United States
Environmental Protection
     Waquoit Bay Watershed
     Ecological Risk
     Assessment: The effect of
     land-derived nitrogen loads
     on estuarine eutrophication


                                                                        October 2002
             Waquoit Bay Watershed Ecological Risk Assessment:  The effect
                 of land-derived nitrogen loads on estuarine eutrophication
                                   U.S. Environmental Protection Agency
                       National Center for Environmental Assessment-Washington Office
                                    Office of Research and Development
                                            Washington, DC
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumfe liber content
processed chlorine free.



        This document has been reviewed in accordance with U.S. Environmental Protection
  Agency policy and approved for publication.  Mention of trade names or commercial products
  does not constitute endorsement or recommendation for use.

        A watershed ecological risk assessment of Waquoit Bay, located on the south coast of
 Cape Cod, MA, was performed for managers to better understand the environmental impacts of
 human activities. An interdisciplinary and interagency workgroup identified all the stressors of
 concern (chemical pollution, pathogens, altered freshwater flow, nutrient enrichment/eutrophica-
 tion, physical alteration of habitat, and fishing/shellfishing) and selected assessment endpoints
 (estuarine percent eelgrass cover, finfish diversity and abundance, scallop abundance,
 anadromous fish reproduction, wetland bird and piping plover habitat distribution and
 abundance, and tissue contamination of fish and shellfish). The workgroup later decided to focus
 on nutrient enrichment and its impacts on percent eelgrass cover and scallop abundance.
        A nitrogen loading model (NLM) was used to estimate the amount and sources of
 nitrogen entering the watershed, and an estuarine loading model (ELM) was used to estimate the
 nitrogen available to producers in shallow estuaries. The NLM indicates that atmospheric
 deposition is the largest source of nitrogen, but because more atmospheric nitrogen than
 wastewater nitrogen is intercepted in the watershed, wastewater becomes the largest contributor
 of nitrogen reaching the bay. By comparing increases in nitrogen loads to losses in the area of
 eelgrass cover over the last 60 years, it appears that eelgrass disappears once nitrogen loads reach
 20 kg/ha/yr. Both the increase in nitrogen load and the decrease in eelgrass can be correlated to
 decreases in the annual harvest of scallops. The models provide the opportunity for managers to
 assess a variety of options to reduce nitrogen loads to their estuaries and to achieve the loads that
 could allow the return of eelgrass to the target area.

 Preferred citation:
 U.S. EPA (Environmental Protection Agency). (2002) Waquoit Bay watershed ecological risk
 assessment.  National Center for Environmental Assessment, Washington, DC; EPA/600/R-
02/079.  Available from:  National Technical Information Service, Springfield, VA; PB2003-
 102013 and .


                             CONTENTS (continued)

     Eelgrass Biomass and Production	5-17
     Combined Effect of Nitrogen Loading on Primary Producers .... 5-19
     Zooplankton Egg Production 	5-20
     Shellfish Growth Rates	5-20
     Finfish Abundance 	5-20
     Summary of Effects on Ecosystem Components  	5-20
         5.2.2. Effects on Assessment Endpoints	5-23
     Percent Eelgrass Cover  	5-23
     Scallop Harvest	5-23
     Summary of Effects on Assessment Endpoints  	5-23

         6.1.1. Changes in Exposure: Back-Casting Nitrogen Loads  	6-1
         6.1.2. Temporal Changes in Effects:  Impact on Percent Eelgrass Cover and
              Scallop Harvest	6-5
         6.1.3. Effects of Other Stressors on Eelgrass	6-7

         7.1.1. Reducing Fertilizer Application Rates 	7-4
         7.1.2. Managing Wastewater	7-4

APPENDIX A.  Supplemental Information on the Waquoit Bay Estuarine Complex  	  A-l
   A.I Geological and Hydrological Characteristics 	  A-l
   A.2 Biological Characteristics 	  A-2

APPENDIX B.  Organizations Concerned About the Waquoit Bay	B-l

APPENDIX C.  Public Concerns and Waquoit Bay Stressors	C-l
     C.I Public Concerns	C-l
     C.2 Sources and Stressors in the Waquoit Bay Watershed 	C-2

APPENDIX D.  Attendees at the Waquoit Bay Management Goals Meeting	  D-l

APPENDIX E.  Information on Contamination from the Massachusetts Military Reservation  E-l
     E.I Phosphorus Loading to Ashumet Pond  	E-l
     E.2 Volatile Organic Compound (VOC) Contamination in Plumes  	E-3




        1.2.1. Management Goal and Objectives	1-2
        1.2.2. Stressors	1-3
        1.2.3. Assessment Endpoints  	<	1-3
        1.2.4. Conceptual Model  	1-3
        1.2.5. Analysis Plan 	1-4
   1.3.  RISK ANALYSIS	1-4

   2.1.  THE WATERSHED	2-1


        4.1.1. Sources of Stressors  	4-1
        4.1.2. Selected Stressors	4-3
   4.3.  ANALYSIS PLAN  	4-6
        4.3.1. Comparative Risk Ranking	4-6
        4.3.2. Relationships Between Stressors and Ecological Responses	4-8
     The Importance of Stressors in Waquoit Bay	4-9
     Focus on Nutrient Enrichment as the Dominant Stressor	4-13
        4.3.3. Summary of the Analysis Plan	4-13

        5.1.1. Nitrogen Loads		5-1
     Estimates Using the Nitrogen Loading Model (NLM)	5-1
     Measurements	5-7
     NLM Validation	5-7
     NLM Uncertainty 	5-10
   5.2.  EFFECTS ANALYSIS  	5-12
        5.2.1. Cascade of Effects on Ecosystem Components	 5-13
     Nitrogen Concentrations	5-13
     Phytoplankton Biomass and Production	5-17
     Macroalgae Biomass and Production	5-17


                                  LIST OF TABLES

3-1.   The Waquoit Bay watershed management objectives	3-2

4-1.   Relationship between assessment endpoints and management objectives	4-5

4-2.   Effects matrix for the Waquoit Bay watershed	4-7

4-3.   Relative importance of identified stressors to the Waquoit Bay ecosystem	4-8

5-1.   Inputs, losses, and default terms used by the nitrogen loading model (NLM)  	5-3

5-2.   Measured nitrogen loads to the subestuaries of Waquoit Bay	5-8

5-3.   Sources and 615N values of nitrate in groundwater	5-10

5-4.   Error analysis of NLM variables	5-12

5-5.   Characteristics of subestuaries of the Waquoit Bay watershed	5-14

6-1.   Relative contribution of each of the major sources of nitrogen to the Waquoit Bay
      estuary in 1938 and 1990	6-5

7-1.   Onsite septic system retention efficiencies reported for various alternative systems  ... 7-8

7-2.   Changes in water residence time predicted by dredging simulations 	7-9

                                 LIST OF FIGURES

2-1.   Delineation of the watershed of Waquoit Bay, MA ............................ 2~2

2-2.   Aerial photographs of Waquoit Bay in 1938 (top) and 1990 (bottom) ............. 2-4

2-3.   Number of parcels that are built upon (black) and remain to be built (white) as of
      1990 in the Childs River watershed of Waquoit Bay, MA  ...................... 2-5

2-4.   Area of eelgrass in Waquoit Bay between  1951 and 1992  ...................... 2-6

2-5.   Framework for the ecological risk assessment ................................ 2'9

4-1 .   Conceptual model of the Waquoit Bay watershed ecological risk assessment  ....... 4-2

5-1.   Schematic of the nitrogen loading model  ................................... 5-2

5-2.   Measured versus modeled nitrogen (N) loads to Waquoit Bay  ................... 5-9

5-3.   Values of 615N of nitrate in groundwater versus percent of nitrogen (N)
      from wastewater [[[

5-4.   Dissolved inorganic nitrogen (DIN) concentrations in the three subestuaries of
      Waquoit Bay subject to land-derived nitrogen (N) loads ....................... 5-14

5-5.    Schematic of inputs and exports of the ELM and the NLM  .................... 5-15

5-6    Comparison of dissolved inorganic nitrogen (DIN) concentrations predicted
       by the estuarine loading model (ELM) with measured concentrations in the
       water column of several Waquoit Bay estuaries (p<0.01)  ...................... 5-16

5-7.   Effects of nitrogen (N) loading on biomass and primary production of
       phytoplankton, macroalgae, and eelgrass in Sage Lot Pond, Quashnet River,
       and Childs River (top to bottom) ......................................... -*"1*
 5-8    Partition of total primary production in shallow estuaries into contributions by
       phytoplankton, macroalgae, and seagrasses, all plotted against measured annual
       nitrogen (N) load  [[[ ->"iy

 5-9.   Characterization of Acartia tonsa in Waquoit Bay estuaries .................... 5-21

 5-10.  The effects of differences in nitrogen (N) concentration on the growth rates of
       softshell clams (top) and quahogs (bottom) ................................. 5~22

 5-11.  Percent seagrass cover lost as nitrogen (N) load increases for a multitude of
       temperate and tropical ecosystems ........................................ 5~24

 5-12.  Volume of bay scallop harvest in Waquoit Bay from 1965-1995 ................ 5-24

                            LIST OF FIGURES (continued)

6-2.    Historical changes in nitrogen (N) loading to the watershed and estuary of
       Waquoit Bay  	6-4

6-3.    Decreases in area of eelgrass and volume of scallops harvested over time as a
       function of increasing nitrogen (N) loads  	6-6

7-1.    Relationship between nitrogen (N) loads, dissolved inorganic nitrogen (DIN)
       concentration, and percent eelgrass cover	7-3

7-2.    Historical changes in nitrogen (N) loading predicted by the nitrogen loading
       model	  7-4

7-3.    Reduction in the total nitrogen (N) load that would result from varying the
       amount of fertilizers used in the Waquoit Bay watershed	7-5

7-4.    Reduction in total nitrogen (N) load in the Waquoit Bay watershed (and corresponding
       year) that would result from implementing various wastewater treatment systems ... 7-7

E-l.    Map of the plumes emanating from the Massachusetts military reservation	E-6

                               LIST OF ACRONYMS
Air Force Center for Environmental Excellence
Dissolved Inorganic Nitrogen
Dissolved Organic Nitrogen
Estuarine Loading Model
U. S. Environmental Protection Agency
Geographic Information System(s)
gallons per minute
Installation Restoration Program
Massachusetts Military Reservation
Atmospheric Nitrogen
National Center for Environmental Assessment
Nongovernmental Organizations
Nitrogen Loading Model
National Marine Fisheries Service
National Oceanic and Atmospheric Administration
parts per billion
Royal Dutch Explosive
Residence time
Total Dissolved Nitrogen
U.S. Geological Survey

                        LIST OF ACRONYMS (continued)

VOC      Volatile Organic Compound

WBLMER Waquoit Bay Land Margin Ecosystems Research Project

WBNERR Waquoit Bay National Estuarine Research Reserve


       Risk assessment plays an increasingly important role in determining environmental
policies and decisions at the U.S. Environmental Protection Agency (EPA). In 1998, EPA
published Guidelines for Ecological Risk Assessment to provide a broad framework applicable to
a range of environmental problems associated with chemical, physical, and biological stressors.
As ecological risk assessment evolves, it is moving beyond a focus on assessing effects of simple
chemical toxicity on single species to the cumulative impacts of multiple interacting chemical,
physical, and biological stressors on populations, communities, and ecosystems. Although EPA
has considerable experience in applying the ecological risk assessment paradigm in source-based
approaches, such as those focused on particular chemicals, specific guidance on place-based
approaches (e.g., watersheds and regions) is still  limited. This assessment of the Waquoit Bay
watershed was completed to address a specific environmental problem through application of the
risk assessment methods represented in the Guidelines. Through this assessment, and other
watershed scale assessments like it, the Office of Research and Development is learning how to
develop new tools and approaches to support local environmental decision makers. An important
component of these approaches is active participation by local stakeholders. The Waquoit Bay
watershed assessment provides a good example of partnering between government,
environmental organizations, and others to support environmental decision making with strong
       Waquoit Bay was selected because its watershed contains valued and threatened
ecological resources; there was an abundance of previously collected stressor and effects data; it
is subjected to multiple physical, chemical, and biological stressors; and it has a number of
organizations  working to protect the ecological resources.  This assessment is intended to address
such concerns by analyzing stressors and resulting ecological effects and by stimulating broader
public awareness and participation in decision making for reducing ecological risks.  This
watershed assessment report serves as an example for others to follow on how to use ecological
risk assessment principles in a watershed-scale assessment to improve the use of science in
decision making.
                                        Michael Slimak
                                        Associate Director of Ecology
                                        National Center for Environmental Assessment
                                        U.S. EPA, Office of Research and Development


       The National Center for Environmental Assessment-Washington Office (NCEA-W),
National Oceanic and Atmospheric Administration-National Marine Fisheries Service, EPA
Region I, Boston University Marine Program, and other organizations developed this watershed
ecological risk assessment to help protect the Waquoit Bay watershed.  The document has three
purposes: (1) to provide information to help make more informed decisions on how to protect
the valued ecological resources of the watershed; (2) to provide data and references for future
research in the watershed; and (3) to demonstrate the benefits of applying ecological risk
assessment at the watershed scale. The report is based on the Guidelines for Ecological Risk
Assessment and advice and support from NCEA, while exercising the necessary flexibility to
implement the risk assessment approach at the watershed scale. To serve as an example for
others seeking to increase the use of science in place-based decision making, the document
includes brief descriptions of the process the workgroup followed along with the major analyses
performed.  The Hterature'search supporting the document was completed in May 2000.
       A more concise report of the assessment's findings and methods can be found in Serveiss
et al. (submitted).  Lessons  learned about applying ecological risk assessment to the watershed
scale, including those acquired from this assessment, are described in Serveiss et al. (2000) and
Serveiss (2002). Diamond and Serveiss (2001) provides another example of an EPA-sponsored
ecological risk assessment.  Discussion on how ecological risk assessment principles can be
applied at an even larger spatial scale (e.g., a region) can be found in Landis and Wiegers (1997)
and Wiegers et al. (1998).


      The National Center for Environmental Assessment-Washington Office (NCEA-W)
within the EPA's Office of Research and Development was responsible for preparing this
document. Draft reports were prepared by Boston University's Marine Program (BUMP) at the
Marine Biological Laboratory (MBL) under cooperative agreement No. CR 825851-01-0 and
Purchase Order No. 1W-0332-NAEX with input from the other authors and workgroup members.
The Waquoit Bay risk assessment was prepared by a diverse group of people representing
organizations and agencies interested in the management and protection of the biota of the
Waquoit Bay watershed. Much of the data used in the risk analysis were collected and analyzed
as part of the Waquoit Bay Land Margin Ecosystems Research project, which was funded by a
grant from the National Science Foundation's Land Margin Ecosystems Research initiative, by
EPA's Region I and by the Northeast Fisheries Science Center (NEFSC) of the National Oceanic
and Atmospheric Administration's (NOAA's) Sanctuaries and Reserves Division. The analysis
itself was performed at the Boston University Marine Program with input from the NEFSC of
NOAA's National Marine Fisheries Service (NMFS).
       In a project that has lasted almost a decade, it is difficult to acknowledge all  of the key
players but the following individuals helped bring this endeavor to fruition: EPA project
managers through problem formulation (Suzanne Marcy and John Miller); Workgroup Chairs
(Maggie Geist and Patti Tyler); MBL/BUMP scientists (Jennifer L. Bowen and Ivan Valiela);
Tern Konoza (EPA/NCEA-W) who managed the document production activities and provided
editing and word processing support; Leela Rao (EPA/NCEA-W) who helped revise the final
report- and mostly David Dow (NOAA/NMFS), who provided continuity and historical
perspective between the Workgroup Problem Formulation report and the Risk Analysis report.

EPA Project Officer:
Victor B. Serveiss, U.S. EPA, NCEA-W, Washington, DC

 Authors:                                                     .              „,   ,
 Jennifer L. Bowen, Boston University Marine Program, Marine Biological Laboratory, Woods
 Hole, MA                                                            .        ,
 David Dow, Northeast Fisheries Science Center, National Marine Fisheries Service, Woods
 Hole, MA
 Victor B. Serveiss, U.S. EPA, NCEA-W, Washington, DC
 Ivan Valiela, Boston University Marine Program, Marine Biological Laboratory, Woods Hole,
 Leela Rao, U.S. EPA, NCEA-W, Washington, DC

 Contributors:                                                    „      ,   ..  ..
 Maggie Geist, Association for the Preservation of Cape Cod, Orleans, MA, formerly with the


 Waquoit Bay National Estuarine Research Reserve, Waquoit, MA
 Patti Tyler, EPA Region VQT, Denver, CO, formerly with EPA Region I, Lexington, MA
 Suzanne Marcy, U.S. EPA, NCEA-IO, Anchorage, AK

 EPA's Ecological Risk Assessment Co-chairs:
 Maggie Geist, Association for the Preservation of Cape Cod, Orleans, MA, formerly with the
 Waquoit Bay National Estuarine Research Reserve, Waquoit, MA
 Patti Tyler, EPA Region Vffl, Denver, CO, formerly with EPA Region I, Lexington, MA

 Other Former Workgroup Members:
 Vicki Atwell, formerly with U.S. EPA, Office of Research and Development, Washington, DC
 Jeroen Gerritsen, Tetra Tech, Inc., Owings Mills, MD
 John Miller, U.S. EPA, Office of Water, Washington, DC
 Conchi Rodriguez, formerly with EPA, Office of Prevention, Pesticides, and Toxic Substances,
 Washington, DC
 Chuck Spooner, U.S. EPA, Office of Water, Washington, DC

 EPA Reviewers:
 James Andreasen, U.S. EPA, NCEA-W, Washington, DC
 Patricia Cirone, U.S. EPA, Region X, Seattle, WA
 Brian Hill, U.S. EPA, Office of Research and Development, Duluth, MN
 Lester Yuan, U.S. EPA, NCEA-W, Washington, DC
 Susan Norton, U.S. EPA, NCEA-W, Washington, DC

 Other Reviewers:
Peter DeFur, Virginia Commonwealth University, Richmond, VA
Wayne Landis, Western Washington University, Bellingham, WA
Kenneth Foreman, The Ecosystems Center, Marine Biological Laboratory, Woods Hole, MA


                              1. EXECUTIVE SUMMARY

        Waquoit Bay is a small estuary on the south coast of Cape Cod, Massachusetts. It is
 prized by residents and visitors for its aesthetic beauty and recreational opportunities. The
 watershed and bay and the adjoining marshes, tidal rivers, and barrier beaches provide ideal
 habitat for plant and animal life, including piping plovers and least terns (endangered birds), the
 sandplain gerardia (endangered plant), winter flounder, blue crabs, scallops, and clams and
 anadromous (spawn in fresh water) and catadromous (spawn in salt water) fish.
        Human encroachment is changing the landscape and contributing nutrients and
 contaminants to the bay.  In the Waquoit Bay watershed more than 85% of homes use onsite
 septic systems. The permeable sandy soils allow much of the nitrogen input from the septic tank
 leachate to reach groundwater. The nitrogen in groundwater travels to coastal waters, where it
 stimulates primary production, resulting in thick mats of macroalgae that have now replaced
 once-abundant eelgrass meadows, the preferred habitat for many organisms in the bay. The
 EPA-sponsored ecological risk assessment created a mechanism to bring organizations together
 to integrate the results of various research and planning efforts. Documenting the process, the
 analyses, and the modeling approach provides a means by which resource managers can  assess
 the risk of land-derived nitrogen to their estuaries and evaluate the magnitude of the problem
 caused by nitrogen in comparison to other stressors. This information can be used to make wiser
 remediation decisions related to providing funds for eelgrass planting, or requiring denitrifying
 septic systems, or making land use changes.

       Ecological  risk assessment is a process for collecting, organizing, analyzing, and
 presenting scientific information to make it more useful for decision making. It is a unique form
 of ecological assessment and includes the term "risk" because it presumes that a cause and effect
 relationship exists  and that the relationship can be expressed as a stressor-response curve.
 Although extensively used to predict the impacts of single stressors (e.g., a particular pesticide)
 on a particular species, EPA seeks to demonstrate the use of ecological risk assessment in
 evaluating environmental problems addressed through the watershed approach. The watershed
 approach is based on using partnerships, sound science, and environmental management  in
 decision making. EPA seeks to demonstrate that integrating the watershed approach with
 ecological risk assessment will increase the likelihood that environmental monitoring and
 assessment data will be used to inform decisions.
       The Waquoit Bay watershed was selected as the site of an EPA-sponsored watershed
ecological risk assessment because local, state, and federal organizations were interested in
cooperating in a risk assessment; it has multiple stressors  (e.g., nutrients, toxic chemicals, and
altered freshwater flow); there is abundant data; and the Waquoit Bay National Estuarine
Research Reserve (WBNERR) and EPA Region I were willing to lead the risk assessment
workgroup. Other members of the workgroup included NOAA's Northeast Fisheries Science


Center National Marine Fisheries Service; Boston University Marine Program, Marine
Biological Laboratory; and EPA's Offices of Research and Development, Water, and Pollution
Prevention and Toxic Substances.
       The assessment process began with planning and problem formulation, proceeds through
analysis of exposure and effects, and ends with risk characterization. During planning, the
management goal for the watershed and the purpose, scope, and complexity of the assessment
was established. In problem formulation, ecologically and socioeconomically relevant
assessment endpoints were defined, and conceptual models and plans for the analysis were
developed. In this assessment, a nitrogen loading model (N1M) predicted the amount of nitrogen
reaching the bay from various sources within the watershed. During risk characterization, the
results of the model were combined with historical data on eelgrass area cover and scallop
harvests to provide insights on how these valued resources can be reestablished to historical

1.2.1. Management Goal and Objectives
       The management goal was defined by an interdisciplinary and interagency workgroup ot
scientists  and managers on the basis of a public meeting, a meeting of workgroup members to
develop the goal and objectives, and a meeting with local resource managers to refine the goal
and objectives. The agreed-upon management goal was:

        Reestablish and maintain water quality and habitat conditions in Waquoit Bay
        and associated wetlands, freshwater rivers, and ponds to (a) support diverse, self-
        sustaining commercial recreational, and native fish and shellfish populations and
        (b) reverse ongoing degradation of ecological resources in the watershed.

        The 10 management objectives that more explicitly state the kinds of management results
 that were implied in the general goal statement include:

        1.  Reduce or eliminate hypoxic or anoxic events
        2.  Prevent toxic levels of contamination in water, sediments, and biota
        3   Restore and maintain self-sustaining native fish populations and their habitat
        4   Reestablish viable eelgrass meadows and associated aquatic communities in the bay
        5.  Reestablish a self-sustaining scallop population that can support a viable fishery
        6.  Protect shellfish beds from bacterial contamination that results in bed closures
        7.  Reduce or eliminate nuisance macroalgal growth
        8.  Prevent eutrophication of rivers and ponds
        9.  Maintain diversity of native biotic communities
        10. Maintain diversity of wetlands habitat

        The workgroup developed a list of stressors and endpoints, built a conceptual model that
 tied the stressors to their endpoints, and decided on an analysis plan.

 1.2.2.  Stresssors
        Six stressors from human activities that could impact resources in Waquoit Bay and the
 potential sources of those stressors within the watershed were selected for analysis.  These



Chemical pollution from pesticides and herbicides, emissions, industrial point
sources, and boating activities
Altered freshwater flow from new construction and waste treatment
Nutrient enrichment/eutrophication from agriculture, lawn and garden fertilization,
waste treatment, industrial point sources, and atmospheric deposition of emissions
Physical alteration of habitat from dredging and boating activities
Fishing and shellfishing resulting from harvest pressure from commercial and
recreational fishing
Pathogens from industrial point sources, runoff from impervious surfaces, and waste
 1.2.3. Assessment Endpoints
       Assessment endpoints are the link between scientifically measurable endpoints and the
 objectives of the stakeholders and resource managers (Suter 1989,1993). Endpoints should be
 ecologically relevant, related to the previously defined management objectives, and susceptible to
 stressors (U.S. EPA, 1998).  The workgroup originally selected seven assessment endpoints. The
 endpoints were:

       1.  Estuarine percent eelgrass cover
       2.  Finfish diversity and abundance
       3.  Scallop abundance
       4.  Anadromous fish reproduction
       5.  Wetland bird habitat distribution and abundance
       6.  Piping plover habitat distribution and abundance
       7.  Tissue contamination of fish and shellfish

 1.2.4.  Conceptual Model
       The conceptual model is a broad representation of relationships among human activities
in the watershed (sources) are the stressors believed to result from those sources (as described in
Section 1.2.2), the exposure pathways linking stressors to effects, and the assessment endpoints.
Each of the pathways in the conceptual model was derived from information about the Waquoit
Bay watershed and estuary in the peer-reviewed scientific literature, from ecological theory on

how systems function, and from similar relationships established in other watersheds. The
conceptual model illustrates relationships such as how excess nutrients increase algal growth,
shading and reducing eelgrass habitat, and reducing scallop abundance. This model provides the
foundation for the analysis plan.

1.2.5.  Analysis Plan
       The workgroup conducted a comparative risk analysis to help define which stressors,
assessment endpoints, and relationships should be further examined. Stressors were ranked in
terms of potential risk to all resources in the watershed on the basis of best professional judgment
of the workgroup. The results of the comparative analysis ranked nutrients first. To verify that
nutrients were indeed the largest stressor in Waquoit Bay, the authors later examined each of the
stressors for the intensity of its impact, extensiveness within the watershed, and the likelihood to
increase over time. This analysis also indicated that nutrient loading is the dominant agent of
change in the Waquoit Bay watershed.
       Later the authors agreed to focus on nitrogen loading because phosphorus input, although
important in eutrophication of freshwater ponds, is being analyzed and mitigated by the Air Force
Center for Environmental Excellence.  In addition, the authors agreed that there were sufficient
data available to develop models to evaluate the quantity of nitrogen entering the bay and draw
correlations with critical ecosystem components.
        To evaluate risk from nitrogen loading to assessment endpoints it was necessary to
quantify the loading of nitrogen into the watershed and estuary and to evaluate how a given load
of nitrogen impacts the estuarine ecosystem. The area of eelgrass cover and scallop abundance
were chosen as the assessment endpoints because they were most  susceptible to the identified
 stressor of concern, nitrogen loading, and were not so variable that they easily evaded
 characterization. Additionally, several other characteristics that are necessary for ecosystem
 functioning, including nitrogen concentrations in the estuary, phytoplankton and macroalgae
 biomass and production, zooplankton egg production, shellfish growth rates, and finfish
 abundance were evaluated in less detail. Although it is possible to evaluate the risk to these other
 ecosystem components, their complexity makes risk modeling extremely difficult.

