MCST Master Slide

1
Study of Sensor Inter-calibrationUsing
CLARREO Jack Xiong, Jim Butler, and Steve
Platnick NASA/GSFC, Greenbelt, MD 20771 with
contributions fromMODIS Characterization Support
Team (MCST), NASA/GSFC
CLARREO Workshop, 21-23 October 2008, Washington,
DC
Page 1
2
Outline

• Introduction
• Applications of CLARREO
• Requirements for CLARREO Observations
• Inter-calibration Approaches Using CLARREO
• Lunar observations
• VIS/NIR/SWIR only
• Simultaneous nadir observations (SNO)
• Ground observations (Dome C)
• Inter-comparison of Terra and Aqua MODIS
  calibration
• Sensitivity Study (future work)
• RSR sensitivity study (using Schiamachy with
  LaRC)
• Spectral and spatial sensitivity studies of Earth
  view targets (Hyperion and AVIRIS)
• Summary

Page 2
3
Introduction

• Applications of CLARREO
• Benchmark Observations
• Accurate, stable, SI traceable, spectrally
  resolved
• Inter-calibration for Other Sensors
• Consistent data records for long-term climate
  change
• Requirements for CLARREO Observations
• General Requirements
• Spectral spatial temporal orbital
• Special Requirements
• Maneuver (pointing) capability
• Lunar observations

Page 3
4
Inter-calibration Approaches Using CLARREO

• Lunar observations
• MODIS (T/A), SeaWiFS, Hyperion, VIRS/TRMM
• Simultaneous nadir observations (SNO)
• MODIS (T/A), AVHRR, AIRS, MISR, ASTR, Landsat,
  GLI, VIRS/TRMM
• Ground observations
• Dome Concordia, Antarctica
• MODIS (T/A), AVHRR, AIRS, MISR, ASTR
• Inter-calibration of Terra and Aqua MODIS

Collaboration with USGS (Moon), NOAA, JPL, JAXA,
USGC (SNO/Dome C)
Page 4
5
Inter-calibration Using Lunar Observations
Images form Aqua MODIS (band 1) Lunar
Observations (Oct 04 Jun 05)
MODIS Lunar Observations via spacecraft
maneuvers at fixed phase angle reference to a
lunar model (USGS) using integrated irradiance
Advantages vs disadvantages
Page 5
6
Inter-calibration Using SNO
Terra MODIS
Aqua MODIS
SNO
SNO
AVHRR
For RSB
For TEB
Advantages vs disadvantages
Page 6
7
Inter-calibration Using a Ground Target

• Why using a ground target
• Validate on-board calibration complement other
  cal/val approaches
• Monitor calibration long-term stability
• Support sensor inter-calibration
• Requirements for a ground calibration target
• Spectral and spatial uniformity and radiometric
  stability (minimum environmental impact)
• Site accessibility and data availability
• Ground measurements of radiometric traceability

Aqua MODIS SD Degradation (0.41 to 0.94mm) Any
Impact on Calibration?
Page 7
8
Examples of Using Dome C for Inter-calibration

• Site Description
• Data Selection
• Methodology
• Thermal emissive reference to Automated Weather
  Station (AWS) measurements
• Solar reflective BRDF model based on ground
  measurements over Antarctic snow
• Results from MODIS
• Recent presentation at SPIE Europe Remote Sensing
  (Xiong et al. 2008)
• Future Work

Page 8
9
Site Description

• Located on Antarctic Plateau (75.1 S, 123.4 E)
• One of the most homogeneous land surfaces on
  earth in terms of surface temperature and
  emissivity.
• Uniformity over spatial scales typical of the
  ground footprint of satellite sensors
• High altitude (3200 m) minimal slope
• Low snow accumulation rate
• Extremely dry, cold rarefied atmosphere
• Low fractional cloud coverage
• Low atmospheric aerosol and water vapor content
• Permanently manned Research Station now
  operational
• AWS data available since 1995
• 10-minute averages of meteorological parameters
  (T, RH, WS, WD, P)
• Daily radiosonde measurements
• Frequent satellite overpass

Dome Concordia Antarctica
CEOS Endorsed Site NASA/NOAA/ESA Effort
Page 9
10
Data Selection

