Knowing when (not) to attempt ecological restoration

Here we argue that there are two important steps in the decision process to restore ecological system that are often ignored. First, consideration of restoration is in response to observed change in a system, but ecological systems can fluctuate widely in their normal dynamic. Thus, there is an imperative to interpret ecological change; shifts in community structure that represent typical fluctuations in a properly functioning ecosystem do not warrant restoration, while change associated with phase shift in the system may well demand restoration action. Second, where restoration effort is warranted, it needs to be determined whether management responses are likely to be successful within resource constraints. Where ecological change involves pronounced hysteresis, even massive effort may have little chance in effecting recovery to a preferred ecosystem state. Theory and models indicate that consideration of the characteristic length scales (CLSs) of ecological systems provides an unambiguous interpretation of ecological change, enabling differentiation of typical fluctuations from phase shift, and here we show that CLSs can be calculated for real communities from their species' dynamics, and that their behavior is as predicted from theory. We also show that for ecological systems where local interactions and forcings are well understood, validated simulation models can provide a ready means to identify hysteresis and estimate its magnitude. We conclude that there are useful tools available for ecologists to address the key questions of (1) whether restoration attempts are warranted in the first place and, if they are, (2) whether it is practical to pursue them.