Scientists Ponder: How Well Are AG Practices Helping The Chesapeake Bay?

Editor’s Note: State and federal leaders have acknowledged that the Chesapeake Bay region will not meet its most fundamental 2025 cleanup goal: reducing nutrient pollution in the Bay and its rivers. Now, many people are asking, “How did we get here?” and “What’s next?” This article is part of an ongoing series that tackles that question.

For 40 years, the Bay region has struggled to sufficiently reduce nutrient pollution from farms. The reasons are complex. But it’s important to explore those challenges as the region begins a tough conversation about the future of the Bay restoration effort beyond 2025.

Previous articles in this series discuss difficult trade-offs with agriculture, the challenge of setting realistic goals, the dearth of technical support for farm conservation projects and concerns about the ag data used in Bay computer models.

ANNAPOLIS, Md. – Once a month, someone from the U.S. Geological Survey drives through the narrow winding roads of Virginia’s Rockingham County to a small bridge near the mouth of Smith Creek. There, they collect what could be a lesson for the Chesapeake Bay cleanup effort.

It comes in the form of water samples carefully captured in a one-liter bottle. Attached to aluminum frame, the bottle is lowered and filled five to seven times across the width of the creek to make sure a full cross section of the waterway is captured.

“If you’re not able to sample that entire profile, we wouldn’t be accurately representing the chemistry of what’s happening here,” said James Webber, a USGS hydrologist who was demonstrating the technique on an early November day.

Along with the monthly samples, someone makes the trip out to the bridge during at least eight storms each year, because what flows by the bridge during storms is different from when water levels are low.

The samples reflect what is happening on Smith Creek’s 105-square-mile watershed, which stretches from the forested edge of Massanutten Mountain to the east, then spills west across rolling pastures and croplands and the small town of New Market.

The picture they have created over time provides a cautionary tale for the Chesapeake cleanup effort. Farmers in the Smith Creek watershed have been working hard to reduce water pollution from agriculture, using many of the “best management practices” or BMPs that are recommended by the state-federal Chesapeake Bay Program.

But the samples Webber and others have drawn do not show a decline in nutrient pollution. Instead, it has increased.

Smith Creek is not alone. Several other monitored watersheds across the Bay region also show that the amount of water-fouling nutrients reaching the Bay from farms has increased or remained steady in recent years despite the promotion and use of various BMPs.

Yet the regionwide effort to clean up the Bay has long hinged on the assumption that the widespread use of BMPs will achieve nutrient reduction goals aimed at improving Bay water quality. Whether that assumption is true is far from certain.

Studies in very small streams have found water quality and stream health improvements from BMPs such as fencing cattle out of streams, restoring forests along streambanks or planting nutrient-absorbing cover crops.

Indeed, some areas of the Smith Creek watershed improved even as overall nutrient trends worsened. Mountain Run, a small tributary, was an “impaired” stream because of the poor condition of bottom-dwelling organisms, but sediment reductions stemming from BMPs allowed them to rebound, and the listing was removed.

In general, though, demonstrating the effectiveness of BMPs on nutrient reductions in larger areas has proven elusive. Scientists have long cautioned that their real-world impact is unclear, especially in watersheds larger than just a few square miles.

That’s because the high number of constantly changing activities in larger areas makes it difficult to know with certainty what drives water quality trends. Forests may be cleared or planted. Farm animal numbers may increase or decrease. Farmland may turn into suburbs.

Those changes can dwarf BMP impacts, especially if relatively few are implemented. Further, there is often a “lag time” of years or even decades between a BMP’s installation and its impacts on water quality. And some BMPs simply might not work in a given setting.

Understanding the actual effectiveness of BMPs is critical to the Bay cleanup effort. It’s equally important to the farmers who are continually called upon to spend time and money installing them on behalf of the Bay. And it’s important to policymakers who help cover the costs with public funds.

In the last decade alone, state and federal agencies have spent more than $2 billion on programs to help farmers in the Chesapeake region install conservation practices. And spending is dramatically increasing as the 2025 deadline for the Bay’s cleanup goals approaches.

Officials in the state-federal Chesapeake Bay Program have acknowledged they will miss the nutrient pollution goal by a wide margin. But it is difficult to know whether current programs targeting agricultural areas — the largest source of nutrients to the Bay — are capable of ever reaching their targets or have even made significant progress.

“We lack a coordinated effort to further monitor, interpret and produce findings about the relation[ship] between agricultural conservation practices and water quality response,” wrote officials from several federal agencies in a December 2021 report.

Such a conclusion shouldn’t be a surprise because the same shortcoming has been identified for decades.

The good news is that the Bay Program this year will launch an effort to better assess the impact that BMPs have on streams in small watersheds. But it will likely take about a decade to know the answer.

