Catalogue of Tipping Points found in ESM simulations (Deliverable D2.1)

Exploring Tipping Points and Their Impacts Using Earth System Models Catalogue of Tipping Points found in ESM simulations Deliverable D2.1 1 D2.1 Catalogue of Tipping Points found in the ESM simulations About this document Submission date to the European Commission: 10 December 2025 Dissemination Level: Public (PU) Work package: WP2 Physical and biogeochemical TPs and their driving processes Authors: KONINKLIJK NEDERLANDS METEOROLOGISCH INSTITUUT (KNMI) PP4, Sybren Drijfhout sybren.drijfho[email protected] Contributors: KONINKLIJK NEDERLANDS METEOROLOGISCH INSTITUUT (KNMI) PP4, Joran Angevaare and Pouriya Alinaghi CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE (CNRS-IPSL, CNRS-EPOC, and Affiliated entities: CEA, IRD), PP3, Eike E. Köhn, Lester Kwiatkowski MET OFFICE (METO), PP9, Andy Wiltshire UNIVERSITY OF LEEDS (UNIVLEEDS), PP8, Colin Jones Reviewer: Danish Meteorological Institute (DMI), PP1, Chuncheng Guo and Chiara Bearzotti [email protected]; Disclaimer: Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Climate, Infrastructure and Environment Executive Agency (CINEA). Neither the European Union nor the granting authority can be held responsible for them. 2 D2.1 Catalogue of Tipping Points found in the ESM simulations TABLE OF CONTENTS 1. Abstract 4 2. Work Done 4 2.1 Catalogue on physical SNS in ocean-sea ice-atmosphere in CMIP6 4 2.3 Analysing the idealised runs 5 3. Results Achieved 5 3.1 SNS in the coupled ocean - sea ice - atmosphere system 5 3.2 SNS in ocean biogeochemistry and land surface processes 6 3.3 Hysteresis and reversibility in idealised runs 7 4. Open Access 8 5. Contribution to the TipESM objectives 8 6. References 8 3 D2.1 Catalogue of Tipping Points found in the ESM simulations 1. Abstract A set of algorithms has been developed to detect abrupt changes over decadal timescales, including tipping points and rapid transitions. This set was used to detect very Strong Nonlinear Surprises (SNS-events) using fully automated scripts, based on several stringent criteria. Thereafter, the Coupled Model Intercomparison Project Phase 6 (CMIP6) archive was analysed for the occurrence of SNS in future climate-change projections. Two different types of SNSs were defined: abrupt changes and slower state transitions, too significant to be explained by the forcing without invoking strong internal feedbacks in the climate system. Firstly, data from 54 models were analysed across five shared socio-economic pathways for ocean, sea ice, and atmospheric variables. The algorithm isolated regions of at least 106 km2 and utilised stringent criteria to select SNS events. In total, 73 SNSs were found, categorised into four abrupt-change and seven state-transition categories. Thereafter, the method was used to identify SNS in ocean biogeochemistry, vegetation and land surface hydrology. For 2 of the 4 European models, idealised runs following the TIPMIP ESM protocol used in TipESM were also analysed. Uploading data of the idealised runs was somewhat delayed and not yet available for the other models. In addition to detecting SNS, hysteresis was found for sea ice, land ice, ocean variables, and the Amazon rainforest. Also, significant transitions in weather extremes are expected during the ramp-up phase. These results will be further investigated in WP3. 2. Work Done 2.1 Catalogue on physical SNS in ocean-sea ice-atmosphere in CMIP6 A set of algorithms has been developed to detect abrupt changes, including tipping points and rapid transitions. This set was used to detect very Strong Nonlinear Surprises (SNS-events) using fully automated scripts, based on several stringent criteria. Thereafter, the Coupled Model Intercomparison Project Phase 6 (CMIP6) archive was analysed for the occurrence of SNS in future climate-change projections. Two different types of SNSs were defined: abrupt changes and slower state transitions, too large to be explained by the forcing without invoking strong internal feedbacks in the climate system. 2.2 Catalogue on SNS in ocean biogeochemistry and land surface processes 4 D2.1 Catalogue of Tipping Points found in the ESM simulations The same workflow was applied in a second iteration to ocean biogeochemistry (OBGC) variables and land-surface variables. In many cases, we found that state transitions were more clearly visible in yearly maxima/minima, especially for variables that show very strong seasonality. State changes in vegetation were nearly always connected to changes in hydrology and should be considered as a set of physically coupled variables. Also, soil moisture changes were either coupled to surface hydrology or, in the case of changes in frozen water content, to temperature changes, where seasonality had to be included. 