2.5 Early warning of tipping points in impacts

Summary

Tipping point research has traditionally focused on environmental systems, but there is increased interest in understanding whether the social and coupled social-environmental systems that are impacted by Earth system tipping points themselves exhibit characteristics of tipping points and whether they can be anticipated using early warning signals. While this question is highly relevant in a context of a changing climate, there are two major challenges in developing early warning systems for tipping points in social-environmental contexts: first, social systems may respond unpredictably  to changes in environmental conditions as they adapt to change; and second, datasets for social systems may not always be available for detection of tipping points. 

Evidence is emerging to demonstrate that social-environmental systems exhibit signals of tipping points through autocorrelation, skewness, variance and threshold exceedance. In food security early warning, lag-1 autocorrelation of soil moisture has demonstrated great utility in predicting transitions into and out of food crises up to six months ahead of a transition – with potentially transformative opportunities for humanitarian interventions. In grazing systems, higher variance of vegetation indices have been associated with changes in environmental conditions that lead to more degraded environments. Research has also demonstrated the exciting opportunities to leverage deep learning to detect tipping points in vaccine opinion using social data. Increasing availability of data from Earth observation, machine learning and social networks open up an unprecedented opportunity to improve early warning of tipping points in social-environmental systems.