Study of the anisotropic reflectance behaviour of wheat canopy to evaluate the performance of radiative transfer model PROSAIL5B

Abstract

A field experiment was conducted on wheat to analyze its bi-directional reflectance in relation to sun-target-sensor geometry. To achieve a large variation in crop parameters, two extreme nitrogen treatments were applied. The study reconfirms the strong and consistent anisotropic patterns of wheat bi-directional reflectance in visible (VIS) and near infra-red (NIR) and its magnitude was highest in the principal plane. This anisotropic pattern extended equally in shortwave infra-red (SWIR). The hotspot broadened with crop growth due to increase in leaf area index (LAI), leaf size and planophilic orientation. The shape and magnitude of PROSAIL5B simulated spectra was in close agreement with the observed spectra in the optical region for most of the view zenith and azimuth angle combinations. In the NIR and SWIR, the magnitude of the model simulations showed good match in the principal plane, whereas underestimation was found in the backward scattering direction at higher view zenith angles in the VIS. The typical bowl shape of observed reflectance in principal plane was very well simulated in NIR by the model but failed in other wavebands. The model performed best in the NIR region followed by SWIR and maximum relative error was in VIS. Over the whole optical region and view zenith angles, the model simulations showed an average error of 26%. The model simulations were poor at low LAI indicating the need to improve soil reflectance algorithm in the model. Results of the study have implications for understanding the strengths/shortcomings in the model and its inversion to derive crop biophysical parameters from multispectral sensors.