When the Atlantic Stalls HR

ClimateNorth Atlantic

When the Atlantic Stalls

The Atlantic Meridional Overturning Circulation (AMOC) carries heat north from the tropics, helping keep Europe mild. Measurements point to weakening, and a growing number of studies warn that it could collapse this century. Explore what scientists know and where uncertainty remains.

Change the strength of the current
  • warm surface water flowing north
  • cold deep water flowing south
  • sinking regions
  • monitoring arrays

Loading land outlines for the globe…

Schematic display: the paths are simplified and the speed of the particles follows the selected strength.

What does a weaker current look like?

Move the slider or choose a measurement or scenario.

Circulation strength at 26.5° N

17.0Sv

≈ 14 times the combined flow of all the world’s rivers

The diagram starts at 17 Sv.

01How it works

A conveyor carrying 17 million cubic metres per second

In the upper kilometre of the Atlantic, warm, salty water flows north in the Gulf Stream along the Americas and in the North Atlantic Current towards Ireland, Norway and Iceland. It releases heat into the atmosphere, cools and becomes dense enough to sink in winter. At depths of one to four kilometres it flows back south to the Southern Ocean, where winds help bring it back to the surface.

This circle is called the reverse circulation because the water in it literally changes direction and depth. At 26.5° north latitude, where it has been measured since 2004, is transmitted on an average of 17 wells (17 million cubic metres of water per second) and about 1.2 petvats of heat, corresponding to approximately one million power plants.

Figure 1. Schematic Atlantic section from Antarctica to the Nordic seas, showing latitude and depth. Dotted vertical lines mark the SAMBA (34.5°S), RAPID (26.5°N) and OSNAP (roughly 53–60°N) monitoring arrays. Select a step to see how it works.
17.0 Sv
mean strength at 26.5° N, measurements RAPID 2004.–2024.
1.2 PW
Heat current to the north, approximately a million power plants
≈ 15 %
weakening from mid-20. century according to reconstructions (Caesar et al., 2018.)
−0.9 Sv
linear trend over the decade in the series RAPID; with the correction of the Florida current −0.8 ± 0.7

Golf current is not the same as AMOC. Most of the Golf currents are driven by the wind in the subtropic vortex and it would exist without the AMOC. AMOC is the part of the current that does not return to the vortex, but continues to the north, sinks and returns to the depths. When the media writes about the "golf current break", they think of the break of the AMOC.

02Tipping point

Why the circulation could collapse abruptly

AMOC helps sustain itself. It brings salty water north, where salt makes the water denser and the denser water sinks, driving the circulation. If the North Atlantic becomes too fresh, the water becomes less dense and sinks more slowly. A weaker circulation then brings in less salt, accelerating the weakening.

Because of this feedback, there is a threshold. Oceanographer Henry Stommel described it as 1961. model of only two boxes of water. The model is extremely simplified, but the same behavior (two stable states and sudden transitions) also show far more complex climate models.

Strength 85 %Influence 0.128

Possible balance states

Strength through time and early warning signs

Image 2. The Stommel model with two boxes in a dimensional form (according to Cessi, 1994.), with a random "time" forest. Left: the full line is stable (the top is today's, the bottom is "included" AMOC, the bottom is collapsed), the broken state is unstable. Right: strength through the time and the variation and autocorrelation in the sliding window, with the removed linear trend. Numbers are illustrative and are not forecast.

The experiment shows two things. First, as the flow approaches the turning points, fluctuations become bigger and slower. It is called critical slowdown, and precisely such signs (growth of variation and autocorrelation) Niklas Boers (2021.) and Peter and Susanne Ditlevsen (2023.) are looking for in real data on sea temperature.

Second, after the crash, it is not enough to restore the supply to the old. The circulation remains down until the flow is reduced far below the threshold. This phenomenon is called histeresis and the reason why the crash is considered virtually irreversible on human time scales. In the Niana and Collaborators model (2026) the collapsed AMOC does not recover while the concentration of CO2 is above about 350 ppm and today it is about 430 ppm.

