AVHRR Visible Band Calibration Intercalibration for Climate Studies

1
AVHRR Visible Band Calibration / Intercalibration
(for Climate Studies)
Andrew Heidinger and Michael Pavolonis Changyong
Cao, Aleksandar Jelenak, Jerry Sullivan, Fred
Wu NOAA/NESDIS Office of Research and
Applications Cooperative Institute for
Meteorological Satellite Studies (CIMSS) Madison,
Wisconsin
2nd CORP Science Symposium Madison Wisconsin,
July 13, 2005
2

• Background
• This effort is part of much larger effort within
  ORA. ORA has funded the AVHRR reprocessing as a
  pilot project for Scientific Data Stewardship
  (SDS) thank you!
• We hope to continue as a funded project under the
  NESDIS SDS initiative.
• ORA now has the entire GAC archive (1979-present
  32 TB) from CLASS.
• For the first time, we have demonstrated the
  ability to reprocess this data within ORA. So
  far we have generated time-series of GVI-x and
  PATMOS-x. Polar Winds and SST projects are also
  underway.
• A large part of this effort is to improve the
  quality of the AVHRR data and this involves both
  reflectance and thermal calibration and
  geolocation.

PATMOS-x
GVI-x
Polar Winds
SST
3
Why use the The AVHRR for Climate Studies
The Advanced Very High Resolution Radiometer
(AVHRR) was launched in the 1979 for
non-quantitative cloud imagery and SST. It flies
on the NOAA Polar Orbiting Satellites (POES)
2. AVHRR provides enough spatial resolution (1
or 4 km) to resolve many atmospheric and surface
features
1. AVHRR Provides enough spectral information for
several applications
3. Combined with its long data record (1981-2012)
make the 5-channel AVHRR data-set the best we
have for decadal studies for many key climate
parameters.
4
Why Improve the AVHRR Reflectance Calibration?

• Without onboard calibration, the prelaunch
  reflectance calibrations can be many in error.
  Accurate reflectances are critical for cloud,
  aerosol and vegetation climate records.
• Analysis of existing post-launch calibrations
  (esp those use at NESDIS) has shown room for
  improvement for climate use
• The calibration of many of the early and the
  morning orbiting AVHRRs has received little
  attention
• New sensors with onboard calibration and new
  processing techniques (SNO) warrant a new look at
  this issue.
• There is still disagreement in the AVHRR
  reflectance calibrations from different
  techniques (see right)

D.R. Doelling, 2001 Proceedings AMS 11th
Conference on Satellite Meteorology and
Oceanography, Madison, Wisconsin, October 15
18, pp. 614-617
5
Goals of this Work

• Improve upon the existing AVHRR reflectance
  calibrations for ORA climate work
• Derive a new self-consistent set of calibrations
  for whole series of AVHRR data (NOAA-6,7,8,9,10,11
  ,12,14,15,16,17,18,) that allows for meaningful
  climate work.
• Try and build a consensus calibration through
  multiple collaborations and transfer these new
  calibrations to the community.

6

• My Reflectance Calibration Background
• Starting working on AVHRR in 1999.
• Collaborated with Nagaraja Rao and Jerry
  Sullivan on extending N. Raos Libyan Desert
  Technique to NOAA-12 (a morning satellite).
• Heidinger, A. K., J. T. Sullivan and N. Rao,
  2003 Calibration of visible and near-infrared
  channels of the NOAA-12 AVHRR using time-series
  of observations over deserts.I.J.R.S., 24,
  3635-3649.
• After death of N. Rao, Mike Weinreb asked that I
  develop initial post-launch calibration for
  NOAA-16 (a dual gain instrument). Worked with C.
  Cao on the first SNOs and wrote the following
  paper.
• Heidinger, A. K., C. Cao, and J. T. Sullivan,
  Using Moderate Resolution Imaging Spectrometer
  (MODIS) to calibrate advanced very high
  resolution radiometer reflectance channels, J.
  Geophys. Res., 107(D23), 4702, doi10.1029/2001JD0
  02035.
• Fred Wu (formerly of CIMSS) was later hired for
  AVHRR calibration and I am now only interested in
  the calibrations for climate work.

