Abstract: We investigated main tree-measuring factors and the biomass of aboveground component of two major urban tree species, Cinnamomum camphoraand Magnolia grandiflora, in Shanghai by harvest method. We further analyzed biomass allocation patterns and explored optimal biomass allometric models and predictor variables by nonlinear equation. The results showed that: (1) The biomass of aboveground components of both species was characterized by stem>branch>leaf. Relative contributions of leaves, branches and stems to total aboveground biomass of C. camphoraand M. grandiflora were 17.88% vs. 27.31%, 30.48% vs. 27.93%, and 51.64% vs. 44.76%, respectively. There was no significant difference between leaves and branches for M. grandiflora. (2) Nine allometric models for biomass were developed based on three variables of diameter at breast height, tree height, and crown width. Results of fitness evaluations of the equations showed that model 8, model 3, model 4 and model 7 fit well to leaf, branch, stem, and total aboveground biomass for C. camphora, while model 4, model 9, model 2 and model 4 for M. grandiflor, respectively. Due to the difficulty of access to tree height and crown width, optimal models selected in application should meet both practical and accurate requirements on the basis of statistics and adaptation of models.

Key words: Pinus massoniana, C:N:P stoichiometry, red soil erosion degradation