The ocean is slowly losing its life-sustaining oxygen, according to a new study that analyzes years of existing research on deoxygenation and puts the problem in stark terms. The big question, now, is whether the trend can be reversed in time to avoid dramatic reductions in biodiversity that could ripple through the marine food web, affecting humans as well as ocean ecosystems.

The review, published in Science, found that 2 percent of the ocean’s total oxygen has been lost over the last half century and that the volume of oxygen-free water on the open ocean has quadrupled. Oxygen-minimum zones – waters with less than half the normal concentration of oxygen – have expanded by an area about the size of the European Union.

“If we lost 4.5 million square kilometers of productive area on land, everyone would be appalled,” said Denise Breitburg, a marine ecologist with the Smithsonian Environmental Research Center and lead author of the study. “But what happens beneath the surface of the ocean is out of sight, and easy to either not notice or ignore.”

Breitburg is part of the Global Ocean Oxygen Network, a consortium of scientists created in 2016 by the United Nations Intergovernmental Oceanographic Commission. The network, which supported the study, aims to advise policymakers on a significant environmental challenge that often falls under the radar compared to other issues such as acidifying and warming oceans.

Rising sea temperatures and disappearing oxygen are, in fact, closely linked. Global warming is “the likely ultimate cause” of oxygen loss in the open ocean, according to the study. That is because warmer water absorbs less oxygen and also speeds up the metabolism of organisms in the sea, causing them to consume oxygen faster. Another issue is that warmer surface water mixes less readily with the oxygen-rich waters of the deep sea.

Closer to coasts, there are additional challenges. Fertilizer, sewage and other pollutants found in coastal runoff deliver an influx of nutrients that fuel coastal algae blooms and lead to oxygen-free “dead zones.” Since 1950, the area of the ocean at risk of developing dead zones has increased more than 10 times. Many more dead zones may exist in developing countries where monitoring is sparse, the scientists wrote.

A study published in August found that if the current deoxygenation crisis mirrors one that took place nearly 100 million years ago, the area of oxygen-deprived waters might double over the next 100–350 years. This would threaten a range of ocean life, fisheries, tourism and coastal communities. In coastal waters, oxygen declines and dead zones could shrink the habitats, stunt growth and impede the reproduction of many species, including commercially important ones like shrimp.

“There will always be some areas of the sea that have low oxygen – just as there are deserts on land – the problem is when these areas expand and replace more productive ecosystems,” said Breitburg.

Matthew Long, an oceanographer with the National Center for Atmospheric Research who has authored several studies on deoxygenation but was not involved in the new report, said that he believes awareness of the problem is growing.

“I get the sense that there’s some momentum,” he said. “I think the scientific community is becoming more and more aware that we need to do a better job of communicating this issue so policymakers are aware of the potential impeding crisis.”

The report identified feasible steps to address dead zones near coastlines, such as reducing nutrient runoff through measures such as improving septic systems. However, slowing or halting the larger-scale decline of oxygen due to warming is a harder issue that “will take a global effort,” Breitburg said.

To stop large ocean oxygen losses, fossil fuel emissions will need to be cut sooner rather than later, according to the study, and it’s possible it could be too late in any event. How quickly deoxygenation rates could be reined in if the temperature rise is slowed remains unknown, however.

In the meantime, Long said, societies can try to adapt. The study suggests several strategies, including setting up sanctuaries that protect animals that have fled from low-oxygen areas as well as limiting the harvest of species most affected by deoxygenation. Better data collection would also help, the researchers said, including real-time observations and more advanced forecasts that could potentially predict when and where low-oxygen events could occur years in advance.

It’s likely these and more adaptation measures will be needed. “There is substantial deoxygenation expected even with relatively low emissions levels,” said Long. “The ocean responds on relatively slow timescales … and actions we take now will require decades to centuries to fully manifest in the oceans – reversibility on human timescales is not guaranteed. We do have the potential to cross areas of no return.”