Farmland on MD Eastern Shore aerial

The EPA’s official tallies of nutrient pollution to the Chesapeake Bay from agriculture are based on computer model estimates, not direct measurements. 

The reductions credited by the U.S. Environmental Protection Agency over the past decade, even in Maryland and Virginia, are predominantly from improved wastewater treatment. Most states would have to accomplish unprecedented reductions from agricultural sources over the next five years to meet their commitments. 

I was one of the scientists who participated in the five-year study during the late 1970s that led to the focus on reducing nutrient pollution as key to the recovery of the Bay ecosystem. I was around to see the Chesapeake Bay Program partnership miss its 2000 deadline, and then its 2010 deadline for achieving those reductions. All hands must redouble their efforts to avoid missing the third deadline, but, as a data geek, I have to point out another elephant in the room. 

How confident are we that the actions taken will, in reality, result in the nutrient reductions that are credited? When will the Bay actually see their impact? These are not easy questions to answer. 

Loads of nitrogen and phosphorus from wastewater treatment facilities are well-monitored, giving us confidence that those reductions have been achieved. Estimated loads from other sources are based on reported on-the ground runoff control actions, assumptions about their effectiveness in reducing nutrient losses, and computer simulations of how those actions ultimately affect the amount of nutrients reaching the Bay’s tidal waters. The EPA’s official tallies of nutrient loads from developed lands and agriculture are based on those computer model estimates, not direct measurements.

The model also assumes those actions quickly reduce loads entering the Bay. It does not include multiyear “lags” in delivery as some nutrients linger in soils or move slowly through groundwater. Consequently, the computer model load estimates are difficult to verify with actual monitoring.

Reconciling model estimates with real-world observations is not a new concern. In a 2004 front page Washington Post article, Bay Pollution Progress Overstated, alleged that estimates derived from computer models were based on overgenerous assumptions and did not agree with water quality monitoring data. The next year, the U.S. Government Accountability Office faulted the Bay Program for comingling various kinds of data such as monitoring data, results of program actions and the results of its predictive model.

In 2009, the EPA requested that the National Academy of Sciences evaluate the Bay Program’s accounting of nutrient management practice implementation. It was unable to quantify the likely magnitude or even the likely direction of errors in the Bay Program’s tracking of nutrient reductions resulting from management actions in the watershed.

In May 2020, scientists from the U.S. Geological Survey published an important paper on factors driving nutrient trends in streams of the Bay watershed. They also used a computer model, but one based on the statistical patterns and trends in water quality monitoring at locations throughout the watershed. The estimated loads are normalized for variations in streamflow, but are influenced by whatever lags in management effects acting in the watershed.

Based on this empirical estimation, the USGS scientists concluded that over the 20 years between 1992 and 2012 the declining nutrient fluxes to the Bay were mainly due to wastewater treatment upgrades. The trends also reflected widespread declines in atmospheric nitrogen inputs. On the other hand, their empirical model showed essentially no change in nitrogen loads and increasing loads of phosphorus emanating from agriculture, despite considerable efforts to control nutrients. Surprisingly, estimated fluxes from urban areas declined slightly for nitrogen, despite ongoing urbanization.

In the Bay Program’s management model, the reductions in wastewater loadings in both nitrogen and phosphorus are generally similar to those in the empirical model. This is not surprising, because the discharges are monitored and directly enter surface waters, without appreciable lag times. But the management model shows reductions of both nitrogen and phosphorus from agricultural sources of 24% and 45% respectively from 1985 to 2019. And, it shows loads from developed lands increasing by about 45%, opposite the declining trend suggested by the empirical model.

I am not suggesting that one modeling approach is right and the other is wrong. There are a number of possible reasons for the discrepancies between the Bay Program’s watershed management model and the USGS empirical model. There are, of course, methodological differences in how the two models simulate the way nutrients are transported to the Bay, including the effects of climatic variability. The management model does not include lag times so on-the-ground actions result in immediate reductions. But those reductions may not yet be fully evident in the observed concentrations used by the empirical model. However, it is also possible that the agricultural practices credited in the management model have not been implemented to the extent reported by the states or they may not be as effective as assumed. 

I have been a nag the past 25 years about the need to reconcile model estimates of reductions with real-world observations because it is critically important in order to adapt practices to achieve desired results. There are only five years left in what, starting with the 1987 Bay Agreement, will be a 38-year effort to reduce nutrient pollution to the Bay. It is past time to verify what the credited management actions will achieve in reality, and in what time frame. I am heartened that USGS and Bay Program scientists are working hard to sort out the degree to which differences in the estimations are attributable to lags in response or erroneous assumptions on the effectiveness of actions that have been, and will be, taken.

This is no time to give up on meeting the Bay nutrient reduction targets, as challenging as they may be. In a flood of new papers, my scientific colleagues have documented improvements in water quality and living components of the ecosystem as a result of those nutrient reductions that have actually been measured. This gives us confidence that substantial further improvements will result if and when the reductions needed to meet Bay cleanup goals are actually achieved.

Come 2025, if the Bay Program partnership falls short, it will have to consider more effective approaches. This is particularly true for agriculture, which is responsible for the largest portion of remaining nitrogen and phosphorus loads to the Bay in both management and empirical models. In a recent review of campaigns to reduce coastal nutrient pollution around the world, I found that substantial agricultural load reductions could only be documented where nutrient management was regulated or public subsidies depended on meeting performance standards. Reconciling management assumptions with empirical observations should inform such future pathways. n

Donald F. Boesch, the retired president of the University of Maryland Center for Environmental Science, is a member of the Bay Journal’s Science Advisory Committee.

The views expressed by opinion columnists are not necessarily those of the Bay Journal.

(0) comments

Welcome to the discussion.

We aim to provide a forum for fair and open dialogue.
Please use language that is accurate and respectful.
Comments may not include:

* Insults, verbal attacks or degrading statements
* Explicit or vulgar language
* Information that violates a person's right to privacy
* Advertising or solicitations
* Misrepresentation of your identity or affiliation
* Incorrect, fraudulent or misleading content
* Spam or comments that do not pertain to the posted article
We reserve the right to edit or decline comments that do follow these guidelines.