United States       EPA Science Advisory      EPA-SAB-DWC-03-005
       Environmental       Board (1400A)           May 2003
       Protection Agency      Washington DC        wwiv.epa.gov/sab
&EPA Disinfection Byproducts and
       Surface Water Treatment: A
       EPA Science Advisory Board
       Review of Certain Elements
       of the Stage 2 Regulatory
       Proposals

       A Review by the Drinking Water
       Committee of the EPA Science
       Advisory Board Executive
       Committee

-------
                                     May 21,2003
EPA-SAB-DWC-03-005
Honorable Christine Todd Whitman
Administrator
U. S. Environmental Protection Agency
1200 Pennsylvania Avenue, NW
Washington, DC 20460

       Subject:      Disinfection Byproducts and Surface Water Treatment: A EPA Science
                    Advisory Board Review of Certain Elements of the Stage 2 Regulatory
                    Proposals

Dear Governor Whitman:

       This review was conducted by a panel convened in response to a request by the Office of
Ground Water and Drinking Water (OGWDW) that the EPA Science Advisory Board (SAB)
review several parts of two rules1 that are being proposed together:

       1. The Long Term 2 Enhanced Surface Water Treatment (LT2ESWT) rule.
       2. The Stage 2 Disinfectant/Disinfection Byproducts (S2DBP) rule.

       The panel consisted of the twelve members of the SAB Drinking Water Committee
(DWC) and six consultants.

       During September, 2000, a Federal Stakeholder Advisory Committee (Stage 2 Microbial
Disinfectants and Disinfection Byproducts Advisory  Committee) reached an Agreement in
Principle on recommendations for both these "Stage 2" rules after nearly two years of fact
finding, deliberation, negotiation, and consensus building. The Stage 1 rules promulgated in
1998, had also been developed after a series of formal negotiations with stakeholders. This
report presents the results of the SAB Drinking Water Committee (DWC) review of information
provided by Agency on the Stage 2 rules. The LT2ESWT rule is intended to increase protection
of public water systems against microbial pathogens, with specific focus  on Cryptosporidium.
The S2DBP rule is intended to increase protection of public water systems from disinfection
1 Only partial drafts of the two rules were provided; see Sections 3.3, 4.1 for listing of review materials.

-------
byproducts (DBFs), specifically variability in exposure. OGWDW intends to propose and
finalize the LT2ESWT and S2DBP rules simultaneously so that systems maintain adequate
microbial protection while reducing risk from disinfection byproducts^

       The Agency's charges with the Panel's comments follow in abbreviated form:

LT2ESWT Rule:

Charge: The SAB was asked to comment on 1) the analysis of the occurrence (measured,
modeled) of a disease-inducing protozoan (Cryptosporidium) in drinking water systems, 2) the
validity of a risk assessment both before and after applying the proposed treatment methods in
the LT2ESWTR to those drinking water distribution systems and 3) the proposed treatment
credits (effectiveness in reducing protozoan contamination) by four methods including off-
stream water storage, pre-sedimentation, lime softening and reducing filtered water turbidity
(referred to as microbial toolbox options).

Findings:

1.      The Panel commends the Agency on its groundbreaking work addressing the impact of
       the proposed regulation on endemic disease and agrees that the regulation should address
       this issue. On the other hand, neither the design of the regulation nor the form of the
       economic analysis directly addresses waterborne outbreaks. Historically waterborne
       outbreaks are the primary stimulus for the regulation and they are the arena where
       intervention through improved water treatment has demonstrated its greatest
       effectiveness. Failure to consider the impact of the proposed regulation on reducing
       waterborne disease outbreaks underestimates the benefit of this regulation on public
       health.
2.      There is a large amount of uncertainty in the modeling of the occurrence of
       Cryptosporidium and of the incidence of the disease cryptosporidiosis and the current
       benefits analysis does not give this uncertainty sufficient visibility.
3.      The modeling of Cryptosporidium occurrence appears to be plausible and well done.
       On the other hand:
       a)    The economic analysis is necessarily complex and a greater effort is required for
             effective communication.
       b)    Some statistical issues need to be addressed, and
       c)    Understanding the transmission of cryptosporidiosis should be explored more
             thoroughly.
4.      The Panel also commends the Agency, as well as the stakeholder process, for developing
       the bin classification framework2 as it adds great flexibility to the rule.
' Determination of regulatory action using a simple classification of water sources based on observed Cryptosporidium densities ("bins").

                                            2

-------
Recommendations:

1.      Reducing the likelihood of waterborne outbreaks should continue to be one of the most
       important goals of Agency regulations in water treatment. The Panel recommends the
       Agency conduct a systematic review of the design of the LT2ESWT Rule, assessing its
       effectiveness in addressing outbreaks.  Changes should be considered if necessary.
2.      The magnitude of the uncertainty in estimating the occurrence of Cryptosporidium
       oocysts and estimating the risk of Cryptosporidium infection and the potential
       significance of these uncertainties to the over- or under-estimation of benefits should
       have high visibility in any final documents.
3.      With regard to the modeling of the occurrence of cryptosporidiosis, the Agency should:
       a)      Include better graphics in the documentation to help the reader understand the
               analytical process.
       b)      Conduct and document sensitivity analyses to the prior distributions3 and
               demonstrate the absence of seasonal effects on annual average Cryptosporidium
               concentrations.
       c)      Clarify and justify the  selection of the dose-response function, assumptions about
               oocyst infectivity, assumptions of host susceptibility, and estimates of water
               consumption.
       d)      Provide more information on evidence of endemic disease; discuss the
               significance of secondary transmission; discuss the role asymptomatic infections
               play in disease transmission and address the effect of age on host susceptibility to
               the disease.
       e)      Compare the quantitative microbial risk assessment approach used by the Agency
               to previous quantitative risk assessments for Cryptosporidium described in the
               scientific literature.
4.      For the bin Classifications the Agency asked the Panel to review, our recommended
       credits are as follows: a)  for off-stream and pre-sedimentation - no credits, b) for two
       stage lime softening - 0.5 credits, but only  if all the water is treated in both stages; and c)
       for plants that meet special requirements in each filter - 0.5 credits.

S2DBP Rule:

Charge: The  Agency asked the  SAB  to comment  on:  1) whether the locational running annual
average (LRAA) (a new method of estimating concentrations of DBFs) of total trihalomethanes
(TTHM)4 and haloacetic acids (HAAS), in conjunction with the initial distribution system
evaluation (IDSE) (recommendations to utilities for identifying appropriate monitoring sites) of
 Previous probability assessments of existing data used to estimate occurrence under new conditions.


 These terms refer to by-products of the chlorination process. The Panel believes that the terminology, TTHMs (total trihalomethanes), to
represent the four regulated bromine- and chlorine-containing THMs is not adequate since they do not represent the full spectrum of
trihalomethanes in drinking water. For example, for some time researchers have also been reporting iodinated THMs in finished drinking water.
To avoid confusion regulations that pertain to only the four bromine- and chlorine-containing THMs should refer to these as THM4. A precedent
for this form of nomenclature already exists, e.g. HAAS, HAA6, HAA9. For the sake of clarity this report has attempted to employ that
nomenclature throughout.

-------
the proposed rule more effectively achieves public health protection than the running annual
average (RAA) (current method of estimating concentrations of DBFs) of the Stage 1 DBF rule
and 2) if the IDSE is capable of identifying new compliance monitoring points that target high
TTHM3 and HAAS levels and if it is the most appropriate tool available to achieve this objective.

Findings:

1.      The Panel believes that the proposed DBP2 rules will result in a reduction in the health
       risk to drinking water consumers.
       a)      The principal outcome of these rules will be increased assurance that each
              consumer will be exposed to regulated DBF levels that are at or below the MCLs
              specified.
       b)      A second, important outcome will be a reduction in the average level of the
              regulated DBFs in  many systems.
2.      The Panel does not believe that the current draft of the benefits document does an
       adequate job of reflecting the uncertainties associated with estimating the reduction in the
       health risk to drinking water consumers:
       a)      The Source Water Analytical Tool (SWAT) is used to estimate DBF
              concentrations in distribution systems before and after the rule, but the Agency's
              own work demonstrates that SWAT does not do a good job of this.
       b)      The rule seeks to reduce short term exposure to high DBF levels, but the IDSE is
              used to identify monitoring points with high DBF levels and it does not consider
              diurnal short-term variations.
       c)      Benefits are estimated by assuming that the incidence of DBF-related bladder
              tumors will decrease in proportion to the reduction in the nine regulated DBFs,
              but it is not evident that this will occur because it has not been adequately
              demonstrated that bladder cancer is associated with any of the regulated DBFs.
3.      The Panel believes that substantial further research will be required before the benefits of
       DBF reduction can be adequately quantified.

Recommendations:

1.      The Panel recommends that the Agency promulgate the proposed rule without delay,
       pursuing the IDSE and LRAA as more effective means of controlling exposure to DBFs
       in drinking water than present practice.
       a)      The Agency should give high visibility to the fact that this rule will increase the
              assurance that each consumer will receive water that meets the DBF MCLs.
       b)      The Agency should also give high visibility to the fact that this rule can be
              expected to reduce the average level of regulated DBFs in most systems.

-------
2.      The Panel recommends that the Agency do a more straightforward job of describing the
       uncertainties in the benefits analysis:
       a)     Either the portion of the benefits analysis which used the SWAT should be
             abandoned or the presentation should be revised to reflect the true uncertainties
             associated with the use of this model.
       b)     The Agency should acknowledge that the IDSE does not consider short term
             (diurnal) variations.
       c)     The Agency should be more candid about the limitations it faces in estimating
             improvements in health risk reduction due to the implementation of the new rule
             rather than assuming that bladder cancers will be reduced in proportion to
             reductions in THM4 and HAAS.
3.      For the future, so that it can address the limitations inherent in the use of the surrogates
       (THM4, HAAS) to represent the full spectrum of DBFs present in drinking water, the
       Panel recommends that the Agency:
       a)     Focus its future research program upon identifying causal agents for bladder
             cancer and other adverse health effects (other risks of cancer, impairment of male
             and  female reproduction, effects on developing organisms) associated with
             chlorinated drinking water in epidemiological studies.
       b)     Link future control strategies for DBFs more directly to the reduction of these
             causal agents.

       Thank you for the opportunity to review these proposals We would be happy to continue
to engage with the Agency as it pursues this action.  We look forward to your response to this
report.
                                       Sincerely,
       /s/                                             /s/

Dr. William Glaze, Chair                        Dr. R. Rhodes Trussell,Chair
EPA Science Advisory Board                    Drinking Water Committee
                                               EPA Science Advisory Board

-------
                                      NOTICE
       This report has been written as part of the activities of the EPA Science Advisory Board,
a Federal advisory committee providing extramural scientific information and advice to the
Administrator and other officials of the Environmental Protection Agency. The Board is
structured to provide balanced, expert assessment of scientific matters related to problems facing
the Agency. This report has not been reviewed for approval by the Agency and, hence, the
contents of this report do not necessarily represent the views and policies of the Environmental
Protection Agency, nor of other agencies in the Executive Branch of the Federal government, nor
does mention of trade names or commercial products constitute a recommendation for use.
Distribution and Availability: This EPA Science Advisory Board report is provided to the EPA
Administrator, senior Agency management, appropriate program staff, interested members of the
public, and is posted on the SAB website (www.epa.gov/sab). Information on its availability is
also provided in the SAB monthly newsletter {Happenings at the Science Advisory Board).
Additional copies and further information are available from the SAB Staff [US EPA Science
Advisory Board (1400AO, 1200 Pennsylvania Avenue, NW, Washington, DC 20460; (202) 564-
4533)].

-------
                     U.S. Environmental Protection Agency
                             Science Advisory Board
                              Executive Committee

CHAIR
Dr. William H. Glaze, Oregon Health & Science University, Beaverton, OR

SAB MEMBERS
Dr. Henry Anderson, Wisconsin Division of Public Health, Madison, WI
      Also Member: Environmental Health Committee

Dr. Trudy Cameron, University of Oregon, Eugene, OR
      Also Member: Advisory Council on Clean Air Compliance Analysis

Dr. Maureen L. Cropper, The World Bank, Washington, DC
      Also Member: Environmental Economics Advisory Committee

Dr. Kenneth Cummins, Humboldt State University,  Arcata, CA

Dr. Virginia Dale, Oak Ridge National Laboratory, Oak Ridge, TN
      Also Member: Ecological Processes and Effects Committee

Dr. Domenico Grasso, Smith College, Northampton, MA
      Also Member: Environmental Engineering Committee

Dr. Linda Greer, Natural Resources Defense Council, Washington, DC
      Also Member: Research  Strategies Advisory Committee

Dr. Philip Hopke, Clarkson University, Potsdam, NY
      Also Member: Research  Strategies Advisory Committee
                   Clean Air Scientific Advisory Committee

Dr. Janet A. Johnson, MFG, Inc., Fort Collins, CO
      Also Member: Radiation Advisory Committee

Dr. Roger E. Kasperson, Stockholm Environment Institute, Stockholm,, Sweden
      Also Member: Research  Strategies Advisory Committee

Dr. Raymond C. Loehr,  University of Texas , Austin, TX

Dr. Genevieve Matanoski, Johns Hopkins University, Baltimore, MD
      Also Member: Research  Strategies Advisory Committee

Dr. M. Granger Morgan, Carnegie Mellon University, Pittsburgh, PA
                                         11

-------
Dr. Rebecca Parkin, The George Washington University, Washington, DC
      Also Member: Integrated Human Exposure Committee

Dr. William H. Smith, Yale University, Center Harbor, NH

Dr. R. Rhodes Trussell, MWH-Montgomery Watson Harza, Pasadena, CA
      Also Member: Drinking Water Committee

SCIENCE ADVISORY BOARD STAFF
Mr. A. Robert Flaak, Washington, DC

Ms Betty Fortune, Washington, DC

Ms. Diana Pozun, Washington, DC
                                        in

-------
                     U.S. Environmental Protection Agency
                            Science Advisory Board
                          Drinking Water Committee
          Stage 2 DBF/Surface Water Treatment Rule Review Panel*

CHAIR
Dr. R. Rhodes Trussell, MWH-Montgomery Watson Harza, Pasadena, CA.

DWC MEMBERS**
Dr. David B. Baker, Heidelberg College, Tiffin, OH

Dr. Mary Davis. West Virginia University Health Sciences Center, Morgantown, WV

Dr. Ricardo De Leon, Metropolitan Water District, La Verne, CA

Dr. Sidney Green, Howard University, Department of Medicine, Washington, DC
      Member: Environmental Health Committee

Dr. Barbara Harper. Yakima Indian Nation, West Richland, WA

Dr. Lee D. (L.D.) McMullen, Des Moines Water Works, Des Moines, IA

Dr. Christine Moe, Emory University, Atlanta, GA

Dr. Philip Singer. University of North Carolina, Chapel Hill, NC

Dr. Gary A. Toranzos. University of Puerto Rico, San Juan, PR

OTHER SAB MEMBERS
Dr. Richard Bull, MoBull Consulting, Inc., Kennewick, WA
      Member: Research Strategies Advisory Committee

Dr. Lauren Zeise, California Environmental Protection Agency, Oakland, CA
      Member: Research Strategies Advisory Committee

CONSULTANTS
Dr. Mark Benjamin, University of Washington, Seattle, WA

Dr. L. Mark Berliner, Ohio State University, Columbus, OH

Dr. Paul Boulos, MWH Soft, Inc., Broomfield, CO

Dr. Michael J. Daniels, University of Florida, Gainesville, FL
                                        IV

-------
Dr. Gregory Harrington, University of Wisconsin, Madison, WI

Dr. Charles O'Melia, The Johns Hopkins University, Baltimore, MD

LIAISONS
Dr. David P. Spath, California Department of Health Services, Sacramento, CA

SCIENCE ADVISORY BOARD STAFF
Mr. Thomas O. Miller, Washington, DC

Dr. James N. Rowe, Washington, DC

* Members of this SAB Panel consist of:

      a. SAB Members: Experts appointed by the Administrator to serve on one of the SAB
Standing Committees.
      b.  SAB Consultants: Experts appointed by the SAB Staff Director to a one-year term to
serve on ad hoc Panels formed to address a particular issue.
      c. Liaisons:  Members of other Federal Advisory Committee who are not Members of
Consultants of the Board.
      d. Federal Experts: "Federal Experts" are federal employees who have technical
knowledge and expertise relevant to the subject matter under review or study by a particular
panel.

