Achieving Water Quality Goals in the

Chesapeake Bay: A Comprehensive

Evaluation of System Response

An Independent Report from the Scientific and Technical

Advisory Committee (STAC)
Chesapeake Bay Program
Annapolis, MD

May 2023

Scientific and Technical Advisory Committee (STAC)

About the Scientific and Technical Advisory Committee

The Scientific and Technical Advisory Committee (STAC) provides scientific and technical
guidance to the Chesapeake Bay Program (CBP) on measures to restore and protect the
Chesapeake Bay. Since its creation in December 1984, STAC has worked to enhance scientific
communication and outreach throughout the Chesapeake Bay watershed and beyond. STAC
provides scientific and technical advice in various ways, including (1) technical reports and
papers, (2) discussion groups, (3) assistance in organizing merit reviews of CBP programs and
projects, (4) technical workshops, and (5) interaction between STAC members and the CBP.
Through professional and academic contacts and organizational networks of its members, STAC
ensures close cooperation among and between the various research institutions and
management agencies represented in the watershed. For additional information about STAC,
please visit the STAC website at http://www.chesapeake.org/stac.

Publication Date: May 9, 2023

Publication Number: 23-006

Suggested Citation:
Scientific and Technical Advisory Committee (STAC). (2023). Achieving water quality goals in the
Chesapeake Bay: A comprehensive evaluation of system response (K. Stephenson & D. Wardrop,
Eds.). STAC Publication Number 23-006, Chesapeake Bay Program Scientific and Technical
Advisory Committee (STAC), Edgewater, MD. 129 pp.

i
Report Editors
Kurt Stephenson, Virginia Tech
Denice Wardrop, Chesapeake Research Consortium and Penn State University

Writer Team
William Dennison, University of Maryland Center for Environmental Science
Zachary Easton, Virginia Tech
Mark Monaco, National Oceanic and Atmospheric Administration, National Centers for Coastal
Ocean Science
Kenneth Rose, University of Maryland Center for Environmental Science
Leonard Shabman, Resources for the Future
Kurt Stephenson, Virginia Tech
Jeremy Testa, University of Maryland Center for Environmental Science
Denice Wardrop, Chesapeake Research Consortium and Penn State University

Steering Committee
Brian Benham, Virginia Tech
Anthony Buda, U. S. Department of Agriculture, Agricultural Research Service
William Dennison, University of Maryland Center for Environmental Science
Zachary Easton, Virginia Tech
Ellen Gilinsky, Ellen Gilinsky, LLC
Andy Miller, University of Maryland Baltimore County
Mark Monaco, National Oceanic and Atmospheric Administration, National Centers for Coastal
Ocean Science
Kenneth Rose, University of Maryland Center for Environmental Science
Leonard Shabman, Resources for the Future
Kurt Stephenson, Virginia Tech
Jeremy Testa, University of Maryland Center for Environmental Science

Contributing STAC Members

William Ball, Johns Hopkins University
Lee Blaney, University of Maryland Baltimore County
Craig Beyrouty, University of Maryland
J. K. Böhlke, U. S. Geological Survey
Kathy Boomer, Foundation for Food and Agriculture Research
Charles Bott, HRSD
Chris Brosch, Delaware Department of Agriculture
Katherine Bunting-Howarth, New York Sea Grant and Cornell Cooperative Extension
Shirley Clark, Penn State University
Amy Collick, Morehead State University
Alix Dowling Fink, Longwood University
Celso Ferreira, George Mason University
Lara Fowler, Penn State University
Deidre Gibson, Hampton University
ii
Kirk Havens, Virginia Institute of Marine Science
Ben Hayes, Bucknell University
Jason Hubbart, West Virginia University
Thomas Ihde, Morgan State University, Patuxent Environmental & Aquatic Research Laboratory
Thomas Johnson, U. S. Environmental Protection Agency
Brandon Jones, National Science Foundation
Hamid Karimi, District of Columbia Department of Energy and Environment
Jennifer Keisman, U. S. Geological Survey
Elliot Kellner, U. S. Department of Agriculture, Natural Resources Conservation Service
Catherine Kling, Cornell University
Scott Knoche, Morgan State University, Patuxent Environmental & Aquatic Research Laboratory
Ellen Kohl, St. Mary’s College of Maryland
Erin Letavic, Herbert, Rowland & Grubic, Inc.
Martin Lowenfish, U. S. Department of Agriculture, Natural Resources Conservation Service
Chanceé Lundy, Nspiregreen, LLC
Dave Martin, The Nature Conservancy
Gregory Noe, U. S. Geological Survey
Efeturi Oghenekaro, District of Columbia Department of Energy and Environment
Leah Palm-Forster, University of Delaware
Michael Runge, U. S. Geological Survey
Larry Sanford, University of Maryland Center for Environmental Science
Adel Shirmohammadi, University of Maryland
Eric Smith, Virginia Tech
Jay Stauffer, Penn State University
Tess Thompson, Virginia Tech
Leon Tillman, U. S. Department of Agriculture, Natural Resources Conservation Service
Lisa Wainger, University of Maryland Center for Environmental Science
Gene Yagow, Virginia Tech
Weixing Zhu, Binghamton University

Acknowledgements
Carl Hershner, Virginia Institute of Marine Science
Lew Linker, U. S. Environmental Protection Agency
Gary Shenk, U. S. Geological Survey
Peter Tango, U. S. Geological Survey
Qian Zhang, University of Maryland Center for Environmental Science

STAC Staff
Meg Cole, Chesapeake Research Consortium
Tou Matthews, Chesapeake Research Consortium

iii
Executive Summary

In 1983, the governors of Maryland, Pennsylvania, and Virginia, the mayor of District of
Columbia, and the U.S. Environmental Protection Agency (EPA) administrator signed the first
Chesapeake Bay Agreement. The one-page agreement acknowledged the “historical decline in
the living resources of the Chesapeake Bay” and committed to addressing a major cause of the
decline by pledging “to fully address the extent, complexity, and sources of pollutants entering
the Bay.” Subsequent Bay agreements have expanded the number of partners and the number
of restoration goals, but reducing two key pollutants, nitrogen (N) and phosphorus (P), has
remained a centerpiece of every subsequent Bay agreement.

Over four decades, water quality and pollutant reduction goals have been established and
refined. Under authority provided by the Clean Water Act, the Bay jurisdictions and EPA
adopted Bay water quality standards in 2003. The water quality standards identified living
resources as the designated use of the Bay and defined numeric water quality criteria deemed
necessary to support the designated use. Numeric water quality criteria were set for dissolved
oxygen, water clarity, and chlorophyll a across five different Bay habitats: shallow water
(submerged aquatic vegetation [SAV]), open water (fish and shellfish), deep water (seasonal
fish and shellfish), deep channel (seasonal refuge), and migratory fish spawning and nursery.

Nutrient reduction goals were first written into the 1987 Bay agreement (and quantified in
1992 amendments). When nutrient reduction efforts failed to attain Bay water quality
standards, EPA developed the country’s most expansive total maximum daily load (TMDL) in
2010. The TMDL set nutrient and sediment load targets for the Bay that, if met, were predicted
to achieve the water quality standards. The TMDL established that all management actions
needed to achieve the target pollutant loads (214.9 million lb/yr of N, 13.3 million lb/yr of P,
and 18,587 million lb/yr of sediment) should be in place by 2025. The Bay states and District of
Columbia wrote watershed implementation plans (WIPs), which were approved by EPA,
describing approaches to reduce nutrients and sediment to meet the load targets.

There has been progress in addressing nutrients since the first Bay agreement. The Chesapeake
Bay Program (CBP) watershed model estimated that N loads to the Bay were reduced from 370
million lb/yr in 1985 to approximately 258 million lb/yr in 2021 and that P loads were reduced
from 29 million lb/yr in 1985 to approximately 15 million lb/yr in 2021. Wastewater treatment
plant upgrades provided the majority of these reductions. According to CBP estimates, the
TMDL sediment limits have been met. Achieving these pollutant reductions in the face of
significant population growth and development throughout the watershed is a noteworthy
accomplishment.

However, modeling and monitoring evidence indicates that current efforts to reduce nutrient
loads will not meet the TMDL targets. In addition, the CBP’s ambient water quality monitoring
program indicates that estuary water quality has been slow to respond to realized nutrient and
sediment reductions in many regions of the Bay. The CBP has estimated that 27% of the Bay

iv
area met the water quality standards in 1985. By 2020, that figure had only risen to the mid-
30% range. The consequences for living resources have not been fully evaluated.

This report summarizes the Scientific and Technical Advisory Committee (STAC) evaluation of
why progress toward meeting the TMDL and water quality standards has been slower than
expected and offers options for how progress can be accelerated. The report evaluates the
effectiveness of current actions to reduce pollutants (N, P, and sediment) from wastewater
treatment point sources and from farms and developed lands (nonpoint sources). Chapter 4
provides results from the evaluation of the water quality response in the estuary (dissolved
oxygen, water clarity/SAV) to the realized nutrient and sediment reductions. Finally, chapter 5
summarizes what is known about the response of fish, shellfish, and other living organisms to
changed water quality conditions. Decision-relevant uncertainties at each stage of program
implementation and assessment are identified and their implications for progress considered.

Three overarching conclusions emerged from these evaluations. First, achieving pollutant
reduction and water quality improvements is proving more challenging than expected. Second,
the Bay system faces permanent and ongoing changes in land use, climate change, population
growth, and economic development that will challenge notions of restoration based on
recreating historical conditions. Third, opportunities to meet these challenges exist but efforts
require changes and new approaches to implementation, planning, and decision-making.
Specific findings of this evaluation supporting these conclusions and associated policy
implications are summarized here.

Achieving the pollutant targets of the Bay TMDL

Finding: Existing implementation actions to reduce nonpoint sources of nutrients are

insufficient to achieve the TMDL.
Meeting the TMDL depends to a significant degree on reducing nonpoint sources of pollutants.
Agriculture is the largest remaining source of nutrient loads to the Bay, and urban nonpoint
sources are the fastest growing. To date, CBP partner efforts to reduce nonpoint sources of
nutrients have not produced sufficient levels of best management practice (BMP)
implementation to meet the TMDL, and the implementation that has occurred may not be
producing the pollutant reductions expected.

The CBP acknowledges the challenges of generating enough nonpoint source BMP adoption to
meet nutrient reduction goals, particularly for N. Tens of millions of pounds of N reductions are
needed to achieve the TMDL goal, but a decade of implementation since 2010 has produced
only 3 million lb/yr of nonpoint source N reductions (as estimated by the CBP watershed
model). The difference between water quality practices implemented and practices needed is
termed an implementation gap and has multiple potential causes. Nonpoint source incentive
programs are generally designed to encourage voluntary adoption of BMPs by covering a
portion of the costs of installation. While successful at encouraging the adoption of practices
that generate benefits to landowners (e.g., enhanced soil productivity), such programs do not
provide sufficient incentives for adoption of practices with the largest pollutant reduction

v
potential. Evidence also suggests that nutrient load reduction gains that have come from BMP
implementation efforts are being partially offset by regional increases in imported nutrients.
For example, increases in livestock numbers mean more N and P are imported into a region in
the form of animal feed without corresponding increases in exports of animal products or by-
products (i.e., manure). This nutrient mass imbalance leads to an accumulation of nutrients in
the watershed that in turn may be transported to the Bay in runoff.

Evidence also suggests that the nonpoint source pollutant control efforts may not be as
effective at producing nutrient reductions as expected by the CBP, resulting in a response gap.
The existence of a response gap means that less progress is being made in meeting TMDL
pollutant targets than represented by current accounting systems, and more nonpoint source
controls will be needed to produce needed pollutant reductions. The response gap for
phosphorus may be particularly large. While CBP modeling suggests that P reductions targeted
by the TMDL are nearly achieved, analysis of water quality at riverine monitoring stations finds
limited evidence of observable reductions in P concentrations. Nutrient response gaps have
many potential causes, including long lag times for actions taken on the ground to produce
reductions at water quality monitoring stations. However, response gaps could have a variety of
other causes, including incomplete understanding of how people use nutrients on the
landscape (particularly animal manures), overestimating nonpoint source practice
effectiveness, incomplete or inaccurate information about nutrient inputs, landscape changes,
and insufficient monitoring. Identifying and addressing response gaps is challenging, and this
challenge is exacerbated by the TMDL accounting framework that tasks water quality managers
with counting practices implemented and thereby diverts attention from the question of
whether those practices generate the predicted pollutant reductions.

Together the implementation and response gaps represent significant challenges to the CBP’s
ability to achieve the nonpoint source pollutant load reductions as required by the TMDL.
Uncertainty and complexity of nonpoint source-generating behaviors and processes confound
assessment of these gaps. These challenges are not unique to the CBP, with many large-scale
eutrophication management efforts (e.g., Great Lakes, Gulf of Mexico, Baltic Sea) also facing
similar challenges for reducing watershed-scale nonpoint source pollutants.

Policy implication: There are opportunities to further reduce nutrients from nonpoint sources,
but changes to programs and policies need to be considered.
Additional funding of existing implementation efforts is unlikely to produce the intended
nutrient reduction outcomes. Achieving and sustaining substantial nonpoint pollutant
reductions will likely require development and adoption of new implementation programs and
tools.

Nonpoint source implementation efforts could be improved by shifting the focus from a census
of implemented practices to an accounting of load reductions. Finer spatial scale modeling and
monitoring could further identify high nutrient loss areas and operations and be used to
consider more effective treatment options. Additionally, new financial incentive programs such
as pay-for-performance or pay-for-success programs offer opportunities to reward treatment of

vi
high-loss areas or operations and to encourage adoption of highly effective practices that land
managers may not consider under standard cost-share programs. These approaches would
provide both the identification of high-value opportunities and the incentives for landowners to
take advantage of them.

Achieving large-scale reductions in nonpoint sources of nutrients depends on adequately

addressing regional nutrient mass imbalances. Many regions of the Bay watershed exhibit mass
imbalances, where nutrient imports to a region (animal feed, fertilizer, atmospheric deposition)
exceed nutrient exports from the region (agricultural products harvested, manure transport).
The problem is particularly acute in areas of intensive livestock production. A variety of options
is available to address these imbalances, including implementing technologies that reduce
nutrient inputs, improving manure distribution (from surplus to deficit areas), and exporting
nutrients from the watershed.

Most nonpoint source policies are based on allowing land managers to decide whether or how
to reduce nonpoint source pollution. Such an approach is often reasonable given the number
and diversity of people involved in producing nonpoint pollutants. However, the extensive
history of nonpoint policy illustrates the limits of relying on voluntary actions. New and refined
requirements in case-specific circumstances may be necessary to achieve substantial progress
in reducing nonpoint source loads. Such requirements need not be overly costly to land
managers if land managers are given flexibility in how to meet the pollutant control
requirements and are provided financial assistance (similar to how some states are upgrading
wastewater treatment plants).

Given uncertainties around the complexity and diversity of nonpoint source pollutant
processes, not all alternatives will work as expected. Nevertheless, program change, innovation,
and experimentation are needed. Institutional innovation could be facilitated by considering
ideas such as sandboxing. Sandboxing is a formalized way to test and evaluate the efficacy of
new rules and programmatic approaches to nonpoint source or water quality management
without disrupting the operation of existing implementation efforts. Sandboxing also requires a
commitment from management agencies to make larger programmatic changes if the
sandboxed change demonstratively improves outcomes.

Achieving the water quality standards

Finding: Preliminary analyses suggest that nutrient load reductions have not produced the
expected level of improvement in estuary water quality, and this response gap is particularly
pronounced in the Bay’s deep channel.
Evidence indicates that the nutrient and sediment load reductions realized to date have led to
improved water quality conditions in some portions the Bay, but these nutrient load reductions
have not produced the expected level of increased dissolved oxygen in most of the Bay’s
habitats. This shortfall, or water quality response gap, is particularly pronounced in the Bay’s
deeper waters and could have significant consequences because of the large nutrient
reductions required to achieve the dissolved oxygen criteria in the deep water and deep

vii
channel habitats. Quantification of a response gap for water clarity is not possible because of
the absence of a formal predictive model, but progress in improving water clarity and
expanding SAV remains below the stated goal.

A variety of factors may explain the response gap in water quality conditions to pollutant load
reduction. For example, recent studies suggest that higher water temperatures offset roughly
6–34% of the water quality improvement from N reductions. Furthermore, Bay water quality
response will differ across habitats and may be nonlinear in some. That is, water quality
response to pollutant loads may occur fairly slowly until conditions are sufficient to accelerate
improvements. The thresholds where conditions more rapidly improve are often called tipping
points.

Identifying response gaps and their potential causes is limited by the design of the CBP
monitoring networks. The current estuary monitoring program is more attuned to assessing
attainment of water quality criteria than understanding processes underlying water quality
response. For example, nonlinear interactions (tipping points) have been identified at the scale
of subsystems in the Bay, but monitoring to determine the thresholds associated with either
degradation or restoration is not currently done. Monitoring designs may need to be modified,
and coupled with research and modeling efforts, to better understand the range of conditions
and relationships between stressors and water quality standards attainment.

Policy implication: Additional nutrient reductions will improve water quality, but water
quality criteria may be unattainable in some regions of the Bay under existing technologies.
The CBP is trying to achieve water quality standards in a highly altered environment that will
continue to change in ways with no historical precedent. Climate change is producing increases
in water temperature and changing precipitation patterns that confound efforts to achieve
water quality goals. The deep channel dissolved oxygen level has proven to be relatively
intransigent to load reduction efforts, but this area often serves as the primary policy focus for
CBP work. This reality may necessitate assessing the costs and tradeoffs of attaining numeric
water quality criteria in specific situations and locations and adapting numeric goals if desired.

Managing water quality to enhance living resources

Finding: Significant enhancement of living resources can be achieved through additional

management actions without complete achievement of water quality standards across all
habitats.
The Bay water quality criteria were selected based on chemical and physical conditions
(dissolved oxygen and water clarity) necessary, but not sufficient alone, to support fish and
invertebrate species living in different habitats and at different life stages. For instance, the
presence of adequate dissolved oxygen in a habitat does not guarantee that organisms will fully
populate that habitat. Direct evidence of the impact of water quality changes on various classes
of living resources is mixed, partly because of the confounding multiple changes occurring and
complex ecological interactions, and partly because there have not been substantial system-
wide changes in some criteria, like dissolved oxygen. As a result, quantifying living resource

viii
responses to any specific management and restoration action is a significant analytical
challenge. While the CBP employs a suite of models to predict the impacts of management
actions on chemical conditions in the estuary (particularly nutrient levels and dissolved oxygen),
the CBP does not use models to relate changes in dissolved oxygen and habitat to the
composition or abundance of living resources.

Living resource benefits may occur without full attainment of water quality criteria across all
habitats and in every region of the Bay. The location and timing of water quality improvements
will influence the composition and abundance of living resources. The five habitats
demonstrate different patterns and trajectories of attainment of water quality criteria, and
attaining the criteria is expected to be most difficult in the deep channel habitat. The shallow
water and open water habitats, however, more directly influence the life cycles of most fish
species. Habitat types also differ in their sensitivity to local management actions that can
enhance living resource response. For example, actions in shallow waters such as creating living
shorelines and improving benthic habitat can greatly increase the living resource response to
water quality conditions.

The living resource outcomes that can be expected from incremental attainment of water
quality criteria depend greatly on a host of other factors. Structural aquatic habitat, nearshore
habitat (wetlands, shoreline), commercial and recreational harvest, disease, and water
conditions (temperature, salinity) are all significant drivers of the composition and abundance
of living resources. Research points to the importance of specific habitats (particularly shallow
water) and nearshore conditions for many important species. Improvements in dissolved
oxygen may not increase the abundance of desirable fish species if these other factors are
already limiting populations. Thus, focusing investments on these other factors could improve
composition and abundance of living resources for any given level of water quality
improvement.

Policy implication: The legal requirements of the Clean Water Act (the water quality goal)
divert attention away from considering multiple means of improving living resources (support
of aquatic life as the designated use) as articulated in the Chesapeake Bay Watershed
Agreement.
The TMDL framework presents challenges to focusing management attention and resources on
improving living resource outcomes. The Chesapeake Bay Watershed Agreement lists 10 goals
and 31 desired outcomes. Water quality is only one of the goals, but it is the only legally
enforceable goal (under the Clean Water Act). This means that the benefits of restoration
actions tend to be expressed primarily in terms of nutrient reductions rather than benefits for
living resources. For example, benefits of restoring wetlands or living shorelines are often
framed in terms of the TMDL rather than improved habitat. Yet these investments can
substantially improve Bay living resources. A broader policy challenge for the CBP is how to
allocate restoration funds and efforts to generate the largest living resource impacts for the
most stakeholders.

ix
Policy implication: Opportunities exist to adjust approaches to prioritize management actions
that improve living resource response.
Possible changes to TMDL implementation could help prioritize water quality investments that
have greater and more immediate impacts on living resources. The TMDL as currently
structured directs management attention toward meeting an aggregate nutrient load limit that
is largely driven by dissolved oxygen conditions in the Bay. A tiered approach to TMDL
implementation would identify the locations or habitats expected to achieve pollutant
reduction limits first. Shallow water habitats in specific regions of the Bay may offer significant
opportunities to produce living resource responses. These are also areas with significant
stakeholder engagement because of their status as primary areas of recreational use, their
cultural significance, and their visibility as iconic Chesapeake landscapes. Reevaluation of water
quality criteria may also include consideration of new criteria (e.g., water temperature, toxic
and emerging contaminants of concern) or new frameworks for devising criteria (e.g., indicators
of resilience). Exploring such policy options would be enhanced with additional analytical
capacities and analyses capable of more fully articulating potential living resource responses to
water quality management.

Enhancing adaptive management

Finding: The Chesapeake Bay Program’s current portfolio of adaptive management processes
is inadequate to address the uncertainties and response gaps described in this report.
Moving forward, the CBP enters a new era of management. The Bay of the future will be
different from the Bay of the past because of permanent and ongoing changes in land use,
climate change, population growth, and economic development. Refining restoration goals over
time should be considered as knowledge evolves about what future conditions are possible,
what local communities and the partnership at-large see as priorities, and what is required to
attain those possible futures. Uncertainty is inherent in each of these.

The CBP has built a sophisticated TMDL implementation and accounting process premised on
the use of deterministic predictive and planning models to secure a desired pollutant reduction
and water quality response. The CBP’s decision framework and associated Strategy Review
System (SRS) assesses and evaluates progress toward achievement of specific CBP goals and
adjusts implementation based on these assessments. The water quality goal also adds an
accountability framework. However, the deterministic models providing single estimates of
pollutant loads for all inputs, land uses, and management actions are not well suited for
evaluating and addressing uncertainty. Such modeling approaches make it difficult to assess the
performance risk of different BMPs, inform decision makers of uncertainties, or assess the
robustness of management actions to underlying assumptions or changing environmental
conditions. The CBP has limited capacity to assess the potential of management alternatives for
improving living resources. The critical question is not simply: Are planned actions being
undertaken? Rather, are the actions producing load reductions and improved estuary
conditions?

x
Policy implication: Expanding the scope of adaptive management could address critical
uncertainties and response gaps.
A formalized adaptive management approach currently exists: the SRS, complemented by the
TMDL accountability framework. It is used to refine the existing implementation programs and
accounting structure. However, the system does not provide adequate insights into potentially
necessary policy changes at multiple levels ranging from devising new programs, rules, and
accountability systems to making budgetary and funding decisions and revising goals.
Enhancing adaptive management for water quality improvement suggests the CBP consider
ways to include more decision makers who have influence in broader scale potential changes to
the programs and policies.

To respond effectively to the issues raised in this report, the current adaptive management
process for water quality could be enhanced in several ways. Decision science offers processes
to integrate complex technical analyses with the planning processes used by those with the
authority to make choices about goals, programs, and budgets. A number of tools and
processes are available to identify and reduce decision-relevant uncertainties. Such approaches
aim to identify those uncertainties that pose the greatest risk to achieving management
objectives, identify how much a given outcome could be improved if a given uncertainty was
resolved, and identify the cost of error. These tools can be used for a variety of purposes
including supporting program design, implementation, and prioritization of research needs.

Conclusion

Four decades of efforts to manage nutrient and sediment pollutants have improved water
quality conditions in some portions of the Chesapeake Bay, but results are mixed. Additionally,
changing conditions from population growth, land use, and climate will make future restoration
more challenging. However, opportunities exist to improve the effectiveness of pollution
reduction efforts and accelerate improvements in living resources by building on the data,
knowledge, and experience gained over decades of effort. Capitalizing on these opportunities
will require adoption of new policies, procedures, and programs and expanded capacities to
address uncertainties around system response in decision-making. Finally, achieving reductions
in pollutants and realizing improvements in water quality and living resources in a system as
large, diverse, and complex as the Bay watershed and estuary calls for patience as changes are
planned and implemented and the system responds.

xi
Contents

Executive Summary......................................................................................................................... iv
Contents ......................................................................................................................................... xii
Figures ........................................................................................................................................... xiv
Foreword ....................................................................................................................................... xvi
1. Introduction ................................................................................................................................ 1
2. Approach to Evaluating System Response to Water Quality Management Efforts .................. 5
2.1. Water quality policy in the Chesapeake Bay: WQS, TMDL, and TMDL implementation ..... 6
2.2. Evaluating system response to nutrient and sediment control efforts.............................. 12
Uncertainty in response ....................................................................................................... 15
3. Nutrient and Sediment Response to Management Efforts...................................................... 17
3.1. Nonpoint source implementation policies in the Chesapeake Bay watershed.................. 17
3.2. Nutrient nonpoint source pollutant response to TMDL implementation policy ............... 19
Implementation gap ............................................................................................................ 19
Response gap ....................................................................................................................... 21
3.3. Assessment of gaps and uncertainties in efforts to reduce nonpoint source loads .......... 27
3.4. Lag times and legacy nutrients and sediment .................................................................... 28
3.5. Nonpoint source response gaps ......................................................................................... 29
Nutrient use and conservation behavior ............................................................................. 30
Program incentives of service providers and water quality managers ............................... 32
BMP effectiveness................................................................................................................ 32
Watershed data and monitoring ......................................................................................... 34
3.6. Implementation gaps in achieving nonpoint source pollutant reductions ........................ 35
Nutrient mass imbalances ................................................................................................... 36
Behavioral response to nonpoint source reduction policy.................................................. 37
Targeting nonpoint source investment ............................................................................... 38
Incentives and ability to reach nonadopters ....................................................................... 39
Incentives for innovation in nonpoint source load reductions ........................................... 39
Climate change..................................................................................................................... 39
3.7. Conclusions and implications.............................................................................................. 40
4. Estuary Water Quality Responses to Nutrient and Sediment Load Reductions ...................... 44

xii
 4.1. Water quality criteria and conditions for attainment ........................................................ 44
4.2. Estuary response to realized nutrient and sediment loads ................................................ 46
4.3. Response gaps and uncertainties in achieving WQS .......................................................... 56
Confounding factors and nonlinear interactions ................................................................. 56
Climate change..................................................................................................................... 60
Additional factors................................................................................................................. 61
Ability to estimate or measure loads delivered to Chesapeake Bay ................................... 61
Ability to accurately assess degree of attainment............................................................... 63
4.4. Conclusions and Implications.............................................................................................. 64
5. Living Resource Response to Water Quality Conditions .......................................................... 66
5.1. Conceptual model of living resource response to water quality conditions ...................... 67
5.2. Evidence and analysis of water quality impact on living resources in Chesapeake Bay .... 70
5.3. Conclusions and future directions ...................................................................................... 73
6. Findings and Implications ......................................................................................................... 75
6.1. Summary of findings ........................................................................................................... 76
6.2. Implications: Policy options for improving effectiveness of water quality management .. 78
6.3. Water quality goals and living resource response .............................................................. 80
6.4. Options for improving living resource response ................................................................ 82
6.5. Improving nonpoint source management .......................................................................... 84
6.6. Expanding adaptive decision-making and improving program learning ............................ 88
List of Abbreviations ..................................................................................................................... 93
References .................................................................................................................................... 94

xiii
Figures

FIGURE 1.1.—Achievement of Chesapeake Bay WQS, 1985–2020 (Source: CBP, n.d.-e). ............... 2
Figure 2.1.—Conceptual representation of system response to Chesapeake Bay water quality
goals and management efforts ....................................................................................................... 6
FIGURE 2.2.—Chesapeake Bay DUs (top) and DO WQC for five DU habitats (Source: USEPA,
2003a). ............................................................................................................................................ 8
FIGURE 2.3.—The 92 water segments of the Chesapeake Bay TMDL (Source: CBP, 2008). ............ 9
FIGURE 2.4.—Chesapeake Bay Program WQC and water quality assessment points across five DU
habitats. ........................................................................................................................................ 10
FIGURE 2.5.—Conceptual representation of uncertainty and gaps in system response to
Chesapeake Bay water quality policy. .......................................................................................... 13
FIGURE 3.1—Chesapeake Bay Program modeling of a typical nonpoint source BMP load
reduction. ...................................................................................................................................... 19
FIGURE 3.2.—Progress toward meeting TMDL nutrient reduction goals and expected response to
watershed management efforts as calculated by the CBP CAST model (Source: CBP, n.d.-a). ... 20
FIGURE 3.3.—Conceptual illustration of possible response gaps and response uncertainty in
achieving TMDL nutrient and sediment targets. .......................................................................... 22
FIGURE 3.4.—Flow-normalized TN, TP, and TSS yields at RIM stations for the period 1985–2020.
Figure created by Chris Mason, US Geological Survey Virginia and West Virginia Water Science
Center, Richmond, VA (Data source: Mason & Soroka, 2022). .................................................... 23
FIGURE 3.5.—Flow-normalized total and source sector TN and TP fluxes to the Chesapeake Bay
for 1992, 2002, and 2012 estimated with the Chesapeake Bay CAST model and SPARROW
model (Source: adapted from Ator et al. [2020] by Easton et al. [2023]). ................................... 24
FIGURE 3.6.—Behavioral and physical system response to Chesapeake Bay nonpoint source
control policy. ............................................................................................................................... 27
FIGURE 3.7.—Self-selection challenge under CBP TMDL accounting system resulting from spatial
variability in baseline nutrient losses. .......................................................................................... 31
FIGURE 4.1.—Conceptual water quality response to nutrient and sediment reductions, where
attainment of WQC is a function of nutrient or sediment load reductions. ................................ 45
FIGURE 4.2.—Conceptual diagram of primary degradation factors for water quality in eutrophic
(nutrient-rich) conditions as compared to low nutrient input conditions (Source: Testa et al.,
2018). ............................................................................................................................................ 45
FIGURE 4.3.—Estuarine physical system response to CBP implementation policy........................ 46
FIGURE 4.4.—Measured and estimated N (panel a) and P (panel b) loads and annual flow by
October 1–September 30 water year (Source: CBP, n.d.-e). ........................................................ 47

xiv
FIGURE 4.5.—Mean change in surface TN (a-c) and TP (d-f) over three time periods at tidal
monitoring stations, calculated using Generalized Additive Modeling approach (Source: Murphy
et al., 2022). .................................................................................................................................. 48
FIGURE 4.6.—Percent change in estuarine TN and TP loads and concentrations, late 1980s to
mid-2010s, where each dot represents a Bay segment (Source: Testa et al., 2018). .................. 49
FIGURE 4.7.—Changes in DO in bottom water layer measured during June—September, short
term (left pane) and long term (right panel); starting dates for long-term measurements vary
(Source: CBP, n.d.-d). .................................................................................................................... 50
FIGURE 4.8.—Time series of area-weighted estimated WQC attainment for the five DUs, 1985–87
through 2016–18 (Source: Zhang, Murphy, et al. [2018] with updated data). ............................ 51
FIGURE 4.9.—Expected and realized relationships between TN loads and DO criteria attainment
for open water, deep water, and deep channel habitat, calculated as 3-year running mean
observed values (blue diamonds) and expected responses from estuary model (orange dots) for
the same time periods. Yellow squares are 10-year means of the observed data. ..................... 52
FIGURE 4.10.—Attainment deficit (0%=full attainment, -100%=full non-attainment) for DO WQC
in three DU habitats, 1985–2020, calculated as 3-year running means. Year shown represents
the first year of a 3-year assessment period (Source: Q. Zhang, personal communication). ...... 53
FIGURE 4.11.—Seagrass area (hectares) by salinity zone and in coastal bays from 1971 through
2021. Amount of seagrass abundance in tidal fresh (top panel), moderately salty (mesohaline—
middle panel), and full salinity (polyhaline—bottom panel) areas. Density of cover is indicated
by degree of shading: very dense (70–80% cover), dense (40–70% cover), sparse (10–40%
cover), and very sparse (<10% cover) (Source VIMS, n.d.-b). ...................................................... 55
FIGURE 4.12.—Effects of N and P additions on physical, chemical, and biological elements of the
estuarine system, including algal biomass, bottom water oxygen, and nutrient recycling. Effects
of climate change (save sea level rise) and land cover change are not included (Source: Kemp et
al., 2005). ...................................................................................................................................... 59
FIGURE 4.13.—Estimated extent of Chesapeake Bay hypoxia with and without 35 years of
nutrient reductions (Source: Frankel et al., 2022)........................................................................ 60
FIGURE 5.1.—Conceptual living resource response to attainment of WQC. ................................. 67
FIGURE 5.2.—Factors affecting the composition and abundance of living resources. .................. 69
FIGURE 5.3.—Indices of Chesapeake Bay fish abundance, various species over time (Source:
VIMS, n.d.-a). ................................................................................................................................ 70
FIGURE 6.1.—Levels of policy feedback and learning in expansion of adaptive management. .... 79
FIGURE 6.2.—Conceptual relationship between costs, attainment of WQS, and living resource
response. ....................................................................................................................................... 81

xv
Foreword

In January 2003, the Chesapeake Bay Program’s Scientific and Technical Advisory Committee
(STAC) produced an independent report entitled Chesapeake Futures that captured the state of
knowledge of the restoration effort and presented a likely set of outcomes, or scenarios, based
on that knowledge and projected trends. The report detailed outcomes for land use and
development, forests, agriculture, and the Bay and its fisheries under each of three scenarios:
under recent trends, if the objectives of agreements in place at that time were met, and if
feasible alternatives were put in place. Many of the scenarios were prescient in their predicted
outcomes under the current agreement objectives and proposed alternatives that were feasible
but only partially implemented. Chesapeake Futures did not propose specific policy
recommendations but instead offered the report as constructive advice from a body of experts
in the hope of informing the decision-making of the partnership engaged in restoration of the
Bay. It inspired reflection, thought, discussion, and debate, and outlined a series of feasible
innovations, all in the hope of moving the work forward. This report aspires to stand on the
shoulders of this previous work.

Chesapeake Futures represented the work of many scientists and their collective expertise, all
under the auspices of STAC. This report does the same, but with a specific focus on synthesizing
the robust amount of data and discovery amassed over the last 30 to 40 years of the
Chesapeake Bay Program (CBP) related to Bay water quality management and assessing the
implications for future policy. The effort began as a STAC independent initiative in March 2019,
after Kurt Stephenson, Zach Easton, and Brian Benham proposed the idea of a report that
would identify gaps and uncertainties in system response—physical, chemical, biological, and
socioeconomic—that impact efforts designed to attain water quality standards in Chesapeake
Bay. STAC agreed to the challenge, and as STAC Chair at the time, Benham facilitated the
development of a collaborative process that would engage the entire committee.

As a first step in approaching the long causal chain that links management actions to their
eventual impact on water quality and living resources, workgroups were formed around the
subsystems of this chain: nutrient and sediment reductions (watershed), water quality response
to nutrient and sediment reductions (estuary) and living resource response to water quality
(living resources). Each of these workgroups generated an independent document with a self-
determined scope (i.e., workgroups were afforded flexibility to address issues beyond the
original objectives). Because the content of each document was both unique and substantial,
STAC chose to publish them as stand-alone documents with authorship attribution (see
references provided below). In the second step, a steering committee developed a series of
framing questions to guide the preparation of this report that would meet the original objective
of identifying gaps and uncertainties in achieving the Bay total maximum daily load (TMDL) and
water quality standards. Coeditors Stephenson and Wardrop, supported by a subgroup of the
steering committee (the “Writer Group” that included Leonard Shabman, Zach Easton, Jeremy
Testa, William Dennison, Kenny Rose, and Mark Monaco), were tasked with assembling ideas
and contributions to write a single draft text, drawing material from the aforementioned
resource documents, STAC and CBP reports, the scientific literature, and a limited amount of
xvi
additional analyses performed in collaboration with CBP scientists. Zach Easton made
significant contributions in drafting the initial framing of the report, and Leonard Shabman was
particularly helpful in assisting coeditors Stephenson and Wardrop in summarizing the main
findings and conceptualizing implications within the context of water policy. The resulting
report (this document) was then submitted for several reviews by steering committee
members, the STAC membership at-large, and the U.S. Geological Survey to produce a
consensus report.

This report represents the thought and analyses of many beyond those who formally put pen to
paper, via participation in the critical thinking and discernment processes that are the
foundation of a science-based approach. They are listed as contributors. The editors also wish
to recognize various members of the Chesapeake Bay partnership who provided additional
expertise. We cannot begin to add up and properly attribute the contributions and
commitment that this effort represents. Thank you for allowing us to stand on the shoulders of
the scientific expertise that you represent and take in the view; it’s an extraordinary and
hopeful landscape.

Resource documents:

Easton, Z., Stephenson, K., Benham, B., Böhlke, J. K., Buda, A., Collick, A., Fowler, L., Gilinsky, E.,
Hershner, C., Miller, A., Noe, G., Palm-Forster, L., & Thompson, T. (2023). Evaluation of
management efforts to reduce nutrient and sediment contributions to the Chesapeake Bay
estuary. STAC Publication Number 23-003, Chesapeake Bay Program Scientific and Technical
Advisory Committee (STAC), Edgewater, MD. 55 pp.