In the simulations of van Westen, Börner and Dejkstra (2026) AMOC collapsed at +2 °C global warming at today's rate of CO2 growth. Five times slower growth remained stable at +5,5 °C, as the ocean had time to adjust. The critical speed in this model is about 0,3 °C per decade, which is close to today's melting rate.

03Measurements

What the measurements show

Since April 2004, the British–U.S. RAPID array has measured AMOC across the Atlantic at 26.5°N. It uses a cable on the floor of the Florida Straits, moored instruments from the Bahamas to Africa and satellite wind measurements. Twenty years is short in climate terms, but the record shows how variable the circulation is and that it has weakened on average.

Image 3. The strength of AMOC at 26.5° N, April 2004 – March 2024. The thin line shows monthly averages, the thicker 12-month sliding average, and the interrupted linear trend. The bottom coloring band is the range where the break is usually referred to. Source: RAPID-MOCHA-WBTS v2024.1a (Moat i sur., 2026.); monthly averages were calculated from 12 hour values.
Annual Table Averages

AMOC at 26.5°N combines three flows. The Florida Current carries an average of 31.7 Sv northward, wind-driven Ekman transport adds 3.7 Sv, and 18.4 Sv flows southward in the upper ocean interior. The net flow is 17.0 Sv.

In the winter of 2009/10, it showed how sensitive the system is. The average from April to March fell from 17.1 to 13.0 Sv, and the sea level north of New York grew 128 mm in two years (Goddard and sur., 2015). New analysis (Huang i sur., 2026, preprint) claims that 2009 was a step shift, not just a fluctuation. Volkov and associates (2024) showed that correction for changing the Earth's magnetic field in the cable measurements of the Florida Stream reduced the trend by about 40%.

Long view: traces in temperature, salt and sediment

For the period before 2004. scientists use indirect indicators. The most famous is cold spot south of Greenland: the only larger area on the planet that has cooled since the beginning of the 20. century, and that is where models predict cooling when AMOC is weak. Most indicators point to weakening, but not all, and the discussion is alive.

IndicatorPeriodFindSource
Direct measurements at 26.5° N2004.–2024.Trend −0.8 ± 0.7 Svet after a decade after correction of Florida electricity. At this rate AMOC would fall to 2100. at ~10 St.nil RAPID; Rahman and Caesar, 2026.
Deep Western border current, four rows (16,5-42.5° N)2004.–2023.Consistent drop on all strings. On 26° N.−2.6 ± 0.7 Svet per decade, with an eye 20 on 15 St.Xing and sur., 2026.
Temperature of sea surface (cold spot and warmer water along with the United States)1870.–2016.Weakness from around 3 ± 1 St. 15 %) from the middle of the 20 century.Caesar and sur., 2018.
Paleoclimatic records (sevents, isotopes, plankton). – 400. – todayStable AMOC during the thousandths, then weakening from 19. century and accelerating from mid-20. Today the weakest in at least a thousand years.Caesar and sur., 2021.
Heat content of the Atlantic (15° N – 45° W)of 1950Weakness from 1.31 ± 0.39 St., with accumulation of heat south of the cold spot.Ren et al., 2025. (according to the review of Rahmstorf and Caesar, 2026.)
Temperature at 1-2 km deep in the equatorial Atlanticfrom 1960.Melting which separated from natural variability at the beginning of 2000; weakening from ~2 St.Ren, Xie and sur., 2025.
The Salinity of the South Atlantic1900.–2017.A form of salt that points to the weakening of the environment of 2 St., accelerated by 1980.Zhu i sur., 2023.
The density of the Irminger Sea1950.–2019.Weakness of ~2.2 Sv (13 %), just below the threshold of statistical significance.Chafik i sur., 2022.
Position of the Golf current (satellite altimetry)1993.–2024.A shift from around 55 km north, which is not a result of wind. In the model such a shift precedes the break of ~25 years.van Westen and Deekstra, 2026.
Age of deep water (CFC-12 and SF6 markers)1985.–2021.The waters of the North Atlantic are ‘old’, meaning that deep water is being slowly restored.Guo i sur., 2026.
Florida current (bottom cable)1982.–2023.After correction for the geomagnetic field stable for four decades. AMOC trend in the series of RAPIDs thus decreases by ~40 %.Volkov and sur., 2024.
Change in ocean-atmosphere heat (reanalysis)1963.–2017.The AMOC's ten-year averages have not weakened, and critics object that the method does not take into account ocean heat storage.Terhaar et al., 2025.
Oceanic reanalysis1958.–2025.ORAS5 −1.44, ECCO −1.05, GLORYS −0.28, and SODA +0.97 Svet per decade. Uncertainty is too large for a reliable conclusion.review of Rahmstorf and Caesar, 2026.
Table 1. Independent Indicators of AMOC Change. Rahmstorf and Caesar (2026.) conclude that "the sum of more independent evidence strongly supports" weakening. Terhaar and sur. (2025.) and Volkov and sur. (2024.) are the main works that question him.
04When?

When could a collapse happen, if at all?

This is where the research diverges most. Statistical methods that look for early warning signs in observations give earlier dates. Climate models usually show a more stable AMOC, but many are biased towards stability: they simulate too little salt in the South Atlantic and omit Greenland meltwater.

should be distinguished from the beginning of break from the break itself. When the threshold is crossed, complete shutting down in models lasts 50 to 100 years. The break is usually referred to when AMOC falls below 3-6 Sv, or a fifth to a third of today’s strength.

    Same estimates in table
    Image 4. Published estimates, each with its method and definition of the fracture. The drawing shows the range of uncertainty, and the point of the central assessment where it was published. Keep the pointer on the line for details.

    The sixth IPCC report (2021.) concluded, with average reliability, that AMOC will not collapse abruptly before 2100. This assessment relied on models that have since been considered too optimistic. The authors of recent works mainly claim that the risk is greater than thought. In a review paper signed by 22 researchers from several countries (Dijkstra and sur., 2026.) it is concluded that the current AMOC is in an area where there are two stable states.

    There are also opposite voices. Baker and associates (2025) in 34 models find that the winds of the South Ocean maintain weakened but not extinguished AMOC even in extreme scenarios. Mehling and Dijkstra (2026, preprint) in the high resolution model find that the melted water from Greenland up to 2100. adds only 0.6 ± 0.2 of its weakening. For the society is smaller than it seems: shallow and weak circulation would have almost as severe consequences as total collapse.

    05Impacts

    What a collapse would mean

    In models an AMOC collapse takes decades. For human societies, that would still be exceptionally fast. In a 2024 simulation by van Westen and colleagues, north-western Europe cooled by more than 1°C per decade, while February temperatures in Bergen fell by 3.5°C per decade. That is about 15 times the present rate of global warming.

    • It's colder
    • warmer
    • rain, drought and sea
      Image 5. Consequences of AMOC Breaking According to Model Studies. Numbers come from different models and scenarios; each is listed by source.

      It's the coldest day to come once in ten years

      Figure 6. Cold extremities after AMOC break with about +2 °C of global warming (CESM model, scenario RCP4.5). In Scotland, it is 23 °C colder than in pre-industrial climate, in the Netherlands around 15 °C. Edinburgh would have 164 cold days (lowest temperature below 0 °C) annually, 133 more than before. Source: van Westen and Baatsen, 2025.; Carbon Brief; Utrecht University.

      Change of annual rainfall in monsoon areas

      Image 7. Average of four climate models with the same "sugar" break. Tropical rain belt moves south. Source: Ben-Yami et sur., 2024.

      What about Croatia?

      For Croatia there are no separate studies of the consequences of the collapse. Models show that the cold is the strongest in Northwest Europe and weakens towards the southeast, so the difference between the north and the south of the continent would increase.