** Current Members or Members at the time of the Panel review

-------
                           TABLE OF CONTENTS

1. EXECUTIVE SUMMARY 	 1

2. INTRODUCTION AND CHARGE 	5
      2.1 Introduction	5
      2.2 The Charge	6

3. LONG TERM 2 ENHANCED SURFACE WATER
      TREATMENT RULE	7
      3.1 Introduction	7
      3.2 Charge  Question 1: Analysis of Cryptosporidium occurrence  	7
            3.2.1 Panel Response to LT2ESWTR Charge Question 1—Analysis of
                  Cryptosporidium occurrence	7
                  3.2.1.1 Background 	7
                  3.2.1.2 Panel Conclusions	 12
      3.3  Charge Question 2: Pre-and post-LT2ESWTR Cryptosporidium risk
            assessment	14
                  3.3.1.1  Hazard Identification 	 15
                  3.2.1.2  Dose-Response Assessment	16
                  3.3.1.3  Exposure Assessment (pgs 5-14 - 5-24) 	 19
      3.4  Charge 3: Treatment credits for four microbial toolbox options 	21
            3.4.1 Panel Response to LT2ESWTR Charge Question 3 	21

4. STAGE 2 DISINFECTION BYPRODUCTS RULE 	24
      4.1 Charge  1: Initial Distribution System Evaluation (IDSE): 	24
            4.1.1 Panel Response to S2DBP rule Charge Question 1	24
                  4.1.1.1  Initial Distribution System Evaluation (IDSE) Effectiveness .... 24
                  4.1.1.2 IDSE Appropriateness 	25
      4.2  Charge2: Public Health Protection of S2DBPR	28
            4.2.1 Panel Response to S2DBPR Charge Question 2	28

REFERENCES 	R-l

ATTACHMENT A - ACRONYMS AND ABBREVIATIONS  	  A-l

ATTACHMENT B - SELECTED GLOSSARY OF TERMS	B-l

ATTACHMENT C - BIOSKETCHES OF THE DRINKING WATER COMMITTEE
      MEMBERS 	C-l
                                      VI

-------
                           1. EXECUTIVE SUMMARY
       The Drinking Water Committee (DWC) of EPA's Science Advisory Board (SAB) met to
consider several support documents that are a part of the Agency's Long Term 2 Enhanced
Surface Water Treatment (LT2ESWT) rule and the Stage 2 Disinfectant/Disinfection Byproducts
(S2DBP) rule, both of which are under development by the Agency.  During September, 2000, a
Federal Stakeholder Advisory Committee reached an Agreement in Principle on
recommendations for both these Stage 2 rules after nearly two years of fact finding, deliberation,
negotiation, and consensus building.  The Stage 1 rule promulgated in 1998,  had also been
developed after a series of formal negotiations with stakeholders. This report presents the results
of the SAB Drinking Water Committee (DWC) review of information provided by the Agency
on the Stage 2 rules.

       The 1996 Amendments to the Safe Drinking Water Act (SDWA) require the Agency to
develop National Primary Drinking Water Regulations (NPDWRs) for contaminants which have
an adverse effect on the health of persons and where regulation provides a meaningful
opportunity for public health protection. The Agency is developing a LT2ESWT rule to provide
increased protection for public water systems against microbial pathogens, with a specific focus
on Cryptosporidium. The proposed rule is intended to supplement existing surface water
treatment rules by establishing targeted treatment requirements for systems with greater
vulnerability to Cryptosporidium.   Such systems include those with high concentrations of
Cryptosporidium in their source water and those that do not provide filtration. In addition, the
1996 SDWA Amendments require  the Agency to develop a S2DBP rule.  The intent of the
proposed S2DBP rule is to  reduce the variability of exposure to disinfection byproducts (DBFs)
for people served at  different points in the distribution systems of public water supplies. The
Agency has suggested that  this decreased exposure will result in reduced risks from potential
reproductive and developmental health effects and cancer. To be consistent with the SDWA
requirements for risk balancing, the Agency intends to propose and finalize the LT2ESWT and
the S2DBP rules simultaneously. This coordinated approach is designed to ensure that systems
maintain adequate microbial protection while reducing risk from disinfection byproducts.

       The Panel believes that the  terminology, TTHMs (total trihalomethanes), to represent the
four regulated bromine- and chlorine-containing THMs is not adequate since they do not
represent the full spectrum  of trihalomethanes present in drinking water. For example, for some
time researchers have also been reporting iodinated TFDVIs in finished drinking water. To avoid
confusion,  regulations that pertain to only the four bromine- and chlorine-containing THMs
should refer to these as THM4.  A precedent for this form of nomenclature already exists, e.g.
HAAS, HAA6, HAA9.  For the sake of clarity this report has attempted to employ that
nomenclature throughout.

       This report has two major parts reflecting the structure of the  Agency Charge. The
charge to the SAB Panel for the Long Term-2 Enhanced Surface Water Treatment rule asked the
SAB to comment on: a) the analysis of Cryptosporidium occurrence; b) the pre- and post-
LT2ESWTR Cryptosporidium risk assessment; and c) the proposed treatment credits for four

                                           1

-------
microbial toolbox options. For the S2DBP rule, the Agency asked the SAB to comment on: a)
whether the locational running annual average (LRAA) for total trihalomethanes (TTHM) and
haloacetic acids (HAAS), in conjunction with the initial distribution system evaluation (IDSE),
of the proposed rule more effectively achieves public health protection than the running annual
average (RAA) of the Stage 1 DBF rule and b) if the IDSE is capable of identifying new
compliance monitoring points that target high TTHM and HAAS levels and if it is the most
appropriate tool available to achieve this objective.

       For the LT2ESWTR, because the risk assessment is quite complex, the Panel
recommends that the document include graphics that show how the different elements were
derived and how they relate to each other.  For clarity,  comments and recommendations are
presented separately for the three charge questions related to the risk assessment.

       First, the Panel concludes that the occurrence modeling appears to be both plausible and
well-done.  However, the Panel believes that a number of issues need to be addressed, either by
supplementing the current documents and/or modifying the model.

       The Panel  recommends that the Agency:

       a)     Conduct and document sensitivity analyses to the prior distributions and,
       b)     Demonstrate the absence of seasonal effects on the annual average
       Cryptosporidium concentration.

       Secondly,  for the microbiological risk assessment review, each of the basic elements was
examined in order: hazard identification, dose-response assessment, and exposure assessment.
Then the outcome of the risk assessment was evaluated. Two criteria were considered in the
Panel evaluation:  a) whether  the Agency assumptions were transparent, and b) whether scientific
evidence exists to support the assumptions. Cryptosporidiumparvum has been responsible for
significant  waterborne disease outbreaks, and it is likely that the organism is responsible for
reports of significant endemic disease as well. Both of these outcomes are important. The
current Agency analysis (The Cadmus Group, Inc., 2001b) for the LT2ESWTR does  an excellent
job of addressing the impact  of drinking water quality on the incidence of non-reportable
endemic disease and the health risk reduction that will  result from the reduction of endemic
disease as a result of the proposed regulation. The Agency is to be congratulated for  this
ground-breaking work. On the other hand, in the present draft, neither the design of the
regulation nor the contents of the Agency analysis directly address reportable waterborne
outbreaks.  These  outbreaks are the primary stimulus for the regulation and reducing  their
occurrence should be one of the most important potential outcomes from the regulation as well.
The Panel recommends that the Agency conduct a systematic review of the design of the
LT2ESWTR regulation keeping its effectiveness in addressing waterborne outbreaks  in mind.

       a)     The Panel agrees with the basic information on Cryptosporidium health effects in
              the Hazard Identification section but recommends that the following be included
              in the  analysis: a) evidence  of current prevalence of endemic disease; b)
              information on secondary transmission of cryptosporidiosis; and c) host age and
              frequency of asymptomatic infections.
                                           2

-------
       b)     For the Dose-Response Assessment, the Panel recommends clarification and
              justification of: a) the basis for the selection of the dose response function that
              was used and whether other models were considered; b) the term "infectivity" as
              it is used in the Agency analyses; c) the assumptions about infectivity of oocysts
              used in human dosing experiments, infectivity of oocysts found in environmental
              samples and of the significance of Cryptosporidium genotype when evaluating
              infectivity for humans; and, d) assumptions about variability in host
              susceptibility, both due to possible immunity resulting from previous infections
              and due to other susceptibility factors such as age and health.

       c)     For Exposure Assessment, the estimates of consumption require clarification.

       d)     For the Risk Assessment, the Panel notes that quantitative microbial risk
              assessment is a rapidly developing field. The Agency should a) identify other
              approaches to microbial risk assessment, especially risk assessments for
              Cryptosporidium., that are reported in the literature and consider how they
              compare to their own assessment; b) include a discussion of uncertainties and
              variability; and c) discuss assumptions which may lead to underestimates or
              overestimates of risk and benefits.

       Finally, for the treatment credits for the four microbial toolbox options, the Panel
commends the Agency, as well as the stakeholder process used, for  developing the bin
classification framework for identifying the treatment requirements  for drinking water and the
microbial toolbox containing possible treatment options to guide systems having treatment
needs.  These alternatives add great flexibility  for meeting varying water quality and treatment
options and should result in safe drinking water for the people of the United States. The Agency
charged the Panel with evaluating Agency information on four of the toolbox options:  a) off
stream raw water storage; b) pre-sedimentation, c) lime softening and d) lower finished water
turbidity.  Specifically, the Agency asked the Panel to comment on the credits that have been
proposed for specific toolbox options for Cryptosporidium removal. In summary, the Panel
recommends that no presumptive credits be given for off-stream storage and pre-sedimentation.
It does agree with giving 0.5 log credit for two-stage lime softening if all the water is treated
with both stages, and 0.5 log credit for plants that demonstrate a turbidity level in  each
individual filter effluent less than or equal to 0.15 NTU in at least 95 percent of the
measurements taken each month. Details about these recommendations  are found in the report.

       For the Stage 2 DBF rule, the Panel believes that the proposed Initial Distribution
System Evaluation is capable of identifying new compliance monitoring points that target higher
THM and HAA levels than are currently measured in the existing THM Rule and  Stage 1 DBF
Rule compliance monitoring programs.  However, the IDSE does not consider short-term,
temporal variations that occur at different sites in the distribution system due to varying water
demands and distribution system architecture and operation.  This temporal variability needs to
be acknowledged in the IDSE documentation.  The Panel further believes that the  proposed
standard monitoring program (SMP) for sub-part H systems serving more than 10, 000 people is
reasonable.  The Panel, however, does make some recommendations concerning the proposed
                                           3

-------
sampling requirements. The switch from the running annual average (RAA) approach to the
locational running annual average (LRAA) approach provides a measure of equity not previously
reflected in the standards for disinfection by-products.  The LRAA allows one to state that a
larger segment of the consumers will be provided with drinking water within a particular water
system which will meet the MCL than would be the case using the RAA approach.  The Panel
also agrees that these changes are likely to result in a reduction in health risk due to DBF
exposure, but the Agency has not demonstrated that this reduction in health risk will be in direct
proportion to the reduction in the THM and HAAS concentrations.

       The Committee recommends that in proposing its Stage 2 DBF rule, the Agency:

       a)    Continue to pursue the concept of locational running annual averages (LRAAs) as
             a more effective means of controlling exposure to harmful compounds  in the
             drinking water than system-wide running annual averages (RAAs).
       b)    Identify temporal limitations in the IDSE documentation and require periodic
             reevaluation of selected sites;
       c)    Reallocate sampling locations so that, for both free chlorine and chloramines,
             sampling takes into  account potential high THM and HAA sites;
       d)    Require the measurement and reporting of residual chlorine and individual THM
             and HAA species;
       e)    Provide more guidance to utilities to identify sampling sites with highest HAA
             concentrations;
       f)     Improve the proposed system specific studies (SSS) approach (Chapter 6);
       g)    Reconsider the use of the SWAT (Surface Water Analytical Tool) model and ICR
             (Information Collection Rule) data in economic analyses or risk reduction
             calculations;
       h)    Focus their research program upon identifying causal agents for bladder cancer
             and other potential adverse health effects associated with chlorinated drinking
             water; and,
       i)     Link control strategies for DBFs to reduction of causal factors of health effects.

-------
                      2.  INTRODUCTION AND CHARGE
2.1 Introduction

       The 1996 Amendments to the Safe Drinking Water Act (SDWA) require the Agency to
develop National Primary Drinking Water Regulations (NPDWRs) for contaminants which have
an adverse effect on the health of persons and where regulation provides a meaningful
opportunity for public health protection. The Agency is developing a Long Term 2 Enhanced
Surface Water Treatment (LT2ESWT) rule to provide increased protection for public water
systems against microbial pathogens, with a specific focus on Cryptosporidium.  The proposed
rule is intended to supplement existing surface water treatment rules by establishing targeted
treatment requirements for systems with greater vulnerability to Cryptosporidium.  Such systems
include those with high concentrations of Cryptosporidium in their source water and those that
do not provide filtration.

       In addition, the 1996 SDWA Amendments require the Agency to develop a Stage 2
Disinfectant/Disinfection Byproducts (S2DBP) rule. The intent of the proposed S2DBP rule is
to reduce the variability of exposure to disinfection byproducts for people served at different
points in the distribution systems of public water supplies. The Agency has suggested that this
decreased exposure will result in reduced risks from potential reproductive and developmental
health effects and cancer.

       To be consistent with the SDWA requirements for risk balancing, the Agency intends to
propose and finalize the LT2ESWT and the S2DBP rules simultaneously.  This coordinated
approach is designed to ensure that systems maintain adequate microbial protection while
reducing risk from  disinfection byproducts.  During September, 2000, a Federal Stakeholder
Advisory Committee reached an Agreement in Principle on recommendations for both these
rules after nearly two years of fact  finding, deliberation, negotiation, and consensus building.
Prior to that, the Stage 1 rules for DBFs and surface water treatment also reflected periods of
formal  regulatory negotiations and stakeholder discussions over a period of years stretching from
the early to mid-1990s.

       The Panel believes that the  terminology, TTHMs (total trihalomethanes), to represent the
four bromine- and chlorine-containing THMs is not adequate since they do not represent the full
spectrum of trihalomethanes in drinking water.  For example, for some time researchers have
also been reporting iodinated TFDVIs in finished drinking water.  To avoid confusion regulations
that pertain to only the four bromine- and chlorine-containing THMs should refer to these as
TFDVI4. A precedent for this form of nomenclature already exists, e.g. HAAS, HAA6, HAA9.
For the sake of clarity this report has attempted to employ that nomenclature throughout

       The EPA Office of Ground Water and Drinking Water (OGWDW) representatives
requested that the EPA Science Advisory Board (SAB) review several parts of the LT2ESWT
and the S2DBP rule proposals and  certain support documents and provide advice in response to a

-------
number of charge questions. This report presents the results of the SAB Drinking Water
Committee (DWC) review of these issues.

2.2 The Charge

The Agency charge to the SAB Panel for the Long Term-2 Enhanced Surface Water Treatment
rule asked the SAB to comment on: a) the analysis of Cryptosporidium occurrence; b) the pre-
and post-LT2ESWTR Cryptosporidium risk assessment; and c) the proposed treatment credits
for four microbial toolbox options.

       For the Stage 2 DBF rule, the Agency asked the SAB to comment on: a) whether the
locational running annual average (LRAA) for total trihalomethanes (TTHM) and haloacetic
acids (HAAS), in conjunction with the initial distribution system evaluation (IDSE), of the
proposed rule more effectively achieves public health protection than the running annual average
(RAA) of the  Stage 1 DBF rule and b) if the IDSE is capable of identifying new compliance
monitoring points that target high TTHM and HAAS levels and if it is the most appropriate tool
available to achieve this objective.

-------
             3. LONG TERM 2 ENHANCED SURFACE WATER
                               TREATMENT RULE

3.1 Introduction

       The Agency convened a group of stakeholders, including EPA itself, to hold formal
negotiations on issues related to the LT2ESWT and Stage 2 DBF rules from 1999 to 2000.  Their
Agreement in Principle, which contains recommendations for the proposed LT2ESWT and Stage
2 DBF rules, was published in the Federal Register on December 29, 2000 (US EPA, 2000).

