Collapse, reorganization, and regime identity: Breaking down past management paradigms in a forest-grassland ecotone

Abstract

The identity of an ecological regime is central to modern resilience theory and our understanding of how systems collapse and reorganize following disturbance. However, resilience-based models used in ecosystem management have been criticized for their failure to integrate disturbance outcomes into regime identity. Assessments are needed to understand how well these classifications represent ecosystem responses that occur over management relevant time scales. We tracked post-wildfire forest and grassland dynamics 27 years after wildfire in eastern ponderosa pine savanna. We tested for differences between the assigned identity of a site (forest or grassland) versus classifications based on the site's disturbance history (burned/unburned and fire severity). Under current ecosystem models used to manage these forest-grassland ecotones, forests that experience high severity fire are expected to resemble an unburned grassland following fire, while forests and grasslands that experience low severity fire are expected to resemble unburned forests and grasslands, respectively. Twenty-seven years after wildfire, burned forests and grasslands displayed a high degree of departure from their expected regime identity. Plant and bird communities deviated significantly on sites that experienced low severity fire from undisturbed sites classified under the same ecological regime (grassland or forest). Forest sites that experienced high severity fire were the most unique of all disturbance history classes. Our results demonstrate that structures and communities predicted under resilience-based models used for eastern ponderosa pine management do not emerge over management relevant time scales following disturbance. Over 20% of variation in ecological structures and communities was explained by a single, 27-year-old disturbance. Integrating disturbance legacies will help improve applied models of ecosystem dynamics.