United States
          Environmental Protection    Office of Water       EPA 811-R-92-005
          Agency             (WH-550)          August 1992
&EPA   FRAMEWORK FOR DECISION
          MAKING: AN EPA PERSPECTIVE

-------
          Framework for Decision Making: EPA Perspective


 Stig  Regli1, John E. Cromwell2,  xin Zhang2, Allen B. Gelderloos3,
                       William  D.  Grubbs4,
                Frank Letkiewicz3,  Bruce A. Macler1


     The United States Environmental Protection Agency (EPA) is
developing national regulations to control for disinfectants and
disinfection by-products (D/DBPs)  in public drinking water
supplies.  EPA intends to propose regulations that will apply to
all public water systems using disinfection and serving non-
transient populations — approximately 220 million people. This
regulation would be unlike the current maximum contaminant level
 (MCL)  for total trihalomethanes (TTHMs) which only pertains to
systems serving more than 10,000 people.

     A major issue is to ensure that the  limits set for D/DBPs do
not cause changes in treatment that result in significant
increases in risk from waterborne pathogens or other DBPs not
regulated at this time.  A fundamental goal in setting any new
regulation is to be able to conclude with confidence that the
resultant changes in treatment would lead to a significant
overall decrease in risk to public health. This paper will
describe the analysis that EPA  is undertaking in pursuit of
meeting this goal.

Background

     The Safe Drinking Water Act (SDWA) passed by the Congress of
the United States requires the  EPA to specify a maximum
contaminant level goal (MCLG) for  each regulated contaminant.
MCLGs are set at levels at which no adverse health effects are
expected to occur.

     The SDWA also requires the EPA to set MCLs as close to the
MCLG as is technically and economically feasible to achieve; the
rule must specify such best available technology (BAT).  If it is
not technically or economically feasible  to measure a contaminant
     1  - U.S.  Environmental Protection Agency

     2  - Wade  Miller Associates

     3  - Malcolm Pirnie,  Inc.

     4  - Science Applications  International Corporation

     $  - Abt Associates Inc.

                                 1

-------
at the level of concern, then the EPA must set a treatment
technique requirement in lieu of an MCL.

     Systems are not required to use BAT in order to comply with
the MCL but can use other technologies as long as they meet all
drinking water standards and are approved by the State.   Systems
unable to meet the MCL after application of the BAT can get a
variance, extending the time for achieving compliance.  Systems
that obtain variances must meet a schedule approved by State
health officials.

     MCLGs are set at zero for contaminants that have
carcinogenic potential from a drinking water exposure. In such
cases the EPA attempts to set MCLs at levels which assure that
lifetime risk from cancer through exposure in drinking water is
between less than one in ten thousand to one in one million (10"*
to 10"6) ,  using conservative margins of safety in calculating such
risk.  Risk assessments usually entail: extrapolation from toxic
effects in animal studies using high doses of exposure to predict
the effects of low doses of exposure in humans; use of the
linearized multistage model which is more conservative than most
other models; and use of a 95% upper confidence interval for
defining the risk factor.

     Based on available occurrence, exposure, and health effects
data, twelve chemicals have been identified for possible
regulation (USEPA 199la; USEPA 1992a; USEPA 1992b). Table I lists
the candidate compounds for regulation, possible MCLGs, the
health effect of concern, and the tentative cancer risk
classification.

     Table II indicates the concentration level for DBPs
classified as 87 compounds in drinking water at which excess
lifetime cancer risk of 1 in 10,000 and 1 in 100,000 would occur
as predicted by the upper 95% confidence interval, assuming 2
liters are consumed per individual per day.  Table II also
indicates the range at which the candidate compounds have been
reported to occur in US drinking water supplies (for a more
detailed characterization of D/DBP occurrence see USEPA, 1992a).
The basis for-the MCLGs, cancer risk classification, and the
indicated cancer risks are described elsewhere  (USEPA 1992b).

     The EPA intends to set MCLs for each compound listed in
Table I.  However, three regulatory options are being considered
for trihalomethanes  (THMs) and haloacetic (HAAs) acids.  For THMs
the EPA is considering:  1) MCLs for each of the four THMs, 2)  an
MCL for total trihalomethanes  (TTHMs), and 3) an MCL for each of
the THMs and an MCL for TTHMs  (which may 'be different from the
sum of the individual THMs). Individual THMs are being considered
as their health risks differ significantly; and the technical
feasibility for limiting their formation can vary greatly,

-------
depending upon source water quality.  An MCL for TTHMs is being
considered to act as a surrogate regulation to limit other DBFs
and because of the precedent already established with the current
TTHM standard.

     For HAAs the EPA is considering:  1) MCLs for
trichloroacetic acid (TCA) and dichloroacetic acid (OCA) as
health effects data are available for each of these compounds;
2) an MCL for total HAAs  (THAAs) including mono-, di-, and tri-
chloroacetic acid; monobromo and dibromo acetic acid; and
possibly bromochloroacetic acid;  and 3) the combination of
options 1 and 2.  A limit for THAAs is being considered because
all the compounds can be analyzed simultaneously using the same
analytical method at no additional cost.  A limit for THAAs would
act as a surrogate to limit production of other HAAs and DBFs for
which health risks are not yet determined.  Future MCLs, when
more data become available, may pertain to specific HAAs other
than DCA and TCA, and/or an expanded list of HAAs including
chloro-bromo acetic acids.

 i'.
Issues With Defining BAT for DBFs

BAT must include disinfection but now much?

     The easiest means for avoiding risk from D/DBPs is not to
disinfect water.  Obviously this is not a viable option, at least
not for surface water supplies, since without disinfection
waterborne disease would be rampant.  Although not specified in
the SDWA, the EPA must define BAT for controlling DBFs in
conjunction with criteria to control pathogens.

     The feasibility of complying with any D/DBP standard with a
given BAT is directly affected by the need to also comply with
microbial standards.  Currently only two national standards have
been promulgated in the U.S. to control for microbial concerns:
the total coliform rule, and the surface water treatment rule
(SWTR). Ground water disinfection requirements are under
development.  _

     The SWTR  (USEPA, 1989a) requires that systems using surface
water, or ground water under the direct influence of surface
water, provide at least 99.9%  (3 logs) and 99.99% (4 logs)
removal or inactivation of Giardia cysts and viruses,
respectively.  These performance levels are not determined by
measurement of Giardia or viruses but by assuming removal
efficiencies by filtration and.  inactivation efficiencies by
disinfection based upon design and operating conditions.

     Well operated filter systems that meet recommended design
and operating, and turbidity performance, criteria are assumed to

-------
achieve 2 to 2.5 log removal of Giardia cysts, and a 1 to 2 log
removal of viruses by filtration alone. Inactivation efficiencies
for Giardia and viruses are determined based upon measurement of
the product of "C" (disinfectant concentration in mg/1) and WT"
(disinfectant contact time in minutes); the "CT" for a system
equates to an inactivation efficiency indicated in tables in the
SWTR.

     The SWTR also requires systems to maintain a) a disinfectant
residual of at least 0.2 mg/1 at all times in the water entering
the distribution system and, b) a detectable disinfectant
residual or a heterotrophic plate count of less than 500/ml in at
least 95% of the measurements throughout distribution system.

     The total coliform rule (USEPA, 1989b) requires all public
water systems (surface and ground water supplies) to measure for
total coliforms in the distribution system and to ensure that no
more than 5% of the measurements are total coliform positive.
For any positive total coliform measurement, follow-up
measurement of fecal coliforms or E. coli is required.  No more
than one positive fecal coliform or E. coli measurement is
allowed in any given month.

     The GWDR will establish minimal disinfection requirements
for all ground water systems unless they are able to demonstrate
they are not vulnerable to fecal viral and bacterial
contamination.  The vulnerability of wells for each system is to
be determined based upon the location of fecal contamination
sources, such as septic tanks and sewage lines, with respect to
each well, and hydrogeological features.

     Major issues are whether more microbial standards, and the
disinfection required to meet these standards, are needed to
limit microbial risk or to prevent significant increases in
microbial risk under new D/DBP regulations.  To what extent, if
any, should the definition of BAT for the D/DBP regulations
include additional microbial treatment requirements to address
this concern?

Which technologies should be considered and how should treatment
performance be defined?

     How the EPA defines BAT will directly affect the  level at
which the MCLs are set and the potential for downside  risks.
Because health risks from by-products of alternate disinfectants
are not as well characterized as they are for by-products of
chlorination, it may be appropriate to define BAT as certain
precursor removal technologies with chlorination  for primary
disinfection under this rulemaking  (since MCLs do not  have to be
met using BAT this would not preclude the use of  alternate
disinfectants).  Examples of precursor removal technologies
include: a) conventional treatment optimized for  DBP precursor

-------
 removal;  b)  granular activated carbon (GAG)  as a filter media
 replacement;  c)  GAC following filtration; and d)  membrane
 filtration.

      Other technologies for BAT consideration,  depending upon the
 source water quality, include conventional treatment or precursor
 removal technologies with use of alternate disinfectants, such as
 ozone for primary disinfection, and chloramines for residual
 disinfection.

      The cost and performance of the above technologies will vary
 greatly depending upon design and operation.   The technical and
 economic feasibility of using these technologies to achieve
 different MCLs can be greatly influenced by source water quality
 and system size.  One approach to this dilemma is to define MCLs
 based on what some reasonable percentage (e.g., 90% or 95%) of
 systems could at least achieve if they were to apply BAT to their
 source water.

      If the MCL were based on what all systems (or a very high
 percentage such as 99%) could achieve using the BAT, then the MCL
 would be higher and potentially much less protective for the
 population at large.  On the other hand, if the MCL were set
 based on what a lower percentage of systems (e.g., 75%) could
 achieve, then treatment costs to comply, or the transactional
 costs to enforce compliance, might be unreasonable.

      For logical decision making, a methodology is needed to
 predict performance of different BATs in the U.S. given the
 diversity of source waters and the effect that such diversity has
 on the treatability for OBPs and microbial concerns.  Such a
-methodology should allow for prediction of BAT performance for
 controlling DBFs while also achieving different objectives for
 pathogen control.

 Implications of Systems not Using BAT to achieve MCLs

      Systems are not obligated to use BAT to achieve the MCL.  If
 BAT were defined as one of the OBP precursor removal strategies
 as described a~bove, systems could choose less expensive
 technologies such as adopting choramines or ozone and chloramines
 to achieve the MCL.  In fact, use of such technologies might lead
 to lower levels of risk from chlorinated OBPs  (e.g., levels well
 below the MCL) than if the BAT were used.  However, depending
 upon the source water quality, such treatment changes may
 introduce new risks from other DBFs that are not yet well
 defined. Also, since chloramines are a much weaker disinfectant
 than chlorine, switching to chloramines may significantly
 increase microbial risk. Switching to ozone — a stronger
 disinfectant than chlorine — followed by chloramines to maintain
 a residual in the distribution system may lead to greater
 nutrient availability for bacteria to grow in the distribution

-------
system.  Such effects will largely depend upon the source water
quality, and the treatment following ozonation.


     If the MCLs are defined based on what., conventional
filtration followed by deep bed GAG and chlorine, or what
membrane filtration and chlorine, could achieve, many systems
would be expected to attempt meeting such MCLs using chloramines
or ozone and chloramines due to cost.  If chloramines, rather
than chlorine, were included as a residual disinfectant in the
definition of BAT, then the MCLs for chlorinated DBFs would be
lower and even greater shifts to use of alternate disinfectants
to achieve compliance could be expected.  The potential health
risk implications that might result from different shifts in
technology needs to be considered when defining BAT.

Downside Risks with DBP Precursor Removal Technologies

     Precursor removal technologies can remove substantial
organic material and disinfectant demand.  As more DBP precursors
are removed, and the water is cleaner, the efficiency of
disinfection is expected to increase per dose of disinfectant
applied.  A decrease in disinfectant demand allows for lower
disinfectant doses to be applied and yet still maintain a
residual in the distribution system.  However, a significant
decrease in the level of disinfection between the point of
application and early points in the distribution system could
increase microbial risk unless precautions are taken  (e.g.,
increasing the disinfectant contact time prior to the first
customer).  The significance of such changes in risk depend on
the pathogen concentrations in the source water, the change in
effect that the water quality matrix has on interference with
disinfection, the net change in the value of the CT product with
respect to different populations receiving the finished water in
the distribution system, and the net change in bacterial
pathogens that might grow in the distribution system.

     Precursor removal technologies can also change the ratio of
disinfectant residual with respect to organic  (e.g., humic and
fulvic acids) -and inorganic (e.g., bromide) precursors. This can
lead to a change in the relative amounts of different types of
DBPs that are formed.  For example, depending upon the bromide
concentration in the source water it becomes possible, with
increased precursor removal, to decrease the amount of TTHMs or
THAAs but to increase the concentration of specific brominated
compounds (e.g., bromoform or brominated HAAs).  Such changes in
the relative amounts of the different DBPs could result in no net
gain in risk reduction and possibly even an increase  in risk. No
health effects data are yet available on the brominated HAAs.

-------
Strategy for Defining BAT and Appropriate MCLs

     The EPA is taking the following steps to address the above
issues and to determine appropriate MCLs for OBPs: 1)
identification of candidate BATs for controlling both DBFs and
pathogens; 2) definition of BAT in the context of also achieving
microbial treatment objectives (the SWTR, coliform rule, or other
criteria); 3) evaluation of candidate BAT performance with
respect to pathogen and DBF control for different source water
qualities;  4) consideration of candidate MCLs, based on what
level is achievable by candidate BATs for a "reasonable"
percentage of systems; and 5) consideration of potential net
changes in risk from exposure to DBFs and pathogens  (known and
unknown), and the costs to implement such changes, to determine
the optimum MCLs.  The EPA is attempting to implement this
strategy through use of a disinfection by-product risk assessment
model. (DBPRAM).

Overview of DBPRAM

       Figure 1 illustrates the DBPRAM modeling framework. OBPRAM
simulates raw water qualities for treatment plants throughout the
U.S. and predicts: 1) the most likely treatment technologies that
utilities would adopt to meet different regulatory targets (MCLs
for OBPs and microbial treatment objectives);  2) the occurrence
of trihalomethanes and haloacetic acids and the associated cancer
risk (i.e., that which is currently quantifiable) after treatment
changes are implemented; 3) the occurrence of pathogens — using
Giardia as the target organism — in the finished water after
treatment changes are implemented, and the associated risk from
infection; and 4) the costs for meeting the treatment objectives.
A detailed description of the model, assumptions, and results
determined to date are described in a series of papers  (USEPA,
1992c, Letkiewicz et al., 1992; Gelderloos et al., 1992; Grubbs.
et al., 1992; Cromwell et al., 1992).

Focus of Analysis

     To date the modeling analysis has focused on systems which
currently filter and disinfect surface water.  Since these
systems include those with the highest levels of DBF precursors
and pathogens in the source water, they represent those with the
highest possible changes in risk that could result from treatment
changes to comply with D/DBP regulations.  If regulatory
scenarios can be described that assure a significant decrease in
overall risk in these systems, then scenarios can also be
described that ensure there is at least a positive net benefit in
overall risk for systems with better source water quality  (e.g.,
most ground water systems). Preferably, such regulatory scenarios

-------
would also ensure there is no significant increase in risk for
any particular system, no matter what its source water quality.

     Due to limited data, the analysis has focused on a subset of
filtered surface water supplies: those which do not use lime
softening and which each serve more than 10,000 people. About 103
million people are currently served by systems meeting this
description.  Figure 2 illustrates key input and output
parameters in the DBPRAM and the conceptual approach for
estimating finished water quality from different treatment
configurations.  Probability distributions for raw water quality
parameters affecting THM and HA formation were generated from
different data sets (Letkiewicz et al 1992).  Raw water Giardia
data reflecting variation among cities (LeChevalier et al., 1991)
and within cities (Hibler et al., 1987) were used to simulate
national raw water occurrence (Grubbs et al 1992).  A water
treatment plant model was used to predict Giardia cyst
concentrations and THMs and HAs in the distribution system as a
function of different filtration and disinfection conditions
(USEPAc, 1992, Gelderloos et al 1992).

