~Ec~

BEFORE THE POLLUTION CONTROL BOARD

OF THE STATE OF ILLINOIS

~

242004

STATEOF,LI~ s

ion Contro’ Board

iN THE MATTER OF:

)

)

REVISIONS TO RADIUM WATER

)

QUALITY STANDARDS: PROPOSED

)

R04-21

NEW

35

ILL. ADM1N. CODE 302.307

)

Rulemaking

-

Water

AND AMENDMENTS TO

35

ILL. ADMIN.

)

CODE 302.207 AND

302.525

)

NOTICE OF FILING

To:

See Attached Service List

Please take notice that on November 24, 2004, we filed with the Office of the Clerk of

the Illinois Pollution Control Board an original and ten copies of the attached

REMEDIATION TECHNOLOGY LLC’S RESPONSE TO DOCUMENTS REQUESTED

DURING FOURTHPUBLICHEARING,

a copy ofwhich is served upon you.

Respectfully submitted,

WATER REMEDIATION TECHNOLOGY, LLC

By:

___________

One of.its Attorneys

Jeffrey C. Fort

Letissa Carver Reid

Sonnenschein Nath & Rosenthal LLP

8000 Sears Tower

Chicago, Illinois 60606

(312) 876-8000

THIS FILING IS BEING SUBMITTED ON RECYCLED PAPER

---

BEFORE THE POLLUTION CONTROL BOARD

CLL~~~3

OF THE STATE OF ILLINOIS

NO~24 ~

C0flt~o~

Board

IN THE MATTER OF:

)

)

REVISIONS TO RADIUM WATER

)

QUALITY STANDARDS: PROPOSED

)

R04-21

NEW

35 ILL. ADMIN. CODE 302.307

)

Rulemaking

-

Water

AND AMENDMENTS TO 35 ILL. ADMIN.

)

CODE 302.207

AND

302.525

)

TO DOCUMENTS REQUESTED DURING FOURTH PUBLIC HEARING

Water Remediation Teclmology LLC (“WRT”) through its attorneys, submits the

documents requested to be submitted by WRT during the October 21 and 22, 2004, public

hearings in this proceeding. A summary ofthese requests and documents follows:

REQUEST ONE:

Illinois Environmental Protection Agency (“IEPA”) requested

references for the statement by Mr. Theodore G. Adams concerning 1 pCi/L as representing

background for total radium in surface waters. The following documents are responsive to this

request:

• Agency for Toxic Substances and Disease Registry (“ATSDR”) toxicity summary on

radium. (~çOctober 22 Hearing Exhibit 16.)

• C. T. Hess, J. Michel, T. R. Horton, H. M. Prichard and W. A. Coniglio, The

Occurrence of Radioactivity in Public Water Supplies in the United States, Health

Physics Vol. 48, No.

5

(May), pp.

553-586 (1985).

(See Attachment A hereto.)

• Jacqueline Michel and C. Richard Cothern, Predicting the Occurrence of 228Ra in

Ground Water, Health Physics Vol.

51,

No. 6 (December), pp. 715-721 (1986).

(See Attachment A.)

---

REOUEST TWO:

Counsel for the City of Joliet requested a copy of the permit

applications submitted to the Illinois Department ofNuclear Safety (“IDNS”) for the Village of

Oswego. Copies of those permit applications (which were marked as Exhibit 17 during the

October 22 hearing) were previously submitted to counsel for IEPA and the City of Joliet.

($~~October22 Hearing Exhibit 17.)

REQUEST THREE: Counsel for the City of Joliet requested copies of WRT contract

with the Village of Oswego andlor the Village of Elburn. WRT has been informed by the

Village ofOswego that it has provided the requested contract to counsel for the City of Joliet.

Respectfully submitted,

WATER RE

DIATION TECHNOLOGY, LLC

By:_________

One o

1

ttori2iey~

Jeffrey C. Fort

Letissa Carver Reid

Sonnenschein Nath & Rosenthal LLP

8000 Sears Tower

Chicago, Illinois 60606

(312) 876-8000

11800145

THIS FILING IS BEING SUBMITTED ON RECYCLED PAPER

2

---

CERTIFICATE OF SERVICE

The undersigned, an attorney, certifies that he/she has served upon the individuals named

on the attached Notice of Filing true and correct copies of

TECHNOLOGY LLC’S RESPONSE TO DOCUMENTS REQUESTED DURING FOURTH

PUBLIC HEARING

by First Class Mail, postage prepaid, on November 24, 2004.

~

---

SERVICE LIST

Dorothy Gunn

Clerk ofthe Board

Illinois Pollution Control Board

100 West Randolph Street

Suite 11-500

Chicago, IL 60601

R04-21

Amy Antoniolli

Hearing Officer

Illinois Pollution Control Board

100 West Randolph Street

Suite 11-500

Chicago, IL 60601

Deborah J. Williams

Stefanie N. Diers

Illinois Environmental Protection Agency

1021 North Grand Avenue East

P.O. Box 19276

Springfield, IL 62794-9276

Joel J. Sternstein, Assistant Attorney General

Matthew J. Dunn, Division Chief

Office of the Illinois Attorney General

Environmental Bureau

188 West Randolph

20th

Floor

Chicago, IL 60601

Stanley Yonkauski

Acting General Counsel

Illinois Department ofNatural Resources

One Natural Resources Way

Springfield, IL 62701

Richard Lanyon

Metropolitan Water Reclamation District

100 East Erie Street

Chicago, IL 60611

Roy M. Harsch

Sasha M. Engle

Gardner Carton & Douglas

191 North Wacker Drive

Suite 3700

Chicago, IL 60606-1698

Claire A. Manning

Posegate & Denes

111 North Sixth Street

Springfield, IL 62701

Lisa Frede

CICI

2250 East Devon Avenue

Suite 239

Des Plaines, IL 60018

William Seith

Total Environmental Solutions

631 East Butterfield Road

Suite 315

Lombard, IL 60148

Albert F. Ettinger

Environmental Law and Policy Center

35 East Wacker Drive

Suite 1300

Chicago, IL 60601

John McMahon

Wilkie & McMahon

8 East Main Street

Champaign, IL 61820

Dennis L. Duffield

City of Joliet

Department ofPublic Works and Utilities

921 East Washington Street

Joliet, IL 60431

Abdul Khalique

Metropolitan Water ReclamationDistrict of

Greater Chicago

6001 West Pershing Road

Cicero, IL 60804

---

Ph

—1--

---

jçealth

physics

Vol. 48. No. 5 (May), pp. 553.-5~ 1985

0017—9078/85 53.00

+

.00

in

the

U.S.A.

00 1985

Health Physics Society

Pcrgamon Press Ltd.

THE OCCURRENCE OF RADIOACTIVITY IN PUBLIC

WATER SUPPLIES IN THE UNITED STATESt

C. T.

HESS

University of Maine, Department of Physics, Orono, ME 04469

J. MICHEL

Research Planning Institute, Inc.,

925

Gervais Street, Columbia, SC 29201

T. R. HORTON

Eastern Environmental Radiation Facility, U.S. Environmental Protection Agency, P.O. Box 3009,

Montgomery, AL 36193

H. M. PRICHARD

University of Texas School of Public Health, P.O. Box 20186, Houston, TX 77025

and

W. A. CONIGLIO

Office of Drinking Water, U.S. Environmental Protection Agency, 401 Street S.W., Washington, DC 20460

Abstract—Examination of the

collected data for radionuclide concentration measurements

in public water supplies in the United States show more than 51,000 measurements for gross

a-particle activity and/or Ra,

89,900 measurements

for U, and 9,000 measurements for Rn.

These

measurements were made as part of national and state surveys of radionuclide

concentrations in utility water supplies for Ra and Rn; and the National Uranium Resource

Evolution

(NIJRE) survey

for U which included non-utility water supplies.

Surface water has low values for Ra and Rn but levels comparable to

ground water for U.

Separate isotope measurements were not taken for much of the Ra and U data. Because

226Ra to 228Ra ratios and ~8U to ~U ratios are not fixed in water, further measurements are

needed to establish the specific isotopic concentrations by region. Analysis ofthe state average

values in geological provinces shows the highest provincial areas for Ra are the Upper

Coastal

Plain, the glaciated Central Platform, and the Colorado plateau. For U, the highest areas are

the Colorado plateau, the West Central Platform, and the Rocky Mountains. For Rn, the

highest provinces are New England and the Appalachian Highlands-Piedmont. Regional

hydrogeological and geochemical models are suggested for guiding the formulation of regional

standards and monitoring strategies. Utility supplies serving small populations have the

highest concentration for each radionuclide and have the lowest fraction of samples measured,

which shows a need for further measurements of these small population water supplies. Risk

estimates for the average concentration of Ra in utility ground water give about 941 fatal

cancers per 70.7-yr lifetime in the United States. Risk estimates for the average concentration

of U in utility surface and ground water give about 105 fatal cancers per 70.7-yr lifetime in

the United States. Using 1 pCi/liter in air for 10,000 pCi/I in water, the Rn in. utility water

risk estimate is for 4,400—22,000 fatal cancers per 70.7-yr lifetime in the United States.

t This paper was prepared for and completed at

.

PREFACE

the National Workshop for Radioactivity in Drinking THE sections of this paper are arranged in the

Water under the sponsorship of the Office of Drinking order of introduction, geochemistry and occur-

Water, U.S. Environmental Protection Agency. The rence. A central theme of the report is that

workshop was held in Easton, MD, 24—26 May 1983.

geological setting strongly influences the occur-

553

Attachment A

---

554

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

rence of natural radionuclides in drinking water.

The observed concentrations of U, Ra and Rn

in ground and surface water can be related to

the rock types and the amount and distribution

of U and Th in the materials which constitute

the aquifer and surficial deposits. The United

Table

I.

Summary of

States can be divided into 11 geological prov.

inces, each of which is characterized by doi~j~

nant types of rocks or deposits as well as

ground-water flow systems, discussed in Table

1 and shown in Fig. I (Be8 1; Sc62). These

provinces are discussed in all sections of the

potential host rocks, geologic framework and nature ofground-water flow systems in

the provinces ofthe coterminous United States (Fig.

1)

PROVINCE

GEOLOGICAL

.

FRAMEWORI

NATURE OF GROUND-WATER

FLOW SYSTEMS

1. New England.

New England-—complexly faulted

Flow and head in bedrock of

Adirondack

metamorphic and aetasedimentary

low permeability and overlying

Mountains

rocks intruded by large masses

of granite. Adirondacks——

mountains composed of marble

and schist intruded by granites,

anorthosite, and gabbro.

glacial aquifers greatly

influenced by local

topography and surface—water

features.

2. Appalachian

Highlands.

Appalachian Highlands——mountain

belt of granites and metamorphics

Flow and head in metamorphic and

granite bedrock of low

Piedmont

thrust westward over Paleozic

rocks, Piedmont——non-mountianous

belt of highly complex metamorphic

rocks with abundant granites.

~

permeability, largely controlled

by topography and surface water

features; folded limestone

locally cavernous and highly per-

rneable at shallow depths supporting

large springs; sandstone aquifers

of moderate extent and

permeability; flow systems

generally related to local recharge

in interstream areas and discharge

to surface—water features.

3. Appalachian

Appalachian and Interior Plateaus

Regional flow in low—to—

and Interior

consist of gently dipping, gently

moderately permeable sandstones

Plateaus

folded sandstones, shales,

carbonates, and evaporites.

In southern Missouri exposes old

crystalline rocks,

and carbonates; carbonates locally

of high permeability at shallow

depth due to fractures and solution

channels support large springs.

4. Coastal Plain

Seaward dipping thickening wedge of

sand, sandstones and shales with some

evaporites and limestones; underlain

by a basement of metamorphic rocks,

Regional flow in sand and limestone

aquifers with intervening clay con-

fining layers; predominant flow

direction seaward; discharge upward

through confining layers and to

streams.

5. Glaciated

igneous and metamorphic rocks on the

Regional flow in sandstone and

central

northwest overlain by sandstones,

carbonate aquifers; highly

Platform

carbonates, shales, and evaporites;

deep basin deposits in Michigan and

Illinois.

mineralized water at depth in

basins; glacial aquifers locally

overlie bedrock.

6. Western

Central

Platform

.

Horizontal to gently dipping

sandstones; deep sedimentary

basins and structural high.

Capped with sands and gravels,

Regional flow in layered sandstone

and carbonate aquifers; thick con-

fining beds of shale; deep basins

contain highly saline water.

Extensive fluvial deposit

aquifers Nebraska south into

Texas and glacial aquifers in

North Dakota and South Dakota

overlie sandstones.

r

---

Table 1. (Contd.)

C. T. HESS

et a!.

GEOLOGICAL FRAMEWORK

NATURE OF GROUND—WATER

FLOW SYSTEMS

Igneous and metamorphic folded core

rocks of Rocky Mountains and

intermontane basins of shaies,

carbonates, evaporites, and

sandstones. Intrusive and

volcanic rocks.

Regional flow in layered sandstone

and carbonate aquifers with shale

confining beds in intermontane

basins; local recharge and

discharge controlled by topography

and surface water features in frac-

tured igneous and metamorphic

rocks.

Flat—lying to gently warped layers

of sandstones, shales, limestones

and evaporites with volcanic rocks.

Regional flow in layered sedimen-

tary rocks; chief aquifers are

sandstones and carbonates;

discharge to major streams; highly

saline water at depth in deep

basins and in association with salt

beds.

Elongated blocky mountains of faulted

rock complexes; deep alluvium-

filled intermontane basins; Intrusive

igneous stocks and plugs; extrusive

ash—flow tuffs, rhyolites, and

basalts,

Flow within closed basins; inter—

basin flow between closed

topographic basins through per-

meable bedrock; interbasin flow in

alluvial channels between basins

with integrated surface drainage;

deep regional flow systems in

carbonate and volcanic rocks.

io.

Columbia

Regional shallow structural basin of

Basaltic lava flows range from

Plateaus

basaltic lava flows; locally faulted

and folded; mountain range on the

west consisting of elongated chain of

andesitic volcanic cones,

highly permeable to dense nearly

impermeable, creating regional

aquifers with perched aquifers

separated by confining beds. The

ground water principally discharges

to the major streams; locally

discharges to a few closed basins.

11. Pacific

Consists of several complex elements:

Regional flow In deep intermontane

Mountain

Large uplifted and tilted blocks

sedimentary basins; igneous and

System

of granite with inliers of

aetasediments; folded and faulted

sedimentary rocks; deep elongated

troughs filled with fluvial sediments,

with fluvial sediments.

metamorphic rocks of’ low

permeability support shallow local

flow systems related to topography

and surface drainage.

report

and provide a framework for understand-

ing the variations in

the distribution and activ-

ities of natural radionuclides in water. In fact,

one hypothesis is that certain provinces or sub-

provinces can be characterized as producing.

ground water with specific radionuclide prob-

lems, or conversely, without

specific radionuclide

problems. If this hypothesis can be verified, it

has important applications to the development

of regional guidelines for monitoring require-

ments in the revised regulations (La83).

OCCURRENCE OF Ra ISOTOPES IN PUBLIC

DRINKING WATER

Introduction

Radium has two natural isotopes which are

of concern in public water supplies. Radium-

226 is generated through decay of

238U

and is

an a emitter with a

t112

=

1,622

yr.

This is the

isotope which is commonly referred to as Ra

and has been measured in many water supplies.

The other isotope, 228Ra, is generated directly

by 232Th

decay and is a shorter-lived, weak

~9

PROVINCE

~,

Rocky Mountain

5ys tern

555

~,

Colorado

Plateaus

~,

Basin and Range

---

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

ftJI

NEW ENGLAND-ADIRONDACK MOUNTAINS

APPALACHIAN HIGHLANDS

-PIEDMONT

APPALACHIAN AND INTERIOR PLATEAUS

~~COASTAL PLAIN

1111111

GLACIATED CENTRAL PLATFORM

LI

WESTERN CENTRAL PLATFORM

ROCKY MOUNTAIN SYSTEM

—

COLORADO PLATEAUS

BASIN AND RANGE

~fl~Jj~J

COLUMBIA PLATEAUS

~~PACIFIc MOUNTAIN SYSTEM

FIG. 1. Geological provinces of the United States, according to Beddinger (Be8 I).

emitter

(t112

5.7

yr). There is a third isotope

of Ra which is of possible concern, 224Ra, with

a

t112

=

3.64 days. Its occurrence is not well

known; only a few data are available from

samples at the well head. The U.S. Environ-

mental Protection Agency (EPA) established

interim regulations in

1976

for maximum levels

of radioactivity in drinking water as follows;

“Maximum contaminant levels (MCL) of

combined Ra-226 and Ra-228—5 picocuries

per liter (pCi/I); gross alpha-particle activity—

15 pCi/i excluding, radon and uranium”

(Ep76a).

