SeaWinds Radiometer (SRad)
Brightness Temperature
Calibration/Validation
Mayank Rastogi
EE Masters Defense
March 29, 2005
Major Advisor: Dr. W. Linwood Jones
Presentation Outlines
• Background
• Principles of Scatterometry/Radiometry
• Revealed Tb differences - Need for Asc/Dsc Corrections
• Radiometric Calibration Procedure
• SRad/AMSR Empirical Asc/Dsc Correction
• Validation of Empirical Correction with independent data-set.
• Conclusions
Background
• SeaWinds is a satellite-borne radar scatterometer
used to remotely sense oceanic surface winds
– Launched June 1999 on the low-earth polar orbit
QuikSCAT satellite
• Issue
– When rain occurs the antenna noise increases
– Approx. 10% of wind retrievals are corrupted by rain
• Resolution
– Previously, CFRSL developed a signal processing
algorithm to measure the antenna noise and “flag” rain
contamination
• MS theses: Mehershahi, 2000 & Susanj 2000
Background cont.
• A second SeaWinds instrument launched on
Japan’s Advanced Earth Observing Satellite
in December 2002
– This thesis concerns the evaluation of the
SeaWinds Radiometer (SRad) on ADEOS-II
Thesis Objective
• Radiometric calibration of the SeaWinds Radiometer (SRad) on ADEOS-II– Evaluate previous QuikSCAT Radiometer (QRad)
transfer function
– Tb cross-calibration with the Advanced Microwave Scanning Radiometer (AMSR) also on ADEOS-II
• Validate SRad brightness temperature (Tb) measurements over ocean using independent AMSR equivalent Tb’s
Principles of
Scatterometry/Radiometry
Black Body Radiation
1
1hc2)S(
5
2
kT
ch
e
S() 2ckT
4
Rayleigh-Jean’s Approximation
(valid over microwave region):
Planck’s law:
According to thermodynamic principles, all matter at
temperatures greater than absolute zero absorb and emit non-
coherent EM energy.
Where: h = Planck’s constant
k = Boltzmann’s constant
c = Speed of light
T = Physical temperature
Region of interest
Microwave
Plank’s Radiation Law
Microwave Radiometry• Blackbody power is:
Pblackbody = k *Tphys* Bandwidthmicrowave antenna measures equivalent blackbody power
For non-blackbodies, microwave brightness temperature, Tb, of a medium is:
where:
• Brightness temperature is the equivalent blackbody temperature
is the emissivity, which varies with– EM wavelength
– Incidence angle and Polarization
– Complex index of refraction
Physb TT
Microwave Brightness Temperature over Ocean
• Measured brightness temperature:– Upwelling atmospheric emission
– Downward atmospheric emission reflected by the sea surface
– Ocean emission
Surface
Emission
Upwelling
Atmospheric
Emission
Reflected
Atmospheric
Emission
Microwave
Antenna
Ocean
A
otr dA
R
GPP
4
2
3
2
)4(
Satellite Microwave Scatterometer
• Radar Equation
• o Normalized Radar
Cross-section of the ocean
surface
Geometry of SeaWinds
Scatterometer on ADEOS-II
54°
46°
Swath = 1800 km
SRad Simplified Functional Diagram
Be
Bn
Frequency
Pow
er
Received Spectrum
Radar Echo
Echo Channel
Noise Channel
SeaWinds Received Power Spectrum
antenna noise = noise chan pwr - echo chan pwr
Early On-orbit Check-out of SRad
Revealed Anomalous Tb Differences Ascending & Descending Orbit Segments
• Compared 3-days averages of asc/desc revs
for QRad and SRad
– QRad 3-day average ascending and descending
Tb’s were equal, but
– SRad ascending Tb’s were colder by ~ 5 K
• Both polarizations exhibited this anomaly
SRad Ascending / Descending Tb
Anomaly (3-day ocean avg)
SRad Ocean Tb, H-Pol QRad Ocean Tb, H-Pol
Descending:Mean = 109
Ascending:Mean = 103 Ascending:
Mean = 106
Descending:Mean = 106
SRad Ascending / Descending Tb
Anomaly (3-day ocean avg)
SRad Ocean Tb, V-Pol QRad Ocean Tb, V-Pol
Ascending:Mean = 176
Descending:Mean = 181
Ascending:Mean = 180
Descending:Mean = 180
Ascending-Descending Tb Anomaly
ADEOS-II
(SRad)
QSCAT
(QRad)
ASDS
Radiometric Calibration
Procedure:Cross-calibration of SRad &
AMSR Brightness Temperatures
Why calibrate the instrument ?