     THMs and HAAs have been targeted in the DBPRAM because they
are among the DBPs identified to date that occur at highest
concentrations in most drinking waters.  They are also likely to
be of the greatest health concern, and models are available for
predicting their formation.  Giardia cysts are targeted in the
DBPRAM because more data are available on their occurrence in
source waters than other pathogens; Giardia are much more
resistant to disinfection than most other waterborne pathogens;
and changes in treatment to control for DBPs, assuming that
systems must still meet the minimal requirements of the SWTR, are
likely to result in more significant changes in exposure from
Giardia than from most other waterborne pathogens identified to
date.  Also, a dose response curve for Giardia is available for
estimating risk from exposure (Rose et al 1991).

Treatment Constraints

     The DBPRAM was run to predict the ability for systems to
comply with different MCL targets for TTHMS and THAAs under two
sets of treatment constraints. The first set of constraints,
termed the "SWTR scenario", requires systems to:

     - achieve at least a 3 log reduction in Giardia cysts by
filtration and disinfection prior to the first customer;
                                8

-------
     - maintain a residual of at least 0.2 mg/l free chlorine (or
     1 mg/l total chlorine if chloramines are used)  at the end of
     the distribution system1;

     - maintain the free chlorine residual below 2 mg/l at the
     first customer because of concern from taste;

     - maintain the chloramine residual below 3 mg/l at the first
     customer to remain below the possible MCLG (see Table II)2
     and;

     - maintain a pH of at least 7.6 in the disinfected water
     entering the distribution system if the hardness is greater
     than 100 mg/l; maintain a pH of at least 8.0 if the hardness
     is less than 100 mg/l.3

     The second set of constraints apply to an "Enhanced SWTR"
(ESWTR) scenario. This scenario assures that systems with poorer
source waters disinfect sufficiently to maintain a low pathogen
risk (using Giardia as the target organism) while still complying
with the D/DBP rule.  The constraints in this scenario are
identical to those of the "SWTR scenario11 except that systems are
required to achieve higher levels of treatment, depending upon
the concentration of Giardia cyst in the source water/ as
follows:
     1  The disinfectant residual constraint in the distribution
system is assumed necessary to meet the total coliform rule and
the requirement to maintain a "detectable" residual in at least
95% of the samples taken in the distribution system under the
SWTR.

     2  Since the above analysis  was conducted,  EPA has begun
considering an MCLG for chloramines of 4 mg/l measured as total
chlorine  (USEPAb, 1992) to account for the health effects of
monochloramine which cannot be practically measured by itself.
In future model runs EPA will consider a constraint of 4 mg/l
rather than 3 mg/l measured as total chlorine.

     3  These constraints are to  simulate corrosion control and
compliance with the lead rule.

-------
             Giardig
    cyst  concentration/100  L             Giardia
                                   Removal/Inactivation
              S 1                         3-log
              1-10                        4-log
             10-100                        5-log
            ~~ *	                      6-log
                                          7-log
   10-100
 100-1,000
1,000-10,00
     The modeling approach assumes that under both the SWTR and
ESWTR treatment constraints, filtration plants are operating in
accordance with the SWTR and are achieving 2.5 log removal of
Giardia cysts.  The model assumes that any additional treatment
which is required is achieved through disinfection.  The modeling
approach assumes that plants will operate disinfection slightly
beyond the minimum requirements to ensure continuous compliance
during abnormal conditions. For disinfection through the
treatment plant with chlorine, the chlorine dose is calculated to
achieve at least 20% greater inactivation than required in the
SWTR or ESWTR.

     The only difference between the modeling approach for the
ESWTR and the SWTR scenario is the amount of disinfection
required in the plant depending on the raw water Giardia
concentrations.

Initial Conditions and Compliance Sorting Routine

     One hundred hypothetical treatment plants were individually
evaluated through the treatment model to predict plant effluent
and distributed water quality and to determine whether a system
meets the treatment constraints and DBF MCLs.

        Due to the large number of treatment configurations, and
because systems are currently making changes in treatment to
comply with the SWTR, total coliform, and lead rules, it was not
possible to simulate a baseline of treatment practice. Rather, it
was assumed that all treatment plants initially use chlorine as
the only disinfectant, though plants use alternate disinfectants
often because of source water quality with substantial THM
precursors and the need to meet the current TTHM rule.  The
validity of the model for predicting treatment changes that
     4 EPA guidelines in the Guidance Manual to the SWTR (USEPA,
1991b) are based on a geometric mean but in this modeling
analysis the arithmetic mean is used for purposes of conducting
statistical analysis.

                                10

-------
systems would make to comply with new DBFs standards was in part
confirmed by comparing the percentage of systems predicted to
modify their treatment to achieve a TTHM MCL of 100 yg/1 with the
percentage of plants that actually made such modifications to
meet a target of 100 ng/l MCL (Gelderloos et al 1992, Cromwell et
al 1992).

     The predicted water quality of the 100 plants is evaluated
through a "compliance sorting routine1*, sorting the treatment
model output according to alternative DBF MCLs.  If a system does
not meet the constraints established by the SWTR (or the ESWTR)
and the MCL of interest, the system implements the lowest cost
modification necessary to meet the constraints.  If the new
treatment method is not sufficient to achieve compliance with
either the treatment constraints or the DBF MCL of interest, the
next lowest cost modification is added. This process is repeated
for each system until it complies with the treatment constraints
and the DBF MCL of interest.

     The compliance sorting routine assumes that systems strive
to comply with the MCL with a 20% margin of safety. In other
words, for an MCL of 40 ug/1, it is assumed that systems would
add sufficient treatment to consistently meet a level of 32 ug/1
(40 - 0.2 x 40 = 32). Similarly, if a system needs 0.5 log of
Glardia inactivation to minimally meet the SWTR treatment
constraints (2.5 log removal is assumed by filtration) then a
system would strive to attempt to achieve this level with a 20%
margin of safety, i.e., 0.6 log inactivation (0.5 + 0.5 x 0.2 =
0.6).  If  higher levels of inactivation are needed to meet an
ESWTR, then these levels are adjusted upward by 20%  (e.g.,
achieving a 3 log level of inactivation would require 3.6 log of
inactivation).

     For coagulation/filtration plants using chlorine as the only
disinfectant, the approach assumes the following order of
alternatives, from the lowest cost alternative to the highest
cost alternative:

   • Elimination of pre-chlorination, if practiced;

   • Installation of ammonia feed system to provide chloramines
     as secondary disinfectant;

   • Increasing coagulant dose to improve DBF precursor removal;

   • Switching to ozone/chloramines for disinfection; and

   • Adding GAC  (30 minute empty bed contact time with 6 month
     carbon regeneration frequency) or membranes.

     The implementation of a chlorine/chloramine or an
ozone/chloramine disinfection strategy may be effective for

                                11

-------
complying with both the SWTR and ESWTR, and TTHM or THA limits.
However, these alternative disinfection strategies may not be
effective for other DBFs. For example, if a system has high
bromide concentrations in the source water the use of ozone may
lead to high bromate concentrations.  Therefore, under some
possible regulatory scenarios, the implementation of certain
disinfection alternatives to chlorine may not be sufficient to
comply with all MCLs.  Therefore, the impacts of alternative DBF
MCLs were evaluated under two different compliance approaches
(see Figure 3).  One compliance approach assumes that alternative
disinfection strategies are available, with the other assuming
that alternative disinfection strategies are not available, i.e.,
systems can only chlorinate.

The compliance approach assuming no availability of alternate
disinfectants allows for prediction of the percentage of systems
that would achieve different MCL targets through only use of
precursor removal technologies.  Estimation of the feasible
performance of candidate BATs, such as optimized filtration for
DBP precursor removal, or optimized filtration and GAC, if
applied to source waters throughout the U.S., could form the
basis for defining the MCL, e.g., the level of DBP(s) that at
least 90 percent of the systems could achieve using the candidate
BAT.  The same approach can be applied for defining the
performance of candidate BATs using alternate disinfectants.

Analysis of these two compliance approaches also makes possible
national cost comparisons between allowed use of alternate
disinfectants versus sole use of precursor removal strategies to
achieve different regulatory targets.   Such cost analysis is
discussed elsewhere  (Cromwell et al.. 1992).
Risk Analysis

DBFs

     The mean ~OBP concentrations predicted in the model described
above were used to estimate cancer risk in each system.  Figure 4
illustrates the methodology and the risk factors used to estimate
the cancer cases per year from the predicted DBP drinking water
concentrations. Risk factors for both the upper 95% confidence
interval and the maximum likelihood (MLE) estimate were used in
the analysis.

     Though chlorodibromomethane has been tentatively classified
as a group C carcinogen, a cancer risk factor equivalent to that
for bromodichloromethane is assumed.  A cancer risk factor for
trichloroacetic acid was used based on animal data (USEPAb,
1992), even though it also is tentatively classified as a group C

                                12

-------
carcinogen.5    No  risk  numbers  are  available  for  the  other HAAs
(monochloroacetic,  and mono and dibromoacetic acid) included in
the predicted THAA occurrence levels and thus a zero risk factor
is assumed for them.

     Not included in the analysis are occurrence or risk
estimates for chloral hydrate or bromate, or the chloro-bromo
HAAs.

Giardia

     Figure 5 illustrates the methodology for estimating Giardia
occurrence and risk at the first customer.  The analysis assumed
a reduction in treatment performance of 1 log 5% of the time, all
viable cysts as infective in humans, infections equivalent to
illness for the purpose of identifying outbreaks,  and outbreaks
as identified if the illness rate at the first customer exceeded
1% within a 30 day period.  Other assumptions are also used in
different analyses (Grubbs et al 1992) but are not described
here.  A dose response curve (Rose et al., 1991) was used to
estimate infections from the predicted occurrence levels.

     Not indicated in Figure 5 is the assumption that the
exposure predicted at the first customer represents 10% of the
exposure for the total population:

number of infections per system * infection rate at first
customer x  total system population x 0.10.

This equations assumes die off of Giardia cysts from exposure to
a disinfectant residual as they move through the distribution
system.  People receiving water in the middle of the distribution
system, because of the large CT values achieved through such
points, would be expected to be exposed to much lower Giardia
cyst concentrations than people near the first customer. The 10%
factor was assumed to account for an approximate weighted average
for all systems for such effects.

     Not included in this analysis are risks from other
pathogens.  This exclusion may be a significant limitation
regarding pathogens more resistant to disinfection than are
Giardia cysts (e.g.., Cryptosporidium, possibly some viruses) and
bacterial pathogens which might grow in the distribution system
(e.g., Legionella and mycobacteria).  Except for such concerns,
it appears reasonable to assume that if risks from Giardia cysts
     5 Risk factors were used for these compounds because,
conceivably, the cancer classification for these compounds could
change to 83 at a later date based on additional review of  the
existing data or new evidence from new data.
                                13

-------
are shown to be negligible then the risks from most other
pathogens would also be negligible.  If the risks from Giardia
cysts are shown to be significant, then the risks from exposure
to pathogens less resistant to disinfection than Giardia may also
be significant, but to a lesser degree.


Modeling Results

Aggregate Analysis

     Figures 6-13 illustrate the predicted ability, based on
the DBPRAM, for public drinking water supplies in the U.S. to
meet different MCL targets for THAAs and TTHMs by different
treatment technologies given treatment constraints to control for
pathogenic organisms.  The analyses in these Figures represent
systems serving more than 10,000 people each and 103,000,000
collectively.  These systems currently filter but do not soften.
For the purpose of estimating exposure and associated risk, each
of the 100 hypothetical water treatment plants was assumed to
Represent the same population (1,030,000 people).  A detailed
description of the information in Figure 6 is provided as an
example of how to interpret the information in Figures 7-13 which
are similarly structured.

Figure 6 - SffTR ft/Alternative Disinfection.  This Figure
illustrates the analysis under which systems were allowed to use
alternate disinfectants and DBF precursor technologies to meet
MCL targets for THAAs ranging from 60 to 10 ug/1.

Column one indicates the MCL target level for THAAs  (i.e., mono,
di, and tri-chloroacetic acid, and mono and dibromo-acetic acid).

Column two indicates the degree of treatment applied (see
treatment code description in upper left). Treatment code #1
indicates no treatment has been applied other than that necessary
to meet the SWTR scenario treatment constraints.

Column thr«« indicates the percentage of systems predicted to use
different levels of treatment (from column two) to meet the
target MCL based on lowest cost.  For example, at an MCL of 30
the model predicts that 48% of the systems would meet this MCL
without any treatment change; 12% would eliminate
prechlorination; 22% would eliminate prechlorination and add
chloramines; 14% would eliminate prechlorination, add
chloramines, and increase alum dose to improve precursor removal;
etc. The numbers in column three sum to 100% for each MCL
alternative.

Column four indicates the cumulative percent of systems meeting
the target MCL through the specified level of treatment
(graphically represented in the small figure on the right hand

                                14

-------
 side  of  the page).  For example,  at an MCL of 20  ug/1,  83%  of  the
 systems  would be able to meet this standard (a  level  of  16 ug/1
 with  the 20% margin of safety) without using GAC (treatment code
 #6) .  It  is  important to note that all of  the treatment
 constraints would have to be met through  each level of specified
 treatment.  Thus,  if chloramines  are used,  the system  must  still
 provide  enough CT with either chlorine or ozone  (if chlorine  can
 not be feasibly used,  i.e.,  treatment code /5) prior  to  the
 addition of ammonia,  to achieve  the 0.6 log inactivation of
 Giardia  cysts necessary to minimally meet the SWTR.

 If an MCL for THAAs were set at  10 ug/1,  the model predicts that
 only  89% (last line in column) would be able to  meet  this  MCL
 even  with installation of GAC following filtration.   As  a  result,
 11% of systems would qualify for a variance if they installed GAC
 (see  small  figure in lower left).

 Columns  five and six indicate the predicted mean concentration of
 TTHMs and THAAs for the systems  through the indicated level of
 treatment to meet the target THAA MCL. Note the effect  of
 decreasing  TTHMs as a function of lowering the MCL for HAAs
 without  having any MCL for TTHMs.  The fourth small figure  from
 the bottom  left side of the  page graphically represents  this
 relationship.

 The first lines in each MCL  target grouping indicate  the
 predicted occurrence of TTHMs (55 ug/1) and THAAs (24 ug/1) with
 no treatment beyond the requirement to minimally meet the  SWTR
 scenario treatment constraints using chlorine as the  disinfectant
 and not  having any MCL for either TTHMs or THAAs.  However, this
 mean  value  does not include  systems which are unable  to  meet  the
-treatment constraints.  Lines 2  through 6 under  each  MCL grouping
 indicate the cumulative mean concentration through each  treatment
 tier, i.e., the average of those meeting  and not meeting the  MCL
 through  the indicated level  of treatment.  Thus,  the mean values
 in columns  5 and 6 through line  5 under-represent the expected
 mean  value  for all systems making no treatment  changes to  comply
 with  an  MCL.

 The mean value- in line 6 for each MCL grouping  represents  the
 mean  value  for all 100 hypothetical systems having made  treatment
 changes  to  comply with the target MCL.

