The objective of this study was to find out the prediction model for copal production (Y) by the use of prediction variables. The variables were diameter (X1), bark thicknesses (X2), stand density (X3), slope (X4), and direction of the slope (X5). The study was conducted at Senduro Forest District, Probolinggo Forest Management Unit of Perum Perhutani, Unit II East Java. Data was collected from 80 sample trees of Agathis loranthifolia Salibs which were tapped on August 2006. The result of the study showed that there were several regression prediction models for copal production: doubled linear, multiplicative, exponential and quadratic regression models. Multiplicative regression model with the highest R2-adj value was then chosen as the best prediction model for copal production. Thereby, production estimation model of copal production of Agathis loranthifolia Salis. was LogY = 0.397 + 1.54 LogX1 + 0.496 LogX2 - 0.528 LogX3 + 0.201 LogX4 ; or Y = 2.4945X11.54X20.496X3-0.528X40.201; R-Sq =84.7%, R-Sq(Adj) = 83.7%.
Keywords: multiplicative regression model, copal, diameter, bark thickness, stand density, slope