature Jacobian Lvl 8
Emissivity = 0.9999
Reflectivity = 0.0001
Temperature Jacobian Lvl 8
(W/(cm^2 sr cm-1 K)
(W/(cm^2 sr cm-1 K
Temperature Jacobians: Level 8
Upwelling at 20 km
Emissivity = 0.0001
Reflectivity = 0.9999
3.0E-9
2.5E-8
2.5E-9
2.0E-8
2.0E-9
1.5E-8
1.5E-9
1.0E-8
1.0E-9
5.0E-9
0.0E+0
5.0E-10
0.0E+0
5.0E-10
-5.0E-9
0.0E+0
-5.0E-9
0.0E+0
2.5E-8
Finite Diff
AJ_Lvl_8
dR_dT Difference expanded * 10
AJ Lvl 8
1.5E-8
1.0E-8
5.0E-9
dR_dT Difference expanded * 10
Finite Diff
2.0E-8
650
700
750
800
850
900
Wavenumber
950
1000
1050
2.5E-9
2.0E-9
1.5E-9
1.0E-9
AJ - Finite Diff
(AJ - Finite Diff)
-1.0E-8
600
3.0E-9
-1.0E-8
600
-5.0E-10
1100
650
700
750
800
850
900
Wavenumber (cm-1)
950
1000
1050
-5.0E-10
1100
1.2E-8
1.4E-9
lblrtm_FD_DN_00_008
dR_dT Difference expanded * 10
Downwelling at the Surface
(W/
Temperature Jacobian Lvl 8
lblrtm_AJ_DN_00_008
1.0E-8
e.g. AERI
8.0E-9
6.0E-9
4.0E-9
2.0E-9
1.2E-9
1.0E-9
8.0E-10
6.0E-10
4.0E-10
0.0E+0
2.0E-10
-2.0E-9
0.0E+0
lblrtm_AJ_00_008 - lblrtm_FD_DN_00_008
-4.0E-9
600
650
700
750
800
850
900
Wavenumber
Clough et al.; 14 March 2006
3
950
1000
1050
-2.0E-10
1100
Atmosphere:
CAMEX Case
27 Levels
0-20 km
(W/(cm^2 sr cm-
Emissivity = 0.9999
Reflectivity = 0.0001
Level 2
TOA Water Vapor Jacobian
Level 2
2.0E-8
2.2E-6
dR_dlnQ Difference expanded * 10
1.4E-8
LVL 2 Finite Diff
AJ Level 2
0.0E+0
-2.0E-8
-4.0E-8
-6.0E-8
-8.0E-8
-1.0E-7
Emissivity = 0.0001
Reflectivity = 0.9999
2.0E-6
Finite Diff Lvl 2
AJ_TOA_Lvl2
1.2E-8
1.8E-6
1.6E-6
1.0E-8
1.4E-6
8.0E-9
1.2E-6
1.0E-6
6.0E-9
8.0E-7
4.0E-9
2.0E-9
AJ - Finite Diff
-1.2E-7
650
700
750
800
850
900
Wavenumber (cm-1)
950
1000
1050
1.6E-7
1.4E-7
1.2E-7
1.0E-7
8.0E-8
6.0E-8
4.0E-8
AJ - Finite_Diff
0.0E+0
-4.0E-7
600
-2.0E-9
1100
1.8E-7
4.0E-7
2.0E-8
-2.0E-7
-1.4E-7
600
2.2E-7
6.0E-7
2.0E-7
0.0E+0
2.4E-7
2.0E-7
dR_dlnQ Diff expanded *10
TOA Water Vapor Jacobian
(W/(cm^2 sr cm-1)
Water Vapor Jacobians: Level 2
Upwelling at 20 km
0.0E+0
650
700
750
800
850
900
950
1000
1050
-2.0E-8
1100
1.6E-6
1.4E-6
e.g. AERI
1.8E-7
lblrtm_FD_DN_01_002
dR_dT Difference expanded * 10
Downwelling at the Surface
(W/
Water Vapor Jacobian
lblrtm_AJ_DN_01_002
1.2E-6
1.0E-6
8.0E-7
6.0E-7
4.0E-7
1.6E-7
1.4E-7
1.2E-7
1.0E-7
8.0E-8
6.0E-8
2.0E-7
4.0E-8
0.0E+0
2.0E-8
0.0E+0
-2.0E-7
lblrtm_AJ_01_002 - lblrtm_FD_DN_01_002
-4.0E-7
600
650
700
750
800
850
900
Wavenumber
Clough et al.; 14 March 2006
4
950
1000
1050
-2.0E-8
1100
Atmosphere:
CAMEX Case
27 Levels
0-20 km
What is ‘Truth’?
