Assimilation method to derive spectral ground reflectance of desert sites from satellite datasets

Abstract

This paper is devoted to an original assimilation method used to derive the visible/near-infrared spectral reflectance properties of calibration reference desert areas selected in previous studies for their spatial uniformity and temporal stability. In the first step, an appropriate ground spectral reflectance model with a small number of parameters was defined. To this end, spectral bidirectional reflectance measurements of typical sands were analyzed. Then, an adjustment method using a heterogeneous set of satellite data acquired under similar conditions over a given reference site was applied to recover ground spectral model parameters. The robustness of the method is demonstrated using a simulated satellite level dataset generated with realistic errors and uncertainties. The last section describes application of the method to a real satellite dataset extracted from the SADE database: “Structure d'Accueil de Données d'Etalonnage”, a data repository built by “Centre National d'Etudes Spatiales” (CNES) for collecting and storing data acquired over various ground reference areas including desert reference sites. The recovered ground model was then transferred to the top of the atmosphere (TOA) and compared with an available Global Ozone Monitoring Experiment (GOME) spectral measurement. The comparison shows good agreement on the spectral shape and a total bias of less than ±3%.