Remote sensing in orchards: influence of canopy architecture and observation conditions on uncertainties associated to key canopy characteristics estimates based on 3D models
Résumé
The effect of canopy architecture in the estimation of biophysical parameters from remote sensing data in orchards is studied through the particular case of vineyard canopies. Two different approximations (1D and 3D models) are tested to estimate LAI from canopy reflectance in a total of 31 plots within commercial field plots. The results shows that 1D models are not pertinent to describe canopy reflectance in row structured canopies, while including a 3D description of canopy architecture provided satisfactory results (RMSE=0.19). A simulation study was also carried out to evaluate the contribution of observational parameters (sun position, rows orientation…) in the accuracy of LAI and fIPAR estimations from canopy reflectance. The results highlight the effect of rows orientation in the uncertainties in the estimation of both parameters. When sun illuminates perpendicular to rows, the sensitivity of canopy reflectance to fIPAR and LAI is higher, thus providing more accurate estimations than in the case of parallel illumination. The shadow projected in the soil when sun is perpendicular to rows seems to enhance the sensitivity of canopy reflectance to changes in leaf area and in light interception.