Are early warning signals present for climate tipping points detected in CMIP6?

Many elements of the climate system are believed to be at risk of tipping in the near future due to ongoing climate change. Abrupt shifts or tipping points have found to be prevalent in several of the latest generation of climate models (CMIP6) under a range of future emission scenarios. However, by observing the time series alone it is notoriously difficult to predict an upcoming tipping point. Therefore, so-called early warning indicators are needed to try to forewarn of an approaching tipping point. Two commonly used early warning signals, designed to detect critical slowing down prior to the tipping point, are to observe an increase in autocorrelation and variance. In this presentation, we assess the reliability and performance of these indicators for a range of tipping points, scenarios and models. In examples of the indicators performing poorly, we consider the potential for system specific indicators.