These MCLs were set under the authority of

the

Safe Drinking Water Act to protect health,

taking treatment costs in consideration. In an

effort to minimize the costs of analysis and

monitoring, the EPA established a series of

screening steps to test for compliance with the

interim regulations. These criteria stated that

when the average gross a-particle activity of

four quarterly samples or composites exceeds

5

pCi/I, the same or equivalent sample shall be

analyzed for 226Ra. If the activity of 226Ra

exceeds 3.0 pCi/I, the sample shall be analyzed

for 228Ra. Inherent in these regulations were the

assumptions’ that 226Ra was to be the dominant

radioactive contaminant in drinking water and

the 228Ra/226Ra aétivity ratio was less than 1.0.

The regulations required all systems supplying

25 or more people to be monitored every 4 yr.

Since the interim regulations were established,

much more information on the occurrence of

Ra isotopes is now available from state compli-

ance data and from detailed studies on the

correlation and interrelationships of 228Ra and

226Ra in ground water with specific geological

provinces (Mi80; AsS I; Ki82; Mi82; ‘Kris82).

In light of these new data, the key issues to be

considered for revision ofthe regulations are:

(1) prioritization of specific areas for moni-

toring for 228Ra and 226Ra;

556

---

C. T. HESS

et a!.

(2) reduction in the interval frequency or

complete omission for specific areas for repeat

monitoring; and

(3) decoupling 228Ra analysis from 226Ra, with

criteria

for

when 228Ra

is to be

measured.

The purpose of this paper is to concisely

review the existing information on the geo-

chemistry and occurrence of 228Ra and 226Ra,

and to provide guidelines forregulatory revision.

GeochemistrY of

Ra

isotopes

The distribution of 228Ra and 226Ra in water

is a function of the Th and U content of the

aquifer, the geochemical setting of the aquifer

solids, and the

ti!2

of each isotope. There are

specific geological and chemical processes that

control the Th and U content in aquifers, which

are discussed in detail by Olson and Overstreet

(0164), Cherdynstev (Ch7 1) and Gableman

(Ga77). In fact, Tb and U have very similar

behavior, with one important exception which

is most responsible for their eventual separation.

Thorium has one oxidation state and is immo-

bile at low temperatures. Therefore, Th distri-

bution is controlled by primary geochemical

processes (such as magmatic crystallization) or

secondary physical processes (such as sedimen-

tary enrichment in placer deposits). Uranium

has

two

oxidation states and

the +6 state

(urany))

can form highly soluble complexes which can

be transported long distances by oxidizing

ground water before being removed by adsorp-

tion or reduction to the +4 state. The estimated

average crustal Th/U activity ratio is 1.2—1.5 so

that, in the absence of enrichment or depletion

processes, 228Ra activity should be higher than

226Ra. However, the tendency for ,U enrichment

under certain geochemical conditions results in

regions of higher 226Ra, thus the EPA’s decision

to emphasize 226Ra in the interim regulations.

Radium enters ground water by dissolution

of aquifer solids; by direct recoil across the

liquid-solid boundary during its formation by

radioactive decay of its parent in the solid (both

isotopes have Th as the immediate parent); and

by desorption. The mechanism of a recoil is an

important factor in the higher solubility of

progeny isotopes compared with their parents.

Uranium-234/uranjum-238 activity ratios in

ground water are generally greater than 1.0 and

can be as high as 28 (Gi82). Radium-224J

radium-228 activity ratios in South Carolina

ground water range from 1.2—2.0 (Mo83) and

in Connecticut from 0.8—1.7 (Kris82). However,

when the progeny/parent pair consists of differ-

ent elements, geochemical factors become int-

portant controls of their relative solubility. An

extreme example is 222Rn, the immediate prog-

eny of 226Ra; 222Rn/226Ra activity ratios in water

can be as high as 106. Because of a recoil and

the different solubilities of the Th and U series

isotopes, extensive disequilibrium occurs in

ground water.

Recent studies have suggested that Ra is

rapidly absorbed from ground water. King

el

al.

proposed that the distance of Ra transport

in ground water was less than that of 222Rn

(with a

t112

=

3.8 days) due to continual ad-

sorption of Ra onto the aquifer solids (Ki82).

Krishnaswami

et al.

calculated adsorption and

desorption rate constants’for Ra in Connecticut

aquifers and proposed that Ra removal rates

are rapid, as short as a few minutes (Kris82).

Equilibrium between adsorption and desorption

is also quickly established, but Krishnaswami

et

a!.

concluded that the partition coefficient

strongly favors the solid phase, and almost all

Ra introduced into the ground water studied

resides on particle surfaces in the adsorbed state.

However, the extent of

sorption is controlled

by the geochemical reactivity of the aquifer

material. King

et al.

(Ki82)

note that the average

228Ra and 226Ra activity in the crystalline aqui-

fers of South Carolina was lower than for the

Coastal Plain sediments, even though the Th

and U content of the rock aquifers was higher.

Furthermore, the 222Rn activity in the crystalline

aquifers was 10 times greater than the aquifers

sampled in the Coastal Plain. King

et a!.

con-

cluded that the affinity of Ra for adsorption

sites in the fresh rock surfaces which have

higher cation-exchange capacities was greater

than for the sand and gravel deposits composed

of refractory minerals such as quartz. Thus, Ra

in ground water does not accumulate with

ground water ‘transport in aquifers; it stays very

close to the area in which it is produced.

The insolubility of Ra and Th can be inferred

from studies of potential contamination of

U

557

U

1

U

U

U

U

I

U

u

ru

I

U

U

U

U

U

U

U

U

U

---

I

558

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

ground water due to seepage from U tailings

ponds in New Mexico reported by Kaufmann

et a!.

(Ka76). At one such pond, they estimated

that nearly 3 X 1 0~liters of seepage entered the

shallow aquifer during a 20-yr period. The

wastes in this pond contained approximately

200 pCi/I of 226Ra and 166,000 pCi/I of 230Th.

Thus, nearly a curie of 226Ra and 500 Ci of

230Th were available to leach with the shallow

ground water; yet, in 1975, monitoring wells

located 1 km down-gradient from the pond

showed no evidence of contamination.

Through an understanding of the physical

and chemical processes which control Ra distri-

bution, we can now begin to interpret the new

data base from state compliance reports, and to

develop predictive models for Ra occurrence on

which new regulations should be structured.

These proposed models would characterize cer-

tain geological settings or aquifer types as pro-

ducing ground water with high or low Ra con-

tent. The EPA has begun to develop a predictive

model for the occurrence of 228Ra, with a~pilot

study completed for two geological provinces,

the Atlantic and Gulf Coastal Plain sedimentary

aquifers and the Piedmont rock aquifer of the

eastern United States (Mi82). Information on

areas of

high Ra occurrence is necessary to

provide guidance to states for additional moni-

toring. From a regulatory point of view, areas

of low Ra activity are very important, in that

they could have a different monitoring priority

and schedule. A predictive model for 228Ra

would also be valuable because so few samples

were measured under the present analytical

scheme.

Occurrence of226Ra and

228Ra

in drinking water

All but six states (Illinois, Nebraska, Colorado,

Utah, Montana and Oregon) have reported

known MCL violations for Ra as required by

the interim regulations. There are approximately

200 reported public water suppliers with 226Ra

activities in excess of

5

pCi/I after normal

treatment (Co83c). The following sections dis-

cuss these results and other studies by water

types, geological setting, and isotope.

Surface water

The Ra content of surface water is usually

very low. Radium-226 generally ranges between

0.1 and 0.5 pCi/I and the 228Ra/226Ra activity

ratio is generally greater than unity (Mo69;’

El82). Also, standard water treatment methods

are known to remove Ra (Ep76b). To the best

of our knowledge, no surface-water violations

for Ra have been reported by the states. Thus,

surface-water systems should be separately eval-

uated; perhaps they could be released from

monitoringrequirements forRa once the source

stream was documented as having low natural

radioactivity.

Ground waler

Out of the nearly 60,000 public water supplies

in the United States, about 80

use ground-

water sources. More than 90 of the ground-

water supplies serve less than 3,300 people and

are classified as small or very small. In general,

Ra in drinking water is a small-system problem.

Figure 2 is a compilation of

the areas and

specific sites which have high Ra in ground

water from both state compliance data and

published studies.

The available state compliance data for Ra

comes almost exclusively from samples which

first showed a gross a-particle activity of ?5

pCi/i. Iowa used a screen of 2 pCi/I for gross

a-particle activity. In some areas, states would

analyze additional samples in an area where

high radioactivity was found during the initial

sampling. Radium-228 data were provided for

about one-half of the 200 226Ra values reported.

Statewide summaries of 226Ra and 228Ra data

have been published for Georgia (C183), South

Carolina (Ki82), Iowa (Krie82) and Illinois

(Ro77); Lucas reported results for more than

90 of

the communities in Illinois, Iowa, Mis-

souri, and Wisconsin (Lu82).

There have been several studies on the tem-

poral variability of the activity of Ra isotopes

in ground-water systems. Kriege and Hahne

reported that the mean value for the average

percent deviations of 141 samples during 18 yr

in Iowa was 21 with a relative standard devia-

tion of 15 (Krie82). Michel and Moore found

a maximum variation of 19 during 2 yr in

individual wells (Mi80). Therefore, in single-

well systems, one sample should be representa-

tive of the average annual activity; also the

present requirement for monitoring at 4-yr in-

tervals would not be necessary unless changes

to the system have been made. Systems with

I

I

mu1tip~

contini

contril

Fror

mean

the M

actiVit)

but th

226Ra.

226Ra

sample

these,

but tol

estimal

violati(

provin

---

C. T. HESS

et a!.

559

NO DATA REPORTED

FIG.

2. The approximate locations and general areas of public water supplies which exceed

5

pCi/i of total Ra (228Ra was reported or combined with 226Ra for about one-half of the sites).

Large dots represent individual violations. The dot pattern represents the general area of a

group of violations, with the adjacent number indicating the number of violations in that

group. When the locations were unknown, just the number of violations was indicated

(modified after Cothern and Lappenbusch (Co83)).

multiple wells have the potential problem of

continuously variable Ra based on the relative

contribution’ of each well when sampled.

From the data reported by the states, the

mean total Ra activity for supplies exceeding

the MCL was almost 10 pCi/I. Radium-226

activity was generally greater than 228Ra activity,

but ,these data were initially biased toward high

226Ra. King

el al.

found that the average 228Ra/

226Ra

activity ratio was 1.2 for more than 180

samples throughout South Carolina (Ki82). Of

these, 10 samples had 226Ra less than 3 pCi/i,

but total Ra greater than

5

pCi/i. King

et a!.

estimated that perhaps 40—50 of the total Ra

violations for the Piedmont and Coastal Plain

provinces were missed using the prescribed

screening procedure which couples 228Ra analysis

to 226Ra (Ki82). Kriege and Hahne reported

additional sampling which identified eight vio-

lations for total Ra although the 226Ra was less

than 3 pCi/I (Krie82).

From the available data, thereare two specific

geological regions where more than 75 of the

known Ra violations occur. They are; (1) the

Piedmont and Coastal Plain provinces in New

Jersey, North Carolina, South Carolina and

Georgia; and (2) a north-central region, consist-

ing of parts of Minnesota, Iowa, Illinois, Mis-

souri and Wisconsin. The rest of the violations

are generally scattered clusters, notably along

the Arizona-New Mexico border, Texas, Missis-

sippi, Florida and Massachusetts (Fig. 2). All of.

* ONLY GROSS ALPHA PARTICLE VIOLATIONS

ARE AVAILABLE 144 SITES WITH 15 pCi/Il

---

560

RADIOACTIVITY

IN PUBLIC WATER SUPPLIES

these scattered violations had high 226Ra activ-

ities, as would be expected from the screening

methods used to detect them. Radium-228 ac-

tivities in these systems were very low. We

believe that the current analytical protocol has

detected a large percentage of the systems with

high 226Ra. Cothern and Lappenbusch have

used the compliance data for 226Ra to estimate

that approximately 500 systems will be deter-

mined to exceed the MCL of 5 pCi/I (Co84).

Improvement on this estimate is difficult with

the existing data base, which is comprised mostly

of reported MCL violations for 226Ra. Statistical

analysis of these data is not possible because

they were not randomly sampled. In this respect,

states should be requested to submit all Ra

results to facilitate further analysis. However,

some calculations can be made to corroborate

the previous estimates of MCL violations. Data

from South Carolina showed that approximately

3.0 of the ground-water supplies exceeded the

5.0

pCi/I limit for total Ra (Ki82). (Note that

the prescribed screening procedures detected

only one-half of these violations). Applying that

percentage to North Carolina and Georgia, both

of which have similar hydrogeology, provides

an estimate of 150 violations for all three states.

In Iowa, approximately 10 of the 605 supplies

sampled to date, using a lower screening criteria,

exceed the MCL. Again,

applying this percentage

to all the ground-water systems of Iowa and

half of Illinois, Missouri and Wisconsin yields

120 violations for Iowa, 75 violations for Illinois,

50 for Missouri, and

60 for Wisconsin. We can

estimate violations for the states that have not

reported as follows: 10 each for Utah, Colorado

and Nebraska, and zero for Oregon and Mon-

tana. There are 71 violations reported in all

other states. The total of these known and

estimated violations is 556. Assuming 10—25

of the actual violations are missed during the

prescribed screening procedure (actual data for

Iowa, 8 out of 60 or 13; for South Carolina,

8 out of 30 or 26),

the number of violations

ranges between 600 and 700.

Estimates of population exposure nationwide

can only be broadly made without additional

information on populations served by the MCL

violations, as well as on the results of all analyses.

Lucas estimated that 91 communities in Illinois,

Iowa, Missouri and Wisconsin with a

population

of 599,000 consume water with 226Ra greater

than 5.0 pCi/I (Lu82).

Under the present screening methods, how-

ever, there was concern that 228Ra’ violations

were not being detected. Thus, the EPA recently

funded a study to determine if a predictive

model for the occurrence of

228Ra could be

developed. The Piedmont and Coastal Plain

aquifers were selected as a pilot study area for

development of a model because the mdi-

ochemistry of these provinces had been exten-

sively studied, these were areas of known high

228Ra activities, and nearly 300 values for 228Ra

were available.

The model comprises a multilevel classifica-

tion of aquifer characteristics for each 228Ra

datum. The nature of the data precluded use of

such analytical tools as regression analysis to

establish a quantitative relationship between,

for example, Th content and 228Ra. As a result,

the underlying model structure was evaluated

to assess the existence of differences not only

between the major aquifer types but also between

lower-level classification within a major aquifer

type. A detailed description of the model devel-

opment, parameters, methodology and results

can be found in a report by Michel and Pollman

(Mi82). Only a summary is discussed below.

Also, although the model was developed specif-

ically for 228Ra, values for 226Ra were available

and similar statistics were calculated. There is

much to be learned from these differences in

the results for these two isotopes.

Table 2 summarizes the means and ranges

for those aquifer types which had significantly

different 228Ra distributions.

Note the striking

differences in the means, although the ranges,

are similarin some cases. Arkosic (immature,

feldspar-rich) sand aquifers had mean values

for both Ra isotopes up to an order of magnitude

greater than quartzose sands. Limestones and

metamorphic rock aquifers in the study area

had very low activities of both Ra isotopes.

Table 3 shows a ranking of all structural levels’

used in the pilot study, with symbols indicating

groups of similarity of 228Ra distribution (with

226Ra means are also given without a ranking).

~Classesidentified by the same letter code in the

grouping column represent stibsets of a group

that are statistically indistinguishable from other

classes of the same group. However, groupings

---

C. T. HESS

et a!.

561

Table

2.

Summary of

228Ra

and

226Ra

distribution in ground water by aquifer type for the

Atlantic Coastal Plain and Piedmont provinces

Aquifer Type

Number

of Values

Ra-228

Ra—226

Geometric

Geometric

Mean

(pCi/i)

Range

(pCi/i)

Mean

(pCi/i)

Range

(pCi/i)

Igneous Rocks

(acidic)

42

1.39

0.0—22.6

Metamorphic

Rocks

Sand

Ark Os e

Quartzose

Limestone

75

0.33

0.0—

3.9

0.37

0.0— 7.4

143

1.05

0.0—17.6

1.36

0.0—25.9

92

2.16 0.0—13.5

2.19

0.0—23.0

50

‘

0.27 0.0—17.6

0.55 0.0—25.9

16

0.06 0.0- 0.2

0.12 0.0—

0.3

that overlap indicate that the particular individ-

ual groups are not unique. For example, group

A as a class is not statistically different from

groups B, C and D; groups E, F and G, however,

represent groups of classes with significantly

lower 228Ra activities than group A. Although

some groups were not

significantly different, the

ranking followed the anticipated trends within

major

types. For example, the ranking of igneous

rocks, with syenite granite diorite, follows

Th abundance in these rock types. Arkosic sand

aquifers are ranked in order of Th content of

the source rock, from high to low. This pilot

study showed that specific aquifer types and

geochemical conditions can be characterized as

producing ground water with high or low 228Ra

activities. Its application can be demonstrated

for the aquifers of the Piedmont and Atlantic

and Gulf Coastal Plain provinces; high 228Ra

was likely to occur in aquifers composed of (I)

acidic, igneous rocks and (2) arkosic sands with

sources having high-to-medium Th content.