        The NLM sums the nitrogen loads to the watershed from three major sources:
 atmospheric deposition, septic-derived wastewater, and fertilizer  application. It then subtracts
 losses during transport to yield a value for the quantity of nitrogen arriving at the edge of the
 estuary (or salt marsh). The estuarine loading model (ELM) estimates the concentrations of
 dissolved inorganic nitrogen available to producers in shallow estuaries. These estimates account
 for inputs and losses of nitrogen after it reaches the bay through processes such as demtnfication,
 burial of nitrogen in sediments, speed of water movement out of the bay (turnover rate),
 regeneration of nitrogen from the sediments, nitrogen fixation, and direct atmosphenc deposition
 of nitrogen onto the surface of the bay.


        The loading rates and effects of nitrogen input on various components of ecosystem
 function are variable. Nitrogen loads entering Waquoit Bay were measured by sampling
 groundwater around the periphery of the bay and the subestuaries. Nitrogen loads ranged from
 433 kg N/yr in the pristine Sage Lot Pond to 9879 kg N/yr in the Quashnet River. The total
 amount of nitrogen reaching all of Waquoit Bay was just over 26,500 kg N/yr.
        Nitrogen loading leads to an increase in the concentration of nitrogen that is available in
 the estuary. As a result, the biomass and production of both phytoplankton and macroalgae, two
 of the dominant primary producers in Waquoit Bay, have increased substantially.  The increases
 in these two producers worked together to prevent light from penetrating the bottom of the
 estuaries, resulting in a decrease in the biomass and production of the other dominant primary
 producer—eelgrass (Zostera marina).  Eelgrass is especially important because it provides
 habitat and a refugia for juvenile fish and shellfish.  The effect of nutrient input on higher trophic
 levels also was variable. Evidence suggests that copepod egg production increases as a result of
 the greater production of phytoplankton,  a preferred food choice. However, there was no related
 increase in the number of copepods (free-floating tiny aquatic arthropods) in the Waquoit Bay
 estuaries. This suggests that there was a  lack of response of copepods to increased food supply,
 probably as  a result of the  short residence time of the water within these estuaries. Growth rates
 of two commercially important shellfish  species, the softshell clam (Mya arenaria) and hardshell
 clam (Mercenaria mercenaria), increased as nitrogen load increased, but populations of the bay
 scallop (Argopecten irradians) decreased drastically as nitrogen load increased. Finally, the two
 most common estuarine fmfish species, Atlantic silverside (Menidia menidia) and mUmmichog
 (Fundulus heteroditus), showed no response to nutrient loading in any of the estuaries of
 Waquoit Bay.

       In risk characterization, the exposure and effects analyses are integrated to estimate risks
 to the assessment endpoints. Risk characterization also serves to summarize and describe the
 results of the risk analysis in such a way that results can be readily translated to managers and
 other stakeholders. In the risk characterization for the Waquoit Bay watershed, the authors
 integrated the findings from the series of  models that were created to assess the eutrophication
 exposure and effects and calculated the risk of eutrophication.
       The models were used to characterize changes in nitrogen loading.  Land-use changes
 over, the last 60 years were assessed from aerial photographs, and this information was integrated
 into the model structure to  back-cast nitrogen loads to the late 1930s. As the Waquoit Bay
 watershed became more urbanized, the modeled nitrogen load to the watershed increased from
 slightly more than 10,000 kg N/yr to more than 24,000 kg N/yr in 1990.  This  increase in
nitrogen load can be correlated to decreases in both eelgrass area and scallop harvest with time.
Eelgrass area declined from 60% coverage in 1955 to less than 10% coverage in 1990.  This
decrease in eelgrass area occurred as nitrogen loads exceeded 20 kg N/ha/yr.  Additionally, the

harvest of scallops decreased from over 200,000 landings (L)/yr to well under 20,000 L/yr during
the same time period.

       The historical reconstruction of land-derived nitrogen loads, plus the linkage of these
loads to assessment endpoints such as percent eelgrass cover and bay scallop harvest, provided
some means to identify management priorities and define potential restoration measures.  If
management goals included reduction of nitrogen loads, then resource managers could use the
changes that occurred in assessment endpoints to establish remediation targets. For example, it
stakeholders wanted to restore eelgrass to 30% coverage in the estuary, then this modeling
characterization indicates that managers would need to reduce nitrogen loads to approximately
18 000 kg N/yr to generate conditions that could potentially allow eelgrass to survive. As a
further step these models also could be used to run simulations identifying management options
that could produce the desired targets. The models could be used to assess, for example, the
reduction in nitrogen loads that would occur if fertilizer application rates were lowered or septic
tank retention efficiency were increased. Thus the risk assessment has provided not only an
analysis of the impact of major stressors on the ecological components of Waquoit Bay, but also
provided insights into how the Waquoit Bay community can reestablish  and maintain the Bay s
habitat and water quality.

                                  2. INTRODUCTION

        The Waquoit Bay watershed in Cape Cod, Massachusetts, has many valuable natural
 resources that are being jeopardized by urbanization and other human activities. Nitrogen
 loading was suspected to be the principal cause of environmental problems, such as loss of
 eelgrass and fish populations (including scallops), which are dependent on eelgrass habitat. This
 risk assessment describes how excess nitrogen input is the stressor of concern in the Waquoit
 Bay watershed and then evaluates and models its effects on valued ecological resources.
 Following this introduction, the report describes the process by which the assessment was
 performed and presents the analytical results.
        An interdisciplinary and interagency workgroup used watershed ecological risk
 assessment principles to analyze available information in a manner that would increase the
 likelihood that it would be useful for decision making. The watershed approach (U.S. EPA,
 1996) is based on using partnerships, a hydrologically defined geographic boundary, sound
 science, and sound environmental management in decision making. Ecological risk assessment
 (U.S. EPA, 1998) is a process to collect, organize, analyze, and present scientific information.
 Watershed ecological risk assessment (Serveiss et al., 2000;  Serveiss, 2002) integrates the
 watershed approach with ecological risk assessment to increase the likelihood that environmental
 monitoring and assessment data will be used in  watershed-scale decision making. Documenting
 the process and performing the modeling and other analyses  provides scientific information to
 help managers justify taking actions to address problems.

       Waquoit Bay is a small estuary on the south coast of Cape Cod, Massachusetts (Figure 2-
 1). Its watershed covers about 53 km2 (21  mi2) and includes  freshwater streams and ponds, salt
 ponds and marshes, pine and oak forests, barrier beaches, and open estuarine waters.  The land
 and water are home, spawning ground, and nursery for a diversity of plant and animal life,
 including threatened and endangered species,  such as piping plovers (Charadrius melodius), least
 terns (Sterna albifrons), and the  sandplain  gerardia (Agalinis acuta\ as well as commercially
 important shellfish species such  as blue crabs  (Callinectes sapidus), bay scallops (Argopecten
 irradians), and hardshell clams (Mercenaria mercenaria), and a multitude of recreational and
 commercial fish species, including alewife (Alosa pseudoharengus), and winter flounder
 (Pseudopleuronectes americanus). For more information on the geological, hydrological, and
biological characteristics of the Waquoit Bay estuary see Appendix A.
      The Waquoit Bay region  was  part of the Wampanoag tribal lands when European settlers
arrived in the early 1600s on what is now Cape Cod (Gallagher, 1983). For more than 200 years,
the Waquoit watershed was used primarily for hunting, farming (strawberries and potatoes were
important crops), and maritime industries such as fishing, whaling, and shipbuilding (Faught,

    42°40' N -
  1.  Eel Pond
  2.  Childs River
  3.  Quash net River
  4.  Hamblin Pond
  5.  Jehu Pond
  6.  Sage Lot Pond
  7.  Snake Pond
  8.  Ashumet Pond
  9.  Johns Pond
 10.  Timms Pond
Figure 2-1. Delineation of the watershed of Waquoit Bay, MA. Subwatersheds drmning into
the six estuaries of Waquoit Bay are noted with dashed lines. The inset shows the location of
Waquoit Bay on Cape Cod, MA. Annual measured nitrogen loads to ^ subestuanes are
provided. The nitrogen loads entering the three upper ponds are included as a portion of the loads
to the receiving estuaries.
Source: Modified from Bowen and Valiela (2001a).

 1945). Since the late 1800s, and with the advent of rail service from Boston to Cape Cod,
 the natural beauty of the area has attracted more people. Today, more than 8,000 people live
 within the Waquoit watershed.  The population of the watershed swells in summer months,
 largely as a result of visitors from the greater metropolitan areas of Boston, Providence, and New
 York City. Cape Cod's economic viability depends largely on tourists who are drawn to the
 abundant sandy beaches, seafood restaurants, boating opportunities, and water recreation
 activities. Thus, the economy and the environment of Cape Cod depend on one another.
        These once-rural surroundings have become increasingly suburbanized with the
 development of bedroom and retirement communities. Aerial photographs of a portion of the
 Waquoit Bay watershed taken in 1938 and in 1990 illustrate the population expansion that has
 occurred over the years (Figure 2-2). According to the U.S. Census Bureau, the population of
 Barnstable county, within which the Waquoit Bay watershed is located, increased by 13.9%
 between 1990 and 1999. In Massachusetts this was surpassed only by Dukes County (Martha's
 Vineyard), which increased by 20.7%, and Nantucket County, which increased by 36.5%. By
 comparison, the increase in population of the entire state of Massachusetts was 2.6% during the
 same period (U.S. Census Bureau, 2000).
       As the population increases, so does the pressure on the valuable natural resources that
 attract people to the region. These resources include clean beaches, healthy eelgrass meadows,
 and viable fisheries. People want to live as near to the water as possible; the result is that the
 near-shore land parcels are the most frequently developed (Figure 2-3). The number of buildable
 lots that remain is still substantial, but most future building will take place farther away from
       Resource managers have expressed concern about the condition of the bay's natural
 resources, specifically, the significant decreases in the area of eelgrass meadows in Waquoit Bay
 and its subestuaries, increases in the frequency of anoxic events, groundwater contamination
 from a military reservation located in the far reaches of the watershed, changes in the abundance
 of recreationally important species, diminished aesthetics, and the ecological stress caused by
 greater recreational use of water resources. The next few paragraphs provide more detail on
 these concerns.
       Eelgrass (Zostera marina} is a flowering plant that inhabits the sediments of many
 shallow embayments of the northwestern Atlantic Ocean.  Numerous studies have shown that
 eelgrass meadows provide an important nursery habitat for the juvenile stages of commercially
 and recreationally important fish and shellfish (Valiela et al., 1992; Duarte, 1995). Eelgrass,
 macroalgae, and phytoplankton are the three dominant primary producers in shallow coastal
 systems like Waquoit Bay. Unlike the other two producers, eelgrass can acquire many of its
required nutrients through roots, but because it is rooted to the bottom of the estuary, it requires
sufficient light penetration through the water for photosynthesis.  In Waquoit Bay, as in many
shallow estuaries, increased phytoplankton and macroalgae, fueled by the addition of nitrogen
from human sources, have reduced the amount of light that reaches the eelgrass. As shown in

Figure 2-2. Aerial photographs of Waquoit Bay in 1938 (top)
and 1990 (bottom).  White areas in the 1938 figure are agricultural
lands and beach. The 1990 photo shows the disappearance of
agricultural land and the preponderance of suburban sprawl,
particularly near the coast.
Source- Massachusetts Department of Environmental Protection.

       .52  200
       s.  16
                                                       •  Built upon
                                                       D  Undeveloped
                            7       10       13       16
                      Distance (x100 m from shore)
     Figue 2-3. Number of parcels that are built upon (black) and remain to be
     built (white) as of 1990 in the Childs River watershed of Waquoit Bay, MA.
     Source: Modified from Valiela et al. (1992).
Figure 2-4, in 1951, eelgrass meadows covered most of Waquoit Bay proper and its adjoining
coastal ponds and rivers (Costa, 1988); today, eelgrass is absent from the bay proper (Short and
Burdick, 1996; Hauxwell et al., 2001). It still exists in some of the subestuaries of the bay,
although it has declined significantly in the Quashnet River, Hamblin Pond, and Jehu Pond.
Sage Lot Pond and Timms Pond still have healthy eelgrass populations (Short and Burdick,
1996; Hauxwell et al., 2001). Species that are dependent on eelgrass for habitat, feeding, or
spawning grounds also have suffered population declines. In particular bay scallops, which rely
on eelgrass for larvae settlement and predator refuge, showed a decline in abundance, as
evidenced by scallop harvest, from nearly 200,00017yr in the early 1960s to less than 20,000
L/yr in the mid-1990s (Shumway, 1991; Bowen and Valiela, 2001a).
       Oxygen depletion of the water is another effect of the increased abundance of
phytoplankton and macroalgae in bays. The decomposition of organic matter from decaying
plankton and algae consumes oxygen, and if there is not enough sunlight to produce oxygen
through photosynthesis, a condition known as hypoxia (low oxygen) or anoxia (no oxygen) can
occur. Very low oxygen concentrations can lead to the death of organisms. In 1988 and 1990,
fish kills occurred in Waquoit Bay; the northern beach was covered with  thousands of dead
winter flounder, shrimp, blue crabs, and other estuarine species (D' Avanzo and Kremer, 1994).

   Figure 2-4. Area of eelgrass in Waquoit Bay between 1951 and 1992. Black areas
   represent the extent of eelgrass beds.
   Source: Figures for 1951-1987 by J. Costa et al. (1992) and data from Short and Burdick (1996), published
   inValielaetal. (1992).

Smaller-scale fish kills also occurred in the summer of 2001. Anoxia can occur in freshwater
ponds as well. In Ashumet Pond and Johns Pond, both located in the upper reaches of the
Waquoit Bay watershed, blooms of phytoplankton have changed the color of the water and
depleted oxygen levels in the bottom waters of the pond. In Ashumet Pond, the deep bottom
waters (>12 m) are consistently anoxic during the summer months. Fish kills occurred in
Ashumet Pond in 1985 and 1986. Several expert panels have concluded that nutrient
enrichment—and the resulting anoxia—is the major man-made pollution problem affecting
coastal waters (GESAMP, 1990; NRC, 2000).
       The Massachusetts Military Reservation (MMR), which is located in the northern portion
of the Waquoit Bay watershed, was designated a Superfund site in 1989 due to extensive
groundwater contamination from a number of sources.  Nutrients from a wastewater treatment
plant, chemicals from fuel spills, and chlorinated solvents used at MMR in the past have
contaminated 55 to 60 billion gallons of Cape Cod's sole-source aquifer for drinking water.
These toxic chemicals include benzene, toluene, ethylbenzene, xylene, trichloroethylene,
tetrachloroethylene, and ethylene dibromide.  A human health risk assessment was conducted by
MMR's Installation Restoration Program (IRP) to evaluate the potential risks of these
contaminants to the public drinking water source.
       Because the Air Force Center for Environmental Excellence (AFCEE) has provided
money to the surrounding towns of Falmouth and Mashpee to hook up residents with private
wells to public water supplies when their water wells were threatened by MMR plumes, the
plumes pose no threat to public health. The goal of the cleanup is restoration of the sole-source


 aquifer for drinking water with the target cleanup level being the established maximum
 contaminant level for each contaminant.  The active extraction, treatment, and reinjection
 systems have reduced the contaminant mass in the plumes, as have some of the plumes for which
 monitored natural attenuation was chosen as the remedy.
       The IRP evaluates ecological impacts on aquatic components on the basis of alteration of
 the hydrologic flow to rivers and ponds that accompanies operation of groundwater extraction
 and treatment systems.  Based on the IRP, potential impacts of changed hydrologic flow are
 minimized during design of these groundwater treatment systems. Given that the focus of the
 IRP is the protection of the drinking water source, contaminant levels are being reduced based on
 human health drinking water criteria standards. However, the human health maximum
 contaminant levels for toxic contaminants in potential sole-source aquifers for drinking water are
 two to three orders of magnitude lower than the acute/chronic toxicity thresholds for aquatic
 biota.  Thus, while there are no known ecological impacts from the toxic chemical plume
 discharges, there is an impact on the freshwater pond from phosphorus discharge (Howarth,
       The Ashumet Valley wastewater plume, which originates at MMR, has contributed to
 seasonal anoxia through increases in phosphorus and nitrogen loading in Ashumet Pond.  The
 pond is the subject of a mitigation plan by the  AFCEE that uses an alum treatment to reduce the
 amount of phosphorus released from the sediments in the deep-water portion of the pond.
       With the increase in population in the Waquoit Bay watershed, the number of people who
 use the bay for recreational boating also has increased.  Resuspended sediments from boating
 activities, toxic chemicals from pressure-treated wood in docks, oil and gas leaks from boat
 motors, propeller scarring, shading of eelgrass beds by docks, and erosion of shores along
 Vineyard Sound have been cited by residents as additional potential sources of stress to valuable
 marine resources.  Sewage discharge from small boats was deemed to be of little concern since
 Waquoit Bay was  designated as a "No Discharge Area" in 1994, and any inputs from wastewater
discharge still occurring would be small relative to the inputs from land-derived sources.
Concern about the effects of population growth on Cape Cod has led to several initiatives and
activities, including the following, that predate the risk assessment:

       •   Creation of the Cape Cod Commission, a regional planning agency that has authority
          over all construction that could have regional impact on Cape Cod resources

       •   Work by the Association for the Preservation of Cape Cod, which has been
          instrumental in fighting for the protection of the Cape's drinking water supply and
          determining the impact of watershed nitrogen loading on coastal water quality

       •  Efforts  by the Waquoit Bay Land-Margin Ecosystem Research Project, a multi-
         institutional, interdisciplinary research program that has contributed to the knowledge

         of the impact of land-derived nitrogen loading on coastal waters and on the ecological
         functioning of the bay

         Designation of the Waquoit Bay National Estuarine Research Reserve (WBNERR),
         which also serves to protect the resources of the bay and its adjacent lands and to
         educate the public about their precious natural resources

         Designation of part of the Waquoit Bay watershed as a U.S. Fish and Wildlife Refuge,
         which has resulted in land acquisition and a multi-agency memorandum of
         understanding to develop a coordinated management plan for the reserve

         Designation of the Waquoit Bay area as an Area of Critical Environmental Concern, a
         Massachusetts designation that provides additional regulation for development that
         might impact natural resources

         Designation of the Waquoit Bay watershed as one of the EPA's sponsored ecological
         risk assessments
       The EPA-sponsored ecological risk assessment of Waquoit Bay builds on local and
regional work by creating a mechanism to integrate the results of various research and planning
efforts into management options for local coastal decision makers (Figure 2-5).  The purpose of
an ecological risk assessment is to collect, organize, analyze, and present scientific information
to help decision makers achieve management goals. It is a unique form of ecological assessment
and includes the term "risk" because it presumes that a cause and effect relationship exists and
that the relationship can be expressed as a stressor-response curve.  The Waquoit Bay watershed
ecological risk assessment was initiated because of interest by local, state, and federal
organizations in the watershed, the type of watershed (estuarine), the diversity of stressors (eg.,
nutrients, toxic chemicals, altered hydrology), an abundance of previously collected data, and
willingness by WBNERR and EPA New England to lead the risk assessment workgroup.
       There is much literature on the ecological risk assessment process (e.g., U.S. EPA, 1992,
 1998- Suter  1993) and limited literature on how it can be applied to watershed and regional
 assessments' (Landis and Wiegers, 1997; Wiegers et al., 1998; Norton et al., 2000; Serveiss et al.,
 2000- Serveiss, 2002). In watershed ecological risk assessments, the planning effort often is
 extensive because there  are many interested parties with diverse interests and numerous stressors
 impacting many valued ecological resources through numerous pathways.  Unlike other types of
 smaller-scale assessments, all of these pathways cannot be precisely quantitatively analyzed and
 innovative methods need to be used to analyze information and produce scientifically credible
 and useful conclusions for resource managers.

                     Ecological Risk Assessment
   expos ire
Communicating results
to the risk manager

Figure 2-5. Framework for the ecological risk assessment.

Source: U.S. EPA (1998).



        The Waquoit Bay watershed ecological risk assessment was based on a proposal by
 managers at Waquoit Bay National Estuarine Research Reserve (WBNERR) and EPA Region I,
 who were concerned about the deteriorating ecological quality of the bay. Based on this interest,
 a risk assessment workgroup was established (see page xii). The purpose of the risk assessment
 planning was to establish clear and agreed-upon environmental management goals and
 objectives. These objectives provided the context for the assessment and were supported by the
 diverse members of the Waquoit Bay community.  To achieve this support, the workgroup held a
 public meeting, reviewed the goals of environmental organizations,  and interacted with
 nongovernmental organizations (NGOs), scientists, and managers at the state and local level.
       Public meeting. EPA, in conjunction with WBNERR, held a public forum on  September
 21, 1993. Concerned citizens and representatives from some of the  organizations that have
 interests in Waquoit Bay attended (Appendix B). At this meeting WBNERR/EPA
 representatives asked participants to provide answers to two questions:

    •  What environmental resources do you value in Waquoit Bay?
    •  What is placing those values at risk?

 The attendees also were given a list of biological, chemical, and physical stressors and were
 asked to evaluate them with respect to the valued resources within the watershed. The results of
 the analyses are presented in Appendix C. From the information gathered at this public meeting,
 the workgroup developed a preliminary management goal.
       Goals of management organizations. To further refine the preliminary management goal
 into more specific management objectives, the workgroup reviewed  the goals of local, regional,
 and national resource management organizations and NGOs with responsibilities in the
 watershed (Appendix B).  This review was based on either written documentation published by
 the organizations or on statutes.
       Risk assessment workgroup consensus. Members of local resource management
 organizations were invited to a meeting sponsored by WBNERR on February 24,1995
 (Appendix D). The management goal and objectives, along with the summary of the goals of
 various organizations, were presented to meeting participants. Ultimately, the management goal
and objectives for the risk assessment were modified by the workgroup from comments from the
managers and regulators present at that meeting. The agreed-upon management goal became:

      Reestablish and maintain water quality and habitat conditions in Waquoit Bay
      and associated wetlands, freshwater rivers, and ponds to (a) support diverse, self-
      sustaining commercial, recreational, and native fish and shellfish populations and
      (b) reverse ongoing degradation of ecological resources in the watershed.

       Ten management objectives were developed to characterize the management goals that
were implied in the more general statement. The management objectives were partitioned into
three categories (Table 3-1). The estuarine and freshwater category included three objectives that
were common to both surface water types. Four objectives under the estuarine category and three
objectives under the freshwater category were unique to those waters. The management
objectives reflected generic concerns of the NGOs and local/state/federal managers with interests
in the watershed. This systematic process enabled the workgroup to identify assessment
endpoints and develop an analysis plan for the assessment.
       Table 3-1.  The Waquoit Bay watershed management objectives
Affected area
Estuarine and freshwater
Management Objective
Reduce or eliminate hypoxic or anoxic events
Prevent toxic levels of contamination in water,
7 sediments, and biota
Restore and maintain self-sustaining native fish
populations and their habitat
Reestablish viable eelgrass beds and associated
aquatic communities in the bay
Reestablish a self-sustaining scallop population in
the bay that can support a viable sport fishery
Protect shellfish beds from bacterial contamination
that results in bed closures
Reduce or eliminate nuisance macroalgal growth
Prevent eutrophication of rivers and ponds
Maintain diversity of native biotic communities
Maintain diversity of water-dependent wildlife

                             4. PROBLEM FORMULATION

        During problem formulation, conceptual models, assessment endpoints, and an analysis
 plan are developed. Problem formulation uses available information on sources, stressors,
 ecological resources potentially at risk and ecological effects to 1) identify the ecological
 resources that will be the focus of the risk assessment; 2) develop conceptual models of how
 these resources may be impacted by human activities; and 3) develop a plan for the analysis

        The watershed conceptual model is a broad representation of relationships among human
 activities in the watershed; the physical, chemical, and biological stressors believed to occur as a
 result of those activities; and the ecological effects likely to occur to each of the assessment
 endpoints. An earlier conceptual model developed on the basis of input from meeting
 participants, was found to help resource managers—who typically focus on a resource or a
 problem—to better consider the big picture. The earlier model, which is on display at the
 WBNERR visitor center, has been found to be a useful tool for communicating with the public.
 Figure 4-1 shows a revised model that reflects further insight by authors, additional input
 provided by reviewers, and modifications to further improve clarity.
        The pathways Figure 4-1  were organized around stressors. Each stressor has a colored
 line that illustrates a pathway connecting its sources to possible effects and selected endpoints.
 Each of the components of the model is represented by a different geometrical shape to aid
 interpretation. This is a broad-based model that provides a framework for the risk assessment
 and an overview of ecosystem processes.  The diagram shows only stressors and effects thought
 to be potentially important in the Waquoit Bay watershed. The exposure pathways, from the
 source of stressors to valued resources, are the possible risk hypotheses to consider for analysis
 as part of the risk assessment. In the following sections of problem formulation the sources of
 stressors, the nature of those stressors, and the assessment endpoints that they affect are

 4.1.1. Sources of Stressors
       The changing land and water use patterns in coastal and upland areas of the Waquoit Bay
 watershed are largely responsible for increasing the intensity of stressors in the watershed.  The
 main sources of stress include agriculture, urbanization and industrial development, and marine
       Agriculture. Fertilizers, pesticides, and herbicides from agricultural applications can
leach into groundwater and eventually enter estuaries. In general, some portion of the fertilizers
that are  applied to agricultural crops can be volatilized to ammonium and then deposited in the
form of nitrogen farther downwind of the source, but for  the Waquoit Bay area, the atmospheric
sources  are likely derived from far outside the watershed (Bowen and Valiela, 2001a).

     Agriculture     Urbanization and industrial development  Marine Activities
               \  flow
                          Of habitat
                               Increased //Increased
       Finfish and  /
        shellfish   /
  Habitat loss
        Brook trout
                                 diversity and
                                                          distribution and

Figure 4-1. Conceptual model of the Waquoit Bay watershed ecological risk assessment
Rectangles represent sources of stressors, hexagons are the specific stressors to toe system,
trapezotds represent the effects of those stressors, and the ellipses indicate specific endpomts
that are affected.

