Assessing the impact of Atlantic Meridional Overturning Circulation collapse on the Amazon rainforest.

Abstract

Models and observations suggest that due to unprecedented stress induced by climate change and deforestation, parts of the Amazon rainforest might transition to a savanna-like state. In this thesis we investigate how changes in precipitation patterns induced by Atlantic Meridional Overturning Circulation collapse impact the vegetation cover in the Amazon rainforest region. To achieve this, precipitation is related to vegetation cover using a reaction-diffusion vegetation model. In this model rainforest fronts exist, which can become stationary on the domain for certain precipitation values due to mechanism referred to as wave-pinning. This stationary front represents the rainforest-savanna boundary. Using numerical bifurcation analysis with respect to the mean annual precipitation P_av and the dominant precipitation slope P_s over the spatial domain, the supported stable (front) steady states at each combination of these parameters are determined. Consequently, the possible vegetation states can be determined at a given precipitation field.

Next, we investigate the tipping dynamics of the vegetation model under changes in P_av and P_s . The results suggest that bifurcation-induced tipping is dominant, but under specific conditions rate-induced transitions from savanna to rainforest can occur before passing the bifurcation point. From this, a vegetation state classification framework is developed taking into account the supported vegetation states as well as the dynamics. This framework allows for the determination of the vegetation state after a change in precipitation.

In the final chapter, the developed classification framework is applied to Atlantic Meridional Overturning collapse precipitation model output obtained from the Community Earth System Model and the North Atlantic Hosing Model Intercomparison Project models under pre-industrial forcing. The classification framework suggests that nine out of nine models predict an increase in rainforest south of the equator, while eight out of nine models predict a decrease in rainforest north of the equator. This suggests that Atlantic Meridional Overturning Circulation collapse has a stabilising effect on the Amazon rainforest south of the equator, while it has a destabilising effect north of the equator.