Scientists have begun applying a new technique to help them more accurately estimate the amount of nutrients and sediment that run off the watershed to the Bay, and determine the trend.
The new method reveals that the total amount — the load — of phosphorus reaching the Chesapeake from most of its nontidal tributaries has either been increasing, or shown no change, over the last 25 years. No river showed improvement over the last decade.
That's different from previously reported monitoring data, which generally showed decreasing concentrations of phosphorus in the rivers' waters.
Both these seemingly conflicting trends are accurate. While the phosphorus trend has actually been improving most of the time, the total amount reaching the Bay from most rivers has also increased.
Don't be. There's a reason for the seeming contradiction, which U.S. Geological Survey scientists have been explaining to state and federal officials at recent Bay Program meetings.
It has to do with the difference inherent in measuring two different things in the water: concentration, which is the amount of a nutrient in a given unit of water, and load, which is the total quantity that reaches the Bay.
For years, scientists from the USGS have monitored nitrogen, phosphorus and sediment where the largest nine tributaries reach the tidal waters of the Bay. Using a statistical model, known as Estimator, they have calculated nutrient and sediment concentrations annually and trends over time.
But the goal of the Chesapeake Bay Total Maximum Daily Load, or pollution diet, is to reduce the total amount, or load, of nutrients and sediment entering the Bay. To meet that need, USGS scientists developed a new statistical technique, "Weighted Regression on Time, Discharge and Season," or WRTDS.
Estimator and WRTDS provide slightly different interpretations of the same data.
The reason, scientists said, is relatively simple. Estimator places more emphasis on changes in concentration, and gives more importance to the normal-to-low flows that persist the vast majority of the time in the Bay's free-flowing tributaries.
Because WRTDS is estimating total loads, it does a better job of incorporating the impact of large storms. While less frequent, storms move large amounts of nutrients and sediment downstream. They particularly impact sediment and much of the phosphorus, which is attached to sediment particles. High flows that accompany larger, more powerful storms, move disproportionately large amounts of sediment and phosphorus downstream. That disproportionate surge is better captured by WRTDS.
Therefore, while in-stream concentration measurements show improving trends in most places — which is true the majority of the time — the disproportionately large amounts of phosphorus and sediment reaching the Bay during less frequent storms has resulted in an increasing total load trend for many rivers.
"The models both see it correctly," said Doug Moyer, a USGS hydrologist. "They are just seeing different aspects of how water quality conditions are changing.
"Both trends are very informative, and together they can really provide you with what has been the water quality response based on changes in the watershed," Moyer said.
During low flows, the overall phosphorus trend, viewed as concentrations, has improved, the USGS data show. That's in large part due to a reduction in orthophosphorus, which is a dissolved form of phosphorus. That reduction likely represents the impact of wastewater treatment plant upgrades and the phosphate detergent ban, according to the USGS.
But total phosphorus loads have increased because most phosphorus is attached to sediment, which moves during higher flows. Relating those trends to specific management efforts is difficult because it's unclear how long it takes sediment, and the associated phosphorus, to work its way downstream and into the Bay. While big storms tend to move a lot of sediment and phosphorus, they don't always move it very far before it settles to the river bed or gets deposited on floodplains or in reservoirs, awaiting the next big storm to carry it another step toward the Chesapeake. That's a journey that can take years, or even decades.
(To a degree, the pattern also exists for nitrogen because some forms of that nutrient are particulates. When scientists look only at nitrate, which is all dissolved, they see better trends than when they look at the total amount of nitrogen reaching the Bay. But because the vast majority of nitrogen — unlike phosphorus — is dissolved, the difference is less pronounced.)
"Things that are related to sediment and particulates are slower to change in response to management actions in the watershed because of those lag effects," said Robert Hirsch, a research hydrologist with the USGS who was largely responsible for developing the WRTDS model. "There are a lot of legacy issues."
The total loads — which include both dissolved and particulate forms of nutrients — may not reflect more recent management actions because of the influence of slower moving particulates and sediment, he said.
The new technique to estimate total loads has already provided profound insights into nutrient trends. When applied to the Susquehanna River earlier this year, it revealed that phosphorus loads have been trending upward over the last decade. Those increases appear to be driven by greater amounts of phosphorus leaving the reservoir behind Conowingo Dam. That phosphorus appears to be pushed past the dam during very high flow events, which may be scouring sediment and phosphorus that have been building up in the reservoir behind the dam for decades. As the reservoir fills, it is less able to trap sediment and phosphorus.
"Conowingo is an example of how the new WRTDS statistical method works," Hirsch said. "Without it, we would not have caught the trend that is taking place there."
But the load technique cannot be used everywhere; it requires at least 20 years of monitoring data to capture enough storms and flow conditions to develop reliable trend estimates.
The USGS is applying the load trends analysis first to the Bay Program partnership's river input monitoring network, which measures what's reaching the Bay's tidal waters from its nine largest tributaries. Those monitoring stations have the longest water quality data records, in some cases reaching back to the 1970s. Together, those sites reflect what is happening on nearly 80 percent of the watershed.
While some of the details are different between estimates from the concentration and load techniques, both Estimator and WRTDS reflect a disconcerting trend: In the last 10 years, both show the rate of nitrogen and phosphorus improvement has generally slowed in and around the watershed.
Scientists hope that as they apply and compare both techniques to more monitoring sites, they'll gain a better understanding of what factors drive the trends they are observing.
A report explaining the two approaches, "Comparison of Two Regression-Based Approaches for Determining Nutrient and Sediment Fluxes and Trends in the Chesapeake Bay Watershed," along with a short summary paper, will be posted on the USGS Chesapeake Bay website chesapeake.usgs.gov in early January.