Mystic River Watershed Alternative TMDL Development for
Phosphorus Management - Final Report
January 2020
Prepared for:
U.S. Environmental Protection Agency
Region 1 - Office of Ecosystem Protection
5 Post Office Square
Boston, MA 02109
Prepared by:
Eastern Research Group, Inc.
100 Hartwell Ave.
Lexington, MA 02421
Subcontractors:
Horsley Witten Group
Paradigm Environmental
Nigel Pickering, PhD
PG Environmental
EPA Contract No. EP-C-16-003
Reviewer:
Dr. Jeff Walker
Walker Environmental Research, LLC
CoVtr Photo: The Tufts University sailing program practicing on Upper Mvstic Lake in Medford with informal swimming at the
Bacow Sailing Pavilion. Upper Mystic Lake is a popular destination with additional access at Massachusetts Department of
Conservation and Recreation's Shannon Beach, and the Medford and Winchester Boat Clubs. Photo credit unknown.
-------
Mystic River Watershed TMDL Alternative Development Final Report
Executive Summary
The Mystic River Watershed, located a few miles north of Boston, is a 76-square mile area that
drains into Boston Harbor. Encompassing all or portions of 22 urban and suburban communities,
the watershed is highly developed and faces multiple water quality impairments. The Massachusetts
Department of Environmental Protection's (MassDEP's) water quality assessment indicates that
nutrients and pathogens are the primary causes of "use impairment" in the freshwater portion of
Mystic River watershed, the focus of this report. Cultural eutrophicationthe degradation of aquatic
environments by nutrient pollution caused by human activity and urban developmentis a major
cause of impairments in the watershed as evidenced by excessive algal and macrophyte growth and
harmful cyanobacteria blooms. Regular occurrences of severe algal blooms during the summer
months reduce water clarity and contribute to anoxic bottom waters that do not support aquatic life.
Algal blooms and macrophyte growth degrade tile aesthetic quality of the river, reduce water clarity,
and impair designated uses such as fishing and boating.
Photo Above: The Mystic River Run and Paddle (May 2016). This annual race celebrates the return of the river herring
and draws the public to the Mystic River. This image looks downstream from Route 16 to the Blessing of the Bay
Boathouse in Somerville with the Boston skyline in the background. Photo credit: Ram Subramanian.
Clean Water Act Requirements
Section 303(d) of the Clean Water Act and the U.S. Environmental Protection Agency's (EPA's)
Water Quality Planning and Management Regulations (Title 40 of the Code of Federal Regulations
[CFR] Part 130) require states to develop Total Maximum Daily Loads (TMDLs) for impaired water
bodies. A TMDL establishes the amount of a pollutant that a water body can assimilate without
exceeding the applicable water quality standard. A TMDL consists of the sum of individual waste
2
-------
Mystic River Watershed TMDL Alternative Development Final Report
load allocations for point sources and load allocations for nonpoint sources and natural background
conditions. The Massachusetts Water Quality Standards (WQS), codified at 314 CMR 4.00, identify
the Mystic River as a Category 5 water body on the Massachusetts 303(d) "List of Impaired Waters"
(2014) for phosphorus, arsenic, chlordane, chlorophyll, DDT (dichlorodiphenyltrichloroethane),
dissolved oxygen, E. coli, PCBs (Polychlorinated biphenyls) [in fish tissue], Secchi depth, and
sediment bio-chronic toxicity.
This report addresses those impairments associated with excessive nutrient loading including
phosphorus, chlorophyll, dissolved oxygen, and secchi depth (water clarity).
Alternative TMDL Process
In 2013, the EPA announced a new framework (Vision) for prioritizing and implementing TMDLs
and pollution control strategies. The guidance for this Vision (found here:
https://www.epa.gov/sites/production/files/2015-07/documents/vision 303d program dec 2013.pdf
allows states to adopt strategies tailored to their water quality program goals and priorities. The
Vision acknowledges that alternative restoration approaches may be more immediately beneficial or
practical in achieving water quality standards than a traditional TMDL. The Vision calls on states to
strategically focus efforts and demonstrate progress over time.
EPA is supporting MassDEP in piloting an alternative TMDL designed to address nonattainment of
nutrient related water quality standards over a period of time. The approach, based on rigorous data
gathering, scientific analysis, and modeling, provides guidance to communities based on a scientific
understanding of conditions. The agencies have already begun working with communities to develop
stormwater management (SWM) strategies to begin progress on implementing effective stormwater
control measures (SCMs) to restore the river and degraded lakes and ponds. This "adaptive
management" approach for the Mystic will be an iterative process of implementing control actions
over an extended period of time while progress is monitored, and new information is gathered to
further inform management needs for attaining water quality standards.
Study and Analysis
The objectives of this technical analysis, conducted between 2017 and 2019, were to: estimate annual
loadings of phosphorus; relate phosphorus loads to response variables in critical surface water
reaches of the watershed; estimate the load reductions needed to improve water quality and attain
water quality standards; and introduce a pilot Opti-Tool analysis that demonstrates cost-effective
and opportunistic stormwater load reduction strategies that communities can consider adopting.
Much of the scientific research needed to document existing conditions occurred prior to this
project, during which time EPA Region 1 collaborated with the Mystic River Watershed Association
(MyRWA), MassDEP, U.S. Geological Survey (USGS) and the Massachusetts Water Resources
Authority (MWRA). Baseline water quality monitoring included the collection of composite samples
linked to streamflow. This project builds on past analyses to develop target reductions in
phosphorus inputs in order to improve water quality in the Mystic River.
To assist with developing phosphorus budgets and recommending load reductions, EPA convened a
Technical Steering Committee (TSC) to provide data, expertise and advice. The Technical Steering
Committee met three times annually. In addition to the Technical Steering Committee, the
consultant team also benefited from expert review provided by Dr. Jeff Walker (Walker
Environmental Research LLC). Further refinements to analyses were conducted based on Dr.
3
-------
Mystic River Watershed TMDL Alternative Development Final Report
Walker's input following consultation with the Technical Steering Committee. This interactive
approach resulting in the project team providing the most comprehensive and complete assessment
possible at this time and one that the Technical Steering Committee determined is suitable to
support an adaptive management process for die Mystic River watershed. Technical Steering
Committee members and our expert reviewer are listed in the Acknowledgements.
The components of the analytical approach, discussed in detail in the report, are listed below:
Develop conceptual model of hydrology and nutrient dynamics
Evaluate existing water quality monitoring data
Review modeling endpoint approaches
Estimate watershed phosphorus loading
Evaluate combined sewer overflow (CSO) and sanitary sewer overflow (SSO) data
Conduct BATHTUB modeling and calibrate results
Determine critical period of interest for phosphorus load reduction analysis
Evaluate watershed phosphorus load reduction analysis
Develop nutrient stormwater management strategies using Opti-Tool
The pollutant of concern for this study is phosphorus because it is directly causing or contributing
to the excessive algal biomass. Since there are no numeric criteria available for phosphorus (i.e., no:
specific concentration of phosphorous that represents a violation of standards), a surrogate water
quality target was needed to calculate pollutant load reductions to the river. Chlorophyll-^ was
chosen as the surrogate water quality target. Chlorophyll-^ is the photosynthetic pigment found in
algae and is, therefore, a direct indicator of algal biomass. EPA and the Technical Steering
Committee determined that a seasonal average chlorophyll-a concentration of <10 j-ig/L would be
protective of narrative eutrophication standards in the watershed, from which associated total
Photo Left:
Cyanobacteria bloom
in the freshwater
segment of the Mystic
River between
Arlington and Medford
in June of 2017. Photo
credit: Jack Bitney,
4
-------
Mystic River Watershed TMDL Alternative Development Final Report
Results
Watershed analyses conducted during this study demonstrate that inadequately controlled
stormwater (SW) runoff from developed landscapes are the predominant source of nutrient loads
specifically phosphorus loadsto the surface waters of the Mystic River watershed. Under existing
conditions, this study estimated that to meet the selected chlorophyllwater quality target for
attaining water quality standards in the most impacted segment, the lower Mystic River, will require a
67 percent reduction of stormwater phosphorus loadings from the watershed However, this
estimate assumes all reduction would be achieved through stormwater control measures.
Load reduction estimates were also modeled for future conditions to account for key variables: wet
vs. dry years; future control of combined sewer overflows and sanitary sewer overflows and
sediment load reduction. Overall, the analysis showed that elimination of combined sewer overflows
and sanitary sewer overflows had minimal impact compared to reducing stormwater loads and
internal loads released from bottom sediments of the river system. The difference between wet vs.
dry years is significant, with much greater difficulty meeting water quality targets during dry years.
The stormwater load reductions required to meet water quality targets under future conditions
(which account for baseline stormwater management, combined sewer overflows/sanitary sewer
overflows controls and an estimate of associated reductions in internal loads) were between 59 and
62 percent.
The Path Ahead
Knowing how and where to site cost-effective stormwater controls to reduce phosphorus loads
from stormwater runoff will be critical for meeting state and federal water quality regulations. The
Opti-tool analysis included in this report shows that by optimizing sizing and location of BMPs,
significant cost savings can be realized. ERG and the project team is currently working with
communities in the watershed to develop cost-effective stormwater best management practices with
a focus on green infrastructure solutions and changes to local bylaws/ordinances to streamline the
process. .EPA envisions a sustained collaborative process of working with the communities to
develop realistic and effective strategic stormwater management approaches to effectively advance
watershed restoration efforts.
Photo: Left: River
herring above the fish
ladder in Upper Mystic
Lake in 2012. This was
the first year since the
late 1800s that had
significant passage of
anadromous fish into
the Upper Mystic Lake.
Photo credit: Patrick
Herron, Mystic River
Watershed Association.
5
-------
Mystic River Watershed TMDL Alternative Development Final Report
Acknowledgements
The following individuals participated on the technical steering committee and EPA thanks them for
their valuable contribution to this project.
Name
Agency/Affiliation
Lise Marx
MWRA
David Wu
MWRA
Chris Goodwin
MWRA
Maret Smolow
MWRA
David Taylor
MWRA
Toby Stover
EPA
Erik Beck
EPA
Suzanne Warner
EPA
Mark Voorhees
EPA
Shutsu Wong
EPA
Newton Tedder
EPA
Barbara Kickham
MassDEP
Matt Reardon
MassDEP
Patrick Herron
MyRWA
Andy Hrycyna
MyRWA
Arleen O'Donnell
ERG
Braden Rosenberg
ERG
Matt Reusswig
PG Environmental (ERG)
Adriane Garneiter
PG Environmental (ERG)
Jennifer Relstab
Horsley Witten Group
Nigel Pickering
Nigel Pickering
Jeff Walker
Walker Environmental Research
Jamie Houle
University of New Hampshire
Khali d Alvi
Paradigm
Ryan Murphy
Paradigm
Photo Above: Rowers practicing on the Maiden River, a tributary to the Lower Mystic River in Medford. Photo credit:
Greig Cranna.
6
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table of Contents
Executive Summary 2
Clean Water Act Requirements 2
Alternative TMDL Process 3
Study and Analysis 3
Results 5
The Path Ahead 5
Acknowledgements 6
I. Introduction 16
LA. Background 16
I.B. Purpose 16
I.C. Organization and Overview 17
II. Conceptual Model of the Hydrology and Nutrient Dynamics in the Mystic River Basin 18
IIA. Watershed Overview 18
II.B. Upper Watershed 20
II.C. Central Watershed 20
II.D. Streamflow Impact of Mystic Lakes Dam 21
II.E. Water Quality Impact of Mystic Lakes Dam 21
II.F. Lower Watershed 22
II.G. Hydrologic and Hydraulic Impacts of Amelia Earhart Dam 23
II.H. Water Quality Gradient Along the Mystic River 23
ILL Long Term Changes in Water Quality 24
II.J. Impact of Aquatic Vegetation on Water Quality 25
III. Review of Existing Water Quality Monitoring Data 26
IIIA. Data Gaps and Recommendations for Future Sampling Efforts 26
III.A.l. Ecological/Biological Indicators of Over-Enrichment 26
III.A.2. Streamflow 26
III.A.3. Sediment 26
III.B. MyRWA and MWRA Monitoring Data 27
III.B.l. Data Characterization 27
III.B.2. Data Vivien' 37
III.C. USGS Flow Data 38
III.D. GIS Datasets 38
IV. Review of Modeling Endpoint Approaches 39
7
-------
Mystic River Watershed TMDL Alternative Development Final Report
IV.A. Water Quality Standards Applicable to the Mystic River Watershed
IV.B. Existing Regional/Local Targets
IV.C. Reference Water Body Conditions
IV.D. Stressor-Response Relationships
IV.E. Mechanistic Models
IV.F. Recommendations
V. Watershed Phosphorus Loading Estimates
VA. Stormwater Loads
V.B. Groundwater Loads
V.C. Sediment Loads
V.D. Observed Loads
V.D. 1. Adjustments of Streamflow
V.D.2. Adjustments of Total Phosphorus Concentrations
V.D.3. Observed Total Phosphorus Loads
V.D.4. Calibrated Streamflow
V.D.5. Calibrated Total Phosphorus Loads
VI. Evaluation of Combined Sewer Overflow and Sanitary Sewer Overflow Data for the Mystic
River Watershed
VIA. Data Types and Sources
VI. A. 1. Spatial Data
VI.A.2. I Tolnmetric Data
VI.A.3. Precipitation Data
VI.A.4. Nutrient Concentrations
VI.B. CSO Data Analyses
VI.B.l. CSO Analyses
VI.B.2. CSO Statistical Outliers
VI.C. SSO Data Analyses
VI.C.l. SSO Data Rei'ieii'.
VI.C.2. SSO Data Processing
VI.C.3. SSO Trend Analyses
VI.C.4. SSO Statistical Outliers
VI.C.5. SSO Missing Datasets
VI. C. 6. /otal Phosphorus and Total Nitrogen Loads for Model Calib ratio n
VII. BATHTUB Modeling Approach
VILA. Model Selection
,40
,40
,41
,43
,46
,47
,48
,48
,56
,57
,58
,58
,59
,61
,62
,62
,66
,66
,66
,67
,67
,67
,68
,68
,68
,71
,72
,73
,73
,74
,74
,74
,80
,80
8
-------
Mystic River Watershed TMDL Alternative Development Final Report
VII.B. Model Setup 81
VII.B.l. Segmentation 81
VII.B.2. Model Options 82
VII.B.3. Atmospheric Fluxes 83
VII.B.4. Modeled Land Loads 84
VII.B. 5. External Loads 85
VII.B.6. Internal Loads 85
VII.C. Model Calibration and Validation 86
VII.C. 1. Water Quality Data Availability 86
VII.C.2. Calibration and I Talidation Periods 86
VII.C.3. Receiving Water Parameters 87
VII.C.4. Dispersion 87
VII.C. 5. Nutrient Availability Factors 88
VII.C.6. Internal Loads 88
VII.C.7. Chl-a Model. 88
VII.C.8. Model Calibration 89
VII.C.9. Model I Talidation 94
VII.C.10. Model Sensitivity Analysis 98
VIII.Critical Period of Interest for Phosphorus Load Reduction Analysis 100
VIII.A. Water Quality Data 100
VIII.B. Combined Sewer and Sanitary Sewer Overflow Data 100
VIII.C. Rainfall Data 100
VIII.D. Critical Period Selection 102
VIII.E. Extreme Rainfall Years 102
IX. Evaluation of Watershed Phosphorus Load Reduction Analysis 103
IX.A. Phosphorus Loading Estimates for Critical Period of Interest 103
IX.A. 1. Existing Conditions 103
IX.A.2. Future Conditions 103
IX.B. Scenarios for Evaluation of Phosphorous Reduction 105
IX.C. Modeling Methodology 106
IX.C.l. Model Setip 106
IX.C.2. Model Inputs 106
IX.C.3. Analysis ivith Wet and Dry Year Data 107
IX.D. Model Results with Average Annual Data 107
9
-------
Mystic River Watershed TMDL Alternative Development Final Report
IX.E. Analysis with Wet and Dry Year Data 110
IX.F. Discussion 113
X. Broad-Based Nutrient Stormwater Management Strategies for the Mystic River Watershed
using Opti-Tool 116
X.A. Study Objectives 116
X.B. Pilot Sub-Watershed Selection 116
X.C. Technical Approach 123
X.C.I. S form water Management Categories Development 123
X.C.2. Estimating BMP Footprints and Impervious Drainage Areas 127
X.C.3. Opti-Tool Setup 128
X.D. Management Scenarios 130
X.D. 1. Results: Scenario 1 131
X.D.2. Results: Scenario 2 135
X.E. Summary 138
X.F. Example Projects 140
X.F.I. Berry Brook Project, Dover, New Hampshire 140
X.F.2. The Advancing Green Infrastructure Program, New Haven, Connecticut 141
XI. References 144
Appendix A: Water body and Monitoring Location Summary Table 148
Appendix B: Water Quality Parameters Included in Monitoring Programs 151
Appendix C: Descriptive Figures of Water Quality Data by Water body 152
Appendix D: Identified Extreme Values 159
Appendix E: Modeled Stormwater TP Load and Rainfall-Runoff Results 160
Appendix F: Baseflow Estimates for Aberjona River and Alewife Brook 168
Appendix G: Maps Depicting CSO Drainage Basins 169
Appendix H: Water Quality Data Used in Calibration and Validation of the Bathtub Model 173
Appendix I: Bathtub Model Inputs for Calibration 175
Appendix J: Bathtub Model Inputs and Outputs for Scenarios 179
Appendix K. BMP Design Parameters Used in the Pilot Watershed 187
10
-------
Mystic River Watershed TMDL Alternative Development Final Report
List of Tables
Table III-1. Water Quality Parameters Included in Each Monitoring Program 28
Table III-2. Temporal Range and Duration of Monitoring within Water Bodies 30
Table III-3 Number of Seasons Wiere Few (N < 3) Samples were Collected for a Waterbody-
Parameter Combination 34
Table III-4. Summary of Documented Data Quality Issues 35
Table III-5 Summary of Non-Detect Data.1 35
Table III-6. Summary of USGS Monitoring Stations 38
Table III-7. GIS Data Sources Available for the Mystic River Watershed 39
Table IV-1. Existing Water Quality Targets 41
Table IV-2. Eco-regional Criteria and Observed 25th Percentile (2000 2016) Mystic River
Watershed Values 43
Table IV-3. Simple Linear Regression on Paired TP & Chlorophyll-a Monitoring Results Collected
April - October 44
Table IV-4. Simple Linear Regression on Average TP and Average or 90th Percentile Chlorophyll-a1
45
Table V-l. Opti-Tool-SWMM HRU Model Climatological Input Datasets 49
Table V-2. Land-Use Categories 50
Table V-3. DCIA Adjustment Factors from MA MS4 Permit 51
Table V-4. Opti-Tool Export Rates by HRU 51
Table V-5. GIS Layers Used to Develop HRUs 52
Table V-6. Estimated Reach Detention Times and Attenuation Factors 64
Table VI-1. Volumetric Data Sources for CSOs and SSOs 67
Table VI-2. Statistical Outlier Analysis for CSO Datasets 71
Table VI-3. CSO Volumetric Datasets for Alewife and Mystic River Drainage Basins 71
Table VI-4. Review of SSO Data Frequency, Duration and Discharge Points 72
Table VI-5. Review of SSO Data Volumes 73
Table VI-6. Statistical Outlier Analysis for Annual SSO Discharge Volumes (Gallons/yr.) 76
Table VI-7. Annual SSO Discharge Volumes for All Sub watersheds (Gallons/yr.) 76
Table VI-8. Estimated Annual Total Phosphorus and Total Nitrogen CSO Loads (lbs./yr.) for
Alewife and Mystic Sub watersheds 77
Table VI-9. Estimated Annual Total Phosphorus SSO Loads (lbs./yr.) for all Sub watersheds 78
Table VI-10. Estimated Annual Total Nitrogen SSO Loads (lbs./yr.) for all Sub watersheds 78
Table VII-1. Model Options 83
Table VII-2. Upstream Reach and Local Sub-basin Contributions 85
Table VII-3. Average Water Quality Data Available by Segment, Site and Year 86
Table VII-4. Total Phosphorus Loads by Segment for Calibration Period 92
Table VII-5. Results of BATHTUB Sensitivity Analysis 99
Table VIII-1. Standard Precipitation Index (SPI) Reference Values 101
Table VIII-2. Summary of Rainfall Data Analyses 101
Table IX-1. Modeled Scenarios for Phosphorus Load Reduction Evaluations 106
Table IX-2. BATHTUB Model Runsawith Average Annual Data 107
Table IX-3. Scenario Results3 with Average Annual Data 109
Table IX I. Model Runsawith Wet and Dry Year Data Ill
Table IX-5. Scenario Results3 with Wet and Dry Year Data 111
Table IX-6. Total Phosphorus Load Reductions for Scenario 2A 114
Table X-l. Ground Slope Classification in Pilot Sub-Watershed, Mystic River 118
11
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table X-2. Hydrologic Soil Group Classification in Pilot Sub-Watershed, Mystic River 119
Table X-3. Land Use Classification in Pilot Sub-Watershed, Mystic River 123
Table X-4. Potential Stormwater Management Categories and BMP Types in Opti-Tool 125
Table X-5. Potential BMP Opportunity Areas (Maximum Footprints) in the Pilot Sub-Watershed,
Mystic River 125
Table X-6. BMP-Treated Impervious Area (Drainage Area) Distribution by Land Use Category
Group in the Pilot Sub-Watershed, Mystic River 127
Table X-7. BMP Area (Footprints) Distribution by Land Use Category Group Required to Treat 1
Inch of Runoff from the Impervious Surface in the Pilot Sub-Watershed, Mystic River 128
Table X-8. Scenario 1: BMP Types and BMP Sizes for The Selected Target Solution 1 (Existing
Condition) in Opti-Tool 133
Table X-9. Scenario 1: BMP Types and BMP Sizes for the Selected Target Solution 2 (Future
Condition 1) in Opti-Tool 133
Table X-10. Scenario 1: BMP Types and BMP Sizes for the Selected Target Solution 3 (52% Target)
in Opti-Tool 134
Table X-ll. Scenario 2: BMP Types and BMP Sizes for the Selected Target Solution 1 (Existing
Condition) in Opti-Tool 136
Table X-12. Scenario 2: BMP Types and BMP Sizes for the Selected Target Solution 2 (Future
Condition 1) in Opti-Tool 137
Table X-13. Scenario 2: BMP Types and BMP Sizes for the Selected Target Solution 3 (52% Target)
in Opti-Tool 138
Table X-14. BMP Scenarios Comparison in Opti-Tool 139
12
-------
Mystic River Watershed TMDL Alternative Development Final Report
List of Figures
Figure II-I Schematic Diagram and Map of Mystic River Watershed 19
Figure II-II Instantaneous Streamflow Above and Below Mystic Lakes 21
Figure II-III Annual Distributions of Monthly TP Concentrations at the Aberjona River Outlet
(ABR006), Mystic Lake Dam (UPL001), and Start of the Mystic River Mainstem (MYR071) 22
Figure II-IV Relative Water Surface Elevation at USGS Gages on Alewife Brook and Mystic River
(Sept. 1 - 10, 2016) 23
Figure II-V Chlorophyll-a and TP Concentrations Along the Mystic and Maiden Rivers, 2015 2016
24
Figure II-VI Annual Mean, Median, 90th Percentile of Chlorophyll-a and TP at Three Stations on
the Mystic River, 2000 2016 (June Oct. only) 25
Figure III-I. Mystic River Watershed Drainage Basins and Monitoring Locations 29
Figure III-II. Number of Years in Which Observations Were Made for Eutrophication Related
Parameters in the Watershed 31
Figure III-III Number of Years in Which Observations Were Made for Eutrophication Related
Parameters in The Mystic River 32
Figure III-IV. Temporal Coverage (both number of days and Percentage of Total Available Days
from 2000 2016) of Water Quality Parameters by Water Body 33
Figure IV-I. Reference Condition Approach for Establishing Numeric Water Quality Targets (EPA,
1998) 42
Figure IV-II. April October Total Phosphorus vs. Chlorophyll-a Concentrations in the Upper
Lobe of the Upper Mystic Lake (Monitoring Station UPLUPL) 44
Figure IV-III. April October Average Total Phosphorus vs. Average Chlorophyll-a Concentrations
in the Mystic River 46
Figure IV-IV. April October Average Total Phosphorus vs. 90th Percentile Chlorophyll-a
Concentrations in the Mystic River 46
Figure V-I. Estimated Annual Runoff Volume and TP Load for the Mystic River 54
Figure V-II. Mystic River Watershed Sub-Basin Delineation and Schematic Diagram for Final Model
55
Figure V-III. Baseflow Estimates and Streamflow Measurements at Aberjona River for 2016 (USGS
Gage 01102500) 56
Figure V-IV. Baseflow Estimates and Streamflow Measurements at Alewife Brook for 2016 (USGS
Gage 01103025) 57
Figure V-V. Regression of Flow Measurements at Aberjona River (01102500) and Alewife Brook
(01103025) USGS Stations 58
Figure V-VI. Regression of Flow Measurements at Aberjona River (01102500) and Alewife Brook
(01103025) USGS Stations 59
Figure V-VII. Correction of TP Values from Method 365.1 to 365.4 60
Figure V-VIII. Adjusted Total Phosphorus Concentrations in the Aberjona River 60
Figure V-IX. Adjusted Total Phosphorus Concentrations in the Mystic River 61
Figure V-X. Adjusted Total Phosphorus Concentrations in Alewife Brook 61
Figure V-XI. Phosphorus Attenuation Curves 63
Figure V-XII. Modeled Land vs. Attenuated Reach Phosphorus Loads (2007-2016) 65
Figure VI-I. Annual CSO Volumes over Time 69
Figure VI-II. Annual CSO Volumes versus Annual Rainfall 69
Figure VI-III. Normalized Alewife Annual CSO Volumes versus Annual Precipitation 70
13
-------
Mystic River Watershed TMDL Alternative Development Final Report
Figure VI-IV. Normalized Mystic Annual CSO Volumes versus Annual Precipitation 70
Figure VI-V. Annual SSO Discharge Volumes by Sub-watershed versus Time 74
Figure VII-I. BATHTUB Model Schematic 81
Figure VII-II. BATHTUB Segmentation for the Mystic River 82
Figure VIITII. Annual Precipitation and Lake Evaporation 84
Figure VIITV. Dispersion Effect on Modeled Phosphorus Concentrations 88
Figure VII-V. Predicted vs. Observed TP by Segment for Calibration Period 90
Figure VII-VI. Predicted vs. Observed TP Relationship for Calibration Period 90
Figure VII-VII. Predicted vs. Observed Chl-a by Segment for Calibration Period 91
Figure VII-VIII. Predicted vs. Observed Chl-a Relationship for Calibration Period 91
Figure VII-IX. Calibration 2015 - Total Phosphorus Loads for Mystic River 93
Figure VII-X. Validation 2016 - Predicted vs. Observed TP by Segment 95
Figure VII-XI. Validation 2016 - Predicted versus Observed Chl-a by Segment 95
Figure VII-XII. Validation 2017 - Predicted vs. Observed TP by Segment 96
Figure VII-XIII. Validation 2017 - Predicted vs. Observed Chl-a by Segment 96
Figure VII-XIV. SSO Outlier Replacement 2010 - Predicted versus Observed TP 97
Figure VII-XV. SSO Outlier Replacement 2014 - Predicted versus Observed TP 97
Figure VIIIT. Standard Precipitation Index (SPI) between 2000-2017 102
Figure IX-I. Predicted Phosphorus Concentrations with Average Annual Data 110
Figure IX-II. Predicted Chlorophyll-a Concentrations with Average Annual Data 110
Figure IX-III. Predicted Phosphorus Concentrations with Wet and Dry Year Data 113
Figure IX-IV. Predicted Chlorophyll-a Concentrations with Wet and Dry Year Data 113
Figure X-I. Location Map of Pilot Sub-Watershed, Mystic River (Highlighted in Yellow) 118
Figure X-II. Ground Slope Map of Pilot Sub-Watershed, Mystic River 119
Figure X-III. Soil Map (Hydrologic Soil Group) of Pilot Sub-Watershed, Mystic River 120
Figure X-IV. Land Use Map of Pilot Sub-Watershed, Mystic River 121
Figure X-V. Land Cover Map of Pilot Sub-Watershed, Mystic River 122
Figure X-VI. Stormwater Management Categories Map of Pilot Sub-Watershed, Mystic River 126
Figure X-VII. Treated Impervious Areas by BMP Type and Land Use Type in the Pilot Sub-
Watershed, Mystic River 128
Figure X-VIII. Opti-Tool Model of Pilot Sub-Watershed, Mystic River 130
Figure X-IX. Scenario 1: Opti-Tool Cost-Effectiveness Curve (Optimize Locations Only) of TP
Annual Average Load Reduction for the Pilot Watershed 132
Figure X-X. Scenario 2: Opti-Tool Cost-Effectiveness Curve (Optimize Locations and BMP Sizes)
of TP Annual Average Load Reduction for the Pilot Watershed 136
Figure X-XI. Green Infrastructure Retrofits for Berry Brook Project in Dover, New Hampshire ..141
Figure X-XII. Homeowners and Common Ground High School students plant perennials in a New
Haven Bioswale 143
14
-------
Mystic River Watershed TMDL Alternative Development Final Report
Abbreviations
cso
Combined sewer overflow
Chl-a
Chlorophyll-a
CSS
Combined sewer separation
DCIA
Directly connected impervious area
DCR
Massachusetts Department of Conservation and Recreation
EPA
U.S. Environmental Protection Agency
ERG
Eastern Research Group
HRU
Hydrologic Response Units
HUC
Hydrologic Unit Code
LLRM
Lake Loading Response Model
MassDEP
Massachusetts Department of Environmental Protection
MS4
Municipal Separate Storm Sewer System
MWRA
Massachusetts Water Resources Authority
MyRWA
Mystic River Watershed Association
SSO
Sanitary sewer overflow
TMDL
Total Maximum Daily Load
TN
Total nitrogen
TP
Total phosphorus
USGS
U.S. Geological Survey
WQS
Water quality standards
15
-------
Mystic River Watershed TMDL Alternative Development Final Report
I. Introduction
LA. Background
The Mystic River Watershed is a 76-square mile watershed that drains into Boston Harbor. Located
in the Greater Boston, Massachusetts metropolitan area, it encompasses all or portions of 22 urban
and suburban communities. The watershed faces multiple water quality impacts related to cultural
eutrophication including excessive algal growth, harmful cyanobacteria blooms, and invasive
macrophyte growth. Sources of pollutants from the watershed include stormwater runoff, combined
sewer overflows (CSO), sanitary sewer overflows (SSO), non-point runoff, contaminated sediment,
and three Superfund sites. The watershed suffers from many legacy pollutants as well as present day
pollutant loadings. Several environmental justice communities are located within the watershed, and
although most developable land is built upon, there is high development and re-development
pressure throughout the watershed.
The Mystic River is listed as a Class B water with a Category 5 water quality rating in the
Massachusetts 303(d) "List of Impaired Waters" (2014) for phosphorus, arsenic, chlordane,
chlorophyll, DDT, dissolved oxygen, E. coli, PCB in fish, Secchi depth, and sediment bio-chronic
toxicity. Due to the multiple stressors present in this watershed, development of a traditional Total
Maximum Daily Load (TMDL) to address all pollutants would be a lengthy and complicated task,
especially considering resource limitations. Instead, the Massachusetts Department of
Environmental Protection (MassDEP), U.S. Environmental Protection Agency (EPA) Region 1
(serving New England) and the Mystic River Watershed Association (MyRWA) have embarked on a
plan to pilot an alternate but equally rigorous method ("Alternative TMDL") for determining how to
address impairments, including the nutrient water quality studies documented herein. These studies
as well as other efforts have determined that effective nutrient management will likely go a long way
towards addressing sources of other impairments in the watershed such as bacteria and sediment-
bound contaminants.
I.B. Purpose
The purpose of this project is to support elements of EPA's TMDL Vision process by providing
technical support for watershed restoration efforts to address phosphorus load reduction needed to
meet water quality targets in the Mystic River Watershed. This project provides an opportunity to
achieve multiple TMDL "vision" goals:
Estimate the load reduction needed to meet water quality targets in critical water bodies
within the watershed;
Engage with communities, state and regional governmental agencies and the local watershed
group;
Inform and guide load reduction implementation by municipalities in the watershed;
Integrate actions needed to address multiple Clean Water Act programs, such as point and
non-point pollution.
Findings will inform the development of analytical tools for EPA Region 1 to estimate phosphorus
load reductions that are needed to attain applicable Massachusetts surface water quality standards
(WQS) related to cultural eutrophication. Another project goal is to strengthen regional
collaborations for enhanced watershed nutrient management approaches.
16
-------
Mystic River Watershed TMDL Alternative Development Final Report
I.C. Organization and Overview
This report is a compilation of nine related technical memoranda that Eastern Research Group
(ERG), its sister company PG Environmental, and subcontractors Horsley Witten Group (HW),
Paradigm Environmental and Dr. Nigel Pickering, have developed over the past two years.
Components of the study are listed and described below:
Phase 1
o Conceptual Model of Hydrology and Nutrient Dynamics in the Mystic River Basin
(Section II)
o Review of Existing Water Quality Monitoring Data (Section III)
o Review of Modeling Endpoint Approaches (Section IV)
o Watershed Phosphorus Loading Estimates (Section V)
Phase 2
o Evaluation of Combined Sewer Overflow and Sanitary Sewer Overflow Data for the
Mystic River Watershed (Section VI)
o BATHTUB Modeling Approach and Calibration Results (Section VII)
o Critical Period of Interest for Phosphorus Load Reduction Analysis (Section
VII.C.9)
o Evaluation of Watershed Phosphorus Load Reduction Analysis (Section IX)
o Broad-Based Nutrient Stormwater Management Strategies for the Mystic River
Watershed using Opti-Tool (Section X)
Phase 1 of the project began with a review of the hydrographic and geographic features of the
Mystic River in 2017-2018, and of the vegetation management practices used in the river in the
preceding two decades. This review informed the development of a conceptual model of flows and
nutrient dynamic processes within the Mystic River. Concurrently, the investigators performed an
exploratory review of the available flow and water quality data within the watershed, identification of
gaps in the data record, and identification of potentially useful GIS information. The result was the
identification of information and data sources that might be usefully exploited in planning modeling
of the watershed, presented in Sections II and III of this report. In addition, these sections
document the region of interest (i.e., the freshwater portion of the Mystic River above Amelia
Earhart Dam) used for later modeling activities.
Under the federal Clean Water Act, Massachusetts statutes, and applicable regulations, narrative
nutrient standards are applicable to the watershed. Section IV of this report presents and discusses
several options for translating the applicable narrative water quality standard to one or more numeric
endpoints. A numeric endpoint provides a standard for water quality modelers to ascertain that
water quality conditions comply with applicable standards. This section includes a review of all
approaches considered by EPA, the consultant team and the Technical Steering Committee (TSC),
including approaches that were not ultimately utilized in later modeling and analyses.
Using available water quality data and GIS information as inputs, the project team developed a
modeling approach to estimate an annual time series of total phosphorus and streamflow within the
watershed modeling domain. The methodology employed for estimating nutrient load and flow time
series from combined sewer overflows (CSOs) and from sanitary sewer overflows (SSOs) is
17
-------
Mystic River Watershed TMDL Alternative Development Final Report
documented; time series based on all other sources (e.g., precipitation-driven overland flow,
groundwater, etc.) are also documented in this report.
Early in Phase I, EPA and the project team identified a discrepancy between measurements of total
phosphorus at two different labs used by local monitoring programs to measure in-stream water
quality. The source of the discrepancywhich was consistent and predicable between the two
datasetswas not resolvable during the Phase I or Phase II development period and, therefore,
EPA determined to adjust the observed total phosphorus measurements to make the data from both
laboratories mutually consistent. This rendered both datasets usable for model calibration purposes
and will allow future adjustments to the dataset and model results should EPA make a final
determination regarding the source of the discrepancy. The basis for the total phosphorus data
adjustment is documented in Section V.D.
Sections VII through X cover work completed under Phase II including the setup and calibration of
BATHTUB model, which was used to model water quality response for select reaches of the
watershed. The information and rationale used in selecting a critical period of interest for the model
is described as well as the modeling scenarios evaluated (e.g., baseline conditions, current conditions,
future loading conditions) and the resulting load reduction targets needed to meet water quality
standards.
II. Conceptual Model of the Hydrology and Nutrient
Dynamics in the Mystic River Basin
This section presents a conceptual model of the hydrology and water quality dynamics in the Mystic
River Watershed. The purpose of this conceptual model is to describe the primary sources and sinks
of nutrients, and to highlight the important features (e.g., dams/impoundments) and dynamics (e.g.,
macrophytes) that affect water quality conditions in the Mystic River. The conceptual model is
intended to serve as a foundation for evaluating alternative target endpoints (Section IV) and
modeling strategies (Section V).
II.A. Watershed Overview
The freshwater portion of the Mystic River Watershedthe focus of this study has a total drainage
area of 63 sq. miles. The watershed can be divided into three sub-watersheds referred to as the
upper, central, and lower watersheds. The delineation of these sub-watersheds was based on the
locations of U.S. Geological Survey (USGS) streamflow gages as well as the impacts of the Mystic
Lakes on both the hydrology and water quality of the river.
Figure II-I shows a schematic diagram and corresponding map of the major sub-basins, water
bodies, streamflow gages, and dams across the watershed.
18
-------
Mystic River Watershed TMDL Alternative Development Final Report
Central
Watershed
^ USGS Flow Gage
\y Dam
Major Tributary
[^j Minor Tributaries/Direct Runoff
j] Receiving Water
Boston Harbor
Note this figure does not include all flow gages or dams across the watershed.
Figure ll-l Schematic Diagram and Map of Mystic River Watershed
Legend
V Dam
# USGS Flow Gage
I I Upper Watershed
I I Central Watershed
I I Lower Watershed
19
-------
Mystic River Watershed TMDL Alternative Development Final Report
II.B. Upper Watershed
The upper watershed contains the Aberjona River basin, which has a total drainage area of 25 sq.
miles (40 percent of the total freshwater Mystic River Watershed). A USGS streamflow gage (Station
01102500) is located near the outlet of the Aberjona just before it flows into Upper Mystic Lake.
There is currently one major impoundment on the Aberjona River at the Center Falls Dam in
Winchester, which is located about 0.5 mile upstream of the USGS flow gage. Historically, several
smaller dams were also constructed to support mill operations (Knight, 2016). The exact number
and locations of existing small dams along the Aberjona River requires further research. There are
also numerous impounded ponds (e.g., Horn Pond, Wedge Pond) along the tributaries to the
Aberjona River. These impoundments likely reduce the phosphorus loads originating from the land
surface through settling and vegetative uptake.
The primary source of nutrient loading in the upper watershed is stormwater runoff. There are no
combined sewer areas or wastewater treatment facility discharges. However, much of the watershed
is served by separate sanitary sewer systems and storm sewer drainages systems. Illicit discharges of
sanitary sewage to separate storm drainage systems are not uncommon in suburban/urban
watersheds and are known to exist within the upper watershed, although programs to eliminate them
are underway. Additionally, SSOs are also known to occur infrequently, typically only during major
storm events. Internal loading from sediment fluxes is likely not significant but may occur in some
impounded areas such as above the Center Falls Dam in Winchester.
II.C. Central Watershed
The central watershed encompasses the Upper and Lower Mystic Lakes, which have a total drainage
area of 9 sq. miles (15 percent of total freshwater watershed) excluding the upper watershed. The
two lakes are separated by the Upper Mystic Lake Dam, which was recently rebuilt in 2012. Both
lakes have a maximum depth of about 80 ft. While the Upper Mystic Lake undergoes seasonal
stratification from spring to fall, Lower Mystic Lake is perennially stratified due to entrapment of
saltwater that prevents complete turn over in the fall (Ludlam and Duval, 2001). Water quality and
streamflow data show that the two lakes have a significant effect on both the hydrology and water
quality of water flowing from the mouth of the Aberjona River to the head of the Mystic River.
The primary source of nutrients to Upper Mystic Lake is the discharge from the Aberjona River in
addition to direct runoff from the local drainage basin. Discharge from the Aberjona River first
enters what is referred to as the upper lobe of Upper Mystic Lake. Due to the nutrient loading from
the Aberjona River, the upper lobe, which is shallow and not likely to stratify, frequently experiences
eutrophic conditions, including cyanobacteria blooms and excessive aquatic vegetation.
Outflow from Upper Mystic Lake passes through the dam into Lower Mystic Lake, which also
receives inflow from direct runoff of its local drainage basin and from Mill Brook (drainage area of
5.5 sq. miles). Because of the stratification in the two lakes, internal loading from sediment fluxes
may not be a significant source of nutrients during the growing season because the bottom water
(hypolimnion) is thought to not fully mix with the surface water (epilimnion); however, this question
is still under investigation.
The Massachusetts Department of Conservation and Recreation (DCR) operates the Mystic Lakes
Dam. Information about dam operations such as release schedules or target water levels are not
currently available. However, prior to storm events, additional water is often released to increase
available storage in Upper Mystic Lake and prevent flooding along its shoreline. Because of these
20
-------
Mystic River Watershed TMDL Alternative Development Final Report
operations,, the dam has a significant effect on the flow and water quality dynamics between the
Aberjona and Mystic Rivers.
II.D. Streamflow Impact of Mystic Lakes Dam
Figure II-II shows the instantaneous streamflow over a 7-day period in November 2016 at the
Aberjona River gage above Upper Mystic Lake (01102500) and at the Mystic River gage just below
Lower Mystic Lake (01103010). Lhe Aberjona River gage shows a typical storm hydro-graph during
the night of November 15. Lhe Mystic River gage also shows evidence of an increase in flows due to
stormwater; however, the shape and magnitude of the hydrograph is significantly altered. Lhe peak
flow is lower below the lakes and occurs about 6 hours later. There is also a significant diurnal
pattern in flows at the Lower Mystic Lake outlet due to the release of water during low tides at the
Amelia Earhart Dam located about 5 miles downstream (discussed below) near the mouth of the
Mystic River. Lastly, the total volume of flow during storm events is in fact higher at the Aberjona
River gage than at the Mystic River gage likely due to some of the flow being stored in Upper Mystic
Lake.
