Gunn E FEMC2007

Designing Models for Designing the Forest Estate

Eldon Gunn, Department of Industrial Engineering, Dalhousie University, 5269 Morris St., Halifax, NS, B3J 2X 4, Canada, eldon.gunn@dal.ca

In this paper, we examine the traditional cycle of using a long-term aspatial linear programming model to develop forest strategy and then using a spatial model to “put it on the ground”. We point out that the harvest flows, often seen as characterizing long-term strategy, can usually be achieved in a variety of quite different ways. The sequence of particular stand types harvested in these alternative solutions may bear little resemblance to each other. Attempting to put any one of these solutions on the ground is an essentially arbitrary choice with little reason to believe that it is any better than other alternatives.

Most ecological and economic strategic issues have a large amount of spatial content. If we want the solution to the long term strategic models to reflect these strategic considerations, then these need to be modeled directly. If we want to take certain spatial properties from the solution to the model, this only makes sense if we specify these as constraints on the model in the first place. If we wish to use linear programming models to design strategy for the forest estate, we need to design out models to achieve this end. We point out a variety of ways in which highly spatial linear programming models can be constructed.