 Conversion of naturally vegetated land to agriculture can also result in erosion of upland soils
 and siltation in adjacent streams and estuaries, but for the coarse soils of Cape Cod, runoff and
 siltation are, at most, of moderate importance. Agricultural use of land in Waquoit Bay was
 prevalent early in the past century, but in recent years the area of land dedicated to agriculture has
 decreased dramatically (Bowen and Valiela, 200la).  Cranberry cultivation remains as the last
 notable commercial agricultural land use in the Waquoit Bay watershed.
        Urbanization and industrial development.  Urbanization has resulted in numerous sources
 of ecological stress in the Waquoit Bay system. As human populations on watersheds increase,
 so does the waste generated. Thus, wastewater disposal is one of the major sources  of stress
 resulting from population expansion. An additional stressor associated with urbanization is the
 fertilizer applied to lawns, gardens, and golf courses. New construction can result in increased
 areas of impervious surfaces, which can alter hydrological flow patterns.  The use of aquifer
 water for both drinking and waste disposal can alter hydrologic flow patterns, providing an
 additional  source of stress to the groundwater system. Finally, increased emissions from
 automobiles can  provide an additional input of nutrients and contaminants in downwind
 locations, but on  Cape Cod these local atmospheric sources are small compared to upwind
 sources (Bowen and Valiela, 2001a).
        Industrial sources of stress in watersheds can stem from point-source industrial waste
 disposal, runoff from areas of impervious land, and from industrial emissions generated both
 within and outside the watershed. On Cape Cod, there is minimal industrial land use, and the
 industries that are present are not of the type that lead to severe  ecological  effects or
 contamination. Disposal of hazardous materials and sewage treatment plant waste at MMR have
 resulted in 14 groundwater plumes that are contaminated with chlorinated solvents, fuel
 constituents, and  high concentrations of nutrients.
        Marine activities.  Stresses associated with marine activities include fuel leaks and
 chemical contamination from boating activities. Physical alteration of habitat could  result from
 anchors and boat  propellers. Stressors also are associated with the construction and operation of
 marinas, docks, and piers, all of which disturb sediments, alter habitats, and contribute chemical
 contaminants to the water. Dredging and shoreline modification in the form of jetty  and
 breakwater construction can also disturb sediments and alter estuarine circulation patterns.
 Finally, over-harvesting of stocks can result in the depletion of fishery resources.

4.1.2.  Selected Stressors
       From the descriptions of the potential sources of stress reviewed above, the workgroup
selected six stressor categories that potentially affect the Waquoit Bay watershed to consider for
further study:

       1.  Chemical pollution (from pesticide and herbicide application, emissions, industrial
          point sources, and boating activities)

      2.  Pathogens (from industrial point sources, runoff from impervious surfaces, and
          leaching of wastewater from septic systems)
      3.  Altered freshwater flow (from new construction and potential increased use of
          centralized wastewater treatment)
      4  Nutrient enrichment/eutrophication (from agricultural and lawn and garden
          fertilization, leaching of wastewater from septic systems, industrial point sources, and
          atmospheric deposition of emissions)
      5  Physical alteration of habitat (from dredging and boating activities)
      6.  Fishing/shellfishing (from harvest pressure from commercial and recreational fishing)

      Once the six potential stressor categories were identified, the next step was to select
assessment endpoints to establish the link between scientifically measurable endpoints and the
objectives identified by the resource managers (Suter, 1989; Suter, 1993). Assessment endpoints
are defined by a valued ecological entity (e.g., fmfish) and the specific attributes of interest (e.g.,
diversity and abundance). Specifically, endpoints should be ecologically relevant, they should be
related to the previously defined management objectives, and they should be susceptible to
stressors.  The workgroup selected several assessment endpoints that could be potentially
affected by stressors in Waquoit Bay. Each endpoint is related to several of the original
management objectives (Table 4-1). The endpoints and the reason for their selection are detailed
below.  "                                     .                                .         ,
       Estuarine percent eelgrass coverage. Eelgrass is a rooted vascular plant that is a critical
habitat to  a variety of fmfish and shellfish (Heck et al.,  1989; Thayer et al., 1984) and is
dependent on sufficient light penetration for its survival. Increases m nitrogen stimulates the
growth  of phytoplankton and macroalgae, thereby limiting the amount of light exposure to
eelgrass beds. Because the Waquoit Bay system is groundwater dominated, suspended sediments
do not contribute significantly to water turbidity to cause an effect on eelgrass productivity.
Macroalgae are a major component of shallow estuaries and they respond impressively to
nutrient enrichment  Biomass and composition of macroalgal canopies may become dominant
features that restructure the benthos of nutrient enriched estuaries (Valiela et al., 1997a;
Hauxwell et al   1998, 2001) and block  light penetration, effectively shading out the eelgrass.
Thus, eelgrass is highly sensitive to the  decrease in light availability that results from nutrient
enrichment and it is an effective indicator of eutrophication.
        Finfish  diversity and abundance. Finfish are susceptible to the dissolved oxygen
concenration (DO) in the water, with the DO levels reflecting the balance between daytime
photosynthesis and the respiration of decomposing organic matter. The minimal DO levels
 which stress fish often occur just before dawn as the result of nighttime respiration and limited
 reaeration of the bottom waters by physical mixing processes.

        Table 4-1. Relationship between assessment endpoints and management objectives
;, ^ •£- \ ' *fl; ,< ;>4-'/t-;,t ..->' t" .»>$ "V ; ',/
;,.; ,4^SSn^Bt:S4p*to : X :' -
Estuarine percent eelgrass coverage
Finfish species diversity and
Scallop abundance
Anadromous fish reproduction
Piping plovers and wetland bird
distribution and abundance (both
relate to migratory bird habitat)3
Fish and shellfish tissue
j- ••>


,'M '
^-2( -

, 4 >

'V , •


' :^'S-:


-7 f-


i I"


/': •" '


 The migratory bird assessment endpoint includes both piping plovers and wetland bird species. These endpoints
 address the same management objectives, and so are combined here, but they have different stressors, and so are
 considered independently in the stressor and endpoint matrix.

        Scallop abundance.  Bay scallops were once a commercially important invertebrate in
 Waquoit Bay. However, because of their dependence on eelgrass beds, the decline in scallop
 populations has reduced their commercial importance (Thayer et al. 1984). Bay scallop larvae
 attach to seagrass blades because they prefer the slower currents induced by the eelgrass
 meadows and it provides protection from predators (Shumway, 1991). Thus, bay scallops are
 susceptible to losses of eelgrass habitat. They also are susceptible to the periodic anoxic events
 that occur as a result of eutrophication. In past anoxic events in Waquoit Bay,  there has been
 significant scallop mortality (D'Avanzo and Kremer,  1994).
       No surveys have been conducted in Waquoit Bay to measure the assessment endpoint of
 scallop abundance directly, so scallop harvest was used as a proxy variable to estimate scallop
 abundance. This is based upon the relationship of C=qfN where C is the catch, f is the effort, q is
 the catchability (assumed to be a constant), and N is the abundance. This equation simplifies to
 N=C/f or the relative abundance is estimated from the catch per unit effort. In our case, the only
 available information on effort was the number of permits issued, which has remained relatively
 constant over time. Data were available on landings (in bushels) but not the catch rate (landings
 plus discarded, undersized shellfish). It would have been preferable to measure effort in days
 fished, but such information was not available.  Thus, an approximation of catch per unit of
 scallop harvest effort was used as a surrogate measure of effect for the assessment endpoint,
 scallop abundance.
       Anadromous fish reproduction. Anadromous brook trout and alewife (river herring) are
commercially important species that use the Waquoit watershed as a breeding ground. These
species depend on swiftly flowing water that is high in dissolved oxygen.

       Piping plover and wetland bird habitat distribution and abundance.
Many avian species nest or forage on the beaches and wetlands of Waquoit Bay. As urbanization
pressure increases, these species and other coastal wildlife will be threatened because of habitat
loss In the Waquoit watershed two barrier beach birds, the piping plover (endangered) and the
least tern (threatened), are listed under the Endangered Species Act, although no species of
wetland birds are currently listed. Because the habitats of barrier beach birds and wetland birds
vary, both were considered as assessment endpoints.
       Tissue contamination offish and shellfish. Tissue contamination offish and shellfish was
a concern voiced by many stakeholders in the watershed. Toxic plumes emanating from MMR
carry trace quantities of carcinogenic substances. Although there is little evidence that these
contaminants have been incorporated into the food web of aquatic organisms, the possibility of
contamination was sufficient for the working group to list tissue contaminant concentration as an
assessment endpoint.  Additionally, bacterial contamination of estuarine water has led to some
shellfish bed closures, resulting in further concerns about water quality.
       These endpoints overlap broadly and may not be mutually exclusive, but they served to
start the process of evaluation of risk. In addition to the endpoints just defined, the workgroup
initially considered other assessment endpoints, including freshwater pond trophic status and
freshwater stream benthic invertebrate diversity and abundance. Freshwater pond trophic status
is addressed in more detail by AFCEE (2001). The authors later combined stream benthic
invertebrate diversity and abundance with anadromous fish reproduction.  The authors eventually
opted to focus on the latter because of the ecological and economic importance of those fish

       To formulate an analysis plan, the authors began by ranking the comparative nsk of each
 stressor on each endpoint. To further analyze these interactions, each stressor was ranked based
 on its intensity, extensiveness, and likelihood of increase over time.

 4.3.1. Comparative Risk Ranking
        The workgroup performed a comparative risk ranking that reviewed the stressors in terms
 of their potential risk to all resources in the watershed, based on best professional judgment and a
 fuzzy set decision-making procedure (Harris et al., 1994). The analysis involved ranking the
 stressors by their impact on the freshwater and estuarine assessment endpoints (Table 4-2).
 Through consensus of the workgroup, each stressor endpoint pair was given a score ranging from
 minimal effect (1) to severe effect (5).  Scores were summed across endpoints to develop a
 cumulative ranking for each stressor. The comparative analysis of the cumulative ranking
 suggests that nutrient enrichment, with a total score of 22, is likely to have the largest potential
 aggregate effect.

               Table 4-2.  Effects matrix for the Waquoit Bay watershed2
Stressor **
Chemical pollution
Altered freshwater
Physical alteration
of habitat
Fishing pressure
:"£'..!;.; r. _,*^^;^^i^K> v, v ?



Csfc ,*

f feirds ,




 "Each cell represents the relative effect of a stressor on an endpoint.  The ranking (l=minor, 5=severe) reflects
 experience with the likely effects specifically for the Waquoit Bay watershed.
       The ordinal scale used here should be taken with a degree of skepticism. First, some of
the endpoints are shown as aggregates (e.g. wetland birds, fish/shellfish contamination), but
subcomponents of each aggregate may have quite different responses. For example, shellfish do
not respond like finfish to contamination because shellfish are full-time residents of the bay,
whereas many commercially important finfish are transient, so they are less likely to be affected
by contaminants within the watershed. Additionally, other contaminants come into the estuary
from outside the watershed boundaries.  The transfer of stressors across watershed boundaries
poses a limitation of the ecological risk assessment process, and has an impact on mitigation
efforts. For example, no method exists to locally mitigate the effects of atmospheric deposition
of nitrogen and mercury, both of which are derived far outside the watershed.
       Because there may be differential responses that are lost in the aggregation of endpoints,
the authors felt that the comparative analysis in Table 4-2 was not robust enough to form the
basis for the risk analysis and risk characterization portions of the risk assessment. To reinforce
the conclusion that nutrient enrichment was the dominant stressor, the authors considered a few
more features of stressor effects, including intensity and extensiveness, and the likelihood that
the source of stress will change across time.  These considerations sharpen the focus of the
effects of stressors.

Table 4-3. Relative importance of identified stressors to the Waquoit Bay ecosystem3
  Nutrient enrichment
 Chemical pollution
rieavy metals
Chlorinated solvents
red freshwater flow
  Physical alteration of habitat
      Propeller Scarring
      Shading of Benthos
      (Docks and marinas)
      Benthic Disturbance

• • •-^^^-^^^^—
^^^^-^^— ^^— ^^^^^^^
  Fishing pressure
     -^^^^•^^^•"*  •"••
 "Ranking is based on available evidence and local experience (l=minor effect, 5=severe effect).
 b Intensity in estuaries.
 c Intensity in fresh-water bodies.

 4.3.2.  Relationships Between Stressors and Ecological Responses
        Next, each stressor in the context of the intensity of the effect on the assessment
 endpoints, the extensiveness of the stressor in time and area throughout the Waquoit Bay
 ecosystem, and the likelihood that the source of stress will change over time is evaluated. The
 extensiveness of a stressor was only considered if the intensity was strong enough to have some
 noticeable impact. Thus, if the intensity of a stressor did not surpass a certain threshold as
 defined by evidence of impact, it was not deemed significant enough to warrant consideration
 under extensiveness. The results of the evaluation and a risk ranking of the identified stressors
  are presented in Table 4-3.

------- The Importance of Stressors in Waquoit Bay
        Chemical pollution includes two categories of potential stressors to the Waquoit Bay
 watershed: heavy metals and chlorinated and other hydrocarbons. Some heavy metals
 (methylated forms) can accumulate in the tissues of organisms and can be passed on to predators
 (Riisgard and Hansen, 1990). Potential sources of heavy metal contamination in Waquoit Bay
 include atmospheric deposition of mercury from industrial processes (Golomb et al., 1997), lead
 residue from paints and from past use of leaded gasoline (Legra et al., 1998), and tributyltin,
 which is used in antifouHng paints for boats, docks, and marinas (Wuertz et al., 1991).  Land use
 in the Waquoit Bay watershed is predominantly residential, with very little industry. There is no
 available evidence to suggest that concentrations of heavy metals have increased to toxic levels
 in these waters, thus heavy metals receive the lowest ranking for intensity.  Since there is no
 evidence of an impact from heavy metal contamination, the threshold for consideration of
 extensiveness was not reached. Additionally, there is no point-source effluent entering Waquoit
 Bay resulting in a small local contribution of these contaminants and a low extensiveness
 ranking. Because heavy metals are now regulated, there is no reason to believe that there  will be
 an increase in their concentrations with time.
        Chlorinated solvents are present at rather low concentrations in some of the plumes
 emanating from MMR, and some of the plumes are moving under the Waquoit Bay watershed.
 These solvents are being treated under the direction of AFCEE (Appendix E).  The goal of the
 mitigation effort is to lower the concentrations of these solvents to below the maximum
 allowable contaminant levels (0.002 ppb for ethylene dibromide, one of the fuel-related
 contaminants). Volatile organic carbons (VOCs) are present only at rather low concentrations,
 and they are highly volatile, so even if they reach Waquoit Bay they are not expected to enter the
 food webs.  The authors therefore suspect that the consequent effects of these compounds  will be
 slight, if measurable at all.  As with heavy metals, other sources of organic contaminants are
 heavily regulated, so there is no reason to believe concentrations will increase in the future.
       Altered freshwater flow can occur through the construction of flow-control structures in
 streams, through removal of groundwater from the aquifer, or through an increase in the
 proportion of impervious surfaces in the watershed. Although flow-control structures and
 domestic and public water supply wells could have an intense effect if they were constructed,
 they are so closely regulated as to not warrant consideration as a potential stressor to Waquoit
 Bay. The removal of aquifer water via domestic water wells is a somewhat more extensive
 occurrence, as some of the houses in the watershed have drinking wells. Currently, most of the
 wastewater released from buildings (including houses) on the watershed is treated with the use of
 on-site septic systems. With septic systems, the wastewater recharges the aquifer through direct
 leaching of the wastewater into the ground. There is one small waste treatment plant at MMR
that discharges sewage waste outside the watershed. If centralized waste treatment becomes the
dominant form of wastewater removal instead of septic use, the aquifer will no longer be
recharged with the leachate, and freshwater flow would be impacted.  Additionally,  with
increased urban development, the proportion of impervious surfaces in the watershed would rise,


resulting in decreased groundwater recharge and increased overland runoff. Unless artificial
recharge of groundwater with treated wastewater or excess surface water is implemented to
replemsh overdrafted aquifers, the alteration of freshwater flow has a moderate likelihood of
increasing across time.
       Nutrient enrichment, which is the addition of nutrients to aquatic systems that results in
eutrophication, takes into consideration the effects of both nitrogen and phosphorus enrichment.
Because the supply of dissolved inorganic nitrogen (DIN) and dissolved inorganic phosphorus
determines rates of primary production in estuaries and ponds, these  nutrient concentrations may
be a telling indicator of eutrophication (Valiela et al. 1992, 2000b; Tomasky et al., 1999;
Foreman et al., submitted). There are two primary sources of nutrients in the water columns of
ponds and estuaries. The nutrients can be either new (or allochthonous), meaning that they are
being discharged from land into the receiving water, or recycled (or autochthonous), meaning
they&are recycled within the water column from decaying plants and  algae and from waste
excreted by water column organisms, or they can be regenerated from the sediments. Our focus
is on allochthonous nutrients since in the Waquoit system, neither dissolved oxygen in the water
nor loss of subtidal habitat have significant impacts on nutrient concentrations in the water when
compared to the effect of land-derived nutrient loading (LaMontagne and Valiela, 1995).
Because nitrogen and phosphorus are the nutrients that limit primary production in salt water and
fresh water, respectively, the relative risk posed by these stressors to the ecosystem are ranked
        Nitrogen is limiting to primary producers in most coastal waters (Howarth, 1988). The
intensity of nitrogen overenrichment results in increases in primary production and the
eutrophication of lakes and streams, an effect that has been well documented in reports on
Waquoit Bay (Valiela et al., 1992,1997b) and in other scientific literature (Duarte, 1995; Nixon,
 1995; NRC, 2000). Previous research in Waquoit Bay has shown that nutrient overenrichment
has resulted in the eutrophication of some subestuaries of the Bay (Valiela et al., 1992,1997b),
thus the highest ranking possible is assigned for both intensity and extensiveness. Finally,
regulation of nonpoint-source nitrogen is underdeveloped, so the potential for increasing rates of
 supply of nitrogen to coastal waters is high.
        Phosphorus tends to be limiting to primary production in freshwater systems (Howarth,
 1988).  The effects of phosphorus enrichment in freshwater systems are similar to those of
 nitrogen in estuaries: they both result in increased primary production and eutrophication. In the
 Waquoit watershed, the intensity of phosphorus enrichment is similar to that of nitrogen.
 Phosphorus, however, is unavailable under aerobic conditions because it is immobilized in
 sediments (Stumm and Morgan, 1981). When sediments become anaerobic, phosphorus may be
 released and move downstream (Krom and Bemer, 1980).  Because the sediments in the Waquoit
 system are very sandy, there is virtually no overland flow of water.  As a result, the sources of
 phosphorus enrichment are not fertilizers from lawns and crops, but wastewater systems that
 allow phosphorus to seep into the groundwater. Most groundwater, however, is aerobic;
 therefore the extensiveness of phosphorus enrichment is not as great as that of nitrogen.  The


 largest exogenous source of phosphorus was the use of detergents, much of which is now
 regulated, so the likelihood of increases in phosphorus concentrations is small.
       Physical alteration of aquatic habitat can result from three major pathways:  boating
 activities, docks and marinas, and dredging.  Boating activities produce a moderately
 intense—but extremely localized—impact on the Waquoit Bay watershed. Propeller scarring can
 significantly disturb the sediments of shallow coastal waters, but this occurrence is infrequent
 and is predominantly seasonal.  Likewise, fuel spills from boat motors also are seasonal. Thus,
 boating activities receive a moderate ranking for intensity and the lowest ranking for
 extensiveness. With increasing human populations on coasts, the number of boaters and the
 frequency of occurrence of boating incidents likely will increase.
       Docks and marinas impact coastal waters through shading of benthic communities  and
 through the introduction of toxins from antifouling agents, especially copper, chromium, and
 arsenic.  Docks have been shown to have a minor impact on benthic fauna (Butler and Connolly,
 1996), but they can have some impact on eelgrass beds (Short and Burdick, 1996). Although the
 use of private docks is moderately extensive, they are now closely regulated and are not likely to
 increase in the future.
       Dredging results in intense physical disturbance of benthic communities and it can  alter
 estuarine circulation. In Waquoit Bay, however, there are very few areas that could be dredged,
 so the authors assigned a low rating for extensiveness. As with the construction of docks and
 marinas, dredging is highly regulated and is not likely to increase in the future.
       Fishing pressure in Waquoit Bay can potentially affect stocks of shellfish and finfish.
 Shellfishing is the dominant of the two fisheries in the bay.  The shallow waters make it an ideal
 location for commercial/recreational and subsistence fishing for quahogs, scallops, and softshell
 clams, all of which are extensively harvested. Shellfishing can result in substantial disturbance
 to the benthic community of estuaries, so the authors consider it a moderately intensive activity.
 Continued growth in the coastal zone likely will cause increases in harvest pressure in the near
       In Ashumet Pond many of the native fish are planktivores or detritivores, and the fish
 community has changed over time for a variety of reasons (e-mail dated 7/9/02 from S. Hurley,
 MA Fisheries and Game Division to David Dow, Woods Hole Laboratory). Several of the
 piscivores targeted by recreational fishers (brook, rainbow and brown trout, and smallmouth
 bass) are stocked by the state, with the predominant native species being yellow perch, brown
 bullhead, chain pickerel, and banded killifish and the introduced species being represented  by
 alewife and green sunfish. Possible evidence for impacts  of phosphorus enrichment can be seen
in the loss of the original native brook trout and the increases in brown bullhead in recent years,
but stocking and introductions make it difficult to attribute changes in the fish community to one
factor. The proposed mechanism is that bottom water anoxia/hypoxia prevents cold-water  fish,
such as the stocked brown trout, from surviving over the summer, but many of the native warm-
water fish can move up from the bottom waters during the summer and escape the anoxia.

      Changs in the fish species community composition over a 20-year time period (1967 to
1987) showed that 20 species exhibited decreased abundance while only 7 species mcreased in
abundance (Deegan and Buchsbaum, in press). In general, the piscivores with an offshore
distribution (cunner, pollock, white hake, winter flounder) exhibited the greatest declines, wtme
the inshore resident, planktivorous/benthivorous forage fish species (rainbow killifish, blueback
herrin,  three-spined and four spined stickleback) showed increases.  The offshore species use
the estuary as a nursery area and since many of the piscivores are targeted by recreational users
and commercial fishing operations, these species can be impacted by both offshore harvesting
and inshore habitat degradation in their nursery areas.  Deegan and Buchsbaum (in press)
attribute a greater relative role for these declines in abundance to habitat degradation, loss of
wetlands and submerged aquatic vegetation, low dissolved oxygen concentrations, toxic
pollutants; or nutrient enrichment, while Link and Brodziak (2002) attribute a greater role to
regional commercial and recreational fisheries harvesting. Since offshore and estuanne systems
frequently mix, it is not easy to delineate the relative importance of fisheries harvesting and
habitat degradation on these declines in the abundance of piscivores. The causes for increases in
the abundance of resident fish species are not well understood, but probably reflect changes in
the abundance of piscivores inshore, alteration in the predation pressure, and changes m
competition from other resident forage fish species that are negatively impacted by inshore
habitat degradation.                                                     .    ,„,,,,.
       Habitat loss could remove valuable nursery grounds for the many species of fish that
depend on macrophytes for their survival. It also has been theorized that replacement of native
marsh species with the invasive Phragmties may reduce the habitat quality for finfish that feed
 on invertebrates in marsh habitats (Weinstein and Balletto, 1999). The impact of Phragrmtes
 invasion in Waquoit Bay ponds and estuaries has not been evaluated.
        Finfishing in Waquoit Bay is not as intensive as shellfishing, and hence the authors
 assigned it a lower ranking. At the 1993 meeting, much concern was expressed about the
 abundance of key recreational species, such as bluefish and striped bass; however, most of the
 species are over-fished at the regional level by commercial and recreational fishers outside the
 watershed.  As with shellfishing, finfishing is likely to increase m the future.
        Pathogens in the environment can be divided into two categories: public health and plant.
 Public  health pathogens have a variety of forms, all of which are estimated by using fecal
 coliform concentration  as a proxy for the extent of contamination caused by wastewater input.
 Although fishing beds are occasionally closed in some of the estuaries of Waquoit Bay because
 fecal coliform loadings exceed the limits for seafood consumption standards for shellfish, none
 of the area beaches have been closed to swimmers, indicating that the contamination is not
 intense enough to cause human heath concerns. Because of the regular monitoring of shellfish
 beds it does not appear that shellfish exposure to public health pathogens m estuaries will
 incre'ase over the next several decades.  Additionally, many pathogens that are considered to be
 threatening to public health have no effect on the invertebrates that carry them; thus, they are

 outside the purview of an ecological risk assessment, but they should be considered in a human
 health risk assessment.
        Plant pathogens have appeared sporadically in the past. Specifically, slime mold
 Labyrinthula (Short et al., 1988), which infects eelgrass beds, destroyed much of the eelgrass in
 Waquoit Bay and elsewhere in the 1930s. Continued spread of this fungus could destroy existing
 eelgrass meadows, but evidence of the disease does not presently extend beyond a few locations
 in the Northeast (Short et al., 1986). Focus on Nutrient Enrichment as the Dominant Stressor
        The aggregate score for nutrient enrichment on the stressor/endpoint matrix in Table 4-2
 is 22. None of the other stressors rank as high; in fact all rank substantially lower. The risk
 ranking in Table 4-3 also identifies nutrient enrichment as the most important stressor in the
 Waquoit Bay ecosystem, as it has a maximum rating for all three contexts. These results were
 derived from local experience with the actual effects as well as from an abundance of previously
 collected data. Because of the overwhelming importance of nutrient enrichment as a stressor to
 the Waquoit Bay watershed, the authors focused the risk analysis on the various effects of
 nutrient enrichment on the two endpoints that are linked to the nutrient enrichment stressor:
 percent eelgrass cover and scallop abundance (as represented by the surrogate measure of the
 effect of scallop harvests) (Figure 4-1). The other assessment endpoints selected by the
 workgroup are not directly susceptible to the nutrient enrichment stressor, so the decision was
 made to focus on these two endpoints in the risk analysis and risk characterization.

 4.3.3. Summary of the Analysis Plan
       The results of the previous section indicate that the effects of eutrophication on
 components of the Waquoit Bay ecosystem are of critical concern. In the risk analysis (Chapter
 5), quantitative information on the impact of nitrogen loading on each of the critical components
 is provided. A modeling approach to assess the causes and effects of nitrogen loading in
 Waquoit Bay was used through a nitrogen loading model (NLM) and an estuarine loading model
 (ELM). The focus was on nitrogen loading because phosphorus input, although important in the
 eutrophication of freshwater ponds, is being analyzed and mitigated by AFCEE.  AFCEE has a
 research program in place that aims to mitigate the phosphorus that is being carried in the
 Ashumet Valley plume (Appendix E).
       In Chapter 5, the risk to percent eelgrass cover and scallop harvest is characterized.
These endpoints were selected for further analysis because they are the most susceptible to the
effects of eutrophication and because sufficient historical information exists that correlative
models can be developed to predict their response to increases in nitrogen loading.  Many of the
other components of the ecosystem, such as phytoplankton biomass and zooplankton dynamics,
are too temporally variable to be modeled effectively.


                                   5.  RISK ANALYSIS

        In the analysis phase, this assessment uses a nitrogen loading model (NLM) to estimate
 exposure to the primary stressor and an estuarine loading model (ELM) to predict impacts of
 nitrogen loading on two assessment endpoints. Ecological responses were  characterized by
 comparing effects in three subwatersheds subjected to different nitrogen loads.  This Chapter
 describes the NLM and the ELM and discusses how they were validated and the level of
 uncertainty in these estimates. The predictions from the NLM and the ELM also are included in
 this Chapter.

