Increasing Precision for French Forest Inventory Estimates using the k-NN Technique with Optical and Photogrammetric Data and Model-Assisted Estimators

摘要: Multisource forest inventory methods were developed to improve the precision of national estimates. These rely on combination data and auxiliary information correlated with attributes interest. As these have been predominantly tested over coniferous forests, present study used this approach for heterogeneous complex deciduous forests in center France. The considered included a type map, Landsat 8 spectral bands derived vegetation indexes, 3D variables from photogrammetric canopy height models. On subset area, changes estimated two successive models also used. A model-assisted inference framework, using k nearest-neighbors approach, was predict 11 field simultaneously. results showed that among tested, metrics improved dendrometric estimates more than other variables. Relative eciencies (RE) varying 2.15 volume 1.04 stand density obtained all Canopy increased RE 3% 26%. Our confirmed importance as demonstrated value change increasing structural such quadratic mean diameter.