493 Parameterizing the Biological Condition Gradient in the northeast United States using a bayesian network approach

Wednesday, May 20, 2009: 4:15 PM
Vandenberg B
Roxolana Kashuba , Nicholas School of the Environment, Duke University, Durham, NC
Thomas F. Cuffney , North Carolina Water Science Center, U.S. Geological Survey, Raleigh, NC
Gerard McMahon , North Carolina Water Science Center, U.S. Geological Survey, Raleigh, NC
Kenneth Reckhow , Nicholas School of the Environment, Duke University, Durham, NC
Jeroen Gerritsen , Tetra Tech Inc., Owings Mills, MD
Susan P. Davies , Maine Department of Environmental Protection, Augusta, ME
The Biological Condition Gradient (BCG) is a method of systematically defining levels of ecosystem health using characteristics of ecosystem structure and function that respond to increasing stress.  BCG levels, or Tiers, help to standardize interpretation and communication of the condition of aquatic biota (collected with differing sampling methods and in diverse ecological settings), facilitate detection of incremental changes in ecological condition, and enable linking of management actions to meaningful environmental outcomes.  We translated the BCG conceptual framework into a Bayesian network of quantifiable nodes, relationships and probabilities to describe the effect of urbanization on macroinvertebrate biological condition in the Northeast U.S.  This was done by integrating Federal and State data with expert elicitation from experienced New England biologists to create a set of probabilistic linkages connecting urbanization metrics to interpretation of macroinvertebrate biological condition (via BCG Tiers).  By incorporating not only information available from data but also expert knowledge, and uncertainty associated with both data and experts, these probabilistic linkages are able to thoroughly characterize the system of interest.  Managers can interactively use this parameterized Bayesian network to calculate the probability of attaining desired aquatic ecosystem BCG tiers assuming different levels of urban stressors, environmental conditions and management options.