IDENTIFICATION OF RETENTION AREAS USING AIRBORNE LIDAR DATA. A CASE STUDY FROM CENTRAL SWEDEN

This paper presents a method for identifying retention areas in forest stands using publicly available ALS (Aerial Laser Scanning) data. Retention areas/trees are the cause of large inaccuracies in compartmental timber volume calculations when updated with remote sensing data. Tree height was selected as the most explanatory parameter for identification. The calculation of the threshold value for each compartment was based on data from the FMS (Forest Management System) or on the evaluation of the statistical distribution of LiDAR data in the compartment. The calculation was applied directly to the 3D point cloud, where points with the corresponding height were classified and processed into the resulting vector layer. Both methods were tested and validated on a reference dataset. The statistical approach proved to be more reliable (OA 89%) due to frequent errors or outdated data in the FMS (OA 82%). After removing dead retention trees (standing tree torsos) from the validation dataset, the OA of both methods increased (FMS approach 90%, statistical approach 94%).