Testa, J. M., Dennison, W. C., Ball, W. P., Boomer, K., Gibson, D. M., Linker, L., Runge, M. C., &
Sanford, L. (2023). Knowledge gaps, uncertainties, and opportunities regarding the response of
the Chesapeake Bay estuary to proposed TMDLs. STAC Publication Number 23-004, Chesapeake
Bay Program Scientific and Technical Advisory Committee (STAC), Edgewater, MD. 61 pp.

Rose, K., Monaco, M. E., Ihde, T., Hubbart, J., Smith, E., Stauffer, J., & Havens, K. J. (2023).
Proposed framework for analyzing water quality and habitat effects on the living resources of
Chesapeake Bay. STAC Publication Number 23-005, Chesapeake Bay Program Scientific and
Technical Advisory Committee (STAC), Edgewater, MD. 52 pp.

xvii
1. Introduction

The 2014 Chesapeake Bay Watershed Agreement (CBWA) (CBP, 2014) contains 10 broad
management goals, most of which are directed at supporting viable populations of living
resources throughout the Bay watershed. Federal, state, and local programs to secure those
living resource goals, as measured by policy attention and resource commitment, have been
directed primarily toward achieving water quality standards (WQS) in the tidal waters of the
Bay. The standards follow from the Clean Water Act (CWA) and begin with specification of
designated uses (DUs) to support specific living resources, and that support is manifested in
ambient numeric water quality criteria (WQC) for dissolved oxygen (DO), chlorophyll a (Chl a),
water clarity, and submerged aquatic vegetation (SAV), that in turn are expected to be attained
by limiting nitrogen (N), phosphorus (P), and sediment loads to the Bay. After 2010 this load
limit, the total maximum daily load (TMDL), or “pollution diet”, became a legal obligation for
the states with areas that drain into the Bay and for the District of Columbia (Linker et al., 2013;
USEPA, 2010).

The 2010 Chesapeake Bay TMDL-driven nutrient and sediment load targets were based on the
predicted response of the numeric criteria in the tidal tributaries and in the Bay itself. The
predictions were made using the suite of computer models developed and refined over the
years since the Chesapeake Bay Program (CBP) was created (Hood et al., 2021). Bay
jurisdictions develop watershed implementation plans (WIPs) that describe water quality
improvement actions they will take to meet the TMDL. The CBP modeling suite translates
actions into nutrient and sediment loads to the Bay and then calculates whether the numeric
WQC will be met.

Over the past 40 years (including two decades of effort prior to the TMDL), the Chesapeake Bay
jurisdictions have made notable progress in reducing nutrient and sediment loads to the Bay.
Based on reports of point and nonpoint source load reduction practices put in place from 1985
through 2021, the CBP watershed model CAST (Chesapeake Assessment Scenario Tool, the
primary tool for calculating load reductions from the practices installed to meet the TMDL)
predicts that the current level of implementation is sufficient to achieve 73%, 91%, and 100% of
the N, P, and sediment load reductions specified in the TMDL (CBP, n.d.-a). To date, the
majority of nutrient reductions have been attributed to wastewater treatment plant (WWTP)
upgrades (point sources) and to large and sustained reductions in atmospheric N deposition
(CBP, n.d.-a). Nonpoint sources have contributed a smaller share of the needed reductions. The
CAST model calculates that total urban and agricultural N and P nonpoint loads decreased 15%
and 29%, respectively, between 1985 and 2021. Nonpoint sources, however, are the single
largest remaining source of nutrients to the Bay. Of the controllable nutrient loads in 2021
(excluding loads from natural sources), approximately three-quarters originate from agricultural
and urban nonpoint sources.

1
However, the CBP’s extensive ambient water quality monitoring system in the estuary indicates
that water quality has been slow to respond to improvement efforts. Water quality standards
have been attained in some portions of the Bay. Some species of SAV, an important Bay living
resource and water quality criteria, have expanded in several regions of the Bay in recent years
(Lefcheck et al., 2018). Nonetheless, the reported 30 years of load reductions appear to have
had limited effect on achievement of WQS. A composite index developed by the CBP
summarizes WQS attainment across the entire Bay (Zhang, Murphy, et al., 2018). Using this
index, the CBP reported in 2020 that 29.6% of the Bay attained WQS during 2018–20, only up
from 26.5% in 1985–87 (CBP, n.d.-e). This is an average annual improvement rate in WQS
attainment of approximately 0.25% over the 30 years since Bay restoration activities began,
although there is year-to-year variability (figure 1.1). At this rate of progress, full attainment of
the WQS is an uncertain and distant possibility.

FIGURE 1.1.—Achievement of Chesapeake Bay WQS, 1985–2020 (Source: CBP, n.d.-e).

This report provides the results of the CBP Scientific and Technical Advisory Committee (STAC)
effort to evaluate whether and why improvements in water quality have failed to meet
expectations (called gaps here). One possible cause for a gap is that water quality
improvements could just take time. Well-recognized lags occur in the system that lengthen the
time between when actions are taken to reduce pollutants and when pollutant reductions and
water quality outcomes are fully realized. However, other explanations also are considered. For
example, the physical, chemical, biological, or socioeconomic systems may not be responding
as scientists and water quality managers expect. The divergence between expected and
realized outcomes can arise from incomplete or incorrect data or knowledge of system
relationships, ranging from how policy translates into behavior and then pollutant control

2
effectiveness to how estuary water quality responds to changing nutrient loads and how living
resources respond to changes in water quality conditions. If water quality response to pollution
control efforts is limited by more than just lag times, then policy changes may be needed.

With the 2025 TMDL WIP implementation deadline approaching, the CBP STAC undertook this
effort to identify key areas of uncertainty and improve learning about efforts to attain the WQS
underlying the TMDL process. The goal is to identify where CBP partnership water quality
programs and policies may not be yielding anticipated system responses and to identify
possible reasons for these gaps. Consequently, this report is purposefully unbalanced: we focus
on response gaps and opportunities to improve our collective understanding of how to achieve
water quality load targets and standards and improve living resource response, rather than
celebrate what we already know or have already achieved. Recognizing and actively addressing
response gaps and uncertainties can improve our understanding of system response and
facilitate the development of new approaches and techniques to improve the effectiveness of
CBP water quality improvement efforts.

The specific objectives of this report are to:

● identify gaps between the expected and realized physical, chemical, biological, and
socioeconomic responses to management actions, and identify recent scientific
developments that can advance efforts to attain WQS;
● characterize the critical uncertainties in system response to management actions, and
identify strategies that improve understanding of system response relevant to the
attainment of WQS; and
● identify strategies for better integrating scientific and technical analysis into
management efforts in order to aid decision-making under uncertainty.

In pursuing these objectives, STAC strives to direct science in service of policy. The scientific
community should not be expected to define what restoration means, decide what water
quality goals are pursued, or decide which management actions and policies should be used to
achieve those goals. However, the work of the scientific community must be relied upon to
improve our collective understanding of the consequences of both the means and ends of
policy, how the system responds to efforts to achieve a specific water quality goal, and what
can be gained and at what cost from different WQS. Consequently, the report offers
implications and options, rather than recommendations, for policies and programs regarding
setting and achieving water quality goals in the face of uncertainty. For example, the report
does not recommend a particular pollutant reduction policy be pursued but rather aims to
frame the discussion as: if X pollutant reduction result is desired, then Y and Z options may
improve the chances of achieving X. This report also focuses attention on those uncertainties
(called decision-relevant uncertainties) and system responses that have direct and immediate
implications for achieving the TMDL and specifying WQS.

The main body of the report is organized around system response. Chapter 2 summarizes Bay
water quality policy and describes the terms and approach used in the report to evaluate

3
system responses to policy goals and efforts. Chapter 3 describes Bay management efforts to
achieve the TMDL and summarizes our understanding of how nutrient and sediment loads are
responding to those management efforts. Chapter 4 reviews how the numeric measures of Bay
water quality used to set the TMDL are being measured and how the measures are responding
to reduced nutrient and sediment loads. The Bay water quality measures and the TMDL limit
set to achieve these measures are expected to support populations of specific living resources.
Chapter 5 reviews the available analyses of living resource response, noting that the link
between meeting the current WQS and the abundance of specific living resources is not
explicitly identified by the CBP and is highly uncertain. Chapter 6 describes the implications of
the findings for future Bay water quality management, including identifying options for
supporting Bay living resources and for research and monitoring to support those options.

4
2. Approach to Evaluating System Response to Water Quality Management
Efforts

The organization of the report follows the conceptual logic underlying attainment of
Chesapeake Bay WQS (see figure 2.1). The left side of figure 2.1 shows the Bay restoration
policy goals and programs as stated in the 2014 CBWA (dark blue box). This report focuses on
the specific water quality goal that became legally enforceable under the CWA in 2010 when
the Bay was listed as an impaired waterbody (from sediment and nutrients P and N) and the
Bay TMDL was developed. Reading from top to bottom in figure 2.1, the CWA planning process
for the Bay began with specifying WQS, including DUs with the numeric WQC deemed
necessary to support each DU. Because monitoring and assessment found that the WQS were
not being met, a TMDL was prepared specifying the reductions in pollutant stressors N, P, and
sediment needed to achieve the WQS. What followed was the design of federal, state, and local
regulations and incentive programs to secure nutrient and sediment reductions from point and
nonpoint sources. Bay states and the District of Columbia outlined general plans in WIPs. As
actions called for in the plan were taken, progress toward implementing practices to meet load
reductions was tracked and reported.

How social, physical, chemical, and biological systems respond to the water policy is shown on
the right side of figure 2.1. STAC divides the causal chain that links pollutant reduction actions
to achievement of WQS into three groups (reading from bottom to top in figure 2.1). First, CBP
models (text box 2.1) predict how the management actions taken in response to regulations
and incentives, such as wastewater treatment at point sources or best management practices
(BMPs) on farm, forest, and urban lands, reduce Bay pollutants (total N, total P, and total
suspended sediment, or TN, TP, and TSS) (orange box in figure 2.1). Second, nutrient and
sediment reductions to the Bay are predicted to change water quality conditions (e.g., DO) at
different locations throughout the estuary (light blue box in figure 2.1). Third, improvements in
estuary water quality (as measured by the WQC) are expected to support fish, shellfish, and
other living organisms in the Bay (green area in figure 2.1), although specifically how living
resources will be supported by improved water quality is currently not predicted by the CBP.

Section 2.1 elaborates on the current structure of the Bay water quality program (left side of
figure 2.1). Section 2.2 describes the general approach used to assess how the system is
responding to CBP water quality policy across the causal chain (right side of figure 2.1). This
background forms the basis of chapters 3, 4, and 5 (corresponding to each box on the right side
of figure 2.1), which describe challenges to generating sufficient behavioral change and practice
adoption (implementation gaps), identifying where the system may not be responding as
predicted (response gaps), and addressing the uncertainties associated with the causes of those
gaps.

5
Figure 2.1.—Conceptual representation of system response to Chesapeake Bay water quality goals and
management efforts

2.1. Water quality policy in the Chesapeake Bay: WQS, TMDL, and TMDL implementation

Chesapeake Bay states and responsible federal agencies have established numerous restoration
goals for the Bay through a series of agreements since the inception of the CBP. For instance,
the 1987 Chesapeake Bay Agreement established general restoration goals (non-numeric)
including protection and enhancement of living resources (specifically identified as aquatic
vegetation, habitat, shellfish, wetlands, and waterfowl/wildlife), water quality, public

6
education, and public access (CBP, 1987). The most recent agreement, signed in 2014, includes
10 overall goals and 31 outcomes under those goals (see dark blue box in figure 2.1) (CBP,
2014). The focus of this report is the Chesapeake Bay water quality goal, but it is only one
among a larger set of Bay restoration goals.

Since CBP inception, water quality improvement has been the primary focus of Bay restoration
efforts. The 1987 Agreement stated the water quality goal as targets for nutrient load
reduction, 40% reduction in nutrients N and P entering the Bay from 1985 baseline levels,
which were to be achieved by the year 2000. The current Bay water quality goals that focus on
the expected water quality conditions were first developed in the early 2000s (Tango & Batiuk,
2013; USEPA 2003a, 2003b). Prior to the required development of the Bay TMDL in 2010, the
WQS were aspirational. In 2010 achievement of the WQS was formally incorporated into a legal
and regulatory structure under Section 303d of the CWA. Once baywide water quality
management came under the requirements of the CWA planning process, the Bay program had
to adapt the CWA water quality planning process and then a TMDL process to the Bay as a
whole.

The WQS consist of DUs for the estuary and numeric WQC to determine achievement of the
DUs (yellow box 1 in figure 2.1). While DUs could include recreation, navigation, or water
supply, the DU chosen for the Bay program was the protection of aquatic living resources.
However, because a variety of habitats across the Bay support differing assemblages of aquatic
organisms, the U.S. Environmental Protection Agency (EPA) convened a process through which
state, federal, academic, and multistakeholder representatives agreed that the DUs would be
made specific to major habitat types occurring across all of the Chesapeake Bay and its tidal
tributaries (USEPA, 2003a). This meant that the DU of supporting aquatic living resources would
be applied to five different habitats: shallow water (Bay grass use), open water (fish and
shellfish use), deep water (seasonal fish and shellfish use), deep channel (seasonal refuge use),
and migratory fish (spawning and nursery use) (figure 2.2).

Text box 2.1. Chesapeake Bay models referenced in this report

Three Chesapeake Bay models are referenced in this report as those used by the CBP to generate the
expected estimates in pollutant loads and water quality outcome responses (Hood et al., 2021). The
CBP uses two versions of the watershed model to generate estimates of nutrient loads flowing into
the Bay from the watershed. CAST estimates average annual loads that would be generated under 10
years of typical weather conditions (typical defined as 1991–2000). The CBP uses CAST to set TMDL
planning targets, design implementation plans, and track implementation progress. The dynamic
version of the watershed model estimates daily nutrient loads and is used to provide loads estimates
to the estuary model (Hood et al., 2021; CBP, n.d.-c). The CBP uses the estuarine model to estimate
attainment of tidal Bay DO, Chl a, and water clarity criteria under different nutrient and sediment
loads. The estuarine model is made up of a hydrodynamics component that measures transport in
the estuary and biogeochemistry component that translates nutrients into water conditions.

7
FIGURE 2.2.—Chesapeake Bay DUs (top) and DO WQC for five DU habitats (Source: USEPA, 2003a).

Numeric WQC are defined to secure attainment of the DUs in each habitat type. These WQC
include numeric values for DO, water clarity, and Chl a for the designated habitats (Tango &
Batiuk, 2013). Dissolved oxygen criteria have been developed for each of the five habitats
throughout the Bay, while water clarity/SAV and Chl a apply to specific habitats (i.e., shallow
water and select regions of the open water habitat, respectively). Open water habitat is
expected to achieve a 30-day mean DO concentration of 5 mg/L, a 7-day mean of 4 mg/L, and
an instantaneous minimum of 3.2 mg/L. These levels are necessary to support different life
stages of fish and shellfish species. For deep channel habitat, an instantaneous minimum DO of
1 mg/L is designed to support benthic organisms (i.e., worms, clams) (figure 2.2). Water quality
criteria for some habitats, such as deep water and deep channel, are defined based on seasonal
values (DO levels in the summer months, June–Sept.).

Note that Bay WQC are largely based on chemical and physical conditions (e.g., DO and water
clarity). With the exception of the SAV criteria for water clarity, the WQC are not based on
achievement of specific living resource species or populations (such as a specific assemblage of
fish, shellfish, or aquatic animals). Rather the numeric criteria were identified based on species
tolerances for regions throughout the Bay (Monaco et al., 1998; Tango & Batiuk, 2013; USEPA
2003a, 2003b). For example, the 5 mg/L DO criteria in the open water habitat is designed to
support finfish species such as striped bass. This approach ensures that adequate conditions are
present to support organisms in different habitats, but it does not necessarily translate into
actual population and food web responses because of the many other factors that also affect

8
species abundances. Essentially, the WQC are necessary, but not sufficient, conditions to
support aquatic organisms populating different habitats.

A series of specific water quality assessment monitoring protocols have been developed to
determine the degree of attainment of the WQS corresponding to each WQC (Tango & Batiuk,
2013; USEPA 2003a, 2004, 2017) (yellow box 2 in figure 2.1). For assessment purposes, the Bay
is divided into 92 distinct areas called segments (figure 2.3). Each segment typically contains
multiple habitats and their associated WQC. For example, one segment may be assessed for
open water (DO, and in some regions Chl a), deep water (DO), and shallow water habitats (DO,
clarity/SAV). The CBP uses three years of data to determine whether the temporal and spatial
distributions of water quality outcomes within a segment have achieved WQS attainment
(compliance) (Batiuk et al., 2009).

FIGURE 2.3.—The 92 water segments of the Chesapeake Bay TMDL (Source: CBP, 2008).

9
The CWA water quality management process expects that once WQC are established a process
of continual monitoring and assessment follows to establish whether WQS are being met. Thus,
full attainment of the WQS requires meeting multiple DO criteria in all 92 segments of the
Chesapeake Bay with up to 5 different habitats in each segment. The Chl a and water clarity
criteria must be met in a subset of those segments and habitats. The combinations of segments,
habitats, and individual criteria represent a total of 1,052 unique conditions to be met before
reaching full attainment (see figure 2.4). In addition, some of the WQS for the Bay cannot
actually be assessed because of insufficient monitoring. For example, one-day and
instantaneous minimums for DO cannot be assessed with existing monitoring frequencies in
open water DUs.

FIGURE 2.4.—Chesapeake Bay Program WQC and water quality assessment points across five DU habitats.

By 2010 the Bay was deemed impaired, and the CWA Section 303d process then required
defining a limit on the allowable pollutants causing the impairment—the TMDL (yellow box 3 in
figure 2.1). When the Chesapeake Bay TMDL was finalized, it was (and remains) the only TMDL
for a water body as large and complex as the Bay. The Bay TMDL sets total annual limits on TN,
TP, and TSS. If the TMDL nutrient and sediment load targets are reached, water quality is
predicted to attain the WQS for all 1,052 assessment points in figure 2.4.

The TMDL was set using the CBP estuary water quality model that predicts how water quality
conditions (e.g., DO) respond to changes in nutrient and sediment loads. As required by CWA
planning regulations, when setting the TMDL modelers were expected to include
“conservative” model assumptions to accommodate model prediction uncertainty. The TMDL
TN, TP, and TSS load targets are 214.9, 13.3, and 18,587 million lb/yr, respectively (CBP, n.d.-e).
The CBP estimates that achieving the nutrient and sediment reductions required to meet the

10
TMDL could not result in attainment of the criteria in every segment and habitat. Initial water
quality modeling indicated that at least one segment (lower Chester River segment) could not
meet the deep water DO criteria under the TMDL pollution diet given economically feasible
pollution control options. Consequently, EPA approved a Maryland state “restoration variance”
for that segment (USEPA, 2010). Since that time, EPA has granted restoration variances for five
deep water or deep channel segments and nine open water segments.

Chesapeake Bay Program partners established a 2025 deadline for Bay jurisdictions to meet the
TMDL pollution control obligations. The deadline does not require that all WQS be met or that
pollutant reductions be realized by 2025, but rather the management practices that CBP
models predict will be sufficient to meet the TMDL pollutant target loads for nutrients and
sediment must be in place (USEPA 2010).

As part of the TMDL establishment, the jurisdictions designed plans for implementation of
actions predicted to result in load reductions from CWA-regulated point sources and
unregulated nonpoint sources (yellow box 4 in figure 2.1). The approach implemented is
broadly similar to approaches taken in most TMDLs. State and federal authorities rely on a
variety of regulatory and voluntary policies and programs (yellow box 5 in figure 2.1) for both
point and nonpoint sources. Jurisdictions develop, and EPA reviews, WIPs that describe the
specific types of practices that will be used to meet the TMDL.

For point sources, jurisdictions impose numeric nutrient effluent limitations on municipal and
industrial WWTPs above a certain size under the CWA permitting program. Urban nonpoint
source loads are the fastest growing category of nutrient and sediment loads. To reduce these
loads, several state governments established numeric nutrient and sediment permits for
municipal separate storm sewer systems (called MS4s). These permits are unique in that states
place numeric limits on urban nonpoint source loads. Numeric MS4 permits represent a
significant departure from traditional MS4 permits which were based on narrative rather than
numeric requirements. Agricultural nonpoint source pollution is the single largest contributor of
nutrient and sediment loads to the Bay, but the CWA explicitly prohibits federal permitting
requirements for controlling most agricultural runoff. The jurisdictions rely on a mix of
education, financial assistance, and technical assistance programs to induce agricultural
producers to voluntarily adopt BMPs. States operate permitting programs for certain
agricultural operations such as concentrated animal feeding operations (CAFOs) over a certain
size.

The CBP uses the CAST watershed model to inform development of WIPs, credit progress
toward meeting TMDL nutrient and sediment load reduction targets, and track implementation
progress (Hood et al., 2021). State and local jurisdictions’ TMDL obligations for 2025 are
considered met when CAST calculates that enough pollutant reduction practices have been
installed to meet the TMDL, not by whether nutrient and sediment delivery to the Bay reaches
a specific level. The presumption is that if CBP partners implement needed nutrient and
sediment pollution control measures as estimated by CAST by 2025, then pollutant reductions

11
called for under the TMDL will be realized at some unspecified future date. The difference
between when pollutant control measures are installed and when pollutant reductions are
achieved is attributed to the time lag associated with nutrient and sediment movements
through the watershed and to the Bay.

For nonpoint sources, EPA expects the states to offer “reasonable assurance” that nonpoint
load reductions will be achieved. A state offers reasonable assurance by including in its WIP
only those land use practices (BMPs) that meet three criteria: the practices must (1) exist, (2) be
technically feasible at a level required to meet allocations, and (3) have a high likelihood of
implementation.

The CBP also implements a version of adaptive management for the CBWA outcomes through a
decision framework (CBP, n.d.-b). This process is intended facilitate decision-making under
uncertainty by continuously monitoring and evaluating progress toward achievement of specific
CBP program outcomes and adjusting implementation based on these assessments. The CBP’s
Goal Implementation Teams (GITs) are responsible for implementing the decision framework.
More generally, the GITs are responsible for overseeing and promoting the implementation of
plans to achieve the goals of the CBP. The implementation of the decision framework at the GIT
level is articulated in the 2-year cycle Strategy Review System (SRS) that establishes the link
between the factors that could impact the partnership’s ability to achieve an outcome and the
actions it is taking to manage them. For water quality goals under the CBWA, the CBP
partnership has an accountability process for the Bay TMDL that is described in the
management strategy for the 2025 WIP and standards attainment outcome (CBP, 2020). While
these outcomes come under the SRS adaptative management reviews, the TMDL accountability
process has additional steps to address regulatory requirements.

2.2. Evaluating system response to nutrient and sediment control efforts

The overarching goals of this report are to evaluate how the Chesapeake Bay system and its
physical, biological, and social subsystems are responding to water quality improvement efforts
and to help identify ways that system response to management efforts could be improved,
when appropriate. To conduct this evaluation, STAC relied on both the state of current
knowledge as reflected in the scientific literature and review of portions of extensive CBP data.
In addition, STAC has evaluated many of the issues in previous workshops, reports, and
requested reviews, and this accumulated body of work is a critical part of this overall
evaluation. Based on this body of evidence, this report discusses the gaps and uncertainties
associated with watershed, estuary, and living resource system response.

Federal, state, and local implementation policies are intended to change people’s behavior in
ways that produce nutrient and sediment reductions (orange graph in figure 2.5). The solid
orange line, labeled expected response, represents the CBP’s predicted nutrient and sediment
reductions in response to implementation of management actions and is largely derived from

12
FIGURE 2.5.—Conceptual representation of uncertainty and gaps in system response to Chesapeake Bay water
quality policy.

the CAST model. The orange star represents 100% achievement of the TMDL N, P, and sediment
targets. Conceptually, the pollutant reduction response to management actions (i.e., practices
and technologies installed) eventually begins to diminish (the orange line begins to flatten),
indicating that more effort will be required to achieve each additional gain in pollutant
reduction (i.e., the actions that produce large reductions in pollutant loads have already been
undertaken). For example, an initial upgrade of a single WWTP might result in a one-million-
pound reduction in N, but achieving subsequent reductions from nonpoint sources could

13
require treating tens of thousands of acres of farmland or the installation of tens of thousands
of urban stormwater BMPs.

As the 2025 TMDL deadline approaches, possible implementation gaps exist. For purposes of
this report, an implementation gap is the difference between current implementation
(expected current condition) and the implementation goal (desired condition). Conceptually, an
implementation gap is how much additional implementation is needed to achieve the pollutant
reduction goal (under the assumption that the effectiveness of the implementation is known
and correct). In the orange graph in figure 2.5, the orange dot represents the expected N, P,
and sediment reduction response given the level of implementation achieved to date. The
horizontal distance between the orange dot and the star represents the implementation gap
that impedes achievement of the TMDL targets.

Reductions in nutrient and sediment loads to the Chesapeake Bay are expected to produce
improvements in Bay water quality (light blue graph in figure 2.5), specifically defined by the
DO, water clarity, and Chl a criteria. Estuary science and modeling of the relationships between
these criteria and water quality produce expected water quality response to reductions in
nutrient and sediment loads (solid blue line in figure 2.5). The implementation gap, as defined
above, affects the achievement of the WQS (distance between blue dot and star). The shape of
the curve represents the expectation that the initial increments of nutrient and sediment
reductions are expected to bring more areas (segments) of the Chesapeake Bay into compliance
than later increments. In other words, larger pollutant reductions will be needed to bring the
remaining few segments into attainment. For instance, CBP modeling indicates that millions of
additional pounds of nutrient reductions are needed to achieve DO criteria in parts of four (out
of 92) segments (deep waters and deep channel habitats in CB3MH, CB4MH, CB5MH, and
POTMH segments [USEPA, 2010]). Complete (100%) achievement of the DO criteria is indicated
by the blue star.

Finally, there is a general public expectation that as progress is made toward achieving the
WQS, living resources, particularly specific species of fish and shellfish, will increase in
abundance (increasing response on the green graph in figure 2.5). How living resources will
respond to improved oxygen levels and improved water clarity is not specifically described by
the WQS or by the CBP, however. As noted above, the WQC were designed to avoid specifying a
living resource outcome because of the numerous other factors that affect living resources
besides water quality. Thus, there is no formal expected response defined by the CBP between
improving water quality conditions and living resources (no expected response curve in figure
2.5). The green graph shows possible living resource responses to improving water quality
conditions (discussed in more detail in chapter 5). The possible living resource responses
represented in figure 2.5 can be thought of as the key benefit responses to the costs incurred in
reducing Bay pollutant stressors in the orange graph.

Achieving pollutant reduction and water quality goals will also be challenging if emerging
scientific data and analyses find evidence that biological, physical, and behavioral outcomes do

14
not match expected/predicted outcomes, called a response gap in this report. For example,
experts may estimate the average nutrient removal effectiveness of specific BMPs to be 75%,
but emerging scientific evidence may strongly suggest that overall BMP effectiveness is less
than 75%. An example of such a response gap is illustrated in the orange graph of figure 2.5 as
the difference between a realized outcome represented by the white dot and the expected
outcome represented by the orange dot. While a response gap could be positive or negative,
this report is most concerned with situations when actual response appears to be falling short
of expected (predicted) response. Such a response gap means that any given level of
management effort is producing less reduction than is being claimed under the TMDL and less
improvement in Bay water quality and living resources.

Uncertainty in response

Uncertainty surrounds how management actions translate into pollutant reductions, water
quality improvements, and changes in living resources and makes evaluating the causes of
implementation and response gaps challenging. Translating policy into appropriate behavioral
change, nutrient and sediment load reductions, and water quality response is a complex
biological, physical, and social process. Implementation policy generates changes in the
behaviors of large numbers of people that are often difficult to observe and sometimes
generates unintended behavioral consequences. The pollution reduction effectiveness of
people’s actions is based on a wide variety of factors including the removal effectiveness of
management actions, BMP maintenance, soil conditions, slope, vegetative cover, distance to
water, and weather conditions that make assessing pollution control effectiveness challenging.
Nutrients and sediments in the Bay interact with changing physical habitat, biological activity,
and climate conditions to produce changes in DO and water clarity. Living resources respond to
these conditions, but the size of that response is influenced by other changes that affect fish
and shellfish abundance, such as water temperature, habitat, and harvest rates. In other words,
the causal chain from TMDL policy to living resource response is long and contains factors that
are occurring simultaneously, not always easily observable or measurable, and often outside of
management control.

Thus, this report emphasizes that Bay water quality management decisions will always be made
under uncertainty. The dotted lines above and below the expected response lines in figure 2.5
illustrate that the actual (realized) system response is not precisely known. Sources of
uncertainty exist throughout the system, including, for example, how people respond to
programs designed to improve water quality (e.g., incentive programs that promote BMP
implementation), the effectiveness of load reduction control practices in specific locations, how
water quality measures at different places in the system will respond to changing nutrient
loads, and how different living resources will respond to changes in estuarine water quality.

Two general sources of uncertainty are relevant to making decisions about water quality policy
goals. The first is natural variation or stochasticity (i.e., aleatory uncertainty). Weather and
climate are examples where natural variation generates fluctuations in nutrient loads, estuarine

15
water quality responses, and living resource population dynamics, and are difficult to account
for. Characterizing natural variation and, to the best of our ability, incorporating it into
management decisions are critically important in accounting for uncertainty.

Uncertainty associated with missing, incomplete, or imperfect knowledge (i.e., so-called

epistemic uncertainty) is distinguished from uncertainty associated with natural variation.
Epistemic uncertainty describes the limits to scientific or technical understanding of
relationships or processes impacting water quality. For example, there may be uncertainty
about the relative magnitude of N removal pathways (e.g., for surface water, groundwater, or
atmospheric sources), the effectiveness of BMPs in different settings, or people’s nutrient use
behavior. In other cases, an underlying conceptual relationship may be well understood, but
data availability or modeling capacity presents a barrier to applying the understanding more
broadly.

Both types of uncertainty are contained in the representation of the dotted lines in figure 2.5.
The dotted lines are illustrative because epistemic uncertainty presents challenges with even
identifying the precise bounds around expected responses.

In this report, we also strive to delineate decision-relevant uncertainties rather than all
uncertainties. A decision-relevant is defined as an uncertainty that, if resolved, may change a
management or policy decision (e.g., what pollutant control technology to implement or how to
express water quality goals). The restoration effort will always require decision makers to make
decisions under uncertainty and assess the results in order to learn, meaning that we knowingly
recognize that one kind of risk (i.e., that our selected management and policy actions may need
to be improved or revised) is being accepted to avoid another (i.e., the outcomes of continuing
to make choices with incorrect information). Assessing and managing this balance requires that
we formally assess the efficacy of our actions and their unintended consequences.

Uncertainty is an inescapable reality of future Bay water quality management (Hershner, 2011).
Recognizing the uncertainty associated with expected responses or any gaps is not a reason to
delay efforts to improve water quality. Rather, recognizing response gaps and their potential
causes, as well as the uncertainty that surrounds all expected responses, offers opportunities to
improve our understanding of how the system works, improve policy effectiveness, and
accelerate progress toward meeting goals. With this understanding, recognizing and reducing
uncertainty and gaps and, in response, changing water quality management and policy
decisions (yellow boxes in figure 2.5) are critical to future program success.

16
3. Nutrient and Sediment Response to Management Efforts

A key issue confronting the CBP as the 2025 TMDL implementation deadline approaches is
whether the physical and social systems are responding to implementation in ways sufficient to
meet nutrient and sediment reduction targets. This chapter addresses the following questions:
● Is the physical-social system responding to management efforts to meet TMDL N, P, and
sediment targets in ways consistent with expectations?
● What are possible gaps and uncertainties confronting efforts to reduce N, P, and
sediment delivered to the Chesapeake Bay?

To date, investments in wastewater treatment have produced tens of millions of pounds of

point source nutrient reductions, and atmospheric deposition of N has decreased steadily for
several decades. These are noteworthy achievements (Lyerly et al., 2014). Despite these
successes, achieving the Bay TMDL depends primarily on additional agricultural and urban
nonpoint source load reductions.

Achieving nonpoint source reductions has proven more challenging than anticipated when the
first nutrient reductions targets were established in the early 1990s. The challenge is twofold.
First, voluntary nonpoint source programs struggle to produce the scale of behavioral change
and practice adoption necessary to achieve water quality goals, i.e., an implementation gap.
Second, the nonpoint source programs and practices implemented may not be as effective as
expected at reducing nonpoint source pollution, i.e., a response gap.

Two key objectives of this report are to (1) identify the gaps and uncertainties that inhibit TMDL
attainment and to (2) identify management options that can improve pollutant control efforts
in the face of uncertainty. Existing policies have limited capacity to induce the behavioral
change required to meet TMDL nutrient targets. Meeting TMDL targets will require new
approaches to nonpoint source management. How much additional nonpoint source reduction
can be achieved and at what cost is uncertain. A major challenge confronting the CBP
partnership is addressing a complex system in the face of great uncertainty (Freedman et al.,
2008; Hershner, 2011; Shabman et al., 2007).

3.1. Nonpoint source implementation policies in the Chesapeake Bay watershed

Assessing the effectiveness of nonpoint source programs requires an understanding of the CBP
nonpoint policies themselves. The CBP approach to nonpoint policies is built on two
foundations. First, Chesapeake Bay jurisdictions implement programs designed to induce
agricultural and urban land managers (including landowners and those who manage lands) to
change behavior to reduce nutrient and sediment loads. Outside of MS4s and CAFOs, which are
regulated to various extents, these programs primarily rely upon incentives to encourage land
managers to voluntarily adopt BMPs. These BMPs are intended to reduce nutrient inputs, retain
nutrients and sediment on the landscape, or transform nutrients into less damaging forms (e.g.,
biologically available forms of N converted to inert N). The second foundation is the CBP TMDL

17
modeling and accounting system that calculates pollution reductions and assigns pollutant
reduction credit to state and local jurisdictions for implementing BMPs. The two foundations of
nonpoint source policy are not independent. The structure of the TMDL modeling and
accounting system influences people’s incentives and behavior.

Agricultural nonpoint source policy generally allows land managers to decide how to manage
their operations and whether or how to reduce nonpoint source pollution (Pannell & Claassen,
2020; Shortle et al., 2021). Information and education programs inform agricultural managers
about BMPs and the availability of government funding to encourage their adoption. Because
most BMPs are costly, and in many cases reduce producer incomes (i.e., private BMP costs
outweigh private BMP benefits), state and federal programs cover a portion of the costs for
implementing BMPs (Shortle et al., 2021). Most state and federal financial assistance takes the
form of cost sharing, which typically pays for a portion of the cost of BMP installation and, in
limited circumstances, some annual operation and maintenance costs (Ribaudo, 2001; Ribaudo
& Shortle, 2019).

For over three decades, federal and state governments have been committed to funding
agricultural financial assistance programs to encourage BMP implementation in hopes of
meeting nonpoint source reduction targets. Federal and state efforts have been successful at
increasing funding to support these programs.

In most states, urban nonpoint source programs assign municipalities and larger industrial sites
responsibility for reducing N, P, and sediment loads from urban lands under their jurisdiction
through MS4 permits (details differ by states).1 Municipal separate storm sewer systems can
meet these requirements using a variety of BMPs. They can implement BMPs on public land,
often upgrading existing stormwater infrastructure and stream restoration projects, or in
private developments using common stormwater BMPs (Gonzalez et al., 2016). Given the
competition for urban land, BMP implementation opportunities are often limited and
exceedingly expensive.

The CBP CAST model credits nonpoint source load reductions by either BMP nutrient removal
efficiencies, land use change, or nutrient source reductions. The CAST model calculates nutrient
load reductions via BMP removal efficiencies by multiplying model-based estimates of average
nutrient runoff (lb/ac) by an assigned BMP removal efficiency (e.g., practice reduces runoff by
60%) and the number of acres treated by the BMP (see figure 3.1). Expected runoff is based on
average precipitation over a baseline 3-year period. The model calculates nutrient loads as an
average over relatively large areas of approximately 20,000 acres, called land-river segments,
for different land use types within each segment (e.g., row crop production, hay production,
etc.). These area averages are based on estimates of commercial fertilizers (purchases), animal
manure production (based on estimates of livestock numbers), atmospheric N deposition

1 States also operate stormwater permitsystems for industrial facilities. Some, but not all, urban stormwater loads
originate from lands managed by permitted sources.

18
(provided by an airshed model and reanalysis of observations), and segment physiography. The
CAST model estimates long-term loads delivered to the Bay.

FIGURE 3.1—Chesapeake Bay Program modeling of a typical nonpoint source BMP load reduction.

The BMP pollutant removal efficiencies are generally a single number (e.g., 30% N removal for a
vegetative buffer strip) averaged across the watershed or very large geographic regions. BMP
removal efficiencies are generated based on input from a group of subject area experts
(Stephenson et al., 2018). The CBP currently lists nearly 300 specific agricultural and urban
BMPs, although several BMPs include a number of variations (e.g., CBP lists over 100 cover crop
BMPs). BMPs must be vetted by the expert panels and approved by the CBP before any nutrient
or sediment reductions can be assigned for implementation of these practices. To get credit
toward TMDL compliance, jurisdictions report the number and types of BMPs installed and the
number of acres treated by the BMPs.

