Charles Darwin University

CDU eSpace
Institutional Repository

 
CDU Staff and Student only
 

Allometry for estimating aboveground tree biomass in tropical and subtropical eucalypt woodlands: towards general predictive equations

Williams, Richard J., Zerihun, Aalsew, Montagu, Kevin D., Hoffman, Madonna, Hutley, Lindsay B. and Chen, Xiaoyong (2005). Allometry for estimating aboveground tree biomass in tropical and subtropical eucalypt woodlands: towards general predictive equations. Australian Journal of Botany,53(7):607-619.

Document type: Journal Article
Citation counts: Scopus Citation Count Cited 23 times in Scopus Article | Citations
Altmetric Score Altmetric Score is 3
Google Scholar Search Google Scholar

IRMA ID 72965231xPUB7
Title Allometry for estimating aboveground tree biomass in tropical and subtropical eucalypt woodlands: towards general predictive equations
Author Williams, Richard J.
Zerihun, Aalsew
Montagu, Kevin D.
Hoffman, Madonna
Hutley, Lindsay B.
Chen, Xiaoyong
Journal Name Australian Journal of Botany
Publication Date 2005
Volume Number 53
Issue Number 7
ISSN 0067-1924   (check CDU catalogue open catalogue search in new window)
Scopus ID 2-s2.0-28644450257
Start Page 607
End Page 619
Total Pages 13
Place of Publication Collingwood
Publisher CSIRO Publishing
Field of Research 0607 - Plant Biology
HERDC Category C1 - Journal Article (DEST)
Abstract A fundamental tool in carbon accounting is tree-based allometry, whereby easily measured variables can be used to estimate aboveground biomass (AGB). To explore the potential of general allometry we combined raw datasets from 14 different woodland species, mainly eucalypts, from 11 sites across the Northern Territory, Queensland and New South Wales. Access to the raw data allowed two predictor variables, tree diameter (at 1.3-m height; D) and tree height (H), to be used singly or in various combinations to produce eight candidate models. Following natural log (ln) transformation, the data, consisting of 220 individual trees, were re-analysed in two steps: first as 20 species - site-specific AGB equations and, second, as a single general AGB equation. For each of the eight models, a comparison of the species - site-specific with the general equations was made with the Akaike information criterion (AIC). Further model evaluation was undertaken by a leave-one-out cross-validation technique. For each of the model forms, the species - site-specific equations performed better than the general equation. However, the best performing general equation, ln( AGB)=- 2.0596+ 2.1561 ln( D)+ 0.1362 ( ln( H))(2), was only marginally inferior to the species - site-specific equations. For the best general equation, back-transformed predicted v. observed values ( on a linear scale) were highly concordant, with a slope of 0.99. The only major deviation from this relationship was due to seven large, hollow trees ( more than 35% loss of cross-sectional stem area at 1.3 m) at a single species - site combination. Our best-performing general model exhibited remarkable stability across species and sites, when compared with the species - site equations. We conclude that there is encouraging evidence that general predictive equations can be developed across sites and species for Australia's woodlands. This simplifies the conversion of long-term inventory measurements into AGB estimates and allows more resources to be focused on the extension of such inventories.
DOI http://dx.doi.org/10.1071/BT04149   (check subscription with CDU E-Gateway service for CDU Staff and Students  check subscription with CDU E-Gateway in new window)
 
Versions
Version Filter Type
Access Statistics: 96 Abstract Views  -  Detailed Statistics
Created: Fri, 12 Sep 2008, 08:35:25 CST by Administrator