The United States has spent over a century collecting water-quality data from its rivers, building a scientific record of more than 50 million observations. A University of Florida study found that fewer than 1% of the nation’s approximately 24 million river segments have ever been sampled.

The study, funded by a $428,000 National Science Foundation grant, raises questions about how well scientists understand America’s waterways while identifying opportunities for future monitoring and research.

The study, led by scientists at UF’s Institute of Food and Agricultural Science (UF/IFAS) and the UF Water Institute, examined where, when and how frequently water-quality samples have been collected in rivers across the nation. The findings reveal patterns in the nation’s monitoring efforts, including an opportunity to better understand smaller streams, which tend to be less predictable than larger rivers but receive less sampling attention. That pattern is important because the study found that water-quality variability generally decreases as streams get larger.

“Our study shows us where we have a picture of water quality in US rivers, where that picture is well-defined, and where it is blurrier,” said lead author Nicolas Fernandez, a postdoctoral associate of the UF Water Institute. “By bringing a century of observations together, the research provides scientists and water managers with a clearer understanding of where information is plentiful, where it is limited and where additional monitoring could help answer important questions about the nation’s waterways.”

On a positive note, Florida emerged as a national leader in water-quality sampling, highlighting the state’s contribution to the extensive record. Florida also devotes the highest proportion of its sampling effort to small streams among all states.

Nicolas Fernandez crosses the Red Cedar River, one of the large rivers included in the study, near East Lansing, Michigan. Photo by Nicolas Fernandez.

Photo courtesy UF/IFAS

The research analyzed approximately 50 million observations covering 11 water-quality constituents, including specific chemical, physical or biological substances and properties that were measured in river water samples. Annual sampling increased dramatically from 1925 to 2010 before declining about 40% by 2022.

“We’ve taken tens of millions of water-quality observations from across the country and organized them to understand where the data are, when they were collected and how monitoring varies,” said James Jawitz, professor of landscape hydrology in the UF/IFAS Department of Soil, Water and Ecosystem Sciences.

Fernandez said the findings are not necessarily an argument that every river should be sampled equally. Rather, they offer an opportunity for scientists and water managers to consider how monitoring efforts are currently designed, and how they could be adjusted to provide more useful information.

How, when and how often should waterways be sampled?

The study also reveals a gap in the information available to understand how water conditions vary across the nation’s river network, said Jawitz.

This limited sampling of smaller waterways stood out among the findings to us, said Jawitz. “Almost none of the small rivers have ever been sampled in U.S. history”

The researchers also found a trade-off between how many locations are sampled and how frequently individual sites are monitored. Many locations have only one recorded observation, while a smaller number have been sampled repeatedly over time.

That distinction affects the questions scientists can answer.

“A single measurement provides a snapshot of conditions at one point in time, while repeated observations can help researchers determine whether conditions are stable, improving or deteriorating,” said Jawitz.

The study also revealed patterns in when samples are collected. Across the data set, sampling was concentrated during the summer, around the middle of the month and during the middle of the day.

“These patterns matter because water conditions can vary over time,” said Fernandez. “Understanding when measurements are taken helps researchers interpret the available data and recognize the limitations of drawing conclusions from observations collected during particular periods.”

Researchers were also surprised by a decline in sampling since 2010. Fernandez noted that only a small number of sites have consistent, long-term records needed to track water-quality trends.

“When regular sampling stops at those locations, scientists can lose the ability to determine whether conditions are changing over time,” he said.

The study did not investigate why sampling has declined, but its findings provide a baseline for understanding how monitoring practices have changed and where continued observations could be especially valuable, said Jawitz.

The research team is now building on this foundation to investigate water-quality patterns and the factors that may help explain them. For example, researchers are examining where nutrient concentrations, such as nitrate, are high or low and how those patterns relate to characteristics of the surrounding landscape.

The team is also conducting more detailed research in Florida to investigate where these concentrations are increasing or decreasing and why environmental interventions may produce different results in different locations.