Nitrate Pollution In Rivers Linked To Dry Climates, Not Just Human Activity

Nitrate, a chemical mixture of oxygen and nitrogen often used in farm fertilizers, can rise to toxic levels in surrounding water under very dry conditions. This relationship has traditionally been attributed to agricultural runoff or human waste in the environment, but new research led by scientists at Penn State shows that a hot and dry climate elevates concentrations of nitrate in rivers and streams, even in areas without urban or agricultural development.

Using artificial intelligence (AI) algorithms to overlay data patterns recorded from over 7,000 river sites around the world, the team built a machine learning model that can reliably map nitrate concentrations across regions with scarce data. Their research, which the team said is the first of its kind at a global scale, suggests that nitrate pollution is not driven by human activity alone — climate conditions also play a critical role. They published their findings today (Nov. 12) in Science.

“Right now, roughly a half billion people are living in proximity to rivers where nitrate concentrations exceed levels that can potentially lead to negative health effects,” said Li Li, the Barry and Shirley Isett Professor of Environmental Engineering and corresponding author on the paper. “This finding implies that nitrate pollution is beyond a local management issue: it’s a broader problem directly related to climate, which is becoming increasingly hotter and drier.”

Nitrate is one of the most prevalent forms of inorganic pollution found in rivers, streams, lakes and groundwater worldwide, with about 18% of the U.S. population exposed to drinking water with elevated nitrate levels between 2021 and 2023. In addition to harming ecosystems, high levels of nitrate in drinking water can lead to a host of health conditions, most notably methemoglobinemia — commonly known as blue baby syndrome — a potentially fatal condition where an infant’s blood cannot effectively carry oxygen.

“Generally, we think nitrate hotspots derive from human input like fertilizer used in agricultural or wastewater from cities,” said Li, who holds an additional affiliation with Penn State’s Institute of Energy and the Environment. “There’s not a lot of very-large-scale water quality work, so people don’t traditionally think about climate in terms of affecting nitrate levels. However, our new model is telling us that although human input plays a major role, long-term climate such as average annual temperature and average annual rainfall of an area contributes to environmental nitrate levels, as well.”

Nitrate occurs naturally in low doses, but heavily farmed areas like the Corn Belt, a portion of the U.S. Midwest spanning from Nebraska to Ohio, exhibit some of the highest nitrate concentrations on Earth, said Kayalvizhi Sadayappan, a postdoctoral scholar in civil and environmental engineering and first author on the paper. Current regulatory standards exist to protect against adverse health impacts, but Sadayappan said that nitrate can still carry risks at lower levels.

“For nitrate, we have a regulatory standard set at 10 milligrams of nitrate per liter of water, primarily to prevent acute diseases like blue baby syndrome,” Sadayappan said. “This has been the standard since the 1970s, but scientists are now observing that slower, chronic conditions like cancers or issues related to reproduction can be linked to exposure at much lower levels, like 2.5 milligrams of nitrate per liter of water.”

The model works by aggregating millions of individual nitrate readings taken from over 7,000 river segments, training predictions based off patterns observed in this data. Li explained that a portion of these readings are withheld from the training data set so that predictions can be compared to real data the model has not processed, allowing the team to evaluate the model’s accuracy. The team built approximately 1,000 models, each trained on a selected subset of monitoring sites, to additionally assess model uncertainty. After training, the model can make predictions of nitrate concentrations, even in locations with little data, allowing the team to reconstruct a global map of nitrate concentration in about 2.9 million river segments around the world.

This map has limitations, Sadayappan explained, noting that most nitrate measurements come from regions with temperate climates, such as North America and Europe with rich resources for monitoring, while far fewer data points are available in arid regions like north Africa. This data gap has long complicated the process of modeling global water quality broadly.

“To reconstruct a global map, we extracted information about 26 different variables for individual river segments, including climate characteristics, land-use characteristics and soil information,” said Sadayappan, explaining that this information enables each model to link land characteristics to nitrate concentration.

The team reported that even in parts of the world with little or no agricultural and urban activity, rivers in arid climates showed consistently elevated levels of nitrate pollution. The Sahara Desert in northern Africa, one of the most arid, remote locations on Earth, displayed comparable if not higher nitrate concentrations than humid locations, like southeast Asia, even if those humid environments had extensive urban or agricultural development. According to Li, the best explanation is that naturally occurring nitrate builds up in arid climates over years, since limited water movement prevents it from being washed out.

“Honestly, the finding surprised us a bit,” Sadayappan said. “Human inputs still have the biggest impact, but nitrate levels are elevated even in some of the world’s most remote regions, particularly in small headwater streams that flow only seasonally.”

Li said she sees this finding as transformative. Nitrate pollution has previously been considered a locally managed issue, but increasing climate aridity around the globe could exacerbate the issue of nitrate pollution around the world. According to Li, better understanding of nitrate’s relationship to climate and how this pollution builds up in the environment, as well as the long-term health impacts of potential exposure, could play a fundamental role in future regulatory decisions about acceptable exposure levels.

“There’s a lot of discussion about floods and droughts because their impacts are easier to see, but it can be difficult to notice water quality issues because you can’t visually observe the issues until they get really, really bad,” Li said. “We hope that work in this space will continue to increase awareness and more data collection, especially in data-sparse regions, around this problem of water quality.”

This work was supported by the U.S. National Science Foundation under award number EAR-2346471, IEE support for Person of the Year and the Isett Professorship. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.