3.2. Nutrient nonpoint source pollutant response to TMDL implementation policy

Implementation gap

The current TMDL planning targets for the Phase III WIPs for N, P, and sediment are 214.9, 13.3,
and 18,587 million lb/yr, respectively.2 The expected N and P response to implementation of
practices for the TMDL is shown in figure 3.2. As of 2021, CAST predicts that sufficient practices
have been installed to meet the sediment goal and 91% of the needed P reductions to achieve
TMDL load targets (from a 1985 baseline). Figure 3.2 illustrates the N reduction goal is the most
difficult pollutant target to achieve; CAST estimates that only 73% of the N reduction has been
achieved. Note: figure 3.2 does not reflect any lag times between BMP implementation and
observed changes in pollutant loads.

2 The N load target assigned to the Chesapeake Bay jurisdictions is 199 million lb/yr, with the difference between

199 and 214.9 attributed to atmospheric N deposition.

19
FIGURE 3.2.—Progress toward meeting TMDL nutrient reduction goals and expected response to watershed
management efforts as calculated by the CBP CAST model (Source: CBP, n.d.-a).

A substantial implementation gap exists for N. In 2021, the difference between expected N
reductions and the TMDL target was estimated by CAST to be about 43 million lb/yr (see figure
3.2). Closing this N implementation gap will be difficult. Since 1985, the CAST model estimates
that annual N loads to the Bay have been reduced by 112 million pounds (370 to 258 million
lb/yr, see figure 3.2). Point source WWTPs and reduced atmospheric deposition of N are
responsible for three-quarters of those reductions. Since most large WWTPs are operating at or
near the limits of technology, point sources can make only a modest contribution to closing the
N implementation gap. As of 2021, point sources and atmospheric sources account for only 18%
of the total N load to the Bay. The remaining N load is from either nonpoint sources (64%) or
natural background sources (18%).

To what extent the N implementation gap can be closed depends largely on reducing nonpoint
sources, and nonpoint source reductions have historically been difficult to achieve. The CAST
model calculates that it took 36 years to reduce agricultural and urban nonpoint source N loads
by 27 million lb/yr (1985 through 2021). Recently, nonpoint source reductions have been more
difficult to achieve. Despite concerted efforts since the TMDL was adopted in 2010, the CAST
model estimates that the annual amount of nonpoint source N loads reaching the Bay has been
reduced by only 3.5 million pounds during the period 2009–2021. The 3.5-million-pound
nonpoint source reduction represents just 9% of the total 40 million lb/yr of N reductions
achieved over the same period from all sources.

20
Finally, the N implementation gap is an underestimate of the total reductions needed to meet
the TMDL. The 43 million lb/yr gap does not include additional needed reductions associated
with the infill of Conowingo reservoir (roughly 6 million lb/yr of N) or additional reductions
needed to meet the DO criteria because of climate change (initial estimates of 9 million lb/yr)
(Shenk et al., 2021). In addition, the CBP has also recently discovered unaccounted for sources
of nutrients (i.e., undercounting millions of animals and missing fertilizer sales, discussed
below) that add millions of pounds of additional N loads to CAST estimates (Blankenship, 2022).

Response gap

Estimates of achieving TMDL targets generally assume that expected pollutant responses to
BMP implementation are accurate. Given the complexity and uncertainties of reducing
pollutant loads across a large watershed, this will rarely be the case. A critical technical and
policy question to address is the extent to which nonpoint source policy and management
actions are as effective as expected. Emerging evidence suggests that nonpoint source
management actions have had mixed success in reducing nutrient and sediment loads and that
response gaps may exist between expected and observed pollutant reductions (Ator et al.,
2019, 2020; Chanat & Yang, 2018; Fanelli et al., 2019; Fisher et al., 2021; Keisman, Blomquist, et
al., 2018; Kleinman et al., 2019; Moyer & Blomquist, 2017; Noe et al., 2020; Roland et al., 2022).
As illustrated in figure 3.3, a response gap would mean that the realized nonpoint source
response (red curve) might be less than what is expected by the CBP CAST estimates (orange
curve). Also, the magnitude of the possible response gap could differ by pollutant. Evidence
suggests that the possible response gap for P is particularly large. A response gap would imply
that additional management effort and implementation would be needed to achieve any given
level of pollutant reduction.

Analysis of ambient water quality trends suggests the possibility of nonpoint source response
gaps (Easton et al., 2023). The U.S. Geological Survey (USGS) monitors and analyzes water
quality through a series of monitoring networks in the watershed. Table 3.1 and figure 3.4
illustrate the long-term (1985–2021) loads and trends in TN, TP, and TSS at nine river input
monitoring (RIM) stations at or near the mouth of major Chesapeake Bay tributaries. The CAST
model estimates that these nine tributaries contribute roughly two-thirds of the nutrients to
the Bay, while point and nonpoint sources discharging into the tidal portions of the Bay
comprise the other third. The numbers in table 3.1 are the average annual loads over the 1985–
2021 time period; the colors indicate trends over those time periods. For example, the yellow
shading for TP in the Susquehanna River at Conowingo indicate no detectable trend. The green
shading for the Potomac illustrates the declining trends for TN, TP, and TSS. Long-term trends
show N loads declining in the four largest tributaries, but P loads are decreasing in just three of
the nine tributaries. Figure 3.4 shows flow-normalized TN, TP and TSS load estimates (gold dots)
and trends (black lines) on a per acre basis. The tributaries that show the most consistent and
sustained decreasing trends in nutrient loads are the Potomac and Patuxent rivers. These two
tributaries also had the highest initial proportion of nutrient loads coming from point sources.
Tributaries where point source nutrient loads comprise less than 10% of total nutrient loads

21
(Choptank, Appomattox, Mattaponi, Rappahannock, and Pamunkey) generally show increasing
or no trend in nutrient loads over the period 1985–2020. Similar mixed results can be found
throughout the watershed (Hyer et al., 2021).

FIGURE 3.3.—Conceptual illustration of possible response gaps and response uncertainty in achieving TMDL
nutrient and sediment targets.

TABLE 3.1—Average load (million lb/yr, 1985–2021) and long-term flow-normalized trends in load of TN, TP, and
TSS at RIM stations (Source: Mason & Soroka, 2022).

TN TP TSS
Susquehanna River at Conowingo MD 135 5.44 3,533
Potomac River, Chain Bridge at Washington, DC 48.6 3.27 2,452
James River at Cartersville, VA 11.0 2.21 1,543
Rappahannock River, near Fredericksburg, VA 4.32 0.64 478
Appomattox River at Matoaca, VA 1.46 0.14 39.4
Pamunkey River near Hanover, VA 1.41 0.16 85.0
Mattaponi River near Beulahville, VA 0.65 0.06 14.8
Patuxent River near Bowie, MD 1.51 0.11 49.0
Choptank River near Greensboro MD 0.55 0.04 5.12
Green shaded cells indicate long-term declining loads; red shaded cells indicate increasing
loads; and yellow shaded cells indicate no statistical trend in loads.

22
Analysis of ambient water quality trends generally shows mixed evidence of nonpoint source
pollution reduction effectiveness. Using the empirical Spatially Referenced Regression on
Watershed Attributes (SPARROW) model that relates ambient pollutant levels to landscape
characteristics and nutrient sources, Ator et al. (2019) found little evidence that agricultural
nonpoint source loads declined between 1992 and 2012. Another statistical analysis of
monitoring data found that while P loads were declining in some regions of the Bay watershed,
those improvements were offset by increases in agricultural P sources in other areas (Fanelli et
al., 2019). A STAC workshop report summarized: “Current research suggests that the estimated
effects of conservation practices [BMPs] have not been linked to water quality improvements in
most streams” (Keisman, Blomquist, et al., 2018, p. 9).

FIGURE 3.4.—Flow-normalized TN, TP, and TSS yields at RIM stations for the period 1985–2020. Figure created by
Chris Mason, US Geological Survey Virginia and West Virginia Water Science Center, Richmond, VA (Data
source: Mason & Soroka, 2022).

23
Figure 3.5, adapted from Ator et al. (2020), presents a comparison between CAST estimates of
delivered N and P loads and an empirical analysis using SPARROW. CAST estimates and
SPARROW’s empirically-driven estimates generally agree that total N delivery to the Bay
declined between 1992 and 2012, driven to a large degree by reductions in point source loads.
Phosphorus loads estimated by CAST, however, differ markedly from empirical SPARROW
estimates. CAST suggests consistent reductions in delivered P loads, with the reductions
occurring across most source sectors. SPARROW suggests that delivered P loads were
increasing and that those increases were largely attributable to both agricultural and urban
nonpoint source pollution (Fanelli et al., 2019).

FIGURE 3.5.—Flow-normalized total and source sector TN and TP fluxes to the Chesapeake Bay for 1992, 2002,
and 2012 estimated with the Chesapeake Bay CAST model and SPARROW model (Source: adapted from
Ator et al. [2020] by Easton et al. [2023]).

The P results have policy-relevant implications because state and local management decisions
designed to meet the WQS are based on CAST results. Program managers are currently focused
largely on the N implementation gap, but analysis of ambient monitoring data suggests
significant response gaps exist, particularly with respect to P reduction goals.

Text box 3.1 illustrates nutrient trends and expected responses for two watersheds, the
Patuxent and Choptank. In 1985, point sources were the dominant source of nutrient loads in
the Patuxent while nonpoint sources contributed the vast majority of nutrient loads in the
Choptank. Patuxent has seen large reductions (exceeding 80%) in point source nutrient loads

24
since 1985, and these reductions are reflected in the declining nutrient loads measured in the
river and a close correspondence between expected (modeled response, blue line) and
observed (orange lines) outcomes. Nutrient loads have generally increased or remained
relatively constant in the Choptank, despite considerable effort to reduce nonpoint source
loads (Fox et al., 2021). The Choptank experience also illustrates the uncertainties and
challenges for realizing pollutant reduction in many nonpoint source-dominated watersheds.

Text box 3.1. Watershed nutrient load trends (19852020) in a point and nonpoint
source dominated watershed, CAST model and flow normalized ambient estimates
Patuxent Choptank
Point source-dominated Nonpoint source-dominated
Point source share of 1985 load: 45-55% Point source share of 1985 load: <1%

CBP CAST model CBP CAST model

Flow normalized ambient load Flow normalized ambient load

Trends
Steadily declining N and P loads, driven Observed nutrient loads (orange line)
by 80%+ reduction in point source loads constant or increasing since 1985
since 1985 Response gap
Close correspondence between CBP CAST Differences between CBP CAST model
model response (blue line) and observed response (blue line) and observed
outcomes (orange line) outcome (orange line), particularly for P
(Source: Data provided by Chesapeake Bay Program)

25
The information in text box 3.1 also shows a large potential P response gap in the Choptank.
The CBP CAST model estimates that P loads have declined since 1985, but ambient water
quality monitoring and the associated trends indicate that P loads have increased steadily since
the mid-1990s.

Uncertainty also surrounds attainment of sediment targets. While CAST estimates that
sediment load reduction targets have been achieved, long-term trend data (based on
monitoring) in most tributaries indicate increasing sediment loads since 1985 (table 3.1),
although more recent trends (2011–2020) indicate some tributaries are improving with respect
to sediment goal attainment. Previous reports have concluded that sediment itself is not the
primary contributor to Bay water quality problems, and nutrients should be the focus of
pollution reduction efforts (Miller et al., 2019).

A variety of explanations for response gaps exist. One possible explanation is lag times, which is
the time between when nutrient and sediment reduction efforts (i.e., BMPs, reduced nutrient
applications, etc.) are initiated and when pollutant reductions are detected through monitoring.
Some BMPs may take years or even decades to fully deliver pollutant load reductions to the Bay
given both the time required for some BMPs to become fully functional (e.g., riparian buffers)
and the time required for nutrients and sediment to migrate from their point of origin to the
Bay (Easton et al., 2023). The CBP readily acknowledges, and scientific studies support, that lag
times influence water quality response to reduction of nonpoint sources (STAC, 2013). One
analysis, using the CBP’s dynamic watershed model, estimated that a considerable portion of
the N response gap could be attributable to lag times. Lag times explain less of the P response
gap, providing evidence that simply waiting will not eventually generate the reductions
expected from management actions (Shenk et al., 2022).3 Response gaps, however, could also
be caused by incomplete or incorrect understanding and representations of behavioral and
physical responses to management efforts or the limitations of the existing monitoring system
to accurately reflect trends that may be occurring (Ator et al., 2020).

The challenge of realizing and sustaining large reductions in nonpoint source loads is not unique
to the CBP. Studies of individual BMPs, or studies of BMP nonpoint source reduction efforts
conducted at a fine scale (i.e., edge of field, headwater basin) with intensive monitoring, have
shown BMP implementation can reduce nutrient and sediment loads to streams (Ator et al.,
2020; Böhlke & Denver, 1995; Denver et al., 2010; Jefferson et al., 2017; Li et al., 2017; Staver &
Brinsfield, 1998). Demonstrating the effectiveness of nonpoint source control efforts at larger
watershed scales has proven more difficult (Lintern et al., 2020; Osmond et al., 2012; Sprague &
Gronberg, 2012; Tomer & Locke, 2011). A recent synthesis review of agricultural conservation
programs concluded that “there has been little evaluation of the incentives-adoption-outcome
chain: that is, how well different incentives promote adoption, whether adoption leads to

3 Note, the ability to accurately model lag times is uncertain. Hood et al. (2021) recommended improvements in
the CBP watershed model to better account for lag times.

26
meaningful and measurable changes in outcomes, and what factors shape these links” (Pineiro
et al., 2021, p. 1).

3.3. Assessment of gaps and uncertainties in efforts to reduce nonpoint source loads

Understanding the possible reasons why nonpoint source reductions have proven so
challenging is of critical importance (Ator et al., 2020; Easton et al., 2023). Three possible
causes that can delay or inhibit achievement of the CBP TMDL nutrient and sediment reduction
targets are: (1) lag times, (2) response gaps, and (3) implementation gaps. If lag times are the
sole reason for limited observed nonpoint source response, then jurisdictions can simply wait
until an unspecified future date for the expected reductions to be observed in the watershed
and the Bay. However, improved understanding of system responses to nonpoint source
implementation programs could enhance effectiveness of pollutant control efforts. The
presence of response gaps suggests that just focusing on installing BMPs to meet TMDL targets
does not ensure actual reductions of nutrient and sediment loads. Implementation gaps also
inhibit future reductions and achievement of TMDL targets. Even if response gaps are
eliminated, nonpoint source programs have yet to induce the type and scale of practice
implementation needed to achieve the TMDL.

The uncertainty that characterizes such a complex system complicates our ability to diagnose
the multitude of possible explanations for the limited observed response in nonpoint source
loads. Figure 3.6 illustrates the major elements in the Chesapeake Bay watershed system that

FIGURE 3.6.—Behavioral and physical system response to Chesapeake Bay nonpoint source control policy.

27
may be characterized by lag times, response gaps, and implementation gaps. In figure 3.6, CBP
nonpoint source implementation policies (yellow boxes) are designed to induce behavioral
changes (open orange boxes) that will ultimately reduce nutrient and sediment loads delivered
to the physical system of the Bay (shaded orange boxes). Behavioral changes are expected
among agricultural and urban land managers as well as federal, state, and local program staff
(water quality managers and technical conservation service providers) who are tasked with
TMDL implementation. The physical system is represented by a simplified watershed nutrient
mass balance portion of the diagram.

Nutrient mass balances quantify the (1) inputs to a system (e.g., livestock feed, fertilizer,
atmospheric deposition), (2) the outputs from the system (e.g., biomass exports such as
agricultural products, losses to water or air), and the (3) changes in nutrient storage in the
system (e.g., in soils, floodplains, or groundwater). The flow of nutrients is represented by solid
arrows in figure 3.6. Movement of nutrients and sediment through the system is influenced by
land application of fertilizers and manures, land use, weather patterns, BMPs, and processing
and transport through soils, groundwater, and stream networks. Ultimately the TMDL objective
is to reduce nutrient and sediment loads delivered to the estuary.

3.4. Lag times and legacy nutrients and sediment

A National Research Council report cautioned that achieving Bay water quality goals could be
significantly delayed by lag times and legacy nutrients (NRC, 2011). The accumulated stores of
legacy nutrients and sediment have been identified as important contributors to the lack of
observable water quality improvement (Basu et al., 2022; Kleinman et al., 2019; Noe et al.,
2020; Sharpley et al., 2013; Stackpoole et al., 2019; Van Meter et al., 2021). Legacy nutrients
result from excess nutrient inputs that accumulate and are stored in soils and groundwater
(Van Meter et al., 2016) and legacy sediment in floodplains (Noe et al., 2020, 2022). Reducing
nutrient inputs and implementing certain types of BMPs designed to remove or transform
nutrients may produce immediate reductions at the point of BMP implementation, but the
collective benefits of these BMPs can take years or even decades to find their way through the
coupled surface water-soil-groundwater system, resulting in significant lag times between BMP
implementation and downstream water quality response (Böhlke, 2002; STAC, 2013). Thus,
legacy nutrients continue to be a source of nutrients to surface water bodies even as
contemporary nutrient loads are reduced (presumably due to BMP implementation).

Legacy P presents a substantial water quality management challenge (Kleinman et al., 2019;
Staver et al., 2014). Legacy P is stored primarily in soils, but it can also be stored in groundwater
(Holman et al., 2008; Meinikmann et al., 2015). In areas with intensive livestock production
(e.g., poultry, dairy), animal manures are typically applied to land as fertilizer. Animal manures
contain more P than N relative to plant needs; as a result of applying manures to meet crop
needs for N, P has historically been applied at rates that exceed crop needs, creating a buildup
of P in soils. Most P management has focused on so-called “soil conservation” strategies, as P is
typically tightly bound to sediment and travels to surface water with eroded sediments carried

28
in runoff. In soils with high P levels, however, P loads to surface water may also be in
biologically-available dissolved forms (Kleinman et al., 2019). The increasing importance of
dissolved P losses from legacy P in soils creates challenges when P management strategies
focus solely on preventing sediment loss. The challenge of remediating legacy P is that
significant P stores in soils can serve as a constant source contributing to dissolved P losses
(Kleinman et al., 2019; Sharpley et al., 2013). Removal of P from soils by crops can take decades
even with no additional applications of P (Fiorellino et al., 2017).

Legacy N exists in both groundwater and soils. Groundwater modeling of the Bay watershed
shows elevated concentrations of N (in the form of nitrate) stored in groundwater in several
regions (Greene et al., 2005). Legacy N is an important element of the contemporary N load.
For instance, in the Susquehanna River Basin, legacy N in groundwater (greater than one year
residence time) was found to contribute nearly 50% of the N load entering the Bay (Van Meter
et al., 2017). The travel time for groundwater discharged into surface water ranges from less
than a year to more than 50 years (Lindsey et al., 2003; Meals et al., 2010; Phillips & Lindsey,
2003; Sanford & Pope, 2013; STAC, 2013), indicating that stores of N will continue to contribute
to surface water loads for decades. Fertilizer nitrates can persist in soils for decades (Sebilo et
al., 2013).

Legacy sediment also introduces potential lag times in system response. Legacy sediment is
defined as eroded sediments from land-disturbing activities stored in uplands, flood plains,
stream channels, and impoundments (Miller et al., 2019). It is estimated that historical erosion
rates through the 19th century were considerably higher than contemporary erosion rates due
to improved soil conservation practices and lower rates of land conversion during the 20 th
century (Noe et al., 2020). A large portion of that legacy sediment is still stored on hillslopes,
footslopes, and valley floors throughout the Chesapeake Bay watershed (Jacobson & Coleman,
1986; Smith & Wilcock, 2015; Walter & Merritts, 2008). The extent to which legacy sediment
can be remobilized and transported to the Bay is highly variable over both time and space
(Miller et al., 2019). More legacy sediments may be mobilized during heavier storms (Opalinski
et al., 2022), and the region is already seeing heavier precipitation, a trend predicted to
continue as a result of climate change (Mallakpour & Villarini, 2017). Sediment can be stored
for long periods of time (years to millennia) and be subject to multiple mobilization and
deposition events before entering the estuary (Pizzuto et al., 2014). In some cases this may
mean that continued erosion of legacy sediment will generate elevated sediment loads well
into the future, despite reduced sediment inputs from upland areas (Jackson et al., 2005).

3.5. Nonpoint source response gaps

The limited nonpoint source response observed to date may also be because nonpoint source
programs are not as effective at producing pollutant reductions as estimated. Possible response
gaps may result from inadequate understanding of behavioral responses to policy,
overestimation of BMP effectiveness, and limited or incorrect model input data, among other

29
causes. Sections of figure 3.5 are reproduced in the following discussion to highlight parts of the
system implicated by each possible cause.

Nutrient use and conservation behavior

Relatively little is known about how conservation planning and

behavior varies across agricultural land managers (Pannell &
Claassen, 2020; Reimer & Prokopy, 2014). The BMP adoption
literature consistently refers to the challenges in identifying
factors that explain farmer adoption of conservation practices or
lack thereof (Patterson et al., 2013; Prokopy et al., 2019; Ranjan
et al., 2019; Reimer et al., 2014). Reimer et al. (2014) argued
that understanding of landowner concerns and adoption
behaviors lags understanding of physical processes associated
with estimating nonpoint source loads.

There is limited information on how, when, and where land managers actually apply fertilizer
and manure to farms and fields in the Bay watershed (Yagow et al., 2016). To calculate nutrient
loads under the TMDL, the CBP makes assumptions about the methods, quantity, and timing of
fertilizer and livestock manure applications across the watershed. The CBP collects information
about total fertilizer sales and livestock numbers (from which manure production is estimated)
at state and county levels. Because of the lack of information on farm level nutrient use, the
CBP then assumes manure and fertilizer are applied within counties according to a complex
formula that considers crop needs and nutrient management plans (CBP n.d.-c, section 3).

The CBP provides credits for land managers who develop different types of nutrient
management plans that outline the rate and timing of nutrient applications. How nutrients are
actually applied given farmers’ risk preferences, changing on-farm constraints, and market
opportunities is unknown, though evidence suggests nutrient management plans and actual
use diverge (Osmond et al., 2015). The potential for some land managers to over-apply
nutrients is particularly high for manure sources because manures are costly to transport, are
concentrated in specific regions of the watershed, and have more uncertain nutrient content
than commercial fertilizers. Agronomists have long recognized the risks of overapplying P given
that the N:P ratios of manures (roughly 2:1) are incommensurate with plant needs (roughly 8:1)
(Kleinman et al., 2017). If manures are applied to meet crops’ N needs, then P will accumulate
in soils.

Voluntary, incentive-based agricultural BMP implementation programs require that land

managers self-select into the program. If nutrient use and nutrient or sediment losses differ
among BMP adopters, then CBP accounting could overestimate the effectiveness of nonpoint
source programs. For instance, if land managers interested in conservation practices have lower
nutrient losses overall, then additional BMPs added by this group will produce lower than
average nutrient reductions. This challenge is illustrated in figure 3.7. In this example,

30
estimated N load reductions are calculated based on a subwatershed-wide average of 20 lb/ac
N in runoff. The watershed (200 total acres) has a willing adopter who has already reduced N
losses to 10 lb/ac, but the neighboring land manager has N losses of 30 lb/ac. If BMPs with a
50% removal efficiency are applied by land managers with high adoption rates (10 lb/ac N in
runoff), only 5 lb/ac of N are reduced (10 lb/ac x 50%), but 10 lb/ac N reduction credit would be
awarded (20 lb/ac x 50%).

There is limited research about how nutrient loads and BMP adoption vary across land
managers. Some research has found that nutrient use can vary widely across land managers.
Pearce and Maguire (2020) estimated P mass balance for 58 dairy farms (a nonrandom sample)
in the Shenandoah Valley and found that surplus P (i.e., P imports minus P exported in products
and manure) ranged from -27.6 to +87.1 lb/ac. High-loss operations with below average
adoption rates could be partially responsible for the nonpoint source response gap.

FIGURE 3.7.—Self-selection challenge under CBP TMDL accounting system resulting from spatial variability in
baseline nutrient losses.

31
Program incentives of service providers and water quality managers

Technical service providers are a conduit between the

entity funding BMP implementation and producers or
land managers, providing engineering, installation, and
maintenance assistance, and facilitating financial
assistance. Agency and program incentives center
around installing practices creditable within CAST
(getting practices on the ground). Because all practices
and operations are generally counted the same within a geographic area, there are no
Chesapeake TMDL-related programmatic incentives for service providers and water quality
program managers to locate areas and operators who generate disproportionately high loads.
Within a land-river segment, a practice adopted by a low-loss or high-loss operator counts the
same, while the amount of effort required to gain adoption may differ. This often means that,
within agricultural areas, service providers such as Natural Resources Conservation Service or
Soil and Water Conservation Districts may only work with cooperative land managers, who
may, as a group, already have relatively low nutrient and sediment loss rates, reinforcing the
self-selection bias described above (Stephenson et al., 2022). Inadequate staffing levels of
technical service providers would compound the challenges of devoting time and staff
resources to high-loss operators.

BMP effectiveness

The CBP may systematically overestimate BMP effectiveness.

There is evidence that estimating BMP effectiveness with models
tends to overestimate BMP performance (Liu et al., 2017, 2018).
Osmond et al. (2012) found that models used in the U.S.
Department of Agriculture (USDA) Conservation Effects
Assessment Project consistently overestimated BMP
effectiveness, and a recent review article (Lintern et al., 2020) noted that most studies showing
water quality improvements used model-generated estimates, while field and monitoring
studies showed mixed or little to no improvement from BMP implementation. Text box 3.2
illustrates these challenges for one of the most common agricultural BMPs, conservation tillage.

Easton et al. (2023) summarized a variety of possible reasons why BMP effectiveness might be
overestimated.

Gaps in the literature. The CBP uses expert panels to generate estimates of average
BMP nutrient and sediment removal efficiencies (Stephenson et al., 2018). The panels
rely on existing studies, and there are often gaps in the scientific literature. For example,
some pollutants and loss pathways are better studied than others. Surface runoff is
better characterized than groundwater leaching, and the long-term fate of stored

32
Text box 3.2. Tradeoffs and uncertainties in nutrient removal effectiveness of conservation tillage
practices

Conservation tillage is often touted as a win-win that improves both water quality and agricultural
profitability. Indeed, the CBP found that conservation tillage reduces surface runoff of N by 2 to 14%
and surface runoff of P by 6 to 72% compared to conventional tillage, depending on type of reduced
tillage method and geographic region (Thomason et al., 2016). Research consistently shows that
conservation tillage practices dramatically reduce soil erosion.

Research also demonstrates the complex tradeoffs and loss pathways surrounding one of the most
widely accepted BMPs (Kleinman et al., 2022). Overall, nutrient removal effectiveness of
conservation tillage depends critically on both site-specific soil/topography/environmental conditions
and nutrient management behavior, all of which can vary significantly across sites and land
managers. Some forms of N, such as nitrate, are highly soluble and can be readily leached through
soils to groundwater. Practices like no-till can enhance water-holding capacity and infiltration rates in
soils, which has been shown to increase nutrient loss risk by facilitating the development of soil
macropores that increase leaching (Thomason et al., 2016).

A number of recent watershed-level studies highlight some of the uncertainties surrounding the
relationship between water quality and conservation tillage. Ator (2019) found that N loads in river
networks in the northeastern United States are positively correlated with increased use of no-till
practices. Using a statistical model (SPARROW), Roland et al. (2022) found that increasing
conservation tillage and cover crop practices by 50% in the Chesapeake Bay watershed would reduce
P loads approximately 6% but would increase N loads to the Bay by 0.9%. In a review of 43 studies,
Daryanto et al. (2017) found mixed and inconclusive results for the overall effectiveness of
conservation tillage to reduce nitrate leaching and runoff losses in agricultural systems.

Conservation tillage can also pose risks of increasing P runoff. Since P tends to bind to soil particles,
reducing soil erosion will reduce the loss of P attached to soil particles. However, evidence shows
how conservation tillage practices can increase subsurface loss of dissolved P (Duncan et al., 2019;
Kleinman et al., 2011, 2015, 2020). By reducing or minimizing soil disturbance, conservation tillage
systems can contribute to P stratification in soils. In stratified soils, P becomes concentrated in the
top layer of soil, reducing potential of P to bind to soil particles and allowing dissolved P to be lost via
surface runoff or groundwater leaching. Stratification can occur relatively quickly, especially when
surface application of fertilizers or manures is combined with no-till. Dissolved P has greater
potential to contribute to eutrophication problems than P attached to soil particles.

Substantial research documents the potential for conservation tillage practices to increase dissolved
P losses (Smith et al., 2019). In a study of research findings on P loss from tillage practices, Daryanto
et al. (2017) found that conservation tillage reduced sediment P losses by reducing erosion but
generally increased dissolved P losses. In some cases, increases in dissolved P can offset reductions in
losses of sediment P (Duncan et al., 2019). Recent increases in harmful algae blooms in Lake Erie
have been associated with increases in the amount of dissolved P arising from increases in the use of
no-till practices (Jarvie et al., 2017; Macrae et al., 2021; Michalak et al., 2013). Water quality analysts
have also observed increases in dissolved P levels in multiple river systems in the Chesapeake Bay
watershed that are linked to agriculture (Fanelli et al., 2019; Kleinman et al., 2019).

33
 nutrients is difficult to characterize within the time frame of typical studies (Jefferson et
al., 2017).

Operation and maintenance behavior. Assumptions must be made about how people
will actually operate and maintain a BMP over time. BMPs must be operated and
maintained according to design specifications, and most BMPs require some type of
routine maintenance to sustain their performance. BMPs that are the subjects of
scientific studies are likely to be operated and maintained differently than in real world
settings. Furthermore, evidence suggests that lack of or inconsistent maintenance is a
frequent problem for both urban and agricultural BMPs (Aguilar & Dymond, 2019;
Jackson-Smith et al., 2010). Additional research is needed to better characterize BMP
operation and maintenance under actual conditions (Liu et al., 2017).

Long-term performance. Evaluations of BMP performance typically assess performance

a few years after installation. For many BMPs, performance varies over time. Many
BMPs function by storing nutrients or sediment, either in the soil or in plant biomass.
Yet the storage capacity of a BMP is limited, and the fate of those stored nutrients or
sediments over time is not well characterized (LeFevre et al., 2015). If stored nutrients
are more likely to be mobilized as the BMP ages, then BMP performance will be
overestimated over time (Hopkins et al., 2020; Selbig & Bannerman, 2008).

Behavioral biases. Behavioral research consistently finds that people are prone to
systematic errors in decision-making in certain situations (Ariely, 2008; Kahneman et al.,
2021). Expert panels are commonly confronted with literature that contains wide ranges
of findings (e.g., from BMPs with removal efficiencies of over 90% to BMPs that increase
nutrient losses under some conditions). In situations with considerable uncertainty,
people (experts included) tend to assign causal explanations to randomly produced
outcomes (Stephenson et al., 2018). This possibility is increased in settings with limited
data, situations most often faced by expert panels.

Watershed data and monitoring

The CBP collects and uses a tremendous amount of data about

the watershed to generate estimates of nonpoint source loads,
evaluate scenarios, and monitor TMDL compliance. Data are
collected on land use, soils, weather, physiography, nutrient
inputs, and BMPs installed. In many cases, data are of high
quality, and numerous processes have been put in place to
improve the quality of data. Land use data are an example.
However, collecting data in a highly complex system is expensive,
necessarily selective, and subject to error.

34
Uncertainty exists regarding the accuracy of input data for the CAST model. For instance, some
jurisdictions express concerns that the CBP undercounts BMPs, therefore underestimating
nonpoint source reductions (Royer et al., 2016). To estimate nutrient inputs, the CBP collects
data on aggregate commercial fertilizer sales across the watershed and estimates the quantity
of manure produced in the watershed based on estimates of animal numbers reported from
National Agricultural Statistics Service (NASS) surveys. Recent evidence suggests NASS survey
data can lead to undercounting livestock, and the CBP undercounted total fertilizer sales for
recent years (Blankenship, 2022).

In addition, limited information is collected about key nutrient generating processes. As already
discussed, little data exist on how, when, and where land managers actually apply fertilizer and
manure in the Bay watershed at the farm- or field-scale (Yagow et al., 2016). Soil P levels are an
important intermediate indicator of the effectiveness of management actions because changes
in soil P are indicative of changes in water quality (Kleinman et al., 2019), but farm-level data on
soil nutrient levels (particularly P) are limited because of privacy concerns. Artificial drainage
(surface ditch and subsurface tile drainage) of agricultural fields can increase nutrient loss by
providing direct pathways to water bodies (Bryant et al., 2019; Kleinman et al., 2019) but is not
explicitly accounted for in the CAST modeling system (Bryant et al., 2019). Artificial drainage
systems for agricultural fields are common in coastal regions of the watershed, and their use is
increasing (Bryant et al., 2019). Bryant et al. (2019) estimated that hundreds of miles of
agricultural drainage have been installed by farmers without cost-share assistance, so the
extent of these additions is largely unknown.

Finally, ambient water quality monitoring data are often not collected at the scale that can
provide insights into the effectiveness of management actions (Easton et al., 2023; Ator et al.
2020). The CBP Nontidal Monitoring Network (NTN) was established to assess watershed
response to nutrient and sediment reduction efforts (CBP, 2022). Selection of sites was driven
largely by the location of historical monitoring stations. Furthermore, the NTN monitoring sites
tend to be located on larger streams and rivers (with drainage areas ranging from 50 to over
1000 square miles), thus limiting the ability to distinguish between multiple factors affecting
changes in monitoring data at more localized scales. Jurisdiction partners have limited
incentives to invest in additional and finer-scale ambient instream monitoring.

3.6. Implementation gaps in achieving nonpoint source pollutant reductions

Several technical and management issues affect the ability to secure the type and level of
implementation necessary to achieve nonpoint reduction goals and TMDL targets. Important
factors contributing to implementation gaps include (1) nutrient mass imbalances, (2) limited
participation in voluntary incentive programs, and (3) limited spatial targeting of efforts to
encourage BMP adoption.

35
Nutrient mass imbalances

From a system
perspective, nutrient
mass balances are critical
determinants of the
nutrient status of a given
region (Easton et al.,
2023). If inputs (e.g.,
fertilizers and livestock
feed brought in) exceed
outputs (e.g., crops and livestock harvested and removed), then excess nutrients may find their
way to surface water or groundwater.

Available data indicate that overall nutrient inputs from fertilizer and livestock feed, two
primary nutrient sources, have declined across the entire watershed since the 1980s (Clune et
al., 2021; Keisman, Devereux, et al., 2018; Sabo et al., 2022). Declines have been driven largely
by reductions in fertilizer use generated by improvements in nutrient use efficiency. Reductions
in P fertilizer inputs were particularly pronounced between 1980 and 2012 (Keisman, Devereux,
et al., 2018). Areas with substantial declines in nutrient inputs are closely associated with
improving water quality trends in nonpoint source-dominated watersheds (Sabo et al., 2022).

Nutrient inputs are unevenly distributed across the watershed, however, and in many places,
nutrients have become more concentrated over time (Keisman, Devereux et al., 2018). Regions
with high imports of nutrients in the form of fertilizer and animal feed are associated with mass
imbalances, and these are also areas with high nonpoint source loads. While mass imbalances
can also be high in many urban areas, the most severe and extensive mass imbalances are in
regions dominated by intensive agriculture (Keisman, Devereux, et al., 2018). Increases in these
areas that already receive higher than average nutrient inputs are being driven largely by
increases in livestock numbers (primarily poultry) and agricultural intensification (Ator &
Denver, 2015; Kleinman et al., 2012, 2019). According to USDA Census of Agriculture data,
poultry numbers in the Lower Susquehanna basin and on the Delmarva Peninsula increased
from 2002 to 2017 by 20.5 million (from 37 million, a 64% increase) and 70.8 million (from 143
million, a 66% increase), respectively (Easton et al 2023). Sabo et al. (2022) found surplus
amounts of nutrients in some agricultural areas of the Bay watershed have increased from 2009
through 2019, reversing longer-term declines since the 1980s. Across the Bay watershed,
livestock produce about 10 times more excrement by mass volume than the human population
(Kleinman et al., 2012), and yet law requires some form of wastewater treatment for all human
waste to reduce nutrient content.

Mass nutrient imbalances limit the amount of long-term progress that can be made in reducing
nonpoint source loads (Ator et al., 2020; Beegle, 2013). Addressing existing mass imbalances

36
would require reducing the import of feed and fertilizer, increasing the uptake of nutrients and
their export out of the region or watershed in food and fiber, hauling manure out of the
watershed, adopting manure conversion technologies, or implementing appropriate
wastewater treatment mechanisms (Flynn et al., 2023; Kleinman et al., 2012; Sharpley et al.,
2013; Spiegal et al., 2020). Appreciable reductions in nutrient loads cannot be achieved unless
regional mass imbalances are successfully addressed.

Many conventional agricultural BMPs that are integral elements of WIPs do not appreciably
change the mass balance. Regional mass imbalances in areas of intensive livestock production
are particularly problematic for P. Phosphorus removal efficiencies of cover crops, conservation
tillage, and riparian buffers tend to result from altering transport pathways, increasing
temporary nutrient storage, or altering the form of nutrients (Kleinman et al., 2022). Drawing
down soil P levels within a reasonable time frame requires reducing P inputs and some sort of P
mining (e.g., by P-scavenging plants).

Behavioral response to nonpoint source reduction policy

While local success stories exist, voluntary, conventional (cost

sharing) incentive-based programs targeting the agricultural
sector have not consistently generated the type or scale of
BMP implementation sufficient to produce the reductions
needed to achieve WQS (Liu et al., 2020; Patterson et al.,
2013; Prokopy et al., 2019; Reimer & Prokopy, 2014; Ribaudo
& Shortle, 2019; Shortle et al., 2012, 2021; Stephenson et al.,
2022). Multiple reasons exist that can limit adoption of effective pollutant control practices.

