Sensitivity of Automated Monte Carlo Unmixing to Surface Reflectance Uncertainties Caused by Aerosols, Water Vapor, and Terrain Slope-Aspect
David
E.
Knapp, Carnegie Institution, Dept. of Global Ecology, dknapp@globalecology.stanford.edu
(Presenting)
Gregory
Paul
Asner, Carnegie Institution, Dept. of Global Ecology, gpa@stanford.edu
Amanda
Naslund
Cooper, Carnegie Institution, Dept. of Global Ecology, acoop@stanford.edu
Mercedes
Maria Cunha
Bustamante, Universidade de Brasilia, Depto. de Ecologia, mercedes@unb.br
Michael
M.
Keller, University of New Hampshire, Complex Systems Research Center and USDA Forest Service, International Institute of Tropical Forestry, michael.keller@unh.edu
José
Natalino Macedo
Silva, EMBRAPA- Amazonia Oriental, natalino@cpatu.embrapa.br
Darrel
L.
Williams, NASA Goddard Space Flight Center, Laboratory for Terrestrial Physics, Darrel.L.Williams@nasa.gov
The effects of aerosol, water vapor, and Earth-Sun orientation are problematic in calculating atmospherically-corrected reflectance for biophysical parameter estimates from satellite imagery. With advances in satellite-based sensor technology, the ability to estimate aerosol optical thickness (AOT), water vapor, and terrain orientation has steadily improved. However, as techniques to extract biophysical parameters become dependent on more accurate measurements of surface reflectance from space, the need to determine the sensitivity of algorithms to uncertainty and variation in AOT, water vapor, and terrain slope-aspect increases in importance.
The effects of aerosol, water vapor, and terrain slope and aspect were evaluated to determine how uncertainties in these parameters affect results from the Automated Monte Carlo Unmixing algorithm used to estimate forest gap fraction throughout Amazonia. In testing the sensitivity to atmospheric perturbations, a radiative transfer model was used to simulate various levels of aerosol and water vapor in a Landsat ETM+ image. The analysis demonstrated that the uncertainties in the ground-based endmember reflectances were far greater than surface reflectance uncertainties caused by aerosols, water vapor, and terrain slope-aspect.
Submetido por David E. Knapp em 23-MAR-2004
Tema Científico do LBA: LC (Mudanças dos Usos da Terra e da Vegetação)