150-
Nov 15
Nov 17
Nov 19
Nov 21
Aberjona (01102500)
Mystic @ Lower Lake Outlet (01103010)
Figure ll-li Instantaneous Streamflow Above and Below Mystic Lakes
I I.E. Water Quality Impact of Mystic Lakes Dam
In addition to the change in the streamflow hydrograph, the lakes also have a significant effect on
water quality., Figure II-III shows the distribution of monthly total phosphorus (LP) concentrations
measured by the MyRWA Baseline Monitoring Program at the Aberjona River streamflow gage
(ABR006), the Mystic Lakes dam (UPL001) at the outlet to the Upper Mystic Lake, and below the
outlet of the Lower Mystic Lake at the mouth of the Mystic River mainstem (MYR071). LP
concentrations are highest coming out of the Aberjona River but drop significantly at the outlet of
Upper Mystic Lake likely due to settling of particulate phosphorus and uptake by aquatic plants. In
Lower Mystic Lake, concentrations increase from the levels coming through the dam due to loads
from Mill Brook and other tributaries. Lhis figure demonstrates the critical role the lakes have in
"resetting" the water quality from the upper watershed (Aberjona River) to the Mystic River
21
-------
Mystic River Watershed TMDL Alternative Development Final Report
mainstem. Upper Mystic Lake is thus a major sink of nutrients due to the long residence time that
promotes settling of particulate phosphorus and uptake by aquatic vegetation and algae.
0.100-
o.ooo
Location ID
E^3 ABR006
E^3 UPL001
MYR071
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Year
Figure ll-lll Annual Distributions of Monthly TP Concentrations at the Aberjona
River Outlet (ABR006), Mystic Lake Dam (UPL001), and Start of the Mystic River
Mainstem (MYR071)1
II.F. Lower Watershed
The lower watershed includes the Mystic River, Alewife Brook, Maiden River and numerous smaller
tributaries, which have a total drainage: area of 28 sq. miles (45 percent of the freshwater portion of
the watershed) excluding the central and upper watersheds. Land use in this area is heavily
urbanized, especially in areas around the lower section of the Mystic River. A portion of this area
contributes to the MWRA combined sewer system, which has numerous CSO outfalls along Alewife
Brook and the Mystic River. Over the past decade, many of these outfalls have been closed. Among
the remaining open outfalls, the frequency and magnitude of CSO discharges has been drastically
reduced. For the period of 2000 to 2016, CSO mitigation projects by the MWRA and the cities of
Cambridge and Somerville have reduced annual CSO discharge volumes to the freshwater portion
of the Mystic River by approximately 88 percent (59 to 7 million gallons) for the typical rainfall year.
The primary sources of nutrient loads are from the outflow of Lower Mystic Lake, stormwater
runoff, CSOs, illicit discharges, and internal loading. Due to the long history of CSO discharges to
the low gradient Alewife Brook and impounded Mystic River, the sediments in Alewife Brook and
the Mystic River are likely highly organic with elevated phosphorus levels, which could cause
significant internal loading. The legacy sediments are likely to also drive a high sediment oxygen
demand, especially in Alewife Brook where the entire water column has been observed to become
hypoxic (and occasionally anoxic) during hot, dry periods in the summer.
1 Boxplot hinges are 25th, 50th, and 75th percentiles of the concentration distribution. Upper and lower whiskers
represent largest and smallest values inside 1.5 x Interquartile Range.
22
-------
Mystic River Watershed TMDL Alternative Development Final Report
II.G. Hydrologic and Hydraulic Impacts of Amelia Earhart Dam
Discharge from Mystic River is controlled by the Amelia Earhart Dam, which is operated by DCR.
The dam includes 3 locks, which unintentionally allow saltwater intrusion from Boston Harbor. The
dam is typically operated by pumping water out of the lower basin prior to storm events to increase
available storage and prevent flooding along the shoreline. Water is also typically released each day at
low tide. The exact operational schedule and targets for this dam are unknown.
Along the entire Mystic River, as well as most of Alewife Brook and the lower section of the Maiden
River, there is a very small gradient in elevation of both the sediment and water surfaces (about 1
ft.). As a result, flow velocities are very low in these water bodies and the entire system acts as one
large impoundment.
Figure II-IV shows the relative water surface elevation at three USGS stations along the Mystic
River and Alewife Brook during an 11-day period from Sept 1, 2016 to Sept 10, 2016. The relative
water surface elevation was computed by subtracting the mean elevation at each station from the
instantaneous value. This was necessary because stage data at the Alewife station is not reported
relative to the same vertical datum as the other stations. The rapid drop in elevation on Sept 6 was
likely due to pumping at the Amelia Earhart Dam in preparation for an upcoming storm. All three
stations reflect the same change in elevation indicating that the entire Mystic River and Alewife
Brook behave as a single impoundment.
Relative Water Surface Elevation at USGS Stations in Mystic River Basin
Relative WSE(t) = WSE(t) - mean(WSE)
Sep 02 Sep 04 Sep 06 Sep 08 Sep 10
Date
Figure II-IV Relative Water Surface Elevation at USGS Gages on Alewife Brook
and Mystic River (Sept. 1-10, 2016)
II.H. Water Quality Gradient Along the Mystic River
Water quality in the upper reach ot the Mystic River in the lower watershed is relatively good quality
due to the reduced phosphorus levels in the outflow from Lower Mystic Lake. Phosphorus and
chlorophyll-a levels then gradually increase going downstream along the Mystic River towards the
Amelia Earhart Dam. In the lower reaches ot the lower Mystic River Watershed, excessive
macrophyte growth due to nutrient enrichment constitutes major water quality impairment, as
farther described below.
Figure II-Y shows chlorophyll-a and TP concentrations at five stations along the Mystic River and
one station in the lower basin of the Maiden River during 2015 and 2016. MyRWA collected the
23
-------
Mystic River Watershed TMDL Alternative Development Final Report
data for its phosphorus loading study. The high concentrations, in late summer of 2016, are most
likely caused by an herbicide application on the Mystic River to remove aquatic vegetation. The
result of this application was an instant release of phosphorus to the water column, which,
combined with an increase in light availability, likely spurred significant phytoplankton growth. A
cyanobacteria bloom was also observed during this period.
6
MYR43
^ *
125
X
100
75
3
19
50
ฃ
O
25
0
0.3
CT>
0.2
k
a.
0.1
0.0
-*>>
*A
V .'v.
v'-V \
r*- ir>
c c
N ID IT)
ฃ 5 3
Figure ll-V Chiorophyll-a and TP Concentrations Along the Mystic and Maiden
Rivers, 2015-2016
11.1. Long Term Changes in Water Quality
Over the long term, there has been a gradual decline in chlorophyll-a and TP concentrations along
the Mystic River based on the MWRA data, except for 2016, due to the herbicide treatment
mentioned above. More information on the 2016 data is presented in Section II.J below. Figure
II-VI shows the annual mean, median, and 90th percentile of chlorophyll-a and TP at the three
MWRA stations based on data from June October of each year. The decline in chlorophyll-a is
most pronounced at the lower-most station, MWRA167. In 2012-2014, 90th percentile of
chlorophyll-a was below 25 ppb, which is slightly greater than the target 20 ppb in the Lower
Charles River TMDL. As discussed below, the declines in TP and chlorophyll-a in the lower Mystic
coincide with a steady increase in growth and coverage of aquatic macrophytes during this period.
24
-------
My stic River Watershed TMDL Alternative Development Final Report
MWRA083
MWRA066
MWRA167
100
75
50-
25-
0-
0.20-
_ 0.15-
ni
E
0.10-
0.05-
0.00-
2000 2005 2010 2015 2000 2005 2010 2015 2000 2005 2010 2015
Year
Statistic
Mean
Median
Q90
Figure II-VI Annual Mean, Median, 90th Percentile of Chlorophyll-a and TP at
Three Stations on the Mystic River, 2000 - 2016 (June - Oct. only)
11. J. Impact of Aquatic Vegetation on Water Quality
Aquatic vegetation is a major nuisance on the Mystic River. Over the past 10 years, the coverage of
water chestnut and water hyacinth has grown dramatically. MyRWA currently spends significant
resources to manually remove this vegetation to allow boat passage. Treatment of the vegetation
includes mechanical and manual harvesting, as well as herbicidal treatments, which began more
recently. The vegetation is likely having a significant impact both on the hydraulics of the river by
increasing drag and forcing flow through a narrow channel, as well as on water quality through
nutrient uptake, increased water column shading, and increased particulate settling. Consequently,
the increase in aquatic vegetation in the lower Mystic is likely the primary cause for the declining
trend in chlorophyll-a and TP concentration described above.
In 2016, a major dose of herbicide was applied to the Mystic River to remove vegetation. Shortly
after tins treatment, both phosphorus and chlorophyll-a levels spiked (see Figure II-II-6). This
response indicates that aquatic vegetation is likely a major control on phytoplankton growth along
the Mystic River mainstem and in the lower basin. Presently, most of the aquatic vegetation found in
the lower Mystic can be rooted or free floating and is capable of taking up nutrients from both
bottom sediments and the water column. The combination of increased flow resistance due to the
rooted plants, which increases particulate settling rates, and the uptake of nutrients from the water
column is likely the primary cause for the declining trend in TP. At the typical TP levels observed in
the lower Mystic (<100 (Jg/L), decreases in TP in open water areas (i.e., free of excessive; aquatic
plants) will likely result in less phytoplankton growth and lower chlorophyll-a. Also, the abundance
of vegetation is likely to be a contributing cause to suppressing phytoplankton growth by reducing
25
-------
Mystic River Watershed TMDL Alternative Development Final Report
light availability through shading and causing light limitation. Thus, it is reasonable to infer that the
long-term decline in chlorophyll-a and TP levels shown above in Figure II-VI (excluding 2016 due
to the herbicide treatment) are likely caused by the increasing abundance of aquatic vegetation
removing phosphorus from the water column and limiting light penetration.
III. Review of Existing Water Quality Monitoring Data
The purposes of this section are as follows:
Describe and summarize the known, available water quality monitoring data.
Review the data for precision, accuracy, representativeness, comparability, and completeness.
According to available data, water quality surveillance of the watershed has been ongoing since at
least 1989. Consistent with the scope of this analysis, the examination will be restricted to data
collected from 2000 through 2016. Data for 2017 were not available until late in the project and
were only included for BATHTUB model validation.
III.A. Data Gaps and Recommendations for Future Sampling
Efforts
Based on the review of the available data and discussions with MyRWA and EPA Region 1, the
following data gaps have been identified which could be addressed through future monitoring
efforts.
III.A.1.Ecological/Biological Indicators of Over-Enrichment
Currently, little data is available on excess vegetative growth. Measurements are limited to
chlorophyll-a and do not include macrophyte abundance, percent cover, or broader measures of
species richness. MyRWA and EPA should consider including, at a minimum, percent of
macrophyte cover in the water body during monitoring events for baseline and phosphorus loading.
III.A.2.Streamflow
As discussed below, in the section of this memo on the available USGS flow data, there are few
locations in the watershed where it is currently feasible to make direct flow measurements. To
develop reliable estimates of nutrient loads through the watershed, measurements or reliable
estimates of flows in the watershed will be needed. This task is further complicated by multiple
impoundments. Should methods for reliable direct measurement prove infeasible, other approaches
for estimating flow based on well-established modeling techniques (e.g., using climatological, land
use, and soil type data available in GIS databases) may be explored to estimate precipitation driven
flows.
III.A.3.Sediment
Sediment attributes (e.g., total phosphorus concentrations, sediment oxygen demand) would be
useful for future modeling but was not available for the modeling portion of the project, and it is
recommended to include these attributes in future watershed surveillance efforts, if feasible.
26
-------
Mystic River Watershed TMDL Alternative Development Final Report
III.B. MyRWA and MWRA Monitoring Data
III.B.1.Data Characterization
The MyRWA provided water quality monitoring data to ERG on January 31, 2017 and provided
supplementary data on February 11, 2017. The dataset was composed of samples collected under:
MyRWA's baseline water quality monitoring.
MyRWA's phosphorus loading monitoring survey.
MWRA Boston Harbor water quality monitoring.
MWRA's combined sewer overflow event monitoring.
The baseline monitoring program has been in operation since 2000 and is used to monitor a
variety of trends in watershed water quality. Collected constituents include pathogen indicators,
nutrients, and physical-chemical water quality parameters (e.g., total suspended solids, pH, etc.).
The phosphorus loading monitoring program has been conducted since 2015 and is used to
collect information on parameters that contribute to eutrophication impairments (e.g., phosphorus)
and response parameters, which could potentially be used as indicators of nutrient over enrichment.
The MWRA water quality monitoring in general started in 1989, with the beginning of the CSO
monitoring program. The Boston Harbor monitoring in the Harbor proper began in 1993, and in
the rivers in 1995. This program was created to establish long-term water quality trends in the
Harbor and tributary watersheds for pathogen indicators, nutrients, and physical-chemical water
quality parameters.
CSO monitoring is conducted to evaluate water quality risks associated with the discharge of
untreated sewages and stormwater runoff into the watershed during CSO events. Monitoring is
conducted on an ongoing basis in Alewife Brook, Chelsea River, Little River, and the Mystic River.
Note that monitoring is not restricted to CSO discharge events. The CSO monitoring program
collects data on pathogen indicators and on physical-chemical water quality parameters.
Table III-1 summarizes the water quality parameters collected under each program included in the
dataset.
27
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table 111-1. Water Quality Parameters Included in Each Monitoring Program2
Parameter
Monitoring Program
Baseline
Boston
Harbor
cso
Phos.
Loading
Nitrogen
Total Nitrogen
yj
yj
Particulate Nitrogen
yj
Nitrogen, total dissolved
yj
Ammonia-nitrogen
yj
yj
Nitrite
yj
Inorganic nitrogen (nitrate and nitrite)
yj
yj
Nitrate
yj
Phosphorus
Total Phosphorus
sj
sj
yj
Phosphorus, Particulate Organic
yj
Dissolved Phosphorus
yj
Orthophosphate
yj
yj
Biological
Chlorophyll-a
yj
yj
Pheophytin a3
yj
Oxygen
Dissolved Oxygen
sj
sj
yj
yj
Dissolved Oxygen (Perc. Saturation)
yj
yj
yj
yj
Water Clarity
Attenuation Coefficient
yj
Secchi Disk Depth
yj
yj
Turbidity
yj
yj
yj
Other Physical/Chemical
Carbon, Total Particulate
yj
PH
yj
yj
yj
Salinity
yj
yj
yj
Specific conductance
yj
yj
yj
yj
Total suspended solids
yj
yj
Water Temperature
yj
yj
yj
yj
Pathogen Indicators
Escherichia coli
yj
yj
yj
Enterococcus
yj
yj
yj
Fecal Coliform4
yj
yj
yj
28
-------
Mystic River Watershed TMDL Alternative Development Final Report
The watershed is subdivided into nine sub-basins5. As illustrated in Figure III-I, water quality
monitoring stations: are present in each of the sub-basins and are typically located at or near
confluences between the drainage areas.
l
rt /
Bed1ฐ Woburn Sloneham
Reservation \~n-i '
Aberjona River
^ | t 'V > 'ซ/
-' ฃ
3' Saugus - ฆ ^
^ jง . J ' Middlesex Melrose
' f Winchester Reservation .
/
ฃ ฉ ~ % J
Vs ' ' ~
'-S ' ' Rumnev
type
Upper Mystic Lake
Fresh
~
ซ sT *** (So
Reservation
A Saline
Lower Mystic Lake
\ 0ฐZ)'
category
. A 2. . (9~
J
\ * + Medford 4
j Arlington # #
Estuary
ฎ I 1
Lake
ฆc Mystic River (FreS^a|den River
~ Alewife Brook
TuftS ^ VJS)
ซ University % \ # ^
Revere i
A ฐ ^
River/Stream
* . ฉ \
Somerville MysticJRiver (Salt)
Chelsea
fe ' "--or./ A
~
ป ^ O ^'
% 0 Cambridge
a. (f'
H
Boston Logan _^JnthrฐP ,
International QiS)
Figure III-!, Mystic River Watershed Drainage Basins and Monitoring Locations
Most water bodies in the watershed have been sampled all years from 2000 2016 (Table III-2). Six
water bodies, mostly located in the upper portions of the watershed, have been sampled in 2015 and
2016 as part of the phosphorus loading survey but have not been monitored otherwise. Two water
bodies in the lower portion of the watershed, Mill Creek and the Belle Island Inlet, have been
monitored for approximately the past decade.
2 Particulate nitrogen, particulate phosphorus, particulate carbon, total dissolved phosphorus, and total dissolved
nitrogen are only tested at brackish/saltwater locations downstream of the Amelia Earhart dam.
3 Phaeophytin is not currently tested using our WQ sondes: though there are older data in the dataset.
4 Fecal colifonn is not currently tested in any of the freshwater locations for either MWRA monitoring program, but
like phaeophytin there will be older data in the dataset. It is currently tested at locations downstream of the dam.
' Note that this delineation is provided for general information purposes and to provide the reader with an
approximate sense of patterns of drainage within the watershed and is not intended to provide a definitive
delineation of drainage within the basin.
29
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table 111-2. Temporal Range and Duration of Monitoring within Water Bodies
Water Body
Sample Range
No. Years Monitored
Min
Year
Max
Year
All
Programs
Program Subtotals
Baseline
Phos.
Loading
Boston
Harbor
cso
Aberjona River
2000
2016
17
17
2
-
-
Horn Pond
2015
2015
1
-
1
-
-
Horn Pond Brook
2015
2015
1
-
1
-
-
Winter Pond
2015
2015
1
-
1
-
-
Wedge Pond
2015
2016
2
-
2
-
-
Upper Mystic Lake
2000
2016
17
17
2
-
-
Lower Mystic Lake
2015
2015
1
-
1
-
-
Mill Brook
2000
2016
17
17
1
-
-
Spy Pond
2015
2015
1
-
1
-
-
Winns Brook
2000
2016
17
17
-
-
-
Little River
2000
2016
17
-
-
-
17
Alewife Brook
2000
2016
17
17
-
-
17
Mystic River (Fresh)
2000
2016
17
17
2
-
17
Meetinghouse
Brook
2000
2016
17
17
1
-
-
Maiden River
2000
2016
17
17
2
15
-
Mystic River (Salt)
2000
2016
17
9
-
17
17
Mill Creek
2008
2016
9
9
-
-
-
Chelsea River
2000
2016
17
9
-
-
15
Belle Isle Inlet
2009
2016
8
8
-
-
-
Restricting the examination to parameters that are likely to be of the greatest significance when
assessing conditions related to eutrophication, it is apparent that total phosphorus and dissolved
oxygen data are temporally and spatially well represented, and Secchi depth data are available for
most of the water bodies (see Figure III-III). In particular, multiple monitoring locations on the
main stem of the Mystic River downstream of Lower Mystic Lake possess large historical datasets
for most parameters of potential interest for developing in-stream water quality targets (Figure
III-IV).
30
-------
Mystic River Watershed TMDL Alternative Development Final Report
Aberjona River Alewife Brook Belle Isle Inlet Chelsea River Horn Pond
I
I
ll
1 1
Lower Mystic Lake
1
|
ฆ
Horn Pond Brook
ฆ
Little River
ฆ
1
1
Maiden River
ฆ
Meetinghouse Brook
_ I _ Jill I
** u
o
S? Mill Brook Mill Creek Mystic River (Fresh) Mystic River (Salt) Spy Pond
fl I. iHillIIi
A*. ^ ^ is
Upper Mystic Lake Wedge Pond Winns Brook Winter Pond ^ ^ (j
J l_ . I J_
d f / ^ cf ? ,oฐ cF f J~ ,
-------
Mystic River Watershed TMDL Alternative Development Final Report
MWRA056 MWRAQ57 MWRA059 MWRA066
II II II u
o
MWRA067 MWRAQ83 MWRA177 MYR071
II Mi. II
i i i i i i i i i i i i
E MYR33 MYR43 f? cf ^ ^
15- O ฃ O ฃ
ฆS -X
10-
5-
o-
Refer to Appendix B for additional detail on water quality parameter codes.
Figure Ill-Ill Number of Years in Which Observations Were Made for
Eutrophication Related Parameters in The Mystic River
As shown in Figure III-III, data for water quality parameters relevant to nutrient impairment have
been collected throughout the watershedparticularly in the main waterwaysand at a relatively
high frequency. Sampling frequencies associated with the Boston Harbor monitoring and
phosphorus loading programs occurred bi-weekly on a seasonal basis, with much of the monitoring
for the baseline monitoring program occurring monthly. Refer to Appendix A for additional detail
on sampling frequency broken down by water body, monitoring location, and monitoring program.
32
-------
My stic River Watershed TMDL Alternative Development Final Report
Number of Sample Days
Winter Pond -
Winns Brook-
Wedge Pond-
Upper Mystic Lake-
Spy Pond-
Mystic River (Salt)-
Mystic River (Fresh)-
Mill Creek-
Mill Brook-
Meetinghouse Brook-
Maiden River-
Lower Mystic Lake -
Little River-
Horn Pond Brook-
Horn Pond-
Chelsea River-
Belle Isle Inlet-
Alewife Brook-
Aberjona River-
No. of Days
3000
2000
1000
i of Days
Refer to Appendix B for additional detail on water quality parameter codes.
Figure lil-IV. Temporal Coverage (both number of days and Percentage of Total
Available Days from 2000 - 2016) of Water Quality Parameters by Water Body
In terms of seasonal coverage, the number of observations was relatively evenly split between
summer (June - October) and winter (November May).: The fraction of summer observations
ranged from approximately 35 percent to 70 percent within the watershed and showed typical values
of 53 +/- 19 percent (average +/- standard deviation) for all water bodies with at least 10
observations for a given parameter.
While temporal and geographic coverage is robust within the dataset, specific
season/parameter/water body combinations do exist where relatively limited data coverage is
available. Table III-3 identifies specific instances where data limitations exist for nutrient response
variables at a seasonal level.
33
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table 111-3 Number of Seasons Where Few (N < 3) Samples were Collected for a
Waterbody-Parameter Combination
Parameter
No. of Seasons with n<3*
Jun-Oct
Nov-May
Alewife Brook
Secchi Depth
3
9
Turbidity
-
5
Little River
Dissolved Oxygen
7
7
Dissolved Oxygen (% Saturation)
7
7
PH
7
5
Specific Conductivity
8
7
Water Temperature
7
7
Turbidity
2
3
Maiden River
Dissolved Oxygen
1
19
Dissolved Oxygen (% Saturation)
-
10
PH
-
9
Secchi Depth
-
12
Specific Conductivity
-
13
Water Temperature
-
12
Turbidity
2
20
Mystic River (Fresh)
Secchi Depth
-
2
Total Dissolved Nitrogen
1
4
Total Dissolved Phosphorus
1
4
Turbidity
-
1
Mystic River (Salt)
Attenuation Coefficient
6
1
Chlorophyll-a
2
-
Ammonia-Nitrogen
2
-
Inorganic Nitrogen
2
-
Phaeophytin
2
-
Orthophosphate
2
-
Total Dissolved Nitrogen
2
-
Total Dissolved Phosphorus
1
-
Total Particulate Carbon
1
-
Total Particulate Nitrogen
2
-
Total Particulate Phosphorus
3
-
Total Suspended Solids
3
-
Turbidity
2
3
Where a water body-parameter combination is absent, or a is listed, then n > 3 observations.
Appendix C includes a series of boxplots that summarize the observed characteristics of the dataset
(i.e., minimum, 25th percentile, median, 75th percentile, and maximum). Additionally, a small number
of unusual outlier values have been noted on the figures. Appendix D includes a table of the outlier
observations with any documented data quality issues noted.
34
-------
Mystic River Watershed TMDL Alternative Development Final Report
Dataset attributes that were assessed included the frequency of observations flagged for quality
concerns and the frequency of censored observations (i.e., measurement results below the analytical
method detection limit). Data quality documentation is currently available for only two of the
monitoring programsBaseline and Phosphorus Loading. As shown in Table III-4, approximately
98 percent of the observations from these two programs were free of documented data quality
issues.
Table 111-4. Summary of Documented Data Quality Issues
Flag
Flag Description
Program Subtotals
Grand Total
Baseline
Phos. Loading
No.
Percent of
Total
B
Analyte detected in the blank
-
6
6
0.023
E
Instrument error
-
7
7
0.027
F
Field replicate quality control
failure
308
-
308
1.186
H
Holding time issues
-
15
15
0.058
J
Detected above method
detection limit but below
quantitation limitresult is an
estimate
-
11
11
0.042
K
pH calibration error
-
101
101
0.389
L
Lab duplicate relative percent
difference exceeded
168
2
170
0.655
0
Other unspecified issues
2
-
2
0.008
No
Flag
No documented data quality
issue
22,652
2,697
25,349
97.613
The majority of the non-detect numbers were total suspended solids data and nitrogen data. In some
cases (e.g., total suspended solids and nitrites), non-detect values make up a substantial portion of
the observations. In these instances, accurate estimates of water quality parameters may be difficult
to develop with a high degree of confidence. See Table III-5 for a summary of non-detect data.
Table 111-5 Summary of Non-Detect Data.1
Water body
No of Samples
Method Detection
Limit
Unit
s
Detecte
d
Non-
Detect
Total Suspended Solids
Aberjona River
275
283
2 - 11.4
mg/L
Alewife Brook
148
41
1-7.7
mg/L
Chelsea River
77
4
5-5
mg/L
Maiden River
119
63
1-20
mg/L
Meetinghouse
Brook
70
120
o
1
1
1
mg/L
Mill Brook
106
81
2-20
mg/L
Mystic River (Fresh)
829
116
0.24-10.9
mg/L
Mystic River (Salt)
1048
6
0.24 -10
mg/L
Upper Mystic Lake
50
133
1-8.3
mg/L
35
-------
Mystic River Watershed TMDL Alternative Development Final Report
Winns Brook
92
96
1-12
mg/L
Total Particulate Nitrogen
Mystic River (Salt)
439
1
0.0027
mg/L
Total Particulate Carbon
Mystic River (Salt)
441
1
0.016
mg/L
Total Phosphorus
Aberjona River
582
13
0.05 - 0.05
mg/L
Chelsea River
92
3
0.01-0.025
mg/L
Maiden River
253
2
0.05 - 0.05
mg/L
Meetinghouse
Brook
199
3
0.05 - 0.05
mg/L
Mill Brook
221
3
0.05 - 0.05
mg/L
Mill Creek
92
1
0.01
mg/L
Mystic River (Fresh)
1041
6
0.05-0.11
mg/L
Mystic River (Salt)
175
2
0.01-0.012
mg/L
Spy Pond
16
1
0.005
mg/L
Upper Mystic Lake
233
6
0.005 - 0.05
mg/L
Total Dissolved Phosphorus
Mystic River (Salt)
455
1
0.0034
mg/L
Orthophosphate
Aberjona River
0
7
0.005 - 0.005
mg/L
Alewife Brook
13
3
0.005 - 0.005
mg/L
Horn Pond Brook
1
5
0.005 - 0.005
mg/L
Maiden River
0
14
0.005 - 0.005
mg/L
Mill Brook
3
4
0.005 - 0.005
mg/L
Mystic River (Fresh)
744
35
0.00031 - 0.005
mg/L
Mystic River (Salt)
903
4
0.00031 - 0.030
mg/L
Spy Pond
0
7
0.005 - 0.005
mg/L
Upper Mystic Lake
0
10
0.005 - 0.005
mg/L
Wedge Pond
0
7
0.005 - 0.005
mg/L
Nitrate
Maiden River
4
1
0.1
mg/L
Mill Brook
5
1
0.1
mg/L
Mystic River (Salt)
7
1
0.1
mg/L
Winns Brook
5
1
0.1
mg/L
Inorganic Nitrogen (Nitrite plus Nitrate)
Belle Isle Inlet
43
38
0.1-0.5
mg/L
Chelsea River
60
27
0.1-0.5
mg/L
Mystic River (Salt)
1023
52
0.00028-0.5
mg/L
Nitrite
Aberjona River
6
12
0.1-0.1
mg/L
Alewife Brook
0
6
0.1-0.1
mg/L
Chelsea River
0
6
0.05-0.1
mg/L
Maiden River
0
5
0.1-0.1
mg/L
36
-------
Mystic River Watershed TMDL Alternative Development Final Report
Meetinghouse
Brook
0
6
0.1-0.1
mg/L
Mill Brook
0
6
0.1-0.1
mg/L
Mill Creek
1
4
0.1-0.1
mg/L
Mystic River (Fresh)
0
6
0.1-0.1
mg/L
Mystic River (Salt)
0
6
0.05-0.1
mg/L
Upper Mystic Lake
0
6
0.1-0.1
mg/L
Winns Brook
0
6
0.1-0.1
mg/L
Ammonia
Mystic River (Salt)
865
42
0.00039 - 0.010
mg/L
Upper Mystic Lake
56
1
0.00039
mg/L
Winns Brook
56
1
0.00039
mg/L
1. Locations with zero non-detect observations have been excluded.
III.B.2.Data Review
The investigators reviewed the available data for accuracy/precision, representativeness,
comparability, and completeness and concluded that the data are largely acceptable for use in
assessing appropriate nutrient endpoint targets in the watershed.
Precision measures the reproducibility of repeated measurements and accuracy measures the
"correctness" of an estimate. Overall, the dataset exhibited satisfactory levels of accuracy and
precision where quality was documented. A relatively small fraction of the dataset included
documented data quality issues (see Table III-4), though this conclusion may be revised upon receipt
of quality control data for the CSO and Boston Harbor monitoring programs. Where data quality
concerns have been documented in the Baseline and Phosphorus Loading data, excluding flagged
data is recommended, with the possible exception of the J-flagged data (i.e., observations where the
parameter was positively detected in the water, but at a level where quantitation is less precise than is
usual).
The data are largely representative, both spatially and temporally, of the watershed. Most water
bodies in the watershed have been monitored for the entire duration of interest (2000 2016). Two
water bodies have been monitored for a majority of the period of interest, and the remainder has
been intensively sampled in 2015 2016. In addition, the dataset also captures seasonal variation in
parameters. Table III-3 documents specific water bodies where the available data are limited. In
addition, Table III-5 documents limited instancesparticularly with TSS and nitrites-where a large
fraction of the data are censored and which may influence the ability to draw accurate inferences
regarding water quality parameters (e.g., averages, standard deviations) for these constituents.
However, these limitations are unlikely to present difficulties when developing nutrient modeling
endpoints, as other constituents are present in the dataset and are likely to be more informative for
purposes of developing protective in-stream water quality targets.
Comparability is an expression of the confidence with which one data set can be compared to
another. Based on the available information, the data from the different programs, which compose
the dataset, appear comparable.
37
-------
Mystic River Watershed TMDL Alternative Development Final Report
The available data appears sufficiently complete for use in developing water quality targets. The vast
majority of the data are valid and composed of point estimates (i.e., detect values) that can be used
to develop reference conditions or stressor-response relationships for use in establishing modeling
endpoints and protective instream water quality targets.
III.C. USGS Flow Data
Valid daily flow data are available from two USGS monitoring stations for the Aberjona River and
Alewife Brook (Table III-6). These stations collectively drain 8.96 square miles of the watershed,
which accounts for approximately 12 percent of the watershed's land area. The two stations possess
historical records, which extend prior to 2000 and to 2007, respectively.
Table 111-6. Summary of USGS Monitoring Stations
Gauge ID
No.
Description
Location
(Lat; Lon)
Discharge Data
Availability
HUC
Net Drainage
Area (sq. mi.)
01102500
Aberjona River at
Winchester, MA
42ฐ26'50.5";
71ฐ08'18.9"
<2000 - 2/2017
01090001
24.7
01103010
Mystic River at
Arlington, MA
42=25' 14";
71ฐ08'33"
6/2016 - 2/20171
01090001
Undetermined
01103025
Alewife Brook Near
Arlington, MA
42ฐ24'25";
71ฐ08'04"
10/2007 - 2/2017
01090001
8.36
01103038
Maiden River at
Maiden, MA
42ฐ25'04";
71ฐ04'23"
7/2016 - 2/20171,2
01090001
Undetermined
01103040
Mystic River at RT16
at Medford, MA
42ฐ24'20.6";
71ฐ05'45.6"
10/2015 - 2/20171
01090001
Undetermined
01103050
Mystic River at
Amelia Earhart Dam
42ฐ23'44";
71ฐ04'32"
None3
01090001
62.7
1. USGS has indicated that all or some of the discharge data collected at these stations are unreliable.
2. Stream temperature data are also available.
3. Gauge height data are available.
Three additional monitoring stations are equipped to estimate discharge volumes, however, USGS
has indicated their measurements may be unreliable or inaccurate due to a combination of the
shallowness of the river and low stream velocities. Currently, it is unclear if all discharge data from
these stations are unusable or if some portion of the streamflow data might be utilized.
III.D. GIS Datasets
Geospatial data are available from the sources listed in Table III-7, below. The data layers that were
considered for calculating phosphorus loads in the watershed include: impervious cover, land use,
and hydrologic soil types. In addition, phosphorous loads may be estimated for the Mystic River
using similar export rates developed for the Massachusetts Municipal Separate Storm Sewer System
(MS4) Permit and which are available from EPA Region 1.
38
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table 111-7. GIS Data Sources Available for the Mystic River Watershed
Data layer
Source
Description
Land use/land
cover
Massachusetts Office of
Geographic Information
(MassGIS) or Multi-
Resolution Land
Characteristics Consortium
(MLCD)
Land use/land cover (LULC) layers contain information on the
physical land type, e.g., forest, wetlands, of an area as well as
information on how people are using the land, e.g., row crops,
low-intensity development. Land cover can be determined in
the field or by interpreting remotely sensed imagery (i.e.,
satellite imagery, aerial photos). Two sources for LULC datasets
are MassGIS's Land Use (2005) or NLCD 2011.
Impervious
surface
MassGIS
Impervious surfaces are surfaces that do not allow water to
penetrate, forcing water to runoff. As water runs off, it can
carry pollution from waterways and other surfaces into water
bodies. Common impervious surfaces include roads, parking
lots, rooftops, driveways and sidewalks, and compacted soils.
NRCS HUC
Basins
(8,10,12) (Sub-
watersheds)
MassGIS
A hydrologic unit code (HUC) is the number assigned to a
hydrologic unit, which is a drainage area that nests in a multi-
level drainage system. Its boundaries are defined by
hydrographic and topographic criteria that delineate an area of
land upstream from a specific point on a river, stream or similar
surface water. HUCs are identifiers as assigned to basin
polygons by the USGS.
Soils
MassGIS (NRCS SSURGO) or
NRCS Web Soil Survey
Information on underlying soils can help determine how much
water can be absorbed or how much will runoff. There are
specific hydrologic soil groups identified for areas that are
based on estimates of runoff potential. Soils are assigned to
one of four groups according to the rate of water infiltration
when the soils are not protected by vegetation, are thoroughly
wet, and receive precipitation from long-duration storms. The
soils in the United States are assigned to four groups (A, B, C,
and D) and three dual classes (A/D, B/D, and C/D).
Sewer-shed
Cambridge and Somerville
Drainage system for the local storm sewer (separate and/or
combined) discharging into the Mystic River Watershed. Data
available for Cambridge and Somerville only.
IV. Review of Modeling Endpoint Approaches
Water quality segments within the Mystic River Watershed are currently impaired for eutrophication-
related parametershigh phosphorus, nitrogen, and chlorophyll-a levels have been documented
throughout the watershed. Over enrichment has resulted in algae blooms and periods of excessive,
nuisance vegetation growth. The purpose of selecting water quality targets is to establish a set of
modeling and water quality endpoints (i.e., target water quality conditions) that meet Massachusetts'
water quality standards and are protective of the designated uses established in the WQS (314 CMR
4.00). Attainment of appropriate targets could eventually result in acceptable levels of algal growth
and the cessation of use impairments caused by excessive macrophyte growth.
This section reviews the following approaches for establishing targets in the watershed:
Use of existing regional/local targets.
Use of reference water body conditions.
39
-------
Mystic River Watershed TMDL Alternative Development Final Report
Development of targets based on stressor-response relationships.
Development of targets based on mechanistic models.
IV.A. Water Quality Standards Applicable to the Mystic River
Watershed
Once a tidal river, the Lower Watershed currently functions as a large impoundment due to a dam
located at the basin outlet. Another significant impoundment on the main stem of the river exists at
the outlet of the Upper Mystic Lake (Figure IV-I and Figure IV-II).
In 314 CMR 4.00, Massachusetts establishes two eutrophication related standards applicable to the
watershed: a narrative nutrient standard and a numeric dissolved oxygen standard.
Nutrients. The narrative standard prohibits discharges containing nutrients in
concentrations that would " cause or contribute to cultural eutrophication, including excessive growth of
aquatic plants or algae, and othenvise render water unsuitable for designated uses." [314 CMR 4.05(5)]
Dissolved Oxygen. For Class B waters, the concentration of dissolved oxygen shall be
greater than or equal to 5 mg/L at all times. [314 CMR 4.05(3) (a)(1)]
IV.B. Existing Regional/Local Targets
Several examples exist that establish protective water quality targets for either (1) local or regional
water bodies, or (2) specific water body types. These include the Total Maximum Daily Toadfor
Nutrients in the Tower Charles River Basin, Massachusetts (Charles River TMDL), Massachusetts' numeric
standards and impairment assessment criteria (Massachusetts Department of Environmental
Protection [DEP], 2016), and the criteria included in EPA's Quality Criteria for Water (1986);
informally known as the "Gold Book." These targets are summarized in Table IV-1.
40
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table IV-1. Existing Water Quality Targets
Parameter
Numeric Target
Target Duration/
Frequency
Source
Total
Phosphorus
<100 ng/L
(Free flowing rivers)
Instantaneous
Gold Book (1986)
<50 ng/L
(Entering
lakes/impoundments)
Instantaneous
<25 ng/L
(Exiting
lakes/impoundments)
Instantaneous
Dissolved
Oxygen
>5 mg/L
Instantaneous
MA Surface Water Quality Standards
[314 CMR 4.05(3)(a)(l)]
Dissolved
Oxygen,
Saturation
<125%
Instantaneous
Upper/Middle Charles TMDL (MA DEP,
2016)
Chlorophyll-a
<10 ng/L
Seasonal Average
Lower Charles TMDL (MA DEP & EPA,
2007) and Upper/Middle Charles
TMDL (MA DEP, 2016)
<18.9 ng/L
90th Percentile
IV. C. Reference Water Body Conditions
The reference water body condition method utilizes observations of water quality conditions (e.g.,
total phosphorus and chlorophyll-a concentrations) in water bodies with limited anthropogenic
impactsor, at least, limited eutrophicationto establish "natural" or background nutrient
conditions in regional waters. In principle, reference condition targets should approximate the best
possible attainable water quality in the absence of human activity or if human impacts are entirely
controlled (Dodds and Oakes, 2004).
EPA suggests several methods for establishing reference nutrient conditions for a water body (Buck,
et al., 2000). Of those applicable to the Mystic River Watershed, the first requires the identification
of reference water bodies comparable to the Mystic River water bodies but which display limited or
no human influences on water quality. The 75th percentile water quality condition of the reference
water bodies is estimated, and this percentile is applied as a target in the water body of interest. The
second approach calculates the 25th percentile concentration of the general population of water
bodiesincluding water bodies with clear human impactsto develop a target. Figure IV-I
conceptually illustrates these two approaches.
41
-------
My stic River Watershed TMDL Alternative Development Final Report
1 1 1 1 1 1 1 t 1 1 1 1 1 ! 1 1 1"
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Total phosphorus (jig/L)
Figure IV-I. Reference Condition Approach for Establishing Numeric Water
Quality Targets (EPA, 1998)
EPA has utilized this reference: approach to develop eco-region-based nutrient criteria for
application through the United States. In eco-region XIV sub-region 59 (EPA, 1998), where the
Mystic River is located? EPA was unable to identify suitable reference water bodies and,
consequently, based their eco-regional criteria on the 25th percentile: of the general population of
water bodies, To confirm that no new, appropriate reference water bodies have been identified
following the publication of the eco-regional criteria, the project team reviewed recent survey efforts
undertaken as part of EPA's National Aquatic Resources Survey. The project team was unable to
identify any water bodies that could serve as appropriate reference water bodies for the Mystic River.
Table IV-2 summarizes EPA's eco-region XIV sub-region 59 criteria and, for comparative purposes,
the 25th percentile parameter values for Upper Mystic Lake, the Upper Mystic River, and for the
entire watershed. Monitoring data for the water bodies of interest display 25th percentile
observational values ranging from 20 [Jg/L - 34 (Jg/L for total phosphors and 1.9 (Jg/L - 6.6 (Jg/L
for chlorophyll a.
42
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table IV-2. Eco-regional Criteria and Observed 25th Percentile (2000 - 2016)
Mystic River Watershed Values
Type/Location
Total
Phosphorus
(Hg/L)
Total Nitrogen
(mg/L)
Chlorophyll-a1
(Hg/L)
Secchi Depth
(meters)
Turbidity
(NTU)
Eco-region XIV, Sub-region 59
Rivers &
Streams
23.75
0.59
0.442
-
1.683
Mystic River Watershed 25th Percentile Observations (2000 - 2016)
Upper Mystic
Lake4
20
-
6.6
-
1.7
Upper Lobe of
Upper Mystic
Lake5
25
-
6.0
-
5.1
Mystic River
(Lower Basin)4
30
1.0
1.9
0.84
3.8
Total Mystic
River
Watershed4
34
1.0
2.0
0.70
3.9
1. Fluorometric method
2. Aggregated by subregion
3. Average of reported and calculated values
4. Aggregate of all monitoring station located in the water body
5. Measured at monitoring station UPLUPL
IV.D. Stressor-Response Relationships
Water quality targets developed based on stressor-response relationships utilize empirical
relationships between nutrients and response variables (e.g., chlorophyll-a, excessive macrophyte
growth) to estimate protective numeric, water body-specific targets. This technique includes four
steps: (1) develop a conceptual model, (2) assemble and explore water quality data, (3) develop
statistical relationships between variables, and (4) derive protective targets based on those
relationships.
