| Summary: | Maritime pine (Pinus pinaster Ait.) is one of the most important tree species in Turkey for establish industrial plantations. In this study, stem taper models were developed for maritime pine plantations in İstanbul-Sarıyer region. For this aim, two commonly used and well-known taper models belonging to different groups (segmented and variable-form) were evaluated. In this frame, totally 194 destructively sampled trees were used. The equations were fitted with ordinary nonlinear least squares method. We incorporated a second-order continuous-time autoregressive error structure (CAR(2)) to address of autocorrelation problem and used condition numbers to detect presence of multicollinearity. The predictive performance of the taper models was evaluated using five different evaluation criteria (R2, RMSE, AIC, BIC, and AAE). According to evaluation statistics, Fang et al. (2000) model provided better results than the Kozak (2004) model for height at a specific diameter, merchantable volume and total volume predictions except for diameter at a specific height. However, there are no significant differences among both models in terms of prediction performance of the models. Therefore, it will be more reasonable for the decision maker to decide which model will be preferred in maritime pine plantations.
|
|---|
