Skip to main content
. 2012 May 16;9(75):2527–2539. doi: 10.1098/rsif.2012.0125

Figure 6.

Figure 6.

Early warning signals in simulated and empirical datasets. The first two columns are simulated data from (a) a stable system (stable), and (b) the same system approaching a saddle–node bifurcation (deteriorating). Empirical examples are from (c) Daphnia magna concentrations manipulated towards a critical transition (daphnia), and (d) deuterium concentrations previously cited as an early warning signal of a glaciation period (glaciation). Increases in summary statistics, computed over a moving window, have often been used to indicate if a system is moving towards a critical transition. The increase is measured by the correlation coefficient τ. Note that positive correlation does not guarantee that the system is moving towards a transition, as seen in the stable system, first column.