• SeaWinds not intended for radiometric
measurements - does not have usual “hot”
and “cold” internal brightness temperature
calibration sources.
Calibration Approach
• Use Advanced Microwave Scanning
Radiometer (AMSR) and compare
simultaneous Tb.
Advanced Microwave Scanning Radiometer (AMSR)
• Multi- frequency microwave radiometer
- 6.9, 10.65, 18.7, 23.8, 36.5, and 89.0 GHz
- Vertical and Horizontal polarizations
- Incidence angle = 55º
SeaWinds Scatterometer
• Specialized active microwave sensor
- Ku-band, 13.4 GHz
- Vertical and Horizontal polarizations
- Incidence angle » H-Pol = 46º
» V-Pol = 54º
Get AMSR Tb at 10.65 GHz & 18.7 GHz
Generate Modeled AMSR & SRad Tb using RadTb
& SSMI-F-15 Environmental Parameters
Check Biases of AMSR Tb Vs Modeled AMSR Tb
Does it Need
Correction
Yes No
Procedure Algorithm
Yes No
Find Asc/Dsc Correction
Create Spectral Ratios
Get SRad Tb
Generate SRad Equiv
AMSR Tb using Spectral
Ratios
Check Biases of SRad Vs AMSR
Find Empirical Asc/Dsc Corrections
Validation of Empirical Corrections with independent
data-set
Model AMSR/SRad Tb
RadTb was run with:
• SRad incidence angles (H-Pol=46deg V-
Pol=54deg) – get SRad Tb at 13.4 GHz
• AMSR incidence angle of 55deg with 2 freq
- get AMSR Tb at 10.65GHz & 18.7GHz
• Environmental pars from NCEP and SSMI
- SST, Water Vapor, Cloud Liquid & Wind Speed
13.4 10.65
18.7 10.65
( )( )
( )
T Tspecratio
T T
• Spectral Ratios were found to be a function of
predominantly Water Vapor (WV)
H-Pol Spectral Ratio Vs Water Vapor
AMSR Asc/Dsc Correction
0 500 1000 1500-4
-3
-2
-1
0
1
2
3Average Correction ASC/DSC H-Pol 0429 0624 0922
AMSR Asc/Dsc Correction H-Pol
0 500 1000 1500-1
0
1
2
3
4
5Average Correction ASC/DSC V-Pol 0429 0624 0922
AMSR Asc/Dsc Correction V-Pol
Generate SRad Equiv AMSR Tb using
Spectral Ratios
• SRad Equiv AMSR Tb (at 13.4 GHz) was
generated from AMSR 10.65 & 18.7 GHz
Tb’s using Spectral Ratio
13.4 10.65 18.7 10.65*( )Tb Tb specratio Tb Tb
SRAD/AMSR Tb Differences
• Calculate Tb differences (biases) between
SRad Equivalent AMSR Tb and SRad Tb for
both polarizations and ascending/descending
orbits
• Model Tb biases as a three-term Fourier
series of orbit latitude
– Latitude range for one orbit includes ascending
and descending segments = 2 x 180°
SRad/AMSR Biases Along Longitude
• There were variations in Tb biases with longitude because of the temperature differences along land/ocean boundaries
– Caused by antenna pattern differences between SRad and AMSR
– Land is radiometrically “hot” compared to ocean
• New “conservative” land mask created to eliminate of land effects
SRad/AMSR Biases along Land/Ocean Boundaries
(9-Day Average)
SRad-AMSR Bias Sep 9-Day Avg. H-Pol Asc
200 400 600 800 1000 1200 1400
100
200
300
400
500
600
700-50
-40
-30
-20
-10
0
10
20
SRad-AMSR Bias 9-Day Avg H-Pol Asc using conservative LandMask
200 400 600 800 1000 1200 1400
100
200
300
400
500
600
700-50
-40
-30
-20
-10
0
10
20
SRad/AMSR Biases after applying Conservative
Land Mask
-90° +90° -90°Orbit Latitude:
Del
ta-T
b (
SR
ad -
AM
SR
), K orbit avg
day-2orbit avg
day-3
orbit avg
day-1
SRad Asc/Dsc Tb (H-Pol) Correction
Orbital Pattern, Average for 3-days
Fourier Series
-120 -100 -80 -60 -40 -20 0 20 40 600
500
1000
1500
2000
2500
3000
3500
4000
4500 SRAD/AMSR Bias H-Pol ASC Without Correction
