Early warnings are too late when parameters change rapidly

TL;DR

This paper investigates the limitations of early warning signals in predicting system tipping points when control parameters change rapidly, highlighting delays and dispersion issues that can hinder prevention efforts.

Contribution

It demonstrates the failure of EWSs in fast-changing conditions and proposes EWS-based control actions to rescue systems after tipping occurs.

Findings

EWSs often detect tipping after bifurcation points when parameters change quickly.

Delays and dispersion in warning signals reduce their effectiveness in real-world scenarios.

EWS-based control can rescue systems post-tipping in rapid parameter variation cases.

Abstract

Early warning signals (EWSs) forewarn a sudden transition (or tipping) from a desirable state to an undesirable state. However, we observe that EWSs detect an impending tipping past bifurcation points when control parameters are varied fast; this questions the applicability of EWSs in real-world systems. When a control parameter is changed at a finite rate, the tipping is also delayed, providing a borrowed stability (in the parameter space) before the system tips. In this study, we use the Hurst exponent as EWS in a thermoacoustic system - a horizontal Rijke tube. We find that upon receiving an EWS alert, a quick reversal of the control parameter within the region of borrowed stability cannot always prevent tipping in real-world systems. We show this failure is due to the (i) delay in receiving the EWS alert and (ii) dispersion observed in the warning points received. For fast variation of parameters, where preventive measures fall short, we demonstrate EWS-based control actions to rescue the system after tipping. Our results in a real-world system for a fast variation of parameter highlight the limits of applicability of EWSs in preventing tipping.