A benefit of the stressor-response approach is that nutrient targets are based on functional
relationships between nutrients and attainment of designated uses. This reduces risk associated with
developing excessively stringent targets, as can happen with reference condition approaches.
However, there is substantial risk that, after analysis, no reliable relationship will be discernable. This
can occur when the water body of interest is extremely impairedresulting in a saturated response
signalor when multiple confounding effects, which cannot be sufficiently controlled for,
contribute to the impairment.
For purposes of demonstration, the project team performed a preliminary stressor-response analysis
using Mystic River Watershed data described in Section III of the report. As a preliminary analysis,
the direct relationship was assessed between variables, while not controlling for covariate effects and
other confounding phenomenon (e.g., flow, phosphorus uptake by macrophytes, etc.). In addition,
43
-------
Mystic River Watershed TMDL Alternative Development Final Report
the project team restricted the analysis to Upper Mystic Lake and the main stem of the Mystic River
in the Lower Watershed, and to total phosphorus concentration values as a stressor variable.
Table IV-3 presents the results of the relationship between total phosphorus and chlorophyll-a
concentrations in the upper lobe of Upper Mystic Lake, all sites within Upper Mystic Lake, and in
the Mystic River. Simple linear regressions were performed on paired concentration values (i.e., total
phosphorus and chlorophyll-a on a given day and monitoring location). Based on the resulting
relationship, a total predicted phosphorus concentration was calculated to produce a 10 (Jg/L
chlorophyll-a concentration (i.e., the seasonal average target for the Lower and Upper/Middle
Charles River TMDLs). Data was restricted to the period April October when evidence of algae
growth was greatest. Predicted total phosphorus concentrations varied from 15 (Jg/L to 29 (Jg/L.
Regressions for the Mystic River and the total Upper Mystic Lake were not significant; however,
regression parameters for the upper lobe of the Upper Mystic Lake were significant (p < 0.05;
Figure IV-IV).
Table IV-3. Simple Linear Regression on Paired TP & Chlorophyll-a Monitoring
Results Collected April - October
Water body
Linear Regression Parameters
Total Phosphorus1
(Hg/L)
Slope
Intercept
r-squared
Upper Lobe of Upper Mystic Lake2
1.433
-27.89
0.43
26
Upper Mystic Lake
0.5872
1.092
0.37
15
Mystic River
0.1332
6.164
0.12
29
1. Total phosphorus concentration implied by linear regression. Corresponds to a chlorophyll-a concentration of 10
Hg/L.
2. Monitoring station UPLUPL.
_J
7 40-
>ป
1-
CL
O
O
J=
O
20-
o
o
O
*
Q.
<
o-
20
30 40
April-October Total Phosphorus (|ig/L)
50
Figure IV-II. April - October Total Phosphorus vs. Chlorophyll-a Concentrations in
the Upper Lobe of the Upper Mystic Lake (Monitoring Station UPLUPL)
Regressions based on seasonal averages in the Mystic River resulted in implied seasonal average total
phosphorus concentrations of approximately 33 and 41 (Jg/L to meet Charles River TMDL targets.
However, like the regressions in Table IV-3 for the Mystic River, these relationships were not
44
-------
Mystic River Watershed TMDL Alternative Development Final Report
statistically significant. Table IV-4, Figure IV-III and Figure IV-IV show regressions of seasonal
average total phosphorus concentrations and chlorophyll-a as seasonal average and a seasonal 90th
percentile value, respectively (shaded area indicates 95% confidence interval). Data was aggregated
based on April October measurements in each year. Upper Mystic Lake was not included in this
analysis since only 2 years of data were available.
Table IV-4. Simple Linear Regression on Average TP and Average or 90th
Percentile Chlorophyll-a1
Water body
Linear Regression Parameters
Chlorophyll-a
Target (ng/L)
Total
Phosphorus2
(Hg/L
Slope
Intercept
r-squared
Mystic River:
Seasonal Average Chlorophyll-a
0.16929
4.43453
0.14
10
33
Mystic River:
Seasonal 90th Percentile
Chlorophyll-a
0.3679
3.7722
0.19
18.9
41
1. Averages and 90th percentile values aggregated by year for the period April - October.
2. Concentration predicted by regression equation for Charles River TMDL target.
45
-------
Mystic River Watershed TMDL Alternative Development Final Report
20-
30 40 50 60 70
April-October Average Total Phosphorus (ng/L)
Figure IV-III. April - October Average Total Phosphorus vs. Average Chlorophyll-a
Concentrations in the Mystic River
30 40 50 60 70
April-October Average Total Phosphorus (ng/L)
Figure IV-IV. April - October Average Total Phosphorus vs. 90th Percentile
Chlorophyll-a Concentrations in the Mystic River
IV.E. Mechanistic Models
Mechanistic models use systems of equations to represent ecological and hydrodynamic processes
within a water body. These can be used to predict changes in eutrophication-related processes in
response to changes in the level of water body enrichment. Like stressor-response models,
mechanistic water quality models can be used to develop nutrient targets based on functional
relationships and using site-specific empirical data. While stressor-response relationships are treated
like statistical "black box" processes, mechanistic models are largely deterministic models of system
behavior (U.S. EPA, 2001).
Mechanistic models have substantial input data requirements and would require a higher level of
effort relative to the options discussed above. Models of this type are not in widespread use for
46
-------
Mystic River Watershed TMDL Alternative Development Final Report
developing modeling endpoints or protective nutrient targets, though they have been successfully
implemented in the past (Paul et al., 2011).
IV.F. Recommendations
The project team recommends the use of existing water quality targets designed for nearby water
bodies. While the Mystic River Watershed possesses some unique hydrologic features, the Charles
River shares many similar features (climate, hydrologic features) and the targets developed in its
TMDL are protective of the applicable Massachusetts water quality standards. Thus, the project
team decided to use the lOug/L target for chl-a from the Charles River TMDL and apply it to the
load reduction analysis for the Mystic River. During modeling and calibration, it was determined that
the lOug/L target was not easily applied to stormwater management BMPs. Instead, the team
decided to use a TP target of 30ug/L based on the Charles River TMDL, which references a TP
target of 30ug/L to achieve water quality conditions that correlate to lOug/L for chl-a.
A mechanistic model is not recommended due to high resource and data requirements. Stressor-
response relationships require the selection of a response variable target (e.g., macrophyte cover and
chlorophyll-a) and, consistent with the results of the preliminary analysis documented in this
memorandum, may not produce results of sufficiently robust statistical significance on which to base
the alternative TMDL analysis. A reference condition approachparticularly in the absence of
reliable reference water bodiesmay prove to be unreliable, as the resultant target does not reflect a
functional relationship between nutrients and over-enrichment conditions.
47
-------
Mystic River Watershed TMDL Alternative Development Final Report
V. Watershed Phosphorus Loading Estimates
Average annual phosphorus loading was estimated for the watershed area sub-basins tributary to the
freshwater portion of the Mystic River. These estimates will be used to estimate loading to critical
receiving water bodies that will be modeled using the Lake Loading Response Model (LLRM,
Wagner, 2009) and BATHTUB water quality models (please note, LLRM analysis is ongoing and
results are not presented in this report). Output from the receiving water model will then be used in
subsequent analyses to identify the reductions in phosphorus loads necessary to bring the watershed
into compliance with selected water quality targets (see Table IV-1 for TP and chl-a targets). In
addition, observed loading data corresponding to several USGS gage stations in the watershed have
been compiled for use in calibration of the land load estimates.
V.A. Stormwater Loads
Stormwater loads to the receiving water system are carried across the land surface by runoff during
most precipitation events. Regional pollutant loading export rates (PLERs) are commonly used as a
way to estimate annual phosphorus loads from various Hydrologic Response Units (HRUs), where
each HRU is comprised of a unique combination of land use, cover type (impervious or pervious),
and soil type. PLERs define the TP mass per unit area per time (i.e., lbs./acre/year) that is exported
from each type of HRU. HRUs are a common modeling method to categorize areas of land that
function similarly in terms of their hydrologic fluxes and pollutant loads.
To develop stormwater loads, EPA Region l's Opti-Tool modeling package was utilized. Opti-Tool
incorporates model generated time-series of hourly stormwater runoff volumes and nutrient runoff
quality that reflects pollutant build-up/wash-off processes and that has been calibrated to
stormwater quality and climatic data representative of the New England region. The Opti-Tool
package includes companion HRU Storm Water Management Models (SWMM) that were used to
dynamically simulate rainfall-runoff events for land-use based impervious and pervious HRU
categories that reflect land cover characteristics in the Mystic River Watershed. Using the SWMM
HRU models, Opti-Tool can be used to simulate runoff volumes and pollutant loads (e.g., annual
loads) for defined sub-basins within the watershed for any period of interest.
Opti-Tool includes SWMM HRU models for each HRU category that collectively are used to
represent key watershed characteristics in the Mystic watershed related to surface runoff. Individual
HRU models applied to the Mystic include land use specific for impervious cover (e.g., commercial
impervious) and land use specific for pervious cover and hydrological soil group (HSG) A, B, C, and
D (e.g., high density residential pervious HSG B). Hourly precipitation data representative of the
Mystic watershed and daily maximum and minimum temperatures are used as inputs to conduct
continuous HRU model simulations to generate annualized stormwater runoff volumes and loads.
The HRU model outputs produce land-use category specific PLERs for impervious cover and
pervious cover (according to HSG) and rainfall-runoff total (in/year). The unit PLER and flow can
be generated for any period of interest, including on an annual basis for each individual year, or as
overall average annual values for the period of record analyzed.
Using the resultant unit export rates, the annual stormwater load and/or flow for a given HRU
category within the watershed area of interest is generated by multiplying the total area represented
by the HRU type (e.g., commercial impervious) by the corresponding HRU specific PLER.
OptiTool loads and flows for each land use were used as input to the land loading models. That
48
-------
Mystic River Watershed TMDL Alternative Development Final Report
loading model accumulated loads and flows and then attenuated the loads to provide values at the
outlet of each sub-basin. The basin values were used as the input to the BATHTUB model.
Table V-l presents the climatological datasets used as inputs to the Opti-Tool model to generate
annual total stormwater phosphorus loads and flows.
Table V-1. Opti-Tool-SWMM HRU Model Climatological Input Datasets
Type
Description
Source
Precipitation
Hourly time series of precipitation values at
Boston Gen E Logan International Airport
(WSAF-WBAN ID 725090 14739) from 1992 -
2016
https://gis.ncdc.noaa.gov/maps/ncei/cdo
/hourly
Temperature
Daily time series of maximum and minimum
temperature values at Boston Gen E Logan
International Airport (WSAF-WBAN ID 725090
14739) from 1992 - 2016
https://gis.ncdc.noaa.gov/maps/ncei/cdo
/hourly
The delineation of sub-basins for quantifying flows and loads through HRU accounting within the
freshwater portion of the Mystic River Watershed was initially based on an existing sub-basin
delineation available from the Massachusetts Office of Geographic Information (MassGIS).
Modifications and refinements were made to further sub-divide several basins to better reflect
watershed routing processes and more closely align with critical water body assessment locations.
Some of these modifications were made using a digital elevation model and delineation for quality
control. Further modifications to several impaired lake basins were made based on consultation with
MassDEP staff and a review of previously developed procedurally generated delineations for the
lake basins. Ultimately, all sub-basin delineations were visually compared to FEMA derived flood
plain delineations as an added quality control measure.
Land-use categories within each sub-basin were delineated based on the intersection of three GIS
layers: land use, impervious cover, and soil type. Table V-2 lists the land-use categories used in Opti-
Tool and provides a crosswalk with the GIS layer categories.
49
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table V-2. Land-Use Categories
Opti-Tool Land-Use
MassGIS ID Code
MassGIS Description
Agriculture
1
Cropland
Agriculture
2
Pasture
Agriculture
23
Cranberry Bog
Agriculture
26
Golf Course
Agriculture
36
Nursery
Commercial
15
Commercial
Commercial
29
Marina
Commercial
31
Urban Public/Institutional
Forest
3
Forest
Forest
4
Non-Forested Wetland
Forest
35
Orchard
Forest
37
Forested Wetland
Forest
40
Brushland/Successional
High Density Residential
10
Multi-Family Residential
High Density Residential
11
High Density Residential
Highway
18
Transportation
Industrial
5
Mining
Industrial
16
Industrial
Industrial
19
Waste Disposal
Industrial
39
Junkyard
Low Density Residential
13
Low Density Residential
Low Density Residential
38
Very Low Density Residential
Medium Density Residential
12
Medium Density Residential
Open Land
6
Open Land
Open Land
7
Participation Recreation
Open Land
8
Spectator Recreation
Open Land
9
Water-Based Recreation
Open Land
17
Transitional
Open Land
24
Powerline/Utility
Open Land
25
Saltwater Sandy Beach
Open Land
34
Cemetery
Water
14
Saltwater Wetland
Water
20
Water
The total impervious area (TIA) associated with each Opti-Tool land-use category was adjusted to
directly connected impervious area (DCIA) or effective impervious area using the following formula
(refer to Table V-3 for adjustment factors):
DCIA = C x TIA
50
-------
Mystic River Watershed TMDL Alternative Development Final Report
Unconnected impervious area (i.e., TIA minus DCIA) was then distributed proportionally among
the pervious soil types within the land-use (e.g., if 70 percent of the soil was assigned as type A soils,
then 70 percent of the unconnected impervious area was re-assigned to type A soils). Table V-3
shows DCIA adjustment factors from MA MS4 Permit.
Table V-3. DCIA Adjustment Factors from MA MS4 Permit
Opti-Tool Land-Use
Adjustment Factor (C)
Agriculture
0.004
Commercial
0.570
Forest
0.001
High Density Residential
0.360
Highway
0.440
Industrial
0.670
Low Density Residential
0.110
Medium Density Residential
0.160
Open Land
0.080
Table V-4 integrates the land-use categories with soil and impervious cover categories. Table V-5
lists the data sources for the layers discussed above, all of which were obtained from the MassGIS.
Table V-4. Opti-Tool Export Rates by HRU
Opti-Tool HRU
Average Annual
PLER
(Ibs./acre/year)1
Average Annual
Rainfall-Runoff Rate
(in/year)
Average Annual
Flow-weighted TP
SW concentration
(mg/L)
Agriculture Impervious
1.49
39.51
0.17
Forest Impervious
1.49
39.51
0.17
Highway Impervious
1.38
39.51
0.15
Industrial Impervious
1.79
39.51
0.20
Commercial Impervious
1.79
39.51
0.20
High Density Residential Impervious
2.36
39.51
0.26
Medium Density Residential
Impervious
1.95
39.51
0.22
Low Density Residential Impervious
1.49
39.51
0.17
Open Land Impervious
1.49
39.51
0.17
Agriculture Pervious
0.44
2.50
0.78
Forest Pervious
0.11
2.50
0.19
Developed2 Pervious A
0.03
0.46
0.29
Developed2 Pervious B
0.11
2.50
0.19
Developed2 Pervious C3
0.21
5.64
0.16
Developed2 Pervious C/D
0.30
7.54
0.18
Developed2 Pervious D
0.37
10.30
0.16
1. Based on simulations spanning 1992 2017.
51
-------
Mystic River Watershed TMDL Alternative Development Final Report
2. Developed Pervious land categories include commercial, industrial, residential, highway, and open land (see Table 1-
2 of Attachment F to 2016 MA MS4 General Permit).
3. Areas with undefined or unknown soil types were assumed to be soil type C.
Table V-5. GIS Layers Used to Develop HRUs
Type
Description
Source
Land Use (2005)
The Land Use (2005) data layer is a
Massachusetts statewide seamless digital
dataset of land cover/land use, created
using semi-automated methods, and based
on 0.5-meter resolution digital ortho
imagery captured in April 2005.
http://www.mass.gOv/a nf/research-and-
tech/it-serv-and-support/application-
serv/office-of-geographic-information-
massgis/datalayers/lus2005.html
Impervious Cover
(2005)
The Impervious Surface raster layer
represents impervious surfaces covering
the Commonwealth of Massachusetts. The
surfaces were acquired in April 2005 as
part of the Color Ortho Imagery project.
http://www.mass.gOv/a nf/research-and-
tech/it-serv-and-support/application-
serv/office-of-geographic-information-
massgis/datalayers/im pervioussurface.html
NRCS SSURGO-
Certified Soils
(2014)
The SSURGO-certified soils dataset is
generally the most detailed level of soil
geographic data developed by the National
Cooperative Soil Survey. The data include a
detailed, field-verified inventory of soils
and miscellaneous areas that normally
occur in a repeatable pattern on the
landscape and that can be cartographically
shown at the scale mapped.
http://landscapeteam.maps.arcgis.com/ap
ps/SimpleViewer/index.html?appid=4dbfec
c52fl442eeb368c435251591ec
Drainage Sub-
Basins (2007)
MassGIS has produced a statewide digital
data layer of the approximately 2,300 sub-
basins as defined and used by the USGS
Water Resources Division and the
Massachusetts Water Resources
Commission and as modified by Executive
Office of Environmental Affairs (EOEA)
agencies.
http://www.mass.gOv/a nf/research-and-
tech/it-serv-and-support/application-
serv/office-of-geographic-information-
massgis/datalayers/subbas.html
Elevation
(Topographic)
Data (2005)
These data represent the "bare earth"
elevation of the terrain surface without
vegetation and artificial features. As a
requirement for the orthorectification
process of the 1:5,000 Color Ortho Imagery
(2005), elevation points were compiled
photogrammetrically by human operators
from imagery acquired by Sanborn, Inc. in
April 2005.
http://www.mass.gOv/a nf/research-and-
tech/it-serv-and-support/application-
serv/office-of-geographic-information-
massgis/datalayers/elev2005.html
Sub-basins delineated within the Mystic River Watershed were based on the connectivity to four
critical reaches and seven ponds. The four critical reaches chosen were modeled using the receiving
water model while ponds selected were those identified for future TMDL modeling using the LLRM
model.
52
-------
Mystic River Watershed TMDL Alternative Development Final Report
The three critical reaches are:
Upper Lobe of Upper Mystic Lake.
The Main-body of Upper Mystic Lake.
Upper Mystic Basin
Lower Mystic Basin.
The major lakes/ponds are:
Blacks Nook Pond (MA71005), Cambridge.
Horn Pond (MA71019), Woburn.
Judkins Pond (MA71021), Winchester.
Mill Pond (MA71031), Winchester.
Spy Pond (MA71040), Arlington.
Wedge Pond (MA71045), Winchester.
Winter Pond (MA71047), Winchester.
The sub-basin delineation within the freshwater portion of the Mystic is presented in Figure V-l
(note: the stippled area denotes the area served by a combined sewer). Figure V-2 shows the
expected routing scheme through the sub-basins in the Mystic River Watershed from the headwaters
to the outlet. The delineation allows modelers to characterize flow and pollutant loading to each of
the three critical reaches, seven major lakes/ponds, and at mainstream segments (the Aberjona
River, the Maiden River, and Alewife Brook).
Sub-basins were named for the pond or water body to which they drained (e.g., the upper reaches of
the Aberjona River drainage area were labeled the Judkins Pond sub-basin since this is the water
body located at the bottom of the drainage area) or to main stream segment in which they
encompass. The exceptions to this naming convention are the Aberjona River 1 and Aberjona River
2 sub-basins which are split around the USGS streamflow gauge for the river.
Estimated annual stormwater (SW) phosphorus loads and runoff volumes (1992 2017) for the
entire Mystic River watershed are presented in Figure V-I. Details for the three critical water quality
segments and seven impaired ponds within the watershed are presented in Appendix E.
53
-------
My stic River Watershed TMDL Alternative Development Final Report
>
c
1
u
CO
O)
5
o
c
cc:
1,SQO,000
1,600,000
1,400,000
1,200,COO
1,000,000
BOO,000
600,0G0
400,COO
200,000
0
1992 1994 1996 1998 20CO 2002 2004 2006 200B 2010 2012 2014 2016
5W GW ~ota)
Tl CT| rr. Ci CTl en CTl CTl OOOOOOOOOO^HtHrHt-frHr-J^lrH
aa?)99aaaRRRRRRRRRRRRRRRRRR
ฆ 5W
. GW
rota I
Figure V-l. Estimated Annual Runoff Volume and TP Load for the Mystic River
54
-------
Mystic River Watershed TMDL Alternative Development Final Report
Upper Mystic/
Aberjona River
Upper Lobe of
Mystic Lake
Upper Mystic
Lake
Alewife Brook
Lower Mystic
Lake
Subbasin
Subbasin
Subbasin
(includes Mill R)
Subbasin
Maiden River
Upper Mystic
Basin
Lower Mystic
Basin
Subbasin*
Tributary*
Massachusetts
Bay
Critical
Modeled
Mill
River
Lower Mystic
Lake
Basin
\
Upper Mystic
Lake
Upper Lobe of
Mystic Lake
Subbasin
Modeled
Critical
Subbasin loads are unattenuated while tributary loads are attenuated by the river network
Figure V-II. Mystic River Watershed Sub-Basin Delineation and Schematic
Diagram for Final Model
55
-------
Mystic River Watershed TMDL Alternative Development Final Report
V.B. Groundwater Loads
Groundwater loads to the river system are those that result from water that infiltrates through the
soil and moves via groundwater flow through the underlying aquifer. In general, phosphorus
movement is retarded through soils and aquifers by years or decades because it is highly adsorbed by
clay particles in either media.
Groundwater flow was estimated by analysis of available streamflow records and separating these
components into stormflow and baseflow (e.g., USGS, 1996; Arnold and Allen, 1999, Arnold, et al.,
2005). Baseflow approximates groundwater flow assuming that the riparian evapotranspiration is
minimal.
Streamflow data suitable for estimating stormflow and baseflow components is available at the
Aberjona River and Alewife Brook USGS gages (Figure V-III). Using the Soil & Water Assessment
Tool (SWAT) Baseflow (Bflow) program, baseflow was estimated at both locations and annual
baseflow / total streamflow fractions were computed (Appendix F),. Figure V-III and Figure V-IV
display baseflow and total streamflow for the Aberjona River and Alewife Brook for 2016.
Streamflow Baseflow
u
4)
ro
160
140
120
100
80
60
40
20
0 ~~
01/2016 02/2016 04/2016 05/2016 07/2016 09/2016 10/2016 12/2016
Figure V-III. Baseflow Estimates and Streamflow Measurements at Aberjona River
for 2016 (USGS Gage 01102500)
56
-------
My stic River Watershed TMDL Alternative Development Final Report
50
45
40
0
01/2016 02/2016 04/2016 05/2016 07/2016 09/2016 10/2016 12/2016
Streamflow Baseflow
Figure V-IV. Baseflow Estimates and Streamflow Measurements at Alewife Brook
for 2016 (USGS Gage 01103025)
From 1992 2017 in the Aberjona River, the average annual baseflow contribution to total
streamflow was approximately 65 percent. From 2006 2016, the average in Alewife Brook was
approximately 70 percent. For each HRU in these basins, baseflow is assumed to be an equivalent
fraction,. For HRU's not in either basin, an average (68 percent) of the two estimated values were
applied. Annual groundwater flow contributions were computed for each HRU using the following
equation:
Groundwater Flow = (f x Stormwater Flow) / (1 f)
Where "Stormwater Flow" was the HRU rainfall-runoff total estimated using the Opti-Tool model
and f is the groundwater fraction of total streamflow.
Groundwater phosphorus concentration can be estimated from well sampling data; however, such
data is unavailable. It is difficult to estimate different groundwater concentrations for each HRU
without a well sampling effort devoted to that end. A well sampling effort undertaken in the Boston
metropolitan area found an average region TP concentration of 0.008 mg/L (Flanagan, et al., 2001).
Based on this study, a groundwater TP concentration of 0.008 mg/L was assumed for all basins.
Using groundwater concentration (i.e., 0.008 mg/L) and baseflow for each sub-basin, the
groundwater phosphorus load from the sub-basin was calculated as the product of the two inputs.
Estimated annual groundwater (GW) phosphorus loads and runoff volumes (1992 2017) for the
entire Mystic River watershed were previously presented in Figure V-I. Details for the three critical
water quality segments and seven impaired ponds within the watershed are presented in Appendix
E.
V.C. Sediment Loads
Ponds and impoundments typically act as sinks for nutrients, but they can become net sources to
downstream waters when internal nutrient stores are mobilized and exported (Powers et al., 2015).
In the Mystic River Watershed, there have.: been no direct measurements of nutrient release: rates
from the sediments, which represent a major data gap in this watershed.
57
-------
Mystic River Watershed TMDL Alternative Development Final Report
In the Upper/Middle Charles Nutrient TMDL (EPA/DEP, 2007), which used measured nutrient
release rates from sediments, EPA/DEP estimated that about 22 percent of upstream land loads
were retained on net in sediments. As there have been no direct measurements of nutrient release
rates from the sediments in the Mystic River Watershed, a net attenuation factor will be assumed
through the watershed based on a reach detention time. The initial attenuation factor will be used as
a calibration parameter (see Section VII for discussion of attenuation).
V.D. Observed Loads
The estimation of observed loads was made at sites that have concurrent flows and water quality
data to allow the most accurate estimation of the annual phosphorus load. These sites are the USGS
gages at the Aberjona River (1939-2017), Alewife Brook (2005-2017), and Mystic River (2015-2017),
which all have reliable daily flow records.
V.D.1. Adjustments of Streamflow
Data gaps on the order of several days exist in the streamflow records for the Aberjona River and
for Alewife Brook. In order to develop flow estimates for missing days, a series of regression
relationships were developed between the USGS sites (Figure V-V and Figure V-VI). These
regressions were used to fill missing days in records at the two sites and to develop modeled
estimates of daily average flow rates at the Alewife Brook and Mystic River for dates outside their
respective periods of record.
October 2005 - December 2017
250
200
150
100
50
y-0.1494X+3.8475
.
R1
-0.7653
**
.*>ฆ***
..
kl'/
***
200 400 600 800 1000 1200 1400 1600
Aberjona Flow, CFS
Figure V-V. Regression of Flow Measurements at Aberjona River (01102500) and
Alewife Brook (01103025) USGS Stations
58
-------
Mystic River Watershed TMDL Alternative Development Final Report
June 2016 - December 2017
350
300
250
E 200
1 ISO
u
1. 100
50
0
y-0.9545x+11.264
....
R1
-0.8412
#
1
li
ฆ M
50 100 150 200 250 300 350
Aberjona Flow, CFS
Figure V-VI. Regression of Flow Measurements at Aberjona River (01102500) and
Alewife Brook (01103025) USGS Stations
V.D.2. Adjustments of Total Phosphorus Concentrations
Differences in the TP concentrations from different laboratories for the same water body were
known by MyRWA prior to commencing this study in early 2017. These differences were previously
discussed in the memo "Dataset Assessment for Development and Calibration of the Mystic River Watershed
Loading and Receiving Water Quality Models" (Walker, September 15, 2017).
The TP differences had to be reconciled prior to proceeding with the Alternative TMDL analysis.
The options for reconciliation were: (1) choosing one of the two data sources as the preferred data,
or (2) finding a way to convert from one source to another. Option 2 was preferred since it would
allow more instances where TP and chl-a were both measured at the same site, which is a
requirement for choosing a suitable model calibration period.
To reconcile the data, we went back to the original data in the memo by EPA and MyRWA entitled
"Mystic Raver TP Laboratory Split-Study Results and Discussion (Hrycyna, September 7, 2017). They
reported TP values measured using EPA-approved methods from the two laboratories: MWRA
Deer Island (MWRA) and EPA Region 1 in Chelmsford (EPA). TP results for side-by-side field
samples in the Mystic River Watershed were systematically higher at MWRA (Method 365.4) than
EPA (Method 365.1). Consultation with other laboratory experts in water quality analysis led to the
conclusion that the differences between labs for field samples is a result of better
conversion/digestion of TP to orthophosphate for Method 365.4 versus 365.1. Because Method
365.1 was considered to be have an incomplete conversion to orthophosphate, it is likely an
underestimate of TP.
For this study, Method 365.4 was considered to be "true" value of TP. Although there were some
seasonal differences between April and December in the relationship between the two methods, we
opted to use a single equation for all the data (see Figure V-VII). This approach results in a lower
slope than Walker's (2017) equation because only the April data were available for that analysis. The
final equation to convert TP data using Method 365.1 to Method 365.4 was:
Y = 1.15 X + 21.7
59
-------
Mystic River Watershed TMDL Alternative Development Final Report
where:
Y = MWRA TP concentrations using method 365.4
X = EPA TP concentrations using method 365.1
Biweekly and monthly sampling of phosphorus data at the Aberjona River Mystic River Watershed
Association (MyRWA) monitoring site (ABR006), Mystic River site (MYR071), and Alewife Brook
(ALB003, ALB006) were linearly interpolated on a daily basis in order to produce an estimated daily
time series of TP concentrations (Figure V-VII through Figure V-X).
140
M 120
3
in 100
to
00
T3
O 80
+-ป
(U
CuO
_c
*!7)
3
CL 40
I-
<
cc
^ 20
0
0 20 40 60 80 100
EPA TP using Method 365.1 (|ig/L)
Figure V-VII. Correction of TP Values from Method 365.1 to 365.4
___ Aberjona River
i
"m 600
5 500
0
ro 400
+-ป
ฃ 300
1 200 .
1
12/1999 09/2002 05/2005 02/2008 11/2010 08/2013
05/2016
Figure V-VIII. Adjusted Total Phosphorus Concentrations in the Aberjona River
60
~~
~ >/y
= 1.1543x + 21.7
/ #
R2 = 0.8582
~
4
~
-------
Mystic River Watershed TMDL Alternative Development Final Report
Mystic River
^ 200
o 150
H 12/1999 09/2002 05/2005 02/2008 11/2010 08/2013 05/2016
Figure V-IX. Adjusted Total Phosphorus Concentrations in the Mystic River
__ Alewife Brook
_i
400
H 12/1999 09/2002 05/2005 02/2008 11/2010 08/2013 05/2016
Figure V-X. Adjusted Total Phosphorus Concentrations in Alewife Brook.
V.D.3. Observed Total Phosphorus Loads
Observed loads for TP were calculated for the river reaches where USGS flow gages and
measurements of nutrient concentrations coincide. The measured TP sites and flow measurements
were both available at USGS flow gages, namely, the Aberjona River (USGS 1102500, ABR006),
Lower Mystic Lake (USGS 1103010, MYR071), and Alewife Brook (USGS 1103025,
ALB003/ALB006) gages. At each of the three gages, estimates of concurrent streamflow and total
phosphorus concentration were used to calculate the daily total phosphorus load (daily flow x daily
concentration) and was summed to give a time series of annual loads.
61
-------
Mystic River Watershed TMDL Alternative Development Final Report
V.D.4. Calibrated Streamflow
A comparison of land-based and observed flows was performed for the two-gauge sites that have
long-term flow data (Aberjona River, Alewife Brook). The analysis of 1992-2017 daily flow data
revealed a close match to the observed average flow at Aberjona River but an underestimate of
observed average flow at Alewife Brook. The average observed flow at Aberjona River was
reasonable (21.3 in/yr.) since it compares well with typical flows for eastern Massachusetts rivers
(20-25 in/yr.). In contrast, the average observed flow at Alewife Brook was very low (15.3 in/yr.).
We have no explanation for this apparent low average flow at Alewife Brook but decided not to use
it. Comparison of average annual streamflows from land-based estimates and observed values at the
Aberjona River gauge were reasonable (R2=0.72) with no visible bias. Given the uncertainty of the
Alewife Brook flow data and the decent fit of modeled land-based loads with observed flows for the
Aberjona gauge, we decided not to calibrate streamflow for this study.
V.D.5. Calibrated Total Phosphorus Loads
Observed loads within a reach are often less than the land loads because there has been some
attenuation (or retention) of nutrients within the river system. Attenuation of nutrients in river
reaches occurs because of biological and chemical changes, plant uptake, particulate settling, and
organic settling from algae or aquatic plant senescence. Higher residence time, or detention time,
usually means more nutrient attenuation occurs.
Estimation of attenuated land loads was performed for the period 2007-2016 since this is the
modeling period (see Section VII.C.3). Attenuated cumulative TP loads for each reach were initially
determined from the modeled land values by using an estimate of instream attenuation factors based
on reach detention time (volume/flow, days). These estimates were calibrated to match observed
reach loads (see below).
A review of the literature on phosphorus attenuation in impoundments revealed two similar curves;
Figure 7 of Powers et al. (2015) and Figure 4 of Koiv et al. (2011). Both of these curves are shown in
Figure V-XI. The equation derived from the Powers et al. (2015) data, which indicates slightly higher
attenuation rates than Koiv et al. (2011) was chosen for the Mystic watershed because the calibration
process indicated that high attenuation factors were needed. The Powers equation for the
attenuation factor is:
Attenuation Factor = 0.5598 + 0.1278+loge (Rd/365), Rd = detention time (days)
62
-------
Mystic River Watershed TMDL Alternative Development Final Report
1.0
0.9
0.8
200 400 600
Detention Time (d)
Koiv (2011) - Powers (2015)
800
1000
Figure V-XI. Phosphorus Attenuation Curves
Detention time (days) was computed from the estimated reach volume (m3) and the modeled annual
flow rate (m3/d). Impoundment volumes were estimated from various pond studies (WH, 1987,
1988; DFW, no date; ENSR, 2000; DEP, 2010; and EPA, 2018). For river reaches (i.e. not pond
reaches) that did not have information to compute an attenuation factor, the value was set to a
nominal value of 0.05 (see Table VII-4).
Predicted TP land loads were calibrated to measured loads at three calibration sites (Aberjona,
Lower Mystic, and Alewife USGS gauges) by changing the reach attenuation factors. The calibration
objective was to minimize the error between observed and attenuated TP loads as closely as possible
at the three sites while still maintaining reasonable attenuation coefficients (0.05 to 0.9) that were
also reasonably close to the initial estimates. The process was conducted sequentially down the
watershed, starting at the most upstream calibration point first. All initial attenuation factors above
this point were adjusted by the same factor to best match the observed data. This process was
repeated downstream. Attenuation in all the reaches contributing to Alewife Brook was increased
further to meet the observed load for that calibration site.
With this process, the investigators were able to reduce the error in the annual average loads to zero
at the three sites and still have reasonable attenuation factors, although the ones in the Alewife
Brook watershed seem a little high, possibly because the flow at this gauge is unreasonably low (see
Section V.D.1X). The error at all sites except the three gauges is unknown because there were no TP
load estimates at these other sites. The estimated and final attenuation factors for TP are given in
Table V-6 and a comparison of modeled land versus attenuated sub-basin loads is given in Figure
V-XII. There might also be some monthly error at the two-gauge sites, but this is not relevant to the
receiving water model which is based on annual load inputs.
63
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table V-6. Estimated Reach Detention Times and Attenuation Factors
Reach Name
Reach Volume
(m3)
Cumulative
Reach Flow
(mWr.)1
Detention Time
(d)
Estimated
Attenuation
Factor (-)
Final TP
Attenuation
Factor (-)
Winter
Pond
58,768
254,236
84
0.37
0.70
Horn
Pond
2,951,250
10,898,619
99
0.39
0.85
Wedge
Pond
277,555
11,968,307
8
0.08
0.20
Judkins
Pond
9,065
37,176,274
0
0.05
0.05
Mill
Pond
6,475
37,395,435
0
0.05
0.05
Aberjona
River 1
37,704,020
0.05
0.05
Calibration Site is USGS Streamflow Gauge at Aberjona River (1102500
Aberjona
River 2
38,177,963
0.05
0.05
Upper
Lobe
219,434
38,368,707
2
0.05
0.05
Upper Mystic
Lake
7,385,437
40,449,476
67
0.34
0.26
Lower Mystic
Lake
3,529,612
49,102,911
26
0.22
0.19
Calibration Site is USGS Streamflow Gauge at Lower Mystic Lake (1103010)
Blacks Nook
Pond
15,110
5,969
924
0.68
0.90
Spy
Pond
1,690,000
2,110,121
292
0.53
0.80
Alewife Brook
-
17,560,488
-
0.05
0.48
Calibration Site is USGS Streamflow Gauge at Alewife Brook (1103025)
1 From land loading model
64
-------
Mystic River Watershed TMDL Alternative Development Final Report
9,000
8,000
^ 7,000
6,000
ฆa
o 5,000
ฆ Land Attenuated
Figure V-XII. Modeled Land vs. Attenuated Reach Phosphorus Loads (2007-2016)
65
-------
Mystic River Watershed TMDL Alternative Development Final Report
VI. Evaluation of Combined Sewer Overflow and Sanitary
Sewer Overflow Data for the Mystic River Watershed
The following describes the data types, sources and approaches used to develop estimates of CSO
and SSO volumes and phosphorous loads in the Mystic River Watershed. CSO and SSO data were
quantified in an effort to provide a more accurate estimate of all loads entering the waterbodies
within the watershed. The land loads provided in Section V will be added to the CSO and SSO
loads and modeled in Section VII. As discussed in Section V, annual loads are the expected input
for the receiving water models; therefore, load estimates have been developed on an annual basis.
VI.A. Data Types and Sources
Four primary data types were required to evaluate CSO and SSO contributions:
Spatial data (CSO and SSO drainage areas).
Volumetric data (annual CSO and SSO discharge volumes).
Annual precipitation data.
CSO and SSO discharge concentrations for total phosphorus and total nitrogen.
The CSO and SSO data used for this analysis included GIS data, Excel spreadsheets, reports and
literature. The sources of these data are noted in the sections below.
VI.A. 1. Spatial Data
There are two sub watersheds in the Mystic River Watershed that contain CSO drainage areas:
Alewife Brook and Mystic River. The outfalls included in these analyses are (location in parenthesis):
Alewife Brook: CAM001 (Cambridge), CAM002 (Cambridge), MWR003 (Cambridge),
CAM004 (Cambridge), CAM400 (Cambridge), CAM401A (Cambridge), CAM401B
(Cambridge), SOMOOIA (Somerville).
Mystic River: SOM007A/MWR205A (Somerville). MWR205 (Somerville), which is located
downstream of the Amelia Earhart dam, was evaluated to compare total discharges from the
Somerville Marginal CSO Facility.
The MWRA CSO map is presented in Appendix G (Figure G-XI-1). The data collected for the CSO
drainage basins included a GIS polygon file provided by the City of Cambridge showing areas
contributing to Alewife Brook with a corresponding attribute table noting the year those areas were
separated. Maps showing the City of Cambridge's CSO drainage basins from 2000 and 2017 are
presented in Appendix G (Figure G-XI-2 and Figure G-XI-3). Additional GIS data on CSO
drainage areas were provided by the City of Somerville in Appendix G (Figure G-XI-4). Portions of
Somerville's CSO drainage areas discharge to Alewife Brook while others discharge to the lower
Mystic River directly, depending on the size of the storm event (see attached map).
There were no GIS data available for SSO drainage areas. A spreadsheet from the Massachusetts
Water Resources Authority (MWRA) provided estimated volumetric discharge data and latitude and
longitude for each SSO. Those locations were converted into a GIS point shapefile that was overlaid
onto the Mystic River sub watersheds to assign SSOs volumes to their recipient sub watersheds.
66
-------
Mystic River Watershed TMDL Alternative Development Final Report
VI.A.2. Volumetric Data
The data sources for volumetric data gathered are shown in Table VI-1.
Table VI-1. Volumetric Data Sources for CSOs and SSOs
Data Type
Source
Years
Data Comments
CSO data
MWRA
2000 to 2017
Modeled data; from Annual Reporting
City of Cambridge
2006 to 2017
Annual NPDES Reports
City of Somerville
2016
Annual NPDES Report
SSO data
MWRA
2000 to 2017
Data from MWRA directly; no SSOs for
2016 or 2017. Includes
latitudinal/longitudinal information.
Massachusetts Department of
Environmental Protection
(MassDEP)
2000 to 2017
Data between 2000-2016 received from
Mystic River Watershed Association;
spreadsheet format. No spatial data
provided. Data includes all known
discharges within the Mystic River
Watershed (including MWRA data).
2017 data were digitized from forms
provided by MassDEP for this project.
The data were reviewed to determine completeness (e.g., data gaps, including missing values or no
data) and for consistency (e.g., extreme data). For CSOs, data provided by the cities were also cross
checked with the MWRA data. In the annual CSO reports, modeled CSO activation durations and
volumes were compared to the reported CSO activations. In some instances, the reported data were
not consistent with the modeled data (model was either over or under predicting), which appeared
to be due to either the modeling of the CSO system/outfall or the resolution of the metered data
versus modeled data. Based on the annual reports, improvements to the model appeared to have
been made over time to address some of these issues.
For SSOs, MWRA and MassDEP noted that the reporting of SSO volumes was updated in 2012.
Prior to 2012 the form only included SSO volume ranges, which generally were: <10,000 gallons,
10,000 to 100,000 gallons, 100,000 gallons to 1,000,000 gallons and > 1,000,000 gallons, limiting the
upper limit of reported SSO volumes. Starting in 2012, the form was updated to allow an estimate of
SSO volume and method for estimating.
VI.A.3. Precipitation Data
Hourly precipitation data for Boston Logan International Airport was extracted from the Opti-Tool
model for 1992 to 2016. Additional hourly data for 2017 was downloaded from NOAA's website
(NOAA, 2018).
