Estimation of clear sky albedo over a tallgrass prairie using the advanced very high resolution radiometer.

A study is conducted to explore the use of the Advanced Very High Resolution Radiometer (AVHRR) for estimating albedo (directional-hemispherical reflectance) over the Konza Prairie in Kansas. Two methods are demonstrated: (1) an inversion method that estimates the bidirectional reflectance distribution function (BRDF) of the grassland which can then be integrated over the viewing hemisphere to estimate albedo, and (2) angular sampling of the BRDF to minimize the effects of anisotropic surface reflectance for direct estimation of albedo. Models of the surface-sensor-atmosphere system are constructed to simulate the process of AVHRR data acquisition, and estimation strategies are developed based on these simulations. Using ground measurements for comparison, the study indicates that environmental noises (clouds, atmospheric effects, etc.), coupled with few available measurements within a time window small enough to avoid phenological change at the surface, severely hamper the use of direct inversion methods from the satellite platform. Angular constraint methods produce realistic seasonal profiles, but cannot discern albedo increases due to changes in surface conditions, from changes due to variable solar illumination angles. An algorithm is developed to mitigate these problems, by mapping measured directional reflectance to albedo using pre-computed anisotropy factors, and using a "minimum channel 1 albedo" scheme to eliminate residual atmospheric effects. These estimations are shown to yield accuracies of +/$-$0.01 range for channel 1 albedo, and +/$-$0.03 for channel 2, without considering the effect of cloud contamination on the estimates.PhDAtmospheric and Space SciencesUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/104176/1/9500997.pdfDescription of 9500997.pdf : Restricted to UM users only