Hank Revercomb, Dave Tobin, Fred Best, Steve Dutcher, Bob Knuteson, Dan LaPorte, Joe Taylor, Ken Vin

1
Hank Revercomb, Dave Tobin, Fred Best, Steve
Dutcher, Bob Knuteson, Dan LaPorte, Joe Taylor,
Ken Vinson  University of Wisconsin-Madison,
Space Science and Engineering Center
TES Radiance Validation using Scanning
High-resolution Interferometer Sounder (S-HIS)
observations during AVE
Aura Validation Expt Science Team Meeting CSC
Building, Lanham-Seabrook , MD 19-20 September
2005
2

• TOPICS
• S-HIS summary
• S-HIS Radiometric Calibration
• TES Validation
• TES Radiometric PerformanceIssue and
  Characterization

3
1. S-HIS summary
4
UW Scanning HIS 1998-Present(HIS
High-resolution Interferometer Sounder, 1985-1998)
Characteristics Spectral Coverage 3-17
microns Spectral Resolution 0.5
cm-1 Resolving power 1000-6000 Footprint
Diam 1.5 km _at_ 15 km Cross-Track Scan
Programmable including uplooking zenith view
Applications

• Radiances for Radiative Transfer
• Temp Water Vapor Retrievals
• Cloud Radiative Prop.
• Surface Emissivity T
• Trace Gas Retrievals

5
Scanning HIS Interferometer with Telescope,
Collection optics, Detectors/Cooler
Laser metrology located on bottom of optical bench
6
S-HIS In-Flight Calibration

• Hot and Cold onboard BBs viewed every x-track
  scan (12 sec).

Cooling fins closely couple the Cold Blackbody
to the Pod Ambient Air Temperature.
Ambient BB(225-260 K)
Hot BB(300-310 K)
7
Scanning HIS has flown on Proteus, NASA ER2,
DC-8, as well as the WB57
S-HIS scans cross-track downward looks upward
8
Okavanga Delta Surface Emissivity ( 27 August
2000)
Tb (980 cm-1)
Tb (980) - Tb(1125)

6
-2
Tb(K)
1250 cm-1
750 cm-1
9
S-HIS zenith and cross-track scanning Nadir views
11-16-2002 from Proteus _at_ 14km
10
2. S-HIS Radiometric Calibration
11
Scanning-HIS Radiometric Calibration BudgetTABB
227, THBB310, 11/16/02 Proteus Similar to AERI
(GIFTS) description in Best, et al., CALCON 2003
(2005)
3-sigmaTb error lt 0.3 K for Tb gt220 K
RSS ofErrors in THBB,TABB TRfl ?HBB, ?ABB
10 of non-linearitycorrection
0.2 K
12
Scanning-HIS LW/MW and MW/SW Band
Overlap11-16-2002
LW HgCdTe band
MW HgCdTe band
SW InSb band
LW/MW overlap
MW/SW overlap
13
The NIST Connection Comparisons with NIST
maintained blackbodies conducted with
ground-based AERI. S-HIS employs the same
calibration approaches Direct test of
S-HIS planned using NIST Transfer Radiometer
(TXR) at aircraft flight temperatures
Max Differencelt 0.055C Longwavelt 0.035C
Shortwavebetween 293 333 K
Miami, 1998
14
Calibration Emphasis
Make full use of the fundamental advantage of
high resolution infrared spectra to provide a
new standard of accuracy for weather and climate
applications
inherent advantage of high spectral resolution
for calibration accuracy (Goody Haskins, 1998)
Now concerned with tenths of K, not 1 K!
15
3. TES Validation
16
Aura Validation Expt AVE, Oct/Nov 2004
NASA WB57
S-HIS views cross-track downward zenith
Scanning HIS
Left Wing Pod
17
Aura Validation Experiment (AVE2004) Flight
paths, Oct/Nov 2004
NASA WB57
18
S-HIS Spectra, 10.6 ?m O3AVE, 26 October 2004
19
S-HIS Spectra, 4.67 ?m COAVE, 26 October 2004
Note good uplooking zero
20
S-HIS Spectra, SW/4.3 ?m CO2AVE, 26 October 2004
Apodized
Zenith (good zero)
non-zero agreement
21
S-HIS Spectra, 4.73 ?m O3AVE, 26 October 2004
Apodized
CO2
O3
22
S-HIS Tropospheric Emission Spectrometer (TES)
Bands near 31 Oct overpass
CO2
?5.5 x 16 km
O3
N2O
CH4
CO2
CO
N2O
H2O
CO2
23
31 Oct 2004 TES-SHIS Comparison Flight
S-HIS Flight Legs
TES Samples
24
SHIS 900 cm-1 Brightness temperature for
Selected TES Footprints (10-31-04)
25
SHIS compared to TES Brightness temperature for
scan 0 TES Footprint (10-31-04)
26
TES/SHIS Brightness temperature comparisonsfor
TES scan 6, 10-31-04
27
7 Nov 2004 TES-SHIS Comparison Flight
28
Scene Uniformity Indicator from S-HIS Scan Angle
Variation
29
TES scan 8 _at_ 192449, SHIS _at_ 1856
SHIS 900 1/cm Tb (K)
SHIS scan angles -15.3, -20.9, -25.6 degrees
30
Limitations of current raw comparisons

• TES mean of scans 0 and 5 (3x5 km) are compared
  to the mean of SHIS spectra within 8.5 km of the
  center of the TES footprint. No calculations
  have yet been incorporated to account for
  altitude and view angle differences.
• For this case, the aircraft nadir track does not
  appear to follow the TES nadir track and the mean
  SHIS view angle is 20 degrees or so, and so
  incorporating the calculations done at the
  correct view angles will be important.