       In general, because the risk assessment is quite complex, the Panel recommends that the
document include more graphics to illustrate how the different elements of the model were
derived and how they relate to each other. Exhibit 5.2 (The Cadmus Group, Inc., 200Ib) is
helpful but does not provide sufficient detail.  Additional figures are needed to show what
elements were in the pre-regulation risk assessment versus the post-regulation risk assessment
and how the reduction in risk from the proposed regulation was calculated. Figures 3.1 through
3.4 of this  report are examples displaying the Panel's understanding based on its reading of the
documents provided by the Agency and its discussions with EPA personnel and discussed below.

3.2 Charge Question 1: Analysis of Cryptosporidium occurrence

       The Agency requested SAB comments on the EPA analysis of Cryptosporidium
       occurrence.

       The Agency provided the Panel with a draft document entitled Occurrence and Exposure
Assessment for the Long Term 2 Enhanced Surface Water Treatment Rule. (The Cadmus Group,
200la) that discusses how the Agency estimated the occurrence distribution of Cryptosporidium
in the source and finished water of public water systems prior to implementation of a new
LT2ESWT rule.  Sections  of the document considered to be of particular importance discussed
the data sources used to estimate Cryptosporidium occurrence in source water, along with
analytical methods, data quality issues, and the statistical techniques used to model occurrence
distributions; information on observed and modeled results from the source water occurrence
surveys; information from studies of the physical removal of Cryptosporidium by treatment
processes; finished water occurrence data resulting from the Information Collection Rule (ICR);
a description of how the Agency estimated finished water Cryptosporidium levels prior to
implementation of the LT2ESWTR; and technical information on the statistical models used to
analyze source water occurrence data.

       3.2.1 Panel Response to LT2ESWTR Charge Question 1-Analysis of
       Cryptosporidium  occurrence

       3.2.1.1 Background. The model  developed by the Agency can be thought of in three
parts (Figure 3-1). The first part is designed to address an important limitation of the data
collected in the ICR and ICR Supplemental Survey (ICRSS), namely information on the national
occurrence of Cryptosporidium parvum oocysts at levels below the detection limits (DLs) of the

                                          7

-------
methods used in those surveys.  Thus, the first part simulates national distributions of the
concentration of C. parvum oocysts in the source water.  Using ICR and ICRSS data, the model
is designed to produce an estimate of the national occurrence of oocysts in untreated surface
waters, above and below the ICR and ICRSS DL4s.  Bayesian hierarchical models and Markov
chain Monte Carlo (MCMC) methods are used to accomplish this (Figure 3-2). These models
accommodate the many complex features seen in the data used by the Agency to develop its
national occurrence estimates, including low recovery probabilities, the presence of false
positives,  and the presence of true Cryptosporidium-free source waters.

Figure 3-1. The model developed by the EPA contains three components. The first uses data from the ICR and
ICRSS to produce a national distribution of C. parvum oocysts in untreated surface water.  The second uses that
national distribution and a model of treatment performance to produce a simulation of the national distribution of C.
parvum oocysts in finished water. The third component uses a dose-response model calibrated via human exposure
studies, data on water consumption, and finished water oocyst levels to predict the level of endemic disease.
                              Model 1 - Occurrence of oocysts in raw
                              water - Model uses data from ICR and
                              ICRSS to estimate the national occurrence
                              of C. parvum oocysts in raw water supplies
                              across the nation
                Model 2 - Occurrence of oocysts in Finished water - Model starts with
                data from Model 1 and then uses estimates of removal in treatment to
                produce an estimate of the national occurrence of C. parvum oocysts in
                finished water. Treatment performance is assumed to have a triangular
                distribution about the nominal performance specified. To estimate
                occurrence before regulation, existing treatment is used.  To estimate
                occurrence after regulation, a decision tree is employed where the
                treatment selected depends on the level of influent oocysts
                      Model 3 - Occurrence of endemic disease - Model starts
                      with data from Model 2 and then uses a dose-response
                      model to estimate the occurrence of disease.  The dose-
                      response model is calibrated using data from three available
                      human feeding studies.
  Responses are modeled above the detection limit as a function of turbidity, location, etc., and the model uses this in addition to information
about the number below the detection limit to 'impute' values below the limit.  The MCMC approach also integrates over the uncertainty of the
values below the detection limit (as opposed to 'imputing' a single value). In terms of validation, there is no way to validate 'values' below the
DL as they were not observed, but we can determine how well the model fits for values above the limit and determine whether the proportion
predicted by the model for the number below the limit is consistent with the observed data.

-------
 Figure 3-2.  Model 1: Occurrence of oocysts in raw water - Bayesian hierarchical models and MCMC Methods
              were used to estimate the national occurrence of C. parvum oocysts in raw water.
                       Z Observed oocyst
                       occurrence only shows
                       data above method
                       detection limit (D.L)
1CYTPA _.

m_
m_





i

4

X
^y

\\
2 (

x* 	 |



D 2
                                                             Log[oocyst/L]
                                               3. Analyze data from survey and propose structure for
                                               model that can be used to predict oocyst
                                               concentrations both above and below detection limit
                                               Q=<
          100%!

          75%

          50%

          25%

           0%
. Corrected .
"Estimate
4 Use model and Markov Chain Monte
Carlo (MCMC) methods to produce a
new "simulated" national frequency
distribution of oocysts.
               Log[oocyst/L]
        The second part of the model takes the national occurrence in untreated water from the
first part and uses treatment assumptions to produce an estimate of the national occurrence of C.
parvum oocysts in treated water (Figure 3-3).  To estimate occurrence before regulation,
treatment credits in the existing Interim Enhanced Surface Water Treatment Rule (IEWSTR) are
used.  The proposed regulation assigns water systems into various bins depending on the level of
oocysts in their untreated water. A higher degree of removal is required for systems with
untreated water falling into bins that correspond to higher oocyst concentrations.  To estimate
occurrence after regulation, treatment is assumed to meet the requirements that correspond to the
bin selected for each supply.  For the analysis in this second part, the Agency assumed that
treatment effectiveness is independent of concentration and, based on expert opinion, treatment
effectiveness across the nation is assumed to follow a simple triangular distribution with the
mode at the performance specified by the rule.
                                              9

-------
Figure 3-3. Model 2 - Occurrence of oocysts in Finished water - Treatment performance is assumed
to have a triangular distribution. Before regulation, existing treatment is assumed to meet the IESWTR.
After regulation, a decision tree is employed where the treatment selected depends on the level of influent
oocysts (the bin).
Fraction of
samples <
100% -i






^
_X
V
/

^"—



-4-20 2
                     Log[oocyst/L]
Fraction of
treatment

100% -



MCMC




0

/)0<\J
l//*\\
iff
jf
2

X
%
X
4
                      LogRemoval
Fraction of
samples <
100% -,






--^
jf°
-^/

/
^=



-4-20 2
                                                  5. Make treatment assumptions:Sefore rule,
                                                  assume nominal removal equals credit in
                                                  IESWTR.  After rule, based on raw water "bin"
                                                  choose treatment from toolbox and give credit
                                                  accordingly
                                                 6. Characterize treatment performance:
                                                 Assume treatment performance follows a
                                                 triangular distribution and that mode of
                                                 distribution varies from site to site ± 0.5 logs.
                                                  7. Simulate treatment performance: Use
                                                  MCMC to sample raw water and make removal
                                                  estimate while varying the mode of triangular
                                                  treatment distribution ± 0.5 logs. To produce an
                                                  estimate of the national distribution of oocysts in
                                                  finished water.
                       Log[oocyst/L]
       The third part of the model estimates the national occurrence of disease.  The model uses
the national occurrence of C. parvum oocysts in finished water and combines it with data on
water consumption and on dose-response to produce an estimate of disease.  The model
considers the distribution of infection (and disease) conditional on the concentration of viable
oocysts in the drinking water through the use of an exponential dose-response model.  The
parameters of the dose-response model were estimated using data from three human dosing
studies. A Bayesian hierarchical model is also used here to model the distribution of infectivity
across Cryptosporidium strains.  To predict the occurrence of disease, Monte Carlo  methods are
used to sample oocyst concentrations in finished water and volumes of water consumed and
estimate disease using the dose-response model  (Figure 3-4).
                                             10

-------
  Figure 3-4. Model 3 - Occurrence of endemic disease - Human feeding studies are used to
  calibrate the dose-response model and then MCMC methods are used to sample from finished
  water, determine the liters consumed and estimate the national incidence of endemic disease.
  f^fpy     	/
                                           0 20 40 bO oO 100

                                             Oocyst Dose
       100%!

        75%

        50%
         0%
                 -2
100%1
6 75/0






	 -w
11
/
/


\.
V




0.5 1 1.5 2 2.5
Liters Consumed
            Log[oocyst/L]



MCMC

100% -
Ł 75% -
Fraction Infe
CO Ol
O Ol O








j

/'

I
0 2






0 '
Ooc}
_,,-


jt


) f
^st

,-"
j


0
ose
-—-



0 100

                    ^Nati
National Disease incidence
                                                 8. Dose Response is calibrated to
                                                 human studies. The exponential
                                                 dose-response model is calibrated
                                                 using data from human studies with
                                                 UCP, Iowa and Tamu isolates of C.
                                                 pan/urn.
                                                 9. National Incidence of Endemic
                                                 Disease is Simulated: MCMC is
                                                 used to sample finished water,
                                                 determine volume of water
                                                 consumed, and then predict incidence
                                                 of disease using dose response
                                                 model
       Monte Carlo integration is used throughout the model, and, for the first and third parts of
the model, MCMC methods were used to sample from posterior distributions which are used to
both estimate parameters in the model and to address the uncertainty associated with these
parameters.  In complex Bayesian models, MCMC is the appropriate way to do this. Both parts
two and three of the model must be re-run each time different regulations or different treatment
conditions must be considered.
                                            11

-------
       Immediately below, is a discussion of some specific issues regarding the first piece of the
model, the national occurrence distribution of Cryptosporidium.

       3.2.1.2 Panel Conclusions.  First, the Panel concludes that the occurrence modeling
appears to be both plausible and well-done. However, the Panel believes that a number of issues
need to be addressed, either by supplementing the current documents and/or modifying the
model.

       The Panel recommends that sensitivity analyses of the modeling effort (specifications of
prior distributions) be conducted and documented.  A key component in Bayesian hierarchical
models is the specification of prior distributions, which a priori, characterize the state of
knowledge about the parameters at the higher levels of the model. Little information is
contained about such priors in the  current documentation and it is not evident that the sensitivity
of the occurrence distribution and  the infectivity parameter, k, to these  priors has been assessed.
Sensitivity analyses should be conducted and documented.  Particular concerns arise when the
data are used to assess the model and direct the selection of prior distributions.  While such
practices are sometimes needed in difficult problems, they can result in underestimation  of
uncertainty due to the double use of the data.  The analysts need to be clear about whether or not
such methods were used,  and if so, how the final uncertainties may be impacted. Much of the
concern can be ameliorated through complete sensitivity analysis.

       The Panel also recommends that seasonal effects be more carefully addressed.  In the
Panel's opinion, the absence  of seasonal effects on the annual average Cryptosporidium
concentration has not been demonstrated.  The Agency should address  and clarify its
computation of the average Cryptosporidium concentration  for plants in a system over the 18-
month period for which the data were collected in the Information Collection Rule (ICR).
Averaging concentrations equally  over the 18 months to obtain an annual average will only give
an unbiased estimate of the true annual average if there are no seasonal effects.  But the absence
of seasonal effects has not been demonstrated. The current  approach effectively counts six
months twice in the averaging5. During discussions at the DWC meeting in December 2001,
Agency representatives stated that parameters characterizing seasonality were included in the
model (in the form of the turbidity term). This problem might be solved by averaging the data
by month, and then to using the mean of the resulting twelve monthly averages  as the annual
average.

       The Panel believes that a number of other improvements would also strengthen the
Agency's LT2ESWTR documentation.  Additional model checking should be conducted. The
current Agency report includes some model-checking using the estimated distributions of true
concentrations, but the Panel recommends additional model checking, specifically, an additional
internal check and an external check.  The internal check could use the current output from the
MCMC sampler to sample from the distribution of predicted oocyst counts ("7") (from the
posterior predictive distribution of "F') . To assess how consistent predictions  from the model
5 Averaging is desired for 12 months for an annual average so the data are averaged by month first (6 months of
averaging two values and six months of just one value) and then averaging across months.
                                           12

-------
are with the observed data, about twenty sample distributions can be plotted versus the observed
distribution of counts. The observed distribution ideally should lie within these 20 and should
look similar. For an external check, the current model could be fit to the first 12 months of the
18 month ICR data, then months 13-18 could be predicted by the model and finally these
predictions could be compared to the observed data.

       There are some additional features that could be included in the documentation to
improve the clarity of the Agency's analyses. A map of the sites for both the ICR and
Information Rule Supplemental Survey (ICRSS) data would be helpful to see how similar the
spatial distribution of sites was across the surveys and to also look for spatial similarity in
concentrations for sites close together and/or in the same regions of the country.  In addition, the
Panel recommends that a short paragraph be added  documenting the convergence and mixing
checks on the MCMC sampler. An additional issue of moderate importance is that several
parameters that were included in the modeling of oocyst levels in filtered water are excluded in
the discussion of the model for the unfiltered plants (e.g., turbidity). Justification for this would
improve the clarity of the Agency's analysis.

       The Panel notes that the Agency approach to concisely summarize the occurrence
distribution functions using parametric models, in particular the log normal function, was done
to simplify computations for the individuals conducting the risk analysis.  Documentation could
be made available to confirm that the realizations of the cumulative distribution functions
(CDFs) from the MCMC sampler were well approximated by log-normal cumulative distribution
functions (CDFs).  Second, several ad hoc simplifications were done to sample the CDF for the
risk analysis (see bottom of p. 5-15 of the economic analysis document, The Cadmus Group, Inc.
200Ib). The Panel recommends that these be examined carefully for their plausibility and the
conclusions documented.

       The Panel concluded that there is a large amount of uncertainty in the modeling of the
occurrence ofCryptosporidium.  For example, the occurrence distributions are estimated based
on only one year of data. This will be fine if these distributions are stable over years.  However,
the current data does not allow determination if the particular year in which the data were
collected were aberrant (for example,  due to weather patterns) or if there is some sort of trend in
occurrence over time.  In addition,  for the infectivity modeling, the distribution of infectivity
across strains is estimated based on only three Cryptosporidium strains which may or may not be
a random sample of strains.  The only way this distribution can be estimated is to make a strong
assumption about its form (here it is assumed to be  log-normal).  The ultimate accuracy of the
predicted decrease  in disease from these stochastic  models relies on both the representativeness
and applicability of the observed data and the numerous modeling assumptions that were made
in the course of the three pieces of the model discussed at the beginning of this section.  This
qualification should be noted in the document.
                                           13

-------
3.3 Charge Question 2:  Pre- and post-LT2ESWTR Cryptosporidium risk assessment

       Agency requested SAB comments on thepre- andpost-LT2ESWTR Cryptosporidium
       risk assessment.

       The Agency provided the Panel with partial drafts of documents entitled: 1) Economic
Analysis for the Long Term 2 Enhanced Surface Water Treatment Rule (The Cadmus Group,
Inc., 200 Ib) and 2) Appendices to the Economic Analysis for the Long Term 2 Enhanced Surface
Water Treatment Rule (The Cadmus Group, Inc., 2001c). These documents show how the
Agency estimated the incidence of endemic cryptosporidiosis attributable to drinking water both
prior to and following implementation of the LT2ESWTR. Information in the documents
considered by the Agency to be of particular relevance included:

       a)     a summary of the LT2ESWTR to be proposed, based on the Stage 2 —DBF
             Advisory Committee Agreement in Principle;
       b)     baseline information used to conduct the risk assessment;
       c)     descriptions of how the Agency modeled pre- and post-LT2ESWTR risk of
             cryptosporidiosis;
       d)     a summary of how the Agency predicted the technologies that filtered and
             unfiltered systems would select to comply with the LT2ESWTR;
       e)     descriptions of how the Agency estimated the percentage of plants expected to
             receive 0.5 and 1.0 log additional Cryptosporidium treatment credit under the
             LT2ESWTR;
       f)     details on estimates of the percent of systems that would be assigned to different
             bins as a result of source water monitoring under the LT2ESWTR;
       g)     distributions of risk of illness;
       h)     unit costs for treatment technologies;
       i)     descriptions of the  methodology used to forecast the percentage of plants assigned
             to a given bin that would select a particular technology;
       j)     results of the technology selection forecast;
       k)     total treatment costs for different system categories associated with different
             regulatory alternatives and assumptions about technology availability.

