Directional reflectance of vegetation targets: Simulation of its space measurements by coupling atmospheric and biophysical radiative transfer models

Vegetative surfaces are non-lambertian and deriving their spectral properties from space-borne sensors becomes complicated when off-nadir view angles are taken into consideration. In order to utilize off-nadir observations a complete understanding of directional reflectance is needed. This requires understanding of propagation of solar radiation as radiative transfer (RT) problem through a coupled system of vegetation canopy and atmosphere as a function of view angles. In this paper, a new approach of coupling biophysical (PROSAIL) and atmospheric (6S-code) RT models has been proposed to simulate at-sensor directional reflectance for vegetation target. At-sensor reflectance was simulated using 6S-code for the viewing geometry pertaining to 0, 5 and 26 deg view angles (hypothetical multispectral sensor with forward and backward viewing capabilities) for varying atmospheric conditions with the lower boundary condition parameterized through vegetation canopy reflectance obtained from PROSAIL model. Sensitivity of directional reflectance to input parameters were inferred at top-of-canopy and top-of-atmosphere level. The degree of anisotropy in reflectance pattern due to directional viewing represented by a parameter called g-factor was quantified (3-24% in red, 2-9% in near-infrared and panchromatic channels at top-of-atmosphere) which depends on wavelength and increases with atmospheric turbidity. This modeling study would help understanding capabilities of future space-borne sensors with directional/multi-spectral imaging.