The results from the model were used to map

specific areas (aquifer types) from New Jersey

to Alabama

that would be likely to produce

ground water with high 228Ra. In fact, in the

Piedmont province, granitic rock aquifers youn-

ger than 350 million yr were shown to produce

high-Ra ground water. Older rocks had under-

gone metamorphism which has tended to re-

crystallize Th and U into resistate minerals in

which Ra is more tightly bound. The arkosic

sand

aquifers

were restricted to the upper Coastal

Plain from Virginia to Georgia. These aquifers

are composed of sediments eroded from the

nearby Piedmont rocks and are mineralogically

immature. They contain higher amounts of Th-

and U-bearing minerals than the middle and

lower Coastal Plain sediments which were de-

posited farther from the source rocks. Thus, the

trend in the Coastal Plain aquifers for 228Ra,

whose parent is not subject to secondary trans-

port processes, is one of decreasing activities

with

distance from the Piedmont source. Only

one out of 50 samples from the quartzose

sands

of the middle and lower Coastal Plains aquifers

was greater than 3 pCi/I for

228Ra. In contrast,

226Ra in the middle Coastal

Plain aquifer is

highly variable,

with values from 0—196 pCi/i,

due to the ability of its parent to migrate in

ground

water and undergo secondary enrich-

ment.

Knowledge of the conditions where very low

radioactivity will occur is also very important.

In

the area studied, low 228Ra occurred in

aquifers of (1) metamorphic rocks, (2) quartzose

sands and (3) limestone. Thus, the lower Coastal

Plain, composed

of extensive limestones ansJ

deep quartzose

sand aquifers, is notable for its

total ‘lack-of 228Ra greater than 1.0 pCi/l in

1.80 0.0-15.9

‘I

I

---

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

Table

3.

Ranking ofall structural levels used showinggroupings ofsimilarity in

Ra

distribution.

Radiwn-226 values are given for comparison but are not ranked

Grouping*

Mean

Ra—228

(pCi/i)

No.

Structural Level

Mean

Ra—226

(pci/i)

A

3.03

2 Igneous, acidic, composite Th, 0,75

A

syenite

A

2.49 46 Sand, unconsolidated, arkosic, 2.03

A

,

high—Th source

A

B A

2,14 43 Sand, unconsolidated, arkosic,

2.73

B A

medium—Th source

BA

B A C

1.59 35 Igneous, acidic, composite Th, 2.31

B A C

granite

B 0 A C

0.85

3 Igneous, acidic, composite Th,

0.99

B 0 C

diorite

B D E C

0.52 31 Sand, unconsolidated, quartzose,

1.72

0 E C

medium-Th source

DEC

F 0 E C

0.39

37 Metamorphic, high—grade,

0.32

F 0 E

nonspecific Th

F 0 E

0.34 10 Metamorphic, low—grade,

1.41

F 0 E

nonspecific Th

F D E

0.28

3 Sand, unconsolidated, arkosic,

0.32

F 0 E

iow—Th source

F DE

F 0 E

0.28

13 Metamorphic, medium—grade,

0.24

F 0 E

nonspecific Th

‘

‘

F DE

F 0 E G

0.24 15 Metamorphic, high—grade,

0.24

F 0 E G

specific Th, monazite

F EG

F E C

0.12

2 Igneous, acidic, refractory rh,

0.14

F E C

monazite

FG

F

C

0.09 19 Sand, unconsolidated, quartzose,

0.09

F

G

low—Th source

C

C

0.06 16 Chemical precipitates, limestone

0.12

*A.BCDE

(po.05).

ground water. Radium-226 will be more variable

because of the high solubility of U complexes

in the carbonate system, but

it is generally

detected by the gross a-particle activity screen.

Nevertheless,

there have been only three 226Ra

violations reported for the entire lower Coastal

Plain province from New York to Texas. These

violations were all from one region in Florida.

The second area of high radioactivity is the

north-central region. Much of the ground water

comes from deep aquifers, frequently having

226Ra

activities of

5—25

pCi/I; 228Ra can be as

high as 32 pCi/I (Lu82). There is no apparent

correlation between 228Ra and 226Ra and no

562

---

C. T. HESS

et a!.

563

specific trends in their distribution by aquifer,

depth in the aquifer, or areal extent. Interpre-

tation of the Ra distribution in this area is

~ompIicated by complex hydrogeology and

multiply screened wells in different aquifers.

There is evidence of significant U migration,

both during geological time and the present,

which provides a mechanism for high 226Ra as

well as resulting in a complex distribution and

disequilibrium of U series isotopes (Gi82; Li02).

Possible sources for high 228Ra have not been

identified, but the 228Ra distribution may’ be

able to be explained by analysis of the sources,

depositional setting, and diagenesis of the sedi-

mentary rock aquifers.

Limited work has been done on Ra occur-

rence in the other geological provinces. Radium-

226 has been found to be high in areas of U

mineralization, such as in Texas and the Colo-

rado Plateau in Arizona and New Mexico (Fig.

2) and violations are ‘expected in Utah and

Colorado when these states report. Thorium

enrichment zones, such as veins and placer

deposits, are expected to produce only scattered,

local 228Ra problems, due to its limited transport

in ground water. These areas would be extremely

difficult to locate under the present regulations.

However, aquifers with much lower but dissem-

inated Th and U (such as granites, tuffaceous

rocks, and immature sandstones) are more likely

to have higher background radioactivity and

wider occurrences of both 228Ra and 226Ra in

ground water.

Occurrence of

224Ra

Data on 224Ra activities in ground water are

scarce. However, it appears that the activity of

224Ra is equal to or as much as twice the 228Ra

activity and therefore could be as high as 30—

40 pCi/I. This 224Ra activity is unsupported;

activities of its parent, 228Th, are usually less

than 0.01 pCi/I. Thus it enters ground water by

a recoil during decay of 228Th adsorbed on the

surface of aquifer solids. The radiotoxicity of

224Ra and its progeny is small because of their

extremely short half-lives.

RADIUM CONCLUSIONS

From the state compliance data and other

studies, much more is now known about the

occurrence of Ra isotopes in public drinking

water supplies, and’ this information should be

incorporated into the revised regulations.

(1) Surface water has very low Ra activities;,

the monitoring interval after initial validation

should be significantly lengthened, or perhaps

omitted, for surface-water systems.

(2) Fewer samples may be ‘needed to deter-

mine the average annual activities, particularly

for single-well systems; the monitoring interval

could be lengthened for unmodified systems.

(3) Monitoring requirements for 228Ra should

be decoupled from 226Ra. Instead, separate

guidelines for the occurrence of Ra isotopes are

needed.

-

(4) Regional 226Ra problems are fairly well

known. The few additional occurrences of high

226Ra activity will be difficult to find without

analysis of every system. As important, however,

are those

areas which have low

226Ra. All values

should be compiled regionally or nationally, to

document the 226Ra distribution for each geo-

logical province, with the goal to classify areas

with a high degree of certainty as producing

low 226Ra ground water. The revised regulations

should include separate, less stringent and less

costly monitoring requirements for such regions.

This approach would shift monitoring efforts

toward known or uncertain areas of high 226Ra

and provide more data on the actual distribution

of high

226Ra

activities, which will allow for a

better risk assessment.

(5)

The occurrence of 228Ra is not well

known. It

has been

shown that, using the present

screening procedures, 10—50

of the viola-

tions for total Ra are being missed. More exten-

sive measurements of 225Ra would be difficult

because of the problems with the approved

analytical method. An alternative approach

would be to develop a conceptual, predictive

model for

228Ra occurrence, based on geochem-

ical principles, to identify specific types of aqui-

fers which are likely to have 228Ra problems.

Once verified, this model should be the basis

for developing regional guidelines for

monitoring

in areas

more likely to have high 228Ra. This

same approach can be used for refining and

interpreting occurrence data for U and Rn as

part of the regulatory process of developing

standards for these

isotopes.

(6) Finally, because aqueous radiochemistry

---

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

is a complex, technical field, the EPA should

provide state water-supply personnel with back-

ground and explanatory guides in laymen’s lan-

guage, which will assist them in understanding

radiological problems and in the implementation

of the regulations.

OCCURRENCES

OF U ISOTOPES IN PUBLIC

DRINKING WATER

Introduction

Uranium has three natural isotopes with long

half-lives

(11/2)

that permit transport into potable

water supplies. These isotopes are 238U (99.27

natural abundance),

11(2

=

4.7

iO~yr,

(0.72 natural abundance),

11/2

=

7.04

X

108

yr,

and 234U (.006 natural abundance), 11/2

=

2.54

X 1 0~yr. All of these isotopes emit a radiation

and prodUce a long decay series of progeny.

The group of U isotopes are found in the Earth’s

crust with an abundance of 4 X 1O~ (Hu73)

and are found in rocks and minerals such as

granite, metamorphic rocks, lignites, monazite

sand, and phosphate deposits as well as in U

minerals such as uraninite, carnotite and pitch-

blend

(Ca80).

It is a trace element in coal, peat,

and asphalt and is present in some phosphate

fertilizers at a level of about 100 ~tg/gor 67

pCi/g. Despite its widespread abundance it has

not been shown to ,be an essential element for

man (Hu73). There is no standard for U in

water supplies as a radioactive element since,

until recently, it has been considered by the

U.S. Nuclear Regulatory Commission to be a

toxic heavy metal with the standard for ingestion

relating to its chemical toxicity (3 X iO~pCi/I);

(10

Code’ of Federal Regulations

20, 601, Ap-

pendix B). However, some concentration mea-

surements in potable water have been done in

association with gross a measurements for the

Ra drinking water standard. The U activity

measured was to be subtracted from the gross

a activity measurement to show compliance

with the gross a standard (Dr8 1). A recent

analysis of U in water supplies was conducted

by Oak Ridge National Laboratory (Dr8 1) using

89,944 measurements of U surface, ground and

domestic waters primarily obtained from the

NURE program. The results of this study are

reviewed in this report.

Geochemistry of

U

Although there are geological processes which

enrich U in certain rock formations, it occurs

as a common trace element in most rock types.

Because of the insolubility of U4~,U must be

oxidized in order to be transported in ground

water. The greater solubility of U” is due in

part to its tendency to form uranyl di- and tri-

carbonate anions. Thus, U solubility is a func-

tion of not only the redox potential of water

but also of the pH and the partial pressure of

CO2 in the system. In comparison to Ra, the

stability of the uranyl carbonate complexes and

their long half-life allow for U to be transported

long distances under oxidizing conditions. Ura-

nium is removed from solution by sorption or

reducing barriers, a process which has been

well described in the sandstone-type U deposits

in the western United States (Ga77).

There have been many studies of the isotopic

composition of U in natural waters which have

shown that most contain more activity from

234U than from 238U. The 234U/238U activity

ratio can be as high as 28, but usually ranges

between I and 3 (Ch7 1; Gi82). The

higher

activity of 234U in water is due to its selective

mobilization by

a recoil. The natural abun-

dances of isotopes and the half-lives give 0.33,

0.0 15 and 0.33 pCi/pg of natural U for 238U,

235U and 234U, respectively, or 0.68 pCi/pgtotal.

Thus, isotopic enrichment can cause changes in

the specific activity of the total sample of U.

Total depletion of 234U from the sample and

replacement by an equal activity of 238U will

result in no net change of total activity; however,

the total mass of U would almost double. The

human dosimetry will also be changed since the

a energies are not the same. Methods which

depend on the mass of U will not predict the

correct activity for samples with variable U

isotope enrichment.

Occurrence of

U

in ground and surface water

Uranium concentration in water depends on

factors such as the U concentration in the host

aquifer rock, the presence of 0 and complexing

agents, chemicals in the aquifer, chemical re-

actions with ions in solution and the nature of

the contact between the U minerals and the

water. These factors vary with regions of the

564

United

ground

genic

I

and s~

expect

to staU

mino u~

tions ol

are she

can be

eraged

shown

surprisi

factors

water ti

identica

average

Table 4

in provi

to 2.3 il

groupin;

I

I

1

~

FIG.

---

C. T. HESS

et a!.

United States due to rainfall, geology and

ground-water flow patterns, and to anthropo-

genic factors such as use rate of ground water

and surface water (Sc62). Thus, one would

expect large variations of U content from state

to state. The geological provinces of the coter-

nimbus United States derived from generaliza-

tions of rock types and hydrological flow systems

are shown in Fig. I and Table 1. These zones

can be compared to the U concentrations, av-

eraged by state from the (NURE) measurements

shown in Figs. 3, 4 and

5.

These averages are

surprisingly similar, showing differences of only

factors of 4 higher concentrations in ground

water than in surface water, with many states

identical for ground and surface water. The

average values for each province is given in

Table 4 with average values ranging from .02

in province 2, the Appalachian Mountains, up

to 2.3 in province 8, the Colorado Plateau. By

grouping low, medium and high averages, one

565

sees the four major zones of similar concentra-

tion: Zone I, the Appalachian Mountains and

New England; Zone 2, Appalachian and Interior

Plateau, and Coastal Plain; Zone 3, the Glaciated

Central Platform, Western Central Platform,

Rocky Mountain System, and Colorado Plateau;

Zone 4, the Basin and Range and the Columbia

Plateau and Pacific Mountain System.

The provinces chosen by Beddinger (Be8 1)

may be compared with those chosen in 1962

by Scott and Barker (Sc62). These provinces are

shown in Fig. I and compared by region in

Table 1. Concentrations of U are given in Table

4 for provincial schemes. The data of Scott and

Barker comprise 561 samples collected in the

coterminous

United States from 1954—58 and

are expressed in pg/I (thus they represent 238U

only).

Table

5

shows population vs U concentration

for drinking water sources with more than

10,000 people (Dr8 1). The levels are given up

FIG.

3. Population-averaged U concentration in picocuries per liter for surface water in the

United States.

---

566

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

FiG. 4. Population-averaged U concentration in picocuries per liter for ground water in the

United States.

to greater than 100,000 population. Due to

limitations in the source information (Dr8 1),

no information is available for cities of popu-

lations less than 10,000, showing a need for

more information on small systems.

Relative source contribution of

U

The dietary intake of U in United States food

is variously reported from 0.87—0.94 pCi/day

(Ha73) to 0.2—0.9 pCi/day (UN77) with an

average of 0.4 pCi/day. The comparison with

drinking water

of average concentration of 2

pCi/l and two liters per day consumption gives

4 pCi/day of water-derived U which is 5—10

times greater than the food-derived

U. Air con-

tributions of U are much smaller than the food

and water contributions.

Uranium conclusions

(1) The data shows that elevated levels of U

found are found in surface water as well as in

ground water.

(2) The highest average values of U concen-

tration are found in decreasing order in the

following provinces (see Fig. 1): Colorado Pla-

teau, Western Central Platform, Rocky Moun-

tain System, Basin and Range and Pacific

Mountain System. The state with the highest U

concentration is South DakOta. Modeling these

variations would be very helpful for regional

standards.

(3) High U concentrations in the East are

widely separated, and most values in the East

are low.

(4) Isotopic estimates are needed for 238U

and 234U since they are found in disequilibrium

in water. Regional variations should be modeled.

(5)

More analyses are needed for low popu-

lation systems.

OCCURRENCE OF Rn ISOTOPES

IN PUBLIC

DRINKING WATER

Introduction

There are two isotopes of Rn, with half-lives

long enough to be considered as drinking water

radionuclides. The first is 222Rn which is the

Fic

progeny

‘

half-life o

is the prr

“thoron,”

time dela

water of

allows m~

is not ob~

hencefortl

and can

ingested a

well as by

lungs. WI:

washing, t

from the

into a-em

eny are cI

aerosol pa

smoke, or

and may I:

lung, bring

075

.075

I;

---

C. T. HESS

et a!.

FIG.

5. Population-averaged U concentration in picocuries per liter for domestic water in the

United States.

567

progeny of 226Ra, called “radon,” and has a

half-life of 3.84 days. The second, 220Rn, which

is the progeny of 224Ra, was historically called

“thoron,” and has a half-life of 56 sec. The

time delay from production to consumption of

water of a few hours to a few days for water

allows many decay half-lives for 220Rn, and it

is not observed in water supplies. Radon-222,

henceforth simply Rn, is transported by water,

and can lead to public exposures by being

ingested and exposing the digestive system as

well as by becoming airborne and exposing the

lungs. When water is used for cleaning, dish-

washing, bathing, or clothes washing, Rn escapes

from the water into building air where it decays

into a-emitting progeny. The resulting Rn prog-

eny are charged and will frequently attach to

aerosol particles in the air. The dust, cigarette

smoke, or aerosol particles will then be inhaled

and may become attached to the interior of the

lung, bringing the a particle-emitting Rnprogeny

into close association with the cell lining of the

respiratory system

(Ar75).