       3.3.1 Panel Response to LT2ESWTR Charge Question 2. This SAB review panel
included experts in statistical modeling, in public health microbiology and engineering, but it did
not include specialists in quantitative microbiological risk analysis, a relatively new field. For
the review, each of the basic elements of microbial risk assessment was examined in order:
hazard identification, dose-response assessment, and exposure assessment.  Then the outcome of
the risk assessment was evaluated. Two criteria were considered in the Panel evaluation: a)
whether the Agency assumptions were transparent, and b) whether scientific evidence exists to
support the conclusions.

       Cryptosporidiumparvum has been responsible for significant waterborne disease
outbreaks, and it is likely that the  organism is responsible for significant endemic disease as
well.  Both of these outcomes are  important.  The current form of the Agency's analysis (The

                                           14

-------
Cadmus Group, Inc., 2001b) for the LT2ESWTR does an excellent job of addressing the impact
of drinking water quality on the incidence of endemic disease and the health risk reduction that
will result from the reduction of endemic disease as a result of the proposed regulation.  The
Agency is to be congratulated for this ground-breaking work.

       On the other hand, in the present draft, neither the design of the regulation nor the
contents of the Agency analysis directly address waterborne outbreaks. These outbreaks are the
primary stimulus for the regulation and reducing their occurrence should be one of the most
important potential outcomes from the regulation as well.

       The Panel recommends that the Agency conduct a systematic review of the design of the
LT2ESWTR regulation and evaluate its effectiveness in addressing waterborne outbreaks. This
review should include an examination of the causes of past outbreaks and how the proposed
regulatory framework will address those causes.  The Agency should then consider if any
changes in the framework must be made. Additional consultation with specialists in quantitative
microbial risk assessment could be of benefit to the Agency as it completes its consideration of
Cryptosporidium risks.

       3.3.1.1 Hazard Identification.  The Panel agreed with the basic information on
Cryptosporidium health effects that were presented in this section.  See pages 5-7 - 5-8 of the
Economic Analysis for the Long Term 2 Enhanced Surface  Water Treatment Rule (US EPA
200Ib). There are a few additional areas that should also be included in the analysis:

       a)    Evidence of current prevalence of endemic disease.  The Agency's analysis is
             based on reduction of endemic disease.  Some direct evidence of endemic disease
             levels would greatly strengthen the case. Perhaps the results of serological
             studies could be used to indicate about the prevalence of Cryptosporidium
             exposure/infection in the US.

       b)    Information on secondary transmission of cryptosporidiosis.  The current
             analysis does not consider secondary transmission of the disease. This decision
             should have stronger support in the documentation or should be reconsidered.
             Haas et al. (1999) present data on prevalence of secondary cases of
             cryptosporidiosis from two outbreak investigations that range from 4 - 33%.
             Other data in the published research literature, and perhaps data from the Centers
             for Disease Control may provide the basis for estimating the magnitude of
             secondary transmission [e.g., household via  child (e.g., Newman et al., 1994),
             household via adult (MacKenzie et al., 1995), child care centers, swimming pools
             (Puech et al., 2001; Sorvillo  et al., 2001); Millard, et al., 1994].  Asymptomatic
             infections may play an important role in secondary transmission of infection.
             Failure to consider secondary transmission will likely underestimate the impact of
             the LT2ESWTR on reducing the risks of cryptosporidiosis.

       c)    Age Effects. Information on the prevalence of asymptomatic Cryptosporidium
             infections by age should be included in the hazard identification. The rationale

                                           15

-------
             for including age effects is that, in general, different age groups are more or less
             prone to asymptomatic infections. Thus there may be a high prevalence of
             Cryptosporidiosis in some age groups that may not be detected if only
             symptomatic cases are evaluated.

       3.2.1.2 Dose-Response Assessment.  For the dose-response component of the risk
assessment, the Panel comments on four areas of the assessment: a) selection of a dose-response
function, b) use of the term infectivity, c) the morbidity rate, and d) the mortality rate.

       a) Clarify the Basis for  Selection of a Dose Response Function  The general
exponential model was used to characterize the dose-response relationship based on the data
from three human challenge studies.  Modeling this relationship is important for estimating the
risk of infection at low  doses because it is not economical to conduct large human  challenge
studies to directly measure infection rates. The choice of the exponential dose-response model is
reasonable and has been used in previous cryptosporidiosis risk assessments (Haas et al., 1996,
1999). But it is not clear if other models were considered and fit to the data from the human
challenge studies. The  Panel recommends that the Agency document the models that were
considered and the reasons for selecting this particular one.

       b) Clarify the Use of the Term Infectivity in the Agency Analysis. A number of
aspects of infectivity that are described in the Agency's analysis (pages 5-10) deserve further
discussion.  Among these things are:  I) the use of the proportion of the total oocysts from the
occurrence estimates that have internal structures to determine the fraction of oocysts considered
infectious, ii) the fraction of the  oocysts from the three strains of C. parvum used in the human
challenge studies (IOWA, TAMU and UCP) which were considered infectious and iii) the
relationship between the two, namely the fraction of oocysts that were infectious in the human
studies versus the fraction of the oocysts that were infectious in environmental samples (i.e., the
parameter "v" in the equation below).

       Infectivity of oocysts in the environment: The assumptions about the proportion of
infectious oocysts in the environment determine the variable "v" used in the Agency equation
for estimating morbidity:

       PM = M{l-[e(CvI/k)]n}

       Where:
       M = fraction of infections resulting in morbidity
       C = concentration of oocysts in water (oocysts/L)
       v = fraction of oocysts that are infectious
       I = volume of water ingested each day (L)
       k = infectivity parameter
       n = number of days of exposure
       PM = probability of disease
                                           16

-------
       In the occurrence data, the Agency assumed that only a proportion of oocysts detected in
the environment are infectious and that proportion was determined by use of data from
microscopic examination of the oocysts.  The proportion of Cryptosporidium oocysts in the
environment that are infectious was estimated from the ICR and ICRSS data based on
morphological appearance of oocysts and the proportion of oocysts with internal structures.
These measures are more frequently used as a measure of viability than infectivity. Viability,
usually evaluated by evidence of dye uptake, excystation or the presence of RNA, is a measure
of the organism's ability to continue to survive as a living organism. Infectivity is usually
defined as invasion and replication in a host cell, mouse model or human volunteers (analogous
to infection). The set of organisms that are infectious is a subset of the set of organisms that are
viable. Infectivity, not viability, is the relevant issue where the parameter is concerned.

       The Agency analysis also used data on infectivity from a study by LeChevallier (2000).
The data were expressed as a distribution with a range of 30 - 50%, mode = 40% (page 5-17).
There is some evidence that polymerase chain reaction (PCR) detection of Cryptosporidium
DNA in cell culture will give false positives because some oocysts may not be infectious but it is
still possible to detect their DNA. Thus,  direct detection of DNA by PCR may also pick up
noninfectious oocysts that  stick to the cell monolayer even if they have not infected the cells
(Rochelle et al., 2001; De Leon and Rochelle, 2000). The Panel recommends that a careful
analysis of these issues be  conducted and their impact on the risk reduction estimates be
evaluated.

       Infectivity of oocysts in the dose in the human challenge studies: The analysis of the
human dose-response data assumes that 100% of the oocysts in the dose were infectious.
However, it is likely that not all of the oocysts in the dose are "infectious". During its
deliberations, the Panel discussed new data on cell  culture infectivity and mouse infectivity that
shows that approximately 5% of freshly excreted oocysts from a cow are "infectious" (see Upton
et all994; Rochelle et al. 2001; Rochelle et al. 2002). It is important to clarify how the viability
and/or infectivity of the oocysts used in the dose was evaluated.  Was this based on excystation
rate or on the morphological appearance of intact oocysts? It would also be helpful to verify the
time between oocyst excretion and dosing volunteers (<2 weeks?) because this may affect the
proportion of infectious oocysts in the various doses. The Panel recommends that the Agency
clarify these details on the  conduct of the original study and include this clarification in its own
documentation.

       Use of human infectivity and cell infectivity data for the  analysis:  The Agency risk
analysis incorporates viability determinations (a much weaker technique) and direct PCR-cell
culture technique (which gives false positives).  It is important that the Agency clearly indicate
that human challenge data  are currently limited to three strains necessitating the use of several
major assumptions in the analysis. However, several strains have been studied in cell culture
and in mouse infectivity assays.  Since it is unclear whether these strains will ever be tested in
human volunteers,  it would be of value to compare the data between human, animal and cell
culture lines. It would be useful for the Agency to  consult with a number of researchers who
have conducted infectivity studies on Cryptosporidium to gain a deeper understanding of how
animal and cell infectivity  data might supplement the data on infectivity from human challenge

                                           17

-------
studies. Further, it will be important to make broader use of statistical analysis as the Agency
seeks to compare these differing types of infectivity data. The Panel recommends using the
PCR-cell culture data as a supplement to the human infectivity data and clarify with the
investigators the strengths, limitations and use of these data.

       Proper statistical treatment of human challenge data from multiple isolates: As discussed
above, there are some major concerns with the models for infectivity across strains. There are
data from only three strains available to estimate the distribution of infectivity across strains.  As
a result, the distribution of infectivity derived from fitting the model relies heavily on both the
assumed class of distributions (log normal) used and the assumed prior distribution for the
standard deviation parameter • fwhich characterizes the variability of infectivity across strains.
The Panel believes that the Agency could use a mixture of two distributions for infectivity to
help characterize this uncertainty. The first component of the mixture will be a log normal
distribution (with probability = * ) and the second component will be a log-t distribution with
three degrees of freedom (with probability = 1 - lambda). The latter provides heavier tails and
considers more extreme values for k to be more likely.  Sensitivity analyses regarding the impact
of the prior on sigma should also be performed.

       The importance of genotype: It is correctly recognized that there are anthroponotic and
zoonotic strains of Cryptosporidiumparvum. One limitation of the infectivity data from human
challenge studies is that currently only zoonotic strains (genotype 2) have been tested to date.
However, most of the recognized Cryptosporidium outbreaks (foodborne and waterborne) have
involved human genotypes. A human challenge study with a human genotype strain (genotype
1) is currently in progress and will provide valuable data for future risk assessments. The Panel
recommends that when this data becomes available, the Agency reevaluate this risk assessment
and the dose response model.

       Variability in host susceptibility and the effect of previous infections:  Variability in host
susceptibility was not considered in the analyses of infectivity and morbidity. For example, the
Agency dose-response model takes the number of oocysts as the dose surrogate. Thus the same
approach is used to evaluate risk for infants and adults  . The Panel recommends that the risk
assessment consider explicitly the risk to susceptible populations (e.g., elderly, young,
immunocompromised, etc.).  These groups may be at greater risk of infection and/or disease due
to greater water consumption per unit body weight, less effective immune systems, etc.  Data
from outbreak investigations may provide evidence of the consequences of infection for these
populations.

       Also, the analysis assumed that the exposed population had no previous immunity to
Cryptosporidium. It is likely that the volunteers in the  human challenge study are a mix of naive
and previously exposed individuals, and that differences in host susceptibility and previous
immunity had an effect on the estimates of the dose-response parameter.  The Panel recommends
that the Agency compare its approach to this issue with the approach taken in other studies.
Differences in host susceptibility and previous immunity will have an effect on the estimates of
the  infectivity parameter "A".

                                            18

-------
       c ) Morbidity Rate (pg 5-13). The morbidity rate was defined as the probability of
illness given infection and was estimated using a triangular distribution based on a range from
Haas et al 1996.  This rate may not be accurately estimated if asymptomatic infections were not
detected in the human challenge studies. The greater the rate of asymptomatic infections, the
more the probability of illness given infection will be underestimated.

       In addition, the probability of illness given infection may be underestimated because
these data are based on challenge studies in healthy adult volunteers.  In the general population,
there may be a greater probability of developing illness given infection because the whole
population includes sensitive sub-populations that are more likely to develop symptomatic
illness given infection.

       Individuals with existing antibodies to Cryptosporidium may have a lower morbidity
rate, although, data from Okhuysen et al., (1998) does not seem to support this.  The Okhuysen,
et al.,  experiment was conducted at relatively high doses, and there are no data on the morbidity
rate at low doses in a population with previous Cryptosporidium  infection. The  high doses
employed may have overwhelmed any immune response in a way that low doses would not.  If a
significant fraction of the population carries antibodies, the incidence of disease might be
significantly reduced.

       The mortality rate in AIDS patients that was used in the economic analysis is based on
old data from the 1993 Milwaukee outbreak. Current therapy has markedly reduced
cryptosporidiosis mortality in AIDS cases. As a result, the mortality rate in this  analysis is
probably overestimated.  At the same time, the mortality rate derived from Milwaukee may be
too low for populations with a greater proportion of immunocompromised individuals.

       The Panel recommends that these questions of morbidity  rate, and their potential impact
on the analysis of risk reduction, be more thoroughly analyzed and discussed in the document.

       3.3.1.3 Exposure Assessment (pgs 5-14 - 5-24). Exposure assessment in the Agency's
analysis included estimation of: a) the distribution of total and infectious Cryptosporidium
oocysts in finished water - derived from source water levels and estimated removal/inactivation
from treatment; b) the population served by systems potentially affected by the LT2ESWTR, and
c) the distribution of individual daily average drinking water consumption.  The  Panel has a
number of comments on this assessment.

       a) Estimates of Consumption (pg 5-22) require clarification. There are a number of
questions that arise in a review of the water consumption estimates used in the analysis. These
questions should be more effectively addressed in the documentation.  They include:

       1)      Why were two distributions of consumption used? What is the difference
              between them?
       2)      Why are the median values  (1.045, 0.71) lower than previous estimates of daily
              water consumption?
                                           19

-------
       3)     Why was Distribution 1 used for the main analysis and Distribution 2 used in the
              analysis in the appendix?

       Finally, it is not clear how the daily estimated consumption was extrapolated to annual
exposure in Exhibit 5.8 (pg 5-23). Is individual consumption split between Community Water
Systems and Non-Transient Non-Community Water Systems based on the estimated proportion
of their time spent at home and at work or school or are individuals counted in both categories -
i.e., total consumption counted twice. This estimate could be refined by age group. The Agency
should examine water consumption patterns of the very young and very old because these are the
most vulnerable age groups.

       3.3.1.4 Results of the Risk Assessment.

       The estimates of risk require clarification as to the general approach to quantitative
microbial risk assessment, discussions of uncertainty and significance of assumptions made..

       Quantitative microbial risk assessment is a rapidly developing field.  Previous work
includes risk assessments by Gasman et al., (2000), Haas et al., (1999)(see in NRC 2000), Perz,
et al., (1998), and Teunis, et al., (1999) and an outbreak model done by Eisenberg, et al., (1998).
The Panel recommends that a review of these and other preceding studies (including the sources
of data, assumptions and statistical methods) be added to the document preamble.  To the extent
the approaches by these predecessors differ from the approach used by the Agency, the
significance of the differences should be discussed and the reasoning behind the choices
provided.

       In regard to discussions of uncertainty, the document should include a summary
discussion of uncertainty and variability that is more detailed than that currently presented on pg
5-26.  This discussion should include the following:

       a)     Identifying sources of uncertainty (already included on pg 5-26)
       b)     Magnitude of uncertainty
       c)     Effect of uncertainty on the estimate of risk
       d)     Sensitivity analysis of which sources of uncertainty have the greatest impact on
              the estimate and the implications of this for future research efforts.  It appears that
              uncertainty in estimates of risk and uncertainty in costs have different drivers.
              Uncertainty in estimates of risk was driven by dose-response data. Uncertainty in
              cost was driven by occurrence data (how the systems are classified into bins
              where action is necessary). Hence, it may turn out that uncertainty is much
              greater in cost than in estimates of risk or vice versa.
       v)     Identifying sources of variability (already included on pg 5-26).  Sources of
              oocysts may be different for different communities (watersheds) animal sources
              vs human sources
              1) Magnitude of variability
              2) Effect of variability on the estimate of risk
                                           20

-------
             3) Sensitivity analysis of what sources of variability have the greatest impact on
             the estimate

       In regard to the significance of assumptions, the document should also include a
discussion of which assumptions may lead to an underestimate or overestimate of the risk and
the benefits of the proposed regulation. For example, because the analysis only considered
morbidity and mortality as outcomes, it is possible that the benefit is underestimated because the
benefit of avoided infection was not considered. Avoiding infection in the community will
reduce the potential for secondary transmission and additional cases and deaths. From a public
health perspective,  infection is the key outcome.

