Biomass estimates from the application of different allometric equations and their relationship with forest structure

Abstract

Aboveground forest biomass is generally estimated from allometric equations applied to forest plot data. However, quantifying the bias (i.e., the percentage of over- or underestimation) is challenging when no local, direct biomass measurements (i.e., reference values) are available. In such cases, we propose comparing biomass estimates derived from different allometric equations and analyzing their relationships with forest structure pa�erns along the studied environmental gradient. In subtropical montane forests (i.e., Yungas), the largest differences in biomass estimates were found between equations that either include or exclude tree height (i.e., on average, 94% to 113% more biomass estimated per hectare with equations that do not include height). In contrast, differences due to the method used to estimate tree variables not measured in the field are relatively small (4% when using different databases for wood density by species and <1% when estimating tree height with different models). In the Yungas, the biomass distribution along the altitudinal gradient results from a combination of lower basal area with high wood density species in the foothill forest (~500 m a. s. l.) and higher basal area with low wood density species in the montane forest (~2000 m a. s. l.). Including tree height allows for modeling biomass vertically, showing a slight decrease with altitude, consistent with the observed canopy height reduction above 1500 m a. s. l. In these mountain forests, equations that include tree height used in this study estimate biomass values of ~160 Mg/ha in the foothill forest and ~130 Mg/ha in the montane forest.