Although many ecological changes are gradual and reversible, some extensive changes may occur more abruptly when critical thresholds are crossed. Familiar examples are desertification, shifts from grassland to woody vegetation, or collapse of living resources such as fish populations. These unusually large changes, termed "regime shifts", may be irreversible or reversed only at great cost. Scientists are not yet able to predict when a regime shift is likely to occur in a given ecosystem. The ability to make such predictions could allow managers to intervene before the change occurred. Theory from diverse disciplines suggests that regime shifts are preceded by observable increases in variability of system components. In addition to increasing variance, theory predicts that variance moves to longer periods or lower frequencies ("reddening") while the rate of recovery from perturbations declines ("critical slowing down").

Our project is a whole-ecosystem experimental test of the theory that 1) increased variance, 2) red-shift of variance and 3) critical slowing down of recovery rate across components of a food web are leading indicators of a common type of regime shift in lake ecosystems caused by changes in the structure of the fish community. Variability will be assessed at all trophic levels for three years in both a manipulated lake and a non-manipulated reference lake using high-frequency data acquisition methods including automated in situ instrumentation for analysis of alagal biomass and primary production, zooplankton abundance and size structure, acoustical measure of fish migration, and daily measures of fish growth from otoliths. Modern signal processing methods of time series, spectral and wavelet analysis will evaluate changes in variance, variance sprectra and return rate.

Regime shifts have been identified as a frontier in ecosystem research because they can cause losses to key ecosystem services that can sometimes have severe consequences for human well being. Expanding human demand for ecosystem services makes regime shifts more likely, but does not make specific regime shifts predictable. This project is a first large-scale experimental test of a theory and method to anticipate such changes before they happen.

Website maintained by: Laura Smith
Last update: 12 March, 2010
Comments/Questions?