Analysis of abrupt transitions in ecological systems

• Main Authors: Bestelmeyer, Brandon T., Ellison, Aaron M., Fraser, William R., Gorman, Kristen B., Holbrook, Sally J., Laney, Christine M., Ohman, Mark D., Peters, Debra P. C., Pillsbury, Finn C., Rassweiler, Andrew, Schmitt, Russell J., Sharma, Sapna
• Format: Article in Journal/Newspaper
• Language: unknown
• Published: Figshare 2016
• Subjects: Environmental Science; Ecology; FOS Biological sciences; Pacific; Sea ice
• Online Access: https://dx.doi.org/10.6084/m9.figshare.c.3308247.v1; https://figshare.com/collections/Analysis_of_abrupt_transitions_in_ecological_systems/3308247/1

The occurrence and causes of abrupt transitions, thresholds, or regime shifts between ecosystem states are of great concern and the likelihood of such transitions is increasing for many ecological systems. General understanding of abrupt transitions has been advanced by theory, but hindered by the lack of a common, accessible, and data-driven approach to characterizing them. We apply such an approach to 30–60 years of data on environmental drivers, biological responses, and associated evidence from pelagic ocean, coastal benthic, polar marine, and semi-arid grassland ecosystems. Our analyses revealed one case in which the response (krill abundance) linearly tracked abrupt changes in the driver (Pacific Decadal Oscillation), but abrupt transitions detected in the three other cases (sea cucumber abundance, penguin abundance, and black grama grass production) exhibited hysteretic relationships with drivers (wave intensity, sea-ice duration, and amounts of monsoonal rainfall, respectively) through a variety of response mechanisms. The use of a common approach across these case studies illustrates that: the utility of leading indicators is often limited and can depend on the abruptness of a transition relative to the lifespan of responsive organisms and observation intervals; information on spatiotemporal context is useful for comparing transitions; and ancillary information from associated experiments and observations aids interpretation of response-driver relationships. The understanding of abrupt transitions offered by this approach provides information that can be used to manage state changes and underscores the utility of long-term observations in multiple sentinel sites across a variety of ecosystems.