ENVIRONMENTAL PROTECTION AGENCY FOR FURTHER 40 CFR...

21576 Federal Register / Vol. 65, No. 78 / Friday, April 21, 2000 / Proposed Rules

ENVIRONMENTAL PROTECTIONAGENCY

40 CFR Parts 141 and 142

[FRL–6580–8]

RIN–2040–AC98

National Primary Drinking WaterRegulations; Radionuclides; Notice ofData Availability

AGENCY: Environmental ProtectionAgency.ACTION: Notice of data availability forproposed rules with request forcomments.

SUMMARY: The Environmental ProtectionAgency (EPA) proposed regulations tolimit the amount of radionuclides foundin drinking water on July 18, 1991. Ingeneral, the proposal revised currentNational Primary Drinking Waterregulations (NPDWR); a NPDWR wasproposed for uranium which isunregulated. Since that time, newinformation has become available whichthe Agency is considering in finalizingthese proposed regulations. In addition,the 1996 Amendments to the SafeDrinking Water Act (SDWA) containedprovisions which directly affect the1991 proposed rule.

This document presents additionalinformation relevant to the MaximumContaminant Level Goals (MCLGs), theMaximum Contaminant Levels (MCLs),and monitoring requirements containedin the 1991 proposal. EPA is seekingpublic review and comment on thesenew data. The Agency is also solicitingcomments on several implementationoptions that are being evaluated forinclusion in the final regulations.DATES: Written comments should bepostmarked or delivered by hand byJune 20, 2000.ADDRESSES: Send written comments tothe W–00–12 Radionuclides RuleComment Clerk, Water Docket (MC–4101), 1200 Pennsylvania Ave., NW,Washington, DC 20460 or by sendingelectronic mail (e-mail) to [email protected]. Hand deliveries shouldbe delivered to: EPA’s Drinking WaterDocket at 401 M Street, SW, EastBasement (Room EB 57), Washington,DC 20460. Please submit an original andthree copies of your comments andenclosures (including references). If youwish to hand-deliver your comments,please call (202) 260–3027 between 9:00a.m. and 4:00 p.m., Monday throughFriday, excluding Federal holidays, toobtain the room number for the Docket.Please see Supplementary Informationunder the heading ‘‘AdditionalInformation for Commenters’’ for

detailed filing instructions, includingelectronic submissions.

The record for the proposal has beenestablished under the docket name:National Primary Drinking WaterRegulations for Radionuclides (W–00–12). The record includes supportingdocumentation as well as printed, paperversions of electronic comments. Therecord is available for inspection from 9a.m. to 4 p.m., Monday through Friday,excluding Federal holidays at the WaterDocket, 401 M Street SW, East Basement(Room EB 57), Washington, DC 20460.For access to the Docket materials,please call (202) 260–3027 to schedulean appointment.FOR FURTHER INFORMATION CONTACT: Fortechnical inquiries, contact DavidHuber, Standards and Risk ManagementDivision, Office of Ground Water andDrinking Water, EPA (MC–4607), 401 MStreet SW, Washington, DC 20460;telephone (202) 260–9566. In addition,the Safe Drinking Water Hotline is openMonday through Friday, excludingFederal holidays, from 9:00 a.m. to 5:30p.m. Eastern Standard Time. The SafeDrinking Water Hotline, toll free 1–800–426–4791.SUPPLEMENTARY INFORMATION:

Regulated Entities

Entities potentially regulated by theRadionuclides Rule are public watersystems that are classified as eithercommunity water systems (CWSs) ornon-transient non-community watersystems (NTNCWSs). Regulatedcategories and entities include:

Category Examples of regulatedentities

Industry ................. Privately-owned CWSsand NTNCWSs.

State, Tribal, andLocal Govern-ments.

Publicly-owned CWSsand NTNCWSs.

This table lists the types of entities,currently known to EPA, that couldpotentially be regulated by theRadionuclides Rule. It is not intended tobe exhaustive, but rather provides aguide for readers regarding entitieslikely to be regulated by theRadionuclides Rule. Other types ofentities not listed in the table could alsobe regulated. To determine whetheryour facility is regulated by theRadionuclides Rule, you shouldcarefully examine the applicabilitycriteria in §§ 141.15 and 141.26 of title40 of the Code of Federal Regulations,and the definitions of Community Watersystems and Non-Transient, Non-Community water systems in § 141.2 Ifyou have questions regarding the

applicability of the Radionuclides Ruleto a particular entity, consult the personlisted in the preceding FOR FURTHERINFORMATION CONTACT section.

Additional Information for CommentersTo ensure that EPA can read,

understand and therefore properlyrespond to your comments, the Agencyrequests that commenters follow thefollowing format: type or printcomments in ink, and cite, wherepossible, the paragraph(s) in thisdocument to which each commentrefers. Please use a separate paragraphfor each issue discussed and limit yourcomments to the issues addressed intoday’s Document.

If you want EPA to acknowledgereceipt of your comments, enclose aself-addressed, stamped envelope. Nofacsimiles (faxes) will be accepted.Comments also may be submittedelectronically to [email protected]. Electroniccomments must be submitted as aWordPerfect 8.0 or ASCII file avoidingthe use of special characters and formsof encryption and must be transmittedby midnight June 20, 2000. Electroniccomments must be identified by thedocket name, number, or title of theFederal Register. Comments and dataalso will be accepted on disks inWordPerfect 8.0 or in ASCII file format.Electronic comments on this documentmay be filed online at many FederalDepository Libraries.

Abbreviations and Acronyms Used inThis Notice

OrganizationsAPHA—American Public Health AssociationASTM—American Society for Testing and

MaterialsAWWA—American Water Works AssociationICRP—International Commission on

Radiological ProtectionNBS—National Bureau of StandardsNSF—National Sanitation FoundationANPRM—Advanced Notices of Proposed

RulemakingATSDR—Agency for Toxic Substances and

Disease RegistryBNL—Brookhaven National LaboratoryCFR—Code of Federal RegulationsEML—Environmental Measurements

LaboratoryERAMS—Environmental Radiation Ambient

Monitoring SystemERD—Environmental Radiation DataERIC—Educational Resources Information

CenterFGR–13—Federal Guidance Report 13FR—Federal RegisterFRC—Federal Radiation CouncilNAS—National Academy of SciencesNCHS—National Center for Health StatisticsNESHAP—National Emissions Standards for

Hazardous Air PollutantsNIRS—National Inorganic and Radionuclide

Survey

VerDate 18<APR>2000 22:20 Apr 20, 2000 Jkt 190000 PO 00000 Frm 00002 Fmt 4701 Sfmt 4702 E:\FR\FM\21APP4.SGM pfrm01 PsN: 21APP4

21577Federal Register / Vol. 65, No. 78 / Friday, April 21, 2000 / Proposed Rules

NIST—National Institute of Standards andTechnology

NODA—Notice of Data AvailabilityNPDES—National Pollutant Discharge

Elimination SystemNPDWRs—National Primary Drinking Water

RegulationsNRC—National Research CouncilNRC—Nuclear Regulatory CommissionNTIS—National Technical Information

ServiceORNL—Oak Ridge National LaboratorySAB—Science Advisory BoardRADRISK—a computer code for radiation

risk estimationSWTR—Surface Water Treatment RuleT&C—Technologies and Cost documentUCMR—Unregulated Contaminant

Monitoring RuleUSDOE—United States Department of EnergyUSDW—underground source of drinking

waterUSEPA—United States Environmental

Protection AgencyUSGS—United States Geological SurveyUSSCEAR—United Nations Scientific

Committee on the Effects of AtomicRadiation

Units of MeasurementBq—BecquerelCi—CurieEDE/yr—effective dose equivalent per yearkBq—kiloBecquerelskBq/m 3—kiloBecquerels per cubic meterkg—kilogramkgpd—kilogram per dayMgkd—milligram per kilogram per dayL—literL/day—liter per daymg—milligrammg/L—milligram per litermg/kg—milligram per kilogrammg UN/L—milligram uranyl nitrate per litermg/kg/day—milligram per kilogram per daymg U/kg/day—milligram uranium per

kilogram per daymgd—million gallons per daymL—millilitermrem—milliremmrem/yr—millirem per yearSv—SievertµCi—microCurieµCi/kg—microCurie per kilogramµg or ug—microgramµg/g or ug/g—microgram per gramµg/L or ug/L—microgram per literµg uranium/L—microgram uranium per literµg uranium/kg/day—microgram uranium per

kilogram per dayµR/hr—micro Roentgen per hourµSv/cm—micro Sievert per centimeterNTU—Nephelometric Turbidity UnitpCi—picoCuriepCi/day—picoCurie per daypCi/g—picoCurie per grampCi/L—picoCurie per literpCi/µg—picoCurie per microgram

Other TermsACA—anticentromere antigenALP—alkaline phosphataseAS—alpha spectrometryBAT—best available treatment

BEIR—biological effects of ionizing radiationBMG—β2-microglobulinCWS—community water systemsDL—detection limitEDE—effective dose equivalentFSH—follicle stimulating hormoneGGT—gamma glutamyl transferaseGI—gastrointestinalIE—ion exchangeLDH—lactate dehydrogenaseLET—low energy transferLOAEL—lowest observed adverse effect levelLP—Laser phosphorimetryMCL—maximum contaminant levelsMCLG—maximum contaminant level goalsMDL—method detection limitn—numberNAG—N-acetyl-β-D-glucosaminidasNTNC—non-transient, non-communityNTNCWS—non-transient, non-community

water systemsPBMS—performance based measurement

systemPE—performance evaluationPOE—point-of-entryPOU—point-of-usePQL—practical quantitation levelPT—performance testingPWS—public water systemsRF—risk coefficientRfD—reference doseRO—reverse osmosisRSC—relative source contributionSM—standard methodsSMF—standardized monitoring frameworkSPAARC—Spreadsheet Program to Ascertain

Residual Radionuclide ConcentrationSSCTL—‘‘Small Systems Compliance

Technology List’’Stnd. Dev.—standard deviationTR—target risk levelUIC—underground injection control

Table of Contents

I. Purpose and Organization of this DocumentII. Statutory Authority and Regulatory

BackgroundA. Safe Drinking Water Act of 1974 and

Amendments of 1986 and 1996B. The 1991 ProposalC. Court AgreementD. Statutory Requirements for Revisions to

RegulationsIII. Overview of Today’s Document

A. Health Risk Consistency With ChemicalCarcinogens

B. Drinking Water ConsumptionC. Risk Modeling and the MCLD. Sensitive Sub-Population: ChildrenE. MCL for Beta Particle and Photon

RadioactivityF. Combined Ra-226 and Ra-228G. Gross Alpha MCLH. UraniumI. Inclusion of Non-Transient Non-

Community Water SystemsJ. Analytical MethodsK. MonitoringL. Effective DatesM. Costs and Benefits

IV. References

AppendicesI. Occurence

II. Health EffectsIII. Analytical MethodsIV. Treatment Technologies and CostsV. Economics and Impacts Analysis

I. Purpose and Organization of ThisDocument

In 1976, EPA promulgated drinkingwater regulations for severalradionuclides. In 1991 (56 FR 33050,July 18, 1991), EPA proposed revisionsto the current radionuclides (i.e. betaand photon emitters, radium-226 andradium-228, and gross alpha radiation)and proposed regulations for uraniumwhich is not currently regulated. EPA ispublishing this Notice of DataAvailability (NODA) to inform thepublic and the regulated community ofnew information concerningradionuclides in drinking water. EPA isevaluating these additional data todetermine how they will affect theAgency’s decisions relative to finalregulations to control radionuclides inpublic water systems. The Agency isunder a court agreement to publishthese final regulations by November2000. Information in today’s Documentincludes data about the occurrence,health effects, and treatment options forradionuclides in drinking water, as wellas analytical methods, and monitoringrequirements. This Document alsopresents data concerning the costs andbenefits of several regulatory options.EPA is soliciting public comment on anumber of issues raised by this newinformation. This introduction providesan overview of the document, and someof the information available to EPA andto highlight the risk managementdecisions the Agency is contemplating.Subsequent sections will contain morespecific information, with a focus onwhat is new, relative to each of thetopics listed previously. Finally, tofurther assist the public, the Agency hascompiled seven appendices, includedwith this NODA, with more detailedinformation on each of these topics inaddition to the public docket ofreference materials. EPA seeks commenton the data and information presentedin today’s NODA, particularly whereregulatory options or alternatives arediscussed. Commenters are asked toprovide their rationale and anysupporting data or information theywish to submit in support of commentsoffered.

Table I–1 summarizes the majorelements of the 1976 rule, the 1991proposal and the issues beingconsidered in today’s NODA.



TABLE I–1.—COMPARISON OF THE 1976 RULE, 1991 PROPOSAL, AND 2000 NODA

Provision 1976 Rule (Current Rule) 1991 Proposal 2000 NODA

Affected Systems CWS ...................................................... CWS + NTNC ....................................... CWS + several NTNC options basedon the 1991 proposal.

MCLG ................... no MCLG ............................................... MCLG of zero ........................................ MCLG of zero.Radium MCL ........ Combined Ra-226 + Ra-228 MCL of 5

pCi/L.Ra-226 MCL of 20 pCi/L; Ra-228 MCL

of 20 pCi/L.Maintain current MCL based on cor-

rected estimates of risk of currentMCL.

Beta/PhotonEmitters MCL.

4 mrem: Methodology for deriving indi-vidual concentration limits incor-porated by reference; MCL = sum ofthe fractions of dose from one ormore contaminants; risks estimatednot to exceed 5.6×10¥5.

4 mrem ede (Effective Dose Equiva-lent). Derived concentration limitschanged to reflect new dose limit;Current estimate of associated risksfor these concentration limits are be-tween 10¥4 and 10¥3 for most.

Maintain current MCL based on cor-rected estimates of risk of currentMCL.

Gross alpha MCL 15 pCi/L excluding U and Rn, but in-cluding Ra-226.

‘‘Adjusted’’ gross alpha MCL of 15 pCi/L, excluding Ra-226, radon, and ura-nium.

Maintain current MCL based on unac-ceptable risk level of 1991 proposedMCL.

Polonium-210 ....... Included in gross alpha ......................... Included in gross alpha ......................... No changes to current rule. Monitoringrequired under the UCMR rule. Fu-ture action may be proposed at alater date.

Lead-210 .............. Not Regulated ....................................... Included in beta particle and photon ra-dioactivity; concentration limit pro-posed at 1 pCi/L.

No changes to current rule. Monitoringrequired under the UCMR rule. Fu-ture action may be proposed at alater date

Uranium MCL ....... Not Regulated ....................................... 20 µg/L or 30 pCi/L w/ option for 5–80µg/L.

Three options being considered: 20,40, 80 µg/L and pCi/L

Ra-224 .................. Part of gross alpha, but sample holdingtime too long to capture Ra-224.

Part of gross alpha, but sample holdingtime too long to capture Ra-224.

Same as current rule, but Ra-224 maybe addressed in a future proposal.

Radium monitoring Ra-226 linked to Ra-228; measure Ra-228 if Ra-226>3 pCi/L and sum.

Measure Ra-226 and -228 separately. Measure Ra-226 and -228 separately

Monitoring base-line.

4 quarterly measurements. Monitoringreduction based on results: >50% ofMCL required 4 samples every 4 yrs;<50% of MCL required 1 sampleevery 4 yrs.

Annual samples for 3 years; Std Moni-toring Framework: >50% of MCL re-quired 1 sample every 3 years;<50% of MCL enabled system toapply for waiver to 1 sample every 9years.

Implement Std Monitoring Frameworkas proposed in 1991. Four initial con-secutive quarterly samples in firstcycle. If initial average level >50% ofMCL: 1 sample every 3 years; <50%of MCL: 1 sample every 6 years;Non-detect: 1 sample every 9 years.(beta particle and photon radioac-tivity has a unique schedule—seeSection III, part K).

Beta monitoring .... Surface water systems >100,000 popu-lation Screen at 50 pCi/L/; vulnerablesystems screen at 15 pCi/L.

Ground and surface water systemswithin 15 miles of source screen at30 or 50 pCi/L. Those drawing waterfrom a contaminated source screenat 15 pCi/L.

Same as 1991 proposal with clarifica-tions.

Gross alpha moni-toring.

Analyze up to one year later ................. Six month holding time for gross alphasamples; Annual compositing ofsamples allowed.

As proposed in 1991. Recommendationto analyze within 48–72 hours tocapture Ra-224.

Analytical Methods Provide methods ................................... Method updates proposed in 1991;Current methods were updated in1997.

Current methods with clarifications.

II. Statutory Authority and RegulatoryBackground

A. Safe Drinking Water Act of 1974 andAmendments of 1986 and 1996

Regulations for radionuclides indrinking water were first promulgatedin 1976 as interim regulations under theauthority of the Safe Drinking Water Act(SDWA) of 1974. The standards were setfor three groups of radionuclides: betaand photon emitters, radium (radium-226 and radium-228), and gross alpharadiation. These standards becameeffective in 1977.

The SDWA Amendments of 1986required EPA to establish health-based

regulatory targets, called MaximumContaminant Level Goals (MCLGs), forevery contaminant ‘‘at the level atwhich no known or anticipated adverseeffects on the health of persons occurand which allows an adequate margin ofsafety.’’ The enforceable standard, theMaximum Contaminant Level (MCL),was required to be established ‘‘as closeto the health-based goal as feasible usingthe best available technology, takingcosts into consideration.’’ EPA proposedan MCLG of zero for the radionuclidesin 1991.

In 1983 and 1986, EPA published anAdvanced Notice of ProposedRulemaking (ANPRM) requesting

additional information and commentson radionuclides and numerous organicand inorganic contaminants in drinkingwater. The 1986 SDWA Amendmentsidentified 83 contaminants for EPA toregulate, including the currentlyregulated radionuclides, which lackedan MCLG, and two additionalradionuclides, uranium and radon. TheAmendments also declared the 1976interim standards to be final NationalPrimary Drinking Water Regulations.

In 1996, Congress again amended theSDWA. These amendments includednew and revised provisions that must beconsidered when revising drinkingwater regulations. Among these are the



health protection clause (section1412(b)(9)) which requires that ‘‘anyrevision of a national primary drinkingwater regulation (NPDWR) shall bepromulgated in accordance with thissection, except that each revision shallmaintain, or provide for greaterprotection of the health of persons.’’

The 1996 Amendments also providefor a cost-benefit analysis whenpublishing a proposal for new NPDWRspursuant to section 1412 (b)(6). Whilethe EPA had proposed the radionuclidesrule prior to these Amendments, theAgency nevertheless conducted ananalysis of the costs and associatedbenefits of all of the options describedin today’s Document. These analysesserve to update and revise the costs andbenefits estimated for the 1991 proposedrule. For the uranium standard, theAgency solicits comment on thepossible use of its new discretionaryauthority at section 1412(b)(6) of theSDWA, which allow for a proposedregulatory level to be set higher than thefeasible level, after the Agency has madea determination that the benefits do notjustify the costs at the feasible level.Note that section 1412(b)(6) applies tonew standards (uranium), not to therevision of existing standards (combinedradium-226 and -228, gross alpha, andbeta particle and photon radioactivity).Where we expect to maintain currentstandards at their existing levels, noadditional analysis was undertakenbecause the rule is already in effect.

B. The 1991 ProposalIn 1991, EPA proposed new

regulations for uranium and radon, aswell as revisions to the existingregulations. The proposal included thefollowing features: (1) an MCLG of zerofor all ionizing radiation; (2) revisedMCLs for beta particle and photonradioactivity, radium-226, radium-228,and gross alpha emitters; (3) proposedMCLs for uranium and radon; and (4)revisions to the categories of systemsrequired to monitor, the monitoringfrequencies, and the appropriatescreening levels. EPA receivedcomments on the new data andregulatory options presented in the 1991proposal. However, the proposal wasnever promulgated as a final rule inlarge part because of controversysurrounding the proposed MCL forradon. The 1996 Amendments to theSDWA directed the Agency to withdrawthe proposed MCL for radon, which wassubsequently done on August 6, 1997(62 FR 42221).

Most of the comments EPA receivedon the proposal related to radon.Approximately 120 comments related tonon-radon radionuclides were valuable

and most are still germane to theAgency’s rulemaking efforts. Thosecomments are addressed, as appropriate,in today’s document.

C. Court AgreementThe SDWA (as amended in 1986)

provided a statutory deadline topromulgate a revised radionuclide ruleof June 1989, but EPA failed to meet thisdeadline. An Oregon plaintiff broughtsuit to require EPA to issue theregulations and EPA entered into aseries of consent agreements settingschedules to issue regulations for theradionuclides. EPA issued a proposal in1991. After the SDWA Amendments in1996, EPA agreed to publish a finalaction with respect to the proposedregulation for uranium by November 21,2000. EPA also agreed to either takefinal action by the same date withrespect to radium, beta/photon emitters,and alpha emitters or publish a noticestating its reasons for not taking finalaction on the proposal. This latterscenario would leave the current rule ineffect.

D. Statutory Requirements for Revisionsto Regulations

Both the 1986 and the 1996Amendments to the SDWA state thatrevisions be made to existing drinkingwater regulations periodically. Section1412(b)(9) of the 1986 SDWAAmendments directed that ‘‘nationalprimary drinking water regulations beamended whenever changes intechnology, treatment techniques, andother means permit greater protection ofthe health of persons, but in any event,such regulations shall be revised at leastonce every 3 years.’’ The 1996 SDWAAmendments provide that EPA ‘‘ * * *not less than every 6 years review andrevise, as appropriate, each nationalprimary drinking water regulation,’’ andthat ‘‘any revision shall maintain, orprovide for greater, protection of thehealth of persons.’’

The radionuclides emit ionizingradiation and, absent data indicatingthat there is a threshold level at whichexposure does not present a risk, EPAuses a linear, non-threshold model to seta zero MCLG for radionuclides. Thismeans that any exposure can potentiallycause harm and that risk associated withthe exposure increases proportionally tothe concentration of the radionculide.

EPA’s current estimate of the unitrisks posed by many of theradionuclides covered by today’sdocument has generally increasedrelative to the 1991 estimate. In fact,based on the newest science (FederalGuidance Report 13), the fatal cancerrisks associated with the 1991 proposed

MCL changes for combined radium,gross alpha, and beta particle andphoton radioactivity generally exceedthe Agency’s risk range of 10¥6 to 10¥4.This document discusses and requestscomment on the issues EPA hasaddressed in determining how to bestmeet applicable SDWA provisions foreach of the radionuclide categoriescovered by today’s document.

III. Overview of Today’s DocumentAdditional data since the 1991

proposal suggest a need to retain someportions of the proposal, while retainingmuch of the current rule. Any changesthat are finalized must meet theprovisions for public health protectionin accordance with the 1996Amendments. EPA has presented itsapproach for finalizing the non-radonportions of the 1991 radionuclidesproposal at several public meetings.

In December 1997 EPA held a publicforum (stakeholder meeting) to discussthe requirements and limitations of thenew Amendments pertaining torevisions to the radionuclide regulation.The Agency discussed most of theconcepts presented in this documentand received valuable feedback from thepublic, the regulated community, andother Federal Agencies. In thisDocument, EPA is presenting thecurrent information and options uponwhich the Agency will make itsdecisions regarding revisions to theexisting standards. At the same time, theAgency is requesting additional dataand comments on the approach EPAexpects to take in formulating the finalrule.

The most significant new informationconcerns the occurrence, monitoring,and health effects of radionuclides indrinking water. Recent data suggest amore widespread occurrence of certainradionuclides which may point to aneed for improved monitoring for theseradionuclides in certain areas of thecountry. Conversely, a betterunderstanding of the occurrencepatterns may also indicate the need forless frequent monitoring. The newesthealth effects models, which are basedon improved age-dependent biokineticand dosimetric models of the effects ofionizing radiation on the body and morerecent epidemiological information,reveal that radionuclides generallypresent a somewhat greater risk than theestimates of previous models, includingthe 1991 RADRISK model. EPA’spublication ‘‘Federal Guidance Report13’’ (FGR–13, EPA 1999b) discusses thenewest risk modeling. The resulting riskestimates based on of the new healtheffects models are largely the reason forthe publication of this document. The



1 If one ranked, from lowest to highest, theaverage daily water consumption levels for everyCWS customer in the U.S., the ‘‘90th percentile’’value of 2.2 L/day is the best estimate of the valuefor which 90 percent of the population would drinkthat much or less on an typical day.

following are some aspects of the NODAwhich the Agency would like tohighlight.

A. Health Risk Consistency WithChemical Carcinogens

The risks associated with exposure tochemical carcinogens are usuallyexpressed as the risks of illness. It isEPA policy to issue standards thatmaintain a risk ceiling in the target riskrange of 10¥6 (one in one million) upto 10¥4 (one in ten thousand). Forconsistency between the level ofprotection between chemical andradiological drinking watercontaminants, EPA is consideringutilizing whichever risk provides thegreater protection for MCL changes, a1×10¥4 risk of cancer incidence, or amortality risk at half the incidence,5×10¥5. The risk of death at 5×10¥5 isthe more protective if the mortality ratefrom a particular radionuclide is morethan 50%, which is true for most of theradionuclides. However, for the thyroid,the mortality rate from thyroid cancer isat 10%. Protecting at 1×10¥4 incidencecorresponds to a mortality at 1x10¥5.Conversely, protecting at 5×10¥5

mortality with only a 10% mortality rateallows an incidence, of 5×10¥4, a lessprotective number.

B. Drinking Water ConsumptionEPA received comments in 1991 from

the American Water Works Association(AWWA), the Colorado Water QualityCommission, the Atlantic Richfield Co.,and the Rio Algom Mining Corp.suggesting that consumption of drinkingwater was actually 1.2 liters per day,thus EPA was being too conservative inusing two liters per day.

When establishing an MCL for acarcinogen, the risk which the MCLwould represent is considered as well astreatability and costs. Radionuclideswill have an MCLG of zero, with MCLsbased on standard assumptions of twoliters intake per person per day (2 L/day), an average individual weight of 70kg, and a 70 year life span. EPA now hasdata to indicate that the averageconsumption of tap water is 1.1 litersper day per person and that aconsumption rate of 2.2 L/dayrepresents the 90th percentileconsumption level.1 Basing the MCL ona consumption rate higher than theaverage value is justified since MCLs areintended to be protective of the personsthat comprise the population and not

just ‘‘typical individuals’’. Since aconsumption value of 2 L/day is lessthan the 90th percentile consumptionrate, EPA believes that its assumption of2 L/day for MCL determinations is notoverly conservative and is justifiable.

When computing the national benefitsof a regulation and the estimate ofcancer mortality risks or risk reductions,EPA is now using 1.1 liters per personper day (L/day) of water as the estimateof the average daily consumption ratefor individuals. In effect, this reducespopulation risk estimates byapproximately one half and reduces theestimate of risk reductions byapproximately one half. Since benefitscalculations are based on riskreductions, this reduces monetizedbenefits by approximately one half. Itshould be noted that it is consistent toset health protection levels based on asubset of individuals that face thehighest risks (sensitive subpopulationsand/or the substantial minority of thepopulation have higher waterconsumption levels), while estimatingbenefits based on average individuals(average consumption and sensitivity).EPA believes this approach leads toprotective MCLs and realistic benefitscalculations.

C. Risk Modeling and the MCLThe Agency’s current radionuclides

health effects model is based on FederalGuidance Report 13 (FGR–13, EPA1999b). The Agency’s new health effectsmodel uses state-of-the-art methods,models and data that are based on themost recent scientific knowledge.Compared with the approaches used in1976 and 1991, the revised methodologyincludes substantial refinements(described in appendix II, ‘‘HealthEffects’’). While commenters havepointed out the MCLs in the current ruleare based on ‘‘old science’’, the newestscience indicates that many of the MCLsproposed in 1991 have correspondingrisks that are much greater than theupper limit of the Agency’s acceptablelifetime excess risk range ofapproximately 10¥6 to 10¥4 (one in onemillion to one in ten thousand lifetimeexcess risk of cancer). The risksassociated with each existing andproposed MCL are described in sectionsthat follow. The risk models aredescribed in detail in appendix II(Health Effects) and in the TechnicalSupport Document for theRadionuclides Notice of DataAvailability (EPA 2000a).

Between 1976 and the present,different scientific models have beenused to calculate risks from radiationexposure. Each model derives adifferent concentration of a particular

nuclide for a given level of risk. Forexample, in 1991, the RADRISK modelindicated that consuming drinkingwater with radium-228 at 26 pCi/Lwould lead to an excess lifetime cancerrisk of 1×10¥4. However, using today’smodel (based on Federal GuidanceReport 13), the best estimate of lifetimerisk of Ra-228 at 26 pCi/L is 1×10¥3, arisk value ten times greater than thoughtin 1991.

Likewise, the 1991 proposed MCL forRa-228 at 20 pCi/L was thought tocorrespond to lifetime excess cancer riskof 7.7×10¥5. The most current riskestimate for Ra-228 at 20 pCi/L7.7×10¥4, again ten-fold greater andmuch higher than the Agency’s targetrisk ceiling of 10¥4. For individualsconsuming water with 20 pCi/L of bothRa-228 and Ra-226, the risk was thoughtto be 1.7×10¥4 in 1991. However, basedon the newest science, these individualswould be exposed to lifetime excessrisks of 1×10¥3 risk (one in a thousand),a risk level 10-fold higher than theAgency’s target risk ceiling for drinkingwater MCLs. EPA requests comments onthese issues.

D. Sensitive Sub-Population: Children

The age-specific, sex-specific modelsused by EPA for estimating risk fromionizing radiation implicitly provide forrisk differentiation by gender and age.The computer program suite, DCAL(FGR–13), uses age-specific metabolicmodels to calculate the dose from a unitintake of a radioisotope during eachyear of life from birth to 120 years ofage. Age-specific organ masses are usedfor all ages up to adult, and for adultmales and adult females. Riskcoefficients are given by age and sex foreach year of life from birth to 120 yearsof age. The risk is then calculated bycombining calculated doses and age-sex-specific risk coefficients with age-sex-specific intake data and age-sex-specificsurvival data.

A separate risk analysis for childrenwas performed and is described inappendix II (Health Effects), part C.Risks to children are explicitlyconsidered when setting MCLs forradionuclides. In the case of theregulated water systems (currently,community water systems), children arefully protected. In the case of theunregulated systems of potentialconcern (non-transient non-communitywater systems, NTNCWSs), the analysisis more complicated. Risks to childrenserved by NTNCWSs are discussed inappendix II, part C, number 3.



E. MCL for Beta Particle and PhotonRadioactivity

1. EPA’s Plans for Finalizing the 1991Proposed MCL for Beta and PhotonRadioactivity

This section presents the importantconsiderations that have led EPA toconsider retaining the current MCL forbeta particle and photon radioactivitywhen the 1991 proposal is finalized inNovember of 2000. EPA is, however,also considering finalizing the 1991proposed changes to the monitoringrequirements for beta particle andphoton radioactivity, as described laterin this section. The current MCL is (40CFR 141.16):

(a) The average annual concentrationof beta particle and photon radioactivityfrom man-made radionuclides indrinking water shall not produce anannual dose equivalent to the total bodyor any internal organ greater than 4millirem/year.