      Oceanographer Natalija Dunic of the Institute for Oceanography and Fisheries in Split notes that research has not shown the direct impact of AMOC on the thermal circulation of the Adriatic. The Adriatic circulation depends on local factors such as barrels and saltier water from the eastern Mediterranean. The impact would come across the atmosphere, and there is the most clear signal of drought for South Europe. Van Westen and associates (2025) find that the collapse would increase the drying of all Europe, most in the south. In Spain, the dry season with +2 °C would increase by 60% instead of 40%, as without breaking.

      With very high emissions (about +4 °C) European cooling in the model disappears: the northwest continues to warm up, only slower than the rest of the world, and the southern hemisphere is therefore stronger. The break of the AMOC therefore does not cancel global warming. It changes its schedule and makes it more unpredictable.

      06Timeline

      From the Ice Age to Iceland

      AMOC has collapsed before, in a world with vast ice sheets. What is new is the speed of human-driven change and the growing number of countries treating it as a security risk.

      1. before 115 000 – 11 700 year
        Earth's past

        Dansgaard-Oeschger events

        During the last ice age, the climate around the North Atlantic was rapidly shifting more than 20 times between cold and warm conditions.

      2. about 12 900 years ago
        Earth's past

        Younger Dryas

        At the end of the Ice Age, large amounts of meltwater entered the North Atlantic and AMOC nearly stopped. The North Atlantic and Europe returned to near-glacial conditions for about 1,200 years. At the end of that period, Greenland warmed by as much as 10°C in just a few decades.

      3. about 8 200 years ago
        Earth's past

        The 8.2-kiloyear event

        The emptying of huge glacier lakes in North America for a short time weakened the AMOC, and around the North Atlantic followed one to two centuries colder climate. Since then the circulation has been stable until the industrial era.

      4. 1961
        research

        Stommel's Two Boxes

        Henry Stommel shows that the temperature and salt-driven circulation can have two stable conditions and quickly shift from one to the next.

      5. 1987
        research

        “Uncomfortable surprises in the greenhouse?”

        Geochemist Wallace Broecker in Nature magazine warns that greenhouse gases could "disengage" the Atlantic circulation and popularize the image of a large ocean conveyor belt.

      6. 2004
        measurements

        RAPID begins

        The first set of instruments to continuously measure AMOC across the Atlantic was set at 26.5° U.S.A. The same year comes to the cinema the day after tomorrow, a film that permanently and incorrectly shaped the public's view of the crash.

      7. 2009.–2010.
        measurements

        Unexpected fall

        Annual AMOC average drops from 17.1 to 13.0 St. North of New York sea level jumps 128 mm in two years.

      8. 2014
        measurements

        OSNAP

        A new series across the subpolar Atlantic, from Labrador across Greenland to Scotland, reveals that most of the turnarounds take place east of Greenland, not in the Labrador Sea as thought (Lozier and sur., 2019.).

      9. 2018
        research

        Weakness footprint

        Caesar and associates from the sea temperature pattern (cold spot south of Greenland, warmer water along with the US) conclude that the AMOC from the mid-20. century weakened around 15.

      10. 2021
        research

        Weakest in a thousand years

        The reconstruction from paleoclimatic records shows that AMOC is the weakest in at least a thousand years. Borers finds signs of loss of stability that same year, and the IPCC estimates that AMOC will very likely weaken during the 21 century.

      11. July 2023
        research

        First date assessment

        Peter and Susanne Ditlevsen estimate that the crossing could occur between 2037. and 2109., most likely around 2057. The public often also mentions the range of 2025.–2095.

      12. February 2024
        research

        ‘On the road to the turning point’

        Van Westen, Kliphuis and Dykstra for the first time simulate a breakdown in a complex climate model and suggest a physical warning indicator. The data from the reanalysis indicate that the actual AMOC is moving towards the turning point.

      13. October 2024
        policy and society

        Open letter

        Forty-four climate scientists from 15 countries are telling the Nordic Council of Ministers that the risk of breaking is "very undervalued" and demand risk assessment and action.

