Wood volume estimation strategies for trees from a Dry Forest/Savannah transition area in Piauí, Brazil

Abstract

Prediction of aboveground wood volume is one of most important stages when conducting a forest inventory and
making forest management decisions. This is more difficult for native forest than for plantations given the high
variability of the former (trees of different species and age groups). The objective of this work was to evaluate different strategies to estimate aboveground standing tree wood volume (v) using its diameter at breast height (DBH) and total height (h), comparing the use of form factors, volume equations fitted by ordinary least squares and mixed modelling, as well as the options from the literature. To achieve this, 351 trees were scaled during field campaigns, with 158 felled trees and 193 measured using Criterion equipment. The data collected from each scaled tree was: h, DBH, v and identification of the botanical species. We found that the best application of form factors occurred when the trees were divided by diameter classes. However, regression models, regardless of the fitting technique, presented better volume estimates than form factors. Mixed models, with either the species or diameter class as the random variable, provided the lowest errors when estimating tree volume. Thus, we recommend the use of mixed models as the best strategy to estimate volume of standing trees. The following equation can be used to estimate aboveground wood volume for trees from vegetation types similar to the ones of this study:  ln(v) = −9.06013 + 1.91756 ln(DBH) + 0.69846 ln(h).