        Measures of exposure take on a new definition when considering risk assessment for a
 watershed.  When conducting a large-scale assessment like this one, it is helpful to consider the
 exposure of the entire system as a unit (Suter, 1993). For this Waquoit watershed assessment,
 the measure of exposure for coastal eutrophication is land-derived nitrogen loading to the bay's

 5.1.1. Nitrogen Loads
        The next few sections, describe a model that was created and validated against measured
 data to assess eutrophication exposure. These results may be relevant for estuarine systems that
 lack the resources to undergo the intensive study that the Waquoit Bay system has received. Estimates Using the Nitrogen Loading Model
        Researchers working on the WBLMER project developed a method for estimating the
 magnitude of exposure to nitrogen loading in the form of an NLM that estimates the amount of
 nitrogen delivered to the estuary on the basis of the land use within the watershed (Figure 5-1,
 Table 5-1).  The model sums the nitrogen from three major sources—atmospheric deposition,
 septic-derived wastewater, and fertilizer use—and subtracts all of the losses that occur during
 transport through the various watershed tessera (an individual part of a larger mosaic).  Each
 tesserae is considered as a component along a depth transect from surface to aquifer and is given
 a loss coefficient for each source of nitrogen in the model. After tracking losses and
 transformations of the nitrogen, the model yields an estimate of total nitrogen entering the estuary
 at the seepage face.
       The input terms that are required by this NLM are basic land-use data, which can be
derived using either geographic information systems  (CIS) or aerial photographs. The most
essential piece of information needed to accurately determine the extent of each land-use type is
accurate watershed delineations. For this watershed the authors used local information on water
table contours, hydraulic conductivity, and effective porosity to simulate the tracks  of water flow

                        ATMOSPHERIC DEP.
                                                                     seepage face
                                   i loading model. Inputs of wastewater-, fertilizer-, and
           ally derived nitrogen to tne watershed and the percent losses (shown as numbers) as
           from the three mahi sources percolates through the watershed and enters the estuary
at the seepage face.
Source: Valiela et al. (2000a) with permission of Kluwer Academic Publishers.

with a three-dimensional, finite-difference hydrologic flow model called MODFLOW
(McDonald and Harbaugh, 1988). Once the watersheds were delineated, the land use within each
watershed was estimated from aerial photographs (Brawley et al., 2000). GIS technology has
improved markedly since the start of this modeling effort, and watersheds can be dehnated and
land use tabulated with the latest GIS software.
       The specific information on land use required by the model is described in Table 5-1. it
is important to note that these land-use categories can, to some degree, be modified for use in
other watersheds. For example, cranberry bogs are a dominant agricultural crop on Cape Cod,
but the fertilizer inputs from cranberry bogs could easily be replaced by the fertilizer inputs from
com if corn was the dominant agriculture type in the watershed being modeled. The loss terms
used in an NLM  are related to the nitrogen that is delivered below the watershed surface via two
main routes (Valiela et al., 1997b). Nitrogen derived from atmospheric deposition and femlizer
use percolates to the unsaturated vadose zone after losses in the turf and vegetation (Lajfia et al.,

 Table 5-1.  Input, loss, and default terms used by the nitrogen loading model (NLM)
apf*iytiy mw b'x
Drainage area
Number of houses
Area of receiving
Area of freshwater
Area of wetlands
Area of natural
Area of cranberry
Area of other
Area of golf courses
Area of other types
of turf
Area of other
impervious surfaces
* ^ ( - -> - (
Uptake by forests (65%)
Uptake by the vadose zone (61%)
Uptake by turf (62%)
Uptake by freshwater ponds (66%)
Uptake by wetlands (78%)
Loss in the septic tank (6%)
Loss in the leaching field (35%)
Loss in the septic plume (44%)
Volatilization of fertilizer (38.5%)
Loss in the aquifer (35%)

' -. "" "" $LMi -* " ,
Evapotranspiration (45% of
Nitrogen in deposition (15 kg
Area of lawns (0.05 ha/house)
Area of roofs and driveways
(185.8 and 46.45 m2/house)
Area of roads (3% of watershed
Occupancy rate (1.79 people/
Per capita biological nitrogen
release (4.82 kg/person/yr)
Fertilizer application to lawns
(122.33 kg N/ha/yr)
Fertilizer application to golf
courses (171 kg N/ha/yr)
Fertilizer application to
cranberry bogs (28 kg N/ha/ yr)
Fertilizer application to other
agricultural land (136 kg
Source: Valiela et al. (1997b).

1995). Nitrogen from wastewater undergoes losses and transformations as it moves through
septic tanks and leaching fields and as it moves through the soil layers in a wastewater plume
(Valiela et al., 1997b). Nitrogen from all three sources undergoes losses in the vadose zone and
in the aquifer (Valiela et al., 1997b).  The specific loss terms used in the Waquoit Bay NLM are
summarized next and described in more detail in Valiela et al. (1997b).

Losses considered by the NLM
       Uptake by forests — The retention of atmospheric nitrogen delivered to coastal forests in
the watershed of Waquoit Bay was determined to be 65%. This estimate was found by
comparing concentrations of total dissolved nitrogen (TON) in atmospheric deposition to those in
water collected below the root zone (Valiela et al., 1997b). The estimate of 65% retention of
atmospheric nitrogen by coastal forests is lower than the 80 to 90% retention reported for upland

forests of New England (Aber et al., 1993), but similar to the 68% retained in a riparian forest in
Georgia (Lowrance et al., 1984).
       Uptake in the vadose zone — The fate of nitrogen during travel through the vadose zone
is not well known (Keeney, 1986; Korom, 1992), although losses of nitrogen during transport
through unsaturated vadose layers of sands have been reported in agricultural fields (Cameron
and Wild, 1982; Starr and Gillham, 1993). From these data, it was estimated that 61% of the
nitro-en that reached the vadose zone of forested or cultivated parcels in the Waquoit Bay
watershed was lost in the unsaturated sediments (Valiela et al., 1997b).  The NLM therefore
calculated 39% of atmospheric nitrogen percolated from forested or cultivated areas of the
watershed through the vadose zone. It was assumed that these values were applicable to the
other land-cover types, and the calculation was repeated for the remaining land uses to obtain an
estimated total loss in the vadose zone.
       Uptake by turf— By comparing nitrogen concentrations from the drainage through soil
under lawns and a grass field (79.6 mM TON, from Lajtha et al., 1995), with concentrations in
atmospheric deposition (209 mM TON, Valiela et al., 1997b) onto the Waquoit Bay watershed,
losses of 62% of atmospheric-derived nitrogen in turf plants and soils were calculated.
       Uptake by freshwater ponds — Freshwater bodies located within coastal watersheds
capture groundwater flow and any nitrogen dissolved in that water. The amount of nitrogen that
traverses the freshwater bodies and moves downgradient is less than the nitrogen that enters the
ponds, lakes, and wetlands. Mass balance studies in lakes, ponds, and various types of
freshwater wetlands show that retention of nitrogen entering the waterbodies ranges from 14 to
 100% (Valiela et al., 1997b). The median retention of nitrogen in ponds and lakes is 56%
 (Valiela et al., 1997b). In these studies, retention was underestimated because inputs by dry
 precipitation were not included.
        Uptake by wetlands — Similar to mass balance studies of lakes and ponds, wetlands
 have been shown to retain a median of 77% of the nitrogen that they receive (Billen et al., 1985;
 Billen et al  1991; Johnston et al., 1990; Johnston, 1991; Molot and Dillon, 1993).
        Loss in the aquifer — Because there is insufficient published information with which to
 estimate the magnitude of nitrogen loss in aquifers (Korom, 1992), the loss for Waquoit Bay was
 estimated using a steady-state approach. The WBLMER data were used to calculate that 35% of
 diffuse nitrogen entering the aquifer under the Waquoit Bay watershed was lost within the
 aquifer (Valiela et al., 1997b). This estimate of losses in aquifers was obtained by calculating the
 percent difference between nitrogen concentrations in groundwater near the water table and
 groundwater about to enter the receiving estuaries. These data were compiled from samples of
 groundwater originating under land parcels covered by natural vegetation.
        Loss in the septic tank — To ascertain if losses of nitrogen occurred within septic tanks,
 published nitrogen concentrations in wastewater entering septic tanks were compared to .the
  concentrations in effluent leaving the tanks (Valiela et al., 1997b).  Concentrations of total
  dissolved nitrogen in water entering septic tanks were about 5% higher than those in effluent
  leaving septic tanks.  These results were smaller than the 10 to 20% reductions suggested by


 others (Andreoli et al., 1979). Even when the septic holding tanks were pumped out, removal
 averages only 4% of nitrogen entering septic systems (Kaplan, 1991), and this septage is often
 kept within the watershed through release to septage lagoons.
        Loss in the leaching field — The leaching field disperses the water that has drained from
 the septic tank, allowing it to percolate slowly into the soil.  To calculate the loss of nitrogen in
 the leaching field, two values of estimated nitrogen retention were averaged.  One estimate was
 based on all retention data available to us on the nitrogen content of septic-tank effluent and on
 effluent from leaching fields. The second estimate used data from a group of 12 papers (e.g.
 Walker et al., 1973; Reneau, 1979; Starr and Sawhney, 1980) that provided concentrations of
 nitrogen in both inputs to holding tanks  and outputs from leaching fields for a given septic
 system. The loss in leaching fields obtained from the aggregate data (35%) and the
 system-specific nitrogen retention  (46%), were then averaged to obtain 40% as an estimate of
 nitrogen losses in septic systems of conventional design. Hence, about 60% of wastewater
 nitrogen is likely to travel beyond leaching fields and into the septic plume (Valiela et al.,
 1997b). This could be a true loss caused by denitrification, or merely a dilution effect. However,
 studies have shown there is little evidence of dilution within the leaching fields of this watershed
 (Valiela etal.,1997b).
        Loss in the septic plume — The water from the leaching field percolates  through the
 vadose zone and into the aquifer. The water moves for a distance in a confined plume of higher
 concentration known as the septic plume. Concentrations of dissolved inorganic nitrogen in the
 septic plumes have been found to decrease with distance away from leaching fields (Valiela et
 al., 1997b).  As with losses in the leaching field, such decreases could be a true loss caused by
 denitrification, or an apparent loss  caused by dispersion.  Each possibility was examined by
 comparing decreases of DIN with decreases in concentrations of chloride or sodium when these
 tracers were also reported.  The results showed that chloride concentrations decreased over
 distance (Valiela et al., 1997b), so dispersive mixing did take place, and actual losses of nitrogen
 also took place. The mean  value of septic nitrogen that may be lost during travel in plumes was
       Volatilization of fertilizer — Gaseous losses by denitrification and volatilization from turf
 account for about 39% of fertilizer  nitrogen (Petrovic, 1990).

 Defaults considered by the NLM
       The defaults that are used by the NLM are specific for the Cape Cod region where the
 model was developed. The model is designed in such a way that these defaults can be changed to
 model another watershed. The derivation of the  defaults for the Waquoit Bay watershed are
 described next.
       Evapotranspiration — Regional estimates of evapotranspiration of 45% were derived
from Thomthwaite and Mather (1957), Running  et al. (1988), and Eichner and Cambareri (1992).
       Nitrogen in deposition — Watersheds receive significant amounts of nitrogen from the
atmosphere.  Deposition of atmospheric nitrogen has varied over recent decades and wide


regional differences in chemistry of precipitation have been reported (U.S. EPA, 1982; Shannon
and Sisterson, 1992; Davies et al., 1992; Fricke and Beilke, 1992; Pack, 1980; Stensland et al.,
1986; Ollinger et al., 1993). Due to these differences, it was necessary to calculate nitrogen input
via atmospheric deposition using local deposition data.
       Local data on wet deposition (Lajtha et al., 1995) and a calculation of dry deposition were
used to estimate total atmospheric deposition to forests. Because local dry deposition data were
unavailable, estimates for the ratios of wet-to-dry deposition from published data were
calculated, and a mean ratio of 0.9 was obtained (Valiela et al, 1997b). This agrees with Hmga
et al. (1991), who concluded that wet and dry deposition were of the same magnitude. Thus,
atmospheric deposition was estimated to be twice as high as wet deposition.
       Area of lawns — Lawns in suburban to semirural coastal areas such as Long Island and
Cape Cod average 0.05 ha in area (Koppelman, 1978; Frimpter et al.,  1990; Interdisciplinary
Environmental Planning (IEP) 1986; Eichner and Cambareri, 1992).
       Area of roofs and driveways — Roofs in suburban to semirural coastal areas such as
Long Island and Cape Cod average 185.8 ha in area, and driveways average 46.45 ha in area
(Koppelman, 1978; Frimpter et al., 1990; IEP,  1986; Eichner and Cambareri, 1992).
       Area of roads — Area roads are estimated as 3 percent of the watershed area
(Koppleman, 1978).
       Occupancy rate — The total number of residences or buildings within a watershed can be
estimated from either aerial photos or GIS data (Lindhult and Godfrey, 1988).  The number of
people per house can be obtained from municipal records, census data, or by using data from
other areas of similar ecological setting and socioeconomic background (Valiela and Costa,
 1988- Hayes et al, 1990).  In some coastal areas, many houses are occupied only during summer,
 vacations, or on weekends. In addition, coastal areas such as Cape Cod are home to many retired
people who spend the winter in warmer climates.  In such cases, it is necessary to adjust
 occupancy rates. Census data from Cape Cod in 1990  were used to determine the duration of
 occupancy of individual houses in the Waquoit Bay watershed. Using these data along with GIS-
 based land parcel data and aerial photos, a weighted mean (± sd) was calculated as 1.8 ± 0.6
 people/house/yr. This value is similar to the seasonally adjusted value of 1.9 people/house/vr
 suggested for an area where  23% of the houses were used only during the summer (Weiskel and
 Howes, 1991).
        Per capita biological nitrogen release — On average, 4.8 kg  N/person/yr are released,
 with a range of 1.8 to 5.4 (Valiela et al., 1997b). These values are similar to those found by
 others (1.8 to 7.4 kg N/person/yr) as observed by Vollenweider (1987), Koppelman (1978), and
 U.S. EPA (1980).
        Fertilizer application to lawns - Mean fertilizer use applied to lawns in suburban areas
 such as Cape Cod average 122.33 kg N/ha/yr (Koppleman, 1978; Nelson et al, 1988).
        Fertilizer application to golf courses — The average nitrogen fertilizer content applied to
 golf courses is 171 kg N/ha/yr (Frimpter et al, 1990; Eichner and Cambareri, 1992; Petrovic,


        Fertilizer application to cranberry bogs — Fertilizer use on cranberry bogs was reported
 in Valiela etal. (1978).
        Fertilizer application to other agricultural land — Calculations of nitrogen loading from
 agricultural fertilizers should use local application rates (Loehr, 1974; Stanley, 1988; Frimpter et
 al.,  1990; Hayes et al., 1990; Correll et al., 1992) as much as possible. In suburban areas, where
 agricultural land use is at a small, local scale and crops are mixed, an average of about 136 kg
 N/ha/yr may be appropriate (average obtained from data in Loehr, 1974;  Stanley, 1988; Hayes et
 al.,  1990; Correll et al., 1992). Measurements
        Nitrogen loading rates to the Waquoit Bay estuarine complex were measured by obtaining
 groundwater samples roughly every 50 meters from the periphery of each of the subestuaries in
 the bay (Valiela et al., 2000a).  The timing and frequency of groundwater sampling varied by
 subestuary, but the differences  did not affect the analyses since there appears to be  no seasonal
 variation in groundwater concentrations (Table 5-2). Using standard methods, the  samples were
 analyzed for concentrations of NO3 (QuikChem® method 31-107-06-1-C), NH4 (QuikChem®
 method 31-107-04-1-C), and DON (modified from D'Elia et al., 1977) (Valiela et al., 2000a).
 The subwatersheds were divided into smaller recharge zones to prevent differences in
 groundwater flow or the number of samples taken in each recharge zone from biasing the
 estimates of nitrogen loads. These zones were delineated from land surface features and
 hydrological flow lines using the MODFLOW groundwater transport model  (McDonald and
 Harbaugh, 1988). The authors  used the area of each recharge zone, the average annual
 precipitation rate, and regional  estimates of evapotranspiration to estimate the volume of water
 that  flows through the aquifer and into the estuary. The volume of water  and the average
 concentration of nitrogen in the groundwater samples taken within each recharge zone were
 multiplied to determine the total nitrogen load (Valiela et al., 2000a).
       As shown in Table 5-2,  the measured nitrogen loads to the subestuaries of Waquoit Bay
 are extremely variable, ranging from 433 to 9879 kg N/yr (Valiela et al., 2000a) and are
 determined in part by the size on the land parcel.  The nitrogen load to the entire bay was more
 than 26,500 kg N/yr. NLM Validation
       Modeled predictions of the quantity of nitrogen entering the estuary were  validated in two
independent ways.  First, the NLM predictions were verified against actual measurements of
nitrogen in groundwater about to enter estuaries (Figure 5-2) (Valiela et al., 2000a). The authors
quantified the performance of the NLM by comparing a suite of statistical features that were
derived from the verification plot. Four statistical inferences were used to quantify  performance.

      Table 5-2.  Measured nitrogen loads to the subestuaries of Waquoit Bay
 Waquoit Bay

Source: Modified from Valiela et al. (2000a)
Second r was calculated, which evaluated the precision of the model by assessing the degree of
scarcer around the regression line between measured and modeled data. Third, the accuracy of
the model was tested by using a t-test between the slope of the regression line and^the 1:1 fane ot
perfect fit.  Finally, the authors examined the R2 term that, based on Prairie (1996), gives a
measure of the predictive ability of the model.
       The NLM was extremely responsive to measured data, based on highly significant ±v,
values (116.6, p=0.01). In addition, the NLM precisely predicted measured loads, with a very
small degree of scatter around the regression line, as demonstrated by highly significant  r values
(0 97 p=0 01)  The slope of the regression line comparing model estimates to those that were
measured in the field could not be statistically distinguished from the 1:1 line of perfect  fit,
indicating that the model is highly accurate (t.1.79, not significant). Finally  the R value was
substantially higher than the 0.65 proposed by Prairie (1996) as the threshold for the 7***™
ability of regression models. The results of these four statistical features indicate that despite the
associated uncertainty, the NLM accurately represents the processes involved in 1^-denved
nitrogen loading and can thus be used effectively as a measuring tool (Valiela et al., 2002).
        The second verification of the NLM assessed the ability of the model to accurately predict
the dominant source of nitrogen to the estuary. The NLM was used to predict the percentage of
nitrogen derived from wastewater and compared this prediction to the 6  N signature of the
 groundwater measured in the subestuaries of Waquoit Bay (McClelland and Vahela, 1998).  The
 percent of nitrate as 5'5N derived from wastewater is higher than the percentage derived from

         100000 -L.
                                                 A  Whole watershed
                                                 •  Subwatersheds
                                                 o  Recharge areas
                         Predicted N load (kg yr  -1)
   Figure 5-2. Measured versus modeled nitrogen (N) loads to Waquoit Bay. Plotted
   against a one-to-one line of perfect fit.
   Source: Valiela et al. (2000a) with kind permission of Kluwer Academic Publishers.
fertilizers and from atmospheric deposition (Table 5-3), making it possible to verify the
percentage of nitrogen derived from wastewater predicted by the NLM. Each watershed has land
uses that contribute different proportions of nitrogen derived from the three main sources; these
mixes are sufficiently distinct and can be linked to nitrogen loads entering the estuaries.  The
high correlation between the 515N signature of nitrate measured in groundwater and the
proportion of groundwater nitrogen derived from wastewater as calculated by the nitrogen
loading model, helps verify that the NLM can describe the proportion of nitrogen coming from a
given source (Figure 5-3).
       The reported results on 8i5N as tracers of land-derived nitrogen are not isolated examples.
Hansson et al. (1997) showed that isotopic signatures in organisms in different areas of the Baltic
Sea reflected the nitrogen contributions from sewage, and the signatures were pervasive in the
food web up into the top predators.  Udy and Dennison (1997,1998) showed that isotopic
signatures in macrophytes identified sources and loading of nitrogen to Australian estuaries.

             Table 5-3. Sources and 815 N values of nitrate in groundwater
10 to 20

 2 to 8
Includes nitrate derived from both human and animal waste.
Kreitler (1975), Kreitler and Jones
(1975), Gormly and Spalding (1979),
Aravenaetal. (1993)	
Kreitler (1975), Kreitler and Jones
(1975), Gormly and Spalding (1979)
Kohl et al. (1973), Freyer and Aly
(1974), Mariotti and Letolle (1977),
Macko and Ostrom (1994)
  . ^i^—^—^^—^^—
Unpublished data from Brazilian tropical coastal lagoons show that 2 to 6% header isotopic
signatures in macrophytes were collected from a lagoon with densely populated villages on its
shore relative to macrophytes collected far from human populations. Water samples collected in
coastal lagoons on the coast of Sinaloa, Mexico, revealed that 615N values of ammonium were
about 2 to 3% in areas remote from people but increased to 21 to 29% in water near wastewater
sources  Similar trends were found in water as well as macrophytes collected from freshwater
ponds and coastal lagoons in other Cape Cod regions. Thus, the isotopic approach furnishes both
an independent verification that an NLM can accurately describe the proportion of stressor
coming from wastewater, and that it can provide a means to estimate the incorporation of the
stressor on the biota in a variety of estuarine systems. NLM Uncertainty
       Uncertainty is a necessary component of every model. Traditional means of addressing
issues of uncertainty involve estimating the variation in replicate samples. Since it is impossible
to have replicate watersheds, the uncertainty of the NLM was estimated using a bootstrapping
method in which the loading calculation was obtained using 2,000 estimates of loads, each of
which was based on a different set of means. The calculations were based on the set of observed
values used for each algorithm variable. Then, the 2,000 bootstrapped means were used to
calculate measures of variation (Table 5-4). The standard error of the mean for the NLM,  as
calculated by the above bootstrapping method, was 12%, and the standard deviation of the mean
was 37% (Collins et al., 2000). The error was also propagated and similar results were found.
       Uncertainty in model estimates can be introduced in three  ways: statistical uncertainty,
 inherent uncertainty, and model uncertainty (Collins et al., 2000). Additional data reduce  the
 level of uncertainty in any model, but information is frequently difficult or expensive to acquire.
 One form of sensitivity analysis is a technique based on Monte  Carlo simulations that gives an
 indication of the parameters of the model that are the most uncertain, thereby providing a  means











  tQR2 »CR2
                          % of groundwater N coming from wastewater
             Figure 5-3. Values of 615N of nitrate in groundwater versus
             percent of nitrogen (N) from wastewater. The relationship of the
             815N signature of nitrate measured in groundwater of the Waquoit
             Bay estuarine system to the proportion of groundwater nitrogen (N)
             derived from wastewater was calculated by the nitrogen loading
             model. Each point represents a recharge zone of Waquoit Bay.  The
             open circle is upper value provided by the literature for 615N of
             wastewater-derived N; the dashed line shows the extrapolation for
             data from Waquoit Bay. F=22.6? p=0.01, r=0.91.

             Source: Valiela et al. (2000a) with kind permission of Kluwer Academic

for focusing research on the terms that most strongly influence the model output.  Such analysis
indicated that eliminating the uncertainty in any one variable would result in small changes
(Collins et al., 2000).  The analysis demonstrated that the error associated with the amount of
nitrogen released per person per day accounted for the majority of the  uncertainty in the final
loading estimate. However, lowering the uncertainty in load estimates required lowering the
uncertainty in several other model variables.
       The effects of temporal variability also could introduce uncertainty into the model results.
It takes longer to feel the impact of land-use changes that occur farther back in the watershed

             Table 5-4. Error analysis of NLM variables
 Propagation of
  standard error
   Standard error
   95% CI
   Standard error
   95% CI
Source: Data from Collins et al. (2000).

than it takes for those occurring on the banks of the estuary, because of the length of time it takes
the groundwater to travel. To address this uncertainty, a dynamic version of the NLM was
developed (Brawley et al., 2000).  This dynamic model incorporates spatial and temporal trends
in land use into the NLM framework. The results of this exercise indicate that the variability
associated with the lag due to groundwater travel was less than the uncertainty in the model,
therefore, it is concluded that temporal variability is relatively small.
       The problem of uncertainty associated with model development is not unique to the
NLM. In fact, every model designed to approximate the impact of land use on the nitrogen loads
of estuarine systems is plagued by the same dearth of data. The difference between the NLM and
other  models designed to predict nitrogen loads and concentrations is that the NLM is verified
against actual measured data. It seems plausible that, despite the associated uncertainty, if a
model can reasonably predict nitrogen loads that are occurring in the environment, then it has
served a valuable purpose.

       The effects of nitrogen loading on the ecosystem components outlined in Table 4-4 was
measured to look at changes occurring throughout the Waquoit Bay watershed. For our
purposes, nitrogen load is treated as the exposure, and nitrogen concentration is an effect of that
load.  Thus, a discussion of the cascading effects from nitrogen loading begins with a description
of the nitrogen concentration in the estuary. The analysis of the cascading effects is followed by
a characterization of the effects on the two assessment endpoints upon which the risk assessment
focuses (Section 5.2.2).

        A space-for-time substitution (Picket!, 1989) was used to compare estuaries that differ
 spatially in the degree of exposure to nitrogen loading as a proxy for the loading change that will
 occur in the ecosystem over time.  Because different subestuaries within the Waquoit Bay
 watershed have been urbanized at different rates, the situation was ideal for establishing
 ecosystem changes that could occur over time without having to conduct long-term ecosystem
 studies. The three estuaries that are most frequently used in this comparison are the nearly
 pristine Sage Lot Pond, the moderately urbanized Quashnet River, and the highly urbanized
 Childs River watershed (Figure 2-1, Table 5-5). Next, empirical data from these subestuaries
 were used to show the responses of the ecological endpoints to land-derived nitrogen loads.

