Accommodating heteroscedasticity in allometric biomass models

Allometric models are commonly used to predict forest biomass. These typically take nonlinear power-law forms that individual tree aboveground biomass (AGB) as functions of diameter at breast height (D) and/or (H). Because the residual variance is in most cases heteroscedastic, accommodating heteroscedasticity (i.e., heterogeneity variance) becomes necessary when estimating model parameters. We tested several weighting procedures and a logarithmic transformation for allometric models. further evaluated effectiveness these with emphasis on how they affected estimates mean AGB per hectare their standard errors large areas. Our results revealed some were more effective than others was greater single predictor but less based both D H. Failing effectively accommodate produced small moderate differences errors. However, between (models H versus only), regardless approach. Similar consequences observed respect whether prediction uncertainty or not included When including uncertainty, estimated means increased substantially, by 44–59%. Therefore, avoid possible negative large-area estimation, we recommend: (i) testing procedure models, (ii) incorporating total estimate (iii) an additional variable