A peer-reviewed review argues aquatic deoxygenation belongs alongside climate change and biodiversity loss on the list of processes that keep Earth liveable. The open ocean has already lost more than 2% of its oxygen since 1960, mostly out of sight — and in some regions the signal has been detectable for nearly 70 years.
A review led by the Scripps Institution of Oceanography, published 30 June in Limnology and Oceanography, argues that oxygen loss across oceans, lakes, rivers and coastal waters — collectively "aquatic deoxygenation" — is now severe and interconnected enough with other Earth-system stresses to warrant formal inclusion in the Planetary Boundaries framework.1 That framework, introduced in 2009, currently tracks nine processes, including climate change, biodiversity loss and ocean acidification, that scientists treat as limits on a stable, liveable planet.1
The review does not present new measurements. Its contribution is synthesis: tracing how falling oxygen levels feed back into the other nine boundaries — warming water holds less oxygen, which stresses ecosystems, which alters the carbon and nitrogen cycling that regulates climate in turn.1 The underlying oxygen decline itself is well established. The most-cited global estimate, from a 2017 analysis of five decades of ocean measurements, put the loss at more than 2% of the ocean's total dissolved oxygen since 1960 — roughly 80 billion tonnes.2
Most public attention to ocean oxygen loss goes to coastal "dead zones" — nutrient-fed hypoxic patches near river mouths and cities, driven mainly by agricultural and wastewater runoff rather than by warming directly. A 2019 global review coordinated by the IUCN counted around 700 documented dead zones worldwide, up from just 45 in the 1960s.6
That is a visible, coastal, largely nutrient-driven problem, and it has grown more than fifteen-fold. But it is dwarfed in scale by oxygen loss in the open-ocean interior, which is driven mainly by warming and reduced circulation, cannot be seen from shore, and has no equivalent headline count — only the diffuse, ocean-wide percentage figure above.2 The new review's central argument is that treating these as separate, minor problems has let the larger, slower one go largely unmanaged.1
The oxygen loss described above has two largely separate causes, and they matter for different reasons. In the open ocean, warmer water simply holds less dissolved oxygen, and a warmer surface layer resists mixing with the cooler water beneath it — a process oceanographers call stratification. Both effects reduce how much oxygen reaches mid-depth waters, and both are consequences of rising temperature rather than of pollution.2 A 2026 modelling study of five tropical low-oxygen zones decomposed this further, separating the solubility effect from a second factor tied to ocean ventilation and biological oxygen use — and found the two contribute in different proportions from region to region.3
Along coastlines, the dominant driver is different: nutrient pollution from agricultural runoff and untreated sewage feeds algal blooms that die, sink and are broken down by oxygen-consuming microbes.6 This is the mechanism behind Exhibit 01, and the counting approach used to track it traces back to a 2008 global survey that first showed the number of dead zones doubling roughly every decade since the 1960s.4 The two problems compound each other in warm, poorly-flushed coastal waters, but they respond to different fixes: nutrient pollution can be reduced locally and relatively quickly, while warming-driven oxygen loss cannot.
Both mechanisms above leave a statistical fingerprint that climate models can date. A 2026 study used a 30-member climate model ensemble to identify, for five tropical ocean regions, the year each one's oxygen decline became distinguishable from the natural year-to-year variation the ocean would show with no human influence at all.3 That threshold — researchers call it "time of emergence" — is not the year oxygen loss began; it is the year it became detectable with statistical confidence.
The review itself is a synthesis of existing evidence, not a new measurement campaign, so its "unsafe space" conclusion is a scientific judgement about accumulated risk rather than a single tested threshold — the authors say as much, noting that formally establishing oxygen as a Planetary Boundaries "control variable" needs further work.1 Coastal dead-zone counts also depend on where anyone has looked: a statistical analysis of the global dead-zone map found documented sites are under-represented in the tropics and over-represented in temperate regions relative to coastline length, and IUCN's tropical-ecosystems reviewers separately estimate several hundred additional hypoxic sites remain undocumented there.57 So the true tropical count is almost certainly higher than Exhibit 01 shows. The Exhibit 02 emergence dates are model estimates with a real uncertainty window, not fixed historical facts. And the two mechanisms in this piece have different dominant causes — nutrient runoff for most coastal dead zones, warming and circulation change for the open-ocean decline — so neither figure should be read as a single "climate change did this" number.