You are here:
Early Detection of Regime Shifts in Complex Systems from Fisher Information
Citation:
CABEZAS, H., T. EASON, AND A. GARMESTANI. Early Detection of Regime Shifts in Complex Systems from Fisher Information. Presented at The Second International Science and Policy Conference: Resilience 2011: Arizona State University, Tempe, AZ, March 11 - 16, 2011.
Impact/Purpose:
To inform the EPA
Description:
The central goal of sustainability is the maintenance of environmental conditions, which are favorable to human existence. A critically important element then is the resilience of the dynamic regime that one wishes to sustain. Resilient systems are able to withstand perturbations while maintaining function. However, it is possible for a system to reach a dynamic threshold and shift to another regime. Regime shifts have been demonstrated for a multitude of ecological and social systems, and often have significant ecological and economic consequences. Although much research has been focused on detecting regime shifts, from an adaptive management perspective, it is pertinent that regime shifts be identified before they occur. While rising variance, kurtosis, skewness, and critical slowing down have all been proposed as indicators for impending regime shifts in simple systems, these approaches typically do not signal the shift until it is well underway. Here, we propose the use of Fisher information as a method of detecting impending regime shifts. As a key method in information theory, Fisher information is suited for complex, multivariate systems. In this study, we compared several well-known indicators using model and real systems and found that Fisher information is a leading indicator of regime change in complex real systems. Thus, we believe this research has tremendous implications for social-ecological resilience, and, therefore, sustainability.