Assessing impacts of off-nadir observation on remote sensing of vegetation: use of the Suits model

Abstract

Off-nadir remote sensing of vegetation can cause undesirable variability in measured spectral reflectance resulting from non-Lambertian characteristics of the canopy. The Suits model of radiative transfer in a vegetation canopy was evaluated as a means to simulate this variability. Comparison was made between model calculations and reflectance of a salt marsh cord grass canopy measured under a variety of solar- and viewing-angle conditions using an in situ radiometer. The model was effective in simulating both the sense and magnitude of reflectance changes due to variable angles of observation. However, the model did not reproduce the observed dependence of nadir canopy reflectance on solar zenith angle. A simple subtractive normalization procedure resulted in high correlation of modelled red and infrared reflectance with values measured at observation angles varying from 10 to 60° off-nadir and with solar zenith angles ranging from 18 to 55°. The modelling procedure was extended to simulate view-angle effects on aircraft scanner imagery of a coastal marsh with good results despite significant variability in biomass and leaf-area index of cord grass within the imaged area. Modelling appears to have potential in predicting view-angle effects and in reducing angular variability in remotely sensed data derived from aerial and orbital sensors.