No easy technological fix is available for farms, which are the largest sources of nutrients, in the form of nitrogen and phosphorus, to the Bay. There, they spur algae blooms and lead to oxygen-starved “dead zones” that are off limits to most aquatic life.

Instead, since the early 1980s, cleanup plans have relied on the widespread use of BMPs to control the runoff of fertilizer and manure from the region’s more than 80,000 farms.

The Bay Program recognizes more than 200 BMPs that can be credited toward meeting nutrient reduction goals. Each is assigned an estimate of its nutrient removal effectiveness.

Every year, states report how many BMPs are installed. Computer models use that information — and a wide swath of other data — to estimate the expected amount of nutrient pollution reduced annually.

As early as 2000, when the Bay Program missed its first nutrient reduction deadline, concerns arose that water quality monitoring was showing less cleanup progress than computer models had predicted.

The exact reasons were unknown. At a presentation to state and federal environmental officials, the chair of the Bay Program’s monitoring subcommittee made the case for making greater investments in small watershed research. “Every time you see results you don’t like, you blame it on lag times. Ten years from now, will you still be blaming lag times?” he asked.

Although monitoring was increased, it was not at scales small enough to pinpoint reasons for the outcomes.

Instead, the failure to meet cleanup goals was increasingly seen as a lack of adequate funding, not a lack of knowledge. A series of reports after 2000 from agencies, non-profit groups and others estimated that it would cost billions to achieve Bay goals — far more than was being spent.

The Chesapeake Bay Foundation, which did one of the cost studies, summed up the message as, “we know what we need to do, we just need to do it.” At the time, it sometimes added, “we don’t need more research.”

Nonetheless, evidence was mounting that BMPs might not deliver the expected results. A 2003 review concluded that the nutrient removal effectiveness of many practices was less than the Bay Program credited. Millions of pounds of estimated cleanup “progress” was erased as a result.

The following year, a report from the Bay Program’s Scientific and Technical Advisory Committee, or STAC, warned that BMP effectiveness was likely still overstated and called for more research.

Tom Simpson, a retired soil scientist with the University of Maryland who led the STAC report, said there was always reluctance to provide significant funding for such work.

“I think the Bay Program really was, and probably still is, the best watershed program that we’ve been able to put together in this country,” Simpson said. “But we tended to feel we had all the answers.”

The result was that the Bay Program increasingly created a system that equated spending with progress — the more wastewater treatment plants upgraded and the more BMPs funded, the greater reductions calculated in its computer models. That was true for wastewater upgrades, where reductions could be measured at the end of a pipe. But there was no monitoring system that could clearly link BMP implementation with water quality improvements.

“The political pressure has been that we don’t need any more science. We know what to do, we just need to go out and implement the solutions,” a USGS scientist told the Bay Journal in 2008. “I appreciate that. But you still need science to look at what you are doing to determine its effectiveness from a management standpoint.”

That’s why scientists have often emphasized the need to monitor small watersheds. At a smaller scale, BMPs can be ramped up, and it’s easier to assess which factors might influence nutrient levels, such as land use changes, increases or decreases in farm animals, and myriad other activities.

The USGS oversees a 123-site monitoring network within the Bay’s 64,000-square-mile watershed, but the network is more geared toward assessing trends than understanding what drives them.

Smaller watersheds in that network generally cover 50–1,000 square miles. The impact of 600 acres of nutrient-absorbing cover crops in drainage areas of that size would be overwhelmed by other activities. It would be much easier to detect the impact of 600 acres of cover crops in, for instance, in a watershed that’s only 6,000 acres, or about 10 square miles.

Further complicating the detection of small changes is that water quality is highly influenced by the weather: the more rain, the more runoff and nutrients in streams. As a rule of thumb, USGS scientists say it takes about 10 years of monitoring to account for annual fluctuations in streamflow and detect a nutrient trend.

Still, in very local, tightly controlled studies — typically where drainage areas are measured in acres, not square miles — BMPs have been effective for reducing nutrients and improving stream health.

For decades, the fields at the University of Maryland’s Wye Research and Education Center on the Eastern Shore have produced lush crops of grass each fall. Scientists there have proven that planting rye cover crops on those well-monitored farms in the fall can absorb much of the nitrogen left in fields late in the year, reducing the amount that reaches underlying aquifers by 45%.

The research also shows how other factors, such as the type of grass or grain grown as a cover crop, if the fields are plowed or not, and whether the fields get nutrients from manure or chemical fertilizer, can affect the results.

But such detailed, long-running studies are rare, said Ken Staver, a scientist who has been overseeing the project for decades. Many are done on small plots under carefully controlled conditions, which may not reflect real-world management, and for short periods of time.