      14. January – February 2025
        research

        Counter-arguments

        Terhaar and the heat exchangers between the ocean and the atmosphere find no weakening of the 1960. Baker and associates in the 34 model find that the AMOC is weakening, but not completely extinguished.

      15. August 2025
        research

        Shut down after 2100.

        Drijfhout and collaborators in simulations extended to 2300. find the extinguishing of the northern part of the circulation in most models with high emissions and in some with low. It is triggered by the collapse of winter mixing in the middle of this century.

      16. September 2025
        policy and society

        Iceland: issue of national security

        The Icelandic National Security Council declares the possible collapse of the AMOC as a security risk. For the first time, some climate process was granted that status. The decision was published publicly in November.

      17. October 2025
        policy and society

        Multiple Warnings

        The Global Typing Points 2025 report warns that the threshold could be below 2 °C. The Nordic Council of Ministers gathers experts in Finland to assess the risk to the region.

      18. start 2026.
        policy and society

        Nordic Report

        The report ‘A Nordic Perspective on AMOC Typing’ calls for a stronger reduction of emissions, continuous funding of measurement arrays and early warning system related to decision-making.

      19. March 2026
        research

        Golf current as alarm

        Van Westen and Deekstra show that in the model, the sudden shift of the Golf current towards the north precedes the break around 25 years, and the real Golf current has moved from 1993. around 55 km to the north. Nian and the associates of the same year show that the collapse would turn the South Ocean into a source of CO2.

      20. April 2026
        research

        About half as weak as 2100.

        Portmann and associates, with the observation limits, estimate the weakening from 51 ± 8 % to 2100. with the medium emissions. Xing and its associates record a harmonised drop in the deep current on four measuring strings and Rahmstorf and Caesar publish a review of the evidence of weakening.

      21. August 2026
        research

        Speed is important

        Van Westen, Börner and Dijkstra show that AMOC's stability depends on the rate of warming, not just its amount. With the rapid growth of CO2 AMOC, the model collapsed at +2 °C.

      07Misconceptions

      Common misconceptions

      A break in the AMOC would bring a new ice age.

      No. The earth would continue to heat globally. Nian and his associates (2026) count on the collapse of deep ocean carbon released in the long term to add about 0.2 °C to global warming. Winters in the northwest of Europe and the Arctic would cool down, while the southern hemisphere would heat up more. Summers in Europe would change less than winter, so the season would become more extreme.

      The golf course will be gone.

      A large part of the Golf Stream is driven by the wind and it would remain. It would disappear part of the current that goes further north and sinks, and with it a large part of the heat that reaches Europe. Studies also warn that the Golf Stream would move, which would in itself raise the sea level along the US coast.

      It's gonna happen overnight, like in the movie.

      In models after crossing the threshold, AMOC has been extinguished for decades to centuries. In a simulation of van Westen and associates (2024), it fell from 10 to 2 Sv in a hundred years. For ecosystems, agriculture and infrastructure, however, it is too fast to adapt: temperature trends from over 3 °C per decade do not have precedents in human history.

      Until we see him in the measurements, there's nothing to worry about.

      Direct measurements only last 20 years, and natural fluctuations are high. According to Carbon Brief estimates, the climate signal in the series RAPID cannot be reliably isolated before about 2033. If the threshold is exceeded, the consequences are virtually irreversible. Therefore, researchers develop early warning systems such as the ARIA program worth £81 million.

      If we cut the shows, the AMOC will come back easily.

      When it collapses once, it is probably not so. Due to hysteresis (see 2) recovery is not enough to return the conditions to those in which it crashed. In the Niana model and the collaborator (2026.) recovery is not possible while CO2 is above the 350 ppm.

      08What can be done

      What can be done

      The risk can be reduced substantially.

      Reduce emissions and slow down warming

      In the Dryifhout and Associate (2025.) study, shutting down to 2300. happens in 70. % of the high emissions model, in 37 % with medium and in 25 % with low. The risk depends on the decisions that are being made now. The work of van Wesen, Börner and Dijkstra (2026.) adds that the rate of warming is important and not only the final temperature. Even the temporary excess of temperature targets could trigger an irreversible collapse.

