See Illinois TMDL report - Alliance for the Great Lakes

brewerobstructionAI and Robotics

Nov 7, 2013 (4 years and 6 days ago)

479 views

January 6, 2011
Bacteria TMDLs for Illinois’ Lake Michigan
Beaches Options Summary Report
FINAL Report
Task Order 2010-25
Period of Performance: Contract award through July 1, 2011
Prepared for:
U.S. Environmental Protection Agency
Region 5
Chicago, IL
Prepared by:
RTI International
3040 Cornwallis Road
Research Triangle Park, North Carolina 27709-2194
EPA Contract No. EP-C-08-003
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
(This page left blank intentionally)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
i
Executive Summary
This Options Summary Report (OSR) presents alternative approaches to developing bacteria TMDLs for
the 51 impaired shoreline segments (mainly beaches) along the Illinois Lake Michigan shoreline. The
following materials were reviewed in the preparation of this report: available data from U.S.
Environmental Protection Agency (EPA), Illinois Environmental Protection Agency (IEPA), and
stakeholders; applicable tools and models; and any additional data requirements for TMDL development.
Also, this report summarizes the various approaches to developing bacteria TMDLs for beaches based on
approved beach TMDLs from across the country. The OSR is intended to help the U.S. EPA and the
IEPA evaluate the best option(s) for completing beach TMDLs that fit within Illinois rules and U.S. EPA
regulations, guidance, and expectations for approvable TMDLs.
Published literature, studies provided by stakeholders, and monitoring reports were reviewed in
preparation of this OSR. A number of studies at individual beaches are used to highlight source detection
efforts and possible methods for TMDL development. Along the Illinois shoreline there are several
beaches that have been studied in detail by various agencies (Chicago Park District, the U.S. Geological
Survey, Illinois Department of Public Health, etc.). Study methods have included DNA identification of
sources, intensive monitoring of transects throughout a beach, predictive modeling, and higher-level
statistical and semi-processed-based modeling. Although some bacteria sources are apparent (e.g. gull
droppings) other sources are still less defined or unknown. Additionally, the responses of individual
beaches to E. coli loadings have been shown in some cases to be similar in a defined geographic region.
Therefore, a grouping analysis is suggested as a first step to developing TMDLs along the shoreline.
Four options are provided for pursuing TMDL development at the 51 impaired shoreline segments:

Option 1: Statistically Based Model with High-Level Uncertainty Analysis

Option 2: Use Mainly Existing Data in Hybrid Statistical Plus Hydraulic Process Modeling

Option 3: Provide for Adaptive TMDL Development to Utilize Upcoming Sanitary Surveys

Option 4: Shoreline Process-Based Modeling
While all of the presented options can be used to produce effective load-based TMDLs, there is a balance
of trade-offs among them. Although Option 3 is the most flexible in terms of source detection and TMDL
implementation, its timeline is dependent on completion of the beach sanitary surveys (BSS). The
planned BSSs are being funded through the Great Lake Restoration Initiative and are scheduled to be
completed in 2012. Option 3 also provides the greatest flexibility to consider both point and nonpoint
loading values in the form of concentrations, loadings, and a binary response. Option 1 utilizes the same
statistical methods as proposed for Option 3, but would rely on currently available data. Reliance on
existing data may limit the ability of the model to provide source detection and implementation guidance.
Both of these options allow for a comprehensive assessment of uncertainty within the TMDL calculation.
Option 2 focuses more on modeling of the hydrologic processes controlling the fate and transport of E.
coliwithin the Lake Michigan waters. Because this option is also partly statistical-based, uncertainty
analysis can be readily incorporated to better quantify a margin of safety. The trade-off with this method,
however, is that there is little an implementer or beach manager can do about the hydrologic processes
represented by the model (e.g., resuspension, currents, etc.).
Option 4 would provide a broader picture of the entire shoreline in a single model that would meet
regulatory requirements, although it would likely not quantify individual sources at individual beaches.
The trade-off with this option is that definition of the source(s) of contamination at each beach and the
source reductions required to meet water quality standards would be less well quantified compared to the
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
ii
other options. Also, the available models do not include explicit uncertainty analysis. Any reduction
scenarios would result from a trial-and-error approach to managing the sources input into the model rather
than a reduction scenario derived based on the modeling results.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
iii
Table of Contents
List of Acronyms ........................................................................................................................................ vi
i
1.Introduction ......................................................................................................................................... 1-1
1.1Background ............................................................................................................................... 1-1
1.2Process Followed for OSR Development ................................................................................. 1-1
2.Beaches Summary ............................................................................................................................... 2-1
2.1Sanitary Surveys and Specific Beach Studies ........................................................................ 2-20
2.1.163
rd
St. BSS (Cook County; Cali et al., 2007) ............................................................ 2-20
2.1.2Characterization of E. coli Levels at 63
rd
St. Beach (Cook County; Whitman et
al., 2001) ..................................................................................................................... 2-21
2.1.3Rosewood BSS (Lake County; Adam and Pfister, 2007) ........................................... 2-21
2.1.4Future BSSs within Suburban Cook County and Lake County .................................. 2-22
2.1.5BSSs within Chicago City Limits ............................................................................... 2-22
2.2Existing Modeling Studies ...................................................................................................... 2-23
2.2.1Predictive Modeling for Beach Closures .................................................................... 2-23
2.2.2Fate and Transport Modeling ..................................................................................... 2-25
2.3Source Identification Studies .................................................................................................. 2-27
2.3.1Lake Influences including Shoreline Topography ...................................................... 2-27
2.3.2Animal Sources........................................................................................................... 2-29
2.3.3Point Sources (CSOs, WWTPs, Storm Water Outfalls) ............................................. 2-31
2.3.4Surface Flows/Runoff/Harbors ................................................................................... 2-34
2.3.5Chicago Area Waterway System (CAWS) Influences ............................................... 2-35
2.3.6Watershed Sources ..................................................................................................... 2-38
2.3.7Other ........................................................................................................................... 2-38
3.Additional Considerations for TMDL Assessments ............................................................................ 3-1
3.1Previous Beach TMDLs ........................................................................................................... 3-1
3.2Grouping ................................................................................................................................... 3-7
3.2.1Geographic and Physical Considerations ..................................................................... 3-8
3.2.2Source/Response-Related Grouping Considerations .................................................... 3-9
4.Conclusions from Initial Data and Source Analysis ............................................................................ 4-1
4.1Sources, Fate, and Transport Summary .................................................................................... 4-1
4.2
Data Availability for Assessment ............................................................................................. 4-2
4.3Critical Areas for Further Investigation .................................................................................... 4-2
4.4Existing Methods Used for Bacteria TMDLs Applied to Illinois ............................................. 4-3
5.Deriving the TMDL ............................................................................................................................. 5-1
5.1Point Sources Quantification .................................................................................................... 5-1
5.2Source and Loading Quantification for TMDL Development .................................................. 5-2
5.2.1Loading Models ............................................................................................................ 5-2
5.2.2Response Models .......................................................................................................... 5-7
5.3Uncertainty ............................................................................................................................. 5-10
6.Recommended Options ....................................................................................................................... 6-1
6.1Option 1: Statistically Based Model with High-Level Uncertainty Analysis ........................... 6-1
6.1.1Given the data that exists to date, what are the inputs to the model? ........................... 6-2
6.1.2What is the basic methodology of the model? .............................................................. 6-2
6.1.3What improvements can be made to the model by additional data collection? ............ 6-3
6.1.4What is the specific data output from the model? ........................................................ 6-3
6.1.5How is the output used to derive a TMDL? ................................................................. 6-3
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
iv
6.1.6How is uncertainty addressed in the TMDL? ............................................................... 6-3
6.1.7How can the TMDL be used to drive implementation? ................................................ 6-4
6.2Option 2: Use Mainly Existing Data in Hybrid Statistical Plus Hydraulic Process
Modeling ................................................................................................................................... 6-4
6.2.1Given the data that exists to date, what are the inputs to the model? ........................... 6-4
6.2.2What is the basic methodology of the model? .............................................................. 6-4
6.2.3What improvements can be made to the model by additional data collection? ............ 6-5
6.2.4What is the specific data output from the model? ........................................................ 6-5
6.2.5How is the output used to derive a TMDL? ................................................................. 6-5
6.2.6How is uncertainty addressed in the TMDL? ............................................................... 6-5
6.2.7How can the TMDL be used to drive implementation? ................................................ 6-6
6.3Option 3: Provide for Adaptive TMDL Development to Utilize Upcoming Sanitary
Surveys ..................................................................................................................................... 6-6
6.3.1Given the data that exists to date, what are the inputs to the model? ........................... 6-6
6.3.2What is the basic methodology of the model? .............................................................. 6-6
6.3.3What improvements can be made to the model by additional data collection? ............ 6-7
6.3.4What is the specific data output from the model? ........................................................ 6-7
6.3.5How is the output used to derive a TMDL? ................................................................. 6-7
6.3.6How is uncertainty addressed in the TMDL? ............................................................... 6-8
6.3.7How can the TMDL be used to drive implementation? ................................................ 6-8
6.4Option 4: Shoreline Process Based Modeling .......................................................................... 6-8
6.4.1Given the data that exists to date, what are the inputs to the model? ........................... 6-8
6.4.2What is the basic methodology of the model? .............................................................. 6-9
6.4.3What improvements can be made to the model by additional data collection? ............ 6-9
6.4.4
What is the specific data output from the model? ........................................................ 6-9
6.4.5How is the output used to derive a TMDL? ................................................................. 6-9
6.4.6How is uncertainty addressed in the TMDL? ............................................................. 6-10
6.4.7How can the TMDL be used to drive implementation? .............................................. 6-10
7.References ........................................................................................................................................... 7-1
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
v
List of Figures
2-1. Impaired beaches within Lake County, Illinois. ........................................................................... 2-2
2-2. Impaired beaches within Suburban Cook County, Illinois – Tile 1. ............................................. 2-3
2-3. Impaired beaches within Suburban Cook County, Illinois – Tile 2. ............................................. 2-3
2-4. Impaired beaches within Northern Chicago in Cook County, Illinois. ......................................... 2-4
2-5. Impaired beaches within Southern Chicago in Cook County, Illinois. ......................................... 2-4
2-6. Current flow pattern around Chicago 63
rd
St. Beach driven by an external current entering
the computation domain through boundaries A-B and B-C and exiting at A-D (Ge et al.,
2010). .......................................................................................................................................... 2-26
2-7. Current flow pattern around Chicago 63
rd
St. Beach driven by an external current entering
the computation domain through boundary A-D and exiting at B-C (Ge et al., 2010). .............. 2-27
2-8. Six- and 12-month average depth-averaged currents for 1982-83 and 1994-95 (Beletsky
and Schwab, 2001). ..................................................................................................................... 2-28
2-9. Observed mean summer, winter, and annual circulation in Lake Michigan during 1982-
1983. (Isobaths are every 50 m) (Beletsky et al., 1999). ............................................................ 2-29
2-10. E. coli sources in 2002 beach water study (Soucie and Pfister, 2003). ....................................... 2-30
2-11. E. coli sources in 2003 beach water study (Soucie and Pfister, 2003). ....................................... 2-30
2-12. Rosewood Beach ravine (3L) as defined in the 2009 Strategic Sub-Watershed
Identification Process Report (Alliance for the Great Lakes, 2009) ........................................... 2-32
2-13. Municipalities covered by MS4 Permits within Lake and Cook Counties. ................................ 2-33
2-14. CAWS Control Structures (Lanyon, 2010). ................................................................................ 2-36
3-1. Multidimensional scaling depiction of E. coli concentrations (log MPN/100 mL) for
23 Chicago beaches for the years 2000-2005 (Whitman and Nevers, 2008). ............................... 3-8
6-1. An example of using adaptive implementation with Bayes updating where the green line
represents the WQS to be met. ...................................................................................................... 6-7
List of Tables
2-1. Impaired Shoreline Segments ....................................................................................................... 2-5
2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments ......................... 2-8
2-3. Examples of Predictive Regression Models Used for Beach Closures within the Great
Lakes ........................................................................................................................................... 2-24
2-4. Active NPDES Discharges along Lake Michigan Shoreline within Illinois ............................... 2-34
2-5. Reversals to Lake Michigan in Millions of Gallons ................................................................... 2-37
3-1. Indiana Lakeshore TMDL (Approved) (Tetra Tech, 2004) .......................................................... 3-1
3-2. Luna Pier Beach, Lake Erie TMDL (MDEQ, 2007) ..................................................................... 3-4
3-3. Duck Neck Beach on Chester River, Maryland (MDE, 2009) ..................................................... 3-5
3-4. Ventura County Beaches, California (California Regional Water Quality Control Board,
2007) ............................................................................................................................................. 3-6
5-1. Methods for Incorporating Identified Sources into TMDL Development .................................... 5-1
5-2. Bacteria Modeling Matrix from Texas TMDL Task Force Report (Jones et al., 2007) ............... 5-3
5-3. Model Computations for TMDL Development ............................................................................ 5-4
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
vi
List of Acronyms
ANN artificial neural networks
BIT Bacteria Indicator Tool
BMP best management practice
BSS Beach Sanitary Survey
Caltrans California Department of Transportation
CAWS Chicago Area Waterway System
CEAM Center for Exposure Assessment Modeling
cfu colony forming units
CPD Chicago Park District
CRCW Chicago River Controlling Works
CSO combined sewer overflow
CWA Clean Water Act
E. coli Escherichia coli
EFDC Environmental Fluid Dynamics Code
FA future allocation
GIS geographic information system
GLERL Great Lakes Environmental Research Laboratory
GLRI Great Lakes Restoration Initiative
HSPF Hydrologic Simulation Program - FORTRAN
IEPA Illinois Environmental Protection Agency
IDPH Illinois Department of Public Health
LA load allocation
LCHD Lake County Health Department
LDC load duration curve
m3 cubic meters
MDE Maryland Department of the Environment
MGD million gallons per day
mL milliliter
MOS margin of safety
MPN most probable number
MR multiple regression
MS4 Municipal Separate Storm Water Sewer System
MWRDGC Metropolitan Water Reclamation District of Greater Chicago
NCDC National Climatic Data Center
NOAA National Oceanic and Atmospheric Administration
NPDES National Pollutant Discharge Elimination System
NRC National Research Council
NRDC National Resources Defense Council
OLD O’Brien Lock and Dam
OSR Option Summary Report
POM Princeton Ocean Model
PSSS Pollution Source Shoreline Survey
s seconds
STORET STOrage and RETrieval System
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
vii
SWAT Soil and Water Assessment Tool
SWMM Storm Water Management Model
TARP Tunnel and Reservoir Plan
TMDL Total Maximum Daily Load
U.S. EPA United States Environmental Protection Agency
USGS United States Geologic Survey
WASP Water Quality Analysis Simulation Program
WDR waste discharge requirements
WLA waste load allocation
WPS Wilmette Pumping Station
WQS water quality standard
WRP water reclamation plant
WWTP waste water treatment plant
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
viii
(This page left blank intentionally)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
1-1
1. Introduction
1.1 Background
Lake Michigan beaches and their coastal waters are a highly valued societal and ecological resource.
These beaches are widely popular, highly used, and therefore frequently monitored by stakeholders and
local government. From May through September, the Lake County Health Department (LCHD) Lakes
Management Unit, Chicago Park District (CPD), and others sample Lake Michigan beaches five to seven
days a week for bacteria. The Illinois Environmental Protection Agency (IEPA) uses these monitoring
data to assess the attainment of Lake Michigan beaches for its designated use of primary contact
recreation. Beach closures occur when Escherichia coli (E. coli) bacteria exceed the water quality
standard (WQS) of 235 colony forming units (cfu) per 100 milliliters (mL) as a single sample maximum
and 126 cfu/100 mL for a 30-day geometric mean. This criterion is based on guidelines established by the
U.S. Environmental Protection Agency (U.S. EPA) for recreational waters. These WQSs are evaluated
against actual and predicted E. coli concentrations.
Fifty-one Lake Michigan shoreline segments are monitored by LCHD and CPD for E. coli, and all are in
nonattainment of their designated use, primary contact recreation. The term ‘shoreline segment’ is used in
place of ‘beach’ because not all 51 segments have swimming access – those few segments are not
considered beaches. However, all Lake Michigan nearshore waters are considered to have a use of
primary contact; therefore, all of the segments are assessed in the same manner for that use. Within
Illinois, water quality is found to be “not supporting” of primary contact use when, on average, (1) there
is one bathing area closure per year of less than 1 week’s duration or (2) there is one bathing area closure
per year of greater than 1 week’s duration or more than one bathing area closure per year. Given that
these 51 segments are in nonattainment, they are included on Illinois’ 303(d) list. The Clean Water Act
(CWA) and U.S. EPA regulations require that States develop Total Maximum Daily Loads (TMDLs) for
all waters on the section 303(d) lists. A TMDL is the sum of the allowable amount of a single pollutant
that a water body can receive from all contributing point and nonpoint sources and still support its
designated uses. The 51 shoreline segments on Illinois’ 303(d) list span a wide range of conditions from
urban Chicago beaches and hardened shoreline to more natural, undisturbed beaches along the northern
edge of Illinois’ shoreline.
Beach TMDLs are notoriously challenging to develop because it is difficult to determine the appropriate
flow conditions, and therefore the loading capacity, for beaches on large bodies of water like Lake
Michigan. Likewise, it is difficult to associate sources with the beach impairment because there is no
easily defined “upstream” and “downstream.” Some states, such as Michigan, have developed beach
TMDLs using a simple concentration-based (organisms/volume) approach. This approach is anchored on
the premise that all water bodies and point source dischargers must meet the established WQS for
bacteria; however, identifying the contributing sources is still difficult. Other states (such as Indiana) have
used complex modeling to determine the loading capacity and associated load and waste load allocations
(Tetra Tech, 2004). This type of modeling can be costly and may not add value. This Options Summary
Report (OSR) is designed to provide enough information to select the best methods to formulate TMDLs
for the wide range of conditions found across the 51 impaired shoreline segments.
1.2 Process Followed for OSR Development
This report summarizes the various approaches to developing bacteria TMDLs for beaches based on
approved beach TMDLs from across the country. Also presented are innovative approaches to developing
bacteria TMDLs that are appropriate for the conditions found in Illinois. The following materials were
reviewed in the preparation of this report: available data from U.S. EPA, IEPA, and stakeholders;
applicable tools and models; and any additional data requirements for TMDL development. The OSR is
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
1-2
intended to help the U.S. EPA and IEPA evaluate the best option(s) for completing beach TMDLs that fit
within Illinois rules and U.S. EPA regulations, guidance, and expectations for approvable TMDLs. The
outcomes from this report include options for conducting detailed source assessments and critical area
identification.
The bacteria sources and available data will guide the exploration and selection of the applicable tools and
models and suggest multiple tool/model combinations to best estimate source contributions to beach
bacteria levels. For example, a single application using a model such as the Environmental Fluid
Dynamics Code (EFDC) model, a model commonly used for bacteria TMDLs at beaches, will likely not
meet all requirements for the TMDLs in this study because of the importance of storm runoff as a source
to the Illinois beaches. This TMDL will require investigating sources such as urban drainage areas,
animals (particularly gulls), regrowth in or resuspension from beach sands, surface drainage (along the
north end of the Illinois shoreline), runoff, and transport within the lake. We investigate methods ranging
from multiple linear regression, to Bayesian networks, to coefficient-based models (e.g., the Bacterial
Indicator Tool [BIT]), to simple process models, to dynamic process models (e.g., HSPF [Hydrologic
Simulation Program-Fortran] or EFDC) before determining which method or combination of methods
best represents the priority sources/variables needed and what data are available to formulate a
comprehensive TMDL. In short, the final options recommended in this report summarize what works best
by taking an adaptive approach to TMDL development.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-1
2. Beaches Summary
Within Illinois, there are 51 segments along 63 miles of Lake Michigan shoreline that are impaired for
primary contact recreation due to elevated levels of indicator bacteria. (Note: There are additional beaches
along Illinois’ shoreline; however, either they are not monitored and therefore the water quality
conditions cannot be assessed or they are in compliance with the WQS.)
The shoreline segments within Illinois fall into four general categories:
1. City of Chicago beaches: These beaches are all within high population areas and are high use
beaches. They are managed (monitored and closed) by one single entity, the CPD.
2. Suburban Cook County beaches: These beaches are also high use beaches but are managed by the
individual municipalities in which they lie. They are typically small beaches that lie at the end of
streets where they meet Lake Michigan.
3. Lake County beaches: While all of these beaches are monitored by the LCHD, they are controlled
by individual municipalities. The beaches to the north are within a state park.
4. Hardened Shoreline Segments within Chicago: These areas have no swimming access and are not
regularly monitored, but at some time in the past they were monitored and found to be impaired.
Typically, these segments consist of impervious areas such as seawalls and shoreline walkways.
Given that the majority of the Lake Michigan coastline in Illinois is developed land associated with the
city of Chicago and its suburbs, there are a variety of conditions that can complicate source detection and
quantification and, therefore, TMDL development:

delineation of the source/drainage areas for each beach is complicated by sewers, storm water
runoff, and other aspects of human development;

a wide range of bacteria sources exist (e.g., dog parks at beaches, storm water detention basins,
combined sewer overflow [CSO] areas, and septic systems);

beaches fall into several categories depending on their location and physical design (e.g., an
urban beach bounded by breakwaters/piers versus a state park beach without physical
boundaries); and

monitoring and modeling of the shoreline segments ranges from intensely monitored and modeled
beaches to shoreline segments without recent monitoring data or without frequent monitoring
data.
For this report, attempts were made to summarize all facets of the aforementioned data for the 51
impaired shoreline segments. Figures 2-1 through 2-5 locate the segments within Lake and Cook
counties. Table 2-1 provides a list of the impaired shoreline segments with supporting characterization
information. Table 2-2 summarizes the basic information on the segment and on the monitoring and
modeling at the beaches. For the “Beach Act Reporting” columns in Table 2-2, note that a swimming ban
is put in place for beaches in Lake County and suburban Cook County when one sample exceeds 235
cfu/100 mL. For beaches in Chicago, a swimming adivsory is put in place when the mean of two samples
exceeds 235 cfu/100 mL and a swimming ban is put in place when the mean of two samples exceeds 1000
cfu/100 mL. Both swimming bans and swimming advisories (numbers in parentheses) are presented as
both conditions are considered closures by Illinois when assessing use attainment.
The Great Lakes in general have been heavily studied for a long time, with water quality issues in the
1970s bringing scientific investigation to the forefront. Lake Michigan, specifically the southern end of
the lake, has been well studied over the last decade by a variety of government, academic, and local
agencies. The following sections summarize the studies most pertinent to the development of TMDLs
along the Illinois shoreline in general terms.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-2
Figure2-1. Impaired beaches within Lake County, Illinois.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-3
Figure2-2. Impaired beaches within Suburban Cook County, Illinois – Tile 1.
Figure2-3. Impaired beaches within Suburban Cook County, Illinois – Tile 2.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-4
Figure2-4. Impaired beaches within Northern Chicago in Cook County, Illinois.
Figure2-5. Impaired beaches within Southern Chicago in Cook County, Illinois.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-5
Table2-1. Impaired Shoreline Segments
AUID
BeachID
SiteID
1
Illinois Department
of Public Health
(IDPH) Name
Assessment Beach Name
Name Note
2
Type
Tier
3
Length
(m)
4
County/
Organization
IL_QH-01 IL913512 365 North Point Marina
Beach
North Point Beach Swimming 1 317 Lake/ LCHD
IL_QH-03 IL677426 349 Illinois Beach State
Park North Beach
IL Beach State Park North Swimming 1 977 Lake/ LCHD
IL_QH-04 IL087773 376 Waukegan North
Beach
Waukegan North Beach Swimming 1 2219 Lake/ LCHD
IL_QH-05 IL234945 377 Waukegan South
Beach
Waukegan South Beach Swimming 1 339 Lake/ LCHD
IL_QH-09 IL215601 345 Illinois Beach State
Park South Beach
IL Beach State Park South Swimming 1 5648 Lake/ LCHD
IL_QI-06 IL195441 356 Lake Bluff Sunrise
Beach
Lake Bluff Beach (Sunrise) Swimming 1 406 Lake/ LCHD
IL_QI-10 IL634222 357 Lake Forest Forest
Park Beach
Lake Forest Beach (Forest
Park)
Swimming 1 809 Lake/ LCHD
IL_QJ IL730475 342 Highland Park
Rosewood Beach
Rosewood Beach Swimming 1 292 Lake/ LCHD
IL_QJ-05 IL782704 340 Highland Park Avenue
Boating Beach
Park Avenue Beach Swimming 1 204 Lake/ LCHD
IL_QK-04 IL942128 337 Glencoe Park Beach Glencoe Beach (Glencoe
Park Beach)
Swimming 1 172 Cook/ Glencoe
Park District
IL_QK-06 IL108354 384 Winnetka Tower
Beach
Tower Beach (Winnetka
Tower Beach)
Swimming 1 167 Cook/ Winnetka
Park District
IL_QK-07 IL595016 382 Winnetka Lloyd Park
Beach
Lloyd Beach (Winnetka Lloyd
Park Beach)
Swimming 1 172 Cook/ Winnetka
Park District
IL_QK-08 IL750698 383 Winnetka Maple Park
Beach
Maple Beach (Winnetka
Maple Park Beach)
Swimming 1 76 Cook/ Winnetka
Park District
IL_QK-09 IL928218 381 Winnetka Elder Park
Beach
Elder Beach (Winnetka Elder
Park Beach)
Swimming 1 121 Cook/ Winnetka
Park District
IL_QL-03 IL984895 354 Kenilworth Beach Kenilworth Beach Swimming 1 122 Cook/ Kenilworth
Water & Light
Dept.
IL_QL-06 IL637664 378 Wilmette Gillson Park
Beach
Gilson Beach (Wilmette
Gillson Park Beach)
Swimming 1 445 Cook/ Wilmette
Park District
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-6
Table2-1. Impaired Shoreline Segments (continued)
AUID
BeachID
SiteID
1
Illinois Department
of Public Health
(IDPH) Name
Assessment Beach Name
Name Note
2
Type
Tier
3
Length
(m)
4
County/
Organization
IL_QM-03 IL505764 328 Evanston Greenwood
Beach
Greenwood Beach (Evanston
Greenwood Beach)
Swimming 1 372 Cook/ Evanston
Health Dept.
IL_QM-04 IL327651 329 Evanston Lee Beach Lee Beach (Evanston Lee
Beach)
Swimming 1 222 Cook/ Evanston
Health Dept.
IL_QM-05 IL291926 330 Evanston Lighthouse
Beach
Lighthouse Beach (Evanston
Lighthouse Beach)
Swimming 1 253 Cook/ Evanston
Health Dept.
IL_QM-06 IL287401 367 Northwestern
University Beach
Northwestern University
Beach
Swimming 1 272 Cook/ Evanston
Health Dept.
IL_QM-07 IL601796 327 Evanston Clark Beach Clark Beach (Evanston Clark
Beach)
Swimming 1 213 Cook/ Evanston
Health Dept.
IL_QM-08 IL636205 331 Evanston South Beach South Boulevard Beach
(Evanston South Beach)
Swimming 1 245 Cook/ Evanston
Health Dept.
IL_QN-01 IL705276 359 Leone Beach Touhy (Leone) Beach (Loyola
Beach)
Considered part of
Leone Beach by CPD
Swimming 1 881 Cook/ CPD
IL_QN-02 359 - part Loyola (Greenleaf) Beach Considered part of
Leone Beach by CPD
Swimming 1 Cook/ CPD
IL_QN-03 IL923491 353 Kathy Osterman
Beach
Hollywood/ Ostermann Beach
(Kathy Osterman Beach)
Swimming 1 525 Cook/ CPD
IL_QN-04 IL228136 334 Foster Avenue Beach Foster Beach Swimming 1 297 Cook/ CPD
IL_QN-05 IL132842 360 Montrose Beach Montrose Beach Swimming 1 837 Cook/ CPD
IL_QN-06 IL748682 352 Juneway Terrace Park
Beach
Juneway Terrace (Juneway
Terrace Park Beach)
Swimming 1 57 Cook/ CPD
IL_QN-07 IL621748 372 Rogers Avenue Park
Beach
Rogers Beach (Rogers
Avenue Park Beach)
Swimming 1 53 Cook/ CPD
IL_QN-08 IL120964 343 Howard Street Park
Beach
Howard Beach (Howard
Street Park Beach)
Swimming 1 80 Cook/ CPD
IL_QN-09 IL603994 351 Jarvis Avenue Park
Beach
Jarvis Beach (Jarvis Avenue
Park Beach)
Swimming 1 217 Cook/ CPD
IL_QN-10 IL259912 370 Pratt Blvd And Park
Beach
Pratt Beach (Pratt Blvd and
Park Beach)
Considered Hartigan
Beach by CPD
Swimming N/A (1) 193 Cook/ CPD
IL_QN-11 IL274491 366 North Shore Avenue
Beach
North Shore/Columbia (North
Shore Avenue Beach)
Considered Hartigan
Beach by CPD
Swimming N/A (1) 235 Cook/ CPD
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-7
Table2-1. Impaired Shoreline Segments (continued)
AUID
BeachID
SiteID
1
Illinois Department
of Public Health
(IDPH) Name
Assessment Beach Name
Name Note
2
Type
Tier
3
Length
(m)
4
County/
Organization
IL_QN-12 IL798802 321 Hartigan Beach Albion Beach Considered Hartigan
Beach by CPD
Swimming N/A (1) 61 Cook/ CPD
IL_QN-13 IL586992 375 Thorndale Thorndale Beach Considered part of
Kathy Ostermann
Beach by CPD
Swimming N/A (1) 58 Cook/ CPD
IL_QO-01 IL666876 363 North Avenue Beach North Ave. Beach Swimming 1 1691 Cook/ CPD
IL_QO-02 IL103378 335 Fullerton (Theater On
The Lake)
Fullerton Beach (Fullerton
(Theater on the Lake))
Fullerton St. Shoreline No access 3 208 Cook/ CPD
IL_QO-03 366 - part Webster Beach Considered North
Avenue Beach by
CPD

IL_QO-04 366 - part Armitage Beach Considered North
Avenue Beach by
CPD

IL_QO-05 N/A N/A N/A Schiller Beach Schiller Ave Shoreline No access N/A (3) N/A No Data Available
IL_QP-02 IL296528 368 Oak Street Beach Oak St. Beach Swimming 1 338 Cook/ CPD
IL_QP-03 IL926480 369 Ohio Street Beach Ohio St. Beach Swimming 1 171 Cook/ CPD
IL_QQ-01 IL820929 317 12
th
Street 12
th
St. Beach Swimming 1 325 Cook/ CPD
IL_QQ-02 IL461767 318 31
st
Street Beach 31
st
St. Beach Swimming 1 275 Cook/ CPD
IL_QR-01 IL865711 N/A N/A 49
th
St. Beach 49th St. Shoreline No access N/A (3) N/A Cook/ CPD
IL_QS-02 IL118596 350 Jackson Park Beach Jackson Park/63
rd
St. Beach Swimming 1 666 Cook/ CPD
IL_QS-03 IL814025 371 Rainbow Beach Rainbow Swimming 1 546 Cook/ CPD
IL_QS-04 IL589159 319 57
th
Street Beach 57
th
St. Beach Swimming 1 241 Cook/ CPD
IL_QS-05 IL288152 320 67
th
Street Beach 67
th
St. Beach 67th St. Shoreline No access N/A (3) 286 Cook/ CPD
IL_QS-06 IL581683 374 South Shore South Shore Beach Swimming 1 212 Cook/ CPD
IL_QT-03 IL376700 323 Calumet South Beach Calumet Beach (Calumet
South Beach)
Swimming 1 404 Cook/ CPD
1 “
Part” indicates that the segment, although considered separate in terms of assessment unit, is considered part of another beach by IL DPH; “N/A” indicates the segment is not
recognized as an accessible segment by IL DPH.
2
This column provides information on how individual segments are related to actual monitored beaches according to CPD.
3 “
N/A (#)” indicates that no Tier ranking was provided by IL DPH but an assumed ranking is indicated in the parentheses; blank cells indicate the segment is part of a larger beach for
which there is a Tier ranking.
4 “
N/A” indicates the beach is not indexed or monitored by IL DPH; blank cells indicate the beach is part of a larger beach for which there is a length provided.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-8
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QH-01 NORTHPT North Point Beach IL913512 2000 13 1.4
Source detection study in 2003 - major sources = avian/gulls and
human/sewage
2001 16 2
2002 17 2.7
2003 18 1.8
2004 37 1.0 6/1/04 8/28/04 85 COUNTY X
2005 15 3.5 6/1/05 9/3/05 93 COUNTY X
2006 40 6.1 5/25/06 9/2/06 94 COUNTY X
2007 8 9.9 5/23/07 9/1/07 104 COUNTY X
2008 19 2.7 5/27/08 8/28/08 126 STORET X
2009 4 22 5/26/09 9/8/09 115 STORET X
IL_QH-03 ILBEACHN IL Beach State Park North IL677426 2000 1 1
Olyphant,(2005) developed multiple regression relating hydrometeorological
data and E. coli
2001 5 1
2002 4 1
2003 14 1.3
2004 8 1 6/1/04 8/28/04 85 COUNTY X X X X X X X
2005 6 1 6/1/05 9/3/05 93 COUNTY X X X X X X X
2006 7 1 5/25/06 9/2/06 198 COUNTY X X X X X X X
2007 10 1.6 5/23/07 9/1/07 194 STORET X X X X X X X
2008 8 1.1 5/27/08 8/28/08 128 STORET X X X X X X X
2009 6 2.3 5/26/09 9/3/09 111 STORET X X X X X X X
IL_QH-04 WAUKN Waukegan North Beach IL087773 2000 9 1.2
2001 10 1
2002 17 1.9
2003 19 1.2
2004 19 1
2005 5 1.4
2006 9 1
2007 18 1.6
2008 14 1.2 5/27/08 8/28/08 65 STORET
2009 3 1 5/26/09 9/3/09 62 STORET
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-9
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QH-05 WAUKS Waukegan South Beach IL234945 2000 8 1.3
-Source detection study in 2003 - major sources = avian/gulls and
human/sewage.
-SWIMCAST Beach
2001 19 1.2
2002 19 3.0
2003 15 1.6
2004 19 1 6/1/04 8/28/04 87 COUNTY X X X X X X
2005 10 1.2 6/1/05 9/3/05 93 COUNTY X X X X X X
2006 16 1.6 5/25/06 9/2/06 123 COUNTY X X X X X X X
2007 20 1.6 5/23/07 9/1/07 160 COUNTY X X X X X X X
2008 14 1 6/2/08 8/28/08 102 STORET X X X X X X
2009 11 1 6/1/09 9/3/09 93 STORET X X X X X X
IL_QH-09 ILBEACHS IL Beach State Park South IL215601 2000 9 1.4
Source detection study in 2003 - major sources = avian/gulls and
human/sewage; some contributions from dog and rodent sources as well
2001 10 1.1
2002 12 1.4
2003 16 1.3
2004 20 1 6/1/04 8/28/04 85 COUNTY X
2005 14 1.7 6/1/05 9/3/05 93 COUNTY
2006 22 5.9 5/25/06 9/2/06 198 COUNTY
2007 17 1.4 5/23/07 9/1/07 95 COUNTY
2008 9 1.7 5/27/08 8/28/08 134 STORET
2009 8 2.5 5/26/09 9/8/09 119 STORET
IL_QI-06 SUNRISE Lake Bluff Beach IL195441 2003 7 1.3
Dog beach adjacent. Sampling for 2004-2007 reported as Lake Bluff Dog Beach.
Weekly sampling completed for those years.
2004 6 1 6/2/04 8/25/04 14 COUNTY
2005 5/26/05 8/31/05 15 COUNTY
2006 5 1 5/25/06 8/30/06 15 COUNTY
2007 6 1.5 5/23/07 8/29/07 14 COUNTY
2008 6 1 5/27/08 8/28/08 59 STORET
2009 2 6 5/26/09 9/3/09 59 STORET
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-10
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QI-10 LFFOREST Lake Forest Beach IL634222 2000 3 1
-Source detection study in 2003 - major sources = avian/gulls and
human/sewage; some contributions from dog and rodent sources as well.
Olyphant,(2005) deloped multiple regression relating hydrometeorological data
andE. coli
-SWIMCAST Beach
2001 12 1
2003 12 1
2004
12 1 6/1/04 9/9/04 117 COUNTY
X (July,
Aug)
X X X X X X
2005 9 11.3 6/1/05 9/3/05 155 COUNTY X X X X X X X
2006 14 1.1 5/25/06 9/2/06 199 COUNTY X X X X X X X
2007 16 1.7 5/23/07 9/1/07 96 STORET X X X X X X
2008 9 1.1 6/2/08 8/28/08 60 STORET X X X X X X
2009 8 1.4 6/3/09 9/4/09 91 STORET X X X X X X
IL_QJ ROSEWOOD Rosewood Beach IL730475 2003 14 1.1
-Source detection study in 2003 - major sources = avian/gulls and
human/sewage; some contributions from deer sources as well.
-SWIMCAST beach
-Rosewood Ravine stream enters near beach
-Sanitary survey conducted in 2007
2004 10 1 6/1/04 8/28/04 85 COUNTY X X X X X X
2005 6 1.7 6/1/05 9/3/05 93 COUNTY X X X X X X
2006 7 1.3 5/25/06 9/2/06 168 COUNTY X X X X X X X
2007 11 1.3 5/23/07 9/1/07 151 COUNTY X X X X X X X
2008 9 1.3 6/2/08 8/28/08 106 STORET X X X X X X X
2009 6 1.3 6/1/09 9/3/09 99 STORET X X X X X X X
IL_QJ-05 PARKAVE Park Ave. Beach IL782704 2003 11 1.2
2004 11 1 6/1/04 8/28/04 85 COUNTY X X X X X X
2005 1 1 6/1/05 9/3/05 93 COUNTY X X X X X X
2006 7 1.1 5/25/06 9/2/06 96 COUNTY X X X X X X
2007 12 1.2 5/23/07 9/1/07 95 COUNTY X X X X X X
2008 3 1 5/27/08 8/28/08 82 STORET X X X X X X
2009 2 4 5/27/09 9/3/09 56 STORET X X X X X X
IL_QK-04 GLBEACH Glencoe Beach IL942128 2007 8 2.25 6/29/07 6/29/07 1 STORET
2008 5 1 6/1/08 7/29/08 53 STORET
2009 11 1.2 5/24/09 9/7/09 87 STORET
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-11
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QK-06 WNTOWER Tower Beach IL108354 2003 3 1
2004 6 1
2005 1 1
2006 10 1
2007 7 1.6 6/14/07 9/2/07 70 STORET
2008 6 1.2 6/12/08 8/23/08 65 STORET
2009 1 2 6/12/09 8/31/09 79 STORET
IL_QK-07 WNLLOYD Lloyd Beach IL595016 2003 3 1
2004 5 1
2005 2 1
2006 8 1.25
2007 4 1.5 6/14/07 8/18/07 66 STORET
2008 6/12/08 8/23/08 67 STORET
2009 (3) (1.3) 6/12/09 8/30/09 78 STORET
IL_QK-08 WNMAPLE Maple Beach IL750698 2003 5 1
2004 5 1
2005 3 1
2006 8 1
2007 8 1.1 6/14/07 8/13/07 61 STORET
2008 4 1.5 6/12/08 9/1/08 73 STORET
2009 1 1 6/12/09 8/30/09 78 STORET
IL_QK-09 WNELDER Elder Beach IL928218 2003 10 1.3
2004 13 1
2005 8 1.25
2006 12 1.25
2007 17 1.7 6/14/07 8/18/07 66 STORET
2008 13 1.8 6/12/08 8/23/08 67 STORET
2009 2 1 6/12/09 8/30/09 78 STORET
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-12
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QL-03 KENILWORTH Kenilworth Beach IL984895 2004 4 1
2005 2 1
2006 5 1
2007 12 1 6/23/07 7/29/07 2 STORET
2008 10 1.2 6/6/08 8/31/08 85 STORET
2009 8 1.75 6/5/09 9/6/09 93 STORET
IL_QL-06 WMGILLSON Gilson Beach IL637664 2003 9 1.3
Dog beach adjacent. 2004 22 1
2005 4 1.3
2006 6 1
2007 6 1.5 5/25/07 9/2/07 131 STORET
2008 5 1.6 5/24/08 9/1/08 100 STORET
2009 2 1.5 5/23/09 9/7/09 227 STORET
IL_QM-03 EVGREENWOOD Greenwood Beach IL505764 2003 6 1.2
Dog beach adjacent. 2004 6 1
2005 5 1.6
2006 8 1
2007 10 1.6 7/2/07 7/5/07 4 STORET
2008 16 1.4 6/14/08 8/31/08 152 STORET
2009 5 1 6/13/09 9/4/09 167 STORET
IL_QM-04 EVLEE Lee Beach IL327651 2003 6 1.2
2004 6 1
2005 3 1
2006 10 1
2007 8 1.1 7/28/07 8/19/07 4 STORET
2008 13 1.8 6/14/08 8/31/08 153 STORET
2009 8 1 6/13/09 9/4/09 167 STORET
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-13
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QM-05 EVLIGHT Lighthouse Beach IL291926 2003 4 1
2004 16 1
2005 5 1
2006 7 1
2007 10 1.3 6/27/07 8/24/07 4 STORET
2008 12 1 6/14/08 8/31/08 151 STORET
2009 7 1.3 6/13/09 9/4/09 167 STORET
IL_QM-06 EVNW Northwestern University BeachIL287401 2003 4 1.25
2005 10 1.6
2006 7 1
2007 7 1.3 6/14/07 8/18/07 3
2008 13 1.4 6/17/08 8/31/08 58 STORET
2009 3 2.7 6/17/09 9/4/09 38 STORET
IL_QM-07 EVCLARK Clark Beach IL601796 2007 12 1 6/17/07 8/9/07 3 STORET
2008 13 1.3 6/14/08 8/31/08 152 STORET
2009 8 1 6/13/09 9/4/09 167 STORET
IL_QM-08 EVSOUTH South Boulevard Beach IL636205 2003 7 1.1
2004 21 1
2005 14 1.1
2006 11 1.1
2007 15 1.7 7/29/07 7/29/07 1 STORET
2008 19 1.9 6/14/08 8/31/08 151 STORET
2009 8 1.25 6/13/09 9/4/09 168 STORET
IL_QN-01 CHLOYOLA Touhy (Leone) Beach IL705276 2003 1 2
-CPD Monitored 2004 21 1
2005 11 1.2
2006 10 1.2
2007 5 1.2 5/28/07 8/31/07 142 STORET X
2008 1 (4) 1 (1) 5/27/08 8/29/08 70 STORET X
2009 1 (6) 2 (1) 5/26/09 9/3/09 71 STORET X
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-14
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QN-02 Loyola (Greenleaf) Beach Monitored all as Leone Beach including “Touhy (Leone) Beach” above
IL_QN-03 CHOSTERMAN Hollywood/Ostermann Beach IL923491 2003 3 1.3 late May early Sept USGS X X X X X X
-CPD Monitored 2004 18 1 late May early Sept USGS X X X X X X
2005 12 1.6 late May early Sept USGS X X X X X X
2006 12 1
2007 10 1.2
2008 2 (5) 1 (1.4) 5/27/08 8/29/08 71 STORET X
2009 4 (6) 1.25 (1) 5/26/09 9/3/09 73 STORET X
IL_QN-04 CHFOSTER Foster Beach IL228136 2003 4 2.25 late May early Sept USGS X X X X X X
-CPD Monitored
-Dispersal of gulls via canine harassment was conducted in 2006 and 2007 for
trial purposes (Hartmann et al., 2010).
2004 19 1 late May early Sept USGS X X X X X X
2005 6 1.8 late May early Sept USGS X X X X X X
2006 15 1
2007 12 1.1 5/28/07 8/31/07 143 STORET X
2008 2 (5) 1 (1.6) 5/27/08 8/29/08 70 STORET X
2009 3 (4) 1.3 (1) 5/26/09 9/3/09 72 STORET X
IL_QN-05 CHMONTROSE Montrose Beach IL132842 2004 25 1 late May early Sept USGS X X X X X X
-CPD Monitored 2005 11 1.7 late May early Sept USGS X X X X X X
2006 18 1.3
2007 15 1.4 5/28/07 8/31/07 148 STORET X
2008 7 (10) 1.3 (1.1) 5/27/08 8/29/08 71 STORET X
2009 2 (12) 1.5 (1.3) 5/26/09 9/3/09 71 STORET X
IL_QN-06 CHJUNEWAY Juneway Terrace IL748682 2003 3 3 late May early Sept USGS X X X X X X
-CPD Monitored 2004 20 1 late May early Sept USGS X X X X X X
2005 7 1.1 late May early Sept USGS X X X X X X
2006 1 1
2007 7 1 5/28/07 8/31/07 142 STORET X
2008 (8) (1.3) 5/27/08 8/29/08 69 STORET X
2009 2 (1) 1.5 (1) 5/26/09 9/3/09 71 STORET X
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-15
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QN-07 CHROGERS Rogers Beach IL621748 2003 1 7 late May early Sept USGS X X X X X X
-CPD Monitored 2004 22 1 late May early Sept USGS X X X X X X
2005 7 1.1 late May early Sept USGS X X X X X X
2006 6 1
2007 5 1 5/28/07 8/31/07 142 STORET X
2008 1 (6) 1 (1.2) 5/27/08 8/29/08 69 STORET X
2009 1 (2) 2 (1) 5/26/09 9/3/09 71 STORET X
IL_QN-08 CHHOWARD Howard Beach IL120964 2003 1 2 late May early Sept USGS X X X X X X
-CPD Monitored 2004 18 1 late May early Sept USGS X X X X X X
2005 8 1.25 late May early Sept USGS X X X X X X
2006 6 1
2007 6 1.2 5/28/07 8/31/07 142 STORET X
2008 1 (6) 1 (1.3) 5/27/08 8/29/08 70 STORET X
2009 2 (1) 1.5 (1) 5/26/09 9/3/09 72 STORET X
IL_QN-09 CHJARVIS Jarvis Beach IL603994 2003 2 1.5 late May early Sept USGS X X X X X X
-CPD Monitored 2004 21 1 late May early Sept USGS X X X X X X
2005 7 1.3 late May early Sept USGS X X X X X X
2006 6 1
2007 7 1.1 5/28/07 8/31/07 142 STORET X
2008 (6) (1.7) 5/27/08 8/29/08 71 STORET X
2009 1 (1) 2 (1) 5/26/09 9/3/09 71 STORET X
IL_QN-10 CHPRATT Pratt Beach IL259912 2003 3 1.3 late May early Sept USGS X X X X X X
-CPD Monitored 2004 16 1 late May early Sept USGS X X X X X X
2005 9 1.6 late May early Sept USGS X X X X X X
2006 15 1
2007 10 1 5/28/07 8/31/07 142 STORET X
2008 (5) 1 5/27/08 8/29/08 70 STORET X
2009 1 (5) 2 (1) 5/26/09 9/3/09 71 STORET X
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-16
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QN-11 CHNORTHSH North Shore/Columbia IL274491 2003 2 1.5 late May early Sept USGS X X X X X X
-CPD Monitored 2004 20 1 late May early Sept USGS X X X X X X
2005 6 1.5 late May early Sept USGS X X X X X X
2006 9 1
2007 9 1.1 5/28/07 8/31/07 141 STORET X
2008 (3) (1) 5/27/08 6/26/08 23 STORET X
2009 2 1.5 5/26/09 9/3/09 71 STORET X
IL_QN-12 CHALBION Albion Beach IL798802 2003 2 1.5
-CPD Monitored 2004 17 1
2005 6 1.5
2006 10 1
2007 7 1.3 5/28/07 8/31/07 148 STORET X
2008 1 (3) 1 (1) 5/27/08 8/29/08 70 STORET X
2009 2 1.5 5/26/09 9/3/09 71 STORET X
IL_QN-13 CHTHORNDALE Thorndale Beach IL586992 2003 3 1.3 late May early Sept USGS X X X X X X
-CPD Monitored 2004 19 1 late May early Sept USGS X X X X X X
2005 9 1.7 late May early Sept USGS X X X X X X
2006 11 1
2007 7 1.1 5/28/07 8/31/07 139 STORET X
2008 2 (7) 1 (1.1) 5/27/08 8/29/08 71 STORET X
2009 3 (3) 1.3 (1) 5/26/09 9/3/09 72 STORET X
IL_QO-01 CHNORTH North Ave. Beach IL666876 2003 1 2 late May early Sept USGS X X X X X X
-CPD Monitored
-Indexed Beach (IL666876) covers several Assessment Units (IL_QO-04 –
Armitage and IL_QO-03 - Webster). Individual monitoring is not reported for
these segments, which are considered part of North Ave. Beach.
2004 21 1 late May early Sept USGS X X X X X X
2005 9 1 late May early Sept USGS X X X X X X
2006 8 1
2007 12 1.2 5/28/07 8/31/07 148 STORET X
2008 5/27/08 8/29/08 68 STORET X
2009 2 (3) 1.5 (1) 5/26/09 9/3/09 72 STORET X
IL_QO-02 Fullerton Beach Not Currently Sampled or Monitored - No swimming access
IL_QO-03 Webster Beach Considered part of North Ave. Beach
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-17
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QO-04 Armitage Beach Considered part of North Ave. Beach
IL_QO-05 Schiller Beach Not Currently Sampled or Monitored - No swimming access
IL_QP-02 CHOAK Oak St. Beach IL296528 2003 2 1.5 late May early Sept USGS X X X X X X
-CPD Monitored 2004 19 1 late May early Sept USGS X X X X X X
2005 4 1.5 late May early Sept USGS X X X X X X
2006 7 1
2007 13 1.2 5/28/07 8/31/07 144 STORET X
2008 2 (1) 1 (1) 5/27/08 8/29/08 68 STORET X
2009 1 (2) 2 (1.5) 5/26/09 9/3/09 72 STORET X
IL_QP-03 CHOHIO Ohio St. Beach IL926480 2003 1 2 late May early Sept USGS X X X X X X
-CPD Monitored 2004 15 1 late May early Sept USGS X X X X X X
2005 6 1.2 late May early Sept USGS X X X X X X
2006 9 1
2007 12 1.1 5/28/07 8/31/07 144 STORET X
2008 1 (4) 1 (1) 5/27/08 8/29/08 68 STORET X
2009 3 (4) 1.3 (1.5) 5/26/09 9/3/09 72 STORET X
IL_QQ-01 CH12 12
th
St. Beach IL820929 2003 3 1.7 late May early Sept USGS X X X X X X
-CPD Monitored 2004 42 1 late May early Sept USGS X X X X X X
2005 10 1.3 late May early Sept USGS X X X X X X
2006 17 1
2007 3 1.3
2008 1 (4) 1 (1) 5/27/08 8/29/08 69 STORET X
2009 4 (6) 1.25 (1.2)5/26/09 9/3/09 73 STORET X
IL_QQ-02 CH31 31
st
St. Beach IL461767 2003 5 1.4 late May early Sept USGS X X X X X X
-CPD Monitored 2004 21 1 late May early Sept USGS X X X X X X
2005 13 1.5 late May early Sept USGS X X X X X X
2006 19 1
2007 16 2 7/12/07 8/31/07 80 STORET X
2008 5 (6) 1 (1.2) 5/27/08 8/29/08 70 STORET X
2009 2 (8) 1.5 (1) 5/26/09 9/3/09 72 STORET X
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-18
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QR-01 49
th
St. Beach 2003 late May early Sept USGS X X X X X X
-Not monitored other than during USGS study
-No beach swimming access
2004 late May early Sept USGS X X X X X X
2005 late May early Sept USGS X X X X X X
IL_QS-02 CHJACKSON Jackson Park/63
rd
St. Beach IL118596 2003 7 2.9 late May early Sept USGS X X X X X X
-CPD Monitored
-Dispersal of gulls via canine harassment was conducted in 2007 for trial
purposes and in 2008 full time (Hartmann et al., 2010).
2004 32 1 late May early Sept USGS X X X X X X
2005 17 2 late May early Sept USGS X X X X X X
2006 36 1.0
2007 18 2.5 5/28/07 8/31/07 160 STORET X
2008 (4) (1) 5/27/08 8/29/08 69 STORET X
2009 13 (21) 1.6 (1.1) 5/26/09 9/3/09 76 STORET X
IL_QS-03 CHRAINBOW Rainbow IL814025 2003 2 2 late May early Sept USGS X X X X X X
-CPD Monitored 2004 16 1 late May early Sept USGS X X X X X X
2005 10 1.6 late May early Sept USGS X X X X X X
2006 16 1
2007 16 1.8 5/28/07 8/31/07 146 STORET X
2008 4 (8) 1 (1.3) 5/27/08 8/29/08 72 STORET X
2009 8 (9) 1.1 (1.3) 5/26/09 9/3/09 74 STORET X
IL_QS-04 CH57 57
th
St. Beach IL589159 2003 5 1.8
-CPD Monitored
-Dispersal of gulls via canine harassment was conducted in 2008 full time
(Hartmann et al., 2010).
2004 16 1
2005 12 1.4
2006 16 1
2007 12 1.5 5/28/07 8/31/07 148 STORET X
2008 5/27/08 8/29/08 68 STORET X
2009 8 (12) 1.4 (1.1) 5/26/09 9/3/09 73 STORET X
IL_QS-05 67th St. Beach Not Currently Sampled or Monitored - No swimming access
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-19
Table2-2. Summary of Listed Impaired Beach Descriptions and Monitoring Assessments (continued)
Assessment Units Mapped to Indexed Beaches
Years
BEACH Act
Reporting
Monitoring Records
Water Quality
Sampling
Hydrometeorological Monitoring
Gull Counts
ID305B
Station ID
Water Name
BEACH_
ID
Number of
Closures
Average Duration
of Closures (days)
Monitoring Start
Date
Monitoring End
Date
Number of
Samples
Source
Replicate Samples
Multiple Samples/
Times per Day
Wind Direction
Wind Speed
Air Temp
Water Temp
Wave Category
Wave Height
IL_QS-06 CHSOUTHSHORE South Shore Beach IL581683 2003 4 1.75 late May early Sept USGS X X X X X X
-CPD Monitored 2004 19 1 late May early Sept USGS X X X X X X
2005 13 1.5 late May early Sept USGS X X X X X X
2006 15 1
2007 13 1.5 5/28/07 8/31/07 146 STORET X
2008 2 (9) 1 (1.2) 5/27/08 8/29/08 71 STORET X
2009 2 (11) 1.5 (1) 5/26/09 9/3/09 73 STORET X
IL_QT-03 CHCALUMET Calumet Beach IL376700 2003 3 2.7 late May early Sept USGS X X X X X X
-CPD Monitored 2004 19 1 late May early Sept USGS X X X X X X
2005 12 1.5 late May early Sept USGS X X X X X X
2006 21 1
2007 14 2.1
2008 3 (9) 1 (1) 5/27/20088/29/2008 70 STORET X
2009 7 (7) 1.1 (1.1) 5/26/20099/3/2009 73 STORET X
STORET = EPA’s STOrage and RETrieval System
USGS = U.S. Geologic Survey
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-20
2.1 Sanitary Surveys and Specific Beach Studies
Beach sanitary surveys (BSSs) are used to investigate the sources of bacterial contamination, the
magnitude of the contamination, and the most efficient locations for monitoring and possibly remediation.
Information is collected at the beach as well as in the surrounding watershed. Beach information may
include counting birds on the beach, topography and physical surroundings of the beach, location and
condition of facilities at the beach, and land use. Survey of the water quality should include information
on location of storm water outfalls or other pipes within tributaries to the beach, surface water quality,
and residential septic tank information.
Two beaches within Illinois were included in a 2007 pilot study of Great Lakes beaches sanitary surveys,
63
rd
St. Beach and Rosewood Beach. Two new studies are in the beginning stages of conducting sanitary
surveys at multiple beaches throughout Cook and Lake counties. The following sections provide
descriptions of the completed BSSs, a summary of an additional beach study, and the locations and
sampling and monitoring objectives of the future BSSs.
2.1.1 63
rd
St. BSS (Cook County; Cali et al., 2007)
In summer of 2007, a BSS was conducted for the 63
rd
St. Beach to identify the sources and extent of E.
coli contribution from the adjacent contributing area or beachshed. This survey was conducted by visual
inspection, physical measurement of the beach extent, as well as data sampling and analysis. Between
August and September 2007, beach E. coli samples were taken five days per week and a targeted
watershed survey was completed. The survey includes supplemental data for the 83 CPD routine
sampling events, four survey-driven site visits, and seven survey-driven supplementary sampling events.
This survey indicates that 95% of beachshed is recreational park with manmade ponds and lagoons. The
subdivisions within the park areas are Lagoon/Pond (41.62 acres), Dune/Beach (3.01 acres), Bobolink
Prairie/Grassland (4.00 acres), Savanna/Woodland (1.39 acres), Wooded Island (Paul Douglas Nature
sanctuary in Jackson Park) including Prairie/Grassland (1 acre), Savanna/Woodland (7.5 acres),
Shrubland (7.50 acres), and Jackson Park Golf Course–Prairie/Grassland (3.50 acre). The rest of the
watershed includes Lakeshore Drive and a portion of the Museum of Science and Industry property that
may drain to the lagoon, which drains to the 59
th
St. Harbor and to 63
rd
St. Beach, as well as parking
areas, a fast food restaurant, and several bathrooms and showers.
The site visits were limited to offshore visual inspection; as result, no potential point sources were
identified. Also no outfalls were found along the lakeshore. Sources of bacteria to the beach are strongly
influenced by precipitation, as demonstrated by the strong correlation between rainfall events and
elevatedE. coli concentrations. In the summer of 2007, there were 16 swim bans (E. coli >1,000 cfu/100
mL) and 29 swim advisories (E. coli 235–999 cfu/100 mL) issued for the 63
rd
St. Beach. Ninety-four
percent (15 of 16) of swim bans were preceded by measurable rain in the previous 72 hours, whereas 52%
(16 of 31) of swim advisories were preceded by measurable rain in the previous 72 hours. In addition to
precipitation, elevated E. coli is also related to warm temperature conditions.
The two suggested mechanisms through which precipitation would influence beach E. coli concentrations
are washing off of E. coli from the watershed and resuspension of E. coli accumulated within the beach
sands. During the survey, for example, on September 7, two field samples were collected from puddles
near the beach parkhouse after a rainfall, and both sample results had the maximum E. coli concentration
detectable with the analytical method. In addition, five elevated (1,990–2,420 cfu/100 mL) samples were
also collected in the 59th St. Harbor on the same date. It was raining at the time samples were collected,
and about 0.2 inches of rain had fallen in the previous 24 hours. On the other hand, this survey confirmed
that elevated E. coli populations exist in lake bottom sand/sediments and levels are greater in shallow
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-21
water than deep water. This beach has a relatively shallow water depth (2–5 ft) within the coifing area.
63
rd
St. Beach also has a configuration that may trap water and inhibit the entry and mixing of cleaner
lake water that probably mitigates bacterial concentrations at most other beaches.
Even though these sporadic observations in the adjacent area indicate the existence of E. coli within the
area, one has to identify the original source of E. coli to implement any reduction mechanism. Seagulls
and other birds are likely originators of a large portion of the E. coli load.
2.1.2 Characterization of E. coli Levels at 63
rd
St. Beach (Cook County; Whitman et al.,
2001)
As a response to frequent closures of the 63
rd
St. Beach due to elevated E. coli concentrations, the City of
Chicago and U.S. Geological Survey (USGS) conducted an investigation of E. coli characterization in the
summer of 2001 (Whitman et al., 2001). The goals of the investigation were (1) to identify sources of E.
coli, (2) determine the efficiency of proposed mitigation, and (3) develop an efficient testing and analysis
protocol. As opposed to other sanitary surveys that rely on the source identification by monitoring
outflow and watershed contributions, this investigation relied on very intensive data collection and
analyses of samples and ambient conditions. The data collected at different locations, times, and water
depths were analyzed to understand the extent and movement of E. coli at the 63
rd
St. Beach. The results
of this study highlight the impact of E. coli stored in the sand and the effect of light on E. coli decay and
identify seagull drops as one of the sources of E. coli at the beach. Findings from this study include the
following:

Though highly correlated, E. coli concentrations from sand samples are higher than those from
water samples.

E. coli concentrations exceed the WQS limits in the morning yet are dramatically reduced in the
afternoon. Therefore, sampling time and frequency are essential in monitoring beaches.

As shown elsewhere, samples from shallow water depths contain higher E. coli concentrations
than samples taken from deeper water depths.

Water samples taken in the area where there were more seagulls observed contained higher E.
coli concentrations than other places along the beach. This observation was further supported by
DNA analyses, which related sample E. coli strains primarily to seagulls and other birds.
2.1.3 Rosewood BSS (Lake County; Adam and Pfister, 2007)
LCHD conducted a BSS for Rosewood Beach during the 2007 beach season. During this period, water
samples were collected from the beach and analyzed for E. coli concentration. At the same time, several
hydrometeorological variables (streamflow, wind speed and direction, insolation, air temperature, water
temperature, lake stage, and relative humidity) were monitored. Observations and counts of gulls on the
beach during a total of 150 visits (average 3.7/visit) indicated that gulls are not a significant concern at
this beach. A slope analysis of the beach was also completed. The survey identified that during the
monitoring period 24 out of 141 samples exceeded 235 most probable number (MPN)/100 mL. The
hydrometeorological data were applied to a previously developed predictive regression model named
SwimCast. By combining these monitored data with the previous year’s swimming season records, the
viability of the model was confirmed – the model predictions were correct 90% of the time. Even the 10%
missed predictions were within the 99% confidence limits (Adam and Pfister, 2007).
On November 19, 2007, investigators conducted a source identification assessment along two small
tributaries that drain the 646-acre watershed of Rosewood Beach. The source assessment survey identified
173 pipe structures (potential contaminant sources). These structures, 44 exposed sanitary sewer sources
and 129 storm water sources, were mapped along the two tributaries. Although the survey reportedly
identified all potential contaminant sources, the claim was not verified by the sample data collected
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-22
because there was little to no flow in the tributaries and pipes. Data from one pipe and six locations on the
tributaries were collected. The few samples that were taken from where there was flow showed an
elevated amount of E. coli (e.g., 1,986.3 MPN/100 mL at a small ponded area immediately downstream of
a large storm water pipe). The elevated E. coli concentrations point to potential sources of E. coli to the
lake during wet weather events. These two tributaries are very short in length, both less than a mile to the
lake, so there will be almost a direct transport of E. coli from these unidentified pipes to the lake.
2.1.4 Future BSSs within Suburban Cook County and Lake County
The IDPH and the University of Illinois – Chicago received a joint grant to pursue beach sanitary surveys
at 10 “high-value” beaches throughout Lake County through the Great Lakes Restoration Initiative
(GLRI). High-value beaches are those at which there have been a high number of WQS exceedances/
beach closures, there are significant population impacts, and there is a likelihood of remediation. The
most recent project period is listed as October 1, 2010 to September 30, 2011 (may be adjusted due to
funding delays). Therefore, beach sampling will not begin until the summer of 2011. The following
beaches will be surveyed (listed from north to south):

Waukegan North Point Marina Beach

Illinois Beach State Park Resort

Illinois Beach State Park South

Waukegan South Beach

Great Lakes Naval Base (not on the list of impaired shoreline segments)

Glencoe Beach

Maple Park (Winnetka) Beach

Elder Park (Winnetka) Beach

Kenilworth Beach

Evanston South Beach
The data gained from these sanitary surveys will be of great value to TMDL development along the
Illinois shoreline. These sanitary surveys will be designed to meet all the requirements found in the pilot
program guidance provided by U.S. EPA in 2007. Sampling and monitoring will be designed to assess
usage patterns, weather, water quality, and wildlife influences at the beaches as well as survey the
surrounding areas for sources of bacteria. Previously identified sources have included gulls, storm water
runoff, watershed runoff due to local beach configuration, beach grooming, and algae. The surveys will
target these sources with bacterial sampling. Upon completion of the surveys, plans for remediation will
be developed to address bacterial sources.
2.1.5 BSSs within Chicago City Limits
Many activities that fall within the realm of sanitary surveys have been conducted within Cook County.
However, many of these activities are not formal projects with corresponding reports or peer-reviewed
articles (C. Breitenbach, personal communication, 2010). CPD applied for and received a grant from the
GLRI to perform a number of BSSs. Starting in the fall of 2010 with a completion date of December
2011, CPD will survey nine beaches within their management domain. CPD and IDPH have coordinated
to ensure that the two future sanitary surveys conducted along the Illinois Lake Michigan shoreline will
not overlap. CPD will follow similar guidelines and have similar objectives to the surveys previously
described for suburban Cook County and Lake County. Between the two sets of studies, CPD and IDPH
aim to gain significant insight into bacterial sources at a wide range of beaches along the shoreline.
CPD will survey the following beaches (listed from north to south):
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-23

Hollywood/Ostermann Beach

Foster Beach

Montrose Beach

31st St. Beach

57th St. Beach

Jackson Park/63rd St. Beach

South Shore Beach

Rainbow Beach

Calumet Beach
2.2 Existing Modeling Studies
The local health departments, USGS, and several other Federal, State, and local agencies have all
conducted modeling studies along the Illinois Lake Michigan shoreline. Models completed to date fall
into general categories of predictive modeling for beach closures and modeling of the lake waters to
determine fate and transport of contaminants within the waters. Examples of modeling applications of
actual source derivation or delivery to the lake waters from sources expected along the Illinois shoreline
have not been found within the published literature at this time.
2.2.1 Predictive Modeling for Beach Closures
There are numerous examples of predictive regression models throughout the Great Lakes region where
hydrometeorological variables are used to predict the bacteria concentration at a beach each day (Table 2-
3). Predictive models help determine when a beach should be closed but do not help determine what the
main sources of the bacteria are or what can be done to improve the water quality at the beaches. As such,
these models can be considered response models.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-24
Table2-3. Examples of Predictive Regression Models Used for Beach Closures
within the Great Lakes
Location
Indicator
Bacteria
R2
False
positive
False
negative
Correct
Predictions
Source Intensive
Explanatory
Variables
Source
Ohio bathing
beaches
E. coli
17-58 4-15.1% 3.9-14.6%73-91%
Turbidity, Number
of birds, Previous
day’s E. coli
concentration
Francy et al.,
2003
Huntington &
Edgewater
beaches, Ohio -
Nowcast 2008
E. coli
42/ 37
34/ 45
(count)
25/ 34
(count)
277/ 283
(count)
Turbidity, Wave
height
Francy et al.,
2009
Southern Lake
Michigan urban
beaches
E. coli
32-64 none
3-6
(count)
NA
Chlorophyll,
Stream and lake
turbidity
Nevers and
Whitman,
2005
Southern Lake
Michigan just
north of Chicago
E. coli
65-76 NA NA 87%
Hourly averages of
variables, Stream
loading of E. coli
Olyphant,
2005
Milwaukee area
on Lake
Michigan
E. coli
3-29 NA NA NA
Wind indexes,
Hours since last
rainfall
McLellan and
Salmore,
2003
2.2.1.1 SwimCast (Lake County, Illinois)
This model provides predictions for three beaches (Forest Park, Rosewood, and Waukegan South) on
Lake Michigan in Lake County, Illinois. It relies on explanatory variables that include air and water
temperature, wind speed and direction, precipitation, relative humidity, wave height, lake stage, insolation
(light energy), and other water quality parameters. This system operates approximately between Memorial
Day and Labor Day (LCHD, 2009). Using money from the GLRI, additional SwimCast stations will be
installed at Calumet, Montrose, and Foster Avenue beaches in the 2011. These stations consist of water
sensors in or near the surf zone and meteorological towers that are installed in the water close to the
beach.
One of the SwimCast models in use in 2009 correctly predicted whether E. coli concentrations were
above or below the WQS between 85% and 86% of the time. SwimCast predictions rely on the 99%
confidence interval of the prediction, which indicates the lower and upper bounds of bacterial
concentrations between which the actual bacteria concentration is expected to lie. For SwimCast beaches,
the beach swimming advisories and bans are posted based on where the upper and lower bounds of this
confidence interval lie in relation to the WQS. For instance, if the lower bound of the 99% confidence
interval prediction is above 235 cfu/100 mL then the beach is posted with a red flag and a swimming ban
is put in place. When the upper bound of the 99% confidence interval prediction is above 235 cfu/100 mL
but the average prediction and the lower bound of the 99% confidence interval prediction are below 235
cfu/100 mL, a swimming advisory is posted (NRDC, 2010).
2.2.1.2 Chicago Regional Regressions (Whitman and Nevers, 2008)
Whitman and Nevers (2008) undertook a study to extend typical predictive beach-specific regression
models to a region using 23 beaches in metropolitan Chicago and monitoring data from 2000 through
2005. Their findings indicate that these beaches do respond similarly in terms of temporal fluctuations in
E. coli concentration with simultaneous peaks and troughs. They also examined spatial correlations
between the beaches resulting in a northern and southern grouping of beaches. These spatial correlations
also revealed that beaches that are more closely situated to one another have concentrations that are more
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-25
closely correlated. Significant regression variables for predicting E. coli concentrations in the regional
model include Julian day, wave height, and barometric pressure. It is also important to note that beaches
along the southern coast generally had higher mean E. coli concentrations with a few exceptions. Also,
Julian day explained the greatest amount of variance, with E. coli concentrations increasing monthly
throughout the summer for all years except 2004. These findings show that there are inherent regional
fluctuations along the Chicago coastline, most likely due to the currents within Lake Michigan, and that
groupings of beaches can effectively be used to look at shoreline response to bacterial loadings.
2.2.1.3 Examples of Other Great Lakes Predictive Models for Beach Closures
Ohio Nowcast (USGS): These models, which are recalculated each season and some years sub-
seasonally, provide predictions for two beaches (Huntington and Edgewater) on Lake Erie and for the
Cuyahoga River. Explanatory variables for the models include turbidity, gridded rainfall data, wave
height, and day of the year. Predictions rely on the probability threshold surrounding the model estimate
and not the concentration estimate itself. The advisories from the models are currently available from
approximately Memorial Day to Labor Day. In recent years these models have resulted in correct
predictions approximately 61% to 85% of the time (Francy et al., 2009).
Project S.A.F.E. (USGS): This model provides predictions for four beaches (Ogden Dunes, Wells Street,
Marquette, and Lake Street beaches) on Lake Michigan in Lake and Porter counties, Indiana, based on
current conditions, turbidity, chlorophyll content, and color. The model predicts the likelihood that the E.
coli count will exceed safe limits, and, on that basis, the beach manager chooses whether to issue an
advisory or closing. Results for the swimming season are hosted on the Indiana Department of
Environmental Management’s website (NRDC, 2010).
2.2.1.4 Other Predictive Modeling Notes
Most of predictive models focus on estimating E. coli concentrations using measured hydro-
meteorological variables in the watershed and water body. These models can be augmented by including
land use characteristics of contributing areas. For example, Kelsey and others (2004) developed predictive
regression models enhanced by using land use properties of the watershed area for a receiving estuary.
Land use variables considered included distances to nearest urban and rural land uses, weighted distances
to number of housing units and population, housing and population density in nearest sub-watersheds, and
waterway depth, width, and distance to estuary mouth. Variables that show significant correlation help
indirectly identify possible location of E. coli sources.
Lin and others (2008) combined a hydrodynamic model with artificial neural networks (ANNs) to
develop an accurate and rapid tool for assessing the bathing water status of the Ribble Estuary, United
Kingdom. Fecal coliform was used as the water quality indicator. The hydrodynamic simulation outputs
were used to train the ANN model, which is the same as estimating the regression coefficients. The
advantage of developing the ANN in addition to the numerical model is to have rapid predictions similar
to SwimCast in Illinois.
2.2.2 Fate and Transport Modeling
Just as important as the loading of contamination to the beach is the fate and transport of the bacteria units
once they reach the beach, swash zone, and lake water. Given the Lake Michigan setting, modeling or
simple observations of these processes become much more difficult than in a typical water body such as a
river where water flows in one direction or in a smaller lake where there are no overriding currents with
direction changes. Therefore, applications to date have relied on larger-scale current-based models such
as the Princeton Ocean Model (POM). Two studies are summarized that provide insight into the Lake
Michigan effects on bacteria fate and transport.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-26
2.2.2.1 Loading and Transport at 63
rd
St. Beach (Ge et al., 2010)
Given previous research at 63
rd
St. Beach hypothesizing that the breakwaters surrounding the beach cause
bacteria and sediment to persist within the nearshore waters (Whitman et al., 2001), this study was
conducted to model the hydrodynamic system at this embayed beach. Additional multiple linear
regression analysis was conducted on monitoring data collected for the study to examine the impacts of
onshore waves. These regressions were significant in relating the impact of onshore wave cases with
variables representing the resuspension of sediments and the interaction of the wave swash zone with gull
droppings on the beach, thereby explaining the higher concentrations and variability in E. coli
concentrations in knee-deep waters at the beach (i.e., the area where water quality samples are typically
collected during beach monitoring).
Hydrodynamic modeling using parameters derived from the POM examined cases of longshore currents
coming from both the southeast (Figure 2-6) and the north (Figure 2-7). Both conditions revealed circular
current patterns within the embayment. These gyres were shown to trap bacteria from the swash zone of
the beach within the embayment, eventually releasing it when the currents shifted.
This hybrid modeling approach in which a hydrodynamic model is combined with a statistical regression-
based analysis has important implications for the beaches along Illinois’ shoreline. First, in this case the
results from each component were shown to support one another. Second, each component provided
additional details on the fate and transport of bacteria within the lake waters that could not be detailed by
one component alone. Considering the number of beaches along the shoreline that have breakwaters or
other physical structures bounding their swimming areas, this methodology could add much detail to the
information required for the formulation of TMDLs at individual beaches.
Figure2-6. Current flow pattern around Chicago 63
rd
St. Beach
driven by an external current entering the computation domain
through boundaries A-B and B-C and exiting at A-D (Ge et al., 2010).
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-27
Figure2-7. Current flow pattern around Chicago 63
rd
St. Beach driven by an
external current entering the computation domain through boundary A-D and
exiting at B-C (Ge et al., 2010).
2.2.2.2 Tributary Dominance in Southern Lake Michigan (Thupaki et al., 2010)
The POM was also used to examine the plume impacts of tributaries to Lake Michigan in consideration of
the lake currents. This work was carried out in Southern Lake Michigan along 72 km of shoreline in
Indiana. The primary beaches impacted were Mt. Baldy and Central Avenue, and the three main
tributaries contributing plumes to the lake were Burns Ditch near Burns Harbor, Trail Creek at Michigan
City Harbor, and Kintzele Ditch. Although these results are site specific and the Illinois shoreline does not
have any tributaries as great as the three examined in this study (unless the Chicago River is once again
allowed to flow into Lake Michigan), they do point to the complicated lake dynamics affecting bacteria
fate and transport highlighted in the previous summarized study (Section 2.2.2.1). Important conclusions
from this study indicate that “dilution due to advection and diffusion accounted for a large portion of the
total [E. coli] budget in the nearshore, and the net [E. coli] loss rate within the water column (cfu/m
3 ∙ s)
was an order of magnitude smaller compared to the horizontal and vertical transport rates. However,
consideration must be given to the fact that vertical exchange is strongly coupled with depth-dependent
loss processes such as solar inactivation and settling” (Thupaki et al., 2010).
2.3 Source Identification Studies
Although there have been two BSSs and numerous studies to aid in the prediction of when advisories
should be issued, there is still less understanding in terms of the specific sources of E. coli and their
magnitudes at each of the Illinois shoreline segments. Based on a report from the National Resources
Defense Council (NRDC), the sources of E. coli contamination at Illinois’ Lake Michigan beaches are
unknown 81% of the time (NRDC, 2010). The following sections provide a summary of further research
that highlights more of the source and transport of bacteria to the beaches. The likely sources of E. coli to
the beaches will guide the selection of methods used to best inform TMDL calculations and development.
2.3.1 Lake Influences including Shoreline Topography
The Great Lakes are closed basins so, although they are large enough to be influenced by the earth’s
rotation, they are actually dominated by their coastal processes. The combination of physical factors along
the coasts provides for complex and unique hydrodynamics within the lakes. For this application of
TMDL development along a shoreline, it is particularly important to note that the “physical transport
processes are often the dominant factor in mediating geochemical and biological processes in the coastal
environment” (Rao and Schwab, 2007). In addition, the coastal regions are coupled to a greater or lesser
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-28
degree by exchanges with mid-lake waters involving transport of materials, momentum, and energy. In
terms of physical processes, humans and other natural processes have influence over the nearshore zone
through such occurrences as sewage outfalls, rivers, and nonpoint sources.
As described by Rao and Schwab (2007), the coastal zone is divided into three regimes, (1) the nearshore
area consisting of swash and surf zones; (2) frictional and inertial boundary layers, together known as
coastal boundary layer; and (3) the open lake. Within the swash and surf zones, wave run-up and breaking
waves are dominant sources of mean flows. Waves caused by wind stress are also important in the surf
and swash zones where they act as the main cause of shore erosion. The swash zone forms the boundary
between the surf zone and the backshore – the area not affected by the wave action. The surf zone, the
area where transport processes are extremely intense, is the area of water between the swash zone and the
seaward side of breaking waves.
Wave-generated currents, which transport beach material, are an important factor in beach stability. The
direction of littoral drift is determined by the angle of wave approach and associated longshore currents
and is particularly critical during periods of high lake levels. Bottom friction also plays a significant role
in nearshore dynamics, where the presence of the shoreline acts as a lateral constraint on water
movements, tending to divert currents so that they flow nearly parallel to the shoreline (Rao and Schwab,
2007).
Within Lake Michigan, circulation is highly episodic, with the most energetic currents and waves
occurring during storms. In the shallow waters of the southern basin of Lake Michigan, the water near the
coast moves in the direction of the wind. Such processes and characteristics subject this basin to recurrent
episodes of sediment resuspension, especially during storms (Rao and Schwab, 2007). Figures 2-8 and 2-
9 present average Lake Michigan currents as modeled over various time periods and seasons. In general,
the Illinois shoreline is subjected to a longshore current from north to south.
Figure2-8. Six- and 12-month average depth-averaged currents for
1982-83 and 1994-95 (Beletsky and Schwab, 2001).
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-29
Figure2-9. Observed mean summer, winter, and annual circulation in Lake
Michigan during 1982-1983. (Isobaths are every 50 m) (Beletsky et al., 1999).
So while Lake Michigan itself does not generate any bacteria sources, the currents and resuspension
actions caused by waves and wind stress within the lake are transport processes that can transfer large
sources of bacteria from one area to another or reintroduce settled bacteria into the water column. Thus,
these processes should be accounted for in TMDL development.
2.3.2 Animal Sources
Animal sources of fecal bacteria have been shown in a variety of studies to contribute to the
contamination of swimming beaches. Although the linkage of elevated bacteria concentrations at a beach
is not as clear as, say, elevated levels within the grazing lands of a cow farm, several different studies
along the Illinois shoreline have contributed supporting information to this conclusion. Methods have
ranged from simple counting of gulls on the beaches and nearby lands to using sophisticated DNA-based
identification techniques on water samples collected from the swimming waters. Following are
summaries of two published studies.
2.3.2.1 DNA-Based Study in Lake County (Soucie and Pfister, 2003)
In 2002 and 2003, Lake County conducted source identification studies of E. coli at beaches along Lake
Michigan using ribotyping. In 2002, beaches sampled were North Point Marina Beach, Waukegan
Municipal Beach South, Lake Forest Beach, and Rosewood Beach in Highland Park. Samples were
collected randomly throughout the study period. In 2003, sampling moved to Illinois Beach State Park
South, and a second round was conducted at Lake Forest and Rosewood beaches. Beach water samples
were collected daily, but only those collected in series with samples containing E. coli above the WQS
were submitted for ribotyping. To provide additional ribotyping patterns/isolates to the comparison
library, samples were taken of animal feces at the beaches and from sanitary sewers. The study shows the
presence of various animals at different percent contributions across the five beaches (Figures 2-1 and
2-
11).
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-30
Figure2-10.E. coli sources in 2002 beach water study (Soucie and Pfister, 2003). (NPMB = North
Point Marina Beach; WMBS = Waukegan Municipal Beach South; LFB = Lake Forest Beach; RWB
= Rosewood Beach)
Figure2-11.E. coli sources in 2003 beach water study (Soucie and Pfister, 2003). (IBSPS = Illinois
Beach State Park South; LFB = Lake Forest Beach; RWB = Rosewood Beach)
In 2002, the dominant sources of E. coli were gulls (~50%). Other identified sources included raccoons,
deer, and humans/sewage (in order of increasing contributions). A small portion (<5%) at each of the
beaches was identified to be other animals (e.g., pigs and cows), while a large percentage (~30%) of
samples at each beach were from unknown sources.
In 2003, the dominant source was by far avian (although not specifically identified as gulls in this study),
with an average of 62% of the contribution across the three beaches. Other contributing sources were
rodents, dogs, and humans/sewage, again in order of increasing contribution. In this analysis, a much
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-31
smaller percentage of the contribution was unidentified (~10%), likely due to the development of
additional ribotype patterns.
This study revealed the large contributions of gulls and other birds to the E. coli contamination at
suburban Lake County beaches. However, applying the same methods to a large number of beaches (51)
would be limited by the high cost of the DNA-based methods. Given the beaches already monitored with
these methods, targeted studies should be conducted at beaches found to have different conditions.
2.3.2.2 City of Chicago Ring-Billed Gull Study (Hartman et al., 2010)
An exponentially increasing number of gulls in the Chicago area has led to a variety of damages and
economic losses, including swimming bans in the area. As a result, the City of Chicago undertook the
2009 Chicago Ring-Billed Gull Damage Management Project (pilot studies were completed in 2007-
2008). The objectives of the project were to “(1) reduce the local production of ring-billed gulls, (2)
reduce the severity of conflicts with gulls including the issuance of swim advisories/bans, and (3) evaluate
how limiting the production of gulls affects gull counts at Chicago beaches” (Hartman et al., 2010).
Overall, gull population management through the oiling of eggs under this project reduced the number of
hatching gulls by between 21,000 and 42,000 starting in July of 2009. Techniques used to control the gull
population, in addition to oiling of eggs, included more frequent and localized trash pickup, education of
the public, and beach grooming programs. Canine harassment was also used to reduce the number of gulls
at selected beaches (Foster Avenue in 2006 and 2007, 63
rd
St. Beach in 2007 and 2008, and 57
th
St.
Harbor in 2008). Data collected in 2008 showed that canine harassment was effective at reducing the
number of gulls, and, during the period of full-time canine harassment, swim advisories/bans at 63
rd
and
57
th
St. beaches were nearly eliminated. The following year when canine harassment was not used, the
numbers of swimming advisories/bans were again at elevated levels. Considerations in using canine
harassment include where the harassed gulls will ultimately settle (e.g., could be another beach) and
whether canines can have full access to the beach (e.g., some natural dune areas may be harmed by
canines).
Using statistical analysis to determine the effectiveness of the gull management project, the authors found
that the proportion of water quality tests requiring issuance of a swim advisory or ban declined during
each of the subsequent 3 years following 2006 during which gull management programs were in place.
“The most notable improvement was experienced when comparing swim advisories/bans between 2006
and 2009, when 18 of 19 beaches had a lower proportion of tests in excess of the beach water quality
exceedance threshold. Of those 18 beaches, seven had a statistically detectable reduction at P< 0.1”
(Hartman et al., 2010).
2.3.3 Point Sources (CSOs, WWTPs, Storm Water Outfalls)
Within Lake County there is limited information on point sources to the beaches; however, most of the
beaches have small drainage areas and therefore are less likely to contain significant point sources (M.
Adams, personal communication, 2010). In 2009 as part of the Lake Michigan Ecosystem Partnership’s
investigation into stresses and opportunities in Illinois Lake Michigan watersheds, the Alliance for the
Great Lakes surveyed and mapped shoreline ravines within Lake and upper Cook Counties (Alliance for
the Great Lakes, 2009). As depicted in Figure 2-12 for the ravine entering Lake Michigan at Rosewood
Beach, numerous potential point sources were identified within almost all surveyed ravines. Point
sources within the Rosewood Beach ravine have been fixed in many cases after a separate BSS (discussed
in Section 2.1.3). However, further investigation is needed to determine the sources of the discharges in
the other ravines that enter Lake Michigan near impaired beaches and whether these discharges provide a
source of bacteria to the beach during wet weather, or even dry, conditions. The data collected by the
2009 ravine study provides a starting point for the identification of the sources, which should be further
explored during TMDL development.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-32
Figure 2-12. Rosewood Beach ravine (3L) as defined in the 2009 Strategic Sub-Watershed Identification Process Report (Alliance for the
Great Lakes, 2009)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-33
Unidentified point sources within ravines and along the shoreline may come from individual entities (e.g.,
homes or businesses), storm water systems, or illegal connections to sewer systems. To begin identifying
those point sources that are linked to storm water systems the Municipal Separate Storm Water Sewer
System (MS4) permits for municipalities along the shoreline within Lake and Cook Counties should be
examined. There are approximately 162 municipalities within these two counties that have MS4 permits
(Figure 2-13). It is likely, however, that many of these systems discharge to inland waters and man-made
conveyance systems and not directly to Lake Michigan. For instance, Evanston is primarily a combined
sewer system where most storm water is conveyed via the combined sewer system to the Metropolitan
Water Reclamation District of Greater Chicago (MWRDGC) interceptor, then to the deep tunnel.
Figure 2-13. Municipalities covered by MS4 Permits within Lake and Cook Counties.
The recent addition of electronic filing for notices of intent and annual reports required by the MS4
permits within Illinois provides a central location to begin review of the receiving water for each permit
as well as the steps taken by each municipality to reduce storm water volume and contamination. These
documents can be viewed at
http://dataservices.epa.illinois.gov/NoticesofIntent/MS4QuickSearch.aspx
.
Besides storm water, general National Pollutant Discharge Elimination System (NPDES) permits can also
provide information on potential sources of bacteria to Lake Michigan. The compilation of data to date
shows that there are some mapped discharges within Lake Michigan, but many have been listed as
inactive or minor or are from operations from which little discharge of bacteria is expected. The active
NPDES permits (Table 2-4) shown to discharge to Lake Michigan must be investigated further.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-34
Table2-4. Active NPDES Discharges along Lake Michigan Shoreline within Illinois
NPDES
Facility Name
Owner
Major/
Minor
River Basin
Hydrologic
Unit
Flow Rate
(MGD)
IL0001881 Abbott Laboratories-N. Chicago Private Major Lm/Western Shore 04040002 25.500
IL0002267 Outboard Marine-Waukegan Private Major Lm/Western Shore 04040002 2.230
IL0002755 R. Lavin And Sons,Inc.-Chicago Private Major Lm/Western Shore 04040002 1.400
IL0030244 Nssd Waukegan Stp Public Major Um/Chicago-Calumet 04040002 22.000
IL0002259 Midwest Generation,Llc-Waukegn Private Major Lm/Western Shore 04040002 795.60
IL0002691 Usx-Uss South Works Private Major Lm/Calumet-Burns 07120003 79.530
IL0002763 Commonwealth Edison-Zion Private Major Lm/Western Shore 04040002 1815.0
IL0002364 Winnetka Water & Electric Public Minor Lm/Muskegon 04040002 13.730
IL0002411 Fansteel Inc. Private Minor Lm/Western Shore 04040002 0.016
IL0035173 333 Building Corporation Private Minor Um/Chicago-Calumet 07120003 0.000
IL0035238 Chicago Union Station Private Minor Um/Chicago-Calumet 07120003 0.001
IL0036536 Evanston Cso Public Minor Um/Chicago-Calumet 07120003 0.000
IL0049883 Highwood Wtp Public Minor Lm/Western Shore 04040002 0.000
IL0066435 Abbott Laboratories Private Minor Um/Chicago-Calumet 04040002 4.010
IL0066605 Mcl Management Corporation Private Minor Um/Chicago-Calumet 07120003 0.720
IL0069809 Johns Manville Private Minor Um/Fox R. 04040002 0.000
IL0069981 Wilmette-Greenleaf Csos Public Minor Um/Chicago-Calumet 07120003 0.000
IL0073741 Trigen-Peoples Dist Energy Co Private Minor 07120003 9.036
IL0001996 Chicago-Jardine Wtr Purif Plt Public Minor Um/Chicago-Calumet 07120003 0.000
IL0002429 Chicago South Wtp Public Minor Um/Chicago-Calumet 07120003 0.000
IL0066541 Northwestern University Private Minor Lm/Western Shore 07120003 0.000
MGD = million gallons per day
Discussions with stakeholders confirmed that there are likely few large sources of bacteria to the Lake
Michigan shoreline due to point sources. There are no point sources within the urban Chicago beaches
area except for a small storm water drain for a nearby field at Calumet Beach (C. Breitenbach, personal
communication, 2010). Most CSOs, waste water treatment plants (WWTPs), and storm water outfalls
discharge to the Chicago River and other Chicago-area waterways. Due to the human alterations to this
river system, the river no longer discharges to Lake Michigan unless extreme storm conditions require an
opening of the locks separating the river system from Lake Michigan. These conditions are discussed in
Section 2.3.5.
2.3.4 Surface Flows/Runoff/Harbors
Along the Illinois shoreline there are very few stream outlets aside from the Chicago River. The streams
that drain the shoreline to the north are small in scale, ranging up to 10 km in length with a drainage area
of up to 14.2 km
2. Ravines exist along the shoreline (Figure 2-12) but appear to be more of conduits for
delivery of point source loads to the shoreline (Alliance for the Great Lakes, 2009). As such, their impacts
have been described in Section 2.3.3.
Within Chicago there are nine lakefront harbors stretching from Montrose Harbor in the north (just south
of Montrose Beach) to Jackson Harbor in the south (just south of 63
rd
St. Beach). The CPD Harbors have
space for more than 5,000, boats making it the Nation’s largest municipal harbor system. These harbors
are also a haven for gulls, where large populations can often be found. Boating within these harbors is not
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-35
expected to be a source of E. coli because city mandates require that boat sanitary disposal hatches be
sealed shut (C. Breitenbach, personal communication, 2010). However, there is a possibility for storm
water accumulation within these harbors, which could be a source of bacteria.
Little published information is available on the quantification of nonpoint sources along the beaches,
which, given the lack of point source discharges or other easily identifiable sources, appear to be a
dominant source of bacteria to Illinois’ Lake Michigan swimming beaches. McLellan and Salmore (2003)
conducted a detailed monitoring study of a public beach within Milwaukee that included both dry and wet
weather sampling across multiple shoreline and offshore sites for E. coli. Their findings indicate that, for
both wet and dry conditions, shoreline sites have significantly higher E. coli levels than offshore regions
where the shoreline samples exceeded the single sample WQS 66% of the time. They also found that
these high levels coincided with the presence of birds and storm water at the swimming beaches, but that
the high levels were not correlated with E. coli levels in a connecting harbor. The authors concluded that
local, persistent contamination is likely the major source of high E. coli levels over regional sources.
2.3.5 Chicago Area Waterway System (CAWS) Influences
The CAWS, under control of the MWRDGC, consists of 78 miles of man-made canals and modified river
channels that support commercial navigation. Over 70% of the river volume originates from the discharge
of treated municipal wastewater effluent (i.e., point source) from four water reclamation plants (WRPs).
Additionally, it receives storm water, tributary streams, and runoff from urban and rural areas. CSOs
discharge from Chicago systems (200), suburban systems (89), and the MWRDGC system (27). It also
supports recreational activities (e.g., boating, fishing, streamside recreation) and provides habitats for
wildlife (MWRDGC, 2008). The CAWS was designed to divert water from Lake Michigan into the Des
Plaines and Calumet rivers rather than having the rivers flow into the lake. By U.S. Supreme Court
Decree, the District is allowed specific volumes of Lake Michigan water as discretionary diversion.
Currently this volume is 270 ft
3/s. This diversion is used primarily in the critical summer months to
improve the water quality of the District waterways. However, reversals from the CAWS back into Lake
Michigan can also occur. The number of reversals from the CAWS to Lake Michigan has been reduced
with the onset of the Tunnel and Reservoir Plan (TARP) in 1972.
As succinctly described in a recent MWRDGC bypass monitoring report:
Lake Michigan and the inland CAWS are separated by locks at the mouth of the Chicago
River and the Calumet River, and by gate structures that control the amount of water
withdrawn from the lake, and allow release of excess river water into the lake during
relatively severe storm events [Figure 2-14]. When the collection system receives excess
flow which cannot be diverted into TARP, the CAWS water elevation rises to flood
stage, and it becomes necessary to open the locks and reverse the flow to Lake Michigan.
The District controls the water level through its operation of lakefront structures: the
Wilmette Pumping Station (WPS); the sluice gates at the Chicago River Controlling
Works (CRCW); and the sluice gates at the O’Brien Lock and Dam (OLD). The District
conducts its operations to ensure that release of excess floodwaters into the lake is a last
resort, when all the District WRPs are operating at their maximum capacity and the
waterways are approaching or exceeding flood stage. During the lake diversion events,
the District conducts water quality monitoring to assess the effects of bypassing storm
flows from the CAWS to the lake (MWRDGC, 2010).
There are two types of reversals— gate reversals and lock reversals. The more common is a gate reversal,
which is characterized by a smaller volume of water released through gates adjacent to the lock. In a lock
reversal, the locks are opened to maximize flow. Lock reversals allow a much greater volume of water to
flow back to the lake. They are only necessary in cases of severe storms and have only occurred three
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-36
times between 2000 and 2010. However, gate reversals have occurred several times in the past decade.
The date, duration, and volume of water released are available for each of these reversals (Table 2-5).
Figure 2-14. CAWS Control Structures (Lanyon, 2010).
Given the availability of the reversal data and MWRDGC’s monitoring efforts during the events, attempts
can be made to correlate spikes in E. coli concentrations at nearby beaches with these releases, depending
on the hydraulic conditions of the lake during the surrounding time period. For instance, the June 2009
event monitoring results reveal a current traveling south from the Wilmette Harbor mouth where a large
pulse of the release was shown to impact Lighthouse and Northwestern Beaches in the early morning
hours, with a reduction of E. coli levels back to below WQS by the evening. Therefore, although these
beaches are significantly impacted by a release of storm water from the CAWS, the impact appears to act
as a pulse and is not long in duration.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-37
Table2-5. Reversals to Lake Michigan in Millions of Gallons
Year
Date
O’Brien Lock
and Dam
Chicago River
Controlling Works
Wilmette
Pumping Station
Total Volume
2010 7/24/10 5784.6 750.3 6534.9
2009 6/19/09 191.6 191.6
3/8/09 143.1 143.1
2/26-27/09 78.9 78.9
2008 12/27-28/08 480.8 480.8
9/13-16/08 2669.2 5438.2 2941.7 11049.1
2007 8/23-24/07 224 224
2006 None 0
2005 None 0
2004 None 0
2003 None 0
2002 8/22/02 1296.4 455.4 1751.8
2001 10/13/01 90.7 90.7
8/31/01 75.3 75.3
8/2/01 883.1 139.9 1023
2.3.5.1 Separation of the Great Lakes and the Mississippi River Watersheds
Because of the public visibility of invasive species (particularly Asian carp) advancing up the Illinois
River and into a canal leading toward Lake Michigan, there are calls to begin investigating the options to
disconnect the man-made waterways connecting the Great Lakes and Mississippi River watersheds.
Support for the investigation comes from the general public and elected officials in the Great Lakes
States, including Chicago Mayor Richard Daley. Members of Congress from Great Lakes States have
proposed legislation calling on the U.S. Army Corps of Engineers to expedite a study into hydrological
separation of the two important basins by which the flow of the Chicago River would be re-reversed into
Lake Michigan.
Changing the Chicago River’s flow would also force the MWRDGC to disinfect the treated, but still
bacteria-laden, sewage it now dumps into the Chicago River — something many environmentalists say is
long overdue. Assuming there is a special condition for the NPDES permit biomonitoring requirement at
the water reclamation plants, treatment will need to be added for bioaccumulative chemicals of concern
amounting to billions of dollars in capital and hundreds of millions of dollars in operating expenses. Of
the 316 CSOs discharging to the CAWS, treatment, which would be required if the CAWS were to drain
to Lake Michigan, is only feasible at 253 of them. Treatment systems at these sites are estimated to cost
$2.9 billion to establish, with an annual operating and maintenance expense of $14 million. There are
several key issues to be considered in pursuing a full separation of the systems: (1) pollution of Lake
Michigan from urban discharges, (2) threat to diversion continuation, (3) floodwater relief in extreme
storms, (4) stability of riparian structures, and (5) stagnant canal reaches (Lanyon, 2010).
Given the stage of hydrological separation study development, the high costs likely to be expended in
constructing and maintaining any engineering controls, and the public scrutiny of such an undertaking, the
separation of the watersheds it not expected to occur in the near future. Therefore, due to the need to
establish TMDLs in the coming years, hydrological separation of the Great Lakes and Mississippi River
watersheds will not be an influencing factor in development of the current TMDL options.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-38
2.3.6 Watershed Sources
In an effort to investigate the watershed contribution of E. coli, Whitman and others (2006) conducted
studies for 10 years in the temperate coastal lacustrine and fluvial watershed of Lake Michigan. This
study attempts to identify whether the accumulation of E. coli in the watershed or sand layers persists for
longer periods and eventually appears in the water body as a result of precipitation or swash and/or wave
actions. In regard to Lake Michigan, their study shows that for 63
rd
St. Beach and West Beach, the
groundwater discharging to the lake contains no E. coli. Additionally, well tests rarely show substantial E.
coli content. However, the wells at the beach show some content of E. coli when the sediments are
saturated by the contaminant. The vertical distribution of E. coli observed at the 63
rd
St. Beach reveals
high amounts of E. coli in the first 30 cm and exponential reduction below 30 cm. This observation
suggests a significant impact of resuspension of E. coli by wave and wind action. Whitman and others
(2006) estimated that such resuspension (5 to 6 log E. coli/100 mL) would result in exceedances to the
current U.S. EPA WQS (2.38 log E. coli/100 mL) for beach closures. This indicates that in some cases
swimming bans/advisories may be caused by the existing bacteria load in the shoreline sediments alone
without any additional time-dependent sources.
2.3.7 Other
2.3.7.1 Adopt-a-Beach
The Adopt-a-Beach program run by the Alliance for the Great Lakes conducted a beach health assessment
of over 60 beaches along the Illinois shoreline in 2008 and 2009. Although the program is conducted year
round, most of the visits occurred between April and September. Among the 51 shoreline segments listed
as 303(d) bacteria impaired water, 29 were assessed by the Alliance. For 2008 there were 118 visits,
which increased to 121 visits in 2009. The beach assessment includes observations, sampling of water
quality, and trash tracking and pick-up timing. This effort of the Alliance provides an opportunity to
augment the regular monitoring of beaches by providing a wide range of data on ambient conditions (e.g.,
weather), wild and domestic animal counts, and E. coli measurements. The group has E. coli sampling
analyses for all but 18 visits in 2008 and all visits in 2009. Some of the E. coli measurements are higher
than the 1-day limit of 235 cfu/100 mL. Overall in 2008 among 100 samples, 12 samples (12%) were
measured at greater than 300 cfu/100 mL. In contrast, in 2009, among 121 E. coli measurements 45
samples (37%) had concentrations greater than 300 cfu /100 mL. (Note that the sample results are
rounded to the nearest hundred.)
The litter collection and monitoring effort generated a count of more than 3,000 pieces of litter among 45
categories during just the 2009 visits. For general water quality, the Alliance identified that some
potential sources were present at more than one of the beach locations, including

Poorly maintained or inadequate garbage cans: Garbage — especially leftover food — can attract
gulls, raccoons, and wildlife to the beaches.

High bird counts: Gulls and other birds are a natural part of the Great Lakes ecosystem. Many
once migrated with the seasons, but increasing urbanization has led to an abundance of food and
the development of resident populations of seagulls and geese. Geese and gulls that spend their
days in coastal areas leave behind fecal material high in bacteria.

Floating trash in the water: Floating trash is a direct source of pollution and can increase bacterial
contamination at the beach.

Outfall pipes and other potential sources of pollution to beaches: Volunteers are tracking water
flow from these sources.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
2-39
2.3.7.2 Observations
During a meeting of local stakeholders for the Illinois Lake Michigan beaches, the following observations
were made concerning significant sources of E. coli to swimming beaches:

Within Lake County, E. coli concentrations are likely influenced by gulls, topography of beaches
(particularly with low slopes), and aging sanitary infrastructure.

There are isolated cases where beach closures are caused by CSOs. Although there have been
precautionary closures, the number of beach closures associated with CSOs are few. Gulls,
nonpoint source, and storm water appear to be the greater threat.

CPD research indicates that gulls and geese as well as background sources in the sand are the
biggest causes of bacterial contamination. Concentrations may be associated with some onshore
waves and wind, some runoff from wildlife waste, and other small sources such as wildlife. There
are seasonal issues and geese are migratory.

There is a uniqueness to suburban Cook County. There are urban and suburban sources,
particularly related to cross connection issues. Gulls, dogs, and residential impacts are sources of
E. coli.

An additional source, or contributing factor, of E. coli is litter and food waste, which attracts
gulls.
During a tour of several of the Cook County beaches, source-based observations included algae washed
up along the shoreline and in the swash zone and unleashed dogs along the swimming beaches, not just in
the designated dog areas.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-1
3. Additional Considerations for TMDL Assessments
This section examines the precedent for beach bacteria TMDLs to determine if any previously used
methods are appropriate for this situation. Also examined is the possibility of using segment-specific
factors to group the 51 shoreline segments into a smaller number of similarly behaving/reacting groups
that can be analyzed together. The purpose of this section is to facilitate TMDL development at a large
number of beaches in the most efficient manner.
3.1 Previous Beach TMDLs
Within the realm of TMDL development, most are completed on individual stream reaches or for lakes
and ponds. Very few have been completed for beaches on large water bodies and even fewer of those
have been completed for impairments due to bacteria. Tables 3-1 through 3-4 summarize bacterial beach
TMDLs, including the methods used for source detection, quantification, and allocation where available.
Table3-1. Indiana Lakeshore TMDL (Approved) (Tetra Tech, 2004)
Area:
43 miles of shoreline along Lake Michigan within the state of Indiana. The watershed associated with the shoreline
is part of the Little Calumet-Galien USGS hydrologic cataloging unit (HUC 4040001). This watershed covers 536
square miles and encompasses the northern portions of Lake, Porter, and LaPorte counties.
Waterbody:
Lake Michigan
Beach(es):
19 beaches grouped along shoreline
Known Sources:
Pathogen loading from streams entering Lake Michigan will be quantified by the ongoing TMDLs being developed
for the Little Calumet River/Burns Ditch, Salt Creek, and Trail Creek. Other pathogen sources that will need to be
quantified for the lakeshore model include wildlife, waterfowl, and failing septic systems.
Monitoring:
The Interagency E. coli Task Force has 28 water quality stations in the watershed and most of these, 24, are on
the shoreline. The total number of E. coli records available for this study is 8,396, and the data spans the period
May 1984 to July 2002. Additional data were collected in 2003 but were not able to be used in the modeling effort
because corresponding meteorology data were not available.
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-2
Table3-1. Indiana Lakeshore TMDL (Approved) (Tetra Tech, 2004) (continued)
Source Quantification:
Point Sources:
There are several facilities regulated by the NPDES program within the Lake Michigan watershed
that discharge E. coli; however, none of them discharge directly to the lake. The loads from these facilities are
therefore considered within the tributary loads and the significance of the loads from these facilities is being
investigated as part of the tributary TMDLs. Although there are a number of CSO discharges in the Lake Michigan
watershed, none discharge directly to the shoreline waters. The CSO contribution to the Lake Michigan tributary
loads is being investigated as a part of the tributary TMDLs.
MS4:
The loading of E. coli to the Lake Michigan shoreline from most urban storm water sources are included in
the estimates of the tributary loads. Discharge of storm water from these communities directly into the lake is not
considered a significant source of E. coli.
Nonpoint Sources:
Potentially significant nonpoint sources of E. coli to the Lake Michigan shoreline include
tributary loadings, septic systems, wildlife, and other sources such as swimmers and boaters.
Existing streamflow and water quality data were used to make estimates of the load of E. coli from each of the
tributaries to the Lake Michigan shoreline. The loads were calculated by multiplying individual samples by their
corresponding flows and summing over the 1999 beach season. Loading values for days between E. coli sampling
events were extrapolated.
Site-specific information on the location of areas with high septic vulnerability was not currently available for the
Lake Michigan watershed. Therefore, estimates of the loads of E. coli from these sources were based on
assumptions.
The wildlife population (raccoons, white-tail deer, and gulls) and the amount of E. coli that each organism may
contribute are estimated and then compared to loads estimated from other sources.
Initial loading estimates from beach restroom facilities and swimmers were made using information from the
Indiana Dunes National Lakeshore that indicates that approximately 10,750 persons visit the Lakeshore Beaches
on an average day during the summer (June to August). These initial loading estimates were revised during the
model calibration process and a portion of the final load is assumed to include loads from beach sands and algae.
The relative magnitude of each of these sources is not known at this time and should be the topic of additional
research.
Boaters have also been mentioned as a potential source of E. coli even though federal and state laws generally
prohibit the discharge of untreated sewage from any vessel in Lake Michigan or navigable tributary. Because no
information on the number of boaters, their typical locations along the shoreline, or their pumpout practices was
identified during this study, best professional judgment was used to estimate the load at the Washington Harbor,
Hammond, and Robert A. Pastrick Municipal Marinas. It was assumed that 100 boaters used the harbor per day
and that 10 percent of the generated E. coli waste reached the water.
Modeling:
Indiana Department of Environmental Management and its consultant selected the EFDC model as the modeling
framework to be used to support TMDL development for the Lake Michigan shoreline. The EFDC model (Hamrick,
1992) solves the vertically hydrostatic, free-surface, variable-density turbulent-averaged equations of motion and
transport equations for turbulence intensity and length scale, salinity, and temperature in a stretched, vertical
coordinate system, and either a Cartesian or curvilinear-orthogonal horizontal coordinate system. Equations
describing the transport of suspended sediment, toxic contaminants, water quality state variables, and E. coli may
also be solved by EFDC. EFDC was selected for this project because it best matches the required technical,
regulatory, and user criteria discussed in more detail in the modeling framework report (Tetra Tech, 2003).
Development and application of the EFDC hydrodynamic fate and transport model to address the project objectives
involved a number of important steps:
1. Development of computational grid for the waterbody
2. Configuration of key model components
3. Model calibration and validation
4. Model simulation for existing conditions and allocation scenarios
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-3
Table3-1. Indiana Lakeshore TMDL (Approved) (Tetra Tech, 2004) (continued)
Methods:
To develop E. coli TMDLs for the Lake Michigan shoreline, the following approach was taken:

Simulate baseline conditions

Simulate loading allocation scenario #1 in which all tributary E. coli counts above the WQS were reduced to the
WQS (125 cfu/100mL)

Simulate loading allocation scenario #2 in which all tributary loads were reduced as noted in alternative #1 and all
residential septic loads, swimmer loads at public restrooms, and boating activity loads were reduced by 80%.

Simulate loading allocation scenario #3 in which all tributary loads were reduced as per scenario #1 plus an 80%
reduction was made to swimmer/restroom/beach sands loads only at Central Beach and Mt. Baldy Beach (i.e., no
reductions were made to residential septic loads, swimmer loads at other beaches, or boating activity loads).

Determine the TMDL and source allocations.

Components of the TMDLs for E. coli are presented in terms of organism counts per recreation season.
The model simulation results for the first allocation scenario revealed that the dominant source of E. coli was the
tributaries. When the tributary loads were reduced in scenario #1, the TMDL targets for both the 30-day geometric
mean and never-to-exceed standards were achieved at all beach monitoring locations except for Central Beach and
Mt. Baldy Beach. Reducing the septic system loads by 80% in scenario #2 resulted in the achievement of the WQS at
both Central Beach and Mt. Baldy Beach. Reducing only the swimmer, beach sands, and algae loads at Central
Beach and Mt. Baldy Beach in scenario #3 also resulted in the achievement of the WQS at all beach monitoring
locations.
Load Allocations:
Waste Load Allocations (WLAs):
There are no permitted NPDES facilities requiring WLAs in the Lake Michigan
study area. Therefore, no WLAs are presented in this report.
Load Allocations (LAs)
: LAs were made for the following dominant nonpoint source categories:

Tributaries

Residential Septic Systems

Swimmers, Beach Sands, and Algae

Boating Activity

Wildlife
The LAs are based on scenario #3. The LAs are presented on a recreation-season basis (April 1 to October 31) and
were developed to meet TMDL targets under a range of conditions observed throughout the recreation season.
Margin of Safety:
An explicit margin of safety (MOS) will be incorporated into the Lake Michigan Shoreline TMDL by reducing the water
quality target to provide additional assurance. The E. coli target will be set five percent lower than the numeric criteria
in WQS.
The model results for scenario #1 indicate that if the E. coli WQS of 125 cfu/100mL is achieved in all tributaries, then
WQSs will also be achieved at all beach monitoring locations except for Central Beach and Mt. Baldy Beach. To
protect the WQSs at these two beaches, an 80% reduction in swimmer loads was stipulated in scenario #3. There is
a substantial MOS between the WQS and the model results of scenario #3 at all 22 beach monitoring locations. The
MOS relative to the WQS, was calculated as:
MOS = (1 – C / Cwqs) x (100%)
where C = maximum E. coli concentration from scenario #3 model simulation
Cwqs = WQS concentration
The MOS varies in magnitude depending on beach location. For example, the MOS was calculated as 12.7% at
Central Beach, 22.3% at Burns Ditch mouth, 35.6% at Mt. Baldy, and 73.4% at State Park East Beach.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-4
Table3-2. Luna Pier Beach, Lake Erie TMDL (MDEQ, 2007)
Area:
City of Luna Pier Beach in Monroe County, Michigan NHD Reach Code: 04100001001145
Waterbody:
Lake Erie
Beach(es):
Lake Erie Luna Pier Beach
Known Sources:
Pathogen loading from CSOs from the large urban areas
Monitoring:
Lake Erie Luna Pier Beach was sampled for pathogens weekly at five stations from May through September 2005.
Precipitation data was obtained from a nearby station. The 30-day geometric mean WQS was exceeded the
majority of the sampling season at all five stations. Daily maximum concentrations ranged from <20 E. coli /100 mL
in September (Station 3) to 5005 E. coli /100 mL in July.
Source Quantification:
Two types of sources identified: local sources within the immediate watershed and remote sources carried to the
TMDL reach by the currents of the western basin of Lake Erie.
Point Sources:
Eleven local facilities were identified. Eight potential illicit drain connections were identified but not
confirmed. These local sources are minor in significance when compared with permitted sanitary wastewater
discharges and CSOs from the larger urban areas, which line the western basin of Lake Erie, including the
metropolitan areas of Detroit, Michigan; Monroe, Michigan; and Toledo, Ohio. Among the cities, Toledo has 69
CSOs, which discharge to the Ottawa and Maumee Rivers before entering Lake Erie approximately 5 miles south
of Luna Pier Beach (Toledo Waterways Initiative). Prevailing surface currents may then carry contaminants to the
beach. The dominant direction of surface water current is to the northeast
Non Point Sources:
Wild life (primarily Birds)
Modeling:
This TMDL is based on a target concentration represented by a loading capacity (LC). Unlike other pollutants the
critical condition associated with low frequency of flow is not required for pathogen levels. For point source
discharges of treated human sewage, pathogen levels are restricted to a monthly average limit of 200 fecal
coliform /100 mL regardless of stream flow.
The TMDL is equal to the target concentration of 130 E. coli /100 mL as a 30-day geometric mean and daily
geometric mean of 300 E. coli /100 mL in all portions of the TMDL reach for each month of the recreational season
(May through October). Expressing the TMDL as a concentration equal to the WQS ensures that the WQS will be
met under all flow and loading conditions; therefore, a critical condition is not applicable for this TMDL.
Methods:
The WQS of 130 E. coli /100 mL for the point source is attained regardless of flow condition. There is no method
described that relates the concentration of pollutant in the point sources to the receiving water through statistical or
physical models.
Load Allocations:
WLAs:
The WLA for 11 permits is equal to 130 E. coli per 100 ml as a 30-day geometric mean and 300 E. coli per
100 ml as a daily maximum during the recreational season between May 1 and October 31.
LAs:
Because this TMDL is concentration-based, the LA is also equal to 130 E. coli per 100 ml as a 30-day
geometric mean and 300 E. coli per 100 ml as a daily maximum. The TMDL is based on the assumption of equal
contribution regardless of land use and hence depend only on the land area portion.
Margin of Safety:
This TMDL uses an implicit MOS because no rate of decay was used. And using the WQS as the TMDL is
considered to be conservative; hence there is no explicit MOS.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-5
Table3-3. Duck Neck Beach on Chester River, Maryland (MDE, 2009)
Area:
Chester River in the Upper Chester River Basin in Kent and Queen Anne’s counties, Maryland basin number
02130510
Waterbody:
Upper Chester River Basin
Beach(es):
Duck Neck Beach
Known Sources:
The loadings from potential sources (human, livestock, pets, and wildlife) were assessed based on the pollution
source shoreline survey (PSSS).
Monitoring:
In this study, the criterion for public beaches adopted by the Maryland Department of the Environment (MDE) is
that steady-state geometric mean density of enterococci shall not exceed 35 cfu/100 mL.
A single station was monitored for bacteria from 2005-2008 for a total of 141 observations. Additional data were
collected in the 2008 summer season by Queen Anne’s County Health Department. Three months of data (June to
August) were collected both upstream and downstream of the beach and on the tributary draining the area near the
beach in the Upper Chester River.
Source Quantification:
Point Sources:
Two WWTP with 0.140 and 0.075 million gallons per day (MGD) permits were identified: Millington
WWTP and Sudlersville WWTP. The monthly log mean fecal coliform permits for both of them are 200 MPN/100
mL. An equation is used to convert the permit values from fecal coliform to enterococci.
MS4:
none
Nonpoint Sources:
Source assessment was performed by conducting a PSSS of the area surrounding Duck Neck
Beach. A PSSS is a tool historically used by MDE’s Shellfish Program to identify potential sources of bacterial
contamination affecting shellfish harvesting areas. The Shellfish Program PSSS mainly focuses on inspecting
residential shoreline homes for septic systems violations. The PSSS survey concluded that wildlife might be the
major bacteria source for Duck Neck Beach.
Modeling:
The impaired area is highly influenced by tides; therefore, to simulate the transport and fate of bacteria in the
Upper Chester River accurately the 3-D EFDC model (Hamrick, 1992) was used. The entire Chester River is
simulated with a total grid of 217 cells. The model was forced by 6 major tidal constituents. The long-term mean
freshwater inflow was obtained from USGS gage station 01493500.
The inverse method was used to estimate the existing load discharged from each subwatershed based on
geometric mean concentration of bacteria obtained from the observations. Sensitivity analyses were done for
upstream and downstream boundary condition of bacteria loads which revealed that the location of the boundary
points are fairly far away so that any large amount of bacteria would not literally reach the beach because of decay
factor.
Methods:
The TMDL is calculated as count per day and considers a future allocation (FA) factor, where applicable.
TMDL = WLAs + LAs + MOS + FA
The maximum steady-state geometric mean values of enterococci from two-year rolling data is 98.2 MPN/100 mL
for the period from 2006 to 2007. Therefore, the baseline load (current load) from the Duck Neck Beach watershed
is estimated based on the maximum concentration of 98.2 MPN/100 mL.
The allowable load is computed using the MDE standard 35 MPN/100 mL. The 3-D model was used to compute
the allowable load for each subwatershed by reducing the existing loads from the watershed so that the bacteria
concentrations in the receiving water meet the appropriate WQSs.
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-6
Table3-3. Duck Neck Beach on Chester River, Maryland (MDE, 2009) (continued)
Load Allocations:
WLAs:
There are no permitted NPDES facilities in the Lake Michigan study area. Therefore, no WLAs are
presented in this report.
LAs:
LAs were made in general though reduction percentage of the entire area by 50.28% over the contributing
watersheds.
Margin of Safety:
MOS is implicitly accounted in the computation of the 3-D model. This is achieved by using the most conservative
decay factor of 0.7. Literature values of the decay factor range from 0.7 to 3.
Table3-4. Ventura County Beaches, California (California Regional Water Quality
Control Board, 2007)
Area:
Ventura County Beaches, California
Waterbody:
Channel Island Harbor and Ventura Harbor
Beach(es):
3 beaches: Kiddie Beach, Hobie Beach and Harbor Cove Beach
Known Sources:
Bacteria sources in the Harbor Beaches of Ventura County include anthropogenic and non-anthropogenic sources
and point and nonpoint sources.
Monitoring:
Daily and weekly monitoring during summer dry weather, winter dry weather, and wet weather.
Source Quantification:
Point Sources:
There are nine active, NPDES permits or Waste Discharge Requirements (WDRs) for discharges
to Channel Islands Harbor and Ventura Harbor. Of the nine active NPDES permits and WDRs, discharge
associated with the Ventura Port District’s individual NPDES permit is a potentially significant source of bacteria
loading.
Discharges from general NPDES permits, individual NPDES permits, WDRs, the Statewide Industrial Storm Water
General Permit, and the Statewide Construction Activity Storm Water General Permit are not expected to be a
significant source of bacteria.
MS4:
Discharges from the Statewide MS4 Permit for the California Department of Transportation (Caltrans) are a
potentially significant source of bacteria loading.
Nonpoint Sources:
While a source identification study conducted at the Channel Islands Harbor indicated that
local nonpoint sources are the major contributor in summer dry weather, high bacteria densities and exceedances
during wet weather may be more indicative of urban and agricultural run-off.
Sources of bacterial contamination at the Harbor Beaches of Ventura County include:

Marina activities including waste disposal from boats, boat deck and slip washing, swimmer “wash-off”, and
restaurant washouts;

Natural sources including birds, waterfowl, and feral cats; and

Agricultural sources.
Modeling:
There is no prediction type or physical model developed for this study
Methods:
Loading capacity or baseline for the Harbor Beaches of Ventura County are defined in terms of bacteria indicator
densities (i.e., concentrations). As the numeric targets shall be met at the specific sampling locations, which are
representative of the corresponding beaches, no degradation or dilution allowance is provided. Three alternatives
of implementation of Best Management Practices (BMPs) are discussed. The BMPs are structural, non-structural
(control) and do nothing.
(continued)
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-7
Table3-4. Ventura County Beaches, California (California Regional Water Quality
Control Board, 2007) (continued)
Load Allocations:
WLAs: WLAs are expressed as allowable exceedance days. The allowable number of exceedance days for a
monitoring site for each time period is based on the more stringent of two criteria (1) exceedance days in the
designated reference system and (2) exceedance days based on historical bacteriological data at the monitoring
site. This ensures that bacteriological water quality is at least as good as that of a largely undeveloped system and
that there is no degradation of existing water quality.
For each beach, allowable exceedance days are set on an annual basis as well as for three time periods. These
three periods are:
1. Summer dry weather (April 1 to October 31)
2. Winter dry weather (November 1 to March 31)
3. Wet weather days (defined as days of 0.1 inch of rain or more plus three days following the rain event).
WLAs are assigned to MS4 permitees and Cities in the watershed
LAs: LAs are expressed as the number of daily or weekly sample days that may exceed the single sample targets
identified under “Numeric Target” at a monitoring site.
For the Channel Islands Harbor Beaches, the County of Ventura and the City of Oxnard are assigned LAs. LAs
may be assigned to agricultural lands in the Channel Islands Harbor subwatershed during Regional Board
Reconsideration based on monitoring data from the Conditional Waiver for Dischargers from Irrigated Lands.
For Harbor Cove Beach, the Ventura Port District, the County of Ventura, and the City of Ventura are assigned
LAs. LAs may be assigned to agricultural lands during Regional Board Reconsideration based on monitoring data
from the Conditional Waiver for Dischargers from Irrigated Lands.
All LAs for summer dry weather, single sample bacteria densities at the Harbor Beaches of Ventura County are
zero (0) days of allowable exceedances. The LA for winter dry weather and wet weather single sample bacteria
densities for Hobie Beach, Kiddie Beach, and Harbor Cove Beach are listed in the TMDL document.
The LA for the rolling 30-day geometric mean during any time period or monitoring site at the Harbor Beaches of
Ventura County is zero (0) days of allowable exceedances.
Margin of Safety:
An implicit MOS is included through several conservative assumptions, such as the assumption that no dilution
takes place between the on-shore sources and where the effluent initially mixes with the receiving water, and that
bacteria degradation rates are sufficient to affect bacteria densities in the receiving water.
In addition, an explicit MOS has been incorporated, as the load allocations will allow exceedances of the single
sample targets no more than 5% of the time on an annual basis, based on the cumulative allocations for dry and
wet weather.
3.2 Grouping
The 51 individual shoreline segments listed on the Illinois 303(d) impaired waters list require an
innovative approach to enable the development of all 51 TMDLs given the wide range of conditions along
the Illinois Lake Michigan shoreline. Illinois has three general categories of beaches along the shoreline:
Chicago urban beaches, suburban Cook County beaches, and suburban Lake County beaches, in addition
to the few segments that are hardened shoreline without swimming access. As summarized by the
research in Section 2, there are a variety of hydrologic, contributing source, and loading conditions that
can be considered in a regional setting to correlate the water quality between beaches. In the following
sections we examine these conditions and suggest the most efficient grouping method to ensure
productive and quantifiable TMDL assessments.
Whitman and Nevers (2008) undertook a grouping analysis of the 23 Chicago urban beaches they
examined due to what they consider “significant insight for potential contaminant sources,
microbiological community composition, and beach management [that] can be gained by considering a
study beach relative to the longer coastline.” Their analysis relied on multidimensional scaling based on
Euclidean distance to compare E. coli concentrations among beaches. As shown in Figure 3-1, two
groupings were uncovered where geographic location was the dividing factor into beaches north and
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-8
south of the Chicago River and Navy Pier. Of these two groups, the northern beaches were shown to be
more tightly grouped together. Although the southern beaches consisted of a second group, their grouping
is looser except for the relation between Calumet and Rainbow beaches. Overall, the authors found that
the Pearson R for E. coli counts at any two beaches decreased as the geographical distance between the
two beaches increased. The northern and southern beaches followed this pattern when considered together
or separately. Geographic location may be more of a latent variable where the distance could be a
surrogate for factors such as socioeconomics, rate of visitation, wind direction, or currents.
Figure3-1. Multidimensional scaling depiction of E. coli concentrations (log MPN/100 mL)
for 23 Chicago beaches for the years 2000-2005 (Whitman and Nevers, 2008).
This research provides a method that can be used when considering the existing data available for the
beaches along the Illinois shoreline. By defining groups of beaches that respond similarly to bacteria
loadings and/or have similar sources, the amount and time-scale of any additional modeling needs can be
reduced. In the following sections we explore some other considerations that may be used in examining
groupings of beaches.
3.2.1 Geographic and Physical Considerations
The physical makeup and the location of beaches have been shown to play a large role in the bacteria fate
and transport dynamics governing water quality exceedances. Beaches may have similar sources and
loadings, yet the physical makeup of the beach exacerbates the problems (i.e., the occurrence of WQS
exceedances and elevated bacteria levels). These conditions need to be deciphered because, although
source loads can be managed in the same way at other beaches, physical changes to the beach may need
to be made to ensure complete reduction in bacteria concentrations. In Lake County and suburban Cook
County, many of the beaches are situated at the ends of streets and have few obstacles for the longshore
current. Without breaks from the longshore currents, E. coli concentrations may spike and then be
followed by equally rapid decreases as seen with the CAWS reversals at Lighthouse and Northwestern
beaches (Section 2.3.5). “South Chicago beaches have numerous breakwaters, and they are also situated
where the coastline begins to curve more sharply to the east, which may promote bacterial settling and
slower return to baseline levels after an extreme weather or bacteria event” (Whitman and Nevers, 2008).
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
3-9
Considering this information, the following factors should be considered in grouping analysis:

Physical structures (e.g., breakwaters)

Beach orientation

Natural areas surrounding beach

Geographic proximity to other beaches

Proximity to CSOs, ravine outlets, or CAWS control structure

Drainage area/size of beachshed.
3.2.2 Source/Response-Related Grouping Considerations
Physicality of a beach is only one side of an equation. The main sources of bacteria to a beach will also
govern the way the bacteria loads can be measured, managed, and remediated. Main drivers in source
apportionment and identification include

Hydrologic conditions (e.g., major stream inflow, runoff from adjacent area, or point sources)

Land use in the immediate and surrounding areas (e.g., developed urban, suburban)

Local animal populations and access, specifically the presence of gull populations

Location of human facilities in regard to the beach area (e.g., rest room facilities or septic
connections).
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
4-1
4. Conclusions from Initial Data and Source Analysis
This section summarizes the currently available data and study methods for TMDL development. Also
highlighted are the most likely sources of bacteria to the beaches, data gaps in these assessments, and the
feasibility of applying any of the previously completed studies to larger areas or other beaches along the
Illinois shoreline.
4.1 Sources, Fate, and Transport Summary
A variety of studies have been undertaken to elucidate the sources of bacteria to swimming beaches—
from specific DNA-based analysis, to correlation analyses using a range of monitoring data, to physical
modeling of a specific beach or region. The following quotations taken from the studies examined
highlight the main findings to date:

“[D]espite the absence of contributions from sewer overflows, but for extraordinary
circumstances when the locks are opened, nonpoint storm water likely is a significant contributor
of bacteria to Chicago’s beach water” (Whitman and Nevers, 2008).

“High waves often result in higher concentrations of E. coli in the water because they resuspend
bacteria laden beach sand, stranded algae, and fecal material from shorebirds and other wildlife.
This effect is magnified during seiches, when additional swash area may be included in
resuspension” (Whitman and Nevers, 2008).

Although not within the study area, but within the general region, McLellan and Salmore (2003)
found “Locally high E. coli counts coincided with bird presence, and preliminary sampling of
storm water from paved surfaces during this study indicated a potentially large contribution to
elevatedE. coli levels at South Shore Beach; runoff samples (n = 2) collected directly from the
parking lot adjacent to the beach following a 0.5 in rainfall demonstrated E. coli loads of
>100,000 E. coli cfu/100 mL. This is consistent with previously reported bacterial loads in runoff
from impervious surfaces.”

At the embayed 63
rd
St. Beach Ge and others (2010) found “Onshore waves significantly (P <
0.05) increased bed shear stresses, responsible for sediment resuspension nearshore, and total
runup on the beach face. Cases with onshore waves were thus major occasions when the beach
water received bacteria (E. coli) loading from foreshore sand and submerged sediment.”
Additionally, the authors found that “Once in the beach water, bacteria were efficiently entrained
into current flows inside the embayment and released out of the embayment when the external
currents were favorable (i.e., longshore).”

Concerning the 2009 Chicago Ring-Billed Gull Damage Management Project the findings were
thus: “The reduction in the number of gulls using Chicago beaches has likely contributed to a
reduction in conflicts, including a decrease in the proportion of swim advisories/bans on 18 of 19
of Chicago’s beaches (without canine harassment) in comparison to 2006” (Hartmann et al.,
2010).

In Lake County in 2002 Soucie and Pfister (2003) found that “In total, 51% of the E. coli
ribotypes isolated from the beach water samples matched gull E. coli ribotypes, 11% matched
sewage/human, 31% were unidentified, and the remaining 7% matched deer, raccoon, pig, and
cowE. coli ribotypes.” In 2003 the authors found “Similar to the 2002 study, avian,
human/sewage, and unidentified sources accounted for 62%, 20%, and 13% respectively of the
isolatedE. coli. The remaining 5% of E. coli matched dog and rodent sources.”
Additionally, in April of 2010, the CPD Board of Commissioners passed a measure that prohibits the
feeding of birds and other animals at bathing beaches due to the research showing that the bacteria from
bird waste can be a significant cause of swim bans (CPD, 2010).
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
4-2
In summary, given the lack of direct point sources in the region, the above studies have provided a large
step in the direction of determining specific bacteria sources in the region from such factors as gulls. Like
the methods applied by Ge and others (2010) and the grouping characterization performed by Whitman
and Nevers (2008), further work needs to be done to identify and quantify the physical barriers and
transport mechanisms imposed on these nonpoint sources by the lake effects.
4.2 Data Availability for Assessment
As shown in Table 2-2, the county and state health departments as well as CPD have undertaken a large
effort in previous few years to obtain frequent water quality monitoring data of the swimming beaches
along the Illinois shoreline. This basis of data, sometimes accompanied by hydrometeorological data, has
allowed for the development of several beach closure predictive models, but can be greatly expanded to
examine additional considerations in terms of source linkages and uncertainty analysis. The beaches with
replicate samples and multiple samples over the course of the day can also add to more detailed studies of
uncertainty, which can be derived from data already in existence.
However, gaps do exist:

First and foremost, explicit identification of sources at each beach has not been completed but
will likely be aided by the upcoming BSSs.

DNA-based source detection methods have been used at only a small number of beaches. While it
would be costly to extend these methods to all beaches, critical areas should be identified at
which these methods should be applied over a series of wet and dry conditions.

Drainage areas for urban and most suburban beach areas are poorly defined or unknown.

As reported, there are few direct point discharges within the urban Chicago beaches; however,
data on point discharges within suburban Cook and Lake counties needs to be assessed and
confirmed that no major sources of bacteria emit from the mapped discharges.

Lake currents have been shown to play a role in the stagnation and/or flushing of bacteria from
swimming beaches but only in certain locations.

Several of the 51 listed impaired shoreline segments have no monitoring data to date.

The relation of trends and responses of E. coli concentrations have been investigated and
quantified to an extent for 23 urban Chicago beaches but not for any of the more northern
suburban beaches.
It is also useful to review the E. coli concentrations and the ambient conditions of the other beaches along
the Illinois shoreline that are not impaired to gain insight into the processes and conditions that enabled
those beaches to maintain compliance with WQS.
4.3 Critical Areas for Further Investigation
Critical areas for further investigation should be chosen out of several categories: (1) beach(es) where
initial or extension investigation has already been conducted and further study would yield conclusive or
significant results in terms of loading sources and magnitudes; (2) beach(es) that are of “high value” in
terms of public use, health, and economic gains; and (3) beach(es) that show signs of serious impairment
that have not been studied and where responses cannot be correlated with any surrounding beach.
Given these categories, the following beaches are suggested as critical areas for further investigation:
1. Illinois Beach State Park South – This beach has been modeled and monitored for SwimCast,
been part of DNA-based source detection work, and is part of a more natural state park area. It is
also part of the upcoming BSS work.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
4-3
2. Waukegan South Beach – SwimCast modeling and monitoring at this beaches has been
conducted for a number of years and yet beach closures remain high in number. This beach is also
part of the upcoming BSS work.
3. Rosewood Beach – One of the few beaches along the Illinois shoreline that is impacted by
another surface water source. Sanitary Surveys in 2007 revealed potential bacteria sources in a
ravine contributing to the beach. Additional study of the ravine in 2009 further identified
potential point sources within the ravine.
4. Glencoe Beach – This is a long straight beach with its own park district and many beach
amenities. This beach will be a part of the upcoming BSSs. It is also near to Rosewood Beach so
comparisons between the two could aid in grouping scenarios.
5. Montrose Beach – Lake Michigan can have largely different effects on north-facing beaches
such as this one as opposed to those that are north-south in orientation. Montrose Beach also has
native dune and beach habitat and a nearby dog beach. This beach will also be studied with an
upcoming BSS and will have SwimCast monitoring equipment installed.
6. North Ave. Beach – A larger, more extensively used beach that contains a series of breakwaters
along a long stretch of beach.The length of this beach requires more intensive monitoring due to
likely complex shoreline dynamics.
7. 63
rd
St. Beach – Extensive investigation at this beach has already highlighted some sources and
physical barriers to the flushing mechanisms typical of Lake Michigan’s longshore currents. This
beach also has newly (within last few years) established dune and beach habitat and will again be
studied with an upcoming BSS.
Given the timing of the upcoming BSSs, this list is only intended as a preliminary selection of critical
area beaches as the data that is produced by the BSSs may suggest another beach should given priority
over the ones on this list. Additionally, the hardened shoreline segments that do not have swimming
access constitute the lowest priority segments for further investigation: Fullerton St., Schiller St., 49
th
St.,
and 67
th
St. segments. As a final note, 31
st
St. Beach is currently undergoing a renovation along the
shoreline.
4.4 Existing Methods Used for Bacteria TMDLs Applied to Illinois
The TMDLs summarized in Section 3.1 are a selection of TMDLs that were found to be the most similar
in conditions (e.g., beaches on large water bodies) to the setting in Illinois. Only the TMDL for Indiana
covered a number of beaches along a significant stretch of shoreline. Two TMDLs were conducted on
Great Lakes and two were conducted on larger coastal water bodies. All four use different methods for
arriving at allocations and also differ on whether the allocations are load- or concentration-based. While
the existing TMDLs do not provide an overall method to follow for developing the Illinois beach TMDLs,
there are certain portions of the methodology that could aid in this situation. For example, in Indiana,
simple loading factors were combined with a detailed hydraulic model. At Luna Pier on Lake Erie, the
load allocations are made on a concentration basis. And in Ventura County, different allocations were
made based on wet and dry weather. Using the information summarized in these TMDLs guided the
selection of the more standard TMDL models outlined in Section 5. However, these approaches also leave
much to be desired as to the inclusion of uncertainty, the impacts of lake currents, and the impacts across
a range of urban and suburban beaches. Therefore, the summarized methods must also be greatly
expanded to meet the needs of more adaptive and Illinois-specific TMDLs.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-1
5. Deriving the TMDL
A TMDL provides the framework for restoring impaired waters by establishing the maximum amount of
a pollutant load that a water body can receive without adverse impact to fish, wildlife, recreation, or other
uses. This load is divided between point (waste load) and nonpoint (load) pollutant loads with an
additional margin of safety. These three components must be quantified.
TMDL = WLA + LA + MOS Eq. 1
where
WLA = waste load allocation (i.e., loadings from point sources);
LA = load allocation (i.e., loadings from nonpoint sources including natural background); and
MOS = margin of safety.
Section 3 of this report summarized some of the general methods that have been used to quantify these
components in previous TMDL work (Section 3.1), and Section 6 provides options on how both these
general methods and some innovative methods can be used for these 51 specific impaired shoreline
segments within Illinois. As an introductory summary, Table 5-1 presents the sources identified to date
and indicates how they can be incorporated into a TMDL. The following sections will explore these
sources and methods further.
Table5-1. Methods for Incorporating Identified Sources into TMDL Development
Sources
Likelihood
Monitoring
Measure
Utilization
Gulls/geese High Animal count Unit per area per day Loading
Sand High Mass sampling Unit per area Loading
Permitted
discharges
Low Discharge reports Unit per volume Concentration
Topography Site-specific Site maps Contributing factor
Contributing
factor
CSOs Low Discharge reports Unit per volume Concentration
Storm water High Targeted sampling Unit per volume Concentration
Wildlife High Animal count Unit per area per day Loading
Septic systems Site-specific Site maps Unit per area per day Loading
Dogs Site-specific Animal count Unit per area per day Loading
Bather load High Visitor count Unit per area per day Loading
Resuspension High Estimation/targeted sampling Unit per volume Concentration
Regional transport Medium Estimation/targeted sampling Unit per volume Concentration
Loading = direct use in a load-based TMDL; dependent on conditions (e.g., runoff, number of swimmers).
Concentration = direct use in a concentration-based TMDL; measured within the source volume.
Contributing factor = use as a multiplier for source loadings.
5.1 Point Sources Quantification
Point source loads for TMDL calculations are often derived directly from permitted discharges. When
calculating a TMDL for a river or lake, the point discharges are direct loads to the water body. For the
impaired beach segments of this study, any point source discharges would not be direct to the beach but
would first enter a contributing tributary to Lake Michigan, a shoreline ravine, or Lake Michigan coastal
waters. From there the load contributing to the beach would depend on the dominant or average lake
currents.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-2
Lake circulation conditions, combined with a further assessment of the point discharges (listed in Table 2-
4), will decide whether the point discharges to Lake Michigan can be eliminated as a substantial source of
bacteria to the beaches due to the diluting effect of the lake or whether a direct point source load should
be applied to certain local beaches with a dilution factor. If any point source is shown to have a likely
significant effect on a beach, a grid-based hydraulic model may also be used to reveal the influence in
terms of magnitude and frequency of the point source.
5.2 Source and Loading Quantification for TMDL Development
For TMDL development along a shoreline, one must consider not only the source and its loading to the
beach but also the beach’s response to the loading to determine the optimal way to meet WQSs and
protect beach goers’ health. For a beach, source and loading investigations can be divided into two
processes: (1) the watershed component collects water from the contributing subwatersheds through
streams/tributaries as well as direct runoff from adjacent subwatersheds that do not have defined streams
and (2) the process(es) within the water body that determines the fate of the bacteria contributed from
each source along the impaired beach (i.e., within the swimming waters). From any modeling situation,
one must also consider how the output will be used to allocate the source loads.
The Texas Commission on Environmental Quality and Texas State Soil and Water Conservation Board
funded a study (Jones et al., 2007) to investigate a wide range of methods for conducting TMDLs for
bacteria within the State of Texas. While conditions within that region of the country differ significantly
from Illinois, the basic modeling summary (Table 5-2) provides an excellent starting point for the
assessment of both the source and loading portions of the Illinois TMDL assessment in addition to the
previous beach TMDL review. The information shown in Table 5-2 also helps rule out several typical
TMDL models from the current assessment. For instance, load duration curves (LDCs) are useful for
rivers and streams where specific loadings can be derived based on flow through the impaired segment.
For beaches there is not a unidirectional flow for which simple loadings can be calculated. Additionally,
the spatial explicit models are also not likely candidates for TMDL development for Illinois beaches
because so many of the beaches have little to no watershed drainage outside of the beach area.
Building on the information provided in Table 5-2, specific models that are applicable to the conditions in
Illinois are listed in Table 5-3, with indication of how they could be used for developing TMDLs. The
models listed are explored in detail in the following loading and response model sections.
5.2.1 Loading Models
Three different categories of loadings models are considered to be applicable for the beach setting and the
sources identified to date: (1) simple loading factor analysis, (2) watershed modeling, and (3) statistical
modeling. Descriptions and examples of models for each category are provided.
5.2.1.1 Simple Loading Factor Analysis
Loadings calculated from models considered within this category typically consist of a bacteria
generation rate, a unit of application, and any decay or limiting factors. The amount of detail or process
considered within the models can range from a simple one-unit analysis to a multiple-unit, multiple-land-
area analysis. However, with simple loading factors, transport mechanisms and geographic conditions and
variations of the sources are often not considered.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-3
Table5-2. Bacteria Modeling Matrix from Texas TMDL Task Force Report (Jones et al., 2007)
Model
LDC
Spatial Explicit Statistical Models
Mass Balance Models
Mechanistic/Hydrologic/WQ
ArcHydro
SPARROW
SELECT
BLEST
BSLC
BIT
HSPF
SWAT
SWMM
WASP
Watercourse Watersheds x x x x x x x x x
River/stream x x x x x x x x x x
Lake/reservoir x x x x x x x
Fresh/saltwater estuarine x x x x x x x
TMDL phase Development x x x x x x x x x x
Implementation x x x x x
Model type Analytical x x x x x x x
Numerical x x x x
Spatial dimension 1-D
x x
x x x x
2-D
x
3-D
x
Time scale Steady-state
x
x x
Time varying
x x x x
Single storm event
x
x x x
Continuous in time
x
x x x x
Watershed characteristics Rural x x x x x x x x x
Urban x x x x x x x x x x
Sediment transport
x x
x x x
In-Stream processes Bacteria regrowth

Bacteria die-off
x
x x
Settling
x x
Resuspension
x
x x
WLA sources WWTF
x x x
x x x
Storm sewers
x x x
x x x
LA sources Septic tanks
x x x x x x x
Direct deposition
x x x x x x
Bed sediment
x
x x
Cost $ $$ $$ $$$
Shaded cells represent cells that are not applicable.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-4
Table5-3. Model Computations for TMDL Development
Model
Allocate Load
Volume for
Loading Capacity
Source-Receptor
Relationship
Level of Effort
Direct Loadings
Loading factors X Low
BIT X Low
Watershed Modeling
HSPF X X X High
SWMM X X X High
Hydraulic Modeling
EFDC X X X High
POM X High
Water Quality Modeling
WASP X High
Statistical Modeling
Regression X X Medium
Bayesian methods X X Medium
Direct Loadings:
Both domestic and wild animals can be significant potential sources of E. coli in the
beaches depending on the population of the animal in the area and the percent of the time they spend
adjacent to the beach. It should be noted that the E. coli from wild animals is associated with the amount
of runoff generated by precipitation. Average load count/day can be calculated by the number of animals
times the E. coli generation rate. These load counts should be weighted by the percent of the time the
animal spends adjacent to the lake. From the computed average daily load, the amount that reaches the
water body greatly depends on the magnitude and rate of precipitation and also other land use and
landscape variables that facilitate decaying of the E. coli before it reaches the water body.
The number of animals in the area can be estimated by the expected number of animals per acre times the
adjacent area. Animal numbers can be estimated through visual observation such as gull counting along
beaches.
BIT
: The BIT is a spreadsheet tool that estimates the bacteria contribution from multiple sources
(
http://water.epa.gov/scitech/datait/models/basins/bs3tbit.cfm
). The tool estimates the monthly
accumulation rate of fecal coliform bacteria on four land uses (cropland, forest, built-up, and pastureland),
as well as the asymptotic limit for the accumulation should no washoff occur. The tool also estimates the
direct input of fecal coliform bacteria to streams from grazing agricultural animals and failing septic
systems. The input data required for this tool are

Land use distribution for each subwatershed (built-up, forest, cropland, and pastureland,
including, to the extent possible, the breakout of built-up land into commercial and services,
mixed urban or built-up, residential, and transportation/communications/utilities)

Agricultural animals in each subwatershed

Wildlife densities for forest, cropland, and pastureland in the study area (built-up land is assumed
not to have wildlife)

Number of septic systems in the study area

Number of people served by septic systems in the study area
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-5

Failure rate of septic systems in the study area.
For this analysis, a conversion factor from fecal coliform to E. coli concentration would be required.
5.2.1.2 Watershed Modeling
Watershed models can range from simple basic relationships like combinations of export coefficients,
time of travel, and first order decay to detailed process-based models with time-steps at less than a day.
They can represent only one type of land use to all ranges of land uses. Some watershed models also have
the ability to incorporate simplistic simulations of water bodies. Given the setting of the impaired
beaches, the two models deemed most applicable to the study area include HSPF and SWMM.
HSPF
: HSPF has been in extensive use since the 1970s and is distributed by U.S. EPA’s Center for
Exposure Assessment Modeling (CEAM). This watershed hydrologic model has been commonly used for
TMDL development for a variety of conventional water quality parameters in Texas (for indicator
bacteria and for dissolved oxygen) and also in other states (Jones et al., 2007). The required data include
land use, watershed and subwatershed boundaries, location and data for rainfall gages and surface water
quality monitoring stations, detailed descriptions of stream geometry and capacity, detailed information
about sources within the watershed, sedimentation and resuspension characteristics, and bacteria die-off
rates, to name a few.
Advantages:

This model offers continuous simulation of watershed response.

This model allows for use of different land uses that potentially contribute to contamination (note
that there is minimal variation in land uses over the expanse of the Illinois Lake Michigan
shoreline)

Scenarios of reduction of loads at any location can be simulated.
Disadvantages:

The model requires monitoring of flow and bacteria concentration for a period of time to calibrate
the flow as well as decay rates. In practice one or more subwatersheds are calibrated and
parameters from these calibrated watersheds are used either directly or subjectively reduced or
increased by factors known to influence the parameters. (Note that there are no major tributaries
to the shoreline at which flow should be monitored.)

In general, the model is dominantly forced by precipitation or snow melt, which generates high
flow. In low flow cases, when the point bacteria sources are critical, the model is less useful
unless the model is rigorously calibrated to simulate both low-flow and high-flow regimes.

HSPF does not have the capability to simulate the estuary/lake or reservoir hydrodynamics and
hence needs to be linked to other models.

Uncertainty analysis is not feasible.
Level of Effort: High in time, costs, data, and expertise.
SWMM
:The U.S. EPA Storm Water Management Model (SWMM) is a dynamic rainfall-runoff
simulation model used for single event or long-term (continuous) simulation of runoff quantity and
quality from primarily urban areas. The runoff component of SWMM operates on a collection of
subcatchment areas that receive precipitation and generate runoff and pollutant loads. The routing portion
of SWMM transports this runoff through a system of pipes, channels, storage/treatment devices, pumps,
and regulators. SWMM tracks the quantity and quality of runoff generated within each subcatchment and
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-6
the flow rate, flow depth, and quality of water in each pipe and channel during a simulation period
comprised of multiple time steps.
Advantage:

The model can be used in densely populated areas to account for bacteria inflows from urban
areas.
Disadvantages:

This model requires detailed input of specification of urban infrastructure.

The model needs to be linked to other water quality model such as WASP to simulate water body
water quality.

At this time it does not appear that point sources of storm water are large contributors, and so the
runoff portions of this must be relied upon more – with little opportunity for calibration.