3.4 Charge 3:  Treatment credits for four microbial toolbox options

       The Agency requested SAB comments on the treatment credits for four specific
       technologies included among its microbial toolbox options.

       The Agency provided the Panel with drafts of portions of the preamble to the
LT2ESWTR, including: a) a Microbial toolbox overview (US EPA  200la), b) Off-stream raw
water storage (US EPA, 200 Ib), c) Pre-sedimentation (US EPA 200 Ic), d) Lime softening (US
EPA, 200Id), and e) Lower finished water turbidity (US EPA 200le).

       These draft  documents were intended to provide the Panel with an understanding of the
role and context of toolbox options in the LT2ESWTR and specific  information on each of the
four toolbox options that the Agency asked the Panel to comment upon.

       3.4.1 Panel Response to LT2ESWTR Charge Question 3

       a) Comments on the Four Options. The Panel commends  the Agency, as well  as the
stakeholder process used, for developing the bin classification framework for identifying the
treatment requirements for drinking water and the microbial toolbox containing possible
treatment options to guide systems having treatment needs. These  alternatives add great
flexibility for meeting varying water quality and treatment options and should result in safer
drinking water for the people of the United States.

       The Agency charged the Panel with evaluating Agency information on four of the
toolbox options: 1) off stream raw water storage; 2) pre-sedimentation, 3) lime softening and 4)
lower finished water turbidity.   Specifically, the Agency asked the Panel to comment on the
credits that have been proposed for specific toolbox options for Cryptosporidium removal.  The
proposal requires that utilities monitor the oocyst densities in their raw water supplies. It then
classifies each supply into one of several bins depending on the oocyst densities observed, each
bin having different treatment removal requirements. The proposal  then identifies a "toolbox" of
several actions that utilities can take in order to get credit for removal.  Removal credits are
generally expressed as the logarithm of the reduction required.  For  example, a 1 log credit
would correspond to 90% removal.
                                          21

-------
       In summary, the Panel recommends that no presumptive credits be given for off-stream
storage and pre-sedimentation. It does agree with giving 0.5 log credit for two-stage lime
softening if all the water is treated with both stages, and 0.5 log credit for plants that demonstrate
a turbidity level in each individual filter effluent less than or equal to 0.15 NTU in at least 95
percent of the measurements taken each month.  Details about these recommendations follow.

       Off-Stream Storage:  The data utilized by the Agency in determining the appropriate
credit for off-stream storage were derived from experiences in the United States as well as peer-
reviewed literature from elsewhere in the world. The data show that there is variability in the
removal of active oocysts in different reservoirs, due primarily to sedimentation, but also due to
inactivation within the environment, both of which are governed to some degree by temperature
and by residence time in the facility. After reviewing the supporting documentation, the Panel
does not  feel there are adequate data to demonstrate the proposed credits for off-stream storage
and therefore recommends that no presumptive credits be given for this toolbox option.
However, the Panel agrees that a particular utility should be able to take  advantage of any
removal achieved by this option by sampling after the off-stream storage facility for appropriate
bin placement.

       Pre-sedimentation: With regard to pre-sedimentation, many water treatment plants
located on surface waters having large variations in water quality utilize  pre-sedimentation as a
treatment technique to remove large quantities of suspended material prior to input to an existing
conventional treatment plant or lime softening operation. The real purpose of pre-sedimentation
is to provide for more consistent water quality prior to the conventional or lime softening
treatment. In reviewing the literature provided by the Agency, not only on Cryptosporidium, but
also  on spore removal with both pilot as well as full-scale plants, it seems that the data are
insufficient to support a 0.5 log presumptive credit for pre-sedimentation. As a result, the Panel
feels that no credit should be given for pre-sedimentation. Additionally, the Panel feels
performance criteria other than overflow rate need to be included if credit is to be given for pre-
sedimentation.  As with off-stream storage, the Panel does agree that a utility should be able to
take  advantage of this removal by sampling after the pre-sedimentation treatment process for
appropriate bin placement.

       Lime-softening: The Agency proposes a 0.5 log credit toward Cryptosporidium treatment
with lime softening plants that utilize two-stage softening. Based on the data provided, it
appears that a 0.5 log of additional Cryptosporidium removal is an average number for a two-
stage lime softening plant. Based on the data, single stage as well as two-stage lime softening
generally outperforms conventional treatment due primarily to the heavy precipitation that
occurs in lime softening reactors particularly when magnesium precipitation occurs.  By treating
water through a second precipitation reactor, additional removal should occur.  However,
depending on how the second reactor is utilized and the  chemical feeds to the second reactor, the
removal efficiencies vary significantly as presented in the literature. Therefore, the Panel
supports  an additional 0.5 log removal for two stage lime softening only  if all the water passes
through both stages.  If a portion of the water bypasses the first stage, the Panel feels there
should be no additional removal credit given.
                                           22

-------
       Lower Finished-Water Turbidity: Finally, the additional credits for lower finished water
turbidity seem to be consistent with what is known in both pilot and full-scale operational
experiences for Cryptosporidium removal.  As was contained in the Enhanced Surface Water
Treatment Rule, lowering effluent turbidity in the treated water results in lower concentrations of
Cryptosporidium. Therefore, it would be consistent to assume that even further lowering of
turbidity would result in further reductions in Cryptosporidium in the effluent from filtration
processes.  It is also logical to assume that individual filter effluent turbidity meeting a specific
criterion will provide for better water quality than for combined filter effluent meeting the same
requirement.  However, limited data were presented to show the exact removal that can be
achieved using these two operational benchmarks. Based on the data provided, the Panel
recommends that a 0.5 log credit be given to plants that demonstrate a turbidity level in each
individual filter effluent less than or equal to 0.15 NTU in at least 95 percent of the
measurements taken each month. No additional credit should be given to plants that demonstrate
a combined filter effluent turbidity of 0.15 NTU or less.

       b) Other Issues. The Panel's understanding of the approach used in developing the
microbial toolbox is as follows.  The additional log removals in the table of bin requirements are
based in part on the assumption that conventional filtration plants in compliance with the Interim
Enhanced  Surface Water Treatment Rule (IESWTR) achieve an average of 3 logs removal of
Cryptosporidium.  The Panel also understands that this assumption indicates that all
conventional treatment plants can be expected to remove a minimum of 2 logs of
Cryptosporidium. Furthermore,  it is the Panel's understanding that an objective of the rule is to
achieve an average oocyst concentration in treated surface waters of 10"4 oocysts/1 or lower.
Given the  oocyst concentrations in bins 2, 3,and 4, and considering an average removal of 3 logs
for conventional treatment, the additional removal requirements in bins 2, 3, and 4 are expected
to provide an average treated water oocyst concentration of 10"4 oocyst/1 or lower.

       This approach differs from past regulatory approaches to Giardia and Cryptosporidium
treatment credits and from present regulatory approaches to Giardia control.  Current regulations
for Giardia control provide 2.5 logs of removal credit when conventional treatment is used.  It is
the understanding of the Panel that this removal credit for Giardia is based on the minimum
removal (not the average removal) achieved by these plants.

       These differences between the IESWTR and LT2ESWTR regulations in the bases for
assuming removal credits for Giardia and Crypotosporidium are not readily apparent and should
be clarified and supported in the new regulations. Appropriate guidance will be needed for
consistent implementation of these two regulations.
                                           23

-------
              4. STAGE 2 DISINFECTION BYPRODUCTS RULE

4.1 Charge 1: Initial Distribution System Evaluation (IDSE):

       The Agency requests SAB comment on whether the IDSE is capable of identifying new
       compliance monitoring points that target high TTHM and HAAS levels and whether it
       is the most appropriate tool available to achieve this objective.

       The Agency provided the Panel with two draft documents on the Initial Distribution
System Evaluation that is to be proposed in the S2DBP rule. Information provided in support of
Charge question 2 below in this section also bears some relevance to this question. The
documents provided by the Agency include:

       a)      "E. Initial distribution system evaluation (IDSE) " (US EPA, 200If) a draft
             overview of the IDSE intended for the preamble of the rule; and
       b)     Stage 2 Disinfectants and Disinfection Byproducts Rule Initial Distribution
             System Evaluation Guidance Manual (The Cadmus Group, Inc., 200Id) which
             provides recommendations for how utilities should proceed to determine
             monitoring sites to reflect the highest levels of TTHM and HAAS occurrence
             within the distribution system.

       4.1.1 Panel Response to S2DBP  rule Charge Question 1.

       4.1.1.1  Initial Distribution System Evaluation (IDSE) Effectiveness The Panel
believes that the proposed IDSE is capable of identifying new compliance monitoring points that
target higher DBF levels than are currently monitored in the existing compliance monitoring
programs for the THM Rule and Stage 1 DBF Rule. However, the IDSE may not identify the
highest levels to which consumers in a given distribution system are exposed. The basis for the
latter statement is that the IDSE does not consider short-term, temporal variations that occur at
different sites in the distribution system due to varying (e.g. diurnal) water demands and
distribution system architecture and operation.  Distribution systems are, by their nature, highly
dynamic. Varying water demand patterns (e.g. low density and high density residential water
use, industrial and commercial water use, irrigation) and operating conditions (e.g. pumping
patterns and storage tank operations) normally lead to appreciable temporal and spatial
variations in hydraulic residence times (water age) and water quality throughout the system that
are not captured by the proposed IDSE.  Hence, it is unlikely that a single grab sample  taken at
any site at any time will yield a representative DBF concentration for that site, and that  grab
samples taken at a number of sites will identify sampling sites with the highest DBF
concentrations.

       Further, rates of disinfection byproduct formation and degradation are temperature-
dependent and may change on a seasonal basis. Coupling this with the fact that water demand
patterns, and therefore hydraulic residence times, also may change with season may mean that
peak DBF levels migrate from the remote parts of the system during colder months to interior
                                          24

-------
portions of the system during warmer months. Furthermore, this behavior will probably not be
consistent for all.

       Therefore, the Panel believes that it is important that site selection be re-evaluated
periodically. In rapidly growing utilities changes in the distribution system architecture and flow
patterns are common. As a result, the sites with high DBF levels often change.  Significant
changes also occur in systems that are not rapidly growing as components fail and/or
improvements are made. If sample locations are not updated with time to reflect these changes
in distribution system behavior, then the sample locations may lose their relevance over time.
Further, the IDSE is only a 12-month program, and utilities and primacy agencies have no
assurances that the 12-month period over which the IDSE is performed will indeed be typical of
normal system operations.  The Panel recommends that temporal limitations be identified in the
documentation and that periodic re-evaluation of selected sites be required so that changes in the
system and/or its use will be addressed.

       4.1.1.2 IDSE Appropriateness. The Agency also asked if the IDSE is the most
appropriate tool to reach the objective of identifying new compliance monitoring points that
target higher THM4 and HAAS levels.  The Panel believes that the proposed standard
monitoring program (SMP) for sub-part H systems serving more than  10,000 people, in which 8
samples are collected at 2-month intervals, is reasonable. The Panel does recommend, however,
that the 8 samples be re-allocated so that, for both free chlorine and chloramines, 3 samples be
taken at potential high THM4 sites, 3 samples be taken at potential high HAAS sites, and only 1
sample each be taken at an average site and at the point of entry to the system. If indeed the
objective is to locate and monitor the sites with high THM4 and high HAAS concentrations,
more samples need to be allocated to this objective.  One point of entry site is sufficient to gauge
the initial concentration of DBFs entering the system, and only one "average" site should be
sufficient to maintain connectivity to the existing compliance monitoring program.  The Panel
also believes that the "average" site for the IDSE should be one of the average locations in the
existing Stage 1 DBF compliance monitoring program. This would mean that every 6 months
(twice during the IDSE), utilities would only have to take 7 samples as part of the IDSE, with the
eighth sample being one of the compliance monitoring samples.

       The Panel also recommends that the IDSE should require the measurement and reporting
of residual chlorine (free or combined) concentrations at the time of DBF sample collection, and
that individual THM and HAA species be reported in addition to the aggregate concentrations.
The Panel also suggests that the IDSE recommend that complementary pH, temperature, and
heterotrophic plate count be measured and recorded concurrently with DBF measurements.
Such information will prove to be valuable to the utilities, the primacy agencies, and the Agency
in the future.

       With respect to time of sample collection, there is no reason to believe that THM4 or
HAAS levels will be highest in the morning. In view of the dynamic and highly complex nature
of water distribution systems, it is equally likely  that THM4 or HAAS levels at some locations
will be highest in the evening. The Committee recommends that the reference to time of sample
                                          25

-------
collection be omitted from the Guidance Manual (e.g. p. 2.9 of Guidance Manual) and be left to
the discretion of the utilities and their respective primacy agency.

       The Panel also recommends that the Agency provide more guidance to the utilities with
respect to identifying potential sampling sites with the highest HAAS concentrations. The only
reference in which some guidance is provided is on page 5-18, line 39 of the Guidance Manual,
although that guidance is not especially clear. It might be expected that, at least in waters with
temperatures supporting microbial activity, HAAS levels may decrease when free chlorine
residuals decrease below 0.2-0.3 mg/1 or combined chlorine residuals decrease below 0.5 mg/1.
This may not be the case in cold waters in which microbial activity is minimal; in such cases,
high HAAS sites may coincide with high THM4 sites.  Distribution system dynamics, water age,
chlorine residual data, and heterotrophic plate count data should be examined in selecting sample
sites.

       The Panel also recommends that the Agency require that the selection of monitoring sites
be justified rather than simply recommending that they be justified (p. 1-4, line 14), and that the
IDSE report provide justification for the selection of sites (p. 5-24, line 16) (The Cadmus Group,
Inc., 200Id).

       The Panel believes that the proposed system specific studies (SSS) approach described in
Chapter 6 of the Guidance Manual needs improvement if sound guidance is to be provided to the
utilities. Water consumption (demands) should be more accurately simulated in the network
model, given the availability of such information. It is important to realize that different types of
water users will consume water at different times and rates during the day. Water demands
should be classified and allocated based on their water use type (domestic, industrial,
commercial, etc.) and each type of water user should be assigned an individual water use pattern
over a 24-hour (or other) period. Estimates of demand distributions could be obtained by using
land use information or by using a water meter or assessor's parcel number location
methodology (geocoded meter location).  For example, the land use computation method
consists of intersecting demand area polygons with land use polygons using water duty factors to
create water demands for selected analysis nodes. The geocoded meter location method consists
of grouping water billing data into demand areas around analysis nodes by using a spatial
reference of water meters, yielding a credible demand distribution as demands are allocated per
customer billing accounts (and automatically taking into account vacant parcels and large water
users). Other spatial demand allocation methods include assigning geocoded customer meters to
the nearest analysis node or to the nearest pipe  and then split the demand among the bounding
analysis nodes.  Some care will be required to ensure that demands are accurately allocated
according to actual spatial consumption.

       4.1.1.3 Other Considerations. The Panel has a number of concerns that it considers to
be of significance but which do not easily fit into the other two charge questions on the S2DBP
Rule. These are discussed in the following paragraphs.
                                           26

-------
       Clarification of Assumptions: A number of assumptions and policy decisions were made
in the development of the form of the Stage 2 DBF Rule and the IDSE. These need to be stated
at the outset and made clear throughout the documentation in support of the rule. These include:

       a)     the decision to continue to regulate THM4s and HAASs collectively as group
             parameters rather than as individual species;
       b)     the decision to continue to regulate only five of the HAAs (HAAS) rather than all
             nine bromine- and chlorine-containing HAAs (HAA9);
       c)     recognition of the fact that, for purposes of simplicity, the IDSE overlooks short-
             term temporal variability in the selection of sites for locating and monitoring
             maximum levels of THM4s and HAASs;
       d)     recognition of the fact that sampling and monitoring costs were key
             considerations in designing the requirements for the standard monitoring program
             for the IDSE;
       e)     recognition of the fact that, although the Source Water Analytical Tool (SWAT)
             model was developed for modeling the effects of treatment on DBF formation and
             was not developed to model changes in individual or aggregate DBF
             concentrations in distribution systems, it was the only tool that the Agency had
             for purposes of the benefits analysis in support of the Stage 2 Rule.