 5.2.1.  Cascade of Effects on Ecosystem Components Nitrogen Concentrations
        In this assessment of Waquoit Bay, land-derived nitrogen load is the exposure that is
 studied. This section describes the first effect of that loading as the measured nitrogen
 concentration.  The use of the measured nitrogen to validate the Estuarine Loading Model
 (ELM), which predicts the concentrations of nitrogen available to primary producers, is also
        Measurements - Nitrogen concentrations in the estuary differ markedly from groundwater
 concentrations.  Once groundwater traverses the seepage face, biotic and abiotic processes occur
 in the estuary that influence the concentration of nitrogen available to primary producers. We
 wanted to assess whether increasing the nitrogen load to estuaries would result in increasing
 concentrations of nitrogen in the water. Indeed, water samples collected monthly from the
 surface and the bottom of each estuary from 1991 to 1995 indicate that as nitrogen loads
 increase, the mean concentration of nitrogen in estuarine water also increases (Foreman et al.,
 submitted) (Figure 5-4).
        Estimates using the Estuarine Loading Model (ELM) - The NLM calculated the nitrogen
 load that enters the estuary at the seepage face.  To then address the concentration of nitrogen in
 the estuary over an annual time scale, an additional estuarine loading model (ELM) was
 developed.  The ELM calculates mean annual concentrations of DIN available to producers in
 shallow estuaries (Valiela  et al., in press). The ELM requires inputs of land-derived nitrogen to
 the estuary obtained by the NLM. It then accounts for direct atmospheric deposition to the
 estuary surface as determined for the NLM, N2 fixation, denitrification and burial of nitrogen in
 wetland and subtidal sediments, regeneration from sediments, and water residence time to
 estimate DIN concentrations in the estuary (Figure 5-5).
       ELM Validation - Estimates of mean annual DIN concentrations calculated by the ELM
 were verified against values measured in Waquoit Bay estuaries (Figure 5-6). Like the NLM, the
ELM is extremely responsive to measured data (F=75.6, ^=0.01). It should be noted that
 although there are only five data points in this regression, each represents a subestuary of
Waquoit Bay, and each data point, in turn, represents the mean of surface and bottom samples

     Table 5-5. Characteristics of subestuaries of the Waquoit Bay watershed3
Area of the watershed (ha)
Number of houses
 % of land that is agriculture
 Measured nitrogen load (kg N/ha/yr)
See Figure 2-1 for watershed locations.
200      400       600

              Figure 5-4. Dissolved inorganic nitrogen (DIN)
              concentrations in the three subestuaries of Waquoit Bay
              subject to land-derived nitrogen (N) loads.

  WW A.D. Pert.
                                         N Fix.  A.D.      N Fix.  A.D.

                                           Dnf. Bur.        Drif. Bur. Net
Figure 5-5. Schematic of inputs and exports of the ELM and NLM. The ELM combines the
NLM estimated inputs from land, nitrogen fixation, and direct deposition and subtracts losses
due to denitrification and burial. It then predicts the amount of dissolved inorganic nitrogen
(DIN) that is available in the estuary. The DIN can then be linked to the dominant primary
producers if there is sufficient data available.

TDN = total dissolved nitrogen
PP = phytoplankton
MA = macroalgae
EG = eelgrasses

       14 r   y = 1.017x +0.902**,  r2 = 0.941
  >»   12 I-

  Q>   10

  1     8
    2        4        6

DIN concentration
                                             predicted by ELM
Figure 5-6. Comparison of dissolved inorganic nitrogen (DIN) concentrations
predicted by the estuarine loading model (ELM) with measured concentrations in
the water column of several Waquoit Bay estuaries (p<0.01).
taken monthly from 10 locations within each estuary from 1991 to 1995. The precision of the
ELM is demonstrated by the highly significant r values (0.997, p=0.01). Finally, if the ELM
predictions of average annual DIN concentrations were the same as those measured in the
estuary, all the data points would lie on the 1:1 line of perfect fit.  A linear regression fitted to
these data is
statistically indistinguishable from the 1:1 line (Figure 5-6), as demonstrated by a nonsignificant
result from a t-test between the slope of the regression and a slope of 1 (t=0.97, not significant).
Finally, the predictive ability of the ELM is high, based on Prairie (1996), with an R2 of 0.94
       The HTM, therefore, furnishes a reasonable fit to actual measurements, even though it
contains many terms that are associated with uncertainties—and many untested assumptions
regarding various inputs, losses, and pools of nitrogen—and how these measurements interact in

 estuaries.  Uncertainties characterized the steps taken to construct the ELM; at every stage we
 made simplifications, best guesses, and reasonable assumptions. In fact, all models suffer from
 similar uncertainties, but one contribution of such models is that they, despite the simplifications
 and guesses, help predict natural responses. If model predictions parallel observations, there is
 some reassurance that we have captured a sufficiently complex representation of the workings of
 the systems under study and that the model may be useful for further applications.  Thus, these
 comparisons furnish some confidence that the ELM is sufficiently competent to be useful, even
 though many assumptions were involved in its development. Phytoplankton Biomass and Production
        Mean annual phytoplankton biomass and mean annual production both increased as the
 nitrogen load to the estuary increased (Figure 5-7, top).  It is concluded that there is a link
 between land-derived nitrogen and the response of phytoplankton in the receiving estuaries and
 that increases in urbanization on coastal watersheds result in increased phytoplankton biomass
 and production. The increase in phytoplankton stock is not surprising in view of the widespread
 nitrogen limitation of such producers in shallow coastal  waters (Boynton et al., 1982; Vitousek
 and Howarth, 1991; Valiela, 1995; Aguilar et al.,  1999; Downing et al.,  1999; Tomasky et al.,
 1999).  The range of land-derived loads observed  in the Waquoit Bay estuaries spans about two-
 thirds of the measured loads to various estuaries worldwide (Nixon, 1992). Hence, across most
 of the range of loads to be expected in shallow estuaries, we would  predict increases in
 phytoplankton biomass and production as nitrogen loads increase were predicted. Macroalgae Biomass and Production
       Biomass and production by macroalgae also increased as nitrogen loads from land
 increased (Figure 5-7, middle). Many references cite widespread nitrogen limitation of
 macroalgal growth in shallow estuaries and lagoons (Harlin and Thome-Miller, 1981; Valiela et
 al., 1992; Peckol et al., 1994; Duarte, 1995; Valiela et al., 1997a). As with phytoplankton, we
 predicted that increases in nitrogen supply to estuaries would increase the standing crop of
 macroalgae. Eelgrass Biomass and Production
       In contrast to the response of phytoplankton and macroalgae, both biomass and
 production by eelgrass sharply decreased as total nitrogen load increased beyond 350 kg N/ha/yr
 (Figure 5-7, bottom). It is important to note that eelgrass may be more sensitive to nitrogen
loading than would be predicted from these data.  Additional features of eelgrass response to
nitrogen loading are considered in Section 5.2.2, where the effects of exposure on the assessment
endpoints are characterized.









 g 200-

 f 16.


                                         2.4    a

         200        400         600

1.8   3

1.2   1


-0.6   ^


 -  2








Figure 5-7. Effects of nitrogen (N) loading on biomass and pmnary

production of phytoplankton, macroalgae, and ee^rass in Sage Lot

Pond, Quashnet River, and Childs River (top to bottom).

Source: Modified from Valiela et al. (2000b) using N-load data from Valiela et al.


                                   200         400         600
                                       N load (kg ha'1 y1)

               I     | Phytoplankton   |_.   | Macroaigae          Seagrasses

         Figure 5-8. Partition of total primary production in shallow estuaries
         into contributions by phytoplankton, macroaigae, and seagrasses, all
         plotted against measured annual nitrogen (3V)  load. Data are from
         Waquoit Bay, and from Buttermilk Bay and Bass Harbor, two other Cape Cod
         Source: Adapted from Valiela et al. (2000b). Combined Effect of Nitrogen Loading on Primary Producers
       These data are presented in synthetic form in Figure 5-8. As nitrogen loads increased, it
 was predicted that the proportion of primary production carried out by the three selected
 producers would shift, with eelgrass quickly disappearing and production becoming dominated
 first by macroaigae, then by phytoplankton.
       The impact of land-derived nitrogen loads, however, is mediated by other factors, in
 particular by water-residence times (Ketchum, 1951; Pace et al., 1992). Residence time is a
 measure of how quickly water is flushed from the estuary, and it is inversely related to the water
 turnover rate. In systems with short residence times (Tr), such as Waquoit Bay, the effects of
 nitrogen loading are reduced in two ways. First, a short Tr dilutes the concentration of available
 nitrogen when there is a rapid exchange with seawater (e.g., Vineyard Sound).  Second, if the Tr
is short enough, then phytoplankton that would typically increase in biomass and production in
response to the added nitrogen source are unable to significantly multiply before they are swept
out of the estuary. Thus, our predictions in Figures 5-7 and 5-8 refer to shallow estuaries  and
lagoons with residence times of less than 5 days. For waters with a longer Tr, the responses of

the phytoplankton to land-derived nitrogen loads would be considerably more intense, with larger
biomass and higher production rates. Zooplankton Egg Production
       Short 1> such as those found in the Waquoit subestuaries, may have an even more
pronounced effect on organisms that have longer generation times, such as zooplankton.  Egg
production of the female Acartia tonsa, the dominant copepod, clearly responds to the food
abundance provided by the phytoplankton, which increase because of larger nitrogen loads
(Figure 5-9 top). Thus, the copepods in each estuary are coupled to the ambient food supply
conditions in that estuary.  This response did not translate into a parallel response of abundance,
as there were no significant differences in the number of copepods among estuanes (Figure 5-9,
bottom). It is inferred from these responses that the short T,s of the Waquoit subestuanes means
that Acartia tonsa population abundance is not directly controlled by food supply, although
clearly that is what controls egg production. Shellfish Growth Rates
       The response of shellfish to increases in nitrogen loading is variable and depends largely
on the species being considered. Research performed in subestuaries of Waquoit Bay and in
other nearby estuaries indicated that the growth rates of softshell clams (Mya arenand) and in
quahogs (Mercenaria mercenaria) increase as nitrogen concentrations increase (Figure 5-10)
 (Weiss et al  2002; Evgenidou and Valiela, in press), although there has been no trend in
 reported hardest of these species to coincide with the increases in growth rates. Additional
 research on the ribbed mussel (Geukensia demissd) shows similar patterns. Ongoing research in
 several Cape Cod estuaries indicates that bay scallop populations decline when they are
 transplanted to estuaries that have high nitrogen-loading rates (Shriver et al., 2002). More
 features of the dynamics of bay scallops are considered in Section Finfish Abundance
        Recent research on finfish in Waquoit Bay indicates no clear relationship between
 nitrogen load and the abundance or growth rates of the two most common estuanne finfish
 species Atlantic silverside (Menidia menidia) and mummichog (Fundulus heteroclitus) (Tober et
 al  2000; Griffin and Valiela, 2001). This is probably due, in part, to the lack of a direct
 relationship between food supply and growth rates for copepods, which are a dominant food
 source for these estuarine fish.  These species are not commercially harvested so that fishing
 pressure is not responsible for the lack of a relationship.

 5 2 1.9.  Summary of Effects on Ecosystem Components
        The previous analysis suggests that in the Waquoit Bay system, the likely responses to
  nitrogen loading include a high risk of increases in phytoplankton and macroalgae biomass, a

     40 n
        0       40       80       120 0       200     400
      Mean chl a cone, (mg m~3)     Nitrogen load (kg ha
' 4-,
£ 3-
"D 0
C£. ~

-5 n.
3 o
3 °

3 O
9 8
n ..-.. — i 	 n-^ 	 1

600    0
                 Nitrogen loading rate (kg ha "1 y ~1)
Figure 5-9.  Characterization ofAcartia tonsa in Waquoit Bay estuaries.
Top: egg production in estuaries exposed to different nitrogen rates (right) and
chlorophyll concentrations (left) (Cubbage et al., 1999). Bottom: abundance of
copepodids (left) and adults (right) relative to land-derived nitrogen load in Waquoit
estuaries (Lawrence 2000).

Source: Valiela et al. (2001).

 loss of seagrass habitat, a mixed risk of effects on shellfish, and a smaller risk of effects on fish
 and zooplankton.

 5.2.2.  Effects on Assessment Endpoints
        The assessment focuses on two measures, percent eelgrass cover and scallop harvest,
 which are discussed in the sections below. Percent Eelgrass Cover
        Previously, it was reported that eelgrass biomass and production decrease as nitrogen
 loads increase, and in the space-for-time substitution presented at the bottom of Figure 5-7, this
 decrease occurred before nitrogen loads reached 350 kg/ha/yr. To more completely assess the
 critical load at which eelgrass begins to decline, data from various estuaries were compiled that
 extend the Waquoit Bay results (Valiela and Cole, 2002). It is evident that for a wide variety of
 estuaries, eelgrass is highly sensitive to increases in nitrogen loads (Figure 5-11) and that the
 aerial extent of eelgrass was sharply reduced at loads greater than 20 kg/ha/yr.  The meadows
 disappeared completely by the time nitrogen loads exceeded 100 kg/ha/yr. Scallop Harvest
        The bay scallop was once one of the most commercially valuable species in the Waquoit
 Bay estuarine system.  Scallops are a benthic invertebrate whose preferred habitat is eelgrass.
 Records have been maintained by the Town of Falmouth Shellfish Warden on the volume of
 scallops harvested (the surrogate measure of effect for this assessment endpoint) over the last
 several decades (Figure 5-12). The quantity of scallops harvested has decreased drastically over
 the last several years, from almost 200,000 IVyr in 1965 to less than 20,000 L/yr in more recent
 years (Bowen and Valiela, 2001a). Summary of Effects on Assessment Endpoints
       As exposure to nitrogen loading increases, there appears to be a loss of eelgrass area,
 especially as nitrogen loads exceed 20 kg/ha/yr. In addition to the loss of eelgrass area, there also
 has been a reduction in scallop harvest.  Scallop larvae attach to seagrass blades and prefer the
 slower currents induced by the presence of eelgrass.  Additionally, eelgrass beds provide
protection from predators (Shumway, 1991).  There is evidence that as the  area of eelgrass
decreases so does the yield of the bay scallop (Bowen and Valiela, 2001a).

                % sg/oss = 0.693 Nload + 14.211  F=14.6*
                      200        400        600

                      N load (kg  N ha-V1)
Figure 5-11.  Percent seagrass cover lost as nitrogen (N) load increases for
a multitude of temperate and tropical ecosystems.

Source: Adapted from Valiela and Cole (2002).


   w   0.1
 Figure 5-12. Volume of bay scallop harvest in Waquoit Bay from 1965-

  Source: Bowen and Valiela (2001a).

                             6. RISK CHARACTERIZATION

        Risk characterization integrates the measures of exposure and the measures of effect to
 estimate risk to the assessment endpoints. It also serves to summarize and describe the results of
 the risk analysis in such a way that it can be readily translated to managers and other
 stakeholders. In the risk characterization for the Waquoit Bay watershed, the NLM and ELM
 models were used to correlate the temporal changes in exposure to the effects of that exposure on
 the assessment endpoints. Finally, the models were used to assess a variety of options that can be
 employed to reduce exposure within the watershed.

        The issue of scale, both spatial and temporal, is critical to understanding and predicting
 the dynamics of ecosystem perturbation. To understand the effects of eutrophication on coastal
 watersheds, we must consider how these changes have occurred over time. To  this end, we
 applied the previously described models to explore historical changes in nitrogen loading and the
 effects of that loading on the two assessment endpoints.

 6.1.1. Changes in Exposure: Back-casting Nitrogen Loads
        The relative contribution of nitrogen to an estuary from atmospheric deposition, fertilizer
 use, and wastewater disposal depends on the specific mosaic of land covers present on the
 watershed surface. There has been a worldwide transition within watershed mosaics from
 naturally vegetated landscapes to human land uses, both agricultural and urbanized. In many
 inland areas, there has been a similar transition to agricultural land covers, and hence, fertilizer
 inputs become a major feature of the loading to watersheds and their receiving waters (Correll
 and Ford, 1982; Jordan and Weller, 1996; Jordan et al., 1997). In coastal areas, urban sprawl has
 become an increasingly dominant feature of the land-cover mosaic, and wastewater is becoming
 a major nitrogen source. Near the end of the 20th century, as much as 37% of the world
 population resided within 100 km of a shoreline (Cohen et al., 1997), and human populations
 increased markedly in the near shore of every estuary around the world (Nixon,  1986; Valiela et
 al., 1992).  It necessarily follows that the wastewater released from these increasingly urbanized
 coastal zones has and will increase. The net results of the geographic land-cover transitions, both
 toward agricultural or urbanized landscapes,  have been that, first, nitrogen loads have increased,
 and second, the relative contribution from each of the three major sources—atmospheric
 deposition, fertilizer use, and wastewater disposal—has changed.
       As shown in Figure 6-1, the Waquoit Bay watershed has undergone such a land-cover
transition over the last 60 years. The area of natural vegetation on the watershed diminished
about fourfold from 1938 to 1990. This change was the result of conversion of vegetated land to
human uses. In this case, the geographic transition was not just an urbanization  of the watershed,
but a more complex rearrangement of land-use categories. Turf associated with lawns, parks,






 "E    800 -r
       600 --
       Naturai veg.
        Human uses
         1930  1940  1950  1960  1970   1980   1990   2000

Figure 6-1. Changes in land uses in the Waquoit Bay watershed from
*10'lQ tf\ 10OO
Top- changes'in the area of natural vegetation.  Bottom: breakdown of
human land uses into three major components:  turf (including lawns,
parks, and golf courses), impervious surfaces (including roads, roots
driveways, runways, and parking lots), and agriculture (predominantly
horticultural crops and cranberries).

Source: Valiela and Bowen, 2002 (Reprinted from Environmental Pollution with
permission from Elsevier Science).

 and golf courses increased as the number of inhabitants in the area increased. Land devoted to
 agriculture decreased while land covered by impervious surfaces (roofs, driveways, roads)
 increased.  The NLM and information gathered from aerial photographs was used to estimate the
 change in exposure that has occurred in the last 60 years.  Information on the atmospheric
 deposition  of nitrogen was used to perform the historical analysis needed for the risk
 characterization. These were based on a reconstruction of measured nitrogen deposition in the
 region from the turn of the century to the present, as well as a reconstruction of changes in
 fertilizer use during the same time period (Bowen and Valiela, 2001b). Specifically, atmospheric
 deposition  was reconstructed using compiled historical data that were adjusted to account for
 regional variability and extrapolated over the time course to include dry and organic depositions
 of nitrogen. Fertilizer use was reconstructed by estimating the changes in fertilizer application
 rates for the northeastern United States as a proxy for those on Cape Cod.  As a result of the
 changes in  land use, the total nitrogen load to the watershed of Waquoit Bay has increased
 (Figure 6-2, Table 6-1).
        Throughout the 50-year period, atmospheric deposition consistently contributed the
 largest portion of the nitrogen load. The atmospheric supply of nitrogen after 1970 appears to
 have stabilized, at least in this region. In 1938 nearly all nitrogen delivered to the watershed
 (95%) was  by precipitation, but by 1990, atmospheric sources had dwindled to 59% (Table 6-1).
 Although the implementation of the Clean Air Act in the early 1970s did not directly address  •
 emissions of nitrogen, the indirect effects of increased fuel efficiency in automobiles and
 reductions in other pollutants emitted from industrial processes may have prevented further
 increases in the atmospheric nitrogen delivered to the land surface in this area (Figure 6-2, top).
 Additionally, the reconstruction indicates that there was a decrease in the amount of ammonium
 in rainwater and an increase in the amount of nitrate, at least for the Cape Cod area. This likely
 represents a shift from an agricultural base (where ammonium dominates) to an industrial base
 (where nitrate dominates). Thus, over the whole century, the two forms of nitrogen are in
 balance and the total increase in nitrogen deposition is not as great as expected if one were
 looking at nitrogen oxides alone.
       Fertilizer- and wastewater-derived nitrogen clearly increased over the 50-year period.
 Wastewater nitrogen contributions reached 22% of inputs by 1990, and 19% of inputs were
 contributed  by fertilizer use (Figure 6-2, Table 6-1).
       Nitrogen loads to estuaries differ markedly from nitrogen loads to watersheds, owing to
 significant interception of nitrogen within the watersheds.  Nitrogen losses within watersheds and
 during passage through the various land-cover types were considerable. Almost 90% of
 atmospheric nitrogen was intercepted within the watershed, compared with 79% of fertilizer
nitrogen and only 65% of wastewater nitrogen (Valiela et al., 1997b). This differential through-
put means that, as the nitrogen in groundwater is about to seep into receiving estuarine waters,
the relative proportions of nitrogen from the different sources and land-cover types differ
markedly from the proportions that entered the watershed.


     200 I
      160 -
      120 -
    80 -
  ^    40
           TO WATERSHED








            TO ESTUARY
                                               1 Wastewater
                                                T Atmospheric

                          I960    1970    1980    1990    2000
Figure 6-2. Historical changes in nitrogen (N) loading to
watershed and estuary of Waquoit Bay. Modeled loads to the
watershed (top) and the estuary (bottom). Tlie loads are broken down into
the three major sources of N: atmospheric deposition, wastewater, and
fertilizer application. Note the order-of-magmtude difference between N
inputs to the watershed (top) and to the estuary (bottom). Data are plotted
with the 12% standard error that is associated with the nitrogen loading

Source: Bowen and Valiela (2001a).

        Table 6-1. Relative contributions of each of the major sources of nitrogen to the
        Waquoit Bay estuary in 1938 and 1990a
£V% .„'•',{ V/^r^x . vi'^j

'? Soapee^nlteo^e» "^
To the watershed
Atmospheric deposition
Wastewater disposal
Fertilizer use
To the estuary
Atmospheric deposition
Wastewater disposal
Fertilizer use
•feft^ttro^io^^ ^
,-' * *
:ifl^kgy4 ,


Ji: • % ,




-* . , ' •• ~
-3k \ - '


 aThe propagated error of the modeled N load is 14% (Valiela et al., 1997b). The percent contribution from each
 source of nitrogen is slightly different from those published in Valiela et al. 1997b and those used in Figure 2-5. The
 differences result from die need to use regional trends to incorporate historical changes.  In some instances these
 regional trends were slightly different from the Waquoit Bay specific information used in the original publication.
 The difference between the regional approach, and the Vaiiela et al. (1997b) results fall within the standard error of
 the model.
 Source: Bowen and Valiela (2001a).

       Wastewater contributions to the receiving estuaries were a minor part of the nitrogen
 loads early on, but they had become the main source by the 1990s (Figure 6-2, bottom),
 surpassing both fertilizer and atmospheric deposition as contributors to the load. The total load
 of nitrogen to the Waquoit Bay estuary increased from slightly more than 10,000 kg N/yr (11.6
 kg  N/ha/yr) in  1938 to more than 24,000 kg N/yr (28.2 kg/ha/yr) in 1990. This more than
 doubling of nitrogen loads can be largely attributed to an increase in the wastewater-derived
 nitrogen, which accounted for only 7% of the total load to the estuary in 1938 but had jumped to
 43% by 1990.

 6.1.2.  Temporal Changes in Effects: Impact on Percent Eelgrass Cover and Scallop
       Temporal changes in eelgrass areas during the last 60 years in Waquoit Bay demonstrate
 a sharp decline (Figure 6-3, top). Before the 1950s, eelgrass was still recovering from the near-
complete loss caused by the wasting disease of the early 1930s (Cottam, 1933; Cotton,  1933;
Renn, 1935). Note also that during the 1930-1950 period, nitrogen loads were lower than the 20
kg N/ha/yr that was suggested in Section 5.3.1 as the upper limit of eelgrass survival.  The
building boom on Cape Cod during the 1960s resulted in an increase in nitrogen loads.  By the








   12        16        20        24

       Modeled historical N load (kg N ha1 y1)

        200 -i
                             Eelgrass (ha)

    re 6-3. Decreases in area of eelgrass and volume of scallops harvested
ov  toe as auction of increasing nitrogen (N) loads. Top: perootof
area covered with eelgrass in Waquoit Bay estimated by renalphotographs
Seen ^40 and 1990 (Costa 1988, Short and Burdick 1996) ptotted ag^nst
modeled historical N loads. Bottom: the number of reported scallops harvested
in Waquoit Bay as a function of eelgrass area.  Scallop data are from the
Shellfish Warden's annual reports, in the Town of Falmouth Annual Report.

Source: Adapted from Bowen and Valiela (2001a).

 early 1970s, the nitrogen load exceeded 20 kg N/ha/yr, and eelgrass meadows were notably
 smaller in area. The loss of eelgrass habitat continued through 1990. The historical
 reconstruction indicates that the nitrogen loads corresponding to the near-complete destruction of
 eelgrass meadows ranged only between 15 and 30 kg N/ha/year (Figure 6-3, top).
        The extrapolation was carried one step further to look at the secondary effects of eelgrass
 decline on the decrease in scallop harvest. Because the presence of seagrass is required for the
 maintenance of many taxa, including commercial shellfish and fmfish species, a change in
 eelgrass cover implies drastic changes in the rest of the estuarine food webs in affected estuaries.
 During the time span when eelgrass meadow area decreased, the annual harvest of bay scallops in
 Waquoit Bay decreased (Figure 6-3). We can therefore claim that urban development can be
 demonstrably linked to drastic restructuring of estuarine ecosystems.

 6.1.3. Effects of Other Stressors on Eelgrass
        Other stressors are potentially damaging both to existing eelgrass beds and to efforts at
 reintroducing eelgrass to estuaries, although these stressors are minor in comparison to nitrogen
 loading because they are restricted to very small regions of the bay or they occur only
 sporadically.  For example, remaining beds of eelgrass may be further impacted by natural
 events. In 1991, Hurricane Bob, a category 3 hurricane, made landfall on Cape Cod. The storm
 surge washed over a spit on Washburn Island in Waquoit Bay, resulting in the burial of an
 eelgrass bed on the inside of Eel Pond (Valiela et al., 1996,1998). Although this was an
 extremely severe effect on that eelgrass meadow, the impact was limited spatially, and the
 meadow recovered fully during the next growing season.
        The eelgrass population of Waquoit Bay was nearly destroyed during the early 1930s as a
 result of an infection of eelgrass wasting disease. The epidemic struck many northeastern U.S.
 estuaries earlier last century and is thought to have been caused by the slime mold Ldbyrinthula
 (Short et al., 1988). Evidence of a pathogenic strain of Labyrinthula has been  seen recently in
 scattered locations.  Continued spread of this fungus could destroy existing eelgrass meadows,
 but evidence of the disease does not presently extend beyond a few locations in the Northeast
 (Short et al., 1986), and it does not exist presently in Waquoit Bay.
       Human activities on estuaries can impact eelgrass meadows through physical
 disturbances. Docks and marinas built over eelgrass beds limit light through shading, resulting in
 the eventual destruction of the eelgrass meadow.  Dredging activities result in a direct loss of
 eelgrass through removal of shoots. Dredging also has an indirect effect because of the
 resuspension of sediment particles, which increases water turbidity and decreases the amount of
 available light for eelgrass growth. Propeller scour from passing boats and mooring scars are
 also localized stressors. The impact of shellfishing on eelgrass meadows has not been examined
extensively, although it could have a significant, but local, impact (verbal communication from J.
Hauxwell, Wisconsin Department of Natural Resources, to I. Valiela, Boston University Marine

       It is important to note here that ecosystems are highly complex and variable systems that
can and do change in species composition, distribution, and abundance. It is an established fact
that nitrogen loading is the major stressor on eelgrass and that decreasing the load of nitrogen to
Waquoit Bay may result in water quality conditions that could support eelgrass. This does not,
however guarantee that eelgrass will reestablish itself or maintain itself if replanted.  Many
efforts have been made recently to restore eelgrass beds in estuaries where meadows no longer
       Research indicates that one of the major impediments to reconstruction of eelgrass habitat
is the lack of genetic diversity in the colonized plants.  Williams and Davis (1996) used several
measures of genetic diversity (percentage of polymorphic loci, allele richness, expected and
observed heterozygosities, and proportion of genetically unique individuals). They found that
overall genetic diversity was significantly reduced in transplanted eelgrass beds.  In addition,
Smith et al (1989) compared the habitat value of natural meadows and recently transplanted
eelgrass meadows for the commercially important bay scallop.  Stocking adult scallops in
recently transplanted meadows was not successful, indicating a lag time may be needed between
transplantation and functional habitat use. Thus, the management options outlined in the next
chapter do not guarantee the return of eelgrass with the lowering of nitrogen loads; they simply
reduce loads to levels where eelgrass might be able to survive.

