Model fitting
AGB was regressed against the various form of predictors (i.e., DBH, H
and WD) and three allometric models showed significant performance (p
< 0.05) on their F-test (Model 1, 3 and 6)(Table 3). Among the
three allometric equations which showed significant performance, the
selection of the best-fit model was conducted based on their P-value,
adjusted r2, AIC, RMSE. The first two models (models 1
and 3) are nested to model 6 and hence the complete model 6is selected
against the nested models 1 and 3 since the p-values of coefficients of
the complete model are significant (p < 0.05). Model6 selected
as the best-fit model is indicated below.
\(lnAGB=\ c+\alpha lnDBH+xlnWD\ +\varepsilon\), with
values of coefficient inserted the model form is:\(\ lnAGB=1.806+1.419\times lnDBH+0.628\times lnWD\)……………………
(Equation 2)
\begin{equation}
{(r}^{2}=0.951,\ \varepsilon=0.219,\ AIC=-1.026,\ RMSE=0.208)\nonumber \\
\end{equation}Where AGB is in kg, DBH in cm and WD in cm 3.
The ANOVA also showed that the best-fit model was found to be
statistically significant (F=283.042, p < 0.000).