Indicators of regime shifts in ecological systems: What do we need to know and when do we need to know it

Because novel ecological conditions can cause severe and long‐lasting environmental damage with large economic costs, ecologists must identify possible environmental regime shifts and pro‐actively guide ecosystem management. As an illustrative example, we applied six potential indicators of impending regime shifts to S. R. Carpenter and W. A. Brock's model of lake eutrophication and analyzed whether or not they afforded adequate advance warning to enable preventative interventions. Our initial analyses suggest that an indicator based on the high‐frequency signal in the spectral density of the time‐series provides the best advance warning of a regime shift, even when only incomplete information about underlying system drivers and processes is available. In light of this result, we explored two key factors associated with using indicators to prevent regime shifts. The first key factor is the amount of inertia in the system; i.e., how fast the system will react to a change in management, given that a manager can actually control relevant system drivers. If rapid, intensive management is possible, our analyses suggest that an indicator must provide at least 20 years advance warning to reduce the probability of a regime shift to <5%. As time to intervention is increased or intensity of intervention is decreased, the necessary amount of advance warning required to avoid a regime shift increases exponentially. The second key factor concerns the amount and type of variability intrinsic to the system, and the impact of this variability on the power of an indicator. Indicators are considered powerful if they detect an impending regime shift with adequate lead time for effective management intervention, but not so far in advance that interventions are too costly or unnecessary. Intrinsic “noise” in the system obscures the “signal” provided by all indicators, and therefore, the power of the indicators declines rapidly with increasing within‐ and between‐year variability in measurable variables or parameters. Our results highlight the key role of human decisions in managing ecosystems and the importance of pro‐active application of the precautionary principle to avoid regime shifts.