Influences of canopy architecture on relationships between various vegetation indices and LAI and Fpar: A computer simulation

Abstract Vegetation Indices (VIs) are often used to estimate important biophysical parameters, like LAI and Fpar, from vegetation canopy reflectance data. In this study, a three‐dimensional model (Diana) is utilized to generate architecturally realistic tree and crop canopies in various development stages and to calculate bidirectional reflectance factors in the principal plane. We investigate the influence of various factors like soil brightness, optical properties of canopy elements, leaf angle distribution, spacing distance between plants and solar and view geometries on relationships between VI and Fpar, LAI and percentage ground cover (GC) in order to determine optimal VI and viewing conditions for the estimation of Fpar and LAI/GC. These simulation studies suggest that: (1) in most cases, Vis using off‐nadir reflectances are more informative and useful than those based on nadir reflectances; (2) the optimal VI and sun/view geometries are usually different for inferring different parameters, depending on canopy architecture; (3) LAI can be practically estimated by VI only for homogeneous canopies while GC and Fpar can be inferred even for an inhomogeneous canopy; and (4) when optical properties of vegetation elements vary within a canopy, neither LAI/GC nor Fpar can be estimated by means of VI method with an acceptable accuracy.