2.3 Analysing the idealised runs The workflow for SNS has not yet been applied to the idealised TipESM runs, as some of the data is still not available. We will do this in 2026 under WP3 task 3.1. In the meantime, analysis of the idealised runs performed by UKESM1.2 and EC-Earth3-ESM has started focusing on hysteresis effects and state transitions in extreme weather. These idealised runs from both models followed the TIPMIP ESM experiment protocol, which specifies an emissions-driven warming trajectory of ∼0.2 K per decade. Zero CO2 emission runs were then branched off the ramp-up run on January 1st of years directly after the year in which the 31-year mean global mean surface air temperature (GMSAT) exceeds 1.5 K, 2 K, 2.5 K, 3 K, 4 K, 5 K, and 6 K above the equivalently-averaged preindustrial value centered on the start time of the ramp-up run. Zero-emission simulations were continued for∼500 years. 3. Results Achieved 3.1 SNS in the coupled ocean - sea ice - atmosphere system In the catalogue for SNS in the coupled ocean-sea ice-atmosphere system within the CMIP6 archive, 73 SNS were found that were categorised into four abrupt change and seven state transition categories. Of the identified SNSs, 45% relate to sea-ice cover, 19% to ocean currents, 29% to mixed-layer depth, and 7% to atmospheric systems like the Intertropical Convergence Zone. For each category, probability density functions for time-windows of maximal change indicate SNS occurring earlier and at lower global temperature rise than previously assessed (e.g., Armstrong McKay et al., 2022), in particular SNS associated with winter Arctic sea ice disappearance, North Atlantic deep mixed-layer collapse, and subsequent transition of the Atlantic Meridional Overturning Circulation (AMOC) to weak states without North Atlantic Deep Water formation involved. This catalogue emphasises the possibility of SNS events already below 2°C of global warming, even more than the previous assessments based on CMIP5 data (Drijfhout et al., 2015), 5 D2.1 Catalogue of Tipping Points found in the ESM simulations For Arctic sea ice, there is a clear shift of the onset of a year-round ice-free Arctic to lower temperature increases, implying this could occur already at the end of the 21st century. This shift is associated with reduced bias in CMIP6 sea-ice thickness during the historical period. A collapse of convective deep mixing also occurs faster and earlier in CMIP6 than in CMIP5 and may already develop between now and 2050. One reason for this is likely a reduced stratification in the North Atlantic Subpolar Gyre (SPG). It is nontrivial to assess the role of model bias in each individual case, but in general, model bias is stronger in models that do not show a collapse than in models that do. Associated with the mixed-layer collapse, the AMOC features earlier shutdowns (a collapse of the North Atlantic Deep Water cell, leaving a residual largely wind-driven overturning that does not extend northward of 30°N. In some cases, the shutdown even manifests in an SSP126 scenario that complies with the Paris Agreements. We also showed that a collapse of deep mixing in the SPG precedes the AMOC shutdown and is a clear example of cascading tipping points. The interaction between deep convection, regional sea-ice cover, air/sea fluxes, and changes in ocean and atmospheric surface temperature, sea surface salinity, also shows elements of a cascade, but we would argue these are not separate tipping points but a cascade of tipping variables within a common tipping point/element. 3.2 SNS in ocean biogeochemistry and land surface processes The OBGC catalogue describes large-scale declines in chlorophyll, plankton carbon concentration, oceanic dissolved nitrate and primary production, driven by the collapse of the AMOC and convective deep mixing. We see increases of mixed-layer depths in the North Atlantic subtropics and Arctic, where they are associated with the decrease in sea-ice cover. This increase is followed by a sharp decrease in some models, depending on timing and extent of AMOC reduction. In some models, weakened upwelling causes a collapse of nutrients along the west coast of South America. In the catalogue of land surface processes, we identified four cases in which parts of northern South America, including the Amazon Basin, undergo dramatic vegetation shifts. In a high-emission scenario, the models project a strong decline in leaf area index across the Amazon basin. This decline is primarily driven by a reduction in tree fraction, which decreases till the year 2300 to roughly half its value in the historical period. The spatio-temporal evolution of precipitation over the region suggests that reduced rainfall over the Amazon may be driving the observed forest dieback, likely triggered by changes in large-scale circulation that reduce moisture transport into the region. A similar story holds for the Congo Basin, where trees are replaced by grasses in response to declining precipitation and evapotranspiration. In Ethiopia and the southern Sahel (north Nigeria and Southern Chad), bare soil is replaced by grass. A similar decrease in bare 6 D2.1 Catalogue of Tipping Points found in the ESM simulations soil, but now replaced by trees, occurs over northern China, Mongolia, and South of the Himalaya. Also, north of the Arctic Circle, we see large-scale growth of Boreal Forests. 