(b) Except for the radionuclides listedin Table A, the concentration of man-made radionuclides causing 4 mremtotal body or organ dose equivalentsshall be calculated on the basis of a 2liter per day drinking water intake usingthe 168 hour data listed in ‘‘MaximumPermissible Body Burdens andMaximum Permissible Concentrations

of Radionuclides in Air or Water forOccupational Exposure,’’ NBSHandbook 69 as amended August 1963,U.S. Department of Commerce. If two ormore radionuclides are present, the sumof their annual equivalent to the totalbody or to any organ shall not exceed4 millirem/year.

TABLE A.—AVERAGE ANNUAL CON-CENTRATIONS ASSUMED TOPRODUCE A TOTAL BODY OR ORGANDOSE OF 4 MREM/YEAR.

Radionuclide Critical organ pCi perliter

Tritium ............. Total body ....... 20,000Strontium-90 .... Bone marrow .. 8

Following these instructions leads toa unique list of concentration limits for168 other man-made radionuclides. Thislist is included in today’s document inappendix II, ‘‘Health Effects.’’

The 1991 proposed MCL for betaemitter and photon radioactivity was 4mrem-ede (effective dose equivalents),with the footnote:

‘‘NOTE. —The unit mrem-ede/yr refers tothe dose committed over a period of 50 yearsto reference man (ICRP 1975) from an annualintake at the rate of 2 liters of drinking waterper day.’’

Following these instructions leads toa unique list of concentration limits for230 radionuclides. EPA has determinedthat there is no way to update the 4mrem dose basis (1976) for the betaparticle and photon radioactivity MCLwithout the extensive process of a newproposal. While some stakeholders havesuggested that reverting to the existingrule for beta particle and photonradioactivity (‘‘beta emitters’’) is relyingon ‘‘old science,’’ it should be pointedout the newest risk estimates, based onthe peer-reviewed Federal GuidanceReport 13, indicates that the risksassociated with the 1991 proposed MCLof 4 mrem-ede (effective doseequivalents) are above the 10¥4 risklevel (10¥3 to 10¥4) for many of the betaemitters. Figure 1 shows the mostcurrent risk estimates for the betaemitter concentration limits derivedunder both the current and proposedMCLs. As the figure shows, the currentMCL results in concentration limitswith risks that fall within the Agency’srisk range goal of 10¥6 to 10¥4 (whilesome are slightly above and someslightly below, all round to valueswithin these orders of magnitude).

BILLING CODE 6560–50–U



BILLING CODE 6560–50–C



In summary, the Agency fullyrecognizes that the dose-based MCL of4 mrem/year is based on older scientificmodels. However, the Agency hasdecided to retain the current MCL giventhat:

• Federal Guidance Report 13 (FGR–13, EPA 1999b) demonstrates that the1991 proposed MCL of 4 mrem-ede/yearresults in concentration limits that areoutside the 10¥6 to 10¥4 range;

• FGR–13 demonstrates that thecurrent MCL of 4 mrem/year results inconcentration limits that are within the10¥6 to 10¥4 range;

• the fact that there is no evidence ofappreciable occurrence of man-madebeta emitters in drinking water;

• the 1996 Safe Drinking Water Actrequires EPA to evaluate all NPDWRsevery six years (‘‘Six Year Review’’).

EPA believes that Six Year Review isthe appropriate vehicle for updating thebeta particle and photon radioactivityMCL.

2. Beta Particle and PhotonRadioactivity Monitoring

Currently, surface water systemsserving more than 100,000 persons arerequired to monitor for beta particle andphoton radioactivity using a screeninglevel of 50 pCi/L, while systems that aredetermined to be vulnerable by the Stateare required to monitor using ascreening level of 15 pCi/L. In 1991,EPA proposed that all ground water andsurface water systems within 15 miles ofa potential source, as determined by theState, be required to monitor using ascreening level of 30 or 50 pCi/L. EPAis considering retaining the currentmonitoring requirement of a 15 pCi/Lscreen for water systems drawing waterfrom contaminated sources. EPA solicitscomment on these issues. EPA is takingcomment on screening levels of 30 or 50pCi/L for systems within 15 miles of apotential source.

3. Lead-210 and Radium-228The 1991 proposal included lead-210

(Pb-210) and radium-228 (Ra-228) in thelist of regulated beta and photonemitters, both of which are naturally

occurring. An 1991 the Agency wasconsidering raising the Ra-228 MCL to20 pCi/L, which is high enough tosignificantly contribute to gross betalevels. However, since the Agency isretaining the current combined Ra-226and Ra-228 standard of 5 pCi/L, Ra-228will no longer be a significantcontributor to gross beta. For the reason,the Agency sees no value in includingRa-228 in the list of beta/photonemitters.

New risk analyses indicate that Pb-210 is of concern well below the currentand proposed screening levels for betaand photon emitters. In order to assessthe occurrence of Pb-210 to determine ifit is present at levels high enough towarrant separate monitoring, EPA hasincluded it on the list published in theUnregulated Contaminant MonitoringRule (UCMR) (64 FR 50556, Friday,September 14, 1999). USGS alsomonitored for Pb-210 in its study withEPA of 100 locations. The reader isreferred to appendix I and the TechnicalSupport Document (EPA 2000a) forfurther information regarding this study.Since Pb-210 specific monitoring wasnot proposed in 1991, EPA cannotaddress this concern without a newproposal. After occurrence data hasbeen reviewed from the UCMR, EPAmay propose appropriate actions.

F. Combined Ra-226 and Ra-228

1. MCL ConsiderationsThe combined radium-226 and -228

NPDWR has long been a contentiousissue. A number of water systemsbelieve the current MCL is too stringentand have not installed treatment ortaken other measures to comply. EPAfirst proposed the possibility ofincreasing the current 5 pCi/L limit forcombined radium-226 and -228 in 1991.The proposal suggested a new level of20 pCi/L for Ra-226 and Ra-228separately along with a proposed limitof 300 pCi/L for radon-222 . Thiscombination was proposed in part dueto the disproportionate costs ofremoving radium compared to radon.The proposal was met with opposition,

largely due to the controversysurrounding the radon component. Inthe ensuing deliberations, debatesregarding the radon component of theproposal interfered with promulgationof the proposal. In the 1996Amendments to the SDWA, Congressdirected EPA to remove the radoncomponent from the proposal.Consequently, the Agency has onceagain considered the issues surroundingthe allowable concentration of radium-226 plus radium-228 in drinking water.

EPA is considering retaining thecurrent MCL for combined radium-226and -228 at 5 pCi/L for the followingreasons. First, the unit risks for Ra-226and Ra-228 are believed to be muchgreater than estimated in 1991, such thatraising the combined Ra-226 and Ra-228MCL up to 20 pCi/L for eachradionuclide would result in a lifetimeexcess cancer risks that are ten-foldhigher than the Agency’s acceptable riskrange of 10¥6 to 10¥4. And second, EPAis required to consider the MCL for Ra-226 and Ra-228 apart from any NPDWRfor radon, both by the 1996 SDWAAmendments and the later courtstipulated agreement. Both points arediscussed further here.

First, in 1976 the estimate of risk fromeither Ra-226 or Ra-228 at 5 pCi/L wasbetween 5×10¥5 and 2 ×10 ¥4,averaging 1×10¥4. In 1991 the RADRISKmodel calculated that a 1×10¥4 riskcorresponded to Ra-228 at 26 pCi/L andRa-226 at 22 pCi/L.

Table III–1 shows the change inestimated risks from 1976 until thepresent. ‘‘Current Risk Estimates’’ arecalculated using the 1999 model, FGR–13 (EPA 1999b). The table allows acomparison between the calculated riskduring each phase of the evolution ofthe radionuclides NPDWRs, includingthe current best estimate of risk basedupon FGR–13 (EPA 1999b). Details ofwhy the models have changed and theadditional data taken into considerationare found in the appendix II and theTechnical Support Document (EPA2000a).

TABLE III–1.—CHANGES IN ESTIMATED RISKS FOR VARIOUS RA-226 AND RA-228 LEVELS

Year model used

Radium-228 Radium-226

Concentra-tion pCi/L

Previous riskestimate

Current riskestimate




2000 FGR–13 .................................................... 5 2 × 10¥4 2 × 10¥4 5 7.3 × 10¥5 7.3 × 10¥5

2000 FGR–13 .................................................... 2.5 1 × 10¥4 1 × 10¥4 6.85 1 × 10¥4 1 × 10¥4

1994 FGR–11 .................................................... 11 1 × 10¥4 4.5 × 10¥4 10 1 × 10¥4 1.5 × 10¥4

1991 RADRISK .................................................. 26 1 × 10¥4 1 × 10¥3 22 1 × 10¥4 3.3 × 10¥4

1991 RADRISK proposed MCL ......................... 20 7.7 × 10¥5 7.7 × 10¥4 20 9.1 × 10¥5 2.9 × 10¥4

1991 RADRISK .................................................. 5 1.9 × 10¥5 2 × 10¥4 5 2.3 × 10¥5 7.3 × 10¥5



TABLE III–1.—CHANGES IN ESTIMATED RISKS FOR VARIOUS RA-226 AND RA-228 LEVELS—Continued

Year model used

Radium-228 Radium-226







1976* .................................................................. 5 1 × 10¥4 2 × 10¥5 5 1 × 10¥4 7.3 × 10¥5

*The risk of either radium-226 or radium-228 at 5 pCi/L was believed to be between 5×10¥5 and 2×10¥4 in 1976. The average would havebeen 1×10¥4.

The 1991 estimated riskcorresponding to 20 pCi/l of Ra-226 inaddition to 20 pCi/l of Ra-228 wasthought to be 1.7 × 10¥4. However, thecurrent risk estimate based on FGR–13(EPA 1999b) for 20 pCi/l of Ra-226 inaddition to 20 pCi/l of Ra-228 is 1×10¥3

(one in a thousand), an order ofmagnitude (ten times) above theacceptable risk of 1×10¥4.

In contrast, maintaining the currentstandard would allow a maximumlifetime risk of 2×10¥4 (within theoriginal risk range of the 1976regulation). This represents a one in5,000 lifetime mortality risk and wouldonly be present if 5 pCi/L in thedrinking water were all radium-228, arelatively rare occurrence situation. Ifthe radium present were all radium-226,the risk would be 7×10¥5, just belowEPA’s risk ceiling. Since:

• the risks associated with the currentMCL of 5 pCi/L are already at the upperend of the Agency’s allowable risk rangeof 10¥5; and

• the 1991 proposed MCLs for Ra-226and Ra-228 have risks as high as1×10¥3, ten-fold higher than theAgency’s allowable risk, the Agencybelieves that maintaining the currentMCL for combined Ra-226 and Ra-228 isthe appropriate action.

Regarding treatment feasibility, EPA’sdetermination that water systems canfeasibility treat and quantify combinedradium at 5 pCi/L is supported by casestudies of systems that had combinedradium levels in excess of the MCL andthat later came into compliance throughtreatment. In addition, EPA has casestudies of systems that have come intocompliance through purchasing water,blending, and developing new wells(EPA 2000a).

Since risk estimates for Ra-228 aresignificantly higher than thought in1991, EPA has evaluated the riskreductions, costs, and benefits ofdecreasing the allowable level ofradium-228 to 3 pCi/L and hasdiscussed the results in the TechnicalSupport Document (EPA 2000a). Theconcern is that a system with 5 pCi/L ofRa-228 with insignificant levels of Ra-226 would be in compliance with thecombined radium MCL, but would have

an associated lifetime excess cancermorbidity risk of 2 × 10¥4, whichexceeds the risk ceiling on 1 × 10¥4.While this is true, the occurrence datareported in appendix I suggest that thissituation should be rare. Since EPA didnot propose this action in the 1991proposal, EPA cannot address thisconcern in the finalization of thisproposal. However, EPA will considerthis situation further and will laterdetermine if a regulatory action isappropriate.

An unintended effect in the 1991proposal was that the costs and benefitswere not evenly distributed to allaffected persons (individuals). In the1991 proposal, an MCL was proposedfor radon at 300 pCi/L and a revisedMCL for radium from 5 pCi/L combinedfor both radium-226 and radium-228 to20 pCi/L each. Benefits and costs wereconsidered together for both radon andradium on a national basis. Compared toradium, radon is easier and cheaper toremove from water due to its airstrippability. Since the risks avoidedwere higher and the treatment costslower for the radon MCL, it wasreasoned that the radon rule was muchmore cost-effective than the combinedradium rule. However, since radium andradon do not tend to co-occur,individuals that would have benefittedfrom the radon rule were not the sameindividuals that would have facedhigher risks under the proposed radiumMCLs. EPA believes that such a trade-off is no longer appropriate. Amongother considerations, the 1996Amendments to SDWA explicitlyseparated the radon rule from the rulefor the other radionuclides.

In summary, EPA based its proposedincrease in the radium standard on therisk models that existed at that time andon a population risk trade-off withradon. The models in use in 1991indicated that radium posed less of arisk than originally believed in 1976.However, current risk models (FGR–13,EPA 1999b) suggest that the combinedradium standard of 5 pCi/L presents aneven greater health risk than thought in1976. Given the much higher currentestimate of risks associated with theproposed Ra-226 and Ra-228 MCLs of 20

pCi/L and the statutorily requiredwithdrawal of radon-222 from theproposal, the Agency believes that theMCLs for radium proposed in 1991 areno longer appropriate. EPA requestspublic comment on retaining the currentradium standard of 5 pCi/L forcombined Ra-226 and Ra-228.

2. Separate Radium Analysis

The 1991 proposal recommendeddecoupling the monitoring of radium-228 from radium-226. The currentradionuclides rule requires analysis ofRa-228 only when Ra-226 levels areabove 3 pCi/L. The rule recommendsanalysis of Ra-226 and/or Ra-228 whengross alpha exceeds 2 pCi/L where Ra-228 may be present, and requiresanalysis of Ra-226 when gross alphaexceeds 3pCi.L.

Ra-228 may be present with minoramounts of Ra-226 or in the absence ofRa-226. In general, the mobility of aparent radionuclide may be verydifferent from that of a daughterelement, depending on the geochemistryof the elements involved. However, theoccurrence of a radionuclide may stillbe governed by the occurrence anddistribution of its parent (see EPA2000a). Since radium-226 arises fromthe uranium decay series and radium-228 arises from the thorium series, it islogical to expect them to occurindependently of one another. Also, theparents of Ra-226 (uranium isotopes)and Ra-228 (thorium isotopes) have verydifferent geochemical behaviors.Uranium is fairly mobile in oxidizingground waters, while thorium is ratherinsoluble. In contrast, the daughterradium isotopes are more mobile inreducing waters and are relativelyimmobile in oxidizing waters. Since Ra-226 is part of the uranium series(relatively mobile parent) and Ra-228 ispart of the thorium series (immobileparent), Ra-226 can and does mobilizein waters containing Ra-228 morefrequently than the reverse situation.These observations indicate that Ra-226and Ra-228 may be expected tosignificantly co-occur, but that thecorrelation will not be strong enough touse the occurrence of one to predict theother with acceptable certainty. Recent



studies support this conclusion (EPA2000a).

This conclusion indicates that thecurrent monitoring screen for Ra-228based on Ra-226 is not reliable.Therefore as proposed in 1991, EPA isconsidering requiring separatemonitoring and analysis of both radium-226 and radium-228 in the final rule.The Ra-228 and Ra-226 results would besummed to determine compliance withthe radium MCL. This will provide amore accurate assessment of systemscontaining little or no radium-226, butpossessing a significant enoughconcentration of radium-228 to exceedthe standard.

G. Gross Alpha MCLThe gross alpha standard promulgated

in 1976 considered natural and man-made alpha emitters as a group ratherthan individually. At the time, theanalytical costs made it impractical toidentify each alpha-emitting nuclide ina given water sample. The existing grossalpha MCL includes radium-226, butexcludes radon-222 and uranium(because these latter nuclides were to beregulated at a later date). The 1991 riskestimates indicated that the inclusion ofRa-226 was not warranted. However,today’s risk estimates, based on FGR–13(EPA 1999b), suggest that the Ra-226unit risk is large enough to warrant toinclude it in gross alpha, as in thecurrent standard. In today’s Document,the Agency is considering maintainingthe current MCL for gross alpha,believing it to be protective. EPA willconsider proposing changes to the rulein the future.

EPA believes that the term ‘‘grossalpha’’ may be confusing. ‘‘Gross alpha’’implies counting the total alphaemissions and is the appropriate namefor that particular analytical method.The standard excludes uranium andradon from the total or gross count. Justas the proposal suggested the term‘‘adjusted gross alpha’’ with theexclusion of radium-226, EPA believesthe term ‘‘net alpha’’ or ‘‘the alphastandard’’ might better describes thecurrent standard which excludes suchalpha emitters as radon, uranium. EPArequests public comment on the namechange.

The gross alpha MCL was originallyestablished at 15 pCi/L to account forthe risk from radium-226 at 5 pCi/L (theradium regulatory limit) plus the riskfrom polonium-210, the next mostradiotoxic element in the uraniumdecay chain. In 1976, the risk resultingfrom exposure to 10 pCi/L of polonium-210 was thought to be equivalent to therisk resulting from exposure to 1 pCi/Lof radium-226. Looked at another way,

the 1976 gross alpha standard equatedto 6 pCi/L of radium-226 (5 pCi/L ofradium-226, plus the10 pCi/L ofpolonium-210 which itself was equal to1 pCi/L of radium-226). Since the riskassociated with the combined radiumstandard was believed to be in the rangeof 5×10¥5 to 2×10¥4, this assumptionplaced the gross alpha standardreasonably within that range as well.

The gross alpha standard proposed in1991 remained at 15 pCi/L, butexcluded radium-226 (because it wasproposed at 20 pCi/L). The new limitwas termed ‘‘adjusted gross alpha.’’ Ineffect, it allowed an increase of 5 pCi/L of non-radium alpha emitters indrinking water from 10 to 15 pCi/L byoccupying the 5 pCi/L originallyrepresented by the radium. In the 1991proposal, the allowable non-radiumgross alpha contribution in that samewater sample i.e. Po-210, would be 15pCi/L. Because this latter scenariorepresents more risk than the scenarioevaluated for the current regulation,EPA no longer supports an ‘‘adjustedgross alpha’’ limit of 15 pCi/L.

In the future, EPA may consider aproposal to exclude radium-226 fromthe gross alpha MCL as proposed in1991 (because of the existence of aseparate standard for radium-226), butto maintain protection, limiting thegross alpha standard to 10 pCi/L.Reducing the limit has the advantage ofeffectively reducing exposure topolonium-210 and radium-224. Inaddition, excluding radium-226 frombeing in both the gross alpha andradium standards may avoid confusion.EPA examined the possibility of thischange in the context of the potential foradded treatment costs versus themarginal benefits to be derived.However it appears that retaining thestandard at 15 pCi/L is protective ofpublic health at a reasonable cost. ApicoCurie cap of 15 represents differentrisks for various nuclides, but this is notunlike other regulated carcinogens orthe other radionuclides. The risksrepresented by two components, namelyradium-224 and Polonium-210, arediscussed next.

1. Polonium-210Current risk estimates suggest that the

risk resulting from exposure topolonium-210 is ten times greater thanoriginally believed in 1976 compared toradium. However, existing occurrencedata indicates that its presence indrinking water is relatively rare. To gaina better understanding of the publichealth risk posed by polonium-210 indrinking water, EPA included thisradionuclide in the Agency’sUnregulated Contaminants Monitoring

Rule (64 FR 50556, Friday, September17, 1999). The Agency may consider afuture proposal to develop a separatelimit for polonium-210 within (orseparate from) a potentially revisedgross alpha standard.

EPA believes that current technologycan limit polonium-210 to 4 pCi/L orbelow, although precise quantificationat this level may present a challenge.Because of its energetic alpha emissions,a gross alpha measurement mayoverestimate the actual concentration ofpolonium-210 in the sample by a factorof two. With current gross alphameasurement, if the total alpha were 15pCi/l contributed by polonium, theactual concentration of polonium couldbe much less, depending on thecalibration standard. At present, sincethere is no specific drinking waterregulation for polonium-210, there is noEPA-approved method for measuringpolonium to determine compliance witha drinking water standard. Should EPAdecide to develop a separate limit forpolonium-210, the Agency will ensurethat the approved analytical method fordemonstrating compliance is in placeand includes a calibration standardappropriate for polonium’s energeticalpha, thereby reducing the possibilityof overestimating its presence. EPArequests information relative to anyknown occurrence of polonium and theneed for a proposal of a separate limit.Recently, USGS co-operated with EPAand the American Water WorksAssociation in monitoring forradionuclides, including Po-210 (103wells in 27 States). The study andfindings are described in EPA 2000a).USGS will publish the study in the nearfuture. In this study, Po-210 levels werefound above 1 pCi/L in less than twopercent of the wells. Since the wellswere targeted for high radiumoccurrence, this may not be typical. Thereader is referred to appendix I(Occurrence) and the Technical SupportDocument (EPA 2000a) for furtherinformation.

2. The Occurrence of Radium-224 andits Impact on Alpha

Recently, the short lived isotope ofradium has been found in somedrinking water supplies. Extensivemonitoring in the State of New Jerseyover the past several years and follow-on survey by EPA and the USGS hasdemonstrated that radium-224 may bepresent in significant quantities inground water, especially where itsdecay chain ancestor radium-228 ispresent. Although it is included in the(gross) alpha MCL, it was not targetedspecifically for several reasons: (1) Itwas not believed to be a health risk, (2)



it was not known to be prevalent and (3)sampling it at a representative pointwithin the distribution system ratherthan the entry point to the systemallowed decay. However, newer FGR–13risk estimates (EPA 1999b), coupledwith the greater occurrence, and the1991 proposal to sample at the entrypoint to the distribution system, nowmake radium-224 a concern.

Radium-224 is a naturally occurringradioisotope, which is part of thethorium decay chain. It emits alphaparticles and has a half-life of 3.66 days.The decay of its progeny via alpha andbeta decay also happens very quickly. Inapproximately 4.1 days, an originalradium-224 atom has decayed to stablelead-208 by emission of an equivalent of4 alpha and 2 beta particles. A grossalpha analysis will detect 3 alphaparticle emissions including daughtersin equilibrium with the parent Ra-224.If a sample analysis is done within 72hours, preferably 48 hours, anappropriate back-calculation can beperformed of the gross alpha count ofthe sample water. Otherwise thelaboratory will significantlyunderestimate the radium-224 and otheralpha emitters that may have beenoriginally present in the sample.

Under the current rule, utilities areallowed to collect quarterly samples,composite and analyze at the end of theyear. In 1991, EPA proposed a holdingtime of 6 months for gross alpha.However, neither the annual compositeunder the current rule or the proposedholding time of 6 months canappropriately capture the presence ofalpha-emitting radium-224, or itsprogeny in a gross alpha analysis. TheAgency intends therefore, to issue aseparate proposal to change the holdingtime for gross alpha analysis to accountfor the presence of radium-224 in thesample.

At this point in time, the Agencystrongly recommends to States andutilities that an alpha analysis beperformed within 48 to 72 hour aftersample collection to capture thecontribution of the alpha particlesarising from radium-224. In this NODA,the Agency is reiterating andunderscoring its recommendation tothat effect as outlined in a memorandumof January 27, 1999 from CynthiaDougherty, Director of the Office ofGround Water and Drinking Water (EPA1999a). For systems to whom a rapidanalysis might be a burden, a reasonablescreening tool for the presence of Ra-224under many geochemical circumstancesis the presence of its radiologicalancestor, Ra-228. Since systems willmonitor for Ra-228, the result can serveas a general proxy for the presence or

Ra-224 for the purposes ofprioritization. It is not definitive andwould not be an acceptable substitutefor a rapid analysis of gross alpha or Ra-224. In the absence of Ra-228, a systemmay not need to place as high a priorityon rapid gross alpha or specific Ra-224analysis. Since, as explained earlier,each Ra-224 atom contributesapproximately three daughter alphaparticles to the gross alpha count, asimple first approximation of Ra-224’scontribution to gross alpha would bethree times the Ra-228 concentration inpCi/L. For the purposes of prioritizingmonitoring for Ra-224, grandfatheredgross alpha data added to three timesthe result of the Ra-228 measurementwould be a reasonable firstapproximation of the gross alphaincluding Ra-224 and its daughtersavailable from a rapid gross alpha test.However, EPA reiterates that thisapproximation is not a substitute forrapid analysis of gross alpha or Ra-224.

EPA is not considering requiring aseparate MCL or analysis for radium-224when the rule is finalized in Novemberof 2000. The definition of gross alphawill continue to include Ra-224. EPA iswilling to consider comments on theneed to apply sub-limits to Po-210 orRa-224 within the MCL of 15 or asseparate standards. Proposing a separatelimit for radium-224 at 10 pCi/L withinthe alpha MCL of 15 pCi/L is a futurepossibility, as is a separate MCL forradium-224. The latter would require aseparate, specific, rapid analysisspecifically for radium-224, rather thanrelying on the gross alpha test and alphaMCL. Such actions would require a newproposal or proposals.

As part of the alpha standard, theAgency does not consider Ra-224 asignificant risk. The lifetime mortalityrisk associated with exposure to 10 pCi/L of radium-224 is approximately 5 ×10¥5 or one in 20,000. Because radium-224 and its progeny have very shorthalf-lives, the total alpha countrepresents the radium-224 and itsprogeny. Consequently, there areeffectively three alpha particle countsfor every atom of radium-224 present.The health risk of radium-224 alreadyincludes the impact of these progeny inthe body (the committed dose).Therefore while the gross alpha countmay be at 15, the impact of theemissions is approximately related toRa-224 at 5 pCi/L and the risk of 2.5 ×10¥5 or excess mortality of one in40,000.

H. UraniumUranium is not currently regulated by

the 1976 radionuclides drinking waterstandards. The 1986 SDWA

Amendments included uranium as oneof the 83 contaminants listed to beregulated in drinking water. Two healtheffects are associated with exposure touranium: cancer, resulting from theradioactive emissions, and kidneytoxicity, resulting from the exposure tothe uranium itself. The mass of theuranium is measured in micrograms (µg)while the radiation activity is measuredin picoCuries. In 1991, EPA proposed alimit on uranium of 20 µg per liter (µg/L) to protect against kidney toxicity. Thecorresponding radioactivity limit wasassumed to be 30 pCi/L. At that time,the Agency also proposed an MCLG ofzero, based on absence of an identifiabledose-response threshold. EPA hasreevaluated both the health impact levelfor kidney toxicity and the cancer risksfrom radiation and costs of regulation.As discussed briefly next, the bestestimate of the cost per cancer case orcancer death avoided at 20 µg/L isrelatively large. However, it should alsobe noted that this cost per case avoidedexcludes the reduction in kidneytoxicity risk. At the present time, kidneytoxicity for uranium must be treated asa non-quantifiable benefit (see appendixII, ‘‘Health Effects’’ and the TechnicalSupport Document, EPA 2000a).

Today’s NODA presents newinformation which supports a regulatorylevel of 20 µg/L, based upon protectionfrom kidney toxicity. The derivation ofthis number is based on newer, morecomplete studies which have alsoresulted in a lower uncertainty factor,now 100-fold. In addition, thecontribution to ingestion from drinkingwater relative to food or inhalation, therelative source contribution (RSC), hasbeen recalculated. Drinking water isnow considered to contribute 80 percentof a person’s total daily uranium intake.This has the effect of permitting 80% ofthe reference dose (RfD) to be occupiedby the drinking water component ofdiet. Both a lower uncertainty factorcoupled with a lower food intake andhigher proportional contribution fromdrinking water to total intake, mightsuggest the allowance of a higherregulatory limit; however, the morerecent studies have offset this byrevealing a lower observed effect levelfor kidney toxicity. The recalculated‘‘safe level’’ for kidney toxicity remains20 ug/L. The derivation of theuncertainty factor is based on the typesof uranium health data available. EPA’spolicy for uncertainty factors forestimating LOAELs is summarized in 63FR 43756 (August 14, 1998, ‘‘DraftWater Quality Criteria MethodologyRevisions: Human Health’’). Thederivation is described in appendix II.



Uranium is also classified as acarcinogen because of its radioactivity,and resulting emissions of ionizingradiation. The two most prevalentisotopes of uranium, uranium-234 anduranium-238, have very different half-lives which result in different amountsof radiation emitted per unit mass.Uranium-234 emits far moreradioactivity than U–238, but is muchless abundant in aquifer materials.Uranium-238 emits less radioactivity,but is far more prevalent than U–234.The average ratio of uranium activity tomass in rock is 0.68 picoCuries per µg.Issues involving the activity to massration follow later in this section.

Complicating the Agency’s decisionmaking about a uranium standard is thefact that the monetized benefit of kidneytoxicity cannot be calculated at lowconcentrations because data are lackingin terms of the level at which kidneydisease is actually manifested. Thecalculated 20 µg/L level represents anintake which would result in no effectover 70 years by drinking two liters perday. Conclusions based on the toxicityof uranium to the kidney are basedprimarily on observed adverse effects atthe cellular level, but which have notnecessarily resulted in a recognizeddisease. It is difficult to monetize thebenefits derived in such a situation, andEPA does not currently have amethodology for estimating benefits forkidney toxicity from uranium. In thecase of reducing the risk of non-fatalcancer resulting from uranium, EPA canmonetize these benefits based onavoided ‘‘cost of illness.’’ Thismethodology is discussed in some detailin the Technical Support Document(EPA 2000a) and elsewhere (EPA2000b).

Thus, for kidney toxicity, the benefitto society are considered as ‘‘non-quantifiable benefits.’’ Kidney toxicityavoidance benefits can be expressed interms of ‘‘avoidance of exposure,’’ butcannot be quantified in terms ofavoidance of a specified number ofcases of disease or fatalities (and theassociated monetized benefits), as withcancer. In addition, it appears thatexcess uranium concentrations tend tobe found in small water systems. Thissuggests that while many systems willbe impacted, the affected populationswill be small. In terms of cancer risk,the number of statistical cases avoidedfor MCLs of 20 and 40 µg/L are low (0.2to 2 cases for 20 µg/L and 0.04 to 1.5cases for 40 µg/L). In terms of exposureavoided for kidney toxicity, around 500thousand to two million persons areexposed above 20 µg/L and 50 thousandto 900 thousand persons are exposedabove 40 µg/L. See appendix V and the

Technical Support Document (EPA2000a) for details.

Although uranium is treatable tolevels well below the 1991 proposedMCL of 20 µg/L (5 pCi/L was evaluated),EPA determined that levels below 20pCi/l were not feasible under theSDWA, after taking the costs oftreatment into consideration. Section1412(b)(6) of the 1996 SDWA permitsthe Agency to evaluate whether thebenefits of regulating at various MCLsjustify the costs. Possible exercise of thisauthority is discussed in more detaillater in this section.

The MCLG that was proposed foruranium in 1991 was zero because ofconcerns about the lack of a knownthreshold for the carcinogenicity ofionizing radiation. The MCL that wasproposed in 1991 (20 µg/L) was basedon uranium kidney toxicity, aspreviously described. Thecorresponding risk of cancer at aconcentration of 20 µg/l is nowestimated to be approximately 5 × 10¥5.In terms of the cost per cancer caseavoided and kidney toxicity reduced,the cost of regulation is still relativelyhigh (see Table V–2 in appendix V).

