Identifying an Optimal Bald Eagle Monitoring Program for Southwest Alaska National Parks

The Southwest Alaska Inventory and Monitoring Network includes bald eagle monitoring as part of their Vital Signs Monitoring Plan. Lake Clark National Park and Preserve, Katmai National Park and Preserve, Kenai Fjords National Park, and Wrangell – St. Elias National Park and Preserve monitor bald eagles annually, albeit slightly differently among parks. Since monitoring decisions involve multiple objectives and stakeholders, there was a need for a structured approach to identify an optimal monitoring program. We used a structured decision making process and an iterative, four-round Delphi Process to collect information about long-term bald eagle monitoring from experts. We collected information about important stressors to bald eagles, and information about various monitoring metrics. We also held an in-person meeting with members of the expert panel to designate fundamental objectives for decisions about the long-term bald eagle monitoring, which are: 1) Minimize cost; 2) Minimize effort; 3) Maximize amount of accurate information collected about bald eagles; 4) Maximize the ability to detect change in bald eagle populations. We used a consequence table to compare monitoring metrics and reduce the list of metrics to consider for the program. Panelists weighted the four fundamental objectives by importance using a swing-weighting technique. Objectives weights are calculated using averages of panelist response: Maximize accurate information: 33.1%; Maximize ability to detect change: 32.3%; Minimize effort: 17.6%; Minimize cost: 17.1%. A Bayesian Decision Net, which uses linear value modeling, compares alternative monitoring scenarios using information collected during the Delphi Process and the weight of fundamental objectives to determine the most optimal scenario. Our model identified a comprehensive monitoring scenario, which includes all feasible monitoring metrics, as the most optimal decision, followed by the current monitoring scenario. We performed a cross-stakeholder sensitivity analysis and an additional sensitivity analysis by varying objective weights. We also performed a sensitivity analysis using a twofunction decision model, combining similarly weighted objectives into two objectives. We found that the cost and effort of the comprehensive monitoring scenario must be 4.4 times greater than the cost and effort of the current scenario, for the current monitoring scenario to become the most optimal decision