|Vande Castle, John - UNIV. OF NEW MEXICO|
Submitted to: Remote Sensing in Hydrology Symposium
Publication Type: Proceedings
Publication Acceptance Date: June 5, 2000
Publication Date: September 25, 2001
Interpretive Summary: This study assesses the effects of physical structure and composition of shrub-coppice dune landscapes on anisotropy in the NASA MODIS-ASTER Airborne Simulator (MASTER) solar channels, combined with an investigation of the viability of simulating multiangular datasets using off-nadir airborne imaging radiometry from a single overpass by means of data segmentation. Segmentation data are structural (plant density and cover) and spectral measures derived from high resolution aerial photography and classified Landsat 7 Enhanced Thematic Mapper imagery. The directional signal was found to contribute a variation of at least 5 - 10% reflectance, necessitating angular corrections. For most areas the signal appears to differ only slightly with changes in landscape structure and composition because the viewing/illumination geometry is poor and because reflectance is dominated by bright soils.
Technical Abstract: A MASTER scene was acquired over a shrub-coppice mesquite dune test area in the USDA-ARS Jornada Experimental Range, New Mexico, on 9/27/99. Calibration was performed at the Airborne Sensor Facility at NASA Ames Research Center. The ground-projected field-of-view is approximately 3.5m at nadir. The atmosphere was clear and dry. Aerosol optical thickness was estimated at only 0.02-0.04 at 550nm using sun photometer measurements. Atmospheric profiles of temperature, water vapor and pressure were obtained from rawinsonde data from the White Sands C-station radiosonde launch at 11:00 GMT. Normalized spectral response functions were interpolated and a series of coefficients were calculated for surface reflectance retrieval using 6S v.4.2. Larger field-of-view data were simulated by convolution of the image with the point spread function of the Landsat MSS. Scanned aerial photography was used to calculate plant density and area statistics for small and large plant types and these data were used to segment the MASTER scene. Landsat 7 ETM+ data from 9/28/99 were also used in segmentation. The difference between backscattering and forward-scattering is 5 - 10% reflectance. The lack of variation in the angular signatures with cover density and type is likely to be due to the sun-sensor geometry and very bright soils, accounting for up to 75% of areal cover. The accumulation of multi-angular signatures using data from a MASTER overflight requires a geometry closer to the principal plane, or use of other multidirectional data to adequately characterize the surface Bidirectional Reflectance Distribution Function.