| Abstract: | Summary Conservation management in agricultural landscapes involves identification and prioritization of assets, and interventions to reverse or arrest decline. Planning requires synthesis of hydrological, ecological and agronomic information and intuitions. We provide a case study involving the Lake Warden Wetland System, a Ramsar‐listed site on the south coast of Western Australia threatened by salinity and flooding. As the relative merits of management options (including engineering‐based solutions and catchment revegetation) may be sensitive to climate change, we captured our knowledge and understanding of the effectiveness of options under different climate change scenarios using Bayesian belief networks. We insulated against overconfidence by an info‐gap analysis that describes the trade‐off between aspiration and immunity to uncertainty. Only engineering‐based solutions offer reasonable prospects for achieving stated conservation goals in the Lake Warden Wetland System within a 25‐year time horizon. Marginal gains derived from co‐investment in revegetation varied among the assets. We advocate explicit treatment of uncertainty and risk‐based approaches to decision‐making to equip managers with a means of progressing conservation goals. The complementary insights offered by Bayesian belief networks and info‐gap analysis provide a sound basis for managers to assess the extent to which candidate management actions are robust to uncertainty. |