      Measure and warn on time

      The measurement sequences of the RAPID, OSNAP and SAMBA depend on short-term funding. The Nordic report from 2026. recommends their permanent funding and early warning system that combines observations and models. The UK ARIA is funding 27 teams with the aim of predicting milestones and up to a decade in advance.

      Prepare

      Iceland has put the possible collapse of the AMOC on the agenda of the National Security Council in September 2025. The ministries now assess the supply of food and energy, infrastructure and transport. "This is a direct threat to our national resilience and security," Reuters Environment, Energy and Climate Minister Jóhann Pall Jóhannsson said. British and Irish parliaments on AMOC discussed 2024. and 2025.

      09Sources and methodologies

      Sources

      Data RAPID: “Data from the RAPID AMOC observing project is based by the Natural Environment Research Council, U.S. National Science Foundation (NSF) with support from NOAA. They are freely available from rapid.ac.uk.”

      1. Moat B. I. and sur. (2026.). RAPID-MOCHA-WBTS, 26° N., 2004.–2024. (v2024.1a). BODC. doi:10.5285/48d0bf43...
      2. Carbon Brief (2026.). AMOC: Is global warfare key Atlantic ocean currents towards ‘collapse’?
      3. Ditlevsen P. i Ditlevsen S. (2023). Warning of a forthcoming collapse of the AMOC. Nature Communications . doi:10.1038/s41467-023-39810-w
      4. van Westen R. M., Kliphuis M. and Dijkstra H. A. (2024). Physics-based early warning signal shows that AMOC is on typing course. Science Advances . doi:10.1126/sciadv.adk1189
      5. van Westen R. M. i sur. (2025). Physics-based indicators for the offset of an AMOC collage under climate change. JGR Oceans . doi:10.1029/2025JC022651
      6. van Westen R. M. and Baatsen M. (2025). European temperature extremes under different AMOC scenarios in the CSM. Geographical Research Letters . doi:10.1002/2025GL114611
      7. van Westen R. M., van der Wiel K., Falkena S. and Selten F. (2025.). Changing European hydroclimate under a colloped AMOC in the CESM. HESS. hesss.copernicus.org
      8. van Westen R. M., Börner R. and Dijkstra H. A. (2026). Failure to track a stable AMOC state under rapid climate change. Nature Climate Change . doi:10.1038/s41558-026-02730-w
      9. van Westen R. M. and Dijkstra H. A. (2026). Abrupt Gulf Stream path changes are a precursor to a collapse of the AMOC. Communications Earth & Environment . doi:10.1038/s43247-026-03309-1
      10. Drijfhout S. i sur. (2025). Shutdown of North Atlantic overturning after 2100 following deep mixing collapse and CMIP6 projects. Environmental Research Letters . doi:10.1088/1748-9326/adfa3b
      11. Portmann V., Swingedouw D., Khattab O. and Chavent M. (2026.). Observational constraints project a ~50 % AMOC weakening by the end of this century. Science Advances . doi:10.1126/sciadv.adx4298
      12. Nian D. i sr. (2026). Collapse of the AMOC would lead to a substantial oceanic carbon release and additional global warming. Communications Earth & Environment . doi:10.1038/s43247-026-03427-w
      13. Xing Q. i sur. (2026.). Meridionally consistent Decline in the observed western bondry contribution to the AMOC. Science Advances . doi:10.1126/sciadv.adz7738
      14. Rachmstorf S. i Caesar L. (2026). Opinion: The AMOC is waking – time to take the record seriously. EGUsphere (preprint). doi:10.5194/ecusphere-2026-2110
      15. Dijkstra H. A. i sur. (2026.). Multi-stability of the present-day AMOC. WIRES Climate Change . doi:10.1002/wcc.70049
      16. Dijkstra H. A. i van Westen R. M. (2026.). The probability of an AMOC collapse onset in the twinty-first century. Annual Review of Marine Science . doi:10.1146/annurev-marine-040324-024822
