The Clarence Cannon National Wildlife Refuge (CCNWR) in the Mississippi River flood plain of eastern Missouri provides high quality emergent marsh and moist-soil habitat benefitting both nesting marsh birds and migrating waterfowl. Staff of CCNWR manipulate water levels and vegetation in the 17 units of the CCNWR to provide conditions favorable to these two important guilds. Although both guilds include focal species at multiple planning levels and complement objectives to provide a diversity of wetland community types and water regimes, additional decision support is needed for choosing how much emergent marsh and moist-soil habitat should be provided through annual management actions.

To develop decision guidance for balanced delivery of high-energy waterfowl habitat and breeding marsh bird habitat, two measureable management objectives were identified: nonbreeding Anas Linnaeus (dabbling duck) use-days and Rallus elegans (king rail) occupancy of managed units. Three different composite management actions were identified to achieve these objectives. Each composite management action is a unique combination of growing season water regime and soil disturbance. The three composite management actions are intense moist-soil management (moist-soil), intermediate moist-soil (intermediate), and perennial management, which idles soils disturbance (perennial). The two management objectives and three management options were used in a multi-criteria decision analysis to indicate resource allocations and inform annual decision making. Outcomes of the composite management actions were predicted in two ways and multi-criteria decision analysis was used with each set of predictions. First, outcomes were predicted using expert-elicitation techniques and a panel of subject matter experts. Second, empirical data from the Integrated Waterbird Management and Monitoring Initiative collected between 2010 and 2013 were used; where data were lacking, expert judgment was used. Also, a Bayesian decision model was developed that can be updated with monitoring data in an adaptive management framework.

Optimal resource allocations were identified in the form of portfolios of composite management actions for the 17 units in the framework. A constrained optimization (linear programming) was used to maximize an objective function that was based on the sum of dabbling duck and king rail utility. The constraints, which included management costs and a minimum energetic carrying capacity (total moist-soil acres), were applied to balance habitat delivery for dabbling ducks and king rails. Also, the framework was constrained in some cases to apply certain management actions of interest to certain management units; these constraints allowed for a variety of hypothetical Habitat Management Plans, including one based on output from a hydrogeomorphic study of the refuge. The decision analysis thus created numerous refuge-wide scenarios, each representing a unique mix of options (one for each of 17 units) and associated benefits (i.e., outcomes with respect to two management objectives).

Prepared in collaboration with the U.S. Fish and Wildlife Service, the decision framework presented here is designed as a decision-aiding tool for CCNWR managers who ultimately make difficult decisions each year with multiple objectives, multiple management units, and the complexity of natural systems. The framework also provides a way to document hypotheses about how the managed system functions. Furthermore, the framework identifies specific monitoring needs and illustrates precisely how monitoring data will be used for decision-aiding and adaptive management.