The concept that ecosystems do not respond smoothly to changing pressures has existed for the last 30 years (May 1977), but the analyses of ecological tipping points and regime shifts remain conceptual in nature (Scheffer et al. 2001, Lenton et al. 2008). In the same period statistical frameworks for analysing such phenomena have been developed in other disciplines, such as climate research and econometrics (Zeileis et al. 2003, May et al. 2008), and their use in ecology has been promoted recently through the THRESHOLDS integrated project (FP6, contract 003933-2, Andersen et al. subm.). As a theoretical framework to explain regime shifts improves (Steele 2004), it becomes apparent that regime shifts are triggered by exceeding a particular critical threshold in the driving variable (May 1977). One example is how changes in climate drivers can force marine ecosystems into a new structural and functional state. Regime shifts are often persistent in that reverting the ecosystem to the original state often requires far greater changes in the driver than what prompted the initial shift. In the worst cases, reversion may prove impossible over managerial time scales. Anticipating where these tipping points are is critical in setting targets to conserve marine ecosystems (Groffman et al. 2006). Unfortunately, most research on regime shifts has been retrospective in nature, focussing on the their detection and analysis after tipping points have been passed (e.g. Mantua 2004). The prediction of ecological thresholds remains elusive, a major bottleneck for the applicability of the concept.