A simple Bayesian Belief Network for estimating the proportion of old-forest stands in the Clay Belt of Ontario using the provincial forest inventory

Abstract

The differences between boreal forest landscapes produced by natural disturbance regimes and landscapes produced by harvesting are important and increasingly well documented. To continue harvesting operations while maintaining biodiversity and other ecosystem services, government policies and certification processes are pushing for practices that preserve landscape features within their range of natural variability. One major shortcoming in the implementation of such a strategy is the lack of complete spatial or temporal information about these landscape features, such as the proportion of old stands, which are believed to act as a coarse filter for conservation if they remain representative enough of natural conditions. The objective of this study was to quantify the proportion of old stands in a very large landscape by combining fragmentary knowledge from two different sources, i.e., a provincial forest inventory and existing fire history reconstructions using a Bayesian Belief Network. This study was conducted over a 6.5 Mha landscape located within the Clay Belt of the province of Ontario, Canada, and suggests that more than 72.4% of this area is occupied by stands where no fire occurred during the last 150 years. The implications for management and potential for future research are discussed.