•
Spectral Residuals are Key!
•
Consistency within a band system
– Focus of this presentation
•
Consistency between bands
– AIRS ν2 and ν3 bands to investigate consistency for CO2
•
Consistency between species
– TES: temperature from O3 and H2O consistent with CO2 ; N2O
•
Consistency between instruments
- SHIS
- MIPAS
- AIRS
- ACE
- TES
5
Spectral Radiance Measurements
•
The retrieval of atmospheric parameters with InfraRed Remote Sensing is a
Poorly Posed Problem
•
High Accuracy in the Forward Model and the Measured Radiances
is Essential
•
Truth at the Level Required is not readily Available
- sonde accuracies; spatial and temporal sampling
- laboratory measurements
6
Aura Validation Experiment :Nov. 7, 2004 : Gulf of Mexico
AVE Observations to Investigate Line-by-line Calculations
SHIS : ~ 2 km
Spectral resolution: 0.48 cm-1
TES Underflight
Altitude of 18 km
TES : 8 x 5 km
Spectral resolution: 0.06 cm-1
Nadir
19:25 UTC
Brightness Temperature (K) @ 1105 cm -1
SHIS scans - averaged nine ~2 km circles
TES nadir scan - average of the sixteen 0.5 x 5km
rectangular pixels from overpass
7
TES - SHIS Radiance Comparison
• TES Convolved to SHIS ILS
• {TES - LBLRTM(TES Geometry)}
- {SHIS - LBLRTM(SHIS Geometry)}
TES CO2 Filter: 2B1
Error in CO2 above SHIS
TES Ozone Filter: 1B2
Error in O3 above SHIS
TES Water Filter: 2A1
8
Reduction of Residuals in ν2 Region
Simultaneous Retrieval of Temperature and Line Parameters
•
For each Vibration-Rotation Band (8 → 3)
– Line Strength (3 parameters)
– Line Width (2 parameters)
•
Global
– P - R Line Coupling (Hartmann) (scaling)
–
Q
Line Coupling (Hartmann) (scaling)
– Halfwidth Temperature Dependence
– CO2 Continuum (scaling)
9
Corrections to CO2 line intensities and widths
•
Linda Brown
Form of correction to line intensities:
Sret = SHITRAN (1+ a0 ) (1+ a1m + a2 m 2 )2
•
Form of correction to line widths:
W ret = W HITRAN (1 + b0 + b1 | m |)
•
•
m is related to the upper state local quantum number.
For the P branch
m=−J"
For the R branch
m=J"+1
Retrieval parameters: a0, a1, a2, b0 , b1
10
Line Coupling Parameters for the 5 < 2 Band
630
0.25
650
670
690
Wavenumber (cm-1)
710
730
750
770
790
0.07
0.00
-0.25
-0.50
0.05
Q Branch Hartmann et al.
Q Branch Hoke et al.
0.03
P, R Branch Ys
-0.75
-1.00
0.01
-1.25
-1.50
-0.01
-1.75
P Branch Hartmann et al.
-2.00
R Branch Hartmann et al.
-0.03
-2.25
-2.50
-0.05
-2.75
-3.00
-3.25
715
-0.07
717
719
721
Wavenumber (cm-1)
11
723
725
Effect of Line Coupling
12
Sensitivity Studies (Jacobians)
•
a0, a1, a2, b1 perturbed to
change intensities/widths in
HITRAN line file.