In its current benefit-cost analysis,EPA also evaluated regulatory options ofuranium MCLs of 40 µg/L and 80 µg/L.EPA estimates that a level of 40 µg/Lwould correspond to a cancer risk ofapproximately a 1 × 10¥4, thusproviding cancer risk protection withinthe Agency’s traditional risk range. Alevel of 40 µg/L would represent aslightly higher risk of kidney toxicity.At a level of 80 µg/L, the cancermortality risk is approximately 2 ×10¥4, which is above the Agency’sacceptable risk range. At 80 µg/L, theprojected total national costs decreasesignificantly, but the estimates of cancercases avoided drops to values close tozero (i.e., benefits diminishconsiderably), indicating that the costper cancer case avoided may not besignficantly lower at an MCL of 80 µg/L than at an MCL of 40 µg/L. From ahealth effects perspective, the toxichealth effects on the kidneys or otherorgans or systems in the body atexposure levels of 80 µg/L is unknownand is four times EPA’s best estimate ofthe ‘‘safe level’’ with respect to kidneytoxicity.

In terms of benefits and costs, TableV–2 (appendix V) shows the range ofcompliance costs and net benefits forthe uranium MCL options of 20 µg/L, 40µg/L, and 80 µg/L. While annualcompliance costs drop significantly asthe MCL increases from 20 up to 80, theestimate of cancer cases avoided dropsconsiderably also. In fact, it is not clearwhether the cost per case avoided

increases or decreases with increasingMCL because of the uncertaintiesinvolved. The corresponding estimate ofcases avoided for MCLs of 20, 40, and80 pCi/L are 2.1, 1.5, and 1.0 casesannually. Based solely on cancerincidence, it may be appropriate forEPA to consider using an MCL higherthan 20 µg/L for uranium, since it isarguable that the benefits do not justifythe costs at this level. However, in termsof kidney toxicity, 20 µg/L may bejustified. EPA solicits comment on thisissue.

Health effects from uranium also needto be evaluated in the context of theeffects of various uranium species andtheir activity levels. A mortality risklevel of 5×10¥5 translates to 23 pCi/L ofU–238, 22 pCi/L of U–235, and 21 pCi/L of U–234 in drinking water. An ‘‘alphaspec’’ analysis of the water woulddetermine the fractions of each presentand a sum of the fractions below 100%would meet the MCL. However, thislevel is costly to obtain. Doubling theradioactivity limit to 46, 44 and 42 pCi/L for U–238, U–235, and U–234respectively corresponds to a mortalityrisk level of 1 × 10–4, which may bemore acceptable, considering the costs.Likewise, a doubling of risk to 2 × 10–4

would again double the picoCurie limitsof each isotope to 92, 88, and 84 pCi/L respectively. However, at these higherrisk levels, the calculated protectivelimit for toxicity to the kidney may beexceeded, depending upon theuncertainty factor used.

By contrast, the relative dissolvedconcentration of the various isotopes ofuranium will differ markedly from onelocale to another. The 1991 proposalutilized a conversion factor of 1.3picoCuries per microgram of uranium toconvert a 20 µg/L proposed MCL inmass units to activity units inpicoCuries (however, 1991 costestimates were based on the moreaccurate conversion ratio of 0.9).Analysis of NIRS data suggest that itwould have been more appropriate touse the 1.3 pCi/µg conversion factor fortotal uranium where concentrations areless than 3.5 pCi/L and a 0.9 conversionfactor for concentrations above 3.5 pCi/L (Telofsky 1999). Converting thederived MCL option of 20 µg/L frommass to activity using a ratio of 0.9 forlevels above 3.5 µg/L yieldsapproximately 18 pCi/L . A statisticalevaluation of uranium data reveals that,based on a linear regression of the data,the appropriate activity based MCL for20 µg/L would be 17.3 pCi/L rounded to17 pCi/L. Coupled with the knowledgethat the concentration of uraniumisotopes varies from place to place, theAgency is led to consider an MCL that



2 Throughout this discussion, exposures and riskswere only considered for populations potentiallyaddressable by regulation, i.e. systems withradionuclides present in excess of the proposedMCLs for community water systems.

3 It is important to remember that the riskassessment for NTNC water systems does notconsider exposure risk from private wells whichmay serve some customers at home. EPA recognizesthat the radionuclide levels in some private wellsmay exceed the MCLs for CWSs, but this is a non-controllable factor since private wells are notregulated by the Safe Drinking Water Act.

is protective in any location against bothtoxicity (µg/L) and cancer (pCi/L),whichever presents the greatest risk.This can be determined by conductingisotopic analysis to determine therelative amounts of each isotope in anyone water system. Once theconcentration ratio is known, aregulated entity may choose to measuremass or activity and select whicheveranalytical method or methods is mostcost effective.

For example, if the uranium standardwere 20 µg/L or pCi/L, a gross alphameasurement screen for uranium couldbe used in the following way (EPA2000c and 2000d). The analysis breaksout as follows: if the result is below thedetection limit for gross alpha, neitheruranium measurements by mass oractivity would be necessary sinceneither 20 pCi/L nor 20 µg/L could beexceeded. If the gross alpha test isbetween 3 and 5.5 pCi/L, the mass of 20µg/L could be exceeded if all the activitywere coming from uranium-238.Therefore a fluroimetric test for uraniummass concentration (µg/L) or an alphaspectrometry test for the activities (pCi/L, converted to µg/L using standardisotopic conversion factors) of thevarious isotopes present would benecessary to determine the uraniumconcentration in µg/L. Because grossalpha tests may underestimate uraniumby a factor of as much as 3.62, if thegross alpha test exceeded 5.5 pCi/L(20÷3.62), it is indicative that the 20pCi/L limit may be exceeded, and anisotopic analysis must be done. EPAsolicits comment on these issues.

EPA is soliciting information andcomment on the data and theappropriate course of action the Agencyshould pursue, given the factors of risklevels, national cost, number of cancersavoided, cost per case, cost per death,and kidney toxicity. EPA is currentlyevaluating three regulatory options:

• Regulate at 20 µg/L and 20 pCi/L(protective of kidney toxicity using theAgency standard 100-fold uncertaintyfactor for this type of LOAEL with anassociated cancer risk of approximately5 × 10¥5 or five in one hundredthousand);

• Regulate at 40 µg/L and 40 pCi/L(this is twice the safe level with respectto kidney toxicity and would reduce themargin of exposure between the effectlevel and the proposed regulatorystandard; with an associated cancer risklevel of 1 × 10¥4 or one in ten thousand,which is the Agency’s usual uppercancer risk target);

• Regulate at 80 µg/L and 80 pCi/L(this is four times the safe level withrespect to kidney toxicity and wouldfurther reduce the margin of exposure

between the effect level and theproposed regulatory standard; with anassociated cancer risk level of 2 × 10¥4

or two in ten thousand, which is abovethe Agency’s usual upper cancer risktarget).

In summary, EPA believes that 20 µg/L is feasible and is the Agency’spreferred option, but may not havebenefits that justify the costs. Were ahigher level to be chosen, EPA would beexercising its discretionary authorityunder section 1412(b)(6) to select a levelabove the feasible level. It should benoted, however, that there may beconsiderable non-quantifiable benefitsof avoiding exposure to cancer andkidney toxicity. Also, as discussedpreviously, there is little available dataor information about the effects ofkidney toxicity at relatively highexposures and thus, the benefitsattributable to avoided illness cannot bequantified. Thus, the costs may bejustified at a more stringent level thanwould be suggested in light of thecurrently quantifiable benefits alone. Inaddition, the Agency generally does notestablish regulatory levels outside of itstarget risk range and, in fact, prefers toset levels at the more protective end ofthat range (1 × 10¥6), wherever possible.Further, we usually follow Agencyguidelines on use of uncertainty factors.For these reasons, the Agency does notfavor an MCL option of 80 µg/L, butsolicits comment on this and thepreviously-described regulatory options,together with any supporting rationaleor data commenters wish to provide.

I. Inclusion of Non-Transient Non-Community Water Systems

Today’s document is solicitingcomment on several approaches forcovering Non-Transient Non-Community (NTNC) water systems.Although current radionuclideregulations do not apply to NTNC watersystems, in 1991 EPA proposedextending the radionuclides NPDWRs toinclude them. Several approachesrepresenting varying degrees of controlare being currently considered forfinalization because, although muchmore has been learned about NTNCwater systems and their customers since1991, there is still very little knownabout the distribution of the highestlevels of radionuclides in their watersupplies. Based on the Agency’soccurrence estimates, control of someradionuclides in NTNC water systemsmay not present a meaningfulopportunity for health risk reduction.This issue arises as a consequence of the1996 Amendments to SDWA whichallow the Agency to consider whetherthe benefits of extending coverage to

this category of water systems wouldjustify the costs (section 1412(b)(6)(A))and whether such regulation wouldprovide a meaningful opportunity forhealth risk reduction (section1412(b)(1)(A)(iii)). The TechnicalSupport Document (EPA 2000a)presents a ‘‘what if’’ analysis for costsand benefits for NTNCWSs.

While it is feasible to controlradionuclides in NTNC water systems,extending regulation to these systemsneeds to be considered in light of thenew SDWA requirements. This analysisrequires a balancing of both quantitativeand non-quantitative factors. Based onthe risk modeling discussed in theTechnical Support Document (EPA2000a), the ninetieth (90th) percentilelifetime risk of cancer incidence in anindividual consuming water from aNTNC water system in the absence of aregulation is not expected to exceedthree in 100,000 2. The cost per cancercase avoided to achieve reductions inthese risks would considerably exceedthe hundred million dollar mark ifcoverage of the rule were extended toNTNCs. The associated cost per caseavoided ranges are well above the rangeof historical environmental riskmanagement decisions.

Relative to community water systems,NTNC systems have much lowerassociated risk levels because mostindividuals served by these systems areexpected to receive only a small portionof their lifetime drinking water exposurefrom this source 3. This conclusionholds even using very conservativeassumptions for modeling the NTNCexposure scenarios. For example, in thecase of school children exposure, theAgency has conservatively assumed allimpacted children would attend onlyschools served by NTNC water systems,have twelve years of perfect attendance,and get half of their daily waterconsumption at school. For the averagethirteen year old, this scenario implieshalf of a liter (over sixteen ounces) everyschool day. Even under this veryconservative set of assumptions, thewater consumed by an individualstudent is estimated to represent less



4 Day care exposure is similarly conservativelyestimated by assuming five years of perfectattendance, fifty weeks per year and five days perweek. Factory workers are assumed to perfectlyattend and work at the same facility for forty-fiveyears. All of these assumptions are undercontinuing investigation and will likely be reviseddownward in the future as the Agency is able togather further information.

5 As an example, the lifetime risk per pCi/L of Ra–228 to a child whose exposure begins under the ageof five is more than ten times greater than thelifetime risk of an individual whose exposurebegins between the ages of 25 and 30.

6 For example, the possibility that a child spentfive years in a day care center, then twelve yearsin schools, and then forty-five years working in a

factory served only by NTNC water systems withhigh radionuclide levels.

7 In other words, the expected number of NTNCsystems nationwide would be less than twenty. Itis because these levels are so rare that the level isfairly speculative. As discussed in the appendix,the Agency believes its estimates of occurrence arereasonable, based on levels observed in smallground water community water systems.

than five percent of lifetimeconsumption 4.

On the other hand, much remains tobe learned about the NTNC watersystems. Little is known about theextent to which users of the differentNTNC water systems use other watersystems. It is conceivable that someareas in the country exist whereindividuals are subjected to exposure ata number of different non-communitysystems (e.g., day care center plusschool plus factory, etc.). In suchcircumstances, individuals would beexposed to proportionately higher risksif the water systems all had elevatedlevels. For some individuals, theexposures could approach levelsobserved in corresponding communitywater systems.

This concern is somewhat alleviatedby the fact that NTNC systems generallyserve only a very small portion of thetotal population. For example, over

ninety-five percent (95%) of all schoolchildren are served by community watersystems, not NTNC systems. Only asmall percentage of children are servedby NTNC water systems and, of thatgroup, less than one percent (or lessthan one in 2000 of the overall studentpopulation) would be expected to haveindividual radionuclides in their waterabove the proposed regulatory levels.Likewise, less than 0.1 percent of thework force population receive waterfrom an NTNC water system. With suchlow portions of the total populationexposed to any particular type of NTNCsystem, the overall likelihood ofmultiple exposure cases in the NTNCpopulation should also be small.

Nevertheless, because children aremore sensitive to radionuclidesexposure 5, multiple water systemexposure scenarios were considered inthe modeling effort 6. Tables III–2 andIII–3 present individual risk estimates

for average and most sensitivepopulations among the NTNC watersystems. All of these factors contributedto the Agency’s evaluation of whether ornot to extend regulation to NTNC watersystems and are discussed further in theappendix.

Review of Table III–3 shows that 90thpercentile individual risk patterns forNTNC water system users exposed touranium or radium-226 are relativelylow. These 90th percentile figuresrepresent risks estimated using thepreviously described conservativeexposure scenarios, maximum waterconsumption patterns, and what areeffectively 99.9th percentile occurrenceestimates 7 from the NIRS data. Evenwith these conservative factors, lifetimecancer risks do not exceed the one in10,000 level which has traditionallyformed the upper bound of allowablerisk in Agency decision-making.

TABLE III–2.—SELECTED SECTOR AND OVERALL NTNC, INDIVIDUAL RISK PATTERNS

[Lifetime cancer risk for individuals using average consumption levels]

Sector Alpha Radium 226 Radium 228 Uranium

School Students ................................................................. 2×10¥5 0.9–1.1 ×10¥5 2–3×10¥5 0.7–0.9×10¥5

Day Care Children ............................................................. 2–3×10¥5 0.6–0.7×10¥5 2–3×10¥5 0.8–1x10¥5

Factory Worker .................................................................. 1–2×10¥5 1×10¥5 2–3×10¥5 1×10¥5

All NTNC Water Systems .................................................. 0.3–0.4×10¥5 0.2–0.3×10¥5 0.5–0.7×10¥5 0.2×10¥5

Note that Radium 224 is being used as a surrogate for alpha emitters.

TABLE III–3.—SELECTED SECTOR AND OVERALL NTNC, INDIVIDUAL RISK PATTERNS

[Lifetime cancer risk for individuals using 90th percentile consumption levels]

Sector Alpha Radium 226 Radium 228 Uranium

School Students ................................................................. 0.5–0.6×10¥4 0.3×10¥4 0.6–0.7×10¥4 0.3×10¥4

Day Care Children ............................................................. 0.7–0.8×10¥4 0.2×10¥4 0.6×10¥4 0.3–0.4×10¥4

Factory Worker .................................................................. 0.5–0.6×10¥4 0.3–0.4×10¥4 0.7–0.8×10¥4 0.5–0.6×10¥4

All NTNC Water Systems .................................................. 0.2×10¥4 0.1×10¥4 0.2–0.3×10¥4 0.1–0.2×10¥4

Radium-228 and gross alpha poseapproximately twice the threat of theother two radionuclides. While sensitiveindividual estimates still fall below theone in ten thousand range, they may notin a scenario in which other drinkingwater sources are similarly high.However, as stated previously, theAgency views it as somewhatimprobable that this system overlapoccurs to a significant extent.Nevertheless, it could be an issue insome rural communities. While such

infrequent and highly site-specificconditions are very difficult to addressefficiently in a National-level regulation,the Agency believes that exemptingNTNC water systems from theradionuclide NPDWRs, given the degreeuncertainty about the occurrence levelsand extent of system customer overlap,may be inappropriate. For these reasons,the Agency believes it may beappropriate to take a somewhat differentapproach with respect to NTNC watersystems than previously practiced.

EPA is considering extending partialcoverage of the radionuclide NPDWRsto NTNC water systems under severalpossible scenarios. Under the first threeoptions, NTNC systems would besubject to targeted radionuclidemonitoring requirements, in whichselected NTNC systems would followthe radionuclides monitoringrequirements for community watersystems. The targeting strategy would bebased on small community water systemoccurrence for the same radionuclides.



The States (or primacy agency) woulddetermine which NTNC systems arelikely to be using contaminated watersystems, based on CWS monitoringresults. These systems may then berequired to monitor and meet CWSMCLs for gross alpha and combinedradium and other relevantradionuclides. EPA is considering:

• Requiring targeted NTNC systems tomonitor and meet the CWS MCLs for allor selected radionuclides, wheretargeting is determined by the Statebased on whether the NTNC system isusing source water for which CWSshave reported MCL violations forradionuclide in question;

• Requiring targeted NTNC systems tomonitor and post notice if the systemexceeds the CWS MCL, using the samedefinition of targeting as in the firstoption;

• Issuing guidance that recommendsthat targeted NTNC systems monitorand meet the CWS MCLs, using thesame definition of targeting as in thefirst option.

The Agency requests comments onthese options and any supportingrationale for such a decision. TheAgency is also interested in receivingcomments on other options such asextending full coverage of the rule toNTNCs and not extending any aspect ofthe radionuclides NPDWRs to NTNCsystems. The Agency will decide, aspart of the upcoming finalization of the1991 proposal, to incorporate what itconsiders to be the most appropriateoption in view of available the data andinformation.

J. Analytical MethodsToday’s NODA provides a brief

update of the methods-related itemswhich have occurred since the 1991proposed rule. For a more thoroughdiscussion of the analytical methodsupdates, the public is referred toappendix III of this NODA and to theAnalytical Methods section of theTechnical Support Document for theRadionuclides Notice of DataAvailability (EPA 2000a).

1. Radionuclides Methods UpdatesOn July 18, 1991 (56 FR 33050; EPA

1991), the Agency proposed to approvefifty-six methods for the measurement ofradionuclides in drinking water(excluding radon). Fifty-four of the fifty-six were actually approved in the March5, 1997 final methods rule (62 FR 10168;EPA 1997a). In addition to these fifty-four, EPA also approved 12radiochemical methods, which weresubmitted by commenters after the 1991proposed rule. Currently, an overalltotal of 89 radiochemical methods are

approved for compliance monitoring ofradionuclides in drinking water. Thesemethods are currently listed in 40 CFR141.25.

The March 5, 1997 Federal Registeralso approved suitable calibrationstandards for the analysis of gross alpha-emitting particles and gross beta-emitting particles. These specificmethods-related items are addressed insome detail in the Technical SupportDocument for the Radionuclides Noticeof Data Availability (EPA 2000a) and ineven greater detail in the 1997 finalmethods rule (62 FR 10168, EPA 1997a)and the 1991 proposed rule (56 FR33050; EPA 1991).

This NODA also notifies the publicabout the use of the gross beta methodfor the screening of radium-228. In the1991 proposed rule (56 FR 33050; EPA1991), the Agency would have allowedthe use of the gross beta-particle activitymethod to screen for the presence ofradium-228 at the proposed radium-228MCL of 20 pCi/L. For the combinedradium-226 and 228 standard of 5 pCi/L (the current standard), the Agency cannot recommend the use of the grossbeta-particle activity method forscreening of radium-228. Instead, aspecific analysis for radium-228 wouldbe necessary. Although several methodsare currently approved for the analysisof radium-228 in drinking water, theAgency requests comments from thepublic and supporting documentationregarding other radium-228 methods ormethod variations which may be able toreach greater sensitivity at the 2 pCi/Llevel.

2. The Updated 1997 LaboratoryCertification Manual

In the 1991 proposed rule (56 FR33050; EPA 1991), EPA cited the 1990laboratory certification manual’sguidance for sample handling,preservation, holding time andinstrumentation. In response to the 1991proposed rule, a commenter questionedwhy the holding time for radioactiveiodine was six months, when the half-life of iodine-131 is eight days. TheAgency recognized this typographicalerror and changed the holding time toeight days in the updated 1997certification manual (EPA 815–B–97–001; EPA 1997d). Table III–2 in theappendix shows the updated guidancefor sample handling, preservation,holding times, and instrumentation thatappeared in this manual. Table III–2 inthe appendix also includes additionalrecommendations for radiochemicalinstrumentation (footnoted by thenumber 6). The Agency is seekingcomment about the additionalrecommendations found in Table III–2.

3. Recommendations for Determiningthe Presence of Radium-224

To determine the presence of theshort-lived radium-224 isotope (half life∼3.66 days), the Agency recommendsusing one of the several optionsdiscussed in the appendix III. Althoughthese measurement options are onlyrecommendations, the Agency stronglyurges water systems to check for thepresence of radium-224 in theirdrinking water supplies. Comments aresolicited from the public about theoptions listed in appendix III or anyother appropriate methods of detection.

4. Cost for Radiochemical Analysis

Revised Cost Estimates forRadiochemical Analysis.

In the 1991 proposed rule (56 FR33050; EPA 1991), EPA cited costestimates for radiochemical analyses.The Agency updated these costsestimates by surveying a small numberof radiochemical laboratories (no morethan 9 laboratories) (EPA 2000a). Therevised cost estimates are shown inTable III–3 (appendix III). Because thisinformation is based on a limitednumber of laboratories, the slightincrease in costs from 1991 to 1999 maybe due to either statistical uncertainty orpossibly others factors such as inflation.

After the 1991 proposed rule, therewere several comments regardinganalytical costs. One commenter statedthe costs of analysis for radium-226,radium-228, radioactive strontium andtotal strontium were unrealistically low.The Agency can neither agree nordisagree. As noted earlier, EPA revisedthe cost estimates for radiochemicalanalysis. Both the 1991 costs estimatesand the revised cost estimates were fromsmall surveys and may not be trulyrepresentative of the actual costs forsome radiochemical analyses.Comparison of the estimated costs from1991 with the revised cost estimatesindicate the costs for some analyses tosimilar, while for other analyses, cost doappear to be higher. The Agency solicitscomments and factual data that wouldclarify this matter.

Several commenters stated that smallsystems, which are likely to need onlya few analyses, cannot take advantage ofrates for volume sample analyses. TheAgency agrees that individual smallsystems may not be able to takeadvantage of lower bulk analysis costs.To alleviate cost burdens, small systemsmay want to consider pooling theiranalytical needs with other smallsystems to negotiate for bulk rates.



5. Externalization of the PerformanceEvaluation Program

Due to resource limitations, on July18, 1996 (61 FR 37464; EPA 1996b),EPA proposed options for theexternalization of the PE studiesprogram (now referred to as theProficiency Testing or PT program).After evaluating public comment, in theJune 12, 1997 final notice EPA (62 FR32112; EPA 1997b):decided on a program where EPA wouldissue standards for the operation of theprogram, the National Institute of Standardsand Technology (NIST) would developstandards for private sector PE (PT) suppliersand would evaluate and accredit PEsuppliers, and the private sector woulddevelop and manufacture PE (PT) materialsand conduct PE (PT) studies. In addition, aspart of the program, the PE (PT) providerswould report the results of the studies to thestudy participants and to those organizationsthat have responsibility for administeringprograms supported by the studies.

EPA has addressed this topic inpublic stakeholders meetings and insome recent publications. For moreinformation, readers are referred to theaforementioned Federal Registernotices. More information aboutlaboratory certification and PT (PE)externalization can be accessed at theOGWDW laboratory certificationwebsite under the drinking waterstandards heading (www.epa.gov/safewater). At this time, it is difficult toascertain how and if externalization ofthe PT program will affectradiochemical laboratory capacity andthe cost of radiochemical analyses. Inthe absence of definitive cost estimates,the Agency solicits public comments onthis subject.

6. The Detection Limits as the RequiredMeasures of Sensitivity

In 1976, the National PrimaryDrinking Water Regulations defined thedetection limit (DL) as ‘‘theconcentration which can be countedwith a precision of plus or minus 100percent at the 95 percent confidencelevel (1.96 σ, where σ is the standarddeviation of the net counting rate of thesample).’’ Table III–4 in the appendixcites the detection limits or the requiredsensitivity for the specific radioanalysesthat were listed in the 1976 rule and arealso cited in 40 CFR 141.25. In the 1991proposal (56 FR 33050; EPA 1991), EPAproposed using the method detectionlimit (MDL) and the practicalquantitation level (PQL) as measures ofperformance for specific radioanalyticalmethods. Acceptance limits based onthe PQLs, which were derived fromperformance evaluation studies, werealso proposed in the 1991 rule. Some

commenters found the use of acceptancelimits confusing and the relationship tothe actual method performance was notclear. With perhaps the exception ofuranium, the Agency will not goforward with the proposed acceptancelimits, PQL, or MDL. Because uraniumhas never been regulated, it did not havea detection limit in the CFR and one hasnever been proposed. In 1991, EPA didpropose a PQL of 5 pCi/L with anacceptance limit of +/¥ 30%. Althoughit is believed that a detection limit foruranium would be very similar to thePQL, because a detection limit has neverbeen proposed, the Agency may have toadopt the PQL for uranium until adetection limit is proposed. For theother radionuclides, which areregulated, the Agency believes thecurrent 1976 detection limitrequirements are most appropriate. Theexisting definition of the detection limittakes into account the influence ofvarious factors (efficiency, volume,recovery yield, background, countingtime) that typically vary from sample tosample. Furthermore, the detection limitis computed for each individual sampleand does not represent an idealized setof measurement parameters. Therefore,the detection limit reflects the expectedrandom uncertainty for a given sampleanalysis.

7. Performance Based MeasurementSystem

On October 6, 1997, EPA published aNotice of the Agency’s intent toimplement a Performance BasedMeasurement System (PBMS) in all ofits programs to the extent feasible (62FR 52098; EPA 1997c). EPA is currentlydetermining how to adopt PBMS into itsdrinking water program, but has not yetmade final decisions. When PBMS isadopted into the drinking waterprogram, its intended purpose will be toincrease flexibility in laboratories inselecting suitable analytical methods forcompliance monitoring, significantlyreducing the need for prior EPAapproval of drinking water analyticalmethods. Under PBMS, EPA willmodify the regulations that requireexclusive use of Agency-approvedmethods for compliance monitoring ofregulated contaminants in drinkingwater regulatory programs. EPA willprobably specify ‘‘performancestandards’’ for methods, which theAgency would derive from the existingapproved methods and supportingdocumentation. A laboratory would befree to use any method or methodvariant for compliance monitoring thatperformed acceptably according to thesecriteria. EPA is currently evaluatingwhich relevant performance

characteristics under PBMS should bespecified to ensure adequate dataquality for drinking water compliancepurposes. After PBMS is implemented,EPA may continue to approve andpublish compliance methods forlaboratories that choose not to usePBMS. After EPA makes finaldeterminations about theimplementation of PBMS in programsunder the Safe Drinking Water Act, theAgency would then provide specificinstruction on the specifiedperformance criteria and how thesecriteria would be used by laboratoriesfor compliance monitoring of SDWAanalytes.

K. Monitoring

1. Features of Today’s NODAAEPA’s 1976 regulations for

radionuclides in drinking watercontained separate monitoringrequirements for radiums, alphaemitters, and man-made beta andphoton emitters. In 1991, EPA proposedto make modifications to the 1976regulations to expand the scope ofcoverage to include non-transient non-community water systems, to change themonitoring location and monitoringfrequencies, and to incorporatemonitoring requirements for radon anduranium. A summary of the 1976requirements and proposed changes in1991 are presented in this section.

In today’s document EPA issuggesting merging the currentrequirements and the 1991 proposedrequirements into a unified systemwhich is consistent with theStandardized Monitoring Framework(SMF), the current rule, and theproposed changes which are stillgermane. EPA is soliciting comment onmonitoring at the entry points to thedistribution system, as proposed in1991, to ensure equal protection for allcustomers, under the sampling scheduleof the SMF. EPA believes that this willincrease consistency betweenmonitoring requirements forradionuclides and the other regulatedcontaminants. As described in sectionIII, part I (‘‘Inclusion of Non-TransientNon-Community Water Systems’’), EPAis considering several options for NTNCwater systems, some of which wouldrequire monitoring. Because somemonitoring provisions of the 1991proposal were based on the proposedMCLs and not the current MCLs, theirapplication to the current levels mayentail a slightly different construct thanin 1991. To the extent comments revealaspects of the framework which need tobe addressed separately from the 1976rule or 1991 proposal, EPA will return



to the current rule’s framework andpropose to correct deficiencies via aproposal which will address analyticalmethod issues as well, such as methodsfor Ra-224, Po-210 and Pb-210.

2. Standardized Monitoring FrameworkPer the current rule, once the

contaminant concentration in the wateris established by the average results forfour consecutive quarterly samples or bysuitable grandfathered data, a systemwould be categorized as to whether itwas above or below 50% of the MCL forthat contaminant. In accordance withthe SMF, as proposed in 1991, EPA issuggesting a tiered frequency for alphaemitters, combined radium, anduranium. This would entail one sampleevery three years for compliant systemswith annual average contaminant levelsabove 50% of the MCL. For compliantsystems with annual average levelsbelow 50% of the MCL for thesecontaminants, one sample would berequired every 6 years; non-detects, onesample every 9 years. EPA believes thissystem would align with thestandardized monitoring framework,and would provide regulatory relief forsystems with low to very low levels(without needing a waiver as called forin 1991). It would also provide morecareful screening for systems withmultiple sources of water entering thedistribution system, by requiring asample at each of these points to beprotective of all of the customers withineach water system. For beta particle andphoton radioactivity, EPA is consideringrequiring four consecutive quartersevery four years, the requirement underthe current rule, for vulnerable systemsbecause of their proximity tocontamination sources.

EPA believes this monitoring schemeis less burdensome on systems in thelong term than either the existing orproposed regulations. It providesslightly more protection than thecurrent rules by more frequentmonitoring for contaminants above halfthe MCL, and less frequent monitoringfor the vast majority of systems belowhalf the MCL. EPA believes this is morerealistic and less burdensome, whilerecognizing the potential for variabilityof naturally occurring radionuclidelevels in ground water over time. Suchvariability (e.g., a change in pH bynitrogen fertilizer application leading toa higher solubility of radium) was seenin New Jersey and is further discussedin appendix I.

Small ground water systems comprisethe vast majority of systems withradionuclide contamination problems.Since most small systems have only oneentry point, an entry point monitoring

requirement will not have an impact.For systems with radium above 50% ofthe MCL, with three or fewer entrypoints to the distribution system,monitoring at each entry point onceevery three years would have an equalor smaller impact (in terms of thenumber of samples analyzed) than the1976 requirement of monitoring fourtimes every four years.

3. Entry Point MonitoringEPA recognizes that sampling

conducted at the monitoring locationspecified in the current rule may under-represent the risk to some consumers.Results can vary depending on the usageof each water source and changes in themonitoring location within thedistribution systems. For systems withmore than one water source, monitoringwithin the distribution system mayyield different results. In the currentrule, sampling is conducted ‘‘at a freeflowing tap’’ within the distributionsystem. The current rule also recognizesthe potential problems by providing thatsystems with two or more sources ofwater with different concentrations ofradionuclides monitor the source water,as well as water from a free flowing tap,when ordered by the State. Entry pointmonitoring, a feature of more recentNPDWRs, provides a better measure ofwater quality for residents near the startof the distribution system thanmonitoring within the distributionsystem (e.g., the middle of the system)where water is subject to blending ifthere are other sources. Therefore, EPAproposed in 1991 to change the locationfor compliance monitoring to the entrypoints to the distribution system,consistent with other NPDWRs.

