January 23, 2026 — Scientists from Louisiana State University are using artificial intelligence to help predict and address water quality issues faster and more accurately, with hopes their tools could help shrink the oxygen-depleted “dead zone” that develops annually in the Gulf of Mexico.
The researchers have developed an AI tool that fills in the data blanks for areas where water testing isn’t available or can’t be verified. To test their technology, they used it to predict how much nitrogen and phosphorus are present in the rivers and streams of Puerto Rico and what amounts are in the soil. The two elements are primary ingredients of agricultural fertilizers, and they can deplete the oxygen content of water by fueling algae growth.
The AI-informed models offer insight into how nitrogen and phosphorus can be retained in soil instead of washing downstream where they can fuel hypoxic dead zones.
“That’s a big difference that can also help planners and decision makers say, ‘Where do we need to restore throughout the watersheds? … How can we increase the capacity of those places that retain nutrients like this?’” said Mariam Valladares Castellanos, lead author of the paper and postdoctoral researcher at LSU. Her team’s findings were published in the academic journal Science of The Total Environment in October.
Puerto Rico, at 3,500 square miles, is relatively small when compared with the Mississippi River Basin, which covers 31 states. But Valladares Castellanos said the island is a really good research subject because its mountains and coastline provide a wide variety of watershed scenarios to teach the AI model.