“There’s tons of studies where people go out and do something for a little while and then pull out,” Staver said. “But any kind of long-term dataset where you’ve sort of captured the variability and climate conditions — there are just not that many.”

In the 1,779-acre branch of the Green Run watershed at the headwaters of the Eastern Shore’s Pocomoke River, a study conducted by the Maryland Department of Natural Resources in the late 1990s replaced the chicken manure used to fertilize farm fields with easier-to-apply chemical fertilizer.

That reduced nitrogen applications by half, because much of the nitrogen in manure is not in forms readily available to plants and remains in the soil. The project also planted cover crops in the fall.

The experiment resulted in a 30% reduction in nitrogen into the local stream, compared with a small branch of Green Run where farms maintained business as usual. Phosphorus exports remained unchanged though.

It wasn’t clear, however, how much of the improvement was driven by the change in fertilization and how much by cover crops.

“It was hard to know why the nitrogen numbers went down, because you did two things at once,” said Staver, who was involved with follow-up studies at the site.

A variety of other studies have shown that BMPs such as fencing livestock out of waterways or planting streamside buffers are highly effective at improving stream health and reducing bacteria levels, even if nutrient trends are uncertain.

In 2010, the U.S. Department of Agriculture partnered with the USGS to establish three “showcase watersheds,” including Smith Creek, where BMP implementation would be ramped up thanks to an influx of Bay funding in the 2008 Farm Bill. Stream monitoring would assess the impact.

Since then, the rate of BMP implementation increased fourfold in the Smith Creek watershed. But monitoring shows that the total amount of nutrients and sediment nonetheless increased in the last decade.

A number of possible reasons exist, noted Webber of the USGS. For example, many of the BMPs were not considered “high impact” practices for nutrients.

But the biggest reason may be that the BMPs were simply outweighed by a growing number of cattle and chickens in the watershed, which increased the amount of manure being generated.

Webber said he did not think lag times fully explained the lack of improvement. Work by the USGS suggests that the average age of groundwater in the Smith Creek watershed appears to be 10 years, he said. Also, some of the nutrient increases were occurring mostly during high flows, which are mainly fed by surface runoff, not groundwater.

Smith Creek isn’t unique. Results from the two other showcase watersheds showed that most nutrient and sediment loads did not improve during the last decade.

The Upper Chester River, which drains part of Maryland and Delaware on the Eastern Shore, shows increases in phosphorus and sediment, with no trend in nitrogen.

Conewago Creek in Pennsylvania shows decreasing sediment while nutrient trends show a mix of increases, decreases and no change at different places. Decreasing trends might be related to a wastewater plant upgrade, according to Webber.

On Maryland’s Choptank River, Tom Fisher of the University of Maryland Center for Environmental Science has also had a difficult time detecting the influence of BMPs in a long-running study of the largely agricultural watershed.

Over nearly two decades, he and his colleagues intensively monitored 15 smaller subwatersheds. They found increasing or stable nitrogen trends in 62% of them and increasing or stable phosphorus trends in 96% — even though state and federal BMP programs have been heavily promoted and implementation increased.

In a separate study, Fisher and his team funded additional BMPs in three sub-watersheds for several years. They saw improvements in two, but the changes were smaller than expected.

Fisher said the findings suggest that the BMPs were less effective than hoped. “My impression is that we don’t have enough of the right BMPs, and they’re not positioned in the best possible places to remove nutrients,” he said.

Jim Lewis, a farmer and Maryland cooperative extension agent who worked with Fisher on his studies, said that in some cases, BMPs were almost certainly as effective as assumed, but in others they likely were not.

Sometimes, cover crops were planted too late to be effective or not seeded at densities needed to maximize their impact. Also, he said it was hard to get large numbers of farmers to actively participate. Many BMPs have little direct benefit to them, and even when implementation costs are covered, they often require more work and can reduce income, as things like buffers take land out of production.

“You can get one farmer to participate, and then they could do everything great,” Lewis said. “But if the neighbor doesn’t, then that counteracts what the one farmer did well.”

Certainly, lag times are causing some delay, although their importance and duration varies from place to place and are different for nitrogen and phosphorus. It’s also possible that most places don’t have enough BMPs in use to make a definitive impact. If those are the major factors, then moving forward with existing programs and increasing the number of BMPs should eventually improve water quality.

“From a management perspective, the best-case scenario is that BMPs are working, but lag times and monitoring limitations are delaying and/or masking a water quality response,” a STAC report said last year. But, the report added, “the evidence suggests that BMPs and policies designed to implement those BMPs are not as effective as expected.”

The Bay Program’s estimates of BMP effectiveness are based on the best professional judgement by teams of experts. But the STAC report said there are often important gaps in the studies available to support their decisions. For instance, nutrient movement lost to surface runoff is easier to research, and typically better studied, than nutrients that sink into the groundwater, which is the primary way nitrogen leaves fields. And studies are often limited to certain soil types or geographic settings and may not capture the full range of climate variables.

