Evaluation, improvement, and application of surface BRDF and albedo retrievals from MODIS observations

Accurate knowledge of surface albedo at appropriate spatial and temporal scales is essential for the estimation of the radiative forcing arising from land cover change and for the prediction of future climate under various scenarios. This dissertation focuses on quality assessment and improvement of the global albedo product derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument aboard NASA's Terra spacecraft. The quality of the albedo retrievals was evaluated in two ways: (1) by examining the global distribution of quality assurance (QA) fields calculated by the product's algorithm; and (2) by comparing retrieved albedos with those observed at ground stations and by other satellite instruments. The QA statistics demonstrate satisfactory global performance of the retrieval algorithm. While cloud is the main reason for resorting to backup magnitude inversion, shortwave albedos retrieved from the backup and main algorithms agree on average within 0.033 reflectance unit. Compared to field measurements, the satellite product generally meets an absolute accuracy requirement of 0.02 and realistically accounts for solar zenith angle dependence. MODIS-derived surface albedos are also consistent with top-of-the-atmosphere albedos derived from Clouds and Earth's Radiant Energy System (CERES) observations and with the historical global albedos derived from other instruments. When the number of MODIS clear-sky observations is small or the MODIS observations are close to the cross principal plane, observations from the Multiangle Imaging Spectroradiometer (MISR,) are shown to improve MODIS retrievals. A synergistic algorithm using MODIS to modify a priori information retrieved from MISR observations, reduces the relative reflectance prediction error by 10 percent in the red and green wavebands relative to MODIS-only retrievals. Finally; MODIS albedo retrievals were examined for snow-covered land surfaces. Results show that snow changes the spectral property of the surface reflectivity and causes high heterogeneity in surface albedo. Further, the magnitude of the albedo increase due to presence of snow depends on vegetation type. The mean shortwave albedo at local solar noon for snow-covered forests is lower than 0.30 due to the effect of canopy shading, and reaches 0.57 for snow-covered grassland and barren land covers.