The structure of practice-based cost-share incentive programs can limit choices about the type
and extent of adoption. Research generally finds that conventional financial assistance
programs encourage the adoption of practices with significant private benefits and low upfront
costs (Claassen et al., 2014; Pineiro et al., 2021), such as conservation tillage (which lowers
operating costs) and cover crops (which increase long-term soil productivity). Perceived risk to
farm profitability is a major factor limiting adoption (Duke et al., 2022; Wilson et al., 2014).
Unsurprisingly, adoption rates tend to be much lower for practices with higher upfront costs
and few private on-farm benefits, such as stream buffers, denitrifying bioreactors, stream
fencing, and manure treatment. Many such BMPs, however, can generate substantial nutrient
reductions at relatively low per unit ($/lb) costs to the public (Easton et al., 2019; Price et al.,
2021). Yet, if financial returns are a primary motivator, a land manager will not install and
operate a BMP with few agronomic benefits, even if a portion of the costs are shared (Prokopy
et al., 2019; Ribaudo, 2015). There is limited evidence that temporary cost-share payments lead
to sustained adoption (Pannell & Claassen, 2020). Similar logic applies to landowner adoption of
structural urban stormwater controls (Gonzalez et al., 2016). Sharing the costs of a practice
with high upfront costs is often insufficient to generate voluntary investments in BMPs with

37
high public benefits. The problem is compounded when the purported public benefits accrue
downstream.

The adoption of BMPs can also have unintended consequences for nonpoint source pollution.
Taking land out of production to install riparian buffers may induce more intensive upland
cropping practices, partially offsetting the pollutant reduction gains from the buffer (Bonham et
al., 2006). Practices and cost-share payments that improve financial outcomes may expand and
intensify agricultural production, potentially offsetting the pollutant reductions gained from
BMP implementation (Fleming et al., 2018; Lichtenberg & Smith-Ramírez, 2011). Cover crops
are an example. Research consistently finds that cover crops (winter ground cover) improve soil
conditions, significantly reduce soil erosion, and potentially reduce nutrient runoff (Blanco-
Canqui, 2018; Koudahe et al., 2021). However, the nutrient reduction potential of cover crops is
contingent on the land manager’s nutrient application behavior. Experience and success with
cover crops may lead farmers to plant more cash cover crops (winter wheat, for example),
intensifying production and encouraging additional nutrient applications, partially offsetting
some of the water quality benefits associated with cover crops.

Research also suggests other factors that limit adoption (Easton et al., 2023). Cultural norms are
one example. Members of plain sect communities in Pennsylvania and Virginia, a significant
portion of the agricultural communities in these regions, do not participate in government
assistance programs, including cost-share programs. These farmers, as well as small, land-
constrained, or resource-limited agricultural operations often have systematically lower BMP
adoption rates, at least under typical cost-share programs (Prokopy et al., 2019). Finally, many
agricultural land managers lease or rent land, which also lowers BMP adoption (Collins et al.,
2022; Ranjan et al., 2019).

Targeting nonpoint source investment

The CBP offers few tools or incentives to identify and treat site-specific high-loss areas (Easton
et al., 2019; Stephenson et al., 2022). While CAST can identify high-loading areas at a relatively
coarse spatial scale (averaged over thousands of acres, see figure 3.1), it does not reflect
localized high-loss areas at field or farm scale (Easton et al., 2020; Lintern et al., 2020).
Researchers have noted that areas of high nutrient and sediment loss are site-specific and
highly localized (Easton et al., 2017; Easton et al., 2008), and many studies suggest that 5–20%
of the land area generates 50–90% of runoff and nonpoint source loads (Heathwaite et al.,
2000; Qui, 2009; Rao et al., 2009; Wagena & Easton, 2018; White et al., 2009; Xu et al., 2019).
Improving the ability to spatially identify and target BMPs to sites and operations with higher
pollution potential could improve water quality and reduce costs of pollution reduction efforts
(Choi et al., 2020; Giri et al., 2012; Kast et al., 2021; Khanna et al., 2003; Lintern et al., 2020; Xu
et al., 2019; Yang & Weersink, 2005).

38
Incentives and ability to reach nonadopters

Existing cost-share program rules also can limit the program effectiveness of service providers
(Collins et al., 2022). Highly localized hotspots may mean that a large percentage of loads could
be treated on a relatively small portion of the farm operation. Most cost-share program rules
require whole farm plans and would prevent targeted treatment for land managers reluctant to
treat the entire operation.

Similarly, strict technical standards for BMP installation have been cited as a barrier to BMP
adoption (Collins et al., 2022). For instance, land managers willing to fence livestock away from
streams or install riparian buffers are ineligible for program benefits if their efforts do not meet
all technical criteria, even if pollutant control impacts are positive. Also, cost-share funding
rules prevent service providers from offering additional financial assistance to reluctant or
resource-limited land managers even if such assistance could produce large quantities of low-
cost pollutant reductions. While such programmatic conditions serve several important
functions, they also represent a barrier to increasing adoption among reluctant, skeptical, or
resource-limited land managers.

Incentives for innovation in nonpoint source load reductions

Technological innovations to reduce nutrient inputs, increase efficiency of nutrient use, and
treat existing sources can accelerate progress toward meeting nonpoint source reduction goals.
However, barriers limit development and adoption of such technologies (Stephenson et al.,
2022). In CBP nonpoint source programs, all BMPs must be approved by the CBP through the
BMP protocol process to be credited toward meeting the TMDL. For new BMPs, the approval
process can take years, creating a high cost of entry for new, and potentially innovative, BMPs.

In addition, the existing accounting framework and incentive systems provide limited incentives
for people to invest in actions to improve the certainty of achieving nonpoint source pollutant
outcomes. Typically, practices are assigned a fixed pollutant reduction credit, but actual
pollutant removal performance is often both highly variable and uncertain. A number of
technologies offer opportunities to directly or indirectly measure pollution outcomes (Rose et
al., 2015; Stephenson et al., 2018; Stephenson & Shabman, 2017a), but their application is rare
because TMDL accounting automatically credits reductions associated with installed practices.
Activities that provide certainty (e.g., measurement of nutrient load from a practice or
intermediate indicators of effectiveness such as changes in nutrient levels in soils) add costs to
practices. The structure of incentives, however, provides people little reason to pay for greater
assurances that pollutant reductions occur.

Climate change

Climate change poses an array of challenges to meeting the Bay TMDL. Some of these
challenges, such as increased precipitation and streamflow, are widely recognized for their

39
potential to increase nutrient and sediment delivery to the Bay (Hanson et al., 2022; Ryberg et
al., 2018; Sinha et al., 2017). Indeed, the TMDL Phase III WIPs now require all Bay jurisdictions
to account for the additional nutrient and sediment loading expected from climate change.

The primary climate-related drivers affecting the Bay watershed are air temperature,
precipitation, and sea-level rise (Najjar et al., 2010). Changes in these drivers are expected to
alter key processes within the Chesapeake Bay and its watershed, including the duration,
frequency, and magnitude of precipitation; evapotranspiration; soil moisture; streamflow; and
temperature. Climate change will also affect watershed water quality by indirect means, with
changes in agricultural land use (e.g., different cropping mixes, increases in artificial drainage)
and increased agricultural intensification (e.g., double cropping) in response to longer growing
seasons and changes in rainfall patterns. This could also fundamentally alter nutrient mass
balances, and consequently the cycling and export of nutrients, in ways that are not fully
understood.

3.7. Conclusions and implications

Achieving nutrient and sediment reduction targets under the TMDL is going to be more difficult
than is generally acknowledged. Because of the success of efforts to reduce point source N and
P and atmospheric N deposition, agricultural and urban nonpoint sources are the largest
remaining sources of controllable nutrient loads. Tens of millions of pounds of nutrient
reductions are needed to meet the TMDL, but a decade of intensive management efforts have
produced only modest (as predicted by CAST) reductions in nonpoint source loads. Statistical
analyses of monitoring data provide some evidence that nonpoint source practices are not
producing the reductions expected. Given the growth of animal production in the watershed,
conservation practices and efficiency improvements in agricultural production have helped
prevent significant increases in nutrient pollutants. Despite successes of point source and
atmospheric reductions, the TMDL targets are not likely to be achieved without increasing
efforts to address nonpoint source loads—in particular the gaps that limit progress toward
achieving nonpoint source load reductions.

Nonpoint source policy relies heavily on voluntary, practice-based incentive and education
programs to induce behavioral change. Empirical evidence suggests that nonpoint source
management programs have not produced the type and scale of behavioral changes needed to
close the implementation gap.

Data indicate that efforts to reduce nonpoint source loads are not as effective as expected.
These response gaps make achieving the Bay TMDL more difficult. Phosphorus is of particular
concern, and while CBP reports significant progress toward achieving TMDL P goals, statistical
analyses of ambient monitoring data often show no change or increasing P loads. Dissolved P
loss, a form of P that contributes the most to algae growth, is increasing in many regions of the
watershed, particularly in regions associated with nutrient mass imbalances. The difference
between realized and expected nonpoint source outcomes could be due to the large stores of

40
legacy nutrients and sediment in the watershed that delay fully realizing nonpoint source
reductions. Evidence also suggests a variety of reasons why our understanding and
characterization of the effectiveness of nonpoint source control programs may be incomplete
or inaccurate.4

STAC concludes that additional effort and funding through existing programs alone is unlikely to
produce the type and magnitude of changes needed to meet TMDL targets. Given the scale of
the challenge, additional amounts of technical assistance and effort will be necessary to help
close the implementation and response gaps. However, new and innovative options need to be
considered to improve the effectiveness of nonpoint source programs, especially toward
programmatic and behavioral changes.

Ways to address these gaps are not always obvious and surely will not be easy. Indeed, the
challenge is not unique to the CBP; nonpoint source management has long been a “wicked”
problem (Patterson et al., 2013; Shortle and Horan, 2017; Wiering et al., 2020). The inability to
produce nonpoint source reductions on the scale sought by water quality management
programs is one of the most fundamental and common challenges confronting large-scale
efforts to address eutrophication (Boesch, 2019; Wiering et al., 2020). Effective nonpoint
source control at large watershed scales has proven difficult for water quality programs in the
United States (Lintern et al., 2020; Osmond et al., 2012; Sprague & Gronberg, 2012; Tomer &
Locke, 2011).

Recognizing these challenges, there are still opportunities to improve program effectiveness
(Easton et al., 2023). Given uncertainties, not all alternatives are guaranteed to work as
expected, but adaptive management, innovation, and experimentation are likely needed to
address the nonpoint source challenge (NRC, 2001; NRC, 2011). Below are some opportunities
to improve program outcomes.

Improved targeting of nonpoint source investments. Research consistently highlights the

potential of targeting nonpoint source investments to reduce nonpoint source loads and costs
(Choi et al., 2020; Easton et al., 2019; Giri et al., 2012; Kast et al., 2021; Khanna et al., 2003;
Lintern et al., 2020; Xu et al., 2019; Yang & Weersink, 2005). Distribution of nonpoint source
loads is highly uneven across watersheds, but existing tools, TMDL accounting rules, budget
allocations, and incentives limit options for targeting nutrient hotspots. Targeting investments
to areas of the landscape that produce disproportionately large loads offers opportunities for
increasing program effectiveness. This could be accomplished in a number of ways. Developing
and using finer-scale (field-level) tools to identify high-loss areas is one example. Additionally,
developing new financial incentive programs, such as pay-for-performance or pay-for-success
programs, could reward treatment of high-loss areas or operations and encourage adoption of
potentially more effective practices (Talberth et al., 2015). Nonpoint source program outcomes

4 Ranking or isolating the relative contributions of the various explanations for the response gap is beyond the
scope of this report.

41
could be improved by designing programs and incentives to shift attention from installing
practices to maximizing reductions.

Focus on nutrient mass imbalances. More effective and systematic approaches to addressing
nutrient mass balance issues offer opportunities for substantial, sustained reductions in
nonpoint source nutrient loads. A mass balance approach describes inputs (e.g., fertilizer and
feed) and outputs (e.g., grain or meat export, loss to water bodies) to and from the system,
reactions or transformations (e.g., denitrification), and storages (e.g., build-up of P in soil) in the
system. Most BMPs do not substantially alter mass balances. Evidence suggests that policies
designed to alter regional mass balances have proven particularly effective in improving water
quality. The largest regional mass imbalances, particularly P, are associated with intensive
livestock agriculture. Given the trends towards increasing animal numbers and more intensive
livestock operations, traditional BMPs like cover crops or no-till that do not substantially alter
inputs or transformations are unlikely to substantially alter nutrient losses. Reducing the
nutrient content of livestock feed through precision feed management as a way to manage
nutrient imports can address mass imbalances (Cerosaletti et al., 2004). Manure utilization,
treatment, and conversion technologies (e.g., thermochemical or microbial transformation)
offer the potential to significantly alter the mass balance by transforming manure-derived
nutrients into less biologically available forms or converting manure into forms more easily
transportable. Manure transport programs that export excess nutrients to nutrient deficit areas
can be strengthened, for example by moving to more restrictive P-based nutrient management
plans (Saha et al., 2022).

Requiring more. A key element of many nonpoint source policies is that land managers
voluntarily decide whether or how to participate in nonpoint source programs. Such an
approach is often reasonable given the diversity and number of people involved in producing
nonpoint pollutants. However, the extensive history of nonpoint policy illustrates the limits to
voluntary adoption, regardless of the type of incentive programs used. To make substantial
progress in reducing nonpoint source loads, new and refined policies in strategic nonpoint
source sectors may be required. Such requirements need not be overly costly to the land
manager if well-designed and accompanied by financial assistance.

Encourage innovation. Finally, improving implementation effectiveness could be facilitated by

encouraging and providing opportunities for innovation. Given that exemptions in the CWA
limit federal authority over nonpoint sources, responsibility for addressing nonpoint sources
rests to a large degree with state and local governments. The complexity and challenges of
nonpoint source management also imply there is no one best way to manage nonpoint sources.
The CBP could encourage innovations in nonpoint source management by supporting state and
local governments to develop new approaches for managing nonpoint sources. Sandboxing is a
formalized way to test and evaluate the efficacy of new rules and programmatic approaches to
nonpoint source management (Higgins & Male, 2019; O’Sullivan, 2021). In a sandbox, a
government agency allows people to develop and implement an innovative rule or approach to
achieving a common objective without disrupting the operation of existing programs.

42
Government agencies must grant permissions for sandboxing and be willing to provide
technical support to adequately test the efficacy of new approaches.

43
4. Estuary Water Quality Responses to Nutrient and Sediment Load Reductions

In this chapter, we examine the progress, ongoing challenges, and future opportunities in the
evaluation of estuary water quality responses to reductions of nutrient and sediment loads
from the watershed. We address whether nutrient and sediment load reductions from the
watershed initiate and sustain the physical, chemical, and biological responses necessary to
meet the stated WQS and, ultimately, protect the aquatic living resources for which the WQS
were developed. This chapter addresses the following questions:
● Is estuary water quality responding in ways consistent with the expected response to
nutrient and sediment load reductions achieved to date?
● What are the major uncertainties in assessing attainment of Bay WQS, including DO,
water clarity/SAV, and algal biomass (Chl a)?
● What are the major uncertainties in efforts to attain Bay WQS (DO, water clarity/SAV,
Chl a)?

The expected response curve in figure 4.1 anticipates the response of estuary water quality to
nutrient and sediment load reductions. The expected response curve rises to 100% attainment
of WQC (represented by the black dashed line at the top of the graph) when the expected load
reductions set by the TMDL are achieved. Expected responses are determined by the CBP
estuarine models (Hood et al., 2021), which predict variations in Bay circulation and water
quality due to changes in input loads, atmospheric forcing, and climate change effects
(temperature, precipitation, and sea-level rise). Given the complexity of estuary biological and
chemical processes, changing environmental conditions, and diversity of habitats, significant
scientific uncertainty exists around the ability to explain (from monitoring results) and predict
(from the estuarine models) the level of WQC attainment from any given reduction of nutrient
and sediment loads to the Bay (dotted curves in figure 4.1).

Chapter 3 established that we have not yet realized the nutrient and sediment load reductions
from the watershed required to meet the TMDL targets due to implementation and response
gaps. In this chapter a response gap (figure 4.1) is defined as the difference between the
expected response of estuary water quality to a given reduction of nutrient and sediment loads
(yellow dot in figure 4.1) and the realized water quality response to those reductions (red dot in
figure 4.1). If actual water quality response is below the expected response (as shown in figure
4.1), attainment is more difficult; likewise, if realized response is above the expected response,
attainment is easier than current science suggests. Accordingly, management concern is
focused on response gaps that may make achieving the WQC more difficult.

4.1. Water quality criteria and conditions for attainment

In developing the WQS for Chesapeake Bay, eutrophication was identified as the primary
challenge to achieving conditions suitable for aquatic life. Figure 4.2 shows primary degradation
factors for water quality that result from eutrophic (nutrient-rich) conditions. (Note that the
figure does not include climate change effects such as sea level rise and water temperature

44
FIGURE 4.1.—Conceptual water quality response to nutrient and sediment reductions, where attainment of WQC
is a function of nutrient or sediment load reductions.

increase). The Bay WQC reflect the need for more water clarity, more SAV, more DO, and less
Chl a in the interest of achieving the designated use of the Bay, which is to support aquatic life.
Reducing nutrient and sediment loads improves water clarity and, thus, improves light
conditions for SAV. Chlorophyll a is an indicator of algae growth, which must be reduced to
address both the DO and water clarity criteria. The WQC are predicted to be achieved if the
TMDL is met (USEPA, 2010).

FIGURE 4.2.—Conceptual diagram of primary degradation factors for water quality in eutrophic (nutrient-rich)
conditions as compared to low nutrient input conditions (Source: Testa et al., 2018).

45
4.2. Estuary response to realized nutrient and sediment loads

The factors illustrated in figure 4.2 underlie the responses of the Chesapeake Bay estuarine
system to nutrient and sediment load reductions, and figure 4.3 illustrates specific processes
associated with these factors. The shaded orange box in figure 4.3 represents the nutrient and
sediments loads entering the Bay from the watershed, as generally described in Chapter 3. CBP
implementation policies (working through the processes illustrated in figure 3.5) are
responsible for generating load reductions, and the CBP partners are responsible for monitoring
and assessing Bay water quality (yellow boxes in figure 4.3). The blue boxes in figure 4.3
represent the physical, biological, and chemical processes that affect or are affected by nutrient
and sediment loads.

FIGURE 4.3.—Estuarine physical system response to CBP implementation policy.

Figure 4.4 shows N and P loads for the period 1985–2020. Data on nutrient and sediment loads
to the Bay come from four main sources. Most of the nutrient and sediment load is monitored
via the RIM stations located on the nine major tributaries in the watershed, and operated via a
partnership between USGS, MD Department of Natural Resources, and VA Department of
Environmental Quality. The flow-weighted annual loads are estimated by USGS using the
monitoring results from these RIM stations. This load, termed “river input”, represents loads
emanating from about 78% of the watershed. In order to estimate total annual loads of N and P
delivered to the estuary from the remaining 22% of the watershed (located below the RIM
stations), the CBP adds measured loads of N and P contributed by WWTPs and computer-
simulated estimates of N and P loads from nonpoint sources located below these RIM stations.
Remaining nutrient load data comes from estimates of the atmospheric deposition of N falling

46
FIGURE 4.4.—Measured and estimated N (panel a) and P (panel b) loads and annual flow by October 1–
September 30 water year (Source: CBP, n.d.-e).
on tidal waters. Annual loads to the estuary vary considerably due to annual variations in
precipitation. Within this variability, declines in N loads to the Bay over the past 40 years have
been relatively modest, based on the source data for figure 4.4. For example, mean annual N

47
loads averaged 368 million lb/yr over the period 1985–94 and fell to 320 million lb/yr for 2011–
20, representing a 15% decline; this same data shows smaller declines for P and sediment. 5

Realized reductions in loads entering the Bay are generally translating into lower nutrient
concentrations throughout the estuary. The majority of estuary monitoring stations have seen
reductions in surface water N and P concentrations when measured over time periods of 10,
20, or more than 30 years (especially since 1985) as shown by the downward pointing blue
triangles in figure 4.5. Based on long term trends, 82% and 79% of stations saw significant
decreases in N and P concentrations, respectively (Murphy et al., 2022).

FIGURE 4.5.—Mean change in surface TN (a-c) and TP (d-f) over three time periods at tidal monitoring stations,
calculated using Generalized Additive Modeling approach (Source: Murphy et al., 2022).

5 The 15% decline in N loads between the time periods 1985–1994 and 2011–2020 calculated from this data

source represents the best estimate of what the Bay “sees” in terms of nutrient loads. In contrast, simulated loads
from CAST and jurisdiction-reported data on wastewater discharges suggest a 30% reduction in N loads (370 to
258.9 million lb/yr, 1985–2020) (see figure 3.2). The difference between estimated and realized loads to the Bay
can be attributed to possible lag times and nonpoint source response gaps (as discussed in chapter 3).

48
Despite the complexities of developing accurate estimates of nutrient and sediment loads,
spatial variation in TN and TP concentration trends in tidal waters generally reflect the
assumption “where nutrient loads decline, estuarine concentrations will decline.” To test this,
an analysis by Testa et al. (2018) compared trends of both computed average annual loads
(kg/yr) and estuarine concentrations (mg/L) between two periods (1989–91 and 2012–14)
selected to represent historic and recent periods with similar average annual flows and also
representative of long-term average conditions (this effectively removed the effects of variable
river flow from the comparison of recent and historical loads). Concentrations of TN and TP in
the estuary were derived from the monitoring results from estuarine water quality monitoring
stations (n=140) and were aggregated into each of the 92 CB segments, and pollutant load
estimates were compiled from the same sources as described for figure 4.4 for each segment.
Points representing a segment’s annual TN or TP concentration trend and annual load trend
were placed in one of four quadrants (figure 4.6) where the bottom left quadrants show results
for segments where both N or P loads and concentrations in the estuary have declined (green
dots), and the upper right quadrant shows results for segments where both loads and
concentrations increased (red dots). Segment loading changes correctly predicted the direction
of change in estuary concentration in 88% (TN) and 66% (TP) of cases (upper right and lower
left quadrants), but the remaining cases show unexpected patterns (gray dots in figure 4.6).
These results illustrate the challenges in directly relating load reductions to decreasing
concentrations.

FIGURE 4.6.—Percent change in estuarine TN and TP loads and concentrations, late 1980s to mid-2010s, where
each dot represents a Bay segment (Source: Testa et al., 2018).

While trends in N and P concentrations show correlations with trends in loads, these do not
necessarily translate into improving trends in constituents represented by WQC. The CBP, MD
Department of Natural Resources, and VA Department of Environmental Quality have sampled
water quality on a bi-monthly or monthly basis at more than 130 stations located throughout
the mainstem of the Chesapeake Bay and the tidal portions of numerous tributaries on the
western and eastern shores since the mid-1980s. These data are used to produce estimates of
conditions (improving/no change/degrading) for nutrients, DO, Secchi depth (a measure of

49
water clarity), Chl a, and water temperature. Scientists evaluate short- and long-term trends
using a Generalized Additive Modeling approach. The most recent report shows that, over the
long term, most stations have seen no change or degrading conditions in DO in the bottom
layer (82% of stations), annual Secchi depth (84% of stations), and spring season surface layer
Chl a (76% of stations), indicating that load reductions are not translating into improved water
quality conditions in the deep waters of the Bay (CBP, n.d.-d). However, there is substantial
spatial variability in the results, as evidenced in the long-term mean DO concentrations
depicted in figure 4.7.

FIGURE 4.7.—Changes in DO in bottom water layer measured during June—September, short term (left pane) and
long term (right panel); starting dates for long-term measurements vary (Source: CBP, n.d.-d).

The question of whether these reduced nutrient concentrations are being manifested in the
attainment of WQS remains. In order to measure progress in meeting the WQS, a multimetric
attainment indicator was developed by Zhang, Murphy, et al. (2018) to estimate combined
standards attainment (DO, Chl a, and water clarity/SAV). Given the spatial variability in water
quality trends, examination of this indicator at finer spatial scales yields additional insights into
the connection between load reduction efforts and water quality response. Figure 4.8 presents
the estimated percent attainment of indicators across different habitats of the Bay over the
period 1985–2018 and shows differential patterns over time. Indicator attainment in the
migratory and spawning habitat (blue line) was closer to 80% from mid-1980s through about

50
2000, but has declined afterwards towards attainment ranging from 60% to 80%. Indicator
attainment in the adjacent open water DU (orange line) reveals a pattern of attainment below
50% before the early 1990s, but with values ranging from 50% to 70% since. Indicator
attainment in the deep water (gray line) ranges from 20% to 40%, while attainment in the deep
channel is lowest (ranging from zero to just over 10%). Increased frequency of non-zero values
for the DO attainment indicator in the deep channel in figure 4.8 is notable, since it may
suggest that nutrient load reduction efforts are beginning to initiate an ecosystem response.
Figure 4.8 also shows that attainment of clarity/SAV in the shallows has slowly improved, with
very few areas attaining before about 2001, but some increasing closer to 10% afterwards.
Finally, Chl a in open water has no discernable pattern and ranges from zero (the majority of
the time) with a few higher values over the period of assessment. It is especially instructive to
compare attainment of DO criteria alone across the habitats; for the 30-year period of 1985–87
to 2014–16, annual changes in attainment of DO criteria are estimated to be -0.66% in
migratory and spawning habitat, 0.00% in the deep channel, 0.10% in deep water, and 0.61% in
open water (Zhang, Murphy, et al., 2018). These patterns suggest that specific habitats might
differ in the timing of their attainment of WQS; for example, the progress of attainment
depicted in figure 4.8 would suggest that the deep channel may be the last to reach attainment
of DO standards.

FIGURE 4.8.—Time series of area-weighted estimated WQC attainment for the five DUs, 1985–87 through 2016–
18 (Source: Zhang, Murphy, et al. [2018] with updated data).

51
The lack of indicator attainment, especially that of DO WQS, across all habitats leads to the
question of whether estuary water quality is responding in ways consistent with expected
response to nutrient and sediment reductions achieved to date. Figure 4.9 illustrates one way
to explore this question, albeit with a few notable caveats. In figure 4.9, the orange circles show
estimates of expected DO standard attainment across a range of N loads for open water, deep
water, and deep channel regions and represent expected responses of DO WQS attainment to
load reductions provided by the CBP estuary model.6 Load estimates were generated for both N
and P, but figure 4.9 shows only the N values for simplicity. The observed response (attainment
of DO WQS from monitoring results at a given estimated load) is represented by the blue
diamonds and yellow squares in figure 4.9. The blue diamonds represent 3-year running mean
of percent attainment of DO standards within each habitat during the period 1985–2020 (y
value) and the corresponding 3-year running mean of annual watershed N load (x value) as
obtained from the data sources referenced in figure 4.4 and described above. Yellow squares
represent the same data in 10-year averages.

FIGURE 4.9.—Expected and realized relationships between TN loads and DO criteria attainment for open water,
deep water, and deep channel habitat, calculated as 3-year running mean observed values (blue
diamonds) and expected responses from estuary model (orange dots) for the same time periods. Yellow
squares are 10-year means of the observed data.

The data in figure 4.9 suggest that the observed levels of DO standards attainment at any given
N load are generally below the expected response (blue and yellow points are below the orange
points), a potential estuary response gap. Three patterns are significant. First, a potential
response gap exists in each of the three habitats (open water, deep water, and deep channel).

6These estimates were generated during preparation of the Phase III WIP planning targets (CBP Principals Staff

Committee, December 19-20, 2017). They represent loads resulting from each of the 12 management planning
scenarios and model year assumptions used in the TMDL development (USEPA, 2010, Section 6) but updated using
the Phase 6 model. Points representing Model and Observed Data use somewhat different assumptions including:
(1) Model points represent the estimated attainment (y-axis) for the hydrology period of 1993–1995, while the
corresponding estimated load (x-axis) are for the hydrology period of 1991–2000; (2) Observed points represent
the estimated attainment for the actual hydrology of each 3-year period while the Model is always the 1991–2000
hydrology (that is, the Observed is showing the response to management and weather, while the Model is just
management response in wet years); (3) Temperature is rising over the Observed period, but not for the Model;
and (4) Phosphorus load assumptions are not the same for Observed and Model.

52
Second, the response gap between expected (orange dots) and observed attainment for 3-year
averages (blue diamonds) increases in size at lower N load levels. Third, the response gap at
lower N load levels is significantly larger for the deep water and deep channel habitats as
compared to the open water habitat, with the deep channel habitat exhibiting the largest
response gap and percent attainment remaining below 20%. This suggests responses in DO
attainment will vary by habitat as loads are reduced and need to be better understood. The lack
of progress towards attainment in the deep channel suggests that management assumptions
regarding the magnitude, speed, and timing of this response should be reconsidered.

The extent to which attainment of the DO WQC in the deep channel is lagging behind
attainment in the open water and deep water can be seen in figure 4.10, which plots the
attainment deficit (Zhang, Tango, et al., 2018) for these three habitats over three-year
assessment periods from 1985 to 2021. While figures 4.7 and 4.8 provide a binary classification
of attainment (full attainment or nonattaining), figure 4.10 provides additional understanding
of actual conditions since it quantifies the nonattainment of DO standards over both space and
time for all tidal segments. Among the three habitats, only open water showed a statistically
significant long-term decrease in nonattainment (0.04% annually). In the deep water and deep
channel habitats, neither the long-term nor short-term trends were statistically significant. In
addition, the relationships between attainment of WQS in one habitat and its influence on the
attainment of WQS in another are poorly understood. For example, water clarity in the shallows
may allow for a benthic algae response that contributes to attainment of DO standards in the
deep channel. Such a response could imply that attainment of WQS in the shallow habitats may
accelerate response in other habitats including the deep channel.

FIGURE 4.10.—Attainment deficit (0%=full attainment, -100%=full non-attainment) for DO WQC in three DU
habitats, 1985–2020, calculated as 3-year running means. Year shown represents the first year of a 3-
year assessment period (Source: Q. Zhang, personal communication).

53
Pronounced water quality improvements associated with nutrient reductions have occurred in
localized regions of the Bay, however. For example, clear signs of substantial water quality
improvement in some Chesapeake Bay regions have been associated with upgrades to WWTPs
in specific portions of the Back River, Potomac River, James River, Patuxent River, and Patapsco
River basins that have provided measurable and substantial reductions in nutrient loads,
ambient nutrient concentrations, and algal biomass (Boynton et al., 2014; Fisher et al., 2021;
Testa et al., 2022). The effects of these point source improvements were observed in waters
local to the WWTP facilities, which are generally sited in brackish and tidal freshwater regions
of tributaries. The identification of a response gap is much harder for water clarity and SAV
because of the lack of a modeled expected response. The SAV goals were based on the
historical presence of SAV for most areas. Baywide SAV trends for the tidal fresh, mesohaline,
and polyhaline salinity zones are presented in figure 4.11; long-term decreases in nutrient loads
have led to a significant increase in SAV, despite a high amount of interannual variability in
loads and SAV coverage (Lefcheck et al., 2018; VIMS, n.d.-c). Most notably, SAV has shown clear
recoveries in tidal fresh (salinity levels <0.5 parts per thousand [ppt]) and oligohaline (0.5 to 5.0
ppt) regions of the Bay (e.g., Susquehanna Flats) over the last several decades (Gurbisz & Kemp,
2014; Gurbisz et al., 2017) and more recently in mesohaline (5.0 to 18.0 ppt) regions of the Bay.
The intense precipitation of 2018–2019, combined with declines in eelgrass, interrupted this
positive trajectory in the polyhaline (18.0 to 30.0 ppt) Bay regions that are experiencing limited
light availability and warming water temperatures (Lefcheck et al., 2017).

The variability in SAV cover in different salinity zones over the past several years (figure 4.11)
illustrates the difficulty in trying to measure progress on shorter time scales. The SAV declines
observed in 2019 were followed by gains in 2020–2021. The recent gains have not offset the
2019 loss, so there is not yet evidence that SAV has returned to a positive trajectory.
Nevertheless, current SAV coverage remains well below the WQC goal of 185,000 acres.
According to the 2021 SAV Report (Patrick et al., 2022), 27,528 hectares (68,025 acres) of
underwater grasses were mapped in the Chesapeake Bay in 2021, representing 52% of the CBP
2025 restoration target of 130,000 acres and 37% of the 185,000-acre goal. While this does not
represent a quantitative gap between an expected and realized response (i.e., the response gap
as defined previously), it does represent a significant gap between current and desired SAV
levels.

In summary, response gaps appear to be likely for DO in the open water, deep water, and deep
channel habitats, with the largest response gap in the deep channel, and may be inhibiting
progress toward achieving WQS. Quantification of a response gap for water clarity/SAV is not
possible because of the absence of a formal predictive model, but progress remains well below
the stated goal. Furthermore, while clear examples of TMDL-associated nutrient load
reductions across the watershed have led to well-documented and substantial declines in
estuarine nutrient availability and improvement in some measures of habitat (SAV) and water
quality (DO), many tidal waters throughout the Bay show limited DO and water clarity response

54
despite substantial reductions in regional and local stressors. The following section explores
these response challenges.

FIGURE 4.11.—Seagrass area (hectares) by salinity zone and in coastal bays from 1971 through 2021. Amount of
seagrass abundance in tidal fresh (top panel), moderately salty (mesohaline—middle panel), and full
salinity (polyhaline—bottom panel) areas. Density of cover is indicated by degree of shading: very
dense (70–80% cover), dense (40–70% cover), sparse (10–40% cover), and very sparse (<10% cover)
(Source VIMS, n.d.-b).

55
4.3. Response gaps and uncertainties in achieving WQS

The lack of sustained progress toward achieving WQS across all habitats strongly suggests that
estuary water quality is not responding to load reductions achieved to date in ways consistent
with current assumptions. This led STAC to examine why these gaps exist and what learning
may help us close the gaps (i.e., improve the effectiveness of water quality actions). Response
gaps, identified quantitatively for DO and qualitatively for SAV/water clarity, may result from
limited ability to accurately assess the degree of attainment due to gaps in monitoring,
problems estimating or measuring loads, or failure to account for other factors that may be
confounding the desired response (including complex estuary processes or the influence of
climate change). Each of these is addressed in the following sections, and parts of figure 4.3 are
reproduced to highlight where in the estuarine system each challenge may be manifested.

Confounding factors and nonlinear interactions

Uncertainties around underlying

estuary processes may also be
partly responsible for water
quality response gaps. Processes
presented in figures 4.2 and 4.3
can operate differentially between
habitats and within specific
locations of each. The variability in
attainment of WQC shown in
figure 4.8 suggests that the
relationship between nutrient and
sediment load reductions and
achievement of WQS is unique to
each habitat.

Nonlinear interactions may explain response gaps, especially when internal processes in the
Bay create patterns where water quality improvements are slowed. Interactions between
nutrients and a host of other factors, such as light availability, algae, turbidity, DO
concentrations, redox conditions, and ecosystem engineers like oysters and SAV habitats (as
shown in figure 4.3) may slow or speed up responses to nutrient reductions (Cerco & Noel,
2007; Kemp et al., 2005; Newell, 1988; Testa & Kemp, 2012). For example, the depletion of DO
leads to changes in chemistry that slow nutrient removal and speed up nutrient recycling,
essentially increasing the number of times that any N or P molecule can be used to sustain
eutrophication. In these cases, oxygen depletion supports a self-sustaining cycle; Testa and
Kemp (2012) and Ni et al. (2020) associated this cycling with a potential doubling of the volume
of hypoxic water generated per unit of TN load in the past half century. Feedback mechanisms
such as this may help to explain particular spatial and temporal patterns, even during years of
reduced N loading.

56
In addition, the abundance and condition of living resources (e.g., oysters) and habitats in the
terrestrial-aquatic interface (e.g., tidal marshes) add additional controls and feedback
mechanisms to both water clarity and nutrient cycling processes. The absorption or removal of
nutrients from tidal waters by both oyster communities and tidal marshes has been
documented (Cerco & Noel, 2007; Parker & Bricker, 2020). Losses of these elements from the
Bay ecosystem over time would make the Bay more sensitive to a given nutrient load from the
watershed. Furthermore, recovery of biological communities within the ecosystem that
modulate the timing and amount of available nutrients and organic carbon will affect nutrient
concentrations (Basu et al., 2022).

A significant consideration in conceptualizing response curves by habitat is that feedback

mechanisms illustrated in figure 4.2 can result in ecological tipping points (or thresholds),
generally defined as ecosystem states where small changes in environmental conditions result
in large or rapid shifts in ecological status or function. When such tipping points exist, the
relationship between nutrient and sediment reductions and attainment of WQS will be
nonlinear and may create wide intervals of uncertainty around expected response. Tipping
points are a reality at the scale of localized regions of the Bay and have been demonstrated in
water clarity, DO, and SAV relationships (Boynton et al., 2009; Ganju et al., 2020; Gruber &
Kemp, 2010; Kemp et al., 1990; Testa & Kemp, 2012). For a specific example, see the case study
of Mattawoman Creek in text box 4.1, where a multiyear lag existed between nutrient load
reduction and ecosystem recovery, but once the recovery began, Mattawoman Creek
proceeded rapidly from a degraded algae-dominated system to a restored, SAV-dominated
system.

Tipping points can indicate progressive conditions on either a degradation or restoration

trajectory (for an excellent summary, see Kemp et al. [2005]), and factors that can contribute to
the type of nonlinear response seen in the recovery in Mattawoman Creek are shown in figure
4.12. The left side of figure 4.12 illustrates how positive feedback interactions tend to reinforce
and accelerate the eutrophication process (i.e., explain response gaps), while the right side
shows how they can reinforce the restoration process by enhancing water quality
improvements once they are initiated. Exceeding critical thresholds in the direction of
restoration is of particular interest now. A recent analysis (Lefcheck et al., 2018) linked nutrient
reductions to the resurgence of SAV in the Chesapeake Bay, providing a lesson on staying the
course, allowing time for benthic communities to recover and cross tipping points that enable
self-sustaining processes to initiate. We note that figure 4.12 includes only one impact of
climate change, namely sea level rise. It does not account for water temperature increases or
hydrologic changes from either climate change or land cover shifts; it also does not include
oyster biology.

An implication of all these uncertainties and challenges is that closing the response gaps
identified in this report may not be feasible in all of the habitats and segments of the estuary.
For example, the very low DO levels in the deep channel may be a historic feature, and DO

57
expectations for this habitat may not align with the dynamics of the rest of the Bay ecosystem.
Changing conditions for SAV, including temperature shifts discussed below, result in species
shifts due to habitat suitability ranges and may make expectations of SAV cover that are based
on historical conditions no longer attainable.

Text box 4.1. Shallow waters and tipping points at work: Mattawoman Creek

Mattawoman Creek, a small tributary branching off the upper Potomac River, has shown a dramatic
restoration in recent decades. Nutrient levels were historically high in this shallow tributary due to
high nutrient inputs from the adjacent Potomac River and small WWTPs within the Mattawoman
Creek watershed. As a result, water clarity was poor (Secchi depth < 0.6 meters), SAV was absent, and
the water was filled with an abundance of microscopic algae. In the mid- to late-1990s, nitrogen
reductions began in earnest, and an extended drought period in 1999–2002 contributed to drops in N
loads. This extended period of reduced nutrient loads produced a decline in algal biomass and a
correlated increase in water clarity. The increase in water clarity supported the resurgence of SAV,
assisted by the presence of an invasive
exotic species (Hydrilla) which can take
advantage of short-term periods of
water clarity for establishment
(Boynton et al., 2014).

The improvements in the ecosystem of

this shallow creek following sustained
nutrient reductions reveal how rapid
recovery can occur in regions of
Chesapeake Bay where modest
improvements in water clarity support
SAV expansion, while large-scale
recovery may lag several years Long-term patterns of chlorophyll a, Secchi depth, and SAV
following nutrient load reductions. This coverage in Mattawoman Creek, 1985–2010. (Source: Boynton
case study is an example of a water et al., 2014).
clarity feedback as expressed in figure
4.12 (Kemp et al., 2005). Comparable recoveries occurred in other regions of Chesapeake Bay,
including the enormous Susquehanna Flats SAV bed.

58
FIGURE 4.12.—Effects of N and P additions on physical, chemical, and biological elements of the estuarine system,
including algal biomass, bottom water oxygen, and nutrient recycling. Effects of climate change (save
sea level rise) and land cover change are not included (Source: Kemp et al., 2005).

59
Climate change

Climate change is a global challenge for ecosystem

restoration with ever-present effects in the
Chesapeake Bay. Climate change is no longer a future
threat: the Bay has already warmed, has higher sea
levels, and is receiving altered patterns and
magnitudes of precipitation (Batiuk et al., 2023;
Fleming et al., 2020; Hinson et al., 2022; Najjar et al,
2010). Climate change effects interact with nutrient
management challenges to further test our
understanding of system response.

Dissolved oxygen concentrations are highly sensitive to climatic processes both directly and
indirectly, and climate change effects are likely to make it increasingly difficult to meet DO
criteria. For example, changes in wind direction, precipitation, and water temperature may
result in climatically forced variability in stratification and the associated supply of DO to
bottom waters (Du et al., 2018; Hinson et al., 2022; Scully, 2010); none of these factors are
explicitly recognized in figure 4.9. Additionally, it appears that warming over the past 35 years
has limited the otherwise positive oxygen response to the TMDL (Frankel et al., 2022; Ni et al.,
2020). A study by Frankel at al. (2022) quantified the impact of watershed N reductions on Bay
hypoxia during a recent period including both average discharge and extremely wet years
(2016–2019) (figure 4.13) and concluded that if the N reductions since 1985 had not occurred,
annual hypoxic regions (i.e., areas where O2 < 3 mg/L) would have been much greater. The
discrepancy between our expected responses and current conditions may be largely accounted

FIGURE 4.13.—Estimated extent of Chesapeake Bay hypoxia with and without 35 years of nutrient reductions
(Source: Frankel et al., 2022).

for by warming in the Bay that has offset roughly 6–34% of the improvement from N reductions
(Frankel et al., 2022). The effects of climate change on the Bay will be felt far beyond oxygen
levels and will include shoreline erosion, alteration of productivity, disease impacts, species
migrations, among many other effects (Batiuk et al., 2023).

60
Additional factors

Climate change, nonlinear actions, and other confounding factors will continue to challenge
predictions of what water quality conditions will be in the future. Furthermore, some habitats
are more sensitive and vulnerable to these changes than others. For example, shallow water
habitats may be particularly sensitive to changes in warming. Shallow water habitats may also
be more capable of achieving nonlinear responses (tipping points) from management actions.
Given the need for enhanced understanding of the impacts of both nonlinear interactions and
climate change and the importance of these habitats to living resources, nearshore shallow
water habitats may be particularly effective test beds for addressing uncertainties and
improving water quality response (text box 4.2).

Ability to estimate or measure loads delivered to Chesapeake Bay

The exercise of assessing response gaps, such as those suggested

by figure 4.9, highlights the importance of accurately measuring
nutrient and sediment loads. Without accurate measures of loads,
prediction of responses is more difficult. The size and complexity
of the Bay make accurate determination of loads a challenge:
monitoring does not allow all loads to be measured directly.
Several pieces of information are used to estimate total loads to
the estuary, including the RIM results, discharges from WWTP,
and estimates of non-point source loads. Estimates of the loads
originating from the watershed above the RIM stations are judged
to be fairly accurate. RIM loads are estimated to account for about 60% of the total load of
nutrients and sediments to the estuary. Additionally, discharge data are available from WWTP
point sources below the RIM stations. From these two information sources, it is estimated that
84% of the load is observed through measurement of flow and effluent concentration. The
remaining nonpoint sources of nutrients and sediments (occurring below the RIM stations) are
provided using the CBP watershed model. In addition, nutrients and sediment may be
transported from coastal terrestrial environments into adjacent aquatic environments during
tidal inundation; a recent modeling effort estimated that the amount of dissolved inorganic N
contributed during a perigean spring tide (seasonally high tide or king tide) event in one Bay
segment exceeded its annual N load allocation (as specified in the appropriate WIP) by 30%
(Macías-Tapia et al., 2021).

61
Text box 4.2. Nearshore shallow water habitats as test beds for restoration and addressing
uncertainty

One habitat deserves particular attention because of its positive response to stressor reductions, its
potentially outsized role in Bay restoration, and its relatively high attainment of applicable WQC
(Figure 4.8). This habitat encompasses the land-sea interface, including the nearshore regions of the
watershed, tidal marshes, and shallow, nearshore habitats in tidal waters, including much of the
migratory and spawning habitat; for the purposes of this report we combine these regions under the
umbrella term of nearshore shallow water. This habitat supports extensive expanses of SAV beds,
oyster reefs, and wetlands, and is heavily influenced by an area termed the terrestrial-estuarine
transition zone, or T-zone. The T-zone is defined as “the area of existing and predicted future
interactions among tidal and terrestrial or fluvial processes that result in mosaics of habitat types,
assemblages of plant and animal species, and sets of ecosystem services that are distinct from those
of adjoining estuarine, riverine, or terrestrial ecosystems” (Goals Project, 2015, p. 59). Notably, the T-
zone extends much farther inland than coastal wetland-upland boundaries and includes areas
affected by tidal surges, well inland of tidal freshwater wetlands. On the western shore of the
Chesapeake Bay, the T-zone extends to the fall line, whereas on the eastern shore, the flat
topography imposes limited constraints to the T-zone’s inland migration as sea level rise occurs. On
the western side of the Bay, the T-zone currently extends more than 30 miles inland from the
mouths of the major tributaries.

These transitional zones have recently been recognized as critical but overlooked landscape areas
connecting watersheds and coastal waters because their waterways and adjacent river corridors
represent critical hydrologic connections that concentrate watershed discharge, regulate exchange,
and influence the distributions of living resources (Boomer et al., 2019). The dynamic hydrochemical
gradients created by shifting circulation patterns and changes in flow velocities, along with a cascade
of biogeochemical constraints along the T-zone waterways, create a diversity of habitats that
influence nutrient and sediment delivery to the Bay’s mainstem. Most of these processes primarily
affect the dynamics of P, arguably the most critical element driving algae blooms and eutrophication
in the low salinity regions of estuaries (Carpenter, 2008; Froelich, 1988), and there is clear evidence
for benthic recycling hotspots in eutrophic triblets (Boynton et al., 2009; Testa et al., 2019). While we
have a conceptual understanding of how these hydro-chemical gradients influence biogeochemical
processes and nutrient dynamics, less is known about when and where they occur and the direction
and magnitude of biogeochemical processes within a complex, nested estuarine and triblet system
like the Chesapeake Bay (Statham, 2012; Wang et al., 2016; Ward et al., 2020).

These combined T-zone/shallow tidal habitats can serve as test beds for integrating the land-sea
interface into investigations of tipping points and climate change using monitoring, modeling, and
research approaches. The planned CBP advances in the Phase 7 modeling suite, which include
increases in spatial resolution and process representation in these shallow systems, will facilitate this
new research and management focus. Coupling this emphasis with increased engagement with
stakeholders who spend appreciable amounts of time interacting with these habitats can help to
identify a wider range of restoration outcomes to meet WQS and prioritize living resources,
recreational uses, and habitat restoration.

62
Ability to accurately assess degree of attainment

The sheer amount of monitoring required to measure WQS at all 1,052

assessment points (figure 2.4) presents an enormous challenge. The
traditional CBP partnership’s long-term water quality monitoring and
assessment program was deemed “marginal” for measuring and
reporting on all published criteria to produce a complete accounting of
Bay health (USEPA, 2003b), and efforts since then have included
adjustments to the monitoring network and analyses to estimate
attainment from available data. However, despite involvement of
multiple partners, the monitoring and assessment program has never
been focused at the temporal and spatial scales necessary to provide a
full assessment for ANY of the 92 segments in the Bay (CBP, 2022). Thus,
the program has essentially been unable to effectively monitor and
assess all required WQS, and their responses to CBP implementation, for all habitats associated
with the TMDL (the yellow boxes in figure 4.3).

A multimetric indicator was developed (Hernandez Cordero et al., 2020) to estimate attainment
of WQC using available monitoring information until a more complete accounting could be
supported by more comprehensive monitoring and assessment protocols. However, without
additional monitoring information and associated research efforts, the CBP cannot fully assess
progress towards (1) attaining WQS; (2) understanding the relationship between load reduction
efforts and achieving WQS; and (3) achieving additional water quality conditions to support
crabs, oysters, and other fisheries in the estuary. Additional monitoring efforts would be
needed to also make connections to other habitat conditions across management-relevant
temporal and spatial assessment scales needed by diverse CBP partner interests and GITs (i.e.,
Water Quality Goal Team; Sustainable Fisheries Goal Team; Scientific, Technical Assessment
and Reporting Climate Resiliency Workgroup) (CBP, 2022).

In addition, water quality monitoring is not carried out evenly across habitat types, leading to
differential understanding of the processes controlling responses to nutrient and sediment
reductions. For example, 17 of the 31 outcomes expressed in the 2014 CBWA (CBP, 2014) are
related to shallow water habitats, yet that habitat type is not adequately monitored, modeled,
or studied. Construction and implementation of a new estuarine model with improved
capability in shallow waters, generally defined as the upper 2 meters of open water habitat, is
underway. However, monitoring of shallow waters is challenging. First, there is limited boat or
road access to smaller waterways (termed triblets) draining catchments that typically range
from 50 to 150 square kilometers in size and support extensive expanses of SAV, oyster reefs,
and wetlands. Second, the sheer number of triblets, along with the variability of physiochemical
conditions within each triblet system, presents a challenge for developing a compelling
framework for collecting and interpolating observation data. Recent technological advances
along with a model-based sampling framework could address the lack of data for these
systems.

63
Monitoring protocols are based primarily on status and trends concepts, and thus are more
attuned to accountability objectives surrounding achievement of WQS than to understanding
ecosystem processes that link water quality to load reduction efforts. In other words, current
monitoring may answer the question of whether we have met criteria for a given nutrient or
DO concentration, but if we haven’t met those criteria, the monitoring program does not
provide measurements that would allow a greater understanding of why.

4.4. Conclusions and Implications

Annual loads of nutrients and sediment to the estuary vary considerably due to variations in
precipitation, but declines in N, P and sediment loads over the past 40 years (though modest in
some cases) have generally resulted in lower nutrient concentrations throughout the estuary.
Even without sustained periods of low loads, the majority of estuary monitoring stations have
seen reductions in surface water N and P concentrations over the long term, and spatial
variation in TN and TP concentrations in tidal waters generally reflect “where nutrient loads
decline, estuarine concentrations will decline.” However, these trends in N and P
concentrations do not necessarily translate into improving trends in WQC variables; most
stations have seen no change or degrading conditions in DO in the bottom layer, annual Secchi
depth, and spring-season surface layer Chl a. When attainment of WQS is examined at the scale
of individual habitats, clear differences emerge, with attainment of WQS relatively intransigent
for DO in the deep water and deep channel, suggesting that these habitats may be the last to
reach attainment of their DO standards.

The lack of attainment of WQS across all habitats has implications about the assumptions of
how estuary water quality will respond to the nutrient and sediment reduction targets. Based
on the likely estuary response gaps for DO in the open water, deep water, and deep channel
habitats, the current load reduction targets may not be adequate to attain WQS in all areas of
the estuary. The largest potential response gap is in the deep channel, which implies additional
nutrient reductions may be required to attain any given level of DO response. Further, the CBP
could consider putting additional focus on nutrient and sediment reduction in shallow and open
waters since the likely response gap is smaller in these habitats, and because attainment of
WQS in the shallow habitats may accelerate response in other habitats including the deep
channel. Quantification of a response gap for water clarity/SAV is not possible because of the
absence of a formal predictive model, but progress remains well below the stated goal. Various
assessments of progress toward WQS attainment indicate that complete attainment is not
likely to happen soon.

A critical element in the resolution of these uncertainties is monitoring information. The

monitoring program has never achieved assessment of WQS at the temporal and spatial scales
necessary to provide an assessment of attainment for any of the 92 segments in the Bay.
Ongoing additions and revisions to the monitoring program are aimed toward being able to
assess attainment in selected segments in 2026 (CBP, 2022). Even with these changes, the

64
monitoring program would not provide answers to questions about how the system is working.
Current monitoring protocols are more focused on accountability objectives than on
understanding processes. For example, the existence of tipping points in subsystems of the Bay
is well-documented, but monitoring to determine the thresholds associated with either
degradation or restoration is not currently done.

The CBP could consider enhancing the estuary monitoring network to better document estuary
processes in addition to improved assessment of WQS attainment. The enhanced monitoring,
coupled with associated research, could cover the range of local conditions where proximate
and distant factors can be characterized and causal relationships between stressor reductions
and WQS attainment can be determined. In addition, monitoring protocols would need to be
aligned with modeling approaches. The most effective approach for identifying relationships
between stressor reduction efforts and WQS attainment may be to structure monitoring
around subsegment or smaller spatial scales. At these smaller scales, clear signs of successful
water quality remediation in some Chesapeake regions, such as reductions in nutrient
concentrations and algal biomass, have been associated with upgrades to WWTPs, but other
areas with similar reductions do not demonstrate the same type of response. Similarly, recent
patterns in TN load and concentration reductions have been linked to a resurgence of SAV in
many small-scale regions of the Bay (Lefcheck et al., 2018), but the relationships between
stressor reduction and living resource response are unique to salinity regimes, SAV species
distributions, and water temperature.

The largest source of uncertainty about the relationship between nutrient reductions and WQS
attainment in each habitat may be the existence of tipping points, which have been
demonstrated in localized regions in the Bay. Tipping points exist in water clarity, DO, and SAV
relationships, and can indicate progressive conditions on either a degradation or restoration
trajectory; crossing the implied thresholds in the direction of restoration is of particular interest
now.

The shallow waters present an opportunity for the CBP to reduce critical uncertainties about
the effects of load reductions on living resources in an area of high engagement by stakeholders
who live or recreate in these areas. In these habitats, DO and SAV respond to many variables
beyond nutrient and sediment reductions, and understanding these dynamics is critical for both
identifying effective management actions in the shallow waters themselves and understanding
the relationships between shallow water habitats and the attainment of WQS in other habitats
(e.g., deep water and deep channel DO). The CBP could put more emphasis on the shallow
waters, both in terms of load reductions and scientific understanding, to accelerate and better
understand attainment of WQS and benefits to living resources.

65
5. Living Resource Response to Water Quality Conditions

The support and enhancement of aquatic living resources (the designated use of the
Chesapeake Bay, in the language of the CWA) is the ultimate goal of efforts to reduce nutrient
and sediment loads and achieve WQS in the Bay. As described in the previous chapters, most
regions of the Chesapeake Bay have seen long-term declines in nutrient levels but more mixed
results in achieving long-term improvements in water quality. Chapter 4 described potential
response gaps and deficits in attaining the WQS, particularly DO standards in the deep channel.
An important question at the center of Bay water quality management efforts is: How will Bay
living resources respond (e.g., increased abundance, diversity, and resilience of desired aquatic
organisms) to efforts to improve water quality conditions? Answering this question would
enable us to provide insight into the direct policy-relevant question: What water quality
management actions could accelerate improvement in Bay living resources?

The DUs in the WQS are generally stated as supporting “recreationally, commercially and
ecologically important species” in the various Bay habitats (USEPA, 2003a).7 The response of
living resources to water quality conditions is most easily expressed for those species with
relatively simple life cycles and for which WQS can be easily derived based on their most basic
requirements; an example is SAV and its companion WQS of water clarity. For the remaining
species of living resources (e.g., blue crabs, oysters, finfish), the responses are more
complicated.

The influence of multiple factors and stressors makes detecting and isolating living resource
responses to changing water quality conditions challenging. In recognition of this challenge
when establishing the Chesapeake Bay WQS, drafters organized the WQS around the water
quality conditions (e.g., criteria for DO) necessary to support an individual of a species in
specific habitats. It was understood that meeting the specific WQS would, in turn, provide
water quality conditions to support specific living resources, but generally not to generate a
specific living resource response.8 The specific numeric WQC (measures) selected for the Bay
were based on laboratory studies of species tolerances and habitat affinities for regions
throughout the Bay (Monaco et al., 1998; Tango & Batiuk, 2013; USEPA 2003a, 2003b). Bay
WQS can be viewed as necessary but not inclusive of all the conditions needed to support
different groups of species in different habitats.

This chapter begins by briefly identifying some basic conceptual ideas that help structure our
understanding of, and frame public expectations about, how much improvement in living
resources can be expected from the water quality changes targeted by the WQS. Unlike for
pollutant loads and estuary water quality conditions, the CBP does not use a comprehensive

7 Specifically, the DUs aim to support particular finfish (including migratory and nonmigratory species, game fish,

prey species across multiple habitats), crabs, oysters (open water, shallow water, deep water), and underwater bay
grasses (shallow water) and to maintain populations of sediment-dwelling worms and small clams (deep channel).
8 The exception is SAV criteria.

66
analytical framework to systematically evaluate the relationship between water quality
management actions and living resource response. The next section provides a brief description
of the conceptual relationship between living resource response and CBP water quality
management efforts. This is followed by illustrative examples of statistical and ecological
modeling analyses that have examined the Bay living resource response to water quality and
other factors. The chapter concludes by noting analytical options for more fully examining
responses of Chesapeake Bay living resources to water quality, habitat, climate, and other
factors, and other additional management actions.

5.1. Conceptual model of living resource response to water quality conditions

Figure 5.1 shows a conceptual representation of the possible responses of living resource
abundance to attainment of Bay WQS (DO, Chl a, and water clarity). The horizontal axis shows
the percent attainment of multiple WQS across the 5 habitats and all 92 segments in the Bay.
The overwhelming majority of those assessment points involve DO. Furthermore, existing
empirical evidence and water quality models suggest that increases in attainment (moving left
to right on the horizontal axis) will not occur evenly across the estuary and across habitats. As
described in chapter 4, as attainment moves toward 100%, the last remaining nonattainment
areas will likely be the deeper water habitats (deep water and deep channel). The vertical axis
shows one conceptual metric of living resource response, overall abundance. Other living
resource metrics, such as biological diversity, could also be used to illustrate similar conceptual
relationships.

FIGURE 5.1.—Conceptual living resource response to attainment of WQC.

67
Figure 5.1 reflects the general expectation that achievement of WQS will increase living
resource abundance over a certain range. While the CBP does not identify a specific living
resource response from water quality changes (i.e., no expected response as the term is used in
this report), reducing areas of low DO expands habitat, improves organism growth rates and
reproductive success, and can decrease mortality (Breitburg et al., 2018). The pattern of
attainment discussed in chapter 4, in which the percent attainment is currently highest for the
open water habitat and lagging most for the deep channel, suggests that percent attainment of
WQS may proceed across habitats somewhat sequentially; that is, a high level of attainment
will occur sooner for the open water habitat, and attainment in the deep channel habitat will
occur last. Since some regions and habitats of the Bay are used more extensively by aquatic
organisms than others (shallow water habitats, for example, support a wide range of fish
species across key life stages), the rate of improvement in living resources is likely to change as
attainment progresses and nonattainment areas shrink toward the deep-water habitats (in
figure 5.1, moving left to right on the horizontal axis). Of note, living resource abundance may
eventually decrease as nutrient loads continue to decrease and WQS approach full attainment.
For example, lower nutrient loads needed to achieve 100% attainment of DO standards may
restrict primary production at the base of the food chain in the more productive shallow water
areas and, at some point, act to limit or diminish the biomass of some species further up the
food chain (Breitburg, Craig, et al., 2009; Breitburg, Hondorp, et al., 2009).

Considerable uncertainty will accompany any effort to predict how fish and shellfish
populations respond to changes in water quality alone (represented by dashed green lines in
figure 5.1). Estimating the abundance of various species throughout the Bay ecosystem is
challenging and subject to measurement error. Complex biological systems respond to variation
caused by stochastic events like weather as well as a host of other factors besides water
quality, and accounting for all the major sources of variation in living resource response is
challenging.

Figure 5.1 shows two possible responses of living resource abundance to achievement of Bay
WQS. The response may resemble curve H, which shows relatively rapid improvements in living
resources in response to water quality improvement, while response L shows a more limited
living resource response to improving water quality conditions. Whether the living resource
response resembles H or L depends on a number of other biological, physical, and management
factors besides water quality.

Figure 5.2 shows a simplified system diagram of the major factors that influence the abundance
and composition of aquatic living resources. Starting at the left, a number of estuarine water
quality conditions impact the aquatic habitat of living resource communities. The factors in
yellow are explicitly managed under Bay WQS (DO, water clarity, Chl a). Water temperature can
have a large influence on the type and abundance of living resources (Ihde & Townsend, 2017).
Water temperatures are increasing in the Bay, and increases are projected to continue with
climate change (Batiuk et al., 2023) Water temperature increases are most pronounced in
shallow water habitats, and these areas are important to forage fish and many fish life cycles.

68
Other major water quality conditions, such as salinity and pH, can also have a pronounced
influence on the habitats used by living resources.

FIGURE 5.2.—Factors affecting the composition and abundance of living resources.

The water and nearshore structural habitat conditions also directly impact living resources
(middle section of figure 5.2). For example, coastal wetlands benefit a wide range of aquatic
animals.9 Conversely, shoreline hardening is associated with diminished abundance of a
number of important recreational and forage fish species (Kornis et al., 2017). Fabrizio et al.
(2021) found that shallow water habitats are particularly important for a number of key forage
fish species. In addition to habitat, additional factors affect living resources (right side of figure
5.2). One important factor impacting living resource abundance is fisheries harvest. For
important commercial and recreational fish species, harvest is often a primary driver. Ecological
and food-web interactions will also influence living resources.

Species respond differently to water quality and habitat conditions, and tradeoffs between
different species occur as these conditions change. Species, as well as their specific life stages,
exhibit varying tolerances across different water quality conditions (DO, water temperature)
and habitat conditions. For example, habitat modeling studies suggest some species (striped
bass, bay anchovy, oysters) may be much less sensitive to changes in the volume of hypoxic
waters than bottom dwelling Atlantic sturgeon (Schlenger et al., 2022).

9 The CBWA includes numeric wetland restoration goals but not specifically for tidal wetlands.

69
5.2. Evidence and analysis of water quality impact on living resources in Chesapeake Bay

Figure 5.3 shows historical indices of fish abundance for four well-known species in the Bay:
blue crab, striped bass, summer flounder, and bay anchovy. The graphs show the average
weight of fish caught per standardized unit of fishing effort (i.e., trawl) from 1988 through
2016. The data series shows significant annual variation in these species and different trends
across each. For example, the population of striped bass varies annually, and there is also a
long-term trend showing a decline in population abundance. However, there is significant
analytical challenge associated with isolating the causal factors that could explain observed
changes in living resources and to predict the impact of the various factors that influence the
composition and abundance of living resources (Hood et al., 2021; Rose et al., 2023). Extensive
data exist to investigate various causal factors and aquatic organism abundance, and analytical
approaches generally fall into one of three categories: habitat models, statistical models, and
process-based models.10 Examples of each category applied specifically to the Chesapeake Bay
follow.

FIGURE 5.3.—Indices of Chesapeake Bay fish abundance, various species over time (Source: VIMS, n.d.-a).

10 Habitat models may use statistical methods for fitting, but they predict habitat quality (typically in units of area
or volume) as the response variable. We use the term “statistical models” to refer to models that predict response
variables about biological quantities, such as organism abundances, growth rate, average weight per individual, or
survival rate, from their correlations to explanatory variables. Process models are simulation (not statistical)
approaches and predict biological quantities from assumed process rates.

70
Habitat-based assessments estimate the relative suitability of a habitat for a species based on
the observed relationship between the species and environmental conditions. A variety of
habitat-based models have been and can be used to relate living resource response to water
quality conditions. For example, Schlenger et al. (2022) used a habitat volume model to
estimate the sensitivity of 12 species in the Bay (at different life stages) to habitat changes
resulting from changes in water temperature, salinity, and DO. The study differentiated
between required habitat (the region outside of which mortality would occur due to salinity,
temperature, and/or DO levels) and optimal habitat (the region outside of which salinity,
temperature, and/or DO levels would cause physiological stress, leading to decreased growth,
production, or impairments to other metabolic processes). The modeling effort provided
quantitative understanding of the factors that limit habitat and their importance to individual
species/life stages. Salinity and temperature were responsible for the largest changes in
optimal habitat, while the results for DO were mixed across optimal vs. required habitat.
Dissolved oxygen had a major impact on the required habitat for some species/life stages (e.g.,
Atlantic sturgeon) but had a relatively minor influence on optimal habitat across other species.
The impact of DO improvement on habitat may vary with water temperature; a habitat model
for Atlantic sturgeon found that achieving the Bay DO criteria would increase total habitat by
13% for an average weather year, but a 1 oC increase in water temperature would reduce
habitat by 65% (Niklitschek & Secor, 2005).

Recent work illustrates how statistical analyses of monitoring data can be used to provide
evidence linking living resources to water quality. Fabrizio et al. (2021) examined changes in
habitat and the abundance of four common Bay forage species: bay anchovy, juvenile spot,
juvenile weakfish, and juvenile spotted hake. Habitat conditions included distance to shore,
sediment, water temperature, salinity, and DO. Habitat conditions such as water depth, water
current, and temperature were identified as key habitat conditions for all four species, while
DO delineated suitable habitats for juvenile spot in the winter and bay anchovy in the winter
and spring. While the study highlighted the importance of shallow water, nearshore habitats for
these forage fish, it found mixed and limited statistical evidence of a relationship between
specific aspects of habitat conditions and observed seasonal abundance over 17 years (Fabrizio
et al., 2021). Woodland et al. (2021) conducted statistical analyses of the relationships between
abundance of classes of forage species (fish and benthic species) and water quality variables
such as DO, water temperature, and salinity but did not include measures of nearshore habitat
conditions. This study found that forage species abundance was consistently related to water
temperature, particularly the rate of warming in the spring and summer (gradual warming is
associated with increased abundance). More complex and varied responses were found to
water quality conditions associated with WQS. Abundance of different classes of forage species
were both positively and negatively correlated to Chl a levels. The abundance of most forage
groups increased with DO levels in the mainstem section of the Bay, but forage group
abundance showed a more varied response (both positive and negative relationships) to DO in
the Bay tributaries (Woodland et al., 2021).

71
A variety of living resource process-based models have been developed to predict ecological
response to different environmental conditions. Network analysis and Ecopath with Ecosim
models have been developed or explored for the Chesapeake Bay (Baird & Ulanowicz, 1989;
Christensen et al., 2009; Ma et al., 2010; Monaco & Ulanowicz, 1997; Townsend, 2014).
Ecopath with Ecosim is the most-applied trophic ecosystem model in the world and is used in
other locations to assess responses to restoration (Vasslides et al., 2017). These simulations
allow quantitative connections to be made between water quality and commercially and
recreationally important species, and they can be used to assess the benefits of and trade-offs
between water quality restoration goals and fisheries management goals, with some
restrictions (Townsend, 2014). Townsend (2014) describes the limitations and challenges of
coupling Ecopath with Ecosim-based fishery models to existing Bay water quality models to
assess the effect of water quality management on managed fish stocks.

Another process-based model that couples predator-prey relationships with biochemistry

conditions is the Chesapeake Atlantis Model (Fulton et al., 2011). The Chesapeake Atlantis
Model has been used to estimate the higher trophic level impacts of fully achieving the TMDL
requirements for N and sediment under present day climate conditions, as well as under
warmer water temperatures, and simulations were also combined with habitat loss and gain
(restoration) scenarios (Ihde et al., 2016; Ihde & Townsend, 2017). Results suggest water
temperature is a major factor in explaining abundance and that “achievement of the TMDL-
prescribed reductions of N and sediment had only a moderate effect on modeled guilds” (Ihde
& Townsend, 2017, p. 7). Fulford et al. (2010) used a network simulation model to examine the
differential impacts of increased oyster production and decreased nutrient loads on
phytoplankton, zooplankton, and species of forage fishes. The model suggested a 50%
reduction in nutrient loads would increase benthic habitat (reduced hypoxia), decrease
phytoplankton production, but have negligible predicted effects on bay anchovy and benthic
fish biomass.

The degree to which specific aspects of water quality and habitat (e.g., DO, wetlands) are
identified as the causes of detected changes in the living resources varies widely among these
analyses. While many of these can be utilized in specific instances, collating results of
independent analytical efforts is difficult because (1) the species, statistical and modeling
methods, spatial coverage (e.g., regions of the Bay), and temporal coverage (e.g., which years)
vary greatly across analyses; (2) existing analyses addressed study-specific questions and
hypotheses but application of results to assess specific restoration actions have been mostly, at
best, suggestive (i.e., speculative); and (3) in situ analyses specific to water quality and habitat
rarely are specifically designed to advance the understanding of the role of the TMDL and other
CBWA-related restoration actions in affecting the living resources response. To date, there has
not been a comprehensive examination of living resources responses in situ that also attempts
to relate the responses to CBP actions.

72
5.3. Conclusions and future directions

Research shows that the timing, location, and extent of changes in DO conditions influence
both the species which are impacted and the magnitudes of the impacts. Direct evidence of the
impact of water quality changes on various classes of living resources varies and is frequently
unclear, partly because of the confounding multiple changes and their effects across complex
ecological interactions, and partly because there have not been substantial system-wide
changes in some criteria (e.g., DO).

Future conditions in the Bay will be different from historical conditions. The Bay is warming,
and precipitation patterns appear to be changing (Hinson et al., 2022; Najjar et al, 2010;
Fleming et al., 2020). Research consistently reports that water temperature (Batiuk et al., 2023)
and salinity will have major influences on living resource communities, independent of water
quality conditions such as DO. Thus, the living resource outcomes that can be expected from
incremental achievement of WQS are conditional on these other factors. Research also
highlights key roles that other factors, many which can be directly managed, have on explaining
abundance, composition, and distribution of aquatic species. For example, research points to
the importance of habitat (particularly shallow water) quality and nearshore conditions for
living resources.

The CBP has devoted substantial resources to the development of analytical and statistical
modeling of pollutant and water quality outcomes. By comparison, quantitative assessment of
living resource responses to water quality and other factors has received less attention outside
of reporting of annual indicators of population health. Ecosystem-level models have been
developed for the Bay system but have not been used to their fullest potential to assess how
different water quality management actions might impact living resources (Hood et al., 2021;
Rose et al., 2023).

While numerous scientific studies have been conducted, particularly for individual species, they
collectively offer limited help with analyzing how different classes of living resources and their
interactions might respond to future and alternative water quality and nearshore conditions
(Rose et al., 2023). Specifically, additional analytical and modeling capacity that focuses on
living resource responses could shed light on questions such as:
● What is the expected (projected) response of living resources to the level of attainment
of WQS and habitat conditions in the Bay: (a) without the TMDL and habitat and
sustainable fisheries targets, (b) if present conditions of attainment and partial
achievement of the habitat and fisheries goals continue, and (c) if water quality targets
under the TMDL and/or the vital habitat and sustainable fisheries targets are fully
achieved?
● What combination of conditions, in addition to and including achievement of existing
WQS, would need to occur to achieve desired responses in living resources?
● What are the decision-relevant uncertainties of existing analyses, and how can they help
guide future monitoring and modeling efforts? This information could be used to inform

73
 natural resource managers about future research needs and their decisions about
quantitatively defensible indicators and measures for tracking living resource responses.

Rose et al. (2023) described a modeling framework that could help address these questions for
the Chesapeake Bay. This approach differs from the present approach by using predicted in situ
responses of populations and food webs, rather than tolerances and preferences of individual
organisms, to inform water quality and habitat targets. To date, restoration progress has been
determined by achievement of the WQS (which were informed by living resource information)
and status and trend indicators as part of the CBWA (CBP, n.d.-e). Without attribution to
specific causative factors, documented changes in living resources (positive and negative
trends) become difficult to interpret as reflective of restoration efforts or not. The approach
described by Rose et al. (2023) could enable broader conclusions and more refined statements
about the role of the TMDL and other restoration actions in the CBWA. The analyses would also
provide information on realistic expectations to managers and the public about species
responses to changes in water quality and habitat. Grounding expectations is especially
important when the responding species (like many in the Chesapeake Bay) are long-lived, have
complex life cycles, and are affected by multiple factors. Further analyses would also be useful
as a guide for interpreting living resource responses to the TMDL and habitat actions, designing
future indicators, informing adaptive management, and possibly refining restoration goals.

74
6. Findings and Implications

Progress toward achieving Bay pollutant reductions and WQS has been slower and more
challenging than expected. The question is: Why? Policy and management actions are intended
to reduce nutrient and sediment pollution to improve Bay water quality conditions and
ultimately enhance Bay living resources. Key objectives of this report were (1) to identify gaps
and uncertainties in the causal chain of responses that link policy and management actions to
these goals, and (2) to offer approaches to address these gaps and uncertainties going forward.

This chapter briefly summarizes the findings from chapters 2–5 and offers options for how
progress can be accelerated. Efforts have fallen short of achieving TMDL nutrient targets. In
many cases, implementation and response gaps confound efforts to reduce nonpoint sources.
Pollutant load reductions , if successful, still take time to work through soils, aquifers, and
stream channels of a large watershed, leading to time lags between implementation and
observed reductions in loads reaching tidewater. If and when loads reaching the Bay are
reduced, the Bay water quality response indicated by standard metrics (DO, water clarity/SAV,
and Chl a) is mixed. For example, nutrients have shown signs of improvement, but water quality
responses have been less than expected. The fact that water quality conditions have not
deteriorated given significant economic and population growth, land use change, and a
changing climate in the past three decades is a notable accomplishment.

The remainder of this chapter offers suggestions on what can be done to improve system
response. What options and opportunities exist for water quality management efforts to get
more nutrient reductions from implementation actions (particularly nonpoint sources), and
what actions might improve the response of living resources to water quality improvement
efforts? STAC does not offer prescriptive solutions, but rather describes choices facing CBP
partners and identifies promising options for improving system response to management
efforts. STAC offers multiple options to improve the effectiveness of nonpoint source programs,
prioritize water quality and TMDL management efforts that have the greatest potential to boost
fish and shellfish habitat and populations, and enhance decision-making about these options
when outcomes are uncertain.

The findings and implications are offered based on three themes:

First, achieving load reductions and water quality improvements is proving more challenging
than expected. To date, efforts to reduce nonpoint sources have not produced sufficient levels
of BMP implementation to meet the TMDL, and the implementation that has occurred may not
be producing the pollutant reductions expected. Evidence indicates that the nutrient and
sediment load reductions realized to date have led to improved water quality conditions in
some areas of the Bay, especially in areas with substantial localized reductions, but other areas
have mixed results.

75
Second, the Chesapeake Bay system observed in the past will not be the same system in the
future. Future conditions will be influenced by permanent and ongoing changes in land use,
climate change, population growth, and economic development. Recognizing these changing
conditions will challenge notions of restoration that are based on recreating historical
conditions.

Third, new approaches to water quality management will be required to adequately address
uncertainty, changing future conditions, and response gaps. Opportunities exist to improve
pollutant management, water quality, and living resource outcomes, but many require changes
and new approaches to implementation, planning, and decision-making. Given uncertainty,
change requires a willingness to accept the risk that some efforts will fail and a desire to
enhance capacities to learn from successes and failures. Nothing in this report should be
interpreted as suggesting “backsliding” or retreating from commitments to improve water
quality in the estuary. However, water quality policy will need to evolve over time based on an
understanding of what future conditions are possible, what local communities and the
partnership at-large see as priorities, and what is required to attain those possible futures.

6.1. Summary of findings

Four major findings emerge from chapters 2–5 and are summarized as follows.

The slow rate of water quality change in the estuary suggests that achievement of WQS in the
Bay is uncertain and remains in the distant future. To date, attainment of the Bay WQS stands
at about 30%. Evidence indicates that load reductions achieved to date have led to improved
conditions in some portions of the estuary, but overall results are mixed. Evidence also
indicates that the nutrient reductions achieved to date have not produced the magnitude of DO
response expected. For example, deep channel and deep water habitats of the Bay exhibit
significant potential response gaps for DO, while SAV coverage in shallow and open water
habitats falls short of stated goals. Multiple possible reasons exist for limited progress in
achieving the WQS, including climate change and insufficient reductions to stimulate an
accelerated response (tipping point).

Existing efforts to reduce nonpoint sources of nutrients are likely insufficient to achieve the
TMDL. The TMDL and its nutrient and sediment reduction targets are the primary policy drivers
to achieve Bay WQS. Progress has been made in reducing nutrient loads, particularly in relation
to point source discharges and atmospheric N deposition. However, the CBP jurisdictions have
identified in their WIPs that additional load reductions will be focused on nonpoint sources in
agricultural and urban areas. These pollutant reduction efforts must induce sufficient and
effective behavioral change from hundreds of thousands of people and land managers within
the watershed whose behaviors contribute to nonpoint source loads from agriculture (the
largest nonpoint source) and developed land (the fastest-growing nonpoint source). Making
substantial progress toward meeting the TMDL will require an increased rate of agricultural and
urban nonpoint source load reductions relative to what has been achieved in the past.

76
The pathway to achieving TMDL nonpoint reduction targets is unclear, but options are
available. The CBP acknowledges the challenges of generating enough nonpoint source BMP
adoption to meet nutrient reduction goals, particularly for N. This implementation gap, the
difference between control practices implemented and practices needed, is widely
acknowledged for N. In addition, statistical analyses of ambient monitoring data indicate that
nonpoint source management actions (BMPs) are not generating the amount of reductions
expected (response gap) for N and P. The P response gap is especially large and has many
potential causes, most of which are associated with regions of the watershed with significant
mass P imbalances (which, in turn, are associated with intensive livestock and human
populations).

Additional funding alone is unlikely to achieve desired nutrient and sediment reductions.
Achieving and sustaining substantial future pollutant reductions will require a willingness to
develop and adopt new implementation approaches and technologies. The basic structure of
the CBP’s implementation programs has been in place for several decades: numeric effluent
limitations for permitted sources, voluntary practice-based cost-share programs for agriculture
and unregulated urban sources, and a TMDL accounting system that tracks success through a
central modeling system. This report describes the accumulated evidence of implementation
and response gaps that limit our ability to secure reductions in nonpoint source loads within
the current structure. Making substantial progress in reducing nonpoint source discharges will
require changes in program structures, incentives, and requirements.

Improving water quality to meet the Bay WQS may not be sufficient to generate desired
changes in the composition and abundance of Bay living resources. Water quality criteria were
chosen to support and enhance Bay aquatic living resources. The question of what type of living
resource response can be expected from water quality improvements is complex and uncertain.
Compared to water quality assessment, the CBP has devoted much less effort and resources to
assessing the impacts of water quality on living resources in different habitats (open water,
shallow water, deep water, etc.). However, a major advantage with the Chesapeake Bay is that
much of the data and information (e.g., physical and biological information) needed to expand
this capacity are available, well-established, and vetted. Expectations for how living resources
will respond to water quality changes should be conditioned upon the range of additional
factors that also impact their composition and abundance, including climate change, land use
and economic change, harvest, human population growth, and natural variations outside the
influence of management. For example, not only does Bay warming make attainment of DO
goals more difficult, warming water temperatures will affect habitat and aquatic species mix
and ecosystem interactions. Changes in nearshore habitat and land use will also affect fish
composition and abundance in a myriad of ways beyond just altering nutrient and sediment
loads.

Current CBP adaptive decision-making processes have limited capacity to effectively address
the full range of questions required for effective water quality policy. Effective water quality

77
policy answers the following questions: (1) what is; (2) what if; and (3) what should be. The CBP
has built a sophisticated adaptive decision-making process focused on TMDL implementation,
the SRS, and an accounting process premised on the use of predictive, deterministic planning
models. The current SRS process, TMDL accountability framework, monitoring program, and
modeling are aimed toward answering questions 1 and 2. The existing analytical models and
planning processes have limited capacities to systematically address the critical uncertainties
and causes of the response gaps for these questions. Current adaptive management processes
have limited decision-making scope to address question 3.

6.2. Implications: Policy options for improving effectiveness of water quality management
The approaching 2025 federal TMDL deadline offers an opportunity to reassess and enhance
the potential of policy to improve water quality and living resource response to pollution
control efforts in the face of uncertainty. The remainder of this section provides three
overarching areas that could be considered to improve water quality management
effectiveness.

Refocusing water quality management efforts on improving living resource response. Current
WQS focus on achieving water quality conditions (DO and water clarity) that broadly support
living resources. This focus does not necessarily mean that progress toward achieving WQS will
directly translate into enhancing living resource populations and responses. Options are
discussed to address the question: Can Bay water quality goals and the way attainment is
measured be revised to increase attention to, and potential for, other water quality
investments to improve living resources?

Improving effectiveness of nonpoint source management. Options are discussed to address

the question: What policy and implementation options offer potential to deliver substantial and
sustained reductions in nonpoint source loads?

Enhancing adaptive management to improve the CBP’s ability to learn and respond to
uncertainties and response gaps. Given the knowledge, uncertainties, and complexities
described throughout this report associated with assessing and responding to response gaps,
options are discussed that address the question: What processes and analytical approaches are
available to improve learning, especially as it pertains to the first two areas?

The practical necessity of TMDL implementation confines management efforts to making

decisions and choices within established accounting rules, operational programs, and well-
defined programmatic goals (in figure 6.1, see feedback arrow 5). The 2025 TMDL deadline,
however, is an opportunity to consider a broader set of options to improve water quality policy
(figure 6.1, feedback arrows 1 through 4) This chapter considers options for designing and
implementing new pollutant control approaches and TMDL policies (figure 6.1, feedback arrow
4). Changes may also include policy options for how water quality goals could be amended (e.g.,
changing criteria, assessments, etc.) to increase the potential for enhancing abundance,
diversity, and resilience of desired aquatic species (arrows 3, 2, and 1 in figure 6.1). Discussions

78
of options and alternatives emphasize the need to consider options beyond existing programs
and goals, but ones that are still possible within the legal and programmatic constraints of the
CWA.

FIGURE 6.1.—Levels of policy feedback and learning in expansion of adaptive management.

79
6.3. Water quality goals and living resource response

The objective of the WQS and associated TMDL is to improve Bay living resources. Water
quality standards are largely expressed in terms of achieving key water properties (DO and
water clarity) to support living resources, but not directly in terms of enhancing species
populations. The approach of the WQS means management attention is directed to reducing
pollutants to achieve the measurable criteria (e.g., DO). The policy challenge that needs to be
considered is this: What can be done to improve living resources from our water quality
management efforts given limited public resources and budget constraints?

Figure 6.2 shows the conceptual cost and benefit tradeoffs associated with different levels of
attainment of Bay WQS. Panel A of figure 6.2 shows a conceptual representation of costs to
attain the 1,052 combinations of WQC (DO, water clarity, and Chl a), estuary segments (92
spatial segments), and habitats (up to 5 different habitats per segment). The cost to bring a
greater portion of the Bay into attainment with the WQS is expected to increase at an
increasing rate for a couple of reasons. First, more pollutant reductions will be necessary to
bring additional segments and habitats of the Bay into attainment. For example, an additional
20 million lb/yr of N reduction may bring many open water and shallow water habitats into
compliance, but another 20 million lb/yr reduction may bring only a few additional, generally
deep water, habitats into attainment (attainment of DO criteria in deep water
segments/habitats is expected to take the longest).

Second, the pollutant control cost to achieve each additional unit of pollutant reduction ($/lb) is
increasing as low-cost control options are used first (Kaufman et al., 2021). Kaufman et al.
(2014) estimated that the costs to reduce agricultural nonpoint source loads in Pennsylvania
would increase sharply (exceeding $100/lb/yr for N) when approaching TMDL targets. Urban
nonpoint source controls are often estimated to be an order of magnitude more expensive
($100 to over $1,000/lb/yr for N) than other nonpoint source controls and will still be needed
to meet TMDL targets (Price et al., 2021). Maryland estimates a cost of $1.195 billion for
municipal stormwater systems to achieve the required additional reduction of 85,000 lb/yr of
N, or approximately $1,100/lb/yr (MDE, 2019).11 As a relative comparison, past point source
reductions at municipal WWTPs have been achieved for less than $40/lb/yr (Stephenson &
Shabman, 2017b).

Figure 6.2 (panel A) also shows the implication of both uncertainties and gaps described in
chapters 3 and 4. The costs to obtain any given level of WQS attainment are not precisely
known given uncertainties in producing pollutant reductions and attaining WQS (dotted blue
curves). The existence of estuary and nutrient management response gaps means that more
reductions will be needed to achieve any given WQS. Given changes in the watershed, climate
change, and the presence of implementation and response gaps identified in this report, full
achievement of the existing WQS may be technically infeasible

11
Annualizing $1.19 billion over 20 years at 5%.

80
(particularly for the deep water and deep channel habitats), regardless of the amount of
management effort and resources spent (i.e., the blue cost curve may not cross 100%
attainment). Further, pursuing an unattainable water quality goal can divert resources from
other actions that could be more impactful in terms of the responses of living resources.

FIGURE 6.2.—Conceptual relationship between costs, attainment of WQS, and living resource response.

As implementation costs increase, so does public interest in what improvements will be

achieved with this expenditure of resources. The public benefits associated with achieving WQS
arise largely from the potential improvements in aquatic living resources (designated uses) (see
panel B in figure 6.2).

81
As described in chapter 5, living resources are expected to respond favorably to improved and
expanded habitat conditions (e.g., more usable areas with adequate DO), but the amount of
improvement is uncertain. The location and timing of the water quality improvements will
affect living resource response. For example, improvements in DO in shallow water habitats
that support both nursery habitat and forage fish may generate larger living resource responses
than similar levels of water quality improvement in deeper water habitats. Currently, the TMDL
management focuses on achieving a single baywide reduction goal for N, P, and sediment. The
final Bay TMDL load targets were established to bring all segments and habitats into
compliance, and these targets were driven by the deep water DO criteria, the most difficult to
attain. Thus, costs are expected to increase sharply and living resource improvements diminish
as nonattainment area shrinks to the deeper water habitats.

The living resource response to improvements in DO and water clarity depends on many other
factors that also impact the abundance and composition of living resources. Whether living
resource response to water quality improvement will follow the H or L response curve in figure
6.2, panel B, depends on a variety of factors such as water temperature, salinity, structural
habitat, disease, and harvest. The response of living resources to DO improvements may be
modest (response L) if these other factors are already limiting carrying capacity for fish,
shellfish, and crustacean populations. Achieving a greater living resource response to water
quality improvements will require appropriate types and levels of investments in these
additional factors, some of which are represented as goals and outcomes in the CBWA (e.g.,
sustaining fisheries and recovering habitats). Supporting and increasing progress towards these
other CBWA goals could be constrained if management and legal attention becomes too rigidly
and narrowly focused on the TMDL.

Finally, some of these factors that will significantly influence the composition and abundance of
living resources are beyond CBP management approaches and policies. Climate change will
increase Bay water temperatures and alter historical salinity patterns and habitat (e.g., tidal
wetlands affected by sea level rise; replacement of historic SAV species with new ones). These
factors should be considered and evaluated in any future discussions surrounding the TMDL
and water quality standards.

6.4. Options for improving living resource response

This section examines options for improving potential living resource response to water quality
management efforts. Several options are presented, beginning with options that would require
the least amount of change in practice implementation and associated policies. First, changes
to the way the TMDL is implemented may enhance living resource benefits without requiring
any changes to the WQS (feedback arrow 4 in figure 6.1). Second, modification of and
refinements to the existing WQS (feedback arrow 1 in figure 6.1) may also be pursued.

Consider a tiered approach to structuring the TMDL and achieving WQS. Currently, the
overriding approach driving water quality management is planning and implementing practices

82
to reach the pollution targets for N, P, and sediment, which would meet WQS in all habitats, in
all 92 segments. The current focus is on achieving criteria in the deep water habitats, with the
assumption that attainment there will be the most difficult to achieve. The CBP could consider
where pollutant reductions should occur first to accelerate the potential for living resource
improvements.

Tiered attainment of the TMDL and WQS based on critical habitats (such as shallow and open
waters) or locations could provide early and more substantial living resource benefits and still
contribute to baywide water quality improvements. Chapter 5 established living resource
response likely varies by the location and timing of water quality improvement, so response
could potentially be accelerated by prioritizing water quality improvement efforts in critical
habitats. The implication is that with a focus on shallow and open water habitats full attainment
of the WQS (including the deep water habitats) in the near future may not be necessary to
achieve significant potential gains in living resource response.

There are several approaches to prioritizing TMDL implementation and water quality
attainment. The following list is illustrative, but not comprehensive. Achievement of TMDL
targets could be prioritized according to location (segments) or habitat type for most living
resources. Establishing different deadlines and staggered TMDL pollutant reduction targets may
focus implementation efforts in areas projected to have the largest potential for improving
living resource conditions. Granting temporary or permanent variances for segments or
habitats where achievement is technically infeasible or exceptionally costly could allow
resources to be shifted to areas where attainment is possible and with a higher living resource
impact.12

It should be stressed, however, that the CBP has devoted limited analytical effort to inform
such policy discussions about ecosystem living resource responses to TMDL pollutant
management (Hood et al., 2021). Additional investment in analytical capabilities and models
could make important contributions to a tiered or prioritized approach to the TMDL (for more
details, see the discussion of analytical tools in section 6.6 and Rose et al. [2023]).

Consider revisions to existing WQC. Potential revisions to the existing WQC (DO, Chl a, water
clarity) could strengthen the link between the criteria and living resources. Reevaluation of
WQC may also include consideration of new criteria or new frameworks for devising criteria.
Criteria with potential impacts on living resources could include localized water temperature
anomalies, fine sediment versus total suspended sediment loads, or impacts from chemicals of
emerging concern for living resources. For example, increasing water temperatures will likely
alter ecosystem structure and function, such as a shift in aquatic ecosystems from diatom-
dominated to green algae- or cyanobacteria-dominated (Batiuk et al., 2023). Excessive fine-
grained sediment is responsible for degrading habitat in both the watershed and estuary,

12 EPA has granted a number of water quality variances for specific habitat segments. See chapter 2 for a brief

discussion.

83
including underwater grasses (Kemp et al., 2004, Lefcheck et al., 2017). The potential effects of
toxic contaminants and contaminants of emerging concern on the abundance and composition
of living resources may be significant, and yet information on sources and occurrence has been
limited.

New frameworks for devising criteria could emphasize system resilience, particularly for the
biological criteria such as SAV. Much of the management of Chesapeake Bay, and estuarine
systems around the world, has traditionally focused on ecosystem conditions. But in the face of
changes due to human population growth and land use change, combined with the impacts of
climate change, this focus will necessitate a shift to managing for resilience. Resilience is
defined as the ability to recover from the disturbances that are increasing in frequency and
magnitude, and the concept is applicable to both human and natural systems. Criteria that
reflect resilience characteristics can be developed and would provide a more appropriate target
than historical condition. The SAV criteria were established based on historical estimates of SAV
acreage (which may no longer reflect SAV potential in a warmer Bay). Other SAV criteria,
besides total area, could consider the recovery of SAV in the wake of episodic events or the
speed of recovery. Ultimately, consideration of other designated uses such as “fishable and
swimmable” or recreational use could imply the development of new WQC.

6.5. Improving nonpoint source management

Improving the approaches to reduce nonpoint sources of nutrients is the central TMDL
challenge for achieving WQS under the current TMDL framework. Nonpoint sources are diverse,
involving tens of thousands of people making land use and nutrient use decisions across
agricultural and urban landscapes. Tracking nutrient changes across the landscape is often
challenging, and isolating cause and effect is difficult. Inducing behavioral and technical changes
of sufficient type and scale to achieve nonpoint pollutant reduction goals is a fundamental
challenge faced by large-scale eutrophication reduction efforts worldwide (e.g., Great Lakes,
Gulf of Mexico, Baltic Sea, etc.).

An overriding theme of possible policy reforms described below is to shift incentives and
behavior away from counting practices toward crediting pollutant reduction and water quality
outcomes. The current CBP partner approach for nonpoint source reduction is built on a
practice-based, cost-share incentive structure and the CBP modeling and accounting system to
credit nonpoint source load reductions. While providing accountability, current voluntary
incentive systems focus attention on installing practices to get credit within the CBP TMDL
modeling framework. This system enables EPA to count jurisdictions’ contributions towards
TMDL compliance, but it separates reporting implementation of practices from assessment of
pollutant reductions. Chapter 3 described multiple examples of the potential to improve
nonpoint source reductions through a better understanding of the sources and locations of high
nutrient loss areas, directing management attention and assistance to who does what and
where to reduce those pollutant loads, and rewarding pollutant reduction outcomes. A focus on

84
practice installation can limit the willingness and ability of water quality managers to capitalize
on these opportunities.

Policy reform can take many different forms (Alberini & Segerson, 2002; Ribaudo & Shortle,
2019; Shortle et al., 2021). Given the diversity of people, production systems, and land use
decisions involved in nonpoint source pollution generation, policy change and reform will not
be a simple answer or single approach. Policies to improve nutrient reduction effectiveness for
agricultural regions with large nutrient mass imbalances will look different from policies
targeting fertilizer applications by urban homeowners. Opportunities for incremental
adjustments to nonpoint management programs exist, including better model input data (e.g.,
animal numbers, fertilizer applications, etc.), increasing cost-share amounts, and increasing the
number of service providers to encourage greater BMP implementation (Chesapeake Bay
Commission, 2017; Collins et al., 2022). However, making substantial progress in nonpoint
source management will require programmatic and policy change.

The following are illustrative, but not exhaustive, policy and programmatic options that offer
promise for reducing nutrients from nonpoint sources.

Address mass imbalances. More effective and systematic approaches to addressing nutrient
mass balance issues offer opportunities for substantial, sustained reductions in nonpoint source
nutrient loads.13 Only limited progress can be made if mass imbalances are not adequately
addressed. Most BMPs do not substantially alter mass balances. Many traditional BMPs (e.g.,
cover crops, no-till, many stormwater practices) that do not substantially reduce nutrient inputs
have limited long-term capacities to alter nutrient losses in areas with mass imbalances.
Evidence suggests that policies designed to alter regional mass balances have proven
particularly effective in improving water quality. The P detergent ban and wastewater
treatment technology investments to increase denitrification are examples from point source
management.

The most significant regional mass imbalances are associated with intensive animal agriculture
and row crop production (Beegle, 2013; Kleinman et al., 2012; Spiegal et al., 2020).
Concentrated livestock production and the subsequent land application of manure is a
contributor to the pronounced response gap surrounding the effectiveness of P reduction
efforts. Evidence suggests intensive livestock operations are closely linked to increasing P
(particularly dissolved P) . Addressing regional mass imbalances related to livestock manure
involves improving program implementation and technologies by reducing nutrient inputs,
improving manure distribution (i.e., from surplus to deficit areas), and exporting manure from
the watershed (Flynn et al., 2023; Saha et al., 2022). Specific efforts that could be strengthened
include reducing the nutrient content of livestock feed, increasing feed efficiency, improving
manure transport programs, siting of livestock facilities, using P-based nutrient management,

13 A mass balance approach describes nutrient inputs (e.g., fertilizer and feed) to and outputs (e.g., grain or meat

export, loss to water bodies) from the system, reactions or transformations (e.g., denitrification), and storages
(e.g., buildup of P and N in soil and groundwater) in the system.

85
expanding manure treatment and conversion technologies, and enhancing advanced nutrient
management incentives.

Improve incentives and capacity for identifying who does what when and where. Improving
the incentives and capacities to better identify where nutrient-reducing actions are needed
could improve program effectiveness (reducing both implementation and response gaps).
Chapter 3 noted that the distribution of nutrient losses is uneven and can vary significantly
across the landscape and across land managers. Existing nonpoint source programs and CBP
accounting systems generally do not provide incentives or the capacity to adequately address
and capitalize on this diversity. Research consistently shows that improving targeting of
nonpoint source control investments can generate more, and lower cost, reductions (Choi et
al., 2020; Giri et al., 2012; Easton et al., 2019; Kast et al., 2021; Khanna et al., 2003; Lintern et
al., 2020; Xu et al., 2019; Yang & Weersink, 2005). Improved technical tools and new incentive
programs could better direct efforts to identify low-cost, high-impact opportunities for nutrient
and sediment reduction (Easton et al., 2019; Fleming et al., 2022).

New outcome-based incentive programs could be designed to place more emphasis on

achievement of pollutant reduction outcomes (Collins et al., 2022; Easton et al., 2019; Fleming
et al., 2022). Payment for environmental service programs (e.g., pay for performance, pay for
success, rewards for success) compensate land managers for quantifiable pollutant reductions
or for attainment of observable benchmarks directly or indirectly linked to pollutant reduction
outcomes (e.g., soil nutrient levels, ambient water quality conditions, etc.). Such systems
reward land managers for pollutant reduction achieved rather than paying a portion of costs to
install a BMP. These alternative financial incentives could encourage agricultural producers and
service providers to search out high-impact, low-cost options for nutrient and sediment
reductions. For example, several nutrient removal technologies have high up-front costs but
produce potentially large and low cost-per-unit nutrient reductions (e.g., manure treatment,
bioreactors, riparian buffers, etc.). Pay-for-performance programs could fully compensate land
managers for implementation.14

Pay-for-performance programs depend critically on how performance is defined (Fleming et al.,

2022). A variety of options exist. Using CAST predictions is one way to define nutrient and
sediment reduction performance, but chapter 3 described a number of limits to CAST’s ability
to identify fine-scale, high-loss areas and land managers and incentivize treatment. Other
measures of performance should be explored and developed, including finer-scale models to
identify critical nutrient and sediment sources. Payments or reward programs could be based
on quantifiable changes in indicators or proxies of nutrient and sediment reduction
performance (e.g., reductions in soil P levels). Indeed, some pilot programs have established
reward payments to groups of landowners for achieving improvements in ambient water
quality outcomes (Maille et al., 2009).

14 Performance incentive systems are based on the premise that land managers can profit from water quality

actions (financial returns above costs). This represents a fundamental departure from traditional cost-share
programs.

86
Differential crediting of nonpoint source investments would facilitate the identification and
treatment of high-loss areas and agricultural operations (Easton et al., 2019). The CBP
accounting framework generally counts and credits loads based on averages (e.g., average
loading rates, average BMP effectiveness, average nutrient application rates). Differential
crediting could be accomplished in several ways, including using a finer-scale targeting or
modeling system or granting flexibility to state and local partners to develop, test, evaluate,
and quantify the performance of BMP alternatives that target high-loss areas or operations. For
example, new CBP stream restoration crediting protocols support greater TMDL credit for
projects located where streambank erosion is greater. This approach will necessitate finer-scale
modeling and measurement capabilities that can identify areas of the landscape contributing
disproportionately high nonpoint source nutrient and sediment loads.

More efforts are needed to close the implementation gap. New incentive systems can help
improve program effectiveness, but the extensive history of nonpoint source policy illustrates
that limits to voluntary adoption exist (Liu et al., 2020; Ribaudo, 2015; Ribaudo & Shortle, 2019;
Shortle et al., 2012). Achieving larger nonpoint source reductions may necessitate additional
requirements on nonpoint source management in selective situations (Kling, 2013). Such
requirements can be accompanied with financial assistance and do not necessarily have to be
broad, inflexible, or exceedingly costly. Performance-based requirements, for example,
establish clear outcomes to be achieved but grant flexibility in how to achieve those outcomes.
In vertically integrated production systems, such as poultry and swine, responsibility for
manure management that will address mass imbalances in a watershed could be shared
between integrators and their contract producers rather than, as is now the case, borne just by
the producers. This shared responsibility would incentivize integrators to work with producers
in seeking and then implementing innovative manure treatment, transport, and use strategies
for the relevant watersheds.

Encourage institutional innovation through sandboxing. Improving implementation

effectiveness will require institutional innovation in addition to innovation in pollution control
methods, modeling, and monitoring. Institutional innovations can range from incremental
improvements within existing programs to consideration of new programs, rules, and
accountability systems. This report’s suggestions for making substantial progress in reducing
nonpoint source loads will require exploring more than incremental adjustments, but major
changes in implementation approaches cannot and should not be made based only on
conceptual arguments. Each of the policy options described above, while offering promise, also
faces numerous design questions and implementation challenges (Howard, 2020; Palm-Forster
et al., 2016; Shortle et al., 2021).

Sandboxing is a formalized process that begins with conceptual development of new rules and
programmatic approaches to nonpoint source load reduction or water quality improvement,
followed by trial and evaluation. Sandboxing allows refinement of institutional reform details

87
before any programmatic changes are made and can proceed without disrupting the operation
of existing programs and operations (Higgins & Male, 2019; O’Sullivan, 2021).

The regulatory agency or programmatic organization would work with the sandbox proposer to
operate in an isolated environment of sufficient size and autonomy to implement and test the
efficacy of a potential change. A sandbox would require a commitment of time and resources
by the CBP and its partners to effectively operate the trial (sandbox). The authorizing agency or
organization would grant permission to form the sandbox under the condition that the
proposer provides both a well-developed plan and a way to evaluate outcomes from the new
rule or programmatic change being considered. To promote innovation, the authorizing agency
or organization would also make commitments to pursue reforms if the sandbox trial produces
demonstrative improvements in the agreed-upon outcomes. Two examples of possible
sandboxing applications are described in text box 6.1.

Supporting innovation and change recognizes that some proposed innovations may sound
attractive in concept but will prove to be ineffective or too difficult to implement in practice.
The possibility of failure should not discourage pursuit of such opportunities, with the caveat
that learning occurs as a result (i.e., the reason for failure is understood). The most effective
learning often starts with failure. The potential reward, however, is the discovery of otherwise
unrealized improvements in water quality outcomes.

6.6. Expanding adaptive decision-making and improving program learning

As described in chapter 2, the CBP has implemented a decision framework intended to

continuously monitor and evaluate progress toward achievement of specific CBP program goals
of the CBWA. The SRS process and the TMDL accountability framework adjusts implementation
based on defining implementation goals, describing factors influencing goal attainment, and
assessing management gaps. Adaptive management processes have occurred primarily within
the implementation phase (feedback arrow 5 in figure 6.1) and have been justifiably focused on
assessing whether planned and implemented management actions are being undertaken to
achieve the TMDL, including implementing and tracking practices, refining modeling and
accounting tools, and assessing staff and budgetary requirements to implement practices (CBP,
2020).

In order to respond effectively to the issues raised in this report, the partnership will need to
expand the current sphere and capacity of adaptive management processes for the TMDL and
attainment of WQS to include the scope and scale of program changes described in this report.
This includes consideration of changes across the entire range of water quality policy
management (figure 6.1, arrows 4, 3, 2, and 1), from understanding how living resources may
respond to water quality management to addressing uncertainties in water quality
management, particularly with respect to nonpoint sources. The question is not simply: Are the
jurisdictions implementing the planned actions? Rather, the questions must be asked: Are the
actions producing the anticipated pollution reductions and attainment of WQS? If so, why? If

88
not, why not? This shift in thinking will require the CBP to reconsider how decision makers are
to be involved and how to apply a more expansive adaptive management process. Some
options to evolve the adaptive management processes are described in the rest of this section.

Expand participation in adaptive management. An expanded adaptive management process

will involve increasing the participation of decision makers across all levels of the program. The
jurisdictions and EPA are responsible for implementing these activities for the TMDL. However,
the policy and management opportunities for improving approaches to meet the TMDL that are
outlined in this report involve greater resource and management commitments. The
consideration and implementation of changes to WIPs, funding support, and water quality goals
will require expanded involvement of those participants who have authority for authorizing and

89
funding programs. Within the Chesapeake Bay organizational hierarchy, this would include
members of the Chesapeake Executive Council and the Principals’ Staff Committee, which
include leaders of federal and state programs.

Use decision science for enhanced adaptive management. Enhancing adaptive management
requires addressing the challenge of how to bring technical, programmatic, and scientific
knowledge about the issues raised in this report to the people with the authority to make the
choices regarding water quality goals, funding, and regulatory and administrative changes.
Decision science can be used to effectively present and consider policy alternatives and options
which produce complex and uncertain outcomes. Decision scientists have developed processes
under a variety of labels (e.g., structured decision making, shared vision planning, collaborative
modeling, etc.) that effectively combine planning processes and technical analyses (Bourger,
2011; Gregory et al., 2012; Runge et al., 2020). This includes the analytical tools and processes
needed to identify decision-relevant uncertainties and the most-probably-effective actions
given the uncertainties. Because the resources for pollution control implementation are limited,
strategy development also requires prioritization of uncertain actions with the objective of
maximizing potential outcomes at lowest possible cost.

Expand analytical tools to support decision-making under uncertainty. The CBP has limited
capacity to systematically evaluate uncertainties and response gaps described in this report.
Deterministic models providing single estimates of pollutant loads for all inputs, land uses, and
management actions are not well suited for evaluating and addressing uncertainty. Such
modeling approaches make it difficult to assess the performance risk of different BMPs, inform
decision makers of uncertainties, or assess the robustness of management actions to
underlying assumptions or changing environmental conditions.

An expanded adaptive management process will need to better incorporate tools and
processes to identify and reduce decision-relevant uncertainties (Marchau et al., 2019). For
example, the expected value of information (Runge et al., 2011) can be used to identify which
uncertainties in nonpoint source management are most important to resolve in order to
improve water quality. Such approaches aim to identify those uncertainties that pose the
greatest risk of not achieving management objectives, identify how much a given outcome
could be improved if a given uncertainty were resolved, and identify the cost of error. Robust
decision-making and modeling tools seek to identify solutions and management strategies that
perform well under many possible assumptions, rather than optimal solutions that minimize
risk given a stringent set of assumptions. Modeling of watershed processes that involves more
explicit characterization of uncertainties in system process and model parameters can be used
for a variety of purposes, including supporting program design, implementation, and
prioritization of research needs.

Improve capacities to assess living resource response. Opportunities may exist to accelerate
and improve living resource response to water quality management actions. Development of
additional analytical and modeling capabilities would enable broader conclusions and more

90
informed policy discussions about the role of the TMDL in improving living resources (Rose et
al., 2023). The process of constructing models of multiple species’ (ecological) responses to
policy-relevant water quality conditions would generate useful information about identifying
and prioritizing among living resource goals and targets. Analyses to address these
management questions can be used to identify the tradeoffs among species (or classes of
species) for alternative combinations of restoration actions, provide an assessment of progress
for species of interest and their food webs, and inform what magnitude of changes in water
quality and habitat are needed for certain sets of responses. Such analyses would be useful to
inform discussions about refining the WQC and help establish clear public expectations about
what living resource responses might be achieved from water quality improvement efforts.

If it is determined that more comprehensive and robust answers to the management questions
about living resource response are needed, obtaining those answers will involve changes for
assessing living resource responses. Such changes would require a concerted and coordinated
effort, and options are summarized in Rose et al. (2023). Multiple issues will need to be
considered, such as which species to address, what data are available, what models to use in
new analyses, and how to leverage the extensive analyses already done. Rose et al. (2023) offer
a framework for guiding the decision about how to assess living resource responses that can
also be used to formulate a strategic analysis plan and interpret results.

Target monitoring and research to support adaptive management. Monitoring networks have
been used primarily to identify water quality trends, inform and calibrate CBP watershed
models, and track attainment of WQS. Ambient monitoring networks have not been designed
to address critical uncertainties regarding nonpoint source program effectiveness, to assess the
efficacy of particular technology or policy approaches, or to address critical uncertainties in
estuary response.

There is widespread acknowledgment that additional monitoring is necessary (CBP, 2022), but
an effective adaptive management process needs to identify key questions and appropriate
monitoring metrics to assess system response and inform future actions. This is difficult
because most water quality improvement strategies incorporate multiple actions, and
determining effects of specific actions is difficult. Designing monitoring networks and
associated research efforts for better evaluating current program efficacy and the implications
for future program goals will be a critical undertaking. While acknowledging these challenges,
additional monitoring and research are critical to assess the following situations.

Finer-scale watershed monitoring could improve assessment of the efficacy and uncertainties
related to nonpoint source management, for example, finer-scale water quality monitoring and
monitoring of intermediate indicators of change (e.g., groundwater, soil conditions, etc.).
Monitoring for BMP implementation effectiveness to improve water quality would also need to
include systematic evaluation of people’s nutrient use, BMP adoption, and land use behavior.
For example, nutrient use behavior and BMP maintenance are either assumed or poorly

91
understood but have important influence on water quality changes. Relatively little has been
invested in understanding behavioral change under different implementation programs.

For better understanding estuary response, elucidating relationships between stressor

reduction efforts and achievement of WQS on a subsegment or smaller spatial scale is
particularly important for shallow water habitats. It is notable that 17 of the 31 outcomes in the
CBWA refer to conditions in shallow waters, which is an area of high engagement by
stakeholders. The shallow waters present an opportunity to resolve critical uncertainties in the
relationship of load reductions to living resource impacts. Dissolved oxygen in these habitats
responds to a number of variables beyond nutrient and sediment reductions, and
understanding these dynamics is critical for both identifying effective management actions in
the shallow waters themselves and understanding the relationships between shallow water
habitats and WQS in other habitats (e.g., deep water DO).

Tipping points in water clarity, DO, and SAV relationships have been shown to exist at the scale
of subsystems in the Bay. Monitoring and research to either discern tipping points or determine
the thresholds associated with the various tipping points can indicate progressive conditions on
either a degradation or restoration trajectory; crossing the thresholds in the direction of
restoration is of particular interest now.

