..

 

 

en: . ‘ .. ..Ir..,:...«§ww..€5_ Am, ..,.,r, ‘ , “.fiua._:n.fi»~...§_n.w

5a. ...$...S§fi

l

 

THEF-
THESIS

SITY LIBRARIES

lllllllllllllllllllllllllllllllllllll llli I ll

3 1293 01564 4580

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

A GEOGRAPHY OF HUMAN RISK TO DOMESTIC RADON EXPOSURE
IN LENAWEE COUNTY, MICHIGAN

presented by

Tarek Antony Joseph

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Geography

 

fiWVWW

Major profesg
Chairperson of Geography Dept.

Date /Zle/9é

MS U i: an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

LEEHARY
Mlchlgan State
University

 

 

 

PLACE IN RETURN BOX to romovo this chockout trom your record.
TO AVOID FINES return on or bdoro doto duo.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Io An Afflrmdlvo ActIonIEqud Opportunlty InctIMIon
Waco-ct

A GEOGRAPHY OF HUMAN RISK TO DOMESTIC RADON EXPOSURE
IN LENAWEE COUNTY, MICHIGAN

By

Tarek Antony Joseph

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Geography

1996

ABSTRACT

A GEOGRAPHY OF HUMAN RISK TO DOMESTIC RADON EXPOSURE
IN LENAWEE COUNTY, MICHIGAN

By

Tarek Antony Joseph

This study assesses overall human risk posed by domestic radon exposure in
Lenawee County, Michigan. Many of the maps produced to show radon risk rely only on
geology (specifically bedrock). This study utilizes a holistic approach which emphasizes
the relationships between humans and the environment, and views disease as multi-causal
and multi-factored. While it is beyond the scope of this study to examine disease
outcomes specifically, the goal is rather to assess those broad factors of risk that may
produce disease outcomes.

Data were collected on the three primary components of this study: geology,
housing structural characteristics, and human behavior. Geologic strata include not only
bedrock, but surficial features and soil associations as well. Of interest is the degree to
which these strata serve as filtering layers for the movement of radon. The data were
collected during the winter heating season of 1994 when 230 households were surveyed
utilizing a stratified random sample. In addition to administering a questionnaire and
making a visual inspection of each house, a radon measurement was also taken.

Results fi'om this study indicate a higher-than-average radon risk for residents of
Lenawee County. Great variations exist in the spatial patterns of risk to radon in Lenawee

County. Areas of potentially higher risk, labeled “hot spots” are located and identified.

Forty-one percent of the 230 houses surveyed in this study have measurement results that
meet or exceed the EPA action level of 4.0 pCi/L. This percentage of “at risk” housing is
relatively high when compared with state wide survey results which reveal that 12% of the
2082 homes measured have results that meet or exceed the EPA action level of 4.0 pCi/L.
Overall summary statistics from this study reveal a median of 3.4 pCi/L, a geometric mean
measurement value of 3.3 pCi/L, and a range of 0.3 - 48.4 pCi/L.

This study also empirically tests working assumptions of geologic risk potentials
that were initially developed. Disaggregation and reclassification of the bedrock strata
occurred based on field findings. In addition to geologic strata, various housing structural
characteristics and intervening human variables are identified and analyzed to better

characterize overall human risk to radon in Lenawee County.

Copyright by
Tarek Antony Joseph
1996

DEDICATION

In Memory Of
ROBERT MICHAEL GEORGE

I963 - 1993

And To My Parents

RICHARD and SHIRLEY JOSEPH

ACKNOWLEDGMENTS

I wish to recognize and express my gratitude to a number of individuals and
institutions who have contributed to the successful completion of this dissertation. First, I
wish to thank and convey my deepest appreciation to my Advisor and Committee Chair,
DrJohn M Hunter, for his guidance, patience, and support. Dr. Hunter’s genuine
enthusiasm, insight and commitment have been great sources of encouragement to me
throughout my years at Michigan State. It has been a great honor and privilege to work
with him. My sincere thanks also to the members of my Guidance Committee: Dr. Daniel
Brown, Dr. Richard Groop, Dr. Daniel Jacobson, and Dr. Gary Manson, for their
numerous contributions, high degree of accessibility, and excellent support.

This research would not have been possible with out the assistance of Susan K.
Hendershott, Indoor Radon Specialist, Michigan Department of Public Health, Division of
Radiological Health. From very early in the planning stages of this project, Ms.
Hendershott took a great interest in helping this study develop and I am most thankful to
her. My profound thanks go to the Michigan Department of Public Health, Division of
Radiological Health for the provision of the radon measurement devices necessary for
home testing, and Environmental Protection Agency literature for distribution to the

members the homes surveyed.

co

I am sincerely gratefiJl to the Lenawee County Health Department, Environmental
Health Division, and its entire staff. Larry Stephens, Health Officer, and Mike Kight,
Environmental Health Director, took personal interest in this study and were most
gracious and willing to make provisions necessary for this research to be conducted in
Lenawee County. I wish to acknowledge Sanitarians Maria Leary and Paul Nelson who
were particularly helpful. My thanks also to my field assistant Greg Braun. Having been
raised in Lenawee County, Greg possessed great knowledge of the county and residents
that he was eager to share during and after my time in the field.

To the citizens of Lenawee County, I offer great thanks. During the course of my
fieldwork, I truly enjoyed meeting and interacting with the friendly people I encountered.
Without the willingness of residents to participate, this study could not have been
completed.

In addition to my Guidance Committee, there are additional individuals at
Michigan State University whom I wish to thank: Dr. David Campbell for his advice and
encouragement. I was fortunate to work with Dr. Campbell for a number of semesters
both as a Research Assistant and Teaching Assistant and I am grateful to him for those
opportunities; Dr. Bruce Pigozzi, who gave generously of his time and advice on statistical
questions; Dr. Judy Olson, former Department Chair, for encouragement and support in
the form of Graduate Office Fellowships; Mike Lipsey for various computer-related help;
Marilyn Bria, Sharon Ruggles, Judy Slate, and Towonna Williams for their continual
assistance; and Bill Enslin and the Center for Remote Sensing for use of computing
facilities. My sincere thanks also go to Mike Conley, Nancy Yattaw, and Jim Wilfong for
their help in the MSU Center for Remote Sensing.

iii

These thanks would be incomplete without the mention of my former professor,
Dr. Daniel Weiner, West Virginia University, and formerly of The University of Toledo. 1
have him to thank for inspiring me and introducing me to the discipline of geography.