There is no federal standard for Rn in water,

although studies on ingestion doses (Hu65; As79)

and inhalation doses have been done (Pr8l;

He83a; He79). There have been standards for

Rn in mine air, and for Rn from soil gas in

buildings placed on mine tailings in the United

States and Canada (Us79; At77). Sweden has

standards for Rn from soil gas in areas of alum

shale and granites (Ak81). Some of these stan-

dards are given in Table 6. In the past, Rn and

its progeny have been excluded from the drink-

ing water standards, and considered only to be

an interference in the Ra measurements.

Information about levels of Rn have been

obtained in state studies and by a federal study

done by the U.S. Environmental Protection

Agency Environmental Radiation Facility in

Montgomery, AL, and by the University of

Texas at Houston.

C

a

J

---

568

Beddinger Provinces

1. New England—-

Adirondack Mts.

2.

Appalachian

Highlands Piedmont

3. Appalachian

Interior

Plateaus

4.

Coastal Plain

5. Glaciated Central

Platform Aquifers

6. Western Central

Platform

7. Rocky Mountain

Systeat

8. Colorado Plateaus

9. Basin and Range

10. Columbia Plateaus

11. Pacific Mountain

System

Radon-Radium geochemistry

Radon is a water soluble inert

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

Table

4.

Uranium concentration in the provinces

Beddinger’ s

uranium concentrations

(pCi/i)

Arithmetic

Mean

gas and its

occurrence is controlled by physical variables

such as pressure, temperature, eniissivity of ,Rn

from rocks, as well as by time, and by the

geochemistry of its parent 226Ra. High activity

of Rn is associated with granitic rocks (St8 1),

U minerals (Ta78), such as uraninite, carnotite

and with tailings from phosphate fertilizer pro-

cessing (Us79) and U mines.

Transport of

Rn

in waterfrom rocks

As discussed in the gecichemistry section, the

occurrence of Rn in water is controlled by

chemical concentration of Ra in the host soil

Standard

Deviation

0.46

0.03

0.020

0.013

0.137

0.27

0.108

0.19

1.04

2.17

2.1

2.2

1.99

1.27

2.31

1.4

2.15

1.53

0.52

0.389

1.41

1.81

Scott’s uranium

concentrations

Arithmetic(pCi/i)

Mean

II

.34

III

.34

I .14

V .34

VI

.71

VI

1.5

VII-.VIII 1.15

IX

.54

VIII

1.1.5

X 0.14,

on rock and by emissivity of Rn into the water.

The physical condition of the rock matrix ap-

pears to play a greater role in Rn production

than does the concentration of parent Ra. Sev-

eral investigators have examined the mecha-

nisms influencing the release of Rn from rock

grains and the transport of Rn ‘through an

aquifer (An72; Ra83; Ta64; Ta80). Experimental

and theoretical considerations indicate that dif~

fusion along microcrystalline imperfections

dominates the release of Rn into the surrounding

interstitial waters. The movement of Rn in

water is governed by water transport rather than

diffusion in most cases, i.e. cases in which

the percolation velocity is greater than l0~

cm/sec.

typ

TaI

tral

util

alsc

low

hig~

utili

Ca;

S we

---

C. T. HESS

et aL

Table 5. Population vs concentration of

U

distribution for drinking water sources that serve

more than

10,000

people. The numbers in parentheses are the number of cities sampled

followed by the total in the category

pOPULATION

0.05

—0.5

0.51

-1.0

1.1

-2

2.1

—3

3.1

—4

4.1

—5

5.1

—6

6.1

—7

7.1

—8

8.1

—9

9.1

—10

10./0

—11

greater than

11 pci/i

10,001

—

50,000

292

(532/2407)

75

46 14 17

4

75 4

2

3

50,001

—

75000

36

(63/224)

4

4 3 3

10

1 1

(12pci/i)

75,001

—

100,000

15

(31/101)

5314

3

greater

than

100,000

117

(174/276)

15 10 7 8 3

7 4 2

(30.2pCi/i)

+assumed

secular equilibrium of

U—234

and U—238.

The variation of Rn concentration with rock

type for well water in Maine is illustrated in

Table 7. This variation of a factor of 20 illus-

trates variation in rock types in Maine. Public

utility water measured at the wells, which is

also presented in the Maine list, is a factor of

5

lower than the state average. This is due to the

higher capacity of the gravel aquifers used for

utility water supplies. These gravel aquifiers

allow more water

to mix with the same amount•

of Rn, leading to lower concentrations of Rn.

Occurrence of

Rn

in public water supplies

Concentrations ofRn in various water sources

conform to the log normal distribution. Table

8 shows the results of a blind sampling of public

water supplies in the central United States. A

reanalysis of other published data (Table 9)

Table

6.

Standards for

Rn

in air and proposed action to be taken

United States (Us79)

Canada (At77)

Sweden (Ak81)

In phosphate—mining regions in Florida

4 pCi/la take remedial action

2 pCi/l: reduce to as low as is

reasonably achievable

In uranium—mining regions

30 pCi/l:

4 pCi/i:

2 pCi/l:

11 pCi/i:

4

pCi/i:

2 pCi/l:

take prompt remedial action

take remedial action

investigate

for existing buildings

for houses undergoing remodeling

for newly constructed houses

569

---

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

Table

7.

Average

Rn

values in private and public water supplies

Arithmetic

Mean pCi/i

Maximum pci/i

Number

of Samples

Maine (He79,83b)

10,000

1,000,000 2,000

In granite zones

22,000

300,000

136*

In sillimanite

grade zone

13,600

100,000

35*

In chlorite

1,100

2,500

56*

In public utilities

2,000

11,700

64

*Rock grade determined by geologist for each private well.

Table

8.

The distribution ofRn in municipal water supplies in the central United States

STATE

N.

Well

G.M.

(pci/i)

Waters

G.S.D.

Distr

N.

ibution S

G.M.

(pCi/i)

ystems

G.S.D.

Arkansas

6

135

2.66

20

47

1.95

Indiana

10

151

2.13

23

70

220

Iowa

33

175

2.13

31 111

,

2.45

Louisiana

61

151

1.84

22

93

2.54

Minnesota

28

252

2.08

28 183

2.65

Nebraska

47

262

1.85

21 178

3.68

New Mexico 27

287

2.39

20 220

2.16

Oklahoma

7

117

1.74

6 134

1.17

COMPOSITE 209

197

2.10

174 115

2.75

Texas

278 131

2.70

G.M.

=

Geometrical Mean

G.S.D.

=

Geometrical Standard Deviation

N.

=

Number of water supplies sampled

570

I;

---

C. T. HESS

et al.

571

Table 9. Radon-in-water results by state and source. Results are geometric means in units in

pCi/I.

Parentheses values are numbers of samples

PUBLIC

STATE

PRIVATE

WELL

PUBLIC WATER

SUPPLY*

GROUND-WATER

SUPPLY

PUBLI

WATER

C SURFACE

SUPPLY

**

AL

120 (22)

8 (31)

70 (182)

ND (8)

AR

230

(

2)

1400

(

1)

12

C

22)

ND (1)

AZ

‘ ——

——

250 (124)

ND (6)

CA

43

(

6)

790

(

2)

470

(

15)

ND (2)

CO

——

--

230

C

76)

——

DE

-—

—-

30

(

72)

——

FL

6000 (34)

320

C

2)

30 (327)

——

GA

2100

(

2)

44 (32)

67 (225)

43 (2)

IA

--

—-

220

C

85)

ND (2)

ID

——

—-

99

(155)

——

IL

—-

-—

95

314)

——

IN

—-

--

35 (185)

——

KS

——

——

120

C

47)

74 (2)

KY

1500 (10)

ND (18)

32 (104)

ND (5)

MA

1000

(

8)

7

C

2)

500 (212)

38

(2)

ME

7000 (24)

990 (71)

MN

1400

C

1)

600

(

1)

130 (233)

——

MO

ND

C

2)

—-

24 (138)

ND (2)

MS

——

260

C

2)

23 (104)

——

MT

4300

C

8)

—-

230

( 71)

ND (6)

NC

15 (29)

27

C

2)

79 (404)

ND (4)

ND

——

440

C

2)

35 (133)

——

NH

1400 (18)

9 (12)

940

C

52)

ND (6)

NJ

——

——

300 (

38)

——

NM

59

(14)

45

C

8)

55 (171)

ND (18)

NV

——

—-

190

C

57)

——

NY

1500

C

4)

34 (20)

52 (292)

ND

C

1)

OH

-—

——

79 (165)

—-

OK

——

——

93 ( 83)

——

OR

450 (18)

——

120

69)

ND

( 4)

PA

910 (16)

—-

380

105)

——

RI

6500 (69)

5200

C

6)

2400

575

)

ND (10)

SC

1100 (28)

——

130

384 )

ND (14)

SD

4200

(

2)

59

(

2)

210

C

155

)

——

TN

ND

C

2)

ND

C

2)

12

(

98)

——

UT

‘

——

——

150

C

195)

——

VA

560 (42)

——

350

C 284)

ND

C

4)

VT

210 (23)

840

(

4)

660

C

71

)

‘

13 (16)

WI

730 (40)

28

C

4)

150

(

278

)

ND (12)

WY

——

--

330

(

32)

ND

C

2)

US

920 (434)

68 (224)

130 (6298)

1 (131)

*May include both ground water and surface

**ND —

Not detected above background levels

shows

a similar trend. Some sources appear to

be samples from a single log normal distribution,

others from two, or perhaps three distributions,

as indicated by the sharp breaks or bends in the

plots. For this reason geometrical averages are

used for the sample shown below.

The occurrence of Rn in public ground-water

supplies in the United States is shown in Fig. 6

---

572

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

FIG.

6. Geometric average Rn concentration in picocuries per liter for public ground-water

supplies in the United States.

and Fig. 8 (taken at the tap) and Table 9

(mainly utility samples). Radon activities are

thousands of times higher than U or Ra, prob-

ably due to absorbtion of the Ra and U by the

host rock (Ki82). The results of these geometric

means show highest values in mountain states

especially in the Appalachians with the highest

states Rhode Island, Maine, New Hampshire,

Vermont, Massachusetts, Pennsylvania and

Virginia. California is the highest western state.

The high Rn values associate with the granitic

areas in the Appalachian Highlands Piedmont

Provinces (Fig. 1). Midwestern and

coastal plane

values are lower and mountain states in the

west are higher. These results of geometrical

averages show the private supplies are higher

by a factor of 3—20 times the public ground-

water samples. This factor results from the use

of low capacity wells for private supplies while

public supplies use high capacity sand or gravel

wells. The higher states in the private well list

are Rhode Island, Florida, Maine, South Dakota,

Montana and Georgia. Larger numbers of sam-

ples would be desirable to strengthen these

conclusions. Public surface supplies have Rn

concentrations less than 100 pCi/I. Table 10

shows a breakdown of the Rn concentrations

in water by state and by population of the town

using the well. Highest Rn concentrations are

found in the less than 100 category (see

Maine and United States for examples). The

United States geometric population average is

187 pCi/I.

Major sources of indoor

Rn

Radon produced from Ra in the surficial soil

and rock (Ak8 I) is released into houses from

water, soil gas, fuel gas, and construction ma-

terials and outdoor air. Both water and soil gas

can be transported into buildings through cracks,

drain holes, as well as water and fuel gas supply

pipes (Sc82). The Rn in the ground water is

released as it is mixed with air in such indoor

uses as cleaning, bathing, dish and clothes wash-

U)

Lzi

C,)

ci:

C)

U-

0

(ii

0

z

I(

(

F;

ing and tc

will mix

throughou

enters bui

stoves. Th

depend or

the volum

rates of th

value for F

has been

regions,

1.

United Sto

---

-~

C. T. I-LESS

et al.

COMBINED RADIUM CONCENTRATION (pCi/I)

FIG. 7. Number of utility water supplies above

5

pCi/I with a measured Ra concentration.

30

573

ing and toilet flushing (He83a; Pa79). Soil gas

will mix into building air and then diffuse

throughout the house. Radon from fuel gas

enters building air from unvented heaters or

stoves. Thus, the Rn concentration in air will

depend on the sum of all Rn sources (Ge78),

the volume of the building, and the ventilation

rates of the building (Fl80; Ne8 1). The average

value for Rn in house air due to all these sources

has been estimated at 0.3—2.2 pCi/i in normal

regions,

1.1—1.67 in anomalous regions of the

United States (Br8 I).

The soil.

Radon diffuses from the soil through

cracks in foundations, unventilated crawl spaces,

basement drains, and other pathways into the

living space. Direct outgassing from the soil is

the dominant source of indoor Rn in most

cases contributing 0.03—1.5 pCi/l in normal

regions, and 0.3—15 pCi/l in anomalous regions.

If most Rn enters structures through the base-

ment or foundation, Rn concentrations would

tend to decline markedly with story above

ground, as shown in Table 11. The limitations

of building materials and ground waters as

30

U)

w

U)

C-)

U-

0

w

0

2

5

0

0

5

$0

15

20

25

---

574

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

.

0

z

U

D

0

U

U.

—

0

10

II)

-~~1~

-i

I

20.00

5000

100.00

+

CONCENTRATION (pCi/I)

x

FIG.

8. Radon concentrations in picocuries per liter for all public ground water supplies in

the United States.

1.00

5.00

10,00

15.00

F

sources of Rn (see below) coupled with the

tendency toward single distributions in given

areas combine to suggest that the soil is generally

the largest source.

Ground water.

Ground waters containing Rn

can add substantially to the amount of Rn in

the air of a dwelling. Much of the dissolved Rn

can escape when water is used for various

domestic purposes inside a dwelling. The

amount of Rn in indoor air due to the use of

water depends a great deal on architectural and

life-style related variables. The most sensitive

dwellings will be small, relatively tight structures

in which large amounts of water are routinely

used in household appliances. A model for the

average increment to the indoor atmosphere

can be expressed as:

Ca

24RV ~

where

Ca

and

C~,

are the concentrations (pCi/I)

of Rn in the air and water, respectively,

R

is

the air change rate (hr~),

V

is the dwelling

volume,

w1

represents the average amount of

water (1) used daily in the ith domestic appli-

cation, and

e

denotes the transfer efficiency, or

the fraction of Rn released to the air for the ith

application.

A number of investigators have made semi-

empirical determinations of

Ca/Cw;

most are on

the general order of l0—~pCi/I

in the air per

pCi/i in the water. Table 12 shows a number

of such estimates and the underlying assump-

tions used in their derivation. Using Table 12

the range of values for Rn in air from water in

the United States ranges from

0.2 pCi/I average

for Rhode Island to as low as 1.2

X

i0’~pCi/I

average for Tennessee.

The relative importance of water as a source

of indoor Rn will of course depend upon the

amount of Rn in the water and the magnitude

of other sources. Upper limit calculations. on

one recent data set (Table 12; Pr83a) show

water as the source of up to

35

of the net

(indoor minus outdoor) Rn observed in a set of

81 bedrooms in single-family houses in the State

of Maine. The highest Rn levels were seen in

the basement, which

suggests that the soil is a

major source of indoor Rn. The long-term

average levels noted in the bathroom were also

higher than those noted in the other living areas,

however. Both the integrated indoor air Rn

100.00

0,

ID

10

ID

0’

ID

ID

N

---

C. T. HESS

el aL

575

Table

10.

Radon-in-water results by state and population. All results are geometric means in

units of

pCi/I.

Parentheses values are in numbers ofsamples

Population Ranges

STATE

100

100—1000 1000—5000 5000—10,000

10,000 Unknown

AL

——

83 (lO)*

59 (76)

82

(46)

68 (44) 170

C

6)

AR

——

32

C

6)

6

(

3)

——

12 C 2)

8 (13)

AZ

——

240

C

2) 200 (68)

350 (22)

340 (30) 160

(

2)

CA

—-

—-

——

——

——

470 (15)

CO

——

130 C 6) 220 (.54)

400

C

8)

300 C 8)

——

DE

——

100

C

4)

39 (44)

ii (12

)

23 (12)

FL

—-

320

(

2) 290

(

4)

49

(78

)

24(243)

——

GA

57

(

4) 42 (12) 130 (56) 190

(27

)

52 (45) 39 (81)

IA

—-

1120 (13) 230 (36) 150 (24)

66

(10) 75 C 2)

ID

6

C

3) 210 (14) 130 (83)

36 (25)

100 (30)

——

IL

——

71

C

1)

81 (30)

80 (98)

hOC

185)

——

IN

——

——

45 (71)

17 (58)

50 (56)

——

KS

KY

260 C 2)

——

230 (12)

20 (10)

43 Cli)

42 (76)

320 C 4)

58

36 C 8)

5

C 8)

(10)

210(10)

——

MA 3300

C

2)

——

380 (47) 540 (67)

510

(88) 58D( 8)

ME

MN

670

C

3)

—-

1600 (23)

10 (22)

690 (33) 2700 C 7)

450

150 (76)

68 (43)

140

C 4)

C9O)

400C 1)

33OC 2)

MO

ND

C

2) 58 (54)

26 (54) 1-100

C

4)

78

C

6)

ND(18)

MS

—-

——

15 (45)

51 (26)

23

(33)

——

MT

740

C

4) 280

C

6)

270 (49)

160

C

8)

18

C

4)

——

NC

ND

NH

11,000

C

6)

——

1700

C

2)

250(111)

13 (10)

1200

C

4)

45 (229)

16 (32)

21

39 (112)

.

