Fuzzy cognitive mapping as a tool to define management objectives for complex ecosystems

Defining objectives for ecological rehabilitation requires consideration of how an ecosystem responds to management. Validated quantitative models of physical, chemical, and biological processes are the best way to project such impacts; however, time, data, and model limitations often make these approaches impractical. An alternative is to encode expert knowledge about interactions among ecosystem components in a fuzzy cognitive map (FCM), which then translates that subjective, qualitative information into predictions of the effects of management on an ecosystem. Herein, we present the steps involved in constructing an FCM of an ecosystem, interpreting FCM output using multivariate statistics, and portraying the information in an easily communicated fashion. To illustrate these ideas, we rely on a complex (>160 variables) ecosystem model built for the Lake Erie watershed under the auspices of the Lake Erie Lakewide Management Plan (LaMP). Based on our experiences in building this model, we also offer recommendations for increasing the efficiency of the model‐development and interpretation process. Use of the FCM method in this case promoted constructive interaction among dozens of scientists, managers, and the public, as well as providing insights concerning the potential effects of broad classes of management actions upon the Lake Erie ecosystem. The analysis focused the attention of participants on four broad alternatives for the Lake. One represents present conditions, and another results from a decrease in nutrient inputs but an increase in stresses from land use and human disturbance. The two others involve reduced stress from nutrients and land use, with one having relatively more nutrients and less human disturbance and fishing. The latter ecosystem alternative was tentatively endorsed by LaMP management, and all four alternatives will be reviewed by the public.