Multi-product versus single-product timber rotation : comparative by regression and computer simulation

The study was done in cooperation with staff and assigned personnel of the Division of Land and Forest Resources, TVA, utilizing the simulation power of the agency's WRAP program. WRAP is an acronym for Woodland Resource Analysis Program.

The major purposes of the study were:

1. To develop regression models to examine multi-product versus single-product timber rotations through the financial-optimization procedures of WRAP.

2. To develop regression models to determine the relationship among the financial inputs to WRAP and the income stream generated in WRAP output.

The study had four stages: 1) The generation of individual - stand simulated data through WRAP analysis for natural stands of loblolly pine (Pinus taeda L.); 2) Model building - i.e., definitions of independent and dependent variables and the form of the equations; 3) Regression analysis using both forward and backward stepwise procedures; and 4) Sensitivity testing of the regression equations including graphic illustrations to predict optimum rotations and present worths by different variables.

Twelve separate equations were found by regression analysis using both forward and backward stepwise programs of the SAS package.

"Effective interest rate" was the most important independent model variable in predicting either optimum rotation or present worth. Relative prices of sawtimber versus pulpwood were also very important under the assumptions of the study.

The optimum rotation equations provided by this study varied from 18 to 55 years and should allow users to predict in advance the single product versus multi-product rotations which will be in the WRAP output. This knowledge should allow a better assignment of silvicultural treatments by the user. Price breaks among current product prices were also found for different optimum rotations.