97

C

5)

10

960 (24) 1000 (14)

550

(16)

C

6)

C

8)

8100(10)

——

——

NJ

—-

——

-—

360 C 6)

1200 (12)

120 (20)

NM 1300

(

2) 620

C

7) 48 (89)

42 (16)

33 (49) 450

(

8)

NV

——

220 C 2) 240 (36)

72 C 9)

530 C 2)

180

(

8)

NY

——

——

56

(.59)

31 (85)

71(113)

56 (35)

OH

——

300 C 2) 160 C 4)

56 (76)

100 (83)

——

OK

—-

260 C 1)

65 (33)

190 (21)

79 (26)

96

C

2)

OR

—-

210 C 2) 110 (48)

23 C 6)

320

C 6)

180 C 7)

PA 1900 C 2)

——

260 (34) 450 (32)

440 (37) 910 (16)

RI 1700( 91) 3900(325) 980 (71) 1300 (30

)

1200 (58)

——

SC

SD

1400 (32)

310 (16)

870 (30)

300 (41)

92 (229)

160 (85)

74 (60

210

(

8

)

60 (31)

)

200

C

5)

410 C 2)

——

TN

UT

——

——

160 C 4)

260

C

2)

19 (48)

140 (104)

6 (25

200 (39

)

5 (21)

)

170 (48)

——

150 C 2)

VA

85 C 6) 880 (56) 140 (151) 320 (12

)

720

14) 2300 (45)

VT

120 C 1)

——

540 (24)

-—

1000

C

2)

750 (44)

WI

——

——

150 (169)

190 (61)

130

48)

——

WY

US

——

990(175)

880 C 6)

620(777)

230 (19)

98(2446)

530

(

6)

54 C 1)

92(1098)

88(1464)

——

140(338)

*Number of data points used to calculate geometric mean.

**ND

—

not detected above background levels.

concentrations and the concentrations of Rn in

water were distributed log-normally, and a

ratio

of 4 pCi/I (bedroom air)

per 1 05 pCi/l (water)

was found by regression. The application of this

factor to the geometric mean of the water

distribution led to a predicted increment of023

)

---

576

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

Table

11.

Indoor

Rn

concentrations

(pCi/I)

in “background” U.S. dwellings

Number of

Site

Dwellings

Geom.

Mean

Geom

S.D.

Eql.

Ratio

Comments Reference

New York

New Jersey

Basements

18

1.7 2.0 0.48 Ordinary

Ge80

1st Floor

18

0.83 2.0 0.49 Houses

2nd Floor

9

0.77 1.8 0.45

Central Maine

Ordinary

Houses

Pr83a

Basements

77

2.46 2.4

lstFloor

82

1.40 2.4

Bedroom (2nd) 81

1.12 2.4

Bathroom

81

1.62 2.4

Outside

67

0.46 2.4

Houston, Texas

Houses,

Apts.

Pr83a

Bedroom

103

0.39 2.5

Bathroom

103

0.58 2.5

Outdoors

81

0.22 2.6

Eastern

Pennsylvania

Sa81

Living Area 36

Summer

1.22 4.6

Winter

4.40 3.7

Basement

36

Summer

3.40 6.4

Winter

5.90 4.0

pCi/I in the air vs a net bedroom concentration

of 0.66 pCi/i. This is an upper limit estimate,

based on the assumption that the magnitudes

of other sources of Rn are independent of the

amount of Rn in the water. In fact, the concen-

tration of Rn in the basements was found to be

correlated with the concentrations of Rn in

water at the p

=

0.01 level.

Construction materials.

A number of recent

publications have addressed the problem of

elevated indoor Rn concentrations arising from

the use of building materials containing 226Ra

(Br8 1). A portion of the Rn arising from con-

struction materials is able to diffuse into the

living area, where the ultimate concentration

increment will depend on the volume of the

dwelling and the ventilation rate ranging from

.003 pCi/I to 0.3 pCi/I in the United States

(Br8l).

Perhaps

the best publicized case involving

construction materials occurred in Grand Junc-

tion, CO, where U-mill tailings were once fre-

quently used as fill materials around founda-

tions. More than 5,000

buildings were associated

with tailings material to some extent, and 3,000

of those buildings were built on top of a layer

of tailings. When Rn levels in some of the

dwellings were found to be markedly elevated

a general survey was conducted, and eventually

a remedial action program was implemented

for those structures exceeding national standards.

(See Table 6).

A similar situation arose in the phosphate-

mining area of Florida. Dwellings were built on

mining lands reclaimed at least in part with

phosl

well

conta

derive

made

Fu

noted

until

to inc

vestig;

natun

stiinu

of stir

a nuci

matio:

causec

home

(Jo73)

were

I

which,

---

C. T. HESS

et a!.

577

Table

12.

Factor relating

Rn

in indoor air to

Rn

in water

(pCi/ 1

in air per

10 pCi/I

in water)

Factor

Reference

Conditions

CKa8O)

Calculated value for typical

Finnish single and double—family

structures, based on

experimentally determined radon

releases.

Calculations based on 4 occu-

pants, experimentally deter-

mined Rn releases.

Volume

15D00 1

34000 1

34000 1

Air Change Rate

0.25 hr~

0.501.00

hr~hr—1

Observation obtained by

regression from 3 months

integrated data from 80 houses

in Maine, causality not

strictly implied.

Air and water grab samples in

13 structures in Halifax, Nova

Scotia.

phosphate-rock residues. Radon concentrations

well over 10 pCi/i have been noted in buildings

containing materials such as gypsum wall board

derived from phosphate residues or cinderblocks

made of fly ash or blast furnace slag (US79).

Fuel gas.

Radioactivity in natural gas was

noted as early as 1904 (Sal 8), but it was not

until the 1970s that the potential contribution

to indoor Rn concentrations was seriously in-

vestigated. Interest in the health implications of

naturally occurring Rn in gas deposits was

stimulated by investigations of the practicality

of stimulating natural gas yields by detonating

a nuclear warhead in the appropriate rock for-

mation (Bu66). Estimates of dose increments

caused by the combustion of natural gas in the

home were made by Barton (Ba73), Johnson

(Jo73) and Gesell (Ge74). Similar calculations

were made for liquified petroleum gas (LPG),

which, because of boiling point considerations,

contains a higher concentration of Rn than the

natural gas from which it is made.

The increment to the indoor environment

depends on the amount of gas of LPG burned

in unvented ranges or heaters, the size and

infiltration characteristics of the dwelling, and

of course, the concentration of Rn in the fuel.

The contribution from this source is usually

quite small (0.15 pCi/I) due to the low use of

unvented heaters (Ge77).

Occurrence of

Rn

in indoor air

There are a number of situations in which

indoor Rn levels are especially elevated. These

situations occur when the structure contains a

stronger than usual source of Rn, or when. the

structure has especially low ventilation and in-

filtration rates or both. Rising heating and air

conditioning costs in the last several years have

14

(Ge8O)

50

10

5

4

(Pr83a)

1

(Mc8O)

---

578

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

ehcouraged people to reduce air infiltration

rates.

A number of trends can be discerned in the

recent literature. Within a given locale, indoor

Rn concentrations tend to be distributed log-

normally, and sample means vary markedly

from area to area (Ge83). it is becOming in-

creasingly apparent that local geological factors

play

a major,

if not dominant role in determin-

ing the distribution of indoor Rn concentrations

in a given area.

Concentrations of

Rn

in the indoor environ-

ment.

One of the most extensive studies of Rn

in dwellings on record is a recent survey of

12,000 Swedish homes (Hi8l). All the dwellings

involved were ones in which elevated Rn levels

were expected. The results as shown in Table

13 and 14 are reported in working levels, and

the presumed associated Rn concentrations,

based on an equilibrium ratio of 0.5, are added

in parentheses.

The measurements summarized above are

not meant to be representative

and are probably

considerably higher than the true area average.

Nevertheless, the number of dwellings involved

and the high values observed combine to dem-

onstrate the extent of the Rn problem in certain

areas. Nearly half of the 12,000 dwellings were

found to exceed current U.S. occupational stan-

dard for U mines (adjusted for constant occu-

pancy vs a 40-hr week).

Another extensive survey involved nearly

10,000 randomly selected houses in 14 Canadian

cities (Mc80). Single grab samples were obtained

from basements or the ground floor during the

summer months. In general, data from a given

city were better fit by the log-normal distribution

than by the normal distribution. The geometric

means ranged from 0.14 pCi/I in Vancouver to

0.88 pCi/I in St. Lawrence, Newfoundland. The

geometric standard deviations ranged from

2.78—6.77.

Table 11 summarizes a number of surveys

conducted in the United States in areas not

known or suspected to involve anomalies due

to mill tailings or unusual mineralization: The

data presented are either the average of a number

of grab samples taken within a single dwelling,

or were developed by long-term measuring de-

vices. The equilibrium fraction

(f)

of Rn prog-

eny is given where available.

Radon conclusions

(1) Radon concentrations in water are highest

in ground water especially in granite areas.

Radon concentrations

iii

surface water are

very low.

(2) Higher concentration occurs in small sys-

tems. Domestic supplies have higher concentra-

tions of Rn in water than public wells. Utility

systems are lower by a factor of 10 than private

wells.

Table

13.

Radon levels measured in air in houses

Average

pCi/i

Maximum

pCi/i

Number

of Houses

Maine CHe82b)

3.0

63.0

85

New York

Albany Area (F180)

3.1

26.0

21

New York City Area (Br79)

1.0

5.9

21

California

San Francisco (Ne81)

0.4

1.1

28

Pennsylvania

Eastern (SaBl)

10

36

Maryland (Mo8l)

3.7

27,0

56

---

C. T. HESS

et a!.

579

Table

14.

Radon progeny in Swedish dwellings

(‘Hi8 1)

Radon Progeny

(Working Levels)

Radon

(pci/i,

f=O.5)

Number

of

Dwellings

Percent of

Dwellings

0.000

—

0.054

0.054

—

0.108

0.108

—

0.270

0.270

0

—

11

ii

—

22

22

—

54

54

6326

4050

1545

162

52

34

13

1

(3) The highest average Rn concentrations

in water are found in the provinces in

decreasing

order: New England, Appalachian Highlands-

Piedmont, Pacific Mountain System, and Rocky

Mountain System.

(4) Ventilation affects Rn concentration in

air with an approximate value of

I.

X iO~for

the ratio of Rn concentration in air to Rn

concentration in water for a house with one air

change per hour. Soil gas Rn contributes a

sizeable portion of the total Rn in air.

(5) Additional measurements of Rn in sys-

tems serving less than a thousand users are

required in order to better quantify exposures

to the group that potentially represents the

highest population dose. Because the number

of such systems is quite large (—37,000), these

measurements should be obtained from a rep-

resentative sampling program guided by geolog-

ical models.

CALCULATION OF WATER UTILITY RISKS

FROM RADIONUCLIDES IN WAThR BASED

ON NATIONAL OCCURRENCE DATA

Each nuclide present in drinking water will

present a risk to the utility users of the water

which is related linearly to the occurrence,

concentration, population exposed, and the in-

dividual risk rate (Ma83). The case of Ra in

drinking water data allows a direct calculation.

The average concentration of Ra in drinking

water from utilities is 1.6 pCi/I (Ho83), and the

population consuming this water is 70 million

people which is the half of the U.S. population

which uses ground water provided by utilities.

Since the individual risk rate is hypothesized to

be by linear dose response 8.4 X 10-6 excess

cancers/lifetime person pCi/I (using (Ma83), we

can calculate

1.6 pCi/I X 8.4

X I ~6

excess cancers

70

lifetime person I pCi/l

941 people

X 106 people

lifetime in U.S.

Even the elimination of bone sarcomas at low

concentration will leave the sinus carcinomas

which are half of this number (Ma83). An

additional 30 X 1 06 people are exposed at less

than 0.5 pCi/I surface water provided by utilities.

The distribution of occurrence of Ra concentra-

tions permits estimates of the number of fatal

cancers averted when the standard is placed at

a

particular concentration. Since the standard

is at

5

pCi/I, it seems reasonable to estimate the

fraction of cancers averted by the standard. The

average for the supplies of greater than

5

pCi/I

is 8 pCi/I. This concentration is multiplied by

an estimate of the population which uses those

supplies obtained from rounding the average

ground water utility population from the U.S.

Environmental Protection Agency summary. See

Fig. 7 for the number of utilities vs Rn concen-

tration.

8 pCi/I

1000 persons

500

supply

8.4 X 10—6 excess cancers

1

pCi/I person lifetime

—~

~ excess cancers

lifetime

Number

f Houses

85

21

21

28

36

56

---

580

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

Population risk can also be done for U in

water by using a similar calculation. For U the

population-weighted average radioactivity con-

centration is 0.8 pCi/I for the whole U.S. pop-

ulation (Dr81). This permits this result for the

risk using individual risk factors from Ma83,

Wr83 and Co83.

6

X l0~cancers

0.8

pCi/I x pCi/I

person X 220

105 persons

X I0~people exposed

=

lifetime in U.S.

Since there is no drinking water standard for

U, at this time, we must use the fraction of

people at each occurrence level to estimate the

number of cases avoided by a standard for water

utilities

Using values from Drury

et a!.,

a water utility

users’ risk estimate table can be compiled (Dr8 I)

(See Table

15.)

The population risk for Rn in utility water

supplies can be calculated using the Rn concen-

tration in utility water, the exchange rate from

water into air, the cancer rate per working level

month per million population and the exposed

population in millions. From the work done by

the Committee on the Health Effects of Radon

in drinking water (Cr83), we can obtain the

individual risk factor which includes the ex-

change rate for Rn from water into air and the

cancer rate per working level month per million

population exposed to air for a lifetime of 20

yr. This factor is 3 X I 0~ lung cancers per

pCi/I Rn in water. This factor is increased to 4

x iO~cancers when stomach and whole body

cancers are included (Cr83). The Rn in water

concentration data for the whole United States

are geologically

controlled and are generally a

mixture of low values around 100—200 pCi/I

and high

values of 10,000—1,000,000 pCi/i.

This

extreme range of values leads to arithmetic

averages which are strongly influenced by the

highest fewpoints. The geometric mean ofthese

data will average

the numbers with less weight

for the high

values. This geometric mean will

be lower than the arithmetic mean. We have

decided to calculate the risk with both of these

means.

Using data from the geometrical and arith-

metic means of Rn concentrations for utilities

of different sizes, and the geometrical and arith-

metic risk factor, we can calculate the population

at risk for the utilities serving less than 100

population, 100—1000 population, 1000—5000

population, 5000—10,000 population, 10,000—

100,000 population, and 100,000 and above

population. Lifetime risk is shown for both

geometric averages and arithmetic averages in

the right column of Table 16.

OTHER NATURAL RADIOISOTOPES OF

POSSIBLE CONCERN

The natural radioisotopes already discussed

above (U, 228Ra,

226Ra and 222Rn)

are of greatest

concern because of their long half-lives and the

Table 15. Cases preventedfor

U

concentrations in public water

pci/i

Supplies

Number

Supplies

People

Exposed*

Average U

Concentration

Cases

1 pCi

23.6

10,808

10.8 x 106

7.5

48

.5

pCi/i

7.88

3,609

3.6 x 106

17.5

38

10 pCi/i

3.46

1,584

1.5 x 106

30.3

27

20 pCi/l

1.33

609

0.6 x 106

54.8

20

*1000/supply

I

---

C. T. HESS

et al.

Table

16.

Assessment of water risks for

Rn

in public ground water

Population

•

Number of

Utilities

Sampled

Mean Concentration

Of Radon in Water

(pCi/i)

U.S. Population

Using This Water

(millions)

Lifetime

Risk

GEOM

-

ARITH

GEOM

—

ARITH

100

100

—

1000

1000

—

5000

5000

—

10,000

10,000

—

100,001

100,000

88

377

1223

549

704

32

990

—

6500

620

-

4100

98

—

390

92

—

350

92

—

290

52

—

150

1.03

7.4

14.0

8.4

28.3

14.3

59.13

407

—

2678

1835

—

1213i

548

—

2184

309

—

1176

1018

—

3283

297

—

858

4414

—

2231~

Population—weighted Average 187.0 pCi/i

—

944.0 pCi/i.

