|Rango, Albert - Al|
Submitted to: International Geoscience and Remote Sensing Symposium Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 6/24/2002
Publication Date: 6/24/2002
Citation: CHOPPING, M.J., RANGO, A., GOMEZ-LANDESA, E. THE IMPORTANCE OF EARLY MORNING LOCAL OVERPASS TIMES FOR BRDF RETRIEVAL, MODELING OF SPECTRAL REFLECTANCE AND FAPAR ESTIMATION. PROCEEDINGS OF THE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM. 2002. V. VI. P. 2264-2266. Interpretive Summary: Interpretive summary not required.
Technical Abstract: A High Resolution Picture Transmission (HRPT)/Local Area Coverage (LAC) data set collected and processed for the NASA Earth Observing System Prototype Validation Exercise (PROVE) campaign for the period May 10-June 3, 1997, was used to retrieve estimates of the bidirectional reflectance distribution function (BRDF) over central-southern New Mexico and parts of Chihuahua, Mexico, by inverting a linear, semi-empirical, kernel-driven model (LISK) composed of isotropic, geometric-optical and volume scattering kernels (unity, LiSparse-MODIS and Ross-Thin, respectively). The surface reflectance estimates used to adjust the model were derived from 17 orbits over a 25-day period at the end of the dry season. The inverse matrix which provides the solution to the inversion problem was used to obtain noise inflation factors (or "weights of determination") for reflectance modeled at viewing angles in the range +/-60 deg for solar zenith angles of f0 deg, 30 deg, and 60 deg in the solar principal plane. A reflectance image was modeled with a view zenith angle of 45 deg and an SZA of 60 deg in the principal plane, suggested in the literature as optimal for retrieval of the fraction of fAPAR. Noise was found to be important (noise amplification factors exceeded unity by a large margin) where observations from early morning overpasses were missing as a result of cloud and cloud-shadow contamination. In these cases, BRDF model parameters and derived reflectance values are unlikely to be reliable, necessitating the injection of a priori knowledge into the model inversion procedure.