                          7. MANAGEMENT IMPLICATIONS

        At the outset of the risk assessment process, the workgroup proposed a management goal
 that aimed to reestablish and maintain water quality and habitat conditions in Waquoit Bay, to
 support diverse, self-sustaining commercial and recreational fisheries and native fish and
 shellfish populations, and to reverse ongoing degradation of ecological resources of the
 watershed. To achieve this goal, the workgroup initially evaluated the various sources of stress
 and the effects of that stress on the ecology of Waquoit Bay. The results of this  analysis indicate
 that nitrogen loading is the biggest threat to the ecology of Waquoit Bay.
        The results of the risk assessment indicate that the bulk of excess nitrogen enrichment
 stems from nonpoint pollution sources, namely, land-use activities within the watershed.
 Because 35% of the U.S. population lives within 100 km of the coast, this problem has
 ramifications nationwide. It poses a challenge for local and state managers who are responsible
 for land-use decisions in the coastal zone and for federal and state agencies that are responsible
 for natural resources (NOAA's NMFS and the U.S. Fish and Wildlife Service) or for pollution
 control (U.S. EPA, NOAA, and Massachusetts Coastal Zone Management).
        The Waquoit Bay watershed ecological risk assessment provides a tool for helping local
 managers recognize the consequences of decisions about land use within the watershed on the
 water quality of coastal embayments. This document should aid managers in resource planning
 and provide useful information supporting zoning decisions. The models developed here are
 applicable to watersheds that are dominated by land use in rural and suburban  areas. These
 models can analyze watersheds that lay above sandy soil in which groundwater flow dominates
 the hydrology and in which the receiving water has a relatively short residence time (less than a
       The tools provided by this risk assessment allow local managers to anticipate the impact
 of future land-use changes within the watershed and explore various treatment options to
 minimize future impact as well as reduce present levels of nitrogen loading to Waquoit Bay.
 Through a series of possible scenarios, managers can use the combination of models to generate
 solutions that will restore water quality to levels that existed in the 1960s.

       The historical reconstruction of land-derived nitrogen loads, plus the linking of these
 loads to assessment endpoints,  such as percent eelgrass cover and bay scallop harvest, provides
 some means for identifying management priorities and defining potential restoration measures.  If
 management goals include reduction of nitrogen loads, even a cursory look at the results of this
 work suggest some general approaches that could be used to remediate the increased nitrogen
 loads. A first step in evaluating the potential effects of remediation options is to examine the
relative nitrogen loads from wastewater disposal, fertilizer use, and atmospheric deposition
 (Table 6-1). Model estimates using 1990 data indicate that wastewater contributed 43% of the
total nitrogen in Waquoit Bay.  Wastewater inputs would thus seem to have a higher priority than


fertilizer use; at most, management of the fertilizer inputs could be expected to lower nitrogen
loads by 19%.                                                      ,. .       ,.
       There are many restoration measures that could reduce the loads of nitrogen from
wastewater and fertilizer use. Management of residential wastewater and fertilizer use is the
most practical measure available locally.  Zoning restrictions, improvement of m situ septic
systems, and installation of small sewage treatment plants could all lower the inputs of
wastewater nitrogen. Restrictions on the use of fertilizer and turf area would also to a smal er
degree, lower nitrogen inputs to Waquoit Bay, but this management option would be especially
effective in more agriculturally intensive areas.  The choice of the appropriate management
option depends on the target endpoint a community wants to achieve.
       Although direct management of atmospheric deposition is impossible on a local scale,
preservation of green, open space can make a contribution to management of atmospheric
nitrogen loads. Atmospheric nitrogen deposition is still the largest nitrogen source to fte
Waquoit Bay watershed. Interception of atmospheric nitrogen is highest where naturid
vegetation covers land parcels, and this measure should be a part of the managers  tool kit when
thinkin^ about the coastal zone.  Preservation of green space, especially vegetated buffers along
 streams may slow the increase in nitrogen enrichment to receiving waters by intercepting
 a^pheric deposition and by limiting the supply of nitrogen that would result from conversion
 ^ land uses, such as agricultura, or residential. Other means of remediating Imogen inputs
 from atmospheric deposition require legislative actions that affect a multistate area, because the
 airshed that contributes nitrate pollution to a region is significantly larger than the area that «
 under local management control. However, if actions that could mitigate sources wi to the
 watershed prove insufficient, citizens in the Waquoit watershed will have no choice but to
 consider legislative or cooperative actions to mitigate sources outside the watershed.
        In the initial part of the risk characterization, linkages between nitrogen loadmg and the
 defined assessment endpoints were developed. Here we develop an approach to using these
 relauonships to help managers make informed decisions.  Through the historic* ^con—
 we have established that nitrogen loads and concentrations have increased over time.
 and stakeholders can examine the historical record and select, using Figure 7-1 as
 the percent eelgrass bed coverage they want to adopt as a management target  Onc
  status is selected, we can use Figure 7-2 to find the year that the nitrogen load corresponded to
  the selected eelgrass target. Then we can investigate what management practices could be
  S o ed o ow!r nitro Jen loads to the target  lev*. For example, suppose the restoration fcrget
  isTo return to condition; that could support the growth of eelgrass on 30% of the main estuary
  (Figure 7-1).  The corresponding load would be about 18, 000 kg N/yr. To achieve the restoration
  target, the nitrogen loads from the watershed to the estuary need to resemble those existing
  around 1970 (Figure 7-2) . The management  goal in that scenario would be to reduce the
  nitrogen loads by  approximately 6,500 kg N/yr. The task then becomes to explore using
  "mltions with the NLM^LM,  the management practices that might be put in place to achieve
  those restoration measures.

      80 n
  f  40^
                                 16          20          24
            Modeled historical N load (10 3 kg N  y ~1)
       80 n

          24 -
              1940    1950    1960    1970    1980    1990   2000

    Figure 7-2.  Historical changes in nitrogen (N) loading predicted by the
    nitrogen loading model. If the management target is a return to 30% eelgrass
    cover, then managers must reduce N loads to 18,000 kg N/yr, loads that are
    comparable to those of around 1970.
7.1.1. Reducing Fertilizer Application Rates
      Reducing the application rates of fertilizer, or the amount of land that is fertilized, will
reduce the nitrogen loading to Waquoit Bay, although at most the reduction will be only 17% of
the total nitrogen load (Figure 7-3).  Notice that fertilizers from lawns and golf courses combined
contribute almost the same amount of nitrogen as that derived from agriculture, and that even the
complete abatement of all fertilizer use in the Waquoit Bay watershed would reduce the nitrogen
load only to the level of the early 1970s.  The benefits of this option must, of course, be weighed
against the economic costs of decreased agricultural output and the potential loss of tourism
resulting from less attractive greens on local golf courses.

7.1.2. Managing Wastewater
       About 25% of the population of the United States disposes of its wastewater via in situ
septic systems of conventional design. In many coastal areas, the use of septic systems is
widespread.  For example, in Cape Cod only 2% of the land surface is serviced by wastewater
plants (Mitchell, 1999); septic systems are used in the remaining areas (Heufelder and Rask,
1996,1997). Wastewater disposal through septic systems has in recent decades become the
major source of nitrogen entering the estuaries of Waquoit Bay (Table 6-1). The most likely


                                                 Fertilizer applied to:
                                                      Golf courses

                                                      All sources
Z  18
_g>    100
                80       60       40       20        0

                % of present dose that is applied
                                             Decrease land area of:
                                                      Golf courses

                                                      All sources
       100      80       60       40       20

             % of present land that is fertilized
Figure 7-3.  Reduction in the total nitrogen (N) load that would result
from varying the amount of fertilizers used in the Waquoit Bay
watershed.  The reduction in total nitrogen load in the Waquoit Bay
watershed that would occur if varying percentages of fertilizer were
withheld (top) or the area of fertilized land were reduced (bottom).

Source: Data from Bowen and Valiela (to be submitted).

option for managing wastewater nitrogen is altering the nitrogen retention in septic systems.
       Potential changes in nitrogen loads that would result by assuming different degrees of
nitrogen retention in septic systems were assessed. The NLM was used to run simulations on
various nitrogen retention values ranging from the 39% nitrogen removal achieved by septic
systems of conventional design presently being allowed by the Commonwealth of Massachusetts
up to complete retention (Figure 7-4).
       The first set of simulations took into consideration all of the buildings on the watershed.
The next set of simulations considered only those buildings up to 1,000 m, 500 m, and 200 m
from shore.  The purpose of these simulations was to ascertain whether buildings closer to the
shore disproportionately contributed to the nitrogen load.  Replacing or retrofitting septic systems
in the entire watershed might be an expensive and politically daunting alternative. Management
of wastewater nitrogen loads from smaller areas closer to shore might be a more feasible
        The simulations suggest that if retention of wastewater nitrogen in all septic systems were
to increase above  values characteristic of current systems (39%) (Valiela et aL, 1997b), the
nitrogen loads to Waquoit Bay would decrease from values extant in  1990 to those of 1958,
roughly by 42% (Figure 7-4).  If only those septic systems located within 200 m of shore
(roughly 2 years' travel time within this aquifer) were considered, the reduction in nitrogen load
would amount to values seen in 1975, a reduction of about 17%. The simulations concerning
other distances yielded intermediate values.
        The results of the simulations, shown in Figure 7-4, suggest that areas of watersheds
closer to shore contribute proportionately more to estuarine nitrogen loads than areas farther from
shore.  This result may derive from two sources. First, buildings in this watershed (and
elsewhere) are more numerous closer to water (Valiela et al., 1992).  Second, groundwater
plumes bearing nitrogen from septic systems may reach the shore only in those areas within 200
m of shore; farther away, the plumes disperse into bulk groundwater  and may lose more nitrogen
during  travel. Regardless of the mechanism, this result points to strategies that target near-shore
septic systems as  being potentially more effective than efforts to manage all of the septic systems
in a watershed.
        So far, we have discussed the benefits of complete retention of wastewater nitrogen
within  in situ systems. In reality, it is hard to find a practical way for in situ systems to be so
effective. Some alternative designs have lower nitrogen retention percentages (Table 7-1), and
the corresponding nitrogen loads are lower (Figure 7-4). Given the measures of uncertainty
associated with the various alternative septic-system designs (note the ranges given in Table 7-1),
it is not altogether evident that there are significant differences in nitrogen retention among the
alternative designs.


^ 20-
g, is-
1 16-
_ iff w
03  £
5 w •= J2 ®
l| o> i 1
1| 1 1 1
Q LL [_. 0) Q^
(1990) ! j
^fe^*^ V
(1982) ^^\. if
(1975) \X\
(1969) \\
                                                         Septic systems from:
                                                          0-200m from shore

                                                          0-500m from shore

                                                          0-1000m from shore

                                                           Entire watershed
1 • '"" i 	 '
	 1 	 •
                    % N released from on-site systems
      Figure 7-4. Reduction in total nitrogen (N) load in the Waquoit Bay watershed
      (and corresponding year) that would result from implementing various
      wastewater treatment systems.
      Source: Bowen and Valiela (to be submitted).
       These scenarios are by no means the only options.  They merely suggest that management
of certain design features may achieve selected targets and that the models we developed are
flexible and useful tools for addressing the issues raised by managers. Further, our approach
aims to ensure that stakeholders make the selections for scenarios and options and that we use the
models and science to provide evaluations for stakeholders to consider. Clearly, stakeholders
must decide, and they will surely need to include criteria other than the loadings produced. Our

      Table 7-1.  Onsite septic system retention efficiencies reported for various
      alternative systems
 Massachusetts Title
 5 septic system
 Peat filters
 Trickling filters

 Recirculating Sand
Various published estimates

Six in situ systems on East Coast
Various systems

Multiple samples from four
Maryland systems

Mean from six in situ systems
in Massachusetts
Valiela et al.
Heufelder an(
Rask (1996)

Piluk and
Peters (1995)_
Rask anc
Source: Bowen and Valiela (2001a).
view is that results from the assessment of restoration should enhance optimal decisions, which
must be made by stakeholders.                   .
       Stakeholders frequently suggest that one option for improving water quality in coastal
embayments is to shorten the water residence time through either dredging or altering the inlet
morphology and/or depth. This approach is based on the inverse relationship between the
nitrogen loading rate and water residence time in determining the nitrogen concentrations within
the water column. Hydrodynamic model simulations by T. Asaji at Applied Science Associates,
foe indicate that increasing the cross-sectional area of both the Eel Pond and the Waquoil.Bay
inlets by 100% would reduce the residence time of the system from 45.42 hours to 40.46 hours
CTable 7-2). This would lower the residence time by only 11%, with a 100% increase in the area
of both inlets. If the inverse relationship between residence time and nitrogen loading is
accurate, then the nitrogen loads would be decreased by only 11 %-within the margin of error of
the model. Dredging is extremely expensive, causes adverse environmental impacts, and
requires much permitting, fo addition, to keep inlets open, dredging needs to be repeated at
regular intervals, requiring more permits and more money. Dredging, therefore, does not seem
like a desirable approach to resolving water-quality issues in Waquoit Bay.

        Table 7-2.  Changes in water residence time predicted by dredging simulations
" -\ " ' •''*' -V T I: \ ' J ,1
> t . '» , %-<. ,- f
r ; Dredging scenario i* :
Present condition
Increase Eel Pond by 50%
Increase Eel Pond by 100%
Increase Waquoit Bay by 10%
Increase Waquoit Bay by 25%
Increase Waquoit Bay by 50%
Increase Waquoit Bay by 100%
Increase both by 50%
Increase both by 100%
Tidal volume increase
1.; ^' <#> ^'"s <
-\ > V^e^€fri^ftie^KJiWs)t
 Source: Bowen and Valiela (to be submitted).

       When public meetings were held, the workgroup noted that the terminology used in the
 risk assessment created confusion, because neither the public nor the attending managers or
 scientists were familiar with concepts such as stressors, ecological effects, assessment endpoints,
 measures of effect, or exposure pathways. Even with the use of the conceptual model at the
 February 1995 meeting with managers, it was difficult to communicate these concepts.
 Difficulties with terminology were exacerbated by issues of scale, since most of the legislative
 mandates, regulations, and jurisdictional boundaries for local/state/federal managers are not set at
 the watershed level.
       The concerns of the public are more localized, and some people had a hard time with the
 concept of evaluating stressors and their impacts on valued resources at the watershed level. For
 example, the focus of stakeholders who are involved in the Superfund cleanup at MMR are
 human health risks (toxic contamination of their sole-source  aquifer for drinking water and the
 risks of swimming and eating fish from local ponds), whereas constituents near Waquoit Bay are
 concerned about ecological risks (anoxia/fish kills, loss of eelgrass and bay scallops, changes in
 the size and abundance of recreational finfish species, diminished aesthetics, etc.).
       Given the limited public outreach program permitted  by the  resource limitations of this
assessment, getting a consensus on the conceptual model and the resulting risk analysis from
constituents who have diverse localized concerns, scientists who study separate components of

the system, and local/state/federal managers who have differing mandates and jurisdictions!
boundaries was difficult.
       The risk analysis report is lengthy and complex, so some type of translation endeavor will
have to be undertaken to make the project's findings accessible to local citizens and local/state/
federal managers. Uncertainty is a key component of the risk assessment process, and given the
technical nature of the analysis, it is a challenge to convert these data into information that will
be useful to managers. Because the costs that local citizens will bear to diminish nitrogen input
from septic systems and fertilizer use are likely to be significant, it will be critical in the nsk
management phase to clearly articulate the benefits and uncertainties associated with the
NLM/ELM models and to provide socioeconomic information on how the costs will be
distributed among different stakeholders. A sociopolitical process will need to be established to
engage citizens and their elected representatives in resolving these cost/benefit issues.
       The ecological risk assessment is an information-intensive process for evaluating risks to
environmental receptors. The unevenness of the information available for the Waquoit Bay
watershed limits the ability to conduct a comparative analysis of the risks from multiple man-
made stressors. In our case, we had a lot of information on nutrient enrichment and its chemical
and biological effects in ponds and the estuary, but we lacked good information on, for example,
the reproductive status of brook trout and the quality of their breeding habitat, the threats to
wetland bird breeding habitat, and the impacts of volatile organic contaminants on fish health m
Johns Pond.                                                            .           ,  ,
       In addition, some of the stressors that impact the ecological resources  in the watershed
have a regional source rather than a source within the watershed. These would include regional
 commercial/recreational fishing pressure and recruitment, which influence the estuanne finfish
 assessment endpoint, and atmospheric mercury and nitrogen input from the regional airshed
 which contributes to the eelgrass and tissue contaminant assessment endpomts.  To deal with
 these regional stressors, an entirely different level of management would be needed to augment
 the risk management activities of local/state/federal managers.  The public desires a seamless
 process to address these issues, but in practice this is rarely possible, and local actions may result
 in limited benefits for the costs incurred.
        Before the Waquoit Bay ecological risk assessment was begun, nitrogen loading from the
 watershed led to water quality problems and diminished populations of valued biotic resources.
 By developing the coupled NDvI/ELM models, this study provided a methodology for local/state/
 federal managers to better understand nitrogen impacts so they can make wiser decisions in
 addressing the problem and engaging local citizens in the dialogue on mitigation strategies
 (including validating model predictions, long-term monitoring, and adaptive management). The
 model will be available on the WBNERR web site (Geist, 1996). It will be important for
 managers to communicate to the public that these model predictions for nitrogen loading will
 lead to improved water quality, but they may not lead to recovery of the eelgrass beds or bay
  scallop populations, as discussed in Section 6.1.2 and 6.1.3.

One of th                                             maj°r eC°lo§ical s^ssor



     APPENDIX A: Supplemental Information on the Waquoit Bay Estuarine Complex

        The Waquoit Bay watershed covers approximately 53 km2 (21 mi2) and spans parts of the
 towns of Falmouth, Mashpee, and Sandwich on the south coast of Cape Cod, MA. The
 watershed was first delineated by Babione (1990) and further refined by Cambareri et al. (1992)
 and Sham et al. (1995).  The watershed covers 8 km (5 mi) from the head of the bay to the
 regional groundwater divide.  The bay and its tributaries encompass  a total surface water area of
 3.9 km2 (389 ha or 1.5 mi2).  The major surface water components of the watershed include
 Waquoit Bay, two major rivers and several smaller streams, freshwater ponds, and freshwater
 wetlands. Within the Waquoit Bay watershed are seven subwatersheds (Childs River, Sage Lot
 Pond, Quashnet River, Eel Pond, Head of the Bay, Hamblin Pond, and Jehu Pond) and three
 freshwater ponds (Ashumet, Johns, and Snake).  These subwatersheds and adjacent ponds
 provide diverse habitats, including barrier beaches, eelgrass beds, saltwater and freshwater
 marshes, coastal sand dunes, and brackish and freshwater ponds.

 A.I.  Geological and hydrological characteristics
       The Waquoit Bay watershed lies entirely within the Mashpee pitted outwash plain
 (LeBlanc et al., 1986), a geologically young landform composed of glacial materials deposited on
 top of bedrock toward the end of the Wisconsinian Glacial Stage, about  12,000 years ago
 (Oldale, 1992). Outwash plains were created by broad meltwater streams that size-sorted the
 drift materials, depositing the heavier boulders and pebbles near the glacial margin and gravel
 and sands farther away. Because Cape Cod is geologically young, the glacial materials have not
 been altered significantly, resulting in a generally sandy, porous soil throughout the area. Clay
 and silt lenses are also found in deeper sediments to the south.
       The outwash plain is "pitted" as a result of the numerous kettle ponds dotting the
 landscape. Kettle ponds mark the sites where  blocks of ice were buried by sediment-laden
 meltwater streams beyond the glacial margin.  Johns Pond and Ashumet Pond are two examples
 of freshwater kettle ponds in the watershed. Waquoit Bay itself may have originated as a kettle
 pond.  The southern margin of the bay was flooded by sea-level rise at the close of the
Wisconsinian Glacial Stage. At this point the ice sheet retreated, the low-lying coastal areas
were inundated, and the water table inland was raised due to hydrostatic pressure at the saltwater-
freshwater interface.  The action of winds, waves, and currents continually eroded and displaced
the loose glacial sand and gravel, contributing to the formation of coastal sand dunes, sea cliffs,

barrier beaches, and salt marshes. These processes continue to alter the dynamic shore (Oldale,
       The geology of Waquoit Bay controls the region's hydrology, which is typical of a glacial
outwash plain. The bay, 1.2 km (4,000 ft) wide and 3.4 km (11,000 ft) long, is a shallow-water
system (average depth of 1.5 m). It receives input from freshwater streams and ground-water
flow, with tidal exchange to Vineyard Sound through two dredged and maintained channels and a
recent breach caused by overwash during a hurricane in August 1991 (Valiela et al., 1996). Fifty
percent of the water entering Waquoit Bay comes from the Quashnet and Childs rivers, 23%
comes from direct precipitation, and 27% comes from ground-water seepage (Cambareri et al.,
1992).  The rivers derive most of their water from ground-water discharge, draining the shallow
surface aquifer. Groundwater is forced to the surface as the permeable aquifer thins from north
to south in the watershed.
        The unconsolidated sediments of Cape Cod make ideal aquifers.  The permeable aquifer
is about 46 m (150 feet) thick near Snake Pond and thins to 9 m (30 feet) near Waquoit Bay
(Garabedian et al., 1991, Cambareri et al., 1992). The porous soils support rapid percolation of
rain, nutrients, and contaminants into the groundwater.  In recognition of the unique ground-
water characteristics of Cape Cod, the U.S. Environmental Protection Agency declared this
region a sole-source aquifer in 1982, a designation that facilitated protection of the water supply.
The Cape Cod aquifer can be subdivided into six groundwater lenses, and generally, groundwater
 does not flow between lenses.  The Waquoit Bay watershed lies within the Sagamore or western
 Cape lens of the Cape Cod Aquifer (Guswa and LeBlanc, 1981).

 A.2  Biological characteristics
        The waters of the southward-flowing, cold Gulf of Maine and the northward-flowing,
 warm Gulf Stream mix off of the coast of Cape Cod to form  a biological transition zone between
 the Virginian (temperate) and Acadian (boreal) biogeographic provinces (Ayvazian et al., 1992).
 This overlap produces more diverse communities than occur in either province. The Waquoit
 Bay estuarine complex benefits from this increased diversity. The watershed also lies near the
 Atlantic coast flyway, an important migratory corridor for many coastal and arctic-nesting birds,
 particularly shorebirds.
        The flora of the watershed include scrub oak and pitch pine forests .(Bailey, 1995).
 Forests covered 2,650 ha (6,548 acres) of the watershed in 1990. Among the state-protected
 plant species found in the watershed are the endangered sandplain gerardia (Agalinis acuta); the
 threatened bushy rockrose (Helianthuemum dumosum)- the knotroot foxtail (Setaria geniculata),


 which is of special concern; and the butterfly weed (Asclepias tuberosa), little ladies' tresses
 (Spiranthes tuberosa), eastern lilaeopsis (Lilaeopsis chinensis), New England blazing star
 (Liatris borealis), thread-leaved sundew (Droserafiliformis), vetchling (Lathyrus palustris), and
 wild rice (Zizania aquatica), which are on the watch list (Geist, 1996) to attempt to keep them
 from becoming threatened or endangered species.
        The fish in Waquoit Bay include freshwater, estuarine, and marine species.  The part-time
 residents represent a composite of estuarine spawners, such as winter flounder
 (Pseudopleuronectes americanus), longhom sculpin (Myoxocephalus octodecemspinosus), scup
 (Stenotomus chrysops}, and tautog (Tautoga onitis); marine species that are estuarine visitors,
 such as the sand lance (Amodytes americanus), summer flounder (Paralichthys dentatus), and
 American pollack (Pollachius virens); nursery species or young-of-the-year, such as winter
 flounder juveniles, mullets (Mugil cephalus), juvenile tautogs, menhaden (Brevoortia tyrannys},
 Atlantic silversides (Menidia menidia), bluefish (Pomatomus saltatrix), and bay anchovy
 (Anchoa mitchilli); and adventitious species that have a more southern distributions but that lack
 an apparent estuarine dependence, such as ladyfish (Elops saurus), halfbeak (Hemiramphus
 brasiliensis), and crevalle jack (Caranx hippos).  Alewives (Alosa pseudoharengus) and
 blueback herring (Alosa aestivalis) cross Waquoit Bay on their annual spawning migrations to
 fresh water, and larger fish such as bluefish and striped bass (Morone saxitalis) enter in pursuit of
 smaller prey fish.  Many primarily marine fishes use the estuary in the winter as a spawning and
 nursery ground. Bluefish, tomcod (Microgadus tomcod), white hake  (Urophycis tenuis), and
 pollack inhabit the bay as juveniles but are rarely present as adults.
       Shellfish species harvested in the estuary include bay scallops (Argopecten irradians),
 which are found in the eelgrass habitat, and hardshell (Mercenaria mercenaria) and softshell
 (Mya arenarid) clams, generally found in the sand and mud habitats, respectively. The biota of
 the estuary also include a variety of temperate and boreal species of planktonic and benthic algae
 and invertebrates, which provide food resources for the fmfish as well as the terrestrial and avian
 wildlife in the watershed.
       Numerous shorebirds use the barrier beach and coastal salt marsh as an important
 stopover on their spring journey north to breeding grounds in Canada and on their fall journey
 south to the southern United States and Central and South America. Shorebirds that appear in
 abundance in the spring and fall on Waquoit Bay's barrier beaches include black-bellied
 (Squatarola squatarola) and semipalmated (Charadrius semipalmatus) plovers; sanderlings
 (Crocethia alba); dunlin (Calidris alpina); semipalmated (Ereunetes pusillus), least (Pisobia
fusicollis), and western sandpipers (Pisobia minutilla); ruddy turnstones (Arenaria interpres);


willets (Catoptrophorus semipalmatus); lesser (Totanus flavipes) and greater (Tringa
melanoleuca) yellowlegs; and short-billed dowitchers (Limnodromus griseus).  Sharp-tailed
sparrows (Ammodramus cudacutus), black-crowned night-herons (Nycticorax nycticorax), snowy
egrets (Leucophoyx thula), and mute swans (Cygnus olor) are found in the saltmarshes.  Several
species of birds that use the waters as nesting or feeding grounds are state- and federally
protected species. The piping plover (Charadrius melodus), listed as threatened, and the least
tern (Sterna antillarum), listed as being of special concern, nest on South Cape Beach and
Washbum Island. The roseate tern (Sterna dougalli), an endangered species, forages in the water
and rests on the beach proper (Geist, 1996).
       The Childs and Quashnet rivers provide a relatively rare  and shrinking habitat for several
anadromous and catadromous finfish species (Baevsky, 1991). Brown trout (Salmo trutta),
brook trout (Salvelinus fontinalis), alewife (Alosa aestivalis), and white perch (Morone
americana) use these rivers as spawning grounds either within the rivers themselves or within
John's Pond (McLarney, 1988; Hurley, 1990). American eels (Anguilla rostrata) use these rivers
to reach spawning grounds in the open sea. These species require very specific ranges of water-
quality parameters (temperature, pH, dissolved oxygen, salinity) for development (Hunter, 1991).
Under the care of the northeast chapter of Trout Unlimited, the ecological integrity and stability
of the Quashnet River have improved  significantly.
       The good-quality water of the Quashnet River also provides habitat for a variety of
 macroinvertebrates that serve as a food source for the finfish communities (Pennak, 1989). As
 part of Trout Unlimited's restoration project, macroinvertebrate species were reintroduced to the
 Quashnet from other freshwater streams. A survey done in 1982-1983 found  species
 representing the Trichoptera (caddisfly), Diptera (true flies), l^pidoptera (butterflies and moths),
 Ephemeroptera (mayflies), and Plecoptera (stoneflies) orders. Stoneflies, and to some extent
 mayflies and caddisflies, are good indicators of healthy water quality, as they require fairly high
 levels of dissolved oxygen.