92
List of Abbreviations

BMP best management practice

CBP Chesapeake Bay Program
CAFO concentrated animal feeding operation
CAST Chesapeake Assessment Scenario Tool
CBWA Chesapeake Bay Watershed Agreement
Chl a Chlorophyll a
CWA Clean Water Act
DO dissolved oxygen
DU designated use
EPA U.S. Environmental Protection Agency
GIT goal implementation team
MS4 municipal separate storm sewer system
N nitrogen
NASS National Agricultural Statistics Service
NTN Nontidal Monitoring Network
P phosphorus
RIM river input monitoring
SAV submerged aquatic vegetation
SPARROW Spatially Referenced Regression on Watershed Attributes
SRS strategy review system
STAC Scientific and Technical Advisory Committee
TMDL total maximum daily load
TN total nitrogen
TP total phosphorus
TSS total suspended sediment
USDA U.S. Department of Agriculture
USGS U.S. Geological Survey
WIP watershed implementation plan
WQC water quality criteria
WQS water quality standards
WWTP wastewater treatment plant

93
References

Aguilar, M. F., & Dymond, R. L. (2019). Evaluation of stormwater control measure performance
uncertainty. Journal of Environmental Engineering, 145(10), 04019060.
https://doi.org/10.1061/(ASCE)EE.1943-7870.0001590
Alberini, A., & Segerson, K. (2002). Assessing voluntary programs to improve environmental
quality. Environmental and Resource Economics, 22, 157–184.
https://doi.org/10.1023/A:1015519116167
Ariely, D. (2008). Predictably irrational: The hidden forces that shape our decisions. Harper
Collins.
Ator, S. W. (2019). Spatially referenced models of streamflow and nitrogen, phosphorus, and
suspended-sediment loads in streams of the northeastern United States. U.S. Geological
Survey Scientific Investigations Report 2019-5118. https://doi.org/10.3133/sir20195118
Ator, S. W., Blomquist, J. D., Webber, J. S., & Chanat, J. G. (2020). Factors driving nutrient trends
in streams of the Chesapeake Bay watershed. Journal of Environmental Quality, 49(4),
812–834. http://doi.org/10.1002/jeq2.20101
Ator, S. W., & Denver, J. M. (2015). Understanding nutrients in the Chesapeake Bay watershed
and implications for management and restoration—The Eastern Shore. U.S. Geological
Survey Circular 1406. http://dx.doi.org/10.3133/cir1406
Ator, S. W., Garcia, A. M., Schwarz, G. E., Blomquist, J. D., & Sekellick, A. J. (2019). Toward
explaining nitrogen and phosphorus trends in Chesapeake Bay tributaries, 1992–2012.
JAWRA Journal of the American Water Resources Association, 55(5), 1149–1168.
https://doi.org/10.1111/1752-1688.12756
Baird, D., & Ulanowicz, R. E. (1989). The seasonal dynamics of the Chesapeake Bay ecosystem.
Ecological Monographs, 59(4), 329–364. https://doi.org/10.2307/1943071
Basu, N. B., Van Meter, K. J., Byrnes, D. K., Van Cappellen, P., Brouwer, R., Jacobsen, B. H., Jarjo,
J., Cunha, M. C., Nelson, N., Bhattacharya, R., Destouni, G., & Olsen, S. B. (2022).
Managing nitrogen legacies to accelerate water quality improvement. Nature
Geoscience, 15(2), 97–105. https://doi.org/10.1038/s41561-021-00889-9
Batiuk, R. A., Breitburg, D. L., Diaz, R. J., Cronin, T. M., Secor, D. H., & Thursby, G. (2009).
Derivation of habitat-specific dissolved oxygen criteria for Chesapeake Bay and its tidal
tributaries. Journal of Experimental Marine Biology and Ecology, 381, S204–S215.
https://doi.org/10.1016/j.jembe.2009.07.023
Batiuk, R., Brownson, K., Dennison, W., Ehrhart, M., Hanson, J., Hanmer, R., Landry, B., Reichert-
Nguyen, J., Soueidan, J., Tassone, S., & Vogt, B. (2023). Rising watershed and bay water
temperatures: Ecological implications and management responses. STAC Publication
Number 23-001, Chesapeake Bay Program Scientific and Technical Advisory Committee
(STAC), Edgewater, MD. 505 pp.
Beegle, D. (2013). Nutrient management and the Chesapeake Bay. Journal of Contemporary
Water Research & Education, 151(1), 3–8. https://doi.org/10.1111/j.1936-
704X.2013.03146.x
Blanco-Canqui, H. (2018). Cover crops and water quality. Agronomy Journal, 110(5), 1633–1647.
https://doi.org/10.2134/agronj2018.02.0077

94
Blankenship, K. (2022). Chesapeake Bay & the case of the missing chickens. Bay Journal, July 13.
https://www.bayjournal.com/news/pollution/chesapeake-bay-the-case-of-the-
missingchickens/article_a514fd52-02aa-11ed-891f-dfd96bff1134.html
Boesch, D. F. (2019). Barriers and bridges in abating coastal eutrophication. Frontiers in Marine
Science, 6, 123. https://doi.org/10.3389/fmars.2019.00123
Böhlke, J. K. (2002). Groundwater recharge and agricultural contamination. Hydrogeology
Journal, 10(1), 153–179. https://doi.org/10.1007/s10040-001-0183-3
Böhlke, J. K., & Denver, J. M. (1995). Combined use of groundwater dating, chemical, and
isotopic analyses to resolve the history and fate of nitrate contamination in two
agricultural watersheds, Atlantic coastal plain, Maryland. Water Resources Research,
31(9), 2319–2339. https://doi.org/10.1029/95WR01584
Bonham, J. G., Bosch, D. J., & Pease, J. W. (2006). Cost-effectiveness of nutrient management
and buffers: Comparisons of two spatial scenarios. Journal of Agricultural and Applied
Economics, 38(1), 17–32. https://doi.org/10.1017/S1074070800022045
Boomer, K., Boynton, W., Muller, A., Muller, D., & Sellner, K. (2019). Revisiting coastal land-
water interactions: The triblet connection. STAC Publication Number 19-005,
Chesapeake Bay Program Scientific and Technical Advisory Committee (STAC),
Edgewater, MD. 37 pp. https://www.chesapeake.org/stac/wp-
content/uploads/2019/11/STAC-Report_Triblets-FINAL.pdf
Bourger, L. (Ed.). (2011). Converging waters: Integrating collaborative modeling with
participatory processes to make water resources decisions. U.S. Army Corps of
Engineers, Institute for Water Resources. https://hdl.handle.net/11681/44821
Boynton, W. R., Hodgkins, C. L. S., O’Leary, C. A., Bailey, E. M., Bayard, A. R., & Wainger, L. A.
(2014). Multi-decade responses of a tidal creek system to nutrient load reductions:
Mattawoman Creek, Maryland USA. Estuaries and Coasts, 37, 111–127.
https://doi.org/10.1007/s12237-013-9690-4
Boynton, W. R., Testa, J. M., & Kemp, W. M. (2009). An ecological assessment of the Corsica
River estuary and watershed: Scientific advice for future water quality management.
Report to Maryland Department of Natural Resources. University of Maryland Center for
Environmental Science Technical Report Series No. TS-587-09.
https://gonzo.cbl.umces.edu/documents/Publications/FULL%20CORSICA%20TEXT%20FI
NAL%2023OCT09.pdf
Breitburg, D. L., Craig, J. K., Fulford, R. S., Rose, K. A., Boynton, W. R., Brady, D. C., Ciotti, B. J.,
Diaz, R. J., Friedland, K. D., Hagy, J. D., Hart, D. R., Hines, A. H., Houde, E. D., Kolesar, S.
E., Nixon, S. W., Rice, J. A., Secor, D. H., & Targett, T. E. (2009). Nutrient enrichment and
fisheries exploitation: Interactive effects on estuarine living resources and their
management. Hydrobiologia, 629(1), 31–47. https://doi.org/10.1007/s10750-009-9762-
4
Breitburg, D. L, Hondorp, D. W., Davias, L. A., & Diaz, R. J. (2009). Hypoxia, nitrogen, and
fisheries: Integrating effects across local and global landscapes. Annual Review of
Marine Science, 1(1), 329–349. https://doi.org/10.1146/annurev.marine.010908.163754
Breitburg, D., Levin, L. A., Oschlies, A., Grégoire, M., Chavez, F. P., Conley, D. J., Garçon, V.,
Gilbert, D., Gutiérrez, D., Isensee, K., Jacinto, G. S., Limburg, K. E., Montes, I., Naqvi, S.

95
 W. A., Pitcher, G. C., Rabalais, N. N., Roman, M. R., Rose, K. A., Siebel, B. A., … Zhang, J.
(2018). Declining oxygen in the global ocean and coastal waters. Science, 359(6371).
https://doi.org/10.1126/science.aam7240
Bryant, R., Baldwin, A., Cahall, B., Christianson, L., Jaynes, D., Penn, C., & Schwartz, S. (2019).
Best management practices for agricultural ditch management in the phase 6
Chesapeake Bay Watershed Model. L. Collins, J. Hanson & C. Gill (Eds.). CBP/TRS-326-19.
https://d18lev1ok5leia.cloudfront.net/chesapeakebay/documents/ag_ditches_bmp_pa
nel_report_draft_for_cbp_review_04sep2019.pdf
Carpenter, S. R. (2008). Phosphorus control is critical to mitigating eutrophication. Proceedings
of the National Academy of Sciences, 105(32), 11039–11040.
https://doi.org/10.1073/pnas.0806112105
Cerco, C.F., & Noel, M.R. (2007). Can oyster restoration reverse cultural eutrophication in
Chesapeake Bay? Estuaries and Coasts, 30, 331–343.
https://doi.org/10.1007/BF02700175
Cerosaletti, P. E., Fox, D. G., & Chase, L. E. (2004). Phosphorus reduction through precision
feeding of dairy cattle. Journal of Dairy Science, 877, 2314–2323.
https://doi.org/10.3168/jds.S0022-0302(04)70053-3
Chanat, J. G., & Yang, G. (2018). Exploring drivers of regional water-quality change using
differential spatially referenced regression—A pilot study in the Chesapeake Bay
watershed. Water Resources Research, 54(10), 8120–8145.
https://doi.org/10.1029/2017WR022403
Chesapeake Bay Commission. (2017). Boots on the ground: Improving technical assistance to
farmers. https://www.chesbay.us/library/public/documents/Policy-Reports/CBC-TA-
Report-Boots-on-the-Ground.pdf
Chesapeake Bay Program (CBP). (n.d.-a). 2025 Watershed Implementation Plans. Chesapeake
Progress. https://www.chesapeakeprogress.com/clean-water/watershed-
implementation-plans
Chesapeake Bay Program (CBP). (n.d.-b). Chesapeake Decisions
https://www.chesapeakebay.net/what/what-guides-us/decisions
Chesapeake Bay Program (CBP). (n.d.-c). Model documentation.
https://cast.chesapeakebay.net/Documentation/ModelDocumentation
Chesapeake Bay Program (CBP). (n.d.-d). Tidal water quality change: 2021 results. Integrated
trends analysis team.
https://d18lev1ok5leia.cloudfront.net/chesapeakebay/documents/2021-Tidal-Trends-
Summary-Document.pdf
Chesapeake Bay Program (CBP). (n.d.-e). Water quality standards attainment and monitoring.
Chesapeake Progress. https://www.chesapeakeprogress.com/clean-water/water-quality
Chesapeake Bay Program (CBP). (1987). Chesapeake Bay Agreement.
https://www.chesapeakebay.net/what/publications/chesapeake-bay-agreement-1987
Chesapeake Bay Program (CBP). (2008). Chesapeake Bay segmentation scheme (for 303d
listing—92 segments). Maps.
https://www.chesapeakebay.net/what/maps/tag/segments

96
Chesapeake Bay Program (CBP). (2014). Chesapeake Bay Watershed Agreement.
https://www.chesapeakebay.net/what/what-guides-us/watershed-agreement
Chesapeake Bay Program (CBP). (2022). Enhancing the Chesapeake Bay program monitoring
networks—A report to the Principals’ Staff Committee.
https://d18lev1ok5leia.cloudfront.net/chesapeakebay/documents/FINAL_Enhancing_th
e_Chesapeake_Bay_Program_Monitoring_Networks_A-
Report_to_the_Principals_Staff_Committee_10.13.22-1_2022-10-18-211409_qtmr.pdf
Chesapeake Bay Program (CBP). (2020). 2025 WIP and Water Quality Standards Attainment and
Monitoring Outcomes.
https://d18lev1ok5leia.cloudfront.net/chesapeakebay/documents/wq_revised_clean_m
s_2021-22_v1.6.2021.pdf
Choi, D. S., Ready, R. C., & Shortle, J. S. (2020).Valuing water quality benefits from adopting best
management practices: A spatial approach. Journal of Environmental Quality, 49(3),
582–592. https://doi.org/10.1002/jeq2.20005
Christensen, V., Beattie, A., Buchanan, C., Ma, H., Martell, S. J. D., Latour, R., Preikshot, D.,
Sigrist, M., Uphoff, J. H., Walters, C. J., Wood, R. J., & Townsend, H. (2009). Fisheries
ecosystem model of the Chesapeake Bay: Methodology, parameterization, and model
exploration. U.S. Department of Commerce NOAA Technical Memorandum NMFS-
F/SPO-106. https://repository.library.noaa.gov/view/noaa/4180
Claassen, R., Horowitz, J., Duquette, E., & Ueda, K. (2014). Additionality in US agricultural
conservation and regulatory offset programs. USDA-ERS Economic Research Report 170.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2502846
Clune, J. W., & Capel, P. D. (Eds.). (2021). Nitrogen in the Chesapeake Bay watershed—A century
of change, 1950–2050 (ver. 1.1, December 2021). U.S. Geological Survey Circular 1486.
https://doi.org/10.3133/cir1486
Collins, L., Stephenson, K., Palm-Forster, L., Power, L., Gibson, A., Arbuckle, J., Handen, A., &
Read, D. (2022). Overcoming the hurdle: Addressing implementation of agricultural best
management practices (BMPs) through a social science lens. STAC Publication Number
22-002, Chesapeake Bay Program Scientific and Technical Advisory Committee (STAC),
Edgewater, MD. 107 pp. https://www.chesapeake.org/stac/wp-
content/uploads/2022/10/FINAL-Report_Overcoming-the-Hurdle_22-002.pdf
Daryanto, S., Wang, L., & Jacinthe, P. A. (2017). Meta-analysis of phosphorus loss from no-till
soils. Journal of Environmental Quality, 46(5), 1028–1037.
https://doi.org/10.2134/jeq2017.03.0121
Denver, J. M., Tesoriero, A. J., & Barbaro, J. R. (2010). Trends and transformation of nutrients
and pesticides in a coastal plain aquifer system, United States. Journal of Environmental
Quality, 39(1), 154–167. https://doi.org/10.2134/jeq2009.0107
Du, J., Shen, J., Park, K., Wang, Y. P., & Yu, X. (2018). Worsened physical condition due to
climate change contributes to the increasing hypoxia in Chesapeake Bay. Science of The
Total Environment, 630, 707–717. https://doi.org/10.1016/j.scitotenv.2018.02.265
Duke, J. M., Johnston, R. J., Shober, A. L., & Liu, Z. (2022). Barriers to cover crop adoption:
Evidence from parallel surveys in Maryland and Ohio. Journal of Soil and Water
Conservation, 77(2), 198–211. https://doi.org/10.2489/jswc.2022.00062

97
Duncan, E. W., Osmond, D. L., Shober, A. L., Starr, L., Tomlinson, P., Kovar, J. L., Moorman, T. B.,
Peterson, H. M., Fiorellino, N. M., & Reid, K. (2019). Phosphorus and soil health
management practices. Agricultural and Environmental Letters, 4(1), 190014.
https://doi.org/10.2134/ael2019.04.0014
Easton, Z. M., Bock, E., & Stephenson, K. (2019). Feasibility of using woodchip bioreactors to
treat legacy nitrogen to meet Chesapeake Bay water quality goals. Environmental
Science & Technology, 53(21), 12291–12299. https://doi.org/10.1021/acs.est.9b04919
Easton, Z., Scavia, D., Alexander, R., Band, L., Boomer, K., Kleinman, P., Martin, J., Miller, A.,
Pizzuto, J., Smith, D., & Welty, C. (2017). Scientific and Technical Advisory Committee
review of the phase 6 Chesapeake Bay Watershed Model. STAC Publication Number 17-
007, Chesapeake Bay Program Scientific and Technical Advisory Committee (STAC),
Edgewater, MD. 47 pp. https://www.chesapeake.org/pubs/379_Easton2017.pdf
Easton, Z., Stephenson, K., Benham, B., Böhlke, J. K., Buda, A., Collick, A., Fowler, L., Gilinsky, E.,
Hershner, C., Miller, A., Noe, G., Palm-Forster, L., & Thompson, T. (2023). Evaluation of
Management Efforts to Reduce Nutrient and Sediment Contributions to the Chesapeake
Bay Estuary. STAC Publication Number 23-003, Chesapeake Bay Program Scientific and
Technical Advisory Committee (STAC), Edgewater, MD. 53 pp.
Easton, Z. M., Stephenson, K., Collick, A., Fleming, P. M., Kellner, E., Martin, J., Ribaudo, M., &
Shenk, G. (2020). Increasing effectiveness and reducing the cost of non-point source best
management practice implementation: Is targeting the answer? STAC Publication
Number 20-002, Chesapeake Bay Program Scientific and Technical Advisory Committee
(STAC), Edgewater, MD. 53 pp. https://www.chesapeake.org/stac/wp-
content/uploads/2020/02/FINAL_STAC-Report_BMP-Targeting.pdf
Easton, Z. M., Walter, M. T., & Steenhuis, T. S. (2008). Combined monitoring and modeling
indicate the most effective agricultural best management practices. Journal of
Environmental Quality, 37(5), 1798–1809. https://doi.org/10.2134/jeq2007.0522
Fabrizio, M. C., Tuckey, T. D., Bever, A. J., & MacWilliams, M. L. (2021). The extent of seasonally
suitable habitats may limit forage fish production in a temperate estuary. Frontiers in
Marine Science, 8, 706666. https://doi.org/10.3389/fmars.2021.706666
Fanelli, R. M., Blomquist, J. D., & Hirsch, R. M. (2019). Point sources and agricultural practices
control spatial-temporal patterns of orthophosphate in tributaries to Chesapeake Bay.
Science of The Total Environment, 652, 422–433.
https://doi.org/10.1016/j.scitotenv.2018.10.062
Fiorellino, N., Kratochvil, R., & Coale, F. (2017). Long-term agronomic drawdown of soil
phosphorus in mid-Atlantic coastal plain soils. Agronomy Journal, 109(2), 455–461.
https://doi.org/10.2134/agronj2016.07.0409
Fisher, T. R., Fox, R. J., Gustafson, A. B., Koontz, E., Lepori-Bui, M, & Lewis, J. (2021). Localized
water quality improvement in the Choptank estuary, a tributary of Chesapeake Bay.
Estuaries and Coasts, 44, 1274–1293. https://doi.org/10.1007/s12237-020-00872-4
Fleming, B. J., Archfield, S. A., Hirsch, R. M., Kiang, J. E., & Wolock, D. M. (2020) Spatial and
temporal patterns of low streamflow and precipitation changes in the Chesapeake Bay
watershed. JAWRA Journal of the American Water Resources Association, 57(1), 96–108.
https://doi.org/10.1111/1752-1688.12892

98
Fleming, P. M., Lichtenberg, E., & Newburn, D. A. (2018). Evaluating impacts of agricultural cost
sharing on water quality: Additionality, crowding in, and slippage. Journal of
Environmental Economics and Management, 92, 1–19.
https://doi.org/10.1016/j.jeem.2018.08.007
Fleming, P. M., Stephenson, K., Collick, A., & Easton, Z. (2022). Targeting for nonpoint source
pollution reduction: A synthesis of lessons learned, remaining challenges, and emerging
opportunities. Journal of Environmental Management, 308, 114649.
https://doi.org/10.1016/j.jenvman.2022.114649
Flynn, K. C., Spiegal, S., Kleinman, P. J. A., Meinen, R. J., & Smith, D. R. (2023). Manureshed
management to overcome longstanding nutrient imbalances in US agriculture.
Resources, Conservation and Recycling, 188, 106632.
https://doi.org/10.1016/j.resconrec.2022.106632
Fox, R. J., Fisher, T. R., Gustafson, A. B., Koontz, E. L., Lepori-Bui, M., Kvalnes, K. L., Bunnell-
Young, D. E., Gardner, J. R., Lewis, J., Winsten, J. R., Fisher, K. A., & Silaphone, K. (2021).
An evaluation of the Chesapeake Bay management strategy to improve water quality in
small agricultural watersheds. Journal of Environmental Management, 299, 113478.
https://doi.org/10.1016/j.jenvman.2021.113478
Frankel, L. T., Friedrichs, M. A. M., St-Laurent, P., Bever, A. J., Lipcius, R. N., Bhatt, G., & Shenk,
G. W. (2022). Nitrogen reductions have decreased hypoxia in the Chesapeake Bay:
Evidence from empirical and numerical modeling. Science of The Total Environment,
814, 152722. https://doi.org/10.1016/j.scitotenv.2021.152722
Freedman, P. L., Reckhow, K. H., Shabman, L., Benaman, J., Schwer, R., & Stiles, T. (2008). A new
approach to adaptive implementation in the TMDL program. Water Practice, 2(1), 1–9.
Froelich, P. N. (1988). Kinetic control of dissolved phosphate in natural rivers and estuaries: A
primer on the phosphate buffer mechanism. Limnology and Oceanography, 33(4, part
2), 649–668. https://doi.org/10.4319/lo.1988.33.4part2.0649
Fulford, R. S., Breitburg, D. L., Luckenbach, M., & Newell, R. I. E. (2010). Evaluating ecosystem
response to oyster restoration and nutrient load reduction with a multispecies
bioenergetics model. Ecological Applications, 20(4), 915–934.
https://doi.org/10.1890/08-1796.1
Fulton, E. A., Link, J. S., Kaplan, I. C., Savina-Rolland, M., Johnson, P., Ainsworth, C., Horne, P.,
Gorton, R., Gamble, R. J., Smith, A. D. M., & Smith, D. C. (2011). Lessons in modelling and
management of marine ecosystems: The Atlantis experience. Fish and Fisheries, 12(2),
171–188. https://doi.org/10.1111/j.1467-2979.2011.00412.x
Ganju, N. K., Testa, J. M., Suttles, S. E., & Aretxabaleta, A. L. (2020). Spatiotemporal variability of
light attenuation and net ecosystem metabolism in a back-barrier estuary. Ocean
Science, 16(3), 593–614. https://doi.org/10.5194/os-16-593-2020
Giri, S., Nejadhashemi, A. P., & Woznicki, S. A. (2012). Evaluation of targeting methods for
implementation of best management practices in the Saginaw River watershed. Journal
of Environmental Management, 103, 24–40.
https://doi.org/10.1016/j.jenvman.2012.02.033
Goals Project. (2015). The Baylands and climate change: What we can do. Baylands ecosystem
habitat goals science update 2015. San Francisco Bay Area Wetlands Ecosystem Goals

99
 Project, California State Coastal Conservancy.
https://www.sfei.org/sites/default/files/biblio_files/Baylands_Complete_Report.pdf
Gonzalez, G., Mosley, A., & Stephenson, K. (2016). An analysis of stormwater utility incentive
programs in the Chesapeake Bay. Watershed Science Bulletin, December.
http://owl.cwp.org/mdocs-posts/an-analysis-of-stormwater-utility-incentive-programs-
in-the-chesapeake-bay/
Greene, E. A., LaMotte, A. E., & Cullinan, K.-A. (2005). Ground-water vulnerability to nitrate
contamination at multiple thresholds in the mid-Atlantic region using spatial probability
models. U.S. Geological Survey Scientific Investigations Report 2004-5118.
https://doi.org/10.3133/sir20045118
Gregory, R., Failing, L., Harstone, M. Long, G., McDaniels, T., & Ohlson, D. (2012). Structured
decision making: A practical guide to environmental management choices. Wiley-
Blackwell.
Gruber, R. K., & Kemp, W. M. (2010). Feedback effects in a coastal canopy-forming submersed
plant bed. Limnology and Oceanography, 55(6), 2285–2298.
https://doi.org/10.4319/lo.2010.55.6.2285
Gurbisz, C., & Kemp, W. M. (2014). Unexpected resurgence of a large submersed plant bed in
Chesapeake Bay: Analysis of time series data. Limnology and Oceanography, 59(2), 482–
494. https://doi.org/10.4319/lo.2014.59.2.0482
Gurbisz, C., Kemp, W. M., Cornwell, J. C., Sanford, L. P., Owens, M. S., & Hinkle, D. C. (2017).
Interactive effects of physical and biogeochemical feedback processes in a large
submersed plant bed. Estuaries and Coasts, 40(6), 1626–1641.
https://doi.org/10.1007/s12237-017-0249-7
Hanson, J., Bock, E., Asfaw, B., & Easton, Z. M. (2022). A systematic review of Chesapeake Bay
climate change impacts and uncertainty: Watershed processes, pollutant delivery and
BMP performance. CBP/TRS-330-22. https://bit.ly/BMP-CC-synth
Heathwaite, L., Sharpley, A., & Gburek, W. (2000). A conceptual approach for integrating
phosphorus and nitrogen management at watershed scales. Journal of Environmental
Quality, 29(1), 158–166. https://doi.org/10.2134/jeq2000.00472425002900010020x
Hernandez Cordero, A. L., Tango, P. M., & Batiuk, R. A. (2020). Development of a multimetric
water quality indicator for tracking progress towards the achievement of Chesapeake
Bay water quality standards. Environmental Monitoring and Assessment, 192, 94.
https://doi.org/10.1007/s10661-019-7969-z
Hershner, C. (2011). The certainty of confronting uncertainty in the Chesapeake Bay restoration
effort. Choices, 26(3).
Higgins, P., & Male, T. (2019). Sandboxing nature: How regulatory sandboxes could help restore
species, enhance water quality and build better habitats faster. Environmental Policy
Innovation Center.
https://www.policyinnovation.org/blog/sandbox?rq=sandboxing%20nature
Hinson, K. E., Friedrichs, A. M., St-Laurent, P., Da, F., & Najjar, R. G. (2022). Extent and causes of
Chesapeake Bay warming. JAWRA Journal of the American Water Resources
Association, 58(6), 805–825. https://doi.org/10.1111/1752-1688.12916

100
Holman, I. P., Whelan, M. J., Howden, N. J. K., Bellamy, P. H., Willby, N. J., Rivas-Casado, M., &
McConvey, P. (2008). Phosphorus in groundwater—An overlooked contributor to
eutrophication? Hydrological Processes, 22(26), 5121–5127.
https://doi.org/10.1002/hyp.7198
Hood, R. R., Shenk, G. W., Dixon, R. L., Smith, S. M., Ball, W. P., Bash, J. O., Batiuk, R, Boomer, K.,
Brady, D. C., Cerco, C., Claggett, P., de Mutsert, K., Easton, Z. M., Elmore, A. J.,
Friedrichs, M. A. M., Harris, L. A., Ihde, T. F., Lacher, L., Li, L., Linker, L. C., … Zhang, Y. J.
(2021). The Chesapeake Bay Program modeling system: Overview and
recommendations for future development. Ecological Modeling, 456, 109635.
https://doi.org/10.1016/j.ecolmodel.2021.109635
Hopkins, K. G., Bhaskar, A. S., Woznicki, S. A., & Fanelli, R. M. (2020). Changes in event-based
streamflow magnitude and timing after suburban development with infiltration-based
stormwater management. Hydrological Processes, 34(2), 387–403.
https://doi.org/10.1002/hyp.13593
Howard, G. (2020). Additionality violations in agricultural payment for service programs:
Experimental evidence. Land Economics, 96(2), 244–264.
https://doi.org/10.3368/le.96.2.244
Hyer, K. E., Phillips, S. W., Ator, S. W., Moyer, D. L., Webber, J. S., Felver, R., Keisman, J. L.,
McDonnell, L. A., Murphy, R., Trentacoste, E. M., Zhang, Q., Dennison, W. C., Swanson,
S., Walsh, B., Hawkey, J., & Taillie, D. (2021). Nutrient trends and drivers in the
Chesapeake Bay watershed. U.S. Geological Survey Fact Sheet 2020-3069.
https://doi.org/10.3133/fs20203069
Ihde, T. F., Kaplan, I. C., Fulton, E. A., Gray, I. A., Hasan, M., Bruce, D., Slacum, W., & Townsend,
H. M. (2016). Design and parameterization of the Chesapeake Bay Atlantis Model: A
spatially explicit end-to-end ecosystem model. U.S. Department of Commerce NOAA
Technical Memorandum NMFS-F/SPO-166.
https://repository.library.noaa.gov/view/noaa/17357
Ihde, T. F., & Townsend, H. M. (2017). Accounting for multiple stressors influencing living
marine resources in a complex estuarine ecosystem using an Atlantis model. Ecological
Modelling, 365, 1–9. https://doi.org/10.1016/j.ecolmodel.2017.09.010
Jackson, C. R., Martin, J. K., Leigh, D. S., & West, L. T. (2005). A southeastern piedmont
watershed sediment budget: Evidence for a multi-millennial agricultural legacy. Journal
of Soil and Water Conservation, 60(6), 298–310.
Jackson-Smith, D. B., Halling, M., de la Hoz, E., McEvoy, J. P., & Horsburgh, J. S. (2010).
Measuring conservation program best management practice implementation and
maintenance at the watershed scale. Journal of Soil and Water Conservation, 65(6),
413–423. https://doi-org.proxy2.cl.msu.edu/10.2489/jswc.65.6.413
Jacobson, R. B., & Coleman, D. J. (1986). Stratigraphy and recent evolution of Maryland
Piedmont floodplains. American Journal of Science, 286(8), 617–637.
https://doi.org/10.2475/ajs.286.8.617
Jarvie, H. P., Johnson, L. T., Sharpley, A. N., Smith, D. R., Baker, D. B., Bruulsema, T. W., &
Confesor, R. (2017). Increased soluble phosphorus loads to Lake Erie: Unintended

101
 consequences of conservation practices? Journal of Environmental Quality, 46(1), 123–
132. https://doi.org/10.2134/jeq2016.07.0248
Jefferson, A. J., Bhaskar, A. S., Hopkins, K. G., Fanelli, R., Avellandeda, P. M., & McMillan, S. K.
(2017). Stormwater management network effectiveness and implications for urban
watershed function: A critical review. Hydrological Processes, 31(23), 4056–4080.
https://doi.org/10.1002/hyp.11347
Kahneman, D., Sibony, O., & Sunstein, C. R. (2021). Noise: A flaw in human judgement. Little,
Brown Spark.
Kast, J. B., Kalcic, M., Wilson, R., Jackson-Smith, D., Breyfogle, N., & Martin, J. (2021). Evaluating
the efficacy of targeting options for conservation practice adoption on watershed-scale
phosphorus reductions. Water Research, 201, 117375.
https://doi.org/10.1016/j.watres.2021.117375
Kaufman, Z., Abler, D., Shortle, J., Harper, J., Hamlett, J., & Feather, P. (2014). Agricultural costs
of the Chesapeake Bay total maximum daily load. Environmental Science & Technology,
48(24), 14131–14138. https://doi.org/10.1021/es502696t
Kaufman, D. E., Shenk, G., Bhatt, G., Asplen, K., Devereux, O., Rigelman, J., Ellis, H., Hobbs, B. F.,
Bosch, D. J., VanHoutven, G. L., McGarity, A., Linker, L. C., & Ball, W. P. (2021).
Supporting cost-effective watershed management strategies for Chesapeake Bay using a
modeling and optimization framework. Environmental Modeling & Software, 144,
105141. https://doi.org/10.1016/j.envsoft.2021.105141
Keisman, J., Blomquist, J., Böhlke, J. K., Davis-Martin, J., Dennison, W., Friedrichs, C., Murphy,
R., Phillips, S., Testa, J., Trentacoste, E., & Weller, D. (2018). Integrating recent findings
to explain water-quality change: Support for the mid-point assessment and beyond.
STAC Publication Number 18-005, Chesapeake Bay Program Scientific and Technical
Advisory Committee (STAC), Edgewater, MD. 27 pp.
http://www.chesapeake.org/pubs/394_Keisman2018.pdf
Keisman, J., Devereux, O. H., LaMotte, A. E., Sekellick, A., & Blomquist, J. (2018). Manure and
fertilizer inputs to land in the Chesapeake Bay watershed, 1950–2012. U.S. Geological
Survey Scientific Investigations Report 2018-5022. https://doi.org/10.3133/sir20185022
Kemp, W. M., Sampou, P., Caffrey, J., Mayer, M., Henriksen, K., & Boynton, W. R. (1990).
Ammonium recycling versus denitrification in Chesapeake Bay sediments. Limnology
and Oceanography, 35(7), 1545–1563. https://doi.org/10.4319/lo.1990.35.7.1545
Kemp, W. M., Batleson, R., Bergstrom, P., Carter, V., Gallegos, C. L., Hunley, W., Karrh, L., Koch,
E. W., Landwehr, J. M., Moore, K. A., Murray, L., Naylor, M., Rybicki, N. B., Stevenson, J.
C., & Wilson, D. J. (2004). Habitat requirements for submerged aquatic vegetation in
Chesapeake Bay: Water quality, light regime, and physical-chemical factors. Estuaries,
27, 363–377. https://doi.org/10.1007/BF02803529
Kemp, W. M., Boynton, W. R., Adolf, J. E., Boesch, D. F., Boicourt, W. C., Brush, G. Cornwell, J.
C., Fisher, T. R., Glibert, P. M., Hagy, J. D., Harding, L. W., Houde, E. D., Kimmel, D. G.,
Miller, W. D., Newell, R. I. E., Roman, M. R., Smith, E. M., & Stevenson, J. C. (2005).
Eutrophication of Chesapeake Bay: Historical trends and ecological interactions. Marine
Ecology Progress Series, 303, 1–29. https://doi.org/10.3354/meps303001

102
Khanna M., Yang, W., Farnsworth, R., & Önal, H. (2003). Cost-effective targeting of land
retirement to improve water quality with endogenous sediment deposition coefficients.
American Journal of Agricultural Economics, 85(3), 538–553.
https://doi.org/10.1111/1467-8276.t01-1-00454
Kleinman, P., Blunk, K. S., Bryant, R., Saporito, L., Beegle, D., Czymmek, K., Ketterings, Q., Sims,
T., Shortle, J., McGrath, J., Coale, F., Dubin, M., Dostie, D., Maguire, R., Meinen, R., Allen,
A., O’Neill, K., Garber, L., Davis, M., Clark, B., … Smith, M. (2012). Managing manure for
sustainable livestock production in the Chesapeake Bay watershed. Journal of Soil and
Water Conservation, 67(2), 54A–61A. https://doi.org/10.2489/jswc.67.2.54A
Kleinman, P. J. A., Fanelli, R. M., Hirsch, R. M., Buda, A. R., Easton, Z. M., Wainger, L. A., Brosch,
C., Lowenfish, M., Collick, A. S., Shirmohammadi, A., Boomer, K, Hubbart, J. A., Bryant, R.
B., & Shenk, G. W. (2019). Phosphorus and the Chesapeake Bay: Lingering issues and
emerging concerns for agriculture. Journal of Environmental Quality, 48(5), 1191–1203.
https://doi.org/10.2134/jeq2019.03.0112
Kleinman, P. J., A., Osmond, D. L., Christianson, L. E., Flaten, D. N., Ippolito, J. A., Jarvie, H. P.,
Kaye, J. P., King, K. W., Leytem, A. B., McGrath, J. M., Nelson, N. O., Shober, A. L., Smith,
D. R., Staver, K. W., & Sharpley, A. N. (2022). Addressing conservation practice
limitations and trade-offs for reducing phosphorus loss from agricultural fields.
Agricultural & Environmental Letters, 7(2), e20084. https://doi.org/10.1002/ael2.20084
Kleinman, P. J. A., Sharpley, A. N., Buda, A. R., Easton, Z. M., Lory, J. A., Osmond, D. L., Radcliffe,
D. E., Nelson, N. O., Veith, T. L., & Doody, D. G. (2017). The promise, practice, and state
of planning tools to assess site vulnerability to runoff phosphorus loss. Journal of
Environmental Quality, 46(6), 1243–1249. https://doi.org/10.2134/jeq2017.10.0395
Kleinman, P. J., A., Sharpley, A. N., McDowell, R. W., Flaten, D. N., Buda, A. R., Tao, L.,
Bergstrom, L., & Zhu, Q. (2011). Managing agricultural phosphorus for water quality
protection: Principles for progress. Plant and Soil, 349, 169–182.
https://doi.org/10.1007/s11104-011-0832-9
Kleinman, P. J. A., Sharpley, A. N., Withers, P. J. A., Bergström, L., Johnston, L. T., & Doody, D. G.
(2015). Implementing agricultural phosphorus science and management to combat
eutrophication. AMBIO, 44(Suppl 2), 297–310. https://doi.org/10.1007/s13280-015-
0631-2
Kleinman, P. J. A., Spiegal, S., Liu, J., Holly, M., Church, C., & Ramirez-Avila, J. (2020). Managing
animal manure to minimize phosphorus losses from land to water. In H. M. Waldrip, P.
H. Pagliari & Z. He (Eds.), Animal manure: Production, characteristics, environmental
concerns, and management, (pp. 201–228). American Society of Agronomy.
https://doi.org/10.2134/asaspecpub67.c12
Kling, C. L. (2013). State level efforts to regulate agricultural sources of water quality
impairment. Choices, 28(3), 1–4.
Kornis, M. S., Breitburg, D., Balouskus, R., Bilkovic, D. M., Davias, L. A., Giordano, S., Heggie, K.,
Hines, A. H., Jacobs, J. M., Jordan, T. E., King, R. S., Patrick, C. J., Seitz, R. D., Soulen, H.,
Targett, T. E., Weller, D. E., Whigham, D. F., & Uphoff, J. (2017). Linking the abundance
of estuarine fish and crustaceans in nearshore waters to shoreline hardening and land

103
 cover. Estuaries and Coasts, 40, 1464–1486. https://doi.org/10.1007/s12237-017-0213-
6
Koudahe, K., Allen, S. C., & Djaman, K. (2021). Critical review of the impact of cover crops on soil
properties. International Soil and Water Conservation Research, 10(3), 343–354.
https://doi.org/10.1016/j.iswcr.2022.03.003
Lefcheck, J. S., Orth, R. J., Dennison, W. C., Wilcox, D. J., Murphy, R. R., Keisman, J., Gurbisz, C.,
Hannam, M., Landry, J. B., Moore, K. A., Patrick, C. J., Testa, J., Weller, D. E., & Batiuk, R.
A. (2018). Long-term nutrient reductions lead to the unprecedented recovery of a
temperate coastal region. Proceedings of the National Academy of Sciences, 115(14),
3658–3662. https://doi.org/10.1073/pnas.1715798115
Lefcheck, J. S., Wilcox, J. S., Murphy, R. R., Marion, S. R., & Orth, R. J. (2017). Multiple stressors
threaten the imperiled coastal foundation species eelgrass (Zostera marina) in
Chesapeake Bay, USA. Global Change Biology, 23(9), 3474–3483.
https://doi.org/10.1111/gcb.13623
LeFevre, G. H., Paus, K. H., Natarajan, P., Gulliver, J. S., Novak, P. J., & Hozalski, R. M. (2015).
Review of dissolved pollutants in urban storm water and their removal and fate in
bioretention cells. Journal of Environmental Engineering, 141(1), 04014050.
https://doi.org/10.1061/(ASCE)EE.1943-7870.0000876
Li, C., Fletcher, T. D., Duncan, H. P., & Burns, M. J. (2017). Can stormwater control measures
restore altered urban flow regimes at the catchment scale? Journal of Hydrology, 549,
631–653. https://doi.org/10.1016/j.jhydrol.2017.03.037
Lichtenberg, E., & Smith-Ramírez, R. (2011). Slippage in conservation cost sharing. American
Journal of Agricultural Economics, 92(1), 113–129. https://doi.org/10.1093/ajae/aaq124
Lindsey, B. D., Phillips, S., Donnelly, C. A., Speiran, G. K., Plummer, N., Böhlke, J. K., Focazio, M.
J., Burton, W. C., & Busenberg, E. (2003). Residence times and nitrate transport in
ground water discharging to streams in the Chesapeake Bay watershed. U.S. Geological
Survey Water Resources Investigations Report 2003-4035.
https://doi.org/10.3133/wri034035
Linker, L. C., Batiuk, R. A., Shenk, G. W., & Cerco, C. F. (2013). Development of the Chesapeake
Bay watershed total maximum daily load allocation. JAWRA Journal of the American
Water Resources Association, 49(5), 986–1006. https://doi.org/10.1111/jawr.12105
Lintern, A., McPhillips, L., Winfrey, B., Duncan, J., & Grady, C. (2020). Best management
practices for diffuse nutrient pollution: Wicked problems across urban and agricultural
watersheds. Environmental Science & Technology, 54(15), 9159–9174.
https://doi.org/10.1021/acs.est.9b07511
Liu, H., Zhang, W., Irwin, E., Kast, J., Aloysius, N., Martin, J., & Kalcic M. (2020). Best
management practices and nutrient reduction: An integrated economics-hydrologic
model of the western Lake Erie Basin. Land Economics, 96(4), 510–530.
https://doi.org/10.3368/wple.96.4.510
Liu, Y., Engel, B. A., Flanagan, D. C., Gitau, M. W., McMillan, S. K., & Chaubey, I. (2017). A review
on effectiveness of best management practices in improving hydrology and water
quality: Needs and opportunities. Science of The Total Environment, 601–602, 580–593.
https://doi.org/10.1016/j.scitotenv.2017.05.212

