2.2.3.1 Amazon dieback

The potential for a tipping point in the Amazon – also known as ‘Amazon dieback’ – relates to the close coupling between the land ecosystem and the atmosphere, with the rainforest playing an important role in maintaining precipitation (and hence soil moisture) at levels sufficient to support rainforest (Betts, 1999).  There is recycling of rainfall from eastern to western part of the Amazon basin (Zemp et al., 2017), so loss of forest in the east could exert further impacts in the west. If forest cover were to be sufficiently reduced, either due to direct, human-induced deforestation or the impacts of climate change (or, more likely, a combination of both), there is the potential for the regional climate to move to an alternative state in which rainforest can no longer be supported, which would prevent the future return of forest and potentially further increase the loss of forest (Hirota et al,. 2021).

Reduced forest cover in the Amazon is also observed to lead to higher temperatures, particularly daily maximum temperatures, both locally at the site of forest loss and in adjacent regions up to 60 km away, due to reduced transpiration, decreased aerodynamic roughness causing reduced dissipation and weakened horizontal transport of heat (Cohn et al., 2019). A drier, hotter climate would lead to an increase in wildfire and soil erosion, which could lead to an expansion of savanna vegetation at the expense of rainforest (Flores and Holmgren, 2021; Flores et al., 2020).  

Passing a tipping point in the Amazon would therefore lead to impacts in the immediate region, and could potentially also lead to impacts elsewhere by influencing moisture transport into and/or out of the region (including via the South American monsoon, Boers et al., 2017) and by altering large-scale atmospheric circulation patterns with potential teleconnections to distant parts of the world such as to the Tibetan Plateau (Liu at al,. 2023).

In the Amazon Assessment Report 2021 (Science Panel for the Amazon, 2021), the chapter on assessing the risk of tipping points concluded: “Local-scale forest collapses could trigger cascading effects on rainfall recycling, intensifying dry seasons and wildfire occurrence, and leading to massive forest loss at continental scales, particularly in the southwest of the basin” (Hirota et al,. 2021).

Loss of the forest would have substantial impacts on biodiversity, and the reduced evapotranspiration would lead to reduced precipitation and hence reduced water availability, with potentially large societal impacts. Loss of the forest would also lead to increased high temperature and greater risk of heat stress.

Amazon dieback would be a major threat to the biodiversity of the rainforest (Gomes et al., 2019; Esquivel-Muelbert et al., 2017). Four potential alternative states to the current closed-canopy primary rainforest have been identified as possible consequences of passing an Amazon tipping point: (i) a closed-canopy seasonally dry tropical forest state; (ii) a native savanna state; (iii) an open-canopy degraded state; and (iv) a closed-canopy secondary forest state (Hirota et al., 2021). Clearly, any of these would have major implications for species of rainforest trees and other plants.

They would also impact animal species, many of which could disappear from the system if they are not favoured by open habitats and their movement becomes restricted by loss, degradation or fragmentation of forest (Barlow et al., 2016; Laurance et al., 2004). Seed dispersal by fruit-eating species may become limited if such species avoid open disturbed habitats, thus reducing tree recruitment and forest regrowth, especially where disturbances are most severe (Turner et al., 1998). Studies in the tropical Atlantic Forest indicate that 30 per cent tree cover is a threshold in which many forest-adapted animal species are replaced by disturbance-adapted species (Banks-Leite et al., 2014).

The health and wellbeing of people in the Amazon region would also be put at increased risk by forest loss. Wang et al., (2023) suggested that increased wildfire frequency and severity associated with Amazon die-back would put regional communities at risk and lead to increased air pollution. Moreover, heat stress is extremely dangerous to humans, increasing the risk of heat-related illnesses and death, especially for vulnerable groups such as children, the elderly and those with underlying health conditions (de Oliviera et al., 2020), and for other exposed groups such as those working in conditions of extreme heat (Spector et al., 2019). The risks of heat stress on humans and other mammals are projected to increase with global warming (Bezner-Kerr et al., 2022;Cissé et al., 2022), and since tropical forests maintain lower temperatures compared to deforested land due to higher levels of evaporation (Ruv Lemes et al., 2023), forest loss following an Amazon tipping point could increase heat stress risks further. De Oliviera et al., (2021) used the BESM-OA2.5 climate model to project the impact on human heat stress risk of a total replacement of tropical Amazon forest with savannah in two climate change scenarios. Heat stress risk was quantified using Wet Bulb Globe Temperature (WBGT) which accounts for the effects of both temperature and humidity on heat stress risk, with high humidity increasing the risk of heat stress as it reduces the body’s ability to cool through sweating. In a scenario reaching approximately 2.5°C global warming by the end of the 21^st century, average daily WBGT values in the hottest month were 30-31°C (high heat stress risk) across most of Amazonia with intact forest. These were elevated to 34-37°C (extreme heat stress risk) when forest was replaced by savannah (Figure 2.2.5), exposing more than 6 million people to extreme heat stress risk.

Extreme events including droughts in the Amazon region are disruptive to the food and transport systems of Indigenous peoples and communities who depend on local resources (Pinho et al., 2015).

Interactions between the Amazon forest and the atmosphere via the water cycle play a crucial role in the impact of forest loss on river flows, with potentially major implications for socioeconomic impacts. Importantly, although land ecosystem-hydrology models that do not account for feedbacks with the atmosphere project forest loss to increase river flows due to reduced evaporation, the opposite is projected when vegetation-atmosphere interactions are considered – reduced precipitation arising from widespread decreases in evaporation are projected to lead to reduced river flows (Stickler et al., 2013). Lapola et al., (2018) suggest that lower river water levels resulting from Amazon dieback would affect transportation, food security and health, which ultimately may influence migration from rural areas to large Amazonian cities. In a coupled climate-vegetation model and hydrology model with a potential 40 per cent decline in forest cover by 2050, river discharge in the Xingsu basin was projected to decrease by 6-36 per cent, leading to hydrological power generation to fall to approximately 25 per cent of maximum installed capacity (Stickler et al., 2013).

Lapola et al., (2018) estimate that Amazon dieback would lead to economic damages of between $US957bn and $US3,589bn (net present value as of 2018) over 30 years, mainly due to changes in the provision of ecosystem services (Figure 2.2.6). For comparison, the Gross Brazilian Amazon Product is approximately $US150bn per year.

Amazon dieback could magnify global warming and its associated impacts by accelerating the rise in atmospheric CO₂. The Amazon is estimated to contain between 150 and 200 GtC in biomass and soil organic matter (Wang et al., 2023), equivalent to about 15 to 20 years of current global anthropogenic CO₂ emissions.  As an estimated upper bound of the contribution of the potential Amazon tipping point to the magnification of global climate change impacts, Betts et al., (2008) projected global warming by 2100 to be increased by 0.3°C in an extreme scenario of total Amazon forest die-back in which forest was almost entirely replaced by either grassland or desert.