Scientists Have Mapped Earth's Climate Tipping Points. New Research Finds Attention to Them Is Wildly Uneven

Climate scientists have spent a quarter-century studying the handful of Earth systems that could tip into abrupt, self-reinforcing and largely irreversible change. But a new analysis of more than 20,000 scientific papers finds that attention to these "tipping elements" has been strikingly uneven — and not necessarily aligned with how close each one may be to a dangerous threshold.

The study, published in the Proceedings of the National Academy of Sciences, examined 20,736 publications from 2000 through 2025 covering 14 recognized climate tipping elements, including the Greenland and Antarctic ice sheets, the Atlantic Meridional Overturning Circulation (AMOC), the Amazon rainforest, coral reefs and permafrost. The Greenland Ice Sheet alone accounted for roughly 20% of the entire dataset — 4,141 papers — while abrupt thaw of boreal permafrost appeared in just 139 papers, or 0.7% of the total, and the North Atlantic Subpolar Gyre, a current system in the ocean off Scandinavia and Greenland, turned up in only 649, or 3.1%.

What the numbers show

The research team, led by He-Ming Wang and Bing-Yu Chen of Northeastern University in Shenyang, China, with co-authors including Feng-Mei Ma of Australia's Commonwealth Scientific and Industrial Research Organisation (CSIRO), Nico Wunderling of the Potsdam Institute for Climate Impact Research, Timothy Lenton of the University of Exeter's Global Systems Institute, and Asaf Tzachor of Reichman University and the University of Cambridge, mined publication records to count how often each tipping element appeared in the scientific literature and cross-referenced that with each element's estimated "warming threshold" — the amount of global temperature rise thought to risk triggering it.

They found no systematic relationship between a lower threshold and more research attention. Boreal permafrost and the Subpolar Gyre both carry comparatively low estimated thresholds, yet both are among the least-studied systems in the dataset. Meanwhile Greenland and the AMOC, which also have relatively low thresholds, attracted the most attention by far. The mismatch, the authors argue, means some of the riskiest possible tipping points are also among the hardest for policymakers to assess, simply because the underlying science is thin.

The team also looked at how often studies engaged with the mechanics that make tipping points dangerous in the first place — abrupt transitions, hysteresis and irreversibility — rather than treating a tipping element as just another gradual climate variable. Only 8.1% of the 20,736 publications explicitly addressed those risk-relevant dynamics. For boreal permafrost thaw and low-latitude coral reef collapse combined, the authors found just 22 such studies; for the Subpolar Gyre, 53.

Why tipping points matter

The idea that parts of the climate system could cross thresholds into rapid, hard-to-reverse change has moved from a niche scientific concern two decades ago to a recurring feature of climate-risk assessments and international negotiations. The Global Tipping Points Report, compiled by more than 100 researchers and released ahead of recent UN climate conferences, has warned that several tipping elements are already at risk of being crossed within the current warming trajectory. Lenton, one of the new study's co-authors, helped popularize the modern tipping-points framework in an influential 2008 PNAS paper that first catalogued the systems now at the center of climate-risk debates.

That framework assigns each tipping element a rough warming threshold — a global temperature rise beyond which the system becomes significantly more likely to shift into a new, self-sustaining state. Those thresholds carry wide uncertainty bands, and the authors of the new study are careful to note that a threshold is not a predictable deadline.

"A warming threshold is not a timer counting down to a known date," Tzachor wrote in an essay for The Conversation describing the findings.

Tzachor also stressed that publication counts are a measure of scientific attention, not of how well-understood or scientifically "settled" a given risk is: "Our analysis measures scientific attention. It does not measure scientific certainty." He added that some systems "receive thousands of studies; others, with similarly low estimated thresholds, receive hundreds or fewer."

Who this affects

The imbalance has practical consequences beyond academia. Governments, insurers and international bodies increasingly rely on the peer-reviewed literature to estimate the likelihood and consequences of abrupt climate shifts when setting emissions targets, adaptation budgets and risk disclosures. A tipping element with a thin research base is harder to model, harder to monitor with early-warning systems, and harder to justify funding for — even if its underlying physical risk is comparable to a heavily studied system like the Greenland ice sheet. The authors frame this as a governance problem as much as a scientific one: attention gaps could quietly shape which climate risks get prioritized at climate negotiations and in national risk assessments, independent of the actual physical danger involved.

What happens next

The study's authors are not arguing that research on Greenland or the AMOC — both extensively studied because of their outsized influence on sea levels and regional weather — should be scaled back. Instead, they call for more funding and monitoring directed at the comparatively neglected systems, particularly boreal permafrost thaw and the Subpolar Gyre, and for more studies that directly examine abrupt-change dynamics rather than treating tipping elements as ordinary, gradually shifting variables. The paper does not resolve why the imbalance exists, leaving open questions for follow-up research: whether it reflects easier access to satellite data for ice sheets, funding patterns tied to visibility, or the relative difficulty of studying slower-moving systems like permafrost and ocean currents. The authors suggest their publication-count method could be applied to other scientific fields where attention to risk may not track the underlying hazard.