Place: 4274 Chamberlin (Refreshments will be served)
Speaker: Steve Carpenter, UW Center for Limnology
Abstract: Ecosystems occasionally undergo rapid massive changes - desertification, algae blooms of lakes, replacement of coral reefs by fleshy algae turf, trophic cascades, economic collapse of fisheries, and shrub invasion of rangelands are a few examples. Some regime shifts have big impacts on human life-support and are therefore important for environmental policy. Thus the detection and prediction of regime shifts has emerged as a research topic in basic and applied ecology. Theory shows that certain regular changes in time series should be measurable before an incipient regime shift - autoregression coefficients near one, variance spectra shifted to low frequencies, and rising variance, skewness and kurtosis, for example. However, not all regime shifts show these indicators, and some show opposite responses. While leading indicators show promise as tools for field science and ecosystem management, at present it is difficult to diagnose the characteristics of incipient regime shifts from the indicators alone. Field trials, and modeling to connect theory with the characteristics of particular ecosystems in the field, are needed to improve our understanding of these signals.