              APPENDIX B:  Organizations Concerned About Waquoit Bay

 Ashumet - John's Pond Association
 Ashumet Valley Property Owners Association
 Association for the Preservation of Cape Cod
 Atlantic States Marine Fisheries Commission
 Barnstable County Department of Health and the Environment
 Buzzards Bay National Estuary Program
 Cape and Islands Coastal Waters Steering Committee
 Cape Cod Beagle Club
 Cape Cod Commission
 Cape Cod Cooperative Extension Service
 Cape and Islands Self Reliance Corporation
 Citizens for the Protection of Waquoit Bay
 Davisville Association
 Falmouth Condo Trust
 Falmouth Conservation Commission
 Falmouth Rod and Gun Club
 Green Briar Nature Center
 League of Women Voters, Falmouth
 Mashpee Briarwood Association, Inc.
 Mashpee Conservation Commission
 Mashpee Harbor Master
 Mashpee Shellfish Department
 Massachusetts Coastal Zone Management
 Massachusetts Department of Fisheries, Wildlife, and Environmental Law Enforcement
 Massachusetts Audubon Society
 Massachusetts Department of Environmental Protection
 Massachusetts Military Reservation
 Massachusetts Heritage Society
 Menauhant Harbor Association
 Monomoscoy Improvement Trust
 National Science Foundation's Land-Margin Ecosystems Research Program
 NOAA's National Marine Fisheries Service
 NOAA's National Estuarine Research Division
 Otis Installation Restoration Program
 Seacoast Shores Owners Association
 Shorewood Beach Owners
 Sierra Club - Cape Cod Group
 South Cape Beach Advocates
The Nature Conservancy
The 300 Committee
Town of Mashpee

Town of Falmouth
Town of Sandwich
Trout Unlimited
U.S. Fish and Wildlife Service
U.S. Geological Survey
U.S. Army Corps of Engineers
U.S. Department of Agriculture Soil Conservation Service
U.S. EPA, Region 1
Wampanoag Tribal Council
Waquoit Bay Watershed Citizens Action Committee
Waquoit Bay Watershed Inter-municipal Committee
Waquoit Bay National Estuarine Research Reserve

                APPENDIX C: Public Concerns and Waquoit Bay Stressors

       At a public forum on September 21, 1993, of local environmental officials, scientists, and
 stakeholders, the following information was generated on environmental concerns and resources
 that should be protected and the stressors and ecological effects of those stressors on Waquoit
 C.I. Public Concerns
 Open space
 Indigenous wildlife
 Scenic views
 Flyway integrity (migrating waterfowl)
 Recreation (swimming)
 Noneconomic values
 Traditional life styles
 Historical/political perspective
 "Historical" bay ecosystem structure
 Food resource safety
 Clean water
 Clean air
 Upland-marsh ecotone
 Habitat  .
 Groundwater quality
 Flushing rates
 Air quality
 Washburn Island
 Human health and domestic animal health
 Recreational "atmosphere"
 Water quality
Marine organisms
River herring
"Quality of life" including:

Access to natural beauty
Freedom to enjoy
Human serenity
Natural noise
Night sky/darkness
Visual beauty
Pleasant sensory experiences

      C.2 Sources and Stressors in the Waquoit Bay Watershed
Septic systems,
fertilizers, atmospheric
Septic systems

 Macroalgal growth
 Introduced predators
 Introduction of exotic
 Marinas and piers
                        Shading by
         Ecological effects
 Increase in macroalgae and
 phytoplankton growth

 Introduction of pathogens and fecal
 coliforms to surface water
                         Shading by
                         Increase in
  by macroalgae
 i  T
 •    _^^-^j-
  Mute swan
                                                    Shellfish bed contamination
                            Alteration of substrate, light
                            Major fauna! alterations in benthic
                            and fish communities
        OllU 11011 V^lHAJJ-Hi-mJ*-"-"^
        Alteration of macroalgal species
        composition, loss of habitat for
        submerged aquatic vegetation, loss
        of spawning sites for fish, loss of
        hiding places and protection for fish,
        loss of scallop larvae settling habitat
        /~"U«-«™a i-n «r«at
        C.2  Sources and Stressors in the Waquoit Bay Watershed (continued)
:'--::/ ,:':•.: SOUTCC '•' .::>. '•
Gasoline and motor oil
from automobile and
boat engines
Massachusetts Military
Reserve, Otis Air Force
Landfill leachates
: Stressor
Seawalls and
Raking and
plunging for
Type ;
Ecological effects
Major alteration of shoreline
dynamics, sediment resuspension,
coastal erosion, sediment build-up,
change in flushing rates
Destruction of vegetation, sediment
resuspension, increased turbulence
and mixing in water column
Disturbance of sediment,
resuspension of nutrients, increasing
Loss of marsh-uplands ecotone,
increase in surface water runoff,
increase in sediment loading, altered
ground-water flow
Sediment disturbance and increase in


APPENDIX D: Attendees at the Waquoit Bay Management Goals Meeting
Tom Cambareri
Bruce Carlisle
Joe Costa
David Dow
Perry Ellis
Tom Fudula
Jeroen Gerritsen
Steve Hurley
Chuck Lawrence
Sandy McClean
Carl Melberg
JoAnne Muramoto
Mark Patton
Pam Polloni
Bob Sherman
Jan Smith -
Patti Tyler
Mary Varteresian
Brooks Wood
Rick York
Cape Cod Commission
Massachusetts Coastal Zone Management
Buzzards Bay National Estuary Program
National Marine Fisheries Service
Mashpee Harbor Master
Mashpee Planning Department
Tetra Tech Inc.
Massachusetts Division of Fisheries and Wildlife
Cape Cod Commission
Citizens for the Protection of Waquoit Bay
U.S. Fish and Wildlife Service
Falmouth Conservation Commission
Otis Installation Restoration Program
League of Women Voters, Falmouth
Mashpee Conservation Commission
Massachusetts Coastal Zone Management
U.S. EPA, Region 1
U.S. Fish and Wildlife Service
Monomoscoy Improvement Trust
Mashpee Shellfish Department


                   APPENDIX E: Information on Contamination from
                         the Massachusetts Military Reservation

 E.I Phosphorus Loading to Ashumet Pond
       Between 1936 and 1995 the sewage treatment plant at the Massachusetts Military Reserve
 (MMR) discharged treated sewage into infiltration ponds on the southern edge of the facility that
 contributed both toxic contaminants and high phosphate levels to the Ashumet Valley Plume.
 The U.S. Geological Survey (USGS) established a study to examine the transport of the
 dissolved phosphorus contained in the Ashumet Valley Plume.  This study showed that the bulk
 of the phosphorus is trapped on sediment particles. The Air Force Center for Environmental
 Excellence (AFCEE) agreed to mitigate the high phosphorus levels in the Ashumet Valley Plume
 as part of the Installation Restoration Program (IRP), even though this is not a Superfund issue.
 Because the bulk of the phosphorus is trapped in the sediments in the saturated zone upgradient
 of Ashumet Pond, AFCEE decided to fund studies that would develop a phosphorus budget for
 Ashumet Pond, determine the existing water quality and the limiting nutrients for phytoplankton
 growth, and develop in-pond mitigation strategies for phosphorus. In September 2001, AFCEE
 and its contractors completed a successful alum addition to the deep hole in  Ashumet Pond in
 order to trap the phosphorus in the sediments and thus reduce the dissolved phosphorus loading
 to the hypolimnion (bottom waters) during summer stratification. The buffered dose of alum and
 sodium aluminate did not change the pH or alkalinity in the receiving water, and the dissolved
 aluminum concentration in the water was below the detection limit (1 ppm).  No fish kills
 accompanied this treatment process.
       The results of the AFCEE supported other investigations showing that phosphorus
 discharges from the Ashumet Valley Plume into Ashumet Pond in the Fisherman's Cove region
 reached at an annual loading rate of 180 kg P/yr, with maximum dissolved phosphorus
 concentrations of roughly 2 mg/L.  Other external loading rates were 51-63 kg P/yr for other
 groundwater and 13-22 kg P/yr for surface water inflows. This  loading represented new
 phosphorus into the system, which could be used in a Vollenweider Loading Model to estimate
 the resulting chlorophyll levels and dissolved phosphorus concentrations in the water column.
 The chlorophyll concentrations measured in the summer of 1999 were less than  1 u-g/L, and the
 total dissolved phosphorus levels were less than 15 ppb, suggesting that the system is
      Experiments in nutrient uptake (measured by alkaline phosphatase induction, which
estimates the potential for converting organic  phosphorus to inorganic phosphorus) were
conducted in the phytoplankton after the spring bloom. It is unknown whether silicon or

phosphorus is limiting to the spring diatom bloom. Cyanobacteria dominate the phytoplankton in
the surface waters during the summer, with a seasonal succession from phosphorus limitation to
nitrogen fixers and back to phosphorus limitation, based on the available dissolved inorganic
nitrogen and phosphorus concentrations in the water. Thus a dynamic interaction apparently
exists between the dissolved nutrient levels and the species that dominate the phytoplankton in
the epilimnion (surface waters). Because zooplankton cannot graze many cyanobacteria
effectively, this was considered an ecological stressor that might not be eliminated by reducing
the phosphorus loading in the bottom waters. Cyanobacteria blooms can also cause taste and
odor problems in surface waters.
       Anoxia develops in the deeper, stratified portion of the pond between May and October,
resulting in recycling of phosphorus from the sediments back into the water column at an annual
rate of 194 kg P/yr. For the shallow (less than 7.6 m) sediments in the pond that remain oxic
throughout the summer, the recycling rate is 195 kg P/yr.  Within the epilimnion, the annual
recycling rate is 3,370 kg P/yr. Thus the recycled phosphorus greatly exceeds the annual external
loading rate of new phosphorus.  The recycled phosphorus is not included in the Vollenweider
model.  Presumably the particulate phosphorus that settles into the bottom waters from the
surface is balanced by the dissolved phosphorus recycled from the sediments.
       Seasonal anoxia resulted in fish kills in Ashumet Pond in July 1985 and May 1996.
Seasonal anoxia is a result of bacterially mediated degradation of particulate organic matter in the
hypoliminion.  This degradation uses up dissolved oxygen in the bottom water and, because of
 water column stratification, it cannot be replenished. From an ecological risk assessment
 perspective, this anoxia/hydroxia is an ecological stressor, even though the surface water is
 classified as mesotrophic. The amount of anoxia/hypoxia resulting from the morphology of the
 kettle hole ponds and the amount due to the nutrient loading and phytoplankton production in the
 surface water are unknown. The USGS has prepared a rough phosphorus budget for Ashument
 Pond (oral communication between D. LeBlanc, USGS, Marlboro, MA, and David Dow
 NOAA/NMFS/ NEFSC, Woods Hole Laboratory, Woods Hole, MA regarding poster presented
 at the Ashmut Pond phosphorus project update meeting, Barnstable County Fairgrounds, MA,
 August 2,2001, by D. LeBlanc) suggesting that the new phosphorus loading sources are
 distributed as follows: precipitation, 10%;  surface runoff, 10%; MMR Ashumet Valley Plume,
 50%; and other groundwater inputs, 30%.

 E.2 Volatile Organic Compound (VOC) Contamination in Plumes
        MMR was designated as a Federal Superfund site in 1989, and since that time 14 ground-
 water plumes have been identified that emanate from the base and extend into the surrounding
 communities (Bourne, Falmouth, Mashpee, and Sandwich). The VOCs at MMR mainly come
 from fuel spills, which produce benzene, toluene, ethylbenzene, xylene and ethylene dibromide
 contaminants, or from chemical spills, which produce chlorinated solvents, such as
 trichloroethylene (TCE), tetrachloroethylene, carbon tetrachloride, etc. The benzene, toluene,
 ethylbenzene, and xylene are often biodegraded by aerobic metabolism near the source areas,
 whereas the ethylene dibromide in aviation fuel is left behind in detached plumes downgradient
 from the source area.  The chlorinated solvents are subject to slow anaerobic metabolism via
 reductive dechlorination in the groundwater, but they tend to be persistent, like the ethylene
 dibromide in the fuel spill plumes. The chemical  spill-10 plume discharges into Ashumet Pond,
 with a portion passing through the isthmus connecting Johns and Ashumet ponds and discharging
 into Johns Pond, where it is referred to as the TCE Plume.  The storm drain-5 plume discharges
 small quantities of trichloroethylene into Johns Pond as well. Fuel spill-1 plume is a detached
 plume that discharges ethylene dibromide into the Quashnet River downstream from its Johns
 Pond source. The elliptical shapes in Figure E-l illustrate the locations and extents of these
       The old fire-training-area site (site 1) is the source area for the Ashumet Valley Plume,
 and the primary constituents within this ground-water plume have been identified as VOCs and
 inorganics. Fire-training-area-1 is a 3-acre parcel of land where fire-training exercises were
 conducted from 1958 to 1985 and materials burned here consist of different fuel types, waste
 oils, solvents, thinners, transformer oils, and spent hydraulic fluids.  Southeast Regional
 Groundwater Operable Unit plumes originate from fire-traming-area-2, landfiIl-2, petroleum fuel
 storage area, and storm drain-5. Fire-training-area-2 and landfilI-2 comprise a 20-acre area of
 land located on top of a former industrial/ municipal landfill that was used for fire-training
 exercises. Compounds disposed of in the landfill or burned on the fire-training area consist of
 fuel, waste oils, waste petroleum distillate solvents, and domestic refuse.  The petroleum fuel
 storage area is an active facility that is involved in the delivery of various types of fuel.  It was
the site of a 2,000 gallon fuel spill in the 1960s.
       AFCEE leads the mitigation of these underground pollution plumes and the associated
source areas  under the Superfund program. The following is a brief description of the mitigation
programs that are either underway or planned for both the plumes and their source areas, which
are often multiple for a given plume. The plume mitigation goal is to reduce the contamination


Figure E-l.  Map of the plumes emanating from the Massachusetts Military Reservation.


 to below maximum contaminant levels (5 ppb for VOCs and 0.020 ppb for ethylene dibromide),
 which are much lower than the acute/chronic toxicity levels for biota, and then to conduct a risk
 assessment and feasibility study to see whether cleanup should proceed to nondetect levels. The
 source removal program is on a separate, less-advanced timeline than the plume cleanup.
       The chemical spill-10 plume is the largest, most contaminated plume at the MMR
 Superfund site. For the chemical spill-10 plume plume, the proposed treatment system will
 feature a combination of shallow well points and four deep-extraction wells that withdraw 750
 gpm, followed by reinjection of the treated water into an infiltration trench.  The extracted water
 is treated with granulated activated carbon treatment and is reinjected back into the groundwater.
 The portion of the chemical spill-10 plume downgradient of the Sandwich Road extraction,
 treatment, and reinjection discharges into the western side of Ashumet Pond, while a portion
 flows under the pond, emerging as the TCE Plume on the isthmus connecting Ashumet Pond to
 Johns Pond. It is unknown whether the Ashumet Pond trichloroethylene discharge has any toxic
 effects on the benthos, but risk assessments suggest that it does not pose a threat to human health.
 The TCE Plume has concentrations of 263 ppb on the eastern side of Ashumet Pond, with
 maximum concentrations near the Johns Pond upwelling point of 1502 ppb trichloroethylene and
 6 ppb tetrachloroethylene. This plume is only 75 feet wide and has a maximum trichloroethylene
 level of 2,200 ppb, so it is very narrow, but highly contaminated. The maximum concentration in
 the pond water above the discharge point is 3.5 ppb trichloroethylene.
       The storm drain-5 south plume also discharges into Johns Pond, but its maximum
 trichloroethylene level is only 60 ppb, with the average near 10 ppb trichloroethylene. Even
 though it has a larger discharge footprint within Johns Pond than does the TCE Plume,
 trichloroethylene is not detectable in the pond water above the bottom. Storm drain-5 south has
 two recirculating wells within the plume that operate at 120 gpm. A proposed extraction,
 treatment and reinjection well along Hooppole Road would connect to the Sandwich Road
 Treatment System. A portion of storm drain-5 south will discharge into Johns Pond over the
 next 10-12 years, as the combined recirculating/ extraction, treatment and reinjection system can
 only capture 55% of the contaminant mass. There is also a system of 10 extraction wells and
 eight reinjection wells that captures the northern portion of the storm drain-5 north plume.
       In Johns Pond, the brown bullheads have exhibited papillomas around their mouths/jaws
 and adenocarcinomas in the liver, which has created controversy about whether there is a link
between the trichloroethylene discharges from storm drain-5 south and the TCE Plume or
whether these tumors stem from viruses or genetic factors. Even though trichloroethylene is not
bioaccumulated in fish, it can bioconcentrate 15- to 30-fold over the concentrations found in the


water.  Past studies of these fish detected polycyclic aromatic hydrocarbons in the fish flesh, and
induction of the mixed-function oxidase enzyme system that metabolizes these contaminants.
No tests were conducted on the induction of trichloroethylene-metabolizing enzymes in these
fish, so there is no direct evidence for an impact of the trichloroethylene discharged into Johns
Pond on the biota. Fuel spill-1 plume discharges into springs that feed the upper portions of the
Quashnet River, and ethylene dibromide has been detected in the river water at 1.4 ppb, with a
maximum concentration within the groundwater of 10 ppb. The Kl cranberry bog area on the
Quashnet River is an important breeding habitat for eastern brook trout. Because of the concern
about ethylene dibromide contamination of cranberry bogs in the discharge region, an elaborate
bog separation program has been put in place, along with an a pilot treatment system. This
system has 138 shallow wells points that extract 400 gpm and one deep-extraction well that
withdraws 200 gpm.  Water from these wells will undergo onsite granulated activated carbon
treatment and then be reinjected.  The surface discharge has a diffuser in order to maintain the
dissolved oxygen at levels (above 8 ppm) necessary for brook trout breeding. A plan has been
developed by AFCEE and approved by the U.S. Environmental Protection Agency and the
Massachusetts Department of Environmental Protection to place three additional extraction wells
in the southern toe of the plume and use the existing treatment plant.
       Snake Pond is the farthest upgradient freshwater pond in the Waquoit Bay watershed, and
 it has  a fish advisory in effect because of mercury levels. The advisory was issued by the
 Massachusetts Department of Public Health on the basis of human health concerns and
 recommends no fish consumption by children younger than 12 years or by pregnant/nursing
 women. The presumed source of the mercury is the atmosphere. It  is unknown whether these
 higher mercury levels have an impact on the fish themselves or on other biota, even though the
 scientific literature suggests possible effects on growth for some freshwater species  (Stafford and
 Haines, 2001). Groundwater underneath Snake Pond has elevated levels of ethylene dibromide
 (above'the maximum contaminant level of 0.020 ppb).  The ethylene dibromide contaminants are
 associated with the fuel spill-12 plume and explosives, such as Royal Dutch Eplosive (RDX),
 from contamination emanating from the J-3 Range in the Impact Area at Camp Edwards.
 However, these contaminants have not been detected in either the water or the sediments of
 Snake Pond, even though models predict that the RDX will upwell into the pond. Thus, no well-
 defined ecological stressors from human activities have been identified in Snake Pond.

Aber, JD; Magill, R; Boone, JM; et al. (1993) Plant and soil responses to chronic nitrogen
additions at the Harvard Forest, Massachusetts. Ecological Applications 3:156-166.

AFCEE (Air Force Center For Environmental Excellence). (2001) Final Ashumet Pond
Phosphorus Management Plan. Brooks Air Force Base, San Antonio, XX.

Aguilar, C; Fogel, ML; Paerl, HW. (1999) Dynamics of atmospheric combined inorganic
nitrogen utilization in the coastal waters off North Carolina. Marine Ecology Progress Series

Andreoli, A; Bartilucci, N; Forgione, R; et al. (1979) Nitrogen removal in a subsurface removal
system. Journal of the Water Pollution Control Federation 51:841-855.

Aravena, R; Evans, ML; Cherry, JA. (1993) Stable isotopes of oxygen and nitrogen in source
identification of nitrate from septic systems.  Ground Water 31:180-186.

Ayvazian, SG; Deegan, LA; Finn, JT. (1992) Comparison of habitat use by estuarine fish
assemblages in the Acadian and Virginian zoogeographic provinces. Estuaries 15:368-383.

Babione, M. (1990) Land use change in the watershed of Waquoit Bay, Massachusetts. Division
HI Exam, School of Natural Science, Hampshire College, MA.

Baevsky, YH. (1991) Physical and water-quality characteristics affecting trout-spawning habitat
in the Quashnet River, Cape Cod,  Massachusetts. USGS Water Resources Investigations Report
91-4045. USGS,Marlborough,MA.

Bailey, RG. (1995) Eastern broadleaf forest (oceanic) province.  In: Description of the ecoregions
of the United States.  U.S. Department of Agriculture, Forest Service, Miscellaneous Publication
1391, 2nd Edition. Washington, DC.

Billen, G;  Somville, M; De Becker, E; et al. (1985) A nitrogen budget of the Scheldt
hydrographical basin. Netherlands Journal of Sea Research 19:223-230.

Billen, G;  Lancelot, C; Meybeck, M. (1991) N, P, and Si retention along the aquatic continuum
from land  to ocean. Pp. 19-44 in R. F. C. Mantoura, J.-M. Martin, and R. Wollast, editors.
Ocean Margin Processes in Global Change.  New York: John Wiley & Sons.

Bowen, JL; Valiela, I. (2001a) The ecological effects of urbanization of coastal watersheds:
Historical  increases in nitrogen loads and eutrophication of Waquoit Bay estuaries. Canadian
Journal of Fisheries and Aquatic Sciences 58:1489-1500.

Bowen, JL; Valiela, I. (2001b) Historical changes in atmospheric nitrogen deposition to Cape
Cod, Massachusetts, USA. Atmospheric Environment 35:1039-1051.

Bowen, JL; Valiela, I. Nitrogen loads to estuaries: Using loading models to assess the
effectiveness of management options to restore estuarine water quality. Estuaries: to be

Boynton, WR; Kemp, WM; Keefe, CW. (1982) A comparative analysis of nutrients and other
factors influencing primary production,  p. 69-80. In: Estuarine comparisons, Kennedy, VS (ed).
New York: Academic Press.

Brawley, JW; Collins, G; Kremer, JN; et al. (2000) A time-dependent model of nitrogen loading
to estuaries from coastal watersheds. Journal of Environmental Quality 29:1448-1461.

Butler, AJ; Connolly, RM. (1996) Development and long-term dynamics of a fouling assemblage
of sessile marine invertebrates. Biofouling 9:187-209.

Cambareri, TC; Eichner, EM; Griffith, CA. (1992) Sub-marine groundwater discharge and nitrate
loading to shallow coastal embayments. In: Proceedings, Eastern Regional Groundwater
Conference, Newton, MA.  Oct. 13-15,1992.  National Groundwater Association, Westerville,

Cameron. KC; Wild, A. (1982) Comparative rates of leaching of chloride, nutrients, and tritiated
water under field conditions. Journal of Soil Science 33:649-657.

Cohen, JE; Small, C; Mellinger, A; et al. (1997) Estimates of coastal populations. Science

Collins, G; Kremer, JN; Valiela, I. (2000) Assessing uncertainty in estimates of nitrogen loading
to estuaries for research, planning and risk assessment. Environmental Management

Carroll, DL; Ford, D. (1982) Comparison of precipitation and land runoff as sources of estuarine
nitrogen.  Estuarine, Coastal, and Shelf Science 15:45-56.

Correll, DL; Jordan, TE; Weller, DE. (1992) Nutrient flux in a landscape: effects of coastal land
use and terrestrial community mosaic on nutrient transport to coastal waters.  Estuaries

 Costa, JE. (1988) Distribution, production, and historical changes in abundance of eelgrass
 (Zostera marina} in Southeastern Massachusetts. Dissertation. Boston University, Boston, MA.

Costa, JE; Howes, BL; Giblin, AE; et al. (1992) Monitoring nitrogen and indicators of nitrogen
to suggest management action in Buzzards Bay. In: Mckenzie, DH; Hyatt, DE; McDonald, VJ;
eds.  Ecological Indicators 1:499-534.

Cottam, C. (1933) Eel grass dying of mysterious disease. Science News Letters 24:73.

Cotton, AD. (1933) Disappearance of Zostera marina. Nature 132:277.

Cubbage, A; Lawerence, D; Tomasky, G; et al. (1999) Relationship of reproductive output in
Acartia tonsa, chlorophyll concentration, and land-derived nitrogen loads in estuaries in Waquoit
Bay, MA.  Biol Bull 197:294-295.

D'Avanzo, C; Kremer, J. (1994) Diel oxygen dynamics and anoxic events in an eutrophic estuary
of Waquoit Bay Massachusetts. Estuaries 17:131-139.

