Patthanan Porkar. Biomass and carbon equations in four tree species in degraded mixed deciduous forest Sai Yok Kanchanaburi. Master's Degree(Appropriate Technology for Resources and Environmental Development). Mahidol University. Mahidol University Library and Knowledge Center. : Mahidol University, 2008.
Biomass and carbon equations in four tree species in degraded mixed deciduous forest Sai Yok Kanchanaburi
Abstract:
The condition of most mixed-deciduous secondary forest in the western part of Thailand
has become degraded due to fire and logging activities. The pressures brought by climate change
have increased the urgency for biomass assessment to measure the forests potential as a carbon
sink and source. Much attention is given to development of methods for measuring how much
biomass there is in natural forests. Therefore improving the quality and precision of the models for
carbon in natural forests is very important. This study aims to investigate the relation of
independent functions such as dry weight biomass (B) and dependent functions such as diameter at
breast height (D), height (H) and wood specific gravity ( ρ ) and formulate biomass equations for
four common tree species (Sterculia pexa, Millettia brandisiana, Grewia eriocarpa and Bridelia
ovata). Destructive sampling of 33 trees was performed to collect biomass (B) and tree dimension
(D, H and ρ ) data in the four species. The empirical relationships between tree dimensions and
biomass were sought for the purpose of on allometric regression. Four regression models that had
different independent variables in four patterns (D, ρ D, D2H and ρ D2H) in four species were
compared. Results showed strong correlations among the variables B, D and H, but not wood
specific gravity, which had low a correlation with other variables. Moreover, wood specific gravity
showed differences among the four species. Variability of ρ (wood specific gravity) indicated that
identification of species is required for accurate modelling. The best regression models for
estimating biomass had two forms:ln(B) = c +αln(D) , and ln(B) = c+α ln( ρ D2H). Dry weight
of an individual tree as estimated by the regression model using ρ D2H, had an average estimated
error in the range of 0.09-2.66%. This was just as suitable as using only D which had an average
estimated error in the range of 0.28-1.77% but it had higher collection cost data in the field.
However, wood specific gravity was necessity in model because it was needed for converting tree
volume to dry weight biomass. Therefore using ρ D2H as independent variable in model is most
appropriate. Linear regression results indicated significant differences among the four species. This
suggests that it is not possible to use only one linear regression for all four species. The results
show that the formulating specific-species regression models is important to access biomass and
carbon in mixed-deciduous degraded forest.