Estimating stand volume in pure or mixed beech and spruce stands using sentinel-2 satellite images and LiDAR data

The most costly and labor-intensive part of the plan is the inventory. In this study, stand volume will be estimated with fewer land measurements by using regression equations developed with the help of terrestrial data and remote sensing methods. Thus, it will be possible to estimate the stand volume in labor-intensive lands which are not accessible or require hardwork under difficult conditions. Regression Analysis was performed between the terrestrial data in 44 sample areas and the band values obtained from Sentinel-2 images, vegetation index values and height values of the LiDAR point cloud and the calculated volume values. Considering the volume values calculated separately as Single and Double Input after the analysis; in the models where single input volume values are estimated, the model including only band values (B6-B7) gives an R2 adj.=0,663 value, while in the models where double input volume values are estimated (B6-B7), R2 adj.=0,607. In vegetation indices, the highest value for single entry volume (SAVI-RVI) was obtained as R2 adj.=0,611 and for double entry volume (SAVI-RVI) R2 adj.=0,561. The values obtained by processing the LiDAR data were R2 adj.=0,482 for single-entry volume, R2 adj.=0,534 for double-entry volume, R2 adj.=0,540 and R2 adj.=0,568 for stand top height. In the analysis including all values, the single-entry volume R2 adj.=0,719 and the double-entry volume value R2 adj.=0,652. It is understood that predictions can be made with high confidence level with the highest R 2 results. Key Words: Stand Volume, Remote Sensing, Sentinel 2, LiDAR, Regression Models