Title: Tree height integrated into pantropical forest biomass estimates
Authors: FELDPAUSCH TedLLOYD JBRIENEN RLEWIS SGLOOR EMONTEAGUDO ALOPEZ-GONZALEZ GBANIN LABU SALIM KALEXIADES MALMEIDA SAMARAL IANDRADE AnaARAGAO L. E. O. C.MURAKAMI AARETS E. J. M. M.ARROYO LAYMARD GBAKER TBANKI OBERRY N. J.CARDOZO NSTROPP CARNEIRO JULIANAPHILLIPS O.l.
Citation: BIOGEOSCIENCES vol. 9 no. 8 p. 3381-3403
Publisher: COPERNICUS GESELLSCHAFT MBH
Publication Year: 2012
JRC N°: JRC69737
ISSN: 1726-4170
URI: http://www.biogeosciences.net/9/3381/2012/bg-9-3381-2012.pdf
http://publications.jrc.ec.europa.eu/repository/handle/JRC69737
DOI: 10.5194/bg-9-3381-2012
Type: Articles in periodicals and books
Abstract: Pantropical above-ground tree biomass and carbon estimates commonly lack integrated height (H) to parameterise estimates to local forest allometry. Using height (H):diameter (D) models derived from a new tree architecture database (n=43,000 concomitant H and D measurements), we here provide a first pantropical evaluation of the effects of incorporating H on forest biomass estimates by geographic locale (plot, region, and continent). We examine the following questions: 1) Which is the best H-model form and geographic scale to include in biomass models to significantly reduce tree-by-tree uncertainty in estimates of destructive biomass? 2) How much does including height reduce uncertainty in plot-level forest biomass estimates tropics-wide? 3) How does inclusion of H in biomass estimates modify pantropical plot-level and continental-scale estimates of forest biomass? We compiled published and unpublished pantropical destructively sampled biomass data, and permanent sample plot data. Together with our tropical tree H database these were used to evaluate the effect of H on estimating biomass of real (destructively) sampled data (n=21 plots; 1816 trees) and on estimating biomass in permanent forest sample plots (n=327) across the tropics. Inclusion of H in biomass models significantly alters biomass estimates relative to estimates based on D+ρW (wood specific gravity) alone. The mean relative error in biomass estimates across diameter classes when including H was half (mean 0.06) that when excluding H (mean 0.13). The power- and Weibull-H asymptotic model provided the greatest reduction in uncertainty. We show fundamentally different stand structure across the four forested tropical continents. African forests store a greater portion of total biomass in large-diameter trees and trees are on average larger in diameter. This contrasts to forests on all other continents where smaller-diameter trees contain the greatest fractions of total biomass. Total biomass per hectare is greatest in Australia, the Guyana Shield, and Asia and lowest in W. Africa, W. Amazonia, and the Brazilian Shield (descending order). The Weibull-H model was selected as preferable to the other model forms because it reduced uncertainty in small-diameter classes that contain the bulk of biomass per hectare in most forests. When propagating error associated with estimating true biomass to the biomass distribution found in these permanent plots, plot level errors when H was not included in the allometric equations varied from 6.6 to 112.4 (mean, 41.8) Mg ha-1. Including Weibull-H estimates reduced these errors considerably: they then ranging from -2.5 to 23.0 (mean, 8.0) Mg ha-1.
JRC Directorate:Sustainable Resources

Files in This Item:
There are no files associated with this item.


Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.