3.3 Hysteresis and reversibility in idealised runs The collapse of the North Atlantic subpolar gyre deep mixing, and phytoplankton bloom is found to be reversible under zero emissions in the UKESM1.2, but only when the temperature increase does not exceed 1.5-2°C, and reversible under negative emissions when global warming does not exceed 5°C. However, recovery timescales, on which conditions in the SPG return to near their pre-industrial state, are projected to take on the order of centuries to millennia. Under zero emissions achieved at and above 2.5°C of global warming, both nutrients and chlorophyll continue to decline, and therefore, the reversibility of the bloom decline cannot be achieved. A global warming threshold of 4.1°C ± 1.3°C is found to tip the Ross Sea from cold to warm water conditions. Tipping is preceded by an extended period of continental shelf freshening, which is already being observed. Cooling the climate through negative carbon emissions eventually brings the ocean conditions back to their original cold state, but the impact on the Antarctic Ice Sheet and global sea level is irreversible on multi-centennial timescales. Amazon tree fraction is rising throughout the ramp-up as cumulative emissions and atmospheric CO2 concentration increase, promoting tree growth. This effect slows until tree fraction peaks at around 1500 GtC of emitted CO2, approaching 4°C of global warming, after which tree fraction declines. Reversibility of the Amazon tree fraction during the positive emission (ramp-up), zero emission (warming stabilisation), and negative emission (ramp-down) is sensitive to the global warming level reached, the timing of initialisation of negative emissions, as well as the negative emission rate. In 1pctCO2-cdr runs the Arctic becomes a hotspot for delayed reversibility in ocean acidification. It is projected to become undersaturated with respect to aragonite at 500 ppm during the ramp-up, but only returns to supersaturated conditions when atmospheric CO2 levels fall below 385 ppm. The acidification hysteresis mainly derives from the large initial undersaturation of the surface Arctic with respect to dissolved inorganic carbon, which is gradually lost during ramp-up as summer sea ice retreats but is not restored during ramp-down as sea ice grows back. 7 D2.1 Catalogue of Tipping Points found in the ESM simulations Also in the Arctic Circle, in the ramp-up phase of idealised TIPESM runs, a sudden increase in annual precipitation occurs, coinciding with sudden increases in wet days and minimum yearly temperature, and with a sudden increase in ocean surface temperature, probably related to a decrease in sea-ice cover. 4. Open Access Three catalogues with TPs have been published on GitHub, in compliance with open access and open science requirements and with the obligations of the data management plan of TipESM: https://github.com/TipESM-EU 5. Contribution to the TipESM objectives This deliverable contributes to achieving the following project objective. SO2 To identify climate TPs in the CMIP6 archive and dedicated runs from TipESM (SO1) and previous EU projects, and improve our understanding of the processes underpinning such TPs and their impacts across the Earth system. KPIs: A new catalogue of the likelihood of crossing TPs as a function of global warming, combined with an assessment of the driving processes and the impact of each tipping event on the Earth system. (Deliverables: D2.1, D2.2 and D2.3). With this deliverable and the three catalogues available in open access, it helps identify the tipping points and improves understanding of the processes behind them. Several extended analyses have been submitted to scientific journals or have already been published; they are all reported in the list below. Ref: WP1 and WP2 6. References Angevaare, J. R. and Drijfhout, S. S.: Catalogue of Strong Nonlinear Surprises in ocean, sea-ice, and atmospheric variables in CMIP6, EGUsphere, https://doi.org/10.5194/egusphere-2025-2039 , 2025. 8 D2.1 Catalogue of Tipping Points found in the ESM simulations Armstrong McKay, et al.: Exceeding 1.5 C global warming could trigger multiple climate tipping points, Science, 377, eabn7950, 2022. Drijfhout, S., et al. B: Catalogue of abrupt shifts in Intergovernmental Panel on Climate Change climate models, Proceedings of the National Academy of Sciences, 112, E5777–E5786, 2015. Drijfhout, S.S. et al 2025 Shutdown of northern Atlantic overturning after 2100 following deep mixing collapse in CMIP6 projections. Environ. Res. Lett. 20 094062 DOI 10.1088/1748-9326/adfa3b Gibbs, L. et al 2025 Zero Emissions Commitment depends on warming level. Submitted to Nature Geosciences. Köhn, E. E. et al., Persistence of Arctic acidification under negative emissions. Submitted to Nature Climate Change Oliver, S. et al. The North Atlantic sub-polar gyre and phytoplankton bloom under potential future climate scenarios. ESS Open Archive. 2025. DOI: 10.22541/essoar.175855713.32401547/v1 Smith, R. S., et al.: Response of ice sheets, sea-ice and sea level in climate stabilisation and reversibility simulations using a state-of-the-art Earth System Mode, https://doi.org/10.5194/egusphere-2025-4476 , 2025. 9