       Use of the SWAT Model: In the risk reduction analysis, the SWAT model is used to
predict monthly DBF concentrations both under current conditions and under conditions where
plant modifications have been made to meet the requirements of Stages 1 and 2 (sections 3.7.2
and 5.4.1.1) (The Cadmus Group, Inc., 2001e). This use of the SWAT model would be
appropriate if it could be relied upon for valid predictions in such applications.  Unfortunately,
the Agency has not demonstrated that this is the case. Large discrepancies exist between SWAT
predictions and ICR data, and these discrepancies raise serious questions regarding both the
accuracy of the SWAT model and the adequacy of attempts to characterize DBF concentrations
of dynamic systems with such a limited number of samples (four sites with four samples per
year).

       Two aspects of data presentation in the Stage 2 DBPR Economic Analysis serve to
illustrate how the discrepancies are under-represented: a) the use of cumulative frequency
distributions (pages 3-31 and A-18  through A20)(The Cadmus Group, Inc., 200le), and b)
miscalculation of "mean predicted errors" (page A-34 and Exhibit A.21).  The problem with the
use of cumulative frequency diagrams is that such plots have the same shape even when paired
values have little agreement. Plants with low THM4 or HAAS from the SWAT model  are not
necessarily the same plants with low THM4 or HAAS plants from the ICR data. This
discrepancy is totally lost when the data are presented as cumulative frequency curves. In the
calculation of the "mean predicted error," "the absolute value of the difference between "SWAT
annual plant mean" and "ICR annual plant mean" should have been used instead of signed
values, or an R2 value should have been calculated. The way the calculation was done, positive
deviations canceled out negative deviations thereby grossly underestimating "mean predicted
errors." The graphical results of pages A-23 to A33 convey a much greater sense of the
discrepancies between the SWAT predictions and the ICR data.  The magnitude of these

                                          27

-------
discrepancies diminishes the value of the subsequent use of either SWAT or ICR data in
Economic Analyses or risk reduction calculations.

       The limitations to the model's accuracy arise from the inherent limitations of the existing
state of the art for predicting DBF concentrations from water quality data and/or the inherent
limitations in the available database, and hence cannot be easily fixed. Under the circumstances,
the contribution that the model can make to an evaluation of the risk reduction from the Stage 2
rule is marginal at best. The Panel recommends that either this portion of the analysis of the risk
reduction be eliminated or that the presentation be altered to reflect the uncertainties associated
with use of the model.

       Monitoring  Frequency Under the IDSE: Though this is a relatively minor point, it should
be made clear, in all documents relevant to the Stage 2 Rule, that quarterly monitoring of DBFs
means every 3 months.  For example, Table 5.4 and page 192 (US EPA, 2001h) do not
unequivocally indicate that the basis for the LRAA calculation is sampling at 3-month intervals
rather than once each quarter as in the current THM Rule and Stage 1 Rule.

4.2 Charge 2: Public Health Protection of S2DBPR.

       4.2.1 Panel  Response to S2DBPR Charge Question 2.

       The Agency requests SAB comment on whether the locational running annual average
       (LRAA) standards for total trihalomethanes (TTHMs) and haloacetic acids (HAAS), in
       conjunction with the Initial Distribution System Evaluation of the proposed S2DBP
       rule, more effectively achieves public health protection than the  current running
       annual average (RAA) standards, given the existing knowledge of DBF occurrence and
       the available health effects data.

       The Agency is concerned with reproductive, developmental, and carcinogenic effects
       which are associated with TTHMs andHAAs.  The Agency intends to reduce the
       variability of exposure to DBFs for people at different points in the distribution system,
       and therefore reduce risks.

       The Agency provided the Panel with documents that gives the Agency's case for why it
believes there is a health concern for disinfection byproducts. Documents provided to the Panel
in support of the Health concerns determination include:

       a)     A draft of preamble section "///. Public Health Risk " (US  EPA, 2001 g) that
             briefly discusses reproductive and developmental epidemiology information
             received after the Stage 1 DBF rule;
       b)     Quantification of Bladder Cancer Risk from Exposure to Chlorinated Surface
             Water (US EPA, 1998) which provides details on the population attributable risk
             concept used to quantify the estimated number of cancer cases that would be
             attributable to the consumption of chlorinated drinking water;
                                          28

-------
       c)     Reproductive and Developmental Effects of Disinfection By-Products (Reif et al.,
             2000) which provides a critical review of the epidemiologic literature pertaining
             to reproductive and developmental effects of exposure to disinfection byproducts
             in drinking water;
       d)     Review of Animal Studies for Reproductive and Developmental Toxicity
             Assessment of Drinking Water Contaminants: Disinfection By-Products (DBFs)
             (Tyl, 2000), which provides a review of the animal reproductive and
             developmental toxicity data on disinfection byproducts; and
       e)      "V.  Discussion of Proposed Stage 2 DBPR Requirements " (US EPA, 2001 h)
             which explains how the chloroform MCLG was developed.

       One document was provided to support evaluation of charge question 2 in the area of
"Occurrence/Reduction of Peaks":

       a)     Excerpts from the Economic Analysis for the Stage 2 Disinfectants and
             Disinfection Byproducts Rule (The Cadmus Group, Inc., 200le) which indicates
             the extent to which the Agency estimates that DBF peaks might be reduced by the
             proposed S2DBPR.

       One document was provided to support evaluation of charge question 2 in the area of
"Monitoring Requirements and Compliance Determination:":

       a)     G. Monitoring requirements and compliance determination. (US EPA, 2001i).

       The Agency issued a Stage 1 DBF regulation  that requires regulated water systems to
meet a standard of 80 ug/1 Total Trihalomethanes (THM4) and 60 ug/1 for five Haloacetic Acids
(HAAS) as well as other DBFs during  1998.  Consistent with the original THM rule, the
regulation requires that systems  implement a Running Annual Average (RAA) approach to
monitoring for these contaminants and that they be kept at or below these levels.  In arriving at
these standards, the Agency recognized, as does this Panel, that the regulated THM4 and HAAS,
which are prominently identified in the rule, are not the only DBFs in these classes that could be
in drinking water, nor are these classes the only possible DBFs in chlorinated or other drinking
water systems. However, the Agency and a large group of stakeholders who were involved in an
extensive series of negotiations, agreed that it was appropriate to focus on these DBFs in the
policy embodied in the Stage I standard.  They further agreed that it was reasonable to assume
that the controls that would be implemented for reducing levels, and therefore risks, of those
regulated DBFs, would also reduce risks from other DBFs that are, as yet, to be identified and/or
studied for health effects.

       The panel is generally supportive of the THM4 and HAAS actions under consideration.
Although the epidemiology data  associating cancer with chlorinated drinking water has resulted
in relatively small odds ratios, the observations have  now been consistent across a broad number
of studies with varying degrees of increasing sophistication, especially for bladder cancer.
While the odds ratios are small, the numbers of attributable cases are large relative to other
environmental issues of concern  (Morris et al., 1992; Poole, 1997).  Therefore, the epidemiology

                                          29

-------
data can be taken to indicate that there is a problem that needs to be taken seriously.  The THM4
and HAAS standards reviewed by the Panel are a constructive interim step towards addressing
this problem.

       The Panel also agrees that establishing an LRAA would be expected to reduce exposure
to the nine compounds that are regulated. As detailed in section 4.1.1.1 of this document, which
discusses the dynamics of water movement through the distribution system and on-going
production and degradation of disinfection by-products, it is uncertain that the requirements of
the IDSE will result in a sufficiently complete distribution system characterization to be
confident that the locations with the highest exposure will be identified and therefore that all the
households will gain the protection of the new standards. Nevertheless, the variability in
exposure to regulated DBFs, from one point in the system to another, will be reduced,
particularly at the extreme locations that the IDSE does identify, and the consumers at those
locations will have lower levels of exposure to the measured DBFs.

       The principle outcome of the LRAA/IDSE proposal will be increased assurance that each
consumer  will be exposed to THM4/HAA5 levels that  are at or below the MCLs specified. The
existing RAA allows locations with THM4/HAA5 levels above the MCL to be averaged with
other locations in the system that do not.  The LRAA identifies locations in the system with
consistently high concentrations of the regulated DBFs and requires that they meet the MCL.
Thus the new proposal substantially reduces the probability that a given consumer will be
exposed to THM4/HAA5 levels above those specified  in the regulation.  The Panel recommends
that the Agency give greater visibility to this benefit.

       A second, but important outcome of the LRAA/RAA proposal will be reduced overall
average level of the regulated DBFs in many systems.  This will occur because, when systems
use precursor removal  as their strategy, THM4/HAA5  levels must be reduced throughout the
system in  order to bring sampling points with high THM4/HAA5 levels into compliance.

       Assessments of health  risk reduction from this rule have emphasized reductions in
bladder cancer risk. It is important to address bladder cancer because epidemiological data
consistently indicate that lifetime consumption of chlorinated surface water poses a bladder
cancer risk (Cantor et al., 1998; Doyle et al.,  1997; King and Marrett, 1996;  McGeehin et al.,
1993;JVIorris et al., 1992; Poole, 1997; Vilanueva et al., 2001; Vena et al., 1993).  There are
other serious putative health effects that have been identified from epidemiology studies or
toxicological studies of individual disinfection byproducts (Cantor et al., 1999; Hildesheim et al.,
1998; King et al., 2000; Reif et al., 2000; Tyl, 2000). These include risks of other cancers (brain,
colon, rectal), impairment of male and female reproduction, and effects on developing
organisms. Additionally, it should be noted that the brominated THM and HAAs might account
for some of the  colon cancer as they can produce colon cancer in rats. Collectively, the risks
calculated from these toxicological studies are 1-2 orders of magnitude less than the bladder
cancer risks indicated by the epidemiology studies. The bladder cancer may well be due to
agents other than the THM4 and HAAS species (Bull et al., 2001) While based on more limited
evidence,  reductions in reproductive health risks are considered to be a benefit of the rule;
however the lack of data preclude quantification of this benefit.

                                           30

-------
       On the other hand, the panel cautions that the Agency has not satisfactorily demonstrated
that promulgating the S2DBP rule will result in the reduction in bladder cancer risk which has
been projected. The following are the reasons for this statement:

a)     The disinfectant by-product mixture produced when water is chlorinated is extremely
       complex, and within a given system, varies considerably.
b)     The specific by-products resulting in increased bladder cancer have not been identified,
       but are unlikely to be accounted for by the aggregate THM4 or HAA56 concentrations.
c)     It has not been demonstrated that actions taken to control the collective THM4 and
       HAAS concentrations will also control other known and unknown by-products.
d).     Treatment technologies may emerge that target only the regulated by-products, without
       addressing the rest of the DBF mixture.
e)     Some technologies aimed at reducing the target DBFs might result in new DBFs of
       unknown significance7.

       In summary, it is the Panel's opinion that cancer and reproductive health risks are likely
to result from water chlorination.  However, the Agency has not demonstrated that the health risk
reductions that accrue from the proposed rule will be proportional to the reductions in the THM4
and HAAS concentrations. Some health benefits in addition to those specifically attributable to
these classes of DBFs could accrue, but only to the extent that the measures that water systems
take to reduce these byproducts also reduce the concentrations of other byproducts. It should be
remembered that changing treatment has some  potential to  change the by-product mixture
produced and some of the new compounds generated could be more harmful.  Nevertheless, the
Panel believes that some risk reduction will occur and that  speculation such as that discussed
above should not delay the promulgation of the present rule.
 For example, the target DBFs being regulated may not be good surrogates for the compounds that produce the reproductive toxicities. The risks
identified in the epidemiology studies are much greater than those suggested by the studies of these individual by-products in animals. It is
important to note, that the target DBFs do not include the most potent reproductive toxicant among the DBFs examined to date,
bromochloroacetic acid..

 The recent identification of N-nitroso-N-dimethylamine (NDMA) as a by-product of chloramination is an example. NDMA belongs to a class of
chemical carcinogens which contains some members that are known to produce bladder cancer in rats. NDMA is between 3 and 4 orders of
magnitude more potent as a carcinogen than the THM4 and HAAS (U.S. EPA,1997). Perhaps the most common method used for controlling
THM4 and HAAS formation is to use chlorine combined with ammonia for residual control.  Recent work has shown that this combined chlorine
can result in increased NDMA formation (Najm and Trussell, 2002, Choi and Valentine, 2002, Mitch and Sedlak, 2002).
                                               31

-------
                                   REFERENCES
Bull, R.J., Krasner, S.W., Daniel, P.A., and Bull, R.D. (2001). Health Effects and Occurrence of
       Disinfection By-Products.  Awwa Research Foundation and American Water Works
       Foundation, Denver, CO, pp. 1-13.

Cantor, K.P., Lynch, C.F., Hildesheim, M.E., Dosemeci, M., Lubin, J., Alavanja, M., and Craun,
       G. (1998). Drinking water source and chlorination by-products. I. Risk of bladder
       cancer.  Epidemiology, 9:211-228.

Cantor, K.P., Lynch, C.F., Hildesheim, M.E., Dosemeci, M., Lubin, J., Alavanja, M., and Craun,
       G. (1999). Drinking water source and chlorinated by-products.  III. Risk of brain cancer.
       Am J Epidemiology, 150:552-560.

Casman, E., Fischhoff, B., Palmgren, C., Small, M., and F. Wu (2000).  Integrated Risk Model
       of Drinking Waterborne Cryptosporidiosis Outbreak.  Risk Analysis, v 20, pp 493-509,
       2000.

Choi, J. H.; Valentine, R. L. (2002). "Formation of-nitrosodimethylamine (NDMA) by reaction
       of monochloramine in a model water: a new disinfection by-product."  Wat. Res. V36,
       pp817-824.

De Leon, R. and P. Rochelle. (2000) Quantitative Cell Culture-based Infectivity Assay for
       Cryptosporidium parvum in Water. Report of a STAR grant. Metropolitan Water
       District of Southern California. Water Quality Laboratory, La Verne, CA; R-825146.

Doyle, T.J., Zheng, W., Cerhan, J.R., Hong, C.-P., Sellers, T.A., Kushi, L.H., and Folsom, A.R.
       (1997).  The association of drinking water source and chlorination by-products with
       cancer incidence in postmenopausal women in Iowa: A prospective cohort study.  Am J
       Public Health, 87:  1168-1176.

Eisenberg JN, Seto EY, Colford JM Jr, Olivieri A, Spear RC. (1998). An analysis of the
       Milwaukee Cryptosporidiosis outbreak based on a dynamic model  of the infection
       process. Epidemiology. 9(3):228-31.

Haas, C.N., C. Crockett, J.B. Rose, C. Herba, A.Fazil. (1996). Infectivity of Cryptosporidium
       parvum oocysts. Journal WWA 88(9): 131-136.

Haas, et al.  (1999). NRC (2000) Committee to Review the New York City Watershed
       Management Strategy. (2000) Watershed Management for Potable Water Supply:
       Assessing New York City Approach.  Water Science and Technology Board, National
       Research Council,  National Academy Press, Washington, DC.
                                         R-l

-------
Hildescheim, M.E., Cantor, K.P., Lynch, C.F., Dosemici, M., Lubin, J., Alavanja, M. and Craun,
       G. (1998). Drinking water source and chlorination by-products: II. Risk of colon and
       rectal cancers. Epidemiology 9:29-35.

King, W.D. and Marrett, L.D. (1996).  Case-control study of water source and bladder cancer.
       Cancer Causes Control, 7:596-604.

King, W.D., Marrett, L.D., and Woolcott, C.G. (2000).  Case-control study of colon and rectal
       cancers and chlorination by-products in treated water.  Cancer Epidemiology, Biomarkers
       & Prevention. 9:813-818.

LeChevallier, M.W., J.L. Clancy, Z. Bukhari, S. Bukhari, T. Hargy, J.S. Rosen, J. Sobrinho,
       M.M. Frey (2000). Source Water Assessment: Variability of Pathogen Concentrations
       (paper presented at the 2000 WQTC).

MacKenzie WR, Schell WL, Blair KA, Addiss DG, Peterson DE, Hoxie NJ,  Kazmierczak JJ,
       Davis JP (1995). Massive outbreak of waterborne cryptosporidium infection in
       Milwaukee, Wisconsin: recurrence of illness and risk of secondary transmission. Clin
       Infect Dis 21(l):57-62

McGeehin, M.A., Reif, J.S., Becher, J.C., and Mangione, EJ. (1993).  Case-control study of
       bladder cancer and water disinfection methods in Colorado.  Am J Epidemiol, 138:492-
       501.

Mitch, W. A., Sedlak, D. L. (2002).  "Formation of-nitrosodimethylamine (NDMA) from
       dimethylamine during chlorination." Environ. Sci. Technol.V36, pp588 -595.

Millard P, K Gensheimer,  DG Addiss, DM Sosin, GA Beckett, A HouckJankoski, A Hudson
       (1994).  An outbreak of cryptosporidiosis from fresh pressed apple cider JAMA
       272(20): 1592-1596]

Morris, R.D., Audet, A.-M., Angelillo, IF., Chalmers, T.C., Mosteller, F. (1992). Chlorination,
       chlorination by-products, and cancer: a meta-analysis.  Am J Public Health; 82: 955-63.

Najm, I, Trussell, R. R. (2001).  "NDMA Formation in Water and Wastewater." Journal
       AWWA,  V93, pp 92-99, 2001, February, 92-99.

National Research Council (NRC, 2000). Watershed Management for Potable Water Supply:
       Assessing the New York City Strategy. Chapter 6. Tools for Monitoring and Evaluation.
       Committee to Review the New York City Watershed Management Strategy, Commission
       on Geosciences, Environment and Resources. National Academy Press, Washington DC.

Newman RD, Zu SX, Wuhib T, Lima AA, Guerrant RL, Sears CL (1994).  Household
       epidemiology of Cryptosporidium parvum infection in an urban community in northeast
       Brazil.  Ann Intern Med 120(6):500-5

                                          R-2

-------
Okhuysen, PC, CL Chappell, CR Sterling, W Jakubowski and HL DuPont (1998).
       Susceptibility and serologic response of healthy adults to reinfection with
       Cryptosporidium parvum.  Infection and Immunity. 66(2): 441-443.

Perz, JF, JK Ennever and SM LeBlancq (1998). Cryptosporidium in tap water: Comparison of
       predicted risks with observed levels of disease.  American Journal of Epidemiology.
       147(3):289-301.

Poole, C. (1997).  Analytic meta-analysis of epidemiologic studies of chlorinated drinking water
       and cancer: Quantitative review and reanalysis of the work published by Morris et al.,
       Am J Public Health 1992; 82:955-63.  September 30,  1997. Cincinnati, National Center
       for Environmental Assessment.

Puech MC, McAnulty JM, Lesjak M, Shaw N, Heron L, Watson JM. (2001). A statewide
       outbreak of cryptosporidiosis in New South Wales associated with swimming at public
       pools. Epidemiol. Infect.  126(3):389-9]

Reif, JS, Bachand, A, and Andersen, M (2000). Reproductive and Developmental Effects of
       Disinfection By-Products.  Prepared for Health Canada.  Colorado State University, Dept.
       of Environmental Health, Fort Collins, CO, October 31,  2000.

Rochelle, PA, DM Ferguson, DM Johnson, and R. DeLeon (2001). Quantitation of
       Cryptosporidium parvum infection in cell culture using a colorimetric in situ
       hybridization assay.  J. Eukaryot. Microbial. 48:565-574.

Rochelle, PA, MM Marshall, JR Mead, AM Johnson, DG Carrick, JS Rosen and R. DeLeon.
       (2002). Measurement of Cryptosporidium parvum infectivity: in vitro cell culture
       compared to a mouse assay. Applied and Environmental Microbial. submitted.

Sorvillo, JF, MPH, K Fagaceae, PhD, M Torrey, MPH, R Kebabjian, RS, W Tokushige, L
       Mascola, MD, S Schweid, M Hillario, SH Waterman, MD (2001). Epidemiologic Notes
       and Reports Swimming-Associated Cryptosporidiosis — Los Angeles County. MMWR
       39(20);343-345

Teunis, PFM and AH Havelaar (1999).  Cryptosporidium in drinking water: Evaluation of the
       ILSI/RSI quantitative risk assessment framework. Rijksinstituut voor Volksgezondheid
       en Lilieu, RIVM report number 284550006.

The Cadmus Group, Inc. (2001a). Occurrence and Exposure Assessment for the Long Term 2
       Enhanced Surface Water Treatment Rule. Report under Contract 68-C-OO-l 13, August
       27, 2001, Fifth Draft.

-------
The Cadmus Group, Inc. (200 Ib). Economic Analysis for the Long Term 2 Enhanced Surface
       Water Treatment Rule 9. Report under Contract 68-C-OO-l 13, August 27, 2001, Fifth
       Draft.

The Cadmus Group, Inc. (200 Ic). Appendices to the Economic Analysis for the Long Term 2
       Enhanced Surface Water Treatment Rule.  Report under Contract  68-C-OO-l 13, August
       27, 2001, Fifth Draft.

The Cadmus Group, Inc. (200 Id). Stage 2 Disinfectants and Disinfection Byproducts Rule
       Initial Distribution System Evaluation Guidance Manual. Report under Contract 68-C-
       99-206, August 22, 2001, Fourth Draft.

The Cadmus Group, Inc. (200 le). Excerpts from the Economic Analysis for the Stage 2
       Disinfectants and Disinfection Byproducts Rule. Report under Contract 68-C-99-206,
       August 24, 2001, Seventh Draft revised.

Tyl, RW (2000). Review of Animal Studies for Reproductive and Developmental Toxicity
       Assessment of Drinking Water Contaminants: Disinfection By-Products (DBFs).
       Research Triangle Institute. October 12, 2000.

Upton, SJ, M Tilley and DB Brillhart (1994). Comparative development of Cryptosporidium
       parvum in 11 continuous cell lines.  FEMSMicrobial. Letters. 118:233-236.

USEPA (1997).  Integrated Risk Information System. N-Nitrosodimethylamine; CASRN 62-75-
       9.

US EPA (1998). Quantification of Bladder Cancer Risk from Exposure to Chlorinated Surface
       Water. US EPA, Office of Water, Office of Science and Technology.  August 26, 1998.

US EPA (2000). Stage 2 Microbial and Disinfection Byproducts Federal Advisory Committee
       Agreement in Principle. FR65, No. 251:pp83015-83024, December 29, 2000.

US EPA. (200la). Microbial toolbox overview. Draft document prepared by the US EPA Office
       of Ground Water and Drinking Water and transmitted to the SAB  via an email from Dr.
       D. Schmelling, September 4, 2001.

US EPA. (200Ib).  Off-stream raw water storage. Draft document prepared by the US EPA
       Office of Ground Water and Drinking Water and transmitted to the SAB via an email
       from Dr. D. Schmelling, September 4, 2001.

US EPA (200Ic). Pre-sedimentation.  Draft document prepared by the US EPA Office of
       Ground Water and Drinking Water and transmitted to the SAB via an email from Dr. D.
       Schmelling, September 4, 2001.
                                        R-4

-------
US EPA (200Id). Lime softening. Draft document prepared by the US EPA Office of Ground
      Water and Drinking Water and transmitted to the SAB via an email from Dr. D.
      Schmelling, September 4, 2001.

US EPA (200 le). Lower finished water turbidity.  Draft document prepared by the US EPA
      Office of Ground Water and Drinking Water and transmitted to the SAB via an email
      from Dr. D.  Schmelling,  September 4, 2001.

US EPA (2001f). Draft of preamble section . Initial distribution system evaluation (IDSE)
      Undated.

US EPA (200Ig). Draft of preamble section //. Public Health Risk  Undated.  Pages 77-116.

USEPA(2001h). Draft of preamble section . Discussion of Proposed Stage 2 DBPR
      Requirements. Undated. Pages 155-202.

US EPA (2001i). Draft of preamble section . Monitoring requirements and compliance
      determination.  Undated.

Venu, I.E., Graham, S., Freudenheim, J., Marshall, J., Zielezyny, M.,  Swanson, M. and Sufrin,
      G. (1993). Drinking water, fluid intake, and bladder cancer in Western New York. Arch.
      of Environ. Health, 48:191-198.

Vilanueva, C., Kogevinas, M. and Grimalt, J. (2001).  Chlorination of drinking water in Spain
      and bladder cancer. Gac. Sanit. 15:48-53.
                                         R-5

-------
                              ATTACHMENT A


                    ACRONYMS AND ABBREVIATIONS

BAT        Best Available Treatment
cdf         Cumulative Distribution Frequency
CWS       Community Water System
DBF        Disinfection Byproducts
DWC       Drinking Water Committee
EPA        U.S. Environmental Protection Agency
HAAS      Haloacetic Acids
HAN       Haloacetonitriles
ICR        Information Collection Rule
ICRSS      Information Collection Rule Supplemental Survey
IDSE       Initial Distribution System Evaluation
IESWTR    Interim Enhanced Surface Water Treatment Rule
LRAA      Locational Running Annual Average
LS          Lime Softening
LT2ESWTR Long Term 2 Enhanced Surface Water Treatment Rule
MCL       Maximum Contaminant Level
MCLG      Maximum Contaminant Level Goal
NTNCWS   Non-transient Non-community Water Systems
PCR        Polymerase Chain Reaction
POTW      Publically Owned Treatment Works
RAA        Running Annual Average
SAB        U. S. EPA Science Advisory Board
SDWA      Safe Drinking Water Act Amendments of 1996
SWAT      Surface Water Analytical Tool
S2DBPR    Stage 2 Disinfection/Disinfectant Byproduct Rule
THM       Trihalomethanes
TTHM      Total Trihalomethanes
                                       A-l

-------
                                ATTACHMENT B

                      SELECTED GLOSSARY OF TERMS
Bayesian hierarchical models - Statistical hierarchical models with Bayesian parameter
estimation techniques which fit probability models to a set of data and summarizes the results
with a probability distribution, to determine parameters of a hierarchical model to predict the
occurrence distribution.

Bin classification framework - The LT2ESWTR incorporates specific treatment requirements
for protection against Cryptosporidium involving assignment of systems into different categories
(bins) based on the results of source water Cryptosporidium monitoring. Additional treatment
requirements depend on the bin to which the system is assigned (see Microbial Toolbox
Options).

Cryptosporidium - Microbial pathogen, Cryptosporidium parvum, associated with waterborne
disease (i.e., cryptosporidiosis) and known to infect immunocompetent and
immunocompromised humans.

Endemic disease - Disease levels that are natural or "on-going" in the "normal" population and
do not usually reach the attention of medical observers as would an epidemic.

Information Collection Rule (ICR) - EPA rule promulgated in 1996 pursuant to SDWA
requirements which required approximately 300 large public water systems to conduct 18
months of sampling for water quality and treatment related to DBF formation and the occurrence
of microbial pathogens. Data on DBF formation in small systems was obtained through 1) a
survey of approximately 120 treatment plants in systems serving fewer than 10,000 people and
2) information received from seven states on small systems.

ICR Supplemental Surveys (ICRSS) - EPA obtained additional pathogen occurrence data
through ICRSS which involved 127 water treatment plants, including 40 small systems, and
comprised one-year of bi-monthly sampling for Cryptosporidium, Giardia, and other water
quality parameters (small systems did not measure protozoa).

Initial distribution system evaluation - Studies conducted by Community Water Systems
which are intended to select new compliance monitoring sites that more accurately reflect sites
representing high TTHM and HAAS levels.  The studies are based on either on system specific
monitoring or other system specific data that provides equivalent or better information on  site
selection.

Locational running annual average (LRAA) - RAAs (see below) calculated for each sample
location in the distribution system which must be below the compliance levels (MCLs) in  each
quarter of the year.

                                         B-l

-------
Log credits - The logarithmic range of credit given for water system treatment and management
options employed, e.g., reducing pathogen loading into the plant, pretreatment processes,
additional pathogen barriers, etc.

Log removal - The logarithm of the reduction in microbial density due to an action. For
example 90% removal corresponds to 1 log removal.

Markov Chain Monte Carlo - A method to obtain a sample from the posterior distribution of
the parameters in a Bayesian hierarchical model that involves both Monte Carlo integration, to
handle high-dimensional, intractable integrals and construction of Markov chains to draw these
samples.  The posterior distribution is used to make inferences about parameters in the model
and to do predictions.

Microbial tool box options - Water systems will choose technologies to comply with additional
treatment requirements from a 'toolbox'  of options, e.g., pretreatment of water or improved
disinfection.

PCR (Polymerase chain reaction) - The process of rapidly amplifying a defined region of DNA
by sequential steps of denaturation and replication.

Priors or  Prior Distributions - Previous probability assessments of existing data used to
estimate occurrence under new  conditions.

Posterior Probabilities - Estimates of occurrence under new conditions produced using prior
distributions.

Running annual average (RAA) - Quarterly measurements of various sampling points in a
water distribution which are averaged over the year to provide a average which  is compared
against the Maximum Contaminant Levels (MCLs) for TTHM and HAAS.

Surface Water Analytical Tool - Model used in conjunction with the ICR data to predict the
impact of potential new standards for DBFs and/or pathogens on shifts in treatment technologies
among water systems and resulting DBF exposure profiles.

Waterborne Disease Outbreak- A waterborne disease outbreak occurs when two  or more
persons experience a similar illness after consumption or use of water intended for drinking and
epidemiologic evidence implicates the water as the source of illness. This outbreak is reported
by the authorities. Also, a single case of chemical poisoning constitutes an outbreak if
laboratory studies indicate that the water has been contaminated by the chemical. Only
outbreaks  associated with water intended for drinking are included.
                                          B-2

-------
                                ATTACHMENT C

       BIOSKETCHES OF THE DRINKING WATER COMMITTEE MEMBERS
                             Science Advisory Board (SAB)
                         U.S. Environmental Protection Agency
Dr. Mary E. Davis: Dr. Mary E. Davis is a Professor of Toxicology in the Department of
Physiology and Pharmacology at West Virginia University Health Sciences Center. Her research
interests are in the mechanisms of toxicity, focusing on renal and cardiovascular systems and
liver and emphasizing agents of environmental and occupational interest, including
halomethanes and disinfection by-products. She earned a doctorate in Pharmacology from
Michigan State University in 1977.

Dr. Davis is a member of the Editorial Board of Toxicology and Applied Pharmacology, and has
served on the Editorial Board of Toxicology. She served as Treasurer for the Society of
Toxicology. Dr. Davis previously served on two NRC Subcommittees on the health effects of
disinfectants and their by-products and use of physiologically-based pharmacokinetics in risk
assessment. She served as an external reviewer of EPA's risk assessment of the WTI hazardous
waste incinerator and of EPA's proposed guidelines for human health risk assessment protocol
for hazardous waste incinerators. In addition to serving on the DWC, Dr. Davis has been the
SAB Liaison to the National Drinking Water Advisory Council (NDWAC) and was a member of
the SAB's Chloroform Review Panel.

Dr. Ricardo DeLeon: Dr. De Leon is the Laboratory Manager for the Microbiology Unit of the
Water Quality Laboratory of Metropolitan Water District of Southern California. The
Microbiology Unit consists of the Compliance, Development and Reservoir Management Teams.
His area of expertise is water microbiology, methods development for detection of
microorganisms in water, inactivation of pathogens by disinfection and removal by treatment
technology. He is currently working primarily on drinking water but his  expertise also includes
water reuse and public health issues associated with water. He has been working in the area of
water microbiology since 1983.