7
Proposed Methodology Simultaneous Nadir
Observations (SNO)

• Our goal is to use Simultaneous Nadir
  Observations (SNO) to improve both the relative
  and absolute calibration. This method provides
  new information and constraints not used in past
  studies,
• Polar orbiting satellites intersect each other at
  high latitudes. This occurs for satellites even
  in very different orbits.
• We have primarily analyzed MODIS and AVHRR SNO
  data though could do VIRS, ATSR and geo imagers.
• We also study AVHRR to AVHRR SNOs to fix the
  relative calibration from one instrument to
  another.

Taken from Changyong Cao http//www.orbit.nesdis
.noaa.gov/smcd/spb/calibration/intercal/
8
Example Imagery from a SNO
Sensor 1 data projected on to Sensor 2 strip.
These points comprise the SNO
5 Nadir Strip from Sensor 1
5 Nadir strip from Sensor 2
9
Example of one Julys SNOs for ch1 and ch2 of
TERRA and NOAA-16
(note y-axis should be ch1 not ch2 on left-hand
plot)
SNOs from MODIS and AVHRR allow us to transfer
MODISs calibration to the AVHRR directly.
10
Example of SNOs for one month of PATMOS-x data
(July 1992)
For July 1992, NOAA-11 and NOAA-12 gave 78
grid-cells that met SNO criteria. Note dark
counts are removed so line should pass through
origin.
This data provides a constraint on the ratio of
NOAA-11 to NOAA-12 calibration slopes. Does not
provide any information on absolute calibration
by itself.
11
ORA (C. Cao and others) has automated SNOs from
AVHRR, AMSU and HIRS and offers a real-time
monitoring capability.
12
A New Libyan Desert Reference for pre-MODIS
Calibration In addition to SNO data (AVHRR/MODIS
and AVHRR/AVHRR), we also employ a New Libyan
Desert Reflectance Reference value to provide an
estimate of the absolute calibration for the
afternoon satellites.

• This new Libyan Reflectance Reference was
  constructed using the following steps
• Selection of Stable Target (Used same as N. Rao)
• Acquisition of MODIS data over Target
• BRDF modeling
• Spectral Adjustment due to Water Vapor (MODIS
  MODTRAN)
• Spectral Adjustment from Hyperion data

This method gives us the absolute calibration for
the pre-MODIS AVHRR data.
MODIS TRUE COLOR
13
Ch1 Calibration Slopes from All Satellites and
All Methods

• Reflectance Calibration Slope x ( Count
  Dark Count)
• Note the relative agreement between the
  calibration slopes derived from different methods.

14
Comparison of Ch 1 Equations for NOAA-7,9,11,14,16
14
9
7
11
16
RCS Rao, Chen and Sullivan VS Vermote and El
Saleous TC Tahnk and Coakley WU Fred Wu
(Operational NOAA)
15
Testing the New Reflectance Calibrations

• Because the ORA effort also involves generation
  of climate products, we can test the impact of
  new calibrations on climate records.
• Aerosol optical depths should test the
  consistency of the low end

Testing Long Term Consistency
Testing Absolute Accuracy
AOT Tables from A. Ignatov
16
Testing the Consistency for bright scenes

• Greenlands Reflectance should be stable and
  tests the high end stability
• Need to compare to absolute standards published
  by Tahnk and Coakley.

17
Conclusions

• ORA is undertaking activity to improve the AVHRR
  1b data. An efforts are underway to communicate
  these improvements to the wider community through
  metadata / ancillary files.
• We are seeking a consensus calibration and are
  seeking collaborators currently working with El
  Saleous and Vermote from NASA.
• Using the SNO technique, we can transfer MODISs
  calibration directly to the AVHRR
• SNOs also provide a direct method to ensure
  reflectance continuity across AVHRR transitions
  without assumptions about vicarious targets.
• SNOs coupled with a new MODIS-based vicarious
  desert calibration target appear to have produced
  an AVHRR reflectance calibration that is
  consistent for all AVHRRs (and is consistent
  with MODIS).