I have been most fortunate to keep and make some very good friends over the
course of the last four years at MSU, and I am thankful to them all. Fellow graduate
students, colleagues, and friends hax'e provided opportunity for intellectual debate and
comic relief: Mark Guizlo, Mark D. Olson, ‘Mark Pires, Linda Beck, Charlie Rader, Nancy
Rader, Renato Cerdena, Johan Liebens, Jenny Olson, Susan George, Fred Shaheen, Doug
Thabit, Cathleen McAnney, Mike Emch, Sheryl Welte, Morris Thomas, Alison Philpotts,
and Jay Samek.

While I have gained many friends here at MSU, I have also lost a very good one
during this time. Bob George, my life long friend, passed away suddenly in late 1993
shortly before I began my field work. This dissertation is dedicated to his memory.

Finally, as inadequate as it may sound, I offer my heartfelt thanks and appreciation
to my parents, Richard and Shirley Joseph, who have seen me through the completion of
this degree. They have been a continual source of encouragement and support for me not
only during my MSU years, but throughout my entire life. This dissertation is dedicated to

them as well.

iv

TABLE OF CONTENTS

LIST OF TABLES ..................................................................................... vii
LIST OF FIGURES .................................................................................... ix
1. INTRODUCTION .................................................................................. 1
2, LITERATURE REVIEW ....................................................................... 4
Radon Overview ............................................................................. 4
The Geology of Radon .................................................................... 8

Radon and Housing ......................................................................... 10

Lung Cancer, Disease, and Health Effects of Radon ........................ 12

3. STUDY OBJECTIVES AND RESEARCH QUESTIONS ...................... 18

Study Objectives ............................................................................. 18

Research Questions ......................................................................... 20

4. METHODOLOGY ................................................................................. 23

Host-Agent -Environment Interaction Model ................................... 23

Radon In Michigan .......................................................................... 25

Radon in Lenawee County .............................................................. 26

Description of Study Area ............................................................... 26

Contacts .......................................................................................... 28

Seasonal Variation of Radon ........................................................... 30

EPA Radon Action Level ..................................................... 30

Spatial Design of Field Sample ........................................................ 31

Sample Size ......................................................................... 31

Sample Design and Selection Proceedures ........................... 32

Overview ................................................................. 32

Sample Design ......................................................... 33

‘ Geologic Risk Classification System ......................... 35

Household Selection, Contact and Visitation ................................... 44

Questionnaire ...................................................................... 49

Environmental Protection Agency Training ..................................... 49

Standard Field Methods and Procedures .......................................... 50

5. SURVEY RESULTS BY LOCAL GOVERNEMNT

ADMINISTRATIVE DISTRICTS .................................................. 51
Presentation of Survey Results ........................................................ 51
Overall County Summary Statistics ...................................... 51
Results By Township ........................................................... 54
Research Question Addressed ......................................................... 59
6. SURVEY RESULTS BY GEOLOGY AND SOILS ............................... 61
Results By Geologic Risk Potential ................................................. 61
Results by Assumed Bedrock Risk Potential ........................ 61
Results by Assumed Surficial Features Risk Potential ........... 62
Results by Assumed Soil Association Risk Potential ............. 63
Results by Overall Geologic Risk Potential Codes ................ 64
Results by Bedrock F orrnation Afier Dissagregation ............ 66
Research Questions Addressed ............................................ 71
7. SURVEY RESULTS BY HOUSING CHARACTERISTICS ................. 82
Results By Housing Structural Characteristics ................................. 82
Interaction Between Bedrock Risk Groups and Housing Features....90
Research Question Addressed ......................................................... 94
8. HUMAN BEHAVIORS THAT AUGMENT CANCER RISKS .............. 102
Intervening Human Behavior Variables ............................................ 102
Research Question Addressed ......................................................... 106
9. SUMMARY AND CONCLUSIONS ....... . .............................................. 109
Summary and Conclusions ............................................................... 109
Host-Agent-Environment Interaction Model Revisited ......... 113
Principal Components Analysis ............................................ 115
Recommendations ........................................................................... 1 16
BIBLIOGRAPHY ...................................................................................... 1 18
APPENDIX ................................................................................................ 123
A. Lenawee County Residential Radon Survey Contact Letter ......... 123
B. Lenawee County Residential Radon Survey Information Sheet... 124
C. Lenawee County Residential Radon Survey Questionnaire .......... 126
D. EPA RMP Training Certificate ................................................... 128
E. Lenawee County Residential Radon Survey Result
Follow Up Letter ................................................................ 129
F. Radon Reduction Methods .......................................................... 130

Table 4.1

Table 4.2

Table 5.1

Table 5.2

Table 5.3

Table 6.1

Table 6.2

Table 6.3

Table 6.4

Table 6.5

Table 6.6

Table 6.7

Table 6.8

Table 6.9

LIST OF TABLES

Allocation of Sample Size by Township ............................... 34
Key to Geologic Radon Risk Codes ..................................... 37
Overall Summary Statistics .................................................. 51

Results Comparison Between Lenawee County
Residential Radon Survey and the Michigan

Indoor Radon Survey ........................................................... 54
Results by Township ............................................................ 55
Results by Assumed Bedrock Risk Potential ........................ 62
Results by Assumed Surficial Features Risk Potential ........... 63
Results by Assumed Soil Association Risk Potential ............. 64
Results by Overall Geologic Risk Potential Codes ................ 65

Original Assumptions of Bedrock Risk Potential
as Against Measured Results... ............................................. 68

Measured Results by Bedrock Formation
Ranked by Mean .................................................................. 69

Empirically Derived Bedrock Risk Categories Based
on Measurement Results ...................................................... 72

Frequencies -- Bedrock Risk Potential
by Measurement Result ....................................................... 77

Frequencies -- Surficial Feature Risk Potential
by Measurement Result ........................................................ 79

vii

Table 6.10

Table 7.1

Table 7.2

Table 7.3

Table 7.4

Table 7.5

Table 7.6

Table 7.7

Table 7.8

Table 7.9

Frequencies -- Soil Association Risk Potential

by Measurement Result ........................................................ 81
Results by Location of Measurement Device ........................ 82
Results by Age of Housing Structure ................................... 83

Results by Age of Housing Structure and Location
of Measurement Device ....................................................... 84

Results by Board of Review Housing

Table 7.10

Table 7.1 1

Table 7.12

Table 7.13

Table 8.1

Table 8.2

Table 8.3

Table 8.4

Table 8.5

Table 8.6

Assessment Values .............................................................. 85
Results by Presence of Sump Pump ..................................... 86
Results by Floor Material ..................................................... 87
Results by Presence of Floor Cracks ...................... . ............. 88
Results by Wall Material ...................................................... 89
Results by Presence of Wall Cracks ..................................... 90
Bedrock Risk Groups and Presence of Basement ................. 91
Bedrock Risk Groups and Presence of Sump Pump ............. 93
Bedrock Risk Groups and Presence of Floor Cracks ............ 95
Frequencies -- Presence Of Sump Pump

By Measurement Result ....................................................... 99
Number of Households with Smokers Present ...................... 102
Results by Smoking Status ................................................... 103
Number of Households with Basement Sleeping .................. 103
Results by Use of Basement for Sleeping ............................. 104
Number of Households Where Basement

Television Viewing Occurs .................................................. 105
Results by Use of Basement for Television Viewing ............. 105

viii

Figure 2.1
Figure 2.2
Figure 2.3
Figure 4.1
Figure 4.2

Figure 4.3

Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8

Figure 5.1

Figure 5.1-A

Figure 5.2
Figure 5.3

Figure 6.1

Figure 9.1

LIST OF FIGURES

Major Radon Entry Routes .................................................. 5
Uranium Decay Chart .......................................................... 7
Radon Risk Estimates .......................................................... 14
Host—Agent-Environment Interaction Model ........................ 24
Regional Screening Measurement Projection ........................ 27

Lenawee County Census Tracts and

Minor Civil Divisions ........................................................... 29
Bedrock Risk Potential Classifications ................................. 38
Surficial Formation Risk Potential Classifications ................. 41
Soil Association Risk Potential Classifications ...................... 42
Overall Geologic Radon Risk Potential Codes ...................... 45
Overall Geologic Radon Risk Potential Codes with

Township Boundaries ..................................... . ..................... 46
Overall Measurement Results Histogram .............................. 52

Overall Measurement Results Histogram

Log Linear Transformation .................................................. 53
Measurement Results by Township ...................................... 57
Distribution of Houses At or Exceeding

EPA- Action Level by Township ........................................... 58
Disaggregated Bedrock Formations with

Measurement Results ........................................................... 70
Host-Agent-Environment Interaction Model Revisited ......... 1 14

ix

CHAPTER 1: INTRODUCTION

Radon is a colorless, tasteless, and odorless radioactive gas produced during the
natural decay of uranium. The presence of radon in the indoor environment is a growing
threat to human health. Numerous medical and environmental studies have linked radon
with human health risks, primarily in the form of lung cancer (Radford, 1985; Klotz et a1,
1989; Svensson et a1, 1989). The United States Environmental Protection Agency (EPA)
estimates that 8,000 to 20,000 lung cancer deaths per year can be attributed to indoor
radon exposure in the United States.

For over fifty years radon has been implicated as a cause of lung cancer in miners,
however, research on human exposure to radon in a domestic setting is relatively new.
Research of the past decade has produced findings which have led radon to be considered
as a potentially serious public health concern. There is much that needs to be learned
about radon as an indoor risk to human health.

A growing amount of research is being conducted on the phenomenon by a
number of disciplines, however there is a void in the geographic literature concerning
radon. Following the disease ecology tradition of medical geography, this research project
is designed as a comprehensive study that takes into account both environmental and

human factors to establish a geography of overall radon risk. In this tradition, human

2
health is considered within the framework of human environmental interaction. Disease is

seen as an outcome of complex variables and multiple factors converging in space and
time. Therefore medical geography “seeks to understand as holistically as possible the
correlates of health and disease in order to advance knowledge of disease etiology and to
promote adaptable population/behavior/environment interactions” (Earickson et a1, 1989:
425)

Radon risk potential has been estimated and mapped on a national scale in the
United States by the EPA. These maps, however, rely strictly on geology (specifically
bedrock) as the single risk determinant. This is adequate in order to generalize and
determine potential radon "hot spot" areas, but is insufficient in evaluating overall
potential risk. There are other factors and criteria that determine levels of risk.

This study provides, through a holistic approach, a geography of human risk to
indoor radon in Lenawee County. The study combines both human and environmental
risk factors to determine overall risk. Geologic risk considers bedrock, surficial features
and soil associations. Housing structural characteristics are examined, as well as
socioeconomic risk factors and personal behaviors.

It is beyond the scope of this study to examine specific disease outcomes resulting
from radon exposure. This study is based on the disease ecology approach which views
disease as multi-factored and multi-causal. Recognizing this, the focus of this study is to

assess overall risks that could lead to a disease outcome.

3
According to the EPA, no two homes have the exact same causes for radon entry;

therefore, the agency stresses that testing is the only way to determine if radon is present
in a home. Recognizing great variation in the conditions leading to human radon exposure
in the domestic setting, this study is an analysis of the most common factors of overall

risk.

CHAPTER 2: LITERATURE REVIEW

Radon Overview

For decades radon was known to pose a health risk to uranium miners and workers
in related industries. Over the last 10 to 15 years however, attention has shified away
from occupational exposure toward household risk in the general public. Human exposure
to radon has increased over the past 50 years with the transformation from an outdoor,
agricultural society to a largely indoor, business oriented society. Research indicates that
on the average, Americans spend 90 percent of their time indoors with 65 percent of that
time spent at home (American Lung Association, 1990). Because of this high degree of
human exposure to indoor air, research has increased on a variety of indoor pollutants,
especially radon.

Radon can enter a house and contaminate the indoor air by moving through the
soil and rock on which a house is built. Radon can seep into a home through dirt floors,
cracks in foundation floors and walls, floor drains, and sump pumps. Figure 2.] presents
major radon entry routes.

In the United States and Canada, the unit of measurement for radon is picocurie
per liter of air (pCi/L). A picocurie is a unit of radioactivity equal to 2.22 nuclear

transformations per minute of a given radioactive isotope.

 

 

F IGUREOZJ: MAJOR RADON ENTRY ROUTES

 

 

 

 

 

 

 

 

IIIIIIIIIIIII'II I I'x
'

  

 

 

 

 

 

 

 

 

 

( A
Y T. :
1'. w Al‘s” ;‘
3.5:: .r' , ".-‘.J\\:. h. Of“... . t. r} all
.'u. .-. O . . 8""

 

 

 

:O 'E ':':':‘ .
MAJOR RADON ENTRY ROUTES

.Cracks in concrete slabs

Spaces behind brick veneer walls

that rest on uncapped hollow-block loundation
. Pores and cracks in concrete blocks

. Floor-wall joints

. Exposed soil. as in a sump

. Weeping (drain) tile, ll drained to open sump
. Mortar joints

. Loose littlng pipe penetrations

. Open tops ol block walls

. Building materials such as some rock

. Water (trom some wells)

XL_IDWIMDO P)

Source: Reducing Radon in Structures, EPA, 1990

 

 

6
Radon is produced during the radioactive decay chain of uranium in which

uranium-238 decays to produce radium-226, which in turn produces radon-222. As this
radioactive decay chain continues beyond radon, other radioactive isotopes of several
elements are formed. Known as "radon daughter products" or "radon progeny", these
four short-lived radioactive isotopes (polonium-218, lead-214, bismuth-214, and
polonium-214) are not gases but are particulate solids which are more chemically active
than radon. These airborne radon decay products are of particular threat to human health
due to their availability to be inhaled and lodged in human tissue (Nazarofi‘ and Nero,
1988). As these radioactive daughters continue to breakdown, they release small bursts of
energy which can damage lung tissue and lead to lung cancer. The uranium-238 decay
process, illustrated in Figure 2.2, finally ends with the formation of a stable isotope of
lead-210.

The single most important factor in the presence of indoor radon is strength of the
source. If there is geologic potential for indoor radon, housing structural features can be
factors. Correlations have been made between radon concentrations and housing
structural characteristics (Mettlin, 1988). In areas where radon exists, the potential for
indoor entry becomes greater when cracks or small openings are present in foundation
floors and walls of a home. These types of features are often common in older housing
stock. Newer homes, however, can also pose a potential threat. Energy efficient homes
with "tight" construction can cause entering radon to accumulate inside since it is more
difficult for radon to escape into the outdoor air. For these and other reasons, it is

difficult to make broad based predictions of potential threat from radon because every

 

 

FIGURE 2.2: URANIUM DECAY CHART

      

218
POLONIUM
3 minutes

Source: Reducing Radon In.Structures. EPA, 1990.

 

 

8
house is unique. Accordingly, the EPA recommends that all households be tested for

radon. Despite EPA recommendations, growing public awareness of radon risk, and
health department radon information campaigns, fewer than 6% of US. home owners test

their houses for radon (Field et a1, 1993).

The Geology of Radon

A number of factors can contribute to the presence of indoor radon, but none are
more important than the strength of the source (EPA 1993: 2-1). The ultimate source of
radon is uranium, found primarily in bedrock and sometimes in reduced amounts in soil
made from rocks with a high uranium content. All rocks contain some amounts of
uranium, but some have elevated levels. These types of rocks include black shales,
volcanic rocks, and sedimentary rocks that contain phosphate (Otton, 1992; Gunderson,
1992).

The US EPA has used type of bedrock to produce radon potential risk maps for
the country. While this type of mapping is adequate for making broad scale
generalizations and viewing potential "hot spots", it fails to consider the filtering potential
of surficial features and soils. Variation in particle size, permeability, porosity, and
moisture content influence the movement of radon in the sub-surface. As a gas, radon
moves through the ground by following the paths of least resistance from areas of higher
to lower pressures.

Studies have revealed correlations between bedrock with high uranium content
and elevated indoor measurements in homes on the above surface (Christoides and

Christodoulides, 1993; Schumann, 1992; Crameri et a1 1989; Reimer and Gunderson,

9
1989). The characteristics of the actual housing structure can also influence the

movement and flow of radon.

Crameri et a1 (1989) conducted a study to examine the distribution of indoor radon
concentrations in different geological regions in Switzerland. Based on type of bedrock,
the country was divided into four geologic zones. Testing was conducted in 400 single-
family homes within the different regions: Region 1, the Molasse Basin is a hill country
where the subsoil is generally of sedimentary origin -- mean radon measurement reading:
1.62 pCi/L. Region 2, the Southeastern Alps is characterized by several sheets of
crystalline basement rocks with relatively high uranium content -- the mean radon
measurement reading was 8.6 pCi/L. Region 3, the Swiss Central Alps represents the area
of the highest indoor radon readings. This area has extensive external basement massives
consisting of crystalline and metamorphous rocks -- the mean radon measurement reading
was 54 pCi/L. The final region 4, Penninic and Helveic Nappes, has rocks and soil of
intermediate uranium content, the survey was very limited in this region and measurement
results were clearly lower than the average found in the two other alpine regions. The
authors conclude that regional differences in geology must be taken into consideration
when examining indoor radon levels and risk.

Reimer and Gundersen (1989) conducted an indoor radon study with particular
attention focused on geology and soil gas. Results from their study, carried out in the
Reading Prong near Boyertown, Pennsylvania, revealed that soil gas radon and indoor
radon concentrations showed a striking correlation when plotted on a geologic map. The
authors conclude that soil gas concentrations follow uranium concentrations in the

underlying bedrock.

10
The ultimate source of radon is uranium found most commonly in bedrock. The

presence of radon in the ground below a house does not necessarily mean elevated radon
levels exist within the home. A number of other factors must be taken into account when

examining indoor radon, including housing structural variables.

Radon and Housing

Three factors must be present for the existence of indoor radon: source, pathway,
and entry point(s) (EPA, 1993). Naturally occurring radon could not be present in a home
with the absence of any of these three components.

In the domestic setting, the most common source of radon is uranium contained in
underlying bedrock. During the uranium decay chain, as radon is formed, it moves
following the paths of least resistance: from high pressure areas, to areas of lower
pressure. It is most common for air pressure in the home to be lower than the relative
surrounding soil pressure. This pressure differential (relatively lower pressure within the
home, and higher pressure in the soil) can serve as vacuum or suction force to draw radon
into a structure (EPA, 1993; Demers, 1991). Substructure particle size, moisture content,
porosity, and permeability can determine radon pathways. Since radon moves through the
paths of least resistance, typically larger particle size and greater porosity facilitates its
movement.

Finally, in order for radon to enter a home, an entry point must exist. These
common entry points, detailed in Figure 2.1, include cracks in foundation floors or walls,
any exposed soil, and sump pump openings. These potential entry points or openings vary

with housing construction, age, and maintenance (Rannou, 1990).

11
A study by Cavallo et al (1994) examined the impact of basement depressurization

on the flow of radon into a structure. The pressure within a home basement was found to
effect the radon entry and concentration level in the home. Testing was conducted in a
research house before and after mitigation procedures were performed. This “double
testing” allows for evaluation of various mitigation techniques.

Christofides and Christodoulides (1993) studied residential radon concentrations in
Cyprus. In their study of 89 houses, the authors took into consideration the construction
of a typical Cyprus house and variations in geology. Most homes are made of a reinforced
concrete frame, with walls of fired red clay hollow bricks, and tiles on concrete floors.

The authors conclude that radon concentrations in households are correlated to the local
geology.

As part of another radon study, Svensson et al (1989), compared the type of
dwelling in relation to radon exposure. The hypothesis was that "the greatest variation in
radon concentration could be expected to occur among dwellings with ground contact, i.e.
one—family houses or apartments on the ground floor in multifamily houses without
basements" (Svensson et a1, 1989: 1861). Differences in geologic potential were
considered as radon measurements were taken in 303 dwellings in Sweden. Each dwelling
surveyed was classified as having ground contact or not, and information on building
materials, foundation, and ventilation were recorded. Results showed that those dwellings
with ground contact had an average indoor radon concentration of 4.32 pCi/L, which was
twice the average concentration of dwellings without ground contact.

Household structural characteristics are an important consideration in studies of

indoor radon. If radon is present around the foundation of a home, the construction and

12
maintenance of the dwelling can affect the entrance of radon into the structure.

Once inside a house, radon and its decay products can be carried by air currents,
and become attached to aerosols, droplets, dust, or surfaces. Once inhaled by humans,
radon and its decay products continue to breakdown during the process of radioactive

decay, causing damage while lodged in pulmonary tissues.

Lung Cancer, Disease, and Health Effects of Radon

Although it is beyond the scope of this research to examine the health effects of
radon, it is necessary to discuss the potential human health impact.

Radon is a known carcinogen as established by the World Health Organization, the
National Academy of Sciences' Biological Effects of Ionizing Radiation Committee, the
lntemational Commission on Radiological Protection, and the National Council on
Radiation Protection and Measurement. In 1988, the United States Surgeon General
issued an advisory to the public calling indoor radon gas a national health problem, stating
radon causes thousands of deaths each year, and suggesting that households be tested
(EPA, 1993).

Therefore, indoor radon continues to receive growing attention and is now
considered with environmental tobacco smoke, as the two indoor pollutants of greatest
health concern withregard to respiratory carcinogenesis (Samet, 1993).

Evidence suggests that exposure to radon and its radioactive decay products
increases the risk of developing lung cancer. Magnitude of this risk is directly related to
the intensity and duration of exposure (American Medical Association Council on

Scientific Affairs, 1991).

l 3
The smoking of cigarettes increases health risk by acting synergistically with radon

to multiply potential risk .for lung cancer (Demers, 1991). The risks from combined
exposure to radon and smoking is greater than the sum of the risks from exposure to
either acting alone (EPA, 1993). The smoke provided by cigarettes increases potential for
radon decay particle solids to remain air borne where they can more readily be inhaled by
humans. The EPA in May 1992 revised its radon risk estimates. Figure 2.3 provides the
updated radon risk estimates for the smoker and the non-smoker.

The scientific basis for radon risk estimates comes mainly from studies and
epidemiologic research on lung cancer mortality in underground miners. There is some
difficulty however, estimating domestic risk based on occupational exposure. In the
former case, usually exposure occurs in lower doses and a greater distance from the radon
source. In the latter case, exposure can occur at unusually high levels and in extremely
close proximity to the source.

In assessing human radon risk, there are three main categories of study. The first
type of study is miner or occupational. The second type of study residential or domestic.
The third type of study is animal or laboratory.

Studies of miners have been conducted in the United States, Canada, China,
Europe, and Australia. Studies are associated most commonly with workers in the metal,

fiuorspar, shale, or uranium industries. A consistent finding in a large number of miner

l4

 

FIGURE 2.3: RADON RISK ESTIMATES

urn II“ If VII SMOKE

 

 

 

 

 

 

Radon "1.me mfiudcancerfiunradon WTTODO:

tent m tothislavelover mm... mm“...
allietlme...

20pOI/l. m135peoplccoddget s—lOOtimesmerisltaMming fixmhome
Wm

land/l. Aboutnpeoplecoudget s-IOOtimestllerlsltoldyingina fixyourhome
Wm Mnefire

890/1. MSIpeoplecoudut Firmware
Warmer

Ina/L Amunpeoplecoddget s—lOOtimestlredskofMinan Fumhome
imam eirplarteaash .

cm finngbeteeenhnd

ZpCi/t AboutISpeoplecow‘et e-Zfimesvtefiskofmina 4pCi/L
WW catered:

L3pCi/l. Abotaneoplecowget (Average-immanent) (Mm radarleiels
mm belatr2pCJ/L

atpcr/t About3peopleaoddget «mammalian» ”MW
lungcanw

 

Note: It you are a lormer snroker, your risk may be lower.

Mill III“ II Ill I“! “WEI SMBIEI

 

 

 

 

 

 

Radon NLOOOpeoplemnever mfldwtcerlramadonmn WHATTOOO:
Level Mmmedtothis cumin...
leveloverab'ietime...

ZOpCi/l. mapeounouldget s—Thenskolbeingkiledinaviolentcrime Frxyowhome
Wanner

IOpCi/l. M4pcoplecoddget Fixyourhome
Wm

Spa/L m3peoplecouldget s—lOtimestheriskofdyingman Fuyourhome
lrmgcancer airplanecrash

”Ci/I M2peoplecodd¢ct s-Theriskafdromim Fuyourhome
W ConsiderfixingoenveenZand

2901/1 Aboutlpersmoouldget «mwofmmehanefire 490/"
Warner

L3pCt/l. Lessmanlpersmcwdget (Averagemdoomdonlevell (Reducingradonlevels
lungcancu belong2pCi/t

0.4pCi/L Lesstnantpersoncouldget (Averageoutdoomdonlevell "“li
lmgcancer

 

 

NotedtyouareelormeramoherJourrtskmhentgner.

Source: Home Buyer's and Seller's Guide to Radon, EPA. 1993.

 

 

 

 

15
studies is an increase in lung cancer risk with exposure to radon. An occupational study of

phosphate workers exposed to radon revealed a dose-response trend of increasing lung
cancer risk with increasing duration of employment (Block et al, 1988). Some studies
have gone on to examine exposure and dose rates, addressing higher short term exposure
versus low cumulative exposure.

A study by Xiang-Zhen et al (1993) of tin miners in China exposed to radon and
radon decay products is the largest occupational study of radon ever conducted. The
study group consisted of 17,143 workers with 98] lung cancer events. Eighty percent of
the workers were employed underground and thus exposed to radon. Results from the
measurement of risk revealed that "risk declines with increasing rate of exposure,
indicating that long durations of exposure at low radon concentrations may be more
harmful than short durations at high radon concentrations" (Xiang-Zhen et al, 1993: 130).
Findings of this nature are of particular importance to assessment of residential radon
exposure.

As discussed, a certain level of uncertainty exists in making radon risk projections
from occupational studies to residential dwellings. This is, in large part, due to the fact
that workers in underground mines are not representative of the general population, nor
are their working conditions. Stidley and Samet ( 1993) examined 15 published studies of

indoor radon and lung cancer. In these studies, contradictory results occur:

16
seven studies found a positive association between indoor radon and lung cancer, six

studies found no association, and two studies found a negative association. The authors
conclude that these’contradictory results occur due to shortcomings of studies, limitations,
and methodological problems.

The two types of residential studies most commonly conducted are ecologic and
case-control. In the former, an attempt is made to relate the number of lung cancer deaths
for a region with substitute measures of exposure for that region. The latter type of study
involves comparison of individuals with lung cancer (cases) and individuals without lung
cancer (controls) for differences in home exposures.

Regardless of the type of study conducted, results have varied. A number of
studies have shown an association between indoor exposure to radon and increased lung
cancer risk in humans (Pershagen et al, 1994; Klotz et al, 1989; Svensson et al, 1987; Lees
and Steel, 1987). Not all studies, however, produce this finding.

A study in the State of Washington revealed evidence of an inverse association
between residential radon exposure and lung cancer (N euberger et a1: 1993). This
particular research was a descriptive study designed to examine lung cancer death rates
compared to county levels of radon. Information was gathered from death certificates
which listed cause of death, smoking habits, and occupation. Previous radon measurement
results were used to. divide counties into high, medium, low, and very low radon risk
groups. Each county was assigned a degree of risk and then causes of death were

examined within each county. Results of the study revealed a significant lung

I 7
cancer excess found in the lowest radon counties. Equally, rates of lung cancer decreased

in the higher radon counties.

Further examination of this State of Washington study reveals that a number of
important variables were not considered in this survey: length and dose of exposure,
latency periods, and housing structural characteristics. It is also important to note that
individual indoor radon measurements were averaged for each county to produce a single
general range. Therefore it is possible that the residence of a deceased person may never
have been tested for radon. This makes it extremely difficult to examine the site specific
relationship between radon and lung cancer within this study.

Results from animal studies have been far more consistent. There is consensus
among scientists that radon causes adverse health effects in laboratory animals. Numerous
experimental animal studies in dogs and rats have shown that radon is a lung carcinogen
(Hofrnann et al, 1993; Cross et al, 1982).

Animal studies have provided valuable information on the effect of exposure rate
on cancer risk, and the potential effect of simultaneous exposure to radon and other
contaminants on the radon lung cancer relationship. Studies involving radon and rats have
been particularly revealing. Laboratory rats exposed to low levels of radon have been
found to have an increased number of respiratory tumors. Another finding is that rats
exposed to radon progeny and uranium dust simultaneously have been found to have
elevated lung cancer risk at exposure similar to those found in homes. The risk decreased
as radon exposure decreased (EPA, 1993). These animal studies involving radon continue
to expand the growing body of knowledge concerning human health risks from radon

exposure.

CHAPTER 3: STUDY OBJECTIVES AND RESEARCH QUESTIONS

The purpose of this study is to provide a geography of overall human risk to
domestic radon exposure in Lenawee County, Michigan. The research questions address

the study objectives.

Study Objectives

Primary objectives of this study are to:

0 better characterize the spatial variation of radon. The single most important factor
contributing to the presence of indoor radon is strength of the source. Because radon
emanation from bedrock types is not uniform, knowledge of type and location of

bedrock is crucial to developing a geography of risk.

0 better char_acterize and assess risk by utilizing a holistic approach within the disea_se_
ecology tradition of medical geography. Although it is beyond the range of this study
to examine disease outcomes specifically, this study examines the factors and
conditions which contribute to risk of disease. The disease ecology approach seeks to
examine the inter-relationships between humans and the environment. Disease is
viewed as a complex and multi-factored outcome of these inter-relationships.

Following this approach, this study examines the three major factors of domestic radon

l8

19
risk: the physical environment in the form of overall geology, housing structural

characteristics, and human behavioral variables.

identify potentially high and lowdon risk areas based upon bedrock. surficial
geology, and soil pipe. Knowledge of the physical characteristics of the study area is
necessary to determine potential risk. Location of source (bedrock) has been
determined and strength has been estimated. Surficial features and soil were examined
as filtering layers for movement of radon from the bedrock source. From these

variations, potential high and low risk areas were developed.

