GEOSCAN Search Results: Fastlink


TitleAn Evaluation of Kernel-Driven Bidirectional Models Using PARABOLA Measurements
AuthorLi, Z; Moreau, L; Cihlar, J; Deering, D
SourceCanadian Journal of Remote Sensing 23, 2, 1997 p. 120-130,
Alt SeriesEarth Sciences Sector, Contribution Series 20042138
PublisherInforma UK Limited
Mediapaper; on-line; digital
File formatpdf
Released2014 07 31
AbstractThe kernel-driven biderectional models (KDBM) are employed increasingly to account for the bidirectional dependence of reflectance. This study presents an evaluation of various KDBMs using surface bidiectional measurements acquired by PARABOLA. Numerous tests were conducted by fitting the KDBMs to reflectance data available in a limited and non-limited viewing domain for fixed and variable solar zenith (SZ) angles at the surface and top of the atmosphere (TOA). The performance was evaluated by comparing the fitness of observed bidirectional dependence and albedos from observations and KDBMs. It varies considerably among the KDBMs when samples are confined to such a limited domain as the principle plane wherein the NOAA/AVHRR data used in this study cover the northern regions. When the angular samples are distributed throughout the entire viewing domain, the majority of KDBMs lead to good estimates of albedo with relative uncertainties less than 10%, but their fits to the bidirectional dependence may differ drastically. Moderate errors ensued when a KDBM tuned to a specific SZ is applied to other SZ angles. Most KDBMs perform even better when applied to TOA measurements. The best performance, for the sites we investigated, is achieved by the combination of Roujean et al.'s surface scattering kernel and ross' volume scattering kernel with a relative uncertainty in albedo estimate merely 2%.

Date modified: