BEFORE THE ILLINOIS POLLUTION CONTROL BOARD
IN THE MATTER OF:
WATER QUALITY STANDARDS AND
EFFLUENT LIMIT A TIONS FOR THE
CHICAGO AREA WATERWAY SYSTEM
AND THE LOWER DES PLAINES RIVER:
PROPOSED AMENDMENTS TO
35 Ill.
Adm. Code Parts 301, 302, 303 and 304
)
)
)
)
)
)
)
)
R08-9
(Rulemaking - Water)
NOTICE OF FILING
To:
ALL COUNSEL OF RECORD
(Service List Attached)
PLEASE TAKE NOTICE that on the 10
th
day of April, 2009, I electronically filed with
the Office
of the Clerk of the Illinois Pollution Control Board, Metropolitan Water
Reclamation District of Greater Chicago's Responses to EPA's Technical Review
Comments Regarding the Report entitled: "Dry and Wet Weather Risk Assessment of
Human Health Impacts of Disinfection vs. No Disinfection of the Chicago Area Waterways
System," dated April 2008.
Dated: April 10,2009.
Fredric
P. Andes
David
T. Ballard
BARNES
&
THORNBURG LLP
One North Wacker Drive. Suite 4400
Chicago, Illinois 60606
(312) 357-1313
METROPOLITAN WATER RECLAMATION
DISTRICT OF GREATER CHICAGO
By:
/s/ David T. Ballard
One
of Its Attorneys
[This filing submitted on recycled paper as defined in 35 Ill. Adm. Code 101.202]
PROOF OF SERVICE
The undersigned, a non-attorney, certifies, under penalties of perjury pursuant to 735
ILCS
511-109,
that I caused a copy of the forgoing,
Notice of Filing of Metropolitan Water
Reclamation District
of Greater Chicago's Responses to EPA's Technical Review
Comments Regarding the Report entitled: "Dry and Wet Weather Risk Assessment
of
Human Health Impacts of Disinfection vs. No Disinfection of the Chicago Area Waterways
System," dated April
2008,
to be served via First Class Mail, postage prepaid, from One North
Wacker Drive, Chicago, Illinois, on the 10
th
day of April, 2009, upon the attorneys of record on
the attached Service List.
/s/ Barbara
E. Szynalik
Barbara
E. Szynalik
[This filing submitted on recycled paper as defined in 35 Ill. Adm. Code 101.202]
2
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
SERVICE LIST
R08-9 (Rulemaking - Water)
Richard J. Kissel
Roy M. Harsch
Drinker, Biddle, Gardner, Carton
Suite
3700
191 N. Wacker Drive
Chicago, IL 60606-1698
Deborah J. Williams, Assistant Counsel
Stefanie
N. Diers, Assistant Counsel
IEPA
Division of Legal Counsel
1021
North Grand Avenue East
P.O.
Box 19276
Springfield,
IL 62794-9276
Kevin G. Desharnais
Thomas W. Dimond
Thomas V. Skinner
Mayer,
Brown LLP
71 South Wacker Drive
Chicago, IL 60606-4637
Robert VanGyseghem
City of Geneva
1800 South Street
Geneva,
IL 60134-2203
Matthew J. Dunn, Chief
Office of the Attorney General
Environmental
Bureau North
Suite 1800
69 West Washington Street
Chicago,
IL 60602
Bernard Sawyer
Thomas Granao
Metropolitan
Water Reclamation District
6001
W. Pershing Road
Cicero, IL 60804
Claire
A.
Manning
Brown,
Hay
&
Stephens LLP
700 First Mercantile
Bank Building
205 South Fifth St., P.O.
Box 2459
Springfield, IL 62705-2459
Katherine D. Hodge
Monica T. Rios
Matthew C. Read
Hodge Dwyer
&
Driver
3150 Roland Avenue
P.O.
Box 5776
Springfield,
IL 62705-5776
Jerry Paulsen
Cindy Skrukrud
McHenry County Defenders
132 Cass Street
Woodstock, IL 60098
Lisa Frede
Chemical Industry Council
of Illinois
Suite 100
1400
E. Touhy Ave.
Des Plaines, IL 60019-3338
James
L. Daugherty, District Manager
Thorn Creek Basin Sanitary District
700
West End Avenue
Chicago Heights, IL 60411
Tracy Elzemeyer, General Counsel
American
Water Company Central Region
727 Craig Road
St. Louis,
MO 63141
[This filing submitted on recycled paper as defined in 35 Ill. Adm. Code 101.202)
3
Electronic Filing - Received, Clerk's Office, April 10, 2009
Keith 1. Harley
Elizabeth Schenkier
Chicago Legal Clinic, Inc.
4th
Floor
205 West Monroe Street
Chicago, IL 60606
W.C. Blanton
Blackwell Sanders LLP
Suite 1000
4801 Main Street
Kansas City,
MO 64112
Traci Barkley
Prarie Rivers Networks
Suite 6
1902 Fox Drive
Champaign, IL 61820
James Huff, Vice President
Huff
&
Huff, Inc.
Suite 3300
915 Harger Road
Oak Brook, IL 60523
Cathy Hudzik
City
of Chicago - Mayor's Office of
Intergovernmental Affairs
City Hall - Room 406
121 N. LaSalle Street
Chicago, IL 60602
Irwin Polls
Ecological Monitoring and Assessment
3206 Maple
Leaf Drive
Glenview, IL 60025
Marc Miller, Senior Policy Advisor
Jamie
S. Caston, Policy Advisor
Office of Lt. Governor Pat Quinn
Room 414 State House
Springfield, IL 62706
Frederick D. Keady, P.E., President
Vermilion Coal Company
1979 Johns Drive
Glenview, IL 60025
James
E. Eggen
Director of Public Works
&
Utilities
City of Joliet, Department of Public
Works
&
Utilities
921 E. Washington Street
Joliet, IL 60431
Ann Alexander, Sf. Attorney
Natural Resources Defense Council
Floor 23
2 N. Riverside Plaza
Chicago, IL 60606
Beth Steinhorn
2021 Timberbrook
Springfield, IL 62702
Dr. Thomas
J. Murphy
DePaul University
2325 N. Clifton Street
Chicago, IL 60614
Vicky McKinley
Evanston Environment Board
223 Grey Avenue
Evanston, IL 60202
Kenneth W. Liss
Andrews Environmental Engineering
3300 Ginger Creek Drive
Springfield, IL 62711
[This filing submitted on recycled paper as defined in 35 Ill. Adm. Code 101.202]
4
Albert Ettinger, Senior Staff Attorney
Jessica Dexter
Environmental Law
&
Policy Center
Suite 1300
35 E. Wacker Drive
Chicago, IL 60601
Tom Muth
Fox Metro Water Reclamation District
682
State Route
31
Oswego, IL 60543
Jack Darin
Sierra Club
Illinois Chapter
Suite 1500
70 E. Lake Street
Chicago, IL 60601-7447
Marie Tipsord, Hearing Officer
John Therriault, Assistant Clerk
Illinois Pollution Control Board
100
W. Randolph Street
Suite 11-500
Chicago, IL 60601
Stacy Meyers-Glen
Openlands
Suite 1650
25 East Washington
Chicago, Illinois 60602
CHDSOI 532756vl
Bob Carter
Bloomington Normal Water
Reclamation District
P.O. Box 3307
Bloomington, IL 61702-3307
Kay Anderson
American Bottoms RWTF
One American Bottoms Road
Sauget, IL 62201
Kristy
A.
N. Bulleit
Brent Fewell
Hunton
&
Williams LLC
1900 K Street,
NW
Washington, DC 20006
Lyman
C. Welch
Manager, Water Quality Programs
Alliance for the Great Lakes
17 N.
State St., Suite 1390
Chicago, IL 60602
Mark Schultz
Regional Environmental Coordinator
Navy Facilities and Engineering Command
201 Decatur Avenue
Building
lA
Great Lakes, IL 60088-2801
[This filing submitted on recycled paper as defined in 35 Ill. Adm. Code 101.202]
5
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
Metropolitan Water Reclamation District of Greater Chicago
BOARD OF COMMISSIONERS
Terrence J. O'Srien
President
Kathleen Therese Meany
Vice President
Gloria Alitte Majewski
Chairman 01 Finance
Frank Avila
Patricia Horton
Barbara
J.
McGowan
Cynthia M. Santos
Debra Shore
100 EAST ERIE STREET
CHICAGO, ILLINOIS 60611-3154
312'751'5600
Patricia Young
Louis Kollias, P.E., BCEE
Director of Research and Development
312'751-5190
Me Andrew Tschampa
Acting
Chief Water Quality Branch
United States Environmental Protection Agency
Region V
77 West Jackson Boul.evard
Chicago, Illinois 60604-3590
March 13, 2009
E COpy
Subject: Response to EPA I.zeview of Dry and Wet 'Weather H.isk Assessment of
HU111an Health Impacts of Disinfection vs. No Disinfection of the
Chicago Area Waterway System
\Ve wish to express our sincere appreciation for the time, effort and expertise that EPA
brought forward
by
reviewing the
dry
and
wet
weather risk assessrnent of the Chicago Area
Waterway System (CA \VS) report. The revie\v comlTlents brought forth were sent to Geosyntcc
Consultants (Geosyntec)
for a detailed and thorough asseSSlTlent of the comments. Responses to
the technical comrnents were received from Geosyntec in a letter dated March 1 1, 2009.
/\ttachcd are the letter and the response document that describes how each comment \vas
addressed in the final report.
Geosyntec found a number of the comments valuable in providing guidance to strengthen
the
presentation of the science in the report. These comments prove valuable in our efforts
[0
publish the research in pcer-reviev.,rccl journals. We concur with the EPA's comment [hat
quantitative microbial risk assessment is an area of research where the ground is not as \vcll trcad
as that in
chemical risk assessment. We are confident that the risk assessment
peri~)fmed
by
Geosyntec represents thc best effort the current state of the science can provide. 'fhere are
inherent uncertainties and assumptions in .microbiaJ risk assessment methodology; and, therefore.
the
District has undertaken a companion epidemiological (Chicago lIealth Environmental
Exposure
&
Recreation Study [CHEERS]) assessment of the health risk to incidental contact
recreating population on the CAWS v;bich is necessary to complete and verify the results of the
quantitative rnicrobial risk assessment study. To date, no study has validated any
qU~lI1[itative
microbial health risk study. The CLII:::ET<S will be the first study to bridge the science of
microbial risk assessrnent with direct public health assessment for secondary contact recreation.
Electronic Filing - Received, Clerk's Office, April 10, 2009
I\1r. Andrew 'T'schampa
March 13, 200')
Subject: Response
to EPA Review of Dry and Wet WCJther Risk Assessment of
Human Healtb lrnpacts of Disinfection vs. No Disinfection of the
Chicago Area Waterway System
We believe the responses presented
by
Geosyntec provide clarification
on the
comments
noted by the reviewer.
If
you have any questions regarding the enclosed letter, please calJ me at
(312) 751-5190.
LK:GR:ss
F!'nclosure
cc w/enc.:
Marcia Willhite, Illinois EPA
Very truly yours,
Louis
Kollias,
Director
rvloni(oring and Research
Ephraim King, USEPA Office of Water Washington D.C.
cc wlo enc.:
Chriso Petropoulotl, Geosyntee Consultants Chicago
Lanyon/Felclman/BilllClrarwto/O'Connor/RijallGlY1TJph
Electronic Filing - Received, Clerk's Office, April 10, 2009
Geosyntec ()
consultants
1:;"1 :\ LaS.,llc
~t
Suite :30n
Chicago, lL GOGO;,>
I'll :112-f):iH-IJ':iOO
HX JL!-G:3iH)S7b
IVW\V,gCOSYIl[CC.COlll
Via E-Mail and U.S. Mail
11 March 2009
Dr. Thomas C. Granato
Assistant Director
of Research
&
Development
Metropolitan Water Reclamation District
of Greater Chicago
6001
W. Pershing Road
Cicero, Illinois 60804-4112
Subject:
Responses to
EPA's Technical Review Comments Regarding the
Report entitled: "Dry and Wet Weather Risk Assessment of Human
Health Impacts of Disinfection Vs. No Disinfection of the Chicago
Area Watcnvays System," dated April 2008
Dear Dr. Granato:
Geosyntec Consultants (Geosyntec) is enclosing responses
to EPA's technical review
comments regarding the subject repoli (see Enclosure). The responses follow the
corresponding EPA comment(s).
In addition, the responses refer to EPA's Technical
Review Comments Regarding the Interim Phase I Report, dated November 2006,
"Dry
Weather Risk Assessment of Human Health Impacts of Disinfection Vs. No Disinfection
of the Chicago Area Watenvays System", which are included as an attachment to the
Enclosure.
If you have any questions or comments regarding the enclosed report please call me at
(312) 658-0500.
Enclosure
engineers
I
scientL;rs I innovators
Very truly yours,
;/h
//Vl/-
f
'.
/;1
'-V,
,_
/
.J'
I
r
)
.1fJ"
,!
./7-~.::-:;r1:?;<""'--/c/2.A
___
Chriso Petropoulou,"Ph.D., P.E., BCEE
Associate
Electronic Filing - Received, Clerk's Office, April 10, 2009
Electronic Filing - Received, Clerk's Office, April 10, 2009
EPA Review of Dry and Wet Weather Risk Assessment of Human Health Impacts of
Disinfection
vs. No Disinfection of the Chicago Area Waterways System
This docllment provides EPA's comments on MWRDGC's Dry and Wet Weather Risk
Assessment. We praise MWRDGC for their willingness
to undertake the effort and expense
associated
v'lith the data collection and analysis in this repOli. We understand that quantitative
microbial risk assessment
is an area of risk assessment whcre the ground is not as well tread as that
in chemical risk assessmcnt and appreciate MWRDGC's challenge in developing the
repOli. The
work that MWRDGC
is doing in the area of risks f1'0111 exposure to fecal contamination from
secondmy contact recreation
is of interest to EPA and we believe it is criticallyimpOltant to ensure
that it
is accurate, transparent and scientifically defensible. We have provided numerous comments
to help MWRDGC improve the report so that it can achieve those goals and would like to offer to
discuss and answer any questions you may have regarding our comments.
This Agency review
is summarized into two main parts; a process-oriented section under "General
Comments"; and, a technical evaluation under "Technical Comments."
General Comments
Risl{ Assessment versus Risk Management and Policy getting
This
repOlt confuses the purposes of risk assessment with risk management and policy setting (e.g.,
see
p.
xiv,
"Microbial Risk Assessment Objectives" pp.
xxix
-
xxx,
"Wastewater Disinfection" and
"Microbial Risk Assessment"). The lack
of clear delineation between these various functions
severely hampers the importance
of transparency in the risk assessment process. In this case, the
goal
of a microbial risk assessment is to estimate the potential for human disease associated with
exposure to waterborne pathogens or a medium
in which the microbes occur. This risk estimate
should be derived
in a transparent fashion and be scientifically defensible. As stated in the ILSI
Revised Framework/c)r Microbial Risk Assessment
(ILSI, 2000) in regards to transparency: "methods
and assumptions should be clearly stated
~U1d
understandable to the intended audience ... " and the
"audience should be able
to evaluate the adequacy of the data and methods from the provided
information. "
Response: The text in the last sentence of the above jJaragraph infers that the 2008 Geosyntec
report does not meet the
ILS! requirements regarding transparency: "methods and assumptions
should be
clearly stated (lnd understandable to the intended audiellce ...
/I
{[lui the "audience should be
able to evaluate the adequacy of the data and methods from the provided information."
However, a review of tlte 2007 Interim Geosyntec Report conducted by the US EPA Office of
Research af1{1 Development for us EPA Region
5,
Office of Water, states the following (see
Attachment A): "The general approach described
for the QMRA also seems appropriate. The
authors
do a thorough job of explaining and justifying their selections
(~f
dose-response
functions
al1d their parameters.
Generally, citations from peer reviewed literature (Ire
provided to support their decisions."
Therefore, it appears that some EPA reviewers believe that the QMRA provides
ty(lI1~l)(lrel1cy
while others disagree. In orderfor Geosyntec and MWRDG to address the EPA comments, we
need to receive consistent and specific comments that we can address.
However, the stated main objective of the MWRDGC dly and wet weather risk assessment "was to
evaluate the human health impact
of continuing the ClU"rent practice of not disinfecting the effluents
ii'om the District's" wastewater treatment plants
(p. xiv, Executive Summmy). This oqjective is
Electronic Filing - Received, Clerk's Office, April 10, 2009
clearly a policy aneVor risk management decision that should be informed by the risk assessment.
While the risk assessment process should be iterative in nature and requires input from risk
managers even in the initial problem formulation phase, it should not be used to simply justify,
a policy decision.
As such, this risk assessment appears compromised in its function and
purpose and the report's conclusions appear suspect.
Response: The stated objective was formulated in MWRDGC's Request for Proposal, dated
January
2005 with input fr0111 variolls stakeholders, including the IlIil10is Environmelltal
Protection Agency. Tlte risk assessment did no/ include any objectives
to justffy a policy
decision. The study objective was formulated
to evaluate, estimate and compare recreational
healtlt risks in tlte Clticago Area Waterway System with alld without ejJluellt
disi1~fectiol1.
The
same objective
was stated in tlte 2007 Interim Dry Weather Report and the EPA reviewers of
tlte subject report did not express any concern about the objective (see Attachmel1t A).
Need for Clear Problem Formulation
Another
maj01~
criticism of this report is the lack of a coherent problem formulation and
development
of a transparcnt conceptual model. This criticism was identified upon review
of the dry weather risk assessment and was never satisfactorily addressed. The problem
formulation is iterative in nature and
of critical importance in the risk assessment process
and should include input from both risk managers and assessors.
Response: The reviewer's (f5;sertion that the QMRA lacks
([
tramparent conceptual model and
a thorough ul1certaill(vlvariability (Inalysis is incorrect. Section
5.2
of the report presents the
conceptual exposure model
of the recreational lise oftlte waterway. Section
5.4.7
of the report
discusses Sensitivity and Uncertainty analysis. Tables
5-16
and
5-17
present pertinent results.
Tlte iterative problem formulation process was
110t within tlte scope of work of tlte Geo:o,Yl1tec
QMRA.
Additionally, a sampling schematic would be helpful to track the various sample methods,
as well as, a table and corresponding justification for the parameters chosen in thc risk
assessment. Having both would greatly improve transparency.
Response: Figures
2-1
and
2-2
presents the sampling locatiolls during the
dry
(lnd wet weather
samples. Table
2-2
presents the
dry
(lnd wet weather samples. Table
2-1
presents a SUI1111Ull)J
of the pathogenic microorganisms selected for the microbial risk assessment and rationale for
their selection. Section
2.3.2
discusses in detail the sample collection equipment, materials
(lnd procedures and Section
2.4.1
presents the microbial methods of analysis. Furthermore,
Tables
5~1
though
5-8
sllmnwrize all parameters chosen for the microbial risk assessment.
Geo'~J!fltec
believes that adequate schematics and tables were provided in the report and aI/ tlte
information IIsed in the QMRA is clearly
and transparently presented.
Additionally, as stated in the report, roughly 70% of the annual flows into the waterways are
from
1I11disinfected sewage treatment plant discharges. This number would most likely be
higher in dry weather and lower during wet weather (i.e., the contribution of precipitation
to the waterways versus the volume
of un disinfected effluents). Conversely, approximately
30% of the annual flows into the waterways are unspecific (e.g., urban runoff CSO
overflows, direct precipitation, etc.). This significant component is mostly ignored by the
risk assessment other than to
make a qualitative attempt to discuss pseudol11onads. The
approximately 230 CSOs
on the waterways were not covered, nor sampled during wet
weather events (Region
5, verbal communication). This component could have been
identified and discussed had a coherent problem formulation, including a transparent and
2
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
clear conceptual model, been employed in the risk assessment process.
Response:
011
the contl'm)', risks were developed lIsing waterway data that accollnts for all
sources to the
watenvc()Js. Section
2.2.2
of the report discusses the Wet Weather Objective of
the Microbial Risk Assessment. One of the ,vet weatller objectives of the microbial risk
assessment was to evaluate the impact
oj combined sewer overflows (CSO.\) 011 tlte microbial
quality
of tlte Clzicago Area Waterway System (CA WS). The pUI1Jose of microbial sampling
during both
(by
and wet weather was to measure the microbial concentrations in the CAWS,
where recreational activities take place. During wet weather, tlte
CAWS receives microbial
loads due
to slI1:face I'unoff, storm drains, overland runo.ff, land lise activities (such as
agriculture and cOils/ruction), erosion, and habitat destructioll and CSOs, including
discharges
from three major pumping stations (North Branch, neal' the North Side WRP,'
Racine Avenue, neal' the Stickney WRP;
and
125(/1
Street, near the Calumet WRP). During wet
weather sampling, samples were collected vel)) near tile pumping stations at locations
determined by the sampling boat captain to be safe. Because
of tlte turbulent flow conditions
induced from the pumping station discharges, it was not possible to sample at the exact point
of discharge. Table
2-3
in the report presents the CSO volumes diSCharged during wet weather
sampling. The
pumping statiolls contribute relatively large volume of CSOs in the waterway.
Therefore, the sampling pel:formed neal' the pumping stations during the wet weather
sampling events
Itas accounted for the contribution of CSOs
011
the microbial qualizr of the
C4WS.
Need for Peer Review
For the report and its conclusions to be considered "scientifically defensible," we strongly
recommend that it
be subject to the same type of external peer review that you arc
conducting for your secondary contact epidemiological study
(el.lEERS).
We feel the
process
of an objective peer review (including incorporating changes in order to address
peer review comments) would allow MWRDGC to strengthen the validity
of the report and
its conclusions.
Response: Tlte QMRA study was conducted by experts using EPA-approved methods ([nd
state-oj the science techniques. The results o.f the study are scientifically defensible. As
indicated earlier, the EPA reviewers' general comment
011
the
2007
Interim DIY Weather
Report acknowledged that world-renowned experts were consulted to cOllduct the QMRA. The
reviewers
further commented that the general approach described for tlte QMRA was
appropriate and the authors did a thorough job of explaining and justifying the selections of
dose-response fUllctions and their parameters witlt citations !i'om peer reviewed literature ('lee
Attachment A). EPA '5 review comments on the Interim DI)' Weather Report ([nd re:o.pollses
submitted by the project team referencing the sections of the Final Report where the changes
were incorporated are provided in the
Attachment A. MWRDGC is pursuing peer review of the
findings of the stmzp by jJublishing tlte results in peer-reviewed scientificjoumals.
Tn addition, it is unclear whether the subcontractors on the Geosyntec team (as listed in the
Executive Summary) have reviewed the final report and would agree with the use and
interpretation
of data they provided. They should be given this opportunity or a more
accurate description
of their contribution to the report should be provided.
Response: The Geosyntec Team, wltich includes Geosyntec Consultants (Ge05)!f1tec) and its
subcontractors: Patterson Environmental Consultants
(PEC),' Cecil Lue-Hing
&
Associates
(CLHA); Dr. Charles Gerba
of the University of Arizona (UA); Hoosier Microbiological
Laboratory, Inc. (HML);
al1d Clancy Environmental Consultants, Inc. (CEC) worked
seamlessly to
pe~iorm
the Microbial
Ri~'k
Assessment study ([nd to prepare the report. Tlte roles
3
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
of each team member were defined ([t the proposal stage of the project. Also, these roles ((re
described in the 2005 and 2006 Sampling and Analysis and Quality Assurance Project Plans,
which are referenced in tlte April
2008 report. Geosyntec had overall responsibility foJ' the
management
of the project and foJ' peljol'ming the microbial risk assessment. At the Ol1set of
tlte study, Dr. Gel'ba provided ol1-site training to the District personnel
011
sample collection
procedures. The subcontractor lahoratories used
for this study (Ire veJY reputable {Ind have
assisted
in the development of EPA-approved methods. The laboratories analyzed the microbial
samples (lnd submitted laboratory reports sllmmarizing the (fl1alyticalresults that were included
ill the Final Report Appendices and formed tlte database for the QMRA. CEC analyzed the
Cryptosporidium and Giardia samples
and provided pertinent laboratory reports.
HML
analyzed the bacteria
and culturable virus samples and provided pertinent analytical reports.
The Environmental Virology LaboratOlY, Department
of Soil, Water and Environmental
Science at the University
of Arizona that peljorl11ed the analysis of adenovirus and 110rovirusfor
tlds study under the direction
of Dr. Gerba. However, it was flot the role of the t/tree
subcontractor laboratories to review the Final Report. The project Iwd three peer reviewers:
Drs. Charles
P. Gerba, Cecil Lue-Hing al1d James W. Patterson, served il1 tlte senior scientific
advis01Y committee
for tile project and provided direction ((l1d peer review
011
evely aspect ojthe
JVork pel/ormed.
PUl1JOSe of Disinfection Chapter
The "Disinfection" section (Chapter 4) of this report serves only to obfuscate the purpose of
this risk assessment. While the discussion of disinfection efficacy, indicator organisms and
pathogens was relatively accurate, it seems tangential
to the actual purpose of estimating
the potential for human disease associated with exposure to waterborne pathogens or a
medium in which the microbes occur.
Re.\pol1se: The main objective of the Microbial Risk Assessment Study was to evaluate tile
h timan health impact
of continuing the current practice of 110t disinfecting the effluents from
the District's Nortlt Side, Stickney and Calumet water reclamation plants versus initiating
disinfection
of the effluent at these three plants. This objective was formulated il1 the
MWRDGC Request for Proposal (RFP) for tltis study. Therefore, the GeosYl1tec Team
peljorl11ed a desk-top study
of peer-review technical literature on wastewater pathogen and
indicator disinfection and
sU11l11Utrized the findings in Section
4
of the report. Disinfection
effectiveness
of chlorination/dechlorin([tion, ultraviolet oxidation and ozollation was
sUI1U1U1l'ized, because these (Ire the technologies cllrrently evaluated by MWRDGC for tlte
Nortlt Side, Stickney (tnd Calumet facilities. Tlte range
of
disiJ~fection
effectiveness reported
for eaclt selected pathogen for the QMRA stU({v was lIsed to estimate the expected pathogen
removal, under the disinfection scenario.
Also of superficial relevance to this human health risk assessment is the discussion of
potential risks hom disinfection byproduct CDBP) formation and exposure. The authors
state that human health effects associated DBPs tend to be chronic in nature and therefore
the development
of a risk assessment for exposure to chcmical constituents, including
DBPs, is far morc complex than the microbial risk assessment. First, even less
is known
about the chronic effects on human health from single ancl/or repeated exposures
to
pathogens, However, data have shown that chronic fatigue syndrome can be linked to
chronic infection by enteroviruses (Kerr,
2008,.J Clin. Pa/hoi,
61: 1-2; Chia, 200S,.J.
Clin.
Pathol
61
:43-48).
Response: This stU((v addresses microbial risks on(v and it does 110t address chemical risks
4
Electronic Filing - Received, Clerk's Office, April 10, 2009
quantitatively or qualitatively. The point of the statement in the report was to acknowledge
the chemical risks of
dish~fection
by-products. The text on page
91,
PaJ'(lgrap/z
3
of tlte
report states that: "Risk assessments
of wastewater
disi1~fectiol1
should consider microbial
al1d chemical quality. The health effects of
di~i1ifectal1ts
are generally evaluated by
epidemiological studies and/or toxicological studies using laboratory (ll1imals."
The
qual1tf{ication
of chemica/risks due to disinfection by-products was olltside the scope of work
of tlzis study. Also, the chronic effects of pathogens on
11ll11lall
health were 110t evaluated in
tlzis study.
Second, a properly conducted microbial risk assessment, including all of the components
necessary to develop a transparent and scientifically defensible evaluation, can
be as
complex a procedure as the development
of a chronic toxicity human health risk
assessment.
Re~pol1se: Geo~yntec
believes that the QMRA study was conducted properly and includes all
of tlze necessary components.
-
There are differences in the structure and approach between a chemical and microbial risk
assessment, but either can range from simple (e.g., in the case
of a qualitative or screening
level deterministic point estimate assessment) to complex (e.g., in the case
of a
probabilistic risk estimation that includes the dynamic nature
of prior immunity and
secondary pathogen spread). That the authors felt that this microbial risk assessment lacked
needed complexity only underscores the need for a proper problem formulation, conceptual
model, and thorough uncertainty/variability analysis. Indeed, it is important to account for
system variability that can lead to changes in exposure and microbial risk because short
periods
of exposure to high pathogens levels can result in greater risk
(Clean Water: What
is Acceptable Microbial Risk?
Amen Soc. Microbial., 2007).
Response: The reviewers comment makes unsubstantiated assertions about what the authors
felt.
Geo!>yntec does notfeel that this QMRA lacked needed complexity. There is
/10
statement
in the report describing such feelings. Also, tlte reviewer's assertioll that the QMRA lacks a
conceptual
model and a thorough uncertainty/variability is incorrect. Section
5.2
of the report
presents the conceptual exposure
model of the recreatiol1al use of the waterway. Sectioll
5.4.7
of the report discllsses Sensitivity and Uncertainty analysis. Tables
5-16
and
5-17
present
pertinent results.
General Issues in Chapter 5
The use of an outdated risk assessment model (e.g., Chapter 5) further hampers
transparency and confidence in this report's conclusions. See the ILSI "Revised Framework
for Microbial Risk Assessment" enclosed with this review (ILSI, 2000).
Respol1se: Geosyntec llsed tlte same risk assessment as in the reference provided in tlte reviewer's
comment. Dr. Gerba in
0111'
team contributed in the development of the 1LS1 model and Ite
cOJ{{irmed that tile model used
ill
the QMRA study is identical to the 1LS11110del.
Chapter 5 also contains numerous inaccurate statements and broad sweeping statements
based on assumptions with little or no justification. For example, gastrointestinal illness is
the principal adverse outcome associated with exposure to
lecally contamh1([fed
(i.e.,
human and or animal waste) waters, not
just water containing microbes (note: all ambient
waters arid many drinking waters contain microbes). To date, rates
of gastrointestinal
illnesses have been most strongly correlated with indicators
of fecal contamination 111
5
Electronic Filing - Received, Clerk's Office, April 10, 2009
epidemiology studies conducted 011 predominantly POTW-impacted (note: with
disinfection) waters, hence the general acceptance of this category of illness as the
'principal adverse outcome'. The pathogens of concern vary by fecal source, but many can
cause gastrointestinal illnesses
of varying severity.
Response: This comment makes 111a11Y broad alUl unsubstantiated claims. Text in Section 5.1
refers to microbial contaminated wlIter, not just water containing microbes as the reviewer
claims.
The text
il1 Section 5.1 of the report refers to microbial pathogens tltat can
contaminate the water
([l1d
cause gastrointestinal illness. Text in Section 5.1 also states that
fecal-oral transmission associated with gastrointestinal illness is the primary effect evaluated
in this stu{{Y.
The authors also state that there is correlation betvieen different pathogens. This
uncorroborated statement is an inaccurate broad conclusion. Which human pathogens are
present
in a waterbody is determined by the source(s) of those pathogens and degree of
treatment those pathogens undergo during their fate and transport.
For example,
Pseudomonas
sp. tends to occur in urban runoff in high numbers (EPA, 1977,
JVJicrobes in
Urban Sformwater;
Pitt, 2002,
Storm water Effects Handbook,
chapt. 3), but is only one of
the pathogens of concern from this particular source. Indeed, the authors do not attempt to
justify or explain how they compare risk with the different pathogens and potential disease
endpoints
in mixed source waters (I.e., are less variably-occurring pathogens with
potentially lower relative illness severity equal to or different from variably occurring
human and zoonotic pathogens with potentially higher relative illness severity). Given that
-30% of the average annual inputs into the waterways can be from non-POTW sources,
more results and discussion is needed on this topic.
Response: We concur with the reviewer's comment about Pseudomol1as occurring ill urban
runoff in high numbers.
The results of the QMRA study indicate that the SOllrces of
Pseudomol1as aerllgil10sa during wet weather are sources otlter than tlte WRP effluents.
However, we disagree with the remaining comment. The QMRA
StlU(Y accollnted for the
effect
of wet weather by collecting (I/1d analyzing s({mples from tlte CA WS during wet weather
events. Sections
3.1, 3.2
({l1d
3.3
of the report disCliSS the wet weather bacteria, protozoa and
virlls results in detail.
In addition, a comparison between
{by
and wet weather results is
provided.
The sometimes-controversial issue of what constitutes the secondary contact portion of the
designated recreational use underpins this risk assessment.
The report attempts to
characterize secondary contact activities (e.g., wading)
in contrast to what can be normally
considered primary contact recreation (e.g., swimming). All 'high' risk secondary contact
activities were combined into the 'canoeing' category or as the report describes, "low
contact boating". These activities include: canoeing, kayaking, sculling, and jetskiing.
Additionally, while observed occurring during the UAA survey,
\;"Iading and swimming
activities were not included at all in this assessment. We recommend more appropriate
categorization for some
of the activities in the "Jow contact boating" category (e.g.,
kayaking, sculling) as we believe they may carry a higher degree of likely incidental or
accidental ingestion than canoeing (i.e., closer to that
of primary contact). These activities
would then be assigned greater consumption values based on the higher exposure. While
one can debate the differences between the consumption values, hence the exposure,
for the
6
various activities in the 'high' risk 'canoeing' group, it is important that the analysis reneet
the full range
of exposures for such activities and not underestimate them.
Response: Exposure was divided into
3
exposure categories; high medium and low exposllJ'e
groups. Representative activities were ascribed to each of these categories,' canoeing, fishing
and boating. For each exposure categol)', input distributions
were deve/ope(1 for lise in the
QMRA. The QMRA accounted for the
full range of expected exposures for all activities in
this categoJ}' by lIsing exposure dumtion and ingestion distributions, which {Ire discussed in
detail in Section
5.2
of the report. J(ayaking (Illd sculling were evaluated tiS high exposure
activities. Tlte input range for the high exposure "canoeing" group includes the potentialfor
ingestion that ranges/rom minimal contact with the CAWS
to exposure levels that Me similar
to those lIsed for swimming ingestion levels. Therefore, we believe that the high exposure
categ(1)' (i.e., canoeing) adequately captures the potentia/for higher incidental intake ofw(tter
while recreating.
Stylistic Comments
A couple of stylistic issues hamper the transparency of the report. First, the executive
summary is rather long and the presentation of the results does not occur until page
xxiv.
An effective executive summary states the bottom line up front. Additionally, typically an
effective executive summary is much shorter in length. Second, having the various tables
and figures embedded in the chapters when they are referenced would facilitate
comprehension
of the report as a whole. This is even more important when the report is
only viewed in the electronic (-pdf) format. Given the current state of desktop publishing
and the relative ease inserting the tables and figure in the text that these software packages
allow,
EPA would suggest making this change prior to submitting this report to an external
peer review.
Response: The style of the reportfollolVs a typical Geo!Jyntecfo1'11wt. Tlte same style was llsed
Jor the Interim Dry Weather Report (lnd the EPA reviewers
of tltat report did not have any
concerns about tlte style
of the report.
Technical Comments
/))lJ10psis
(~ll11qjor
comments:
• Variability of concentration of pathogens in water appear not to have been adequately
addressed
in the risk assessment nor was sensitivity analysis of that key variable
reported.
Re!Jpollse: Tlzis comment is misleading and inaccurate. Section
5.4.
J
of tlte report discllsses
the bootstrapping method that
was llsed in tlte QMRA. Bootstrapping is a widely accepted (lnd
extensively lIsed procedure
in statistical analysis {{nd represents a process of selecting a
random il1putfrom a dataset. This tecltnique is useful
in Monte Carlo analysis wlten tlte exact
distributional Jorm
oj (In input v(lriable is either unknown or unable to be represented Ivith
((
continuolls distribution. Bootstrap samples are random selections from the empirical data
with replacement. Bootstrap methods provide robust estimates of variability in Monte Carlo
assessments as the probabilities associated witlt drawing extremes
ill
the distributioll is
mimicked by the presence of extreme values in lite empirical data.
Assumptions are not provided. For example, the report should provide a table that clearly
lays out, for each pathogen assessed, the assumptions and descriptive parameters used.
7
Response: Section
5.3
and Tables
5-5
through
5-7
in tlte report present each pathogen
assessed,
the assumptions and descriptive parameten' used.
It
is not appropriate to combine thc \vet and dry vlcather analyses, as that will
underestimate the risk from the wet weather events.
Re.\ponse: It appears that in this comment EPA implies that the ell)' ([nd wet weather results
were arbitrarily combined, which is incorrect.
The
(I1:v
and wet weather results were
integrated to simulate tlte climatic conditions expected within
{/
recreational season, based 011
({ctual weather and pllmping station discharge occurrence dafa as described
ill
Section
5.4.4
of
the report and Table
5-8.
To represent risks from recreational exposure across tile entire
recreatiol1al season, tlte input pathogen cOllcentrations used in the risk assessment should
account
for tile probability of encollntering pathogen concentrations related to different
weather conditions. The proportion
of days under each weather condition in
(f
recreational
year was developed
fr0111 historical records of CSO and
rab~rllll
records.
The input
distribution used in tlte simulations
for selecting weather specific pathogen cOllcel1tmtions is
shown in Table
5-8.
A conservative assumption was made in tltis (ll1a(vsis that recreational use
and weather conditions are not correlated.
C0111mon experience would suggest tltis is not tlte
case
as people tend to 5pend less time recreating during rain events. Tlte assllmption in the
QMRA was that recreational lise may resume sltortly
({{tel' /'ain events wlten waterway
concentrations
(Ire still strongly
i1~flllel1ced
by the preceding weather pattel'lls.
•
Based on information presented in the report, it is difficuit to get a clear picture
regarding the quality
of the data (e.g., assumptions not provided, no description of
method recovery, no probability density functions (PDFs) used to describe viability
nor if viability data was uscd in the estimates of pathogen concentrations,
inappropriate mllnber
of significant figures),
Response: Tlte reviewer's assertion fltat tltere is
110
description
011
method recovery is
inaccurate. Where appropriate, method recovet:y was discussed. For example, Section 2.4.3
of the report presents tlte quality control (QC) data of (11/ microbial results, including
O),ptosporidiul11 and Giardia QC data. Tlte following
QC samples were analyzed for
O:~'ptosporidium
and Giardia: Matrix spike (MS), ongoing precision and recovery (OPR), and
method blanks.
The
reviewer's assertion that no recovery data is presented
1101' corrected for is inaccllrate.
Section
2.4.3
of the report summarizes the recoveJ)' /'(/Ies of the MS and OPR s([l'nples. MS
results were within the acceptance criteria
,~pec{fled
in EPA Method
1623.
In addition, all
recoveries were well within the acceptance criteria specifiedfor OPR samples' in Method
1623.
No oocysts oJ' cysts were detected in method blanks analyzed indicating no contamination in
the spiking
01'
sample processing procedures.
EPA Method
1623
does not allow the lise of MS recovery I'esults to adjust the samples. Text in
Section
2.4.3
of the report states that the MS sample results were not lIsed to adjust
0J'pt05poridiul11 and Giardia recoveries at any sampUng locatioll.
PDFs were
not llsed for viability because a velJ' small percentage of samples had viable Giardia
cysts. The average
viabiliZJl was used to ([(!just the detactable concentrations of Giardia
ill
the
samples.
Report does not provide information on the duration of the wet weather discharges
(events). This is critical in understanding the exposure to recreators, in essence,
8
what is the time to return to 'background' conditions versus when recreation may
resume?
Re:,ponse: Table
2-3
ill the QMRA provide both, tlte pumping station discharge volumes
ill
millions of gallons and tlte duration of the discharges. In additioll, Section
5.4.3
and Figure
5-7 (lnd Section
5.4.4
and Table
5-8
discuss tlte integration of dly (lnd wet weather data in the
QMRA. TIte assumption in the QMRA was conservative in that recreational use may resume
shortly {ifler rain events when waterway concel1tmtions are still strongly
i1~f1l1el1ced
by the
preceding weather patterns.
Data regarding the removal of pathogens through secondary treatment appears to differ
from published data - no discussion presented to explain this.
Re:,pol1se: It is flOt clear what this comment
r~fers
to. Tlte removal efficiency of pathogens
through the second{l/J' treatment was
110t assessed in this study. Specifically,
110
il1fluent
untreated wastewater samples were collected.
Therefore, the reviewer's assertion
is
unsubstantiated and false. The QMRA microbi(fl
conc~nt/'{{tions
are based Oil all extensive
microbial characterization
of the District's final effluents. The QMRA results indicate that the
pathogens are generally lower
thall
tilat observed in several other sewage discharges reported
in the literature. The analytical microbiological results reflect the actllal concentrations
measured in the
WRP effluents.
This report (as provided on MWRDGC's website) is missing Appendices B-D and, thereiore, we
could not view the analytical data that serve as the basis for much of the analysis.
Re:,pol1se: Mr. Louis Kollias, Director of Research and Development of the Metropolitan
Water Reclamation District
of Greater Chicago, provided a copy of the April 2008 Geosyntec
report, entitled: "Dry
and Wet Weather Risk Assessment of Human Health Impacts of
Disinfection Vs. Non-Disinfectiol1 of the Chicago Area Waterways System," to Mr. Allen
Melcer, Water Quality Branch, EPA Region
5.
In Ilis transmittal letter, Mr. Kol/ias
specifically acknowledged that tIle raw data can be made available upon request. To this
date,
MWRDGC has 110t received a requestfrom EPA for the raw data.
General
In describing the results of a quantitative microbial risk assessment there are two key
issues:
1) characterizing the estimated risk(s) against some benchmark or relative measure,
and 2) identifying uncertainties where possible so as to better inform those interpreting the
results. This report uses a draft benchmark risk for recreational water use
of 14 illness per
1000
exposed recreators - which is neither adopted nor policy of the U.S. EPA. That value
was discussed in an EPA draft guidance document that was never formerly issued. EPA's
current criteria are based on 8 cases
of highly credible gastroenteritis per 1000 for
freshwaters, and 19 cases per 1000 for marine waters. None
of the targeted
thresholds/benchmarks are presented with statistically-derived confidence intervals or
ranges that
reneet the uncertainties associated with their values, including the values
reported in the final study.
Re:,pol1se: Table 5-10 of the report presents
((
Slll11nulIY ofvariolls EPA acceptable swimming-
associated gastroenteritis rates
per 1,000 swimmers. Because EPA does not currently have
microbial water quality criteria
for second([JY contact recreation, Geosyntec considered all
historical criteria. Footnotes to Table 5-10 provided clarifications
and citations of the sources
of the information presen.ted. The rate of acceptable swimming-associated gastroenteritis of
9
14 illnesses per 1,000 swimmers was a limit that EPA proposed in May 2002. Table
5-9
in the
report
summarizes tlte total expected illnesses under dry, wet and combined
(l1:v
(Jnd wet
weather events. Although the designated lIses of the CA WS do 110t include swimming (lnd
other
primmy contact activities, the results in Table
5-9
indicate that the total expected
illnesses
of recreational users in the CAWS are below EPA's current criteria
(~f
8
ifllless of
highly credible gastroenteritis per 1000 swimmers.
In additioll, the reviewer's assertion that "/lOlle of tlte tmgeted threshold!Jjbel1cltmarks are
presented with statistically-derived confidence intervals or ranges that reflect tlte uncertainties
associated witlt their values"
is ;'lcorrect. Sectioll
5.4.7
of lite report discllsses the sensitivity
and uncertainty analysis that was pelformed on the microbial risk assessment results. Results
of the sensitivity evaluation (Ire presented
011
Table
5-16.
Table
5-17
presents all alternative
sensitivity evaluation.
Whether the waters are natural or not is not considered relevant when determining if the
human exposure from recreation presents unacceptable risks. Given that 70
% of the annual
flows in the CWS
.are fr0111 discharges of secondary treated municipal wastewater effluent
ftom the District's WRPs (this review has assumed largely from Calumet, North Side and
Stickney) - focus should have been given to the performance
of these wastewater facilities
with respect
to frequency and duration of unacceptably high pathogen loads (which could
have been back-calculated from the target 'acceptable' risk level). That is, what
is the
duration
of the wet weather discharges?
Re!Jpol1se: Geosyntec agrees with the reviewer's comment that whether the waters are l1atural
OJ'
110t is not considered relevant when determining
It
uman exposure from recreation. In fact,
for the QMRA study,
(by
and wet weather sluface water samples were collected from the
CA WS, that receives contributions of both:
(1)
discharges of secol1(/(u), treated municipal
wasteHYlter
effluent from the District's WRPs at the North Side, Stickney and Calumet; altd
(2)
wet
weather inputs. Therefore, the QMRA study {[ccounted for the issues discllssed in the
reviewer's commel1t.
Table
2-3
in the report presellts the CSO volumes discharged (in millions of ga!low,) during
wet weather
sampling and the duration of the discharges. Tlte pumping stations contribute
relatively large volumes of CSOs in tlte waterway for relatively long periods of time. Therefore,
the
sampling pel.iorl11ed near tlte pumping stations during the wet weather sampling event'!
has ([ccolillted
for tlte contribution of CSOs 011 the microbial quality of the CAWS.
Also,
during wet weatlzel; additional
sampling locations were lIsed to include the entire stretch (if
each waterway segment in the s([mpling program as illustrated
il1
Figure 2-2.
The range of microorganisms studied seems appropriate, yet the number of pathogen
samples appears unacceptably low (detects in only a few
of 10-12 samples per WRP, of a
total
of 50 wet and 75 dry weather samples collected) to simply take mean estimates, rather
than predict probability density functions (PDFs)
of pathogen concentrations and their
uncertainties. Further, the use
of geometric means in the report is useful to provide an
estimate
of the central tendency of microbial concentrations, but loses information about
uncertainties that could have been achieved by describing concentrations as PDFs and
Monte Carlo sampling to estimate infection risks. Lastly, it would seem inappropriate
to
combine wet and dry weather pathogen sample results due to the few detects collected, if a
goal was to describe risks under dry versus wet conditions.
Respol1se: Geosyntec cOllcurs with the reviewer's comment t/tat the range of microorganisms
studied
seems ([ppropriate. However, the reviewer's assertion that the number of pathogens
10
appears
1Il1accept([b~v
low, is vague alld unsubstantiated. The sampling results reflect tlte
actual concentrations measured
ill
the CA WS and the WRP effluents. The analysis of 125
samples
(75
dlJ! weather and 50 wet weather
saJ11ple~)
provides a vel)' robust database of
microbial pathogens and indicatoNi.
Text in Section
5.4.1
of tlte report discusses the bootstrapping method tlUlI was used in the
QMRA. Bootstrapping
is a widely accepted and extensively used procedure in statistical
allalysis and represents
({
process of selecting a random input/rom a dataset. This technique
is useful in Monte Carlo allalysis when the exact distributional/orm of
([/1
input variable is
either unknown or unable to be represented with a continuolls distribution. Bootstrap samples
are
J'{t11dom selections from the empirical data with replacement. Bootstrap methods provide
robust estimates
of vari{[bility
il1
Monte Cario assessments as the probabilities associated with
drawing extremes in the distribution is mimicked by the presence
of extreme values in the
empirical data.
Geo~J!ntec
believes that the bootsrapping technique captured tlte variability
il1
the concentration of pathogens.
Geosyntec disagrees with the reviewer's comment that
"it would seem inappropriate to
combine wet and eli)) weather pathogen sample results due to the few detects collected,
if (l
goal was to describe risks under d/J! versus wet conditions." Table
5-9
ill the report
summarizes the total expected illnesses under
dry, wet and combined dlY (llld wet weather
events. Therefore, dry
{Ind wet weather risks were estimated and reported in tlte QMRA study.
In ({ddition, combined dry and wet weather risks Ivere estimated to represent the entire
recreational season that includes both
£II)! and wet wemlter events. It appears that ill this
comment EPA implies that the
tll)l
and wet weather results were arbitrarily combined, which is
incorrect. Tlte
(by
and wet weather results were integrated to simulate the climatic conditions
witltin a recreational season, based
011
actual weather and pumping station discharge
occurrence data as described in Sectioll
5.4.4
of the report (lnd Table
5-8.
To represent risks
from recreational exposure across tlte entire recreational season, the input pathogen
concentrations llsed in the risk assessment should {(ccount
for the probability of encountering
pathogen cOllcentrations related
to different weather conditiolls. Tlte proportion of days
under each weather condition in a recreational year (April through Novembel'l was developed
from historical records of CSO and rainfall records. The input distribution llsed in the
simulations
for selecting weather specific pathogen concentrations is shown ill Table
5-8
of
the report. A conservative lIsslimption was made in tltis analysis that recreational use (tnd
weather conditions are
110t
correlated. Common experience would suggest this is not the case
as people tend to spent/less time recreating during rain events. The assumption ill the QMRA
was that recreational use may resume shortly after rain events when waterway concentrations
are still
strong~v
influenced by the preceding weather pattel'lls.
As stated in the executive summary, the four main objectives of the wet weather
QMRA were, in summary:
1. Evaluation of wet weather impact on outfall microbial quality
2. Evaluation of CSOs impact on CWS
3. Health risk from CWS under wet weather conditions
4. Risk reduction from disinfecting WRP wet weather effluent
Points where at least
parts of these main objectives are not met are discussed below.
11
Statistical Analyses
The merging of pathogens data for dry and wet weather may be inappropriate, depending on
the question being addressed. Comments sllch as (page xxi) that
'The
Salmonella
spp. dry
weather results had statistically insignificant detections and therefore an
ANOV A analysis
of both the dry and wet weather results was not performed' are not really satisfactory, as a
non-detect means that the concentration was below a certain concentration, which could
have
been compared against the distribution of detects under wet \veather conditions.
Response: It appears that in tltis comment EPA implies
11Ult
the dry and wet weather results
were merged, which
is incorrect. The
(by
and wet weather results were integrated to simulate
tlte climatic conditions within
{{
recreational season, based 0/1 actual weather and pumping
station discharge occurrence
data as described in Section
5.4.4
of the report altd Table 5-8.
Statistical estimates may be biased in cases where (In ANOVA is conducted with highly
censored datasets. Salmonella
!.JJjJ. was detected in only
13%
of the
dry
weather samples and
therefore
(til ANOVA analysis o.ftlte results was not pelformed. However, the geometric mea11_
values for the Salmonella !.pp. censored datasets (i.e., datasets containing below detection
reslllt!.~
were computed llsing
({
maximum likelihood method. Salmonel/a !.pjJ. concentration
data witlt censoring greater than
80% were considered statistically insignificant, and therefore
no geometric mean values were computed (see Table
3-2(1 in the report). The April 2008
Report presents aI/ Salmonella
spp. results. Although, the ANOVA statistical test was not
pelformed because o.f the reasons outlined above, a direct comparison
of the results can be
pelfonned by
(IllY
reviewer o.fthe report.
One related factor that appears to be missing is the waterway recovery time, how long after
a
wet event does it take the recreational water bodies to reach 'baseline' conditions? This
raises the question as to how dry and wet weather samples periods were defined -which
does not
appear to be reported?
Re::,pol1se: Section
2.3.2,
Page
17
of the report discusses wet weather sampling protocol. In
addition, Sectiou
5.4.3
of the Report discusses tlte integration of dry and wet weather results
ill
the QMRA. Figure
5-4
presents
(111
illustration of the attenuation o.fpathogen concentrations
between wet
and dlJi sampling events that was used to derive estimates of the pathogen
concentrations between wet and
dl)1 weather events. Section
5.4.3
of the report discllsses the
estimation
([l1d incorporation of the estimates o.{ microbial concentratiolls between wet and (1)1
weather in the microbia/risk assessment.
Information regarding the analysis of pathogen samples is not sufficient. Section 3 provides
adequate details
of the raw data collected, but Section 4 summary concentration
tables/figures appear not to indicate the sample sizes involved. In Section 3, the actual
numbers of positive samples used to estimate concentration was really too low to give
meaningful values as simple means. Given all the data available, far better estimates of
means and their uncertainties could have been achieved, which could have been carried
through to the QMRA results.
Response: The reviewer incorrectly assumes that geometric mean pathogen concentrations
were used in the QMRA.
As discllssed ill Section
5.4.1
of tlte report the bootstmpping method
was llsed in the QMRA. Bootstrap
metlzod~
provide robust estimates
(~f
variability
il1
Monte
C({r/o assessments as the probabilities aSSOci(lted with drawing extremes in the distributioll is
mimicked by the presence of extreme values in the empirical data.
12
Electronic Filing - Received, Clerk's Office, April 10, 2009
Section
4
is
([
summary of information pl'esenteti in peel' review literature regarding
disinfection of pathogens in wastewater samples. Disinfection efficiency data is sllmmarized
and available pertinent information is presented in the te.xt and table footnotes. Such
information includes tlte (vpes of tests (bench- or pilot-scale) or reagellts, and reagent dosages.
l>arasitic Protozoa
Some of the 10v/ positive rates for pathogens were (from page xxi):
Dry Weather:
North Side:
Giardia
outfall (5/5), upstream (411 0), downstream (? 11 0)
Cryplosporidium
outfall (3/5), upstream (1110), downstream (6/1 0)
Stickney:
Giardia
outfall (515), upstream (4/10), downstream (8/1 0)
Crypto.sl')oridium
outfall (3/5), upstream (Ill 0), downs!ream
(311 0)
Calumet:
Giardia
outfall (4/5), upstream (011 0), downstream
(411 0)
Cryp/mporidium
outfall (1/5), upstream (I /1 0), downstream (4/1 0)
There appear to be some translation errors
or missing data, for example, in Table 3-3a there
are only five up and downstream samples reported, but in the executive summary (p xxi)
positives are reported out
of 10 samples? Presumably there was data collected for dry
weather in addition to 2005 data reported in Table 3-3? However, as Appendix C was not
included with the report (nor for that matter Appendices B-D)
it was not possible to check
against the original data provided
by CEC.
Response: The reviewer miscounted the number of samples in Table 3-3a; the table clearly
indicates that samples at the North Side oll((all
and waterway segment were collected
011
5
different dates (event5): 7/28/05; 8/4/05; 8/18/05; 8/25/05; 9/01/05. During each event, 2
upstream (stlljace and I-meter depth) and
2
downstream samples (slIIface and i-meter depth)
l!Jere collected. Therefore, a total of 10 upstream and 10 dowl1stream samples were collected
at each waterway. The reviewer's statement/question:
"Pl'esumab~}}
there was data collected
for dry weather in addition to 2005 data reported
ill
Table
3-3?"
is false. All data collected was
reported.
Mr. Louis Kollias, Director
of Research alUl Development of the Metropolitan Water
Reclamation District
of Greater Chicago, provided
(I
copy of the April 2008 Geosyntec report,
entitled:
"DIJ)
and
Wet Weather Risk Assessment of Human Health Impacts of Disinfection
1/:5'.
Non-Disb~(ection
of tile Chicago Area Waterways System," to Mr. Allen Mefcer, Water
Quality Branch,
EPA Region
5.
In Itis transmittal letter, Mr. Komas spectfically
acknowledged that tlte
/'{(w data can be made available upon request. To tltis date, MWRDGC
has
110t
received (f requestfrom EPA for the raw data.
Nonetheless, secondary-treated sewage effluent will always have some
Giardia
and
Crypto.sporidiul11
in it, and based on the 20-liter samples being processed it in unlikely to
have non-detects
if recoveries were >50%; this raises a major concern in that no recovery
nor corrections for recoveries were reported when estimating pathogen risks. Similarly, for
the environmental waters assayed for parasitic protozoa (typically 18.9 liters assayed
according to Section 3), no recovery data is presented nor corrected for.
13
Re!}pol1se: Section
2.4.3
of the report presents the qualiZF control (QC) data of all microbial
results, including
Crypto!}jJoridiu11l and Giardia QC data. The following QC samples were
analyzed for O:vptosporidiu11l alld Giardia: Matrix Spike, ongoing precision ({nd recovel}'
(OPR), aud method blanks.
The reviewer's assertion t!tat
110
reCOVel}' data is presented
1101'
correctedfor is inaccllrate (lnd
false. Section
2.4.3
of the report summarizes the recovelY rates of the MS (Ind OPR samples.
MS results were within the acceptance criteria specified in EPA Method
1623.
In addition, all
recoveries were well within the acceptance
cl'itel'irr specified for OPR samples in Method 1623.
No
oocys'-~'
or cysts were detected in method blanks analyzed indicating
110
contamination in
the spiking or sample processing procedures.
EPA Method
1623
does
110t
require or allow the lise of MS recovelJl results to adjust the
samples. Text in Section
2.4.3
of the report states that the MS sample results were not lIsed to
adjust
OJ)pto,~poridilll11
and Giardia recoveries at allY sampling 10catiol1.
Again, in the
abs~l1ce
of the original data it is hard to make any more of a comment on the
'viability' testing
of oocysts, other than to say that if only a few oocysts were examined, as
indicated by the dry weather positive counts,
it would not be appropriate to report two
significant figures for the precision
of the viability statistic reported, such as 21 % or 26 %
when the error in such estimates is likely to be at least 50%. Also, with only three of 125
Oypfosporidiul11
samples (75 dry weather and 50 wet \l>,/eather) testing 'viable' (2.4
%
as
presented on page xxiv) it calls into question how sensitive the viability assay is with so
few oocysts being assayed - another uncertainty not discussed.
Response: The reviewer misc/tal'(lcterizes and misime/prets tile results. Overall, this comment
is inaccurate
([11£1 inco/lerent. The report does not report the precision of viability. It reports
the percentage
of total cysts that (Ire viable, based 011 pl'opidiul11 iodide (PI) staining. Section
3.2.3 of the report discllsses the Giardia viability results.
Also, the reviewer misc/taracterizes the OJiptospol'idiul11 results
and refers to
3
of
125
sml1ples
testing 'viable.' In fact, the text on page xxiv refers to
'iI~rectious
foci'
110t
'viable'
Oyptosporidiu11l.
Section
3.2.2
of tlte report discusses "Detection of Infectiolls
Oypto,~poridillm Oocyst.~·
Using Cell Culture." The infectivity test for Oypto!}porifiium is
completely different than the 'viability' test.
In summary, with poor accuracy (and unreported) in parasitic protozoan viability and no
reporting
of recoveries, there is considerable uncertainty introduced into the datasets used
which has not been expressed when using and reporting risks from these data,
Re!}ponse: The reviewer's assertiol1 that tlte accuracy is poor and unreported is inaccurate alld
false. Section
2.4.3
of the report summarizes tlte l'eCOVel}' rates of the MS (lnd OPR samples.
MS results were witltin the acceptance criteria !)]Jec(fied il1 EPA Method
1623.
In ([ddition, all
recoveries were well within the acceptance criteria ,\pecifiedfor OPR samples
il1
Method 1623.
Enteric viruses
In the executive summary (p xxiv) under virus results, the terms 'enteric viruses,
adenovirus and Calicivirus' are used, presumably 'enteric viruses' should read
'enteroviruses' here and elsewhere in the report when cnteroviruses were indeed the target
group (noting concerns
if only cytopathicity was the endpoint in cell line assays),
Based on Tables 3-4, enteric viruses were assayed from 1
GO-I. samples, but no protocol was
14
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
given.
It
is unclear
if
the full 100-L concentrate was used for each of the three virus groups
assayed (i.e., 300-L collected for all virus assays), or
if 100-L was split, so in essence a
lesser volume equivalent
of the concentrate was assayed for the three different virus groups?
Given the \vay the data are presented, for example in Table 3-5, a
<I MPNIl 00 L implies
that all 100 liters
\overe assayed for each. However, since there is no protocol provided in the
report (and Appendix D was not available), one cannot determine
how the sample analysis
was performed. The concern here is data correction bias that occurs when smaller volumes
are assayed than what is reported, also
no uncertainties were presented with the MPN
values given
in Table 3-5. This concern is a major issue for the Norovirus elata, where the
PCR assay claims (Tables 3-7, 3-8)
to only have utilized some 0.2 liters of the original
water sample, but is reported on a 100-L basis. The MPN in various tables (e.g. 3-6, 3-7)
present results
\~/ith
three significant figures, far too many than what the assay can justify.
Response: Section
2.3.2.1
of the report discllsses virus sampling. Text in Section 2.3.2.1
states that appl'oximate{l' 300-L
of upstream and downstream samples were filtered at each
location during dry and wet weather sampling. 111 addition, approximately
.1
OO-L samples
were filtered at tlte outfall. The {(ctual volumes collected were recorded in the sample
collection forms ill Appendices
A-I alld A-2 of the report. Also, Appendices B-1 and B-2 and
D-1 aud D-2
of the report include the laboratol)} bench-scale forms that indicate the sample
volumes al1alyzed
for vi1'US samples. Appendices B-1 and B-2 include the total cuiturable
enteric virus results by Hoosier Microbiological Labora/ol}}. Appendices D-1 and D-2 include
the adenovirus and
l1ol'ovirus
results by the University of Arizona, The reviewer's assertion
that Appendix D was
110t available is incorrect. MI'. Louis Kollias, Director of Research ([11d
Development of the Metropolitan Water Reclamation District of Greater Chicago, provided a
copy
of the April 2008 Geosyntec report, entitled: "Dry and Wet Weather Risk Assessment of
Human Health Impacts of
Disb~rection
Vs. Non-Disinfection of the Chicago Area Waterways
System,"
to Mr. Allen Melcer, Water Quality Branch, EPA Region 5. In his transmittal letter,
Mr. Kollias specifically acknowledged titat the
f'((W
data can be made (Ivailable
lipan
request.
To this date, MWRDGC has 110t received a requestfr0l1l EPA for the raw data.
In (lddition, the reviewer's concern t!tat only 0.2 Liters of sample was utilized for llorovirus
analysis
is unjustified. The volume of 0.2 Liters of sample analyzed is significantly gre((ter
than EPA's estimated water ingestion volume
for swimmers of 30mlllnd significantly greater
of the incidental ingestion volumes for the recreatioltal lIses considered in this microbia/risk
assessment including, boating, canoeing ([l1dfislting (see Section 5.2.2
of tlte report, Rtposllre
Inputs).
In the PCR assays used, as no method data was available, it is unknown what level of
amplicon confirmation was used, e.g. was sequence confirmation undertaken, probing or
none? For cell lines showing a cytopathic effect (e.g. PCLlPRF/5 for adenoviruses) on
Table 3-6, footnote 1 states that only
31 of the 42 virus infected cell line samples were
confirmed as adenoviruses by PCR. Hence, was the adenovirus
MPNIl OOL adjusted on that
percentage'?
It appears that the total MPN value was simply translated into adenovirus MPN
'vvithout any adjustment given the same values presented in Tables 3-6 and 3-8 (and only
42/50
PCR confirmed in Table 3-8).
Re:,pol1se: Tlte reviewer's comment is incorrect. First, there was
110
a(/justment
011
tlte
adenovirlls concentration based
011
the ratio (31142) of samples that were confirmed as
adenoviruses by PCR. For tlte samples with PCR confirmation of aden a viruses, tlte total
concentration
of sample was assumed to be adenovil'Us, which is a conservative assumption
for the risk assessment.
15
Electronic Filing - Received, Clerk's Office, April 10, 2009
The summary enteric viruscs data Tables 3-9 & 3-10 ha\/c
l~\r
too many significant tlgures
given
the lack of precision in the assays used along with the data management issues
associated with the actual volumes assayed versus the
100-I.,
reported volume (sometimes
four significant figures are reported, when
1-2 are all that can likely be justified). Overall,
the outfall concentrations
of enteric viruses reported appear low, particularly for a 110n-
disinfected wastewater, compared to what has been published in the literature. Based on the
E. coli
& fecal coliform concentration data (Table 3-1), thc wastewater seems to have only
lost about 2 logs through treatment as expected
fl'om normal raw sewage. lIenee, virus
numbers seem
to be some orders of magnitude less than expected for undisinfected
efi1uents, which has potential significant ramifications
for disinfection studies and risk
assessments using this data.
Re!Jponse: The virus analytical results under both dry and wet weather results amlfrom two
different laboratories (HML
and UA) indicate that the virus concentrations are vel))
low. TI,e occurrence and cOllcentration of protozoa, cuiturable virllses, adellOvil'lIses
and 1101'0virllS were generally equal to or lower than observed in other studies by Dr.
Gerba and others 011 wastewater discharges and slI1face waters ill general during dry
weather conditions (Gerba,
2008; Rodriquez et al., 2008; Rose et al.,
1988, 1991,1996).
These studies involved both disinfected and
11011-disb~fected
treated wastewater, ([l1d
streams into which they were discharged. Some of these studies were conducted in
Europe where disinfection
of treated w((stelv(tter discharges is Ilsually 110t practiced.
The eliltumble viruses
J'vel'e
also lower than observed in a study of a recreational
stream in Arizona cOlUlucted by
Dr. Gel'ba's laboratolJ) il1 which bathers were the on(}}
source (Rose at al.,
1987).
The Geo.5yntec Team, including Dr. Gerba, believes that
the results are representative
of the CAWS.
Referel1ces
to this response:
Gerba,
C.
P. 2008. Virlls occurrence and survival in lite environmental waters.
111:
Human
Viruses in
Water. A. Bosch,
ed. pp. 91-108. Elsevier, Amsterdam.
Rodriquez, R. A.,
P. M. Gundy (tnd
C
P. Gerba. 2008. Comparison of BGM and PLCIPRCI5
cell lines for total cultul'able viral assay of treated sewage. Appl. Environ. Micro bioI.
74:2583-2587.
Rose, J.B., R.L. Mul/inax, S.N. Singh, M.
V.
Yates, and
C.
P. Gerba.
1987.
OCCllrrence of
l'otavil'lIses and entel'ovil'lIses ill recreational waters of
Oa/(
Creek, Arizolla. Water
Research
21 :1375-1381.
Rose, J.B., CPo Gerba altd
ij~
Jakubowski.
1991.
Survey of potable W{{tel' supplies for
O:ppto.\pol'idilll11 and Giardia. El1virol1. Sci. Technol. 25:1393-1400.
Rose, J. B., L.
1.
Dickson, S. R. Farrah alld R. P. Crrrnalwil.
1996.
Removal of pathogenic
and indicator micI'oogallis111s by full-scale water reclamation facility. Water Res.
30:2785-2797.
Smith, H.
V.
and A. M. Grimasol1. 2003. Giardia and OJlptosp0 ridiul1l. The Handbook of
Water and Wastewater Microbiology. D. MaJ'{[ and N. Horan. pp.
695-756.
Elsevier,
London.
16
Electronic Filing - Received, Clerk's Office, April 10, 2009
Disinfection
The potential disinfection effects
of ozonation, UV and chlorination given in Table ES-l are
generally lacking any ranges - so again minimal uncertainty has been assigned to these
data. Furthermore, actual efficacy under operating conditions would be expected to increase
the range in performances
of these unit operations.
In summary, the disinfection chapter docs not actually
present operational data nor
performance ranges required to undertake a sensitivity analysis
or thorough risk assessment
- hence it adds little
to the document.
Response: Tlte reviewer's comment is incorrect and provides
(111
unjustified (tnd unfair
criticism
of the disinfection section (Section
4)
of the report. Section
4
is a sumnulIY of (In
exhaustive literatllre search and provides
if~form{/tion
presented
ill
peer review literature
regarding disinfection
of pathogens in wastewater samples.
Disb~fection
efficiency data is
summarized and {Ivai/able pertinent information is presented in the text
and table footnotes.
Such information includes the types of tests (ben clt- or pi/oj-scale) or reagents, and reagent
dosages. The il{(ormation was lIsed
to derive a range of expected pathogen disinfection
effectiveness lIsing Uv, c11/0I'inatiol1ldech/orination and ozonatioll. No treatability studies
were conducted as part
oftlte QMRA study to determine site-!J]Jecific disinfection effectiveness.
Microbial Risk Assessment
Given the above comments, it is clear that the intended microbial risk assessment was
largely focused at
what would be called a screening level largely using point mean
estimates in a deterministic manner.
Response: The reviewer's comment is grossly inaccurate. The QMRA did Itot use mean
estimates. Text in Section
5.4.1
of the report disclIsses tlte bootstrapping method that was used
in tlte QMRA. Also a probabilistic, /lot a deterministic methodology llsing distributions
of
exposure parameters was lIsed
il1
tlte QMRA. Section
5.2
discllsses the methodology used
Yet there are some surprising attempts to incorporate some elements of a stochastic
assessment,
such as in the PDF describing ingestion rates Cfable 5-4). No reference is
provided to
j
lIstify either the values presented in Table 5-4 nor the precision implied by the
number
of significant figures presented (generally three, sometimes four).
Response: The reviewers comment is false. The reviewer reluctflntly acknowledges that QMRA
has elements
of
([
stochastic assessment, but calls them "surprising." It is
110t
clear wit at that
characterization refers too. Tlte reviewer claims that there are
no references for the
i1~(orm(ftioll
presented
011
Table
5-4.
11tis statement
is
incorrect. Table
5-4
summarizes the
illformation di\'cussed in detail
il1 Section
5.2.2,
where multiple references (fre presented.
rt is stated (bottom of page 130) that a one-dimensional probabilistic risk assessment was
undertaken (i.e., taking on board variability, but not also uncertainty). However, as stated
above,
PDFs do not appear to have been utilized in describing pathogen concentration
variations; indeed, it is unclear to this reviewer what all the assumptions are as they appear
not to be listed.
For example,
(1)
were median values or averages used? (2) what standard
deviations and
assumed distributional forms were used or each PDF or how were
parameters fitted for
each PDF? (3) ho\"" were viability estimates incorporated into the
results? (4)
if 'normal' pathogen loads in raw sewage 'were used and their dilutionlremoval
was based on
E. coli
or other indicators in stream waters - how vvould that change the
17
estimated pathogen ranges? (5) what ranges were assessed in the sensitivity analyses and
on what basis were they selected? and (6) how many iterations Vicre undertaken in the
Monte Carlo simulations?
The only PDFs for input parameters appear to be ingestion
volume (from Figure 5-2, which has no source identified as to where these numbers come
from) and canoeist duration activity (Figure 5-3).
Response: Tlte reviewer's comment is inaccllrate. Tlte QMRA did 110t lise mean estimates.
Text in Section
5.4.1
of the report discusses the bootstrapping method that was used in the
MRA. Also a probabilistic, not a deterministic methodology using distributions of exposure
parameters lVas used in the QMRA. Sectioll
5.2
discusses the methodology used. Table 5-4
S1l11111Utrizes tlte inforl11atiol1 discussed in detail
il1
Section
5.2.2,
where multiple references are
provided. Also, tlte reviewer
is asking the /lumber of iterations used. Section
5.4.5
of the report
discusses the !lumber
of simulations used. Specifically, text on page
126, 101"1
Paragraph
indicates that 1,000,000 iterations were pelformed.
Furthermore, there are various key questions not addressed in this assessment, such as:
.
What were the risks during wet weather alone (to take a worst case scenario)
given
it was not noted how long it takes to return to 'baseline' conditions? Rather
than using some mix (Figure 5-4) to estimate pathogen concentrations between
\vet and dry conditions, and not even using any variability
of that in the
assessment.
Response: Table
5-9
clearly presents the wet weather risks
If method recovery was included for each of the pathogen groups, what would be
the implications to the estimated risks?
Response: Method reCOVeJ)i correction is not required or ((lIowed ill the EPA-approved
method., lIsed
for the analysis.
Therefore, it is not scientifically defensible to derive
speculative estimates that are
not based
011
proven
t
validated methods.
•
"What about sediment load of pathogens and resllspension of those to added risk?
Response: The sampling accollntedfor sediment re-smpension of pathogens. Section
2.3.1
of the
report discusses sediment re-sw.pension due
to barge trajjic and sampling when these conditions
occurred.
What levels of indicators could be predictive of 'safe' recreational waters.
Response: This assessment was olltside the scope of tlte QMRA. The CHEERS (Chicago
Health Environmental Exposure
&
Recreatiol1 Study) being conducted by the District will
answer this comment.
18
Electronic Filing - Received, Clerk's Office, April 10, 2009
Electronic Filing - Received, Clerk's Office, April 10, 2009
no _ ZU'II F9
n
....
vz
'" _
* ... _
Metropolitan Wafer Reclamation District
of
Greater Chicago
100 EAST ERIE. STREET
CHICAGO,
ILLINOIS
60611~S1S4
312-751,5600
Louis Kollias.
p.e .. ace!:
170059838137
T~~
J.
O'Dn..n
p~t
Kathl;cln
Th~
MBIl11)'
VICO
Piesfdellr
GlQI'lQ AIi!tO.
MJj~~1
C/l31M/AA
01
F/lmlt~
FIlIIIk Avlla
Pam;!;aHolWn
Bmbars
01, McGDW8n
Cynll'lla M.
Sanl""
OubraSt!vRl
Pllllioia Young
P.01/28
Olrecfof of
Rese~t'dt
and DeveTopment
31~'7S1'5190
May 1S
T
2008
FILE COpy
Mr. Allen Melcer
United
States Envjronmcntal Protection Agency
Region V. Water Quality Branch
77 West Jackson Boulevard
Mail Code:
WQ~16J
Chicago, n.. 60604-3507
Dear Mr. Melcer:
~
(
Subject: Final Report Entitled "Dr)' and Wet Weather Risk Assessment of
,
Human Health Impacts of Disinfection vs. No Disinfection of the
oS
Chicago Area Waterways
System," and Response to COll1.l!lents on
d
Interim
Draft Report
~
!f
_J
-
-;S ..>....,.,
The Metropolitan Water Reclamation o;strict of Greater Chicago (District) is pleased to
~ ~
-1 provide you the final report entitled
"Dry
'and Wet Weather Risk Assessment of Human Health
~
Y:5- Impacts of Disinfection vs. No Disinfection of the
~ica?o
Area Waterways System (CAWS)."
~ ~
t
The report was prepar.ed
by
the Geosyntec team whJch Includes Geosyntec Consultants; Cecil
~ ~-",:r
Lue.Hing
&
Associates; Dr. Charles Gerba of the: University of Arizona; Hoosier Microbiology
~
1<11
~
Laboratory; and Dr, Jennifer Clancy
of
the Clancy Environmental ConsulLants Inc. The District
.
~ ~ ~
is confident that the microbial risk assessment performed by the Geosyntec team represents the
;ii
W
best effort the current state of 'the
scienc~
can p!ovid.e. The report acknowledges
uncert~inties
..
~ ~
D
that are inherent in any risk assessment methodology. To address these uncertainties and to
'-2;j
'validate the microbial risk assessment report, the District has embarked
0))
a companion
\.C Q
epidemiological study to ascertain health impacts of recreational use of the CAWS.
One paper copy of the report is enclosed. The raw data are not included
in
the final report
and can. be made available upon request,
In
addition, a copy of the final report is posted on the
District website (www.rnwrd.org)andforconvenientaccess.click on "UAA Study" listed under
"Public lnterest," and then click on the eighth bullet. Also attached to this letter is a copy of the
itemiz.ed responses to your
comxmmlS
dated March 20, 2007 on
the
Interim Draft Report. The
comments
Were
reviewed by
'Lh~
Geosyntec team and the responses to the comments presented
_reflect the changes made
t<)
the final document. We very much appreciate the reviewers' time
anc;l efforts and have found their comments useful in imprOVing the quality of the final report.
,.
r
!
i
i
I
I.
I
I
I
;
;
,.
I:
i
I.
I.
I
! .
i
I
I
I
I
I-
I
I
I
1
I,
I
i
i
i
I
!
~
! ,
Electronic Filing - Received, Clerk's Office, April 10, 2009
.~EP-03-20as
14:19
RESRCH STI CKNEY
17138588381217
P.132/28
Mr.
Allen Melcer
2
May 28, 2008
Subject: Final Report Entitled ''Dry and Wet Weather Risk Assessment of
Human Health Impacts
of
Disinfection vs. No Disinfection of the
Chicago Area Waterways System," aJld Response to Comments on
Interim Draft Report
We would like to thank you all for your valuable contributions to this report.
If
there are
any questions, please feel free to contact Dr. Thomas Granato, Assistant Director of Research
and Development, Environmental Monitorb;tg and Research Division, at (708)
588~4059
or
.
e-mail Thomas,GranatQ@mwrd.org.
~tr
LK:TG:GR:rag
Attachments
. cc: R. Lanyon
F
.. Feldman
W. Stuba
.
T. Granato
G. Rijal
62;1
lid az
).Vli
BOOl
o
'i
~
.:JO
• Hill
yery truly
youI;S,
~0tI~~
Louis Ko1lias
Director
Research
and
Development
I'
i
1
i
1
I
I
I
I
I.'
I
I
I.
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
Geosyntec{)
55
We~(
Wacker Dl'ivc
SlIile \\00
Cb;engo. II. li0601
consultants
PI! :1I2.G5H-0500
,'" :1l2,G5U-1l57G
WW\v.gc()synwe,com
Via E-Mail and U.S. Mail
23 May 2008
Dr. Thomas C. Granato
Assistant Director of
Research
&
Development
Metropolitan Water Reclamation District
of Greater Chicago
6001 W. Pershing Road
Cicero, Illinois 60804-4112
Subject:
Responses to EPA's Technical Review Comments Regarding the
Interim Phase I Report; dated November 2006, "Dry Weather Risk
Assessment of Human Health Impacts of Disinfection Vs. No
Disinfection of the Chicago Area Waterways System"
Dear Dr. Granato:
Geosyntec Consultants (Geosyntec) is enclosing responses to EPA's technical review
comments
regarding
the subject report. Geosyntec's responses refer to the April 2008
Final Report entitled, "Dry and Wet Weather Risk Assessment of Human Health Impacts
of Disinfection Vs. No Disinfection of the Chicago Area Waterways System," (Final
Report),
which is incorporated to the responses by reference. The responses follow the
corresponding EPA comment(s),
If you have any questions 01' comments regarding the enclosed report please call me at
(312) 658-0500.
Enclosure
Very truly yours,
/./
/)
/
.-"
:f
(//1/7
J:tJ
/
lh",,:?}7
C':-1--~<-:-r-.....
Clu-iso Petropolllou, Ph':D., P,E., BCEE
Associate
j.
I
I
I
I
I
i
i
!
;.
ENCLOSURE
Responses to EPA's Technical Review Comments Regarding the Interim Phase I
Report, dated November
2007~
"Dry Weather Risk Assessment of Human Health
Impacts of Disinfection Vs. No Disinfection of the Chicago Area
Waterways System"
Dry
Weather Risk Assessment of Human Health Impacts of Disinfection vs. No
Disinfection of the Chicago Area Waterways System
Review conducted for: US EPA Region 5, Office of Water,
Review conducted by: US EPA Office of Water, Office of Science and Technology
NOTE: In an effort to avoid duplicalion, these points are in addition to comments sent
by
ORD
already.
OSTIHECD
agrees with ORD's comments.
Summary:
A
Quantitative Microbial Risk Assessment (QMRA) of the Chicago Area Waterways (CAW) was
conducted to evaluate the risk of illness posed to recreational users ofthe CAW with the current
practice of not disinfecting the effluent atthree wastewater treatment plants with discharges into
the CAW. Using monitoring data for pathogenic
m~croorganisins
and integrating over dose
response functions, exposure times and ingestion rates, the conclusion was made that the risk for
gastrointestinal illness was well under the
8~101l000
currently .deemed "acceptable!" by the US
EPA 1986 Ambient Water Quality Criteria, an.d thattht;!re was therefqre no need for additional
disinfection to adequately protect public health.
This QMRA was only done for the Phase I "dry" weather season, and does not present results for
the wet season. So presumably any conclusions would be only applicable to the
dry
season until
the wet season analysis is completed.
Response: We
conCUr
with the. reviewer's
comment.
The Interim Rep()ri summarizes the dry
weather microbial risk assessmflnt results am/any conclusions are only appiicable to the
dry
season.
However, the April 2008 Final Report entttled; "Dry and Wet Weather Risk
Assessmellt of Human Health Impacts of Dis inJection Vs. No Disinfection of the Chicago Area
Waterways
System," (Final Report) integrates both the dry and wet weather microbial risk
assessment results in a comprehensive outcome;
Health and Ecological Criteria Division
•
Introductory material biases risk
assessment
A few statements made in the Introduction were either opinion or unsupported fact (e.g.,
page 2, paragraph 2: The
year~round
implementation of chlorination ..... ). There is no need to
focus on chlorination, since there are alternatives available. No citations were given to support
these upfront conclusions. Additionally, there is no mention of the benefits of disinfection of
human sewage effluents, chlorinated or otherwise. Mentioning this in the introduction as it is
serves only to bias the reader.
Response: The report includes the following citation for the statements made:
."Metropolitan Sanitary District of Greater Chicago (MSDGC), 1984, Wastewater
Disinfection: A Review
of Technicala.nd Legal Aspects in Illinois. Department of
Research and Development. Report No. 84-17. July."
However
t
the above-mentWned paragraph has been removed from the IntroductWn of the
Final Report. In addition, a section has been added (Section
4) in
the Final Report tllat
provides a comprehensive overview
of disinfection technologies, includilig:
(1)
chlorinatiofzJdechlorination,
(2)
ozona/ion, and
(3).
UV. Advantages and disadvantages of each
technology are discussed, including disinfection effectiveness, and disinfectWn by-product
formation.
Another example: page 3, paragraph 3, The CWS is not a coastal recreation water. This
statement follows evidence for increased and encouraged use of the waterways for recreational
activities. Wbilethe CWS is not, by definition, a coastal recreation water, it is a 'water of the
United States' as defined by the Clean Water Act.
Response: We conCllr with the reviewer's comment; The subject sentmee has been removed
. from tlteFinat Report.
•
Data presented.are for dry weather only
The risk assessment's main conclusion that the risk for GI illness was well under EPA's
recommended 1986 recreational A WQC is a bit premature given that no wet weather datil was
available atthe tiine this report WitS puhlished. Rain events can be a major driver forinflux of
microbes into a surface water body, so until the wet weather data is analyzed, any broad sweeping
conclusions in this report should be taken in context.
Respo1l$e: We con(;ur with the reviewer's comment. The Interim Report summarizes the dry
weather data only. However., the Filial Report integrates both the dry and wet weather data in
.
a comprehensive outcome in thernierohial risk assessment.
•
Enterococcus etlUrrteration method: most appropriate?
The author's used EPA method 1106.2 to enumerate Enterococcus. Method 1600 is the
re.commended method to use for this purpose.
Response: At the tjme of the planning and implementation of the study, EPA Method 1106.2
was
the EPA-approved methodJor Enterococcus.
•
Risk assessment lacks necessary components
While this report contains a fair amount of 'upfront' material, there is a concern over the
lack of a coherent problem formulation. This would include a listing of paranleters evaluated in
the assessment and why each parameter was chosen. A range of estimates with the rationale for
picking one deterministic point over another would be helpful.
Response: The 2006 Interim Phase I Dry Weather Report has the information mentioned in
the reviewer's comment. This information is also included in SectWn
5,
of the Final Report.
More specifically, Section
5.2
of the Final Report discusses
ill
detail tlte parameters evaluated
as part of the exposure assessment, including:
(1)
waterway use and receptor group
categorization and
(2)
exposure inputs. Tlte rationale for parameter selection is also provided.
Also, the exposure input parameters used were based on distribution functions and not single
deterministic point values. SectWn
5.2.2
of the Final Report discusses in detail the types of
exposure input distributions that were used to develop estimates for the following parameters:
(l)incidental water ingestWn rates and
(2)
exposure duration. In addition, Sectum
5.3
of tile
2
Electronic Filing - Received, Clerk's Office, April 10, 2009
Final Report provides the basis and rationale for the selection of dose response parameters
used in the microbial risk assessment
analysis for each of the pathogens of concern, including:
Enteric Virus,
Calicivirus, Adenovirus, pathogenic E. coli (estimated), Pseudomonas
aeruginosa,
SalinoneUa, Cryptosporidium and Giardia.
.
Also, this impaots the Jack of a sensitivity analysis nientioned by Tim Wade. In order for this
report
to impart confidence in its conclusions, an effort
to
spell out each parameter and the
rationale behind that choice would
be welcome (e.g., why choose the pathogens they did). Given
the propensity
for choosing assumptions that minimize risk at each step of the risk
a~sessment,
more credibility would be gained by also stating why those assumptions were chosen.
Response: Section
5.4.7
of the FinatReport includes a detailed discussion regarding Sensitivity
and Uncertai.nty .Analysis.
The. sensitivity a1l(llysis was performed. to identify the contribution
oj eachinputdistrihutwn to the variance of tke resultingrlsk estimates. In addition,
uncertailltyfactors aml their impact
in the risk estimates are clearly identified and discussed.
Also, for the sake of clarity: fecal coli forms,
E.
coli
and Enterococci are NOT pathogens. All
three are fecal pollution indicator organisms. Theygiye
110 direct evidence.of the presence of
pathogens. While there are pathogenic strains
of
E. coli,
these strains are not enumerated by the
method used.
Response: We agree with the reviewer's comment about fecal colifarms, E.coliand
Enterococci. The analytical results afthese bacteria were only used to characterize the
miCrobitll quality of the waterway. The microbial risks of the waterway were estimated based
on
bacteriapathogens,vi~se$,
and protozoa. Although strains of pathqgenic E. coli were not
determined during this
study, we relied .on results published in the/cchnical literature and
made conservative
asSU1tZptions
to
estimate the percent the pathogenic E. coli as a percentage
oj the totaiE. coli detected. Seetion S.3.4 of the Final Report includes a detailed discussion
regarding the
dose response of p(1thogenic E.coli (estimated), Pseudomonas aerugilWsa,
Salmonella,. Enteric
Virus, Calicivirus,Adenovirus, Cryptosporidium and Giardia.
•
Indicator correlations are not appropriate
The authors state that they attempted
to identify a correlation between fecal coli forms and
other pathogen concentrations
(page 33, paragraph 3). If this correlation could be discerned, then
the historic fecal coliform concentration data could be extrapolated to generate concentration
statistics for other pathogens. This
is highly inappropriate and takes up a fair amount of the
report. Fecal indicator bacteria, such as the fecal coliform group, only indicate the presence of
fecal pollution. They
do not indicate the presence of pathogens; that has always been an
inference. Additionally, fecal indicator bacteria do not correlate with pathogen loads, only fecal
pollution loads.
GiVen the myriad
of
potential fecal pollution sources listed in the report, each
with a different spatial
and temporal influx to the waterways, the indicator to pathogen ratio
would be quite variable
and would be difficult to elicit based on five sample points over a six-
week period.
One would expect a correlation between
E.
coli
(as measured in this report) and fecal coJiforms,
since
E. coli
is a subset of the fecal coliform group. Thi$ would be different if one were
enumerating the toxin-producing strains
like
E.
coli
0157:H7, which are not necessarily
enumerated
by the method usedin this report. Also, the con'elation of Enterococci and fecal
coliforms would also be expected since both are of fecal origin and excreted by warm-blooded
3
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
animals. Given the source of these organisms here, it is no surprise that as the concentration of
one increases, so does
the other.
Response: We agree with the reviewer's comments that during dry weather there is poor
correlation
between indicator bacteria and pathogens. However, the ultimate purpose of the
analysis was to determine correlations between pathogens and imlicators under both dry and
wet weather conditions in order to ascertain iftJie weather or any other factor can affect such
correlations.
The~·taiistical
correlations between bacteria pathogens and indicators have been
removed from the body oj the report and are included in Mtachment A oj the Final Report.
TIre statistical analysis in Appendix A indicates that the cQrrelati(Jn oj bacteria in wet weather
samples is statistically more significant compared to dry weather samples.
•
GI illness as the sale endpoint of ri"sk
This isa major weakness in the risk assessment. On page 90, paragraph I, the authors
state that
GI illness is the principal adverse outcome associated with exposure to
microbiologically contaminated water. This
is not necessarily true. As noted byORD in their
epioemioJogical studies, the greatest correlations are noted between
fecal indicator concentrations
and
Gl illness rates, but that does not mean that other endpoints and other metrics are not just .as
viable. Inhalation is another major route of infection, but is somewhat poorly correlated to fecal
indicators (which are of GI origin). Pseudomonas and adenovirus were found, so the authors
should have explored the inhalation route
to properly examine the risk associated with recreating
on this water.
If
there was a problem formulation, then the various ro.utesof exposure could have
been discussed and compartmented for risk analysis. Canoeists, boaters,
jet skiers, etc. an are
affected
by this route of exposure. AlSO, respiratory illnesses can he easily transmitted to other
persons.
Response: Section
5.1
of the Fiilal Report descrilJes.in detail th.e Hazard Identification
component
of the microbial risk assessment study. As stated in this sectiOn, exposure
to microbial contaminated' water may result in both gastrointestinal and non..,
gastrointestillalillness. However, there are no kn(lwn dose resp()nse models for the
non-gastrointestinal exposure routes. The risk of gastrointestinal illlteSS was selected
as the sentinel effect for conducting
the quantitative risk assessment. However, non-
gastrointestinal illnesses
were addressed qualitatively. Section
5.4.6
of the Final Report
presents a qualitative assessment of the non-G! risks associated with Pseudomonas
4eruginosa.
While I have no data at hand to properly discuss this point, there is a notable lack of
. discussion of the
food intake route of exposure. Given the levels of fecal pollution in this
waterbody and the fact
the authors discuss increased fishing on the waterways, I wonder whatthe
fish intake route would
add to the overail risk. Is there evidence for pathogen concentration in
fish tissues here?
If
this were a chemical contamination issue, these additional exposure pathways
would be included
in the toxicological analysjs.
Response:
Fish
consumption was not part oj this microbial risk assessment study. Pathogens
present
ill
the fish 'would most likely he destroyed during ihecookingprocess. Also, fish
consumption
is.
typically regulated withjis.h advisories.
4
Electronic Filing - Received, Clerk's Office, April 10, 2009
Overall, this risk assessment does not do an effective job at presenting the actual risk of exposure
to undisinfected sewage effluerit present in the CAWs. More transparency would aid the reader
in
the confidence of the conclusions.
Response: We believe that we have conducted a very comprehensive systematic study to
characterize the microbial quality and associated risks of the CWS, under both dry and
wet weather conditions. The samples
were collected and analyzed during the
recreational season, over a
two~year
period;
dry
weather samples were collected during
the Z005 recreational season
and wet weather samples Were collected during the 2006
recreational season. This study focused
all
the detection of microorganisms typically
present in the feces
of humans and other
warm~blooded
animals as indiCators of fecal
pollution. Hence, a group
,of
EPA~approvedindicator
microorganisms, such as E. coli,
enterococci, and fecal coliform was selected for this study. In addition to the indicator
microorganiSms, pathogens representative
of those present in the wastewater that are
also
oj public health concern were selected.
Overal~
one hundred and twenty five
(125) samples were collected and analyzed during the dry and wet weather events.
Risk assessment inputs were drawn extensively from
site~specifu:
dgta an.d were
developed using state-of.the-science methodology to accurately represent recreational
user exposure conditions and risks.
RecreiitioiUlI survey studies were used to provide
insight on the types and frequency
of recreational exposure expected in the watenvay.
For quantitative risk analysis, the
.DM$wdy
w~s
used as the primary source for
exposure lise datafor the CWS. Exposure parameters were developed as distributional
parameters for each ret;eptor scenario
.~
inputs.. to the exposure model, These
parameters in.clude incidental ingestWn rates and exposure duration. Selection
of
input distributions relied on literature derived sources, site-specific use information
and professional judgment using conservative assumptions. Dose-response data was
developed from regulatory documents, industry white papers
and peer reviewed
literature. Concentrations
of pathogens in the waterway were selected for each
simulation jrom the entire dataset
of dry and wet weather samples collected. The
proportion
oj dry and wet weather samples utilized were weighted to account for the
proportion
of dry and wet weather daysina typical Chicago recreational season.
5
I
I
I
Dry Weather Risk Assessment of Human Health Impacts of Disinfection vs. No
Disinfection of the Chicago Area Waterways System
Review conducted for: US EPA Region 5, Office of Water,
.Review conducted by: US EPA Office of Research and Development
Summary,
A Quantitative Microbial Risk Assessment (QMRA) ofthe Chicago Area Waterways (CAW) was
conducted to evaluate the risk of illnes's posed to recreational users of the CAW with the current
praCti~
of not disinfecting the effluent at three wastewater treatment plants with discharges in,to '
the CAW: Usingmonitonng data for pathogenic microorganisms and integrating over dose
response functions,exposure times and ingestion rates, the conclusion was made that the risk for
gastrointestinal illness was well under the 8-10/1000 currently deemed "acceptable" by the US
EPA 1986 Ambient Water Quality Ctitena, and that there was therefore no need for additional
disinfection to adequately pI:otect public health
:rhisQMRA was only done for the Phase I "dry" weather season, and does not presentresultsfor
the wet season. So presumably any conclusions would be only applicable to the dry season until
the wet season analysis is completed.
Respo.nse: We concur with the reviewer's comment. The Interim Report summarizes the dry
weather microbial riskasseSsme"t results and any concluswllS are only applicable to. the dry
seaso.n.
However,
tI~e
April 20()'8 Final Report entitled, "Dry and Wet Weather Risk
AssessmentofHurMn Health Impacts o.f Disinfection Vs. No. Disinfection o.f the Chicago. Area
WaterWays System," (Final Report) integrates hath the dry and wet weather microbial risk
assessment results in acomprehensiveoutco.me.
National Health and Environmental Effects Research Laboratory (NHEERU:
Note: This lab's review ctoesnot assess in detail the adequacy of the microbial methods, QA
procedures and sampling techniques.
Comments:
The QMRA was conducted by a consulting group, GeoSyntec Consultants, based in Chicago,
with analytical assistance from Dr. Charles Gerba at University of Arizona, and Dr. Jennifer
Clancey of Clancey Envirolllnental, among others.
The microbial sampling and characterization Seems thorough and adequate. World-renowned
experts were <:onsulted and retained to conduct the analyses for pathogenic microorganisms and
details of the sampling scheme. rationale and methods are well described.
The general approach described for the QMRA also seems appropriate. The authors do a
thorough job ofexpJaining and justifying their selections of dose-response functions and their
parameters. Generally, citations from peer reviewed literature are provided to support their
decisions.
However, there are some fundamental problems in the application. presentation and interpretation
of the results of the QMRA. These are detailed below:
Electronic Filing - Received, Clerk's Office, April 10, 2009
•
No justification was provided for the organisms measured or pathogens
considered in the QMRA
•
The risks presented are only for a few gastrointestinal pathogens. Risks were not
presented for Hepatitis
A, Shigella, Camplyobacter, to name a few. Therefore
risks presented
will be biased low.
Response: Section
2.1
of the Final Report presents the rationale for indicator and
pathogenic microorganism selection. This study did not account for all pathogens that
may be present in CWS recreational water. This study focused on
the
detection of
microorganisms typically present
ill
the feces of humans and other
warm~blooded
animals, as indicators offe(;(ll pollution. Hence, a group of EPA-approved indicator
miCroorganisms,
such as E. coli, enterococci, and fecal colifonn was selected. In
addition,pqthogens representative of those present in the wastewater that are also of
public health concern were selected. Table-2-1 in SectiOn
2.1
of the Final Rep_ort
presents asltmmaryof the microorganisms selected for this microbial risk assessment
study and rationale for their selection. The rationale for selecting the pathogens jor
-lhismicrobial
rlskasses~me1it
study.included the Jollowingcriteria:..
• The pathogens selected are. associated
with
documented outbreaks of disease,
including
gastrointesti1Ul1 andr-espiratorydiseases and injections
•
There. are EPA-approved
method~or
laboratory standard operating procedures
(SOP$) available for the measurement
of the selected pathogens.
•
Only gastrointestinal illness was considered
Since
Pseudomonas
and adenovirus were found, descriptions of non GI Illness should
also
be provided to present a clear picture of the actual risk associated with recreating in
. [he CAW
Response: Section 5.1 of the Final Report describes in detail the Hazard Identification
component
of the microbial risk assessment study. As stated in this se{;tion, exposure to
microbial contaminated water may result in both gastrointestinal and non-
gastrointestinal illness.
However, there are no known dose response models for the
non-gastrointestinal exposure routes. The risk Q/gastrointestinal illness was selected
as the sentinel effect for conducting the quantitative risk assessment. However, non-
gastrointestinal illnesses
were addressed qualitatively. Section
5.305
of the report
discusses the dermal risks and eye and ear infections caused by Pseudomonas
aeruginosa. Although Pseudomonas aeruginosa is not a pathogen that is linked to
gastrointestinal illness, this pathogen has been linked to recreational illness outbreaks
involving dermal (folicutitis),
eye, and ear (otitis extemia) injections. For this reason,
the levels of Pseudomonas aeruginos(l wereevaillated under the sampling program/or
this risk assessment. However, quantitative evaluation
of/he risk jor this pathogen is
problematic. There are no published dose-response relatlonships Jor Pseudomonas
aerugillosa. Without a clear dose-response relationship there is no
way to establish the
expected illness level associated with any particular waterway concentration. The
2
Electronic Filing - Received, Clerk's Office, April 10, 2009
dermal pathway for estimating exposure to Pseudomonas aeruginosa is also
problematic.
Ear and eye
infections associated with contact
by Pseudomonas
aeruginosa contaminated water are typically associated with full immersion activities.
Since these types
of activities are not permitted or designated uses of the CWS the
incidence
of ear and eye exposures are expected to be low and as the result of
accidental or intentional misuse of the wateni'ay. Pseudomonas related foliculitis
commonly requires a break in the
skirz from a preexisting cut, open sore or scrape as
an entry point for infectum. Immunocompetent individuals without skin abrasions
rarely develop foliculitis by exposure to intact skin. For these reasons, a quantitative
evaluation
of risks is notfeasible.
Section
5.4.6
ojthe Fin(J.l Report presents a qualitative assessment of the non-OJ risks
associated with
Pseu£kJtlUJnas
aeruginos~
.
•
Conservative assumptions were not.made
In nearly every case, when simplifications and assumptions were made in such a way to
ultimately minimiZetheestimated,risk.
Response: We believe that cOllServative assumptions were made in estimating the microbial
risks in tlte CWS. Section
5.4;7
()f tlte Final Report discusses in detail tlie Sensitivity and
U1Jc;ertainty Analysis()f the Microbial Risk Assessment and provides the following examples:
• Secondary transmission rates used are generally at the high end of those
reported
in the technical literature. ThfJreJore, the assumptions on
secondary
tra~smission
are conservative and the resulting secondtlry tUness
rates may be biased high.
• The measured pathogen concentration$ llnder dry wetzthercdnditions are
limited to sampling locations near the
W'RPs and they were used as
representative concentrations
of the entire waterway downstream of the
WRP. Under dry weather
conditions,- these concentraiWllS willbebiased
high relative to concentrations at rocatiollS more distant from the WRP.
• The measured concentrations
of E. coli ate assumed to represent the most
virulent strain; the percentage
of pathogenic E. coli was tonservaUvely
assumed
to represent
2.7%
oJ the total measured concentrations. For other
organisms, such
as adenovirus, all the organisms are assumed to represent
the pathogenic strain leading to gastrointestinal illness. This assumption
may overestimate the illness associated with exposure to these organisms.
• Virus concentrations measured
by the assay systems may overestimate viral
risk. Viral assays are not specific to the pathogenic virus in question and
may detect less pathogenic viral strains.
• Recreational use may
be inversely correlated with. wet weather. CWS
recreational use was assumed to occur randomly over the course of the
recreational season. The
majority of the iUnesses were associated with wet
3
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
weatlur events.
If
the frequency oj exposure on wet weather days is lower
than average then the resulting risk estimate may be biased high.
• Some receptors with frequent use
oj the CWS mo:y have lower sensitivity to
some pathogells due to acquired immunity.. Repeated exposure topathogens
in water
is known to produce tolerance in individuals through immune
related mechanisms. Dose.response parameters used
in the assessment are
generally derived
!rom({naive" individuals and represent upper-end
estimates
of infectivity for the general populatiOll. Since repeated exposure
to the waterway is likely for a significant subset of the recreational
population, th¢ risk
oj illness Jor these individuals is probably over-
estimated
by this risk assessment.
For example, high Calicivirus measures were dismissed as an artifact and an outlier.
Response: Section
3.3.3
oj the Final RllPort discusses alfCalicivirus results in detail
During dry weather,norovirus was .only detected in
5
samples or about
7%
of the 75
samples; Dilnngtlte-l'Vorth SidedryWliiithersaiiijiling, only one iJutjallsample(1-oJ25
samples
[4%J)
had a detectable norovirus concentratiOn of 35,000 PCRMPN/lOOL (see
Tables
3.7
and
3-9
in the Final Report). The .greater concentration of
Caltciviru~
or
norovirus observed in this sample could be attributed to
the/act that only duplicates
per dilution in
.the MPN assay could be performed because of reassay difficulties
reducing the precision
of this analysis. In atiditir.m, of the five norovirus sampkis with
MPNassays,. (his sample was the
Qnly one tkat had a positive resillt in the highest
dilutioit. The combination
of thesejacto.rsc.ould have resulted in the relatively high
MPNvalue of this sample; As stated in the report, the high Calicivirus concentration
in the subjectsample
is likely an artifact oJthesejactors and it appears to be an outlier.
High infectivity parameters for adenovirus were dismissed because they usually cause
respiratory illness.
Response: The reviewer's comment mischaracterizes how adenovirus microbial risks
were estimated. Section
5.1
of the Final Report clearly states that some adenovirus
strains are primarily associated with respiratory illness.
However, fec{ll.oral
transmission associated with gastrointestinal illness
is the primary effect evaluated in
this study.
As a conservative assumptWn all detected adenovirus
.
was assumed to
contribute to gastrointestinal illness.
.
The lower infectivity of echovirus was considered instead of rotavirus.
Response: The reviewer's comment misckaracterizes the selection ofihe echovirus dose
response as a surrogateJoradenovirus. Section.S.3.3 oj the Final Report sUItes that
several
dose~response
relationships are reported for adenovirus but none of these are
specifically for Ad40 or Ad41, subtypes primarily associated with gastrointesti'nal
illness. This willtead
to an overestimateQf the true risks fQr gastrointestinal illness.
Therefore, the dose-response
for echovirus
12
was selected as a surrogate jor total
4
enteric viruses. This approach was recommended by Dr. Charles Gerba of the
University
of Arizona.
The notable exception to this is secondary transmission where some apparent
conservative assumptions
were made, but since it is not clear how secondary transmission
was modeled and. since there was no sensitivity analysis conducted
it
is impossible to
evaluate how these assumptions ultimately affected the results.
Response: Section S.4.2 of the Final Report presents a detailed discussion on Disease
Transmission
Mode~
including secondary attack rates. As stated in the report the
secondary
attack rates/or various organisms depend on the virulence of the organism
in question, the amount
"of organisms
an
infected in4ividua/ sheds. and the
enviromnental stability
of the organisms. Table
5
M
6
of the Final Report presents a
summary
of secondary attack rates used in this analysis. Footnotes to TableS-6
indicate
that thes8cfmdarytransmission rates used
ill.
the microbial risk estimates are
generally
at the" hlghend of those reported in the technical literature. Therefore, the
(lssumptions on secon,dqrytransmission are conservative
and the resulting secondary
··············ill1tess·ratesm-ay~be:biased-high.·········
.. ......................
._ .. _ .... .
There is also some question about the activities consjdered. Why wasn't fuJi body jet
skiing considered? Or other
full body exposures even if they area rare and prohibited,
would still result in risk ofmness.
Response: As stated, in the Introduction of the Final Report (see first paragraph on
page
5),
the UAAStakeholders evaluating the CWS have agreed that swimming and
other primary contact recreation shou14 not
be considered as a viable designated use
for the CWS because
of physical limililtions due to the configuration of the
e.mbankments
and safety hazards. It was not within the scope of work of the microbial
risk assessment to evaluate health risks originating from undesignated uses
of the
CWS.
• Inadequate reporting of risk assessment results and methods
The actual risk assessinent is brief and contains no graphs and few brief tables.
It
is
unclear how microbial pathogen densities were estimated. Were distribution functions
estimated based on the observed results, or were the potential values sampled from the
actual results? Were only viable Cryptosporidium results considered? A table should be
provided listing the details of all parameters and their ranges in used in the risk
assessment. Furthermore, it is not clear how activities were randomly assigned, were they
assigned based on their frequency of occunence, or were they completely random? It is
also not clear how secondary illness was modeled or incorporated into the estimate.
B.esponse: Section 5.0 of tit€! Final Report (pages 94-140) discusses the data used; assumptions
made and detailed procedures involved in the risk assessment calculations, including: (1)
hazard identification,
(2)
exposure asse.ssmellt,(3). dose response assessment, and
(4)
risk
5
Electronic Filing - Received, Clerk's Office, April 10, 2009
characterization. In addition, Tables
5-1
to
5.17
and Figures
5-1
to
5-4
provide pertinent
information that addresses
the reviewer's comments.
Section 3.0
of the Final R,eport presents all the analytical results that were used in the
,nicrpbial risk
estiTlUltes in accordance with the procedures discussed in Section
5.4.3
of the
report.
Section
5.4.2
of the report discusses the
d~'ease
transmission ltWdel, including
secondary illness.
.
For cryptosporidium, the infectiouscollcentrations determiltedby the EPA-approved method
were used
in the microbial risk assessment.
it
Interval estimates were not reported
This is a major failing since only one estimate of the risk was reported. With the
significant amount of assumptions and uncertainty, bounds on these estimates must be
provided (95% bounds). Complete details of the Monte Carlo analysis should be provide
so the distribution of risk can be visualized.
•
No sensitivity analysis was provided
A sensitivity analysis should describe which assumptions most affected the risk estimates
and how they affected the risk estimates. Since so many assumptions. that were made
were not necessarily conservative. this is a vita] aspect to a risk assessment.
Response: Sectwn 5.4.70/ the Firnil Report presents a sensitivity analysis of the
contribution
o/each microbial risk input distribution to the variance of the resulting
risk estimates.
•
Variability and uncertainty were not discussed, evaluated or quantified
Each step of the risk assessment contains variability and uncertainty. Uncertainty could
be considered in the dose.response parameters or in the microbial densities.
Response: Section.
5.4.7
of the Ftnal Report presents a sensitivity analysis of the
contribution of each microbial risk input distribution tolhe variance of the resulting
risk estimates.
In
addition, uncertainties associated with tke risk estimates are also
discussed in this section.
•
Limitations were not discussed
One clear limitation is that only a few pathogens were considered and this methodology
does not characterize the cumulative risk associated with all pathogens potentially present
in an environment. Another clear limitation is the failure to discuss sensitive or
susceptible limitations. illnesses other than GI and the potential for long term sequelae
resulting from infection.
Response: Section
5.4.7
of the Final Report presents a discussion of all above-
mentioned limitations.
As stated in the text, this study did not account for all
6
Electronic Filing - Received, Clerk's Office, April 10, 2009
pathogens that may be present in the CWS recreational water.
However, the
microorgalZisms that were selected
for inclusion in the study include regulatory
indicators
and those that could be measured by EPA-approved methods that were
judged most likely to produce gastrointestinal illness. In addition, Section
2.1
of the
report includes a more complete rationale
on pathogen selection.
Section 5.1
of the Fin-al Report describes in detail the Hazard Identification component
of the microbial risk assessment study. As siatedin this section, exposure to microbial
contaminated water may result in both gastrointestinal
and non-gastrointestinal illness.
However, there are
no known dose response models for the non-gaStrointestinal
exposure routes. The risk
of gastroilltestinal illness was selected as the sentinel effect
Jor conducting the quantitative risk aSsessment.
However, non-gastrointestinal
jllnesses were only addressed qualitatively. Section
5.4.6
of the Final Repottpresents a
qualitative assessment
of the non-Gf risks associated with Pseudomonas aeruginosa.
In summary, while the QMRA methodology
is appropriate, many assumptions are
questionable; important
de~ails
are left out, there is no evaluation of the potential range of
--risks, arid -rio-sensitivitY -iiniilysis; Therefbiethe' QMRX- doesnofprovidesiiffiCienI----
information to support the assertion that there is minimal risk with the current state of no
disinfection. These details should either be provided to support the claims made, or
another, independent risk assessment should be conducted.
Response: The reviewer's comment makes. a lot
of assertions, but does. not provide any
specifics.
Section
5.4.7
of the Final Report presents a
sensitiviJy~g.naiysis
of the
contribution
of each microhial risk input distribution to the variance of the resulting
risk estimates.
In additiOn, uncertainties assocwted with the risk estimates are also
discussed
in this section.
Additional specific comments:
Introduction:
Did all the consultants listed contribute? While Drs. Gerba and Clancy role was clear, that of Dr.
Jack Colford was not.
If Dr. Colford contributed specifically to this study, his role should be
clearly defined.
Response: Dr. Colford was a member of our team and his role was to provide peer review of
the final Dry and Wet Weather risk assessment report. However, due to other professional
qommitments he informed
llS in December 2007 thai he was /lot available to provide these
services for our
report.
Page 2:
" .. no outbreakS .. traceable to treated wastewater ... "
Statement
is misleading because outbreaks are not a reliable health indicator due to problems with
consistent and reliable detection. Furthermore. statements such as these require citation from peer
reviewed literature or other outside sources to avoid the perception of bias.
7
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
Response: The report includes the following citation for the statements made:
"Metropolitan Sanitary District of Greater Chicago (MSDGC), 1984, Wastewater
Disinfection: A Review
of Technical and Legal Aspects in Illinois. Department of
Research and Development. Report No. 84-17. July."
However, this statement was remove4 from the
FilUll
Report.
"The year round implementation of chlorination to disinfect the sewage treatment effluents has
been reported to have adverse environmental effects"
The purpose of statements such as these is unclear and their presence in the introduction of a
presumably unbiased risk assessmentis .concerning. While this rriay be true, citations from peer
reviewed literature are necessary followjng statements such as these to avoid the perception of
bias. Furthermore, benefits of chlorination should also be discussed if the downsides are going to
be presented.
Response:
The
report includes the following citation/or the statements made:
"Metropolitan Sanitary District of Greater Chicago (MSDGC), 1984, Wastewater
Disinfection: A Review
of Technical and Legal Aspects in lllinois. Department of
Research and Development. Report No. 84-17. July."
However, this statement wasremovedfrom the
Final
Report.
In addition, a s.ectWn
has
been added (Section
4)
in tile Final Report that provides a
comprehensive
overview
of
disinfection
technologif]s,
incJuding:
(1)
cltlorinatWnldechlorination,
(2)
ozonation, and
(3)
UV. Advantages and disadvantages of each
technology are discussed, including disi7ifectioneffectivettess,and disinfection by-product
formatioll.
.
.
Page 32:
If
censoring is greater than 80%, all data are statistically insignificant? Even though there was
20% detection?
.
As discussed in Section
3.1.3
of the Final Report, semi.log box plots were created to
graphically demomtrate the central tendencies and variability of the various bacteria
datasets. The text states that no box
plots
were prepared for dry weather Salmonella
results
as most of these datasels were statistically insignificant (i.e., non-detect
frequency
>80%). As explained in the text these results were not excluded, but the
geometric mean values (generated using the maximum likelihood method) are better
indicators
of this trelld for significantly cellsored datasets. However) box plots of
bacteria, including. Salmonella were prepared for wet weather data that had a more
robust data base
of detectable results.
8
Electronic Filing - Received, Clerk's Office, April 10, 2009
Page 33:
What is the point to the detailed analysis of the correlation of indicator organisms? These are not
used
in the risk assessment. Rather energy should have been spent on providing more details of
the actual risk assessment.
Response: The ultimate purpose of the analysis was to determine correlations between
pathogells and indicators unaerboth
dry and wet weather contfitions in order to ascertain if the
weather or any other factor can affect such correlations. To address the reviewer's comment,
the statistical correlations between bacterid pathogens and indicators have been removed from
the body
of the report and are included
in
Attachment A of the Final Report. The statistical
analysis in Appendix A
indiCates that the correlati(m of bacterid in wet weather samples is
statistically more signijicantcomparedto the dry weather samples.
Pag~36:
Although the ECIFCdifferences in upstream vs. downstream samples were not statistically
significant this could be a function of sample size--there is a consistent difference and there
:'couldbemoresophisticated:measures:to~assess
this. Thep"valueshouldbereported;not.sitnply
stated as >0.05.
The difference in the EC:FC ratios with what the District obtained calls into question the
representativeness of the data for the risk assessment.
Response: The lower EC/FC estimates in this study could be attributed to the fact that the
Di$trict's analysis
is based on a much la:rger database that includes several years of sampling
oj the waterway.
Page 41:
"While levels of potentially viable
Giardia
cysts may pose public health risk, it is important to
note that not all viable organisms are capable of infection"
Seems to be a prejudicial statement. Not clear why this is important to note.
Response: This statement was taken. verbatim from the Clancy Environmental Consultants,
Inc. (eEC) analytical laboratory
report. CEe was our expert laboratory for protozoa analysi$.
According to eEC this is a factual statemellt that is important to twte. All eEe analytical
reports are included in Appendices
e.l and C.2 of the Final Report.
Page 42:
"The results indicate that a relatively small number of samples (23%) had detectable
concentrations of enteric virus."
Relative to what? This could be an important contribution to pathogen exposure, but no
infOlmation is provided to support the assertion that it is "relatively" small.
Response:
"Relativ.e'~
refers to the total number of samples.
9
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
Page 44:
Citations need to be provided for statements to the effect of that
blc
the RT PCR does not provide
infectivity information it impedes meaningful health risk evaluation; Certainly
it
puts bounds on
the levels of potential risk
(0%
viable, to
100%
viable). Other sources could be evaluated for
viability of norovirus
in wastewater.
Page
91:
Inhalation not considered important-need citations to support this anti-conservative
simplification and assumption.
For canoeists, kayakers, this could
be an important pathway
Response: Section S.2o/the Final Report discusses exposure assessment pathways.
The text clearly states that the most imporrnnt exposure pathway
is via incidental
iligestion but other routes can also
beimpormnt for some microorganisms, like
e::fPtJsw~
... via
irthglgJfQ!J,
.
fffye()c.d.e.rJfJgl
cJJrttgc:t.~r1J~
...
~c.~t llrs()4i.~f~~S.~§.!Ii~relfltive
..
.c()ntribiitiolz io totalinitike byseveriiljiiitlnvay's"'(iiiCidentalwaleF ingestilili; .iithiilation. .
and dermal
cOri/(ict)
to .determine the reliltivecontriblition of each pathway to taW
exposure to microbiolOgical organisms in surface water while recreating.
Page 92:
Activities such as water Skiing, etc. were excluded because they are not allowed, but do they
occur?
Is the prohibition <,<nforced? An accurate risk assessment would consider these activities if
they occurred especially when evaluating
the potential benefit of disinfection.
JetSkis~classified
as pleasure boating with minimal contact. This is problematic-also "the RA
does not consider jet skis that result in immersion.
$.esponse: As stated in the Introduction of the Final Report (see First Paragraph on
page
5),
the UAA Smkeholders evaluating the CWS have agreed that swimming and
other primary contact recreation should not be considered
as a viable designated use
for the CWS because of physical limitations due to the configuration of the
embankments
and safety hazards. It was not within the scope of work of the microbial
risk assessment to evaluate health risks originating from undesignated uses
of the
CWS.
Page 100:
Using echovirus (less infectious) instead of rotavirus (the most infectious) for the dose response
relation, results
in less conservative (fewer illness) estimates.
Response: Section
5.3.3
of the report dis.cu.sses the Dose Response Assessment of
Adenovirus. As stated in the report, several dose-response relationships are reported for
adenovirus
but none of these are specifically for Ad40 or Ad41, subtypes primarily
a{)sociated with gastrointestinal illness. This will lead to an overestimate of the true
risks
for gastrointestinal illness. Therefore, the dose-response for echovirus
12
was
10
Electronic Filing - Received, Clerk's Office, April 10, 2009
selected as .a surrogate for total enteric viruses. This approach was recommended by
Dr. Charles Gerba of the University of Arizona.
Page 101:
Was genetic immunity/susceptibility to norovirus infection considered?
Response: No special distribution was applied to account for genetic polymorphisms
related to susceptibility. Similarly no adjustment was made
to account for acquired or
natural immUlzity. We do not helieve that the additional uncertainty addu! by
including these factors
is warranted by the increase in accuracy of the results
if
these
factors were considered. For example,
we do not have data to. indicate what
perC8'ittage of the recreational population are repeat visitors and potentially more
resistant vyacquired immunity. Our analysis considers all receptors nai've-andequally
susceptible.
.
Page 102:
By using the mqre conservative GI rnoclel for
adenovir\l~;
total 4ell,lth effects are underestimated.
Should also evaluate respiratory risks with the more infectious model. What is the justification for
using the Jess infectious parameter?
Response: Section
5.3.3
of the Final Report discusses the Dose Response Assessment of
Aden!lvirus.As stated in the report, several dose,.respcmse relationships are reportedfor
adenovirus
but none of these are specifICally for Ad40 or Ad41, subtypes primarily
assocmted with gastrointestinal illness. This
wUl lead to an overestimate of the true
risks
for gastrointestinal illness. Therefore., the dose-response for echovirus
12
was
selected as a surrogate for total enteric viruses. This approach was recommended by
Dr. Charles Gerba ofthe.University of Arizona.
Page 105:
Again the focus on GI results in a conservative estimate of overall risk
R.esponse: Section 5.1 of the Final Report describes in detail the Hazard Identification
component
of the microbial risk assessmetlt study. As stated in this section, exposure to
microbial contaminated water may result in both gastrointestinal and non-
gastrointestinal illness. However, there are
no known dose response models for the
non-gastroilztestinal exposure routes. The risk
of gastrointestinal illness was selected
as the sentinel effect for conducting the quantitative risk assessment; However, non-
gastrointestinal illnesses were addressed qualitatively.
Section
5.4.6
of the Final Report presents a qualitative assessment of the non-GI risks
associated with Pseudomonas aeruginosa.
11
Page 111:
Since Monte Carlo analysis was used, why wasn't a risk distribution (e.g., 50
th
percentile, 90
th
percentile, etc) generated?
Response~'
To simplify the presentation of the resuits, the final exposure distributions
were realized
for a set 0/ recreational receptors and the proportion of that population is
reported. Specifu:qJly, for each of the one million indtviduals evaluated in the Monte
Carlo analysis an exposure dose
WaS computed and the probability of infectUm
computed.
At that point a random number was generated and compared to the
probability
of infection. If the random number was less than thepT(1habilitythen the
individual was assumed to be infected
and subsequent evaluation of the probability of
illness given infection and secondi{ry infection was computed. .The advantage of this
technique is
the easy computation of the proportion of recreaticmal users.in the CWS
that may become
ill
during recreational exposure..
Details on how secondary spread was modeled are not clear.
Response.~
Section
5.4.2
of the Final Repart,discusses the. DiseaseJransmission
mode~
including secondary transmission. As stated
in
the report, to account for secondary
transmission, a dynamic risk
model was developed that considers secondary exposure
through contact with
CWS recreational users.
Estimates of the ittfectivity and
transmission rate as inputs
for the dY1tamic model were derived from the primary
literature
far each of the microorganisms of interest.
Because the number of
individuals exposed through recreation on the CWSis ii relativeiy small proportion of
the total popuiation of the Chicago metropo[{tan area, population levels. oj acquired
immunity
and tllness hy secondary transmission were not impacted. Therefore, the
proposed dynamic
model considers a steady-state level of immunity and estimates
disease incidence only
in the recreational receptor populatiottand their immediate
family. This approach addresses
the important dynamic aspects of disease transmission
from
CWS exposure in the population most at
risk.
Page 117:
How was recreation type selected in the simulation? Were they in proportion to the actual usage?
Response: Section 5.2.1 oflhe Final Report discusses Waterway Use Summary and
Receptor Group Categorization.
As stated in the report, several sources of information
were reviewed to estimate recreational use
and exposure to the CWS. Each of these
studies provides insight
on the types and frequency of recreational exposure expected
in the waterway. For quantitative
risk andlysis
7
the Use Attainability Analysis (UAA)
study was
used as the primary sourcefor exposure use datafor the CWS. The purpose
oj the UAA is to "evaluate existing conditions, including waterway use practices and
anticipated
future uses to determine
if
use classification revisions are warranted"
(Source: Camp Dresser
and McKee, Inc. (CDM), 2007, Use Attaillability Analysis of
the Chicago Area Waterway System. August). The UAA surveys were conducted to
evaluate
the types of recreational use thai are currently being exhibited. on each of the
12
waterway segments. Based on the UAA, several recreational exposure scenarios were
selected for evaluation in the risk assessment.
Page 134:
Risk assessment was only conducted for limited number of GI pathogens.
Response: This study did not account for all pathogens t!tat may be present in the CWS
recrlliltional water. Section
2.1
of the Final Report inciudes a more complete rationale
on pathogen selection. Howevelj the pathogens that
were selected for inclusion in the
study include regulatory indicators and those that could
be measured by EPA approved
methods that were judg(!d most. likely to produce gastrointestinal illness. In addition,
_ ..
Section
5~Jof
the Final Dry and Wet Weather Report, dated April 2008 describes in
detailthe
Ha'QJi'd Iden#jication component of the microbial risk assessment study. As
stated in this section, exposure
to microbial contaminated water may result in both
gasttointestjnal and non-gastrointestinai illness.
However, there are
rtf)
known dose
response models for thenon.;gasttointestinal exposure routes.
The
risk
oj
'gastrointestinal-ulnesswas selected as the sentinel effect for.. conducting the
......
-quantitative risk assessment, However, non-gastrointestinal illnesses -were addressed
qualitatively.
Section
5.4.6
of the Final Report presents a qualitative assessment of the non-Gf risks
associated with Pseudomonasaeruginosa.
National Center for Environmental Assessment (NeEA):
Note: this lao's comments are based on a cursory review only.
Comments
There are some serious surrogacy issues -- e.g., using rotavirus data fol' a norovirus dose-response
is implausible.
Response: Section
5.3.3
ojthe Final Report discusses the Dose Response Assessment of
Adenovirus. As stated irt the report, several d(lse-responserelationships are reported for
adenovirus but none
of these are specifically for Ad40 or Ad41, subtypes primarily
{lssociated with gastr(lirltestinal illness. This
will lead
to
an overestimate of the ttue
risks for gastrointestinal illness. Therefore, the dose-response for echovirus
12
was
selected as a surrogate for total enteric viruses. This approach was recommended
by
Dr. Charles Gerba of the University of Arizona.
Page 133:
Table 4-6 presents a summary of the secondary attack rates that appear quite high. Additional
investigation of the original references are needed to get a better idea of whether or not the values
posted are reasonable.
13
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *
Response: Secondary transmission rates used are generally at.the higJz end of those r.eported
in the technical literature. Therefore, the assumptions OIl secondary transmissioll are
conservative and the resulting secondary illness rates may be biased high.
Page 115-116:
The discussion of the "disease transmission model" and secondary attack rates is very sketchy.
The authors vaguely mention "dynamic models" (which do not seem to be provided anywhere in
the document) and appear to be rather naive about the difficulty of parameterizing such models.
They state that secondary attack rates depend on virulence, shedding rate, and environmental
stability of the organisms. But probably buman contact patterns, characteristics, and age groups
are more important.
It does appear that this risk assessment has weaknesses that could potentially be meaningful
Response: Section
5.4.2
of the Final Report discuss.es the Disease Transmission Model,
including secondary trallsmission. As stated in the report, to account
for secondary
....::tr'a'lismissiOii;
iiliyntiiizic
..iiSli. iiiiiiiefwd:s.devid(jptiFiiiiiFiJiJizsMits. s'eeoiidiiij'-i£epd$ure
through contact with CWSrecreatWnal users.
Estimates of the
.
infectivity and
transmission rate as inputs for the dynamic model were derived from the primary
literature for each
of the microorganisms of interest.
Because the number of
individuals exposed through recreation on the CWS is a relatively small proportion of
the total population of the Chicago metropolitan area, population levels of acquired
immunity and illness by secondary transmission were
not impacted. There/ore, the
proposed dynamic model considers a steady-state level
of immunity and estimates
disease incidence only in the recreational receptor populatiQli and theirimmedklte
family. This approach addresses the importallt dyilamic aspects
of disease transmission
from
cws
exposure in the population most at risk.
National Exposure Research Laboratory (NERL):
Comments
Since the overall goal of the study is. to determine wbetheror not to disinfect the effluent why the
protozoans were included in this study?
The chlorine concentrations that would be used would result in little or no inactivation of the
G/C.
However. CEC's summation of the protozoan results and interpretation and method
limitations were quite reasonable .
. The number of Giardia cysts is lower than some other reports for sewage; bowever, this may
because there are only dry weather events in this portion of the study.
It
should be more cleady emphasized that the number of Cryptosporidium oocyst) from the
samples were below the cell culture detection limit and even if all of the oocysts applied were
infectious it is unlikely that a foci would develop.
The documents treatment of the parasite issue was really not adequate.
14
Electronic Filing - Received, Clerk's Office, April 10, 2009
Response: We believe that the Final Report provides a comprehensive evaluation of the
protozoa in the
CWS~
The fa/rowing aspects of protozoa are discussed in the report:
1. Section
3.2
discusses Pr(Jtozoa Aualytical Results including, inJectious
Cryptosporidium and Viable Giardia Cysts under both dry and wet weather conditions
2.
Section
4.5.2
discusses wastewater protozoa disinfection effectiveness using
Uv,
chlorination and otonation
3. Sections
5.3.7
and
5.3.8
present dose-response models for cryptosporidium and giardia
The risk assessment appears to be a standard boiler plate, which is only as good as the data used
tofonn it.
Response: The lise ojprobabilistic microbial risk asse.ssmentjor estimation ojflfness in
recreatlonal
~sers
is the state-of-the-science approach for estimating risk. Inclusion of
$econdary injection risks within a limited recreational popultition,.joint risk estimation
for multiple pathogens,. and realization of risks to esttmate the proportion of users that
J!r~.ljk~Jy!~become
ill
(lj"~ 1Joy(!l~tl!c~niqll:es
and represent the West thtnkingon
risk
evalu.ation.The ineihods and results from this study have beenTli"esubJiiCi iiJ4paj£er$
pr'eieitiedat
NaiionatconJeienc~s·and
3
peer manusc,.ijiisari
currentlyiiifireparatioll
for peer review stemming from this work.
This assessment uses input data that represent the highest quality and most extensive
contemporaneous bacteria, virus and protozoa data
jor recreational water currently
available. The fact that sampling
was conducted over multiple years from numerOllS
locatiolls along the
w4[erway
in
conditions that encompasses a range of weather
conditions provides some assurance that support illformation on
ceilSUS figures,
meteorological data, and recreational use are developed from highly reliable sources.
While
it is tnte that the results of a risk assessment are fmly as good as the i,!!-put data
used,
the inputs for this study are arguably the best recreation. use microbial risk
databases ever assembled.
15
Electronic Filing - Received, Clerk's Office, April 10, 2009
* * * * * PC # 186 * * * * *