§x_amine the relzLionship between potential rislpand housingstructural characteristics.
Overall condition of the dwelling, as well as specific construction features, were
recorded (for example, poured versus block basement). Housing structural

characteristics were compared with measured radon levels.
examine the relationahip between personflehaviorsJ and level of risk. Variations in
smoking status and amount of time spent at home in the basement were examined and

considered with potential for risk.

provide state and coung health officials with a dfliled analysis of overall risk to

 

@on in the county. All data, analysis, and findings have been made available to state

and county health officials.

20

 

0 provide a valuable public service to county residents by testing houses and creating

increased awareness. Public awareness is a vital step in prevention and control of a
health risk. Not only were survey participants given literature on radon, but an actual
test was conducted of each household with the result made known to the home owner.
Follow up information was also provided to answer questions and help determine

appropriate actions if necessary.

Research Questions

Of central importance to this study are the following questions:

0 What are the spatial patterns of rial; to radon associated with geologic var_iation in the

county? i.e. high and low potentfl risk areas. How do variations in bedrock, surficial

 

geology, and soil combine to produce domestic radon risk. Where are the potential

high and low risk areas within the county?

0 In areas of similar risk potential based upon bedrock, to what degree do surficial
geology and soil pipe impact radon levels? Surficial features and soil can serve as
filters for radon movement in the ground. To what degree is indoor radon potentially

affected by these filtering layers?

21

o What role does housing construction and structural characteristics have in determining

the variation of radon measurement readings? What are the variations in levels of
radon risk based upon the structural characteristics of a dwelling? For example, do

block basements and foundations present more of a risk than a poured foundation?

0 What is the prominence of intervening human behaviors in Lepawee County that may

 

place people at increased ris_k to indoor radon? Cigarette smoking and radon have a
synergistic effect. Typically, smokers are at a higher potential risk when radon is
present. Radon risk is based not only on exposure dose but on duration of exposure as
well. Spending time in the basement for various activities can increase length of
exposure, and thus increase potential risk. To what degree do these intervening risk

behaviors occur in Lenawee County?

0 How do test results fiom this survey compare with the predictions baaed upon the
screeflgs in the Michigan Department of Public Hefl? A very small sample was
chosen for home testing in Lenawee County during the state wide survey in 1987.
How do the results of a larger sample size compare with the state's initial
measurements and estimates?

It is beyond the parameters of this study to investigate disease and health effects of
radon in individuals. Radon has been named the second leading cause of lung cancer by
the EPA and American Lung Association. In the disease ecology approach, disease is

viewed as multi-causal and multi-factored. Following this tradition, the presence of lung

22
cancer within an individual in a surveyed household may be due to a number of factors,

including, but not limited to, domestic radon exposure. Furthermore, long latency periods
are associated with radon exposure and risk is based upon the amount and duration of
exposure. For these reasons, this study seeks to examine and assess the broad factors of

domestic radon risk.

CHAPTER 4: METHODOLOGY

Host-Agent-Environment Interaction Model

In medical geography, the disease ecology approach views disease as multi-causal
and multi-factored. A disease outcome is seen as a complex process where all necessary
components must converge in space and time (Pyle, 1979; Meade et al,l988). This
holistic approach examines the interaction between a human and the total environment.
As Paul (1985: 400) writes, "the principal goal of disease ecology is to understand the
dynamics of disease which vary according to climate, vegetation, mineral traces in water
and bedrock minerals."

The host-agent-environment model is used by medical geographers,
epidemiologists, and other scientists and practitioners to explore the interaction between
factors that can produce an outcome of disease. Figure 4.1 shows the model as it applies
to this research. In this case the agent, which does not necessarily have to be a living
organism, is radiation, or more specifically radon. The host is the human. The
environment is where the agent and the host live or meet, which in this case is the
domestic dwelling. Each of these components must be examined individually and

collectively.

23

24

 

 

FIGURE 4.1.: HOST-AGENT-ENVIRONMENT INTERACTION MODEL

HUMAN RADON

 

HOUSING

 

 

25
There are a number of influential factors acting on each of these model

components. For instance, the presence of radon in a home is dependent primarily on the
strength of the radon source and the ability of the gas to travel through surficial
formations and soil; and also varies according to home construction characteristics.
Furthermore, various behaviors of the human host, such as amount of time spent at home

and smoking status, can contribute to risk levels.

Radon In Michigan

In late 1986 the EPA announced that Michigan had been chosen to be among the
first states to receive EPA assistance in conducting an indoor radon survey. From
October 1987 through May 1988, the Michigan Department of Public Health (MDPH) in
collaboration with the EPA and county health departments, implemented a state-wide
indoor radon survey. Over 2,500 selected measurements were taken in single-family,
owner occupied homes in 79 of Michigan's 83 counties. Over 2,000 households were
randomly selected for testing, while approximately 500 additional houses were chosen
based on a geographic sampling scheme designed to assure better geographic coverage.

Findings of this survey project that 88% of Michigan homes have indoor radon
screening levels less than or equal to the EPA's recommended action level of 4 picoCuries
per liter of air (pCi/L); 12% have screening levels between 4 and 20 pCi/L; and fewer than
1% exceed 20 pCi/L (Michigan Department of Public Health, 1990; Michigan Department

of Public Health, 1988).

26
Figure 4.2, Regional Screening Measurement Projection, represents the percentage

of homes estimated to have screening levels greater than 4 pCi/L for each participating
county. - The three counties with the highest screening measurement frequencies are
Hillsdale, Lenawee, and Washtenaw, where 40-45% of homes are estimated to have
readings above the EPA action level.

As a result of the indoor radon survey, the MDPH has made a number of
recommendations, including additional testing in all counties, especially those areas where

greater than 15% of the homes are estimated to have screening levels above 4 pCi/L.