 Column 7 shows the predicted cancer cases per year, at different
 MCLs  for THAAs, caused by exposure from each of  the THMs,
 assuming the risk factors at the 95% upper confidence interval.
 For example, a THAA standard of  20 ug/1 (with no standard  for
 TTHMs) is predicted to reduce the incidence of  cancer attributed
 to THMs  from at least 33.8 cases per year to 16.1 cases  per year.
 A more precise way to interpret  the data  is to  consider  only  line
 6 under  each MCL grouping.  For  example,  a standard of 60  ug/1
 vs. a standard of 20 ug/1, assuming no TTHM standard  were  in

                                 15

-------
place, is predicted to reduce the incidence from 27.9 to 16.i
cases per year from exposure to THMs.

Column 8 indicates predicted cancer cases per year, at different
THAA MCLs,  attributed to THMs using the risk factors based on
the maximum likelihood estimate. Note the much fever predicted
incidents of cancer versus those for the upper 95% confidence
interval estimates.

Column 9 indicates the predicted cancer cases per year caused by
exposure from THAAs assuming the risk factors are based on the
95% upper confidence interval.

Column 10 indicates the predicted cases of cancer attributed to
THAAs (i.e., only di and trichloroacetic acid since no risk
factors were available for the other HAAs), assuming the risk
factors are based on the maximum likelihood estimate.  The second
small figure on the bottom left represents the combined annual
cancer cases caused from exposure to both THAAs and TTHMs using
the MLE risk factors.

Column 11 indicates the predicted .incidence of infection per
year.  As the target MCL for THAAs decreases, the predicted
incidence of infections increases (see also the small third
figure from the bottom left).  There are several factors
contributing to this phenomena.

The first is the presence of more DBF precursors preceding
disinfection resulting in higher disinfectant demand. As more
precursors are removed at lover MCLs because of the adoption of
precursor removal technologies, a lover disinfectant demand
exists in the vater and less disinfectant can be used to meet the
constraint of providing a disinfectant residual at the end of the
distribution system. In other words, in high oxidant demand
waters the constraint of maintaining a residual in the
distribution system is much more significant in affecting a
higher level of inactivation than the constraint of minimally
having to achieve the 0.5 log Giardia. inactivation criterion.

A second is that many of the systems that switch to chloramines
in lieu of chlorine as a residual disinfectant to lower THM or
HAA formation have high disinfectant demand water.   When using
chlorine, such systems have to maintain a high dosage, and
subsequently a high CT prior to the first customer, in order to
maintain a free chlorine residual in the distribution system.
When these systems switched to a chloramine residual in the
distribution system, a much weaker disinfectant than chlorine,
these systems no longer provided the high level of inactivation
prior to the first customer.  Rather, their level of inactivation
was reduced to only meeting the 0.5 log inactivation criterion.
                                16

-------
The difference in predicted Giardia incidence at high versus low
THAAs may be understated because the SWTR scenario assumes that
systems are designed and operated to minimally meet the
disinfection constraints,  in reality, many systems may provide
higher levels of inactivation than would be minimally required
and still meet the high THAA MCLs.  However, recent water
treatment plant data (LeChevalier et al 1991) appear to support
the validity of the predicted Giardia infections in the high MCL
range (Cromwell et al 1992).

Column 12 shows the predicted incidence of Giardia outbreaks
at the 90th percentile for different THAA MCLs. For example, ten
out of every one hundred systems, those with the highest Giardia
cyst concentrations, are predicted to have an outbreak within
every 3,089,000 years.at an MCL of 60 ug/1, and within every
0.658 years at an MCL of 40 ug/1.

     The large deviations in the predicted outbreak intervals for
different target MCLs are explained by the large deviations
around the mean among the systems with high concentrations of
Giardia in their source water and the concurrent significant
decreases in inactivation that theoretically could occur with
treatment modifications.    The model predictions suggest
potentially very significant increases in the occurrence of
waterborne outbreaks as the MCL for THAAs is lowered, i.e.,
without a parallel requirement for increased stringency of
regulation to also control for Giardia (e.g., an ESWTR).

     It is important to note that the assumptions in this
analysis,  such as the definition of what constitutes an outbreak
or a symptomatic response (i.e., infection versus illness) have
large uncertainties and will substantially affect the predicted
output.  The assumed criteria that define an outbreak, for the
purpose of this analysis, appear to be overly conservative and
the predicted outbreak intervals cannot be considered reliable
estimates.  Probably the most important indication of this
analysis (rather than the specific predicted outbreak intervals)
are the potentially large relative differences in likelihood for
waterborne outbreaks to occur as a function of meeting
increasingly stringent MCLs for DBFs.


Figure 7 - SWTR Without Alternative Disinfection. Under this
modeling scenario systems can only use precursor removal
technologies to meet the MCL target.  The result of this
constraint is that more systems would require use of GAC to
achieve the target MCLs.  If an MCL were set at 10 ug/1 the model
predicts that, even with use of GAC, 17% of the systems would not
be able to meet this limit.
                                17

-------
     At an MCL of 60 ug/1, this scenario predicts a higher number
of Giardia infections  (246,531) than those predicted when  "
alternate disinfectants could be used (213,045, Figure 6).  This
higher number is attributed to more systems reducing disinfectant
demand than in Figure  6, because of greater use of precursor
removal technologies,  and subsequently lowering their
disinfection dose while still being able to meet the minimal
disinfection treatment requirements.

Figure 8 - Enhanced SffTR filth Alternative Disinfection. This
analysis differs from  Figure 6 in that systems must achieve
proportionally higher  levels of Giardia inactivation as the
source water concentration of Giardia cysts increases. Predicted
Giardia infection incidence is reduced to below the 1/10,000
infection per person per year goal of the SWTR through
implementation of the  recommended guidance.  The probability for
outbreaks to occur is  eliminated.

The predicted level of treatment necessary to meet the different
MCL targets (columns 2, 3, and  4) is not much different than
that predicted in Figure 6 without the ESWTR.  The additional
contact time within the plant required by the ESWTR (which is
assumed can be added,  if needed) to meet the inactivation target,
does not significantly change the predicted concentration of
THAAs at the average customer.  As average residence time in the*
distribution system is about 1.6 days, the additional 1 to 3
hours in the plant, if chlorine is used, does not significantly
increase THAA concentration to the average customer.

     These results imply that requiring greater inactivation of
Qiardia for more highly contaminated source waters will not have
a significant impact on the national costs of compliance.  The
cost for increased contact basin size following
coagulation/filtration or GAG to meet the inactivation
requirements of the ESWTR does not significantly influence the
total cost.  However,  if systems do not have space to increase
contact basin size, other options might then be required  (e.g., a
stronger disinfectant  such as ozone which could significantly
increase treatment costs).

Figure 9 - Enhanced SWTR Without Alternative Disinfection. This
scenario is the same as Figure 8 except systems can only use
precursor removal technologies to achieve the different THAA
MCLs.  The results in  Figure 9 are similar to those in Figure 8
except higher levels of precursor removal technology are needed
to achieve the MCLs.   The technologies predicted to achieve the
MCLs are almost identical to those in Figure 7 where no alternate
disinfectants were allowed but less stringent disinfection
requirements were required.  This result is explained by the same
logic as described under the description of Figure 8.


                                18

-------
Figures 10 - 13 - These Figures are analogous to Figures 6 -9
except analysis targets for TTHMs in lieu of THAAs. Note how the
concentrations of predicted THAAs are affected by systems
targeting to achieve different TTHM MCLs, indicated in the
fourth figures from the bottom left.

The results at the target MCL of 100 ug/1 in Figures 10 are of
particular interest. The constraints in the model that led to
these results most closely represent the current situation today
under the existing TTHM standard of 100 ug/1. The main difference
is that some plants have not yet made treatment modifications to
comply with the requirements of the SWTR, total coliform, and
lead rules and that this may influence existing levels of TTHMs.

Figure 10 indicates that 61% of the systems using only chlorine
as their disinfectant would be able to meet a TTHM MCL of 100
ug/1 without making any treatment modifications.  It also
indicates that 21% would eliminate prechlorination, and another
18% would switch to chloramination in order to meet a TTHM MCL of
100.  These results appear to approximate the changes that large
filtered systems actually made to comply with the TTHM rule since
it was promulgated in 1979 (Gelderloos 1992; Cromwell et al
X992).

Figure 10 also indicates predicted mean of 40 ug/1 for TTHMs and
20 ug/1 for THAAs in the distribution systems for all systems
required to comply with a TTHM MCL of 100 ug/1.  These results
also appear to approximate the existing mean occurrence level for
these groups of compounds in large filtered supplies today
(Cromwell et al 1992).
Disaggregate Analysis  (Changes in Risk to Individual Systems)

     The modeling results discussed so far, which pertain to 103
million people, indicate that the risk from Giardia infections
may increase by several hundred thousand per year under
significantly more stringent DBF regulations than currently
exist, unless-revisions are made to the SWTR.  What may be even
more significant are the potential changes in microbial risk that
might occur in populations of individual systems under new DBF
regulations.


     In order to understand the potential for changes in risk at
the system level it becomes essential to examine the source water
quality and level of treatment for individual systems.  Figure 14
presents a plot of influent Giardia concentrations (Yw axis)
versus estimated total log reductions achieved ("X" axis) at the
plant effluent  (first customer) based upon survey results from 46
different plants in the U.S.  (LeChevalier et al 1991).  A

                                19

-------
critical underlying feature of these data, which can be
considered a pre-sWTR baseline, made clear by the random scatter
pattern in this plot, is that there appears to be no correlation
between the influent Giardia concentration and the total log
reduction achieved.

     The vertical line at 3 log reduction indicates the minimal
level of treatment required through the first customer by the
SWTR, although higher levels of treatment may be needed to
maintain the disinfectant residual in the distribution system.
The diagonal line through the points represents the formula of
providing the level of treatment, prescribed in the SWTR Guidance
Manual, for systems to achieve a level of risk of less than one
infection per 10,000 people per year (1/10,000). The points to
the right of the diagonal line represent plants achieving more
total log reduction than needed to meet the 1/10,000 risk target;
points to the left represent plants achieving less total log
reduction than needed to meet the 1/10,000 risk target.

     The lack of correlation between the influent Giardia concen-
tration and the total log reduction achieved could be attribut-
able to numerous factors: l) the lack of monitoring capability
for Giardia and the knowledge of when to treat more or less; 2)
attempts to comply with the current TTHM MCL (significantly, many
of the plants to the left of the line are using chloramines.); 3)
attempts to comply with current standards for coliforms and to
maintain a chlorine residual in the distribution system (e.g.,
Ten State Standards); or 4) other treatment objectives served by
pre-disinfection.

     Figure 14 illustrates the relevance of considering changes
in risk at the plant level rather than the average levels across
all plants. The clear implication of the data in Figure 14 is
that the minimum 3-log reduction requirement of the SWTR does not
appear sufficient to achieve the target 1/10,000 risk level in
most systems.  The higher levels of total reduction specified in
the SWTR Guidance Document appear to be necessary in order to
assure meeting the target risk level.  Another relevant issue is
how the distribution of points indicated in Figure 14 might
change as a function of different DBF regulatory targets.

     The Monte Carlo simulation framework established in the
DBPRAM is extremely well-suited to analyzing the relationship
between influent Giardia levels and total log-reductions achieved
at each individual plant.  Since the influent Giardia
distribution used to drive the simulation was developed from the
above-described data,  some direct similarities to the findings
are to be expected in the model results.  Unfortunately, a
complete calibration check of the model against the survey
findings  (LeChevallier et al 1991) is not possible because the
data were collected in a pre-SWTR compliance environment whereas


                                20

-------
the model is constructed to include the SWTR requirements as
initial constraints.  Some similarities are nonetheless evident.

     As discussed above, there are many possible reasons why the
data in Figure 14 indicate a lack of correlation between the
influent Giardia level and the total log reduction achieved.  In
the over-simplified reality of the DBPRAM, a few assumptions
dominate this relationship.  Figure 15 presents a plot similar to
that of Figure 14 showing simulation results for a scenario that
assumes implementation of the SWTR (with alternate disinfectants)
and a TTHM standard of 100 ug/1.  Of 100 simulated plants, 39
were determined by the DBPRAM algorithm to require treatment
adjustments in order to meet a TTHM MCL of 100 ug/1.  Only these
39 points are plotted in Figure 15.  The plotted points reflect
the relationship between influent Giardia and total log reduction
after compliance with these modeling constraints.

     Many of the 39 points in Figure 15 appear in a vertical line
reflecting the SWTR requirement for a minimum of 3-logs of total
reduction.  By contrast to the pre-SWTR baseline reflected in the
data in Figure 14, some of the points along the 3-log line in
Figure 15 would represent a higher level of total reduction than
in the pre-SWTR condition while others would represent a lower
level.  In the simplified model logic, those plants that were not
previously achieving a 3-log level would have increased their
level of total reduction to equal 3-logs, while some of those
achieving more than a 3-log reduction would have dropped their
level of total reduction to the 3-log minimum.  Notably, all the
points aligned vertically along the 3-log minimum are to the left
of the diagonal line that defines the 1/10,000 risk threshold
that the SWTR sought to achieve, indicating that they are
incurring risks greater than this target level.

     The remaining points plotted in Figure 15 show more than 3-
logs of total reduction.  Some of them are above the 1/10,000
risk threshold — to the right of the diagonal — while others
are below it — to the left of the diagonal.  In tracing the
source of these results through the model logic, it appears that
they exhibit more than the 3-log minimum due to the additional
SWTR constraint of maintaining a disinfectant residual in the
distribution system.  The simulated plants achieving more than a
3-log total reduction at the first customer (see Figure 15)
represent conditions where enough total organic carbon (TOC) is
present to require elevated chlorine dosages to overcome the
chlorine demand of the TOC in the distribution system, but yet
where the resulting contribution to TTHM formation is not
significant enough to require additional TOC removal.

     Figure 16 illustrates the effect of a tighter by-product MCL
of 25 ug/1.  Here 83 out of 103 plants would be required to
change treatment in order to comply.  As shown in the graph, the

                                21

-------
result is that all the points are aligned along the 3-log minimum
line.  Those plants with more than 3-logs of total reduction in
Figure 15 have been guided by the treatment algorithm into TOC
removal strategies which in turn reduced the chlorine demands
that had to be overcome in order to meet the SWTR residual
requirement in the distribution system.  As a result, the model
predicts that all plants required to change treatment to meet a
TTHM standard of 25 ug/1 would reduce their total log reduction
to the 3-log minimum.

     As shown in Figure 16, the convergence to the 3-log minimum
results in a risk level higher than the 1/10,000 target for all
of the 83 plants compelled to modify treatment in meeting a TTHM
MCL of 25 ug/1.  The resulting change in endemic incidence of
Giardia infections will be larger for some plants than for
others.  In particular, there may be some plants for which the
co-occurrence of high influent TOC and high influent Giardia
result in a very large increase in incidence due to significant
drops in chlorine dosages for compliance with SWTR residual
requirement after TOC removal has been triggered by the need to
meet a TTHM MCL.  The increase in incidence of Giardia infections
has been computed based on the change in total log-reduction
between one MCL and another, as computed by the OBPRAM.  Results
of this analysis are presented in Figures 17 and 18.

     Figure 17 plots the cumulative distribution of the 100
simulated plants by their level of increase in annual Giardia
infections per 1 million persons when moving from a TTHM MCL of
100 ug/1 to an MCL of 75 ug/1.  The plot confirms that some
plants would face extreme conditions, resulting in significant
increases in Giardia infections.  Nearly 10 percent of the plants
could experience increases of between 1,000 and 10,000 infections
per 1 million population.  The top end of this range approaches
outbreak proportions.  Figure 18 shows that the impact of a shift
from a TTHM MCL of 100 ug/1 to a TTHM MCL of 25 ug/1 would
produce increases in Giardia incidence on this same order of
magnitude for over 30 percent of the plants.  Figures 19 and 20
present the same results that are plotted in Figures 17 and 18 as
bar charts,  emphasizing the impacts at the extreme ends of the
range.        -

     The ultimate analysis that must be made is a comparison
between the increase in incidence of Giardia infections and the
corresponding decrease in cancer incidence that results from
alternative MCLs.  Combining the above analysis of Giardia
infections with the by-product concentrations simultaneously
projected by the DBPRAM produces such an analysis.  Results are
summarized graphically in the log-log plots presented in Figures
21 and 22.  Figure 21 plots increased incidence of Giardia
infections per 1 million persons versus decreased cancer
incidence per 1 million persons for the .plants having to change


                                22

-------
treatment when the TTHM MCL is changed from loo ug/1 to 75 ug/1
in a SWTR scenario (with alternate disinfectants).