104
Liu, Y., Engel, B. A., Flanagan, D. C., Gitau, M. W., McMillan, S. K., Chaubey, I., & Singh, S. (2018).
Modeling framework for representing long-term effectiveness of best management
practices in addressing hydrology and water quality problems: Framework development
and demonstration using a Bayesian method. Journal of Hydrology, 560, 530–545.
https://doi.org/10.1016/j.jhydrol.2018.03.053
Lyerly, C. M., Hernández Cordero, A. L., Foreman, K. L., Phillips, S. W., & Dennison, W. C. (2014).
New insights: Science-based evidence of water quality improvements, challenges, and
opportunities in the Chesapeake Bay.
https://ian.umces.edu/site/assets/files/11070/new-insights-science-based-evidence-of-
water-quality-improvements-challenges-and-opportunities-in-the-chesapeake.pdf
Ma, H., Townsend, H., Zhang, X., Sigrist, M., & Christensen, V. (2010). Using a fisheries
ecosystem model with a water quality model to explore trophic and habitat impacts on
a fisheries stock: A case study of the blue crab population in the Chesapeake Bay.
Ecological Modelling, 221(7), 997–1004.
https://doi.org/10.1016/j.ecolmodel.2009.01.026
Macías-Tapia, A., Mulholland, M. R., Selden, C. R., Loftis, J. D., & Bernhardt, P. W. (2021). Effects
of tidal flooding on estuarine biogeochemistry: Quantifying flood-driven nitrogen inputs
in an urban, lower Chesapeake Bay sub-tributary. Water Research, 201, 117329.
https://doi.org/10.1016/j.watres.2021.117329
Macrae, M., Jarvie, H., Brouwer, R., Gunn, G., Reid, K., Joosse, P., King, K. Kleinman, P., Smith,
D., Williams, M., & Zwonitzer, M. (2021). One size does not fit all: Toward regional
conservation practice guidance to reduce phosphorus loss risk in the Lake Erie
watershed. Journal of Environmental Quality, 50(3), 529–546.
https://doi.org/10.1002/jeq2.20218
Maille, P., Collins, A. R., & Gillies, N. (2009). Performance-based payments for water quality:
Experiences from a field experiment. Journal of Soil and Water Conservation, 64(3),
85A–87A. https://doi.org/10.2489/jswc.64.3.85A
Mallakpour, I., & Villarini, G. (2017). Analysis of changes in the magnitude, frequency, and
seasonality of heavy precipitation over the contiguous USA. Theoretical and Applied
Climatology, 130(1), 345–363. https://doi.org/10.1007/s00704-016-1881-z
Marchau, V. A. W. J., Walker, W. E., Bloemen, P. J. T. M., & Popper, S. W. (Eds.). (2019). Decision
making under deep uncertainty. Springer.
Maryland Department of the Environment (MDE). 2019. Maryland’s Phase III Watershed
Implementation Plan to Restore Chesapeake Bay by 2025.
https://mde.maryland.gov/programs/Water/TMDL/TMDLImplementation/Pages/Phase
3WIP.aspx
Mason, C. A., & Soroka, A. M. (2022). Nitrogen, phosphorus, and suspended-sediment loads and
trends measured at the Chesapeake Bay River Input Monitoring stations: Water years
1985–2021. U.S. Geological Survey data release. https://doi.org/10.5066/P90CZJ1Y
Meals, D. W., Dressing, S. A., & Davenport, T. E. (2010). Lag time in water quality response to
best management practices: A review. Journal of Environmental Quality, 39(1), 85–96.
https://doi.org/10.2134/jeq2009.0108

105
Meinikmann, K., Hupfer, M, & Lewandowski, J. (2015). Phosphorus in groundwater discharge—
A potential source for lake eutrophication. Journal of Hydrology, 524, 214–226.
https://doi.org/10.1016/j.jhydrol.2015.02.031
Michalak, A. M., Anderson, E. J., Beletsky, D., Boland, S., Bosch, N. S., Bridgemen, T. B., Chaffin,
J. D., Cho, K., Confesor, R. Daloğlu, I., DePinto, J. V., Evans, M. A., Fahnenstiel, G. L., He,
L., Ho, J. C., Jenkins, L., Johengen, T. H., Kuo, K. C., LaPorte, E., … Zagorski, M. A. (2013).
Record-setting algal bloom in Lake Erie caused by agricultural and meteorological trends
consistent with expected future conditions. Proceedings of the National Academy of
Sciences, 110(16), 6448-6452. https://doi.org/10.1073/pnas.1216006110
Miller, A., Baker, M., Boomer, K., Merritts, D., Prestegaard, K., & Smith, S. (2019). Legacy
sediment, riparian corridors, and total maximum daily loads. STAC Publication Number
19-001, Chesapeake Bay Program Scientific and Technical Advisory Committee (STAC),
Edgewater, MD. 64 pp. http://www.chesapeake.org/pubs/399_Miller2019.pdf
Monaco, M. E., & Ulanowicz, R. E. (1997). Comparative ecosystem trophic structure of three
U.S. mid-Atlantic estuaries. Marine Ecology Progress Series, 161, 239–254.
https://doi.org/10.3354/meps161239
Monaco, M. E., Weisberg, S. B., & Lowery, T. A. (1998). Summer habitat affinities of estuarine
fish in U.S. mid-Atlantic coastal systems. Fisheries Management and Ecology, 5, 161–
171.
Moyer, D. L., & Blomquist, J. D. (2017). Summary of nitrogen, phosphorus, suspended-sediment
loads and trends measured at the Chesapeake Bay Nontidal Network stations: Water
year 2016 update. https://www.usgs.gov/media/files/summary-nitrogen-phosphorus-
and-suspended-sediment-loads
Murphy, R. R., Keisman, J., Harcum, J., Karrh, R. R., Lane, M., Perry, E. S., & and Zhang, Q.
(2022). Nutrient improvements in Chesapeake Bay: Direct effect of load reductions and
implications for coastal management. Environmental Science & Technology, 56(1), 260–
270. https://doi.org/10.1021/acs.est.1c05388
Najjar, R. G., Pyke, C. R., Adams, M. B., Breitburg, D., Hershner, C., Kemp, M., Howarth, R.,
Mulholland, M. R., Paolisso, M., Secor, D., Sellner, K., Wardrop, D., & Wood, R. (2010).
Potential climate-change impacts on the Chesapeake Bay. Estuarine, Coastal and Shelf
Science, 86(1), 1–20. https://doi.org/10.1016/j.ecss.2009.09.026
National Research Council (NRC). (2001). Assessing the TMDL approach to water quality
management. National Academies Press.
National Research Council (NRC). (2011). Achieving nutrient and sediment reduction goals in the
Chesapeake Bay: An evaluation of program strategies and implementation. National
Academies Press.
Newell, R. I. E. (1988). Ecological changes in Chesapeake Bay: Are they the result of
overharvesting the American oyster, Crassostrea virginica." In M. P. Lynch & E. C. Krome
(Eds.), Understanding the estuary: Advances in Chesapeake Bay research (pp. 536–546).
Chesapeake Research Consortium Publication 129, CBP/TRS 24/88. http://www.oyster-
restoration.org/wp-content/uploads/2012/06/Newell-1988-filtering.pdf

106
Ni, W., Li, M., & Testa, J. M. (2020). Discerning effects of warming, sea level rise and nutrient
management on long-term hypoxia trends in Chesapeake Bay. Science of The Total
Environment, 737, 139717. https://doi.org/10.1016/j.scitotenv.2020.139717
Niklitschek, E. J., & Secor, D. H. (2005). Modeling spatial and temporal variation of suitable
nursery habitats for Atlantic sturgeon in the Chesapeake Bay. Estuarine, Coastal and
Shelf Science, 64(1), 135–148. https://doi.org/10.1016/j.ecss.2005.02.012
Noe, G. B., Cashman, M., Skalak, K., Gellis, A., Hopkins, K., Moyer, D., Webber, J., Benthem, A.,
Maloney, K., Brakebill, J., Sekellick, A., Langland, M., Zhang, Q., Shenk, G., Keisman, J., &
Hupp, C. (2020). Sediment dynamics and implications for management: State of the
science from long-term research in the Chesapeake Bay watershed, USA. Wiley
Interdisciplinary Reviews: Water, 7(4), e1454. https://doi.org/10.1002/wat2.1454
Noe, G. B., Hopkins, K. G., Claggett, P. R., Schenk, E. R., Metes, M. J., Ahmed, L., Doody, T. R., &
Hupp, C. R. (2022). Streambank and floodplain geomorphic change and contribution to
watershed material budgets. Environmental Research Letters, 17, 064015.
https://doi.org/10.1088/1748-9326/ac6e47
Opalinski, N., Schultz, D., Veith, T. L., Royer, M., & Preisendanz, H. E. (2022). Meeting the
moment: Leveraging temporal inequality for temporal targeting to achieve water-quality
load-reduction goals. Water, 14(7), 1003. https://doi.org/10.3390/w14071003
Osmond, D. L., Hoag, D. L. K., Luloff, A. E., Meals, D. W., & Neas, K. (2015). Farmers’ use of
nutrient management: Lessons from watershed case studies. Journal of Environmental
Quality, 44(2), 382–390. https://doi.org/10.2134/jeq2014.02.0091
Osmond, D., Meals, D., Hoag, D., Arabi, M., Luloff, A., Jennings, G., McFarland, M., Spooner, J.,
Sharpley, A., & Line, D. (2012). Improving conservation practices programming to
protect water quality in agricultural watersheds: Lessons learned from the National
Institute of Food and Agriculture—Conservation Effects Assessment Project. Journal of
Soil and Water Conservation, 67(5), 122A–127A. https://doi.org/10.2489/jswc.67.5.122a
O’Sullivan, A. (2021). Expanding regulatory sandboxing to fast-track innovation. The James
Madison Institute. https://jamesmadison.org/wp-
content/uploads/2021/01/Regulatory_Sandbox_1.27.21.pdf
Palm-Forster, L. H., Swinton, S. M., Lupi, F., & Shupp, R. S. (2016). Too burdensome to bid:
Transaction costs and pay-for-performance conservation. American Journal of
Agricultural Economics, 98(5), 1314–1333. https://doi.org/10.1093/ajae/aaw071
Pannell, D. J., & Claassen, R. (2020). The roles of adoption and behavior change in agricultural
policy. Applied Economic Perspectives and Policy, 42(1), 31–41.
https://doi.org/10.1002/aepp.13009
Parker, M., & Bricker, S. (2020). Sustainable oyster aquaculture, water quality improvement,
and ecosystem service value potential in Maryland Chesapeake Bay. Journal of Shellfish
Research, 39(2), 269–281. https://doi.org/10.2983/035.039.0208
Patrick, C. J., Wilcox, D. J., Whiting, J. R., Kenne, A. K., & Smith, E. R. (2022). 2021 distribution of
submerged aquatic vegetation in Chesapeake Bay and coastal bays. Virginia Institute of
Marine Science.
https://www.vims.edu/research/units/programs/sav/reports/2021/index.php

107
Patterson, J. J., Smith, C., & Bellamy, J. (2013). Understanding enabling capacities for managing
the ‘wicked problem’ of nonpoint source water pollution in catchments: A conceptual
framework. Journal of Environmental Management, 128, 441–452.
https://doi.org/10.1016/j.jenvman.2013.05.033
Pearce, A., & Mcguire, R. (2020). The state of phosphorus balance on 58 Virginia dairy farms.
Journal of Environmental Quality, 49(2), 324–334. https://doi.org/10.1002/jeq2.20054
Phillips, S. W., & Lindsey, B. D. (2003). The influence of ground water on nitrogen delivery to the
Chesapeake Bay. U.S. Geological Survey FS-091-03.
https://pubs.usgs.gov/fs/2003/0091/report.pdf
Pineiro, V., Arias, J., Elverdin, P., Ibanez, A. M., Opazo, C. M., Prager, S., & Terero, M. (2021).
Achieving sustainable agricultural practices: From incentives to adoption and outcomes.
International Food Policy Research Institute Policy Brief.
https://cdm15738.contentdm.oclc.org/utils/getfile/collection/p15738coll2/id/134262/fi
lename/134472.pdf
Pizzuto, J., Schenk, E. R., Hupp, C. R., Gellis, A., Noe, G., Williamson, E., Karwan, D. L., O’Neal,
M., Maquard, J., Aalto, R, & Newbold, D. (2014). Characteristic length scales and time-
averaged transport velocities of suspended sediment in the mid-Atlantic region, USA.
Water Resources Research, 50(2), 790–805. https://doi.org/10.1002/2013WR014485
Price, E., Flemming, T. H., & Wainger, L. (2021). Cost analysis of stormwater and agricultural
practices for reducing nitrogen and phosphorus runoff in Maryland. University of
Maryland Center for Environmental Science Technical Report #TS-772-21.
https://doi.org/10.13140/RG.2.2.28896.74246/1
Prokopy, L. S., Floress, K., Arbuckle, J. G., Church, S. P., Eanes, F. R., Gao, Y., Gramig, B. M.,
Ranjan, P., & Singh, A. S. (2019). Adoption of agricultural conservation practices in the
United States: Evidence from 35 years of quantitative literature. Journal of Soil and
Water Conservation, 74(5), 520–534. https://doi.org/10.2489/jswc.74.5.520
Qui, Z. (2009). Assessing critical source areas in watersheds for conservation buffer planning
and riparian restoration. Environmental Management, 44, 968–980.
https://doi.org/10.1007/s00267-009-9380-y
Ranjan, P., Church, S. P., Floress, K., & Prokopy, L. S. (2019). Synthesizing conservation
motivations and barriers: What have we learned from qualitative studies of farmers’
behaviors in the United States? Society and Natural Resources, 32(11), 1171–1199.
https://doi.org/10.1080/08941920.2019.1648710
Rao, N. S., Easton, Z. M., Schneiderman, E. M., Zion, M. S., Lee, D. R., & Steenhuis, T.S. (2009).
Modeling watershed-scale effectiveness of agricultural best management practices to
reduce phosphorus loading. Journal of Environmental Management, 90(3), 1385–1395.
https://doi.org/10.1016/j.jenvman.2008.08.011
Reimer, A. P., & Prokopy, L. S. (2014). Farmer participation in U.S. farm bill conservation
programs. Environmental Management, 53(2), 318–332.
https://doi.org/10.1007/s00267-013-0184-8
Reimer, A., Thompson, A., Prokopy, L. S., Arbuckle, J. G., Genskow, K., Jackson-Smith, D., Lynne,
G., McCann, L., Morton, L. W., & Nowak, P. (2014). People, place, behavior, and context:
A research agenda for expanding our understanding of what motivates farmers'

108
 conservation behaviors. Journal of Soil and Water Conservation, 69(2), 57A–61A.
https://doi.org/10.2489/jswc.69.2.57A
Ribaudo, M. (2001). Non-point source pollution control policy in the USA. In J.S. Shortle & D.
Abler (Eds.), Environmental policies for agricultural pollution control (pp. 123–150). CABI
Publishing.
Ribaudo, M. (2015). The limits of voluntary conservation programs. Choices, 30(2), 1–5.
Ribaudo, M., & Shortle, J. (2019). Reflections on 40 years of applied economics research on
agriculture and water quality. Agricultural and Resource Economics Review, 48(3), 519–
530. https://doi.org/10.1017/age.2019.32
Roland, V. L., Garcia, A. M., Saad, D. A., Ator, S. W., Robertson, D., & Schwarz, G. (2022).
Quantifying regional effects of best management practices on nutrient losses from
agricultural lands. Journal of Soil and Water Conservation, 77(1), 15–29.
https://doi.org/10.2489/jswc.2022.00162
Rose, K., Monaco, M. E., Ihde, T., Hubbart, J., Smith, E., Stauffer, J., & Havens, K. J. (2023).
Proposed framework for analyzing water quality and habitat effects on the living
resources of Chesapeake Bay. STAC Publication Number 23-005, Chesapeake Bay
Program Scientific and Technical Advisory Committee (STAC), Edgewater, MD. XX pp.
Rose, K. A., Sable, S., DeAngelis, D. L., Yurek, S., Trexler, J. C., Graf, W., & Reed, D. J. (2015).
Proposed best modeling practices for assessing the effects of ecosystem restoration on
fish. Ecological Modelling, 300, 12–29. https://doi.org/10.1016/j.ecolmodel.2014.12.020
Royer, M., Shortle, J., & Cook, A. (2016). An analysis of the Pennsylvania farm conservation
practices inventory for purposes of reporting practices to the Chesapeake Bay Program.
Penn State University College of Agricultural Sciences.
https://files.dep.state.pa.us/water/ChesapeakeBayOffice/Farm%20Survey%20Report%2
0Final%20121516.pdf
Runge, M. C., Converse, S. J., & Lyons, J. E. (2011). Which uncertainty? Using expert elicitation
and expected value of information to design an adaptive program. Biological
Conservation, 144(4), 1214–1223. https://doi.org/10.1016/j.biocon.2010.12.020
Runge, M. C., Converse, S. J., Lyons, J. E., & Smith, D. R. (Eds.). (2020). Structured decision
making: Case studies in natural resource management. Johns Hopkins Press.
Ryberg, K. R., Blomquist, J. D., Sprague, L. A., Sekellick, A. J., & Keisman, J. (2018). Modeling
drivers of phosphorus loads in Chesapeake Bay tributaries and inferences about long-
term change. Science of The Total Environment, 616-617, 1423–1430.
https://doi.org/10.1016/j.scitotenv.2017.10.173
Sabo, R. D., Sullivan, B., Wu, C., Trentacoste, E., Zhang, Q., Shenk, G. W., Bhatt, G., & Linker, L.
C. (2022). Major point and nonpoint sources of nutrient pollution to surface water have
declined throughout the Chesapeake Bay watershed. Environmental Research
Communications, 4(4), 045012. https://doi.org/10.1088/2515-7620/ac5db6
Saha, A., Saha, G. K., Cibin, R., Spiegal, S., Kleinman, P. J. A., Veith, T. L., White, C. M., Drohan, P.
J., & Tsegaye, T. (2022). Evaluating water quality benefits of manureshed management
in the Susquehanna River Basin. Journal of Environmental Quality, forthcoming.
https://doi.org/10.1002/jeq2.20429

109
Sanford, W. E., & Pope, J. P. (2013). Quantifying groundwater's role in delaying improvements
to Chesapeake Bay water quality. Environmental Science & Technology, 47(23), 13330–
13338. https://doi.org/10.1021/es401334k
Schlenger, A., North, E. W., Li, Y., Li, M., & Kemp, W. M. (2022). Applying a three-dimensional
habitat volume model to estimate sensitivity of Chesapeake Bay living resources to
environmental change: A proof-of-concept exercise. Estuaries and Coasts, 45, 393–412.
https://doi.org/10.1007/s12237-021-00967-6
Scully, M. E. (2010). The importance of climate variability to wind-driven modulation of hypoxia
in Chesapeake Bay. Journal of Physical Oceanography, 40(6), 1435–1440.
https://doi.org/10.1175/2010JPO4321.1
Sebilo, M., Mayer, B., Nicolardot, B., Pinay, G., & Mariotti, A. (2013). Long-term fate of nitrate
fertilizer in agricultural soils. Proceedings of the National Academy of Sciences, 110(45),
18185–18189. https://doi.org/10.1073/pnas.1305372110
Selbig, W. R., & Bannerman, R. T. (2008). A comparison of runoff quantity and quality from two
small basins undergoing implementation of conventional and low-impact-development
(LID) strategies: Cross Plains, Wisconsin, water years 1999–2005. U.S. Geological Survey
Scientific Investigations Report 2008–5008. https://doi.org/10.3133/sir20085008
Shabman, L., Reckhow, K., Beck, M. B., Benaman, J., Shapra, S., Freedman, P., Nellor, M., Rudek,
J., Schwer, D., Stiles, T., & Stow, C. (2007). Adaptive implementation of water quality
improvement plans: Opportunities and challenges. Nicholas Institute for Environmental
Policy Solutions Report #NI-R-07-03, Duke University.
Sharpley, A., Jarvie, H. P., Buda, A., May, L., Spears, B., & Kleinman, P. (2013). Phosphorus
legacy: Overcoming the effects of past management practices to mitigate future water
quality impairment. Journal of Environmental Quality, 42(5), 1308–1326.
https://doi.org/10.2134/jeq2013.03.0098
Shenk, G., Zhang, Q., Bhatt, G., Bertani, I., Mason, C., Moyer, D., & Tango, P. (2022). Proposed
integrated watershed TMDL indicator and dashboard products. Presentation to
Chesapeake Bay Program Water Quality Goal Implementation Team, November 14.
https://d18lev1ok5leia.cloudfront.net/chesapeakebay/documents/2022-11-14-
Monitored-TMDL-Indicator-gshenk-WQGIT.pdf
Shenk, G., Blatt, G., Tian R., Cerco, C. F., Bertani, I., & Linker, L. C. (2021). Modeling climate
change effects on Chesapeake Bay water quality standards and development of 2025
planning targets to address climate change. CBPO Publication Number 328-21.
https://cast-
content.chesapeakebay.net/documents/P6ModelDocumentation%2FClimateChangeDoc
umentation.pdf
Shortle, J. S., & Horan, R. D. (2017). Nutrient pollution: A wicked challenge for economic
instruments. Water Economics and Policy, 3(2), 1650033.
https://doi.org/10.1142/S2382624X16500338
Shortle, J., Ollikainen, M., & Iho, A. (2021). Water quality and agriculture: Economics and policy
for nonpoint source water pollution. Palgrave MacMillan.

110
Shortle, J. S., Ribaudo, M., Horan, R. D., & Blandford, D. (2012). Reforming agricultural nonpoint
pollution policy in an increasingly budget-constrained environment. Environmental
Science & Technology, 46(3), 1316–1325. https://doi.org/10.1021/es2020499
Sinha, E., Michalak, A. M., & Balaji, V. (2017). Eutrophication will increase during the 21 st
century as a result of precipitation changes. Science, 357(6349), 405–408.
https://doi.org/10.1126/science.aan2409
Smith, D. R., Macrae, M. L., Kleinman, P. J. A., Jarvie, H. P., King, K. W., & Bryant, R. B. (2019).
The latitudes, attitudes, and platitudes of watershed phosphorus management in North
America. Journal of Environmental Quality, 48(5), 1176–1190.
https://doi.org/10.2134/jeq2019.03.0136
Smith, S. M. C., & Wilcock, P. R. (2015). Upland sediment supply and its relation to watershed
sediment delivery in the contemporary mid-Atlantic Piedmont (U.S.A). Geomorphology,
232, 33–46. https://doi.org/10.1016/j.geomorph.2014.12.036
Spiegal, S., Kleinman, P. J. A., Endale, D. M., Bryant, R. B., Dell, C., Goslee, S., Meinen, R. J.,
Flynn, K. C., Baker, J. M., Browning, D. M., McCarty, G., Bittman, S., Carter, J., Cavigelli,
M., Duncan, E., Gowda, P., Li, X., Ponce-Campos, G. E., Cibin, R., Silveira, M. L., … Yang Q.
(2020). Manuresheds: Advancing nutrient recycling in US agriculture. Agricultural
Systems, 182, 102813. https://doi.org/10.1016/j.agsy.2020.102813
Sprague, L. A., & Gronberg, J. A. M. (2012). Relating management practices and nutrient export
in agricultural watersheds of the United States. Journal of Environmental Quality, 41(6),
1939–1950. https://doi.org/10.2134/jeq2012.0073
STAC (Chesapeake Bay Program Scientific and Technical Advisory Committee). (2013).
Incorporating lag-times into the Chesapeake Bay Program. STAC Publication Number 13-
004, Chesapeake Bay Program Scientific and Technical Advisory Committee (STAC),
Edgewater, MD. 66 pp. https://www.chesapeake.org/pubs/305_Hirsch2013.pdf
Stackpoole, S. M., Stets, E. G., & Sprague, L. A. (2019). Variable impacts of contemporary versus
legacy agricultural phosphorus on US river water quality. Proceedings of the National
Academy of Sciences, 116(41), 20562–20567. https://doi.org/10.1073/pnas.1903226116
Statham, P. J. (2012). Nutrients in estuaries—An overview and the potential impacts of climate
change. Science of The total environment, 434, 213–227.
https://doi.org/10.1016/j.scitotenv.2011.09.088
Staver, K., Kleinman, P., Ator, S., Buda, A., Ketterings, Q., Sims, J. T., & Meisinger, J. (2014). A
review of agricultural P-dynamics in the Chesapeake Bay Watershed Model. STAC
Publication Number 14-005, Chesapeake Bay Program Scientific and Technical Advisory
Committee (STAC), Edgewater, MD. 20 pp. https://www.chesapeake.org/stac/wp-
content/uploads/2021/01/Final-STAC-Phosphorus-Review-Report-8.26.14.pdf
Staver, K. W. and Brinsfield, R. B. (1998). Using cereal grain winter cover crops to reduce
groundwater nitrate contamination in the mid-Atlantic coastal plain. Journal of Soil and
Water Conservation, 53(3), 230-240.
Stephenson, K., Hershner, C., Benham, B., Easton, Z., Hanson, J., Julius, S., & Hinrichs, E. (2018).
Consideration of BMP performance uncertainty in Chesapeake Bay program
implementation. STAC Publication Number 18-003, Chesapeake Bay Program Scientific

111
 and Technical Advisory Committee (STAC), Edgewater, MD. 33 pp.
http://www.chesapeake.org/pubs/390_Stephenson2018.pdf
Stephenson, K., & Shabman, L. (2017a). Nutrient assimilation services for water quality trading
programs: A comparative analysis with nonpoint source credits. Coastal Management,
45(1), 24–43. https://doi.org/10.1080/08920753.2017.1237240
Stephenson, K., & Shabman, L. (2017b). Where did the agricultural nonpoint source trades go?
Lessons from Virginia water quality trading programs. JAWRA Journal of the American
Water Resources Association, 53(5), 1178–1194. https://doi-
org.proxy2.cl.msu.edu/10.1111/1752-1688.12565
Stephenson, K., Shabman, L., Shortle, J., & Easton, Z. (2022). Confronting our agricultural
nonpoint source control policy problem. JAWRA Journal of the American Water
Resources Association, 58(4), 496–501. https://doi.org/10.1111/1752-1688.13010
Talberth, J., Selman, M., Walker, S., & Gray, E. (2015). Pay for performance: Optimizing public
investments in agricultural best management practices in the Chesapeake Bay
watershed. Ecological Economics, 118, 252–261.
https://doi.org/10.1016/j.ecolecon.2015.07.033
Tango, P. J., & Batiuk, R. A. (2013). Deriving Chesapeake Bay water quality standards. JAWRA
Journal of the American Water Resources Association, 49(5), 1007–1024.
https://doi.org/10.1111/jawr.12108
Testa, J. M., Boynton, W. R., Hodgkins, C. L., Moore, A. L., Bailey, E. M., & Rambo, J. (2022).
Biogeochemical states, rates, and exchanges exhibit linear responses to large nutrient
load reductions in a shallow, eutrophic urban estuary. Limnology and
Oceanography, 67(4), 739–752. https://doi.org/10.1002/lno.12037
Testa, J. M., Dennison, W. C., Ball, W. P., Boomer, K., Gibson, D. M., Linker, L., Runge, M. C., &
Sanford, L. (2023). Knowledge gaps, uncertainties, and opportunities regarding the
response of the Chesapeake Bay estuary to proposed TMDLs. STAC Publication Number
23-004, Chesapeake Bay Program Scientific and Technical Advisory Committee (STAC),
Edgewater, MD. XX pp.
Testa, J .M., Harris, L, Boynton, W. R, Bukaveckas, P., Eshleman, K., Murphy, R., Keisman, J.,
Blomquist, J., Lane, M., Karrh, R., & Lyubchich, V. (2018). Progress toward the
restoration of Chesapeake Bay in time and space. Chesapeake Biological Laboratory
Technical Series TS-716-18.
Testa, J. M., & Kemp, W. M. (2012). Hypoxia-induced shifts in nitrogen and phosphorus cycling
in Chesapeake Bay. Limnology and Oceanography, 57(3), 835–850.
https://doi.org/10.4319/lo.2012.57.3.0835
Testa, J. M., Lyubchich, V., & Zhang, Q. (2019). Patterns and trends in Secchi disk depth over
three decades in the Chesapeake Bay estuarine complex. Estuaries and Coasts, 42, 927–
943. https://doi.org/10.1007/s12237-019-00547-9
Thomason, W., Duiker, S., Ganoe, K., Gates, D., McCollum, B., & Reiter, M. (2016). Conservation
tillage practices for use in phase 6.0 of the Chesapeake Bay Program Watershed Model.
CBP/TRS-308-16.
https://d18lev1ok5leia.cloudfront.net/chesapeakebay/documents/CT_6.0_Conservatio
n_Tillage_EP_Revised_Full_Report_12-14-16.2_FINAL_NEW_TEMPLATE.pdf

112
Tomer, M. D., & Locke, M. A. (2011). The challenge of documenting water quality benefits of
conservation practices: A review of USDA-ARS's conservation effects assessment project
watershed studies. Water Science & Technology, 64(1), 300–310.
https://doi.org/10.2166/wst.2011.555
Townsend, H. (2014). Comparing and coupling a water quality and a fisheries ecosystem model
of the Chesapeake Bay for the exploratory assessment of resource management
strategies. ICES Journal of Marine Science, 71(3), 703–712.
https://doi.org/10.1093/icesjms/fst060
U.S. Environmental Protection Agency (USEPA). (2003a). Ambient water quality criteria for
dissolved oxygen, water clarity and chlorophyll-a for the Chesapeake Bay and its tidal
tributaries. EPA 903-R-03-002.
https://nepis.epa.gov/Exe/ZyPDF.cgi/P100YKPQ.PDF?Dockey=P100YKPQ.PDF
U.S. Environmental Protection Agency (USEPA). (2003b). Technical support document for
identification of Chesapeake Bay designated uses and attainability. EPA 903-R-03-004.
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1001WE5.txt
U.S. Environmental Protection Agency (USEPA). (2004). Ambient water quality criteria for
dissolved oxygen, water clarity and chlorophyll-a for the Chesapeake Bay and its tidal
tributaries: 2004 addendum. EPA 903-R-04-005.
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1001WB8.txt
U.S. Environmental Protection Agency (USEPA). (2010). Chesapeake Bay total maximum daily
load for nitrogen, phosphorus and sediment. https://www.epa.gov/chesapeake-bay-
tmdl/chesapeake-bay-tmdl-document
U.S. Environmental Protection Agency (USEPA). (2017). Ambient water quality criteria for
dissolved oxygen, water clarity and chlorophyll-a for the Chesapeake Bay and its tidal
tributaries: 2017 technical addendum. EPA 903-R-17-002.
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100YL1C.txt
Van Meter, K. J., Basu, N. B., & Van Cappellen, P. (2017). Two centuries of nitrogen dynamics:
Legacy sources and sinks in the Mississippi and Susquehanna River Basins. Global
Biogeochemical Cycles, 31(1), 2–23. https://doi.org/10.1002/2016GB005498
Van Meter, K. J., Basu, N. B., Veenstra, J. J., & Burras, C. L. (2016). The nitrogen legacy: Emerging
evidence of nitrogen accumulation in anthropogenic landscapes. Environmental
Research Letters, 11(3), 035014. https://doi.org/10.1088/1748-9326/11/3/035014
Van Meter, K., McLeod, M. M., Liu, J., Tenkouano, G. T., Hall, R. I, Van Cappellen, P., & Basu, N.
B. (2021). Beyond the mass balance: Watershed phosphorus legacies and the evolution
of the current water quality policy challenge. Water Resources Research, 57(10),
e2020WR029316. https://doi.org/10.1029/2020WR029316
Vasslides, J. M., de Mutsert, K., Christensen, V., & Townsend, H. (2017). Using the Ecopath with
Ecosim modeling approach to understand the effects of watershed-based management
actions in coastal ecosystems. Coastal Management, 45(1), 44–55.
https://doi.org/10.1080/08920753.2017.1237241
Virginia Institute of Marine Science (VIMS). (n.d.-a). Indices of abundance.
https://www.vims.edu/research/departments/fisheries/programs/juvenile_surveys/dat
a_products/indices/index.php

113
Virginia Institute of Marine Science (VIMS). (n.d.-b). SAV monitoring and restoration.
https://www.vims.edu/research/units/programs/sav/index.php
Wagena, M. B., & Easton, Z. M. (2018). Agricultural conservation practices can help mitigate the
impact of climate change. Science of The Total Environment, 635, 132–143.
https://doi.org/10.1016/j.scitotenv.2018.04.110.
Wang, P., Wang, H., Linker, L., & Tian, R. (2016). Effects of cross-channel bathymetry and wind
direction on destratification and hypoxia reduction in the Chesapeake Bay. Estuarine,
Coastal and Shelf Science, 178, 168–188. https://doi.org/10.1016/j.ecss.2016.05.009
Walter, R. C., & Merritts, D. J. (2008). Natural streams and the legacy of water-powered mills.
Science, 319(5861), 299–304. https://doi.org/10.1126/science.1151716
Ward, N. D., Megonigal, J. P., Bond-Lamberty, B., Bailey, V. L., Butman, D., Canuel, E. A.,
Diefenderfer, H., Ganju, N. K., Goni, M. A., Graham, E. B., Hopkinson, C. S., Khangaonkar,
T., Langley, J. A., McDowell, N. G., Myers-Pigg, A. N., Neumann, R. B., Osburn, C. L.,
Price, R. M., Rowland, J., … Windham-Myers, L. (2020). Representing the function and
sensitivity of coastal interfaces in Earth system models. Nature Communications, 11,
2458. https://doi.org/10.1038/s41467-020-16236-2
White, M. J., Storm, D. E., Busteed, P. R., Stoodley, S. H., & Phillips, S. J. (2009). Evaluating
nonpoint source critical source area contributions at the watershed scale. Journal of
Environmental Quality, 38(4), 1654–1663. https://doi.org/10.2134/jeq2008.0375
Wiering, M., Liefferink, D., Boezeman, D., Kaufmann, M., Crabbé, A., & Kurstjens, N. (2020). The
wicked problem the water framework directive cannot solve. The governance approach
in dealing with pollution of nutrients in surface water in the Netherlands, Flanders,
Lower Saxony, Denmark and Ireland. Water, 12, 1240.
https://doi.org/10.3390/w12041240
Wilson, R. S., Howard, G., & Burnett, E. A. (2014). Improving nutrient management practices in
agriculture: The role of risk-based beliefs in understanding farmers’ attitudes toward
taking additional action. Water Resources Research, 50(8), 6735–6746.
https://doi.org/10.1002/2013WR015200
Woodland, R. J., Buchheister, A., Latour, R. J., Lozano, C., Houde, E., Sweetman, C. J., Fabrizio,
M. C., & Tuckey, T. D. (2021). Environmental drivers of forage fishes and benthic
invertebrates at multiple spatial scales in a large temperate estuary. Estuaries and
Coasts, 44, 921–938. https://doi.org/10.1007/s12237-020-00835-9
Xu, Y., Bosch, D. J., Wagena, M. B., Collick, A. S., & Easton, Z. M. (2019). Meeting water quality
goals by spatial targeting of best management practices under climate change. Journal
of Environmental Management, 63, 173–184. https://doi.org/10.1007/s00267-018-
01133-8
Yagow, G., Collick, A., Ribaudo, M., Thomason, W., & Veith, T. (2016). Scientific and technical
advisory committee review of nutrient input estimation for the Chesapeake Bay
Watershed Model. STAC Publication Number 16-005, Chesapeake Bay Program Scientific
and Technical Advisory Committee (STAC), Edgewater, MD. 49 pp.
http://www.chesapeake.org/pubs/358_Yagow2016.pdf

114
Yang W., & Weersink, A. (2005). Cost-effective targeting of riparian buffers. Canadian Journal of
Agricultural Economics, 52(1), 17–34. https://doi.org/10.1111/j.1744-
7976.2004.tb00092.x
Zhang, Q., Murphy, R. R., Tian, R., Forsyth, M. K., Trentacoste, E. M., Keisman, J, & Tango, P. J.
(2018). Chesapeake Bay’s water quality condition has been recovering: Insights from a
multimetric indicator assessment of thirty years of tidal monitoring data. Science of The
Total Environment, 637–638, 1617–1625.
https://doi.org/10.1016/j.scitotenv.2018.05.025
Zhang, Q., Tango, P. J., Murphy, R. R., Forsyth, M. K., Tian, R., Keisman, J., & Trentacoste, E. M.
(2018). Chesapeake Bay dissolved oxygen criterion attainment deficit: Three decades of
temporal and spatial patterns. Frontiers in Marine Science, 5, 422.
https://doi.org/10.3389/fmars.2018.00422

115