Differences in Computed Individual-Tree Volumes Caused by Differences in Field Measurements
Abstract Individual-tree volumes are primarily predicted using volume equations that rely on measured tree attributes. In the northeastern United States, the Forest Inventory and Analysis program determines tree volume using dbh, bole height, proportion of cull, and species information. These measurements are subject to variability due to a host of factors. The sensitivity of the volume equations were assessed in relation to changes in each of the input variables. Additionally, data from 3,345 trees that were independently remeasured were used to assess differences in gross and net volumes between operational and audit measurements. Evaluations were conducted for dbh, bole height, and proportion of cull classes, across 18 different species groups. Differences in bole height and proportion cull measurements were found to contribute the most to volume differences. Surprisingly, trees with relatively short bole heights were affected more than trees having taller bole heights. Differences in dbh and species identification contributed little to the volume differences. An analysis of the full data set across all realized volume differences showed no statistical bias in either gross or net volume. These results show the influence that specific field measurements have on accurate estimation of volume, which may be useful for targeting specific attributes where additional training or refined measurement protocols could improve consistency.