Dr. De Leon holds a Bachelor's of Science in Microbiology and a Ph.D.  in Microbiology and
Immunology from the University of Arizona and did post-doctoral training in the Department  of
Environmental Sciences and Engineering of the University of North Carolina. He was also a
faculty member at the University of California, Irvine Campus prior to joining Metropolitan
Water District. He has been the principal or co-principal investigator on 22 research grants on
methods development, disinfection of microorganisms and microbial aspects of water treatment
technology. He has published more than 29 journal articles and book chapters on pathogen
detection in environmental samples. He is currently  serving in the Drinking Water Committee of
the Science Advisory Board to the U.S.  Environmental Protection Agency and on the National
Research Council Committee on Indicators of Pathogens in Water.
                                          C-l

-------
Dr. Barbara Harper: Dr. Harper is an independent consultant in the areas of toxicology, risk
assessment, CERCLA oversight, tribal water quality, and environmental management. She is
affiliated with AESE, Inc (www.aeseinc.com).  AESE's clientele consists entirely of
Tribes/Alaska Natives. She is also an adjunct faculty member of Oregon State University's
Public Health Department.  Dr. Harper is a board-certified toxicologist (Diplomate of the
American Board of Toxicology, 1989).  She received her B.A. degree cum laude with
departmental honors in biology from Occidental College in 1970.  She received her PhD in
genetics from the University of Texas at Austin in 1974.  She was on the faculty of the
University of Texas Medical Branch (UTMB) at Galveston in the Department of Preventive
Medicine and Community Health; Division of Genetic and Environmental Toxicology.  She then
took a position with the Commonwealth of Pennsylvania' s Department of Environmental
Resources, and developed and managed the Special Science and Resources Program.  She
taught risk assessment as an adjunct faculty member at Penn State Harrisburg during this time
period as well. She was recruited by Battelle's Pacific Northwest National Lab as a program
manager in risk assessment in 1993 (Hanford), where she started working on tribal risk issues.
She joined the Yakama Nation ERWM  Program in 1997 and developed methods for tribal risk
assessment methods now in use at DOE and EPA, and continues to develop tribally-relevant
methods for  evaluating cumulative risks and impacts to tribal health  and culture. Her research
interests include contamination offish and other tribal subsistence foods, the associated health
effects, eco-cultural and human health risk method development, nutrition,  anthro-toxicology,
and tribal parameters for subsistence exposure assessment.

Dr. Irva Hertz-Picciotto: Irva Hertz-Picciotto, Ph.D., Professor. Dr. Hertz-Picciotto received
her Master's  of Arts in Biostatistics, a Ph.D. in Epidemiology and a Master's of Public Health
from the University of California, Berkeley.  She has held positions as Assistant, Associate and
Full Professor at the University of North Carolina, Chapel Hill, and most recently joined the
Department of Epidemiology and Preventive Medicine at the University of California, Davis.
Dr. Hertz-Picciotto receives funding for research from the National Institutes  of Health, the
U.S. Environmental Protection Agency, the Medical Investigations of Neurodevelopmental
Disorders (M.I.N.D.) Institute, State of California Office of Environmental Health Hazard
Assessment, the Health Effects Institute, the Hawaii Heptachlor Research and Education
Foundation, the International Life Sciences Institute, and the University of California,
Berkeley.

Dr. Hertz-Picciotto serves on editorial boards for the two major journals in  her field, namely
Epidemiology and the  American Journal of Epidemiology, as well as for Human and Ecological
Risk Assessment.  She served as Chair of the Institute of Medicine/National Academy of
Science's Veterans and Agent Orange:  Update  2000 committee, and is currently Chair of the
IOM/NAS Update 2002 committee. Dr. Hertz-Picciotto is also a member of the Board
of Scientific Counselors of the U.S. National Toxicology Program, the Food Safety in Europe
Working Group sponsored by the International  Life Sciences Institute, and the Carcinogen
Identification Committee of the California Governor's Scientific Advisory Board. She is
currently President of the International  Society  for Environmental Epidemiology, and was
recently a delegate to the NIEHS-sponsored U.S.-Vietnam Scientific Conference on the
                                          C-2

-------
Environmental and Health Effects of the Vietnam War. She founded the Center on
Environmental Health and Susceptibility at the University of North Carolina, Chapel Hill. For
over ten years, she has taught methods for epidemiologic  data analysis in Chapel Hill, and has
taught courses on four continents. Dr. Hertz-Picciotto has published seminal papers on the use
of epidemiology in quantitative risk assessment and is internationally renowned for her work in
this field, as well as occupationally related cancer, environmental exposures, reproductive
outcomes, and methods for epidemiologic research.

Dr. Joseph R. Landolph:  Dr. Joseph R. Landolph is currently Associate Professor of Molecular
Microbiology and Immunology and Pathology and a Member  of the USC/Norris Comprehensive
Cancer Center, in the Keck School of Medicine and Associate Professor of Molecular
Pharmacology and Toxicology, in the School of Pharmacy, with tenure, at the University of
Southern California (USC) in Los Angeles, California.  Dr. Landolph received a B. S. degree in
Chemistry from Drexel University in 1971 and a Ph. D. in Chemistry from the University of
California at Berkeley in 1976, under the guidance of the late Professor Melvin Calvin, where he
studied the metabolism of the chemical carcinogen, benzo(a)pyrene, and its ability to induce
cytotoxicity in cultured mouse liver epithelial cells and morphological transformation in Balb/c
3T3 mouse fibroblasts. Dr. Landolph performed postdoctoral  study in chemical carcinogenesis
and chemically induced morphological and neoplastic cell transformation and mutagenesis at the
USC/Norris Comprehensive Cancer Center at the University of Southern California under the
late Professor Charles Heidelberger from 1977-1980. Dr. Landolph was appointed Assistant
Professor of Pathology in 1980, and Associate Professor of Microbiology, Pathology, and
Toxicology at USC in 1987. Dr. Landolph has served as a grant reviewer for the U. S. E. P. A.
Health Effects Panel, for special RFAs for the N. I.  E. H. S., and as an ad hoc member for the
Chemical Pathology Study  Section and the Al-Tox-4 Study Section of the N.I. H. Dr. Landolph
has also been a member of the Carcinogen Identification Committee reporting to the Scientific
Advisory Committee of the Office of Environmental Health Hazard Assessment of the
California Environmental Protection Agency from 1994-2002. He is the recipient of numerous
awards,  including the Merck Award in Chemistry and the Superior Cadet Award in ROTC from
Drexel University in 1971, the Edmundson Teaching Award in the Dept. of Pathology at USC in
1985, a Traveling Lectureship Award from the U.  S. Society of Toxicology in 1990, and a
competitive  American Cancer Society Postdoctoral Fellowship from 1977-1979.  Dr. Landolph
receives funding from the Nickel Producers Environmental Research Association (NiPERA),
from the National Cancer Institute, National Institutes of Health, from the National  Institute of
Allergy  and Infectious Diseases, National Institutes  of Health, and from the Office of
Environmental Health Hazard Assessment of the Environmental Protection Agency of the State
of California.

       Dr. Landolph's research interests and activities include studies of the genetic toxicology
and carcinogenicity of carcinogenic insoluble nickel compounds, carcinogenic chromium
compounds, carcinogenic arsenic compounds, and carcinogenic polycyclic aromatic
hydrocarbons.  His laboratory is focused on studying the ability of these carcinogens to induce
morphological and neoplastic transformation of C3H/10T1/2 mouse embryo cells and the
cellular and molecular biology of the transformation process.  His laboratory is currently
studying the ability of carcinogenic nickel compounds to induce activation of expression of

                                          C-3

-------
oncogenes and inactivation of expression of tumor suppressor genes in cells transformed by
insoluble carcinogenic nickel compounds, such as nickel subsulfide, crystalline nickel
monosulfide, and green (high temperature) and black (low temperature) nickel oxides. His
laboratory is also studying the molecular biology of chromium compound-induced cell
transformation and the role of valence in cell transformation by various chromium-containing
compounds. Dr. Landolph is an expert in chemically induced morphological and neoplastic
transformation and chemically induced mutation in murine and human fibroblasts.  He is the
author of 32 peer-reviewed scientific publications, 21 book chapters/review articles, and has held
peer-reviewed research grant support from the U. S. E. P. A., the U. S. National Cancer Institute,
and the U. S. Institute of Environmental Health Sciences.

Dr. David L. Sedlak: Dr. David L. Sedlak is Associate Professor of Civil and Environmental
Engineering at the University of California, Berkeley.  Dr. Sedlak received has B.S. degree in
Environmental Science from Cornell University in 1986. He received his Ph.D. degree in Water
Chemistry from the University of Wisconsin in Madison in 1992 and served as a postdoctoral
researcher at the Swiss Federal Institute for Environmental Science and Technology (EAWAG)
from 1992 to 1994. He has received several notable awards including the NSF CAREER Award
in 1997, the Hellman Family Faculty Award in 1996 and the American Chemical Society
Graduate Student Award in 1991.  His areas of research interest include analytical methods for
measuring organic compounds in water, fate of chemical contaminants in water recycling
systems, metal speciation and its effect on metal uptake and reaction, environmental
photochemistry and ecological engineering. David Sedlak receives research funding from
federal (i.e., National Science Foundation) and state (i.e., University of California Water
Resources Program, University of California Toxic Substances Research and Teaching Program)
programs. He also receives funding from a private foundation (i.e., National Water Research
Institute) and several  water industry sponsored foundations (i.e., American Water Works
Association Research Foundation,  Water Environment Research Foundation and WateReuse
Foundation)

Dr. Philip C. Singer: Dr. Philip C. Singer is the Dan Okun Professor of Environmental
Engineering in the Department of Environmental Sciences and Engineering in the School of
Public Health at the University of North Carolina at Chapel Hill. He directed the Water
Resources Engineering Program at UNC for 19 years and currently directs UNC's Drinking
Water Research Center. He has conducted research on  chemical aspects of water and wastewater
treatment and on aquatic chemistry for the past 35 years, and has published more than 160
papers and reports in  these areas. For the past 27 years, Dr. Singer's research has focused on the
formation and control of disinfection by-products in drinking water. In 1993, Dr Singer was
selected for the Freese Lecture by the American Society of Civil Engineers, in 1995 he was
given the A.P. Black  Research Award by the American Water Works Association,  and in 1999
he received the Fuller Award from the North Carolina section of the American Water Works
Association.

Dr. Singer has been active in the American Water Works Association, serving as a past Chair
and Trustee of the Research Division, and has served on the Research Advisory Council of the
American Water Works Association Research Foundation. He was  on the editorial board of

                                          C-4

-------
Ozone Science and Engineering and is a past associate editor of Environmental Science and
Technology. He was a member of the Water Science and Technology Board of the National
Research Council, and served on the National Research Council's Committee on Drinking Water
Contaminants. He is currently on the Board of Directors of the Water Environment Research
Foundation and the U.S. Environmental Protection Agency Science Advisory Board's Drinking
Water Committee.  In 1995, Dr. Singer was inducted into the National Academy  of Engineering.

Dr. Laura Steinberg: Dr. Steinberg is Associate Professor in the Civil and Environmental
Engineering Department of Tulane University. She holds a B.S.E. in Civil and Urban
Engineering from the University of Pennsylvania and an M.S. and Ph.D. in Environmental
Engineering from Duke University. Her research currently focuses on water quality modeling
and natural hazards management. She has recently completed modeling studies of arsenic
concentrations in water distribution systems and transport processes in contaminated sediments,
and is working on spatial statistical modeling of heavy metals and PCB's in contaminated
sediments. During the last two years, she has spent several months in Turkey, investigating the
impacts of the devastating earthquake of 1999 on industrial infrastructure and the environment,
and evaluating the effectiveness of chemical risk management procedures. Dr. Steinberg is the
incoming chair of the American Society of Civil Engineer's National Environmental Policy
Committee, and a past member of the ASCE's National Water Policy Committee. She serves on
the Water Environment Federation's Disinfection Committee, and is a fellow of the Institute of
Civil Infrastructure Systems and a former member of the Chapel Hill, NC Planning Board. She
has consulted to the USEPA's  Science Advisory Board on technology diffusion,  and the
Department of Energy on risk assessment. Prior to her work in academia, Dr. Steinberg was
Environmental Engineering Department Head at the planning and engineering firm of Louis
Berger International, and Business Development Manager at Geraghty and Miller, an
environmental engineering firm. She also had the distinct honor of serving as a US
Congressional Page while attending high school.

Ms. Susan Teefy: Susan Teefy currently serves on the staff of the Water Quality and Treatment
Solutions, Inc. Susan formerly served as the Operations Engineer for the Alameda County
Water District in Fremont California. Since 1992, she has worked with this public water agency
to ensure compliance with drinking water regulations, and analyze and optimize plant
operations. She has held positions of increasing authority with the District, including Manager
of the Water Production Division, which is responsible for the operation and maintenance of
three water treatment plants and the distribution  system. Ms. Teefy has also supervised
ACWD's Environmental Engineering section, where she developed and implemented water
quality monitoring programs and conducted plant optimization studies. Her particular interest is
surface water treatment (particulate removal processes) and ozone disinfection. Prior to working
with the Alameda County Water District, she worked at the East Bay Municipal Utility District
in Oakland California, providing technical support for surface water treatment plant  operations.
Ms. Teefy also worked for the U.S. Environmental Protection Agency, Region 9, in  San
Francisco where she managed the drinking water program  on Indian Lands in California.

Ms. Teefy has a bachelor's degree in civil engineering from the University of California at
Berkeley, and a master's degree in environmental engineering from the University of North

                                          C-5

-------
Carolina at Chapel Hill.  She is a registered civil engineer in the state of California, and a
licensed water treatment plant operator (Grade 5, highest level). In 1985 she was awarded
USEPA's Bronze Medal for outstanding service for significantly improving compliance with
drinking water regulations on California Indian Lands. In 1989 she was the first recipient of the
AWWA Larson Aquatic Research Support (LARS) Scholarship. In 1991 she received AWWA's
Academic Achievement award for her Master's thesis. She has chaired AWWA's California
Nevada Section Research Committee, and currently is a member of AWWA's national
coagulation and filtration committee. Ms. Teefy has been a Project Advisory Committee
member on several projects funded by the AWWA Research Foundation, and a peer-reviewer for
the Journal of AWWA. She has served on AWWARF's Unsolicited Proposal Review
Committee, as well as AWWARF and EPA-convened Expert Panels regarding water treatment
issues. She has given numerous presentations at international AWWA and International Ozone
Association conferences

Dr. Gary A. Toranzos: Gary A. Toranzos is a professor of microbiology in the Department of
Biology, University of Puerto Rico, Rio Piedras Campus. He got his Ph.D. in 1985 at the
University of Arizona in Tucson. His research interests are varied and include water
microbiology, the ecology of enteric pathogens and the development of indicators of risk. He
has published extensively on all the above subjects and is currently working on projects dealing
with bacterial nitrification and denitrification in soils, as well as development of new indicators
of biological contamination in waters. Dr. Gary A. Toranzos receives funding from NASA to
study nitrifying and denitrifying microbial communities in tropical soils. He also has funding
from the USGS (Water Resources Center, University of the U.S. Virgin Islands) to study the
microbial water quality of bathing beaches in Puerto Rico and St. Thomas, U.S.V.I.

He is currently working at the National Science Foundation as a Program Director in the
Division of Molecular and Cellular Biosciences.

He is an elected member of the American Academy of Microbiology, a Fellow  of the American
Association for the Advancement of Science and is serving a term as member of the Technical
Advisory Board of the Water Environment Research Foundation

Dr. Rhodes Trussell: Dr. R. Rhodes Trussell is Director of the Water Knowledge Center and
Senior Vice President at MWH,Inc. He has served in that Role since September 2001. For
several years prior to that he served as the firm's Director of Corporate Development and as a
member of the firm's  Board of Directors.  The bulk of Dr.Trussell's technical career has been
spent advising municipal utilities, both in the US and abroad, concerning problems of drinking
water quality  and treatment. Dr. Trussell is active in American Water Works Association and in
the International Water Association where he serves on the program committee, the Strategic
Council and the editorial board for North America. He also serves on the Water Science and
Technology Board for the National Resource Council where he has served on several specific
Committees, most recently those on potable reuse, the CCL, and indicators for pathogens in
water. Dr. Trussell serves on the Magazine Board for Environmental Science and Technology, as
a member of the Industrial Advisory Board for Engineering program at UC Riverside and as
Chair of the Industrial Advisory Board for the Department of Civil Engineering at UCLA.  Dr.

                                          C-6

-------
Trussell received his B.S.(1966), M.S.(1967), and Ph.D.(1972) in Environmental Engineering
from the University of California at Berkeley. He was elected to the National Academy of
Engineering in 1995 and serves on the Peer Committee for Civil Engineering. He is currently
the Chair of the SAB's Drinking Water Committee.

For the past 30+ years, Dr. Trussell has worked for MWH, Inc. and is solely funded by the
corporation. During the past year he as worked directly on projects for the city of Portland
Oregon, for the East Bay Municipal Water District, for Hong Kong, the City of San Diego, the
City of Long Beach, the Metropolitan Water district of Southern  California, and  the Los
Angeles Department of Water and Power.
                                          C-7

-------