Probability, possibility and evidence: approaches to consider risk and uncertainty in forestry decision analysis

Uncertainty is an important issue in the support of any forestry decision. Forestry decision making today typically involves objectives and information concerning ecological, economic and social issues. The consequences of alternative forest management programmes might be hard to assess, and predictions and assessments always include uncertainty. Forestry decisions also often concern large areas, long time horizons and multiple stakeholders, which further complicates forest management planning and increases uncertainty involved in it. This paper deals with different definitions and classifications of uncertainty, sources of uncertainty, and theories and methodologies presented to deal with uncertainty. The aim is to provide readers with an overview of alternative approach for coping with uncertainty, especially from the viewpoint of forestry and natural resource management applications. Generally taken, there are two main conventional approaches, namely classical (frequentist) and Bayesian probability theory. These lead to either classical or Bayesian decision theory, respectively. In addition, uncertainty can be dealt with, for instance, using the fuzzy set theory. This theory mostly deals with uncertainty due to the ambiguity of concepts. So far, in decision support tools, probability and fuzzy set theory are the most common approaches. However, the possibility theory and the evidence theory, for instance, can also be relied upon when managing uncertainty. These theories deal with subjective beliefs and expert judgements. They are able to deal with partial information and pure ignorance. The counterparts to the classical decision rules based on these theories are presented, as well as some decision support methods designed using the approaches presented. Because of the manifold sources of uncertainty, all these approaches can have application in the support of forestry decisions