VI.A.4. Nutrient Concentrations
Nutrient concentrations for CSO and SSO discharges were estimated to facilitate the calculation of
annual phosphorus and nitrogen loads. TP and total nitrogen (TN) concentrations for CSOs were
based on data from Breault et. al. (2012). The document reported average CSO TP and TN
concentrations of 3.1 mg/L and 9.3 mg/L, respectively, for samples collected by MWRA. SSO TP
and TN concentrations were based on the average annual influent wastewater concentrations for
2016 sampled at the Deer Island Sewage Treatment Plant of 5.23 mg/L and 41.8 mg/L,
67
-------
Mystic River Watershed TMDL Alternative Development Final Report
respectively.6 The analysis used in this memorandum assumes that that SSOs discharges have similar
concentrations to untreated wastewater.
VI.B. CSO Data Analyses
After discussions with MWRA and EPA Region 1, the MWRA CSO data were further analyzed to
evaluate how precipitation and time influenced the CSO data and how average CSO volumes
compared to the annual data. Two data analyses were performed on the datasets: regression analyses
and evaluation of statistical outliers. The purpose was to determine if the average annual CSO
volume for the evaluation of phosphorus load reduction estimates (Section IX) was appropriate
and/or if modifications to the annual CSO volumes would allow a more representative average. In
particular the analysis focused on years where no CSOs were reported (modeled discharge estimates
noted as '0') and CSO data extremes.
VI.B.1. CSO Analyses
CSO volumetric data were plotted against time and annual precipitation depth to evaluate potential
relationships. The volumetric data were also normalized by acreage of CSO drainage basins
contributing to the CSO outfalls. Figure VI-I through Figure VI-IV represent these comparisons.
There were significant linear relationships with time (years) and annual rainfall for the Alewife CSO
drainage basin (p-value < 0.05). There were no similar significant relationships for the Mystic River
CSO drainage basin, which might be due to the complex connection of Somerville's CSO system to
Alewife Brook, the Mystic River, and other sewer systems.
VI.B.2. CSO Statistical Outliers
A statistical analysis was conducted on each of the CSO datasets to determine if there were any
outliers that could be having excessive influence on the volumetric averages. This process first
involved defining the upper bounds of the annual volumetric data to identify the outliers using the
following two equations:
Upper Bound: Q3 + (1.5*IQK)
where Q3 is the third quartile value and IQR is the interquartile range or the difference between the
first and third quartiles (Q1 and Q3). Outliers were defined as values higher than the upper bound.
The results are presented in Table VI-2. The mean of the datasets was also calculated.
6 SSO TP and TN concentrations were based on MWRA's North System influent only, which includes communities
in the Mystic Watershed.
68
-------
My stic River Watershed TMDL Alternative Development Final Report
90,000,000
80,000,000
70,000,000
10,000,000
CSO Volumes in the Alewife and Mystic River basins
~ Mystic River
ฆ Alewife
Linear (Mystic River)
Linear (Alewife)
2004 2006 2008 2010 2012 2014 2016 2018
Year
Figure VI-I. Annual CSO Volumes over Time
CSO Volumes in the Alewife and Mystic River basins
90,000,000
80,000,000
70,000,000
= 50,000,000
ru
3B
E 40,000,000
o
O 30,000,000
to
u
20,000,000
10,000,000
y = 3E+06x- 1E+08
R2= 0.6355
~ Mystic River
ฆ Alewife
Linear (Mystic River)
Linear (Alewife)
20 30 40
Rainfall (inches/year)
Figure Vl-ll. Annual CSO Volumes versus Annual Rainfall
69
-------
Mystic River Watershed TMDL Alternative Development Final Report
Alewife CSO Volumes by Area (gal/ac)
>
y
ro
1
00
01
Q.
HI
ฃ
_D
O
>
O
i/i
u
60,000
50,000
40,000
30,000
20,000
10,000
ฆ Alewife CSO
Linear (Alewife CSO)
ฆ
ฆ _
ฆ
y = 1722.4x -57102
RJ = 0.6337
ฆ .
ฆ
ฆ
ฆ
10
20
30
Rainfall (in/yr)
40
50
60
Figure Vl-lll. Normalized Alewife Annual CSO Volumes versus Annual
Precipitation
Mystic CSO Volumes by Area (gal/ac)
30,000
.> 25,000
u
nj
"rc 20,000
ฆSB
S 15,000
g. 10,000
5,000
Q.
cu
E
o
>
O
to
VJ
~ Mystic CSO
Linear (Mystic CSO)
~
y = 154.52x +1283.4
~
~
<
R2 = 0.0281
~
~
i i
~
i i
* , .
10
20 30 40
Rainfall (in/yr)
50
60
Figure VI-IV. Normalized Mystic Annual CSO Volumes versus Annual Precipitation
70
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table VI-2. Statistical Outlier Analysis for CSO Datasets
Statistical Data Type
Alewife CSO Volumes (gal)
Mystic River CSO Volumes (gal)
First Quartile (Ql)
10,590,000
3,660,000
Third Quartile (Q3)
36,220,000
10,085,000
Interquartile Range (IQR)
25,630,000
6,425,000
Upper Bound
74,665,000
19,722,500
Mean (all data)
27,208,333
8,013,636
Mean (without outliers)
22,656,364
6,580,000
Two data points were identified as outliers: CSO volumes from 2010 in the Mystic River Drainage
Basin and 2008 in the Alewife Drainage Basin. However, the volumetric mean was not modified
because data points were within 10 percent of the upper bound, so was not considered to have
undue influence on the volumetric mean. The CSO volumes from 2006 were not available for the
Mystic River and was replaced with the mean using all data to complete the dataset from 2006 to
2017. The final datasets are provided in Table VI-3.
Table VI-3. CSO Volumetric Datasets for Alewife and Mystic River Drainage
Basins
Year
Alewife CSO Volumes (gal)
Mystic River CSO Volumes (gal)
2006
61,540,000
8,013,636a
2007
15,320,000
5,750,000
2008
77,280,000
10,420,000
2009
12,310,000
920,000
2010
63,590,000
22,350,000
2011
27,780,000
9,240,000
2012
21,830,000
11,760,000
2013
5,430,000
9,750,000
2014
22,450,000
5,120,000
2015
12,620,000
9,360,000
2016
1,300,000
1,280,000
2017
5,050,000
2,200,000
a) CSO volume was not available. CSO volume replaced with mean value of all data (2006-2017).
VI.C. SSO Data Analyses
The SSO data were first reviewed and summarized to characterize the raw data, including discharge
frequency and duration, discharge locations and volumes. Then the data were processed to be able
to assign SSO volumes by sub watershed for further evaluation. Finally, similar to CSO data, SSO
data analyses were completed to understand how precipitation and time influenced the annual
volumes and how average SSO volumes compared to annual data. The purpose of the analyses was
to determine if the average annual SSO volume for the evaluation of phosphorus load reduction
estimates (Section IX) was appropriate and/or if modifications to the annual SSO volumes would
allow a more representative average.
71
-------
Mystic River Watershed TMDL Alternative Development Final Report
VI.C.1. SSO Data Review
A more in-depth review of the data entries was completed to understand the frequency and duration
of SSOs, their discharge points and the magnitude of the reported volumes. The data are
summarized below in Table VI-4 and
Table VI-5.
Overall, of the 774 total SSOs reported during the period of available data, only a small percentage
of them were identified as discharging to a catch basin or directly to a waterbody (60 in all, 8% of
total) despite the large number of them being a result of a rain event (528 in all, 68% of total). The
locations, durations and volumes ranged significantly, which suggests that SSO discharges are
dependent on the type and location of event (i.e., there is not a singular threshold event which will
result in SSO discharges in all sub watersheds). For example, note that both 2010 and 2014 appear as
years with much more frequent SSOs, accounting for approximately 68% and 8%, respectively, of
the total discharges over the period from 2006 to 2017.
Further discussions with MWRA and MassDEP on the SSO data indicate that there can be
variability in the data reporting. Data varies in the way that durations are reported (e.g., what is
identified as the 'start' of the SSO event may be assumed or may be at the time it is 'found'),
volumes are calculated (e.g., some locations are monitored, others are estimated), and number of
SSOs identified (e.g., only 'found' SSOs are reported, may be others not being reported). Therefore,
while the data is representative, it may not be consistent or accurate in all cases.
Table VI-4. Review of SSO Data - Frequency, Duration and Discharge Points
Year
Total
Number of
SSOs
reported
Average
duration of
SSOs (hrs.)
Min/Max
Duration of
SSOs (hrs.)
Community with
Highest SSOs
Reported/# of
SSOs
Number of
SSOs Directly
Discharging to a
Catch Basin
Number
Directly
Discharging
to a
Waterbody
2006
24
N/A
N/A
Arlington
(6)/Winchester (6)
2007
21
N/A
N/A
MWRA (15)
-
-
2008
24
N/A
N/A
MWRA (15)
-
-
2009
11
N/A
N/A
MWRA
(2)/Lexington (2)
2010
524
N/A
N/A
MWRA (107)
-
-
2011
22
N/A
N/A
N/A
-
-
2012
26
N/A
N/A
MWRA (4)
-
-
2013
25
4.3
0.3/22
Wakefield (10)
3
1
2014
65
9.6
0.3/52
Medford (29)
7
43
2015
10
8.9
0.2/72.8
Lexington (2)
-
2
2016
8
14.5
0.5/66.5
Burlington (3)
1
3
2017
14
7.1
0/71.8
Cambridge (5)
-
-
Totals
774
-
-
-
11
49
72
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table VI-5. Review of SSO Data - Volumes
Year
Total
Number of
Number of
Average
Minimum
Maximum
Number of
SSOs as a
entries where
reported SSO
Reported SSO
Reported
SSOs
result of a
SSO volume
volume (MG)
volume (gal)
SSO Volume
reported
rain event
was estimated
(MG)
2006
24
17
18
2.10
10,000
26.9
2007
21
17
19
0.20
50
1
2008
24
20
20
0.30
1,000
1
2009
11
2
9
0.07
4,000
0.6
2010
524
404
303
0.30
3.5
9
2011
22
11
20
0.30
2
2.6
2012
26
1
22
0.20
10
1.9
2013
25
0
23
950
5
<0.1
2014
65
47
32
2.40
10
24.6
2015
10
3
10
0.01
10
<0.1
2016
8
0
8
<0.01
2
0.1
2017
14
6
11
0.09
10
<0.1
Totals
774
528
495
-
-
-
VI.C.2. SSO Data Processing
The two available SSO volumetric datasets (see Table VI-1) were merged before proceeding with the
statistical analyses. The following analysis steps were used to process the MassDEP SSO data:
Remove all MWRA data from the MassDEP datasets.
Identify data points from the MassDEP data located in Mystic River Watershed.
Convert volumetric ranges (see Section VI.A.2) to a single volume. For example, data
identified as "<10,000 gallons" was converted to 10,000 gallons. Similarly, ">1,000,000
gallons were converted to 1,000,000 gallons. The intermittent range, "100,000 gallons to
1,000,000 gallons" was converted to a midpoint value (500,000 gallons).
Assign SSO volumes to the sub watersheds.
In contrast to the MWRA data, the MassDEP datasets were not geospatially located, so the data
were filtered and summarized by the communities in the Mystic River. A unit discharge
(gallons/acre) was calculated for each community and then allocated to each sub watershed by area
to estimate the total SSO volumes by sub watershed. Finally, MWRA and MassDEP SSO volumes
were combined by sub watershed.
VI.C.3. SSO Trend Analyses
SSO volumetric data were plotted over time and normalized by the annual rainfall data to transform
the data for evaluating trends. The data are graphed in Figure VI-V and Figure VI-VI. These plots
highlight the lack of observable or statistical relationships between the rainfall and SSO data in any
of the sub watersheds.
73
-------
Mystic River Watershed TMDL Alternative Development Final Report
VI.C.4.
SSO Statistical Outliers
Similar to the CSO outlier analysis presented in Section VI.B.2, an evaluation of SSO outliers was
completed by comparing the dataset to upper volumetric bounds. The results are presented in Table
VI-4. This evaluation identified several outliers; all sub watersheds had a statistical outlier in 2010,
while other watersheds had outliers in 2006 and 2014. After reviewing the data, these outliers were
removed and replaced by the upper bound value. The final datasets are provided in Table VI-5.
VI.C.5. SSO Missing Datasets
Similar to the CSO datasets, missing (or no data) was evaluated to determine if the values should be
replaced or not. Based on the evaluation of the trend analyses and the statistical outliers, it seemed
likely that some sub watersheds with missing data may have zero annual SSO discharge. Therefore,
no missing data was replaced.
VI.C.6.
Total Phosphorus and Total Nitrogen Loads for Model Calibration
The annual watershed phosphorus and nitrogen loading estimates were developed using the final
CSO and SSO discharge volumes and the TP and TN concentrations noted in Section VI.A.4. The
final TN and TP load estimates are presented in Table VI-6, Table VI-7, Table VI-8,
Table VI-9, and Table VI-10.
SSO Volumes
ro
01
>ฆ
40,000,000
35,000,000
30,000,000
25,000,000
20,000,000
C
_o
"S
01
E
3
o 15,000,000
O
i/>
10,000,000
5,000,000
2004
~
||i
2006 2008
2010 2012 2014
Year
2016
~ Aberjona River 1
ฆ Aberjona River 2
A Alewife
Blacks Nook
> Horn
~ Judkins
i Lower Mystic Lake
- Maiden
- Mill
~ Mystic River
ฆ Spy Pond
Upper Lobe
Upper Mystic Lake
Wedge
Winter
2018
Figure Vl-V. Annual SSO Discharge Volumes by Sub-watershed versus Time
74
-------
Mystic River Watershed TMDL Alternative Development Final Report
SSO Volumes
TO
0)
>
l/l
ฃ
_o
.25
o
E
_3
O
>
O
i/>
i/)
40,000,000
35,000,000
30,000,000
25,000,000
20,000,000
15,000,000
10,000,000
5,000,000
10
20
X
~ . +
I *-~
งฆฆ ii iijh-M-
30 40 50
~ Aberjona River 1
ฆ Aberjona River 2
A Alewife
X Blacks Nook
> Horn
~ Judkins
t Lower Mystic Lake
- Maiden
- Mill
~ Mystic River
ฆ Spy Pond
Upper Lobe
Upper Mystic Lake
Wedge
Winter
60
Rainfall (inches/year)
Figure VI-VI. Annual SSO Discharge Volumes by Sub watershed versus Rainfall
75
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table VI-6. Statistical Outlier Analysis for Annual SSO Discharge Volumes
(Gallons/yr.)
Statistical Data
Type
Mystic
Alewife
Aberjona
River 1
Aberjona
River 2
Blacks
Nook
Horn
Judkins
Lower
Mystic Lake
Maiden
Mill
Spy Pond
Upper
Lobe
Upper
Mystic
Lake
Wedge
Winter
1st quartile (25%)
273
8,622
0
0
0
912
460
3,296
213
0
0
0
0
0
0
3rd quartile (75%)
2,591,755
181,440
46,771
168,117
75
1,311,220
847,633
1,764,945
128,257
25,145
262,014
56,666
459,882
114,294
47,818
Interquartile
range
2,591,482
172,818
46,771
168,117
75
1,310,308
847,174
1,761,649
128,044
25,145
262,014
56,666
459,882
114,293
47,818
Upper bound
(Q3+1.5*IQR)
6,478,979
440,668
116,927
420,293
187
3,276,683
2,118,394
4,407,418
320,324
62,863
655,035
141,665
1,149,705
285,733
119,544
Mean (all data)
4,454,583
241,727
31,859
2,290,032
249
3,205,967
3,687,084
3,252,328
621,826
17,128
286,506
38,599
339,345
83,825
36,425
Mean (w/o
outliers)
1,494,411
67,331
47,810
84,824
52
519,601
261,804
640,451
50,047
25,704
189,330
57,925
301,668
73,825
31,069
Table VI-7. Annual SSO Discharge Volumes for All Sub watersheds (Gallons/yr.)
Year
Mystic
Alewife
Aberjona
River 1
Aberjona
River 2
Blacks
Nook
Horn
Judkins
Lower
Mystic
Lake
Maiden
Mill
Spy
Pond
Upper
Lobe
Upper
Mystic
Lake
Wedge
Winter
2006
111,630
212,747
90,077
164,790
0
3,987,702
2,172,54
5
3,887,253
588,314
48,428
356,831
109,134
795,749
226,172
95,998
2007
462,267
17,056
0
45,000
0
798,129
256,280
59,757
450,000
0
11,752
0
4,200
360
232
2008
2,144,904
150,133
72,754
178,100
0
1,026,864
826,009
732,833
21,010
39,115
235,038
88,147
623,718
177,714
74,334
2009
363
9,945
0
0
55
426,822
131,302
17,436
18,069
0
0
0
0
0
0
2010
9,701,117
475,150
138,580
428,612
219
3,987,702
2,172,54
5
6,177,132
588,314
74,504
722,475
167,899
1,286,235
338,584
141,642
2011
84,268
215,806
38,109
114,719
62
2,164,288
912,507
1,057,509
382
20,489
342,942
46,172
405,270
93,154
38,979
2012
3,932,309
33,861
0
0
181
36
60
0
611
0
0
0
0
0
0
76
-------
Mystic River Watershed TMDL Alternative Development Final Report
Year
Mystic
Alewife
Aberjona
River 1
Aberjona
River 2
Blacks
Nook
Horn
Judkins
Lower
Mystic
Lake
Maiden
Mill
Spy
Pond
Upper
Lobe
Upper
Mystic
Lake
Wedge
Winter
2013
4
20,537
0
0
113
508
519
1
215
0
0
0
0
7
4
2014
7,596,241
5,573
34,645
428,612
3
764,894
462,123
6,177,132
9,939
18,626
85
41,975
257,041
84,733
35,467
2015
2
7,299
0
0
1
653
212
4,437
28
0
0
0
0
0
0
2016
1
0
0
0
0
998
280
656
11
0
0
0
0
0
0
2017
1,653
348
3,464
6,388
2
12,814
28,746
4,176
205
1,863
0
4,197
25,701
8,463
3,540
Table VI-8. Estimated Annual Total Phosphorus and Total Nitrogen CSO Loads
(Ibs./yr.) for Alewife and Mystic Sub watersheds
Year
Alewife CSO TP
Mystic River CSO TP
Alewife CSO TN
Mystic River TN
2006
1,591.1
207.3
4,776.3
622.0
2007
396.3
148.8
1,189.0
446.3
2008
1,999.3
269.6
5,997.9
808.7
2009
318.5
23.8
955.4
71.4
2010
1,645.1
578.2
4,935.4
1,734.6
2011
718.7
239.0
2,156.1
717.1
2012
564.8
304.2
1,694.3
912.7
2013
140.5
252.2
421.4
756.7
2014
580.8
132.5
1,742.4
397.4
2015
326.5
242.2
979.5
726.5
2016
33.6
33.1
100.9
99.3
2017
130.6
56.9
391.9
170.7
77
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table VI-9. Estimated Annual Total Phosphorus SSO Loads (Ibs./yr.) for all Sub
watersheds
Year Mystic Alewife Aberjona Aberjona Black Horn Judkins Lower Maiden Mill Spy Upper Upper Wedge Winter
River 1 River 2 s Mystic Pond Lobe Mystic Lake
Nook Lake
2006
4.9
9.3
3.9
7.2
0.0
174.0
94.8
169.7
25.7
2.1
15.6
4.8
34.7
9.9
4.2
2007
20.2
0.7
0.0
2.0
0.0
34.8
11.2
2.6
19.6
0.0
0.5
0.0
0.2
0.0
0.0
2008
93.6
6.6
3.2
7.8
0.0
44.8
36.1
32.0
0.9
1.7
10.3
3.8
27.2
7.8
3.2
2009
0.0
0.4
0.0
0.0
0.0
18.6
5.7
0.8
0.8
0.0
0.0
0.0
0.0
0.0
0.0
2010
423.4
20.7
6.0
18.7
0.0
174.0
94.8
269.6
25.7
3.3
31.5
7.3
56.1
14.8
6.2
2011
3.7
9.4
1.7
5.0
0.0
94.4
39.8
46.2
0.0
0.9
15.0
2.0
17.7
4.1
1.7
2012
171.6
1.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2013
0.0
0.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2014
331.5
0.2
1.5
18.7
0.0
33.4
20.2
269.6
0.4
0.8
0.0
1.8
11.2
3.7
1.5
2015
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2016
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2017
0.1
0.0
0.2
0.3
0.0
9.3
1.3
0.2
0.0
0.1
0.0
0.2
1.1
0.4
0.2
Table VI-10. Estimated Annual Total Nitrogen SSO Loads (Ibs./yr.) for all Sub
watersheds
Year Mystic Alewife Aberjona Aberjona Blacks Horn Judkins Lower Mystic Maiden Mill Spy Pond Upper Upper Wedge Winter
River 1 River 2 Nook Lake Lobe Mysti
c Lake
2006
38.9
74.2
31.4
57.5
0.0
1,390.1
757.9
1,356.0
205.2
16.9
124.5
38.1
277.6
78.9
33.5
2007
161.3
5.9
0.0
15.7
0.0
278.2
89.4
20.8
157.0
0.0
4.1
0.0
1.5
0.1
0.1
2008
748.2
52.3
25.4
62.1
0.0
358.2
288.1
255.6
7.3
13.6
82.0
30.7
217.6
62.0
25.9
2009
0.1
3.5
0.0
0.0
0.0
148.9
45.8
6.1
6.3
0.0
0.0
0.0
0.0
0.0
0.0
2010
3,384.1
165.8
48.3
149.5
0.1
1,391.1
757.9
2,154.8
205.2
26.0
252.0
58.6
448.7
118.1
49.4
2011
29.4
75.3
13.3
40.0
0.0
755.0
318.3
368.9
0.1
7.1
119.6
16.1
141.4
32.5
13.6
2012
1,371.7
11.8
0.0
0.0
0.1
0.0
0.0
0.0
0.2
0.0
0.0
0.0
0.0
0.0
0.0
2013
0.0
7.2
0.0
0.0
0.0
0.2
0.2
0.0
0.1
0.0
0.0
0.0
0.0
0.0
0.0
2014
2,649.9
1.9
12.1
149.5
0.0
266.8
161.2
2,154.8
3.5
6.5
0.0
14.6
89.7
29.6
12.4
78
-------
Mystic River Watershed TMDL Alternative Development Final Report
Year
Mystic
Alewife
Aberjona
River 1
Aberjona
River 2
Blacks
Nook
Horn
Judkins
Lower Mystic
Lake
Maiden
Mill
Spy Pond
Upper
Lobe
Upper
Mysti
c Lake
Wedge
Winter
2015
0.0
2.5
0.0
0.0
0.0
0.2
0.1
1.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2016
0.0
0.0
0.0
0.0
0.0
0.3
0.1
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2017
0.6
0.1
1.2
2.2
0.0
4.5
10.0
1.5
0.1
0.6
0.0
1.5
9.0
3.0
1,2
79
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
VII. BATHTUB Modeling Approach
This section summarizes the modeling approach and calibration and validation results for the
BATHTUB receiving water model of the Mystic River. The BATHTUB model uses instream
measurements of TP, TN, orthophosphate (OP), inorganic nitrogen (IN), chlorophyll-a (chl-a), and
oxygen depletion in the hypolimnetic and metalimnetic layers (if the data are available) for
calibrating the model.
The calibrated BATHTUB model was used to run nutrient reduction scenarios to identify the
reductions necessary to bring the critical receiving water reaches (see Section V.A.I) into compliance
with water quality targets selected for this project.
VILA. Model Selection
Prior to modeling of water quality of the Mystic River, three receiving water models were reviewed,
namely, the LLRM (Wagner, 2009), BATHTUB (Walker, 2004), and AQUATOX (Clough, 2014)
models (see memo "Options for Modeling the Mystic River Watershed" (ERG/Pickering/PGE, May 26,
2017). These models were selected based on the conditions, data, and modeling needs in the Mystic
River Watershed. They all predict receiving water quality (nitrogen, phosphorus, chlorophyll-a, etc.)
to water inputs and nutrient loads (nitrogen, phosphorus) from all contributing sources (land, point,
and atmospheric loads) in the watershed. They are all scientifically sound and well suited for this
task.
The LLRM and BATHTUB models are both regression-based models that are similar in complexity
and effort, whereas the AQUATOX model is a much more mechanistic model. All three models
predict the nutrient water quality responses in multiple reaches, lakes, and impoundments. The
AQUATOX model was eliminated from consideration because it requires significantly more time
and effort. The other two models are similar in complexity and level of effort however, the
BATHTUB model offers a number of extra features like additional water quality inputs and outputs,
the ability to link a number of reaches or sub-reaches together, and an easy-to-use interface.
The BATHTUB model is an appropriate choice for this study because it is a semi-empirical model
that computes both the mass balance of each segment and utilizes empirical relationships between
the water quality variables (see Figure VII-I). Those empirical relationships have been calibrated to a
large dataset from Army Corps of Engineers (US-ACOE) reservoirs across the country (Walker,
1982; Walker, 1985). The tool is also appropriate to the level of available water quality data in the
Mystic River since it has a limited number of calibration factors that can be adjusted thus avoiding
over-calibration.
Based on input from the technical team and the TSC, the BATHTUB model was selected.
BATHTUB Version 6.2 was used for this modeling task. BATHTUB has been used in in many
other similar studies (Walker, 1996; Walker 2004). In particular, the BATHTUB model was used in
the development of the Lake Champlain Phosphorus Total Maximum Daily Load study (TetraTech,
2015; EPA, 2016).
Since nitrogen was ultimately not used in the final calibration, discussion of nitrogen has been
minimized in the rest of the BATHTUB modeling section.
80
-------
Mystic River Watershed TMDL Alternative Development Final Report
Figure Vll-I. BATHTUB Model Schematic
VII.B. Model Setup
VII. B. 1. Segmentation
The BATHTUB model allows for a number of segmentation options in the set up. A basic setup
would use a single upstream tributary to a single water body. The model also allows for river reaches
to be linked together, each with single or multiple tributary inputs, and multiple sub-reaches within a
reach. The chosen configuration for this project is discussed below.
The final BATHTUB model was divided into 5 segments (see Figure VII-II) as described in the list
below. Each segment is numbered from upstream to downstream, with abbreviated names in
parentheses, and the critical reaches for water quality attainment identified:
1. Upper lobe of Upper Mystic Lake (Upper Lobe, critical).
2. Main body of Upper Mystic Lake (Upper Lake, critical).
3. Lower Mystic Lake (Lower Lake, not critical).
4. Upper part of the Lower Mystic Basin (Upper Basin, critical).
5. Lower part of the Lower Mystic Basin (Lower Basin, critical).
81
-------
Mystic River Watershed TMDL Alternative Development Final Report
VII.B.2. Model Options
The BATHTUB model allows the user to select a number of model options to represent the
receiving water body response to the estimated input loads. The nutrient related options in
BATHTUB include the following:
82
-------
Mystic River Watershed TMDL Alternative Development Final Report
Phosphorus Sedimentation Models
Nitrogen Sedimentation Models
Chlorophyll-a Models
Secchi Depth Models
Longitudinal Dispersion Models
Application of Nutrient Availability Factors
Application of Various Calibration Factors
More information on these options and the different equations used is given in the BATHTUB user
manual (Walker, 2004).
The final choice of model options is primarily a function of the availability of water quality data. For
example, Secchi depth and transparency data were limited in the segments and years modeled, so
those options were dropped. In addition, nitrogen data were sparse and very different from
predicted incoming land-based concentrations, so although nitrogen was included in initial testing of
the model, it was not used in the final BATHTUB setup. The final model options used in this study
are given in Table VII-1.
Table VII-1. Model Options
Model Option
Default Choice
Final Choice
Calibrated
Nitrogen Model
Not computed
Not computed*
-
Phosphorus Model
2nd order, available P
Same as default
No
TN Calibration
Sedimentation rates
Not computed*
No
TP Calibration
Sedimentation rates
Same as default
No
Nutrient Availability Factors
Not included
Not included
No
Chl-a Model
P, light, flushing
Same as default
Yes
Secchi Depth Model
Chi a and turbidity
Not computed
-
Transparency Model
Chl-a, turbidity
Not computed
-
Longitudinal Dispersion
Fischer-Numeric
Same as default
Yes
Internal Loading
Not included
Included
Yes
* See reasons in Section VII.C.3.
VII.B.3. Atmospheric Fluxes
Atmospheric TP loading was considered to have a minor effect on the total watershed load (see
VII.C.10), so we used the BATHTUB default TP value (0.27 lb./ac/yr.) split evenly between organic
and inorganic fractions.
Annual precipitation (PREC, in/yr.) for Logan Airport was derived from the NCEI (2018) daily data
downloaded previously (see Section V.A.1).
Annual lake evaporation (ETw, in/yr.) was more difficult to estimate for the reasons below:
The world-wide adopted FAO Penman-Monteith method is the best method for estimating
potential and lake evaporation but it uses solar radiation that is no longer collected by US
Class I national weather stations
83
-------
Mystic River Watershed TMDL Alternative Development Final Report
Available MassGIS data on potential and lake evaporation derived using similar radiation-
based methods does not apply to the modeling period, and only has monthly averages not a
time series of yearly values.
Methods that do not include solar radiation (like Blaney-Criddle) are usually biased high or
low and would have to be calibrated to a more reliable method.
Since annual lake evaporation is a minor variable in this model that only affects the total flow for
each reach slightly, we used an alternative simpler approach for estimating yearly lake evaporation.
The 1961-2005 annual time series of precipitation and lake evaporation data from the HSPF model
in the Upper/Middle Charles TMDL (2011) was used to create a regression equation (PETW =
41.06 -0.0952+PREC, R2=0.2) to predict annual lake evaporation from precipitation. Although the
R2 value for this relationship is low, it is based on the best available data and evaporation methods
and it does reflect higher lake evaporation for drier years and vice versa (see Figure VII-III
comparison for 2007 to 2017).
Figure VII-III. Annual Precipitation and Lake Evaporation
For the calibration period (2015), the values for annual precipitation and average lake evaporation
were 34.8 and 37.7 in/yr., respectively. In comparison, for the scenario period (2007-2016), the
average values for annual precipitation and lake evaporation were 43.0 and 37.0 in/yr., respectively.
These two time periods represent the calibration and critical period for scenarios (see Sections
VII.C.2 and VIII.D).
VII.B.4. Modeled Land Loads
As described in Section V, cumulative annual flow (in/yr.) and TP loads (lb./yr.) were developed for
each sub-basin and corresponding reach for the period 1992-2017. The annual land-based flows and
loads include contributions from stormwater, baseflow, and CSOs/SSOs. The initial sub-basin
delineation for the Lower Mystic Basin (8.63 mi2) was split into the Upper (6.08 mi2) and Lower
Basins (2.55 mi2) using available elevation, and a spatial understanding of connectivity between
combined sewer areas (CSAs) and the associated receiving waters based on information from
community and DEP knowledge (see Section VI.B).
84
-------
Mystic River Watershed TMDL Alternative Development Final Report
VII.B.5. External Loads
In this section, we estimate the total external loads for the BATHTUB model. The model allows the
input of loads using flow and concentration. We used the BATHTUB "Tributary" feature to input
the combined total load from the upstream reach and the local sub-basin using the scheme laid out
in Figure VII-I. Because land loading is part of the "Tributary" input, explicit export coefficients
were not used for the BATHTUB model.
External load was specified as a flow-weighted average concentration (load/flow) and an annual
streamflow using the Tributary Input option in BATHTUB. The total external load is the sum of the
attenuated upstream reach load plus the unattenuated land load from the local sub-basin (see Table
VII-2). The total segment loads were then converted to an average concentration (i.e. load/flow)
associated with the average annual streamflow for input into the BATHTUB model.
Table VII-2. Upstream Reach and Local Sub-basin Contributions
Segment Name
Reach Name
Sub-basin Name
(attenuated load)
(unattenuated load)
Upper Lobe
Aberjona 2
Upper Lobe
Upper Lake
-
Upper Lake
Lower Lake
-
Lower Lake
Upper Basin
Alewife
Upper Basin
Lower Basin
Maiden
Lower Basin
The BATHTUB model requires the total external contributions to be divided into organic and
inorganic nutrient parts. Because the Mystic River had few nutrient measurements from tributaries,
we used the average measured concentration ratio (60 percent inorganic TP) from the
Upper/Middle Charles Nutrient TMDL (DEP-EPA, 2011). These ratios were used to split the total
concentration into organic and inorganic concentrations. This assumption is justified because the
Charles and Mystic tributary watersheds have very similar land use patterns and potential nutrient
sources.
VII.B.6. Internal Loads
The BATHTUB model can also use available sediment TP release rates for estimating the internal
load instead of adjusting the sedimentation rates. Internal load in river reaches results from nutrient
release from the accumulated organic sediments. Since there have been no direct measurements of
nutrient release rates from the sediments in the Mystic River Watershed, measured release rates from
the Upper/Middle Charles Nutrient TMDL (DEP-EPA, 2011) were used to constrain the initial
model inputs. Average TP release rates from impoundments in the Upper/Middle Charles TMDL
were less than 6 mg/m2/d. TP values were set to a maximum of 6 mg/m2/d during the calibration
process according to the presence of soft sediments (i.e., not actual sediment release rates) as
detected by field monitoring by EPA (2018). Reaches with large areas of sediments were assigned a
high value (6 mg/m2/d) while those with no sediments or no data (e.g. Upper Lake and Upper
Lobe) were assigned a low value (1 mg/m2/d). As part of the calibration process, these internal
loads were adjusted with the expected range (0-6 mg/m2/d) to better match observed water body
TP concentrations.
85
-------
Mystic River Watershed TMDL Alternative Development Final Report
VII.C. Model Calibration and Validation
VII.C.1. Water Quality Data Availability
The average water quality data available for the modeled five segments in all calibration years and
validation years are summarized in the following table. The unaveraged values for water quality data
for the modeled segments are given in Appendix H. All values for TP concentration were previously
adjusted to correct for differences in laboratory methods (see Section V.D.2).
Table VII-3. Average Water Quality Data Available by Segment, Site and Year
Year
Segment
WQ Site IDs
AvgTP
Avg Chl-a
2010
Upper Lobe
UPLUPL
2010
Upper Lake
UPLCTR / UPL001
0.029
2010
Lower Lake
MYR071
0.036
2010
Upper Basin
MWRA083 / MWRA066
0.043
7.086
2010
Lower Basin
MYR33 / MAR003 / MWRA167
0.054
14.582
2014
Upper Lobe
UPLUPL
2014
Upper Lake
UPLCTR / UPL001
0.038
2014
Lower Lake
MYR071
0.034
2014
Upper Basin
MWRA083 / MWRA066
0.045
8.701
2015
Upper Lobe
MYR33 / MAR003 / MWRA167
0.052
16.790
2015
Upper Lake
UPLCTR / UPL001
0.032
8.929
2015
Lower Lake
MYR071
0.036
4.749*
2015
Upper Basin
MYR43
0.056
17.944
2015
Lower Basin
MYR33 / MAR003 / MWRA167
0.059
23.534
2016
Upper Lobe
UPLUPL
0.060
17.103
2016
Upper Lake
UPLCTR / UPL001
0.029
8.709
2016
Lower Lake
MYR071
0.036
2016
Upper Basin
MYR43
0.072
21.273
2016
Lower Basin
MYR33 / MAR003 / MWRA167
0.089
30.636
2017
Upper Lobe
UPLUPL
0.053
13.351
2017
Upper Lake
UPLCTR / UPL001
0.036
13.257
2017
Lower Lake
MYR071
0.038
2017
Upper Basin
MYR43
0.062
22.476
2017
Lower Basin
MYR33 / MAR003 / MWRA167
0.066
27.264
* Estimated from adjacent segments and other variables
VII.C.2. Calibration and Validation Periods
The calibration period was determined by the availability of good quality and representative instream
data in the critical receiving water bodies. Originally, the period 2015-2017 was recommended for
calibration. However, 2015 was found to be a better choice because it is the only common period
with good quality data for all five segments that have typical water quality without the interference
from the macrophytes herbicide treatments in 2016 and 2017. Although this year is drier than
average (see Figure VII-III), it serves as a critical-conditions period because the dry, sunny
conditions provide more ideal conditions for phytoplankton growth due to increased direct sunlight,
86
-------
Mystic River Watershed TMDL Alternative Development Final Report
higher temperatures, and lower flows (increased residence times). Since one of the chl-a attainment
targets was the 90th percentile, it is important to capture optimal growth conditions for model
calibration. For these reasons, 2015 was used as the calibration period for the joint calibration of all
segments.
VII.C.3. Receiving Water Parameters
The BATHTUB model requires the input of the following physical parameters for each receiving
water body: surface area, mean depth, length, epilimnion depth, hypolimnion depth, and non-algal
turbidity. Since most ponds in Massachusetts do not stratify if less than 10 ft deep, epilimnion depth
was estimated as 10 ft or the actual depth if less. Any remaining depth was assigned to the
hypolimnion or set to zero if fully assigned to the epilimnion. Non-algal turbidity is an inverse
measure of Secchi depth that represents the portion of light extinction that is due to factors other
than algae (inorganic suspended solids, color). A value of 1 per 6 ft was used since Secchi depth
measurements for the non-algal season in the segments and years modeled were not available. The
value of 6 feet was an average determined for the non-algal season (outside of May-Sep) in non-
modeled river segments for the years modeled.
VII.C.4. Dispersion
The longitudinal dispersion rate was calculated in the model using the BATHTUB model default
dispersion method (Fischer et.al., 1979; adapted by Walker, 1985). Longitudinal dispersion is a result
of the mixing effect among adjacent segments, both between reaches and within reaches if there are
sub-reaches. High dispersion rates create high mixing conditions and a low range of the values
among segments whereas low dispersion rates create low mixing conditions with the segments all
having distinctly different values.
Observed TP data showed more of a range of data values among the segments than the default
model predicted. Therefore, the longitudinal dispersion rate was calibrated by using a multiplier of
0.2. This approach gave more differentiation of the TP values among the segments than using the
default multiplier of 1.0, allowing the modeled TP (see Figure VII-VFigure VI-V) to better reflect
the U-shaped pattern of observed TP values (see Figure VII-VI).
87
-------
Mystic River Watershed TMDL Alternative Development Final Report
70
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
ฆ Calibrated (Dispersion*0.2) Default (Dispersion*!.0)
Figure VII-IV. Dispersion Effect on Modeled Phosphorus Concentrations
VII.C.5. Nutrient Availability Factors
The BATHTUB model allows the use of organic and inorganic fractions to drive the nutrient
sedimentation model through the use of a Nutrient Availability Factor (see Table VII-1). Organic
phosphorus (OP) controls sedimentation and can influence the predicted TP values allowing more
options for calibration. The OP fraction of TP was set to 40-50% percent for inflow to all segments
based on observed data from the Upper Charles TMDL (DEP-EPA, 2011). Sensitivity trials that
varied this ratio resulted in no model prediction improvement. Ultimately, this model option was not
used in the final calibration run.
VII.C.6. Internal Loads
The TP predictions in the five modeled segments were generally lower than measured values
indicating that there is some internal loading of TP from the sediments. Average TP release rates
from impoundments in the Upper/Middle Charles TMDL averaged less than 19.5 lb./ac/yr. for
aerobic and anaerobic conditions, respectively. (DEP-EPA, 2011). The final calibrated values for the
BATHTUB model were set to 19.5, 3.3, 3.3, 13.0 and 13.0 lb./ac/yr. for segments 1 to 5,
respectively (see Section VII.B.l for names).
VII.C.7. Chl-a Model
Predicted chl-a values using the default chl-a method (P, light, flushing) in BATHTUB model were
always too high (16-58% high with the percent errors larger downstream) even though TP was being
accurately predicted.
The linear chl-a method in BATHTUB (P only) worked slightly better. That approach is consistent
with the chl-a versus TP regressions developed in Table IV-3 (R2 = 0.12 to 0.43).
88
-------
Mystic River Watershed TMDL Alternative Development Final Report
In the final model, we used a chl-a calibration factor. The best calibration for chl-a used the default
chl-a model with a global calibration factor of 0.6 applied to all segments. An atypical relationship
between TP and chl-a in ponds can be the result of light limitation (Filstrup and Downing, 2017).
Calibration factors provide a means for adjusting model predictions to account for site-specific
conditions. These modify the coefficients of the empirical models within the BATHTUB model.
They are usually set to 1.0 and should be modified only with extreme caution and site-specific data.
We justify the use of a calibration coefficient based on some evidence of light limitation the Mystic
River. One possible explanation is that the natural tan color of the Mystic River limits light
penetration into the water column. Another explanation is that the submerged and floating aquatic
plants (see Section II.J) create a similar light penetration issue.
VII.C.8. Model Calibration
In summary, the default BATHTUB model was calibrated to the Mystic River conditions using the
following sequential approach:
Changing the default multiplier for longitudinal dispersion from 1.0 to 0.2
Adding internal TP loading to each modeled segment, and
Changing the default multiplier for the chl-a model from 1.0 to 0.6.
The 2015 data were used to calibrate the model since it had the most available water quality data and
that year was unaffected by transient nutrient and chlorophyll-a changes from herbicide applications
in 2016 and 2017 used for macrophyte control in the Lower Basin. Despite this good set of data, the
chl-a value for the Lower Lake had to be estimated from adjacent segments and other variables.
Since this segment is not a critical reach, it is not a critical value.
A total of 22 calibration runs were performed and tracked both visually and for computed goodness-
of-fit parameters for TP and chl-a. A high correlation coefficient (R2) represents a strong linear
relationship between predicted and observed values while a low root-mean square error (RMSE,
Mg/L) indicates good overall fit to the observed data. The best calibration run was a compromise
between TP and chl-a runs with the goal of having similar R2 values for both parameters. The final
calibration run had an R2 and RMSE for TP of 0.86 and 3.8 (Jg/L, and an R2 and RMSE for chl-a of
0.84 and 1.5 (Jg/L, respectively. Excluding the estimated chl-a value for Lower Lake (not a critical
segment) slightly improved the model fit for chl-a.