FACTOR USED:

0.4 x 10—6 deaths

pCi/i water

581

health risks associated with the activities that

can be present in public drinking water. How-

ever, there are two classes of other natural

radioisotopes that may be of possible concern:

(1) relatively long-lived isotopes whose activity

is derived from the aquifer, termed unsupported

(232Th, 230Th, 210Pb and 210Po) and (2) very

short-lived isotopes which “grow in” once the

ground water is pumped from the aquifer and

thus are supported (primarily 222Rn progeny).

However, very few data are available on the

activities of these isotopes in drinking water,

primarily due to their low solubility and/or the

difficulty of measuring isotopes of short half-

lives.

Table 17 shows the typical ranges of activities

of the longer-lived radionuclides in ground wa-

ter. Activities in surface waters will be extremely

low (except perhaps in hot springs) due to rapid

sorption onto suspended and bottom sediments.

The highest known activity of 232Th and 230Th

in U.S. drinking water is from a well in Cali-

fornia that also contained large amounts of

dissolved organic matter which probably com-

plexed with Th. Most other values are below

0.1 pCi/I. Thorium-230 would be expected to

be slightly higher than 232Th due to generation

by 234U in solution and by a recoil. Likewise,

detectable activities of 210Pb and 210Po would

be expected because of the relatively large

amount of 222Rn present in many ground waters.

The insolubility of these isotopes and their

short-lived precursors in the aquifer is demon-

strated by the fact that more than 99.9 of the

activity generated by 222Rn decay in ground

water is removed within the 2 hr necessary for

equilibrium to be established between 222Rn

and 210Pb. The only known anonymously high

210p0 value in drinking water is the surprisingly

large activity in Louisiana for which the source

has not been determined. In general, these

longer-lived isotopes are not expected to occur

in activities greater than 1.0 pCi/I.

The second class of radioisotopes of possible

concern are 222Rn progeny which reach equilib-

rium with 222Rn within 2 hr. In untreated

I

I

20

I

er

dy

er

es

a

i/I

IS

ic

it

11

e

e

S

I

I

~1

I

I

I

I

I

I

I

,~ :

:

:

LI

:

:

:

:

:

:

LI

:

:

:

---

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

Table

17.

Concentrations of

Th, Pb

and

P0

isotopes in ground water

(pCi/i)

Description

Th—232

Th—230 Pb-211

P0—210 Reference

New Mexico

Grants Mineral Belt

(54 wells)

Paquate—Jackpiie Area

0.01

0.02

0.39

Ka 76

Grants—Bluewater Area 0.01 0-0.04

0—0.66

Ka 76

United Nuclear Area

0—0.03 0—0.09~

0.3—2.3

Ka 76

Ambrosia Lake

0.03 0-0.08

0-3.8

Ka 76

Gullup Area

0.02

0-0.09

0-0.6

Ka 76

Rapides Parish

Louisiana (1 well)

290—607

Mu 82

California well

1.3

1.1

ND

This paper

Arizona well

ND*

ND

0.9

This paper

Connecticut

.

glacial drift

0.02 0.001

Kris 82

glacial drift

0.02 0.004

Krls 82

crystalline rock

0.03 0.06

Kris 82

sandstone

0.12

0.020

Kris 82

sandstone

0.07 0.005

Kris 82

sandstone

0.03

0.004

Kris 82

Leesville, S.C.

sand aquifer

o.oi

This paper

NO

=

not detected

ground-water systems, removal by

adsorption

would not be as rapid as in the aquifer due to

the smallsurface area of the distribution system.

Exposure from consumption of these supported,

extremely short-lived progeny should be evalu-

ated. Radon-220 and its progeny do not pose a

similar problem primarily because the 54.5-sec

half-life of 220Rn is too short to allow diffusion

from the aquifer materials and the initial activ-

ities are much lower.

Acknowledgments—We

would like to acknowledge

the help of the U.S. Environmental Protection Agency

especially Dr. C. Richard Cothern, and Dr. William

L.

Lappenbusch and Mr. J. S. Drury

of Oak Ridge

National Laboratory for helping us to obtain infor-

mation about U occurrence. We would like to ac-

knowledge the assistance of Dr. Willard S. Moore of

the University of South Carolina for previewing the

manuscript and suggesting useful changes. We would

also like to acknowledge the objectionsof Mr. Thomas

Horton to the inclusion of the risk section in this

582

---

paper. We thank the typist, Mrs. Patricia Heal, for

her patience in making revisions to the manuscript.

REFERENCES

Ak81

Akerblom G. V. and Wilson V., 1981, “Radon

gas—A radiation hazard from radioactive bedrock

and building materials,”

Bull. mt. Assoc. Engin.

Geology

23, 51.

An72 Andrews J. N. and Wood D. F., 1972,

“Mechanism of radon release in rock matrices and

entry into groundwaters,”

Institution of Mining

and Metallurgy Bulletin

81, 198—209.

Ar79

Archer V. E., Radford E. P. and Axelson 0.,

1979,

“Radon daughter cancer in man: factors in

exposure response relationships,” in:

Proc. Work-

shop on Lung Cancer Epidemiology and Industrial

Applications to Sputom Cytology,

p. 324 (Golden,

CO:

Colorado School of Mines Press).

As79

Asikainen M. and Kahlos H., 1979, “Anom-

alously high concentrations of uranium, radium

and radon in water from drilled wells in the

Helsinki

region,”

Geochim. et Cosmochim: Acta

43, 1681.

As81 Asikainen M., 1981, “State of disequilibrium

between U-238, U-234, Ra-226, and Rn-222 in

ground water from bedrock,”

Geochim. Cosmo-

chim. Acta

45, 201—201.

At77 Atomic Energy Control Board, 1977,

Criteria

for Radioactive Clean-up in Canada,

Atomic Energy

Control Board ofCanada, Ottawa, Ontario, Canada,

Info. Bulletin 77-2.

Be81 Beddinger M. S. and Sargent K. A., 1981,

“Identification of suitable geohydrologic environ-

ments for the disposal of high-level radioactive

waste,” in:

Proc. Workshop on Siting ofRadioactive

Waste Repositories in Geological Formations,

pp.

151—187 (Paris: Nuclear Energy Agency).

BEIR8O Advisory Committee on Biological Effects

of Ionizing Radiation, 1980,

The Effects on Popu-

lations of Exposure to LOW Levels of Ionizing

Radiation

(BEIR III), National Academy of Sci-

ences, National Research Council, 2100 Constitu-

tion Ave., Washington, DC.

Br79 Breslin A. J. and George A. C., 1979, “Radon

sources, distribution, and exposures in residential

buildings,”

Trans Am. NucI. Soc.

33, 145.

Br81

Bruno R. C., 1981, “Sources of indoor radon

in houses,” Paper presented at

mt. Symp. on

Indoor Air Pollution, Health, and Energy Conser-

vation,

13—16 Oct. 1981, Amherst, MA. (Available

from U.S. Environmental Protection Agency,

Washington, DC.)

Br83 Bruno R. C., 1983, “Sources of indoor radon

in houses: a review,”

Journal of the Air Pollution

Control Association

33, 102—109.

583

Bu66 Bunce L.A. and Sattler F. W., 1966, “Radon-

222 in natural gas,”

Radiological Health Data and

Reports

7, 441. (Available from Eastern Environ-

mental Radiation Facility, U.S. Environmental

Protection Agency, Montgomery, AL 36193.)

Ca80 Canadian Radiation Protection Bureau, 1980,

Section on Uranium from Canadian Radiation

Protection Bureau Guidelines for Canadian Drink-

ing Water Quality

1978 (1980

Support Documen-

tation),

Health and Welfare, Ottawa, Canada.

Ch71 Cherdyntsev V. V., 1971,

Uranium-234

(Je-

rusalem: Keter Press).

C183 Cline W., Adamovitz S., Blackman C. and

Kahn B., 1983, “Radium and uranium concentra-

tions in Georgia community water supplies,”

Health

Phys.

43,

1—12.

Co83a Cothern C. R. and Lappenbusch W. L.,

1983, “Occurrence of uranium in drinking water

in the United States,”

Health Phys.

45,

89—99.

Co83b Cothern C. R., Lappenbusch W. L. and

Cotruvo J. A., “Health effects guidance for uranium

in drinking water,”

Health Phys. 44,

377—384.

Co83c Cothern C. R. and Lappenbusch W. L.,

1983,

Radium-226 and Gross AlphaParticle Activity

Compliance Data for Drinking Water Supplies in

the United States,

U.S. EPA, Office of Drinking

Water WH550, Washington, DC 20460, Internal

Report.

Co84 Cothern C. R., 1984, Personal Communica-

tion, Office of Drinking Water, U.S. EPA, 401 M

Street SW, Washington, DC 20460.

Cr85 Cross F. T., Harley N. H. and Hofmann W.,

1985, “Health effects and risks from 222Rn in

drinking water,”

Health Phys.

48, 649—670.

Dr8l Drury J. S., Reynolds S., Owen P. T., Ross

R. H. and Ensminger J. T., 1981,

Uranium in U.S.

Surface Ground and Domestic Waters.

Health and

Environmental Studies Center Complex, Oak Ridge

National Laboratory, Oak Ridge, TN 37830 EPA-

370/9-81-001.

Dy76 Dyck W., Chatergee A. K., Gammell D. E.

and Murricane K., 1976, “Well water trace element

reconnaissance, Eastern Maritimes, Canada,”

J.

Geochem. Explor. 6,

139.

Ei80 Eichholz 0. G., Matheny M. D. and Kahn

B., 1980,”Control of radon emanation from build-

ing materials by surface coating,”

Health Phys.

39,

301.

El82 Elsinger R. J., King P. T. and Moore W. S.,

1982, “Ra-224 in natural waters measured by

gamma-ray spectrometry,”

Anal. Chim. Ada.

144,

277—28 1.

Ep76a U.S. Environmental Protection Agency,

1976,

National interim Primary Drinking Water

C. T. HESS

et a!.

I:

—

I

---

584

Regulations,

U.S. EPA, Office of Water Supply,

Washington, DC 20460, EPA-570/9-76-093.

Ep76b U.S. Environmental Protection Agency,

1976,

Determination ofRadium RemovalEfficien-

cies in Illinois Water Supply Treatment Processes,

U.S. EPA, Office of Radiation Programs, Washing-

ton, DC 20460.

Ev8 I Evans R. D., Harley J. H., Jacobi W., McLean

A. S., Mills W. A. and Stewart C. G., 1981,

“Estimate of risk from environmental exposure to

radon-222 and its decay products,”

Nature

290,

98.

Fe79 Femleck J. K. and Cadigan R., 1979,

Radium

and Uranium Concentrations and Associated Hy-

drogeochemistry in Ground Water in Southwestern

Pueblo County, Colorado,

p. 54, U.S. Department

of

the Interior, U.S. Geological Survey, Reston,

VA 22092, Report No. 79-974.

Fi75 Filliben J. J., 1975, “The probability plot

correlation coefficient test for normality,”

Tech-

nometrics

17, Ill.

Fi79 Findlay W. 0., 1979, “Application of radon

standards to new and existing housing in Elliot

Lake, Ontario,” in:

Radon in Buildings

(Edited by

E. Colle), National Bureau of Standards Special

Publication 581. (Available from U.S. Government

Printing Office, U.S. Department of Commerce,

Washington, DC.)

F180 Fleisher R. L., Mogro-Campero A. and Turner

L. G., 1980,

Radon Levels in Homes in the North-

eastern United States: Energy-Efficient Homes,

•

General Electric Research and Development Cen-

ter, Schenectady, NY 12301, Report. No.

80CRD288.

Ga77 Gabelman 1. W., 1977, “Migration of ura-

nium and thorium—exploration significance,” in:

Studies on Geology No.

3 (Tulsa, OK: American

Association of Petroleum Geologists).

Ge74 Gesell T. F., 1974, “Estimation of the dose

equivalent to the U.S. population from radon in

liquified petroleum gas,” pp. 347—354, in:

Proc.

8th Midyear Topical Symp. of Health Physics

Society.

(Available as U.S. Atomic Energy Com-

mission Report 741018 from National Technical

Information Services, U.S. Department of Com-

merce, Springfield, VA 22151.)

Ge75 Gesell T. F. and Prichard H. M., 1975, “The

•

technologically enhanced natural radiation envi-

•

ronment,”

Health Phys.

28, 36 1—366.

Ge78 Gesell T. F. and Prichard H. M., 1978, “The

contribution of radon in tap water to indoor radon

concentration,” pp. 1347—1363, in:

Natural Radia-

tion Environment

III, CONF 780422, U.S. Depart-

ment of Energy Special Symposium Series 51.

(Available from National Technical Information

Services, Department of Commerce, Springfield,

VA

22151.)

Ge80 Gesell T. F. and Prichard H. M., 1980, “The

contribution of radon in tap water to indoor radon

concentrations,” pp. 1347—1363, in:

Natural Ra-

diation Environmental

III, Vol. 2. (Available from

National Technical Information Services, U.S. De-

partment of Commerce, Springfield, VA 22151.)

Ge83 Gesell T. F., 1983, “Background atmospheric

222Rn concentration outdoors and indoors: a re-

view,”

Health Phys.

45,

289—302.

Gi82 Gilkeson R. H. and Coward J. B., 1982, “A

Preliminary Report on 238U Series Disequilibrium

in Ground Water of the Cambrian-Ordovician

Aquifer System of Northeastern Illinois,” pp. 109—

118, in:

Isotope Studies of Hydrologic Processes

(Edited by E. C. Perry and C. W. Montgomery)

(Dekalb, IL: Northern Illinois University Press).

He79 Hess C. T.,. Norton S. A., Brutsaert W. F.,

Casparius R. E., Coombs E. G. and Hess A. L.,

1979,

Radon-222 in Potable Water Supplies in

Maine: The Geology, Hydrology, Physics and

Health Effects,

Land and Water Resources Center,

University of Maine, Orono, ME.

He83a Hess C. T., Weiffenbach C. V. and Norton

S. A., 1983, “Environmental radon and cancer

correlations in Maine,”

Health Phys.

45,

339—348.

He83b Hess C. T., Weiffenbach C. V. and Norton

S. A., 1983, “Variations of airborne and waterborne

radon-222 in houses in Maine, U.S.A.,” accepted

for publication in

Environment International;

Paper

presented at the

Int. Symp. on Indoor Air Pollution,

Health and Energy Conservation,

13—16 October

1981, Amherst, MA.

Hi8l Hildingson 0., 1981, “Measurements of radon

daughters in 5,600 Swedish homes,” Paper pre-

sented at the

In!. Symp. on Indoor Air Pollution,

Health and Energy Conservation,

13—16 October

1981, Amherst, MA.

Ho83 Horton T. R., 1983, “Methods and results of

E.P.A.’s study of radon in drinking water,” Paper

presented at 15th

Annual National Conf on Radia-

tion Control, Program Directors,

16 May 1983,

Reno, NV. (Also available as EPS 520/5-83-027

from Office of Radiation Programs, U.S. EPA,

Washington, DC.)

Hu65 Hursh J. B., Morken D. A., Davis T. P. and

Lovaas A., 1965, “The fate of radon ingested by

man,”

Health Phj’s.

11, 465.

Hu73 Hursh J. B. and Spoor N. S., 1973,

Uranium,

Plutonium, Transpiutonium Elements

(Edited by

H. C. Hodge, J. S. Stannard, and J. V. Hursh)

(New York: Springer-Verlag).

ICRP79 International Commission on Radiological

Protection, 1979, “Limits for Intakes of Radionu-

RADIOACTIVITY IN PUBLIC WATER SUPPLIES

ii

r.1

---

clides by Workers,”

!CRP Publication

30.

Annals

of the ICRP

2(3/4) (Oxford: Pergamon Press).

Jo73 Johnson R. H., Bernhard D., Nelson N. S.

and Calley H. W. Jr., 1973, “Assessment of radio-

logical health effects from radon in natural gas,”

in:

Proc. ofNobel Gas Symposium,

U.S. EPA, 401

M St. SW, Washington, DC, EPA-600/9-76-026.

Ka73 Kahlos H. and Asikainen M., 1973,

Natural

Radioactivity of Ground Water in the Helsinki

Area,

Institute of Radiation Physics (Sateilyfysiikan

Laitos), Helsinki, Finland, Report SFL-A 19.

Ka80 Kahlos H. and Asikainen M., 1980, “Internal

radiation doses from radioactivity ofdrinking water

in Finland,”

Health Phys.

39, 108—Ill.

Ka76 Kaufmann R. F., Eadie 0. 0. and Rusell

C. R., 1976, “Effects of uranium mining and

milling on ground water in the Grants Mineral

Belt, New Mexico,”

Ground Water,

14(5), 16—25.

