Several key messages emerge from this section. Firstly, there is considerable evidence for tipping points existing in many parts of the Earth system (see Table 1.7.1 for summary). Several tipping points are likely in the cryosphere, at a large scale in ice sheets and on a more local scale in permafrost and glaciers. In the biosphere, evidence for regime shifts and tipping points exist in many ecosystems such as in tropical forests, savannas, drylands, lakes, coral reefs and fisheries, and are often spatially complex. Tipping points in ocean circulation and monsoons are also likely to exist, but the proximity of their thresholds are subject to high uncertainty. In contrast, some other suggested ESTPs have been assessed as unlikely in this report, including for Arctic sea ice, global-scale permafrost or glacier tipping, some types of lake ecosystem tipping, tropical clouds and climate sensitivity, and the El Niño–Southern Oscillation (ENSO).
Secondly, we know that we could already be very close to some Earth system tipping points. Several cryosphere tipping points cannot be ruled out at 1.5oC of global warming, which will be reached even with aggressive mitigation. These tipping points become likely beyond the 2oC of warming that the Paris Agreement commits countries to stay well below, but which current policies are likely to substantially overshoot (Climate Action Tracker, 2022; Meinshausen et al., 2022; IEA, 2023). In the biosphere, deforestation in the Amazon combined with climate change-induced drying could lead to regional dieback, some drylands are close to degradation tipping points, and coral reef die-off is already occurring in many regions. In the North Atlantic Ocean, convection in the Labrador and Irminger Seas could collapse within Paris Agreement warming levels of well below 2oC, with severe impacts across the North Atlantic region. Early warning signals indicate that several systems, such as parts of the Greenland Ice Sheet, Atlantic meridional overturning circulation (AMOC) and the Amazon rainforest may be losing resilience, which could mean their tipping points are approaching (but exactly when is uncertain).
Thirdly, complex and sometimes uncertain interactions between tipping drivers, components of the Earth system and key feedbacks make tipping dynamics difficult to assess for some systems. For example, parts of the Amazon rainforest could die back as a result of climate change-driven drying as well as direct deforestation and degradation, but while their combination makes tipping likely sooner, the thresholds for the combined effect of these processes is difficult to estimate. For many of the systems considered, key feedbacks and processes that could be involved in tipping are not well understood and so are either represented simplistically or left out of models (such as fire feedbacks, land use change, and spatial variability in the Amazon), making future projections more uncertain. Short observational records for some systems make early warning signals less reliable as well. Many tipping systems closely interact through the climate, and evidence – particularly from palaeorecords – suggests that most interactions mean one system tipping it tends to destabilise connected systems. Model limitations mean there are large uncertainties around these potential tipping cascades, but as warming approaches the levels where some key tipping points become likely, the possibility of cascades is a growing risk that requires new approaches to assess.
Together, this evidence provides strong motivation for both rapidly reducing human-driven pressures on the Earth system, from eliminating greenhouse gas emissions (GHG) and deforestation to increasing social-ecological resilience (see Section 3) and preparing adaptation plans for the societal impacts of Earth system tipping points (see Section 2) should some tipping points occur despite mitigation efforts.
Equally, uncertainties around tipping point thresholds and interactions makes mitigation even more critical, as we cannot rule out tipping happening sooner than we currently expect.