Predicted versus observed values of TP are given in Figure VII-V and Figure VII-VI. Similar plots
are for chl-a are given in Figure VII-VII and Figure VII-VIII. The error bars in these plots are +/-
one standard deviation.
Table VII-4 gives a breakdown of the mass balance for TP in each modeled segment excluding
transfers from one model segment to the next. Since the focus of this mass balance is the five
modeled segments of the BATHTUB model, we report attenuated land loads in this table. A pie
chart of total loads to the Mystic River Watershed for the calibration period is given in Figure
VII-IX.
A complete set of the BATHTUB inputs is provided in Appendix I.
89
-------
Mystic River Watershed TMDL Alternative Development Final Report
80
^ 70
3 60
ฃ
O
+3 50
ro
c
OJ
u
c
o
u
O.
40
30
20
10
i
1
i
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
Predicted Observed
Figure Vll-V. Predicted vs. Observed TP by Segment for Calibration Period
Observed TP (|ig/L)
Predicted 1:1 Line
Figure VII-VI. Predicted vs. Observed TP Relationship for Calibration Period
90
-------
Mystic River Watershed TMDL Alternative Development Final Report
35
^ 30
3
c
o
5
to
ฃ 20
0)
u
c
ฐ 1c
U 15
25
to
IE
u
10
I
(
I
)
t t
c
J
Upper Lobe Upper Lake Lower Lake Upper Basin
Predicted C Observed
Lower Basin
Figure VII-VII. Predicted vs. Observed Chl-a by Segment for Calibration Period
25
bo
3
to
20
-a
a>
15
T3
0)
10
10 15
Observed Chl-a (|ig/L)
20
25
Predicted
1:1 Line
Figure VII-VIII. Predicted vs. Observed Chl-a Relationship for Calibration Period
91
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table VII-4. Total Phosphorus Loads by Segment for Calibration Period
Name
External Load*
(lb./yr.)
Internal Load
(lb./yr.)
Atmospheric Load
(lb./yr.)
Total Load
(lb./yr.)
Upper Lobe*
2734.6
676.5
9.3
3420.4
Upper Lake
218.3
459.1
37.7
715.1
Lower Lake
1000.8
306.0
25.1
1332.0
Upper Basin
2250.2
740.9
15.2
3006.4
Lower Basin
2937.1
1610.8
33.1
4580.9
Mystic River
9141.0
3793.3
120.4
13054.8
Name
External Load
(%)
Internal Load
(%)
Atmospheric Load
(%)
Total Load
(%)
Upper Lobe*
80.0
19.8
0.3
100.0
Upper Lake
30.5
64.2
5.3
100.0
Lower Lake
75.1
23.0
1.9
100.0
Upper Basin
74.8
24.6
0.5
100.0
Lower Basin
64.1
35.2
0.7
100.0
Mystic River
70.0
29.1
0.9
100.0
* External load includes input from Upper Mystic/Aberjona River, other segments exclude load transfers
92
-------
Mystic River Watershed TMDL Alternative Development Final Report
Loads are in lb./yr. External load - stormwater + groundwater +_CSO/SSO load
Figure VII-IX. Calibration 2015 - Total Phosphorus Loads for Mystic River
93
-------
Mystic River Watershed TMDL Alternative Development Final Report
VII.C.9. Model Validation
The 2016 and 2017 years were used to validate the BATHTUB that was model calibrated to the
2015 data. These years were excluded from the calibration because of herbicide treatment of
macrophytes in the Upper/Lower Basin segments of the model. These treatments generally kill the
vegetation, release nutrients into the water column, open up the water surface to allow greater light
penetration, and result in high algal growth (see Section II.J). According to MyRWA (pers. comm.,
2019), the treatment in 2016 was a contact herbicide that has rapid results, while the treatment in
2017 was a systemic herbicide that is slower acting but more prolonged. In the light of this
information, we expected the predicted TP and chl-a values to be lower than the observed results
for the Upper and Lower Basin, with more of a difference in 2016 than 2017.
In 2016, the predicted values were similar to observed values in the upper segments but lower in the
Upper and Lower Basin segments. Differences were in the range of 13-23 (jg and 7-13 (Jg/L for TP
and chl-a, respectively. (Figure VII-VIII and Figure VII-IX). In 2017, the predicted values were
higher that observed values in the most segments but much larger in the Upper and Lower Basin
than in 2016, in the range of 9-13 (Jg/L for TP and 8-11 (Jg/L for chl-a, respectively (Figure VII-X
and Figure VII-XI). This was expected because the systemic herbicide used in 2017 acts more slowly
than the contact herbicide used in 2016. The reason for under prediction of TP and chl-a for the
upper model segments in 2017 is unclear.
The 2010 and 2014 years were used to validate the outlier replacement technique used to develop
the annual flow and loads for SSOs (see Section VI.C). These two years had significant occurrences
and volumes of SSOs occur throughout the Mystic River Watershed. The BATHTUB was run in
two modes, with and without SSO outlier replacement, to test which outlier approach gave predicted
results closer to the observed TP values. In 2010, the predicted TP values using outlier replacement
closely matched the observed values (-1.1 to 3.9 (Jg/L), while when outliers were included gave
higher than observed TP values (3.1 to 8.1 (Jg/L) (Figure VII-12). In 2014, the results were similar
but had more spread. Predicted TP values using outlier replacement matched the observed values on
average (-5.8 to 8.3 ug/L), while with outliers included gave higher than observed TP values (-1.9 to
14.7 (Jg/L) (Figure VII-XIII).
These 2016 and 2017 validation runs confirm that the calibrated model was able to perform as
expected for those years. The 2010 and 2014 validation runs confirm that the outlier replacement
approach used for SSOs was not an inappropriate approach.
94
-------
Mystic River Watershed TMDL Alternative Development Final Report
00
3
C
o
5
to
100
90
80
70
60
= 50
0)
u
ฃ 40
o
u
O. 30
ฐ I s
o
20
10
0
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
Predicted O Observed
Figure Vll-X. Validation 2016 - Predicted vs. Observed TP by Segment
35
^ 30
M
2.
"c 25
.ฐ
ฆฃ
ฃ 20
4->
c
CD ฆ
u
C 15
O
u
to
10
0
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
Predicted Observed
Figure VII-XI. Validation 2016 - Predicted versus Observed Chl-a by Segment
95
-------
Mystic River Watershed TMDL Alternative Development Final Report
70
^ 60
H.
"c 50
o
to
ฃ 40
C
0)
C
o 30
u
O.
I_ 20
10
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
Predicted O Observed
Figure VII-XII. Validation 2017 - Predicted vs. Observed TP by Segment
30
< 25
M
3
c
O 20
'+ฆป
ro
4-ป
c
u 15
ง
u
? 10
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
Predicted C Observed
Figure VII-XIII. Validation 2017 - Predicted vs. Observed Chl-a by Segment
96
-------
Mystic River Watershed TMDL Alternative Development Final Report
70
T 60
3
c
o
5
to
50
ฃ 40
c
0)
u
c
O 30
u
O.
I
20
10
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
ฉObserved Predicted (outliers replaced) Predicted (with outliers)
Figure VII-XIV. SSO Outlier Replacement 2010 - Predicted versus Observed TP
70
T 60
3
c
o
to
50
i: 40
ฃ
0)
u
o 30
u
Q.
I
20
10
9
Upper Lobe Upper Lake Lower Lake Upper Basin Lower Basin
ฉObserved Predicted (outliers replaced) Predicted (with outliers)
Figure VII-XV. SSO Outlier Replacement 2014 - Predicted versus Observed TP
97
-------
Mystic River Watershed TMDL Alternative Development Final Report
VII.C.10. Model Sensitivity Analysis
A model sensitivity analysis was performed on the calibrated BATHTUB model by varying six
important variables (model parameters and TP input loads). Although BATHTUB has a built-in
sensitivity analysis tool, we chose not to use it because it only allows analysis of two parameters, it
does not allow the variation of input loads, and it does not have a consistent sensitivity range for all
variables to allow cross-comparison of sensitivity among all items.
The sensitivity analysis that we performed multiplied the values by 50% or 200% of the original
calibrated values of the following parameters/loads. Note that only the first two parameters (*) are
included in the built-in tool in BATHTUB.
Longitudinal dispersion*
Sedimentation rate of TP+
Atmospheric deposition TP load
Internal/sediment load of TP
Segment TP load (by varying concentration)
Segment TP load (by varying flow)
To conduct this sensitivity analysis, two additional runs were necessary for each calibrated variable,
performed by multiplying the test value by 50% or 200%. For each parameter/load, the output TP
and chl-a values were recorded for the calibrated run and the two sensitivity runs. A normalized
sensitivity range was calculated using the following formula:
Sensitivity Factor = (Value 200% Value 50%) / Value calibrated
Table VII-5 gives the results of the sensitivity analysis. The table lists the predicted TP and chl-a
results for the three runs for each parameter/load and also shows the sensitivity factor both
numerically and graphically.
From these results, we can conclude the sensitivity for tested parameters/loads was the following:
Longitudinal dispersion low
Sedimentation rate of TP high
Atmospheric deposition TP load very low
Internal/sediment load of TP - moderate
Segment TP load (by varying concentration) - high
Segment TP load (by varying flow) - moderate
This analysis provides confirmation that several professional judgements made in calibrating the
BATHTUB model consistently chose the correct approach:
Use of default values for atmospheric deposition loads
Calibration of longitudinal dispersion before internal load
Use of internal load instead of sedimentation rate to adjust segment TP
Not calibrating segment input flow to observed gauge values because of data ambiguity
The result for varying segment input TP load by concentration versus flow is worth additional
discussion. Lowering the flow, and consequently the load, has a muted response on lowering TP and
98
-------
Mystic River Watershed TMDL Alternative Development Final Report
chl-a. This outcome is likely because flow has a negative feedback effect via residence time. Even
though the load is lower, the lower flow results in a higher residence time in each segment, such that
the TP/chl-a concentrations do not decrease as much compared to same change in load by lowering
the input concentrations.
Table VII-5. Results of BATHTUB Sensitivity Analysis
Longitudinal Dispersion
50%
100%
200%
50%
100%
200%
Sen sitivity
Sensitivity
Sensitivity
Sensitivity
Segment
TP
TP
TP
Chl-a
Chl-a
Chl-a
TP
Chl-a
TP
Chl-a
Upper Lobe
57.6
49.2
42.9
19.0
16.2
14.0
-29.9%
-30.3%
1
Upper Lake
33.1
33.4
33.7
7.9
7.9
8.0
1.7%
1,6%
Lower Lake
32.5
32.6
35.4
7.7
7.7
8.4
9.1%
8,9%
1
Upper Basin
49.6
49.7
47.9
18.0
18.1
17.4
-3,4%
-3.7%
1
Lower Basin
66.1
66.1
65.0
22.7
22.7
22.3
-1,7%
-1.6%
1
Sedimentation Rate
50%
100%
20 D%
50%
100%
200%
Sen sitivity
Sensitivity
Sensitivity
Sensitivity
Segment
TP
TP
TP
Chl-a
Chl-a
Chl-a
TP
Chl-a
TP
Chl-a
Upper Lobe
54.1
49.2
44.5
17.8
16.2
14.6
-19.4%
-19.8%
1
Upper Lake
42.4
33.4
25.4
9.9
7.9
6.0
-50.8%
-49.3%
ฆ
Lower Lake
42.4
32.6
24.3
9.9
7.7
5.7
-55.7%
-54.3%
ฆ
Upper Basin
56.7
49.7
43.3
20.8
18.1
15.5
-26.9%
-29.1%
I
Lower Basin
73.7
66.1
58.1
25.1
22.7
20.0
-23.6%
-22.296
1
Atmos Deposition
50%
100%
200%
50%
100%
200%
Sensitivity
Sensitivity
Sensitivity
Sensitivity
Segment
TP
TP
TP
Chl-a
Chl-a
Chl-a
TP
Chl-a
TP
Chl-a
Upper Lobe
49.1
49.2
49.4
16.2
16.2
16.3
0.6%
0,6%
Upper Lake
33.3
33.4
33.6
7.9
7.9
8.0
1.1%
1,1%
Lower Lake
32.4
32.6
32.9
7.7
7.7
7.8
1.4%
1,4%
Upper Basin
49.5
49.7
50.0
18.0
18.1
18.2
1.096
1,1%
Lower Basin
65.9
66.1
66.5
22.6
22.7
22.9
1.096
0,9%
Internal Load
50%
100%
200%
50%
100%
200%
Sen sitivity
Sensitivity
Sensitivity
Sensitivity
Segment
TP
TP
TP
Chl-a
Chl-a
Chl-a
TP
Chl-a
TP
Chl-a
Upper Lobe
44.6
49.2
58.1
14.6
16.2
19.1
27.5%
27.7%
1
1
Upper Lake
30.4
33.4
38.9
7.2
7.9
9.2
25.5%
24.5%
1
1
Lower Lake
30.0
32.6
37.2
7.1
7.7
8.8
21.8%
21.2%
1
1
Upper Basin
44.7
49.7
59.4
16.1
18.1
21.8
29.6%
31.7%
1
ฆ
Lower Basin
57.7
66.1
82.3
19.9
22.7
27.6
37.3%
34,096
ฆ
ฆ
Segment Load (cone)
50%
100%
20 D%
50%
100%
200%
Sen sitivity
Sensitivity
Sensitivity
Sensitivity
Segment
TP
TP
TP
Chl-a
Chl-a
Chl-a
TP
Chl-a
TP
Chl-a
Upper Lobe
32.2
49.2
80.7
10.2
16.2
25.6
98.6%
95,096
Upper Lake
25.1
33.4
46.3
5.9
7.9
10.7
63.3%
59.9%
ฆ
Lower Lake
24.7
32.6
44.8
5.8
7.7
10.4
61.7%
59.1%
ฆ
Upper Basin
34.2
49.7
77.7
11.8
18.1
28.3
87.6%
91.5%
m
Lower Basin
46.4
66.1
101.4
15.9
22.7
32.6
83.36
73.7%
m
Segment Load (flow)
50%
100%
200%
50%
100%
200%
Sensitivity
Sensitivity
Sensitivity
Sensitivity
Segment
TP
TP
TP
Chl-a
Chl-a
Chl-a
TP
Chl-a
TP
Chl-a
Upper Lobe
54.6
49.2
47.7
18.4
16.2
15.1
-14.1%
-20.3%
1
1
Upper Lake
32.8
33.4
36.3
8.0
7.9
8.1
10.3%
1.3%
1
Lower Lake
31.0
32.6
36.5
7.5
7.7
8.2
17.096
7,9%
1
1
Upper Basin
54.8
49.7
49.2
20.4
18.1
17.3
-11.296
-17.1%
1
1
Lower Basin
76.1
66.1
62.4
26.5
22.7
20.6
-20.8%
-25.9%
1
1
99
-------
Mystic River Watershed TMDL Alternative Development Final Report
VIII. Critical Period of Interest for Phosphorus Load
Reduction Analysis
This section documents the approach for evaluating and selecting the critical period of interest for
the phosphorus load reduction analyses in the BATHTUB model discussed in Section IX. The
approach involved evaluating available water quality data availability; identifying events or actions
that may influence the water quality data; and identifying a range of representative climatic
conditions.
VIII.A. Water Quality Data
As discussed in Section III and noted in Table III-2 water quality monitoring data was available
between 2000 and 2017. Six water bodies (located primarily in the upper watershed) were sampled
more intensively in 2015 through 2017. The review of eutrophic related parameters, Chlorophyll-a
(chl-a) and TP, including statistical analyses completed in Section IV indicated that there has been a
gradual decline in chl-a and TP concentrations in the Mystic River. This appears to be primarily due
to the growth of macrophytes, particularly in the last 10 years, which appear to be removing
phosphorus from the water column and/or suppressing phytoplankton growth due to reduced light
availability. The only exceptions to this trend are the instant releases caused by the use of herbicide
applications in 2016 and 2017.
Based on these considerations, data between 2000 and 2015 was considered to be most
representative of the available water quality datasets.
VIII.B. Combined Sewer and Sanitary Sewer Overflow Data
As noted in Table VII-1 in Section VII, CSO and SSO data was available for the years 2000 to 2017,
with the exceptions noted for missing 2006 CSO data in the Mystic River basin and missing SSO
event data.
VIII.C. Rainfall Data
Annual rainfall data was available from NCEI (2018). As noted in Section VII, the annual lake
evaporation was not able to be directly estimated, so the 1961-2005 annual time series of
precipitation and water surface evaporation data (PETW) from the HSPF model in the
Upper/Middle Charles TMDL (2011) was used. Evaluations of the rainfall data and potential
evapotranspiration were reported in inches per year and meters per year.
A review of the annual precipitation conditions was conducted using the standardized precipitation
index, which is a statistical method for assessing rainfall (SPI; McKee, 2993). The SPI normalizes the
data to provide a better understanding of whether a year was wet (positive SPI values, greater than
average precipitation) or dry (negative SPI values, less than average precipitation). A summary table
of the SPI values and the representative condition is provided below in Figure VIII-I.
Table VIII-1. The complete summary of the rainfall data between 1990 and 2017 is provided in
Table VIII-2 SPI between 2000 and 2017 is shown in Figure VIII-I.
100
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table VIII-1. Standard Precipitation Index (SPI) Reference Values
SPI Value
Drought/Wetness Condition
2 and above
Extremely wet
1.5 to 1.99
Severely wet
1.0 to 1.49
Moderately wet
-0.99 to 0.99
Near normal
-1.0 to-1.49
Moderately dry
-1.5 to-1.99
Severely dry
-2.0 and less
Extremely dry
Table VIII-2. Summary of Rainfall Data Analyses
Year Precip PETW SPI value Condition
(in/yr.) (in/yr.)
1990
46.5
36.6
0.5
Near normal
1991
42.3
37.0
-0.1
Near normal
1992
43.7
36.9
0.1
Near normal
1993
43.2
36.9
0.0
Near normal
1994
47.6
36.5
0.7
Near normal
1995
35.1
37.7
-1.2
Moderately dry
1996
48.7
36.4
0.9
Moderately wet
1997
28.3
38.4
-2.2
Severely dry
1998
51.3
36.2
1.3
Very wet
1999
37.8
37.5
-0.8
Near normal
2000
45.6
36.7
0.4
Near normal
2001
30.8
38.1
-1.8
Severely dry
2002
41.1
37.1
-0.3
Near normal
2003
44.4
36.8
0.2
Near normal
2004
44.6
36.8
0.2
Near normal
2005
43.7
36.9
0.1
Near normal
2006
52.9
36.0
1.4
Moderately wet
2007
39.5
37.3
-0.5
Near normal
2008
54.5
35.9
1.7
Very wet
2009
43.5
36.9
0.0
Near normal
2010
49.7
36.3
1.0
Moderately wet
2011
52.4
36.1
1.4
Moderately wet
2012
36.8
37.6
-1.0
Moderately dry
2013
40.4
37.2
-0.4
Near normal
2014
45.3
36.7
0.3
Near normal
2015
34.8
37.7
-1.2
Moderately dry
2016
33.1
37.9
-1.5
Severely dry
2017
43.5
36.9
0.0
Near normal
101
-------
My stic River Watershed TMDL Alternative Development Final Report
5PI
ฆ2.0 -1.8
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Figure Vlll-I. Standard Precipitation Index (SPI) between 2000-2017
VIII.D. Critical Period Selection
The critical period of interest for the phosphorus load reduction analyses is intended to be
representative: of critical climatic conditions related to the water quality endpoints (chl-a and TP),
which may lead to excessive algal growth and cyanobacteria blooms in the Mystic River. Based on
discussions with the technical team, a minimum of a 10-year period was recommended in order to
capture the critical conditions that could lead to eutrophication. Initially, the period of 2008 to 2017
was identified to utilize the last 10-years of meteorological data and most recent water quality
conditions in the water bodies. However, 2017 SSO data from MassDEP was not available at the
time, so the critical period of 2007 to 2016 was selected for further review.
During the critical period, there was one very wet year (2008), two moderately wet years (2010,
2011), two moderately dry years (2012, 2015) and one severely dry year (2016). An average analysis
of the rainfall and potential evapotranspiration are 43.0 and 37.0 inches/year, respectively. In
comparison to the period of record (2000 to 2017), the average annual rainfall and potential
evapotranspiration are 43.1 and 37.0, respectively. Overall, 2007 to 2016 was selected as the
representative climatic period for further phosphorus load reduction analyses with the calibrated
BATHTUB model.
VIII.E. Extreme Rainfall Years
An additional evaluation of the data was also done to identify extreme rainfall years that may be used
to compare against average annual data during the selected critical period. This included looking at
both SPI values and 10th and 90th rainfall depth percentiles, which were used in the Upper/Middle
and Lower Charles River TMDLs to: evaluate the potential for exceedances for water quality and the
margin of safety. During the critical period, the 10th percentile rainfall depth is 34.1 inches, while the
90th percentile rainfall depth is 52.6 inches. Comparatively, over a longer term, from 1990-2016, the
10th percentile rainfall depth is 34.4 inches and the 90th percentile rainfall depth is 51.6 inches.
Comparing these values to Figure VIII-I above, 2008 would be identified as an extreme wet year,
while 2016 would be considered an extreme dry year.
102
-------
Mystic River Watershed TMDL Alternative Development Final Report
IX. Evaluation of Watershed Phosphorus Load Reduction
Analysis
This section documents the BATHTUB modeling methodology, loading estimates and the scenarios
developed to evaluate the watershed phosphorus loading reductions necessary to address
eutrophication within the Mystic River Watershed and attain water quality target (chl-a) and
secondary indicatory (TP) identified in Section IV. The calibrated BATHTUB model discussed in
Section VII was used for this project; the critical period of interest was identified in Section VIII for
the phosphorus load reduction analysis, which is the 10-year period from 2007 to 2016. Our analysis,
as described below, suggests that CSO and SSO management Ent, while important, have far less
impact on annual phosphorus loads compared to stormwater.
IX. A. Phosphorus Loading Estimates for Critical Period of Interest
Five different types of load estimates were developed for each reach and corresponding tributary as
shown in Figure VII-II:
Stormwater loads.
Groundwater loads.
Sediment nutrient efflux loads.
Combined sewer overflow loads.
Sanitary sewer overflow loads.
IX.A. 1. Existing Conditions
The watershed phosphorus loading estimates from Section V were averaged for the 10-year period
from 2007 to 2016 and summarized by the five main segments (reaches) in the BATHTUB model:
Upper Lobe, Upper Lake, Lower Lake, Upper Mystic Basin and Lower Mystic Basin. This includes
both attenuated loads in the reaches and unattenuated loads in the tributaries. No adjustments were
made to attenuation factors or to the average loads for the "existing conditions" scenario.
The existing conditions loads for the critical period of interest were summarized to determine the
relative influence of each on the system. The analysis also included atmospheric loads, which are
included in the BATHTUB model (see Section VII).
IX.A.2. Future Conditions
In discussions with the TSC, further evaluation of phosphorus loadings under future conditions was
requested in order to account for the ongoing work being done in the watershed to address both
CSO and SSO overflows to downstream water bodies. The ERG team consulted with EPA,
MassDEP, and MWRA to determine what expected changes should be reasonably applied to the
loads under future conditions. The following are the assumptions used for the future condition
scenarios discussed in Section IX.B. Refer to Section VI for further detailed discussion on the
development of the CSO and SSO volumes and loads.
CSO Loads
In BATHTUB, the CSO loads and flows from the Cambridge combined sewer areas and parts of
Somerville combined sewer areas contribute to Alewife Brook tributary, the CSO loads and flows
from the rest of the Somerville's combined sewer area contributes to the Lower Basin segment. No
103
-------
Mystic River Watershed TMDL Alternative Development Final Report
CSO flows go to the Upper Basin segment directly except via Alewife Brook. The annual CSO
Loads used for the calibrated BATHTUB model were developed using modeled CSO volumes from
MWRA's annual reports and a representative TP concentration of 3.1 mg/L, as discussed in Section
VI.A.4. Under future conditions, the CSO volumes are assumed to meet the long-term control plan
(LTCP) estimates for the typical rainfall year. As noted under MWRA's 2017 annual reporting of
CSO discharge estimates (MWRA, 2018), the annual LTCP CSO volumes for the Alewife Brook and
Upper Mystic River are 7.29 MG and 3.48 MG, respectively. To calculate the phosphorus loads,
volumes were multiplied by the representative TP concentration.
SSO Loads
Similar to CSO Loads, SSO loads were developed by using estimated volumes obtained from
MWRA and MassDEP and representative TP concentration of 5.23 mg/L, as discussed in Section
VI.A.4. Under future conditions, a 50 percent reduction in SSO volumes is assumed for ongoing
SSO mitigation work being done within the Mystic River Watershed. Volumes were multiplied by
the representative TP concentration to determine the phosphorus loads.
Stormwater Loads from Combined Sewer Separation
In addition to CSO reductions, EPA also reached out to the cities of Cambridge and Somerville to
determine potential for combined sewer separation (CSS) within areas still connected to CSOs at
Alewife Brook and the Upper Mystic River (above the Amelia Earhart Dam). As of November 2018,
the City of Cambridge has no immediate plans to evaluate or complete CSS and the City of
Somerville is in the process of completing an alternative analysis to determine if CSS is a viable
option. Upon discussion with the Technical Steering Committee, the ERG team developed three
future conditions loading estimates: 0 percent CSS, 25 percent CSS, and 100 percent CSS. The 100
percent CSS condition is intended to provide a bookend for the maximum phosphorus load
reductions that may be required and is not an expected outcome of ongoing or future efforts by the
cities of Cambridge and Somerville.
Modeling of CSS for BATHTUB translates to adding the combined sewer land area added to the
total acreage for each sub-basin HRU (discussed in Section V), which generate extra stormwater and
groundwater loads and annual flow volumes. The combined sewer drainage areas GIS data were
provided by the cities of Cambridge and Somerville (refer to Figure G-XI-1 through Figure G-XI-3
in Appendix G). The modeled land uses within the CSO drainage areas were extracted by overlaying
the drainage areas and the MassGIS land use and soils data.
It is the ERG team's understanding based on conversations with MassDEP the majority of the CSO
drainage areas draining to outfall SOM007A/MWR205A are treated at the Somerville Marginal
facility and discharging through outfall MWR205, which is located downstream of the Amelia
Earhart Dam. A volumetric comparison of the two outfalls during the critical period indicates that
approximately 90 percent of the volume discharges through outfall MWR205 to the Massachusetts
Bay, while 10 percent is discharging into the Upper Mystic River at SOM 007A/MWR205A.
Consequently, for the purposes of evaluating future conditions, the ERG team has assumed that 10
percent of the land uses that were assumed to be contributing to outfall SOM007A/MWR205A are
discharging to the Mystic River above the dam. Further, because the Somerville Marginal CSO
Facility provides screening and disinfection only, there are no reductions in TP concentrations
assumed for future conditions.
To evaluate the impact of the three CSS future conditions, the percent CSS (0, 25, or 100 percent)
was multiplied by the land uses contributing to either the Alewife Brook tributary or the Lower
104
-------
Mystic River Watershed TMDL Alternative Development Final Report
Basin reach. The land areas for each were added to the HRUs, which were then summarized for the
BATHTUB model. In addition, the CSO loads were reduced according to the CSS future condition
(25 or 100 percent), so that for the scenario with 100 percent CSS, the CSO loads were zeroed.
Other Stormwater Loads
No additional changes to the land uses or stormwater loads were assumed (e.g., new development or
redevelopment of land).
Groundwater Loads
Since groundwater loads are computed from stormwater loads, changes due to the CSS future
conditions results in changes to groundwater loads. The scale of the impact is dependent on the
percent CSS noted above and proportional to the change in stormwater loads.
When stormwater reductions were made in the various scenarios, the groundwater loads were not
changed.
Sediment Nutrient Efflux Loads
Under future conditions the ERG team assumed that there would be a decrease in sediment
nutrients as a result of management and land loads For modeling purposes, the ERG team
estimated that the sediment load is reduced by 50 percent of the estimated stormwater phosphorus
load reduction (e.g., Stormwater TP Load Reductions = 60 percent, Sediment Load Reductions = 30
percent). This methodology is slightly more conservative than the method used for the
Upper/Middle Charles Nutrient TMDL (EPA/DEP, 2011). Under that method, there was a 25
percent sediment nutrient efflux load reduction that could be assumed with the reduction of
phosphorus loads and sediment loads were approximately 50 percent of the total phosphorus load
reduction (including CSO/SSO loads, groundwater, etc.) required to meet the water quality target
(e.g., Total TP Load Reductions = 50 percent, Sediment Load Reductions = 25 percent).
IX.B. Scenarios for Evaluation of Phosphorous Reduction
The ERG team developed four scenarios to evaluate the watershed conditions necessary for the
water quality target to be met: existing conditions (1 scenario) and future conditions (3 scenarios)
(Table IX-1). As identified in Section IXA, future conditions include reductions in CSOs and SSOs,
in addition to evaluating conditions with CSS. Each of these four scenarios were run with a baseline
model (#1 ,2, 3, 4) and a water quality (WQ) target model (# 1A, 2A, 3A, 4A) to provide
comparisons between the results.
105
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table IX-1. Modeled Scenarios for Phosphorus Load Reduction Evaluations
Scenario #
Scenario Name
CSO Load
Reductions3
SSO Load
Reductions'3
CSSC
Stormwater
Load
Reductions
Sediment
Load
Reductions
1
Existing Conditions - Baseline
1A
Existing Conditions - WQ Target
X
X
2
Future Conditions 1 - Baseline
X
X
2A
Future Conditions 1 - WQ Target
X
X
X
X
3
Future Conditions 2 - Baseline
X
X
25%
3A
Future Conditions 2 - WQ Target
X
X
25%
X
X
4
Future Conditions 3 - Baseline
X
X
100%
4A
Future Conditions 3 - WQ Target
X
X
100%
X
X
a) CSO volumes to the Alewife Brook and Mystic River (Lower Basin) reduced to meet the LTCP target.
b) SSO volume reductions at 50 percent across all tributaries and sub basins.
c) CSS percentage indicates percent of combined sewer area that is assumed to be separated. The separated
land uses are added to the sub-basin area and generate additional stormwater flow and loads.
For each of the WQ target scenarios (#1A, 2A, 3A, 4A), multiple model runs were completed to
evaluate the reductions in stormwater loads and sediment efflux necessary to meet the water quality
target in all model segments. In Section IV, chlorophyll-a (chl-a) was identified as the WQ target,
specifically the seasonal average chl-a (<10 (Jg/L). TP of 30 |ag/Lwas used as a secondary indicator.
This analysis focused on the seasonal average chl-a because the regressions of the TP and seasonal
average chl-a concentrations performed in Section IV indicate a linear relationship, such that
attainment of the seasonal average chl-a (10 (Jg/L) would provide attainment of the TP
concentration needed to meet water quality goals. Iterations were conducted for each WQ target
scenario until the seasonal average chl-a was met. Predicted TP concentrations are reported in the
results and was compared to the water quality secondary indicator (30 (Jg/L) for reference.
IX. C. Modeling Methodology
IX.C.1. Model Setup
The BATHTUB model used for the load reduction analysis was calibrated by Dr. Nigel Pickering in
November 2018.
IX.C.2. Model Inputs
The BATHTUB model allows the user to change a number of model inputs. The BATHTUB input
parameters that are adjusted for each segment in the model to achieve the various scenarios include
drainage area, flow, total nitrogen, inorganic nitrogen, total phosphorous, orthophosphate, and the
total phosphorus internal loading rate. The parameters are changed for the total segment, including
both unattenuated loads to the reach (from adjacent sub basin area) and the attenuated loads from
inflowing stream (tributaries). The flow and loads were average annual values calculated over the
critical period from 2007 to 2016. The input data for BATHTUB was converted to metric units (see
Section VII) from the units derived from the loading spreadsheet model (e.g., lb./year, ac-in/yr.,
106
-------
Mystic River Watershed TMDL Alternative Development Final Report
mg/L). None of the calibrated model coefficients, model options, calibration factors, segment
morphometry, atmospheric loads, and global variables were changed for this analysis.
Once the input parameters were changed in the BATHTUB model to reflect the average annual
scenario conditions, the BATHTUB model was run. Model output was exported into an Excel
spreadsheet. The model outputs that were recorded for each scenario run include the average chl-a
and Total P concentrations for each model segment: Upper Lobe (1), Upper Lake (2), Lower Lake
(3), Upper Basin (4), and Lower Basin (5).
IX.C.3. Analysis with Wet and Dry Year Data
The model scenarios described above use the average annual flows and loads calculated over the
critical period (2007-2016). In order to evaluate whether the water quality target would be met
during extreme precipitation, additional model scenarios were developed to evaluate responses to
the water quality target during a wet year and for a dry year. The selection of the years was based on
precipitation and relative drought/wetness condition (as defined by the standard precipitation index
and the 10th and 90th percentile) noted in Section VIII as well as the annual total nutrient loads based
on data from Section V. While 2008 was considered to be the wettest year in the critical period, 2010
had the highest annual loads for total phosphorus and was considered to be moderately wet. 2016
was a severely dry year and had the lowest annual loads for total phosphorus. For extreme
precipitation analyses, 2010 and 2016 were selected for wet year and dry year analyses, respectively.
IX.D. Model Results with Average Annual Data
A total of 29 runs were performed for the water quality target scenarios (Table IX-2). A summary of
the key results is shown in Table IX-3; further detailed inputs and outputs are outlined in Appendix
J. The BATHTUB output spreadsheet files, which are not submitted as part of this memorandum,
can be made available upon request.
The starting point for the stormwater phosphorus load reduction runs was determined based on the
magnitude of reductions that is needed to meet the water quality target without sediment load
reductions, which is approximately in the 70 to 80 percent range. The modeling of the stormwater
phosphorus load reductions process started with reducing the percent reductions for stormwater
each run until the predicted water quality conditions transition from exceeding the target to meeting
it. Once the water quality target was met with only stormwater reductions, the sediment efflux
reductions were incorporated. The sediment efflux was reduced by half of the estimated stormwater
phosphorus load reduction. The iterative process then continued, lessening both the percent
reductions for stormwater loads and sediment efflux (half of the stormwater reductions) until the
water quality target was met. The ending point for the stormwater and sediment efflux reduction
runs was when the water quality target was exceeded for all segments. Text in red font in Table IX-2
indicates the run that met the chl-a water quality target.
Table IX-2. BATHTUB Model Runsawith Average Annual Data
Scenario
Run #
Stormwater
Reduction (%)
Sediment Efflux
Reduction (%)
1
1
0
0
la
2
80
0
3
78
0
107
-------
Mystic River Watershed TMDL Alternative Development Final Report
4
78
39
5
70
35
6
68
34
7
67
33
8
66
33
2
9
0
0
10
80
0
11
73
0
2a
12
73
37
13
64
32
14
62
31
15
61
31
3
16
0
0
17
75
0
18
73
0
3a
19
73
37
20
63
32
21
61
31
22
60
30
4
23
0
0
24
75
0
25
71
0
4a
26
71
36
27
62
31
28
59
29
29
58
28
a) Red font indicates the run that met the water quality target for chl-a.
Table IX-3 is a summary of the key results for each scenario: the base run and the run that met the
water quality target with stormwater and sediment efflux reductions. This table shows the predicted
phosphorus concentration and chl-a concentration. In addition, the model input, TP Load, is shown
for comparison across scenarios. Cells highlighted in grey indicate segments that do not meet the
water quality target. Figure IX-I and Figure IX-II further show the key results by segment for
phosphorus and chl-a concentrations, respectively. More detailed scenario information is available in
Appendix H.
108
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table IX-3. Scenario Results3 with Average Annual Data
Scenario
Parameter
Upper Lobe
Upper
Lower
Upper
Lower
- Run #
Lake
Lake
Basin
Basin
Total P Load (Ib./yr.)
3,907
416
1,558
3,298
3,849
1-1
Predicted P Cone. (ng/L)
43.4
33.1
33.6
46.2
57.7
Predicted chl-a Cone. (ng/L)
13.9
7.6
7.7
16.4
19.4
SW P Load Reduction (%)
67
P Sediment Efflux Reduction (%)
33
la-7
Total P Load (Ib./yr.)
1,490
161
625
1,553
1,608
Predicted P Cone. (ng/L)
20.8
18.6
19.4
25.3
31.1
Predicted chl-a Cone. (ng/L)
5.9
4.1
4.3
8.1
10.0
Total P Load (Ib./yr.)
3,892
410
1,526
2,968
3,707
2-9
Predicted P Cone. (ng/L)
43.3
33.0
33.3
44.2
55.8
Predicted chl-a Cone. (ng/L)
13.9
7.6
7.7
15.6
18.8
SW P Load Reduction (%)
62
P Sediment Efflux Reduction (%)
31
2a -14
Total P Load (Ib./yr.)
1,655
175
663
1,354
1,632
Predicted P Cone. (ng/L)
22.4
19.8
20.4
24.9
31.1
Predicted chl-a Cone. (ng/L)
6.5
4.4
4.6
7.9
9.9
Total P Load (Ib./yr.)
3,892
410
1,526
3,033
3,704
3-16
Predicted P Cone. (ng/L)
43.3
33.0
33.3
44.2
55.6
Predicted chl-a Cone. (ng/L)
13.8
7.6
7.6
15.6
18.7
SW P Load Reduction (%)
61
P Sediment Efflux Reduction (%)
31
3a-21
Total P Load (Ib./yr.)
1,691
178
677
1,393
1,652
Predicted P Cone. (ng/L)
22.7
20.0
20.6
24.9
31.1
Predicted chl-a Cone. (ng/L)
6.6
4.5
4.6
7.9
9.9
Total P Load (Ib./yr.)
3,892
410
1,526
3,227
3,697
4-23
Predicted P Cone. (ng/L)
43.2
32.9
33.2
43.4
54.5
Predicted chl-a Cone. (ng/L)
13.8
7.5
7.6
15.3
18.3
SW P Load Reduction (%)
59
P Sediment Efflux Reduction (%)
29
4a-28
Total P Load (Ib./yr.)
1,763
186
705
1,490
1,680
Predicted P Cone. (ng/L)
23.4
20.4
21.1
25.0
31.1
Predicted chl-a Cone. (ng/L)
6.9
4.6
4.7
7.9
9.9
a) Grey highlighted cell indicates the predicted value does not meet the chl-a target.
109
-------
Mystic River Watershed TMDL Alternative Development Final Report
Predicted P Concentrations with Scenarios using Average Annual Data
ฆ
1-1
i
la -7
2-9
2a-14
*
3-16
I
3a - 21
ซ
4-23
4a-28
Upper Lobe Upper Lake
Lower Lake
Upper Basin Lower Basin
Segment
Figure IX-i. Predicted Phosphorus Concentrations with Average Annual Data
20
Predicted Chl-a Concentrations with Scenarios using Average Annual Data
16
'J- 14
1
c
.2 12
ra
c
01
110
u
ra
j;
U 8
t
1-1
la - 7
2-9
4
2a- 14
3-16
ฆV 8
0)
TJ
ill
9
3a-21
4-23
4a-28
! !
Upper Lobe
Upper Lake Lower Lake Upper Basin
Segment
Lower Basin
Figure IX-II. Predicted Chlorophyll-a Concentrations with Average Annual Data
IX.E. Analysis with Wet and Dry Year Data
An additional eight runs were conducted to evaluate whether the chl-a water quality target would be
met during extreme wet and dry years. The starting point for these runs under each scenario was the
110
-------
Mystic River Watershed TMDL Alternative Development Final Report
run that met the target (refer to Figure IX-II and Figure IX-III). The average annual flow and loads
were replaced by either the wet or dry year annual flow and load values, as appropriate. The same
stormwater reduction and sediment efflux reductions were applied as was used to meet the chl-a
water quality target in the scenario when using the average annual flow and load values. Table IX-4
shows the additional runs that were conducted for this analysis. Text in red font in indicate Table
IX-4 the run that met the chl-a water quality target.
Table IX-4. Model Runs3 with Wet and Dry Year Data
Scenario
Run #
Wet/Dry Year Data
Stormwater
Reduction (%)
Sediment Efflux
Reduction (%)
la
7
Wet (2010)
67
33
la
7
Dry (2016)
67
33
2a
14
Wet (2010)
62
31
2a
14
Dry (2016)
62
31
3a
21
Wet (2010)
61
31
3a
21
Dry (2016)
61
31
4a
28
Wet (2010)
59
29
4a
28
Dry (2016)
59
29
a) Red font indicates the run that met the water quality target for chl-a.
Table IX-5 is a summary of the key results for each run. Cells highlighted in grey indicate segments
that do not meet the water quality target. Figure IX-III and Figure IX-IV further show the key
results by segment for phosphorus and chl-a concentrations, respectively. More detailed scenario
information is available in Appendix H.
Table IX-5. Scenario Results3 with Wet and Dry Year Data
Scenario Parameter Upper Lobe Upper Lower Upper Lower
- Run # Lake Lake Basin Basin
la-7
(Wet)
SW P Load Reduction (%)
67
P Sediment Efflux Reduction (%)
33
Total P Load (Ib./yr.)
2,153
301
1,089
2,758
2,445
Predicted P Cone. (ng/L)
18.9
17.9
19.8
26.2
30.5
Predicted chl-a Cone. (ng/L)
5.2
3.8
4.3
8.3
9.5
la-7
(Dry)
SW P Load Reduction (%)
67
P Sediment Efflux Reduction (%)
33
Total P Load (Ib./yr.)
