Dominant Height and Site Index Models for Acacia mangium Willd. Plantation

Accurate estimate of site productivity (site quality) is essential for making sound forest management decisions. Site index, defined as dominant height of a stand at a base age, is often used to measure site qualite. In this study, seven algebraic difference equations (ADE) based on the models proposed by Chapman-Richards, Lundqvist-Korf and McDill-Amateis were tested for fitting site index models for Acacia mangium Willd. plantations in South Sumatra,Indonesia. The models were fitted using re-measured, permanent plot data that involves non-overlapping growth intervals. The issues of serial correlation and heteroscedasticity in regression errors were tested, but not significant for any model. Numerical and graphical analyses were used to evaluate the performances and predictive abilities of the models. Results indicated that the models having an explicit asymprote parameter generally had greater accuracy and precision of parameter estimates and dominant height predicitions, and could be applied for producing polymorphic site index curves. The selected model was derived from the Lundqvist-Korf function, which provided the best compromise between statistical and biological considerations and produced the most adequated site index curves. A base age of 6 years was proved to be best for predictions of the selected model. The model had a superior performance in terms of predictive ability and flexibility in use when compared to previously developed models for A.mangium.