Climate Tipping Points: One Has Been Crossed, Most Have Not

Text type: EDT (Editorial/Analysis) | State of facts: August 7, 2026

In October 2025, over 160 researchers from 87 institutions across 23 countries presented the Global Tipping Points Report, shortly ahead of COP30 in Belém. One finding made it into almost every headline: one tipping point of the Earth system is considered crossed. Since then, several more recent studies have appeared that sharpen the picture in some areas — and make it more nuanced in another.

Both belong in an honest assessment. Because the two most common errors in dealing with tipping points are strikingly similar: those who deduce doomsday from a single finding, and those who turn scientific uncertainty into an all-clear, both confuse risk with certainty.

Three Points in Time That Do Not Coincide

A tipping point is a critical threshold beyond which even small additional perturbations trigger a qualitative, self-reinforcing, and largely irreversible system change. There is no single global “point of no return”, but rather around a dozen very distinct tipping systems, each with its own thresholds. And these thresholds are uncertainty ranges, not exact values.

Crucial here is a distinction that is almost always lost in public debate: triggering, commitment, and visible collapse are three distinct points in time. A destabilization can be initiated long before a reversal becomes practically impossible — and the fully observable restructuring can follow centuries later. For ice sheets, these three moments are separated by centuries or millennia. Therefore, “a tipping point has been crossed” explicitly does not mean that anything will collapse tomorrow. It means that the course is set well before the visible outcome manifests.

Equally important: it is not just peak temperature that matters, but the duration of overshoot and regional compounding stressors such as deforestation, overfishing, or pollution.

When looking at the current level of warming, precision pays off, because two correct figures are in circulation that do not measure the same thing. Human-induced warming reached 1.37 °C in 2025 according to the Indicators of Global Climate Change report. The observed annual mean according to Copernicus/ERA5 was 1.47 °C — this higher value includes natural variability such as El Niño. Both are correct; they should not be conflated.

The First Crossed Tipping Point: Coral Reefs

Warm-water coral reefs have reached their central threshold range of around 1.2 °C (range 1.0 to 1.5 °C). In the fourth global bleaching event, NOAA recorded heat stress across approximately 84 percent of global reef area (as of September 2025) in at least 83 countries and territories — the largest such event ever measured.

“Tipped” here does not mean “disappeared”, but rather: transition to systems dominated by algae and sponges, with isolated refugia. The time horizon is years to decades, not centuries. Reefs are thus the system where the three points in time are closest together — and the only one where the restructuring is already actively underway.

Bar chart of estimated temperature thresholds for eight tipping systems with uncertainty ranges and current level of warming

The thresholds are ranges, not lines — and today’s level of warming sits at the lower edge of several of them.

The Ice Sheets: Slow, but Decided Early

For the Greenland ice sheet, research identifies a threshold range of about 0.8 to 3 °C with a central value just above 1.5 °C. In the event of total loss, this would correspond to approximately seven meters of sea-level rise over millennia. When major dynamical changes will set in is difficult to predict due to the internal variability of ice streams.

For Antarctica, a study by Ricarda Winkelmann and colleagues (Nature Climate Change, February 2026) mapped a sequence of critical thresholds for the first time: the initial basin-wide thresholds in West Antarctica lie at 1 to 2 °C and affect around 40 percent of marine-grounded ice volume in the long term. The Amundsen Sea Embayment with Thwaites and Pine Island Glaciers exhibits the lowest thresholds and may already have been crossed at current levels. For the marine basins of East Antarctica (Wilkes, Aurora), the same study identifies 2 to 5 °C.

These figures explicitly do not originate from the 2025 report, but from an independent, more recent paper — as do the subsequent Amazon and AMOC findings published between February and May 2026.

AMOC: The System with the Most Honest Uncertainty

The Atlantic Meridional Overturning Circulation is an established tipping system, but its timing remains intensely disputed — and recent research has not become clearer, but more divergent.

Van Westen and colleagues demonstrated in 2025 that the tipping point persists even in an eddy-resolving ocean model, meaning it is not an artifact of coarse model resolution. In contrast, a more recent study by Mehling and colleagues (2026), also using a strongly eddy-resolving model, finds only moderate, approximately linear weakening driven by Greenland meltwater through 2100 — without prominent abrupt transitions. Two methodologically similar approaches, two different answers.

