A succession of theories: purging redundancy from disturbance theory.

The topics of succession and post-disturbance ecosystem recovery have a long and convoluted history. There is extensive redundancy within this body of theory, which has resulted in confusion, and the links among theories have not been adequately drawn. This review aims to distil the unique ideas from the array of theory related to ecosystem change in response to disturbance. This will help to reduce redundancy, and improve communication and understanding between researchers. We first outline the broad range of concepts that have developed over the past century to describe community change in response to disturbance. The body of work spans overlapping succession concepts presented by Clements in 1916, Egler in 1954, and Connell and Slatyer in 1977. Other theories describing community change include state and transition models, biological legacy theory, and the application of functional traits to predict responses to disturbance. Second, we identify areas of overlap of these theories, in addition to highlighting the conceptual and taxonomic limitations of each. In aligning each of these theories with one another, the limited scope and relative inflexibility of some theories becomes apparent, and redundancy becomes explicit. We identify a set of unique concepts to describe the range of mechanisms driving ecosystem responses to disturbance. We present a schematic model of our proposed synthesis which brings together the range of unique mechanisms that were identified in our review. The model describes five main mechanisms of transition away from a post-disturbance community: (i) pulse events with rapid state shifts; (ii) stochastic community drift; (iii) facilitation; (iv) competition; and (v) the influence of the initial composition of a post-disturbance community. In addition, stabilising processes such as biological legacies, inhibition or continuing disturbance may prevent a transition between community types. Integrating these six mechanisms with the functional trait approach is likely to improve the predictive capacity of disturbance theory. Finally, we complement our discussion of theory with a case study which emphasises that many post-disturbance theories apply simultaneously to the same ecosystem. Using the well-studied mountain ash (Eucalyptus regnans) forests of south-eastern Australia, we illustrate phenomena that align with six of the theories described in our model of rationalised disturbance theory. We encourage further work to improve our schematic model, increase coverage of disturbance-related theory, and to show how the model may link to, or integrate with, other domains of ecological theory.