Data from: Developing allometric models to predict the individual aboveground biomass of shrubs worldwide

Aim Existing global models to predict standing woody biomass are based on
trees characterized by a single principal stem, well-developed in height.
However, their use in open woodlands and shrublands, characterized by
multistemmed species with substantial crown development, generates a high
level of uncertainty in biomass estimates. This limitation led us to i)
develop global predictive models of shrub individual aboveground biomass
based on simple allometric variables; ii) to compare the fit of these
models with existing global biomass models; and iii) to assess whether
models fit change when bioclimatic variables are considered. Location
Global. Time period Present. Major taxa studied 118 species. Methods We
compile a database of 3243 individuals across 49 sites distributed
worldwide. Including basal diameter, height and crown diameter as
predictor variables, we built potential models and compared their fit
using generalized least squares. We used mixed effects models to determine
if bioclimatic variables improved the accuracy of biomass models. Results
Although the most important variable in terms of predictive capacity was
stem basal diameter, crown diameter significantly improved the models fit,
followed by height. Four models were finally chosen, with the best model
combining all these variables in the same equation (R2 = 0.930, RMSE =
0.476). Selected models performed as well as established global biomass
models. Including the individual bioform significantly improved the models
fit. Main conclusions Basal diameter, crown diameter and height measures
could be combined to provide robust AGB estimates of individual shrub
species. Our study supplements well-established models developed for
trees, allowing more accurate biomass estimation of multistemmed woody
individuals. We further provide tools for a methodological standardization
of individual biomass quantification in these species. We expect these
results contribute to improve the quality of biomass estimates across
ecosystems, but also to generate methodological consensus on field biomass
assessments in shrubs.