The impact of plot numbers and plot configuration on the accuracy of pre-harvest stand estimates obtained by terrestrial laser scanning.

Abstract

Estimation of merchantable timber volume in a stand, based on pre-harvest inventories, requires accurate measurements of parameters such as diameter, height, taper and stocking, and appropriate sampling. Data were collected using terrestrial laser scanning with a multi-scan mode, in a stand of Sitka spruce (Picea sitchensis (Bong.) Carr.) scheduled for clearfelling. The stand was divided into four blocks. Two to five plots per block were established in the stand and were scanned. For a number of sample trees in each plot, detailed manual measurements of diameter were taken at half-metre intervals along the stem, for comparison with parameters derived from multi-scanning. Pre-sale measurements were also carried out in each block, based on Coillte’s (The Irish Forestry Board) Standard Operational Procedures (SOPs), a standard methodology for pre-sale inventories in Ireland, based on the tariff system. The mean diameter at breast height (DBH) values derived from the point cloud data from the three scan positions in each plot were in good agreement with the manually measured DBH values (i.e. root mean square error (RMSE) of 1.72 cm and a bias of 0.3 cm). The volumes derived after adjusting the height values obtained from multi-scan point cloud data with plot-based DBH-height regression models showed the closest match with those produced by the segment method of calculating individual tree volume. The estimated stand volume using data from the three scan positions after correcting for occlusion, and the stand volume estimated based on the use of multi-scan data from the intersecting areas of the scan circles, resulted in volumes within 6.9% and 8.5%, respectively, of the volume of the stand as measured using Coillte’s SOPs.