968
79
353
760
998
Predicted P Cone. (ng/L)
24.9
20.3
20.1
26.3
35
Predicted chl-a Cone. (ng/L)
7.5
4.7
4.6
8.6
11.6
2a -14
(Wet)
SW P Load Reduction (%)
62
P Sediment Efflux Reduction (%)
31
Total P Load (Ib./yr.)
2,328
304
1,053
1,884
2,124
Predicted P Cone. (ng/L)
20.0
18.7
20.2
23.2
27.4
Predicted chl-a Cone. (ng/L)
5.6
4.0
4.4
7.1
8.4
111
-------
Mystic River Watershed TMDL Alternative Development Final Report
2a-14
(Dry)
SW P Load Reduction (%)
62
P Sediment Efflux Reduction (%)
31
Total P Load (Ib./yr.)
1,090
90
398
944
1,177
Predicted P Cone. (ng/L)
26.9
21.4
21.2
29.0
38.4
Predicted chl-a Cone. (ng/L)
8.3
5.0
4.9
9.7
12.9
3a-21
(Wet)
SW P Load Reduction (%)
61
P Sediment Efflux Reduction (%)
31
Total P Load (Ib./yr.)
2,377
310
1,073
1,939
2,155
Predicted P Cone. (ng/L)
20.4
19
20.4
23.3
27.5
Predicted chl-a Cone. (ng/L)
5.7
4.1
4.4
7.2
8.4
3a-21
(Dry)
SW P Load Reduction (%)
61
P Sediment Efflux Reduction (%)
31
Total P Load (Ib./yr.)
1,115
92
407
965
1,186
Predicted P Cone. (ng/L)
27.2
21.6
21.4
28.9
38.3
Predicted chl-a Cone. (ng/L)
8.4
5
5
9.7
12.9
4a-28
(Wet)
SW P Load Reduction (%)
59
P Sediment Efflux Reduction (%)
29
Total P Load (Ib./yr.)
2,474
322
1,112
2,079
2,206
Predicted P Cone. (ng/L)
21
19.5
21
23.4
27.7
Predicted chl-a Cone. (ng/L)
5.9
4.2
4.5
7.2
8.5
4a-28
(Dry)
SW P Load Reduction (%)
59
P Sediment Efflux Reduction (%)
29
Total P Load (Ib./yr.)
1,164
96
425
1,015
1,188
Predicted P Cone. (ng/L)
28.1
22
21.8
28.7
37.9
Predicted chl-a Cone. (ng/L)
8.7
5.1
5.1
9.6
12.7
a) Grey highlighted cell indicates the predicted value does not meet the chl-a target
112
-------
Mystic River Watershed TMDL Alternative Development Final Report
Predicted P Concentrations with Scenarios using Wet and Dry Year Data
1
~
f
ซ
t
ซ
I
ฆ
ซ
Upper Lobe
Upper Lake
Lower Lake
Segment
la - 7 (Wet)
la-7(Dry)
2a -14 (Wet)
2a -14 (Dry)
3a - 21 (Wet)
3a-21 (Dry)
4a - 28 (Wet)
4a - 28 (Dry)
Figure IX-III. Predicted Phosphorus Concentrations with Wet and Dry Year Data
III
Predicted Chl-a Concentrations with Scenarios using Wet and Dry Year Data
=2
a
.
c
,0
* la - / (Wet)
|
:
i 1m
1 9
c
2d 14jWM>
U
2a - 14 (D'v'j
Z
D
5a - 71 (Wft)
1 6
5a 21 (Dry)
|
t
f
-1a - 78 (Werl
a.
f
1
ซ la - JS (Dry)
6
Upper lobe Upper lake
tower take Upper Basin lower Rastrt
5ซgmซnt
Figure IX-IV. Predicted Chlorophyll-a Concentrations with Wet and Dry Year Data
IX.F. Discussion
In each base run for the four scenarios, the Upper Lake and Lower Lake segments meet the chl-a
water quality target without any stormwater and sediment load reductions. The Lower Basin
113
-------
Mystic River Watershed TMDL Alternative Development Final Report
segment requires the most significant reductions to meet the water quality target (i.e., it is the critical
segment in this analysis). Overall, under existing conditions, a 67 percent reduction is required for
stormwater phosphorus loads with a 33 percent reduction in sediment efflux. The stormwater load
reductions required under future conditions (scenarios 2A, 3A and 4A) were between approximately
59 and 62 percent with a 30 to 31 percent required sediment efflux reduction. Table X provides a
total phosphorus load comparison between existing conditions and scenario 2A. Since the focus of
this table is the entire watershed, we report unattenuated land loads here. Under future conditions
scenarios 3A and 4A with CSS, the overall phosphorus loads, and flows increase, though not
proportionally, which appears to be resulting in lower overall phosphorus concentrations and lower
required reductions compared to scenario 2A.
Table IX-6. Total Phosphorus Load Reductions for Scenario 2A
Item
Stormwater
Groundwate
r
cso/sso
Internal
Atmospheri
c
Total
Existing
Conditions
Total P Load
(lb./yr.)
14,887
1,141
1,696
3,793
120
21,638
Scenario 2A
P Load
(lb./yr.
9,974
1,141
412
1,271
120
12,919
Reduction
(%)
67%
0%
24%
34%
0%
60%
For this analysis, stormwater and sediment efflux reductions were made consistently across all
tributaries and sub-basins. However, further analysis can determine what the minimum required
reductions are for the first three segments versus the last two segments (Upper Basin and Lower
Basin) in order to meet the water quality target. For example, in scenario 1 A, run #7, the water
quality target for chl-a is just met in the Lower Basin segment with a concentration of 9.9 (Jg/L.
However, in the Upper Lake and Lower Lake model segments, the concentrations are over half the
water quality target at 4.1 and 4.3 (Jg/L, respectively. It is evident that the Upper Lobe, Upper Lake,
and Lower Lake model segments do not require as significant reductions compared to the Upper
Basin and Lower Basin model segments. An evaluation of high intensity land uses (e.g.,
commercial/industrial uses, high density/multi-family residential, transportation, etc.) such was done
in the Upper/Middle and Lower Charles River TMDL evaluations, including analysis of relative
loads and required reductions by segment, could provide additional clarification in the future as the
TMDL implementation plan is established to meet the water quality targets.
Further analysis with wet and dry year data explores whether the chl-a water quality target can be
met during extreme conditions. During a wet year (such as 2010) the chl-a target appears to be met
without having to further adjust the stormwater reductions or sediment efflux reductions as
compared to the average annual loads. It appears that the wet year provided dilution of the loads,
and perhaps overall improved water quality, so much that less stormwater reductions would be
needed to achieve the target. In contrast, a dry year (such as 2016) would not meet the chl-a target
unless further stormwater reductions were applied. However, because 2016 was an extreme dry year
(below the 10th percentile threshold), the exceedance of water quality targets under these conditions
would be statistically infrequent. More detailed modeling with additional water quality data could
114
-------
Mystic River Watershed TMDL Alternative Development Final Report
help to provide greater confidence in the recommended phosphorus load reductions and the
potential exceedances of chl-a.
115
-------
Mystic River Watershed TMDL Alternative Development Final Report
X. Broad-Based Nutrient Stormwater Management
Strategies for the Mystic River Watershed using Opti-Tool
Highly urbanized areas often have limited opportunities for implementing large-scale Stormwater
Control Measures (SCMs) for treating stormwater runoff. Distributed green infrastructure (GI)
practices can provide cost-effective solutions that achieve load reduction numeric targets while
effectively integrating within urbanized landscapes. In New England, almost 50 percent of daily
rainfall events are less than 0.3 inches. The relatively small size of distributed GI facilities
substantially increases the feasibility to provide treatment to runoff from impervious surfaces in
constrained developed spaces and achieve meaningful water quality benefits in receiving waters.
Strategically optimizing the selection and placement of distributed SCMs within highly urbanized
settings can also help to develop management strategies that are more cost-effective than the
traditional approach of sizing BMPs at fixed locations to treat a design storm. The Opti-Tool, which
was developed for the United States Environmental Protection Agency (USEPA) Region 1, is a
continuous simulation model that can be used to optimize the selection and placement of distributed
GI practices at a watershed scale. This case study demonstrates how the Opti-tool can be used to
help with stormwater management planning in urban New England settings and highlights the value
of conducting strategic planning to address stormwater impacts for achieving water resource goals.
It presents an analytical framework that can be readily customized and applied in other settings to
inform the stormwater management planning effort. This section also provides examples of GI
implementation efforts in other locations where distributed GI practices were also found to be cost-
effective stormwater management strategies.
X.A. Study Objectives
Two stormwater management scenarios were formulated to evaluate two different stormwater
management approaches that meet the required annual TP load reduction target for the Mystic River
Watershed. These scenarios were configured and optimized using the Opti-Tool:
1. Design-Storm Objective: Optimize distributed BMP locations by land use type with fixed
sizes to capture 1 inch of runoff for a design storm and develop cost-effectiveness curve
(CE-Curve).
2. Mix-Storm Objective: Optimize distributed BMP locations and sizes by land use type to
reflect flexible sizing approach and develop CE-Curve.
This section presents a step-by-step, high-level technical approach to identify structural controls
associated with cost-effective stormwater management strategies. The Mystic River Watershed in
Massachusetts was selected as a watershed to test the sensitivities of the two stormwater
management scenarios and to identify the most cost-effective management approach that achieves
the phosphorus reduction objectives.
X.B. Pilot Sub-Watershed Selection
The Mystic River Watershed consists of 15 sub-watersheds, which drain into three primary river
segments, and seven large impoundments (ponds/lakes). This stormwater management portion of
the project uses one Mystic River sub-watershed to implement a high-level, generalized approach
and step-by-step guidance that is transferable to other sub-watersheds.
116
-------
Mystic River Watershed TMDL Alternative Development Final Report
The pilot sub-watershed that was selected is representative of the overall land use distribution in the
Mystic River Watershed. This comparison was made by computing the root mean square error
(RMSE) of the percent area distribution of each land use category between individual sub-
watersheds and the entire Mystic River Watershed. The four watersheds with the minimum RMSE
(indicating a close match to the overall watershed distribution) were: Maiden. Tudkins. Lower Mystic.
and Mystic River. Each of the four were intersected with municipal boundaries. The Mystic River
sub-watershed was selected for the pilot study because of its central location within the watershed. It
also represents a large portion of one community, the City of Medford. The pilot sub-watershed is
highlighted in yellow in Figure X-I.
The pilot sub-watershed comprises 5,151 acres of land and 179 acres of water bodies. Combined
Sewer Overflow (CSO) basins within the pilot sub-watershed drain 1,010 acres of land. Half of the
sub-watershed area is on low-slope topography (i.e., less than 5%) and 12% of the sub-watershed
area has slope larger than 15% (Table X-l, Figure X-II). The dominant soil type in the pilot
watershed is hydrologic soil group (HSG) C, which makes up 70% of the total watershed area (Table
X-2, Figure X-III). Soil map units with no HSG attribute data were assumed to be C soils (Group C
soils have relatively low infiltration rates). The dominant land use type in the pilot watershed is high-
density residential (46%) followed by forest (24%) and commercial (15%) land uses (Figure X-IV).
The pilot watershed is 49% impervious and the impervious portion of the watershed is mostly high-
density residential (28%), followed by commercial (11%) and transportation (highways) (4%). The
pervious portion of the watershed is mostly forest (23%) followed by high density residential
pervious (18%) and commercial pervious (4%), as shown in Figure X-V and Table X-3.
One-fifth of the pilot sub-watershed area falls within a CSO basin. That area was excluded from the
analysis because the CSO drainage area does not include separate stormwater drainage systems that
discharge to the receiving water. Thus, no stormwater controls were explored in CSO areas.
117
-------
My stic River Watershed TMDL Alternative Development Final Report
Streams
Lakes
AbeijonaRiverl
I I AbeijonaRiver2
Alewife
BlacksNook
Horn
Judkins
LowerMysticLake
Maiden
MysticRiver
UpperLobe
UpperMysticLake
Spy Pond
Wedge
fed! Winter
\//\ Cambridge CSO Basin 2017
, I//1 Somerville CSO Basins 2017
/ * <.*> ฆ
Figure X-l. Location Map of Pilot Sub-Watershed, Mystic River (Highlighted in
Yellow)
Table X-1. Ground Slope Classification in Pilot Sub-Watershed, Mystic River
Ground Slope Classification
Percent Slope
Area (acres)
Low
0% - 5%
2,852.92
Moderate
5% -15%
1,861.32
High
>15%
615.65
Total area
5,329.89
118
-------
Mystic River Watershed TMDL Alternative Development Final Report
I I Mystic River Subwatershed
VZA Somerville CSO Basin
Slope Classification
0% - 5%
4 5% -15%
f Wk >15%
JTErmh*... J
3 mile
Figure X-ll. Ground Slope Map of Pilot Sub-Watershed, Mystic River
Table X-2. Hydrologic Soil Group Classification in Pilot Sub-Watershed, Mystic River
Soil Classification
HSG
Area (acres)
No Data
C
3,601.45
HSG-B
B
560.90
HSG-C
C
85.04
HSG-C/D
C/D
283.45
HSG-D
D
799.05
Total area
5,329.89
119
-------
Mystic River Watershed TMDL Alternative Development Final Report
I I Mystic River Subwatershed
Y/A Somerville CSO Basin
Hydrologic Soil Group
No Data
-IsmAbi..
3 mile
Figure X-lll. Soil Map (Hydrologic Soil Group) of Pilot Sub-Watershed, Mystic
River
120
-------
Mystic River Watershed TMDL Alternative Development Final Report
High Oensity Residential, 46%
Highway, 6%
Industrial, 4%
Forest, 24%
Commercial, 15%
* I I Mystic River Subwatershed .
V7/1 Somerville CSO Basin
Landuse Classification
$ Agriculture
ฃ ฆ Commercial
ฆH Forest
H High Density Residential
Highway Pปt
Industrial
Low Density Residential
4 ฆ Medium Density Residential
Open land
Figure X-IV. Land Use Map of Pilot Sub-Watershed, Mystic River
121
-------
Mystic River Watershed TMDL Alternative Development Final Report
High Density Residential, 28%
I I Mystic River Subwatershed
V/A Somerville CSO Basin
Land Cover Type
>5 It Impervious
Pervious
Pervious Cover
Commercial, 4%
Forest, 23%
High Density Residential, 18%
Highway, 2% ,ndustrial> 1%
Open land, 3%
Impervious Cover
Commercial, 11%
3 mile
Figure X-V, Land Cover Map of Pilot Sub-Watershed, Mystic River
122
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table X-3. Land Use Classification in Pilot Sub-Watershed, Mystic River
Land Use Classification
Area (acres)
Percent Impervious
Percent Pervious
Commercial
769.79
11%
4%
Forest
1,225.94
0%
23%
High Density Residential
2,370.66
28%
18%
Highway
294.14
4%
2%
Industrial
235.22
4%
1%
Low Density Residential
0.55
0%
0%
Medium Density Residential
3.09
0%
0%
Open land
251.20
2%
3%
Water
179.30
Total
5,329.89
49%
51%
X.C. Technical Approach
The Opti-Tool provides the ability to evaluate options for determining the best mix of structural
BMPs to achieve water quality goals. Structural BMPs are permanent structures, provide stormwater
storage capacity, and rely upon vegetation and soil mechanisms in order to perform as intended. The
tool incorporates long-term runoff responses (HRU timeseries) for regional climate conditions that
are calibrated to regionally representative stormwater data and annual average pollutant load export
rates from nine land uses. The tool uses regionally representative BMP cost functions and regionally
calibrated BMP performance parameters for four pollutants, including total phosphorus, to calculate
long-term cumulative load reductions for a variety of structural controls. Structural controls
simulated by the tool include low impact development (LID) and green infrastructure (GI) practices,
such as infiltration systems, bio-filtration, and gravel wetlands.
The technical approach for applying the Opti-Tool is organized into three general steps:
1. Develop stormwater management categories for SCMs known to be highly effective at
removing phosphorus (e.g., shallow filtration, infiltration, biofiltration) based on the site
suitability analysis of GIS layers;
2. Estimate the available opportunity by BMP type (i.e., physical footprint area) within each
management category and summarize the upstream impervious drainage area that can be
managed for each management category, and
3. Set up and run the Opti-Tool application to identify the most cost-effective combination of
BMP options that achieve the desired management objectives.
X.C.1. Stormwater Management Categories Development
Spatial data analyses were conducted during Phase 1 of the project to characterize watershed
features and identify the corresponding stormwater management categories that were suitable for
application with the Opti-Tool for the pilot study area. The GIS data used for the evaluation of
stormwater management categories for the Mystic River Watershed include: municipal boundaries,
watershed sub-basins, land use coverage, impervious cover, Hydrologic Soil Group (HSG),
wetlands, Digital Elevation Model (DEM) for ground slopes, Activity and Use Limitation (AUL)
and MGL Ch. 21E sites (for contaminated land), and property ownership. All data are from
123
-------
Mystic River Watershed TMDL Alternative Development Final Report
MassGIS data layers except the watershed sub-basins, which were derived using FEMA catchments
and DEM data as described earlier in this report.
The following assumptions were made to develop the stormwater management categories and
estimate available BMP opportunity:
1. Areas with no depth to groundwater data were assumed to have a depth to groundwater
greater than 2.5 feet.
2. Areas with no HSG identified for soils were assumed to be classified as HSG C. This was
typically the predominant soil type in urban areas of the Mystic River Watershed.
3. The extent of potential contamination from AUL and/or MGL Ch. 21E sites was
approximated using the parcel in which the site was located.
4. Wetland areas were not included within the management categories because they were not
considered candidates for implementation of stormwater practices.
5. Both public and private land areas were assessed for identifying the GI opportunity areas
based on the site suitability criteria.
Areas with impervious cover (IC) were also explored for certain management practices (i.e., porous
pavement) that not only replace IC but can also treat stormwater from adjacent IC (for example,
within parking lots), but at a higher unit cost. Table X-4 presents siting criteria for potential
stormwater management categories (SMC), which were derived from the GIS data analysis. Table
X-5 shows the maximum footprint areas of each SMC that is available in the pilot watershed and the
spatial locations of those SMCs are shown in Figure X-VI.
124
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table X-4. Potential Stormwater Management Categories and BMP Types in Opti-
Tool
Cover Type
Ground
Slope (%)
AUL/21E
HSG
Management
Category
BMP Type(s) in Opti-
Tool
Pervious
Area
LO
I
II
V
Is a AUL/21E
Site
A/B/C/D or No Data
(HSG C assumed)
Shallow
filtration
Biofiltration (e.g.,
Bioretention with
underdrain option)
Not a AUL/
21E Site
A/B/C or No Data
(HSG C assumed)
Infiltration
Surface Infiltration
Basin (e.g., Rain
Garden)
D
Biofiltration
Biofiltration (e.g.,
Gravel Wetland)
> 15
-
-
Less likely for
onsite BMP
-
Impervious
Area
<= 5
-
A/B/C/D or No Data
(HSG C assumed)
Shallow
filtration
Porous Pavement
>5
-
-
Less likely for
onsite BMP
-
Table X-5. Potential BMP Opportunity Areas (Maximum Footprints) in the Pilot
Sub-Watershed, Mystic River
Stormwater Management Category
Area (acres)
Percent Area (%)
Biofiltration
437.92
8%
Shallow filtration (pervious area)
513.13
10%
Infiltration
1,087.20
20%
Shallow filtration (impervious area)
1,477.06
28%
Less likely for BMP
1,814.58
34%
Total
5,329.89
100%
125
-------
Mystic River Watershed TMDL Alternative Development Final Report
1 I Mystic River Subwatershed
V/A Somerville CSO Basin
Stormwater Management Categories
Biofiltration
H Infiltration
Less likely for BMP
PorousPavement
Shallow filtration
Oake/ Qckrtc/ f V
Figure X-VI. Stormwater Management Categories Map of Pilot Sub-Watershed,
Mystic River
For this study, stormwater management siting was primarily evaluated for areas with pervious cover,
outside of wetland areas, and with ground slopes of less than 15%. Porous pavement on impervious
cover with ground slopes of less than 5% was also evaluated with the assumption of drainage
impervious area ratio of 2:1, meaning one acre of porous pavement was assumed to treat its own
footprint plus one additional acre of adjacent impervious land. Only 10% of suitable impervious
land (based on the siting criterion) was identified for this practice. While removal of some
impervious land cover and implementation of porous pavement are viable options within this
watershed, those practices are typically costlier than stormwater practices located on pervious land.
Therefore, a multiplier of 3X (unit cost) for installing porous pavement was used in the Opti-Tool to
126
-------
Mystic River Watershed TMDL Alternative Development Final Report
account for installing underdrains and connections to the drainage system. In comparison, a
multiplier of 2x (unit cost) was used for other BMPs for installing new BMPs in developed areas.
X.C.2. Estimating BMP Footprints and Impervious Drainage Areas
The distribution of the BMP opportunity areas (i.e., BMP footprints) was estimated by land use
category group. This distribution represents the maximum available BMP footprint in the pilot
watershed, based on GIS spatial data analysis, and does not necessarily represent the actual
opportunity areas. The total impervious areas by land use group were proportionally distributed to
the BMP drainage areas based on the available percentage of opportunity area of that specific BMP
type by land use type as determined through the Management Category analysis (Table X-6 and
Figure X-VII). For example, if the opportunity area of Bio-filtration was 20% of the total available
opportunity area in commercial land, then 20% of the impervious area in the commercial land was
treated by Bio-filtration practices located on commercial land. For this case study, no field
verification was performed, and maximum opportunity areas were set to limit the BMP footprints
needed to capture up to 1 inch of runoff from the impervious drainage areas (Table X-7).
Table X-6. BMP-Treated Impervious Area (Drainage Area) Distribution by Land
Use Category Group in the Pilot Sub-Watershed, Mystic River
Land Use Type
Biofilt ration
Infiltration
HSG-B
Infiltration
HSG-C
Porous
Pavement
Total
(Land Use)
(acre)
(acre)
(acre)
(acre)
(acre)
Commercial
167.92
20.95
114.01
55.02
357.9
High Density Residential
9.74
31.78
714.35
116.01
871.88
Highway
86.81
13.67
46.76
-
147.24
Industrial
-
28.17
61.24
18.84
108.25
Open land
5.25
17.26
16.09
7.86
46.46
Total (BMP-treated)
269.72
111.83
952.45
197.73
1,531.73
127
-------
Mystic River Watershed TMDL Alternative Development Final Report
Treated Impervious Areas by BMP Type and Landuse Type
ฆ Commercial ฆ Industrial ฆ High Density Residential ฆ Open land ฆ Highway
Figure X-VII. Treated Impervious Areas by BMP Type and Land Use Type in the
Pilot Sub-Watershed, Mystic River
Table X-7. BMP Area (Footprints) Distribution by Land Use Category Group
Required to Treat 1 Inch of Runoff from the Impervious Surface in the Pilot Sub-
Watershed, Mystic River
Land Use Type
Biofilt ration
Infiltration
HSG-B
Infiltration
HSG-C
Porous
Pavement
Total
(Land Use)
(acre)
(acre)
(acre)
(acre)
(acre)
Commercial
10.00
0.87
4.75
27.51
43.13
High Density Residential
0.58
1.32
29.76
58.00
89.66
Highway
5.17
0.57
1.95
-
7.69
Industrial
-
1,17
2.55
9.42
13.14
Open land
0.31
0.72
0.67
3.93
5.63
Total (BMP-treated)
16,06
4,65
39.68
98.86
159.25
X.C.3. Opti-Tool Setup7
The following steps were performed to set up the Opti-Tool for the pilot sub-watershed.
1. Establish baseline condition: The climate data was extended to develop unit-area HRU
timeseries for the critical period of interest (Jan 2007 Dec 2016), which was used as the
7 Opti-Tool User Guide provides the step-by-step instructions on how to setup the Opti-Tool project.
128
-------
Mystic River Watershed TMDL Alternative Development Final Report
boundary condition to the BMP simulation model. The Opti-Tool provides a utility tool that
runs the SWMM models, calibrated to Region 1 specific land use average annual loading
export rates, and generates the HRU hourly timeseries in the format needed for the Opti-
Tool. These are the same loading HRUs being used in the watershed and water quality
modeling work done as part of the Mystic River Eutrophication analysis to determine
needed TP load reductions.
2. Management objective: Identify the most cost-effective stormwater controls (types and
sizes) for achieving a wide range of TP load reductions at the watershed scale.
3. Optimization target: Develop cost effectiveness curve (CE-curve) for TP average annual
load reduction.
4. Land use information: Estimate the area distribution for the major land use groups within
the pilot watershed. Assign the corresponding unit-area HRU timeseries for each land use
group in the model.
5. BMP information: Eighteen BMP types were selected on five major land use categories
based on the Management Category analysis and BMP specifications were set using the
default parameters and BMP cost function available in the Opti-Tool, see Appendix K).
Assign impervious drainage areas to be treated by each BMP type in the model.
6. Run optimization scenario: Define the simulation period (2007 - 2016), the pollutant of
concern (TP), the objective function (minimize cost), and create an input file for the
optimization run. Run the optimization using the continuous simulation BMP model to
reflect actual long-term precipitation conditions that includes a wide range of actual storm
sizes to find the optimal BMP storage capacities that provide the most cost-effective
solution at the watershed scale. Each optimization run generates a CE-Curve showing the
optimal solutions frontier for a wide range of TP load reduction targets.
Figure X-VIII shows the main interface of the Opti-Tool, the left panel guides the user to follow
steps in the chronological order. The right panel allows the user to place BMPs on the map and
enter design specifications for each BMP type.
129
-------
Mystic River Watershed TMDL Alternative Development Final Report
X.D. Management Scenarios
In this study, two management scenarios were created and optimized using the Opti-Tool. The most
cost-effective solution from each scenario was selected that met the TP average: annual load
reduction target for the pilot watershed. Two numeric targets; 67% and 62% (load reduction target
for existing condition [Section 1\. \. 11 and tuture condition 1 [Section IX.A.2], respectively) were
evaluated for each management scenario- The existing condition's target requires the load reduction
from stormwater only whereas the load reduction target for future condition 1 assumes additional
reductions to CSO volumes that meet the long-term control plan (LTCP) estimates for the typical
rainfall year and 50% sanitary sewer overflow (SSO) volume reductions for the ongoing SSO
mitigation work being done within the Mystic River Watershed. The load reduction targets for
tuture condition 2 and future condition 3 were not optimized in this study as it requires changing the
baseline by shifting land areas from the CSS areas. The details on the existing and future conditions
are discussed in Section IX of this report.
Scenario 1: Sizing the BMPs to capture one inch of surface runoff from the impervious drainage
areas and spatial optimization for the strategic locations (at the land use level) in the pilot
watersheds. The BMP sizes were fixed, and the optimization engine explored the best mix of BMP
types and strategic locations to identify the cost-effective solutions.
Scenario 2: Sizing the BMPs to capture from one-tenth to an inch of surface runoff from the
impervious drainage areas and spatial optimization for the strategic locations (at the land use level) in
the pilot watersheds. The BMP sizes were variable (increment of one-tenth of an inch to a maximum
130
-------
Mystic River Watershed TMDL Alternative Development Final Report
of 1 inch) and the optimization engine explored the best mix of BMP types, sizes, and strategic
locations to identify the cost-effective solutions.
X.D.1. Results: Scenario 1
Figure X-IX shows the CE-Curve for Scenario 1 from the model simulation in the Opti-Tool. The
Scenario: The scenario term is used to describe the BMP siting criteria; scenario 1 captures a typical one-inch
storm si^e (i.e., one-inch of runoff depth from the IC) whereas scenario 2 captures a range of storm si ^es from
one-tenth to a one-inch storm.
Condition: The condition term represents the target sources; existing condition taigets the stormwater load
reduction only whereas future condition 1 also taigets the CSO and SSO had reductions in addition to the
stormwater had reduction.
Solution: The solution term represents the optimised mixture of different BMP types, si^es, and locations
that meet the given numeric had reduction target.
Cost: The coit estimates are intendedfor planning levelpurposes and are intended to highlight relative cost
differences among the scenarios.
curve is an interactive plot showing the target solution (red triangle for the existing condition and
orange triangle for the future condition 1) and all the iterations performed during the optimization
process. The grey-dots on the curve are the inferior solutions and the blue-diamonds form the cost-
effectiveness curve for a wide range of load reduction targets. Based on the given target reduction,
Opti-Tool searches for the closest solution and provides the information on the selected BMPs
under that target solution (BMP ID, BMP type, surface area, storage depth, treated the impervious
area, runoff depth, annual maintenance hours, and BMP cost).
The results of Scenario 1 (where BMPs are sized using a typical design criterion of capturing one inch
of runoff from the impervious drainage area) show that it would cost $106.37 million to meet a 67%
TP average annual load reduction target for the existing condition whereas it would cost $66.12
million to meet a 62% TP average annual load reduction target for the future condition 1 for the
pilot watershed (Figure X-9). The cost estimates are based on regional unit cost information for the
control types, a 35% add-on for engineering and contingencies and a site factor multiplier to account
for anticipated difficulties associated with installations. For this analysis, a multiplier of 2Xwas
assumed for all controls except for porous pavement for which a 3X multiplier was applied. These
cost estimates can be considered conservative because they do not reflect the potential for
significant cost offsets that could be achieved through the installation of SCMs as part of other
development/redevelopment, urban renewal and roadwork related projects.
Though the optimization engine did not optimize the BMP sizes, it still shows a significant co st
saving in optimizing the strategic locations (where to place a BMP and what BMP combination to
use). Table X-8 shows the selected BMP types in the optimal solution that meet the TP load
reduction target for the existing condition. Table X-9 shows the selected BMP types in the optimal
solution that meet the TP load reduction target for the future condition 1. The optimizer preferred
the infiltration BMPs because they provided the highest volume reduction and associated water
quality benefits compared to the more expensive practices such as biofiltration and porous
pavements.
131
-------
My stic River Watershed TMDL Alternative Development Final Report
AllSolutions ~ Best Solutions A Target Solution 1
Target Solution 2
Target Solution 3
% TP Reduction
Annual Average Load
Figure X-IX. Scenario 1: Opti-Tool Cost-Effectiveness Curve (Optimize Locations
Only) of TP Annual Average Load Reduction for the Pilot Watershed
The shape of the Civ Curve itself provides valuable information for informing strategic stormwater
management planning. As indicated, the slope of the curve is relatively mild from 0" <> to 58% in TP
reduction and then increases sharply for higher TP reductions. The incremental cost of $ 40 million
to move from a TP reduction target of 62% to 67% is substantially higher than the $5 million
incremental cost increase associated with the same incremental increase in percent TP reduction of
53% to 58% on the flatter part of the curve. This curve highlights the potential high value of
investing in other measures including nonstructural control such as leaf litter management, high-
efficiency street cleaning, catch basin cleaning and fertilizer management that could achieve TP
reductions of 10-20% or higher and shift the target for the structural control retrofit program from
the steep part of the curve to the flatter portion. For example, assume that a 15% TP reduction
could be accomplished through nonstructural controls and that the target for structural control can
be reduced from 67% to 52% resulting in an estimated cost for structural controls of approximately
$40 million (less than half).. Not only is it far less costly to move from the steep portion of the curve
to the flatter portion but the total amount ot impervious cover area requiring treatment for a 52 %
reduction (909 acres) is substantially less than the 67% (1,379 acres) and 62% (1,211 acres) options
as indicated in Table X-10, Table X-9, and Table X-9 respectively.
132
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table X-8. Scenario 1: BMP Types and BMP Sizes for The Selected Target
Solution 1 (Existing Condition) in Opti-Tool
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth
(in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP1
Infiltration-B
High Density Residential
31.78
1.00
863,013
$1,439,764
BMP2
Infiltration-B
Commercial
20.95
1.00
568,931
$949,148
BMP3
Infiltration-B
Industrial
28.17
1.00
764,973
$1,276,203
BMP4
Infiltration-B
Open land
17.26
1.00
468,813
$782,121
BMP5
Infiltration-B
Highway
13.67
1.00
371,150
$619,189
BMP6
Infiltration-C
High Density Residential
714.35
1.00
19,401,598
$32,367,663
BMP7
Infiltration-C
Commercial
114.01
1.00
3,096,513
$5,165,908
BMP8
Infiltration-C
Industrial
61.24
1.00
1,663,216
$2,774,741
BMP9
Infiltration-C
Open land
16.09
1.00
437,036
$729,107
BMP10
Infiltration-C
Highway
46.76
1.00
1,270,010
$2,118,756
BMP11
Biofiltration
High Density Residential
-
-
-
-
BMP12
Biofiltration
Commercial
167.92
1.00
4,559,847
$18,847,683
BMP13
Biofiltration
Open land
5.25
1.00
142,556
$589,242
BMP14
Biofiltration
Highway
86.81
1.00
2,357,192
$9,743,223
BMP15
Porous Pavement
High Density Residential
-
-
-
-
BMP16
Porous Pavement
Commercial
55.03
-
13,573,865
$28,963,024
BMP17
Porous Pavement
Industrial
-
-
-
-
BMP18
Porous Pavement
Open land
-
-
-
-
Total
1,379.27
1.00
49,538,712
$106,365,771
Table X-9. Scenario 1: BMP Types and BMP Sizes for the Selected Target Solution
2 (Future Condition 1) in Opti-Tool
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth
(in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP1
Infiltration-B
High Density Residential
31.78
1.00
863,013
$1,439,764
BMP2
Infiltration-B
Commercial
-
-
-
-
BMP3
Infiltration-B
Industrial
28.17
1.00
764,973
$1,276,203
BMP4
Infiltration-B
Open land
17.26
1.00
468,813
$782,121
BMP5
Infiltration-B
Highway
13.67
1.00
371,150
$619,189
BMP6
Infiltration-C
High Density Residential
714.35
1.00
19,401,598
$32,367,663
BMP7
Infiltration-C
Commercial
114.01
1.00
3,096,513
$5,165,908
BMP8
Infiltration-C
Industrial
61.24
1.00
1,663,216
$2,774,741
BMP9
Infiltration-C
Open land
16.09
1.00
437,036
$729,107
BMP10
Infiltration-C
Highway
46.76
1.00
1,270,010
$2,118,756
133
-------
Mystic River Watershed TMDL Alternative Development Final Report
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth
(in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP11
Biofiltration
High Density Residential
-
-
-
-
BMP12
Biofiltration
Commercial
167.92
1.00
4,559,847
$18,847,683
BMP13
Biofiltration
Open land
-
-
-
-
BMP14
Biofiltration
Highway
-
-
-
-
BMP15
Porous Pavement
High Density Residential
-
-
-
-
BMP16
Porous Pavement
Commercial
-
-
-
-
BMP17
Porous Pavement
Industrial
-
-
-
-
BMP18
Porous Pavement
Open land
-
-
-
-
Total
1,211.24
1.00
32,896,167
$66,121,134
Table X-10. Scenario 1: BMP Types and BMP Sizes for the Selected Target
Solution 3 (52% Target) in Opti-Tool
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth
(in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP1
Infiltration-B
High Density Residential
31.78
1.00
863,013
$1,439,764
BMP2
Infiltration-B
Commercial
20.95
1.00
568,931
$949,148
BMP3
Infiltration-B
Industrial
28.17
1.00
764,973
$1,276,203
BMP4
Infiltration-B
Open land
-
-
-
-
BMP5
Infiltration-B
Highway
-
-
-
-
BMP6
Infiltration-C
High Density Residential
714.35
1.00
19,401,598
$32,367,663
BMP7
Infiltration-C
Commercial
114.01
1.00
3,096,513
$5,165,908
BMP8
Infiltration-C
Industrial
-
-
-
-
BMP9
Infiltration-C
Open land
-
-
-
-
BMP10
Infiltration-C
Highway
-
-
-
-
BMP11
Biofiltration
High Density Residential
-
-
-
-
BMP12
Biofiltration
Commercial
-
-
-
-
BMP13
Biofiltration
Open land
-
-
-
-
BMP14
Biofiltration
Highway
-
-
-
-
BMP15
Porous Pavement
High Density Residential
-
-
-
-
BMP16
Porous Pavement
Commercial
-
-
-
-
BMP17
Porous Pavement
Industrial
-
-
-
-
BMP18
Porous Pavement
Open land
-
-
-
-
Total
909.25
1.00
24,695,027
$41,198,686
134
-------
Mystic River Watershed TMDL Alternative Development Final Report
X.D.2. Results: Scenario 2
The results of Scenario 2 (where BMPs are sized to capture from one-tenth to an inch of runoff from
the impervious drainage area) show that it would cost $102.83 million to meet a 67% TP average
annual load reduction target for the existing condition whereas it would cost $51.13 million to meet
a 62% TP average annual load reduction target for the future condition 1 for the pilot watershed
(Figure X-X). For this scenario, the optimization engine optimized the BMP sizes and the strategic
locations (i.e., where to place a BMP, what BMP size to pick, and what BMP combination to use).
The results show more cost saving ($3.5 million for existing condition and $15 million for future
condition 1) for optimizing the BMP sizes as compared to only optimizing the strategic locations.
Table X-ll shows the selected BMP types in the optimal solution that meet the TP load reduction
target for the existing condition. Table X-12 shows the selected BMP types in the optimal solution
that meet the TP load reduction target for the future condition 1 for Scenario 2. However, similar to
scenario 1, the TP reduction targets of 67% and 62% are located on the portion of the curve with
the steeper slope indicating much higher incremental cost increases for increasing TP reduction
targets. Again, assuming that nonstructural controls could achieve a 15% TP reduction, the
estimated cost of achieving a 52% reduction using structural controls is $20 million and equal to 'A
of the estimated cost ($40 million) for achieving the same reduction using 1-inch design capacities.
Table X-13 summarizes the results of achieving 52% TP reduction for scenario 2. As indicate,
optimized sizing of the structural controls for 52% are notably smaller than was determined for the
62% and 67% reductions.
The optimizer mostly picks the infiltration BMPs because they provide the highest volume reduction
and water quality benefits compared to more expensive practices such as biofiltration and porous
pavement. For Scenario 2, the optimizer picks a combination of BMP sizes ranging from 0.1 to 1.0
inches, showing that BMPs designed to manage smaller-size storms in New England region may
provide more cost saving and increased feasibility for implementation in the highly urbanized
watershed such as the Mystic River.
135
-------
My stic River Watershed TMDL Alternative Development Final Report
All Solutions ~Best Solutions A Target Solution 1 t Target Solution 2 Target Solution 3
% TP Reduction
Annual Average Load
Figure X-X. Scenario 2: Opti-Tool Cost-Effectiveness Curve (Optimize Locations
and BMP Sizes) of TP Annual Average Load Reduction for the Pilot Watershed.
Table X-11. Scenario 2: BMP Types and BMP Sizes for the Selected Target
Solution 1 (Existing Condition) in Opti-Tool
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth (in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP1
Infiltration-B
High Density
Residential
31.78
0.70
604,107
$1,007,830
BMP2
Infiltration-B
Commercial
20.95
0.60
341,364
$569,497
BMP3
Infiltration-B
Industrial
28.17
0.80
611,981
$1,020,967
BMP4
Infiltration-B
Open land
17.26
0.40
187,524
$312,847
BMP5
Infiltration-B
Highway
13.67
0.80
296,922
$495,354
BMP6
Infiltration-C
High Density
Residential
714.35
1.00
19,401,581
$32,367,635
BMP7
Infiltration-C
Commercial
114.01
0.90
2,786,886
$4,649,358
BMP8
Infiltration-C
Industrial
61.24
0.80
1,330,580
$2,219,806
BMP9
Infiltration-C
Open land
16.09
0.90
393,340
$656,208
BMP10
Infiltration-C
Highway
46.76
0.70
889,019
$1,483,150
136
-------
Mystic River Watershed TMDL Alternative Development Final Report
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth (in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP11
Biofiltration
High Density
Residential
9.74
0.80
211,654
$874,853
BMP12
Biofiltration
Commercial
167.92
1.00
4,559,847
$18,847,683
BMP13
Biofiltration
Open land
5.25
1.00
142,555
$589,236
BMP14
Biofiltration
Highway
86.81
0.90
2,121,495
$8,768,993
BMP15
Porous Pavement
High Density
Residential
-
-
-
-
BMP16
Porous Pavement
Commercial
55.03
-
13,573,865
$28,963,024
BMP17
Porous Pavement
Industrial
-
-
-
-
BMP18
Porous Pavement
Open land
-
-
-
-
Total
1,389.01
o
1
O
47,452,719
$102,826,441
Table X-12. Scenario 2: BMP Types and BMP Sizes for the Selected Target
Solution 2 (Future Condition 1) in Opti-Tool
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth (in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP1
Infiltration-B
High Density
Residential
31.78
0.60
517,806
$863,855
BMP2
Infiltration-B
Commercial
20.95
0.30
170,682
$284,749
BMP3
Infiltration-B
Industrial
28.17
0.70
535,483
$893,346
BMP4
Infiltration-B
Open land
17.26
0.40
187,524
$312,847
BMP5
Infiltration-B
Highway
13.67
0.80
296,922
$495,354
BMP6
Infiltration-C
High Density
Residential
714.35
0.70
13,581,106
$22,657,344
BMP7
Infiltration-C
Commercial
114.01
0.80
2,477,232
$4,132,763
BMP8
Infiltration-C
Industrial
61.24
0.50
831,613
$1,387,379
BMP9
Infiltration-C
Open land
16.09
0.40
174,818
$291,648
BMP10
Infiltration-C
Highway
46.76
0.70
889,019
$1,483,150
BMP11
Biofiltration
High Density
Residential
9.74
0.70
185,198
$765,496
BMP12
Biofiltration
Commercial
167.92
0.80
3,647,877
$15,078,146
BMP13
Biofiltration
Open land
5.25
0.90
128,299
$530,313
BMP14
Biofiltration
Highway
86.81
0.20
471,443
$1,948,665
BMP15
Porous Pavement
High Density
Residential
-
-
-
-
BMP16
Porous Pavement
Commercial
-
-
-
-
BMP17
Porous Pavement
Industrial
-
-
-
-
BMP18
Porous Pavement
Open land
-
-
-
-
Total
1,333.99
(0.2-0.9)
24,095,022
$51,125,054
137
-------
Mystic River Watershed TMDL Alternative Development Final Report
Table X-13. Scenario 2: BMP Types and BMP Sizes for the Selected Target
Solution 3 (52% Target) in Opti-Tool
BMPID
BMP Type
Land Use
Treated
Impervious
Area (acres)
Runoff
Depth (in.)
