Almost nothing can be seen about the systems that make weather forecasts possible. A lot of people don’t give the underwater gliders, seafloor sensors, and moored buoys that float somewhere in the Southern Ocean much thought. They’re not known by many people. But forecasters, farmers, emergency managers, and insurance actuaries will all feel it eventually, sometimes before they know what hit them.
Scientists are warning about this right now, as the Trump administration works to break down the Ocean Observatories Initiative, a huge network of $368 million systems and platforms on the ocean floor that is run by the US National Science Foundation. Along both coasts of the United States, the OOI goes out into the North Atlantic and Southern Oceans. It has been used to look into everything from subduction zone earthquakes to heat waves in the ocean. It also sends important data in real time to the Global Ocean Observing System, which is the international system that makes modern weather prediction possible.
A study that came out last month in Nature Climate Change gives us a sobering look at what losing this network really means in real terms. If US observations were taken out of the equation, the error for annual ocean heating rates would grow by 163%. In this case, the study found that losing American platforms would be statistically worse than throwing away 80% of all ocean data collected around the world at random. It’s not a rounding mistake. That’s a loss of signal.
Sabrina Speich, a co-author of the study and an expert in global ocean monitoring at the Ecole Normale Supérieure in Paris, put it simply: the ocean’s heat content is the best way to tell what’s going on with the climate system as a whole, not just the ocean itself. She said that the vertical temperature profiles that give us this information are some of the simplest measurements that science can make. If you lose them, you’ll lose something that can’t be quickly replaced or imitated from space. From space, you can’t see the deep ocean.
Take a moment to think about that detail. These measurements aren’t strange or experimental. They are simple, tried-and-true readings that have been used to make predictions for decades. And predictions are important for more than just climate science. Farmers in the US and South America have months to decide what crops to plant before El Niño early warning systems go off. The El Niño of 2023–2024 was one of the five strongest on record, and it directly caused the world’s temperature to rise to a record high level last year. If the network of observers gets worse before another expected El Niño year, there will be less time to get ready, the warnings will not be as accurate, and the economy will be more vulnerable to the shocks that come after.
John P. Abraham, an engineering professor at the University of St. Thomas in Minnesota and co-author of the study, said that the move was “penny-wise, pound-foolish.” Since 1980, the US has been hit by more than 400 weather and climate disasters that cost at least $1 billion each. These costs added up to $177 billion just in 2024. In this situation, getting rid of the sensors that help make forecasts more accurate doesn’t seem like a smart way to save money; it seems more like something else. He said, “This is not about saving money.” “This is about killing climate science research.”
The National Science Foundation said that the changes were not a complete cancellation but a “descope,” which means that parts of the program were cut back. It’s not yet clear how much observing capacity will be left over after the cut. At the same time, the European Union said it would invest €92 million in OceanEye, a project to monitor the oceans. More than half of this money will go to the Global Ocean Observing System. The EU was careful in how it talked about it. Officials said this had been planned for a long time and wasn’t a direct response to Washington’s decision. But it’s hard to ignore the timing.
As I watch all of this happen, I have the feeling that the results won’t be very loud. They’ll come slowly, in the form of errors in forecasts that are a little bigger than expected, models of storm paths that move a little wider, and quieter closings of planning windows for agriculture. When the damage is clear, the sensors and the institutional knowledge that they held will be gone. Because of this, important infrastructure often fails without a warning, in the space left by the lack of one.
