×

You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Title
Mesures physiques et signatures en télédétection

57
COMPARISON OF DIFFERENT ATMOSPHERIC CORRECTION
ALGORITHMS WITH THE ATSR/ERS-1 INFRARED RADIOMETER
C. FRANÇOIS and C. OTTLE
C. E. T. P.
CENTRE UNIVERSITAIRE
10,12 Avenue de l'Europe
78140 Vélizy, FRANCE
ABSTRACT
The remote sensing estimate of the surface temperature from the thermal infrared radiance emitted by the
terrestrial surface requires atmopheric corrections. The ATSR/IR infrared radiometer (on board ERS-1) is
a multi-channel (11 and 12 pm) and double-viewing radiometer (0° and 55°). Then, three different
atmospheric correction methods may be used to retrieve the surface temperature : the One-Channel
method, the Split-Window technique and the Double-Viewing method. The last two methods (which does
not need ancillary information) have been compared with the One-Channel method and with ground truth
data. Applications on ATSR data show for both Split-Window and Double-Vie wing methods a good
agreement with a mean error lower than 0.5 K in the case of SST retrieval.
KEY WORDS : Atmospheric corrections, ATSR, Split-Window, Double-viewing method, SST, LST
1 - INTRODUCTION
The ATSR/IR observes the Earth's surface along two curved swaths (about 512 km in width), one being
the Nadir swath (a nearly vertical path through the atmosphere : the nadir viewing angle varies from 0° to
nearly 22° at the edge of the swath), and the other the Forward swath (an inclined path forward of the
subsatellite point: the forward viewing angle varies between 52.4 ° and 55°). The spatial resolution is
about 1 km x 1 km at Nadir and about 1.5 km x 2 km for the forward pixels (due to the viewing
perspective), each view being re-sampled to a 512 x 512 pixels image. The radiometer thermal infrared
spectral domain is situated in the atmospheric window, where the atmospheric absorption (mostly due to
the water vapor absorption) is the weakest: 11 pm and 12 pm.
The surface temperature may be derived from the satellite brightness temperatures if the
atmospheric effects (absorption, diffusion, emission) are corrected. In order to develop atmospheric
corrections which does not require ancillary information (radio-sonde data for instance) two different
methods have been studied : the Split-Window technique and the Double-Viewing method, which uses
the dual-angle capability of the ATSR.
After the presentation of these two methods, different applications with ATSR data will be
displayed.
2 - THE SPLIT-WINDOW TECHNIQUE
2.1. Methodology
The Split-Window technique is based on the differential absorption between the 11 and 12 pm channels.
From the radiative transfer equations written for both channels, the surface temperature may be derive as
a linear combination of the 11 pm and 12 pm satellite brightness temperatures :
T s = anTbli + ai2Tbi2 + c (1)
The an, ai2 and c coefficients are chosen to minimize the errors in the surface temperature
determination.
There are different ways to calculate the Split-Window coefficients : theorically from the
radiative transfer equation, linearizing the Planck's equation (Sobrino et al. 1991, Becker, and Li 1990,
Prata, 1993), or statistically using a wide range of atmospheric situations (Ottl6 and Vidal-Madjar, 1992,