You must have a javascript-enabled browser and javacript and stylesheets must be enabled to use some of the functions on this site.
 
        

 

Spectral index and band setting for aerosol remote sensing over land

Richard Santer(1) and Jérome Vidot(1)

(1) Université du Littoral, 32 avenue Foch, 62 930, France

Abstract

Aerosol remote sensing over land is based on the use of pixels covered by vegetation. Because of the absorption of the photosynthesis pigments, these pixels (DDV, Dense Dark Vegetation) are quite dark in the blue and in the red. The DDV pixels are selected by a spectral index, the ARVI (Atmospheric Resistant Vegetation Index). ARVI combined spectral bands at 440 nm, 670 nm and 865 nm after a Rayleigh correction. In parallel to this measured ARVI, the MERIS surface reflectance model refers to a theoretical ARVI (TARVI). TARVI is computed both with standard values of the DDV reflectances and outputs of a radiative transfer code which account for the aerosol scattering with a standard aerosol model (continental model and 23 km for the horizontal visibility). The resistance of the TARVI to different aerosol models is first studied and, second, the impacts of the uncertainties of the TARVI on the aerosol product are evaluated. Alternatively to the ARVI, we can use the good MERIS spectral coverage of the red edge to define another spectral index, the MTCI is one of them. By using a small spectral range in the red-NIR, we expect to reduce the impact of the atmospheric effect. Aerosols are remote sensed at 440 nm and 670 nm. If the DDV reflectance appears to be quite well defined at 440 nm, it is not the case at 670 nm. The extraction of the aerosol signature at 670 nm combines this disadvantage with the reduction of the aerosol scattering with wavelength. Alternatively, 412 nm and 490 nm can be used and we explore this alternative. To do so, we first have to generate new look up tables and make algorithm modifications. Comparison between the initial algorithm and this new one is done on MERIS images for which we have a data set of ground based measurements.