D'Elia, CF; Steudler, PA; Corwin, N. (1977) Determination of total nitrogen in aqueous samples
using persulfate digestion.  Limnology and Oceanography 22:760-764.

Davies, TD; Pierce, CE; Robinson, HJ; et al. (1992) Towards an assessment of the influence of
climate on wet acidic deposition in Europe. Environmental Pollution 75:111-120.

Deegan, LA; Buchsbaum, RN. (In press) The effect of habitat loss and degradation on fisheries.
In: The Decline of Fisheries Resources in New England: Evaluating the Impact of Overfishing,
Contamination, and Habitat Degradation. Buchsbaum, RN; Robinson, WE; Pederson, J (eds).
Massachusetts Bays Program, University of Massachusetts Press, Amherst, MA.

Diamond,  JM; Serveiss, VB. (2001) Identifying sources of stress to native aquatic species using a
watershed ecological risk assessment framework. Environmental Science and Technology

Downing,  JA; Osenberg, CW; Sarnelle,  O. (1999) Meta-analysis of marine nutrient enrichment
experiments: Variation in the magnitude of nutrient limitation. Ecology 80:1157-1167.

Duarte, CM. (1995) Submerged aquatic  vegetation in  relation to different nutrient regimes.
Ophelia 41:87-112.

Eichner, EM; Cambareri, TC. (1992) Nitrogen Loading. Cape Cod Commission Technical
Bulletin 91-001. Barnstable, MA.

Evgenidou A; Valiela, I. (In press) Response of growth and density of a population of Geukensia
demissa to land-derived nitrogen loading to Waquoit Bay, Massachusetts. Estuarine, Coastal,
and Shelf Science.

Faught, MC. (1945) Falmouth, Massachusetts: Problems of a Resoit Community. New York:
Columbia University Press.

Foreman, K; Tomasky, G; Soucy, LA; et al. (Submitted) Responses of phytoplankton to land-
derived nitrogen loads in Waquoit Bay, MA. Aquatic Ecology.

Freyer, HD; Aly, AIM. (1974) 15N variations in fertilizer nitrogen.  Journal of Environmental
Quality 3:405^06.

Fricke, W; Beilke, S. (1992) Indication for changing deposition patterns in central Europe.
Environmental Pollution 75:121-127.

Frimpter, MH; Donahue, JJ; Rapacz, M.  (1990) A mass-balance model for predicting nitrate in
groundwater.  New England Water Works Association 104:219-232.

Gallagher, J. (1983) The South Cape Beach, Mashpee, Massachusetts: An Intensive Survey of
the Balance of the Park, Phase 1  (B). Massachusetts Department of Environmental Management,
Boston, MA.

Garabedian, SP; LeBlanc, DR; Gelhar, LW; et al. (1991) Large-scale natural gradient tracer test
in sand and gravel, Cape Cod, Massachusetts, 2, analysis of spatial moments for a nonreactive
tracer. Water Resources Research 27:911-924.

Geist, MA. (1996) The Ecology of the Waquoit Bay National Estuarine Research Reserve.
Waquoit Bay National Estuarine Research Reserve, Waquoit Bay, Massachusetts.

GESAMP (Group of Experts on  the Scientific Aspects of Marine Pollution). (1990) The State of
the Marine Environment. IMO/FAO/UNESCOAVMOAVHO/IAEAAJN/UNEP. Rep. and Stud
39. Blackwell, London, UK.

Golomb, D; Ryan, D; Eby, N; et al. (1997) Atmospheric deposition of toxics onto Massachusetts
Bay: metals. Atmospheric Environment 31:1349-1359.

Gormly, JR; Spaulding, RF. (1979) Sources and concentrations of nitrate nitrogen in the
groundwater of the central Platte region, Nebraska.  Ground Water 17:291-201.

Griffin, MPA; Valiela, I. (2001) d15N isotope studies of life history and trophic position of
Fundulus heteroclitus and Menidia menidia. Marine Ecology Progress Series 214:299-305.

Guswa, JH; LeBlanc, DR. (1981) Digital models of groundwater flow in the Cape Cod aquifer
system, Cape Cod, MA. USGS Water Resources Investigations Open File Report 80-67.

Hansson, S; Hobble, JE; Elmgren, R; et al. (1997) The stable nitrogen isotope ratio as a marker
of food-web interactions and fish migration.  Ecology 78:2249-2257.

Harlin, MM; Thome-Miller, B. (1981) Nutrient enrichment of seagrass beds in a Rhode Island
coastal lagoon. Marine Biology 65:221-229.

Harris, HJ; Wegner, RB; Harris, VA; et al. (1994) A method for assessing environmental risk: A
case study of Green Bay, Lake Michigan, USA. Environmental Management  18:295-306.

Hauxwell, J; Behr, P; McClelland, J; et al. (1998) Grazing and nutrient control on macroalgal
biomass in Waquoit Bay, MA.  Estuaries 21:347-360.

Hauxwell, J; Cebrian, J; Furlong, C; et al. (2001) Macroalgal canopies contribute to eelgrass
(Zostera marina) decline in temperate estuarine ecosystems. Ecology 82:1007-1022.

Hayes, S; Newland, L; Morgan, K; et al. (1990) Septic tank and agricultural non-point source
pollution within a rural watershed. Toxicological and Environmental Chemistry 26:137-155.

Heck, KL; Able, KW; Fahey, MP; et al. (1989) Fishes and decapod crustaceans of Cape Cod
eelgrass meadows:  species composition, seasonal abundance patterns, and comparisons with
unvegetated substrates. Estuaries 12:59-65.

Heufelder, G; Rask, S. (1996) Alternative Septic System Update #9. Barnstable County
Department of Health and the Environment, Barnstable, MA.

Heufelder, G; Rask, S. (1997) Alternative Septic System Update #10.  Barnstable County
Department of Health and the Environment, Barnstable, MA.

Hinga, KR; Keller, AA; Oviatt, CA. (1991) Atmospheric deposition and nitrogen inputs to
coastal waters. Ambio 20:256-260.

Howarth, R. (1988) Nutrient limitation of net primary production in marine ecosystems.  Annual
Review of Ecology and Systematics 19:89-110.

Hunter, CJ. (1991) Better trout habitat: A guide to stream restoration and management.
Washington, DC: Island Press.

Hurley, ST. (1990) Fisheries sampling report: Quashnet River, Falmouth, Mashpee.
Unpublished report, Massachusetts Division  of Fisheries and Wildlife, Blizzards Bay, MA.

Interdisciplinary Environmental Planning (IEP), (1986) Hydrological analisis and impact
assessment- proposed Yarmouth public golf course. IEP, Inc., Sandwich, Mass.

Johnston, CA. (1991) Sediment and nutrient retention by freshwater wetlands: Effects of surface
water quality. Critical Reviews in Environmental Control 21:491-565.

Johnston, CA; Detenbeck, ME; Niemi, GJ. (1990) The cumulative effect of wetlands on stream
water quality and quantity: a landscape approach. Biogeochemistry 10:105-141.

Jordan, TE; Weller, DE. (1996) Human contributions to total nitrogen flux.  BioScience

Jordan TE; Correll, DL; Weller, DE. (1997) Relating nutrient discharges from watersheds to land
use and streamflow variability. Water Resources Research 33:2579-2590.

Kaplan, OB. (1991) Septic systems handbook.  Boca Raton, FL: Lewis Publishers.

Keeney, D. (1986) Sources of nitrate to ground water. Critical Reviews in Environmental
Control 16:257-304.

Ketchum, BH. (1951) The exchanges of fresh and salt water in tidal estuaries. Journal of Marine
Research 10:18-38.

Kohl, DH; Shearer, GB; Commoner, B. (1973) Variation of dI5N in corn and soil following
application of fertilizer nitrogen. Soil Science Society of America Proceedings 37:888-892.

Koppelman, LE. (1978) The Long Island comprehensive waste treatment management plan, vol.
H: summary documentation. Nassau-Suffolk Regional Planning Board, Haupauge, NY.

Korom, SF. (1992) Natural denitrification in the saturated zone: a review. Water Resources
Research 28:1657-1668.

Kreitler, CW. (1975) Nitrogen-isotope ratio study of soils and groundwater nitrate from alluvial
fan aquifers in Texas.  Journal of Hydrology 42:147-170.

Kreitler, CW; Jones, DC. (1975) Natural soil nitrate: the cause of the nitrate contamination of
groundwater in Runnels County, TX. Ground Water 13:53-61.

Krom, MD; Berner, RA. (1980) Adsorption of phosphate in anoxic marine sediments.
Limnology and Oceanography 25:327-337.

 Lajtha, K; Seely, B; Valiela, I. (1995) Retention and leaching losses of atmospherically derived
 nitrogen in the aggrading coastal watershed of Waquoit Bay, MA. Biogeochemistry 28:33-54.

 LaMontagne, MG; Valiela, I. (1995) Denitrification measured by a direct N2 flux method in
 sediments of Waquoit Bay, MA. Biogeochemistry 31:63-83.


Landis, WG; Wiegers, JA. (1997) Design considerations and a suggested approach for regional
and comparative ecological risk assessment.  Human and Ecological Risk Assessment

Lawrence, DL. (2000) Estuarine calanoid copepod abundance in Waquoit Bay MA: Effects of
season, salinity, and land-derived nitrogen loads.  Master's Thesis. Boston University.

LeBlanc, DR; Guswa, JH; Frimpter, MH; et al. (1986) Groundwater resources of Cape Cod,
Massachusetts.  USGS Hydrologic Investigations Atlas, Volume 692.

Legra, JC; Safran, RE; Valiela, I. (1998) Lead concentration as an indicator of contamination
history in estuarine sediments. Biological Bulletin 195:243-244.

Lindhult, MS; Godfrey, PJ. (1988) Using Geographic Information Systems to predict the trophic
states of lakes. In: Proceedings from the Symposium on Coastal Water Resources Wilmington,
NC. American Water Resources Association, Bethesda, MD.

Link, J; Brodziak, J. (eds). (2002) Report on the Status of the NE US Continental Shelf
Ecosystem. NEFSC Ecosystem Status Working Group. Document # 02-11, Northeast Fisheries
Science Center, Woods  Hole, MA.

Loehr, RC. (1974) Characteristics and comparative magnitude of nonpoint sources. J Water
Pollut Control Feder 46:1849-1872.

Lowrance, R; Todd, RL; Asmussen, LE. (1984) Riparian forests as nutrient filters in agricultural
watersheds. BioScience 34:374-377.

Macko, SA; Ostrom,  NE.  (1994) Pollution studies using stable isotopes. Pp. 45-62. In: Stable
Isotopes in Ecology. Lajtha, K; Michener, RS (eds). London: Blackwell Scientific.

Mariotti, A; LeTolle, R. (1977) Application of nitrogen isotopes to hydrology and hydrogeology:
Analysis of the peculiar case of the Melarchez Basin (Siene-et-Marne, France). Journal of
Hydrology 33:157-172.

McClelland, JW; Valiela, I. (1998) Linking nitrogen in estuarine producers to land-derived
sources. Limnology and Oceanography 43:577-585.

McDonald, MG; Harbaugh, AW. (1988) A Modular Three-Dimensional Finite Difference
Groundwater Flow Model. US Geological Survey, Techniques of Water Resources
Investigations, Book 6,  U.S. Government Printing Office, Washington, DC,

McLarney, WO. (1988) Who says they don't make trout streams anymore? Trout, Autumn:

Mitchell, S. (1999) Homeowner's Guide to Title 5.  Association for the Preservation of Cape
Cod, Orleans, MA.

Molot, LA; Dillon, PJ. (1993) Nitrogen mass balances and denitrification rates in central
Ontario lakes.  Biogeochemistry 20:195-212.

Nelson, ME; Horsley, SW; Cambareri, TC; et al. (1988) Predicting nitrogen concentrations in
groundwater an analytical model. In: Proceedings from the Conference on Eastern Regional
Groundwater Issues. September 27-29,1988, Stamford, CT. National Well Water Association.
Dublin, OH.

Nixon, SW. (1986) Nutrients and the productivity of estuarine and coastal marine ecosystems.
Journal of the limnological Society of South Africa 12:43-71.

Nixon, SW. (1992) Quantifying the relationship between nitrogen input and the productivity of
marine ecosystems. Proceedings of the Advanced Marine Technology Conference, Tokyo,
Japan. 5:57-83.

Nixon, SW. (1995) Coastal marine eutrophication: A definition, social causes, and future
concerns. Ophelia 41:199-219.

Norton, SB; Cormier, S; Smith, M; et al. (2000) Can biological assessments discriminate among
types of stress? A case study from the eastern corn belts plains ecoregion. Environmental
Toxicology and Chemistry I9(4):1113-1119.

NRC (National Research Council). (2000) Clean coastal waters.  Committee on the causes and
management of coastal eutrophication, ocean studies board, and water science and technology
board. Washington, DC: National Academy Press.

Oldale, RN. (1992) Cape Cod and the Islands:  The Geologic Story. East Orleans, MA:
Parnassus Imprints.

Ollinger, SV; Aber, JD; Lovett, GM; et al. (1993) A spatial model of atmospheric deposition for
the northeastern U.S. Ecological Applications 3:459-472.

Pace, ML; Findlay, SEF; Lints, D. (1992) Zooplankton in advective environments: The Hudson
River community and a comparative analysis.  Canadian Journal of Fisheries and Aquatic
 Sciences 49:1060-1069.

 Pack,DH. (1980) Precipitation chemistry patterns: a two-network data set.  Science

Peckol, P; DeMeo-Anderson, B; Rivers, J; et al. (1994) Growth, nutrient uptake capabilities, and
tissue constituents of the macroalgae, Cladophora vagabunda, and Gracilaria tikvahiae, related
to site specific nutrient loading rates. Marine Biology 121:175-185.

Pennak, RW. (1989) Freshwater Invertebrates of the United States, 3rd ed.: Protozoa to mollusca.
New York: Wiley.

Petrovic, AM. (1990) The fate of nitrogenous fertilizer applied to turfgrass. Journal of
Environmental Quality 19:1-14.

Pickett, STA. (1989) Space-for-time substitutions as an alternative to long-term studies. In:
Long-term Studies in Ecology: Approaches and Alternatives. Likens, GE (ed.), New York:
Springer-Verlag; pp. 110-136.

Piluk, RJ; Peters, EC. (1995) Small recirculating sand filters for individual homes. Anne
Arundel County Health Department. Annapolis, MD. Available at:

Prairie, YT. (1996) Evaluating the predictive power of regression models. Canadian Journal of
Fisheries and Aquatic Sciences 53:490-492.

Rask, S; Heufelder, G. (1998) Alternative Septic System Update #5. Barnstable County
Department of Health and the Environment.  Barnstable, MA.

Reneau, RB. (1979) Changes in concentrations of selected chemical pollutants in wet, tile-
drained'soil systems as influenced by disposal of septic tank effluents. Journal of Environmental
Quality 8:189-196.

Renn, CE. (1935) A mycetozoan parasite of Zostera marina. Nature 135:544-545.

Riisgard, HU; Hansen, S. (1990) Biomagnification of mercury in a marine grazing food chain:
 algal cells (Phaeodactylum tricornutum) musselys (Mytilus edulis) and flounder (Platichtys
flesus) studied by means of a stepwise reduction CVAA method. Marine Ecology Progress
 Series 62:259-270.

 Running, SW; Nemani, RR; Hungerford, RD. (1988) Extrapolation of synoptic meteorological
 data in mountainous terrain and its  use for simulating forest evapotranspiration and
 photosynthesis. Canadian Journal of Forestry Research 17:472-483.

 Serveiss, VB. (2002) Applying ecological risk principles to watershed assessment and
 management. Environmental Management 29(2): 145-154.

Serveiss, VB; Norton, DJ; Norton, SB. (2000) Watershed ecological risk assessment: the
watershed academy. U.S. Environmental Protection Agency on-line training module at:

Serveiss, VB; Bowen, JL; Dow, D; et al. Waquoit Bay watershed ecological risk assessment.
Environmental Management: submitted.

Sham, C-H; Brawley, J; Moritz, MA. (1995) Quantifying nitrogen loading from residential septic
sources to a shallow coastal embayment. International Journal of Geographic Information
Systems 9:463^73.

Shannon, JD; Sisterson. (1992) Estimation of S and NOx-N deposition budgets for the United
States and Canada. Water, Air, and Soil Pollution 63:211-235.

Shriver, AC; Carmichael, RH; Valiela, I. (2002) Growth, condition, reproductive potential, and
mortality of bay scallops, Argopecten irradians, in response to eutrophi-driven changes in food
resources. Journal of Experimental Marine Biology and Ecology 279:21-40.

Short, FT; Burdock, DM. (1996) Quantifying eelgrass habitat loss in relation to housing
development and nitrogen loading in Waquoit Bay, Massachusetts. Estuaries 19:730-739.

Short, FT; Mathissen, AC; Nelson, JJ. (1986) Recurrence of the eelgrass wasting disease at the
border of New Hampshire and Maine, USA.  Marine Ecology Progress Series 29:89-92.

Short, FT; Ibelings, BW; DenHartog, C. (1988) Comparison of a current outbreak of eelgrass
disease to the wasting disease in the 1930s.  Aquatic Botany 30:295-304.

Shumway, SE. (1991) Scallops: Biology, Ecology and Aquaculture. Amsterdam: Elsevier Press.

Smith, I; Fonseca, MS; Rivera, JA; et al. (1989) Habitat value of natural versus recently
transplanted eelgrass for the bay scallop. Fishery Bulletin 87:189-196.

Sogard, SM; Able, KW. (1991) A comparison of eelgrass, sea lettuce, macroalgae, and marsh
creeks as  habitats for epibenthic fishes and copepods. Estuarine, Coastal and Shelf Science

Stafford, CP; Haines, TA. (2001) Mercury contamination and growth rate in two piscivore
populations. Environmental Toxicology and Chemistry 20:2099-2101.

Stanley, DW. (1988) Historical trends in nutrient loading to the Neuse River estuary, NC. In:
Proceedings from the Symposium on Coastal Water Resources, Wilmington, NC. American
Water Resources Association, Bethesda, MD.

Starr, RD; Sawhney, BL. (1980) Movement of nitrogen and carbon from a septic system
drainfield Water, Air and Soil Pollution 13:113-123.

Starr, RC; Gillham, RW. (1993) Damnification and organic carbon availability in two aquifers.
Ground Water 31:934-947.

Stensland, GJ; Whelpdale, DM; Oehlert, G. (1986) Precipitation Chemistry.  In: Acid
Deposition: Long-Term Trends. Committee on Monitoring and Assessment of Trends in Acid
Deposition, National Research Council. Washington, DC: National Academy Press; pp.

Stokes, JC. (2000) General information on wastewater and alternative onsite sewage systems.
Ad Hoc Committee on Alternative Septic Systems.  New Jersey Pinelands Commission. New
Lisbon, New Jersey. Available at: http://www.state.nj.us/pinelands/428rept.htm.

Stumm W; Morgan, JJ. (1981) Aquatic Chemistry, 2nd ed.  New York: Wiley.

Suter, GW. (1989) Ecological endoints. In: Ecological Assessment of Hazardous Waste Sites: A
Field and Laboratory Reference Document. Warren-Hicks, W; Parkhurst, BR; Baker Jr, SS.
(eds.), EPA 600/3-89/013. Corvallis Environmental Research Laboratory, OR.

Suter, GW. (1993) Ecological Risk Assessment. Boca Raton, FL: Lewis Publishers.

Thayer, GW; Stuart, H. (1974) The bay scallop makes its bed of seagrass. Marine Fisheries
Review 36(7):27-30.

Thayer, GW; Kenworthy, WJ; Fonseca, MS. (1984) The ecology of eelgrass on the Atlantic
coast:  A community profile. FWS/OBS-84-2. US Fish and Wildlife Service, Washington, DC.

Thornthwaite, CW; Mather, JR. (1957) Instructions and tables for computing potential
evapotranspiration and the water balance. In:  Publications in Climatology.  Drexel Institute of
Technology, Philidelphia, PA; 10:185-311.

Tober, JD; Griffin, MPA; Valiela, I. (2000) Growth and abundance of Fundulus heteroclitus and
Menidia rnenidia in estuaries of Waquoit Bay, MA exposed to different rates of nitrogen
loading. Aquatic Ecology 34:299-306.

Tomasky, G; Barak, J; Valiela, I; et al. (1999) Nutrient limitation of phytoplankton growth in
Waquoit Bay, Massachusetts, USA: a nutrient enrichment study. Aquatic Ecology 33:137-155.

Udy, JW; Dennison, WC. (1997) Growth and physiological responses of three seagrass species to
elevated sediment nutrients in Moreton Bay, Australia. Journal of Experimental Marine Biology
and Ecology 217:253-277.

Udy, JW; Dennison, WC. (1998) The use of the seagrass Zostera capricorni to identify
anthropogenic nutrient sources in Moreton Bay. In: Moreton Bay and Catchment. Tibbetts, IR;
Hall, NJ; Dennison, WC (eds.).  The University of Queensland, Brisbane, Australia.

U.S. Census Bureau. (2000) Population Estimates Program, Population Division, Washington,
DC. Data obtained from http://www.census.gov/population/estimates/.

U.S. EPA (Environmental Protection Agency). (1980) Design manual, onsite waste treatment and
disposal systems. EPA/625/1-80/012. Washington, DC.

U.S. EPA. (1982) National air pollution emission estimates 1940-1980. EPA/450/4-82/001.
Washington, DC.

U.S. EPA. (1992) Framework for ecological risk assessment. EPA/630/R-92/OQ1, Office of
Water, Washington, DC.

U.S. EPA. (1996) Watershed approach framework.  EPA/840/S-96/001. Office of Water.
Washington, DC.

U.S. EPA. (1998) Guidelines for ecological risk assessment. EPA/630/R-95/002Fa.
Washington, DC.

Valiela, I. (1995) Marine ecological processes. 2nd ed. New York: Springer-Verlag.

Valiela, I; Bowen, JL. (2002) Nitrogen sources to watersheds and estuaries: role of land cover
mosaics and losses within watersheds. .Environmental Pollution 118:239-248.

Valiela, I; Cole, ML. (2002) Comparing evidence that salt marshes and mangroves protect
seagrass meadows from land-derived nitrogen loads.  Ecosystems 5:92-102.

Valiela, I; Costa, JE. (1988) Eutrophication of Buttermilk Bay, a Cape Cod coastal embayment:
concentrations of nutrients and watershed nutrient budgets. Environmental Management 12:

Valiela I; Teal, JM; Volkmann, S; et al. (1978) Nutrient and particulate fluxes in a salt marsh
ecosystem: tidal exchanges and inputs by precipitation and groundwater. Limnology and
Oceanography 23:708-812.

Valiela, I; Foreman, K; LaMontagne, M; et al, (1992) Couplings of watersheds and coastal
waters: sources and consequences of nutrient enrichment in Waquoit Bay, Massachusetts.
Estuaries 15:443-^4-57.

Valiela, I; Peckol, P; D'Avanzo, C: et al. (1996) Hurricane Bob on Cape Cod. American
Scientist 84:154-165.

Valiela, I; McClelland, J; Hauxwell, J; et al. (1997a) Macroalgal blooms in shallow estuaries:
controls and ecophysiological and ecosystem consequences. Limnol Oceanography

Valiela I; Collins, G; Kremer, J; et al. (1997b) Nitrogen loading from coastal waters to receiving
estuaries: new method and application.  Ecological Applications 7:358-380.  Model available at

Valiela, I; Peckol, P; D'Avanzo, C; et al. (1998) Ecological effects of major storms on coastal
watersheds and coastal watersheds: Hunicane Bob on Cape Cod. Journal of Coastal Research

Valiela, I; Geist, M; McClelland, J; et al. (2000a) Nitrogen loading from watersheds to estuaries:
verification of the Waquoit Bay nitrogen loading model. Biogeochemistry 49:277-293.

Valiela, I; Tomasky, G; Hauxwell, J; et al. (2000b) Operationalizing sustainability: management
and risk assessment of land-derived nitrogen loads to estuaries. Ecological Applications

Valiela, I; Bowen, JL; Cole, ML; et al. (2001) Following up on a Margalevian concept:
interactions and exchanges among adjacent parcels of coastal landscapes. Scientia Marina

Valiela, I; Bowen, JL; Kroeger, KD. (2002) Assessment of different models for estimation of
land-derived nitrogen loads to shallow estuaries. Applied Geochemistry 17: 935-953.

Valiela, I; Bowen, JL; Stieve, EL; et al. ELM, an estuarine nitrogen loading model: formulation,
verification of predicted concentrations, and links to producers. Ecological Modelling: in press.

Vitousek, PM; Howarth, RW. (1991) Nitrogen limitation on land and in the sea: how can it
occur? Biogeochemistry 13:87-115.

Vollenweider, RA.  (1987) Scientific concepts and methodologies pertinent to lake research and
lake restoration. Schweizerische Zeitschrift F4r Hydrologie 49:129-147.

Walker, WG;  Bouma, GJ; Kenney, DR; et al. (1973) Nitrogen transformations during  subsurface
disposal of septic tank effluent in sands: I. Soil transformations.  Journal of Environmental
Quality 2:475^80.

Weinstein, M; Balletto, J. (1999) Does the common reed Phragmities australis affect essential
fish habitat? Estuaries 22:793-802.

Weiskel, PK; Howes, BL. (1991) Quantifying dissolved nitrogen flux through a coastal
watershed. Water Resources Research 27: 2929-2939.

Weiss, E; Carmichael, RH; Valiela, I. (2002) The effects of N loading on the growth rates of
quahogs and softshell clams through changes in food supply. Aquaculture 211:275-289.

Wiegers, JK; Feder, HM; Mortensen, LS; et al. (1998) A regional multiple stressor rank-based
ecological risk assessment for the fjord of Port Valdez, AK. Human and Ecological Risk
Assessment 4(5):1125-1173.

WilHams, SL; Davis, CA. (1996) Population genetic analyses of transplanted eelgrass beds reveal
reduced genetic diversity in Southern California.  Restoration Ecology 4:163-180.

Wuertz, S; Miller, ME; Doolittle, MM; et al. (1991) Butyltins in estuarine sediments two years
after tributyltin was restricted. Chemosphere 22:1113-1120.




United States
Environmental Protection
National Center for
   Environmental Assessment
Washington, DC 20460

Official Business
Penalty for Private Use
Please mate all necessary cnanges en r>e MIOW labe..
cetacri or copy, a->c return to me aOCress in the upper
ett-nand comer.

W you do "01 wish to receive :fiese reports CHECK HEREU
Cetacn. or cooy :h s ccver, and resum ;o ne add^ss in the
upper le*t-riar,c comer
                                                                                     PRESORTED STANDARD
                                                                                      POSTAGE & FEES PAiD
                                                                                         PERMIT No. G-35
October 2002