What is solidifying independently of this is the weakening itself: under observational constraints, Portmann and colleagues project a weakening of 51 ± 8 percent by 2100 — significantly stronger than the previous multi-model ensemble mean. A halved overturning circulation is severe. But it is not a collapse, and a robust collapse probability before 2100 cannot currently be specified.

The Amazon Is Not a Pure Climate System

Here lies perhaps the most actionable insight of the year. A modeling study by Nico Wunderling and colleagues (Nature, May 2026) shows: without additional deforestation, the Amazon’s climate threshold would lie at roughly 3.7 to 4.0 °C. If 22 to 28 percent deforestation is factored in, large-scale transitions across 62 to 77 percent of the forest area become possible at just 1.5 to 1.9 °C. Currently, around 17 to 18 percent has been deforested.

The Amazon is thus not a simple toggle switch operated solely by global temperature. Anti-deforestation measures serve as immediately effective climate action here — and that is a decision, not a physical constant.

What Is Not a Classic Tipping Point

Just as important as confirmed systems are those that have been refuted. For permafrost, current research finds no sufficiently strong global feedback mechanism: thawing proceeds predominantly gradually and scales with the extent and duration of warming. Local abrupt processes such as thermokarst are real; a global “methane bomb” is not. The report classifies Arctic summer sea ice with high confidence as a non-tipping element; its loss is more reversible upon cooling than other processes. Likewise, for ENSO and the jet stream, evidence does not support independent tipping points.

This is not an all-clear, but good news of a different kind: systems that respond proportionally to warming also respond proportionally to its limitation.

Cascades: A Risk Scenario, Not Destiny

A systematic review by Wunderling and colleagues (2024) identified more destabilizing than stabilizing interactions between tipping systems. The AMOC is a particularly interconnected node: Greenland meltwater acts upon it, while a weakened AMOC in turn affects Amazon precipitation and monsoons.

Network diagram of destabilizing interactions between tipping systems, with the AMOC as a particularly strongly connected node

Most known couplings increase risk elsewhere — but a domino cascade has not begun because of it.

An inevitable cascade is not in motion. It is a risk scenario whose probability of occurrence increases with warming.

What Follows

Taking all evidence together, three sober guidelines emerge: keep peak warming below 2 °C. Following an overshoot, return swiftly below 1.5 °C. Long-term, stabilize toward 1 °C. Minimizing the duration of overshoot is more important than the absolute peak height.

Every tenth of a degree is measurable: according to a model analysis by Möller and colleagues (2024), the risk of crossing tipping points rises between 1.5 and 2 °C by about 1 to 1.5 percent per additional tenth of a degree — and by around 3 percent above 2.5 °C. Conversely, every avoided tenth of a degree and every year shortened in overshoot represents a real, quantifiable gain.

And the most effective levers are known and available: reduce emissions, halt deforestation, protect peatlands and coastal ecosystems. The risk grows with every tenth of a degree and with every year spent above it. Yet precisely therein lies the message: both can be influenced.

References

Image Credits

Both graphics in this article are AI-generated (labeled in accordance with Art. 50 EU AI Act):

Temperature thresholds of tipping systems — Data graphic (SVG), created with Google Gemini (via Antigravity) according to editorial specifications. Data basis: the threshold ranges cited in the article and the current level of warming mentioned.
Cascade network of interactions — Schematic graphic (SVG), created with Google Gemini (via Antigravity) according to editorial specifications. Content basis: Wunderling et al. (2024), systematic review of interactions between tipping systems.

Transparency note: The research foundation for this article was an in-depth investigation using GPT. All core statements were subsequently verified independently by Claude against primary sources via web research (20 individual claims; the source-check report is available internally to the editorial team). Both graphics were created using Gemini (Antigravity). English translation and localization performed by Gemini (Antigravity) in accordance with the Um:bruch Editorial Policy. Editorial responsibility and final text review lie with Um:bruch.

In case of discrepancies between language versions, the German original prevails.