Rapid transitions in ecosystem structure, or regime shifts, are a hallmark of alternative stable states. However, regime shifts can occur even when feedbacks are not strong enough to cause alternative stable states. Despite the theoretical and practical importance of distinguishing between different types of threshold responses, empirical evaluations of alternative stable states on management-relevant scales are rare. NTL-LTER researchers have observed rapid transitions between invasive rusty crayfish (Orconectes rusticus) and native sunfishes (Lepomis spp.) in two LTER lakes. A transition from sunfish to rusty crayfish dominance occurred as rusty crayfish invaded Trout Lake, WI, and the reverse transition occurred following an 8-year experimental removal of rusty crayfish from Sparkling Lake, WI. We fit a stage-structured population model of species interactions to 31 years of time-series data from each lake. This model allowed identification of significant drivers of both rusty crayfish and sunfish abundance, and parameter estimates were used to assess the likelihood of shifts between alternative stable states causing observed transitions.
Water level was identified as an important environmental driver of community composition, with drought conditions reducing rusty crayfish recruitment and allowing sunfish dominance. Predation of rusty crayfish by non-sunfish littoral fishes did not significantly influence crayfish abundance. Reciprocal negative interspecific interactions between rusty crayfish and sunfish led to alternative stable states in the model, where each species was capable of excluding the other within a narrow range of environmental conditions (Fig. 3). However, rapid transitions caused by a threshold response lacking alternative equilibria were also possible given uncertainty in parameter estimates. Simulated forward and backward transitions between species dominance occurred at different environmental conditions (i.e., hysteresis) even when the parameters used for simulation did not predict alternative stable states as a result of slow species responses to environmental drivers. Thus, alternative stable states are possible, but by no means certain. Explanations for rapid transitions in this system coupled with our results highlight the difficulties associated with distinguishing alternative states from other types of threshold responses. However, whether regime shifts are caused by alternative stable states may be relatively unimportant in this system, as the range of conditions over which transitions occur is narrow, and under most conditions the system is predicted to exist in only a single state.