• MODIS Collection 5 Level 1B data
• Multiple MODIS observations each day (8) at
  different angles of incidence.
• Only near-nadir overpasses used (nadir track
  within /- 50 km of Dome C). One granule every
  2-3 days.
• 20x20 pixel average centered on Dome C
• No cloud screening applied for TEB. All granules
  used.
• Uniformity screening applied for RSB to eliminate
  any granules showing greater than 2
  non-uniformity in reflectance over the 20x20
  pixel area

Page 10
11
Methodology and Approach (TEB)
AWS surface temperature measurements are used as
a proxy to track any trends in the relative bias
between MODIS Terra Aqua. DTMODIS BTMODIS
TAWS Relative Bias DTTerra
DTAqua Relative Bias calculated for each MODIS
band and only for days with measurements from
both Terra Aqua. Applications to other
sensors relative spectral response (RSR),
spatial resolution (ground footprint)
Page 11
12
Methodology and Approach (RSB)
A BRDF model developed by Warren et al. (JGR
1998) based on near-surface reflectance
measurements over the Antarctic snow R
(?,?,f) c1 c2cos(p- f) c3cos2(p- f)
c1, c2 and c3 are functions of cos(?) and
cos(?) c1 a0 a11 - cos(?), c2 a21
- cos(?), c3 a31 - cos(?) ai b0i
b1i cos(?) b2i cos2(?) (i 0, 1, and
2) where ? is the incident solar zenith angle, ?
is the viewing zenith angle, and f is the
relative azimuth angle. A ratio of the observed
reflectance factor r to modeled reflectance
factor R is calculated by ?r r / R Spectral
BRDF of Antarctic snow from Hudson, Warren et al
(JGR 2006)
Page 12
13
Sensor (11 and 12 mm) and AWS Observations
Good correlation between sensor and AWS
observations (focusing long-term behavior, not
individual observations)
Page 13
14
DTMODIS BTMODIS TAWS
11mm
12mm
Terra Black diamonds Aqua Blue squares
Long-term draft (lt10mK) for bands 31 and 32 high
quality on-board TEB calibration
Relative Bias DTTerra DTAqua (time)
Excellent calibration consistency (11mm
0.0252.984K 12mm 0.0133.010K)
Relative Bias DTTerra DTAqua (temperature)
No obvious temperature dependent bias (lt20mK)
Page 14
15
Inter-comparison of Aqua MODIS and AIRS at 11
mm (using Dome C observations)
Dome C data (near nadir)
Old version New version
One orbit June 20, 2006 (near nadir footprints)
190K 330K
Page 15
16
Sensor (0.65 and 0.86 mm) Observations
strong correlation between sensor observations
and solar zenith angle
Page 16
17
Sensor (0.65 and 0.86 mm) Observations versus
Modeled Values
Model parameters derived using sensor first-year
observations
Averaged fitting residual 1.3 1.9
Page 17
18
Terra MODIS observations (0.65mm) over Dome C
(2002 - 2003)
Page 18
19
Terra MODIS observations (0.86mm) over Dome C
(2002 - 2003)
Page 19
20
Terra and Aqua MODIS observations over Dome C
Page 20
21
Sensitivity Study (future work)

• RSR Sensitivity Study
• Extend from our previous study reported at
  April/May CLARREO workshop (e.g., preferred
  inter-calibration scene types vs. spectral band)
• Work with LaRC using Schiamachy data
• Spectral and Spatial Sensitivity Studies
• Hyperion observations, Dome C and other targets
• 0.4 to 2.5mm, 30m IFOV (_at_705km altitude)
• AVIRIS observations
• 0.4 to 2.5mm, 1 mrad IFOV

Page 21
22
http//eo1.usgs.gov/hyperion.php
Page 22
23
http//aviris.jpl.nasa.gov/html/aviris.instrument.
html
Page 23
24
Summary

• Sensor Inter-calibration Using CLARREO
• Improve current sensor inter-calibration
  approaches with highly accurate, stable, and
  spectrally resolved observations
• Resolve calibration differences or establish
  calibration consistency among sensors with
  on-orbit SI traceable measurements
• Ground Target Characterization Using CLARREO
• Extend consistent data records, using
  observations from previous, current, and future
  missions/sensors, for studies of long-term
  climate changes
• Dome C site can be used to track sensor long-term
  stability and calibration consistency among
  sensors (challenges in VIS/NIR/SWIW)
• Other Approaches
• Lunar observations SNO

Page 24