Reference
spectrum
•
Values chosen to induce 10%
change in intensity/width at J’’
~ 40.
a0
•
Perturbation spectra calculated
for each parameter, each band.
a1
a2
Example: 2‡1 band
b1
Wavenumber [cm-1]
13
SHIS AVE Retrieval Results
Temp &
Spectroscopy
Temp
Only
GMAO
Scatter plots of Brightness
Temperature vs. Residuals
(Obs - Calc.) binned in 20 cm-1
intervals from the CO2 v2 region
Retrieval Spectral Region
14
Detail View Before Spectroscopy
•
Further reduce residuals in the CO2 v2 spectral region
15
SHIS AVE Temperature Retrieval Results
Temp Only
16
Temp &
Spectroscopy
Averaging Kernels
17
SHIS AVE CO2 Retrieved Parameters
Band
Band
center
(cm-1)
32
618.03
2 1
667.38
5 2
720.81
Retrieved Parameter
ALL Bands
P/R Line Coupling
1%
1%
a0
b0
a2
- 2%
a2
b0
- 4%
3%
CO2
CNTMN
2x
- 50%
Sret = SHITRAN (1 + a0 ) (1+ a1m + a2m2)2
Wret = WHITRAN(1 +b0 +b1 |m|)
18
What is ‘Truth’?
•
Spectral Residuals are Key!
•
Consistency within a band system
– Focus of this presentation
•
Consistency between bands
– AIRS ν2 and ν3 bands to investigate consistency for CO2
•
Consistency between species
– TES: temperature from O3 and H2O consistent with CO2 ; N2O
•
Consistency between instruments
- SHIS
- MIPAS
- AIRS
- ACE
- TES
19
AIRS: new spectroscopy
Atmosphere: sonde/adjusted ecmwf
20
‘Truth’ isn’t ubiquitous- Yet!
•
Spectral Residuals are Key!
•
Consistency within a band system
– Focus of this presentation
•
Consistency between bands
– AIRS ν2 and ν3 bands to investigate consistency for CO2
•
Consistency between species
– TES: temperature from O3 and H2O consistent with CO2 ; AIRS: CO2 , N2O
•
Consistency between instruments
- SHIS
- MIPAS
- AIRS
- ACE
- TES
21
Back-up slides to follow………
22
CO2 Line Shape
LBLRTM Chi Function for CarbonDioxide
1.0
CO2 Continuum
impact
Clough et al., 1978
Cousin (a) N2 296K
0.8
Cousin (c) N2 238K
Chi Function
Chi_LBLRTM_v_8
0.6
0.4
0.2
Lorentz (impact)
lblrtm_v6.01
0.0
0
5
10
15
20
Wavenumbers from Line Center (cm-1)
25
Chi function: The spectral function by which the
Lorentz line shape (impact) must be
multiplied to obtain the “true” line shape.
Sub-Lorentzian chi function accounts
for the duration of collision effects.
23
lblrtm_v8.2
Continuum: Line contributions 25 cm-1
beyond line center
Perturbations to line intensities/widths: 32 band
Reference
spectrum
a0
a1
a2
b1
24
Perturbations to line intensities/widths: 21 band
Reference
spectrum
a0
a1
a2
b1
25
Perturbations to line intensities/widths: 42 band
Reference
spectrum
a0
a1
a2
b1
26
Perturbations to line intensities/widths: 52 band
Reference
spectrum
a0
a1
a2
b1
27
Perturbations to line intensities/widths: 63 band
Reference
spectrum
a0
a1
a2
b1
28
Perturbations to line intensities/widths: 74 band
Reference
spectrum
a0
a1
a2
b1
29
Perturbations to line intensities/widths: 85 band
Reference
spectrum
a0
a1
a2
b1
30
Perturbations to line intensities/widths: 84 band
Reference
spectrum
a0
a1
a2
b1
31
Averaging Kernels
32
Correlation matrix (H-1)
33
LBLRTM Validation : TES CO2 v2 Filter
TES Obs.
TES - LBLRTM
TES(SHIS ILS) - LBLRTM
SHIS - LBLRTM
CO2 v2
34
SHIS AVE CO2 Width and Intensity Retrieval Results
Band
Band
center
(cm-1)
3 2
618.03
a0
a1
a2
b1
-0.0053
-0.00051
0.000004
0.00055
2 1
667.38
a0
a1
a2
b1
0.077
0.00022
-0.000039
-0.00077
5 2
720.81
a0
a1
a2
b1
0.015
0.00028
-0.000016
-0.00075
Retrieved
value
Sret = SHITRAN (1 + a0) (1 + a1m + a2m2)2
Wret = WHITRAN(1+b1|m|)
35