Irreversible investment in wetlands preservation: Optimal ecosystem restoration under uncertainty

The question of how to manage a lacustrine wetland is analyzed given the uncertain potential for long-term lake level changes resulting from global warming and the uncertain biological processes involved in creating wetlands. Three management options are considered: do nothing; construct a dike that removes hydrological connections with the lake ("closed dike"); and build a dike that maintains a hydrological connection with the lake, but can be converted to a closed dike under adverse conditions ("open dike"). For all practical purposes, dike construction represents an irreversible choice. The model, a stochastic dynamic program, is used to optimize the timing and type of protective structure under a range of management goals. A wetland can either be optimal for fish or optimal for mammals and waterfowl, but not both. Because credible estimates of the economic values of wetland services do not exist, we treat those values as parameters in a multiobjective analysis and show the decisions implied by alternative valuations. The model is applied to the case of Metzger Marsh, a Lake Erie coastal wetland near Toledo, Ohio, where the decision was made in 1993 to construct an open dike. We find that the optimal decision is robust with respect to varying assumptions about the formation of barrier beaches and the probability of climate change, but that the decision is not robust to assumptions concerning the health of an unprotected Metzger Marsh. The most important source of uncertainty is the biological health of an unprotected wetland.