A recursive multiple regression model for predicting yields of grade lumber from lodgepole pine sawlogs

A study was conducted to determine the efficacy of an alternative to least squares methods for fitting regression models to lumber grade yield data. A method based on a simultaneous system of equations was used to specify the regression model. The system of equations was fitted using techniques suitable for models with limited dependent variables. The regression model was used in a study of value loss from beetle-killed timber for four sawmills with unique product lines. The method was shown to give accurate predictions of the yields of individual grades. The rate of decline in recoverable lumber value from dead lodgepole pine trees (with years-since-death) is faster for board mills compared to dimension and stud mills.

You must be logged in to download any documents. Please login (login accounts are free) or learn how to Become a Member