Carver M FEMC2007

Determining Hydrologic Hazards in BC using Bayesian Belief Networks To Support Decision-Making in Mountain-Pine-Beetle Response

Martin Carver, Water Stewardship Division, Ministry of Environment, P.O. Box 9362 Stn Prov Govt, Victoria, BC, V8W 9M2, Canada, Martin.Carver@gov.bc.ca
Greg Utzig, Kootenai Nature Investigations, Ltd., 602 Richards St, Nelson, BC, V1L 5K5, Canada, g13utzig@telus.net
Randy Sulyma, Fort St. James Forest District, Ministry of Forests and Range, Box 100, Fort St. James, BC, V0J 1P0, Canada, Randy.Sulyma@gov.bc.ca

Throughout British Columbia, anthropogenic disturbances combine with intrinsic watershed characteristics to yield a mosaic of hydrologic hazards with the potential to diminish aquatic values. These hazards are of concern in the management of a variety of water-related resource values including aquatic habitat, drinking water, downstream infrastructure and recreation. Measures taken to protect these values often have significant impacts on potential timber supply. In the past, hydrologic hazards have been modelled only on a selective and regional basis. In this study, Bayesian Belief Networks are used to model seven distinct hydrologic hazards with a potentially provincial application. The hazards are: peak flow increase, bedload increase, suspended sediment increase, low flow decrease, stream temperature increase, organic input decrease (fine or coarse) and adverse water chemistry. Phase One of the present study estimates the first three of these hazards using available digital information.

The framework approach has been developed on a pilot basis for two study areas within areas affected by the Mountain Pine Beetle. As pilot areas for testing and validating model construction and assumptions, Supply Block C of the Prince George TSA (Fort St. James) and the Kootenay Lake TSA offer strongly contrasting aquatic resource concerns and physiographic, hydrologic and ecological characteristics. Due to varying data quality and availability, modelling of the hydrologic hazards has required innovative interpretation of existing data to allow the modelling of complex ecological relationships.

When linked to the outputs from forest estate modelling, the hydrologic model can be used as a prediction tool for potential hydrologic impacts resulting from a range of scenarios of forest development and natural disturbance. In Phase Two of the project, the low-flow and stream-temperature hazards will be modelled and, along with the three hazards developed in Phase One, applied to the larger area affected by the Mountain Pine Beetle. In addition, water values will be incorporated to determine overall risk. It is expected that the models may provide decision-support for an array of resource planning and public safety decisions, particularly in light of potential hydrologic changes associated with Global Climate Change.