4. Grandfathering DataIn the implementation guidance,

which will be available on OGWDW’shome page (http://www.epa.gov/ogwdw), EPA is suggesting that sampleswithin the latest compliance period,beginning June, 1996, be eligible for usein determining the baseline formonitoring frequency. While EPAprefers this approach, others may bepossible. Please provide data andsupporting rationale if you comment onthis issue. The application of thisprovision would extend to all classes ofradionuclides for which data areavailable.

The Agency solicits comment on twodifferent approaches for the betamonitoring requirements. The firstoption is to not allow any reducedmonitoring and the second would be toallow reduced monitoring similar to thealpha emitters. If systems must collectsamples on a quarterly basis (no

reduced monitoring) thengrandfathering of data is not necessary.If the Agency decides to allow reducedmonitoring, the Agency believes thatStates may use historical data tosupplement their vulnerabilityassessments but should not usegrandfathered data to satisfy the initialmonitoring requirements because asufficient baseline needs to beestablished in those systems consideredvulnerable to man-made radioactivity.

Grandfathered data would be used tocomply with the initial monitoringrequirements for gross alpha, radium-226/228, and uranium, under somecircumstances. Data collected after June1996, during the most recentcompliance period, would beconsidered for grandfathering. It wouldbe the State’s responsibility todetermine if grandfathered data issufficient to satisfy the initialmonitoring requirements established bythis rule. At the State’s discretion,systems with one entry point to thedistribution system (EPTDS) could usegrandfathered data to satisfy the initialmonitoring requirements. Systems thathave multiple entry points to thedistribution system could usegrandfathered data collected after June1996 to satisfy the initial monitoringrequirements, provided that the datawere collected at the EPTDS.

EPA is also considering that, at theState’s discretion, systems with up tothree entry points to the distributionsystem could also use grandfathereddata to satisfy initial monitoringrequirements, even if not collected fromEPTDS, if the State makes a writtenfinding that the circumstances of thesystem and their review of historic datajustify such action. While the Agencycannot prescribe every possible scenariothat a State may encounter, an exampleof circumstances that might supportsuch a finding could be: a system thathas three wells (and EPTDS), that aresimply from different parts of a well-field, using the same aquifer, with goodhistorical data showing uniform, low tono radionuclide occurrence from allwells, perhaps from the raw water aswell as distribution system samples.

5. Sample CompositingIn general, compositing of samples is

an effective means of decreasinganalytical costs to systems. Compositingis permitted for alpha emitters and betaand photon emitters in the current rule.It is also allowed for radium-226 and-228 to the extent gross alpha was usedas a screen for Ra-226 and, in turn, Ra-228. In the 1991 proposal, gross betacompositing was prohibited.Compositing for other nuclides was



allowed for up to five sampling pointswithin one system; if the result for thecomposite was more than 3 pCi/l for anynuclide, individual non-compositedsamples were to be analyzed. Thisprovision stemmed from the lowestMCL in the 1991 proposal, adjustedgross alpha at 15 pCi/l. Because of thepossibility that one of the five samplestaken might be at 15 pCi/L even if theother four were at zero, the ruleenvisioned one fifth (3 pCi/L) of theMCL as the maximum allowed result toassure no single well could exceed theMCL. The principle of limiting theresult of five composited samples fromseparate entry points to one fifth of theMCL (or four composites to one fourththe MCL etc.) is still valid as a generalmatter, and should be followedwhenever compositing is done.

A 3 picoCurie limit in the proposalwould have been conservativelyprotective for a five sample compositedfor the proposed separate MCLs radium-226 and radium-228 at 20 pCi/L each,since 1⁄5 of each MCL is greater than 3pCi/L. However, because EPA isconsidering retaining the currentradium standard at 5 pCi/L combined,adding the results of five compositedentry points samples for Ra-228 to theresults of 5 composited entry pointsamples of Ra-226 must yield a result ofone tenth (1⁄10) of the MCL to be assuredthat the combined Ra-226 and Ra-228concentration could not exceed the MCLat any one entry point. Because onetenth of the MCL (0.5 pCi/L) is belowthe detection limit for Ra-226 and Ra-228, compositing of separate entrypoints cannot apply in case of Ra-226 orRa228. However, annual compositing ofsamples from the same entry point mayapply.

EPA requests comment on thefeasibility and practical utility ofcompositing separate entry points(spatial compositing) versuscompositing samples over time from thesame entry point (temporalcompositing). EPA believes that the useof one or the other (but never bothsimultaneously) may be appropriateunder some circumstances. Greatercertainty in the analytical result isobtained by taking the average of fourseparate (non-composited) results fromone sampling location than by using asingle result of composited samples.However, where an MCL is sufficientlyabove the detection limit such thatanalytical results are not subject tosignificant error near the MCL,compositing may be a cost savingmeasure. Additionally, when historicaldata indicate that contaminant levels arenegligible (e.g., non-detects) for a watersystem, compositing among wells in a

system or between systems having onepoint of entry may be advisable at Statediscretion. However, because of thecosts of re-sampling and re-analysis ofall points to confirm an MCL violation,or to qualify for decreased monitoring,it may not be in the systems best interestto initially composite in the absence ofhistorical data.

6. Increased and Decreased MonitoringAdditionally, the Agency is

considering having the final rule allowsystems that are currently on a reducedmonitoring schedule to remain on thatreduced schedule as long as the systemqualifies for reduced monitoring basedon the most current analytical result.Systems for which the most currentanalytical result indicates a higher levelthan allowed for that monitoringschedule would resume monitoring at afrequency consistent with the mostrecent result. For example, a systemwith an annual average below half of theMCL could reduce monitoring to onesample every 6 years. If, while on thisreduced frequency, the system collects asample with an analytical result abovehalf the MCL, the system would have toincrease monitoring again to once every3 years. It could revert to its previousreduced frequency of once every sixyears if the subsequent analytical result(of the sample taken three years later)was less than half the MCL. EPA alsobelieves it is prudent to requirequarterly samples to be collected at least60 days apart, to capture seasonalvariations. EPA solicits comment on thisand other monitoring provisions.

7. Compliance DeterminationsCompliance would be determined

based on the annual average of quarterlysamples collected at each entry point forall classes of radionuclides. If theannual average of any entry pointexceeds an MCL, the CWS would be inviolation. If NTNC systems are subjectto MCLs, the same situation wouldapply to them. An immediate violationwould occur for any sample analyticalresult or combination of sampleanalytical results that would place thesystem in violation before four quartersof data are collected (e.g., the firstsample is greater than 4 times the MCLor the average of the first two samplesis greater than twice the MCL). If asystem has a sample that exceeds theMCL while on reduced monitoring, itwould need to begin quarterlymonitoring the following quarter.Compliance would be based on theaverage of the four consecutive quartersof data beginning with the initial resultthat exceeded the MCL. If a system failsto collect all samples required during

any year, compliance would becalculated based on available data.Under the current rule, quarterlymonitoring is continued until theannual average concentration no longerexceeds the MCL or until a monitoringschedule as a condition to a variance,exemption, or enforcement actionbecomes effective.

The following is a summary of certainfeatures of the monitoring requirementsfor each regulated radionuclide orradionuclide group.

8. Combined Radium-226 and -228

Standardized monitoring: EPAcontemplates application of thestandardized 3, 6, 9 year cycles to thecombined radium standard dependingon whether analytical results forcompliant systems are greater than (3) orless than (6) half the MCL or are a non-detect (9), as previously discussed.Decreased and increased monitoringwould be based on the result of theanalysis of the most recent requiredsample(s).

Entry point monitoring: Monitoring atentry points to the distribution systemwould be a requirement per the 1991proposal unless EPA receives commentswith compelling reasons for not doingso.

Sample Compositing: To decrease theburden of monitoring at distributionentry points, EPA is contemplatingallowance of sample compositing forradium-226 or radium-228, but onlywhen results will be indicative of thetrue level at a single entry point(temporal compositing). According tothe proposal, systems would be requiredto analyze for Ra-228 separately fromgross alpha or Ra-226. The Agency seesno reason why four separate samplesfrom a single entry point (collected 60days apart) could not be either analyzedand averaged or composited in thelaboratory and analyzed, to determinefuture monitoring frequency. Therefore,EPA is suggesting for public commentthat systems take the average analyticalresults from four individual samples, orthe composite of four samples from eachentry point, in order to determine futurefrequency.

As discussed previously, EPA doesnot contemplate allowing compositingof multiple entry points for derivation ofcombined radium results. EPA requestscomment on any element of theforegoing discussion.

9. Alpha Emitters

Standardized monitoring: Same as forcombined radium (see previousdiscussion).



Decreased and increased monitoring:Same as for combined radium (seeprevious discussion).

Entry point monitoring: Same as forcombined radium (see previousdiscussion).

Sample Compositing: The current ruleallows compositing of four samples in alaboratory or the averaging of fourseparate analyses. Under the 1991proposal and the current rule, systemswould be allowed to compositeannually for samples taken from singleentry points (temporal compositing)and, under the 1991 proposal, tocomposite samples representing up tofive entry points with a six monthholding time (spatial compositing).

10. UraniumStandardized monitoring, monitoring

frequency, and entry point monitoring:Same as for combined radium (seeprevious discussion).

Sample Compositing: For systemswith gross alpha levels that are highenough to warrant uranium monitoring,annual composites for a single entrypoint would be allowed. Compositing offive samples representing five entrypoints would be permitted. If the resultwas greater than one fifth of the MCL,the individual samples would have tobe analyzed or re-sampling and analysisof the new individual samples wouldhave to occur.

11. Beta and Photon EmittersStandardized monitoring framework,

decreased monitoring: Monitoring forbeta and photon emitters would followthe same schedule as in the current rule.Decreased monitoring is not envisionedfor beta and photon emitters since onlyvulnerable systems would monitor,although EPA is taking comment on thepossibility of decreased monitoringaccording to the standardizedmonitoring framework as outlinedpreviously.

Screening levels: EPA recognizescertain problems with the current andproposed system. The proposedrequirement of a 30 pCi/L screen forgross beta and photon emitters had theeffect of no longer requiring Sr-90monitoring because the proposed limitwas above the screen of 30 pCi/L. Underthe current MCL, there is only onecontaminant that has a concentrationlimit near the 50 pCi/L screening level(Ni-63). There are five contaminantswith concentration limits at or near 30pCi/L and seven with limits below

thirty. A screen level of 50 pCi/L wouldpotentially miss the 12 contaminantswith concentration limits below 50 pCi/L and a screening level of 30 pCi/Lwould potentially miss the 7contaminants with concentration limitsbelow 30 pCi/L. Systems that aredrawing water from sources with knownbeta particle and photon radioactivityare required to use a screening level of15 pCi/L under the current rule. The1991 proposal retained this feature.

EPA thinks it is advisable to retain theproposed monitoring for sites within 15miles of a source of beta photonemitters. The screening level in theoriginal rule only affected surface watersystems serving over 100,000, or othersystems at State discretion, and thescreening level for gross beta reflectedthis limited regulation. However, aknown source of particular beta andphoton emitters should be monitored forthe specific radionuclides present atthat source which may be a healthconcern below the screen, but wouldnot be triggered by the screen. EPAwould give States discretion onrequiring specific monitoring forcontaminants from specific sources.

In addition, a 15 pCi/l screening levelis currently required for systems usingwater contaminated by effluents fromnuclear facilities. These systems mayalso be required by the State to monitorfor individual nuclides on a case by casebasis. Since both screens may missradionuclides of concern, EPA believesthis issue is important and may need tobe addressed in a future proposal. Inaddition, since many beta particle andphoton emitters have half-lives that aretoo short to be detected under thecurrent holding time, the issue ofsample holding time may have to be re-visited in a future proposal.

Holding time: Another issue has abearing on the screening level for whichEPA is requesting comment. There are asignificant number of beta and photonemitting radionuclides with short halflives, including those 13 nuclides ofconcern below the screening levelsbeing considered. Because annualsample compositing is allowed underthe current rule for beta and photonemitters, a screen above 30 pCi/L woulddetect a greater number of nuclideswhich (due to decay) may have beenabove a screen of 50 at the time ofsampling, but are now between 30 and50 pCi/L by the time of analysis. Ascreen level at 30 pCi/L would be more

sensitive a screen for beta particle andphoton radioactivity. The Agencyrequests comment on the selection ofscreening levels.

Sample Compositing: Annualcompositing is permitted for beta andphoton emitters in the current rule. Inaddition, for systems utilizing watercontaminated by effluents from nuclearfacilities, a quarterly compositing of fiveconsecutive daily samples was to beanalyzed for iodine-131, with morefrequent monitoring at State discretionif it was detected in the finished water.EPA believes this compositing for singlenuclide determinations is still valid.However, the 1991 proposed ruleexcluded compositing for beta andphoton emitter samples. It also limitedholding times to 6 months for singlesamples or 12 months for compositesper the lab cert manual. A screen above50 pCi/L , but with a sample holdingtime of 6 months without compositingmay be a reasonable approach,considering screening options, holdingtimes, and compositing issues. EPAsolicits comment on these beta andphoton emitter monitoring issues.

Entry point monitoring: EPA solicitsopinion on requiring beta photonmonitoring at entry points to thedistribution system for vulnerablesystems. EPA believes this isappropriate as it is for other nuclides,especially as an early warning ofcontamination from a localized sourceof man-made beta photon emitters.

12. Monitoring for Non-Transient Non-Community (NTNC) Systems

If EPA finalizes an option thatrequires monitoring for some or allNTNC systems, EPA wishes to make themonitoring requirements consistentbetween CWSs and those NTNC systemsrequired to monitor. See the previousdiscussion for CWS monitoring fordetails. As with CWSs, monitoringunder the SMF would be required atentry points to the distribution system,based on a nine-year cycle, consisting ofthree, 3-year monitoring periods, withprovisions for reduced monitoring asappropriate. If the radionuclidesNPDWRs for CWSs are fully extended toNTNCWSs, the monitoring frameworkswould be the same.

Table III–4 summarizes themonitoring frequencies for CWSs andNTNC systems, under the options thatrequire monitoring:



TABLE III–4. COMPARISON OF THE MONITORING FRAMEWORKS: THE EXISTING RULE, THE 1991 PROPOSAL, AND THEAPPROACH DESCRIBED IN THE NODA

Current rule (1976) 1991 proposal 2000 NODA

Radium Alpha Emitters and Uranium

Initial baseline: 4 consecutive quarterly samples Initial baseline: one sample per year for 3years.

Initial baseline: 4 consecutive quarterly sam-ples taken within 3 years from effectivedate or grandfathered data in previous com-pliance period.

If average > MCL=treat, etc ............................... Same as 1976 .................................................. Same as 1976.If one or more samples >MCL, do quarterly

sampling until average < MCL.Same as 1976 .................................................. Same as 1976.

If >50% of MCL, 4 Quarters every 4 years ........ If >50% of MCL, one sample every 3 yrs orwaiver to every 9 yrs.

Same as 1991 with no waiver.

If < 50% of MCL, 1 sample every 4 years ......... If <50% of MCL, one sample every 3 yrs orwaiver to every 9 years.

If < 50% of MCL, one sample every 6 years.

If no detect, 1 sample every 4 years ................. If no detect, one sample every 3 yrs or waiverto every 9 yrs.

If no detect, one sample every 9 years.

Beta and Photon Emitters

Quarterly gross beta monitoring. Vulnerablesystems and surface water systems >100,000 pop. Screen of 50; screen of 15 forcontaminated water I–131 quarterly, Sr–90and H–3 annual Sr–89 and Cs134 if above15.

Vulnerable systems (surface and groundwater) within 15 miles of source of manmade emitters do gross beta screen pro-posed at 30.

Same as 1991:Vulnerable systems within 15miles of source of man made emitters willmonitor with screen of 50 or 30. Same as1976: Screen of 15pCi/L for systems usingcontaminated waters. Same contaminantsas 1976 with corrections per NBS HB–69.

13. Polonium-210 and Lead-210

Risk estimates based on FederalGuidance Report No. 13 indicate thatcurrent screening levels for gross alphaand gross beta may not be adequate tocapture all contaminants of concern.Specifically, based on the new health-effects information contained in FGR–13(EPA 1999b), EPA believes it may beappropriate to require systems toperform isotopic analyses for additionalradionuclides that may present asignificant threat to human health. As aresult of this information, EPA isrequiring some systems to do analysesfor polonium-210 (a naturally occurringalpha emitter) and lead-210 (a naturallyoccurring beta emitter) under theRevisions to the UnregulatedContaminant Monitoring Regulation(UCMR) (64 FR 50556, Friday,September 17, 1999), to be implementedafter analytical methods for thesecontaminants have been approved.

14. Reporting Requirements

On May 13, 1999, EPA proposedsubpart Q (64 FR 25964) to revise theminimum requirements public watersystems must meet for publicnotification of violations of NPDWRsand other situations that pose a risk topublic health from the drinking water.EPA anticipates the final PublicNotification Rule (PNR), under part 141,subpart Q to be published in early 2000.After the final PNR is published,subsequent EPA drinking waterregulations that affect public

notification requirements will amendthe PNR as part of each individualrulemaking.

The proposed PNR divides the publicnotice requirements into three (3) tiers,based on the type of violation. ‘‘Tier 1’’applies to violations and situations withsignificant potential to have seriousadverse effects on human health as aresult of short-term exposure. Notice isrequired within 24 hours of theviolation. ‘‘Tier 2’’ applies to otherviolations and situations with potentialto have serious adverse effects onhuman health. Notice is required within30 days, with extensions up to threemonths at the discretion of the State orprimacy agency. ‘‘Tier 3’’ applies to allother violations and situations requiringa public notice not included in Tier 1and Tier 2. Notice is required within 12months of the violation, and may beincluded in the consumer confidencereport at the option of the water system.

Today’s NODA requests comment onwhether community water systems(CWS) should provide a Tier 2 publicnotice for MCL violations under theradionuclide NPDWRs and to provide atier 3 public notice for violations of themonitoring and testing procedurerequirements. If NTNC water systemsare required to monitor and notify, thenthey would be required to provide a Tier2 notice if the systems exceed the MCLs.EPA requests comment on theimplementation of public notificationrequirements by the effective date of theMCL and on the Tier 2 public notice

requirement for quarterly repeat noticesfor NTNC systems that continue toexceed the CWS MCL(s) under the‘‘monitoring and notification-only’’option. EPA believes States will phasein monitoring of NTNC systems basedon results of CWS systems in the sameproximity. The agency requestscomment on whether or not the sameincrease or decreased monitoringrequirements which pertain to CWSsshould apply to NTNC water systemsi.e. the 3, 6 and 9 year monitoring basedon being above 50% of the MCL, below50%, or non-detect.

As in the current rules, an analyticalresult that exceeds the MCL wouldtrigger additional confirmation samples,which in turn could trigger quarterlymonitoring. For man-made beta andphoton emitters, EPA is suggesting tofinalize the proposal regarding ascreening level of 30 or 50 pCi/L for‘‘vulnerable systems,’’ which aredefined as being within a 15 mile radiusof a source of this class of radionuclides.For Pb-210, EPA will be collecting datato make a future determinationregarding additional monitoring for thisnatural beta emitter.

Tables III–5 and III–6 summarize thecurrent and proposed monitoringrequirements and those suggested bytoday’s document.



TABLE III–5.—INITIAL (ROUTINE) MONITORING REQUIREMENTS

1976 1991 proposal 2000 NODA

GROSS ALPHA

CWSs: Four consecutive quarters at represent-ative point(s) within the distribution systemevery four years.

CWSs and NTNCWSs: Annual monitoring ateach entry point for first three years.

CWSs and NTNCWSs 1: Four consecutivequarters of monitoring at each entry point,anytime during first 3 years.

RADIUM

CWSs: Four consecutive quarters at represent-ative point(s) within the distribution systemevery four years. Initial monitoring is for ra-dium-226. If radium-226 exceeds 3 pCi/L,analysis for radium-228 is required. A grossalpha measurement can be substituted for ra-dium 226 and/or uranium monitoring if thegross alpha measurement is below the appli-cable MCL(s).

CWSs and NTNCWSs: Annual monitoring foreach radium isotope (radium-226 and ra-dium-228) at each entry point, for threeyears.

CWSs and NTNCWSs: Four consecutivequarters of monitoring for each radium iso-tope (radium-226 and radium-228) at eachentry point, any time during first 3 years.2

URANIUM

None ................................................................... CWSs and NTNCWSs: Annual monitoring ateach entry point for three years.

CWSs and NTNCWSs: Four consecutivequarters of monitoring for uranium to deter-mine compliance with both mass and activ-ity either by gross alpha or specific mass oractivity analysis at each entry point, everythree years.2

BETA AND PHOTON EMITTERS

CWSs serving > 100,000 persons and usingsurface water (and other systems designatedby the State): Four consecutive quarters forgross beta, tritium and strontium-90 at rep-resentative point(s) within the distribution sys-tem. Determine major constituents if exceedscreen of 50pCi/L Systems using water con-taminated with effluent from nuclear facilities:Quarterly monitoring 1 for gross beta and io-dine-131, strontium-90 and tritium. If grossbeta level is above 15 pCi/L, the same orequivalent samples must be analyzed forstrontium-89 and cesium-134.

Vulnerable systems only CWSs andNTNCWSs: (as designated by State): Twogross beta screening levels were discussedin the 1991 Proposal. Using a screen of 30pCi/L, quarterly monitoring for gross beta isrequired, along with annual tritium moni-toring. Using a screen of 50 pCi/L, quarterlymonitoring for gross beta is required, alongwith annual tritium and strontium-90 moni-toring.

Vulnerable systems only CWSs andNTNCWSs: (as designated by the State):Two gross beta 14 screening levels arebeing considered. Using a screen of 50 or30 pCi/L, quarterly monitoring for grossbeta is required, along with annual moni-toring for tritium and strontium-90 as in1976. Vulnerability based on proximity (15miles ) to source per 1991. Screen of 15 forcontaminated waters as in 1976.3

NOTE: 1 This assumes that monitoring will be required at NTNC systems. If this is not the case, these requirements would not apply to NTNCsystems.

2 A gross alpha measurement can be substituted for radium-226 and/or uranium monitoring if the gross alpha measurement is below the appli-cable MCL(s).

3 Quarterly monitoring for gross beta would be based on the analysis of monthly samples or the analysis of a composite of three monthly sam-ples. For iodine 131, a composite of five consecutive daily samples shall be analyzed once per quarter. Additional monitoring may be required toidentify specific isotopes if gross beta measurement exceeds the screening level.

TABLE III–6.—REDUCED MONITORING REQUIREMENTS


GROSS ALPHA

CWSs: One sample every four years if annualaverage from previous results (four consecu-tive quarterly samples) is less than 1⁄2 MCL.

CWSs and NTNCWSs: One sample everythree years, if previous monitoring results(from three years of annual monitoring) arebelow MCL. If system is reliably and con-sistently below MCL, the system could re-ceive a waiver, and monitor once everynine years.

CWSs and NTNCWSs: One sample everythree years if previous monitoring results(‘‘previous results’’) are reliably and consist-ently at or below MCL; one sample everysix years if previous results are reliably andconsistently at or below 1⁄2 MCL; or onesample every nine years if previous resultsare reliably and consistently at or below theMDL.



TABLE III–6.—REDUCED MONITORING REQUIREMENTS—Continued


RADIUM

CWSs: One sample every four years if annualaverage from previous results (four consecu-tive quarterly samples) is less than 1⁄2 MCL.

CWSs and NTNCWSs using ground water:One sample every three years, if previousmonitoring results (from three years of an-nual monitoring) are below MCL. If systemis reliably and consistently below MCL, thesystem could receive a waiver, and monitoronce every nine years.

CWSs and NTNCWSs: One sample everythree years if previous results are reliablyand consistently at or below MCL; onesample every six years if previous resultsare reliably and consistently at or below 1⁄2MCL; or one sample every nine years ifprevious results are reliably and consist-ently at or below the MDL.

URANIUM

None ................................................................... CWSs and NTNCWSs: One sample everythree years, if previous monitoring results(from three years of annual monitoring) arebelow MCL. If system is reliably and con-sistently below MCL, the system could re-ceive a waiver, and monitor once everynine years.

CWSs and NTNCWSs: One sample everythree years if previous results averagebelow MCL; one sample every six years ifprevious results average at or below 1⁄2MCL; or one sample every nine years ifprevious results average below the MDL.

BETA AND PHOTON EMITTERS

CWSs serving > 100,000 persons and usingsurface water (and other systems designatedby the State): Every four years, systemsmust collect samples from four consecutivequarters for gross beta at representativepoint(s) within the distribution system. Sys-tems using water contaminated with effluentfrom nuclear facilities: No reduced monitoringis allowed.

Vulnerable systems only (as designated byState): Since only vulnerable systems arerequired to monitor, no reduced monitoringis allowed.

Vulnerable systems only (as designated bythe State): Since only vulnerable systemsare required to monitor, no reduced moni-toring is allowed.

15. Laboratory Capacity Issue ‘‘ PossibleExtension of Initial Monitoring Period

As discussed earlier in the analyticalmethods section (III.J), the PerformanceEvaluation Program (now known as theProficiency Testing Program) has beenexternalized. Although the Agency isunsure at this time how externalizationmay affect laboratory capacity, EPArecognizes that it may be animplementation issue for at least threereasons:

• The recent externalization of theradionuclides Performance Evaluation(PE) studies program may cause short-term disruption in laboratoryaccreditation;

• Requiring NTNCWSs to monitorunder the Standard MonitoringFramework will add approximately20,000 systems to the universe ofsystems that are already required tomonitor;

• And the radon rule will beimplemented simultaneously with theradionuclides rule.

NIST is in the process of approving aprovider for PT samples forradionuclides. States also have theoption of approving their own PTsample providers. Should laboratorycapacity issues related to externalization

present implementation problems forthe initial monitoring period (threeyears), EPA will consider allowing anadditional year (four years total) for theinitial monitoring period. During thespecified time period, systems would berequired to analyze four consecutivequarterly samples to determinecompliance. If the final rule ispromulgated in November of 2000, thenew monitoring requirements wouldbegin to be enforced in November of2003. If EPA implements a one yearextension, water systems would haveuntil December 31 of 2007 to completethe required initial monitoring. Thisscenario would allow the ‘‘one third ofsystems per year’’ strategy inherent inthe Standard Monitoring Framework tobe applied, while allowing oneadditional year, if necessary, to addressany laboratory capacity issues. EPAsolicits public comment on this matter.

L. Effective DatesMuch of the rule that will be finalized

in November will involve retainingcurrent elements of the radionuclidesNPDWR. Those portions of the final rulethat are unaffected by the upcomingregulatory changes are already in effect.MCLs for gross alpha, beta particle andphoton radioactivity, and combined

radium-226 and -228 will be unchangedand are already in effect. Regardingwater systems that are currently out ofcompliance with the existing NPDWRsfor gross alpha, combined radium-226and -228, and/or beta particle andphoton radioactivity, States withprimacy and EPA will renegotiateenforcement actions that put systems oncompliance schedules as expeditiouslyas possible.

Under the Safe Drinking Water Act,final rules become effective three yearsafter promulgation (November of 2003,assuming that the rule becomes final inNovember of 2000). The followingdiscussion assumes a promulgation dateof November 2000. For reasonsdescribed in the monitoring section ofthe NODA (section III, part K) and theAppendices (appendix V), initialmonitoring will be required tocompleted by December 31, 2007. Underthe Standard Monitoring Framework,systems have three years to completethe initial monitoring cycle of fourconsecutive quarterly samples.However, for reasons described in themonitoring section of the NODA(section III, part K) , systems will havean additional year to complete theinitial monitoring cycle, which willcorrespond to an end date of December



31, 2007. This includes initialmonitoring for uranium, the newmonitoring requirements for radium-228, and new initial monitoring underthe requirements for entry points.Compliance determinations and futuremonitoring cycle schedules are alsodiscussed in the monitoring sectionscited. MCL violations resulting from thenew requirement for separate Ra-228monitoring will be treated as ‘‘newviolations’’ and will be on the sameschedule as other new violations (e.g.uranium).

M. Costs and BenefitsThe Safe Drinking Water Act provides

for EPA to consider both public healthand the feasibility (taking costs intoconsideration) in establishing drinkingwater MCLs. In addition the newAmendments require EPA to evaluatethe costs and benefits of potentialrevisions to the current standards. Asnoted earlier, the Agency conducted ananalysis of the costs and associatedbenefits of each of the options describedin today’s document. These analyseswere performed consistent with therequirements for a Health RiskReduction and Cost Analysis set forth inthe 1996 Amendments to the SDWA(section 1412(b)(3)(C)).

First, all public water systems that arecurrently treating and are in compliancewith the 1976 standards will have noadditional cost if the rule remains thesame as it is now. At the same time, EPArecognizes that it may be costly tosystems which have delayedcompliance. However, to the extent therule remains the same, costs necessaryto comply with the existing rule, as wellas public health benefits associated withit, have accrued to that 1976 rule. If EPAchanges nothing, the existing 1976requirements must be met. EPAconsiders only those costs associatedwith accommodating revisions to thecurrent regulations to be new costs.Costs incurred, or those that shouldhave been incurred to comply with aprevious regulation, are not factoredinto current considerations.

Second, EPA has reexamined thecosts of the 1991 proposal regardingmonitoring for any changes which maybe warranted based on new data. EPA iscontemplating several changes whichwere part of the 1991 proposedregulation and which may increasecosts. These include: (1) Promulgatingan NPDWR for uranium; (2) applyingthe radionuclide NPDWRs to non-transient, non-community (NTNC)systems; (3) requiring monitoring at thepoint of entry to distribution systems,and ; (4) requiring separate monitoringfor radium-226 and radium-228.

EPA is also recommending rapidsample analysis for alpha emitters todetect the presence of short livedradionuclides such as radium-224, butis not contemplating requiring it as partof the revision to the radionuclides rule.The Agency will pursue the issue of atimely analysis of gross alpha to reflectshort half lived Ra-224 in a separateproposal.

Costs and benefits for the variousoptions are presented in appendix V oftoday’s document, in the TechnicalSupport Document (EPA 2000a), and inthe draft Health Risk Reduction andCost Analysis (EPA 2000b). Today’sNODA solicits comment on whether theincremental risk reduction may justifythe costs for certain of the revisionsdescribed in the NODA. EPA requestspublic comment on such questions andon the extent to which its discretionaryauthority provided by section 1412(b)(6)of the SDWA should be used. ThisNODA also requests public inputregarding the need for furtheradjustments to the limits based on thecost and risk data presented in today’sNODA.

IV. References

Parsa, Bahman. ‘‘Contribution of Short-Lived Radionuclides to Alpha-ParticleRadioactivity in Drinking Water and TheirImpact on the Safe Drinking Water ActRegulations’’. Radioactivity &Radiochemistry (Vol. 9, No. 4). pp. 41–50).1998.

USEPA. Drinking Water Regulations;Radionuclides. Federal Register, Vol. 41, No.133, p. 28402. July 9, 1976.

USEPA. National Primary Drinking WaterRegulations; Radionuclides; Proposed Rule.Federal Register, Vol. 56, No. 138, p. 33050.July 18, 1991.

USEPA. Presumptive Response Strategyand Ex-Situ Treatment Technologies forContaminated Ground Water at CERCLASites: Final Guidance. EPA 540/R–96/023.(EPA 1996a)

USEPA. Performance Evaluation StudiesSupporting Administration of the CleanWater Act and the Safe Drinking Water Act.Federal Register, Vol. 61, No. 139, p. 37464.July 18, 1996. (EPA 1996b)

USEPA. National Primary Drinking WaterRegulations; Analytical Methods forRadionuclides; Final Rule and ProposedRule. Federal Register, Vol. 62, No. 43, p.10168. March 5, 1997. (EPA 1997a)

USEPA. Performance Evaluation StudiesSupporting Administration of the CleanWater Act and the Safe Drinking Water Act.Federal Register, Vol. 62, No. 113, p. 32112.June 12, 1997. (EPA 1997b)

USEPA. Performance Based MeasurementSystem. Federal Register, Vol. 62, No. 193,p. 52098. October 6, 1997. (EPA 1997c)

USEPA. Manual for the Certification ofLaboratories Analyzing Drinking Water. EPA815–B–97–001. 1997. (EPA 1997d)

USEPA. 1997. Memorandum fromAdministrator Carol M. Browner of EPA to

Chairperson Shirley A. Jackson of theNuclear Regulatory Commission. February 7,1997. (EPA 1997e)

USEPA. Office of Solid Waste andEmergency Response (OSWER). 1997.Memorandum from Stephen Ludwig andLarry Weinstock (Office of Radiation andIndoor Air) to Addressees. ‘‘Establishment ofCleanup Levels for CERCLA Sites withRadioactive Contamination’’. OSWER No.9200.4–18. August 22, 1997. (EPA/OSWER1997a).

USEPA. Office of Solid Waste andEmergency Response (OSWER). 1997.Memorandum from Timothy Fields toRegional Administrators. ‘‘The Role ofCSGWPP’s in EPA Remediation Programs’’.OSWER No. 9283.1–09. April 4, 1997. (EPA/OSWER 1997b).

USEPA. Memorandum to WaterManagement Division Directors, Regions I–X,from Cynthia C. Dougherty, Director, Officeof Ground Water and Drinking Waterregarding Recommendations ConcerningTesting for Gross Alpha Emitters inDrinkingWater (January 27, 1999). (EPA1999a)

USEPA. Cancer Risk Coefficients forEnvironmental Exposure to Radionuclides,Federal Guidance Report No. 13. USEnvironmental Protection Agency,Washington, DC, 1999. (EPA 1999b)

USEPA. ‘‘Technical Support Document forthe Radionuclides Notice of DataAvailability’’. Draft. March, 2000. (EPA2000a)

USEPA. ‘‘Preliminary Health RiskReduction and Cost Analysis: RevisedNational Primary Drinking Water Standardsfor Radionuclides’’. Prepared by IndustrialEconomics, Inc. for EPA. Draft. January 2000.(EPA 2000b)

U.S. Environmental Protection Agency.Memorandum to David Huber, EdwinThomas, OGWDW from Scott Telofski, ORIAregarding Uranium Analysis Screening Levelfor Drinking Water Regulations NODA.March 14, 2000. (EPA 2000c)

U.S. Environmental Protection Agency.Memorandum to David Huber, EdwinThomas, OGWDW from Scott Telofski, ORIAregarding Gross Alpha Screening forUranium. March 20, 2000. (EPA 2000d)

U.S. Geological Survey (USGS). ‘‘Radium-226 and Radium-228 in a Shallow GroundWater, Southern New Jersey’’. Fact Sheet FS–062–98. June 1998.

Appendix I—Occurrence

In order to estimate the total national costsand benefits of revising the MCLs it isnecessary to develop updated nationalestimates of the occurrence and exposure tothese radionuclide contaminants in drinkingwater. Occurrence data and associatedanalyses provide indications of the numberof public water supply systems withconcentration of radionuclides above therevised MCL as well as the population servedby these systems. Monitoring and treatmentcosts can be estimated from the occurrencedata.

A. Background

EPA conducted a nationwide occurrencestudy of naturally occurring radionuclides in



public water supplies called the NationalInorganic and Radionuclides Survey (NIRS)(see EPA 1991, proposed rule). The objectiveof NIRS was to characterize the occurrence ofa variety of constituents, including radium-226, radium-228, uranium (mass analysis),

gross alpha-particle activity, and gross beta-particle activity, present in communityground-water supplies (finished water) in theUnited States, and its territories. The surveyincluded a random sample from 990collection sites. The public water supplies

were stratified into four size categories, andthe samples were chosen to best represent thesame stratification present in the totalpopulation of community water supply inexistence at the time, as shown in———Table I–1.

TABLE I–1.—COMPARISON OF NIRS TARGET SAMPLE WITH FEDERAL REPORTING DATA SYSTEM (FRDS) INVENTORY

Population category (population range) Number ofFRDS sites*

Percentage ofFRDS sites

Number ofNIRS sites

Percentage ofNIRS sites

Very small (25–500) ................................................................................ 34,040 71.4 716 71.6Small (501–3,300) ................................................................................... 10,155 21.3 211 21.1Medium (3,301–10,000) ........................................................................... 2,278 4.8 47 4.7Large and very large (10,001–>100,000) ................................................ 1,227 2.6 26 2.5

Total .............................................................................................. 47,700 100.1 1,000 100.0

* Based in FRDS inventory for fiscal year 1985 from Longtin, 1988.

Results of NIRS were used to develop theproposed radionuclide rule in 1991 (56 FR33050; EPA 1991). There has not been acomparable national survey for radionuclidessince. Since the publication of the proposed1991 revision to the MCLs, the United StatesGeological Survey has collected additionaldata on various radionuclides in groundwaterto augment the data of the NIRS. Thesestudies are summarized subsequently, and ingreater detail in the Technical SupportDocument (EPA 2000a).

Szabo and Zapecza (1991) detail thedifferences in the occurrence of uranium andradium-226 in oxygen-rich and oxygen-poorareas of aquifers. Because the chemicalbehavior of uranium and radium are vastlydifferent, the degree of mobilization of the

parent and product are different in mostchemical environments.

Recently, high concentrations of radiumwere found to be associated with groundwater that was geochemically affected byagricultural practices in the recharge areas bystrongly enriching the water with competingions such as hydrogen, calcium, andmagnesium (Szabo and dePaul, 1998).Radium-228 was detected in about equivalentconcentrations as radium-226 in the aquiferstudy in New Jersey (Szabo and dePaul,1998).

B. USGS Radium Survey

A 1998 USGS survey (see EPA 2000a) wasdesigned to target areas of known, orsuspected, high concentrations of radium-224as inferred by associated radium occurrencedata, geologic maps, and other geochemical

considerations. Thus, the survey is likelybiased toward the extreme high end of theoccurrence distribution for radium-224 andco-occurring contaminants such as radium-228. Approximately half of the samples werebelow the minimum detectable concentrationof radium-226 and radium-228 in spite of thefact that public water systems were targetedin areas where high concentrations of radiumwere expected. Table I–2 shows that, of the104 samples, 21 exceeded the MCL forcombined radium, and about 5 percentexceeded 10 pCi/L of radium-224, thoughseveral of these samples with pH less than4.0 also contained detectable concentrationsof thorium isotopes as well. Concentrationsexceeded 1 pCi/L in about 10 percent of thesamples analyzed for lead-210 and 3 percentfor polonium-210.

TABLE I–2.—PERCENT OF SAMPLES EXCEEDING SPECIFIED CONCENTRATION

RadionuclideTotal num-ber of sam-

ples

Percent of samples exceeding given concentration (pCi/L)

1 2 3 5 7 10

Ra-224 ......................................................................................... 104 30 26 20 15 9 5Ra-226 ......................................................................................... 104 33 22 17 10 5 2Po-210 .......................................................................................... 95 3 1 1 1 0 0Pb-210 .......................................................................................... 96 10 3 1 1 0 0

Radium-224 occurs in many of the wellssampled at concentrations that highlight thelimitations of the present monitoring schemefor the gross alpha-particle standard. Inaddition, the contribution of radium-224 andits short-lived daughter products to grossalpha emissions was estimated with datafrom a concurrent study of ground-watersupplies by the USGS in cooperation withthe state of New Jersey (Szabo et al., 1998).In that study, gross alpha emissions weremeasured before the decay of radium-224 andafter sufficient time had elapsed for radium-224 decay (about 18–22 days). In this way,the difference between the initial gross-alphameasurement and the final measurement isindicative of the contribution of radium-224and all other alpha emitting isotopes thatwould decay within this time frame. Theresults indicate that the contribution ofradium-224 and its short-lived daughter

products is approximately three times theconcentration of radium-224. While thisanalysis was developed with a small data setin a restricted geographic range, it is basedon a physical process and has importantimplications for such things as projections ofradium-224 occurrence in association withgross-alpha concentrations. These results arealso important in light of both the costlinessand difficulty of the radium-224 analysis.

Concentrations of radium-228 were highlycorrelated with radium-224. Although thiscorrelation was based on a limited number ofdata points, there is a physical basis to thecorrelation since both nuclides originate fromthe same decay chain. Therefore, there ispotential for using radium-228 as a proxyindicator for the much shorter lived andinfrequently sampled radium-224. Inaddition, the isotopic ratios of radium-226 toradium-228 were below 3:2 in many samples

indicating that the gross alpha-particle screenthat is currently used for combined radium(radium-226 + radium-228) compliancewould be inadequate in many situations.

Polonium-210 and lead-210 are derivedfrom the uranium-238 decay series; the decayseries that produces radium-226. However,the survey was designed to assess radium-224; therefore results are possibly biased toareas that would more likely have isotopes inthe thorium-232 decay series. In addition, thecorrelations of radium-226 with radium-224and radium-228 are only 0.51 and 0.61respectively; consequently, the wells thatwere sampled may not be located in areasexpected to have polonium-210 or lead-210.Within these constraints, the new data helpto fill the gap in occurrence information thatexisted for these isotopes. Polonium-210 wasfound in concentrations exceeding 1 pCi/L inonly two wells. At this time, these



observations could not be associated withunique geochemical controls (as has beenaccomplished in a previous study in Florida;Harada et al., 1989) and further investigationswould be necessary to infer anything moreabout the national distribution andoccurrence of polonium-210.

Approximately 12 percent of the samplesexceeded a lead-210 concentration of 1 pCi/L; however only one sample was greater than3 pCi/L. The greatest frequency of detectionwas in the Appalachian PhysiographicProvince of the northeastern United States,especially in of Connecticut andPennsylvania. The geochemical mechanismthat controls lead-210 dissolution is also notwell established and needs further study,though lead is less soluble than radium. Inaddition, lead-210, like polonium-210, isderived from a different decay chain thanradium-224 and it was therefore notconsidered in designing the study. Onepossible explanation for the frequentdetection of lead-210 in concentrationsgreater than 1 pCi/L in the Appalachianregion may be the high concentrations ofradon-222 in ground water in this region(Zapecza and Szabo, 1986). As the radon insolution decays through a series of very shorthalf-lived products to Lead-210, a smallfraction of the lead-210 may not be sorbedonto the aquifer matrix; thus, the higher theinitial radon-222 concentration, the morelikely measurable amounts of lead-210 wouldbe found in the ground water. Thishypothesis could not be tested howeverbecause radon-222 was not analyzed in thisstudy.

C. USGS Beta/Photon Data Collection Effort

The major source of data for man-maderadionuclides is the Environmental RadiationAmbient Monitoring System (ERAMS) whichis published quarterly in the EnvironmentalRadiation Data (ERD) reports. The ERDreports provide concentration data on grossbeta-particle activity, tritium, strontium-90,and iodine-131 for 78 surface-water sites thatare either near major population centers ornear selected nuclear facility environs.

An additional data collection effort wascompleted by the U.S. Geological Survey inthe summer of 1999 (see EPA 2000a) toanalyze targeted beta-particle emittingradionuclides from a small number of publicwater systems that had shown relatively highlevels of beta/photon emitters during theoriginal NIRS survey. Of the 26 public watersystems contacted for this effort none couldascertain which wells in their systems wereoriginally sampled as part of NIRS.Consequently, although all efforts were madeto include as many of the original systems aspossible, it is presently unknown if the wellssampled match those in NIRS. Theradionuclide analyses for this data collectioneffort included; short-term (48 hour) grossbeta-particle and gross alpha-particleactivities, long-term (30 days) gross beta-particle and gross alpha-particle activities,tritium, strontium-89, strontium-90, cesium-134, cesium-137, iodine-131, uranium-234,uranium-235, uranium-238, radium-228,radium-226, lead-210, and cobalt-60.

Gross beta-particle activities were all below50 pCi/L in water collected from public water

systems that were sampled previously duringthe National Inorganics and RadionuclideSurvey (NIRS) and had been found to containgross beta-particle activity in excess of 20pCi/L. To the extent possible, all sampleswere collected from the original public watersystems surveyed for NIRS where gross beta-particle activities were 20 pCi/L or greater.However due to the amount of time that hadelapsed since the NIRS samples werecollected, correlation with the originalsampling point could not be verified forevery water supply sampled.

Though the number of samples was limited(26 samples), a few conclusions can bereached. Concentrations of gross beta-particleactivities will rarely exceed 50 pCi/L in watercollected from public water systems (and didnot do so in this study). A significantpercentage (15% or 4 samples) of the 26samples analyzed, however, contained grossalpha-particle activities at or in excess of the15 pCi/L MCL indicating that concern overthe presence of elevated concentration ofgross alpha-particle activity in ground wateris justified. Long-term (30-day) gross beta-particle activity analyses did not indicatesignificant ingrowth of beta-particles in anyof the samples, though this result is qualifiedby the absence of significant quantities ofuranium-238 in any of the samples collected.Naturally occurring potassium-40 andradium-228 are a significant source of grossbeta-particle activity to many of the samplesin agreement with results of Welch et al.,1995., Minor concentrations of naturally-occurring lead-210 are also detectedoccasionally. No manmade radionuclide wasdetected in concentration above themaximum detectible concentration (MDC) inany of the samples. The presence of naturallyoccurring beta-particle emittingradionuclides must be taken into accountwhen evaluating the source of high grossbeta-particle activity in ground water as firstsuggested by Welch et al., 1995.

D. References

Harada, Kow, William C. Burnett, Paul A.LaRock, and James B. Cowart, 1989.Polonium in Florida groundwater and itspossible relationship to the sulfur cycle andbacteria. Geochemical et Cosmochimica ActaVol. 53, pp. 143–150.

Longtin, Jon, 1988. Occurrence of Radon,Radium and Uranium in Groundwater.Journal American Water Works Association,pp. 84–93.

Szabo, Z., and V.T dePaul, 1998. Radium-228 and radium-228 in shallow groundwater, southern New Jersey. U.S. GeologicalSurvey Fact Sheet FS–062–98.

Szabo, Z., V.T. dePaul, and B. Parsa, 1998.Decrease in gross alpha-particle activity inwater samples with time after collection fromthe Kirkwood Cohansy aquifer system insouthern New Jersey: Implications forregulations. Drinking Water 63rd annualmeeting American Water Works AssociationNew Jersey Section. Atlantic City, NJ.

Szabo, Z., and O.S. Zapecza, 1991,Geologic and geochemical factors controllinguranium, radium-226, and radon-222 inground water, Newark Basin, New Jersey:Gundersen, L.C.S. and Wanty, R.B., eds.,Field studies of radon in rocks, soils, and

water, U.S. Geological Survey Bulleting 1971,p. 243–266.

USEPA. National Primary Drinking WaterRegulations; Radionuclides; Proposed Rule.Federal Register. Vol. 56, No. 138, p. 33050.July 18, 1991.

USEPA. ‘‘Technical Support Document forthe Radionuclides Notice of DataAvailability’’. Draft. March, 2000. (EPA2000a)

Welch, A.H., Szabo, Z., Parkhurst, D.L.,Van Meter. P.C., and Mullin, A.H., 1995,Gross-beta activity in ground water: naturalsources and artifacts of sampling andlaboratory analysis: Applied Geochemistry, v.10, no. 5, p. 491–504.

Zapecza, O. S., and Z. Szabo, 1986. Naturalradioactivity in ground water—a review. U.S.Geological Survey National Water Summary1986, Ground-Water Quality: HydrologicConditions and Events, U.S. GeologicalSurvey Water Supply Paper 2325. pp. 50–57.

Appendix II—Health Effects

The following information summarizes thesalient changes in risk assessmentinformation and risk characterizationmethodology during the past two decades.The Technical Support Document (EPA2000a) also provides additional information.

A. Use of Linear Non-Threshold Assumption

In estimating the health effects fromradionuclides in drinking water, EPAsubscribes to the linear, non-threshold modelwhich assumes that any exposure to ionizingradiation has a potential to producedeleterious effects on human health, and thatthe magnitude of the effects are directlyproportional to the exposure level. TheAgency further believes that the extent ofsuch harm can be estimated by extrapolatingeffects on human health that have beenobserved at higher doses and dose rates tothose likely to be encountered fromenvironmental sources of radiation. The risksassociated with radiation exposure areextrapolated from a large base of human data.EPA recognizes the inherent uncertaintiesthat exist in estimating health impact at thelow levels of exposure and exposure ratesexpected to be present in the environment.EPA also recognizes that, at these levels, theactual health impact from ingestedradionuclides will be difficult, if notimpossible, to distinguish from naturaldisease incidences, even using very largeepidemiological studies employingsophisticated statistical analyses. However,in the absence of other data, the Agencycontinues to support the use of the linear,non-threshold model in assessing risksassociated with all carcinogens.

B. Continuous Improvements in Models, DataBase

As various scientific institutions havecontinued to collect data on the observedeffects of radiation from the cohort of bombsurvivors, patients with medical exposure,and workers with occupational exposure;continuous improvements have been possiblein models to extrapolate effects and toestimate the risks of small exposures toradiation from the natural environment orman-made sources. The data have led to



changes in risk estimates as summarizedhere.

1. Basis of 1976 Estimates of Risk

• Risk of bone cancer from radium dialpainters.

• Autopsy radioassay (see EPA 2000a).Body burden from natural intake or radium,about 1 pCi/day.

• Estimate annual dose rate in severalorgans from natural radium in rad/year.

• BEIR I risk numbers for radium dialpainters yields risk/year per rad/year.

• Calculate risk over lifetime.a. 1976 Estimates of the Risks from

Radium-226 and Radium-228. In general,EPA followed the Federal Radiation Council(FRC) recommendation that radium ingestionlimits for the general population should bebased on environmental studies and not themodels used to establish occupational doselimits (see EPA 2000a). In setting the MCL,EPA considered bone cancer and other softtissue cancers to be the principal healtheffects associated with radium ingestion. Tocalculate body burdens, doses, and risks fromingestion of radium-226 and radium-228, in1976, EPA relied on data from the 1972report of the United Nations ScientificCommittee on the Effects of AtomicRadiation (see EPA 2000a) and the 1972 theNational Academy of Sciences (NAS)Committee on the Biological Effects ofIonizing Radiation, BEIR I Report (see EPA1991, proposed rule). Additional informationand support were found in the InternationalCommission on Radiological Protection,Publication 20 (see EPA 2000a). Theliterature suggests that radium-228 was astoxic as radium-226, and possibly twice astoxic for bone cancers in dogs. Given this,EPA believed that it was prudent to assumethat the adverse health effects due tochronically ingested radium-228 were at leastas great as those from radium-226.

Assuming equal toxicity with radium-226,EPA reasoned that lifetime ingestion of onlyradium-228 at 5 pCi/L would yield lifetimetotal cancer risks equal to those for a lifetimeingestion of only radium-226 at the sameconcentration, i.e., between 0.5 to 2 × 10¥4.By setting the MCL at 5 pCi/L for radium-226and radium-228 combined, rather thanindividually, EPA sought to limit the lifetimetotal cancer risk from the ingestion of bothisotopes in drinking water to 2 × 10¥4 or less.

b. Basis for the 1976 MCL for Gross AlphaParticle Activity. One of the main intentionsof the 15 pCi/L MCL for gross alpha particleactivity, which includes radium-226 butexcludes uranium and radon, was to limit theconcentration of other naturally-occurringand man-made alpha emitters relative toradium-226. Specifically, this limit wasbased on the fact that EPA estimated thatcontinuous consumption of drinking watercontaining polonium-210, the next mostradiotoxic alpha particle emitter in theradium-226 decay chain, at a concentrationof 10 pCi/L might cause the total dose tobone to be equivalent to less than 6 pCi/L ofradium-226.

The 15 pCi/L limit, which includesradium-226 but excludes uranium and radon,was based on the conservative assumptionthat if the radium concentration is limited to5 pCi/L and the balance of the alpha particle

activity (i.e., 10 pCi/L) is due to polonium-210, the total dose to bone would be less thanthat dose associated with an intake of 6 pCi/L of radium-226.

c. Basis for the 1976 MCL for Beta Particleand Photon Radioactivity. In 1976, EPAestimated that continuous consumption ofdrinking water containing beta and photonemitting radioactivity yielding a 4 mrem/yrtotal body dose may cause an individual fatalcancer risk of 0.8 × 10¥6 per year, or alifetime cancer risk of 5.6 × 10¥5, assuminga 70-year lifetime. In setting the MCL forman-made beta and photon emitters, EPAused cancer risk estimates from the BEIR Ireport for the U.S. population in the year1967 (see EPA 1991, proposed rule). For anexposed group having the same agedistribution as the U.S. 1967 population, theBEIR I report indicated that the individualrisk of a fatal cancer from a lifetime totalbody dose rate of 4 mrem per year rangedfrom about 0.4 to 2 × 10¥6 per yeardepending on whether an absolute or relativerisk model was used. Using best estimatesfrom both models for fatal cancer, EPAbelieved that an individual risk of 0.8 × 10¥6

per year resulting from a 4 mrem annual totalbody dose was a reasonable estimate of theannual risk from a lifetime ingestion ofdrinking water. Over a 70-year period, thecorresponding lifetime fatal cancer riskwould be 5.6 × 10¥5, with the risk from theingestion of water containing less amounts ofradioactivity being proportionately smaller.

Based on 1967 U.S. Vital Statistics (seeEPA 1991 and EPA 2000a), the probabilitythat an individual would die of cancer wasabout 0.19, and was thought to be increasedby 0.1 percent from a lifetime doseequivalent rate of 15 mrem per year.Therefore, EPA calculated that the 4 mrem/yr MCL for man-made beta and photonemitters corresponded to a lifetime riskincrease of 0.025 percent to exposed groups.

EPA knew that partial body irradiation wascommon for ingested radionuclides sincethey are, like radium, largely deposited in aparticular organ, or in a few organs. In suchcases, EPA acknowledged that the risk permillirem varies depending on theradiosensitivity of the organs at risk. Forexample, EPA estimated that cancers due tothe thyroid gland receiving 4 mrem per yearcontinuously ranged from about 0.2 to 0.5 peryear per million exposed persons (averagedover all age groups). Considering the sum ofthe deposited fallout radioactivity and theadditional amounts due to releases fromother sources existing at that time, EPAbelieved that the total dose equivalent fromman-made radioactivity was not likely toresult in a total body or organ dose to anyindividual that exceeded 4 mrem/yr.Consequently, EPA did not believe that the4 mrem/yr standard would affect manypublic water systems, if any. At the sametime, the Agency believed that an MCL set atthis level would provide adequate publichealth protection.

2. 1991 Proposal: Basis of Health RiskEstimates

During the years since the publication ofthe 1976 regulations, the Agency obtained agreat deal of additional data and a betterunderstanding of the risks posed to human

health by ingested radionuclides. Many ofthese new studies were presented anddiscussed in the Advance Notice of ProposedRulemaking announcing EPA’s intent torevise the MCLs (51 FR 34836, Sept. 20,1986) and the supporting health criteriadocuments (see EPA 2000a and EPA 1991,the proposed rule).

Among the most important changes madeby EPA in developing the 1991 revisions wasthe adoption of a common calculationalframework, the RADRISK computer code (seeEPA 1991, proposed rule), to estimate therisks posed by ingestion of radionuclides indrinking water. The RADRISK code consistedof intake, metabolic, dosimetric, and riskmodels that integrated the results of a largenumber of studies on a variety of radioactivecompounds and radiation exposuresituations into an overall model to estimaterisks for many different radionuclides.Radionuclide-specific parameters were basedon the results of individual scientific studiesof a specific radionuclide, such as radium;human epidemiological studies; orexperimental animal studies of groups ofchemically-similar radionuclides. Tosummarize, the following are some of thesalient changes.

• Used RADRISK metabolic model insteadof natural uptake equilibrium model. Basedon known intakes.

• Used ICRP report 20 (see EPA 2000a) onalkaline earth elements with Oak Ridgemodeled exponential fit to that model.

• BEIR IV risks for alpha emitters.• Ra–224 data from ankylosing

spondylitis, tuberculosis.• Change in results from Ra–228

calculations (Oak Ridge model of ’84) andICRP 30 (see EPA 2000a) yielded differentresults based on retention and distribution ofeach member of decay chain.

a. Basis for the 1991 MCL for Radium-226and Radium-228. In 1991, EPA proposedrevised MCLs for radium-226 and radium-228 individually at 20 pCi/L each. TheAgency thought at that time that the limit foreach of these radium isotopes was within theAgency’s acceptable risk range of 10¥6 to10¥4. The Agency no longer believes theMCLs proposed in 1991 for radium-226 andradium-228 are within the Agency’sacceptable risk range.

i. Human and Animal Health Effects DataConsidered. In 1991, EPA based its riskestimates for radium using information fromtwo epidemiological study groups. The firstgroup consisted of radium dial painters whohad ingested considerable amounts of radiumpaint (containing various proportions ofradium-226 and radium-228) by sharpeningthe point of their paint brush with the lips.The second group consisted of patients inEurope injected with a short-lived isotope ofradium, radium-224, for treatment of spinalarthritis and tuberculous infection of thebone (see EPA 2000a). The results of thesestudies are described briefly next.

At high levels of exposure to radium,several non-cancer health effects wereobserved in radium dial painters, such asbenign bone growths, osteoporosis, severegrowth retardation, tooth breakage, kidneydisease, liver disease, tissue necrosis,cataracts, anemia, immunologicalsuppression and death (see EPA 2000a).



Exposed radium dial painters alsoexhibited significantly elevated rates of tworare types of cancer, bone sarcomas(osteosarcomas, fibrosarcomas andchondrosarcomas) and carcinomas of headsinuses and mastoids (see EPA 2000a andEPA 1991, the proposed rule). The incidenceof head carcinomas was associated withexposure to radium-226, but not radium-228(see EPA 2000a). This is because these lattercancers were due to an accumulation ofradon gas (radon-222) in the mastoid air cellsand paranasal sinuses caused by the escapeof radon-222 into the air spaces.

ii. Body Burden, Dose, and RiskCalculations. Risk calculations for ingested

radium were made using RADRISK (see EPA1991, proposed rule) based on annual doserates. For this purpose, EPA computed doserates for specific organs and tissues atspecific ages for an annual unit intake of eachradium isotope (see EPA 2000a). Calculationof body burdens was based on metabolicmodels derived from the radium dial painterstudies. Calculations of absorbed doses inspecific organs or tissues included crossirradiation from radium in all other organs.RADRISK included lifetime cancer riskestimates for high- and low-LET (linearenergy transfer) radiation separately forleukemia, osteosarcomas, sinus tumors, andother solid tumors. These estimates were

taken from the BEIR III and BEIR IV (see EPA1991, proposed rule) reports.

Table II–1 compares the methods used byEPA in 1976 and 1991 to calculate organburdens, doses, and risks from radiumingestion. Bone doses calculated for radium-226 in 1991 were about 33 percent lowerthan those assumed in 1976, and the softtissue doses were about 40 percent lower.Risk estimates for bone per unit dose wereabout 65 percent lower in 1991 than in 1976,and the soft tissue risk estimates were about9 percent lower.

TABLE II–1.—COMPARISON OF DERIVATION OF 1976 AND 1991 MCLS FOR RADIUM

Model 1976 1991

Organ and Tissue Burdens ................................ Calculation of body burdens based on envi-ronmental studies and ratio of intakes.

Calculation of body burdens based ontoxicokinetic models derived from studies ofpatients injected with radium.

Dosimetry ........................................................... Calculation of absorbed dose based on organand tissue burden.

Calculation of absorbed dose based on organor tissue burden and cross irradiation termsfrom all other organs.

Risk Coefficients ................................................. Risk estimated using the geometric mean ofthe absolute and relative risk coefficientsfrom the 1972 BEIR I report.

Risk estimated using the absolute risk coeffi-cient from the 1980 BEIR III report.

b. Basis for the 1991 MCL for Gross AlphaParticle Activity. In 1991, EPA proposed toretain the 15 pCi/L MCL for gross alphaparticle activity, but modify it by excludingradium-226, as well as uranium and radon.The exclusion of uranium and radon wasbased on the fact that the Agency anticipatedsetting separate NPDWRs for thesecontaminants with the finalization of the1991 proposal. The proposed exclusion ofRa–226 was based on the 1991 risk estimatewhich suggested that its unit risk was smallenough not to warrant regulation withingross alpha. The 1991 limit was intended tolimit the lifetime cancer risk due to ingestionof naturally-occurring and man-made alphaparticle emitters in drinking water tobetween 10¥6 and 10¥4, the Agency’s targetrisk range for carcinogens. Specifically, thislimit was based on the followingconsiderations:

Using RADRISK modeling, EPA estimatedthat continuous consumption of 15 pCi/L ofmost alpha particle emitters in drinkingwater at 2 L/day would pose a lifetime cancerrisk between 10¥6 and 10¥4.

EPA performed the risk assessment for thealpha emitters using RADRISK (EPA 1991,proposed rule). The model was used toestimate radiation dose to organs, the dosewas used to calculate risk to organs, and therisks to organs were summed to estimateoverall risk. EPA used RADRISK to calculateconcentrations of alpha emitterscorresponding to lifetime mortality andincidence risks of 10¥4, assuming ingestionof two liters of drinking water daily, andpresented those values in appendix C of the1991 proposed rule.

In determining the risks from ingestion ofalpha emitters in drinking water, EPA wasparticularly interested in polonium-210 andisotopes of thorium and plutonium, becausethese radionuclides had been observed in

water and may cause health effects atrelatively low concentrations.

However, the BEIR IV report concludedthat there was no direct measure of risk formost polonium isotopes based on the humandata, and suggested several possible means ofestimating risk. EPA, as discussed, relied onRADRISK in assessing polonium risk. Themodel estimated that continuous ingestion oftwo liters per day of drinking watercontaining 14 pCi/L would pose a lifetimefatal cancer risk of 1 × 10¥4.

EPA also consulted the BEIR IV report foravailable information on the adverse effectsof thorium. Epidemiological studies ofpatients injected with Thorotrast, a contrastagent consisting of ThO2 and used in medicalradiology from the 1920s to 1955, showedclear increases in liver cancer, as well aspossible increases in leukemia and othercancers. However, the BEIR IV reportdiscussed the limitations of these data forassessing the risk due to other forms ofthorium that might have different metabolicbehaviors and effects. Using RADRISK, EPAestimated that, at a lifetime fatal cancer risklevel of 1 × 10¥4, derived drinking waterconcentrations for thorium isotopes rangedfrom 50 to 125 pCi/L, and noted that thoriumconcentrations in drinking water weregenerally near one pCi/L (EPA, 1991f).

EPA relied on the BEIR IV report forinformation on the health effects ofplutonium isotopes and other transuranicradionuclides that were widely distributed inthe environment in very low concentrationsdue to atmospheric testing of nuclearweapons from 1945 to 1963. The BEIR IVreport concluded that plutonium exposurescaused clear increases in cancers of the bone,liver, and lungs in animals, but not inhumans. At that time, the limited availableepidemiological studies had notdemonstrated a clear association between

plutonium exposure and the development ofcancer in human exposure cases. The reportrecommended that assessing the risks ofplutonium exposure should be based onanalogy with other radionuclides and high-LET radiation exposure risks. UsingRADRISK, EPA estimated that, at a lifetimefatal cancer risk level of 1 × 10¥4, deriveddrinking water concentrations for plutoniumisotopes ranged from about 7 to 68 pCi/L, andnoted that plutonium concentrations indrinking water were generally less than 0.1pCi/L (EPA, 1991f).

c. Basis for the 1991 MCL for Beta Particleand Photon Radioactivity. In 1991, EPAproposed to alter the 4 mrem/yr MCL for betaparticle and photon radioactivity. TheAgency modified the standard by basing thelimit on the committed effective doseequivalent (EDE). (An effective doseequivalent approach adjusts the dose that anindividual organ may receive based on itsradiosensitivity. The less radiosensitive anorgan is, the greater the allowable radiationdose.) The MCL was also modified to includenaturally-occurring beta/photon emitters.The 1991 proposed standard was intended tolimit the lifetime cancer risk due to ingestionof naturally-occurring and man-made betaparticle and photon emitters in drinkingwater to between 10¥6 and 10¥4, theAgency’s target risk range for carcinogens.

Using RADRISK modeling, EPA estimatedthat continuous consumption of two litersper day of drinking water containing aconcentration of beta particle or photonemitting radiation corresponding to 4 mremEDE/yr would pose a lifetime cancer risk ofabout 10¥4.

Comparison of the 1976 Regulation and1991 Proposed Regulation. In 1976, EPAbased the MCL for beta particle and photonemitters on a target dose rate of 4 mrem/yr.The annual average activity concentration of



individual radionuclides and mixtures ofradionuclides resulting in a 4 mrem/yr doseto the total body or any internal organ wasthen calculated. This ‘‘critical organ dose’’radiation protection philosophy was basedon the recommendations of ICRP Publication2 (see EPA 2000a).

The Agency was aware that in 1976, whenexposed to equal doses of radiation, differentorgans and tissues in the human will exhibitdifferent cancer induction rates.Consequently, EPA knew that the lifetimecancer risks for individual radionuclideswould vary widely (from near 10¥7 to 5.6 ×10¥5 because the same dose equivalentwould be applied to different critical organs,resulting in different cancer risks. However,at that time, EPA did not have an acceptedmethod for equalizing risks. In addition,since no dose could be greater than 4 mremto every organ, the associated risk was theceiling for the risk of beta/photon emitters indrinking water.

This was addressed in 1991 when EPAproposed to adopt the effective doseequivalent, or EDE, radiation protectionphilosophy recommended by ICRP (1977)(see EPA 1991, proposed rule). The effectivedose equivalent normalizes radiation dosesand effects on a whole body basis forregulation of occupational exposures. TheEDE is computed as the sum of the weightedorgan-specific dose equivalent values, usingweighting factors specified by the ICRP(1977, 1979; see EPA 1991, proposed rule).By changing to a limit of 4 mrem EDE/yr,EPA was able to derive activityconcentrations for individual beta/photonemitters that corresponded to a more uniformlevel of risk. Using 4 mrem EDE and themetabolically-based dose calculations, thederived concentrations for most beta particleand photon emitters increased in 1991 ascompared to the values calculated in 1976(shown in Table II–3). As a result of derivedconcentrations increasing in 1991, thecorresponding risks increased as well. EPAestimated that, for most of theseradionuclides, the corresponding lifetimefatal cancer risk would be 1 × 10¥4, abouttwice as high as the risk level estimated in1976.

d. Basis for the 1991 Proposed MCL forUranium. In 1991, EPA proposed an MCL of20 µg/L for uranium based on kidney toxicityand a corresponding limit of 30 pCi/L basedon cancer risk. The MCLG was proposed atzero because of the carcinogenicity ofuranium, and the MCL was proposed at themost sensitive endpoint, kidney toxicity. TheMCL was based on kidney effects seen in the30 day study in rats (see EPA 1991, proposedrule).

Using RADRISK modeling, EPA estimatedthat uranium in water posed a cancer risk of

5.9 × 10¥7 per picoCurie per liter, assumingcontinuous intake of water of two liters perday. Concentrations in water of 1.7 pCi/L, 17pCi/L and 170 pCi/L corresponded to lifetimemortality risks of approximately 1 × 10¥6, 1× 10¥5 and 1 × 10¥4, respectively. Aconcentration of 30 pCi/L of uranium-238was thought to be equivalent to about 20micrograms/L, the level considered to beprotective against kidney toxicity (thecorresponding mortality was 5 × 10¥5.

In determining the MCL for uranium in1991, EPA proposed to regulate uranium ata level that would be protective of bothkidney toxicity, resulting from the element’schemical properties, and carcinogenicpotential due to radioactivity. Thecarcinogenic effects of uranium were basedon the effects of ionizing radiation generally,the similarity of uranium to isotopes ofradium, and on the effects of high activityuranium.

C. Today’s Methodology for Assessing RisksFrom Radionuclides in Drinking Water

1. Background

Since 1991, EPA has refined the way inwhich it estimates potential adverse healtheffects associated with ingestion ofradionuclides in drinking water. TheAgency’s new approach uses state-of-the-artmethods, models and data that are based onmore recent scientific knowledge. Comparedwith the approaches used in 1976 and 1991,the revised methodology includes severalsubstantial refinements. Specifically, the newrisk-assessment methodology:

• Accounts for age- and gender-specificwater-consumption rates and radionuclideintakes, and for physiological and anatomicalchanges with age in quantifying costs andbenefits;

• Uses Blue Book (see EPA 2000a) forestimating radiogenic risk: ICRP dosimetrymodel, 1990 vital statistics instead of 1980;

• Uses the most recent age-dependentbiokinetic and dosimetric modelsrecommended by the ICRP; Federal GuidanceReport-13 dynamic input-output metabolicmodel;

• Incorporates the latest information onradiogenic human health effects summarizedby the National Academy of Sciences andother national and international radiation-protection advisory committees;

• Includes updated life tables based ondata from the National Center for HealthStatistics that are used to adjust radionucliderisk estimates for competing causes of death;and

• Uses an improved computer program tohandle the complex calculations of radiationdoses and risks.

Overall, EPA believes that theserefinements significantly strengthen thescientific and technical bases for estimatingrisks, and consequently, for deriving MCLsfor radionuclides. A brief overview of thisnew methodology is summarized later in thissection. Interested individuals are referred totwo EPA publications Estimating RadiogenicCancer Risks (EPA, 1994) and FederalGuidance Report No. 13 (EPA, 1999) fordetailed discussions on the revised riskassessment methodology for radionuclides.Electronic copies of both documents areavailable for downloading at EPA’s web site(http://www.epa.gov/radiation/rpdpubs.htm).

Federal Guidance Report No. 13: (EPA,1999) presents the current methods, models,and calculational framework EPA uses toestimate the lifetime excess risk of cancerinduction following intake or externalexposure to radionuclides in environmentalmedia. The report presents compilations ofrisk coefficients that may be used to estimateexcess cancer morbidity (cancer incidence)and mortality (fatal cancer) risks resultingfrom exposure to radionuclides throughvarious pathways.

The risk coefficients for internal exposurerepresent the incremental probability ofradiogenic cancer morbidity or mortalityoccurring per unit of radioactivity inhaled oringested. For most radionuclides, FederalGuidance Report No. 13 presents riskcoefficients for seven exposure pathways:inhalation, ingestion of food, ingestion of tapwater, ingestion of milk, external exposurefrom submersion in air, external exposurefrom the ground surface, and externalexposure from soil contaminated to aninfinite depth. For some radionuclides,however, only external exposure pathwaysare considered; these include noble gases andthe short-lived decay products ofradionuclides addressed in the internalexposure scenarios.

a. Radium. EPA set the current MCL of 5pCi/L for radium-226 and radium-228,combined, based on limiting the lifetimeexcess total cancer risk to between 5×10¥5

and 2×10¥4. In 1991, EPA proposed separate,and revised, MCLs for radium-226 andradium-228 of 20 pCi/L for each. At thattime, EPA believed that the revised MCLscorresponded to lifetime excess fatal cancerrisks of 1×10¥4 each, or 2×10¥4 combined,assuming lifetime ingestion. The moresophisticated model used today calculates arisk for Ra-228 at 5 pCi/l to be 2×10¥4, andthe risk for 5 pCi/l of Ra-226 to be about7.3×10¥5. Retaining a combined MCL at 5pCi/L would produce the following risksshown in Table II–2.

TABLE II–2.—MORTALITY RISK OF RADIUMS FOR CONCENTRATION COMBINATIONS AT THE MCL

Radium-226 Radium-228 Ra-226 + Ra-228

pCi/L Risk pCi/L Risk pCi/L Risk at 5 pCi/L

0 ............................................................................. 0 5 2.0×10¥4 5 2.0×10¥4

1 ............................................................................. 1.5×10¥5 4 1.6×10¥4 5 1.8×10¥4

2 ............................................................................. 2.9×10¥5 3 1.2×10¥4 5 1.5×10¥4

3 ............................................................................. 4.4×10¥5 2 8.1×10¥5 5 1.3×10¥4

4 ............................................................................. 5.8×10¥5 1 4.1×10¥5 5 9.9×10¥5



TABLE II–2.—MORTALITY RISK OF RADIUMS FOR CONCENTRATION COMBINATIONS AT THE MCL—Continued

Radium-226 Radium-228 Ra-226 + Ra-228

pCi/L Risk pCi/L Risk pCi/L Risk at 5 pCi/L

5 ............................................................................. 7.3×10¥5 0 0 5 7.3×10¥5

b. Alpha Emitters. Both the current and1991 proposed MCLs for alpha-emittingradionuclides permit up to 15 pCi/L of alphaparticle radioactivity in drinking water fromindividual and multiple alpha emitters. EPAestablished the current gross alpha MCL of 15pCi/L (including radium-226 and excludingradon and uranium) to account for the riskfrom radium-226 at 5 pCi/L (the radiumregulatory limit) plus the risk frompolonium-210, which the Agency believedwas the next most radiotoxic element in theuranium decay chain. The current riskestimated (FGR–13) indicates that the unitrisk for Ra-226 is large enough to warrant itsinclusion in gross alpha, as thought in 1976.

In 1991, EPA thought that exposure to 10pCi/L of polonium-210 posed a lifetime fatalcancer risk comparable to that fromcontinuous lifetime ingestion of about 1 pCi/

L of radium-226, that is, between 0.5 and2×10¥4. In 1991, EPA based the revised,adjusted gross alpha MCL on revised doseand risk calculations which indicated thatthe 15 pCi/L limit posed a lifetime cancerrisk for most alpha emitters that fell withinEPA’s acceptable risk range of between 10¥6

and 10¥4.The current estimate of risk from

polonium-210 at 7.0 pCi/L is 1×10¥4. Therisk for radium-226 at 6.8 p/L is also 1×10¥4.When the current rule was written, 10 pCi/L of polonium-210 was believed to beequivalent to 1 pCi/L of radium-226;however, the risks are now equivalent. Thuspolonium is ten times the risk it was thoughtto be relative to radium-226. Retaining a 15pCi/L standard including radium-226 ensuresthat the risk of 15 pCi/L will not increase byallowing greater polonium (up to 15 pCi/L)

in addition to the radium-226 in the radiumstandard. As expected, a uniform picoCurielimit results in widely differing risks (EPA2000a).

c. Beta/Photon Emitters. As discussedelsewhere in this document, EPA is able tocalculate the risks from individual beta/photon emitters using the FGR–13methodology. It is now possible to calculatea risk equivalent to the current picoCurielimit for each beta/photon emitter.Appropriate adjustments are then possible inkeeping with the original risk maximum of5.6×10¥5. The derived concentration valuesfor the beta particle and photon emitters from1976 rule and 1991 proposal in comparisonto today’s newest risk model using 5.6×10¥5

mortality are found in Table II–3.




d. Uranium. Since the 1991 proposal, anumber of new studies have been publishedin peer-reviewed journals. A literature searchwas conducted and covered the time periodbetween January 1991 to July, 1998.Databases searched were TOXLINE,MEDLINE, EMBASE, BIOSIS, TSCATS andCurrent Contents (see EPA 2000a). Theresults of the literature search were reviewedand articles were identified, retrieved andreviewed and analyzed. Subsequently, theToxicological Profile for URANIUM (Update)was published extending the database toSeptember 1999 (see EPA 2000a).

i. Health Effects in Animals. The potentialtoxic effects of uranium following oralexposures have been evaluated in recentanimal studies (see EPA 2000a). In a 28-dayrange-finding study, male and femaleSprague-Dawley rats (15/sex/group) wereadministered concentrations of 0, 0.96, 4.8,24, 124, or 600 mg uranyl nitrate/L (UN/L)in drinking water for a period of 28 days.Results of the study showed no significantdose-related effects on body weight gain,food intake, fluid consumption, clinicalsigns, or hematological parameters of treatedanimals when compared with controlanimals. Histologic examinations indicatedno statistically significant differences in theincidence of a particular lesion in animals inthe 600 mg UN/L treatment group whencompared with animals in the control group.However, a slight increase in the number ofaffected animals in the 600 mg UN/L groupwas observed, when compared with thecontrol group.

As discussed in the Technical SupportDocument (EPA 2000a), the long-term effectsof exposure to low-levels of uranium indrinking water has been demonstrated.Female rabbits and male albino rats wereexposed to 0, 0.02, 0.2, and 1 mg/kg uranylnitrate for 12 months or 0.05, 0.6, 6, and 60mg/L uranyl nitrate for 11 months,respectively. Results of the study indicated adecrease in acid phosphatase activity in thespleens of rabbits in the 1 mg/kg group, butnot in rats, when compared to controls. Astatistically-significant (p<0.05) increase inserum alkaline phosphatase activity wasobserved by the eleventh month of exposurein rats in the 6 and 60 mg/L groups, whencompared with controls. A statistically-significant decrease in the content of nucleicacids in the renal and hepatic tissues wasobserved in rats in the 60 mg/L group and inrabbits in the 1 mg/kg group, when comparedwith controls.

ii. Health Effects in Humans. Recentepidemiological studies have evaluated theeffects observed in humans exposed touranium in the drinking water (see EPA2000a). These studies demonstrate therelationship between uranium levels in thedrinking water and urine albumin, anindicator of renal dysfunction, wasevaluated. Three sites were selected for thecontrols (site 1) and the exposed groups (sites2 and 3), with mean uranium water levels of0.71, 19.6 and 14.7 µg/L reported for sites 1,2 and 3, respectively. An index of uraniumexposure was estimated for each study

participant by multiplying the uraniumconcentration in the water supply by theaverage number of cups consumed at eachresidence and the total number of years atthat residence. Based on the results of alinear regression analysis, which includedterms for age, diabetes, sex, smoking, and theuse of water filters and softeners, astatistically-significant association wasreported for cumulative exposure to uraniumand urine albumin levels. However, theauthors noted that for most of the studyparticipants, the urine albumin levels werewithin the range of normal values.

A recent study of a village in Nova Scotia(see EPA 2000a) demonstrated the renaleffects following chronic exposure touranium in the drinking water. Two groupswere evaluated, a low exposure group(uranium levels < 1/L) and a high exposuregroup (uranium levels > 1µg/L). Twenty-fourhour and 8-hour urine samples werecollected and evaluated for uranium,creatinine, glucose, protein, b2-microglobulin(BMG), alkaline phosphatase (ALP), gammaglutamyl transferase (GGT), lactatedehydrogenase (LDH), and N-acetyl-b-D-glucosaminidase (NAG). Statisticallysignificant positive correlations werereported with uranium intake for glucose(males, females and pooled data), ALP(pooled data) and BMG (pooled data). Noother statistically significant differences werereported. Based on these results, the authorsconcluded that the proximal tubule was thesite of uranium nephrotoxicity.

In June 1998, a workshop was held by theUSEPA to discuss issues associated withassessing the risk associated with uraniumexposure and updating the RfD and MCLGfor uranium. The numerous technical issuesassociated with the development of a riskassessment for uranium in drinking waterwere discussed. Based on these discussions,it was apparent that there is a range of valuesfor each factor used in the development ofthe RfD and MCL for uranium. However,based upon the input received at theworkshop and the most current information,EPA believes that the LOAEL for renal effectsin male rats of 0.06 mg U/kg/day reportedcould be used for the development of an RfDfor uranium (see EPA 2000a). The relativesource contribution (RSC) was revised to 80percent (0.8). The total uncertainty factor wasdetermined to be about 100 (about 3 foranimal to human extrapolation, about 10 forintraspecies differences, about 1 for a lessthan lifetime study, and about 3 for the useof a LOAEL), with the body weight of 70kilograms (kg) and daily water consumptionof two liters used in the calculation. Theseassumptions are consistent with the datapresented at the workshop and appear to bereasonable and justifiable. EPA believes thesefactors allow for the calculation of a safelevel of uranium in drinking water (in termsof chemical toxicity).

The application of the total uncertaintyfactor of 100 to the LOAEL of 0.06 mg/kg/dayresults in an RfD of 0.6 ug uranium/kg/day.The RfD can be used to determine the MCLby multiplying the RfD by body weight (70

kg) and RSC (0.8) and dividing by waterconsumption (2 L), resulting in a value of 17ug uranium/L, which can be rounded off to20 /L.

2. Consideration of Sensitive Sub-populations: Children’s EnvironmentalHealth

In compliance with Executive Order 13045‘‘Protection of Children from EnvironmentalHealth Risks and Safety Risks’’ (62 FR 19885,April 23, 1997), risks to children fromradionuclides have been considered. There isevidence that children are more sensitive toradiation than adults, the risk per unitexposure in children being greater than inadults.

Risk coefficients used by the Agency forradiation risk assessment explicitly accountfor these factors. The age-specific, organ-specific risk per unit dose coefficients usedin the lifetime risk model apply theappropriate age-specific sensitivitiesthroughout the model. The model alsoincludes age-specific changes in organ massand metabolism. The risk estimate at any ageis the best estimate for that age. In developingthe lifetime risks, the model uses the lifetable for a stationary population. Use of thelife table allows the model to account forcompeting causes of death and age-specificsurvival. These adjustments make thelifetime risk estimate more realistic.

At the same time, consumption rates offood, water and air are different betweenadults and children. The lifetime riskestimates for radionuclides in water use age-specific water intake rates derived fromaverage national consumption rates whencalculating the risk per unit intake. Since theintake by children is usually less than theintake by adults, it tends to partially mitigatethe greater risk in children compared toadults when evaluating lifetime risk.

D. References

EPA, 1999. Cancer Risk Coefficients forEnvironmental Exposure to Radionuclides,Federal Guidance Report No. 13. USEnvironmental Protection Agency,Washington, DC, 1999. Uranium IssuesWorkshop—Sponsored by United StatesEnvironmental Protection Agency,Washington, DC ; June 23–24, 1998.

USEPA. ‘‘Technical Support Document forthe Radionuclides Notice of DataAvailability.’’ Draft. March, 2000. (EPA2000a)

Appendix III—Analytical Methods

Table III–1 briefly summarizes theregulatory events associated with:

• The testing procedures for regulatedradionuclides approved in 1976;

• Major analytical additions or changesproposed or discussed in the 1991radionuclides rule;

• Testing procedures and protocolsapproved in the March 5, 1997—radionuclides methods rule (62 FR 10168,cited in 40 CFR 141.25); and

• Items discussed in today’s NODA.



TABLE III–1.—BRIEF SUMMARY OF THE REGULATORY EVENTS ASSOCIATED WITH RADIOCHEMICAL METHODS

1976 National primary drinkingwater regulations

July 18, 1991—Radionuclidesproposed rule

March 5, 1997-Radionuclidemethods final rule Today’s notice of data availability

The 1976 NPDWR approved:* Radiochemical methods to

analyze for gross alpha-par-ticle activity, radium-226,total radium, gross beta-par-ticle activity, strontium-89and -90, cesium-134 anduranium

* Defined the detection limit(DL) as the required measureof sensitivity and listed therequired DL for each regu-lated radionuclide

The July 18, 1991—radio-nuclides rule proposed:

* Fifty-six additional methodsfor compliance monitoring ofradionuclides

* Guidance for the sample han-dling, preservation and hold-ing times that were cited inthe 1990 U.S.EPA ‘‘Manualfor the Certification of Lab-oratories Analyzing DrinkingWater’’

* The use of practical quantita-tion limits (PQLs) and ac-ceptance limits as the meas-ures of sensitivity forradiochemical analysis

The March 5, 1997 final rulefor radionuclide methods:

* Approved 66 additional radio-nuclide techniques for grossalpha-particle activity, ra-dium-226, radium-228, ura-nium, cesium-134, iodine-131, and strontium-90

Responded to comments re-garding the analytical meth-ods (excluding radon) re-ceived from the July 18,1991 proposed radionuclidesrule

Updates the public on changes that haveoccurred regarding radiochemical meth-ods of analysis since the 1991 proposedrule. The updates discussed in today’sNODA include:

* A brief discussion of the analytical meth-ods updates which were promulgated bythe Agency on July 18, 1997 final rule.

* Guidance for the sample handling, preser-vation and holding times listed in the1997 U.S.EPA ‘‘Manual for the Certifi-cation of Laboratories Analyzing DrinkingWater.’’

* Recommendations for the analysis ofshort-lived, alpha-emitting radioisotopes(i.e., radium-224).

* Revised cost estimates for radiochemicalanalysis.

* The Agency’s intent to continue to use thedetection limits defined in the 1976 ruleas the required measures of sensitivity.

* Response to some of the comments onthe 1991 proposed radionuclides.

* The externalization of the PerformanceEvaluation Program.

The Agency’s plans to implement a Per-formance Based Measurement System.

A. The Updated 1997 LaboratoryCertification Manual

A revised version of the certificationmanual was published in 1997 (EPA 815–B–

97–001, EPA 1997b). Table III–2 lists theguidance for sample handling, preservation,holding times, and instrumentation whichappeared in this manual. Table III–2 also

includes additional recommendations forradiochemical instrumentation (footnoted bythe number 6), which the Agency isrequesting comment on.

TABLE III–2.—SAMPLE HANDLING, PRESERVATION, HOLDING TIMES AND INSTRUMENTATION

Parameter Preservative 1 Container 2 Maximum holdingtime 3 Instrumentation 4

Gross Alpha ......................................... Concentrated HCl or HNO3 to pH <25 P or G 6 months A, B or GGross Beta .......................................... Concentrated HCl or HNO3 to pH <25 P or G 6 months A or GRadium-226 ......................................... Concentrated HCl or HNO3 to pH <2 .. P or G 6 months A, B, C 6, D or GRadium-228 ......................................... Concentrated HCl or HNO3 to pH <2 .. P or G 6 months A, B 6, C 6 or GUranium natural ................................... Concentrated HCl or HNO3 to pH <2 .. P or G 6 months A 6, F, G 6, or OCesium-134 ......................................... Concentrated HCl to pH <2 ................ P or G 6 months A, C or GStrontium-89 and -90 .......................... Concentrated HCl or HNO3 to pH <2 .. P or G 6 months A or GRadioactive Iodine-131 ........................ None .................................................... P or G 8 days A, C or GTritium .................................................. None .................................................... G 6 months EGamma/Photon Emitters ..................... Concentrated HCl or HNO3 to pH <2 .. P or G 6 months C

1 It is recommended that the preservative be added to the sample at the time of collection. It is recommended that samples be filtered if sus-pended or settleable solids are present at any level observable to the eye prior to adding preservative. This should be done at the time of collec-tion. If the sample has to be shipped to a laboratory or storage area, however, acidification of the sample (in its original container) may be de-layed for a period not to exceed 5 days. A minimum of 16 hours must elapse between acidification and start of analysis.

2 P = Plastic, hard or soft; G = Glass, hard or soft.3 Holding time is defined as the period from time of sampling to time of analysis. In all cases, samples should be analyzed as soon after collec-

tion as possible. If a composite sample is prepared, a holding time cannot exceed 12 months.4 A = Low background proportional system; B = Alpha and beta scintillation system; C = Gamma spectrometer [Ge(Hp) or Ge (Li)]; D = Scin-

tillation cell system; E = Liquid scintillation system; F = Fluorometer; G = Low background alpha and beta counting system other than gas-flowproportional; O = Other approved methods (e.g., laser phosphorimetry and alpha spectrometry for uranium).

5 If HCl is used to acidify samples which are to be analyzed for gross alpha or gross beta activities, the acid salts must be converted to nitratesalts before transfer of the samples to planchets.

6 Additional instrumentation that was not listed in the USEPA 1997 ‘‘Manual for the Certification of Laboratories Analyzing Drinking Water.’’

B. Recommendations for Determining thePresence of Radium-224

To determine the presence of the short-lived radium-224 isotope (half life ∼3.66days), the Agency recommends using one ofthe following several options.

1. Radium-224 by Gamma Spectrometry andAlpha Spectrometry

(a) Gamma Spectrometry. Radium-224 canbe specifically determined by gammaspectrometry using a suitably preparedsample. In this method a precipitate in which

the radium isotopes are concentrated isgamma counted. The primary advantage ofthis technique is specificity for radiumgamma rays, radium-224 included. Otheradvantages of this method include:



• a simple sample preparation wereradium isotopes are concentrated fromsamples 1 liter or larger;

• specificity for the radium-224 isotopebased on a unique gamma energy;

• optimal accuracy and precision if thesample is counted within 72 hours ofcollection (40 hours is recommended);

• and is cost competitive with the grossmethods because a single count rather thatthree counts (see the gross alpha methodsdiscussion) is necessary to measure theradium-224 in a routine sample.

A gamma spectrometry method byStandard Methods is currently pending butfor now the reader is referred to the methodused by Parsa. (Parsa, 1998).

(b) Alpha Spectrometry. The alphaspectrometry method measures alphasemitted by radium-224 and its alpha emittingdaughters. The alpha spectrometry method,used for the USGS occurrence survey (seeappendix I and EPA 2000a), was a slightmodification of an existing method (see EPA2000a). Using an appropriate tracer (e.g. Ba-133), barium and radium isotopes areseparated from other radionuclides andinterferences using cation ion exchangechromatography. A prepared sample,counted for approximately 100 minutes usingalpha spectrometry, can be used to measurethe radium-224 in the sample and is capableof good accuracy and precision. Other alphaspectrometry techniques, similar to themodified method used for the USGSoccurrence survey, should be sufficient forthe detection of radium-224. It is costcompetitive with the gross methods(discussed next) because a single count ratherthan three (for gross methods) is sufficient tofor measurement of radium-224.

2. Gross Radium Alpha (Co-precipitation)Within 72 Hours

The presence of radium-224 can bedetermined indirectly using the radium-224

half-life decay and the gross radium alphatechnique. Gross radium co-precipitationmethods, like EPA 903.0, concentrate radiumisotopes by co-precipitation, separatingradium and radium-like isotopes frompotential interferences. Relative toevaporative methods, the co-precipitationtechnique can be used for larger (> 1 L)sample sizes with a resulting increase in themethod sensitivity. Initial analysis within 72hours after sample collection (40 hoursrecommended for optimal data quality) usingthe co-precipitation methods yield results,reflecting both alpha-emitting radiumisotopes (radium-224 and radium-226). Forthese to produce unambiguous results,radium-224 must be the dominant isotopepresent, i. e. the ratio of radium-224 toradium-226 must be three or greater. If thisis the prevailing composition, the estimatedcontribution of radium-224 to the overallvalue can be ascertained by recounting thesample at 4 or 8 days intervals andcalculating the change in the measuredactivity. The noted change will show adecrease with a 4 day half-life indicative ofRa-224. Formulas are available to calculatethe initial radium-224 concentration presentin the sample when collected. Theadvantages of this technique include:

• enhanced sensitivity (≥1 L samples);• it does not require additional analyst

training;• it is specific for radium isotopes; and• the resulting precipitate can be measured

by a number of techniques, includingproportional counting, alpha scintillationcounting, or gamma counting.

3. Evaporative Gross Alpha-Particle AnalysisWithin 72 Hours

The radium-224 isotope, when inequilibrium with its decay progeny, emitsfour alpha particles. Three of these alphaparticles equilibrate almost immediately(within 5 minutes) after sample preparation

and add to or amplify the sample count rate.This count rate amplification can beexploited for the measurement of radium-224in a sample at low concentration (<15 pCi/L). The presence of the radium-224radioisotope in drinking water may beascertained by performing an initialevaporative gross alpha-particle analysiswithin 72 hours (40 hours recommended)after sample collection. In the absence of anyother alpha-emitting nuclide (e.g., uraniumor radium-226) and if the gross alpha-particlevalue is above the MCL, the sample may bere-counted at 4- and 8-day intervals todetermine if the observed decrease in activityfollows the 3.66 day half-life of radium-224.A decrease in the gross alpha value with a4-day decay rate indicates the likely presenceof radium-224. Formulas are available tocalculate the concentration of radium-224 inthe initial sample. The advantages of thisoption include:

• the method is similar to the generalmethod for evaporative gross alpha;

• it requires no special training of theanalyst; and

• it can be a definitive test if other alpha-emitting nuclides are known to be absent.

The Agency recognizes that analysis withinthe 72-hour time frame creates difficulties inshipping and handling and may increase theprice of the analysis.

C. Revised Cost Estimates for RadiochemicalAnalysis

The cost estimates for radiochemicalanalysis from the 1991 proposed rule and therevised cost estimates are shown in Table III–3.

TABLE III–3.—THE 1991 AND 1999 ESTIMATED COSTS OF ANALYSES FOR RADIONUCLIDES

Radionuclides Approximatecost $ (1991)1

Approximatecost $ (1999)2

Gross Alpha and beta .............................................................................................................................................. 35 45Gross alpha—coprecip. ........................................................................................................................................... 35 45Radium-226 ............................................................................................................................................................. 85 90Radium-228 ............................................................................................................................................................. 100 110Uranium (total) ......................................................................................................................................................... 45 48 (LP)Uranium (isotopic) .................................................................................................................................................... 125 128 (AS)Radioactive Cesium (-134) ...................................................................................................................................... 100 125Radioactive Strontium .............................................................................................................................................. 105 144Total Strontium (-89 and -90) .................................................................................................................................. ...... 153Radioactive Iodine -131 ........................................................................................................................................... 100 131Tritium ...................................................................................................................................................................... 50 60Gamma/Photon Emitters ......................................................................................................................................... 110 142

Source:1 56 FR 33050; July 18, 1991.2 USEPA, 2000a.Abbreviations: LP = laser phosphorimetry; AS = alpha spectrometry.Note: Estimated costs are on a per-sample basis; analysis of multiple samples may have a lower cost.



1 Package plants are skid mounted factoryassembled centralized treatment units that arrive onsite ‘‘virtually ready to use’’. Package plants offerseveral advantages. First, since they combineelements of the treatment process into a compactassembly (such as chemical feeders, mixers,flocculators, basins, and filters), they tend to requirelesser construction and engineering costs. Anotheradvantage is that many package plant technologiesare becoming more automated and thus can be lessdemanding of operators than their fully engineeredcounter-parts (EPA 1998b).

2 Point-of-entry (POE) treatment units treat all ofthe water entering a household or other building,with the result being treated water from any tap.Point-of-use (POU) treatment units treat only the

water at a particular tap or faucet, with the resultbeing treated water that one tap, with the other tapsserving untreated water. POE and POU treatmentunits often use the same technological conceptsemployed in the analogous central treatmentprocesses, the main difference being the muchsmaller scale of the device itself and the flows beingtreated (EPA 1998b).

D. The Detection Limits as the RequiredMeasures of Sensitivity

Table III–4 cites the detection limits or therequired sensitivity for the specificradioanalyses that were listed in the 1976rule and are also cited in 40 CFR 141.25.

TABLE III–4.—REQUIRED REGULATORYDETECTION LIMITS FOR THE VAR-IOUS RADIOCHEMICAL CONTAMI-NANTS (40 CFR 141.25)

Contaminant Detection limit (pCi/L)

Gross Alpha .............. 3Gross Beta ................ 4Radium-226 ............... 1Radium-228 ............... 1Cesium-134 ............... 10Strontium-89 .............. 10Strontium-90 .............. 2Iodine-131 ................. 1Tritium ....................... 1,000Other Radionuclides

and Photon/Gamma Emitters.

1⁄10th of the ruleNIPDWR 1976table IV–2A and 2B

E. References

Parsa, B., 1998. Contribution of Short-livedRadionuclides to Alpha-ParticleRadioactivity in Drinking Water and TheirImpact on the Safe Drinking Water ActRegulations, Radioactivity andRadiochemistry, Vol. 9, No. 4, pp. 41–50,1998. USEPA, 1991. National PrimaryDrinking Water Regulations; Radionuclides;Proposed Rule. Federal Register. Vol. 56, No.138, p. 33050. July 18, 1991.

USEPA, 1997a. National Primary DrinkingWater Regulations; Analytical Methods forRadionuclides; Final Rule and ProposedRule. Vol. 62, No. 43, p. 10168. March 5,1997.

USEPA, 1997b. ‘‘Manual for theCertification of Laboratories AnalyzingDrinking Water.’’ EPA 815–B–97–001. 1997.

USEPA, 2000a. ‘‘Technical SupportDocument for the Radionuclides Notice ofData Availability.’’ 2000.

Appendix IV—Treatment Technologies andCosts

A. Introduction

This section describes updates to EPA’sprevious evaluations of the feasibility andcosts of treatment technologies for theremoval of radionuclides from drinkingwater. Prior to this update, the latestevaluation was the 1992 ‘‘Technologies andCosts document’’ for radionuclides indrinking water (EPA 1992). The updates tothe 1992 radionuclides Technologies andCosts document comprise an updatedTechnologies and Costs Document (EPA1999a) and a radium compliance cost study(EPA 1998a), which are described later inthis section. This section also describes otherrelevant documents, including the1998Federal Register notice of the ‘‘SmallSystems Compliance Technology List’’(SSCTLs) for the currently regulatedradionuclides (63 FR 42032) and itssupporting guidance document (EPA 1998b).Both of the documents supporting the SSCTs

can be obtained on-line at ‘‘http://www.epa.gov/OGWDW/standard/tretech.html’’.

The SSCTLs for the meeting the MCLs forcombined radium-226 and radium-228, grossalpha emitters, and combined beta andphoton emitters are included in‘‘Announcement of Small SystemCompliance Technology Lists for ExistingNational Primary Drinking Water Regulationsand Findings Concerning VarianceTechnologies,’’ published in the FederalRegister on August 6, 1998 (63 FR 42032).The supporting guidance document citedpreviously includes information regardingsmall systems treatment and waste disposalconcerns relevant to radionuclidecontaminants and was made publiclyavailable on September 15, 1998. Furtherevaluations of small systems treatmenttechnology applicability and affordabilityhave been done since the SSCTLs forradionuclides were published, including ananalysis of SSCTs for uranium (EPA 1999b).These evaluations are summarized later inthis section.

B. Treatment Technologies Update

1. Updates on Performance of Technologiesfor Removal of Regulated Radionuclides andUranium

One of the purposes of the update to theradionuclides Technologies and Costs (T&C)document (EPA 1999a) was to update thetreatment technology performance sections ofthe 1992 radionuclides T&C document. Thepeer-reviewed literature revealed no newsignificant sources of information regardingperformance for the previously describedtechnologies, nor did it reveal literatureregarding any new treatment technologies forradionuclides in drinking water. Both the1992 and 1999 radionuclides T&C documentsinclude performance evaluations of the BATsproposed in 1991 for the regulatedradionuclides and uranium (56 FR 33050, Jul.18, 1991) and additional technologies thatwere reviewed as potential BATs for the 1991proposed rule, but that were not proposed asBAT for various reasons.

Although the 1999 T&C documentconcludes that the peer-reviewed literaturedescribes no new technologies since the 1992T&C document was completed, there havebeen some developments that are significant.In particular, both package plant 1

technologies, including those equipped withremote control/communication capabilities,and point-of-entry (POE)/point-of-use (POU)versions 2 of existing technologies have

become more widely applicable for use forcompliance. This is true both because ofimprovements in these technologiesthemselves (NRC 1997) and since the 1996SDWA explicitly allows package plants andPOE/POU devices to be used as compliancetechnologies for small systems (section1412.b.4.E). Package plant technologies andPOE/U technologies are discussed in moredetail in the Technical Support Document(EPA 2000a).

2. Treatment Technologies Evaluated asCompliance Technologies for Radionuclides

The following technologies are reviewed inthe 1999 radionuclides T&C document: (1)for radium, the 1991 proposed Best AvailableTechnologies (BATs), which are limesoftening, ion exchange, and reverse osmosis;and two other applicable technologies withsignificant radium removal data,electrodialysis reversal and greensandfiltration; (2) for uranium, the 1991 proposedBATs, which are coagulation/filtration, ionexchange, lime softening, and reverseosmosis; and two other applicabletechnologies, electrodialysis reversal andactivated alumina; (3) for gross alpha particleactivity, the 1991 proposed BAT, which isreverse osmosis; and one other applicabletechnology, ion exchange; and (4) for betaparticle activity and photon radioactivity, the1991 proposed BATs, which are ionexchange and reverse osmosis. No othertechnology studies pertinent to total beta andphoton activity were found, but this is largelydue to the fact that treatment applicabilitydepends on what specific beta and photonemitters are present and so should beevaluated on a case-by-case basis. Thisconsideration also applies to gross alphaactivity. It is likely that reverse osmosis,being applicable to a broad range of inorganiccontaminants, including radionuclidecontaminants, is the best alternative forsituations where multiple radionuclidesoccur.

3. Data on Additional TreatmentTechnologies

The 1999 radionuclides T&C documentdoes not identify any new treatmenttechnologies for radionuclides, but doesprovide information on two additionalvariants of coagulation/filtration for uraniumremoval: direct filtration and in-linefiltration.

4. Small Systems Compliance TechnologyList and Guidance Manual for the RegulatedRadionuclides and Uranium

The 1996 SDWA identifies three categoriesof small drinking water systems, thoseserving populations between 25 and 500, 501and 3,300, and 3,301 and 10,000. In additionto BAT determinations, the SDWA directsEPA to make technology assessments for eachof the three small system size categories inall future regulations establishing an MCL or



treatment technique. Two classes of smallsystems technologies are identified for futureNational Primary Drinking Water Regulations(NPDWRs): compliance technologies andvariance technologies.

Compliance technologies may be listed forNPDWRs that promulgate MCLs or treatmenttechniques. In the case of an MCL,‘‘compliance technology’’ refers to atechnology or other means that is affordable(if applicable) and that achieves compliance.Possible compliance technologies includepackaged or modular systems and point-of-entry (POE) or point-of-use (POU) treatmentunits, as described previously.

Variance technologies are only specifiedfor those system size/source water qualitycombinations for which no technology meetsall of the criteria for listing as a compliancetechnology (section 1412(b)(15)(A)). Thus,the listing of a compliance technology for asize category/source water combinationprohibits the listing of variance technologiesfor that combination. While variancetechnologies may not achieve compliancewith the MCL or treatment techniquerequirement, they must achieve themaximum reduction that is affordable

considering the size of the system and thequality of the source water. Variancetechnologies must also achieve a level ofcontaminant reduction that is ‘‘protective ofpublic health’’ (section 1412(b)(15)(B)).

In the case of the currently regulatedradionuclides, i.e., combined radium-226 and-228, gross alpha activity, and total beta andphoton activity, there are no variancetechnologies allowable since the SDWA(section 1415(e)(6)(A)) specifically prohibitssmall system variances for any MCL ortreatment technique which was promulgatedprior to January 1, 1986. The Variance andExemption Rule describes EPA’sinterpretation of this section in more detail(see 63 FR 19442; April 20, 1998).

Small systems compliance technologies forthe currently regulated radionuclides,combined radium-226 and -228, gross alphaemitters, and total beta and photon activity,were listed and described in the FederalRegister on August 6, 1998 (EPA 1998a) andin an accompanying guidance manual (EPA1998b). Small systems compliancetechnologies for uranium were also evaluated(EPA 1999a). Small systems compliancetechnologies (SSCTs) for uranium were

evaluated in terms of each technology’sremoval capabilities, contaminantconcentration applicability ranges, otherwater quality concerns, treatment costs, andoperational/maintenance requirements. TheSSCT list for uranium is technology specific,but not product (manufacturer) specific.Product specific lists were determined to beinappropriate due to the potential resourceintensiveness involved. Information onspecific products will be available throughanother mechanism. EPA’s Office of Researchand Development has a pilot project underthe Environmental Technology Verification(ETV) Program to provide treatment systempurchasers with performance data fromindependent third parties.

Tables IV–1 and IV–2 summarize the smallsystems compliance technologies listed inthe 1998 SSCTL for combined radium-226,and -228, gross alpha emitters, total beta andphoton activity. Table IV–1 is shown as itwill be updated when uranium is regulated.Table IV–1 describes limitations for each ofthe listed technologies and Table IV–2 listsSSCTs for each contaminant.

TABLE IV–1.—LIST OF SMALL SYSTEMS COMPLIANCE TECHNOLOGIES FOR RADIONUCLIDES AND LIMITATIONS TO USE

Unit technologiesLimitations(see foot-

notes)Operator skill level required1 Raw water quality range and

considerations1

1. Ion Exchange (IE) ................................. (a) Intermediate .............................................. All ground waters.2. Point of Use (POU) IE .......................... (b) Basic ......................................................... All ground waters3. Reverse Osmosis (RO) ........................ (c) Advanced .................................................. Surface waters usually require pre-filtra-

tion.4. POU RO ................................................ (b) Basic ......................................................... Surface waters usually require pre-filtra-

tion.5. Lime Softening ...................................... (d) Advanced .................................................. All waters.6. Green Sand Filtration ........................... (e) Basic .........................................................7. Co-precipitation with Barium Sulfate .... (f) Intermediate to Advanced ......................... Ground waters with suitable water quality.8. Electrodialysis/Electrodialysis Reversal .................... Basic to Intermediate ................................ All ground waters.9. Pre-formed Hydrous Manganese Oxide

Filtration.(g) Intermediate .............................................. All ground waters.

10. Activated alumina ............................... (a), (h) Advanced .................................................. All ground waters; competing anion con-centrations may affect regeneration fre-quency.

11. Enhanced coagulation/filtration ........... (i) Advanced .................................................. Can treat a wide range of water qualities.

1 National Research Council (NRC). Safe Water from Every Tap: Improving Water Service to Small Communities. National Academy Press.Washington, D.C. 1997.

Limitations Footnotes to Table IV–2: Technologies for Radionuclides:a The regeneration solution contains high concentrations of the contaminant ions. Disposal options should be carefully considered before

choosing this technology.b When POU devices are used for compliance, programs for long-term operation, maintenance, and monitoring must be provided by water util-

ity to ensure proper performance).c Reject water disposal options should be carefully considered before choosing this technology. See other RO limitations described in the

SWTR Compliance Technologies Table.d The combination of variable source water quality and the complexity of the water chemistry involved may make this technology too complex

for small surface water systems.e Removal efficiencies can vary depending on water quality.f This technology may be very limited in application to small systems. Since the process requires static mixing, detention basins, and filtration,

it is most applicable to systems with sufficiently high sulfate levels that already have a suitable filtration treatment train in place.g This technology is most applicable to small systems that already have filtration in place.h Handling of chemicals required during regeneration and pH adjustment may be too difficult for small systems without an adequately trained

operator.i Assumes modification to a coagulation/filtration process already in place.

Table IV–2 lists the Small SystemsCompliance Technologies for the currentlyregulated radionuclides. Technology

numbers refer to the technologies listed inTable IV–1.



TABLE IV–2.—COMPLIANCE TECHNOLOGIES BY SYSTEM SIZE CATEGORY FOR RADIONUCLIDE NPDWRS (AFFORDABILITYNOT CONSIDERED, EXCEPT FOR URANIUM, DUE TO STATUTORY LIMITATIONS)

Contaminant

Compliance technologies1 for system sizecategories (population served)

25–500 501–3,300 3,300–10,000

Combined radium-226 and radium-228 ....................................................................................... 1, 2, 3, 4, 5, 6,7, 8, 9

1, 2, 3, 4, 5, 6,7, 8, 9

1, 2, 3, 4, 5, 6,7, 8, 9

Gross alpha particle activity ........................................................................................................ 3, 4 3, 4 3, 4Total beta particle activity and photon activity, average annual concentration .......................... 1, 2, 3, 4 1, 2, 3, 4 1, 2, 3, 4Uranium ....................................................................................................................................... 1, 2, 4, 10, 11 1, 2, 3, 4, 5,

10, 111, 2, 3, 4, 5,

10, 11

Note: 1 Numbers correspond to those assigned to technologies found in the table ‘‘List of Small Systems Compliance Technologies for the Cur-rently Regulated Radionuclides.’’

C. Waste Treatment, Handling and DisposalGuidance

In the proposed radionuclides rule of July1991, EPA referenced a 1990 EPA draft reportentitled ‘‘Suggested Guidelines for Disposalof Drinking Water Treatment WastesContaining Naturally-OccurringRadionuclides’’ (EPA 1990). That 1990 reportoffered guidance to system managers,engineers, and State agencies responsible forthe safe handling and disposal of treatmentwastes that, in many cases, were notspecifically addressed by any statute. Thatguidance report was later updated in 1994(EPA 1994).

The guidance provided information on thefollowing: (1) Background on water treatmentprocesses and characteristics of wastesgenerated; (2) rationale for radiationprotection, including citation of programsand regulations affecting other sources ofsuch waste; (3) guidelines for severalmethods of disposal of solid and liquid typewastes containing the subject radionuclides;and, (4) the specification of practicalguidance to protect workers and others whomay handle or be exposed to water-treatmentwastes containing radiation abovebackground levels.

The Technical Support Document (EPA2000a) discusses disposal methods andissues, including comments received inreference to the 1990 ‘‘Suggested Guidelinesfor Disposal of Drinking Water TreatmentWastes Containing Naturally-OccurringRadionuclides,’’ and the 1994 update to thisreport.

D. Unit Treatment Cost UpdatesTreatment costs for coagulation/filtration

(including direct filtration and in-linefiltration), lime softening, ion exchange,reverse osmosis, electrodialysis reversal,greensand filtration, point-of-use (POU)reverse osmosis, POU ion exchange, andpoint-of-entry cation exchange were updatedin the appendix to the 1999 radionuclidesT&C document. This update includes land-cost considerations and waste-disposal costestimates. Cost estimates were made usingstandard EPA treatment technology costingmodels. Outputs were updated to currentdollars using standard engineering costingindices, e.g., the Bureau of Labor’s Chemicaland Allied Products Index. Costs forindividual technologies were analyzed interms of water usage, removal efficiency,interest rate, and other variables.

In addition to cost model updates, EPA hasperformed a study of the actual costs oftreatment and other compliance measures forthe radium standard (EPA 1998c), whichprovided a ‘‘snapshot’’ of the costs incurredby water systems in complying with theexisting combined radium-226 and radium-228 MCL. Studies of this nature allow EPAto compare modeled costs used in regulatoryimpact assessments with real-world data forthe purposes of model validation and costestimate amendments. They also allow EPAto check assumptions about the prevalence ofuse of particular water-treatmenttechnologies.

The study comprises data compiled fromcontacts with water-treatment personnel,State representatives, and EPA Regionalrepresentatives within EPA Regions 5 (IL, IN,MI, MN, OH, and WI) and 8 (CO, MT, ND,SD, UT, and WY). Specifically, data wereobtained regarding water systems inCalifornia, Florida, Idaho, Illinois, Indiana,Ohio, Wisconsin, and Wyoming. StateAgencies and EPA Regional offices identified136 systems as having water sources withcombined radium-226 and radium-228 abovethe MCL of 5 pCi/L. Of these, 55 of thesystems were contacted, of which 29 wereeither treating for radium or were in theprocess of selecting a treatment method. Theremaining systems were either further behindin treatment selection plans or pursuingother compliance measures. All of thesystems that were currently treating forradium were in compliance with the MCL.Twenty-six of these systems responded withcost data, of which 17 were small systems(design flow < 1 mgd). Thirty-five percent ofthe small systems reported were usingreverse osmosis which, at an average totaltreatment cost of $3.02 per thousand gallons,was the most expensive treatment technologyidentified. Other treatment options used werelime softening and ion exchange. These hadaverage total treatment costs of $2.36 and$0.73 per thousand gallons, respectively.Unit costs are discussed in more detail in theTechnical Support Document (EPA 2000a).

EPA requests comments on its analysis oftreatment technologies, costs, and treatmentresiduals disposal.

E. References

National Research Council (NRC). SafeWater From Every Tap: Improving WaterService to Small Communities. NationalAcademy Press. Washington, DC. 1997.

USEPA. Office of Drinking Water.Suggested Guidelines for Disposal ofDrinking Water Treatment Wastes ContainingNaturally-Occurring Radionuclides (July1990 draft).

USEPA. National Primary Drinking WaterRegulations; Radionuclides; Proposed Rule.Federal Register. Vol. 56, No. 138, p. 33050.July 18, 1991.

USEPA. Technologies and Costs for theRemoval of Radionuclides from PotableWater Supplies. Prepared by Malcolm Pirnie,Inc. July 1992.

USEPA. Office of Ground Water andDrinking Water. Suggested Guidelines forDisposal of Drinking Water Treatment WastesContaining Radioactivity (June 1994 draft).

USEPA. Announcement of Small SystemsCompliance Technology Lists for ExistingNational Primary Drinking Water Regulationsand Findings Concerning VarianceTechnologies. Federal Register. Vol. 63, No.151, p. 42032. August 6, 1998. (EPA 1998a).

USEPA. ‘‘Small System ComplianceTechnology List for the Non-MicrobialContaminants Regulated Before 1996.’’ EPA–815–R–98–002. September 1998. (EPA1998b).

USEPA. ‘‘Actual Cost for Compliance withthe Safe Drinking Water Act Standard forRadium-226 and Radium-228.’’ Final Report.Prepared by International Consultants, Inc.July 1998. (EPA 1998c).

USEPA Technologies and Costs for theRemoval of Radionuclides from PotableWater Supplies. Draft. Prepared byInternational Consultants, Inc. April, 1999.(EPA 1999a).

USEPA. ‘‘Small System ComplianceTechnology List for the Radionuclides Rule.’’Prepared by International Consultants, Inc.Draft. April 1999. (EPA 1999b).


Appendix V—Economics and ImpactsAnalysis

A. Overview of the Economic Analysis

1. Background

Analysis of the costs, benefits, and otherimpacts of regulations is required under theSafe Drinking Water Act Amendments of1996, Executive Order 12866 (RegulatoryPlanning and Review), and EPA’s internalguidance for regulatory development. These



1 The NIRS database is stratified into fourcategories: systems serving between 25–500persons, 501—3,300 persons, 3,301–10,000 persons,and 10,001–1,000,000 persons. Because of the smallsample size used to describe the larger systems, ourmodel uses only three categories: we combine thetwo categories for systems serving greater than3,301 persons into a single category.

2 Model systems describe the universe of drinkingwater systems by breaking it down into discrete‘‘system size categories’’ by population served.There are nine size categories: 25–100 personsserved; 101–500; 501–1,000; 1,001–3,300; 3,301–10,000; 10,001–50,000; 50,001–100,000; 100,000–1,000,000; > 1,000,000. Within each size category,the systems are described by a single set of ‘‘typicalcharacteristics’’ by source water type (groundversus surface water) and ownership type (publicversus private ownership). These characteristicsinclude the average and design flows and thedistribution of numbers of entry points per system.

3 Unit compliance costs models include watertreatment cost models (e.g., W/W Cost and theWATER model) and models for other complianceoptions, like alternate water well sources andpurchasing water. For a discussion of the standardEPA water treatment cost models, see EPA 1999d.

4 Decision trees are models of the relativeprobabilities that water systems will chooseparticular compliance actions when in violation.The probabilities are estimated based onconsiderations of source water type, system size,water quality, required removal efficiency, unitcosts, treatment issues (e.g., co-treatment and pre-/post-treatment requirements), and residualsdisposal costs and issues.

5 While the treatments installed to eliminate grossalpha and combined radium may also reduceuranium levels, we do not quantify these impactsin this analysis. We make no adjustment for threereasons. First, the NIRS data suggest that systemswith elevated levels of gross alpha or combinedradium rarely report uranium concentrations abovelevels of concern. Second, some types of treatmentused to remove gross alpha or radium are lesseffective in removing uranium. Lastly, radium anduranium occur at higher levels under very differentaquifer conditions: radium tends to occur at highlevels in water with low dissolved oxygen and hightotal dissolved solids, while uranium occurs athigher levels in oxygen-rich waters with low totaldissolved solids (see the Technical SupportDocument, EPA2000a).

requirements are new relative to the 1991proposal for revisions to the existing NationalPrimary Drinking Water Regulations(NPDWRs) for radionuclides.

The actions that are anticipated to haveregulatory impacts are evaluated in thissection. These actions are: (1) the correctionthe monitoring deficiency for combinedradium-226 and radium-228; and (2) theestablishment of a uranium NPDWR with anMCL of 20 µg/L; or (3) the establishment ofa uranium NPDWR with an MCL of 40 µg/L; or (4) the establishment of a uraniumNPDWR with an MCL of 80 µg/L. See‘‘Combined Ra-226 and Ra-228’’ in thetoday’s NODA (section III, part F) for adiscussion of the monitoring corrections thatwill be finalized for the combined radium-226 and radium-228 (‘‘combined radium’’)NPDWR. See ‘‘Uranium’’ in the NODA(section III, part H) for a discussion of theoptions being considered for finalization forthe uranium NPDWR.

2. Economic Analysis of the RegulatoryActions Being Considered for Radionuclidesin Drinking Water

The economic analysis summarized heresupports the finalization of the 1991Radionuclides proposal. The more detailedeconomic analysis (the Health RiskReduction and Cost Analysis, EPA 2000b)may be obtained from the Water Docket, asdescribed in the Introduction to today’sNODA (see ADDRESSES). It provides central-tendency estimates of national costs andbenefits and presents information on the datasources and analytic approaches used,including a qualitative discussion of theanalytical limitations and uncertaintiesinvolved. Further uncertainty analyses willbe performed to support the analysessummarized here and will be reported in thepreamble to the final rule. It should be notedthat these additional uncertainty analyses arenot expected to alter regulatory decisions.

The basic steps in a comprehensiveeconomic analysis include: (1) Estimatingbaseline conditions in the absence ofrevisions to the regulations; (2) predictingactions that water systems will use to meeteach regulatory option (the ‘‘decision tree’’);(3) estimating national costs resulting fromcompliance actions; (4) estimating nationalbenefits resulting from compliance; and (5)assessing distributional impacts and equityconcerns. In today’s NODA, we presentpreliminary estimates of national costs andbenefits for the options evaluated, focusingon monitoring and compliance costs andreductions in cancer risks. Other nationalcosts and benefits (e.g., state administrativecosts and risk reductions from incidentaltreatment of co-occurring contaminants) andpotential distributional impacts are describedqualitatively (see EPA 2000a and EPA2000b).

The first step in the economic analysis,defining the analytical baseline, requires thatwater systems be apportioned into severalgroups based on their predicted levels ofradionuclides and the current monitoringscheme. In the case of the radionuclidesNPDWRs, this provides unusual challenges.This is partly due to the fact that severalcommunity water systems are not complyingwith the existing regulations, which is

reflected in the occurrence database used forthis work (the National Inorganics andRadionuclides Survey, ‘‘NIRS’; see EPA 1991,proposed rule and EPA 2000a). Also, asdiscussed in the Introduction to today’sNODA, there are weaknesses in the currentmonitoring requirements that has lead to asituation in which some water systemshaving combined radium levels greater thanthe MCL of 5 pCi/L will not have knowledgeof this fact (and hence are not presently inviolation of the combined radium NPDWR).Both of these influences, the existingunresolved radionuclides NPDWR violationsand the monitoring deficiencies, must beaccounted for in the analytical baseline.

The regulatory baseline and otheranalytical baselines are benchmarks tomeasure regulatory impacts against.Generating a national-level contaminantoccurrence profile is an important part of thisbenchmarking process. The database used asthe basis for this model, NIRS, is describedin appendix I of today’s NODA (Occurrence).The analysis of regulatory impacts uses thissystem-size stratified baseline occurrencemodel 1 to estimate the percentages of watersystems with contaminant levels withinspecified values (e.g., 30 to 50% above theMCL). This information is then combinedwith other models to estimate the compliancecosts and benefits associated with eachoption. Examples of models relevant tonational costs estimation include ‘‘modelsystems 2,’’ compliance cost equations 3, andthe compliance action prediction model or‘‘decision trees 4.’’ Examples of modelsrelevant to risk reduction and benefitsestimation include the risk models describedin appendix II and the risk reductionvaluation models described in the TechnicalSupport Document (EPA 2000a).

The analytical baseline for combinedradium reflects full compliance with the

existing regulations as written, which havebeen fully enforceable since the 1986reauthorization of the SDWA. This approachassumes that, in the absence of any changesto the radionuclides NPDWRs, EPA and theStates will eventually ensure that all systemsfully comply with the existing regulations.This approach allows us to separate out thepredicted number of systems with combinedradium levels in excess of the MCL that haveknowledge of the violation (‘‘systems inviolation’’) from the predicted number ofsystems that have levels in excess of theMCL, but that would not have knowledge ofthis under the current monitoringrequirements. Since uranium is not currentlyregulated, no such corrections are necessary.It was also determined that treatmentinstalled to remove the other radionuclidesshould not significantly impact the uraniumanalytical baseline.5

B. Approach for Assessing Occurrence, Risksand Costs for Community Water Systems

1. Assessing OccurrenceTo develop estimates of the baseline

radionuclides occurrence profile forcommunity water systems, we began byextrapolating from data obtained throughEPA’s National Inorganics and RadionuclidesSurvey (NIRS). This survey measuredradionuclide concentrations at 990community ground water systems between1984 and 1986. For detailed information onthe design of NIRS, see Longtin 1988. Fordetailed information on how NIRS was usedin this work, see the background documents(EPA 2000a and 2000b).

We made adjustments to the NIRS data toaddress certain limitations, including (1) thesmall size of the sample of systems servingpopulations greater than 3,300 persons; (2)the decay of radium-224 prior to analysis ofthe NIRS water samples; (3) the need toconvert mass measurements of uranium toactivity levels; and, (4) the lack ofinformation on surface water systems. Theanalyses and discussions that followconcentrate on CWSs serving retailpopulations of less than one million persons.Discussions of preliminary and futureeconomic impacts analyses of Non-TransientNon-Community Water Systems (NTSCsystems) and the largest CWSs follow later inthis section. The two occurrence approacheswe examined are described next. For adiscussion of the relative strengths andweaknesses of the two approaches toestimating occurrence, see the TechnicalSupport Document (EPA 2000a).



6 See the Technical Support Document (EPA2000a) and the HRRCA (EPA 2000b).

7 This analysis focuses on changes in cancer risksfrom tap water ingestion. Individuals may beexposed to radionuclides in drinking water throughother pathways (e.g., inhalation while showering),and uranium may have toxic effects on the kidneys;however, we expect that any changes in these types

of risks will be, while not insignificant, muchsmaller than the changes in cancer risks fromingestion, and hence discuss them onlyqualitatively in this analysis.

8 ‘‘Unit risk factors and ‘‘unit risks’’ refer to therisk per pCi/L in drinking water. They are notestimates of cancer incidence per se, but rather areindicators of the ‘‘potency’’ of a radionuclide. To

get estimates of the risks of cancer incidence for anexposed population, the unit risk factors must beused in conjunction with a radionuclide drinkingwater occurrence model. These population risksrefer to the estimated numbers of excess statisticalcases of cancer that a population will face under agiven set of exposure assumptions.

2. ‘‘Direct Proportions Approach’’ toEstimating Occurrence

Because of uncertainties related toextrapolating from the NIRS database tonational-level estimates, we applied twoapproaches for estimating the national-levelcentral-tendency occurrence estimate. First,we assumed that national occurrence isdirectly proportional to the occurrence levelsmeasured in NIRS. For example, if theradionuclide concentration in one percent ofthe samples from NIRS representing aparticular water system size category aregreater than the MCL, we assumed that onepercent of all systems in that size class wouldbe out of compliance at the national level (Itis worth noting that using NIRS toextrapolate to the State or regional level isnot valid, since NIRS was designed to berepresentative at the national-level, but not atthese other levels). In cases where thisapproach predicts ‘‘zero probability’’ of non-compliance for a system size category (i.e., nosamples in NIRS were above the MCL beingconsidered), this approach is flawed, sincethe expectation is that this finding actuallyreflects a small probability, not ‘‘zeroprobability.’’ In other words, in situationswhere ‘‘zero impact’’ is predicted, it is muchmore likely that a very small number of watersystems will be impacted compared to true‘‘zero impact.’’ For this reason, we also useda mathematical model to simulate theoccurrence distribution, in which these ‘‘zeroprobabilities’’ are replaced by estimatedsmall probabilities.

3. ‘‘Lognormal Model Approach’’ toEstimating Occurrence

The second approach recognizes that‘‘true’’ radionuclides occurrence will mostlikely be spread over a range wider than thatobserved in the survey. This approachassumes that ‘‘probability plots’’ of the NIRSdata are lognormally distributed. Aprobability plot compares the radionuclideconcentration for the various samples to theprobability of a given sample having thatlevel or less, where this probability isestimated from the actual occurrence datafrom NIRS. An assumption of lognormalitymeans that a probability plot for thelogarithms of the radionuclide levels wouldbe expected to be linear (fall on a straightline).

Inspection of the NIRS data suggests thatit is distributed in a roughly lognormal

pattern, with most systems reportingconcentration levels well below the MCLs ofconcern. Several other studies also suggestthat the distribution of radionuclideoccurrence in drinking water systems islikely to follow a lognormal distribution 6, sothis assumption should be robust in mostcases. If the NIRS data were perfectlylognormally distributed, both approacheswould lead to similar estimates ofoccurrence. This is usually the case.However, it should be noted that thereinstances of significant deviations betweenthe two approaches. For example, the directproportions approach predicts that 0.4 % ofthe systems serving more than 500 personswill be impacted (61 systems) by an MCL of20 pCi pCi/L for uranium, whereas thelognormal model approach predicts that1.8% of systems will be impacted (255systems), amounting to a difference inprediction of almost 200 impacted watersystems in this size category. There areseveral possible explanations for thisdeviation, but the important point is that theuse of both approaches allows the data gapto be recognized and fully considered.

A statistical software package (‘‘Stata’’) wasused to estimate a lognormal distribution thatbest fits the data for systems in each sizeclass. We then used the fitted log means andlog standard deviations of the resultingdistributions to estimate the number ofsystems out of compliance with eachregulatory option using standard statisticalequations. More detail regarding theoccurrence models and the estimation of thenumbers of impacted systems can be foundelsewhere (EPA 2000a and 2000b).

4. Assessing Risk

After determining the number of systemsout of compliance with each regulatoryoption under consideration, we assessed therisk reductions that would result from thesesystems taking actions to come intocompliance. The approach for the riskanalysis begins with the development ofintrinsic ‘‘risk factors’’ for each group ofradionuclides. These risk factors arecomposites that involve multiplying EPA’sbest estimates of unit mortality andmorbidity cancer risk coefficients (risk perpCi) for each group of radionuclides bystandard assumptions regarding drinkingwater ingestion to determine the risk factorsassociated with drinking water exposure (risk

per pCi/L). We then applied the individualrisk factors 7 to the estimates of the reductionin exposure associated with each regulatorychange under consideration, taking intoaccount the population exposed. Thecalculation of risk factors from riskcoefficients and a discussion of exposureassumptions are detailed elsewhere (EPA2000a). The risk factors (per pCi/L indrinking water) used in the risk reductionanalyses are summarized in Table V–1.

The unit 8 risk factors applied in thisanalysis refer to the aggregated small changesin the probability of incurring cancer over alarge population. These unit probabilities canbe interpreted in two ways: as the unitlifetime excess probability of cancerinduction averaged over age and gender forall individuals in a population or as the riskfor a statistically ‘‘averaged individual.’’ Itshould be noted that no one individual istruly average, since the averaging also occursover gender. Given a model of radionuclideoccurrence, the population risks of excesscancer incidence can be estimated before andafter a given regulatory option for theindividuals comprising the population, withthe difference being equal to the reducedrisk. These reductions in individual cancerincidence probabilities may then be summedover the population to indicate the central-tendency number of ‘‘statistical cancer casesavoided’’ annually. However, it should bekept in mind that for many reasons,including the large variance associated withsuch risk factors, it is impossible to ‘‘checkthis prediction’’ in any meaningful way. Ininterpreting reduced risks for given options,it is arguably best to think of them in termsof reduced average ‘‘individual excess risk,’’rather than ‘‘cases avoided,’’ for the reasonsjust described. For example, it is much easierto understand the idea that an individual’saverage lifetime risk of developing cancerdue to exposure to radionuclides in drinkingwater has been reduced from three in tenthousand to one in ten thousand for anumber of water systems under a givenoption then to understand that an average of0.5 cancer cases are avoided annually at thenational level for that option. The use of‘‘individual excess risk’’ avoids much theconfusion about ‘‘statistical cases,’’ which areconceptually difficult to understand.

TABLE V–1.—AVERAGE INDIVIDUAL RISK FACTORS, AVERAGE WATER CONSUMPTION (1.1 L/PERSON/DAY) (PER PCI/L)

Regulatory option

Morbidity Mortality

Lifetime in-gestion

Annual in-gestion

Lifetime in-gestion

Annual in-gestion

Gross Alpha: changes in monitoring requirements (weighted average of Ra-224 andRa–226) ........................................................................................................................ 5.24E–06 7.48E–08 3.26E–06 4.65E–08

Gross Alpha: changes in MCL (Ra–224 only) ................................................................ 4.77E–06 6.81E–08 2.90E–06 4.15E–08



9 The monitoring deficiency will be corrected byrequiring the separate analysis of Ra-228 forsystems with gross alpha levels below 5 pCi/L.

TABLE V–1.—AVERAGE INDIVIDUAL RISK FACTORS, AVERAGE WATER CONSUMPTION (1.1 L/PERSON/DAY) (PER PCI/L)—Continued

Regulatory option

Morbidity Mortality

Lifetime in-gestion

Annual in-gestion

Lifetime in-gestion

Annual in-gestion

Combined Radium: changes in monitoring requirements (weighted average of Ra–226 and Ra–228) ......................................................................................................... 2.30E–05 3.28E–07 1.63E–05 2.32E–07

Combined Radium: changes in MCL (Ra–228 only) ...................................................... 2.98E–05 4.26E–07 2.12E–05 3.03E–07Uranium: establish MCL (simple average of U–234, U–235, and U–238) ..................... 1.95E–06 2.79E–08 1.26E–06 1.81E–08

5. Estimating Monetized Benefits

In this section, we summarize theinformation used in estimating monetizedbenefits. A description of the methodologyused for these estimates is found in theTechnical Support Document (EPA 2000a),which provides background information on:(1) The economic concepts that provide thefoundation for benefits valuation; (2) themethods that are typically used byeconomists to value risk reductions, such aswage-risk, cost of illness, and contingentvaluation studies; (3) the approach forvaluing the reductions in fatal cancer risksand nonfatal cancer risks; (4) the use of thesetechniques to estimate the value of the riskreductions attributable to the regulatoryoptions for radionuclides in drinking water;and (5) the limitations and uncertaintiesinvolved in the estimation. For more detailon the methodology employed, see theHealth Risk Reduction and Cost Analysis(HRRCA, EPA 2000b).

This benefits analysis is based on two basictypes of valuation: fatal cancer riskreductions and non-fatal cancer riskreductions. Fatal cancer risk reductions arevalued in terms of the ‘‘value of a statisticallife’’ (VSL), which does not refer to the valueof an identifiable individual, but rather refersto the value of small reductions in mortalityrisks over a large population. For example,let us assume that a regulatory option resultsin a risk reduction of ‘‘one statistical fatalcancer case.’’ This refers to the summation ofsmall risk reductions over a large number ofpersons such that the summation equals ‘‘onecase’’ (say, one hundred thousand personseach face a risk reduction of 1/100,000).Using our methodology, the resulting benefitswould be equal to ‘‘one statistical life.’’Continuing the example, if each person werewilling to pay $20 for such a risk reduction(1/100,000), the resulting VSL would be $2million ($20 times 100,000 persons).However, since there is no direct informationon what persons are willing to pay for therisks we are interested in, we must useindirect methods for estimating the VSL. Thecurrently accepted methodology involvestransferring the VSL from studies of the wageincreases that persons ‘‘demand’’ in exchangefor accepting jobs with slightly higherchances of accidental fatality (‘‘wage-riskstudies’’). There are a number of assumptionsinvolved in making this transfer, which arediscussed in more detail in the backgrounddocumentation (EPA 2000a and 2000b).

Valuing nonfatal cancer risk reductions isoften done with ‘‘cost of illness studies,’’which examine the actual direct (e.g.,

medical expenses) and indirect (e.g., lostwork or leisure time) costs incurred byaffected individuals. Unfortunately, thisvaluation does not measure the ‘‘willingnessto pay’’ to avoid nonfatal cancers, but ratherassumes that benefits are equal to theavoided costs. The studies used andassumptions involved are discussedelsewhere (EPA 2000a and 2000b).

Because of the uncertainties involved invaluations, we used an estimate of the rangeof values of reductions in fatal and non-fatalrisks attributable to the radionuclidesregulations using the following estimates(1998 dollars):

Fatal Risk Reduction Valuations (‘‘Value ofa Statistical Life’’, VSL):

Best Estimate: Value of fatal risk reductions= Statistical lives saved * $5.9 million perstatistical life.

Low End Estimate: Value of fatal riskreductions = Statistical lives saved * $1.5million per statistical life.

High End Estimate: Value of fatal riskreductions = Statistical lives saved * $11.5million per statistical life.

Non-Fatal Risk Reduction ValuationsBest Estimate: Value of nonfatal risk

reductions (medical costs only) = Statisticalcases averted * $0.10 million.

Low End Estimate: Value of nonfatal riskreductions (medical costs only) = Statisticalcases averted * $0.09 million.

High End Estimate: Value of nonfatal riskreductions (medical costs only) = Statisticalcases averted * $0.11 million.

6. Estimating the Costs of Compliance

The last component of the analysisinvolves estimating the costs of compliancefor each regulatory option. The options underconsideration will increase the costs ofmonitoring for all regulated systems, as wellas require a small fraction of the systems totake action to reduce the contaminant levelsin their finished water to achievecompliance. Examples of compliance actionsinclude installing treatment, purchasingwater from another system, changing thewater source used (e.g., installing a newwell), blending the contaminated water withother source water that is below the MCL,and, in cases where the contaminated well isnot essential to meet capacity, stoppingproduction from the contaminated well. Thecost analysis models both new capital costsand, and when appropriate, incrementaloperations and maintenance costs for thisvariety of compliance options. The inputsused in the cost analysis and a comparisonof the modeled costs for treatment, alternatesource, purchased water to case studies can

be found in the Technical Support Document(EPA 2000a) and elsewhere (EPA 1998a).

C. Summary of Annual Costs and Benefits

1. Estimates of Costs and Benefits forCommunity Water Systems

The following results reflect the regulatoryoptions that are currently being considered.Results for the other options that wereanalyzed (correction of monitoringdeficiencies for gross alpha and changes toMCLs for gross alpha and Ra-228), but thatEPA does not plan to adopt, are located inthe Technical Support Document (EPA2000a). In addition to EPA’s preferredoptions, we have included all results in theTechnical Support Document to allowinterested stakeholders to comment on theseother options, if desired.

Table V–2 shows the summarized resultsfor EPA’s analysis of risk reductions, benefitsvaluations, and costs of compliance (see EPA2000b for a break-down of the summary bywater system size). The risk reductions andcost estimates are based on the estimatedrange of numbers of community watersystems predicted to be out of compliancewith each of the regulatory options assessed.The ranges shown reflect the two occurrencemodel methodologies previously described,the ‘‘direct proportions’’ and ‘‘lognormalmodel’’ approaches. The ranges inoccurrence predictions necessarily result inranges of estimates for risk reductions,benefits valuations, and compliance costs.There are two ranges shown for values ofcancer cases avoided, the ‘‘best-estimaterange,’’ based on the best-estimate of riskreduction valuations, and the ‘‘low/high-estimate range,’’ which reflects the use of thetwo occurrence models and the uncertaintyin the risk reduction valuations (‘‘low-end’’versus ‘‘high-end’’ estimates). These rangesdo not reflect uncertainty in other modelinputs, like risk factors in the case of riskreduction estimates and treatment unit costsin the case of compliance costs. Quantitativeuncertainty analyses for risk reductions,benefits, and compliance costs will beconducted and reported in the preamble tothe final rule. EPA expects that theseuncertainty analyses will not impact finaldecisions.

Eliminating the combined radium-226/-228monitoring deficiency 9 is predicted to leadto 210 to 250 systems out of compliance withan MCL of 5 µg/L, affecting 33,000 to 460,000



persons. Implementing an MCL of 20 µg/L foruranium is predicted to impact 830 to 970systems, affecting 470,000 to 2,100,000persons. An MCL for uranium of 40 µg/L ispredicted to impact 300 to 430 systems,affecting 47,000 to 850,000 persons; 80 µg/Lis predicted to impact 40 to 170 systems,affecting 7,000 to 170,000 persons. Theseestimates for uranium are based on theassumption that the activity-to-mass ratio indrinking water is 1:1. EPA’s current best-estimate for the average activity-to-mass ratiofor the various uranium isotopes in drinkingwater is 0.9. EPA will update this assumptionfor the uranium options in the RegulatoryImpact Assessment supporting the rulefinalization. However, the impact is expectedto be small. For example, using the lognormaloccurrence distribution model for the 40 µg/L option, an assumption of an activity-to-mass ratio of 0.9 results in an estimatednumber of impacted systems of 370, adecrease of only 12–13%.

The estimated risk reduction range for theoption addressing the combined radiummonitoring deficiency is 0.3 to 0.5 cancercases avoided annually, with an associatedannual monetized benefits range of one totwo million dollars. The risk reductionsestimated for the uranium options range from0.2 to 2 cases avoided annually for an MCLof 20 mg/L, 0.04 to 1.5 cases avoidedannually for an MCL of 40 µg/L, and 0.01 to1 case avoided annually for an MCL of 80 µg/

L. The associated annual monetized benefitsfor the uranium options range from 0.6 to 8million dollars (20 mg/L), 0.1 to 6 milliondollars (40 µg/L), and less than 0.1 to 4million dollars (80 µg/L).

Annual compliance costs range from 20 to30 million dollars for the option addressingthe combined radium monitoringdeficiencies. Annual compliance costs for theuranium options range from 30 to 140million dollars for an MCL of 20 mg/L, 6 to60 million dollars for an MCL of 40 µg/L, and5 to 30 million dollars for an MCL of 80 µg/L.

As demonstrated by this analysis theestimated range of central-tendency annualcompliance costs exceed the ranges ofcentral-tendency annual monetized benefitsfor all options. This is not surprising giventhat most of the systems impacted are smallwater systems, which tend to have muchhigher per customer compliance costsrelative to large systems, while the percustomer risk reduction is independent ofwater system size. Except in cases where riskreductions are quite large, it is predictablethat estimated annual costs will outweighestimated annual benefits for small watersystems (given the current methodologies forestimating benefits). However, it should bepointed out that all of the regulatory optionsbeing considered have associated lifetimemorbidity risks near or in excess of one inten thousand, which is the upper bound on

the preferred risk range according to EPA’spolicies on regulating drinking watercontaminants. In the case of uranium, it isalso important to recognize that there may beconsiderable non-quantified (notmonetizable) benefits associated withreductions in kidney toxicity risks. If suchbenefits were quantified, it is likely that thenet benefits would be more favorable for alluranium options.

Some commenters may argue that costsand benefits considerations should lead tothe conclusion that the finalization of thecorrection of the combined radiummonitoring deficiencies and/or theestablishment of a NPDWR for uranium arenot warranted. However, this conclusionwould lead to a situation where customers ofmany ground water systems face lifetimemorbidity risks greatly in excess of theacceptable risk upper limit of one in tenthousand. According to EPA’s policies, theproper use of this flexibility should lead toregulatory decisions that have associatedrisks that are within or acceptably close toEPA’s longstanding goals of limiting excesslifetime morbidity risks to the range of onein a million to one in ten thousand, exceptunder unusual circumstances. EPA solicitscomment on this interpretation of costs andbenefits for the finalization of the 1991radionuclides proposal.




10 Table 2.4, Uncertainty Categories for SelectedRisk Coefficients. Federal Guidance Report 13(1999).

11 A latency period refers to the average amountof time that passes between the beginning ofexposure to a carcinogen or multiple carcinogensand the on-set of fatal cancer. There is considerableuncertainty in estimating a ‘‘typical latency period’’for the options studies here for many reasons,including the large ranges in estimated latencyperiods for given cancer types and the largeuncertainty involved in predicting which type ortypes of cancer will result from exposure to a givenradionuclide in drinking water. It is also uncertainwhat discounting rate would be appropriate in thissituation. Some may argue that discounting isentirely inappropriate (a rate of zero) and othersmay argue that typical financial discount rates areappropriate (3 to 7%).

12 This estimate is based on total capital costsranging from approximately $135,000 to $550,000per MGD of flow. The estimate assumes typicalrelationships between design and average dailyflows and a capital discount rate of 3 or 7%.

2. Uncertainties in the Estimates of Benefitsand Costs

The models used to estimate costs andbenefits related to regulatory measures haveuncertainty associated with the modelinputs. The types and uncertainties of thevarious inputs and the uncertainty analysesfor risks, benefits, and costs are qualitativelydiscussed later in this section.

a. Uncertainties in Risk ReductionEstimates. For each individual radionuclide,EPA developed a central-tendency riskcoefficient that expresses the estimatedprobability that cancer will result in anexposed individual per unit of radionuclideactivity (e.g., per pCi/L) over the individual’slifetime (assumed to be 70 years). Two typesof risks are considered, cancer morbidity,which refers to any incidence of cancer (fatalor non-fatal), and cancer mortality, whichrefers to a fatal cancer illness. For thisanalysis, we used the draft September 1999risk coefficients developed as part of EPA’srevisions to Federal Guidance Report 13(FGR–13, EPA 1999e). FGR–13 compiled theresults of several models predicting thecancer risks associated with radioactivity.The cancer sites considered in these modelsinclude the esophagus, stomach, colon, liver,lung, bone, skin, breast, ovary, bladder,kidney, thyroid, red marrow (leukemia), aswell as residual impacts on all remainingcancer sites combined.

There are substantial uncertaintiesassociated with the risk coefficients in FGR–13 (EPA 1999e): researchers estimate thatsome of the coefficients may change by afactor of more than 10 if plausible alternativemodels are used to predict risks. While thereport does not bound the uncertainty for allradionuclides, it estimates that the central-tendency risk coefficients for uranium-234and radium-226 may change by a factor ofseven depending on the models employed toestimate risk.10 Ranges that reflectuncertainty and variability in the riskcoefficients will be used in a Monte Carloanalysis of risk reductions and benefits, theresults of which will be reported in thepreamble to the final rule.

In addition, as previously described inappendix I, ‘‘Occurrence,’’ the availableoccurrence data do not provide informationon the contribution of individualradionuclides or isotopes to the totalconcentrations of gross alpha or uranium.Therefore, there is uncertainty involved inthe assumptions about which radionuclidescomprise the reported gross alpha or uraniumactivity. These and other uncertaintiesrelated to occurrence information (e.g.,uncertainty in extending the NIRS databaseresults to the national level) will also beincorporated in a Monte Carlo analysis ofbenefits to estimate the range of uncertaintysurrounding the central-tendency estimates.Other inputs that will be used in the MonteCarlo analysis of benefits are the age- andgender-dependent distributions of wateringestion, which are used in estimatinglifetime exposure, and the credible range forthe ‘‘value of a statistical life.’’ This

uncertainty analysis is not expected to alterthe regulatory options discussed in today’sNODA.

b. Effects of the Inclusion of a LatencyPeriod and Other Factors on the Estimate ofBenefits. The expected analytical impacts ofthe inclusion of other factors, e.g., a cancer-latency period, cancer premiums, and non-quantifiable benefits have been discussed inthe recent radon proposed NPDWR (64 FR59295). The relevant points are summarizedbriefly here and in more detail in theTechnical Support Document for theRadionuclides NODA (USEPA 2000a).

There are several potentially importantsources of uncertainty related to thevaluations of risk reductions for theregulatory options examined. Since themortality valuations dominate the estimatedbenefits, factors that affect the VSL are mostimportant. Factors that may affect the VSLinclude discounting due to cancer latencyperiods,11 cancer-related premiums that mayraise the value of statistical life, and othercurrently non-quantifiable benefits. Cancerlatency-related discounting would beexpected to decrease the present VSL, whilecancer premiums would tend to increase thepresent VSL. It is not clear whether aninclusion of all of these factors would beexpected to result in a lower or higherpresent VSL. However, EPA is currentlyworking with the Science Advisory Board(SAB) to determine how to best include thesefactors, whether the inclusion is quantitativeor qualitative.

c. Uncertainty in Compliance CostEstimates. Regarding uncertainty in thecompliance cost estimates, these estimatesassume that most systems will installtreatment to comply with the MCLs, whilerecent research suggests that water systemsusually select compliance options likeblending (combining water from multiplesources), developing new ground waterwells, and purchasing water (EPA 1998a andc, EPA 2000a). Preliminary data (202compliance actions from 14 States) on nitrateviolations suggest that only around a quarter(25%) of those systems talking action inresponse to a nitrate violation installedtreatment, while roughly a third developed anew well or wells. The remainder eithermodified the existing operations (10–15%),blended (15%), or purchased water (15–20%). Similar data for radium violationsfrom the State of Illinois (77 complianceactions) indicate that around a quarter ofsystems taking action installed treatment,while the majority (50–55%) purchasedwater, with the remainder (20–25%) either

installing a new well, blending, or stoppingproduction from the contaminated well orwells. The prevalence of the use of these non-treatment options is a cross-cutting issue forfuture Regulatory Impact Assessments andprobably will not be resolved before theradionuclides NPDWR is finalized. EPA isfollowing up with this study and will reportthe results at a later date.

While these ‘‘other than treatment’’ optionsmay cost as much as or more than treatmentin some cases, they are expected to be lessexpensive on average, which largely explainstheir prevalence as compliance options. Forexample, EPA has recently estimated thecosts associated with developing municipalwells to range from $0.08/kgal to $0.46/kgal,depending on system size, geologic setting,and other site specific parameters (EPA1999b), with an average of $0.23/kgal forsystems serving between 501 and 1,000persons and $0.17/kgal for systems servingbetween 10,001 and 50,000 persons.12 Thesecosts include testing and drilling, steelcasings with cement lining, pumps,including electrical connections andcontrols, and a pump shelter. For smaller,non-municipal PWS systems, we estimatethat wells could cost from 10 to 80 percentof the costs presented for municipal systems.As shown in the Technical SupportDocument (EPA 2000a), these productioncosts are much lower than those for typicaltreatment, especially for small systems.When feasible, selection of such options mayreduce compliance costs significantly. TheTechnical Support Document includes dataon other non-treatment options likepurchasing water and blending.

Preliminary uncertainty analyses suggestthat variability in the unit compliance costsand decision tree assumptions dominate theover-all cost variability. To evaluate thepotential variability in the compliance costestimates, a Monte Carlo analysis willsupport the Regulatory Impact Assessmentfor the final rule. Inputs that influence costvariability include:

• Numbers of total systems in the varioussystem size categories.

• Distributions of entry points per systemin the various system size categories.

• Distributions of populations served bysize category.

• Flow sizes as a function of populationserved.

• Daily household water consumption.• Proportions of systems and sources

exceeding regulatory limits.• Unit costs (capital and O&M) of

treatment technologies and annual costs ofalternate source and regionalization.

• Proportions of non-compliant systemschoosing between treatment, alternate source,and regionalization.

Since per system costs are much higher forvery large systems, the assumptions used inthe larger water system size categories can beexpected to dominate the variability innational costs. Each of these inputs will bemodeled using probability distributions that



reflect the state of the available data. In somecases, input variability will be estimatedfrom SDWIS, the CWSS, or other sources(e.g., distributions of populations served,daily household water consumption, unitcosts) . In other cases, input variability willhave to be based on best professionaljudgement. Again, this uncertainty study isexpected to provide useful information, butis not expected to result in changes to theregulatory decisions described in today’sNODA.

D. Estimates of Costs and Benefits for Non-Transient Non-Community Water Systems

The available data are not sufficient toallow EPA to predict a central-tendencyimpact of the regulatory options on non-transient non-community water systems(NTSC systems). Instead, EPA conducted a‘‘what-if’’ analysis of potential costs andbenefits based on reasonable assumptions ofthe percentage of NTSC systems impacted bythe various options (EPA 2000b). A ‘‘what-if’’analysis allows us to pose hypotheticaloccurrence scenarios and to estimate costs

and benefits for these scenarios. If thescenarios are chosen properly, they shouldbound the reasonable set of potential costsand benefits for NTSC systems. However, theestimates should not be interpreted asrepresenting ‘‘best estimates,’’ which wouldbe based on an occurrence survey ofradionuclides occurring at NTSC systems.The Technical Support Document (EPA2000a) provides details on the inputs andassumptions used for estimating regulatoryimpacts for NTSC systems. The resultingestimates of the percentage of systems out ofcompliance are provided in Table V–3.

TABLE V–3.—ASSUMPTIONS FOR HYPOTHETICAL ‘‘WHAT-IF’’ ANALYSIS FOR NON-TRANSIENT NON-COMMUNITY WATERSYSTEMS (APPROXIMATELY 19,300 SYSTEMS NATIONWIDE)

Regulatory option

Percent of na-tional systemsin states with

elevated levels(1) (percent)

Upper bound:10% of col. (1)

(percent)

Lower bound:1% of col. (1)

(percent)

Gross Alpha at 15 pCi/L .............................................................................................................. 60 6 1Combined Radium at 5 pCi/L ...................................................................................................... 79 8 1Uranium at 20 pCi/L:

Ground water ........................................................................................................................ 54 5 1Surface water ....................................................................................................................... 29 3 0

We calculated risk reductions associatedwith each set of assumptions using the sameanalytic approach as outlined for thecommunity water systems. However, we uselower water intake assumptions because thepopulation affected generally is not at thelocation served full-time or year-round. Therisk factors were estimated using the samerisk coefficients as a starting point (risk perpCi), but use different water consumptionassumptions to calculate lifetime excess riskfactors (risk per pCi/L). A cost model is usedto predict the annual compliance costs forthese systems based on their size classes(EPA 2000); in general, non-transient non-community systems tend to use ground waterand serve small populations.

The results of the analysis are summarizedin Table V–4. If EPA requires non-transientnon-community systems to comply with thegross alpha standard of 15 pCi/L, under theassumptions used in the analysis the numberof systems out of compliance could rangefrom 110 to 1,100 systems. The associatedannual costs range from $1 million to $4million and the statistical cancer cases (fataland nonfatal) avoided annually range from0.01 cases to 0.1 cases. For combined radium,the resulting number of impacted systemsranges from 150 to 1,500 systems with annualcosts ranging from $1 million to $6 millionand an associated number of annualstatistical cancer cases avoided ranging from0.02 cases to 0.2 cases. For a uranium MCL

of 20 µg/L, the results suggest a range ofimpacted ground water systems from 100 upto 1,000 systems with annual costs rangingfrom $1 million to $4 million and anassociated number of annual statisticalcancer cases avoided ranging from less than0.01 cases up to 0.04 cases. The resultingnumber of surface water systems impacted bya uranium MCL of 20 µg/L ranges from lessthan 10 to less than 20 systems. Theassociated national annual costs for surfacewater systems is less than $0.1 million up to0.1 million with annual risk reductions ofless than 0.01 statistical cancer cases.

TABLE V–4.—HYPOTHETICAL ‘‘WHAT-IF’’ RESULTS FOR NON-TRANSIENT NON-COMMUNITY WATER SYSTEMS

Regulatory option Lower Bound Estimate Upper Bound Estimate

Number ofsystems outof compli-

ance

Annual costs(milliondollars)

Statisticalcancer cases

avoided(cases)

Number ofsystems outof compli-

ance

Annual costs(milliondollars)

Statistical can-cer casesavoided

Gross Alpha at 15 pCi/L .................................... 110 1 0.01 1,100 4 0.1Combined Radium at 5 pCi/L ............................ 150 1 0.02 1,500 6 0.2Uranium at 20 pCi/L:

Ground water .............................................. 100 1 <0.01 1,000 4 0.04Surface water .............................................. < 10 0.03 <0.01 < 20 0.1 <0.01

Note: These results are based on hypothetical assumptions regarding the percent of systems likely to be out of compliance with each regu-latory option as discussed in the preceding text. These are not estimates of actual compliance costs or risk reductions, and are provided for illus-trative purposes only.

E. Impacts for Systems Serving Greater ThanOne Million Persons

Based on an Internet search of the availablewater quality information for water systemsserving greater than one million persons(very large systems), there is no directevidence that closing the monitoring

deficiencies for radium will impact thesesystems. However, the internet search wasnot conclusive in ruling out the possibilitythat one or more systems serving greater thanone million persons would be impacted bythese options. For this reason, EPA hasfollowed up with the few systems in questionto determine the likelihood of impact. The

follow-up confirmed that there were noimpacts expected for these systems. Uraniumoccurrence data for these systems wascollected to the extent feasible and there isno evidence of an impact at 20 or 40 µg/L.



F. ReferencesLongtin, J.P. ‘‘Occurrence of Radon, Radium,

and Uranium in Groundwater.’’ JAWWA.Vol. 80, No. 7, pp. 84–93. July 1988.

National Research Council. Risk Assessmentof Radon in Drinking Water. NationalAcademy Press. Washington, DC. 1999.(NAS 1999).

USEPA. National Primary Drinking WaterRegulations; Radionuclides; ProposedRule. Federal Register. Vol. 56, No. 138, p.33050. July 18, 1991.

USEPA. ‘‘Actual Cost for Compliance withthe Safe Drinking Water Act Standard forRadium-226 and Radium-228—FinalReport.’’ Prepared by InternationalConsultants, Inc. for the Office of GroundWater and Drinking Water. Draft. July1998. (EPA 1998a)

USEPA. ‘‘Evaluation of Full-Scale TreatmentTechnologies at Small Drinking WaterSystems: Summary of Available Cost andPerformance Data.’’ Prepared by ICF, Inc.

and ISSI, Inc. for the Office of GroundWater and Drinking Water. December 10,1998. (EPA 1998b)

USEPA. Results from survey of complianceactions taken in the State of Illinois inresponse to the NPDWR for combinedradium. Submitted from EPA Region 5.1998. (EPA 1998c)

USEPA. ‘‘Guidelines for Preparing EconomicAnalysis.’’ Review Draft. November 3,1998. (EPA 1998d)

USEPA. ‘‘Regional Variation of the Cost ofDrinking Water Wells for Public WaterSupplies.’’ Prepared by Cadmus for theOffice of Ground Water and DrinkingWater. Draft. October 1999. (EPA 1999b)

USEPA. ‘‘Drinking Water Baseline Handbook:First Edition.’’ Draft dated March 2, 1999.(EPA 1999c)

USEPA. ‘‘Evaluation of Central TreatmentOptions as Small System TreatmentTechnologies.’’ Prepared by SAIC for EPA.Draft dated January 28, 1999. (EPA 1999d)

USEPA. Cancer Risk Coefficients forEnvironmental Exposure to Radionuclides,Federal Guidance Report No. 13. USEnvironmental Protection Agency,Washington, DC, 1999. (EPA 1999e)


USEPA. ‘‘Preliminary Health Risk Reductionand Cost Analysis: Revised NationalPrimary Drinking Water Standards forRadionuclides.’’ Prepared by IndustrialEconomics, Inc. for the Office of GroundWater and Drinking Water. Draft. January2000. (EPA 2000b)

Dated: April 7, 2000.Charles J. Fox,Assistant Administrator, Office of Water.[FR Doc. 00–9654 Filed 4–20–00; 8:45 am]BILLING CODE 6560–50–U


ENVIRONMENTAL PROTECTION AGENCY FOR FURTHER 40 CFR...

Documents

Transcript of ENVIRONMENTAL PROTECTION AGENCY FOR FURTHER 40 CFR...