Application of Bayesian networks to large-scale predictive ecosystem mapping
Adrian Walton · 2005
Large-scale ecosystem maps are essential tools for managers of forest-related activities. In British Columbia, the prevailing approach for ecosystem mapping has been to use an expert system that captures expert knowledge in the form of a belief matrix. In this project, I replaced the belief matrix with a Bayesian network in an attempt to overcome some of the drawbacks of the belief-matrix approach. I created a Bayesian-network knowledge base applied to an area encompassing Prince Rupert and part of the following three biogeoclimatic units: Coastal Western Hemlock Very Wet Maritime Montane, Coastal Western Hemlock Very Wet Maritime Submontane and, Coastal Western Hemlock Very Wet Hypermaritime Central. Using each knowledge base, I produced a map of grouped site series. Accuracy assessments performed on each of the maps of grouped site series revealed that the maps poorly predicted the spatial distribution of rare and very wet site-series groups.