MEAN = - 5.7346
-100 -80 -60 -40 -20 0 20 40 600
500
1000
1500
2000
2500
3000
3500
4000
4500 SRAD/AMSR Bias H-Pol ASC With Correction
MEAN = 0.2118
SRad/AMSR Bias H-Pol Ascending
Without Correction
Mean = -5.7346
With Correction
Mean = 0.2118
-120 -100 -80 -60 -40 -20 0 20 40 600
500
1000
1500
2000
2500
3000
3500
4000
4500 SRAD/AMSR Bias H DSC Without Correction
MEAN = -3.1126
-100 -80 -60 -40 -20 0 20 40 60 800
500
1000
1500
2000
2500
3000
3500
4000
4500 SRAD/AMSR Bias H DSC With Correction
MEAN = 0.1792
SRad/AMSR Bias H-Pol Descending
Without Correction
Mean = -3.1126
With Correction
Mean = 0.1792
SRad/AMSR Asc/Dsc Tb Biases
Mean (K) STD (K) # pts
H-pol Asc -5.7 3.0 52,253
V-pol Asc -6.4 2.7 66,584
H-pol Dsc -3.1 2.3 52,031
V-pol Dsc -1.6 2.3 67,545
Mean (K) STD (K) # pts
H-pol Asc 0.21 3.0 52,253
V-pol Asc 0.41 2.1 66,584
H-pol Dsc 0.18 2.3 52,031
V-pol Dsc 0.70 2.3 67,545
(a) Original algo - before correction
(b) Revised algo - after correction
• Validation of refined SRad Tb algorithm
through independent comparisons between
SRad and AMSR
• After empirical correction, ocean delta Tb
bias is reduced to approx ± 2 K over latitude
compared to a peak value of -10 K before
correction
SRad Tb Algorithm Validation
100 150 200 250 300 350 400 450 500 550 600-18
-16
-14
-12
-10
-8
-6
-4
-2
0
2
Latitude Index
Delta
Tb(K)
H-Pol Asc Without SRad Correction 090103 L2A Data
100 150 200 250 300 350 400 450 500 550 600-10
-8
-6
-4
-2
0
2
4
6
8
Latitude Index
Delta
Tb(K)
H-Pol Asc With SRad Correction 090103 L2A Data
Without empirical correction
With empirical correction
-55° 0° +55°
Latitude
Del
ta-T
b, K
Del
ta-T
b, K
(SRad - AMSR) H-pol delta-Tb Asc
100 150 200 250 300 350 400 450 500 550 600-10
-8
-6
-4
-2
0
2
4
Latitude Index
Delta
Tb(K)
H-Pol Dsc Without SRad Correction 090103 L2A Data
100 150 200 250 300 350 400 450 500 550 600-8
-6
-4
-2
0
2
4
6
Latitude Index
Delta
Tb(K)
H-Pol Dsc With SRad Correction 090103 L2A Data
Without empirical correction
With empirical correction
+55° 0° -55°
Latitude
(SRad - AMSR) H-pol delta-Tb DscD
elta
-Tb, K
Del
ta-T
b, K
0 100 200 300 400 500 600 700 800-15
-10
-5
0
5
10
15SRad-AMSR Tb Vs Latitude H-Pol Ascending
Latitude Index
Delta
Tb (k
)
0 100 200 300 400 500 600 700 800-15
-10
-5
0
5
10
15
Latitude Index
Delta
Tb (K
)
SRad-AMSR Tb Vs Latitude Descending
SRad Tb Algorithm Validation,
Zonal Average Ocean delta Tb
South Pole North Pole
North PoleSouth Pole
3-Day Ascending orbit avg
3-Day Descending orbit avg
80 100 120 140 160 180 200
80
100
120
140
160
180
200
AMSR Brightness Temp(K)
SRad Tb Algorithm Validation,
Sept. 1, 2003 - Asc/Dsc Segments
Slope Offset
Asc 1.02 -3.98
Dsc 1.01 -0.93
"Blue" = Ascending
"Red" = Descending
AMSR Brightness Temp, K
SR
ad B
rightn
ess
Tem
p,
K
V-pol
H-pol
Conclusions
• Empirical Tb corrections are developed for SRad using AMSR simultaneous ocean Tb comparisons
– Function of orbit position (latitude, Asc/Dsc)
• After applying empirical correction, results demonstrate that good quality SRad brightness temperatures are obtained over the oceans.
SRad/AMSR Biases along Land/Ocean Boundaries
(9-Day Average)
SRad/AMSR Biases after applying Conservative
Land MaskSRad-AMSR Bias 9-Day Avg H-Pol Asc using conservative LandMask
200 400 600 800 1000 1200 1400
100
200
300
400
500
600
700-50
-40
-30
-20
-10
0
10
20