Ka77 Kaufmann R. F. and Bliss J. D., 1977,

Effects

of phosphate mineralization and the phosphate

industry on

Ra-226

in ground water of central

Florida,

U.S. EPA, Office of Radiation Radiation

Programs, Las Vegas, NV, Rep. EPA/520-6-77-

010.

Ki82 King P. T., Michel J. and Moore W. S., 1982,

“Groundwater geochemistry of Ra-228, Ra-226,

and Rn-222,”

Geochim. Cosmochim. Acta 46,

1173—1182.

Ko79 Kobal I., Kristan J., Ancik M., Jerancic S.

and Skofijanec M., 1979, “Radioactivity of thermal

and mineral springs in Slovenia,”

Health Phys.

37,

239—242.

Krie82 Krieg C. R. and Hahne R. M. A., 1982,

“Ra-226 and Ra-228 in Iowa drinking water,”

Health Phys.

43,

543—559.

Kris82 Krishnaswami S., Graunstein W. C., Tur-

ekian K. K. and Dowd J. F., 1982, “Radium,

thorium, and radioactive lead isotopes in ground-

• water: application to the

in situ

determination of

adsorption-desorption rate constants and retarda-

tion factors,”

J. Water. Res. Research,

18,

1633—

1675.

Ku54 Kuroda P. K., Damon P. E. and Hyde H. I.,

1954, “Radioactivity of the spring waters of Hot

Springs National Park and vicinity in Arkansas,”

Am. J. Science

252, 76—86.

Le79 Lee R. 0., Watson J. E. Jr. and Fong S. W.,

1979, “An assessment of radium in selected North

Carolina drinking water supplies,”

Health Phys.

37, 777—779.

Li02 Lively R. S. and Morey 0. B., 1902, “Hydro-

geochemical distribution of uranium and radon in

east-central Minnesota,” pp. 91—108, in:

Isotope

Studies of Hydrologic Process

(Edited by E. C.

585

Perry and C. W. Montgomery) (Dekalb, IL: North-

ern Illinois University Press).

Lu82 Lucas H. F. Jr., 1982, “Ra-226 and Ra-228

in drinking water,” Paper presented at 27th

Anmial

Mtg. Health Physics Society,

27 June—I July 1952,

Las Vegas, NV.

Ma85 Mays C. W., Rowland R. E. and Stehney

A. F., 1985, “Cancer risk from lifetime intake of

Ra and U isotopes,”

Health Phys.

48, 63 5—647.

Mc80 McGregor R. G., Vasudev P., Letourneau

E. 0., McCullough R. S., Prantl F. A. and Tarsi-

guchi H., 1980, “Background concentration of

radon and radon daughters in Canadian homes,”

Health Phys.

39, 285—289.

Mi80 Michel J. and Moore W. S., 1980, “Ra-228

and Ra-226 content of ground water in fall line

aquifers,”

Health Phys.

38, 663—671.

Mi82 Michel J. and Pollman C., 1982,

A Modelor

the occurrence of

Ra-228

in Ground Water,

U.S.

EPA, Office of Drinking Water, Washington, DC,

Environmental Science and Engineering Rep. No.

81-227-270.

Mi58 Miholic S., 1958, “Radioactive waters from

sediments,”

Geochem. et Cosmochem. Acta

14,

223—233.

Mo69 Moore W. S., 1969, “Oceanic concentrations

of radium-228,”

Earth Planet. Sci. Lea. 6,

437—

446.

Mo8l Moschandreas D. J. and Rector H. E., 1981,

“Indoor Radon Concentrations,” Paper presented

at

Int. Symp. on Indoor Air Pollution, Health and

Energy Conservation,

13—16 October 1981, Am-

• herst, MA. (Also available from the Illinois Institute

of Technology, Chicago, IL.)

Mo83 Moore W. 5., 1983, Personal Communica-

tion, Department of Geology, University of South

Carolina, Columbia, SC.

Mu82 Mullin A. 1982, “Abnonnally high alpha

activity in a Louisiana drinking water supply,”

Paper presented at 27th

Annual Meeting, Health

Physics Society,

27 June—I July, 1982, Las Vegas,

NV.

Ne81 Nero A. V. and NazaroffW. W., 1981, “Dis-

tribution ofindoor radon concentrations and source

magnitudes: measurements and policy implica-

tions,” Paper presented at

Int. Symp. on Indoor

Air Pollution, Health, and Conservation,

3—16 Oc-

tober 1981, Amherst, MA. (Also available from

Lawrence Berkeley Laboratory, University of Cal-

ifornia, Berkeley, CA 94720.)

0164 Olson J. C. and Overstreet W. C., 1964,

“Geologic distribution and resources of thorium,”

U. S. Geol. Survey Bull.

1204,

61.

Pa79 Partridge J. E., Horton T. R. and Sensintattar

E. L., 1979,

A Study of Radon-222 Re!easedfrom

C. T. HESS

et aL

J

---

RADIOACTIVITY IN

PUBLIC WATER SUPPLIES

Water During Typical Household Activities,

Eastern

Environmental Research Facility, U.S. EPA,

Montgomery, AL 36193, EPA Tech. Note ORP-

EERF-79-I.

Pr8l Prichard H. M., and Gesell

T. F., 1981, “An

estimate of population exposures due to radon in

public water supplies in the area of Houston,

Texas,”

Health Phys.

41,

599.

Pr83a Prichard H. M., Hess C. T., Nyberg P.,

Weifl’enbach C. V. and Gesell T. F., 1983, “Asso-

ciations between grab sample and integrated radon

measurements in dwellings in Maine and Texas,”

in:

Environment International,

Vol. 8, 155, 1—6,

83—87.

Pr83b Prichard H. M. “Radon-222 in municipal

water supplies in the central United States,”

Health

Phys.

45, 991—993.

Ra83 Rama and Moore W. S., 1983 “Mechanisms

of transport of U-Th series radioisotopes from

solids into ground water,”

Geochemica Cosmo

Chemica Acta.

48, 395—399.

Ró77 Rowland R~~J~uca.s

H. F. Jr. and Stehney

A. F.,J~212~

“High radium levels in the water

supplies of Illinois and Iowa,” in:

Proc. mt. Symp.

on Areas of High Natural Radioactivity,

pp.

65—

73 (Rio de Janeiro: Academia Brasileira de Cien-

cias).

Sa18 Satterly J. and McLennan J. C., 1918, “Ra-

dioactivity of the natural gases of Canada,”

Trans-

actions of Royal Society of Canada

12(Series 3),

153.

Sa8l

Sachs H. M., Hernandez T. L. and Rign

J. W., 1981, “Regional geology and radon vari-

ability in buildings,” Paper presented at

mt. Symp.

on Indoor Air Pollution, Health, and Energy Con-

servation,

13—16 October 1981, Amherst, MA.

Sa78 Sasser M. K. and Watson J. E. Jr., 1978, “An

evaluation of the radon concentration in North

Carolina ground water supplies,”

Health Phys.

34,

667.

Sc62 Scott R. C. and Barker F. B., 1962,

Data on

Uranium and Radium in Ground Water in the

United States, 1954—195

7, U.S. Department of the

Interior, U.S. GeologicalSurvey, Reston, VA 22092,

U.S. Geological Survey Paper 426.

Sc82 Scott A. G. 1982, “Remedial actions, active

and passive,” Paper presented at the

Health Physics

Society Summer School: Radon, A National Health

Problem?,

2 1—25 June 1982.

Sh80 Shannon S. and Livingston D., 1980, Los

Alamos National Laboratory, Los Alamos, NM

(Livingston: U.S. Department of Energy, Grand

Junction, CO).

Sn70 Snihs J. 0., 1970,

The Content of Some

Natural Radioactive Elements, Especially Radon-

222,

In some Potable Waters in Sweden,

National

Institute of Radiation Protection, Fack S-I 04 01,

Stockholm 60, Sweden, SSI 1970-029.

St79 Strain C. D. and Watson J. E., 1979, “An

evaluation of Ra-226 and Rn-222 concentrations

of ground and surface waters near a phosphate

mining and manufacturing facility,”

Health Phys.

37, 779—783.

St81 Stranden E., 1981, “Radon in houses using

stone magazines for heat accumulation,”

Health

Phys.

41,

29.

Ta80 Taniguchi H. and Vasudev P., 1980, “Radon

and radon daughters due to natural uranium

occurrences in a rural Ontario community,” in:

Natural Radiation Environment

III (Edited by

T. F. Gesell), p. 1623. (Available from U.S. De-

partment of Energy, National Technical Informa-

tion Center, P.O. Box 62, Oak Ridge, TN 37830.)

Ta64 Tanner A. B., 1964, “Radon migration in the

ground: a review,” in:

The Natural Radiation

Environment

(Edited by J. A. S. Adams and

W. M. Lowder), pp. 161—190 (Chicago, IL: Uni-

versity of Chicago Press).

Un77 United Nations Scientific Committee on the

Effects of Atomic Radiation, 1977,

Sources and

Effects ofIonizing Radiation,

Report to the General

Assembly (New York: UN).

Us79 U.S. Environmental Protection Agency, 1979,

“Indoor radiation exposure due to radium-226 in

Florida phosphate lands; radiation protection rec-

ommendations and request for comment,”

Federal

Register

44,

38664.

Us80 U.S. Environmental Protection Agency, 1980,

“Proposed cleanup standards for inactive uranium

processing sites; invitation for comment,”

Federal

Register

45,

27370.

Wr83 Wrenn M., Durbin

P. W., Nelson C., Rundo

J., Still E. T. and Willis D., 1983, “Metabolism of

ingested U and Ra,”

Health Phys.

48, 601—633.

586

I

---

Health Physics

Vol.

51,

No.

6 (December), pp. 715—721, 1986

Printed

in the

U.S.A.

0017—9078/86

53.00

+

.00

151

1986

Health

Physics

5oci~ly

Pergamon JolIrnals Ltd.

JACQUELINE MICHEL

Research Planning Institute, mc., 925 Gervais Street, Columbia, SC 29201

and

C.

RiCHARD COTHERN

Office of Drinking

Water

(WH-550),

U.S. Environmental Protection

Agency,

Washington, DC 20460

(Received

7

October 1985; accepted 6 June

1986)

Abstract—A conceptual model is presented to predict the relative probabilityofthe occurrence

of elevated

225Ra in public drinking water supplies (those which

serve more

than 25 people or

15

connections) using ground water sources.

The model is based on

an aquifer classification

scheme,

which

is

developed

from an understanding of

the geochemical

and

radiological behavior

of 228Ra and its parent, 232Th. Using this model, all

aquifer types are classified as

low, medium,

or high probability of having elevated 228Ra. Summaries of the available data are discussed to

show the actual concentrations found in each type of aquifer. As part of the initial application

ofthis approach to develop a nationwide occurrence profile for 228Ra, all counties in the United

States were classified and a map presented to show the distribution of the three classes. Nation-

wide, 71 of the counties were ranked as low, 18 were ranked as medium, and 11 were

ranked as high.

INTRODUCTION

IN 1976, the U.S. Environmental Protection

Agency (EPA) established the national interim

primary regulations for radioactivity in drinking

•

water. These regulations set maximum contam-

inants levels (MCLs) for gross a-particle activity

of

555

becquerels per cubic meter (Bq m3) or

15 picocuries per liter (pCi L~)and total radium

activity (226Ra plus 228Ra) of 185 Bq m3

(5

pCi

L’). The 226 Ra activity measurement was re-

quired only when the gross a-particle activity ex-

ceeded 185 Bq nf3

(5

pCi L~)and 228Ra mea-

surement was required only when the 226Ra ac-

tivity exceeded 111 •Bq m3 (3 pCi L~). This

screening protocol and the combination of Ra

isotopes was proposed because it was believed

that 226Ra would be the dominant isotope in

The thoughts and ideas expressed in this paper are

those of the authors and do not necessarily reflect the

policies ofthe U.S. Environmental Protection Agency.

public drinking water supplies and would avoid

the successively more expensive 228Ra analyses.

Very little was known about the geographic dis-

tribution of 228Ra activity in drinkingwater, even

after the initial monitoring program was com-

pleted. Use of the monitoring results to better

predict the occurrence of 228Ra was difficult be-

cause the monitoring protocol required by the

regulation produced a very biased database.

The monitoring program did identify some re-

gions which had high natural radioactivity and

these have been studiedin detail by several groups

(Ro77; Mi80; Ma81; Ki82; Kri82; C183; Gi84;

Lu85). The results of these studies and an EPA-

sponsored workshop on radioactivity in drinking

water were summarized in Hess

et a!.

(He85).

One recommendation ofthe workshop group was

that monitoring requirements for 228Ra should

be decoupled from 226Ra and separate guidelines

for 228Ra were needed. The coupling of these iso-

topes was not justified because of the differences

in the geochemical behavior of their parent iso-

715

Supplied by the British Library - “The world’s knowledge” www.bl.uk~

PREDICTING THE OCCURRENCE OF 228Ra

IN GROUND WATER*

Attachment A

---

716

topes. In fact, the monitoring scheme required

by the regulations could even be biased away

from areas ofelevated 228Ra where 226Ra hadbeen

selectively removed because of the mobility of

the parent U.

In the development of the Revised National

Drinking Water Regulations, the EPA needs

some way to estimate the occurrence of 228Ra

before formulation of crtieria for the regulations.

Without better occurrence data, it would not be

possible to calculate population risks and treat-

ment costs, or develop monitoring requirements.

However, the existing data on 228Ra were inad-

equate to even attempt any estimates based on

extrapolation of available measurements. That

analytical methods for measuring 228Ra in water

samples are difficult, time-consuming and costly

are facts which limit the effectiveness of a large-

scale monitoring program to actually measure

levels of 228Ra in drinking water. Also, the dis-

tribution of 228Ra has been shown to he log-nor-

mal (Ki82), with most values below the present

detection limit of 37 Bq m3 (I pCi L~).

The EPA is currently sampling about 1,200

randomly-selected ground water sources of

drinking water in the United States. Public

drinking water supplies are those that serve more

than 25 people or 15 connections. There are

about 48,000 such ground water supplies in the

United States. Therefore, the effectiveness ofthis

random program would be limited in predicting

different exposure levels because of the small

number of samples that are expected to occur

above the detection limit. Without a new ap-

proach to estimate the occurrence of 228Ra, hu-

man health and economic effects of revision of

the drinking water regulation would be difficult

to evaluate.

In this paper, an approach is presented which

uses aquifer typeand water-quality characteristics

to classify specific areas according to the relative

probability of havinghigh 228Ra in ground water.

This concept is based on the fact that 228Ra is not

randomly distributed and thus cannot be pre-

dicted using statistical analysis ofnational survey

data alone. There will be too few “hits” of the

high values to be of use in identifying, to any

degree of certainty, the true distribution of228Ra

or detecting systems with high 223Ra using ran-

dom sampling techniques, Instead, the United

States can be divided into smaller units which

are classified on a relative scale oflow, medium,

and high probability of including elevated 228Ra

concentrations. The classifications are produced

from a conceptual model of the most important

factors affecting 228Ra distribution in ground

water.

OCCURRENCE OF

2~°RaBY AQUIFER

TYPE

AND WATER QUALITY

The correlation of high 228Ra concentrations

in ground water with certain types of aquifers has

been observed by many researchers, beginning

with Krause

(Kra59)

who reported high 226P~and

228Ra in the ground water from deep sandstone

formations in midwestern states. Even then,

Krause concluded that the 226Ra/228Ra activity

ratio was not constant, and in order to calculate

the total radiation dose, the 228Ra content of the

water would have to be determined separately.

After the interim standards were set, detailed

studies of of 226Ra and 228Ra were conducted in

Iowa (Kri83), South Carolina (Mi80; Ki82), and

Georgia (Cl83). Also, the monitoring results pro-

vided a new source for 228.Ra values.

In 1982—83, the EPA sponsored two studies to

determine the feasibility of usingaquifer type and

water-quality characteristics to predict the oc-

currence of225Ra in ground water (Mi82; Mi83).

In these studies, an aquifer classification scheme

was developed and tested on three major hydro-

geological provinces, namely the Atlantic Coastal

Plain, the Piedmont, and part of the glaciated

Central Platform. The Atlantic Coastal Plain and

Piedmont are two adjacent provinces which ex-

tend from New York to Alabama and east of the

Appalachian Mountains to the coast. The parts

of the glaciated Central Platform studied included

the states of Wisconsin, Iowa, Illinois, and Mis-

souri. These regions were selected for develop-

ment of the predictive model because they had

the largest existing database of 228Ra measure-

ments, which indicated that there were areas of

both very high and very low 228Ra concentrations

in these regions.

The aquifer classification model was developed

from an understanding of the geochemical and

radiological properties of 228Ra and its parent

232Th within an overall framework of the geolog-

ical processes which affect the composition of

aquifers. Discussions of the model are given else-

where (Mi82; Mi83; He85). The results of these

~~~Supplied by the British Library - “The world’s knowledge” www.bl.uk~

PREDICTING THE OCCURRENCE OF 228Ra IN GROUND WATER

---

J. MICHEL and C. R. COTHERN

717

studies showed that therewere distinct differences

in the relative distribution of 228Ra by aquifer

type and water quality. Table 1 shows the sum-

mary statistics for the specific aquifers studied in

detail during earlier studies. These results were

used to develop the following aquifer classifica-

tion model for predicting the 228Ra content of

ground water.

Aquifers which always had low activities of

228Ra included the following:

(1)

Carbonate aquifers

(limestone and dolo-

mite) had a log mean 228Ra of 11 Bq m3 (0.3

pCi L’) for 120 samples. These rock types have

very low 232Th content because they precipitate

from water, and Th has very low solubility in

water.

(2)

Metamorphic rock aquifers

also had a log

mean of 11 Bq m3 (0.3 pCi L~)for 75 samples,

Metamorphism of all intensities and rock types

tend to recrystallize any loosely-bound Th into

resistant minerals which minimizes the ability of

daughter 228Ra to go into solution. Also, ground

water flow in these rock types is generally through

fractures instead of through pore spaces, which

decreases the rock surface area in contact with

ground water.

(3)

Quartzose sand and sandstone aquifers

generally producedground water with low 228Ra,

particularly if the total dissolved solids (TDS)

were below 1,000 milligrams per liter (mg L~).

The high quartz content of the aquifer material

indicates that the sediments were intensely

weathered or that they were eroded from a source

which initially had low Th content. In either case,

the only Tb contained in the aquifer material is

likely to be in heavy minerals which do not

readily

give

up 228Ra into solution. There are two

special subclasses of this aquifer type. Sand and

gravel aquifers of glacial origin are generally low

in 228Ra log mean of 22 Bq m3 (0.6 pCi L~)

for 17 samples regardless of their composition

because of the large size of the clasts and the low

ratio of surface area to water. Even when these

aquifers are poorly sorted (i.e., contain various

amounts of clay, sand, and gravel), the clay tends

Table

I.

Summary statistics for

228Ra

content ofground wet erfrom specific aquifer types

Aquifer

Type

95

Upper

Confidence

Limits

Igneous Rocks

Bq

pCi L1

(Granites)

Chemical Precipitates

— Coastal Plain Aquifers

- Midwest Aquifers

Metamorphic Rocks

- Piedmont Region

Sand Aquifers

— Coastal Plain Fall Line

Arkoses

—

Coastal Plain

—

Qua rtrose

— Midwest Glacial/Alluvial

Sandstone Aquifers

- Midwest Cambrian -

Ordivician Aquifer

t2

52

1.14

590

16.0

16

14

0.1

7

0.2

1014

22

0.6

107

2.9

75

11

0.3

92

2.5

89

81

2.2

630

17.0

53

11

0.3

89

2.14

135

26

0.7

160

14.2

320

78

2.1

‘l~40

12.0

Supplied by the British Library - “The world’s knowledge” www.bl.uk

N

x

Bq m~3 pCi L1

---

718

to rapidly adsorb radium. Sand and gravel aqui-

fers of alluvial origins (such as those found in

river valleys) also were found to have iow 228Ra,

with a log mean of 26 Bq m3 (0.7 pCi L’) for

118 samples. These aquifer types usually contain

quartz-rich sediments and produce large quan-

tities of water.

(4) Basic, igneous rock aqujfers

(such as basalt)

are classified as having very low 228Ra although

there are few data from this type. Basalts have

low Th content but, more importantly, this rock

type rapidly sorbs 228Ra from solution. Basalts

are so efficient at sorption of metals from water

that they are considered as one of the candidate

rock types for long-term disposal of high-level

radioactive wastes.

There are three main aquifer types which are

predicted to have a high probability ofproducing

ground water with elevated 228Ra.

(I)

Granite rock

aquifers

which had not un-

dergone any metamorphism had a log-mean

228Ra concentration of 52

Bq m3 (1.4 pCi L~)

and

a range up to

850 Bq m3 (23 pCi L~)for

42 samples. Granites are composed of large

amounts of coarse-grained feldspar, a mineral

which contains thorium, not always in resistant,

crystalline forms but often loosely bound at sites

bet~veenindividual grains. Radium-228 produced

from this type of Th distribution is more likely

to go into and stay in solution. Igneous rocks of

similar composition

but fine-grained, such as

light-colored volcanic rocks, would have low

228Ra because of

rapid

resorption of

228Ra

from

solution.

(2)

Arkosic sand and sandstone aquifers

had

medium to high 228Ra, with a log mean of 90 Bq

m3 (2.4 pCi

L’) for 89 samples. Arkoses, by

definition, contain a significant amount of feld-

spathic minerals, which can have high Th. Ifthe

arkosic sediments were derived from granitic

rocks, they can have very high Th. If the sedi-

ments have undergone any weathering which has

broken

down

the feldspar, the 228Ra content of

the ground

water can

be very high. There is a

gradient of the 228Ra content of these aquifers

depending upon the amount and composition of

feldspathic minerals and rock

fragments in

the

sediments of the aquifer.

(3)

Quartzose sandstone aquifers with high

TDS

were found to have a wide range of 228Ra,

including very high concentrations up to 1180

Bq m3 (32 pCi L~).All data for this type of

aquifer came from groundwater supplies in Iowa,

Illinois, and Missouri, which had a mean 228Ra

concentration of 74 Bq m3 (2 pCi L~)andusu-

ally had high 226Ra as well. It was concluded that,

at TDS of greater than 1,000 mg L’ in clean,

quartz sandstones, competition forthe few avail-

able sorption sites on the aquifer solids were the

controlling factor keeping 228Ra in solution. This

aquifer type may

be one of the most overlooked

sources of elevated 228Ra, especially in areas

where U has been removed by ground water

leaching. With U removed, the gross a-particle

activity screen would be low and neither 22~Ra

nor 228Ra would be measured.

The purpose of this paper is to demonstrate a

concept of an aquifer classification scheme to

~nap areas nationwide on a relative scale oflow,

medium and high probability of having elevated

228Ra concentrations. It should be noted that

areas ranked high are not predicted to have only

elevated 228Ra in the ground water. All areas will

have a large percentage of the community water

supplies with very low 228Ra in the ground water.

Instead, these areas ranked high are more likely

to have elevated 228Ra

occur more frequently.

The mapping unit selected was by county, the

smallest unit for which data on population and

number of systems could be compiled. The EPA

is currently conducting a field-monitoring pro-

gram in which 1,200 community drinking water

supplies are being sampled for various parame-

ters, including 225Ra. The results from this survey

will provide the data to develop an occurrence

profile for each class.

METHODS OF STUDY

To delineate aquifer

types

nationwide using

the proposed classification scheme for 3,076

counties, many various sources of information

were used. State and federal agencies responsible

for geological, mineral, andwater resources were

contacted for pertinent literature and geological

and hydrological maps. One of the most difficult

tasks was identification of which geological units

were used as aquifers.

After all the available data were reviewed, the

main aquifers (not just rock types) foreach state

and county were identified, described, and clas-

..~.-—

Supplied by the British Library - “The world’s knowledge” www.bl.uk

_____

PREDICTING .THE OCCURRENCE OF 22~RaIN GROUND WATER

n

---

J. MICHEL and C. R. COTHERN

719

I

sified as having a low, medium, or high proba-

bility ofhaving elevated 228Ra concentrations. For

counties with multiple aquifer types, either areally

or with depth, professional judgment was used

in the final classification. For multiple aquifers

with depth, the county was classified according

to the aquifer ofthe highest classification. In some

cases of

areally

different aquifers within a county,

where information was available on population

distribution from detailed maps, the dominant

aquifer type most probably used for public water

supplies determined the county classification.

Aquifers classified as low probability included

those composed of limestone, metamorphic rocks

(including high-grade metamorphosed granites),

glacial and alluvial sand and gravel, volcanic

rocks, and quartzose sand and sandstone of low-

to-moderate TDS. Aquifers classified as medium

probability included those composed of semiar-

kosic sand and sandstone, low-grade metamor-

phic granites, quartzose sand and sandstone of

moderate TDS, alluvial sand and gravels specif-

ically described as containing feldspathic sedi-

ments, and a relatively small set ofaquifers which

could not be definitely classified as low or high.

Aquifers classified as high probability included

those composed of granite, arkosic sand and

sandstone,

alluvial deposits located immediately

adjacent to granitic rock sources, and quartzose

sandstones with high TDS, particularly in those

areas where high 228Ra has been reported.

RFSULTS AND DISCUSSION

Figure 1 shows the areal distribution of the

three different classes forthe counties in the con-

tinental United States. All other states and ter-

ritories were ranked as having low probabilities.

Nationwide, 7! of the counties were ranked as

low, 18 were ranked as medium, and 11 were

ranked as high. It should be noted that in review-

ing these data, that counties varied in size (by

area)

by over three orders of magnitude and in

population by atleast a factor of 3,000, Counties

are by no means uniform units for direct com-

parison of population at risk or number of sys-

tems impacted. The EPA is currently compiling

HIGH

MEDIUM

LOW

FiG. 1. Distribution of areas of rclative probability of having elevated 283Ra in community

ground water supplies.

Supplied by the British Library - “The world’s knowledge” www.bl.uk~

---

720

this information by county so that the actual

population and number of systems can be deter-

mined at the county level. It should be pointed.

out that the population exposed in any one

county may vary from nearly all to only a small

portion of the total.

Even with a brief review of Fig. 1, some inter-

esting observations can be made. There are large

areas where 228Ra is expected to be very low, with

most values below 37 Bq m3 (1.0 pCi L~)and

probably below 18 Bq m3 (0.5 pCi L’). These

areas include the quartzose sand and limestone

aquifers of the lower Coastal Plain along the At-

lantic andGulf coasts. The U mineralization belt

in Texas is ranked low because much of the U is

leached from volcanic rocks which do not release

Th. The very old carbonate and quartz-rich

sandstone aquifers underlying the general area

between the Appalachians and the High Plains

were

ranked low except for localized areas, such

as the Oachita Mountains in Arkansas and the

St. Francois Mountains in Missouri. Much of the

area ranked low in the Northeast is composed of

metamorphic rock terrains of low ground water

production, overlain by glacial sand and gravel

aquifers, from which most of the ground water

is obtained. In the West and Northwest, low-

ranked areas were predominantly basalt, alluvial

sand and gravel, and clean sandstones.

Areas which were ranked as high included a

band from Pennsylvania to Georgia which rep-

resented mostly the Fall Line sand aquifers which

are locatedat the contactbetween the rock terrain

of the Piedmont province and the sediments of

the Coastal Plain. These aquifers have been

shown to have high 228Ra and 226Ra (Mi80). This

zone also included the granitic rock aquifers in

the adjacent Piedmont which are used by small

suppliers.

Large areas of northern Illinois, Iowa, Min-

nesota, and Wisconsin were ranked as high, be-

cause of granitic terrain or the presence of sand-

stone aquifers with high TDS. Counties ranked

low or medium in this area were those where

ground water was obtained mostly from alluvial

or glacial sand and gravel aquifers or from car-

bonate aquifers. It should be noted that thisarea

is mapped in great detail because it was one of

the detailed study sites used in developing the

aquifer type model.

Parts of Colorado and Montana were ranked

high where

the

granites ofthe Rocky Mountains

might influence the 228Ra content ofthe aquifers.

However, the ground water use in this area should

be small, and therefore, the relatively large area

ranked as high may not represent many users. In

Idaho, the high areas are those underlain by the

extensive Idaho granite. The granite functions as

a poor aquifer, but it could have an impact on

radiological water quality in that the granite is a

significant source of the sediments that fill the

alluvial valleys which are important aquifers.

Much ofCalifornia is underlain by granitic ig-

neous rocks which were ranked high, although

community ground water use in these areas is

probably small. The complicated hydrogeology

of California, which is composed of 394 ground

water basins with no widespread aquifer systems,

made classification of the region difficult andthe

results very weak. The rest ofthe areas ranked as

high risk were scattered occurrences where gra-

nitic intrusions or sediments derived from them

constituted local aquifers.

There were several large areas designated as

having medium probability, the largest ofwhich

was the High Plains area, underlain by a shallow

sand aquifer known regionally as the Ogallala.

This aquifer supplies much of the ground water

forirrigation of farms from Texas to South Dak-

ota. The sediments of this aquifer are thought to

have been eroded from the granite ofthe Rocky

Mountains and deposited by streams draining

east. The sediments can contain significant

amounts of feldspathic minerals locally and the

entire aquifer is ranked as medium. There are

very few 228Ra measurements from this aquifer

to indicate what the actual concentrations are.

The remaining large area of medium proba-

bility (parts of Idaho, Montana, Utah, and Ari-

zona) are counties where granitic rocks were

thought to have contributed some but not sig-

nificantly to the sediments of the alluvial valley

and sandstone aquifers.

No attempt was made to quantify the relative

rankings of low, medium, and high. With the ex-

isting 228Ra data, which are nonrandom and of

variable quality, and for which the aquifer type

is frequently unknown, it is not possible to cal-

culate occurrence statistics with any degree of

certainty. It is hoped that the current EPA ground

Supplied by the British Library - “The world’s knowledge” www.bl.uk

PREDICTING THE OCCURRENCE OF 228Ra IN GROUND WATER

I

---

3.

MICHEL and C. R. COTHERN

721

water survey will provide some indication of the

actual distribution and confirmation of’ the rel-

ative rankings. Ifso, this approach can be refined

if necessary and used in better determining the

population risk, the costs of compliance, and

monitoring guidelines. The map in Fig. I could

be used to identify specific areas requiring differ-

ent monitoring strategies.

REFERENCES

Cl83 Cline W., Adamovitz S., Blackman C. and

Kahn B., 983, “Radium and

uranium concentra-

tions

in Georgia community water supplies,”

Ifealth

Phys.

43, 1—12.

Gi84 Gilkeson R. H., Perry E. C. Jr., Cowart 3. B.

and Holtzman R. B., 1984,

Isotopic Siudies ofthe

Natural Sources ofRadium in Groundwater in liii-

~iois,

University of Illinois at Urbana-Champaign,

Water Resourccs Center, Urbana, IL, Research Rept.

187.

He85 Hess C. T., Michel J., Horton T. R., Prichard

1-I.

M. and Coniglio W. A., 1985, “The occurrence

ofradioactivity in public water supplies in the United

States,”

Health Phys.

48,

553—586.

Ki82 King P. T., Michel 3. and Moore

\V.

S., 1982,

“Groundwater geochemistry of 228Ra, 226Ra, and

222Rn,”

Geochim. Cosmochim. Ada

46, 1173—1182.

Kra59 Krause D. P.,

1959,

228Ra

(Mesothoriuin I)

in Illinois Well Waters,

Argonne National Labora-

tory, Radiol. Phys. Div., Argonne, IL, Semiannual

Rept. ANL-6099, pp.

5

1—52.

KrieS2 Kriege C. R. and Hahne R. M. A., 1982,

“226Ra and 228Ra in Iowa drinking

water,”

Health

Phys.

43,

543—559.

Lu85 Lucas H. F., 1985, “Radium-226 and 223Ra in

water supplies,”

J. Am. Water WorksAssoc. 77, 57—

67.

Ma81

Markos M., 1981, “Distribution and possible

causes of radionuclides in groundwaters of Mis-

souri,” in:

Proc. 15th Annual Conference on Trace

Substances in Environmental Health

(edited by

D. D. Hemphill), pp. 69—77, Environmental Trace

Substances Research Center and Extension Division,

University of Missouri, Columbia, MO.

Mi80 Michel J. and Moore W. S., 1980, “228Ra and

226Ra content of ground water in fall line aquifers,”

Health Phys.

38, 663—671.

Mi82 Michel J. and Pollman C., 1982,

A mode/for

the occurrence of

225Ra

in Ground Water,

U.S. En-

vironmental Protection Agency, Office of Drinking

Water, Washington, DC, Environmental Science

and Engineering Rept. No. 8 1-227-270.

Mi83 Michel

3.

and Poliman C., 1983,

A mode/for

the occurrence

of228Ra

in ground water: application

to the north-central United Stales,

U.S. EPA, Office

of Drinking Water, Washington, DC, Environmen-

tal Science and Engineering Rept. No. 8 1-227-280.

Ro77 Rowland R. E., Lucas H. F. Jr. and Stehney

A. F., 1977, “High radium levels in the water sup-

plies of Illinois and Iowa,” in:

Proc. mt. Synip. on

Areas ofHigh Natural Radioactivity,

pp.

65—73

(Rio

de Janeiro, Brazil: Academia Brasileira de Ciencias).

Supplied by the British Library - “The world’s knowledge” www.bl.uk