      17. Smallers E., van Westen R. M. and Dijkstra H. A. (2024). Probability estimates of a 21st Century AMOC collage. arXiv (preprint). arXiv:2406.11738
      18. Baker J. A. et al. (2025). Continued Atlantic overturning circulation even under climate extremes. Nature . doi:10.1038/s4158-024-0
      19. Terhaar J., Vogt L. and Foukal N. P. (2025). Atlantic overturning inferred from air-sea heat fluxes indications no decline since the 1960s. Nature Communications . doi:10.1038/s41467-024-55297-5
      20. Bonan D. B. et al. (2025.). Observative constraints imprably limited the AMOC weekending future. Nature Geoscience . doi:10.1038/s41561-025-01709-0
      21. Volkov D. L. i sr. (2024). Florida Current transport observations new four decades ofsteady state. Nature Communications . doi:10.1038/s41467-024-51879-5
      22. Mehling O. and Dijkstra H. A. (2026.). Week 21st-centirate AMOC response to Greenland Meltwater and a strong eddying ocean model. arXiv (preprint). arXiv:2602.17235
      23. Huang H. i sur. (2026.). A regimen shift in Atlantic surface currents reveals a step-like decline of the AMOC. arXiv (preprint). arXiv:2604.11605
      24. Caesar L. i sur. (2018). Observed fingerprint of a weekend Atlantic Ocean overturning circulation. Nature . doi:10.1038/s4158-018-0006-5
      25. Caesar L. i sur. (2021.). Current AMOC WEAKEST in last millennium. Nature Geoscience. doi:10.1038/s41561-021-00699-z
      26. Boers N. (2021). Observation-based early-warning signals for a collision of the AMOC. Nature Climate Change . doi:10.1038/s41558-021-01097-4
      27. Ben-Yami M. et al. (2024). Effects of AMOC collapse on monsoon rainfall: a multi-model comparison. Earth’s Future . doi:10.1029/2023EF003959
      28. Ritchie P. D. i sur. (2020). Shifts in national land use and food production in Great Britain after a climate tipping point. Nature Food . doi:10.1038/s43016-019-0011-3
      29. Armstrong McKay D. I sur (2022). Exceeding 1.5°C global warfare could trigger multiple climate typing points. Science . doi:10.1126/science.abn7950
      30. IPCC (2021.). AR6 WG1, chapter 9: Ocean, Cryosphere and Sea Level Change. ipcc.ch
      31. Goddard P. B. i sur. (2015.). An extreme event of sea-level rises along the Northeast coast of North America and 2009-2010. Nature Communications . doi:10.1038/ncomms7346
      32. Lozier M. S. i sur. (2019.). A sea change in our view of overturning in the subpolar North Atlantic. Science . doi:10.1126/science.au6592
      33. Stommel H. (1961). Thermohaline connection with two stable regimens of flow. Tellus . doi:10.1111/j.2153-3490.1961.tb00079.x
      34. Broecker W. S. (1987). Unpleasant surprises in the greenhouse? Nature . doi:10.1038/328123a0
      35. Open Letter by Climate Scientists to the Nordic Council of Ministers (October 2024). vedur.is
      36. Nordic Council of Ministers (2026). A Nordic Perspective on AMOC Typing (TemaNord 2026:504). pub.norden.org
      37. CNN (November 15, 2025). A crucial system of ocean currents may be on course to collapse. This country just declared it a national security threatened. cnn.com
      38. Global Typing Points Report 2025. global-typing-points.org
      39. Carbon Brief (2025). Ocean current ‘collapse’ could trigger ‘profound cooling’ and northern Europe. carbonbrief.org
      40. University of Utrecht (2025.). What will happen to Europe if the Gulf Stream weekends are significant? uu.nl
      41. Index.hr (February 14, 2024). Confirmed that the main Atlantic current is threatened by a recent collapse. What does this mean for Croatia (declaration by Dr. Natalije Dunić, IOR). index.hr
      42. ARIA, the Forecasting Tipping Points program. ARIA.org.uk