A greatly enhanced stream-monitoring network throughout the Bay watershed is helping scientists shed light on one of the Chesapeake region’s most fundamental questions: Is the amount of pollution from the Bay’s rivers increasing or decreasing — and why?

The network, overseen by the U.S. Geological Survey along with state and regional agencies, now samples water quality at 117 sites across the six-state watershed — a nearly fourfold increase from several years ago.

“As we increase the number of sites, it is just staggering how much information we are generating,” said Doug Moyer, a USGS hydrologist.

Using the additional data and a new analytic technique, researchers in recent years were able to figure out that the Conowingo Dam holds the key to worsening nutrient and sediment pollution trends at the mouth of the Susquehanna River — a reversal from previous techniques that showed improving trends.

USGS scientists say that with more studies, they hope the upgraded monitoring will soon help explain the reasons behind baffling trends seen in some of the Bay’s other major rivers. That understanding will be needed this year and next as the state-federal Bay Program partnership undertakes a “midpoint assessment” of the Chesapeake cleanup plan.

Historically, there were so few water monitoring sites — and they covered such large drainage areas with so many land uses — that it was often difficult, if not impossible, to explain why water quality trends were happening. The network, which expanded gradually over the last decade and a half, now provides more information about smaller areas, giving scientists hope that they will be better able to assess the impact of pollution control efforts.

In another upgrade, scientists are using a new analytical technique to estimate trends at many of those sites in “loads,” or actual amounts, of nitrogen, phosphorus and sediment moving downstream. The method should make it easier to assess progress toward meeting the regulatory requirements of the Chesapeake Bay Total Maximum Daily Load which — as that phraseology indicates — caps the amounts of those pollutants allowed in the water.

“This really is a first-of-its-kind network, at least in the United States, where you have a TMDL of this magnitude and a monitoring network like this that can actually look at trends in load,” Moyer said.

Employing the new analytical technique — “weighted regression on time, discharge and season,” or WRTDS — USGS scientists have improved the representation of pollutant measurements from their monitoring stations to better account for the disproportionately large amounts of nutrients and sediment flushed downstream during short-lived stormflow periods.

Previously, the USGS determined water quality trends by looking at the concentration of nutrients and sediment in water samples — but that technique more effectively shows changes during average river flows.

The difference between the two techniques was first evident during recent work on the Susquehanna. Concentration trends had shown improving water quality coming from the Bay’s largest tributary. But when scientists used the new technique, which better accounted for the impact of storms, they discovered that the actual pollutant load had been trending upward in recent years. That was the first clear indication to experts that the reservoir behind Conowingo Dam had essentially filled up and was no longer trapping sediment or nutrients coming downriver.

It takes 10 years of data before the WRTDS can estimate loads at a site, and it was only recently that the USGS had enough data from enough sites to switch over to the new methodology. About 80 sites in the network can provide load trends right now; the others will be coming online in the next several years. Still, that’s an increase from 30 sites that provided trends results a year ago.

Taken as a whole, recently released data show:

  • For nitrogen, 54 percent of the sites are improving, 22 percent are worsening and 19 percent have no detectable trend.
  • For phosphorus, 68 percent are getting better, while 20 percent are degrading and 12 percent show no real change.
  • For sediment, 50 percent of sampling stations show gains, 30 percent show declines and 20 percent have no trend.

The new data present new mysteries for scientists to unravel. While water quality at most upstream sites is improving, many of those trends disappear or are even getting worse where the Bay’s nine major rivers meet tidal waters, especially during the most recent decade.

Specifically, seven of nine major Bay rivers showed long-term improvements in nitrogen loads from 1985 through 2014. Yet only three showed improvements since 2005, and four lost ground.

For phosphorus, the story was worse, with only two rivers showing improvements in the most recent decade — compared with four over the longer 30-year time period — and four rivers trended worse.

The Potomac is the only river that showed both long– and short-term improvements for nutrients and sediments. The Patuxent River saw long– and short-term gains in nutrients, but not sediment.

For the Susquehanna, at least, it’s clear to scientists now that the dam’s failing trapping capacity is behind the worsening phosphorus levels and the lack of a nitrogen trend near the river’s mouth — even though the majority of upstream stations are improving.

“We think we have a pretty good handle on the Susquehanna because it has the confounding nature of the reservoir,” Moyer said.

He and others hope the new analytical techniques will help them decipher what is going on in other rivers as well. “There’s a lot of stories to sort out,” Moyer said.

Meanwhile, the upgraded monitoring network is producing other valuable data, such as the average rates of nutrient and sediment runoff per acre from land upstream of each sampling station. That information helps scientists weigh the relative effectiveness of pollution control efforts in a particular area.

For example, nitrogen runoff into the Conestoga River, which drains much of Lancaster County, PA, has declined by about 5 pounds per acre over the last decade — the greatest per-acre reduction anywhere in the Bay.

But the Conestoga is still receiving 33.4 pounds of nitrogen runoff per acre of land that drains into it, which is also one of the highest pollution rates in the watershed. (The average for the whole Susquehanna river system – of which the Conestoga is a part — is about 8.3 pounds per acre.)

Scientists said they hope the new data and more analysis will help with the midpoint assessment of the TMDL cleanup plan. States could use better information to understand which of their cleanup efforts will be most effective — and where they should be targeted — as they adjust pollution control strategies to go into effect in 2018.

“How can you more strategically place the practices in basins that are not improving up to this date, as well as in places where we really have the highest loads and we need to see the greatest amount of improvement?” asked Scott Phillips, USGS Chesapeake Bay coordinator. “That is where information like the pounds per acre results can really help.”

Beth McGee, a senior water quality scientist with the Chesapeake Bay Foundation, said that, overall, the data offer a glimmer of good news because pollution in most areas is going down — and it offers the potential to make cleanup efforts more effective in the future.

“Overall, we view this positively,” McGee said. “The majority of the stations that they looked at are improving. We think that is good news. The value of this will be when they are doing the next step, which is drilling down within these areas — both ones that are degrading and those that are improving — and trying to figure out why. What are we doing right? And what isn’t working?”

Rich Batiuk, associate director for science with the Environmental Protection Agency’s Bay Program Office, said the upgraded monitoring will allow scientists to better assess over time the full range of human activities that affect water quality — from land use changes to pollution control efforts to climate change — to help explain often-confounding water quality trends not only now, but in the future.

Besides providing new insights about targeting additional actions, Batiuk said the monitoring data also provides a more tangible way to explain to people the impacts of actions taken on the land, in contrast with the computer models used to develop cleanup plans and estimate progress.

“We hear farmers tell us that if we can show them data that they need to do more to help with downstream water quality, they’ll look at it,” Batiuk said. “But, they say, ‘Don’t come to me with modeling output.’ ”

Summaries of the nontidal monitoring data, as well as online tools that allow results to be explored in depth, can be found at the USGS Chesapeake Water-Quality Load and Trend Website: