Introduction

The Amazon forest could reach a tipping point of large-scale dieback in the coming decades. Understanding the actual likelihood of this tragedy requires addressing the various factors shaping Amazonian resilience. The region is undergoing a major regime shift. Ancient stabilizing feedbacks are now weakening and being replaced by novel ones that increase the risk of abrupt shifts.

Objectives

Grounded on the theory of complex dynamical systems, we assess existing evidence for tipping points in the Amazon forest system, as well as the most probable trajectories following ecosystem shifts.

Methods

We review the scientific literature and compile evidence of tipping points from paleo-ecological studies, current observations and modelling. We discuss how the inherent complexity of the system may add uncertainties about the risk of tipping points, but also reveal opportunities for action.

Results

We identify four potential large-scale tipping points in the Amazon forest system involving temperature, rainfall and deforestation as external drivers. We also identify three common ecosystem trajectories associated with distinct disturbances and stabilizing feedbacks. Most likely, ongoing deforestation may cause large parts of the Amazon to shift into an open-canopy degraded state, trapped by fires. Such local ecosystem shifts may become contagious and cause a large-scale forest dieback, but this depends on a combination of mechanisms. In particular, spatial heterogeneity and connectivity play important roles in shaping forest resilience by affecting whether the system will collapse abruptly or gradually; i.e. whether tipping points in fact exist.

Conclusions

We conclude that the very same mechanisms that kept Amazonian forests resilient for millennia are now disappearing; heterogeneities that reduce disturbance contagiousness, ancient connectivities that promote forest recovery, and the biological and cultural diversities that increase forest adaptability. Now, these pillars of Amazonian resilience are being lost. Existing evidence indicates that the most important actions to avoid tipping points are ending deforestation and forest degradation. Nonetheless, climate change will continue to intensify, implying that keeping the Amazon resilient may depend on our capacity to control global greenhouse gas emissions. Decisions regarding the future of this iconic system must involve local and global actors, bridging the ancient ecological knowledge of indigenous and local peoples with updated science to promote its social-ecological resilience to upcoming events.