Validation of Satellite-Derived Vegetation Indices for Estimating Forest Structural Parameters Using Mobile Terrestrial LiDAR in Aleppo Pine Forest Stands 

Abstract

Remote sensing technologies have been crucial in the monitoring and the assessment of forest carbon sequestration, emphasising the need for implementing Adaptive Forest management (AFM) as a vital strategy to decrease vulnerability to climate change impacts. Within this context, obtaining quantitative information on forest structure becomes necessary because AFM depends on the pre-existing forest structure and involves its subsequent modification. Consequently, this modification of forest structure has an impact on forest carbon sequestration. The main objective of this study is to validate the use of different satellite-based indices and algorithms as reliable quantitative estimators of forest structural parameters associated with its potential photosynthetic activity. Our study relies on the application of mobile terrestrial LiDAR for characterising vegetation structure at both individual tree and plot levels in Aleppo pine (Pinus halepensis L.) forests. As documented in the literature, these satellite-based indices were not consistent predictors of photosynthetic performance in evergreen species for most of the year. This was attributed to seasonal reductions in photosynthetic radiation-use efficiency that occurred without substantial declines in canopy greenness. Despite this finding, we hypothesize that the spatial information provided by these remote-based indices remains valid for capturing forest structural parameters relevant for carbon sequestration studies. To achieve this, we collected LiDAR scans from 21 forest compartments, ranging in size from 8 to 30 hectares, and 24 plots of approximately 0.2 hectares each. Thus, we captured detailed information about diameter at breast height, total tree height, and overall stand structure characteristics per hectare (i.e., number of trees, basal area, total timber volume and crown coverage). By comparing these ground-truth measurements with indices and algorithms derived from satellite imagery (i.e., NDVI, EVI, red edge index, NDWI), we evaluated their efficiency as estimators of forest structural parameters. Here, two spatial scales were considered: the 300-m resolution from Sentinel-3 for the forest compartments, and the 10-m resolution of Sentinel-2 for the plots. Our findings illustrated a good correlation between LiDAR-derived structural metrics and various selected indices. Once the robustness of these indices is confirmed, their application for downscaling satellite images related to gross primary productivity (GPP) or net ecosystem productivity (NEP) can be justified. This validation process will enhance our confidence in the use of remote sensing data to extract quantitative information about forest structure and will support its application for AFM purposes.