Natural disturbances and directional replacement of species

William J. Platt, Louisiana State University
Joseph H. Connell, University of California, Santa Barbara


The original concept of succession emphasized directional changes in species composition occurring over time on "blank slates" (new substrates or those that followed catastrophic disturbances). In this paper, we explore relationships between different effects of disturbances on residents and the initiation of directional species replacement on sites in a landscape. We present a conceptual model involving recurrent natural disturbances of sites in a landscape containing two species, one of which arrives at sites early and the other arrives late following disturbance. We predict effects of recurrent catastrophic disturbances (no survivors) and non-catastrophic disturbances (some survivors) on the initiation of directional species replacement, assuming temporal gradients in environmental conditions on sites after disturbance. We predict that directional species replacement will be initiated by catastrophic disturbances, and by non-catastrophic disturbances when the early species does not survive, but reinvades. Most natural disturbances tend to leave survivors and do not remove early species preferentially. As a result, many post-disturbance changes are possible on sites; directional species replacement is only one change, and often may not be the most common. One important consequence of the low likelihood that disturbances will initiate directional species replacements should be the precarious existence of early species dependent on disturbances that open sites for colonization. We examined constraints on disturbances, sites, and life histories that might increase colonization by early species. We explored how refuges and life history characteristics might reduce the likelihood of replacement of early species by late species, and we consider how spatial and temporal variation in disturbance effects might influence persistence of early species. In this way, we propose relationships between effects of non-catastrophic disturbance and the dynamic relationships between species that may be applicable across different ecological systems.