How far can we trust climate change predictions?

TL;DR

This paper argues that due to the complex and sensitive nature of climate systems, numerical models cannot reliably predict tipping points, and observed instabilities suggest an imminent irreversible shift to a new climate state if drastic measures are not taken.

Contribution

It highlights the limitations of current numerical climate models in predicting critical tipping points and emphasizes the importance of observed instabilities as indicators of imminent irreversible change.

Findings

Numerical models cannot reliably detect approaching climate tipping points.

Observed climatic instabilities indicate an imminent critical transition.

Achieving 1.5°C or 2°C warming targets by 2030-2050 is unlikely without drastic emission reductions.

Abstract

Current techniques for predicting climate change are mainly based on "massive" deterministic numerical modeling. However, the ocean-atmosphere system is a so-called "complex system", made up of a large number of interacting elements. We show that, in such systems, owing to the particularly large sensitivity to initial conditions, the approach of a possible tipping over a critical point cannot be evidenced "by construction" using numerical modeling, due to the divergence of computation time in the vicinity of the tipping point. On the other hand, the increasing amplitudes of observed climatic instabilities seem to be an obvious sign of the approach of such a tipping point, easily interpreted as a "critical softening", well known in the theory of dynamical systems, that would bring us irreversibly into a new and totally unexplored equilibrium state, except for a significantly higher temperature and in a much closer time than expected from numerical modeling extrapolations. Thus, maintaining climate warming around 1.5$^o$C or 2$^o$C by 2030 or 2050 appears fairly unrealistic unless worldwide drastic green house gases reduction measures are immediately taken and applied.