William R. Tahnk and James A. Coakley, Jr

Using Sun Glint and Antarctic Ice Sheets to Calibrate MODIS and AVHRR

Observations of Reflected Sunlight

William R. Tahnk and James A. Coakley, Jr

Cooperative Institute for Oceanographic Satellite StudiesCollege of Oceanic and Atmospheric Sciences

Oregon State University

Problem:• Maintaining the calibration of aircraft and satellite-borne radiometersused to measure reflected sunlight

• Verifying the calibration while in flight and subject to the stressesof the flight environment

Objective:

• Use reflected radiances in glint areas to assess relative accuracies of radiances at different wavelengths

• If the accuracy at one wavelength is known, observations in glint areascan be used to check the accuracies of radiances at other wavelengths

Method:

True color image for Terra on Dec 8 2002 over the Indian Ocean at 0700 UTC

• Select sub-regions of reflected radiances to analyze within sun glint areas for various satellites • Selection criteria: Narrow region, free of clouds Region spans a major fraction of the sun glint (edge to center) Good dynamic range in reflectance

• Derive estimates of the slopes and intercepts for the reflectances

Image of sun glint area constructed from 1-km radiances at 3.7 m

a cb

1-km reflectances at 0.64, 1.6, and 2.1 m, and radiances at 3.7 m for the boxed region in the figure above

A Monte Carlo radiative transfer model (MOCARAT) was used to assess the sensitivity of the reflectance relationships at various wavelengths

MOCARAT results:

Values of the slopes and intercepts derived for simulations of the reflectancesfor atmospheres with marine aerosols are rather insensitive to surface wind speed and direction aerosol burden (for < 0.2 and large aerosol particles) solar zenith angle (for < 35º)

The slopes and intercepts of the linear relationships among the reflectancescan then be used to assess, in a relative sense, the calibration of the radiometer

Terra 0.84/0.64-m Aqua 0.84/0.64-m NOAA16 0.84/0.64-m NOAA17 0.84/0.64-m

Terra 1.60/0.64-m Aqua 1.60/0.64-m NOAA17 1.60/0.64-m

Satellite Month/Year Pass Count 0.84/0.64Slope

1.60/0.64Slope

Terra Dec 2001 37 1.098 ± 0.008 1.084 ± 0.013

Terra Dec 2003 47 1.114± 0.008 1.081 ± 0.009

Terra Dec 2004 44 1.105 ± 0.005 1.082 ± 0.009

Aqua Dec 2002 42 1.107 ± 0.011 1.082 ± 0.015

Aqua Dec 2003 61 1.108 ± 0.006 1.074 ± 0.006

Aqua Dec 2004 32 1.110 ± 0.005 1.071 ± 0.007

NOAA-16 Dec 2001 63 1.050 ± 0.037a --

NOAA-16 Dec 2002 50 1.052 ± 0.041a --

NOAA-16 Dec 2003 57 1.163 ± 0.018b --

NOAA-16 Dec 2004 39 1.137 ± 0.033b --

NOAA-17 Dec 2002 77 0.943 ± 0.029a 1.031 ± 0.021a

NOAA-17 Dec 2003 54 0.936 ± 0.026a 1.014 ± 0.012a

NOAA-17 Dec 2004 43 1.146 ± 0.033b 1.203 ± 0.014b

Slopes for the 0.84/0.64 and 1.60/0.64-µm linear relationships for Terra and Aqua MODIS and NOAA-16 and -17 AVHRR passes for the period December 2001-2004

aReflectances obtained using pre-launch calibration coefficients imbedded in the Level 1(B) data streambReflectances obtained using updated calibration coefficients imbedded in the Level 1(B) data stream

Terra 0.84/0.64-m NOAA17 0.84/0.64-mNOAA16 0.84/0.64-mAqua 0.84/0.64-m

Terra 1.60/0.64-m Aqua 1.60/0.64-m NOAA17 1.60/0.64-m

NOAA16 NOAA17

Channel 1 Channel 1Channel 2 Channel 2

Reflectances obtained using the operational calibration coefficients imbedded in the Level 1(B) data stream

Reflectances obtained using calibration coefficients derived with the Antarctic ice sheet method

Note: The solid line represents the NOAA9 calibration reference for Antarctica, from Loeb (1997)

Summary and Results:• Antarctic ice sheets and ocean glint areas were used to check the calibrationof solar reflectance channels on Terra and Aqua MODIS and NOAA16 and NOAA17 AVHRR for the period December 2002-2004

• Terra and Aqua MODIS observations at 0.64, 0.84, and 1.6 m are consistentwith each other and internally consistent through the period analyzed

• With the inclusion of updated calibrated coefficients in the Level 1(B) data stream for NOAA16 and NOAA17, the AVHRR solar reflectance channels are more consistent with MODIS

• The 0.64-m reflectances for both NOAA16 and NOAA17 fall short of theAntarctic ice sheet calibration by about 4%

• The 0.84-m reflectances are accurate within 1% for both NOAA satellites when compared with the Antarctic ice sheet calibration

• The 1.6/0.64 slopes derived for ocean glint regions indicate that the 1.6-m reflectances for NOAA17 (in high gain mode) and calibrated using the updatedLevel 1(B) calibration coefficients are too high by about 6%

References:

Loeb, N.G., 1997: In-flight calibration of NOAA AVHRR visible and near-IR bands over Greenland and Antarctica. Int. J. Remote Sens., 18, 477-490.

Tahnk, W. R. and J. A. Coakley, Jr., 2001: Improved calibration coefficients for NOAA-14 AVHRR visible and near-IR channels. International Journal of Remote Sensing, 22, 1269-1283.

Tahnk, W. R. and J. A. Coakley, Jr., 2001: Updated calibration coefficients for NOAA-14 AVHRR channels 1 and 2. International Journal of Remote Sensing, 22, 3053-3057.

Tahnk, W.R. and J.A. Coakley, Jr., 2002: Improved calibration coefficients for NOAA-12 and NOAA-15 AVHRR visible and near-IR channels. J. Atmos. Ocean. Tech. 19, 1826-1833.

Luderer, G., J.A. Coakley, Jr., and W.R. Tahnk, 2005: Using sun glint to check the relative calibration of reflected spectral radiances. J. Atmos. Ocean. Technol. (in press).

Tahnk, W.R. and J.A. Coakley, Jr., 2005: Calibration of visible and near-IR channels of MODIS and AVHRR through 2004 using Antarctic ice sheets and ocean glint regions. (in preparation).