Estimation of forestry parameters in mountainous coppice stands using airborne laser scanning
Résumé
An area-based method is implemented to predict forest stand parameters from airborne laser scanning data. Multiple regression models are calibrated with 31 field plots inventoried in a hillside dominated by coppice stands, located in the French Alps. Cross-validated prediction accuracies are respectively 13.2, 14.3, 19.4 and 25.5% for dominant height, mean diameter, basal area and stem density. Reducing calibration plot radius greatly influences prediction results. Although median values of forest parameters remain unchanged, field observations variability is higher for smaller plots. Depending on the forest parameter, prediction accuracy is significantly correlated with some distribution statistics (minimum and maximum values) of field observations computed with various radii. Indeed, greater variance or outliers may degrade the degree of fit of regression models.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...