     Figure 22 presents the same picture for the plants that
would have to change treatment if the TTHM MCL is changed from
100 ug/1 to 25 ug/1.  The pattern of the points plotted in these
diagrams indicates that the degree of change in cancer incidence
spans only one order of magnitude (less than one case per 1
million persons per year) while the degree of change in incidence
of Giardia infections spans several orders of magnitude (from one
to more than 10,000 infections per 1 million persons per year).
This outcome indicates that the microbial risk is more sensitive
to treatment changes needed to meet TTHM MCLs.

     Figures 23 through 30 present the same series of graphs as
Figures 15 through 22 except that the entire analysis is
structured in terms of alternative MCLs for Total Haloacetic
acids.  The interpretation is identical to that described above.

     One obvious direction indicated by these results would be to
strengthen the SWTR before proceeding with requirements for
additional by-product controls.  In the ESWTR scenario, it is
assumed that the portion of the SWTR Guidance Document which
recommends an additional log of inactivation for each additional
log of influent Giardia. is converted to a mandatory requirement.
In terms of the graphs presented in Figures 15 and 16, this
requirement would mean that the model logic would result in the
plants being lined up along or to the right of the diagonal line
representing the 1/10,000 target risk level rather than along the
vertical line representing the 3-log minimum required by the
present SWTR.  By definition of the 1/10,000 risk level as the
target, there would therefore be no microbial risk trade-offs.

     The evaluation of this option with the DBPRAM showed that
adherence to the 1/10,000 risk level resulted in very little
additional cancer incidence and in very little modification in
treatment, and therefore very little additional cost.  This
outcome results from the model assumption that additional
inactivation requirements at the first customer would be met with
additional contact time at the plant effluent.  This additional
contact time allows inactivation requirements to be met without
increased chlorine dosages.  While the additional contact time
contributes to inactivation, it does not contribute much in the
way of additional by-product formation because the additional
increment of contact time is trivial in proportion to the total
residence time in the distribution system.
                                23

-------
Model Shortcomings and Next Steps

This discussion will only pertain to model analysis or
modifications anticipated to occur in support of the development
of a proposed rule for D/DBPs.  Much more research will be needed
over a longer time frame to address the unknowns identified in
the shaded areas of Figure 1.

Develop Predictive capability for Bromate

     Under the modeling scenarios with alternative disinfectants
allowed, the DBPRAM currently selects ozone, with chloramines as
a residual disinfectant, as an alternative disinfectant to
chlorine to achieve lower levels of THMs and HAAs.  This
selection may lead to high levels of bromate formation in systems
with high bromide concentration in the source water (AWWA, 1992).

     Since exposure from bromate may result in greater risk than
exposure from THMs or HAAs (Table II), depending upon the bromide
concentration in the source water and the subsequent formation of
bromate upon ozonation, the model may be inappropriately
selecting ozone in the compliance sorting routine.  The model may
also be underestimating the total cancer risk from the selected
distribution of technologies to meet a target MCL for THMs or
HAAs.  EPA is now developing formation equations for bromate and
intends to include these equations in the modeling analysis in
developing the proposed rule*

Extend Modeling Analysis to Smaller Systems

     The current analysis only pertains to systems serving more
than 10,000 people.  Since EPA intends for the new DBP
regulations to pertain to smaller systems, it becomes important
to evaluate the potential risk and cost impacts in such systems.
EPA is now expanding the application of DBPRAM  to systems
serving less than  10,000 people. This analysis will consider the
following factors as being different from larger systems: a)
possible smaller distribution system residence times, b) the
potential for higher treatment failure rates, c) the effects of
not having to already meet an existing TTHM MCL on net changes in
total risk and d)  economies of scale in installing and operating
different technologies  (e.g., using membrane filtration in lieu
of GAC as a precursor removal technology).

Extend Modeling Analysis to Systems Using Lime Softening

     Systems that  use surface water and lime softening may
represent those systems having the greatest difficulty in meeting
microbial and DBP  standards.  Since these systems operate at high
pH, it is more difficult to achieve effective inactivation with


                                24

-------
 most disinfectants.   High pH conditions may also make it
 substantially more difficult to remove DBP precursors.
 Therefore,  systems using lime softening may be the best indicator
 for characterizing potential changes in risk from exposure to
 DBFs and pathogens that might result under new regulations.

      EPA is developing formation equations to predict THMs and
 HAAs for systems using lime softening.  Also, new data are being
 developed to evaluate the inactivation efficiency of disinfection
 at high pHs (Logsdon et al, 1992).  EPA anticipates using this new
 information to expand DBPRAM to predict potential changes in risk
 in lime softening systems under new OBP regulations.

 Model Validation and Refinement

      Some model validation analysis has already been conducted
 (Cromwell et al 1992, USEPAc, 1992) but much more is needed.
 DBP occurrence data being collected at water treatment plants
 will be compared with model predictions to evaluate the accuracy
 of model predictions. Sensitivity analyses are being conducted to
 determine which of the parameters and model assumptions have the
 greatest impact on affecting model predictions.  EPA is also
 considering the use of different TKM and HAA formation equations
 than are now being used in the DBPRAM (AWWA 1992).

      The DBPRAM currently assumes that the data collected from
 one survey (LeChevalier et. al., 1991) is representative of the
 amplitude and variation among communities using filtered surface
 supplies in the US.   The DBPRAM also assumes that the data
 collected from another source (Hibler et. al., 1987) is
 representative of the variation within a community.  EPA will
jreevaluate and possibly modify these assumptions as more data
 become available.  Also, several "what if" scenarios pertaining
 to the probabilities that infections may lead to illness, and the
 rate of illnesses that would likely cause an outbreak to be
 identified, will be considered in future model runs.

 Conclusions

 1. EPA has developed a model that predicts the occurrence of
 trihalomethanes (THMs), haloacetic acids (mono, di, and tri-
 chloroacetic acid; and mono, and di-bromoacetic acid,
 collectively cited as THAAs), and Giardia cysts in finished
 drinking waters in the U.S. as a function of source water quality
 and treatment practice.  The model predicts risk of cancer from
 THMs, and di- and trichloroacetic acid, and risk of infection
 from Giardia cysts.  The model is being used to predict the
 potential changes in treatment and risk that might result from
 different regulatory targets.
                                25

-------
 2. Model  predictions  currently  only  pertain  to  filtered  systems
 that  do not  use  lime  softening,  serve greater than  10,000 people,
 and collectively distribute  water  to 103 million  people  in the
 U.S.   The model  is  currently being expanded  to  include analysis
 for systems'serving less  than 10,000 people  and those which use
 lime  softening.

 3. More data are needed to evaluate  the validity  of the  model
 predictions.   Refinements in the model assumptions  and predictive
 equations are anticipated and,  therefore, the model predictions
 should be considered  tentative.  Two risk estimates for  cancer
 are given: numbers  without parenthesis indicate estimates based
 upon  maximum likelihood risk factors, and numbers within
 parenthesis  indicate  estimates  based upon the upper 95%
 confidence interval.   Risk estimates for Giardia  infections are
 based upon a maximum  likelihood analysis.

 4. Model  predictions  indicate that at the existing  TTKM  MCL of
 100 ug/1, following compliance  with  the SWTR, total coliform
 rule,  and lead rules,  45  (173)  cases of cancer  per  year  will
 occur based  upon predicted exposure  from THMs  [1, (24) cases] and
 di- and tri-chloroacetic  acid [44, (139) cases].

 5. Model  predictions  indicate that if an MCL of 25  ug/1  were  set
 for TTHMs, and alternative disinfectants to  chlorine were not
 allowed,  41% of  the systems  would  be able to meet this standard
.without the  use  of  GAC.   5%  of  the systems could  not meet this
 standard  with the use of  deep bed  GAC following optimized
 filtration for precursor  removal,  with allowed use of alternate
 disinfectants, 97%  of the systems  could meet a  25 ug/1 standard
 without the  use  of  GAC.

 6.  Model predictions indicate  that  if an MCL of  10 ug/1 were set
 for THAAs, and alternative disinfectants to  chlorine were not
 allowed,  43% of  the systems  would  be able to meet this standard
 without the  use  of  GAC.   17% of the  systems  would not be able to
 meet  this standard  with the  use of deep bed  GAC following
 optimized filtration  for  precursor removal.  With allowed use of
 alternate disinfectants,  50% of the  systems  could meet the
 standard  without the  use  of  GAC. 11% of the  systems could not
 meet  the  standard with the use  of  GAC.

 7. Model  predictions  indicate that at a TTHM MCL  of 25 ug/1,  with
 no other  DBF standard and allowed  use of alternate  disinfectants,
 27  (102)  cancer  cases per year  would occur based  upon predicted
 exposure  from THMs  0.4 (10 cases)  and di- and  tri-  chloroacetic
 acid  26.2 (82 cases). At a  THAA MCL of  10 ug/1 (with no other
 DBF standard),  12 (45) cancer cases  per year would  occur based
 upon  predicted exposure  from THMs  [0.4  (10.4)  cases] and di  and
 tri-chloroacetic acid [11.3  (35.3) cases].


                                 26

-------
8.  Model predictions indicate that if systems were to only
minimally comply with SWTR, i.e., not follow guidance of
providing higher levels of inactivation for poorer source water
quality, 340,000 Giardia infections per year would result under
the TTHM MCL of 100 ug/1.  A new standard of 25 ug/1 for TTHMs
alone would result in 170,000 additional Giardia infections per
year.  A new standard of 10 ug/1 for THAAs alone would lead to
226,000 additional Giardia infections per year.

9. Under any new DBF MCL, e.g., 75 ug/1 for TTHMs or 50 ug/1 for
THAAs, some systems would experience increases of 1000 to 10,000
Giardia infections per million people per year with decreases in
cancer risk of 0.1 to 1 case per million people per year (based
upon maximum likelihood estimates.  If MCLs were set at, e.g., 25
ug/1 for TTHMs or 10 ug/1 for THAAs, more than 25% of the systems
are predicted to experience this range of respective net changes
in risk.

10. The modeling results indicate that it is not possible to
conclude with certainty that the total risk will decrease as a
result of any specific regulatory action.

11. If an enhanced SWTR could be implemented to achieve an
infection rate of less than 1/10,000 at the first customer, where
higher levels of treatment for Giardia were required depending
upon the level of Giardia in the source water, the predicted
number of Giardia infections per year would be less than  500.

12. The model predicts that the universe of systems that use
surface water and which practice coagulation and filtration could
use almost the same decision tree of treatment technologies (but
with differences in the level of disinfection) to meet the
existing SWTR and stringent DBF MCLs as they could to meet an
enhanced SWTR and stringent DBF MCLs.  If this conclusion is
correct, the costs for complying with these two different
regulatory scenarios would not be significantly different even
though much greater microbial protection could be provided under
an enhanced SWTR,  if such a rule could be implemented.

13. If an enhanced SWTR could be implemented, any MCL for
controlling THMs or HAAs would lead to greater use of alternate
disinfectants, unless prevented through regulation, since these
technologies are more cost effective for lowering THMs or HAAs
than use of 6AC or membrane filtration.  Since the health risks
from DBFs of alternate disinfectants to chlorine are largely
unknown, it remains possible that greater use of alternate
disinfectants could lead to greater health risks than those from
chlorinated DBFs such as THMs and HAAs.
                                27

-------
 14.  Precluding the use of alternate disinfectants to achieve
 lower levels of THMs and HAAs would lead to substantially greater
 treatment costs than if they were allowed.   Also, use of
 alternate disinfectants allows for substantially lower THM and
 HAA concentrations to be achieved than if only precursor removal
 technologies were allowed.


 References

 Cromwell, J.E.;  Zhang, X.;  Letkiewicz,  F.  J.; Regli, S. and B.
 Macler.  "Analysis of Potential Tradeoffs In Regulation of
 Disinfection By-Products."   Office of Ground Water and Drinking
 Water Resource Center. Washington D.C. USEPA in press, 1992.

 Gelderloos,  A.B.; Harrington,G.W.; Owen, D.W.; Regli, S.:
 Schaefer, J.K.; Cromwell, J.E. and X. Zhang. "Simulation of
 Compliance Choices For Regulatory Impact Analysis". Office of
 Ground Water and Drinking Water Resource Center. Washington D.C.
 US EPA in press. 1992.

 Grubbs,  W. D., Macler, B. and S. Regli.   "Modelling Giardia
 Occurrence and Risk." Office of Ground Water and Drinking Water
 Resource Center. Washington D.C. USEPA.  EPA-811-B-92-005. 1992.

 Hibler,  C.P. Analysis of Municipal Water Samples for Cysts of
 Giardia. Report prepared for Office of Drinking Water, U.S. EPA,
 1987.

 LeChevallier, M. W., Norton, D.N. and R. G. Lee. Occurrence of
-Giardia and Cryptosporidium in Surface Water Supplies. Applied
 and Environmental Microbiology 57:2610-2616. 1991.

 Letkiewicz,  F.J.; Grubbs, W.G.; Lustiik, M.; Mosher, J.; Zhang,
 X. and S. Regli.  "Simulation of Raw Water Quality and Water
 Treatment Characteristics in Public Water Supplies in Support of
 the Disinfection By-Products Regulatory Impact Analysis". Office
 of Ground Water and Drinking Water Resource Center. Washington
 D.C. USEPA.  EPA-811-R-92-001. 1992.

 Logsdon, G,  et al.  The Removal and Disinfection Efficiency of
 Lime Softening Process for Giardia and Viruses. Poster
 presentation, AWWA Annual Conference, Van Couver, Canada.  1992.

 Rose, J.B.,  Haas, C.N., and Regli, S.  Risk Assessment and
 Control of Waterborne Giardiasis. Amer.  Jour. Public Health
 81:709-713.  1991.
                                 28

-------
USEPA.  National Primary Drinking Water Regulations: Filtration;
Disinfection; Turbidity, Giardia Iambiia, viruses, Legionella,
and Heterotrophic Bacteria. Final Rule. Fed.Reg. 54:27486 (June
29,1989).  1989a.

USEPA. National Primary Drinking Water Regulations: Total
Coliforms. Fed. Reg. 54:27544 (June 29, 1989). 1989b.

USEPA. Status Report on Development of D/DBP Regulations. Office
of Ground Water and Drinking Water. USEPA. Wash. D.C.
199la.

USEPA.  Guidance Manual for Compliance with the Filtration and
Disinfection Requirements for Public Water Systems Using Surface
Water Supplies; USEPA, Wash., D.C. 1991b.

USEPA. Occurrence Assessment for Disinfectants and Disinfection
By-Products (Phase 6a) in Public Drinking Water. Office of Ground
Water and Drinking Water Resource Center. Washington D.C. USEPA.
EPA-811-R-92-003. 1992a*

USEPA.  Status Report on Development of MCLGs for Disinfectants
and Disinfection By-products. Office of Science and Technology.
1992b.

USEPA o Water Treatment Plant User's Manual. Office of Ground
Water and Drinking Water Resource Center. Washington D.C. USEPA.
EPA-811-8-B-92-001. 1992C
                                29

-------
                             Table I
            Candidate Compounds For Regulation
trihalomethanes
possible        health effect
MCLG         and cancer status
*(RFD not      "(tentative - not yet
yet approved)  approved by CRAVE)
chloroform               0
bromodichloromethane     0
dibromochloromethane     60
bromoform               0

total THMs               0

haloacetic acids
trichloroacetic acid        100 /*g/l*
dichloroacetic acid         0
               cancer,  B2
               cancer,  B2
               liver, C
               cancer,  B2

               cancer, B2
               liver, C*
               cancer,  B2*
total HAs                 0              cancer, B2*
(mono, di, and tri chloroacetic acid; mono and dibromoacetic acid)
Other

chloral hydrate
bromate
chlorine
chloramines
chlorine dioxide
chlorite
chlorate
5/tg/l
0
4 mg/1*
3 mg/1**
0.8 mg/1*
0.3 mg/1 *
health advisory
liver,  C*
cancer,  B2*
blood,  D*
blood,  D*
blood, neurological, D*
blood, D*
** 4 mg/1 if measured as total chlorine

B2 compounds include possible human carcinogens, C compounds
include possible human carcinogens, and D compounds  include those
with inadequate or no human  and animal evidence of
carcinogenicity.

-------
                       Table II
        Drinking Water Concentrations G*g/l)  for
  Lifetime Cancer Risks (upper 95% confidence interval)
DBF
Bromodichloromethane

Bromoform

Chloroform

Dichloroacetic acid**

Bromate**
10-
                                   -5
known
range of
60
400
600
10
5
6
40
60
1
0.5
«-*
occur
0-100
0-50
0-340
0-80
?
•
** tentative risk estimate

-------
                             Table I
            Candidate Compounds For Regulation

trihalomethanes            possible         health effect
                          MCLG         and cancer status
                                         (tentative)**

chloroform                0               cancer, B2
bromodichloromethane      0               cancer, B2
dibromochloromethane      60 jig/1         liver,  C
bromoform                0               cancer, B2

total THMs                0               cancer, B2

haloacetic acids

trichloroacetic acid         100 /ig/1         liver,  C
dichloroacetic acid         0               cancer, B2

total HAs                 0               cancer, B2
(mono, di, and tri chloroacetic acid; mono and dibromoacetic acid)

Other

chloral hydrate             5 /zg/1           liver,  C
bromate                   0               cancer, B2
chlorine                   4 mg/1          blood,  D
chloramines                3 mg/1*         blood,  D
chlorine dioxide            0.8 mg/1         blood, neurological, D
chlorite                   0.3 mg/1         blood, D

* 4 mg/1 if measured as total chlorine
** B2 compounds include probable human carcinogens, C compounds
include possible human carcinogens,  and  D compounds  include those
with inadequate or no human and animal evidence of
carcinogenicity.

-------
                        Table II
        Drinking Water Concentrations 0*g/l)  for
   Lifetime Cancer Risks (upper 95% confidence interval)
 DBF
IQ"4
        known
10'!     range of
        occurrence
Bromodichloromethane

Bromoform

Chloroform

Dichloroacetic acid**

Bromate**
100
400
600
10
5
10
40
60
1
0.5
0-100
0-50
0-340
0-80
9
•
** tentative risk estimate

-------
                             DBPRAM Modelling Framework
   Monte Carlo Simulation
  of Influent Water Quality
           1
           Batch Mode
   Treatment and Distribution Model
, for Given Combinations of Treatments
Compliance Sorting Routine
 for Alternative DBP MCLs
  and MCL Combinations
                                                    ASSUMPTIONS:
                                                    Model Plant and
                                                Model Distribution System
                                                     Characteristics
                                                CONSTRAINTS:
                                         • Microbial Treatment Objectives
                                          (SWTR vs ESWTR)
                                         • Corrosion Control Objectives
                                         • Taste and Other Objectives
                                            Compliance Vectors:
                                            Percentage of Plants
                                         Requiring Given Treatment
                                      Combinations to Meet Given MCLs
EXPOSURE ASSESSMENT
Source
Water
Pathogens
G
i
a
r
d
i
a
O
t
h
*
r
; S

Distrib.
Sys.
Pathogens







THMs







Halo
Acids
C
h
1
o
r
0

B
r
0
m
o


Other
8
t
o
m
a
t
9
O
t
f)
£
r
$

                                                          I
  Unit Cost Vectors:
Estimated Unit Costs for
   Given Treatment
   Combinations to
  Meet Given MCLs

                                                                            I
                                                                       ''FEASIBILITY-
                                                                       ASSESSMENT
                                                                      Definition of BAT
                                                                       for Given MCL
                                                                       Combinations
                                                      National   ]
                                                        Cost
                                                      Estimate  J
                                                        i
:                                                      Regulatory
                                                    Impact Analysis
                                                    Under E012291
RISK ASSESSMENT
Source
Water
Pathogens
G
i
a
r
d
i
a
0
t
h
6-
f
8

Distrib.
Sys.
Pathogens







THMs







Halo
Acids
C
h
I
0
r
0

B
r
o
m
o


Other
B
r
0
m
a
t
e
o
t
h
e
r
3

                                                                           RISK
                                                                       MANAGEMENT
                                                                         Estimates
                                                                        ^Uncertainty
                                                                         •••Unknowns

                                                                          Decision
              Estimables
                                    Unknowns

-------
                                          Figure 2

                                      Major Parameters
 Mfcrobial
Occurrence
  Giardia

Raw Water Quality Parameters

IA
Hardness

IA
Turbidity

I/X
Alkalinity

|/\i
TOG

\^\
PH

\/\
UV
IA
Temp.

I/V
Bromide

                                       Treatment
                                        Process
                                         Model

-------
     Compliance Decision Tree
     Switch to
        i
      Increase
     Alum Dose
      Switch to
      03/NH2CI
        1
Add
GAC
                 Eliminate Pre-CL
                  (if applicable)
                      I
                1
   Add
Nanofiltration

                Add
                QAC


Increase
Alum Dose
i


                               1
   Add
Nanofiltration

-------
                                                             Figure  4

                                        DBF Cancer Risk Calculation Data and Analysis

                                                      \
     Population Exposed (persons) x OBP Concentration (jig/I) x Annual DW Risk Factor
/ cases/person/year \

         ufl/l
                                                                                        - cases/year
                                                            . Cl Lifetime Risk Factor:
MLE or Upper 95% a
 Lifetime Risk Factor
 (cases/person/lifeti

      mg/kg/day
iime\
               Conversion Factors:

                       1
          70kg x
                                          x 10»ug/rng
    Lifetime DW
    Risk Factor

/ cases/person/lifetime'

          UQ*
         Conversion Factor
     Annual DW
     Risk Factor

/ cases/person/year \
     /     70 years/lifetime     \


      Divide Lifetime DW Risk Factor
      into10MO»,10«to
      obtain DW concentrations i
      for these lifetime risk levels
MLE and Upper 95% Cl Lifetime
Risk Factors Used:
DBP
Chloroform
Bromodichlormethane (BDCM)
DibronKKhtoromethane (DBCM)
Bromotorm
Dichloroacettc acid
Tichkxoacetic acid
MLE
3.5xio*
8.2x10*
8.2x10*
4.4X104
88x10J
2.1xia»
Upper 95%
Cont. Int.
61x1&3
2.5xiaa
2.5x1 a2
7.9x1 a»
2.8x1 a*
6.3x10*
DW Concentrations (ufl/l)
for Lifetime Risks (for Upper 95% Cl)
10"
600
140
140
400
12.5
55
10*
60
14
14
40
1.25
5.5
10*
6
1.4
1.4
4
0.125
0.55
Comments
MLE from R. Bull, 95% Cl from draft HA
MLE from R. Bull, 95% Cl from draft HA
Assumed to be the same as BDCM
MLE from R. Bull. 95% Cl from draft HA
Provided by R. Cantilli (OST)
Provided by R. Cantilli (OST)

-------
                                       Overview ofGiardia Modelling
     Hibler Data
      73 plants
  variation between
  plants: lognormal
LeChevalier Data
    46 plants

variation between
plants: lognormal
         t
  temporal variation
    within plants:
delta negative binomial
1
distribution used
for influent simulation
1


Assuming:
• 48% recovery
• 13% viability
• deletion of
values estimated at
limit of detection
                                                 100
                                           simulated influents
                                                  1
                                           Treatment Model
                                                 100
                                      plant effluents (1 st customer)
                                                  t
                                           summary statistics
 distribution used
   for simulation
  of within plant
temporal variation
mean 90th%
1 I
compute expected
value of Rose dose
response function
compute expected value
and variance of Rose
dose response function
                              estimate
                             of endemic
                              incidence
                                rate
                    estimate
                   of outbreak
                      risk
                                          Assuming:

                                          • 5% failure rate
                                            (foss of 1  log)
                                          • 25% secondary
                                            infection rate
                                          •1% per 30 days
                                            outbreak threshold
                                          • viable cyst-
                                            infective in humans
                                          • infection-illness

-------
                                                                                                                                                                     Ontt3O-Jtn-K
                                                                         Figure 6
MODEL OUm/T(«Ml4M«*>«*4nlr>Q):
                                                            4 - pn-ahkM * •nmaila +
                                                            6 - pn-cMof « •Mnanta * •kin * aura
                                                            0 - pn-ahtoi •» ••onjila * ttom • arm* * QMS
MCt
(THMU)
00
00
00
oo
00
00
60
SO
60
60
60
•V
40
40
40
40
40
40
JO
10
JU
JU
JO
•
JU
30
20
JO
20
*>
to
10
10
10
10
«J
Tit.
Cafe
1
2
3
4
s
•
1
2
S
4
6
•
1
2
3
4
S
•
l
2
J
4
S
•
1
2
J
4
S
•
1
2
3
4
5
•
«at8v*
70
12
10
0
0
•
04
16
17
4
0
•
SO
10
22
6
1
O
44
12
22
14
3
1
3'
0
0
23
7
IT
10
3
1
22
6
80
OmiMW
*<*•*•.
 tiMUnwu tiw (nckKta* ZO% CM
-------
                                                                                                                                                                                        Oretl 30-Jan-ez
                                                                                    Figure 7
     MODEL OUTPUT (witaw w/o nRonkio): 8WTH W/O ALTERNATIVE D»»KCTON
                pm-ehtof » moOtf •kun doM
        pra-ohkw » Aim doM + OAC
MCL
(TMM*)
00
00
00
OP
so
so
so
e»
40
40
40
40
30
30
30
ao
20
JO
20
to
10
10
10
10
TM.
Cod*
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
.3
4
*f*8y».
Endk*
70
12
11
7
04
IS
14
7
SO
18
to
12
4S
12
10
«i
37
a
20
96
IB
3
21
ft
CunuMlM
*<48y».

24
10
14
•
24
10
13
•
0M
33.0
30.2
202
m*
33.6
20.0
26.6
W.7
33.6
26.0
26.0
M.0
33.0
27.0
24.6
mi
330
250
22.2
W.I
33.0
25.2
20.3
14.0
*» - - - 1 "-• -
._WiHUfMll
It
»A£
1.3
1.1
1.1
1.1
1.3
1.1
1.1
1.0
1.3
1.1
1.0
08
1.3
1.0
00
a0
1.3
10
0.0
tt7
1.3
0.0
0.7
««
""W
06«
1S33
1MO
134.0
1»7
1S33
137.3
1204
IMS
1543
1315
1160
tas.o
1533
120.0
1061
00-0
1S33
1W1
030
M.O
153,3
124.2
04.2
•0.0
56 	
IAE
40.4
4S.O
431
41.7
404
44.1
41 3
40.0
404
42.3
375
89.7
40.4
414
336
as-0
404
40.5
301
17*
40.4
300
27.1
M
-tzssA
kttaolkm



(Mint



aw^0»



277.142



•MOM



40HM6



H0.430
•atmMmsUan
^^^^BB5nP~^^^
(0M^ *.Vm. to OuHin^i



219M0



1.0*1



aoe0



0.174



0.094



aoo0
                 IMntOAC
       10   tO   »   M

            fHAAMCL (HfA4
It    W    10    40    00

        tHU MCL (y«AJ
•p«Mnl d *yM«n» kwWIttig Mch l/Mlmtnl tlw (InckjdM 20K ovw-d««tgn tictcK)
•cumuMlM pwcwit a« (yolwra ibl«lo moot MCL * Mch IfMlnMnI llw (toludM 20% mw-daolgn Itdor)
' m**n canccntoallan «l Mch Iradrrant tlw c* oil syMwm; twM mMIng MCL (ndttwwnal mMlhg MCL
        1O% o( papuMkxi «npoo«J M lit* cuMomw
      * lo> Ban p«c«ntll*

-------
                                                                                                                                                                           Doit: ao-Jtn-02
                                                                             Figure 8
   MOGELOUmfTa): ENHANCED8«nRWVM.1BVkTnEIMMirECnON
           Tl - Ml ra«iMig tfl*M» tntfMnl modi
           2 - (IMMMipiv-iMafMkn
           a - dtatoai Bt»-cMo« * add MMnunli
                                                                                     1
iwcvil ol •ytura,
"»«
ion.
THMte)
00
80
80
00
80
00
60
60
60
60
60
00
40
40
40
40
40
40
30
30
30
ao
30
00
ao
ao
ao
ao
ao


to
10
to
10
10
w
tfummmf —
W.
Cote
1
2
3
4
6
•
1
2
3
4
6
•
1
2
























«<*8y*
Endktf
70
11
17
2
0
0
02
10
10
3
1
a
68
10
10
10
2
P
47
12
12
16
10
1
35
10
4
16
10
1?

17
3
0
16
16
•



MC
1.3
1.2
l.t
1.1
1.1
M
1.3
1.2
1.1
1.1
1.1
M
1.3
1.1
1.0
1.0
10
1.0
1.3
1.0
0.0
0.6
O.6
ao
1.3
1.0
0.6
0.7
0.6
ao

1.3
1.0
0.6
0.0
0.4
*«

IW
OS*
1S83
1432
1440
143.4
1434
HM
1683
130.6
138.3
1381
136.3
1«M
1S6.3
, 13&1
131.6
1232
121.7
»»».»
IS83
1326
1249
107.2
101.6
MM
1683
1280
1200
96.9
82.0
OM

1683
127.0
1220
88.6
87.8
18,4

Jto
HUE
60.3
48.1
40.6
48.1
40.1
40.1
60.3
44.8
44.8
43.6
43.6
 *.V«. to OMbmfc





• WM»





MMy





WhBy





hOjttf





MMy






WMy
                                                                                                                           100%
                                                                                                                                                               00
                                I"
                                                                                                              I

                                                                                                              I
                                               A UCt <««M
  IrMlminl Ito (incMM iO% orar-dMOn IKIOC)
M>l* lo im« MCt. M Mch U*«m«nl Uw (iickxjM 2O% avw -dMlgn ticttn)
                  ton* mMttna MCt «nd tx>w na tnMtttg MCt

-------
                                                                                                                                                                               Drdt 3O-Jtm-IU>
                                                                       Figure 9
    MODEL OUTPUT (witac* w/o «*mlnfl): ENHANCED 8WIH ¥«D ALIBWAT1VE DBIIFECIION
     1 - not rapMifl *"•>•• IntfnnnlinodMMlkin
     3 - «IMMte pm-cMofkuMan
     • - rfMxl* pm-flhtof » MM«y Oun M
                *• iftMtdaw * O*C
MCL
(IHAto)
00
80
00
•0
60
60
60
M
40
4O
40
40
30
30
30
•0
AJ
*'
J1J
»
IU
10
10
to
lit.
Goto
1
2
3
4
1
2
3
4
1
2
3
• 4
1
2
3
4
i
J
>
4
i
/
)
4
KalSr*
Endhg-
70
It
12
7
02
10
16
f
SO
to
ts
1»
47
12
21
3D
V,
to
JO
»
i;
3
22
M
«<48ya.

67
44
40
»
57
42
. 30
91
S7
4O
32
M
57
30
28
I/
i^LOn.?
26
21
20
10
26
21
10
1*
26
20
10
M
25
10
18
13
25
10
14
' 0
26
10
13
0
Cm«iki|
FTW fe
06«
340
. 31.2
20.0
*M
34.0
30.7
20.2
•M
34.0
20.0
27.3
tt.0
340
200
263
as.*
340
200
227
f*«
34.0
20.1
200
M4
MLE
1.3
1.2
1.1
».1
1.3
1.2
1.1
1.1
1.3
1.1
10
OL*
13
to
00
ao
13
1.O
00
0.7
1.3
1.O
0.0
at
*ta««,
HMto
06%
150.3
1432
1353
ttt.0.
160.3
1306
120.4
1MB
150.3
1361
1102
MM.4
15O3
1325
100.2
02.0
150.3
120.0
05.1
6»-4
1603
127.0
040
MM)
ULE
60.3
401
43.5
42.1
603
44.0
41.3
400
603
434
30.3
•&•
503
42.0
341
ao.4
50.3
41 4
30.5
W.»
60.3
41.0
27.2
M
-G&SA
htaBUoM



8M



aw



270



•10



MO



MO
•nfctnhiHton*
nJilm^r ~B*
(00») ».Yi» to aulbiwk



brfbkkM
•<^wi



fcigoj _H_
VHPHiy



MhRy



hIMy



WM»



MMy
                                                                                                                              100%

                                                                                                                              90%

                                                                                                                            O 8°*
                                                                                                                            3
                                                                                                                            0 70%

                                                                                                                            tu 00%
                                                                                                                            £
                                                                                                                            (0

                                                                                                                            §40%

                                                                                                                            ra 30%

                                                                                                                            ° 20%

                                                                                                                              10%

                                                                                                                               0%
                                                                                                                                    10    20    30    40    50

                                                                                                                                            THAA MCL (ufl/L)
                                                                                                                                                                  60
                                                                           90M

                                                                           00)1
•pwovil at lydwra kwUIUng Met) IfMlmtnt Itof (kiclxlM 2O% ovw-dMlgn tactoi)
• oumuWIm p*fcanl a« tyttwra ri>l« to fmrt MCL « Mcfi IrMlrranl llw (hdudM 20% «w - dMl^i tcdor)
    n omonhrilan • Met) Iculmanl llw of M syMxra. >IOM mMlIng MCI tndtmMnol mMthg MCI.
        10% of pcpuMion upoad * IIM cudamw
' MlimM* lu eon p*c«nlil*

-------
                                                                                                                                                                          Onrt.
    MODEL OUIPUT(«iilm«*»Mta*IQ): 8W1RWS ALTERNATIVE UBMECnON
                                                                           Figure  10
HDL
J1TOM*
100
100
100
100
100
MP
76
76
76
76
76
n
so
so
60
SO
so
•
»
»
M
»
2S
*
Til.
Onto
1
2
9
4
6
•
1
2
3
4
6
•
1
2
3
4
J
•
1
2
3
4
i
•
• otSy*.
En**'
01
21
18
O
0
0
4»
26
23
3
0
0
33
23
34
e
i
a
i;
10
30
27
13
•
OMWMMB
«<*•*•.
.. .M
1.9
1.0
0.6
0.0
0.0
*•
13
00
0.7
0.0
0.0
»•
1.9
0.0
0.6
0.4
0.4
0.4
H to
am
1S3.3
1381
190.7
130.7
190,7
WM
1633
1921
12810
127.6
127.6
«7.i
1S33
1200
110.0
11D4
100.0
WM
1693
1240
107.7
006
04.0
*1.7
»AE
404
497
440
44.0
44.0
«M
49.4
42.6
41.7
410
41.0
«J>
40.4
40.7
97.6
96.4
96.9
M*
49.4
40.1
34.6
204
27.2
ma
EflMfMfllD*
•Hvmivvv





••MM





«0«M





474.906





•10.146
tilS'faL^





. M«





QiU





ao4





aoo»
                                                                                                                                     20  30  40  SO  00   70   80  90  100
                                                                                                                                              TTHM MCL (ug/L)
ft
                  hg MohliMimM ter (InoUtn 20%ow-dMl(n I«lo0
              41 Mch IfMtoMnl U*c at •* (yitani*: *KM« mMthg MCX "Hi ttim* not tiMttng MCt
      la OOti pMMndto

-------
                                                                                                                                                                                     'Oiftt
                                                                                 Figure 11
    MODEL OUTPUT (urlM K/O KjUnkifl}.  SWTR MOD AtTEHNATTME OmNFECTION
    TratfiMnl Code:
    I -

    9 - •*•*>«• pm-cMot * modNy *ktm
    4 - pra-ohkx •» iftMiKtaM + CMC
 fTTHM.)
too
too
too
wo

 7S
 75
 76
 1C

 60
 SO
 SO
 to

 25
 25
 2S
         T*
          EndkV
17
10
14
SO.
                          
-------
                                                                                                                                                                       DnH.3O-Jtn-aa
                                                                        Figure  12
MODEL aUIPUT(«urtB»«*>**tntoa): ENHMKXO8WmW«LietlMTIMEO«l»FECnaN
MCL'
(I1HMJ)
100
100
too
too
100
WP
7S
7$
76
74
76
1*
60
60
60
60
60
00
26
26
26
2S
86
• 86
Til.
Onto
1
2
3
4
6
•
1
2
3
4
6
•
1
2
3
4
S
' •
1
2
3
4
s
• •
%***•.
KatddhMa'
UMBUI
•i
20
19
3
0
• «
40
24
. 32
4
1
«
32
22
26
17
4
o
14
14
17
27
26
a
OttHuMH»
%
-------
                                                                             Figure  13
    MOOEL OUTPUT (iuitaM«/OMlUnktg): EMMNC^ BOTH MW ALTERNATIVE OBUFECI1ON
    TmdfMT* Opd»:
    1 - IK* raqpMig tirthv Inrtmnl mod
    2 - •*«*»*• pm-cMai»MUan
    a - dMutepn-oMof + modNy akim
    4 - fm-chkx » *im.
 0.7
0.0
1.3
0.0
07
a4
HMte
06*
1503
1380
134A
1M»
15O3
136.0
1230
lit*
1SO.3
13Q4
1042
M.i
1SO3
120.0
aoo
SB.»
MLE
60.3
44.0
43.3
410
603
43.7
30.7
M.I
60.3
41.0
336
37.1
503
41 1
28*
I8.S

htaothm



ata



mi



•12



aw
EpJ*MBk)'
(00«) » Vm. to Oulbr«k



MMT



MM*



***»



MMy
                                                                                                                                  100%

                                                                                                                                  00%

                                                                                                                                  80%

                                                                                                                                  70%

                                                                                                                                  00%
                                                                                                                                  50%

                                                                                                                                  4O%

                                                                                                                                  30%

                                                                                                                                  20%

                                                                                                                                  10%

                                                                                                                                   0%
                                                                           Til #4
                                                                                                                                       2030405000  70  8000  100
                                                                                                                                               TTHM MCL (ug/L)
                                  I-
OOM
ooa
0«B


 00>
                                                                         {::
                                                                         f 0*1
                                                                                                                             TTHU MCL (1^.1
'pwoml a tyttam Iratallhg Mch lt«Mlm«nl llw (IncUlw 20% cvw-dMlp) laclor)
'ounuMlw p«c«nl ol lyttemi Alt) to IDM! MCL * MOh UMlnwnl tlw (InoludM 20% avw-dwlp) Itctoc)
•iDMn oanMntraUan « Mch ItMtiranl Uw al m lyilwra; tiOM mMlhg MCL Btd (KM*nd mMlhg MCL
4 (MurnM 10% al paiwMlan MpcMd 41 lim cuptomw
'••Urn** lor 00h p«ctnlll*

-------
                                  Figure
          Influent Glardia vs. Total Log Reduction (Winter) From LeChevailier
10,000
 1,000
8

I
,0,
£

O
  100
   10
                                  Total Log Reduction

-------
                     SWTR Scenario: With Alternate Disinfectants
                     TTHM MCLs 100 (ug/l)     #. of Plants = 39
10,000
 1,000
§
I

&
3

3
  too
   10
                              Total Log Reduction Achieved
                                                                            10

-------
                                     Figure 16
   10,000
    1,000
3

I
I
3
100
      10
                        SWTR Scenario: With Alternate Disinfectants



                    TTHM MCL s 25 (ug/l)     #. of Plants Treating = 83
                                           7
                                 Total Log Reduction Achieved
                                                                               -10

-------
                       SWTR Scenario: With Alternate Disinfectants
                             TTHM MCL = 100 ug/l to 75 ug/l
o
£
§
I
100%



 90%



 80%



 70%



 60%



 50%
   40% -i-
   30% -i-
   -20% -:-
   10% 1-
    0% -*-

         1
                   10            100           1,000          10.000

                  Increase in Annual Glardia Infections Per 1 Million People
100.000

-------
                                 figure  J.a
                  SWTR Scenario: With Alternate Disinfectants
                        TTHM MCL = 100 ug/l to 25 ug/l
0%
                  10            100           1,000         10,000
                InerMtt in Annual Glardta Infections Pw 1 Million People
100,000

-------
100% —

90% -
80% -
70% -
(ft
1 60% -
I
Percentage of
Si 5
i i
30% -
20% -
10% -
0% -
TTHM MCL = 100 ug/l to 75 ug/l
••






85%






-






5- 7%

2% 1V

             1-10             10-100           100-1,000         1,000-10,000      10.000-100,000
                       Increase in Annual Giardte Infections Per 1 Million People

-------
                                   Figure 20
100%
90%
         1-10
                        SWTR Scenarios: With Alternate Disinfectants
                               TTHM MCL = 100 ug/l to 25 ug/l
                       10-100
                                     100-1,000        1,000-10.000      10.000-100,000
                        in Annual Glardta Infections Per 1 MUlton Peopto

-------
                     SWTR Scenario: With Alternate Disinfectants

                         From TTHM MCL = 100 ug/l to 75 ug/l
                         Change in Risks Per 1 Million People
    10
I   '
   o.ot
                    10            100           1.000
                            Increase in Annual Giardia Infections
10,000
100.000

-------
                                  Figure 22
                    SWTR Scenario: With Alternate Disinfectants

                        From TTHM MCL = 100 ug/l to 25 ug/l
                        Change in Risks Per 1 Million People
I

3
10

1
1


0.1]












f




<







1














1


1





"
























































jT















.01*
1



























=a

















<




























«




















b




















g>
























• •> I

















d


B


















i_
O
<


















n









1














1



















4




















• «








0 	










































/%>
7
•





—






<





^











...
R?
g

<



















i
«






















n




































































































10 100 1,000 10.000 100,
                           inerMM in Annual Glardia Intacttons

-------
                       SWTR Scenario: With Alternate Disinfectants
                         THAA MCL s 60 (ug/l)      #. of Plants = 29
   10,000
   1,000
8
u,
3
9

                                 Total Log Reduction Achieved
                                                                               10

-------
                                   Figure 24
   10,000
   1,000
8
,0,
a
(9
     100
      10
                       SWTR Scenario: With Alternate Disinfectants


                    THAA MCL = 10 (ug/l)    #. of Plants Treating = 80
                                                                             10
                                Total Log Reduction Achtewd

-------
                   SWTR Scenario: With Alternate Disinfectants
                          THAA MCL a 60 ug/l to 50 ug/l
100%



 90%


 80%


 70%


 60%


 50%


 40%


 30%


 20%


 10%


 0%
                   10            100          1,000          10,000
                 Increase in Annual Giardta Infections Per 1 Million People
100,000

-------
                                 figure 26
                  SWTR Scenario: With Alternate Disinfectants
                         THAA MCL = 60 ug/l to 10 ug/l
o%  -t
    1
 10            100           1,000         10,000
Incniie in Annual Gfardta Infections Per 1 Million People
100,000

-------
1 WW /O

90% -
80% -
70% -
| 60% -
CO
1 50% -
Q|
S
I 40%;-
30% -
20% -
10% -

0% -
95%

^M


_





















SWTR Scenarios: With Alternate Disinfectants
THAA MCL = 60 ug/l to 50 ug/l









40/ ««/ 3%
I • ° • i . 1% - i ,J^*m o* -
1-10            10-100           100-1,000         1.000-10,000      10,000-100.000
         IncreaM in Annual Glardia Infection* Per 1 Million People

-------
                                    ngure
100%  —



 90%  --



 80%  --



 70%  --



 60%  «



50%



40% —
            1-10
SWTR Scenario: With Alternate Disinfectants
       THAA MCL = 60 ug/l to 10 ug/l

20% -

10% -

0% -

M»




•










H


12%'









24%



•


• : ; : '• '. • : • : • : • : • : • : | : : : • ; - ; : ; •
-_:'•_: ''.I. •_:'•' •'.}:'•:'•:'•:':•;'•'.
: : : ; :':•':•: ; : - : : :••-"'-•'••'••























4%
— I — fcPife1*^^"-1* — i
                   Inc
   10 '^00          100-1.000        1.000-10,000      10,000-100.000
    in Annual Glardia Infections Pw 1 Million Peopto

-------
SWTR Scenario: With Alternate Disinfectants

    From THAA MCL = 60 ug/l to 50 ug/l
    Change in Risks Per 1 Million People
10-
1
1 ; -
n Annual Cancer Incit
>
A
Decrease 1
c
'
































<



































































































0.01 T
1








































i
t
i




































































































































































i





















>
























































































4











































-










































4






















1







































































































































10 100 1,000 10,000 100,
      Increase in Annual Glardla Infections

-------
                                  Figure  30
                  SWTR Scenario: With Alternate Disinfectants


                         THAA MCL = 60 ug/l to 10 ug/l
                      Chang* in Risks Per 1 Million People
 10
                                                  1
                        mm
                        1CI..I
0.1
0.01
                 10
100          1,000

 in Annual Outfit Infection*
10,000
100,000

-------
               STATUS REPORT ON ANALYTICAL METHODS
             TO SUPPORT THE DISINFECTANT/DISINFECTION
                      BY-PRODUCTS REGULATION

                           August 1992
Background
     The United States Environmental Protection Agency (EPA) is
developing national regulations to cc..trol disinfectants and
disinfection by-products  (D/DBPs) in public drinking water
supplies.  Twelve D/DBPs  have been identified for possible
regulation under this rule12-3, based, on available occurrence,
exposure, and health effects data.  EPA intends to set maximum
contaminant levels (MCLs) for these D/DBPs, and analytical
methods will be specified for use in demonstrating compliance
with each MCL.  This document provides a summary of the
analytical methods that EPA intends to propose as compliance
monitoring methods.  A discussion of surrogate measurements that
are being considered for  inclusion in the regulation is also
provided.

     There are several technical issues that EPA is trying  to
resolve before the methods discussed in this document are
proposed/promulgated.  The public is encouraged to contact  EPA,"
if they have information  that will help EPA in this process.
Information should be addressed to:

               Patricia Snyder Fair,  Chair
               D/DBP Analytical Methods Task Force
               U.S. Environmental Protection Agency
               Office of  Ground Water & Drinking water
               26 W. Martin Luther King Drive
               Cincinnati, OH 45268


Disinfectants:

     Chlorine and Chloraaines.  The measurement of chlorine
residuals will be required as part of the D/DBP Rule in order to
demonstrate compliance with MCLs for chlorine and monochloramine:
The utility will have the option of either measuring free or
total chlorine residuals  for the chlorine MCL.  Compliance  with
the monochloramine MCL will be based on a total chlorine residual
measurement,' because monochloramine cannot be practically
measured by itself.

     There are many analytical methods available for measuring
chlorine residuals, and each has certain limitations.  All
methods are subject to interferences, so the accuracy and

-------
precision  of  the  analysis  will vary with matrix.  LOW  level  free
chlorine measurements  are  subject to  interference errors  fron
chloramines and other  oxidants.

      Several  methods for measuring chlorine residuals were
promulgated with  the Surface Water Treatment Rule (SWTR)*: l)
amperometric  titration; 2) DPD ferrous titrimetric method; 3) DPD
colorimecric  method; and 4) the  leuco crystal violet (LCV)
method.  These were referenced to the 16th Edition of Standard
Methods for the Examination of Water and Wastevater3.   EPA used
these methods as  the starting point for determining methods that
are applicable to the  D/DBP Rule.  Methods listed in the  17th
Edition of Standard Methods4  were also considered.   Table l
summarizes the methods most likely to be proposed with the D/DBP
Rule.  The methods required to demonstrate compliance with the
SWTR will also be updated  to be  consistent with the D/DBP Rule.

     EPA intends  to propose thre« of the four chlorine residual
methods from the  SWTR, when the  D/DBP Rule is proposed.   The LCV
method will not be proposed.   It was not included in the  17th
Edition of standard Methods,  due to its relative difficulty and
lack of comparative advantages.  EPA thinks it is unlikely that
this method is being used, so dropping it should have  little
effect on the regulated community.

     EPA is considering approval of two additional methods for
compliance monitoring: 1)   the syringaldazine (FACTS) method for
free chlorine and 2) the iodometric electrode method for  total
chlorine.  Both methods are included in the 17th Edition  of
Standard Methods.

     Method 4500-C1 E  for  measuring low levels of total chlorine
residual is also available in the 17th Edition of Standard
Methods.  Although it  is not useful for measuring total chlorine
residuals at concentrations near the MCL, it is useful for
determining compliance with the  SWTR requirements for a
detectable residual in the distribution system.  Therefore, EPA
is considering approval of this  method when it updates methods
for the SWTR.   Data obtained using this method can also be used
to demonstrate compliance  with the total chlorine residual MCL in
the D/DBP Rule.

     The 18th Edition of Standard Methods7 will  be available  for
purchase in August, 1992,   so the D/DBP Rule will cite  it  for
descriptions of the chlorine residuals methods.  The method
numbers did not change from the  17th to-the.18th edition, so the
numbers given in Table 1 can be  used with either edition.  Since
EPA is required to review  its regulations every 3 years,  EPA
anticipates incorporating  the latest versions of the methods into
the D/DBP Rule each time it is reviewed.

-------
     Historically, EPA has required all analyses that are done to
demonstrate compliance with an MCL, be performed by a certified
laboratory.  Because chlorine residuals are not stable, these
samples must be analyzed  immediately efri cannot be transported to
an off-site laboratory.   Utility per-ainnel have been performing
chlorine residual analyses either in the field or in the
treatment plant.  EPA feels these measurements should continue to
be made at the water system by any person acceptable to the
State.  It is not EPA's intent to require the certification of
each water system for chlorine residual measurements.

     EPA will continue to allow the use of DPD colorimetric test
kits for field measurements of chlorine residuals,  if they are
approved by the State.

     Chlorine Dioxide.  The SWTR promulgated two methods for
measuring chlorine dioxide residuals:  1) amperometric titration
and 2) the DPD method.  The 16th Edition of Standard Methods was
cited as the reference.  As discussed above for chlorine residual
measurements,  the latest versions of methods for measuring
chlorine dioxide residuals (see Table 1)  will be cited in the
D/DBP Rule.

     The 17th Edition of Standard Methods proposes a second
amperometric titration method (4500-C102  E)  for  measuring
chlorine dioxide residuals.  EPA is considering adding it as an
acceptable compliance monitoring method.

     The methods for chlorine dioxide residuals are indirect
methods, because the concentrations are determined by difference.
Amperometric titration methods are preferred over the DPD method,
if they are used by highly trained personnel.

     It is EPA's intent to continue to allow the measurement of
chlorine dioxide residuals at the water system by any person
acceptable to the State.

     Due to the limitations of the current methodology, EPA is
seeking information on new methodology that may be applicable for
compliance monitoring.  New methods must provide demonstrated
advantages over the current methods and have the potential for
being distributed in a standard format to interested public in
the timeframe of the D/DBP regulation.  New methods must be
useable in the field or by utility personnel.


Disinfection By-products:

     Trihaloaethanes.  There are two methods currently approved
for THM compliance monitoring:  EPA Methods 501.1 and 501.2.
These packed column, gas  chromatography  (GC) methods were

-------
promulgated with  the  1979 THM Rule*.   EPA intends to propose two
capillary column  GC methods  by the end of  1992:  EPA Methods
502.2" and 524.2iJ,  After a  45  day comment period,  the  methods
will  be promulgated,  and they can be used  for THM compliance
monitoring effective  30 days after promulgation.  EPA Methods
502.2 and 524.2 are in general use in many laboratories, because
they  are used to  measure the concentrations of volatile organic
compounds (VOCs)  in drinking water.

      EPA Methods  502.1 and 524.1* are not being  proposed in 1992
and they are not  being considered for the  D/DBP Rule, because
they  use packed column chromatography.  The GC technology  has
progressed to the point that packed columns are becoming
obsolete.  EPA is considering eliminating  packed column GC
methods (EPA Methods  501.1 and 501.2) from the list of approved
THM compliance monitoring methods when the D/DBP Rule is
proposed.  This would not be implemented until the monitoring
requirements of the D/DBP Rule become effective.

      EPA Method 551"  is the only new method that is likely to be
addeu for THM compliance monitoring when the D/DBP Rule is
promulgated.  It  involves adjusting the ionic strength of  the
sample,  extracting the analytes into methyl-tertiary-butyl ether
(MTBE),  and analyzing the extract by capillary column GC with
electron capture detection (ECD).  This method can also be used
to measure the concentrations of haloacetonitriles, chloropicrin,
1,1-dichloropropanone, 1,1,1-trichloropropanone, and chloral
hydrate, if the appropriate dechlorinating agents are used.

      Several laboratories around the country have modified EPA
Method 551 by using pentane, instead of MTBE, as the extraction
solvent.  When this is done, chloral hydrate (CH) is not included
in the *--lysis, because CH is too polar to be extracted by
pentane.   ?A Method  551 permits the analyst to modify GC
columns,  .C conditions, detectors, extraction techniques,
concentration techniques, internal standard or surrogate
compounds, as long as the analyst demonstrates the modified
method still meets the performance criteria established in the
nethod.   Therefor«, if Method 551 is approved for THM compliance
  •mitoring,  approval to use pentane will not be considered
   -=>ssary as lonq is the method performance criteria are met.
. ...»  will eliminate some of the need for approval of alternative
test  procedures (ATPs).

      Many commercial  sources of MTBE are contaminated with
chloroform,  necessitating cleanup prior to use in THM analyses.
MTBE  can be purified  by distillation using appropriate safety
precautions, but  it should not be stored for long periods  of
time, in order to prevent the formation of peroxides.  At  least
some  lots of OmniSolv (EMScience) MTBE labeled "Suitable for
Spectr©photometry, Liquid Chromatography,  Gas Chromatography,

-------
Residue Analysis, Assay by GC: 99.9% pure" have not contained
measurable  levels of chloroform or chlorinated solvents,
indicating  suitable purity MTBE can be produced commercially.
This should encourage  laboratories to make the need for high
purity MTBE known to their local solvent suppliers.           • '
Manufacturers may be willing to prepare and market special lots,
when they are aware of a market for their product.

     The EPA methods that are likely to be approved for
compliance monitoring of THMs are summarized in Table 2.  All of
these methods include the option to use ascorbic acid as a
dechlorinating agent.  This practice is under review, because
ascorbic acid may cause the loss of brominated THMs under some
conditions.  If laboratories are aware of alternatives to the HC1
and ascorbic acid required by EPA Method 524, they are encouraged
to share that information with EPA.  (Sodium thiosulfate and HC1
were originally included in 524,  but they cause interference
problems with some of the early eluting analytes included in the
method.  Thiosulfate and HC1 can still be used, if the sample
does not have to be analyzed for the early eluting compounds.)

Chloral Hydrate.  EPA Method 551" will be proposed as the
compliance monitoring method for chloral hydrate (CH), but
additional work must be done before the method is ready for
general use.  Chloral hydrate is subject to base-catalyzed
hydrolysis, and the current version of the method does not
provide a mechanism for preventing hydrolysis.  This method will
require the addition of a preservative, probably in the form of
acidification.  Work is underway to evaluate preservatives for
CH.  EPA solicits data demonstrating the stability of CH in
drinking water samples.  The data must indicate the
dechlorinating agent used, preservation procedure, sample pH,
storage conditions and holding time.

     The regular dechlorinating agent (NH4C1)  recommended  in
Method 551 cannot be used in some drinking water matrices,
because it interferes with the CH analysis.  Either ascorbic acid
or sodium sulfite can be used under those circumstances.  Since
ascorbic acid causes problems with THM analyses in some matrices,
EPA prefers the use of sodium sulfite.  As part of the effort to
evaluate preservation techniques for CH, EPA will look at how
THMs are affected.  It is anticipated that a technique can be
developed in which THMs and CH could be measured in the same
sample using EPA Method 551.  If successful, this would reduce
the monitoring costs associated with these analytes.

     Some laboratories have expressed concern about the safety of
using MTBE.  EPA recommends that ether extracts should be stored
in an explosion-proof refrigerator/freezer.

-------
     The availability of  standards for CH has been a problem
 laboratories.  However, dilute solutions of CH in acetone are -
 available  from NSI  Environmental Solutions, Inc.  (Research
 Triangle Park, NC)  ^Catalogue # 001179-01-01, 332.00]  This is a«
 EPA certified standard.   A neat standard is available from
 Supelco  (Bellefonte, PA)  [Catalogue # 4-8048].  Other commercial
 sources may become  available during the next year.

     Very  few laboratories are doing CH analyses.  EPA is not
 aware of any commercial laboratories, but this should change over
 the next few years  as utilities include it as a by-product of
 interest in their bench and pilot scale treatment studies.

     Chloral hydrate will be included in the EPA Performance
 Evaluation (PE)  Studies by the end of 1992.  This will provide
 laboratories an opportunity to evaluate how well the method is
 performing for them.  It will also provide EPA with an estimate .
 of how many laboratories are doing CH analyses and how well they
 are doing.

 Haloacetic Acids.   The EPA method that will be proposed for
 haloacetic acid (HAA) compliance monitoring will give the analyst
 several options.  The initial (standard)  version of Method 552"
 specified the following steps in the procedure: 1) extraction
with MTBE after sample pH adjustment to >11.S; 2) discard MTBE
 fraction and adjust pH of sample to <0.5; 3) extraction with
MTBE;  4) concentration and drying of the extract; 5) conversion
of the HAAs to their methyl esters using diazomethane; and 6)
 analysis by capillary column GC/ECD.

     A microextraction option was later added to EPA Method
 55212.   This option  eliminated the cleanup extraction, extract
 concentration, and  extract drying steps.   The 18th Edition of
Standard Methods will include a microextraction method for
measuring HAAs.   EPA is considering it as a compliance monitoring
method, because it  is equivalent to the option described in
       552.
     EPA is also developing a liquid/solid extraction technique .
using ion exchange resins and an acidic methane 1 derivatization
procedure13.  A written procedure should be available for public
distribution by the end of 1992, and it will be designated EPA
Method 552. l'°.

     All of the above procedures will be proposed as compliance
monitoring techniques; all will be covered under EPA Method 552.
Giving the analyst the flexibility to choose extraction technique
and derivatization method should make it easier to begin this
analysis in the laboratory.  The acidic methanol derivatization
procedure will also eliminate the concern expressed by some
states and other entities over the use of diazomethane due tc

-------
safety  issues.

     The  quantisation  procedures  specifed  in the current version
of  EPA  Method  552"  must  be  clarified prior to proposal of the
D/DBP Rule.  The  initial  procedure  (described-above  in the first
paragraph under HAAs)  does  not require that standards be
extracted,  and no correction  is made for analytes which are not
fully extracted from the  aqueous  sample  (e.g., dibromoacetic acid
[DBAA]).   The  microextraction option recommends the  preparation
of  aqueous standards that are carried through the same procedure
as  the  samples.  Using this option, the analytical results are
automatically  corrected  for less  than 100% extraction
efficiencies.  Either  calibration method will be acceptable for
compliance monitoring, if EPA only  sets MCLs for dichloroacetic
acid (DCAA) and trichloroacetic acid (TCAA), because both acids
are extracted  from  water  at > 90% efficiency.  However, EPA is
also considering a  total  HAA  (THAA) MCL.  Samples from utilities
with high levels of bromide ion in  their source water may contain
high levels of the  brominated and mixed broraochloro-acetic acids.
Since these  acids  are not extracted from water as efficiently as
the chlorinated acids, use of the initial HAA analytical
procedure (described above  in the first paragraph under HAAs)
would give a lower  THAA result than the microextraction
procedure.  For this reason, future versions of Method 552 will
specify that aqueous standards be used for calibration purposes.
Analyses  performed  to  meet compliance monitoring requirements  for
a THAA MCL must use aqueous standards for calibration purposes.

     Five of the HAAs  (monochloroacetic acid [MCAA], DCAA, TCAA,
monobromoacetic acid [MBAA], and  DBAA)  have been included in EPA
PE studies, since 'study WS026 (spring,  1990).  A calibration
procedure for the HAAs was not specified with the PE studies,  so
laboratories reported  data from both procedures.  Since the two
procedures are not  equivalent for several of the HAAs, EPA will
specify that HAA PE samples be analyzed using aqueous standards
in future PE studies.

     EPA  Method 552 specifies the HAA standards be prepared in
MTBE, but many laboratories are using methanol instead of MTBE.
'Recent work by Yuefeng Xie14 at the  University of Massachusetts,
indicates that over time  the acids  will undergo conversion to
their methyl ester  analogs when stored in methanol.  The
conversion rate varies with analyte and storage conditions.
Unless  the analyst  specifically checks for this conversion by
analyzing a non-derivatized standard, it is unlikely the problem
would be  detected.  The use of ester or mixed free acid and ester
standards will provide inaccurate results, due to differences  in
extraction efficiencies between acids and esters.  EPA has not
studied this problem,  but recommends that if laboratories must
continue  to use methanol, they monitor their methanol standards
for this  conversion, and  prepare  fresh standards when esters are

-------
detected.  As  an  alternative,  laboratories  should prepare  HAA
standards  in MTBE.

     There are  9  haloacetic.acids'(HAAs; that could potentially
be  included in  a  THAA MCL.   However'-~r.e EPA Method 552 only
includes 6 of  the HAAs  (MCAA,  DCAA, TCAA, MBAA, DBAA, and
bromochloroacetic acid  [BCAA]).  Tribromoacetic acid  (TBAA)  is
not included,  because it  is  not reliably measured using the
current techniques.  The  remaining 2 HAAs (bromodichloroacetic
acid [BDCAA] and  dibromochloroacetic acid [DBCAA]) have not  been
tested with Method  552, because standards are not commercially
available.  It  is unlikely that TBAA, BDCAA, or DBCAA will be
included in Method  552  in the  timeframe of this regulation.

