Formulation of allometric equation
Based on the data collected, several equations were developed. Before
establishing the allometric equation, scatter plots were used to see
whether the relationship between independent and dependent variables was
linear. Furthermore, several allometric relationships between
independent and dependent variables were tested. The independent
variables included DBH, Height, and wood density, whereas the dependent
variable was the dry weight of the AGB. Because the data exhibited
heteroscedasticity, a power function was an inappropriate model in this
study, so the data was transformed for linear regression using
logarithmic transformation. The transformation equalized the variance
over the entire range of biomass values, which satisfies the
prerequisite of linear regression (Sokal and Rohlf, 1995; Sprugel,
1983).
To develop an allometric model for the above-ground biomass, Diameter at
breast height ((DBH), Height (H) and Wood density (WD) were taken as
explanatory variables. Given that there are p effect variables
X1, X2, . . . ,Xp, there are
2p − 1 models that include all or some of these effect
variables (Picard et al., 2012). Thus, in this study, the 3
effect variables can result in 23-1 which becomes 7
candidate regression model (Table 1).
Table
1. The seven candidate regression model which serve to select the best
models