BMP
Storage
Capacity
(gallon)
BMP Cost ($)
BMP1
Infiltration-B
High Density
Residential
31.78
0.60
517,806
$863,855
BMP2
Infiltration-B
Commercial
20.95
0.30
170,682
$284,749
BMP3
Infiltration-B
Industrial
28.17
0.20
152,995
$255,242
BMP4
Infiltration-B
Open land
17.26
0.40
187,524
$312,847
BMP5
Infiltration-B
Highway
13.67
0.20
74,230
$123,838
BMP6
Infiltration-C
High Density
Residential
714.35
0.40
7,760,632
$12,947,054
BMP7
Infiltration-C
Commercial
114.01
0.30
928,962
$1,549,786
BMP8
Infiltration-C
Industrial
61.24
0.50
831,613
$1,387,379
BMP9
Infiltration-C
Open land
16.09
0.20
87,409
$145,824
BMP10
Infiltration-C
Highway
46.76
0.20
254,005
$423,757
BMP11
Biofiltration
High Density
Residential
9.74
0.10
26,457
$109,357
BMP12
Biofiltration
Commercial
167.92
0.10
455,985
$1,884,768
BMP13
Biofiltration
Open land
-
-
-
-
BMP14
Biofiltration
Highway
-
-
-
-
BMP15
Porous Pavement
High Density
Residential
-
-
-
-
BMP16
Porous Pavement
Commercial
-
-
-
-
BMP17
Porous Pavement
Industrial
-
-
-
-
BMP18
Porous Pavement
Open land
-
-
-
-
Total
1,241.93
(0.1-0.6)
11,448,300
$20,288,455
X.E. Summary
The results of this pilot study provide quantitative and qualitative technical guidance to support
watershed-based GI management planning. Opti-Tool analysis results help to identify optimal
stormwater controls (including categories of methods and sizing approaches) that could increase the
technical and economic feasibility of retrofitting needed stormwater management strategies into
developed watershed areas. This study highlights the computational power of optimization
algorithms in Opti-Tool for evaluating thousands of iterations for a combination of different BMP
types and BMP sizes at strategic locations. As demonstrated in Scenario 1 (BMPs sized for a typical
design storm), spatial optimization at the watershed scale can provide significant cost savings as
compared to picking locations by best professional judgment. Scenario 2 further demonstrates that
when location and size are optimized, there is potential for further cost savings for the same annual
average load reduction benefit. The results of both scenarios indicate that considerable cost savings
or avoidance may be accomplished through investing in nonstructural controls to reduce the TP
138
-------
Mystic River Watershed TMDL Alternative Development Final Report
reduction target for structural controls from the steep portion of the CE-Curve to the flatter
portion. Table X-14 compares the two scenarios simulated in this case study.
Table X-14. BMP Scenarios Comparison in Opti-Tool
Scenario
ID
Scenario
Description
TP Load
Reduction
Target (%)
Impervious
Area Treated
(acre)
Runoff
Depth (in.)
BMP Storage
Capacity
(Million gallon)
BMP Cost
(Million $)
Scenario 1
BMP size (1 in.)
and optimize
the spatial
locations
67%
1,379
1.00
49.54
$106.37
62%
1,211
1.00
32.90
$66.12
52%
909
1.00
24.70
$41.20
Scenario 2
Optimize BMP
size (0.1 in.
increment and
max size 1 in.)
and the spatial
locations
67%
1,389
o
1
^|-
o.
47.45
$102.83
62%
1,334
(0.2-0.9)
24.10
$51.13
52%
1,242
(0.1-0.6)
11.45
$20.29
The CE-curve for both scenarios shows that the numeric targets for existing condition (67%) and
future condition 1 (62%) are above the knee-of-curve where optimal solutions tend to become
expensive due to the cheaper stormwater controls being exhausted. The performance curve provides
clear guidance on achieving the management objectives in a most-cost effective manner. The future
condition 1 load reduction target solution for Scenario 2 provides a cost saving of $51.7 million as
compared to the existing condition target solution for Scenario 2. By lowering 5% of load reduction
target (from 67% to 62%), it can provide almost 50% of cost saving (from $102.83 to $51.13) for
Scenario 2. The CE-curve provides optimal solutions for a range of load reduction targets, so it can
also be used to pick solutions for the intermediate milestones that show progress towards meeting
the final load reduction target. Additionally, the CE-Curve provides information to make a strong
case for investing in nonstructural and source reduction controls to achieve TP reductions and avoid
the need for installing the most expensive control in the most challenging locations.
This study provides planning level analysis with no site-specific project information but provides
guidance on which land use sources to target and what type of BMPs are suitable and how to size
those BMPs. For example, Scenario 2 of future condition 1 (Table X-12 and Table X-13) show
recipes for meeting the 62% and 52% TP load reduction targets by implementing 14 and 12
different BMP types, respectively (combination of land use, soil, and storage capacity) in the pilot
139
-------
Mystic River Watershed TMDL Alternative Development Final Report
watershed. Next step would be performing the field investigation to identify the feasible sites in the
watershed and selecting the BMP types and sizes based on the guidance provided in Table X-13. For
example, identified suitable sites on high density residential land use should be designed as
infiltration practices to capture 0.6-inch runoff depth for underlying soil B type and to capture 0.4-
inch depth for underlying soil C type. For poor draining soil type D, biofiltration practices can be
designed to capture 0.1-inch of runoff depth. The required BMP storage capacity reflects the storage
volume of the control expressed in terms of runoff depth from the contributing impervious area.
X.F. Example Projects
This section provides two example projects of GI implementation efforts in other locations where
distributed GI practices were also found to be cost-effective stormwater management strategies.
X.F.1. Berry Brook Project, Dover, New Hampshire
Berry Brook Project in Dover, New Hampshire is a partnership between New Hampshire
Department of Environmental Services (NHDES), University of New Hampshire Stormwater
Center (UNHSC) and the City. This unique partnership between regulators, academics, and
committed city staff has reduced best management practice implementation costs, increased
effectiveness, and led to more maintainable stormwater management systems. The project goal is to
filter, infiltrate, and reduce stormwater runoff from Effective Impervious Cover (EIC) as a means
for managing pollutant loading and controlling runoff volumes to Berry Brook. The project has
become a prime example of how scientists and public works departments collaborated to improve
water quality in an urban watershed, using Low Impact Development (LID) and Green
Infrastructure (GI) retrofits, that reduced the effective impervious area in the 185-acre urban
watershed from 30% down to 10%. Below is the list of BMPs implemented for this project and
Figure X-ll shows their locations in the watershed. A detailed report on this project is available on
UNHSC website to download.
12 bioretention systems,
a tree filter,
a subsurface gravel wetland,
one acre of new wetland,
3 grass-lined swales
2 subsurface gravel filters
an infiltration trench system
3 innovative filtering catch basin designs
140
-------
Mystic River Watershed TMDL Alternative Development Final Report
N Berry Brook BMPS
0 0.0450.09 0.18 0.27 0.36
am i iMiles
Legend
New BMPs
I | BB_Watershed
2015 1-foot Orthophotography
Figure X-XI. Green Infrastructure Retrofits for Berry Brook Project in Dover, New
Hampshire
X.F.2. The Advancing Green Infrastructure Program, New Haven, Connecticut
The Advancing Green Infrastructure Program in New Haven, Connecticut is a public-private
partnership that: promotes environmental protection and social justice through the construction of
141
-------
Mystic River Watershed TMDL Alternative Development Final Report
hundreds of right-of-way bioswales to combat water quality pollution associated with stormwater
runoff and combined sewer overflows. Every year, approximately 260 million gallons of combined
sewage enters the waterways surrounding New Haven, contributing to pollution in local waterways
and the Long Island Sound, and negatively impacting ecosystem health and public recreation.
The bioswale program began as a pilot project in 2014, funded through the National Fish and
Wildlife Long Island Sound Futures Fund, with the goal of quantifying how effective green
infrastructure was at stormwater retention in the city. The project was spearheaded by the New
Haven Resources Initiative (URI), a non-profit in New Haven, and made possible by the
partnerships URI fostered in academia, the non-profit sector, and the city and water pollution
control authority. University of Rhode Island (URI), Yale School of Forestry & Environmental
Studies, the City of New Haven's Engineering Department and the Greater New Haven Water
Pollution Control Authority (GNHWPCA) sited eight bioswales in the Westville neighborhood of
New Haven. Considerations in siting were slope, obstacles (trees, telephone poles, driveways,
underground utilities), homeowner agreement and interest, and the desire to place bioswales as close
upstream of catch basins as possible.
Once bioswales were sited, outreach was conducted in the neighborhood to ask homeowners if they
would be willing to adopt the bioswales, ensuring that bioswales were taken care of early in the
process and maintained in the future. Each of the eight bioswales was adapted from the City of New
York standard 3.05 m x 1.5 m right of way bioswale optimized for high capacity stormwater
retention. These bioswales were generally smaller than typical NYC bioswales, and ranged from 8' x
5.2' (the smallest) to 16.5' x 6' (the largest). Four feet of soil was excavated, and a geo-textile was
installed in the base to prevent fine soil and sediment infiltration past the base of the swale. The
bioswale was backfilled with two feet of engineered soil (a mix of New Haven sandy loam and
compost/mulch organic material to facilitate plant growth) and one foot of river stone or gravel.
Soils in the New Haven area are highly permeable, consisting of sandy loams with infiltration rates
near 0.46 m/hr. Along one end of each bioswale, 0.6 0.7 m wide vertical wire fenced gabions were
placed in each bioswale and filled with river stones to encourage fast infiltration if ponding occurred.
New Haven high school students from Common Ground High School, alongside homeowners who
adopted the swales, planted native plants and shrubs, such as winterberry holly, black-eyed susans,
and grasses and chose their plant pallet (see Figure X-XII). The sidewalk curbs were cut to allow
water to infiltrate into the gardens.
This small pilot program turned into a citywide partnership program that recently won Harvard's
prestigious Roy Award for Environmental Partnership (City of New Haven, 2018). Funding for an
additional 275 bioswales has been secured, with 200 built within downtown New Haven to alleviate
flooding and 75 within the combined sewer area to mitigate combined sewer overflow pollution. In
small cities like New Haven that are considering sewer separation, bioswales have proven to be a
cost-effective alternative to this expensive and disruptive procedure, while also improving flooding
hazards and risks in urban areas. Bioretention systems are scalable to larger cities, especially those
that have aging infrastructure. If sited strategically and developed with adequate maintenance, green
infrastructure can serve as a long-term solution to aid in stormwater management.
142
-------
Mystic River Watershed TMDL Alternative Development Final Report
Figure X-XII. Homeowners and Common Ground High School students plant
perennials in a New Haven Bioswale
Photo taken by Kelsey Semrod, https://hixon.yale.edu/practice/bioswales.
143
-------
Mystic River Watershed TMDL Alternative Development Final Report
XI. References
Arnold JG and PM Allen, 1999. Automated methods for estimating baseflow and groundwater
recharge from streamflow records. J Am Water Resource Assoc, 35(2), 411-424.
Arnold, JG, PM Allen, R Muttiah, and G Bernhardt, 2005. Automated base flow separation and
recession analysis techniques. Ground Water, 33(6): 1010-1018.
Berry Brook Project. (2017). Berry Brook Watershed Management Plan Implementation Projects
Phase III. Prepared for: The New Hampshire Department of Environmental Services, NH.
Prepared by: The City of Dover and University of New Hampshire Stormwater Center, NH.
December 2017.
Breault, RF, Sorenson, JR and Weiskel, PK, 2002, Streamflow., Water Quality and Contaminant Loads in
the Lower Charles River Watershed, Massachusetts, 1999-2000. Water Resources Investigations Report 02-
4137, 139 p. U.S. Geological Survey, Westborough, Massachusetts.
City of New Haven. (2018). New Haven's "Advancing Green Infrastructure Program" Wins
Harvard's Roy Award for Environmental Partnership Promoting environmental protection, climate
resilience, and social justice in New Haven, CT. Accessed November 9:
https://www.newhavenct.gov/news/displaynews.htm?NewsID=516&TargetID=67
Clough JS, 2014. AQUATOX (Release 3.1 plus). Modeling Environmental Fate and Ecological
Effects in Aquatic Ecosystems. Volume 1: User's Manual. US Environmental Protection Agency,
Office of Science and Technology, Washington DC.
Demonstration of Opti-Tool: Buzzards Bay Watershed Case Study. Prepared for: U.S. EPA Region
1, Boston, MA. Prepared by: Tetra Tech, Inc. Fairfax, VA. December 30, 2016.
DEP, 2010. Draft Total Maximum Daily Loads of Phosphorus for Selected Boston Harbor (Mystic
Basin) Lakes. Department of Environmental Protection, Division of Watershed Management, MA.
DEP-EPA, 2011. Total Maximum Daily Load for Nutrients in the Upper/Middle Charles River,
Massachusetts. Control Number: CN 272.0. Massachusetts Department of Environmental
Protection, Worcester, MA and US Environmental Protection Agency, Boston, MA. Prepared by
Charles River Watershed Association, Weston, MA and Numeric Environmental Services, Beverly
Farms, MA.
DFW (no date). Various Bathymetry Maps. Department of Fish and Game, MA.
Dodds, WK, & Oakes, RM. 2004. A technique for establishing reference nutrient conditions across
watersheds affected by humans. Limnol. Oceanogr.: Methods 2, 2004, pgs. 333341.
ENSR, 2000. Diagnostic/Feasibility Study of Winter Pond, Winchester, Massachusetts. Prepared for
The Town of Winchester and the Friends of Winter Pond Association.
EPA, 2016. Phosphorus TMDLs for Vermont Segments of Lake Champlain. U.S. Environmental
Protection Agency, Region 1, Boston, MA.
EPA, 2018. Bathymetric Depths and Soft Sediment Presence Measurements in the Mystic River.
Personal communication with Mark Voorhees and Monique Dulac of EPA Region 1.
Filstrup, CT and Downing, JA, 2017. Relationship of chlorophyll to phosphorus and nitrogen in
nutrient-rich lakes, Inland Waters, 7(4), p385-400, D01:10.1080/20442041.2017.1375176.
144
-------
Mystic River Watershed TMDL Alternative Development Final Report
Fischer, HB, List, EJ, Koh, RCY, Imberger, J and Brooks. NH, 1979. Mixing in inland and coastal
waters. Academic Press, New York, N.Y.
Flanagan, SA, DL Montgomery, and JD Ayotte, 2001. Shallow Ground-water Quality in Boston,
Massachusetts Metropolitan Area. USGS Water-Resources Investigations Report 01-4042.
HW, 2015, Stump Brook / Monponsett Pond Hydrologic and Water Quality Assessment. Prepared
for the Massachusetts Division of Ecological Restoration, MA by Horsley Witten Group, Sandwich,
MA.
IDEQ, 2015. Lower Boise River TMDL. Total Phosphorus Addendum. Idaho Department of
Environmental Quality, Boise, ID.
Knight, E. (2016). Aberjona River: Historical Background. Prepared by Ellen Knight, PhD for the
Town of Winchester's Flood Mitigation Program. Updated 6/25/2016.
https://www.winchester.us/DocumentCenter/View/1745 (Accessed June 28, 2017)
Koiv T, Noges T, and Laas A, 2011. Phosphorus retention as a function of external loading
hydraulic turnover time, area and relative depth in 54 lakes and reservoirs. Hydrobiologia 660, pl05
115. DOI 10.1007/s 10750-010-0411-8.
Ludlam, S. D., & Duval, B. (2001). Natural and Management-Induced Reduction in Monimolimnetic
Volume and Stability in a Coastal, Meromictic Lake. Lake and Reservoir Management, 17(2), 7181.
Massachusetts Department of Environmental Protection & U.S. EPA. 2007. Final Total Maximum
Daily Load for Nutrients in the Lower Charles River Basin, Massachusetts. CN 301.0.
Massachusetts Department of Environmental Protection. 2016. Massachusetts Consolidated
Assessment and Listing Methodology Guidance Manual for the 2016 Reporting Cycle. CN 445.0.
McKee, T.B., Doesken N.J, and Kliest J. 1993. The relationship of drought frequency and duration
to time scales. In Proceedings of the 8th Conference of Applied Climatology, 17-22 January,
Anaheim, CA. American Meteorlogical Society, Boston, MA. 179-184.
MWRA, 2018. CSO Discharge Estimates and Rainfall Analyses for Calendar Year 2017. Technical
Memorandum. April 30, 2018.
MWRA, 2016. Industrial Waste Report. Number 32. October 2016, variously paged. Massachusetts
Water Resources Authority, Boston, Massachusetts.
NADP, 2016. Nitrogen from the Atmosphere. ISWS Miscellaneous Publication 207 and NADP
Brochure 2016-01. National Atmospheric Deposition Program, University of Wisconsin, Madison,
WI. http://nadp.slh.wisc.edu/lib/brochures/nitrogenAtmos.pdf.
NADP, 2018a. Map of Nitrate Ion Wet Deposition, 2015. National Atmospheric Deposition
Program, University of Wisconsin, Madison, WI.
http://nadp.slh.wisc.edu/maplib/pdf/2015/N03_dep_2015.pdf.
NADP, 2018b. Map of Ammonium Ion Wet Deposition, 2015. National Atmospheric Deposition
Program, University of Wisconsin, Madison, WI.
http://nadp.slh.wisc.edu/maplib/pdf/2015/NH4_dep_2015.pdf.
NADP, 2018c. Map of Total Nitrogen Deposition, 2015. National Atmospheric Deposition
Program, University of Wisconsin, Madison, WI.
http://nadp.slh.wisc.edu/committees/tdep/tdepmaps/preview.aspx#n_tw.
145
-------
Mystic River Watershed TMDL Alternative Development Final Report
NCEI, 2018. National Centers for Environmental Information, formerly the National Climatic Data
Center, https://www.ncdc.noaa.gov/data-access.
NOAA, 2018. Hourly/Sub-Hourly Data Observation Map.
https://gis.ncdc.noaa.gov/maps/ncei/cdo/hourly. Accessed May 2018. National Oceanic and
Atmospheric Administration, Washington, DC.
Opti-Tool for Stormwater and Nutrient Management: User's Guide. Prepared for: U.S. EPA Region
1, Boston, MA. Prepared by: Tetra Tech, Inc. Fairfax, VA. June 30, 2016.
Park RA and Clough JS, 2014. AQUATOX (Release 3.1 plus). Modeling Environmental Fate and
Ecological Effects in Aquatic Ecosystems. Volume 1: Technical Documentation. US Environmental
Protection Agency, Office of Science and Technology, Washington DC.
Paul, M.J., J. Robbiani, L. Zheng, T. Rafi, S. Bai, and P. von Loewe. 2011. Development of Nutrient
Endpoints for the Northern Piedmont Eco-region of Pennsylvania: TMDL Application Follow-up
Analysis. Prepared by Tetra Tech, Inc., Owings Mills, MD for United States Environmental
Protection Agency, Region 3.
Powers SM, Tank JL, and Robertson DM, 2015. Control of nitrogen and phosphorus transport by
reservoirs in agricultural landscapes. Biogeochemistry 124, p417439. DOI 10.1007/s 10533-015-
0106-3.
Sloto RA and MY Crouse, 1996. HYSEP: A Computer Program for Streamflow Hydrograph
Separation and Analysis. Water-Resources Investigations Report 96-4040. US Geological Survey,
Lemoyne, PA.
TetraTech, 2015. Lake Champlain BATHTUB Model Calibration Report. Prepared for the U.S.
Environmental Protection Agency, Region 1, Boston, MA.
Tipping, E, Benham, S, Boyle, J F, Crow, P, Davies, J, Fischer, U, Guyatt, H, Helliwell, R, Jackson-
Blake, L, Lawlor, AJ, Monteith, DT, Roweg, EC, and Tobermanac, H, 2014. Atmospheric
deposition of phosphorus to land and freshwater. Environ. Sci.: Processes Impacts, 2014 (16), 1608
1617. https://doi.org/10.1039/c3em00641g.
U.S. EPA, 1986. Quality Criteria for Water.
U.S. EPA, 1998. Level III eco-regions of the continental United States (revision of Omernik, 1987).
U.S. EPA, 2001. Using Stressor-response Relationships to Derive Numeric Nutrient Criteria. EPA-
820-S-10-001.
Wagner K, 2009. LLRM Lake Loading Response Model. Users Guide and Quality Assurance
Project Plan. Water Resource Services, Wilbraham, MA.
Walker, W, 1982. An empirical analysis of phosphorus, nitrogen, and turbidity effects on reservoir
chlorophyll-a levels, Canadian Water Resources Journal, 7(1), 88-107,
https://doi.org/10.4296/cwrj0701088. Published online 2013.
Walker, W, 1985. Empirical Methods for Predicting Eutrophication in Impoundments - Report 3:
Model Refinements", Technical Report E-81-9, U.S. Army Corps of Engineers, Waterways
Experiment Station, Vicksburg, MI.
146
-------
Mystic River Watershed TMDL Alternative Development Final Report
Walker W, 1996. Simplified Techniques for Eutrophication Assessment & Prediction: User Manual.
Instruction Report W-96-2. US Army Corps of Engineers, Waterways Experiment Station,
Vicksburg, MI. Updated September 1999.
Walker W, 2004. BATHTUB-Model Version 6.1. Simplified Techniques for Eutrophication
Assessment & Prediction. US Army Corps of Engineers, Waterways Experiment Station, Vicksburg,
MI.
WH, 1987. Clean Lakes Program: Diagnostic/Feasibility Studies of Black Nooks Pond. Wiitman
and Howard, MA.
WH, 1988. Clean Lakes Program: Diagnostic/Feasibility Studies of Wedge Pond. Wiitman and
Howard, MA.
147
-------
Mystic River Watershed TMDL Alternative Development Final Report
Appendix A: Water body and Monitoring Location
Summary Table
Table A-1. Number of Days Sampled from 2000 - 2016 and Maps of Monitoring
Locations.
Monitoring Program
/ Monitoring
Station
Baseline
Phos.
Loadin
g
Bosto
n
Harbo
r
cso
Event
Total
Sample
No.
Aberjona
River
558
39
597
ABR006
187
39
-
-
226
ABR028
J 90
-
-
190
ABR049
184
-
-
-
184
Horn Pond
-
2
-
2
HOPCTR
-
2
-
-
2
Horn Pond
Brook
-
1
-
-
1
HOB002
-
1
-
-
1
Winter Pond
-
5
-
-
5
WIPCTR
-
5
-
-
5
Wedge Pond
-
33
-
-
33
WEPCTR
-
33
-
-
33
Upper Mystic
Lake
183
60
_
_
243
UPL001
183
-
-
183
UPLCTR
-
33
-
-
33
UPLUPL
-
27
-
-
27
Lower Mystic
Lake
2
2
LOLCTR
-
2
-
-
2
~ABR049
~ABR028
waterbody
AHOPCTR
~ Aberjona River
\ \ / ' Stone Zoo'Q j
A Horn Pond
ฆ Horn Pond Brook
WIPCTR ฆHOBt)02
-|- Lower Mystic Lake
^ฆWEPCTR
0 Upper Mystic Lake
)|(- Wedge Pond
~ABR006
Winter Pond
0UPLUPL
[gUPLCTR
0UPLOO1
-(-LOLCTR
Table A-2. Water Body/Monitoring Station and Program
Water body/
Monitoring
Station
Monitoring Program
Total
Monitoring
Events
Baseline
Phos.
Loading
Boston
Harbor
CSO
Event
Mill Brook
188
38
-
-
226
MIB001
188
-
-
-
188
MIB0045
-
38
-
-
38
Spy Pond
-
17
-
-
17
SPPCTR
-
17
-
-
17
Winns Brook
190
-
-
-
190
A-148
-------
Mystic River Watershed TMDL Alternative Development Final Report
Water body/
Monitoring Program
Total
Monitoring
Station
Baseline
Phos.
Loading
Boston
Harbor
cso
Event
Monitoring
Events
WIB001
190
-
-
-
190
Little River
-
-
-
442
442
MWRA174
-
-
-
442
442
Alewife Brook
192
39
-
1,393
1,624
ALB006
192
39
-
231
MWRA070
-
-
-
466
466
MWRA074
-
-
-
469
469
MWRA172
-
-
-
458
458
Mystic River
(Fresh)
190
109
995
2,0531
3,3291
MWRA056
-
-
-
365
365
MWRA057
-
-
-
373
373
MWRA059
-
-
-
367
367
MWRA066
-
-
548
-
548
MWRA067
-
-
-
70
370
MWRA083
-
-
261
85
646
MWRA177
-
-
186
194
3651
MYR071
190
38
-
-
228
MYR33
-
36
-
-
36
MYR43
-
35
-
-
35
Meetinghous
e Brook
191
13
_
_
204
MEB001
191
13
-
-
204
Maiden River
183
73
-
322
578
MAR003
-
36
-
-
36
Mystic River
(Salt)
1701
459
1,079
1,70s1
MWRA015
-
-
-
344
344
MWRA052
-
-
-
92
492
MWRA069
-
-
-
243
243
MWRA137
-
-
459
-
459
MYR275
82
-
-
-
82
MYRMMP
95
-
-
-
95
Mill Creek
93
-
-
-
93
MIC004
93
-
-
-
93
Chelsea River
95
-
-
370
465
CHR95S
95
-
-
-
95
MWRA027
-
-
-
370
370
Belle Isle Inlet
81
-
-
-
81
BEI001
13
-
-
-
13
BEI093
68
-
-
-
68
1.Sub-category values do not sum to this value due to sampling events at different
locations or programs occur same day.
A-149
-------
Mystic River Watershed TMDL Alternative Development Final Report
4tri
A11
(7.,
-^SPPCTR
VVIB001
$Vr071
^ ^'wr4Reboi)1
I^MWRAOSe
lr< vsrsilv \
^.L3C0
^MWRA172
iMm*w
j \
Sarocfvitlc
Harvard
i ฆ . Mity
O
Hnwarr-sij.mrr o Cambridge
.*$67
.1VVRA059
<*ป, ft.ฎ
IcwMfftseE
-*ฆ U4!i.-ซE-:f *
SfcWwl. y
7D .Niwrfl
Mv'svafijSe.ltf ^
ฉ e * * :>arKral-4>&nibHv
O - i'Jtrcrl' 3!JPe^
vvaterbody
Ale wife Brook
A Little River
ฆ Meetinghouse Brook
|- Mill Brook
ERI Mystic River (Fresh)
Spy Pond
Winns Brook
-sr
P>/ซ ;l
s,J
to 'Lu.
Ma: den *<-.
* >
,r
".i v
I.1AR036
ฆ
MWRA176
ฆ
MAR003
wr
TS>
MWl
Itomi
Sn,.
+
WC'004
;i*iwป f
M
131WYRI.WP 4k
UVVRAOl^MMW
Si
BEI093
BEI001
OU^'^OcnslrtL'.cr
j bo:
J 6nK
-------
Mystic River Watershed TMDL Alternative Development Final Report
Appendix B: Water Quality Parameters Included in
Monitoring Programs
Table B-1. Water Quality Parameters Included in Monitoring Programs
Parameter Code
Parameter Name/Description
Speciation
ATTENUATION_COEFFICIENT
Light attenuation coefficient
-
CHLA
Chlorophyll-a
-
DO
Dissolved Oxygen
-
DO_SAT
Dissolved Oxygen, %
saturation
_
ECOLI
Escherichia coli
-
ENT
Enterococcus
-
FCOLI
Fecal Coliform
-
NH3
Ammonia
as N
N02
Nitrite
as N
N023
Nitrate + Nitrite
as N
N03
Nitrate
as N
PH
PH
-
PHAEOPHYTIN
Pheophytin a
-
P04
Orthophosphate
asP
SALINITY
Salinity
-
SECCHI
Secchi Disk Depth
-
SPCOND
Specific conductance
-
TDN
Nitrogen, total dissolved
as N
TDP
Dissolved Phosphorus
asP
TEMP_WATER
Water Temperature
-
TN
Total Nitrogen
as N
TP
Total Phosphorus
asP
TPC
Total Particulate Carbon
-
TPN
Particulate Nitrogen
as N
TPP
Phosphorus, Particulate
Organic
asP
TSS
Total suspended solids
-
TURB
Turbidity
-
B-151
-------
Mystic River Watershed TMDL Alternative Development Final Report
Appendix C: Descriptive Figures of Water Quality
Data by Water body
Note that values marked as "dot" denote observations that extend above or below the nearest hinge
(i.e., the 25th or 75th percentile) by a distance exceeding 1.5 times the inter-quartile range. Unusual
values are marked by a red circle and additional information on these values are provided in
Appendix D.
Figure C-1. Attenuation coefficient results by water body.
C-152
-------
Mystic River Watershed TMDL Alternative Development Final Report
Figure C-2. Chlorophyll-a results by water body.
Figure C-3. Dissolved oxygen results by water body.
C-153
-------
My stic River Watershed TMDL Alternative Development Final Report
DO SAT
60D -
o
400-
a>
3
flj
>
0)
cn
200-
/
/
+ ^ + _
*
Figure C-4. Dissolved oxygen percent saturation results by water body
Figure C-5. Inorganic nitrogen (nitrite plus nitrate) results by water body.
C-154
-------
Mystic River Watershed TMDL Alternative Development Final Report
Figure C-6. pH results by water body.
PHAEOPHYTIN
20-
Figure C-7. Phaeophytin results by water body.
C-155
-------
Mystic River Watershed TMDL Alternative Development Final Report
SECCHI
6-
Figure C-8. Secchi depth results by water body.
Figure C-9. Total nitrogen results by water body.
C-156
-------
Mystic River Watershed TMDL Alternative Development Final Report
Figure C-10. Total phosphorus results by water body.
Figure C-11. Total suspended solids results by water body.
C-157
-------
Mystic River Watershed TMDL Alternative Development Final Report
Figure C-12. Turbidity results by water body
C-158
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Appendix D: Identified Extreme Values
Table D-1. Extreme values identified in Appendix C
Sample
Date &
Time
Monitori
ng
Program
Water
body
Name
Monitor
ing
Location
Paramet
er Name
Sample
Result
Uni
ts
Qu
al.
Fl
ag
ID
Flag
Descript
ion
Commen
t
Analytic
al
Method
Analytical
Method
Description
2/2/09
7:56
Baseline
Chelsea
River
CHR95S
Dissolve
d
Oxygen
26.66
mg
/I
-
E
instrume
nt error
YSI likely
not
calibrated
correctly
D888(B)
Dissolved
Oxygen by
Instrumental
Probe
6/25/03
7:36
CSO
Maiden
River
MWRA1
76
Dissolve
d
Oxygen
(%
Saturati
on)
616.7
%
-
-
-
-
D888(B)
Dissolved
Oxygen by
Instrumental
Probe
5/1/14
6:27
Baseline
Belle Isle
Inlet
BEI093
Total
Phos.
3.65
mg
/I
-
-
-
-
4500-P-E
Phosphorus in
Water by
Colorimetry-
Ascorbic Acid
Method
12/16/1
1 7:21
Baseline
Belle Isle
Inlet
BEI093
Total
Phos.
2.25
mg
/I
-
-
-
-
4500-P-E
Phosphorus in
Water by
Colorimetry-
Ascorbic Acid
Method
7/12/06
6:37
Baseline
Maiden
River
MAR036
Total
Phos.
1.9203
76
mg
/I
-
-
-
-
4500-P-J
Persulfate
Method for
Simultaneous
Determination
of Total
Nitrogen and
Total
Phosphorus
10/13/1
5 10:00
Phosphor
us
Loading
Horn
Pond
Brook
HOB002
Turbidity
4720
NT
U
-
-
-
-
180.1
Turbidity by
Nephelometry
D-159
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Appendix E: Modeled Stormwater TP Load and
Rainfall-Runoff Results
Table E-XI-1. Modeled Stormwater TP Load (Ibs./year; No Calibration)
Year
Critical Water Quality Segments
Impaired Ponds
Upper
Lobe
Basin
Upper
Mystic
Lake
Upper
Basin
Lower
Basin
Blacks
Nook Pond
(MA71005),
Cambridge
Horn Pond
(MA71019),
Woburn
Judkins
Pond
(MA71021),
Winchester
Mill Pond
(MA71031),
Winchester
Spy Pond
(MA71040),
Arlington
Wedge
Pond
(MA71045),
Winchester
Winter
Pond
(MA71047),
Winchester
1992
7,748
8,315
14,693
18,940
2.19
2,360
7,501
7,543
371
2,581
61
1993
5,428
5,737
10,408
13,718
0.73
1,490
5,260
5,291
288
1,646
36
1994
6,426
6,849
12,203
15,899
1.35
1,877
6,227
6,262
318
2,061
46
1995
5,868
6,276
11,211
14,556
1.44
1,726
5,684
5,716
294
1,894
43
1996
7,492
7,990
14,194
18,434
1.70
2,217
7,255
7,297
365
2,430
55
1997
3,749
3,939
7,327
9,771
0.28
951
3,635
3,657
218
1,060
22
1998
9,409
10,142
17,714
22,613
3.04
2,977
9,098
9,146
438
3,257
79
1999
6,852
7,367
13,067
16,801
2.02
2,087
6,628
6,664
341
2,292
55
2000
6,523
6,950
12,458
16,256
1.35
1,882
6,321
6,357
331
2,069
46
2001
4,004
4,218
7,697
10,193
0.40
1,073
3,880
3,904
218
1,190
25
2002
4,692
4,940
9,127
12,131
0.43
1,217
4,549
4,577
267
1,354
29
2003
5,449
5,758
10,419
13,739
0.68
1,500
5,282
5,313
288
1,659
35
2004
6,231
6,636
11,845
15,442
1.27
1,812
6,036
6,071
312
1,991
44
2005
5,451
5,763
10,519
13,878
0.76
1,478
5,283
5,314
295
1,635
36
2006
6,747
7,180
12,814
16,710
1.37
1,960
6,536
6,574
335
2,152
48
2007
7,376
7,926
13,968
17,976
2.16
2,264
7,141
7,181
350
2,475
58
2008
7,469
7,972
14,158
18,402
1.66
2,207
7,236
7,277
366
2,421
54
2009
5,235
5,521
10,146
13,454
0.55
1,381
5,076
5,106
293
1,534
32
2010
8,315
8,931
15,614
20,039
2.40
2,594
8,045
8,089
386
2,837
67
2011
6,559
6,950
12,543
16,483
1.02
1,832
6,356
6,394
342
2,022
44
E-160
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
2012
5,036
5,322
9,616
12,675
0.63
1,392
4,881
4,910
265
1,539
33
2013
5,491
5,844
10,498
13,707
1.09
1,574
5,318
5,349
283
1,734
39
2014
6,141
6,537
11,747
15,345
1.24
1,758
5,950
5,984
314
1,933
43
2015
4,077
4,288
7,947
10,586
0.32
1,045
3,953
3,977
235
1,165
24
2016
3,912
4,115
7,647
10,202
0.27
994
3,795
3,818
229
1,110
23
2017
5,236
5,544
10,061
13,246
0.78
1,442
5,075
5,105
278
1,593
35
Average
Annual
6,035
6,423
11,525
15,046
1
1,734
5,846
5,880
308
1,909
43
Table E-XI-2. Modeled Groundwater TP Load (Ibs./year; Concentration = 8 mg/L;
No Calibration)
Year
Critical Water Quality Segments
Impaired Ponds
Upper
Lobe
Basin
Upper
Mystic
Lake
Upper
Basin
Lower
Basin
Blacks
Nook Pond
(MA71005),
Cambridge1
Horn Pond
(MA71019),
Woburn
Judkins
Pond
(MA71021),
Winchester
Mill Pond
(MA71031),
Winchester
Spy Pond
(MA71040),
Arlington
Wedge
Pond
(MA71045),
Winchester
Winter
Pond
(MA71047),
Winchester
1992
562
597
1,102
1,432
0.15
168
544
547
32
184
4
1993
429
453
842
1,110
0.06
120
416
419
25
132
3
1994
522
552
1,021
1,339
0.10
151
506
509
30
165
4
1995
380
402
746
977
0.08
109
367
370
22
120
3
1996
634
673
1,238
1,612
0.15
189
613
617
36
207
5
1997
229
240
454
604
0.02
60
222
223
15
66
1
1998
744
790
1,450
1,881
0.19
226
719
723
42
247
6
1999
500
532
982
1,274
0.13
150
484
486
29
165
4
2000
490
519
962
1,259
0.10
142
475
477
29
155
3
2001
279
293
548
723
0.04
77
270
272
17
85
2
2002
337
354
666
883
0.04
90
327
329
21
99
2
2003
421
443
825
1,089
0.05
116
408
410
25
128
3
2004
524
555
1,024
1,339
0.11
153
507
510
30
168
4
2005
432
455
849
1,118
0.07
121
418
421
26
133
3
E-161
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
2006
648
686
1,264
1,651
0.14
191
627
630
37
209
5
2007
442
467
864
1,132
0.08
128
427
430
26
140
3
2008
641
678
1,252
1,638
0.13
187
620
624
37
205
4
2009
374
393
739
979
0.04
101
363
365
23
111
2
2010
685
727
1,333
1,733
0.16
207
662
666
38
226
5
2011
512
539
1,004
1,323
0.07
143
496
499
30
157
3
2012
336
354
662
875
0.05
93
326
328
20
102
2
2013
427
450
835
1,096
0.07
122
413
415
25
134
3
2014
492
520
965
1,266
0.09
141
477
479
29
155
3
2015
270
283
534
710
0.02
71
262
263
17
79
2
2016
251
263
497
662
0.02
65
243
245
16
73
1
2017
385
405
757
999
0.05
106
373
375
23
117
2
Average
Annual
459
486
901
1,181
0
132
445
447
27
145
3
1. All resu
ts rounc
ed to the nearest pound. Groundwater loads for Blacks Nook Pond are non-zero but
below the rounding level.