31
Tb (K)
TES scan 8
SHIS LW SHIS MW TES 1B2 TES 2A1
Tb (K)
TES scan 10
Both at S-HIS spectral resolution (0.5 cm-1) and
oversampled
32
Tb (K)
SHIS LW TES 1B2
TES scan 8
Compared on next slide
Tb (K)
TES scan 10
33
Brightness temperature histograms for selected
wavenumber regions 940-970 cm-1
TES scan 8 TES 1B2 SHIS
TES scan 10 TES 1B2 SHIS
Tb (K)
Tb (K)
34
Brightness temperature histograms for selected
wavenumber regions 940-970 cm-1
TES scan 8 TES 1B2 SHIS
TES scan 10 TES 1B2 SHIS
Tb (K)
Tb (K)
35
Tb (K)
SHIS LW SHIS MW TES 2A1
TES scan 8
Compared on next slide
Tb (K)
TES scan 10
36
Brightness temperature histograms for selected
wavenumber regions 1177-1205 cm-1
TES scan 8 TES 2A1 SHIS
TES scan 10 TES 2A1 SHIS
Tb (K)
Tb (K)
37
Tb (K)
TES scan 8
SHIS MW TES 2A1
Tb (K)
TES scan 10
38
TES (blue)-Scanning HIS (red) Radiance
differences Smoothly varying with wavenumber
Scan 8, 7 Nov 04
TES - SHIS
Smoothness not indicativeof Calibrationerrors
39
Summary of TES Radiance Validation from AVE2004

• TES and Scanning HIS agree to order 1K or better
• TES band-to-band differences in the 1B2/2A1
  spectral overlap region is variable, on the order
  of /- 0.5 K from scan to scan
• TES and Scanning-HIS differences on the edges of
  spectral bands 1B2 and 2A1 also vary by similar
  amounts
• Variable TES-TES and TES-SHIS differences are
  largely spectrally smooth, symptomatic of
  interferometric noise (not calibration error)

40
4. TES Radiometric PerformanceIssue and
CharacterizationEvidence that TES scan-to-scan
instability (order 1 K or less) is caused by
sample-position errors that may be correctable
41
Variations of Calibrated TES Blackbody spectra
estimate the size of the effect

• Errors are correlated in wavenumber as depicted
  by the purple curve below (for band 2A1)
• Earth radiance errors will be of similar smoothly
  varying shape, but somewhat smaller
  (corresponding Tb will of course not be smooth)

42
Symptom 1 Phase of the raw spectra oscillate
creating 1-2 radiance unit oscillation of ImPart
of calibrated spectra
mW/m2 sr cm-1
88 sec period (apparent)
43
Symptom 2 Rocking of magnitude spectra,
consistent with sample position error mechanism
2A1 Runs 2261 and 2263 BBODY magnitudes, scanDir0
Differences from mean, Run 2261
DIF
Differences from mean, Run 2263
wavenumber
44
Symptom 3 Magnitude variation is 90º out of
phase with Phase ImPart oscillations
See red curves for Blackbody views
Calibrated spectrum shows the same dependence
45
Symptom 3 Verified 2 Obits of Blackbody Data
Show Circular Lissajous pattern demonstrating
that Magnitude phase oscillations are
consistently 90º out of phase
Lissajous ofMagnitude vs Phase
Phase
Magnitude
Sampling conceals relationship
Run 2146 with 160 s segments consisting of 3 BB
scans (forward and back),4 space scans, and
about 90 seconds idle
46
Simulation Pure frequency sample-position error
creates the observed error properties
Sinusoidal disturbance ? 1sec
Magnitude 90º out of phase from ImPart Phase
Note 4 s sampling can alias actual oscillation
frequency
47
Simulation (2) Magnitude errors show spectral
dependence similar to that observed with TES
1 Hz disturbance
TES
48
Possible mechanism OPD speed variation, coupled
with a laser-IR sample delay mismatch

• Order 100 Hz OPD speed variation is required to
  give the observed magnitude variations (highly
  aliased into observed 88 sec period)
• Measured OPD speed spectra have frequency options
  that could be order 100 Hz, aliased into close to
  0 Hz (Nyquist frequency is about 50 Hz)

49
Conclusions on Oscillations in TES Spectra

• Nature Small rocking of calibrated radiance
  spectra, reaching order 1 K near band edges
• Mechanism Sample position error, possibly
  related to Michelson mirror OPD speed variations
• Identification Calibrated spectra are affected
  least when the non-zero imaginary part of the
  calibrated spectrum is largest assuming no
  problems with phase ambiguity removal
• Characterization Value Assigning error
  estimates for retrieval, etc, and avoiding most
  affected spectra regions
• Correction may be possible Based on the
  sample-position error representation used for
  simulation (we dont have it yet)

50
Summary

• TES radiance validation is progressing well,
  order 1 K agreement with Scanning HIS has been
  shown, and further TES calibration improvements
  are an active endeavor
• The presence of a small, but significant,
  sample-position-error is consistent with both
  Scanning HIS comparisons and with the TES data
  itself.
• Recent error characterization will help optimize
  analyses and may allow a correction scheme to be
  implemented