Using Bayesian Belief Networks to Identify Potential Compatibilities and Conflicts Between Development and Landscape Conservation

Experts with different land use interests often use differing definitions of land suitability that can result in competing land use decisions. We use Bayesian belief networks linked to GIS data layers to integrate empirical data and expert knowledge from two different land use interests (development and conservation) in Maine’s Lower Penobscot River Watershed. Using ground locations and digital orthoquads, we determined the overall accuracy of the resulting development and conservation suitability maps to be 82% and 89%, respectively. Overlay of the two maps show large areas of land suitable for both conservation protection and economic development and provide multiple options for mitigating potential conflict among these competing land users. The modeling process can be adapted to help prioritize and choose among different alternatives as new information becomes available, or as land use and land-use policies change. The current model structure provides a maximal coverage strategy that allows decision makers to target and prioritize several areas for protection or development and to set specific strategies in the face of changing ecological, social, or economic processes. Having multiple options can generate new hypotheses and decisions at more local scales or for more specific conservation purposes not yet identified by stakeholders and decision makers in the region. Subsequently, new models can be developed using the same process, but with higher resolution data, thereby helping a community evaluate the impacts of alternative land uses between different prioritized areas at finer scales