     EPA is aware of three suppliers that are selling or plan to
sell HAA standards:
     1)  Supelco, Inc.  (Bellefonte, PA) is distributing a
     calibration  standard that contains MCAA, DCAA, TCAA,  MBAA,
     DBAA, and  BCAA (catalogue #4-8047).  They.will  also  market
     a BCAA standard in solution form as a custom chemical.
     2)  Absolute Standards, Inc.  (New Haven, CT) is  selling a
     HAA QC sample, containing MCAA, DCAA, TCAA, MBAA, DBAA, 2,4-
    . dichlorophenol (24DCPh),  and 2,4,6-trichlorophenol
     (246TCPh).   A  standard  containing the methyl derivatives of
     the same compounds is also available, thus, providing
     laboratories a mechanism  for checking the derivatization
     efficiency of  their  method.  BCAA will soon be added  to the
     free acid  standard mixture, and should be available for
     release in August, 1992 (catalogue # 30054).
     3)  ULTRA  Scientific (North Kingstown, RI) plans to release
     a standard containing MCAA, DCAA, TCAA, MBAA, DBAA, BCAA,
     24DCPh, and  246TCPh  (catalogue # PHM-552A).  The standard  is
     tentatively  scheduled for release in early August, 1992,
     assuming successful  results from a 60 day time storage
     study.

     EPA believes there will be adequate laboratory capability
available by the  time compliance monitoring for HAAs  is required.
Sixteen laboratories participated in the WS029 PE study by
analyzing the HAA sample.  This number is expected to increase
significantly over  the  next  few years, as laboratories expand
their analyses  to address the  new regulations.

     chlorite,  (Chlorate) and  Broaate.  An  ion chromatography
method will be  proposed as the compliance monitoring  method  for
these anions.   EPA  does not  intend to propose an MCL  for the
chlorate ion (CIO,*), but  a description of the analytical method
is presented here since it can be determined with the chlorite
ion (C102') .  The  current  version of EPA Method 300.0 Part  B15 can
be used for measuring C10:~ and CIO,' in drinking waters.   In  some
cases, a weaker carbonate eluent may be required16, because Clo:

-------
elutes in the -void volume  in high  ionic strength samples.  A
change in the eluent strength  is permitted as part of Method
300.0 Part B, as  long as the analyst demonstrates that the
quality control requirements outlined in the method are met (see
Sections 10.1.1 and 11.1 of the method).

     Chlorite ion is unstable  in many waters1", so a preservation
technique will be required if  the  samples cannot be analyzed
within 15 minutes of collection.   Recent studies'*"  indicate
ethylenediamine (EDA) is a suitable preservative for C10: ,  and it
does not adversely affect analysis of the other anions at the
concentrations typically found in  drinking water.  Therefore, EPA
is considering requiring the addition of EDA to samples analyzed
for C102~ compliance monitoring.

     Measurement of CIO,' in samples containing a free chlorine
residual will also require the use of EDA.  Free chlorine reacts
with CIO," to form CIO,  and chloride.

     Samples containing a chlorine dioxide residual must be
sparged with an inert gas  (e.g., He or Ar) at the time of
collection to eliminate the chlorine dioxide.  Otherwise, C10:
will continue to form C10:' and CIO,' in the samples.   Prior to
sparging, the samples must be  protected from light to prevent
photodecomposition of the chlorine dioxide to form C102' and  CIO,'
19
 •

     EPA is aware of other techniques (e.g., flow injection
analysis [FIA])  that are being used to measure chlorite and
chlorate.  FIA and/or any other applicable procedure will be
considered as a potential compliance monitoring method, if it is
demonstrated to be as accurate and precise as ion chromatography.
The technique would have to be available in a standard format for
public distribution.

     Analyses for bromate ion  in drinking water samples will
require a modification to the  EPA  ion chromatography method, due
to the interference from chloride  present in these matrices,  one
approach is to pretreat the sample with a silver media in order
to remove the chloride20.  When this technique is used, traces of
silver are deposited on the 1C analytical column, necessitating
the removal of silver before using the column for bromide
analyses.  Another alternative is  to change the eluent.  Recent
studies16-11 demonstrate that substituting a borate eluent  for  the
carbonate eluent specified in  the  method not only provides the
resolution required to quantitate  bromate ion in drinking water,
but also gives a more stable baseline.  All the DBF anions can  be
measured using the same chromat©graphic conditions with the
borate eluent.  Sections 10.1.1 and 11.1 of EPA Method 300.0 Part
B permit the use of the borate eluent as long as the analyst
demonstrates that the quality  control requirements outlined in

-------
the method are met.

     EPA  is concerned about the ability to reliably measure
bromate at the levels necessary for.tttis regulation.  Studies'are
underway  in several  laboratories, including EPA, to examine ways
to lower  the detection  limit for this anion.  Concentration
techniques are the most promising avenues of research.  EPA
requests  information on any techniques that are demonstrated to
be effective in drinking water matrices.

     Chlorite, chlorate and bromate ions are now included in EPA
PE studies.  Thirteen laboratories participated in study WS029.


Surrogate Measurements!

     Simulated Distribution System (80S) Test.  EPA may consider
the use of a Simulated Distribution System (SOS) test as a
surrogate measurement of DBP concentrations in the distribution
system.   EPA will recommend a modified version of the procedure
described in Method 5710 E in the 17th edition of Standard
Methods (reference will be updated to the 18th edition when
proposed).  The method is written for use in determining THM
concentrations, but it can be extended to other DBFs as long as
appropriate dechlorinating agents are used after the storage
period.  Results of the test are not valid, if there is no
detectable chlorine residual at the end of the storage period.
If the test is used to determine a "worst case11 for consumer
exposure,  then the sample storage period should be representative
of the maximum distribution system temperature, pH, chlorine
concentration and the longest detention time in the distribution
system.  EPA will accept data comparing results from the SOS test
to actual distribution system samples.  The data must include a
full description of the SOS test conditions, including
dechlorination procedures and analytical methods used to
quantitate the DBFs.  Information concerning the data from
distribution system samples should include temperature, chlorine
residuals, pH, and approximate detention time of the water in the
system.

     Total organic Carbon.  There are 3 methods for measuring
organic carbon listed in the 17th Edition (18th Edition) of
standard Methods, and 2 of them are applicable to requirements of
the D/DBP rule.  The persulfate-ultraviolet oxidation method
(5310 C) and the wet-oxidation method (5310 0) provide the
sensitivity necessary for low level organic carbon measurements,
so they will be recommended if organic carbon is proposed in the
D/OBP rule as a surrogate measurement for DBP precursors.
Depending upon the sample pretreatment, several different organic
carbon fractions are measured by these methods.  If the sample is
filtered through a 0.45pm pore size filter before analysis,
Dissolved Organic Carbon (DOC)  is measured.  Purging the sample

                                10

-------
prior to analysis results  in the measurement of Nonpurgeable
Organic Carbon  (NPOC).  Both of these parameters have been used
as indicators of raw^water quality.  EPA will accept information
concerning the  use of DOC or NPOC as a surrogate for DBP
precursors in source water.

     Total Organic Halida.  Method 5320 B in the 17th Edition
(18th Edition)  of Standard Methods will be recommended for
determining Total Organic Halide (TOX).  The sample must be
dechlorinated and acidified at the time of collection.  Sodium
sulfite crystals or a FRESHLY prepared sulfite solution should be
used for dechlorinatipn.  Following dechlorination, the sample is
acidified to a  pH < 2 using nitric acid in order to preserve the
sample.  If the bottles must be shipped to the sampling site with
the reagents already present in the bottles, then sulfuric acid
should be substituted for nitric acid.  Department of
Transportation  (DOT) regulations must be followed when shipping
bottles containing sulfite and sulfuric acid.

     UV Absorbance.  There is not a standardized method for this
parameter.  The technique originally used to establish a
relationship between raw water UV absorbance and THM formation
involved filtering the sample through a 0.45jim pore size filter
and then measuring UV absorbance at 254 nm.  The filter must be
prewashed to remove water-soluble organics.  EPA will accept
information concerning the applicability of alternative
procedures for  defining this measurement.

     EPA is evaluating its data to determine the relationship
between TOC and UV absorbance.  It may be necessary to measure
both parameters in order to better characterize raw water
quality.

     EPA will accept data demonstrating interferences to the UV
measurement at  254 nm that would prevent its use in specific raw
waters.
      L1 Consents t
     EPA knows the laboratory capacity for doing DBP compliance
monitoring samples is not yet in place.  However, EPA believes
the capacity can be developed by the time monitoring requirements
take effect.  PE samples will be available for all the DBPs
during 1992, so laboratories can take advantage of those
opportunities to demonstrate competency.  Part of the laboratory
certification process will involve successful performance in PE
studies.  Laboratories can request PE samples through their state
Certification Officer and participation is free.

     EPA does not have interlaboratory method performance data
for the non-THM DBP methods.  EPA anticipates that good

                                11

-------
laboratories will be able to achieve quantitative results on
analyses of PE samples that are within at least ± 40 % of the
sample's true value for the non-THM DBFs.  However, this has yet
to be demonstrated.  Some performance data will be generated
using the results from PE studies.  EPA will also examine
alternatives ways to obtain the data, such as round robin, studies
among several laboratories currently analyzing drinking water
samples for DBPs.

     The most recent PE studies that included the THMs indicated
that the majority of the laboratories participating in the
studies were able to achieve results within ± 20 % of the THM
sample's true value.  Therefore, EPA will continue to require
that level of performance for certification for THM analyses.


References:       .

1.   USEPA.  Status Report on Development of D/DBP Regulations.
     Office of Ground Water and Drinking Water. Washington, D.c.
     1991.

2.   USEPA.  Occurrence Document in Support of the Development of
     the D/DBP Regulations. USEPA Publications. Washington, D.C.
     in press. 1992.

3.   USEPA. Status Report on Development of MCLGs for
     Disinfectants and Disinfection By-products. Office of
     Science and Technology. Washington, D.C., 1992.

4.   USEPA.  Drinking .Water; National Primary Drinking Water
     Regulations; Filtration,  Disinfection; Turbidity,  Giardia
   -  lamblia,  Viruses, Legionella, and Heterotrophic Bacteria;
     Final Rule. Fed.  Rea.. 54:124:27486 (June 29, 1989).

5.   Standard Method's for the Examination of Mater and
     Wastewater, 16th Edition,  American Public Health.
     Association, American Water Works Association, and Water
     Pollution Control Federation, 1985.

6.   Standard Methods for the Examination of Water and
     Wastewater, 17th Edition,  American Public Health
     Association, American Water Works Association, and Water
     Pollution Control Federation, 1989.

7.   Standard Methods for the Examination of Water and
     Wastewater, 18th Edition,  American Public Health
     Association, American Water Works Association, and Water
     Pollution Control Federation, 1992.

8.  ' 40 CFR 141, Subpart C, Appendix C.


                                12

-------
9.    USEPA. Methods for the Determination of Organic Compounds in
     Drinking Water, EPA/600/4-83/039,  PB91-231480,  National
     Technical Information Service (NTIS),  December 1988 (revised
     July 1991).                    r  *

10.  USEPA. Methods for the Determination of Organic Compounds in
     Drinking Mater - Supplement II,  Environmental Monitoring
     Systems Laboratory, Cincinnati,  OH,  August 1992.

11.  USEPA. Methods for the Determination of Organic Compounds in
     Drinking Water - Supplement I, EPA/600/4-90-020,  PB91-
     146027, NTIS, July 1990.

12.  Barth, R.C.  & Fair, P.S.  "Disinfection By-Prodcuts:
     Analysis of Haloacetic Acids and Chlorophenols;
     Microextraction Procedure vs. USEPA Method 552." AWWA Water
     Quality Technology Conference (WQTC) Proceedings, November
     1990.                                                    •

13.  Hodgeson, J.W., Collins, J., & Becker, D.  "Advanced
     Techniques for the Measurement of Acidic Herbicides and
     Disinfection Byproducts in Aqueous Samples."  Proceedings of
     the 14th Annual EPA Conference on Analysis of Pollutants in
     the Environment.  May 1991.

14.  Xie, Yuefeng, University of Massachusetts, Amherst, MA,
     personal communication, May, 1992.

15.  USEPA. The Determination of Inorganic Anions in Water by Ion
     Chromatography Method 300.0.  Environmental Monitoring
     Systems Laboratory, Cincinnati,  OH,  August 1991.

16.  Hautman, p.p. & Bolyard, M. "Analysis of Oxyhalide
     Disinfection By-Products and Other Anions of Interest in
     Drinking Water by Ion Chromatography/" International Ion
     Chromatography Symposium Proceedings.  October 1991.

17.  Pfaff, J.D.  & Brockhoff, C.A. "Determination of Inorganic
     Disinfection By-Products by Ion Chromatography," Jour.
     82:4:195  (April 1990).

13.  Hautman, D.P. & Bolyard, M. "Analysis of Inorganic
     Disinfection By-Products Using Ion Chromatography," AWWA
     WQTC Proceedings. November 1991.

19.  Zika, R.G. ET AL. "Sunlight-Induced Photodecomposition of
     Chlorine Dioxide," Water Chlorination Chemistry:
     Environmental Impact and Health Effects. Vol. 5, Lewis
     Publ., Inc., Chelsea, Mich.  (1985).
                                13

-------
20.   Kuo,  C.  and Weinberg, H.S.  "Analysis of Inorganic
      ocprocH-u      IOn Chromatography," AWWA
     WQTC Proceedings.   November 1990.
                              14

-------
Table 1.  Methods Being Considered for Use in Compliance Monitoring of
Disinfectant Residuals Under the D/DBP Rule
Residual
Free Chlorine
Methodology
Amperoraetric Titration
DPD Ferrous Titrimetric
DPD Colorimetric
Syringaldazine (FACTS)
Std Meth*
4500-C1 D
4500-C1 F
4500-C1 G
4500-C1 H
Working
Range (mg/L)
> 0. 1
> 0.1
> 0.1
> 0. 1
Total Chlorine
Chlorine Dioxide
Araperoraetric Titration
Amperonetric Titration
DPD Ferrous Titrimetric
DPD Colorimetric
lodometric Electrode

Amperometric Titration
DPD Method
Amperometric Titration
  (proposed)
4500-C1 D
4500-C1 E
4500-C1 F
4500-C1 G
4500-C1 I

4500-ClOj  C
4500-ClOj  D
4500-C1O,  E
  0.1
  0.2
>
<
>
>
> 0.1
  0
  0
                                                                        1
                                                                        1
> 0. 1
> 0. 1
> 0.1
   Method Number Used in the 17th and 18th Editions of Standard Methods.
                                       15

-------
Table 2.  Methods Being Considered for Use  in Compliance Monitoring of Disinfection
By-Products Under the DBF Rule
                                                                   Working
  Residual              Methodology*          EPA Method (Ref)"*   Range4
Trihalomethanes     PfcT/GC/ElCD & PID             502.2  (9)         > 2.0
                    P4T/GC/MS                     524.2  (10)        > 0.5
                    LLE/GC/ECD                    551  (11)          > 0.5

Chloral Hydrate     LLE/GC/ECD                    551  (11)          > 0.5

Haloacetic Acids    LLE or SPE /GC/ECD            552  (104,11)       > 5.0

Chlorite & Chlorate 1C     ,                       300.0  Part B  (15) >10**

Broroate             1C                            300.0  Part B  (15) >1044

   P&T = purge and trap;  GC = g3S chromatography; E1CD = electrolytic conductivity
detector; PID = photoionization detector; MS = mass spectrometer;  LLE =  liquid/ liquid
extraction; BCD = electron capture detector; SPE = solid phase  extraction; 1C = ion
chromatography
  Reference for method is given in  ().
*   The concentrations listed in this table are estimates of the lowest levels
laboratories can routinely measure with confidence.  For methods  involving
multianalytes  (e.g., THM methods), the range is  given  for  the analyte with the
highest detection/quantitat ion level.
*f Based on experience in EPA's laboratory using a borate eluent.  These levels may
be optimistic  for many drinking water laboratories.
                                          16

-------