Decision Analysis and Sensitivity Testing of Reforestation Strategies1

Abstract

Abstract Decision analysis might provide better guidance in silvicultural decisions if users could understand the relative impact of component assumptions and subjective judgments more easily. A reforestation project in western Oregon that involved four alternatives was structured as a decision analysis problem. The optimal solution ranked the alternatives based on their probabilistically weighted soil expectation value (SEV). The solution was tested to determine its sensitivity across realistic ranges in costs, probabilities, and biological responses. Relative rankings were most sensitive to cost components with wide natural ranges (such as prescribed burning) and to variations in 15-year survival and growth predictions; rankings were least sensitive to probability estimates. Rankings were quite stable even when the detail in the parameters was increased. A systematic sensitivity analysis pointed out the impact of errors or uncertainty in judgment and focused the decision maker's effort in information gathering and reanalysis. West. J. Appl. For. 6(3):73-78.