In some cases, the report said, BMPs may not be well-implemented and maintained, decreasing their effectiveness. In other cases, they may not be installed in places where nutrient problems are the greatest. Also, many BMPs have relatively low

nutrient removal effectiveness — some highly effective practices, such as streamside buffers, have lower adoption rates because they take land out of production, which hurts farm income.

Climate change could also be offsetting BMP effectiveness as the intensity and frequency of storms increases. Storms can overwhelm many practices, minimizing their nutrient removal impact.

Zach Easton, a Virginia Tech professor who worked on last year’s STAC report, said that climate could be contributing to the increase of nutrients in Smith Creek, where he has also done work, as the intensity of storms in the area has increased in the last decade.

And he said that, in areas with large amounts of animal agriculture, the supply of nutrients from manure and fertilizer outstrips what’s removed in farm products such as meat, milk, grain, vegetables or fruit, leaving a major excess on the landscape.

“The mass balance can serve as sort of a masking effect for BMP impacts on water quality,” Easton said. The showcase

watersheds, for instance, are in intense animal agriculture areas, he said, “making it incredibly difficult to detect a BMP signal, even if they are effective.”

Overcoming that is difficult because the economic realities of farming, and a growing population, drives increased production — and therefore increased nutrient demand.

“The main thing you are countering is not ignorance or evil. It’s market forces,” Staver said. “It’s getting practices on the ground at high enough levels to make a difference. Why do we have polluted water? Because market forces encourage behavior that leads to nutrient losses.”

Bay Program computer models illustrate how challenging a task that would be. Since the latest nutrient reduction goals were set in 2010, more farm acres were treated with some type of pollution control practice than in the previous 25 years.

Still, recent computer modeling shows that, regionwide, farms were sending more nitrogen to the Bay at the end of 2022 than when the goals were set. That’s partly because the number of farm animals increased, as did the amounts of fertilizer used to fuel increased crop productivity.

Those figures are disputed by many in the agricultural community, who question some of the data in the model, including fertilizer figures, and say the number of installed BMPs is greatly undercounted.

If correct, though, the figures indicate that of the 71.5 million pounds of nitrogen reductions needed to meet Bay goals, only 24 million pounds had been achieved through 2022. And almost all of those came from wastewater treatment plant upgrades.

About 90% of future nutrient reductions are expected to come from farms. But if more than $2 billion was required over the last decade to simply hold the line, it raises questions about how long it will take to reach the goals and whether they can even be attained.

Last year’s STAC report cautioned that simply providing more funds for existing programs is “unlikely to produce the intended nutrient reduction outcomes.”

It said programs should be changed, but that the Bay Program lacks critical information, including monitoring data, needed to adapt policies.

The issue is of paramount importance as frustration builds over the region’s failure to meet cleanup goals, but it’s not a problem that was unforeseen.

In 2011, the National Academy of Sciences warned that the Bay Program could face a “disillusioned public” if it was not able to explain how BMPs were affecting water quality and called for a small-scale monitoring program to resolve those uncertainties.

In a report, in underlined type, they stated, “A major challenge identified by the team was the need for enhanced monitoring at finer scales to better connect implementation of management practices with water quality and sediment changes in the Chesapeake watershed.”

This year, the agencies expect to launch such research in five small watersheds — generally 10 square miles or less — where BMPs will be increased and water quality closely assessed.

“We didn’t want to go somewhere that was already saturated with implementation and we weren’t going to see a change,” said Lee McDonnell, chief of the science, analysis and implementation branch in the EPA Bay Program Office. “We wanted to be able to see what happened when change occurred in the watershed.”

The project will incorporate help from others, including state agencies, conservation districts, universities, watershed groups and citizen monitoring programs. The hope is that the partnerships will lead to complementary studies that provide more detail about what is happening.

Citizen monitors, for instance, might be able to collect a series of water samples from a single storm event at different places in the watershed.

“One of the things I’m excited about is the community science aspect,” McDonnell said. “Getting that community involvement,

and hopefully that spurs more awareness, more stewardship and maybe brings more BMP money into that area depending on what’s going on.”

He and others hope the work spurs efforts to launch other small-scale projects in the Bay watershed. There is often little trust placed in computer model results, but a much higher level of confidence in monitoring data.

“If we see success and we’re working in partnership with the producers and the watershed groups, hopefully that drives confidence and implementation,” said Ken Hyer, acting coordinator for USGS’ Chesapeake Bay efforts.

The results will take time — several years at the least — but it may, at last, begin to answer a question that has loomed over the Bay effort for decades.