Abstract
The natural disturbance model of forest management is the basis of many of the sustainable forest management systems being proposed for the boreal forest of Canada. Wildfire is the dominant natural agent of disturbance in the boreal mixed-wood forest. The natural disturbance model assumes that timber harvesting systems emulating the annual area burned by natural fire, its spatial distribution, and the amount of residual material can be developed. It is further assumed that natural processes can be emulated closely enough to maintain forest biota at natural or near-natural population levels. This is a coarse filter approach to ecosystem management. In order to emulate the natural rate of disturbance, one needs to quantify it. The annual area burned in the study area, under natural conditions, is characterized as a random draw from a lognormal distribution. A modeling system comprised of an aspatial Monte Carlo simulation model and a linear programming based forest activity scheduling model was developed. The simulation model is used to develop 200 yr forecasts of probability distributions for habitat area of five vertebrate species under a stochastic wildfire regime. These probability distributions are used to construct habitat area constraints for use in an optimization model to help quantify the trade-offs between timber values and maintenance of the range of natural variability in the forest. The model is used to identify the trade-offs between forest harvesting, wildlife habitat, and the degree of similarity between the managed forest structure and the distribution of structures that could be generated by natural disturbance. FOR. SCI. 49(2):209–223.