Table E-XI-3. Modeled Total Streamflow TP Load (Ibs./year) from Stormwater and
Groundwater (No Calibration and No Attenuation)
Year
Critical Water Quality Segments
Impaired Ponds
Upper
Lobe
Basin
Upper
Mystic
Lake
Upper
Basin
Lower
Basin
Blacks
Nook Pond
(MA71005),
Cambridge
Horn Pond
(MA71019),
Woburn
Judkins
Pond
(MA71021),
Winchester
Mill Pond
(MA71031),
Winchester
Spy Pond
(MA71040),
Arlington
Wedge
Pond
(MA71045),
Winchester
Winter
Pond
(MA71047),
Winchester
1992
8,310
8,913
15,795
20,372
2.34
2,528
8,045
8,090
403
2,764
65
1993
5,857
6,189
11,250
14,828
0.79
1,610
5,676
5,710
314
1,779
38
1994
6,948
7,401
13,224
17,238
1.45
2,028
6,732
6,771
348
2,226
49
1995
6,247
6,678
11,957
15,532
1.51
1,836
6,051
6,086
316
2,014
46
1996
8,126
8,663
15,432
20,046
1.86
2,407
7,869
7,914
401
2,637
60
1997
3,978
4,179
7,781
10,375
0.29
1,011
3,857
3,881
233
1,127
24
1998
10,152
10,932
19,165
24,494
3.24
3,203
9,816
9,869
480
3,503
84
E-162
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
1999
7,352
7,899
14,049
18,075
2.15
2,237
7,112
7,150
370
2,457
59
2000
7,014
7,469
13,420
17,516
1.45
2,023
6,795
6,834
360
2,224
50
2001
4,283
4,511
8,245
10,916
0.44
1,150
4,150
4,175
235
1,275
27
2002
5,029
5,294
9,793
13,015
0.47
1,307
4,876
4,906
288
1,453
31
2003
5,870
6,201
11,243
14,828
0.73
1,616
5,689
5,723
313
1,787
38
2004
6,755
7,190
12,869
16,781
1.37
1,965
6,543
6,581
342
2,159
48
2005
5,883
6,219
11,368
14,996
0.83
1,599
5,701
5,735
321
1,767
38
2006
7,394
7,866
14,079
18,361
1.51
2,151
7,163
7,204
372
2,361
53
2007
7,818
8,393
14,832
19,108
2.24
2,392
7,569
7,611
375
2,614
61
2008
8,110
8,650
15,411
20,040
1.80
2,394
7,856
7,900
403
2,626
59
2009
5,609
5,914
10,884
14,433
0.59
1,482
5,438
5,471
316
1,645
35
2010
9,000
9,657
16,947
21,771
2.56
2,801
8,707
8,755
424
3,063
72
2011
7,070
7,490
13,547
17,806
1.09
1,975
6,852
6,892
373
2,179
47
2012
5,373
5,677
10,278
13,550
0.67
1,484
5,207
5,238
285
1,641
35
2013
5,917
6,294
11,333
14,803
1.16
1,696
5,731
5,764
308
1,867
42
2014
6,633
7,057
12,712
16,611
1.34
1,900
6,426
6,464
343
2,088
47
2015
4,347
4,571
8,481
11,296
0.34
1,116
4,215
4,241
252
1,243
26
2016
4,163
4,379
8,143
10,863
0.29
1,059
4,039
4,063
245
1,182
24
2017
5,621
5,950
10,818
14,245
0.83
1,549
5,448
5,480
302
1,710
37
Average
Annual
6,495
6,909
12,425
16,227
1
1,866
6,291
6,327
335
2,054
46
Table E-XI-4. Modeled Stormwater Rainfall-Runoff (in-acre/year) Results
Attributable to Stormwater (No Calibration)
Critical Water Quality Segments
Impaired Ponds
Year
Upper
Lobe
Basin
Upper
Mystic
Lake
Upper
Basin
Lower
Basin
Blacks
Nook Pond
(MA71005),
Cambridge
Horn Pond
(MA71019),
Woburn
Judkins
Pond
(MA71021),
Winchester
Mill Pond
(MA71031),
Winchester
Spy Pond
(MA71040),
Arlington
Wedge
Pond
(MA71045),
Winchester
Winter
Pond
(MA71047),
Winchester
1992
163,532
172,857
299,273
386,996
34
48,984
158,337
159,243
7,458
53,541
1,201
E-163
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
1993
124,957
131,132
228,531
299,548
15
35,074
121,200
121,922
5,923
38,533
811
1994
151,887
159,854
277,229
361,607
23
43,887
147,257
148,122
7,020
48,090
1,026
1995
110,439
116,346
202,510
263,692
18
31,888
107,023
107,645
5,206
34,979
764
1996
184,500
194,666
336,333
435,720
35
55,105
178,670
179,708
8,282
60,185
1,334
1997
66,630
69,539
122,938
162,806
4
17,386
64,691
65,087
3,386
19,230
396
1998
216,286
228,638
394,132
508,494
45
65,734
209,281
210,472
9,710
71,879
1,618
1999
145,562
154,019
266,821
344,342
31
43,804
140,861
141,655
6,735
47,978
1,089
2000
142,610
150,205
261,067
340,099
23
41,244
138,230
139,038
6,647
45,187
980
2001
81,045
84,916
148,487
195,056
9
22,396
78,615
79,087
3,896
24,633
519
2002
98,038
102,569
180,500
238,187
8
26,240
95,165
95,740
4,882
28,956
599
2003
122,371
128,239
223,592
293,853
12
33,921
118,729
119,439
5,854
37,342
771
2004
152,491
160,591
278,158
361,786
25
44,629
147,742
148,606
6,987
48,869
1,064
2005
125,609
131,889
230,356
301,812
16
35,190
121,833
122,558
5,993
38,640
821
2006
188,424
198,535
343,313
445,967
33
55,633
182,546
183,616
8,494
60,793
1,329
2007
128,454
135,124
234,544
305,637
19
37,152
124,466
125,196
5,968
40,747
883
2008
186,341
196,318
340,070
442,459
30
54,472
180,579
181,632
8,558
59,652
1,291
2009
108,768
113,884
200,147
263,889
10
29,298
105,583
106,218
5,392
32,316
668
2010
199,181
210,309
362,159
468,354
38
60,157
192,826
193,935
8,871
65,750
1,455
2011
148,835
156,207
272,223
357,113
17
41,628
144,398
145,255
7,090
45,772
953
2012
97,844
102,650
179,543
235,965
10
26,972
94,958
95,523
4,748
29,693
616
2013
124,104
130,387
226,509
295,882
16
35,490
120,288
120,998
5,811
38,976
833
2014
143,166
150,681
262,035
341,763
22
41,170
138,785
139,599
6,695
45,131
973
2015
78,511
82,072
144,666
191,383
5
20,705
76,244
76,705
3,965
22,898
465
2016
72,989
76,278
134,516
178,297
4
19,059
70,913
71,341
3,724
21,118
420
2017
111,957
117,413
205,258
269,529
12
31,009
108,617
109,266
5,391
34,106
715
Average
Annual
133,636
140,589
244,420
318,855
20
38,393
129,532
130,292
6,257
42,115
908
E-164
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Table E-XI-5. Modeled Groundwater Flow (in-acre/year) Results
Year
Critical Water Quality Segments
Impaired Ponds
Upper
Lobe
Basin
Upper
Mystic
Lake
Upper
Basin
Lower
Basin
Blacks
Nook Pond
(MA71005),
Cambridge
Horn Pond
(MA71019),
Woburn
Judkins
Pond
(MA71021),
Winchester
Mill Pond
(MA71031),
Winchester
Spy Pond
(MA71040),
Arlington
Wedge
Pond
(MA71045),
Winchester
Winter
Pond
(MA71047),
Winchester
1992
310,282
329,668
608,065
790,428
80
92,834
300,072
301,790
17,708
101,464
2,269
1993
237,057
249,894
465,018
612,652
35
66,473
229,692
231,061
14,063
73,023
1,532
1994
288,158
304,720
563,690
739,100
54
83,174
279,075
280,713
16,668
91,133
1,939
1995
209,530
221,810
411,823
539,012
43
60,434
202,825
204,004
12,361
66,287
1,443
1996
350,059
371,192
683,395
890,007
83
104,434
338,607
340,574
19,664
114,054
2,520
1997
126,394
132,440
250,595
333,477
10
32,949
122,600
123,350
8,040
36,442
747
1998
410,390
436,066
800,564
1,038,307
106
124,578
396,619
398,876
23,055
136,215
3,056
1999
276,195
293,776
542,123
703,279
74
83,018
266,953
268,457
15,991
90,922
2,056
2000
270,564
286,351
530,850
695,146
55
78,165
261,968
263,499
15,783
85,633
1,851
2001
153,749
161,796
302,272
399,084
20-
42,445
148,987
149,883
9,250
46,680
979
2002
185,977
195,396
367,697
487,619
19
49,730
180,353
181,443
11,593
54,874
1,132
2003
232,145
244,344
455,090
601,152
28
64,287
225,010
226,356
13,899
70,766
1,457
2004
289,316
306,154
565,466
739,317
59
84,581
279,994
281,631
16,590
92,610
2,009
2005
238,295
251,350
468,759
617,306
38
66,691
230,892
232,267
14,230
73,226
1,552
2006
357,496
378,515
697,750
911,177
78
105,436
345,956
347,984
20,167
115,211
2,515
2007
243,707
257,573
476,959
624,767
46
70,411
235,882
237,266
14,171
77,219
1,668
2008
353,536
374,279
691,310
904,163
72
103,235
342,228
344,224
20,319
113,047
2,442
2009
206,333
216,968
407,645
540,156
23
55,525
200,096
201,300
12,802
61,242
1,262
2010
377,926
401,058
735,670
956,458
91
114,009
365,440
367,542
21,063
124,605
2,754
2011
282,354
297,679
553,937
730,413
40
78,893
273,658
275,283
16,834
86,742
1,802
2012
185,615
195,606
365,464
482,757
25
51,118
179,961
181,031
11,273
56,270
1,164
2013
235,449
248,509
460,735
604,953
39
67,260
227,965
229,310
13,797
73,861
1,573
2014
271,616
287,239
532,893
698,638
52
78,025
263,020
264,564
15,897
85,527
1,840
E-165
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
2015
148,931
156,333
294,781
391,900
12
39,241
144,494
145,368
9,415
43,394
879
2016
138,452
145,290
274,137
365,152
9
36,120
134,392
135,202
8,841
40,020
793
2017
212,390
223,734
417,794
551,403
29
58,768
205,847
207,077
12,800
64,634
1,351
Average
Annual
253,535
267,990
497,095
651,836
47
72,763
245,484
246,925
14,857
79,811
1,714
Table E-XI-6. Modeled Total Streamflow (in-acre/year; Stormflow + Groundwater)
Results
Year
Critical Water Quality Segments
Impaired Ponds
Upper
Lobe
Basin
Upper
Mystic
Lake
Upper
Basin
Lower
Basin
Blacks
Nook Pond
(MA71005),
Cambridge
Horn Pond
(MA71019),
Woburn
Judkins
Pond
(MA71021),
Winchester
Mill Pond
(MA71031),
Winchester
Spy Pond
(MA71040),
Arlington
Wedge
Pond
(MA71045),
Winchester
Winter
Pond
(MA71047),
Winchester
1992
473,814
502,525
907,338
1,177,424
114
141,818
458,409
461,033
25,166
155,005
3,470
1993
362,014
381,026
693,550
912,199
50
101,547
350,892
352,982
19,986
111,556
2,343
1994
440,045
464,573
840,920
1,100,707
76
127,061
426,332
428,835
23,688
139,223
2,965
1995
319,969
338,156
614,333
802,704
61
92,322
309,847
311,649
17,567
101,265
2,207
1996
534,559
565,859
1,019,728
1,325,728
118
159,539
517,277
520,282
27,947
174,239
3,855
1997
193,024
201,978
373,533
496,283
14
50,335
187,291
188,437
11,426
55,672
1,143
1998
626,677
664,704
1,194,696
1,546,801
151
190,312
605,899
609,347
32,765
208,095
4,674
1999
421,756
447,794
808,944
1,047,621
105
126,822
407,813
410,112
22,725
138,900
3,145
2000
413,174
436,556
791,917
1,035,245
78
119,408
400,198
402,537
22,430
130,820
2,830
2001
234,794
246,713
450,759
594,140
29
64,841
227,602
228,970
13,147
71,313
1,498
2002
284,014
297,965
548,197
725,806
28
75,969
275,518
277,183
16,475
83,830
1,731
2003
354,516
372,583
678,682
895,005
40
98,208
343,739
345,795
19,753
108,108
2,228
2004
441,808
466,744
843,624
1,101,102
84
129,210
427,736
430,237
23,577
141,479
3,073
2005
363,904
383,239
699,115
919,117
54
101,881
352,725
354,825
20,224
111,866
2,373
2006
545,920
577,050
1,041,063
1,357,145
110
161,069
528,502
531,600
28,660
176,004
3,844
2007
372,161
392,697
711,503
930,404
65
107,563
360,348
362,462
20,140
117,966
2,551
2008
539,877
570,597
1,031,380
1,346,621
103
157,707
522,808
525,856
28,876
172,699
3,733
E-166
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Year
Critical Water Quality Segments
Impaired Ponds
Upper
Lobe
Basin
Upper
Mystic
Lake
Upper
Basin
Lower
Basin
Blacks
Nook Pond
(MA71005),
Cambridge
Horn Pond
(MA71019),
Woburn
Judkins
Pond
(MA71021),
Winchester
Mill Pond
(MA71031),
Winchester
Spy Pond
(MA71040),
Arlington
Wedge
Pond
(MA71045),
Winchester
Winter
Pond
(MA71047),
Winchester
2009
315,101
330,853
607,792
804,045
33
84,823
305,679
307,517
18,195
93,558
1,930
2010
577,106
611,367
1,097,829
1,424,813
129
174,166
558,267
561,477
29,934
190,354
4,209
2011
431,189
453,885
826,160
1,087,526
57
120,521
418,056
420,539
23,924
132,514
2,755
2012
283,459
298,256
545,007
718,721
35
78,090
274,919
276,554
16,021
85,964
1,780
2013
359,552
378,896
687,244
900,835
55
102,750
348,254
350,308
19,609
112,837
2,406
2014
414,783
437,920
794,928
1,040,401
74
119,195
401,805
404,163
22,592
130,658
2,814
2015
227,442
238,405
439,447
583,284
18
59,946
220,738
222,073
13,381
66,293
1,344
2016
211,442
221,568
408,654
543,449
13
55,179
205,305
206,543
12,565
61,138
1,213
2017
324,347
341,148
623,052
820,932
42
89,777
314,464
316,343
18,191
98,740
2,067
Average
Annual
387,170
408,579
741,515
970,692
67
111,156
375,016
377,217
21,114
121,926
2,622
E-167
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Appendix F: Baseflow Estimates for Aberjona
River and Alewife Brook
Table F-XI-7. Baseflow Estimates for Aberjona River and Alewife Brook
Year
Aberjona River (1992 - 2016)
Alewife Brook (2006 - 2016)
Streamflow
(in-acre/year)
Baseflow (in-
acre/year)
Baseflow
Fraction
Streamflow
(in-acre/year)
Baseflow (in-
acre/year)
Baseflow
Fraction
1992
242,453
160,274
0.66
1993
287,207
193,927
0.68
1994
302,226
203,394
0.67
1995
199,736
145,786
0.73
1996
430,248
270,572
0.63
1997
223,833
162,732
0.73
1998
431,381
268,083
0.62
1999
233,642
155,263
0.66
2000
300,981
193,870
0.64
2001
316,739
195,302
0.62
2002
243,445
155,279
0.64
2003
360,401
238,359
0.66
2004
336,903
208,603
0.62
2005
386,938
266,063
0.69
2006
491,444
299,860
0.61
101,411
65,492
0.65
2007
293,250
184,530
0.63
70,514
47,048
0.67
2008
471,436
297,944
0.63
92,796
59,443
0.64
2009
394,291
266,786
0.68
76,214
54,396
0.71
2010
482,776
269,709
0.56
106,148
64,646
0.61
2011
444,306
290,130
0.65
102,179
71,934
0.70
2012
224,915
131,197
0.58
77,111
57,112
0.74
2013
255,365
189,513
0.74
70,244
53,077
0.76
2014
381,277
248,166
0.65
85,295
61,040
0.72
2015
261,463
179,430
0.69
73,946
54,779
0.74
2016
210,791
142,541
0.68
69,169
52,276
0.76
2017
312,960
210,840
0.67
95,648
72,209
0.75
Avg.
327,668
212,591
0.65
85,056
59,454
0.70
CV
0.28
0.25
0.065
0.16
0.12
0.069
F-168
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Appendix G: Maps Depicting CSO Drainage Basins
SOMOOi
SOM004-
MEDFORD
SOM002A/00:
T\ SOM001-
IOM007 U
CAM401
CAM400
-SOM007i
SOMOOf
/R205
CAM004
CAM401A-
SOMERVILLE
BOS 017-
Figure G-XI-1. MWRA CSO Map (updated in September 2015).
ARLINGTON
CAiViLitlDGE
^ Roadi
1,250
~ Feet
2.500
Legend
| Town Boundaries V7A CSO (Cambridge)
Vteterbodies KX>0i CSO (Somervilie)
Roads CSO Outfalls (MWRA)
Interstate ACTIVE
US Highway CLOSED
State Route SPECIAL (Treated CSO)
Mystic River Subbasins
Alewife Brook
Lower Mystic LaKe
Maiden River
Mystic River
! Spy Pond
MALDEN
G-169
-------
My stic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Alewife Sub-Watershed Eutrophication Study
City of Cambridge: 2000
| Separated Area ฉ CSO Outfall
j Combined Sewer Area Watershed Boundary
[;"ฆ -y ; I Separated (With Loading From Town of Belmont)
] Separated (Wrth Active Common Manholes)
Figure G-XI-2. City of Cambridge CSO Drainage Basins (2000).
G-170
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
BOSTON
Alewife Sub-Watershed Eutrophication Study
City of Cambridge: 2017
| Separated Area ฉ CSO Outfall
I I Combined Sewer Area Watershed Boundary
X] Separated (With Loading From Town of Belmont)
Figure G-XI-3. City of Cambridge CSO Drainage Basins (2017).
G-171
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Legend
Area
MWRA Connection
Combined Sewer Overflow
Somerville Marginal Interceptor (SMI)
Mystic River
Alewife Brook Conduit (ABC)
Alewife Brook
Cambridge Branch Sewer (CBS)
Interactions with SMI and McGrath system
Primary to CBS with overflows to ABC
Alewife Brook, plus interactions with CBS
Figure G-XI-4. City of Somerville CSO Drainage Basins (2017).
G-172
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Appendix H: Water Quality Data Used in
Calibration and Validation of the Bathtub Model
Year
Segment
WQ Site ID
TP
Chl-a
2010
Upper Lobe
UPLUPL
2014
Upper Lobe
UPLUPL
2015
Upper Lobe
UPLUPL
0.052
16.790
2016
Upper Lobe
UPLUPL
0.060
17.103
2017
Upper Lobe
UPLUPL
0.053
13.351
2010
Upper Lake
UPLCTR
2010
Upper Lake
UPL001
0.029
2014
Upper Lake
UPLCTR
2014
Upper Lake
UPL001
0.038
2015
Upper Lake
UPLCTR
0.035
8.929
2015
Upper Lake
UPL001
0.029
2016
Upper Lake
UPLCTR
0.037
8.709
2016
Upper Lake
UPL001
0.022
2017
Upper Lake
UPLCTR
0.036
13.257
2017
Upper Lake
UPL001
2010
Lower Lake
MYR071
0.036
2014
Lower Lake
MYR071
0.034
2015
Lower Lake
MYR071
0.036
2016
Lower Lake
MYR071
0.036
2017
Lower Lake
MYR071
0.038
2010
Upper Basin
MWRA083
0.036
7.158
2010
Upper Basin
MWRA066
0.050
7.015
2014
Upper Basin
MWRA083
0.044
8.379
2014
Upper Basin
MWRA066
0.045
9.023
2015
Upper Basin
MYR43
0.056
17.944
2016
Upper Basin
MYR43
0.072
21.273
2017
Upper Basin
MYR43
0.062
22.476
2010
Lower Basin
MYR33
2010
Lower Basin
MAR003
2010
Lower Basin
MWRA167
0.054
14.582
2014
Lower Basin
MYR33
2014
Lower Basin
MAR003
2014
Lower Basin
MWRA167
0.047
15.574
2015
Lower Basin
MYR33
0.064
23.771
2015
Lower Basin
MAR003
0.059
22.881
2015
Lower Basin
MWRA167
0.054
23.951
2016
Lower Basin
MYR33
0.092
29.825
2016
Lower Basin
MAR003
0.094
26.699
2016
Lower Basin
MWRA167
0.080
35.384
J-173
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Year
Segment
WQ Site ID
TP
Chl-a
2017
Lower Basin
MYR33
0.072
26.674
2017
Lower Basin
MAR003
0.065
24.750
2017
Lower Basin
MWRA167
0.062
30.367
J-174
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Appendix I: Bathtub Model Inputs for Calibration
Calculation Input Worksheet (English units)
Type
Parameter
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
Atmospheric
Precipitation
(m/yr.)
34.8
34.8
34.8
34.8
34.8
Atmospheric
Lake Evaporation
(m/yr.)
37.7
37.7
37.7
37.7
37.7
Atmospheric
Total P Load
(lb./ac/yr.)
0.27
0.27
0.27
0.27
0.27
Atmospheric
Inorg P Fraction (-)
50%
50%
50%
50%
50%
Atmospheric
Inorg P Load
(lb./ac/yr.)
0.13
0.13
0.13
0.13
0.13
Atten*
External
Area (mi2)
24.72
5.87
10.28
Atten*
External
Flow (ac-in/yr.)
226,408
95,529
123,300
Atten*
External
Flow (in/yr.)
14.31
25.43
18.74
Atten*
External
P Load (lb./yr.)
2,701
1,108
2,312
Atten*
External
P Load (lb./ac/yr.)
0.17
0.29
0.35
Sub-basin*
Area (mi2)
0.21
1.90
6.59
5.50
0.97
Sub-basin*
Flow (ac-in/yr.)
1,034
10,963
49,212
56,766
20,881
Sub-basin*
Flow (in/yr.)
7.69
9.02
11.67
16.13
33.64
Sub-basin*
P Load (lb./yr.)
21
225
996
1,150
623
Sub-basin*
P Load (lb./ac/yr.)
0.16
0.18
0.24
0.33
1.00
Ext* + Sub-
basin
Area (mi2)
24.93
1.90
6.59
11.37
11.25
Ext* + Sub-
basin
Flow (ac-in/yr.)
227,442
10,963
49,212
152,295
144,181
Ext* + Sub-
basin
Flow (in/yr.)
14.26
9.02
11.67
20.93
20.03
Receiving
Water
P Load (lb./yr.)
2,722
225
996
2,257
2,935
Ext* + Sub-
basin
P Load (lb./ac/yr.)
0.17
0.18
0.24
0.31
0.41
Ext* + Sub-
basin
P Cone (mg/L)
0.05
0.09
0.09
0.07
0.09
Ext* + Sub-
basin
Inorg P Fraction (-)
60%
60%
50%
50%
50%
Ext* + Sub-
basin
Inorg P Cone
(mg/L)
0.03
0.05
0.04
0.03
0.04
Ext* + Sub-
basin
Org P Cone (mg/L)
0.02
0.04
0.04
0.03
0.04
J-175
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Receiving
Water
Surface Area (ac)
35.6
140.9
92.8
56.9
124.2
Receiving
Water
Depth (ft)
10.0
42.5
30.8
3.8
4.5
Receiving
Water
Length (mi)
0.36
0.8
0.6
3.6
1.3
Receiving
Water
Epi Depth (ft)
10.0
10.0
10.0
3.8
4.5
Receiving
Water
Hypo Depth (ft)
0.0
32.5
20.8
0.0
0.0
Receiving
Water
Volume (ac-ft)
177.9
5987.5
2861.5
215.9
553.5
Receiving
Water
Cum Flow (ac-
in/yr.)
227,442
238,405
287,617
439,912
584,093
Receiving
Water
Retention Time (d)
3.4
110.0
43.6
2.1
4.2
Receiving
Water
Total P Cone
(mg/L)
0.05
0.03
0.04
0.06
0.06
Receiving
Water
Inorg P Fraction (-)
60%
60%
60%
60%
60%
Receiving
Water
Inorg P Cone
(mg/L)
0.03
0.02
0.02
0.03
0.04
Receiving
Water
Org P Cone (mg/L)
0.02
0.01
0.01
0.02
0.02
Receiving
Water
Chl-a Avg Cone
(mr/l)
16.8
8.9
4.7
17.9
23.5
Receiving
Water
Non-algal Turbidity
(1/ft)
0.17
0.17
0.17
0.17
0.17
Receiving
Water
Internal TP load
(lb./ac/yr.)
19.5
3.3
3.3
13.0
13.0
* Atten attenuated, Ext external, Sub-basin unattenuated local sub-basin
J-176
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model Input Worksheet (Metric units)
Case Title
Mystic River Alternative TMDL - Final Calibration
Number of Segments
5
Number of Tributaries
5
Number of Channels
0
Global Variables
Mean
CY
Averaging Period (yrs.)
1
0
Precipitation (m)
0.90
0
Evaporation (m)
1
0
Storage Increase (m)
0
0
Atmos. Loads (kg/km2-yr) Mean CY
Total P
30
0.5
Ortho P
15
0.5
Segment Data
Segment Number-
Segment Name
Outflow Segment Number-
Segment Group Number
1
2
3
4
5
Upper Lobe
2
1
Upper Lake
3
1
Lower Lake
4
1
Upper Basin
5
1
Lower Basin
0
1
Segment Morphometry
Surface Area (km2)
0.14
0.57
0.38
0.23
0.50
Mean Depth (m)
1.52
12.95
9.40
1.16
1.36
Length (km)
0.58
1.29
0.97
5.79
2.09
Mixed Depth (m)
1.52
3.05
3.05
1.16
1.36
Hypol. Depth (m)
0.00
9.91
6.35
0.00
0.00
Observed Water Quality
Non-Algal Turb (1/m)
0.55
0.55
0.55
0.55
0.55
Conservative Subst
0.00
0.00
0.00
0.00
0.00
Total P (ppb)
52.00
32.00
36.00
56.00
59.00
Chlorophyll-a (ppb)
16.80
8.90
4.70
17.90
23.50
Total P - Ortho P (ppb)
31
19
22
34
35
Segment Calibration Factors
Dispersion Rate
1
1
1
1
1
Total P
1
1
1
1
1
Chlorophyll-a
1
1
1
1
1
J-177
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Total P - Ortho P (ppb)
1
1
1
1
1
Internal Loading-Rates (me:/m2-
dav)
Total P
6
1
1
4
4
Tributary Data
Tributary Number-
1
2
3
4
5
Tributary Name
Segment Number-
Tributary Type Code
Upper Lobe
1
1
Upper Mystic
2
1
Lower Mystic
3
1
Upper Basin
4
1
Lower Basin
5
1
Drainage Area (km2)
Flow (hm3/yr.)
Total P (ppb)
Ortho P (ppb)
64.57
23.40
53
32
4.92
1.10
90
54
17.07
5.10
89
45
29.45
15.70
65
33
29.14
14.80
90
45
NonPoint Source Areas (km2)
Not used
0
0
0
0
0
Non-Point Source Export Coefficients
Not used
0
0
0
0
0
Transport Channels
Not used
0
0
0
0
0
Model Coefficients (Mean, CV)
Dispersion Rate
Total Phosphorus
Chl-a Model
TP-OP Model
HODv Model
MODv Model
Minimum Qs (m/yr.)
Chl-a Flushing Term
Chl-a Temporal CY
Availability Factor - Total P
Availability Factor - Ortho P
0.20
1
0.60
1
1
1
0.10
1
0.62
0.33
1.93
0.70
0.45
0.26
0.15
0.15
0.22
0
0
0
0
0
Model Options
Phosphorus Balance
Chlorophyll-a
Dispersion
1
2
1
J-178
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Phosphorus Calibration
1
Error Analysis
1
Availability Factors
0
Mass-Balance Tables
1
Output Destination
2
Appendix J: Bathtub Model Inputs and Outputs for
Scenarios
Table J-XI-8. Detailed BATHTUB Model Inputs and Outputs for Average Annual
Data
Model
Parameter
Scenario -
Upper
Upper
Lower
Upper
Lower
Input/
Run
Lobe
Lake
Lake
Basin
Basin
Output
l-l
371,417
148,218
192,965
la - 7
371,417
148,218
192,965
2-9
371,386
147,525
192,963
Atten Trib Flow
2a -14
371,386
147,525
192,963
(ac-in/yr.)
3-16
371,386
157,040
192,963
3a - 21
371,386
157,040
192,963
4-23
371,386
185,594
192,963
4a - 28
371,386
185,594
192,963
1-1
1,856
20,243
84,185
90,239
30,380
la - 7
1,856
20,243
84,185
90,239
30,380
2-9
1,855
20,238
84,159
90,202
30,185
Sub-basin Flow
2a - 14
1,855
20,238
84,159
90,202
30,185
(ac-in/yr.)
3-16
1,855
20,238
84,159
90,331
31,315
Input
3a - 21
1,855
20,238
84,159
90,331
31,315
4-23
1,855
20,238
84,159
90,719
34,701
4a - 28
1,855
20,238
84,159
90,719
34,701
1-1
373,272
20,243
84,185
238,457
223,345
la - 7
373,272
20,243
84,185
238,457
223,345
2-9
373,241
20,238
84,159
237,727
223,148
Total Flow (ac-
2a - 14
373,241
20,238
84,159
237,727
223,148
in/yr.)
3-16
373,241
20,238
84,159
247,371
224,278
3a - 21
373,241
20,238
84,159
247,371
224,278
4-23
373,241
20,238
84,159
276,314
227,664
4a - 28
373,241
20,238
84,159
276,314
227,664
Atten Trib P Load
(Ib./yr.)
1-1
3,873
1,637
3,130
la - 7
1,476
872
1,186
2-9
3,858
1,352
3,128
J-179
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
2a -14
1,640
644
1,329
3-16
3,858
1,442
3,128
3a - 21
1,676
667
1,358
4-23
3,858
1,603
3,128
4a - 28
1,748
733
1,416
1-1
34
416
1,558
1,661
719
la - 7
14
161
625
681
421
2-9
34
410
1,526
1,616
579
Sub-basin P Load
2a - 14
15
175
663
710
303
(Ib./yr.)
3-16
34
410
1,526
1,618
577
3a - 21
15
178
677
725
294
4-23
34
410
1,526
1,624
570
4a - 28
16
186
705
757
264
1-1
3,907
416
1,558
3,298
3,849
la - 7
1,490
161
625
1,553
1,608
2-9
3,892
410
1,526
2,968
3,707
Total P Load
2a - 14
1,655
175
663
1,354
1,632
(Ib./yr.)
3-16
3,892
410
1,526
3,033
3,704
3a - 21
1,691
178
677
1,393
1,652
4-23
3,892
410
1,526
3,227
3,697
4a - 28
1,763
186
705
1,490
1,680
1-1
695
459
302
741
1,619
la - 7
466
307
202
497
1,084
2-9
695
459
302
741
1,619
P Sediment load
2a - 14
480
317
209
512
1,117
(Ib./yr.)
3-16
695
459
302
741
1,619
3a - 21
480
317
209
512
1,117
4-23
695
459
302
741
1,619
4a - 28
494
326
215
526
1,149
1-1
0.78
1.35
1.31
1.00
1.15
la - 7
0.78
1.35
1.31
1.00
1.15
2-9
0.78
1.35
1.31
1.00
1.15
N cone (mg/L)
2a - 14
0.78
1.35
1.31
1.00
1.15
3-16
0.78
1.35
1.31
0.96
1.15
3a - 21
0.78
1.35
1.31
0.96
1.15
4-23
0.78
1.35
1.31
0.85
1.14
4a - 28
0.78
1.35
1.31
0.85
1.14
Inorg. N cone
1-1
0.47
0.81
0.78
0.60
0.69
(mg/L)
la - 7
0.47
0.81
0.78
0.60
0.69
J-180
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
2-9
0.47
0.81
0.78
0.60
0.69
2a - 14
0.47
0.81
0.78
0.60
0.69
3-16
0.47
0.81
0.78
0.58
0.69
3a - 21
0.47
0.81
0.78
0.58
0.69
4-23
0.47
0.81
0.78
0.51
0.68
4a - 28
0.47
0.81
0.78
0.51
0.68
1-1
0.05
0.09
0.08
0.06
0.08
la - 7
0.02
0.04
0.03
0.03
0.03
2-9
0.05
0.09
0.08
0.06
0.07
P cone (mg/L)
2a - 14
0.02
0.04
0.04
0.03
0.03
3-16
0.05
0.09
0.08
0.05
0.07
3a - 21
0.02
0.04
0.04
0.03
0.03
4-23
0.05
0.09
0.08
0.05
0.07
4a - 28
0.02
0.04
0.04
0.02
0.03
1-1
0.03
0.05
0.04
0.03
0.04
la - 7
0.01
0.02
0.02
0.01
0.02
2-9
0.03
0.05
0.04
0.03
0.04
Inorg. P Cone
2a - 14
0.01
0.02
0.02
0.01
0.02
(mg/L)
3-16
0.03
0.05
0.04
0.03
0.04
3a - 21
0.01
0.02
0.02
0.01
0.02
4-23
0.03
0.05
0.04
0.03
0.04
4a - 28
0.01
0.02
0.02
0.01
0.02
1-1
0
la - 7
67
2-9
0
Stormwater load
2a - 14
62
reduction (%)
3-16
0
3a - 21
61
4-23
0
4a - 28
61
1-1
0
la - 7
33
2-9
0
P sediment load
2a - 14
31
reduction (%)
3-16
0
3a - 21
31
4-23
0
4a - 28
29
Output
1-1
43.4
33.1
33.6
46.2
57.7
J-181
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
la - 7
20.8
18.6
19.4
25.3
31.1
2-9
43.3
33.0
33.3
44.2
55.8
Predicted P cone
(|ig/L)
2a -14
22.4
19.8
20.4
24.9
31.1
3-16
43.3
33.0
33.3
44.2
55.6
3a - 21
22.7
20.0
20.6
24.9
31.1
4-23
43.2
32.9
33.2
43.4
54.5
4a - 28
23.4
20.4
21.1
25
31.1
Predicted chl-a
1-1
13.9
7.6
7.7
16.4
19.4
cone (ng/L)
la - 7
5.9
4.1
4.3
8.1
10.0
2-9
13.9
7.6
7.7
15.6
18.8
2a - 14
6.5
4.4
4.6
7.9
9.9
3-16
13.8
7.6
7.6
15.6
18.7
3a - 21
6.6
4.5
4.6
7.9
9.9
4-23
13.8
7.5
7.6
15.3
18.3
4a - 28
6.9
4.6
4.7
7.9
9.9
Table J-XI-9. Detailed BATHTUB Model Inputs and Outputs for Wet and Dry Year
Data
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
Input
la - 7 (Wet)
574,239
218,660
285,460
la - 7 (Dry)
210,500
88,880
115,436
2a -14
(Wet)
574,105
216,564
285,449
Atten Trib Flow
(ac-in/yr.)
2a -14 (Dry)
210,499
89,101
115,436
3a - 21
(Wet)
574,105
230,094
285,449
3a - 21 (Dry)
210,499
95,041
115,436
4a - 28
(Wet)
574,105
270,695
285,449
4a - 28 (Dry)
210,499
112,870
115,436
la - 7 (Wet)
3,137
34,308
135,055
135,664
42,401
la - 7 (Dry)
942
10,127
45,401
52,853
19,407
2a -14
(Wet)
3,133
34,284
134,942
135,512
41,679
Sub-basin Flow
2a -14 (Dry)
942
10,127
45,401
52,853
19,488
(ac-in/yr.)
3a - 21
(Wet)
3,133
34,284
134,942
135,703
43,217
3a - 21 (Dry)
942
10,127
45,401
52,930
20,221
4a - 28
(Wet)
3,133
34,284
134,942
136,277
47,825
J-182
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
4a - 28 (Dry)
942
10,127
45,401
53,161
22,418
Total Flow (ac-
in/yr.)
la - 7 (Wet)
577,376
34,308
135,055
354,324
327,860
la - 7 (Dry)
211,442
10,127
45,401
141,733
134,842
2a -14
(Wet)
577,238
34,284
134,942
352,077
327,128
2a -14 (Dry)
211,442
10,127
45,401
141,953
134,923
3a - 21
(Wet)
577,238
34,284
134,942
365,797
328,666
3a - 21 (Dry)
211,442
10,127
45,401
147,970
135,657
4a - 28
(Wet)
577,238
34,284
134,942
406,972
333,274
4a - 28 (Dry)
211,442
10,127
45,401
166,031
137,854
Atten Trib P Load
(Ibs./yr.)
la - 7 (Wet)
2,127
1,607
1,579
la - 7 (Dry)
960
350
827
2a -14
(Wet)
2,303
818
1,752
2a -14 (Dry)
1,082
482
932
3a - 21
(Wet)
2,351
854
1,789
3a - 21 (Dry)
1,106
492
953
4a - 28
(Wet)
2,448
955
1,864
4a - 28 (Dry)
1,155
518
995
Sub-basin P Load
(Ibs./yr.)
la - 7 (Wet)
27
301
1,089
1,151
865
la - 7 (Dry)
7
79
353
410
170
2a -14
(Wet)
25
304
1,053
1,066
371
2a -14 (Dry)
8
90
398
462
244
3a - 21
(Wet)
26
310
1,073
1,085
365
3a - 21 (Dry)
9
92
407
473
232
4a - 28
(Wet)
27
322
1,112
1,124
343
4a - 28 (Dry)
9
96
425
496
193
Total P Load
(Ibs./yr.)
la - 7 (Wet)
2,153
301
1,089
2,758
2,445
la - 7 (Dry)
968
79
353
760
998
2a -14
(Wet)
2,328
304
1,053
1,884
2,124
2a -14 (Dry)
1,090
90
398
944
1,177
3a - 21
(Wet)
2,377
310
1,073
1,939
2,155
3a - 21 (Dry)
1,115
92
407
965
1,186
J-183
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
4a - 28
(Wet)
2,474
322
1,112
2,079
2,206
4a - 28 (Dry)
1,164
96
425
1,015
1,188
P Sediment load
(Ib./yr.)
la - 7 (Wet)
466
307
202
497
1,084
la - 7 (Dry)
466
307
202
497
1,084
2a -14
(Wet)
480
317
209
512
1,117
2a -14 (Dry)
480
317
209
512
1,117
3a - 21
(Wet)
480
317
209
512
1,117
3a - 21 (Dry)
480
317
209
512
1,117
4a - 28
(Wet)
494
326
215
526
1,149
4a - 28 (Dry)
494
326
215
526
1,149
N cone (mg/L)
la - 7 (Wet)
0.506
0.794
0.814
0.673
0.782
la - 7 (Dry)
1.383
2.691
2.422
1.683
1.900
2a -14
(Wet)
0.507
0.795
0.815
0.677
0.783
2a -14 (Dry)
1.383
2.691
2.422
1.680
1.899
3a - 21
(Wet)
0.507
0.795
0.815
0.650
0.782
3a - 21 (Dry)
1.383
2.691
2.422
1.607
1.894
4a - 28
(Wet)
0.507
0.795
0.815
0.579
0.777
4a - 28 (Dry)
1.383
2.691
2.422
1.420
1.878
Inorg N cone
(mg/L)
la - 7 (Wet)
0.304
0.477
0.489
0.404
0.469
la - 7 (Dry)
0.830
1.614
1.453
1.010
1.140
2a -14
(Wet)
0.304
0.477
0.489
0.406
0.470
2a -14 (Dry)
0.830
1.614
1.453
1.008
1.139
3a - 21
(Wet)
0.304
0.477
0.489
0.390
0.469
3a - 21 (Dry)
0.830
1.614
1.453
0.964
1.137
4a - 28
(Wet)
0.304
0.477
0.489
0.348
0.466
4a - 28 (Dry)
0.830
1.614
1.453
0.852
1.127
P cone (mg/L)
la - 7 (Wet)
0.016
0.039
0.036
0.034
0.033
la - 7 (Dry)
0.020
0.035
0.034
0.024
0.033
2a -14
(Wet)
0.018
0.039
0.034
0.024
0.029
2a -14 (Dry)
0.023
0.039
0.039
0.029
0.038
3a - 21
(Wet)
0.018
0.040
0.035
0.023
0.029
J-184
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
3a - 21 (Dry)
0.023
0.040
0.040
0.029
0.039
4a - 28
(Wet)
0.019
0.041
0.036
0.023
0.029
4a - 28 (Dry)
0.024
0.042
0.041
0.027
0.038
Inorg P Cone
(mg/L)
la - 7 (Wet)
0.010
0.023
0.018
0.017
0.016
la - 7 (Dry)
0.012
0.021
0.017
0.012
0.016
2a -14
(Wet)
0.011
0.023
0.017
0.012
0.014
2a -14 (Dry)
0.014
0.023
0.019
0.015
0.019
3a - 21
(Wet)
0.011
0.024
0.018
0.012
0.014
3a - 21 (Dry)
0.014
0.024
0.020
0.014
0.019
4a - 28
(Wet)
0.011
0.025
0.018
0.011
0.015
4a - 28 (Dry)
0.015
0.025
0.021
0.013
0.019
Stormwater load
reduction (%)
la - 7 (Wet)
67
la - 7 (Dry)
67
2a -14
(Wet)
62
2a -14 (Dry)
62
3a - 21
(Wet)
61
3a - 21 (Dry)
61
4a - 28
(Wet)
59
4a - 28 (Dry)
59
P sediment load
reduction (%)
la - 7 (Wet)
33
la - 7 (Dry)
33
2a -14
(Wet)
31
2a -14 (Dry)
31
3a - 21
(Wet)
31
3a - 21 (Dry)
31
4a - 28
(Wet)
29
4a - 28 (Dry)
29
Output
Predicted P cone
(Mg/L)
la - 7 (Wet)
18.9
17.9
19.8
26.2
30.5
la - 7 (Dry)
24.9
20.3
20.1
26.3
35.0
2a -14
(Wet)
20.0
18.7
20.2
23.2
27.4
2a -14 (Dry)
26.9
21.4
21.2
29.0
38.4
J-185
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Model
Input/
Output
Parameter
Scenario -
Run
Upper
Lobe
Upper
Lake
Lower
Lake
Upper
Basin
Lower
Basin
3a - 21
(Wet)
20.4
19.0
20.4
23.3
27.5
3a - 21 (Dry)
27.2
21.6
21.4
28.9
38.3
4a - 28
(Wet)
21.0
19.5
21.0
23.4
27.7
4a - 28 (Dry)
28.1
22.0
21.8
28.7
37.9
Predicted Chl-a
cone (ng/L)
la - 7 (Wet)
5.2
3.8
4.3
8.3
9.5
la - 7 (Dry)
7.5
4.7
4.6
8.6
11.6
2a -14
(Wet)
5.6
4.0
4.4
7.1
8.4
2a -14 (Dry)
8.3
5.0
4.9
9.7
12.9
3a - 21
(Wet)
5.7
4.1
4.4
7.2
8.4
3a - 21 (Dry)
8.4
5.0
5.0
9.7
12.9
4a - 28
(Wet)
5.9
4.2
4.5
7.2
8.5
4a - 28 (Dry)
8.7
5.1
5.1
9.6
12.7
J-186
-------
Mystic River Watershed TMDL Alternative Development Phase 2 Technical Memoranda
Appendix K. BMP Design Parameters Used in the
Pilot Watershed
General
Information
BMP Parameters
Biofilt ration
Infiltration-B
Infiltration-C
Porous
Pavement
BMP Dimensions
Surface Area (ac)
Table X-7
Table X-7
Table X-7
Table X-7
Surface Storage
Configuration
Soil Properties
Orifice Height (ft)
0
0
0
0
Orifice Diameter
(in.)
0
0
0
0
Rectangular or
Triangular Weir
Rectangular
Rectangular
Rectangular
Rectangular
Weir Height
(ft)/Ponding Depth
(ft)
0.5
2
2
0.2
Crest Width (ft)
100
100
100
100
Depth of Soil (ft)
2.5
0
0
2.67
Soil Porosity (0-1)
0.2
0.4
0.4
0.23
Vegetative
Parameter A
0.9
0.9
0.9
0.1
Soil Infiltration
(in/hr.)
2.5
2.41
0.52
17.42
Underdrain
Properties
Consider
Underdrain
Structure?
Yes
No
No
Yes
Storage Depth (ft)
1
0
0
1.75
Media Void Fraction
(0-1)
0.4
0
0
0.4
Background
Infiltration (in/hr.)
0
2.41
0.52
0
Cost Parameters
Storage Volume
Cost ($/ft3)
$15.46
$6.24
$6.24
$5.32
Cost Function
Adjustment
BMP Development
Type
New BMP in
Developed
Area
New BMP in
Developed
Area
New BMP in
Developed
Area
Difficult
Installation in
Highly Urban
Settings
Cost Adjustment
Factor
2
2
2
3
Decay Rates
TP (1/hr.)
0.13
0.27
0.27
0.0051
Underdrain
Removal Rates
TP (%, 0-1)
0.43
0
0
0.1
K-187
------- |