Abstract:

Objective: In order to improve the accuracy of biomass estimation of southern subtropical evergreen broad-leaved forest, the biomass distribution among different organs of 4 common tree species were investigated, and the models of single plant and different biomass components(aboveground and underground) were constructed for each of the 4 species. Method: Based on the historical data of forest survey of the southern subtropical evergreen broad-leaved forest in Guangzhou and its actual species composition, 4 common tree species, Castanopsis fissa, C. chinensis, Aleurites montana and Machilus chinensis were selected to measure the biomass of single plant and different components of each tree species using the harvest method, and further to construct the biomass models of each tree species. The influences of adding tree height (H) and wood density (ρ) as second independent variables in the models on the accuracy of the models were discussed. Result: The proportion of leaf biomass of the four 4 tree species accounted for 5.34%-7.28%, and the proportion was significantly lower of M. chinensis than of C. fissa (P < 0.05). The proportion of branch biomass accounted for 16.82%-24.20%, and significantly lower of M. chinensis than of C. chinensis and C. fissa (P < 0.05). The proportion of trunk biomass accounted for 47.22%-58.05%, and significantly lower of C. chinensis than of the other three species (P < 0.05). The proportion of root biomass accounted for 14.25%-22.25%, and significantly lower of C. fissa and A. montana than of C. chinensis and M. chinensis (P < 0.05). The proportion of leaf biomass of A. montana and M. chinensis decreased significantly (P < 0.01) with the increase of DBH. With the increase of DBH, the proportion of branch biomass increased, while the proportion of trunk biomass decreased of the four species. There was no significant correlation between the proportion of root biomass and DBH. The biomass model with DBH as a single independent variable had good fitting accuracy (average R²=0.947). Adding H compounded with D as one variable (D²H) in the models decreases the model fitness of aboveground and total biomass, and slightly improves the fitness of underground biomass. Including H as an independent second variable in the models increases the multiple col-linearity of the model. Adding ρ compounded with D as one variable (D²ρ) in the models decreases the model fitness of single plant and all components, but including ρ as an independent second variable in the models slightly improves the model fitness. Conclusion: The sunlight-loving tree species, A. montana, allocated more biomass to the trunk, while the shade tolerant tree species, C. chinensis, allocated more biomass to the crowns. The shade tolerant tree species have more developed roots. We suggest that the model of D as a single independent variable be used to estimate the biomass of the four species, and H is not recommended to be included in the model as the second variable. Adding ρ as the second independent variable slightly improves the fitting ability of the model. It is suggested to make a trade-off between accuracy and difficulty according to the investigation purpose in practical applications.

Key words: southern subtropical evergreen broad-leaved forest, allometric growth model, wood density, Guangzhou city, root-shoot ratio