Open Collections

Abstract

This thesis explores sampling and system error risks affecting projected stand values. British Columbia's Second Complete Provincial Inventory data for coastal Douglas-fir (Pseudotsuga menziesii var. menziesii), collected between 1961 and 1977, is used to construct a multiple regression equation predicting stand volume point estimates and prediction intervals in order to analyze system error and sampling error from a silvicultural and financial perspective. Incorporating sampling error from the model, represented by the 95% prediction interval, on silvicultural and financial decisions is explored for site indexes 15 m, 25 m and 35 m. System error is also analyzed by performing sensitivity analysis on several key independent variables: basal area; site preparation cost; projected timber price and discount rate. Incorporating sampling error into mean annual increment and long term sustained yield calculations can cause significant differences in harvest quotas. Financially, including sampling error can create considerable variation to projected net present values and internal rates of return. This risk to a projected NPV should be incorporated into the evaluation by calculating a new NPV with a risk adjusted discount rate. The effects of system error from uncertain future timber prices has the most profound impact on projected stand value. In addition, the choice of discount rate profoundly effects net present value. Basal area and site preparation cost variation have less of an influence on projected stand value. Analysis of site indexes lead to the conclusion that only SI 20 m and above are inside the extensive margin and are attractive in terms of private silvicultural investment.