Uncertainty analysis is not feasible.
Level of Effort: Medium in time, costs, data, and expertise.
5.2.1.3 Statistical Models: Regression and Multilevel (Empirical Bayes) Methods
Statistical methods provide a way to characterize the relationships between sources of bacteria and water
body response. Using existing knowledge of these relationships and available data, it is possible to
develop multiple regression (MR) models as well as multilevel or Bayesian models for this purpose. To
date, few models have actually used bacterial sources as predictor variables, but given the right data these
models could be explored and developed.
Multiple regression models use a set of predictor variables (the independent model variables) to predict
the value of another variable (the dependent model variable). In this case the predictor variables could
consist of the number of gulls on a beach each day and the water temperature, and the dependent variable
would be the E. coli concentration measured at the swimming beach. Many examples of regression
models at beaches in the area have been cited throughout this report (e.g., the SwimCast models, work by
Whitman and Nevers, 2008) and are available in the peer-reviewed literature (Frick et al., 2008). Often
MR is used and the independent variables characterize water body and not source conditions. Although
this has been typical of MR applications to date, this precedent does not eliminate the possibility of
including more source-based variables in an explanatory model.
MR is particularly attractive because it allows one to determine the contribution of variance by each
predictor in the model, which enables easy testing and modification as new data and/or predictors become
available in the future. The MR equation and its predictor variables can be evaluated using several
criteria, such as number of observations used, standard error of the model, coefficient of determination
(R
2), Mallow’s Cp, number of model parameters, and behavior of model residuals.
Multilevel (or empirical Bayes) model structure is surprisingly common in environmental and water
quality studies, yet it is rarely exploited in statistical modeling in these fields. A multilevel model is likely
to be preferred over a standard MR model in cross-sectional analyses when there are categories or groups
(e.g., individual samples at each beach, beaches affected by common factors, beaches at different lakes).
This advantage arises from retaining the group or categorical membership in construction of the model.
This situation should be addressed using multilevel modeling for two important reasons – one practical
and the other theoretical. From a practical standpoint, Reckhow (1996) and others have shown that by
‘‘borrowing strength’’ across groups, a multilevel model should outperform a standard regression model,
particularly when sample sizes vary substantially among groups. From a theoretical standpoint, group
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-7
data typically have correlated errors for the group variables, and this results in a violation of the least
squares assumptions (resulting in a misleading reduction in parameter standard errors). Thus, least
squares regression is not actually appropriate for application to a multilevel dataset that includes both
individual and group-level data.
Advantages:

Statistical methods provide measures of uncertainty in calculated model parameters.

These methods explicitly reveal the correlations and predictive power of independent variables to
the chosen dependent variable.

Existing statistical methods in terms of regression models have been the basis for several Great
Lakes beach closure models for the past several years.

The Virtual Beach Software has been built explicitly to aid in regression model development for
beach water quality predictions (Wolfe et al., 2010). Other standard statistical software can be
used for both regression and/or Bayesian methods with no cost to the user.
Disadvantages:

A large amount of monitoring data must be available for all model variables/parameters in order
to formulate statistically significant models.

Fate and transport processes are not explicitly represented but may be indirectly represented by
interaction terms or surrogate variables.
Level of Effort: Varies depending on the amount of existing data and the intensity of chosen statistical
model; High in expertise.
5.2.2 Response Models
Response models are used to show how the fate and transport of bacteria at a beach varies once the
bacteria is in the swimming water at the beach. Responses may vary from a complete and fast flushing of
a source from the beach waters to the settling into the sediments of the shoreline with frequent
resuspension events. Three different types of response models/methods are potentially useful for this
analysis: (1) hydraulic models, (2) water quality models, and (3) statistical methods. Descriptions and
examples of models for each category are provided.
5.2.2.1 Hydraulic Models
Hydraulic models represent the movement of water across a defined area. The area may be defined in up
to three dimensions and is usually represented as a series of grid cells. While hydraulic models may
provide simulation of the transport of some water quality parameters, the simulations are usually of less
detail than of the hydraulic components.
EFDC
: The EFDC model (Hamrick, 1992) solves the vertically hydrostatic, free-surface, variable-density
turbulent-averaged equations of motion and transport equations for turbulence intensity and length scale,
salinity, and temperature. A stretched, vertical coordinate system and either a Cartesian or curvilinear-
orthogonal horizontal coordinate system are used. Equations describing the transport of suspended
sediment, toxic contaminants, water quality state variables, and E. coli may also be solved by EFDC.
Input data to drive the EFDC model include open boundary water surface elevations, wind speed and
direction, atmospheric thermodynamic conditions, open boundary salinity and temperature, volumetric
inflows, inflowing concentrations of E. coli, and E. coli decay rate. Model outputs include water surface
elevation, horizontal velocities, salinity, temperature, and E. coli concentration.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-8
An extensive application of the model is reported in developing a bacteria TMDL for the Indiana portion
of the shoreline of Lake Michigan. EFDC was selected for this project because it best matched the
required technical, regulatory, and user criteria in the modeling framework report (Tetra Tech, 2003).
Advantages:

The principal advantage of such a detailed model is in regard to developing a collective or group
TMDL to a number of beaches along a shoreline.

The temporal and spatial concentration of bacteria can be simulated.

Physical processes such as dilution of bacteria in the large water body is implicitly included.

The use of EFDC for TMDL formulation has been approved in other applications.
Disadvantages:

The model requires an intensive data set that may not readily available.

The model needs to be supplied with the inputs of flow and water quality data either from another
model such as HSPF or from simple calculation of daily rates.

As in the HSPF model, this model also needs calibration of parameters.

Scenarios of reduction of pollutant loads to the water body can be studied but only through the
user varying input parameters after external and independent examination of baseline model
results.

Sensitivity of parameters and uncertain inputs such as concentration at boundary need to be
studied for accuracy of the model.

Uncertainty analysis is not feasible.
Level of Effort: High in time, costs, data, and expertise.
POM
:The POM is a full three-dimensional finite-difference program that uses a second-moment
turbulence closure scheme to solve the heat, mass, and momentum conservation equations of fluid
dynamics. It can simulate a wide range of problems: circulation and mixing processes in rivers, estuaries,
shelf and slope, lakes, semi-enclosed seas, and open and global ocean. The numerical POM most often
applied in the Great Lakes region was originally developed for coastal ocean applications by Blumberg
and Mellor (1987) and subsequently adapted for Great Lakes use at the National Oceanic and
Atmospheric Administration (NOAA) Great Lakes Environmental Research Laboratory (GLERL)
(Schwab and Bedford, 1994; O’Connor and Schwab, 1994). This model relies on time-varying wind data,
heat fluxes at the surface and bottom, and bottom friction and roughness conditions to compute currents,
water levels, and thermal structure. It is applied on a gridded domain.
Ge and others (2010) applied simulated current and wave parameters from the GLERL POM in their
examination of bacterial loading at the embayed 63
rd
St. Beach. Using the modeled parameters, the
authors were able to determine that during days where longshore currents, as represented by the POM,
impacted the beach the circular currents created within the embayment entrain bacteria from the beach
swash zone into the center of the embayment before eventually releasing it out into the open lake waters.
The authors also applied the current velocity parameters to their estimations of bed shear stress, which
determines sediment (and therefore bacteria) resuspension.
Advantages:

The model can simulate the regional and local hydrologic conditions of Lake Michigan that can
impact water quality conditions along the Illinois shoreline.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-9

NOAA has developed a Great Lakes-specific POM application that has been tested and applied
under a variety of conditions.

A beach-specific study developed for 63
rd
St. Beach utilized output of GLERL’s POM application
for Lake Michigan to estimate current circulation and bed shear stress and the impacts of these
processes on bacteria concentrations.
Disadvantages:

Use of the POM in any TMDL application for the Illinois shoreline would likely have to rely on
simulations that have already been conducted. Therefore, the conditions simulated may not match
the observed conditions during the time period in which there is water quality sampling for each
beach.

Intensive input parameters including meteorological, bathymetric, and other lake conditional
observations are required for new applications.

Uncertainty analysis is not feasible.
Level of Effort: Working with the POM requires a high level of effort due to model set up and operation.
However, experts at NOAA and USGS could be called upon to facilitate the application of this model.
5.2.2.2 Water Quality Models
Water quality models can represent a number of parameters and the interactions between them.
Depending on the model, different water bodies may be represented by the model (e.g., rivers or lakes)
although the hydraulics simulated in the model are often of less detail than the interactions and fate of the
water quality parameters.
WASP
: The Water Quality Analysis Simulation Program (WASP) model is also distributed by the U.S.
EPA’s CEAM and is a well-established water quality model incorporating transport and reaction kinetics.
Unlike HSPF, however, WASP is driven by flow velocity rather than rainfall, thus it is usually coupled
with a suitable hydrodynamic model such as EFDC or the Soil and Water Assessment Tool (SWAT).
WASP is typically used for main channels, reservoirs, and bays and estuaries and not for modeling
watershed-scale processes. Problems studied using WASP include biochemical oxygen demand and
sources of bacteria, dissolved oxygen dynamics, nutrients and eutrophication, and organic chemical and
heavy metal contamination. Unfortunately, thorough uncertainty analysis is not feasible with WASP.
Advantages:

WASP can be applied in the estuaries and reservoirs where 3-D modeling of water quality is
necessary.

Additional water quality parameters such as dissolved oxygen can be modeled at the same time.
Disadvantages:

The model requires inputs from the watershed computed using either a simple method or
watershed models such as HSPF.

For Lake Michigan, the hydrodynamic component should be modeled by other 3-D models.
Developing these models and coupling the hydrodynamic model would be a challenge.
Level of Effort: High in time, costs, data, and expertise.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
5-10
5.2.2.3 Statistical Methods: Regression and Bayesian (Multilevel) Methods
Regression and Bayesian (multilevel) methods were reviewed under the previous loadings model section
because these methods can span both components, depending on the data available to model. One existing
example of source-related statistical model is the work completed by Ge and others (2010). In their model
at 63
rd
St. Beach, they used regression techniques to determine positive net contributions of foreshore
sand and gull-droppings sources as well as the mechanisms of sediment resuspension and swash to E. coli
concentration in the knee-deep water for onshore-wave cases.
Advantages, disadvantages, and level of effort remain the same when statistical methods are used to
examine either loading or source components.
5.3 Uncertainty
Sources of uncertainty are varied and complex within the TMDL development process. Jones and others
(2007) provide the following description of the issue:
Because of the nature of the pollutant, bacteria TMDLs and [Implementation] Plans,
while using best available information and applying accepted methods of determination,
will contain uncertainties. Even if sources of error in field sampling, kinetics modeling
and numerical implementation could be eliminated, there is a core uncertainty associated
with the “noise” in the bacterial determination methodologies themselves, as indicated by
imprecision in replicate measurements... Efforts to reduce this uncertainty and to provide
heightened defensibility of the process are both worthy and necessary goals... However,
the brutal reality is that over the near future uncertainties — that are sometimes quite
large compared to other water-quality parameters — will exist in bacteria TMDL and
[Implementation] Plan development. Hence, the need exists for one other potential area
of research — quantification of uncertainty and the associated communication of risk
resulting from the uncertainty associated with TMDLs and [Implementation] Plans.
(Jones et al., 2007)
A key recommendation of the 2001 NRC assessment of the TMDL program stated “The TMDL program
currently accounts for the uncertainty embedded in the modeling exercise by applying a margin of safety
(MOS); [U.S.] EPA should end the practice of arbitrary selection of the MOS and instead require
uncertainty analysis as the basis for MOS determination” (NRC, 2001).
Uncertainty analysis for the beach TMDLs can be undertaken using some of the models as noted above.
In addition, the Analytica® program ProVAsT, described in Gronewold and Borsuk (2009), is
recommended when statistical methods are used to develop the source and load quantifications. This
program can be used to compute the confidence of compliance for WQS, which can provide the
uncertainty basis for the MOS. Further examination of this intensive uncertainty analysis will be provided
with the recommended options, where appropriate.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-1
6. Recommended Options
Based on the compiled data and findings of studies within the Illinois Lake Michigan region, the four
options are offered for conducting a TMDL analysis. The four options take advantage of existing and
future collected data. They approach TMDL development through combinations of and stand-alone
statistical, hydraulic, and process modeling methods. Each option is outlined through responses to a series
of guiding questions.
Each of the options will answer the following questions:
1. Given the data that exist to date, what are the inputs to the model? How would the existing data need to be
manipulated to be input to the model?
a. How are point sources considered?
b. How are nonpoint sources considered?
2. What is the basic methodology of the model? For a statistical model, what relations are represented and
how? For a process model, what processes are represented and how? Can they be altered?
3. What improvements can be made to the model by additional data collection? What should this data
collection effort consist of?
4. What is the specific data output from the model? Provide an example.
5. How is the output used to derive a TMDL? Is the TMDL concentration- or load-based? If the TMDL is
concentration-based, how are sources allocated?
6. How is uncertainty addressed in the TMDL?
7. How can the TMDL be used to drive implementation?
Note that it is recommended that as the first step of the selected option trend and correlation analysis
should be completed on the existing monitoring data to confirm the impairment of each of the listed
segments. While the process of delisting an impaired segment can constitute a large amount of effort,
depending on the option chosen removing even a small number of segments from the list of 51 could
greatly speed and reduce the effort needed to derive segment-specific TMDLs.
Additionally, the predictive studies that have been established in the lake region show a significant
correlation between precipitation and the E. coli concentration at beaches. This correlation signifies the
importance of adjacent watershed contribution of E. coli originating from several sources. Therefore, this
contributing area should be delineated and serve as a key element of any TMDL development option.
These watersheds are small in drainage area and there is no major stream that drains to the beaches.
However, using a high-resolution digital elevation model and a linear length of the beach along the
shoreline as a pour point (line), watersheds or “beachsheds” can be developed. It is noted that such
beachshed was delineated for the Rosewood beach. In some cases, if more than one beach is located in
close proximity to another, a single beachshed may be delineated.
6.1 Option 1: Statistically Based Model with High-Level Uncertainty Analysis
For multiple beach TMDLs, we could use existing data from all of the beaches to develop a simple
universal regression model, or a set of regression models based on grouping analysis, and predict the E.
coli response at all beaches based on proposed load reductions. Alternatively, we could develop separate
models for each beach. Here we propose to use a third option that might be considered “the best of both
worlds” by combining models using an empirical Bayes (or multilevel) analysis (see Reckhow, 1996). In
the language of Bayes theorem, the universal (or grouped, if grouping is justified by data analysis)
regression yields the “prior” model that represents information from all beaches, the beach-specific
regression would yield the site specific model representing knowledge from a particular beach of interest,
and the “posterior” (or final predictive) model would then be used to provide the final predictions based
on both prior knowledge and site-specific analysis.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-2
This option would be performed for selected beaches and, therefore, would benefit from a prior grouping
analysis so that the results could be extrapolated to other similar beaches. The method will provide an
explicit MOS based on the uncertainty analyses that can be undertaken with these models.
6.1.1 Given the data that exists to date, what are the inputs to the model?
As shown in Table 2-2, there is a wide variety of monitoring data for E. coli concentrations. There are
fewer available data on sources, but an initial analysis could be completed at this time. For an MR or
multilevel modeling analysis, all existing source data as well as other relevant forcing functions (e.g., the
hydrological monitoring data identified in Table 2-2) will be considered here as candidate predictor
variables. Data can be used in concentration, loading, or binary (e.g., present or absent) formats; the
statistical model does not differentiate. Therefore, both point source and nonpoint source loadings can be
used wherever data exist and for whatever data exist (i.e., X cfu/ day from a point source or Y number of
gulls counted along the beach). Another option is to use a loading model (either simple or complex) to
provide input to the models.
If grouping is found to be useful for modeling purposes, missing data may be an issue; for example, a
particular grouping may seem appropriate due to beach proximity, but may pose modeling difficulties due
to missing data for some beaches in the group. This situation would be resolved by comparing the group
model fit with/without the missing-data-beaches.
6.1.2 What is the basic methodology of the model?
MR and multilevel model structure will be determined based on a combination of existing knowledge
concerning important predictor variables as well as exploratory data analysis using existing data. Model
parameters will be estimated, and model fit assessed, using statistical routines (e.g., lmer for multilevel
models) in R (
www.r-project.org
).
A basic MR takes the form of:
log(C)i = β0 + β1 x
1 + β2 x
2 + ….+ βk x
k + εi
Eq. 2
where
Ci = E. coli concentration (MPN/100 mL or cfu/100 mL), i = 1, 2, . . .n where n = the number of
observations
β
0
= model intercept
β
k
= coefficient for the k
th
variable, k = 1, 2, . . . p where p = the number of explanatory
variables in the model
x
k
= k
th
explanatory variable, k = 1, 2, . . . p
ε
i = model residual error which represents the deviations of the observed values log(C)i from
their predictions, i = 1, 2, . . . n.
A basic Bayesian model takes the form of

﵌
ﰗ
ﰏ
﯅
﯒ﰏ

ﰗ
ﰏ
﯅
﯒ﰏ
ﯗﰏ

Eq. 3
where
ș= unknown parameter of interest
ʌ(ș) = prior probability density function, summarizing our knowledge of the parameter before
observing the data Y
L(Y|ș) = likelihood function, a function of the unknown parameter ș.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-3
“The likelihood function summarizes the levels of support from the data to various șvalues. The posterior
distribution combines the prior knowledge and the likelihood, representing our knowledge about ș after
observing the data. When important prior information is available and is desired to be included in the
decision process, the Bayesian approach is much more effective than the classical statistics approach”
(Qian and Reckhow, 2007).
6.1.3 What improvements can be made to the model by additional data collection?
Additional data would likely be needed on identified sources to beaches. For example, if statistically
significant correlations are shown between numbers of gulls and E. coli concentrations, then additional
monitoring programs should be undertaken to institute gull-counting programs. Any data that show
correlations with E. coli concentrations within the model would be an added value in terms of the
possibilities to increase model predictive strength and reduce model uncertainty. Other examples of
additional data include estimates of E. coli loadings from surrounding impervious areas, E. coli loads
from stormwater outfalls, and E. coli density in sediment resuspension.
The predictor variables (x
k in Equation 2) that have been studied up until now, which have shown
significant influence, are weather and lake conditions. The MR as described above attempts to include
additional variables that are directly linked to the inflow loads of E. coli as explanatory variables. These
additional variables must be significantly correlated to the monitored E. coli concentration in the water
body in order to ascertain their influence. Rather than weather and lake variables which have been shown
to have some correlation with waterbody E. coli concentrations without providing an indication of source,
the significance of these additional variables can be used to determine a load-based TMDL and its load
allocations.
6.1.4 What is the specific data output from the model?
The first step of this method is to provide explanatory variables to the statistical model. Therefore, there
may be output of a loading model as an intermediate step. The output from both an MR and a multilevel
model would be an estimated E. coli concentration in the swimming waters bounded by a prediction
interval, much like the typical output of the SwimCast models in Lake County. Concentrations are
predicted so that the model predictions can be compared to the WQS. Through model analysis the
magnitude of the explanatory variables corresponding to the threshold E. coli concentration at the WQS
provides information that can be used to formulate a TMDL as described in the next section.
6.1.5 How is the output used to derive a TMDL?
Once the statistical model fit is deemed satisfactory based on goodness-of-fit criteria, it can be used to
predict the allowable load (the TMDL) that achieves compliance with the bacterial WQS by varying the
predictor variables within the model given the sensitivity of the model predictions to each predictor. For
instance, if the current direction interacted with the point load from a WWTP is shown to greatly
influence the predicted E. coli concentrations, then this point load can be varied as input to the model to
determine its load allocation to achieve the WQS. The TMDL can be either load- or concentration-based,
and both LAs and WLAs can potentially be set using the predictor variables.
6.1.6 How is uncertainty addressed in the TMDL?
Uncertainty can be estimated based largely on the model standard error, which is estimated as part of the
model fitting. In addition, the procedure developed by Gronewold and Borsuk (2009) will be used to
provide the confidence of standards attainment. This procedure relies on a hierarchical Bayesian approach
for relating actual in situ or model-predicted pollutant concentrations to multiple sampling and analysis
procedures, each with distinct sources of variability – precisely the conditions found along the Illinois
shoreline where multiple organizations monitor and track data. The distributions created by the statistical
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-4
method are used to assess a confidence of compliance – an alternative MOS that is explicitly calculated
and based on the variability in the monitoring data.
6.1.7 How can the TMDL be used to drive implementation?
The model will be simple enough for expression in a user-friendly manner for scenario assessment by
stakeholders and decision makers. This formulation can facilitate discussion of alternative implementation
options considering the predictor variables within the model that account for the greatest proportions of
variability, i.e., the most significant sources.
6.2 Option 2: Use Mainly Existing Data in Hybrid Statistical Plus Hydraulic
Process Modeling
The second option is composed of a method similar to that conducted by Ge and others (2010) where
hydrodynamic model components are combined with statistical correlations or regressions to support
source hypotheses. This option would require considerably more effort but would provide a more process-
based approach to determining the fate and transport of bacteria within the nearshore waters – processes
that are likely highly important and highly variable given the possible direction and magnitude of changes
in Lake Michigan’s currents. The advantage of this method in combining a statistical approach with
hydrodynamic modeling is the increased ability to include uncertainty measures in the results over a pure
hydrodynamic model or a hydrodynamic model using direct loading calculations.
This option would be performed for selected beaches and, therefore, would benefit from a prior grouping
analysis so that the results could be extrapolated to other similar beaches. This option provides an explicit
MOS linked to given confidence limits of the regression.
6.2.1 Given the data that exists to date, what are the inputs to the model?
There are site-specific data for some beaches, such as 63
rd
St. Beach, where, in addition to a single point
sample of E. coli there are multiple samples across the length of the beach and throughout several
distances from shore. This is the type of detailed monitoring data that would be required for this option.
Because not many beaches have this detailed data, the methodology would have to be applied and refined
at the few beaches that do while additional data collection efforts are undertaken across one or multiple
swimming seasons.
Additionally, existing output from hydrodynamic models such as GLERL’s version of the POM could be
harnessed for wave and velocity parameters along the whole Illinois shoreline. These mined data are
necessary to numerically simulate wave and current parameters for use in estimating bed shear stress and
total runup height. These estimates, in turn, reflect the potential of sediment resuspension and bacteria
input from the foreshore sand, respectively.
6.2.2 What is the basic methodology of the model?
A hybrid hydrodynamic-statistical modeling effort can take many different forms. We envision that a
suitable hybrid method would be one similar to that conducted by Ge and others (2010), where larger
hydrodynamic results are harnessed and focused on a smaller area to create additional wave and current
parameters that can be correlated and/or combined with monitoring data to predict water quality
concentrations. Following the work of Ge and others (2010), there are several steps to this method:
1. Mine existing E. coli and source-related monitoring data and/or collect additional transect
samples over varying time periods. Compile data on beach-specific characteristics such as beach
width, length, and backshore area.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-5
2. Obtain site-appropriate hydrometeorological variables, such as wind speed and direction, air
temperature, wave height, and current speed and direction. Some of these variables can be mined
from existing hydrodynamic models when not available from observations alone.
3. Numerically estimate bed shear stress and total runup height corresponding to all monitored
conditions. (Further detail provided in Ge et al., 2010.)
4. Finalize source-based explanatory variables from monitoring and estimated data for use in
statistical analysis. For example, decompose wind and current vectors into onshore and longshore
components. Interact beach width with source variables such as density of gull droppings.
5. Complete statistical analysis to correlate variables with E. coli concentrations and form final
predictive model using the source-based explanatory variables.
6.2.3 What improvements can be made to the model by additional data collection?
This option would benefit from and likely requires additional targeted source monitoring data along
transects and within the swash zone. Any further information that can be gained on source detection from
the upcoming BSSs can be incorporated into this hybrid model with the potential to improve model
predictions.
With hydrodynamic modeling components, a large improvement would be to run a specific hydrodynamic
simulation for the specific time period and focus area under study rather than relying on larger-scale Lake
Michigan simulations. Collaboration with experts at USGS and NOAA is advised with this option to
obtain the best possible input data and possibly collaborate on future detailed site or regional simulations.
6.2.4 What is the specific data output from the model?
For a TMDL analysis, the goal of the statistical analysis using the derived explanatory variables would be
to find the best possible combination of variables to predict the E. coli concentration in knee-deep water.
(In contrast, Ge and others [2010] formed models to predict E. coli concentrations near and far from the
shore.) Therefore, the final model output is a time-series of predicted E. coli concentrations in the
swimming waters bounded by prediction intervals. Intermediate output would consist of loads derived the
source-specific variables such as bacteria resuspension loads and runoff loads.
6.2.5 How is the output used to derive a TMDL?
Similar to Option 1, the output from this hybrid model is used as an endpoint to reach by varying model
inputs. In this case, the inputs are source-based parameters that are varied according to likely management
strategies and hydrologic conditions. Again, those parameters that show the largest influence over model
variability would be targeted for reductions in forming load allocations. This hybrid approach, which
rightly considers hydrodynamic conditions, may ultimately identify sediment resuspension and wave
runup as major contributors to E. coli concentrations. In that case, load allocations may have to focus on
sources that are not directly measured but are known to contribute bacteria to the water, such as bather
load. The TMDL would be load-based as the inputs to the model are loadings even though E. coli
concentrations are predicted by the model.
6.2.6 How is uncertainty addressed in the TMDL?
Because this option ultimately relies upon regression methods to formulate a predicted E. coli
concentration, uncertainty measures will be based on the prediction intervals around the model estimates
and the standard error of the model. As with some of the beach closure predictive models in use, the
numbers of false positive and false negative results as well as the correct number of predictions in terms
of WQS exceedances can be used to evaluate the model during its calibration period and get a handle on
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-6
the model’s uncertainty in its average estimation capabilities. An explicit estimate of the uncertainty can
be determined with model error terms.
6.2.7 How can the TMDL be used to drive implementation?
Depending on the source-based explanatory variables shown to be significant in predicting swimming
waterE. coli concentrations, various scenarios can be tested within the model to allocate loads among the
sources. If, however, the source-based explanatory variables are more hydrodynamically related (e.g.,
wave runup) than source-related (e.g., gull droppings), more holistic implementation plans may need to be
derived including things like beach grooming. Additionally, some implementation methods have the
potential to be tested in surrogacy within the predictive model. For instance, beach grooming may reduce
theE. coli concentration within the swash zone sediments. Therefore, a scenario can be tested with a
reduced value for this explanatory variable assuming it was a variable in the calibrated prediction model.
6.3 Option 3: Provide for Adaptive TMDL Development to Utilize Upcoming
Sanitary Surveys
In an evaluation of the U.S. EPA TMDL program (NRC, 2001), it was recognized that uncertainty in
TMDL assessments is usually quite large; in effect, the initial TMDL analysis is likely to be in error. The
NRC panel recommended adaptive management as a strategy to address this uncertainty. Adaptive
management, or learning while doing, is feasible here due to the upcoming BSSs. If initial loading
allocations can be set based on data already in existence, the BSSs can be targeted to gather data to
improve the initial TMDL assessment. As more information becomes available, the survey program can
adapt to narrow down the source assessments. This option could present regulatory challenges as it would
require setting TMDL targets initially and then revising them with time. However, adaptive management-
based reassessment of TMDLs is a realistic strategy when considered in the context of growth and change
in land use/cover in a watershed.
This option would be performed for beaches where BSSs are planned and, therefore, would require prior
grouping analyses to inform the TMDLs for the beaches that are not surveyed. BSSs will provide further
monitoring data of sources including gull counts, runoff water quality sampling, and beach/watershed
forcing parameters such as slope and drainage area. The additional data would be used to inform the
posterior models generated using statistical techniques.
6.3.1 Given the data that exist to date, what are the inputs to the model?
This option would proceed immediately utilizing all available data currently within the systems of the
managing agencies as is done with Option 1.
6.3.2 What is the basic methodology of the model?
The initial method would be to follow Option 1 and create an MR or multilevel statistical model that best
predictsE. coli concentrations at the targeted beaches. Over time, as more data come in, the posterior
models are adjusted to fit the new data and refine the water quality predictions following the same
original multilevel or Bayesian methods. An example of this adaptive model framework is provided in
Figure 6-1. In this case, as more monitoring data are collected each year for chlorophyll a the posterior
model is refined to better reflect the conditions defined by the observations, thereby reducing the
uncertainty surrounding the prediction of WQS attainment.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-7
Figure6-1. An example of using adaptive implementation with Bayes updating where the green
line represents the WQS to be met.
6.3.3 What improvements can be made to the model by additional data collection?
Additional data is the key to this option, and activities to collect additional data would benefit the TMDL
and also benefit from the TMDL. That is, as more data are collected, the focus of collection efforts can be
narrowed to more specific targets, time periods, or frequencies. The BSSs will act as a guide but also be
informed by the initial model findings. For instance, it is known that the BSSs will provide more
information on gull populations at the surveyed beaches. Therefore, gull population densities will be
tested in the initial statistical framework. If this source is shown to be related to the E. coli concentrations
in the swimming waters, then the BSSs should continue to target surveying efforts on this source.
6.3.4 What is the specific data output from the model?
Output of this adaptive framework will be similar to what is visualized in Figure 6-1— there are prior and
posterior distributions that relate to the collected monitoring data. How these distributions change over
the course of monitoring periods indicates how the model adapts to new data. The data that go into
forming these distributions also are used to produce time-series E. coli concentrations similar to the
SwimCast model.
6.3.5 How is the output used to derive a TMDL?
An initial TMDL will be set based on the prior statistical model developed using the methods described
for Option 1. Each year new posterior models will be created and the initial load allocations tested. As the
posterior model converges with additional data, the load allocation scenarios based on predictor variables
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1992
0.103
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1993 0.044
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1994 0.028
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1995
0.011
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1996 0.006
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1997 0.016
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1998
0.003
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
1999 0.002
-4-20246
0.00.20.40.60.81.0
Prior
Posterior
2000 0.007
Log Chla Concentration
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-8
will be finalized so that the predicted distribution does not overlap the WQS, creating a situation where
there are no or few WQS exceedances expected with a given amount of certainty.
6.3.6 How is uncertainty addressed in the TMDL?
As with Option 1, this method will rely on the uncertainty analysis procedure developed by Gronewold
and Borsuk (2009). See Section 6.1.6 for further detail.
6.3.7 How can the TMDL be used to drive implementation?
The adaptive nature of this option presents a variety of ways to implement water quality control measures,
if not specific TMDL measures. For example, showing a significant correlation between algal density and
theE. coli concentration in the swimming waters may lead the beach managing organization to institute
beach grooming activities. Over time, as data are collected and the model adapted, the fruitfulness of
these implemented activities can be confirmed or refuted by the inclusion of the predictor variable within
the model. Alternatively, implementation activities can be withheld until a final model is accepted upon
final data collection. At that time the source-based predictors as well as surrogates for sources used in the
model can be expanded into implementation measures that can be tested within the final predictive model.
6.4 Option 4: Shoreline Process-Based Modeling
This approach relies on the accurate application of a process-based model or combination of models for
the shoreline. This complete modeling effort would require both watershed and water quality/hydraulic
representations. The modeling effort would be composed of a selection of models outlined in Section 5. It
will be important to obtain the correct model inputs from onshore as well as in the lake. The existing
monitoring data, and possibly supplemental data, will be used to calibrate the model. Once calibrated, the
model can be used to assess different scenarios of reduction in loads from each of the identified/ assumed
sources. The scenario that satisfies an attainment of the WQS at all the beaches will be adopted and
consequently implemented.
This option would be performed for the full shoreline and, therefore, would not directly benefit from a
prior grouping analysis. The MOS for this option would be implicit.
6.4.1 Given the data that exist to date, what are the inputs to the model?
The process-based model type requires three set of data: bathymetry of the shoreline,
hydrometeorological data, and E. coli loadings.
Bathymetry data: The bathymetry of Lake Michigan can be obtained from the NOAA GeoDAS database
of hydrograph survey (http://map.ngdc.noaa.gov). A portion of the Lake Michigan bathymetry data that
covers the Illinois shoreline (65 miles in length and a certain width into the lake) will be used. The
distance from the shoreline into the main water body (width) will be set to have a boundary condition that
has a minimal effect on computation of E. coli concentrations. The bathymetry data should be converted
to a grid mesh compatible to a specific model. Note that variable grid sizes can be used (smaller grid size
to capture shoreline structures and local loading points and larger grid sizes in the main water bodies to
reduce computation times). The data available from NOAA is provided in a 90-meter grid cell format;
however, this data was resampled and digitized from datasets of varying resolution (as shown on the
following source map: http://www.ngdc.noaa.gov/mgg/image/images/michigan_wallsizebot_300.pdf).
Hydrometeorological data: The meteorological data that are critical to the process models are
precipitation, wind speed and direction, atmospheric pressure, cloud cover, and temperature. The
precipitation data are the most critical in generation of runoff from the beachsheds. Hourly precipitation
data and daily surface data can be accessed from NOAA’s National Climatic Data Center (NCDC). The
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-9
closest complete data set can be obtained from the Chicago O’Hare International Airport station
(COOPID=11547). Additional data from climate and lake monitoring stations must be gathered as inputs
to the model. There are no major identified stream flows in the area (except during releases from the
CAWS). The baseline modeling will likely require data for two or more years.
E. coli Loadings: The E. coli loadings are required at all shoreline boundaries and at the model domain
boundaries. The shoreline boundary loads are from observed point source (if known outflows) and
beachshed loads as a result of runoff. For nonpoint sources, the E. coli loadings are estimated using
watershed models or simple loading techniques. (Using a separate watershed model to estimate loadings
would require additional inputs.) Along the Illinois shoreline, the existence of routine monitoring of
beaches provides an opportunity to estimate loads iteratively as explained in Section 6.4.5. The E. coli
loading at the domain boundary within the lake should also be estimated. Later sensitivity analyses should
be performed, which may require the shifting of the boundary further into the lake to minimize the effect
of uncertainty of this boundary condition.
6.4.2 What is the basic methodology of the model?
The basic principle of the process-based shoreline model is to simulate the hydrodynamic and water
quality processes of the shoreline water body by accounting for the hydrometeorological forcings and the
response of the water body. Hydrodynamic fate and transport modeling studies are based upon four
principles: (1) conservation of momentum, (2) conservation of mass and energy, (3) thermodynamics, and
(4) ecological interactions and processes (Tetra Tech, 2003). While the hydrodynamic component of the
model produces water depth and velocity of the water at each of modeled points (grids), the fate and
transport component uses these values to determine the values of water quality variables, in this case E.
coli concentrations, at those grids.
6.4.3 What improvements can be made to the model by additional data collection?
In order to provide the most realistic and justifiable model, this option requires the most intensive data
inputs. Upon selection of the model(s), the additional data requirements would be determined and would
likely consist of intensive geographic information system (GIS) analysis, additional watershed
characterization, and more intensive source monitoring. Any and all monitoring data for E. coli and
hydrometeorological data would be incorporated where possible, even for beaches and segments in
between the impaired segments.
6.4.4 What is the specific data output from the model?
For a given baseline or future reduced E. coli loading, the shoreline model produces spatially and
temporally continuous E. coli concentration time series at a gridded scale. The model outputs can be
customized to extract outputs for grid cells with a specified time interval as small as the computation time
and also for desired locations, such as a summarization of grid cells defining the limits of each impaired
swimming beach and shoreline segment. Using the time series data, the number of WQS exceedances at
each summarized segment can be determined for each reduction scenario.
6.4.5 How is the output used to derive a TMDL?
Once the model is created for the shoreline, it has to be calibrated using routine monitoring data. The
calibration process includes accurate reproduction of water depths and E. coli concentrations at all
observed points. After calibration, a baseline condition will be simulated. The baseline period will be
selected representing the current average shoreline conditions. The E. coli loadings from each of the
beachsheds and other sources that correspond to the baseline simulation are then considered current loads.
Next, to compute allowable loads, several load-reduction scenarios can be applied (by trial and error)
until the E. coli concentrations at all beaches meet the required WQS exceedance criteria defined by
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
6-10
Illinois. As the model produces continuous time series, both simulated geometric mean as well as
simulated maximum daily E. coli concentrations can be compared to the U.S. EPA/IEPA requirements.
Precaution is necessary to have targeted loading computation mechanisms for each of the beachsheds. We
anticipate using a domain of the model that extends along the entire Illinois shoreline. This domain will
include all the beaches in the area. In the calibration process, it could be possible to identify the influence
of the loading from each of the beachsheds and point sources on the individual beach E. coli
concentrations. It is even possible, with an intensely high level of effort, to iteratively estimate the loading
of each beachshed to match the concentration at each observed location. In some cases, the loading from
one beachshed may have substantial effect on more than one beach. Prior statistical studies will be used to
characterize the extent of impacts of loadings to best focus the source identification studies and model
input estimates. However, the loadings estimated from each beachshed will be the largest source of
uncertainty to a modeling effort of this type and all possible efforts must be made to fully quantify the E.
coli loadings in each one in order to produce the most realistic and reliable results.
6.4.6 How is uncertainty addressed in the TMDL?
The MOS for this option can be included implicitly by using conservative E. coli decay rates.
Alternatively, a sensitivity study of model parameter and boundary conditions can facilitate determination
of a MOS.
6.4.7 How can the TMDL be used to drive implementation?
The model setup for this option requires that E. coli loadings from the landscape (i.e., from the
beachsheds) be quantified and input into the hydrodynamic model for fate and transport processes within
the lake. Therefore, these loadings will be based either on inputs to a watershed model or on simpler
loading factors. The reduction scenarios used to determine the TMDL will vary each of these input source
loadings at their origins—meaning the locations at which they can be managed. For instance one loading
to a beachshed may be the number of gulls depositing fecal matter along the beach. A reduction scenario
may indicate that the number of birds needs to be reduced by 50%. Therefore, this method can provide a
direct linkage to implementation options for the derived TMDL but is highly dependent on how these
loadings will be estimated.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
7-1
7. References
Adams, M., Lake County Health Department, personal communication, October 2010.
Adams, M. and M.A. Pfister. 2007. Rosewood Beach Sanitary Survey Report. Completed for the Illinois
Department of Public Health. Chicago, IL.
Alliance for the Great Lakes. 2009. Stresses and Opportunities in Illinois Lake Michigan Watersheds
Strategic Sub-Watershed Identification Process (SSIP) Report. Prepared for the Lake Michigan
Watershed Ecosystem Partnership. Available at:
http://www.greatlakes.org/Page.aspx?pid=881
.
Beletsky, D., J. H. Saylor, and D.J. Schwab. 1999. Mean Circulation in the Great Lakes. Journal of Great
Lakes Research 25(1): 78–93.
Beletsky, D., and D.J. Schwab. 2001. Modeling circulation and thermal structure in Lake Michigan:
Annual cycle and interannual variability. Journal of Geophysical Research 106(C9):19,745–
19,771.
Blumberg, A.F., and Mellor, G.L. 1987. A description of a three-dimensional coastal ocean model. In
Three dimensional coastal ocean models, coastal and estuarine sciences, N. Heaps, ed., pp. 1–16.
Washington, D.C.: American Geophysical Union.
Breitenbach, C., Chicago Park District, personal communication, November 2010.
California Regional Water Quality Control Board. 2007. Substitute Environmental Documents for the
Harbor Beaches of Ventura County (Kiddie Beach, Hobie Beach, and Harbor Cove Beach)
Bacteria Total Maximum Daily Load. Prepared under the California Environmental Quality Act
(CEQA) Requirements of a Certified Regulatory Program.
CPD (Chicago Park District). 2010. Chicago Park District’s “Welcome to Chicago’s Lakefront” website.
Available at: http://www.chicagoparkdistrict.com/resources/beaches/.
Cali, S., J. De Witt, and A. Abelmann. 2007. Sanitary Survey of 63
rd
Street Beach, Chicago. Completed
by the University of Illinois at Chicago. Available from the Illinois Department of Public Health.
Chicago, IL. December.
Francy, D.S., A.M. Gifford, and R.A. Darner. 2003. Escherichia coli at Ohio bathing beaches—
distribution, sources, wastewater indicators, and predictive modeling. WRIR 02-4285, U.S.
Geological Survey, Columbus, Ohio.
Francy, D.S., E.E. Bertke, and R.A. Darner. 2009. Testing and Refining the Ohio Nowcast at Two Lake
Erie Beaches—2008. U.S. Geologic Survey. Open-File Report 2009–1066, 19 p.
Frick, W.E., Z. Ge, and R.G. Zepp. 2008. Nowcasting and Forecasting Concentrations of Biological
Contaminants at Beaches: A Feasibility and Case Study. Environmental Science & Technology
42(13): 4818–4824.
Ge, Z., M.B. Nevers, D.J. Schwab, and R.L. Whitman. 2010. Coastal Loading and Transport of
Escherichia coli at an Embayed Beach in Lake Michigan. Environmental Science & Technology
44: 6731-6737.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
7-2
Gronewold, A.D., and M.E. Borsuk. 2009. A Software tool for translating deterministic model results into
probabilistic assessments of water quality standard compliance. Environmental Modelling and
Software 24: 1257-1262.
Hamrick, J.M. 1992. A three-dimensional environmental fluid dynamic computer code: theoretical and
computational aspects. SRAMSOE #317, Virginia Institute of Marine Science, The College of
William and Mary, Gloucester Point, VA.
Hartmann, J.W., S.F. Beckerman, T.W. Seamans, R.M. Engeman, and S. Abu-Absi. 2010. Report to the
City of Chicago on Conflicts with Ring-billed Gulls and the 2009 Integrated Ring-Billed Gull
Damage Management Project. Prepared for Chicago Department of Environment. Available at:
http://www.chicagoparkdistrict.com/docs/8e4762a1-a0fe-401c-8d6e-
a68214d2dbb7_document.pdf.
Jones, C.A., K. Wagner, G. Di Giovanni, L. Hauck, J. Mott, H. Rifai, R. Srinivasan, and G. Ward. 2007.
Bacteria Total Maximum Daily Load Task Force Final Report. Prepared for Texas Commission
on Environmental Quality and Texas State Soil and Water Conservation Board. Report TR-341.
Kelsey, H., D.E. Porter, G. Scott, M. Neet, and D. White. 2004. Using geographic information systems
and regression analysis to evaluate relationships between land use and fecal coliform bacterial
pollution. Journal of Experimental Marine Biology and Ecology 298(2):197-209.
Lanyon, D. 2010. Separation of the Great Lakes and Mississippi River Watersheds. Metropolitan Water
Reclamation District of Greater Chicago. Presented October 28
th
, Chicago, IL. Available at:
http://www.mwrd.org/pv_obj_cache/pv_obj_id_2CF1FDE8165D904E274616DA8DEEF868B99
B6000/filename/101028_watershedseparation.pdf.
LCHD (Lake County Health Department). 2009. SwimCast Data. Online data. Lake County Health
Department, Community Health Center, Waukegan, IL. Available at
http://www.lakecountyil.gov/Health/want/SwimCast.htm.
Lin, B., M. Syed, and R.A. Falconer. 2008. Predicting faecal indicator levels in estuarine receiving waters
- An integrated hydrodynamic and ANN modelling approach. Environmental Modelling and
Software 23(6):729-740.
MDE (Maryland Department of Environment). 2009. Total Maximum Daily Loads of Bacteria for Duck
Neck Beach of Chester River in the Upper Chester River Basin in Kent and Queen Anne’s
Counties, Maryland. Submitted to U.S. Environmental Protection Agency, Region 3. Available
at:
http://www.mde.state.md.us/programs/Water/TMDL/DraftTMDLforPublicComment/Documents/
www.mde.state.md.us/assets/document/TMDL_Duck_Neck_07_22_2009_PN.pdf
McLellan, S.L. and A.K. Salmore. 2003. Evidence for localized bacterial loading as the cause of chronic
beach closings in a freshwater marina. Water Research 37: 2700–2708
MDEQ (Michigan Department of Environmental Quality). 2007. Total Maximum Daily Load for E. coli
for Lake Erie Luna Pier Beach, Monroe County. Available at
http://www.michigan.gov/documents/deq/wb-swas-tmdl-lunapier_214458_7.pdf
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
7-3
MWRDGC (Metropolitan Water Reclamation District of Greater Chicago). 2010. Microbiological Report
of Bypass Samples in 2009. Report No. 10-17. Monitoring and Research Department. Chicago,
IL.
MWRDGC. 2008. Description of the Chicago Waterway System for the Use Attainability Analysis.
Report No. 08-15R. Monitoring and Research Department. Chicago, IL. Available at:
http://www.mwrd.org/pv_obj_cache/pv_obj_id_ECEFE06C27756E8C4A4D233A95E69F52BD2
42200/filename/08-15%20Description%20of%20CWS%20Report%20for%20UAA.pdf.
National Research Council (NRC). 2001. Assessing the TMDL Approach to Water Quality Management.
Washington: National Academy Press.
Nevers, M.B. and R.L. Whitman. 2005. Nowcast modeling of Escherichia coli concentrations at multiple
urban beaches of southern Lake Michigan. Water Research 39: 5250–5260.
NRDC (National Resources Defense Council). 2010. Testing the Waters: A Guide to Water Quality at
Vacation Beaches. Twentieth Annual Report. Available at:
http://www.nrdc.org/water/oceans/ttw/ttw2010.pdf.
O’Connor, W.P. and D.J. Schwab. 1994. Sensitivity of Great Lakes Forecasting System nowcasts to
meteorological fields and model parameters. Paper presented at 3rd International Conference on
Estuarine and Coastal Modeling, American Society of Civil Engineers, Waterways, Port, Coastal,
and Ocean Division, Oak Brook, IL.
Olyphant, G.A. 2005. Statistical basis for predicting the need for bacterially induced beach closures:
Emergence of a paradigm? Water Research 39(20): 4953-4960.
Qian, S.S. and K.H. Reckhow. 2007. Combining Model Results and Monitoring Data for Water Quality
Assessment. Environmental Science & Technology41: 5008-5013.
Rao, Y.R. and D.J. Schwab. 2007. Transport and Mixing Between the Coastal and Offshore Waters in the
Great Lakes: a Review. Journal of Great Lakes Research 33: 202-218.
Reckhow, K.H. 1996. Improved Estimation of Ecological Effects Using an Empirical Bayes Method.
Water Resources Bulletin32: 929-935.
Schwab, D.J. and K.W. Bedford. 1994. Initial implementation of the Great Lakes forecasting system: A
real-time system for predicting lake circulation and thermal structure. Water Pollution Research
Journal of Canada 29: 203–220.
Soucie, W.J. and M.A. Pfitser. 2003. E. coli Source Identification on Lake Michigan Beaches in Lake
County, Illinois. Presented at: 2003 Great Lakes Beach Association Annual Meeting, Muskegon,
MI.
Tetra Tech. 2003. Lake Michigan Shoreline TMDL for E. coli Bacteria. Modeling Framework Report.
Prepared for the Indiana Department of Environmental Management, Indianapolis, IN. Available
at: http://www.laportecountybeaches.com/resources/pdf/lkmichTMDLfrmwrk.pdf.
Tetra Tech. 2004. Lake Michigan Shoreline TMDL for E. coli Bacteria. Prepared for the Indiana
Department of Environmental Management, Indianapolis, IN. Available at:
www.in.gov/idem/tmdl_lakemich_report.doc.
Bacteria TMDLs for Illinois’ Lake Michigan Beaches Options Summary Report
7-4
Thupaki, P., M.S. Phanikumar, D. Beletsky, D.J. Schwab, M.B. Nevers, and R.L. Whitman. 2010. Budget
Analysis of Escherichia coli at a Southern Lake Michigan Beach. Environmental Science &
Technology 44(3): 1010–1016.
Whitman, R.L., M.B. Nevers, and M.N. Byappanahalli. 2006. Watershedwide distribution of Escherichia
colialong southern Lake Michigan: An integrated approach. Applied Environmental
Microbiology 72: 7301–7310.
Whitman, R.L., and M.B. Nevers. 2008. Summer E. coli patterns and responses along 23 Chicago
beaches.Environmental Science & Technology 42(24):9217–9224.
Whitman, R.L., T.G. Horvath, M.L. Goodrich, M.B. Nevers, M.J. Wolcott, and S.K. Haack. 2001.
Characterization of E. coli levels at 63rd Street Beach. Prepared for the City of Chicago.
Available at: www.glsc.usgs.gov/_files/reports/63rdStreetBeachReport.pdf.
Wolfe, K. R. Parmar, M. Cyterski, and M. Galvin. 2010. Virtual Beach v 2.0 User Guide. Draft. U.S.
Environmental Protection Agency. Center for Exposure Assessment Models. Athens, GA. 50 pp.