Radon in Lenawee County

Lenawee County participated in the MDPI-I state-wide radon survey in 1988. The
test results from 39 households produced a geometric mean of 4.2 pCi/L, an arithmetic
mean of 7.8 pCi/L, and a median of 3.9 pCi/L. From these results estimates were made
indicating that Lenawee County had 40-45% of homes with radon readings greater than 4
pCi/L. Lenawee County was one of three counties with screenings levels in the highest

potential risk group.

Description of Study Area

Lenawee County is located in south eastern Michigan on the Ohio State border
(indicated in Figure 4.2) and has an area of 654 square miles. Neighboring counties
include Monroe, Washtenaw, Jackson, and Hillsdale. The total population is 91,476

( 1990). There are 22 townships in this largely rural county. The Lenawee County

27

 

FIGURE 4.2: REGIONAL SCREENING
MEASUREMENT PROJECTION

'-

“\\
a
' mw. ‘r‘ r,-

9009. s

000900 ,-

00690 9.9

%&&&39 ..

\ .~... . . .....

9.0.0 e 9.9.

t. eze: e

Percentage of Home: Estimated

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

With Screening Level Greater

Than 4pCt/I
1 N 0-1073 {r * [I j
2 , 10-157. * * l L

 

 

 

 

3 15-207. ~ L1 \ ,
4 30—357. E g? //

5 I 40-457.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

¢
* Non-Participant“ CNN! Lenawee County

 

100 miles

 

Source: Michigan Department of Public Health, 1990.

 

 

 

28
Health Department is located in the city of Adrian. Figure 4.3 shows census tract

boundaries (dark lines) and Minor Civil Division boundaries (light lines) with cities and
towns in lower case and townships in capital letters.

A large segment of the population earns their livelihood by engaging in farming
and raising livestock. The cities of Tecumseh and Blissfield serve as home to a number of
workers who commute daily to the larger urban areas of Ann Arbor/Detroit, Michigan and

Toledo, Ohio respectively.

Contacts

This entire project has been carried out with the firll support and cooperation of
state and local health departments and the citizens of Lenawee County. As with any
outreach project, it was vital to work through the appropriate administrative channels and
community structures, while developing both professional and community ties.

The Michigan Department of Public Health, Division of Radiological Health was
aware of this project as it was being developed, and agreed to provide help in the
implementation of this study. The State’s Indoor Radon Specialist, who encourages local
projects that promote home radon testing, took great interest in this project. The
Michigan Department of Public Health provided the test kit devices and radon educational
literature for this study. All data collected have been shared with state and local health
departments.

At the local level, this study was carried out through the Lenawee County Health

Department. The county was very receptive to learning more about radon through this

29

 

 

FIGURE 4.3: LENAWEE COUNTY CENSUS
TRACTS AND MINOR CIVIL DIVISIONS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_‘ l Morcnci L

l jCement {:61 Clinton]
City C N
woopsrocx CAMBRIDGE FRANKLIN TECUMSEH MACON
Onstead I '
Salmon Tecumseh
[— l .. Britton
Manrtou 1 ‘
Beach- W
Devrl's ROME ADRIAN RIDGEWAY
Lake RAISIN 1 r'
ROLLIN
L r P 1 0:55thch
f C1 Adrian ' '
a ton L...
“1 HUDSON -‘ y PALMYRA BLISSFIELD meme”
.. l] J DOVER MADISON .
Hudson 1:“ CHARTER l
1 igg l
Blissfteltl
MED'NA SENECA FAIRFIELD OGDEN RIGA

 

 

 

 

 

 

Dark lines represent census tracts; light lines minor civil divisions.

 

six miles

 

 

30
study and provided full office support, secretarial support, office space, and an

undergraduate intern to this survey.

Developing relationships with State and local agencies strengthened the credibility
of this project in the eyes of the local community. The citizens of Lenawee County were
supportive and cooperative, and most proved to be very eager to learn more about radon.
Contacts were established with the community in various ways including interviews for
newspaper and magazine articles, contact letters, phone calls, and community speaking

engagements.

Seasonal Variation of Indoor Radon

The amount of radon present in a home varies by season. Typically, measurement
readings taken in the winter months are higher than readings taken in the spring and
summer months. The EPA recommends testing homes during the “closed house” winter
heating season. The measurement results obtained in this study represent a worst case
scenario for the amount of radon in houses.
EPA Radon Action Level

The EPA has determined an indoor domestic radon measurement of 4.0 pCi/L or
greater as an “action level”. If the measurement result in a home obtained from a short
term or long term test meets or exceeds this action level, the EPA recommends specific

actions on the part of the house occupants.

31
Spatial Design of Field Sample

WAC
' Determination of a sample size for this research was linked with the type of
statistical test proposed for data analysis and logistical constraints.

Based on the geologic risk potential codes deve10ped for this survey, 27 (3 x 3 x 3)
possible combinations exist between bedrock, surficial features, and soil associations. Each of
these combinations can represent a cell. Of the 27 possibilities, 19 actually occur in the
geologic radon risk potential ofLenawee County.

In order to avoid unwanted interactions in the results of certain tests, it is most
desirable to have equal numbers of observations in each cell (Johnston, 1991). This is known as
balanced as opposed to unbalanced observations. In this study, since the geologic potential
was estimated based on existing information before sampling, it was possible, in theory, to
balance the cells. As a rule of thumb, based on chi-squared, five or six observations are
necessary as the minimum number of observations for each cell. In the planning stages of this
study the minimum number of five observations was doubled to produce 10 observations per
cell, for a sample size of 190 households (19 x 10). This sample size of 190 households was
viewed as appropriate given possible types of statistical analysis and logistical constraints. No
occupied housing existed in three of the 19 geologic potential risk code areas, therefore the
actual number of geologic potential risk codes where sampling occurred was 16. The total
number of households surveyed for this study was 230, which is an addition of 40 over the

sample size of 190.

32
Sample Design and Selectipn Procedures

Overview

The sample design and methodologies for this study were chosen with the purpose
of obtaining the best possible set of representative field measurements. An operational
plan and framework was developed to guide this research and address the objectives and
questions set forth.

This survey was designed to utilize a stratified random sample. As part of the
operational plan, sampling would occur in every township and census block within the
county. Strata included in the sample are the geologic variables of bedrock, surficial
features, and soil associations, as well as the census variable of total number of occupied
housing units. Other data relevant to this study were gathered fi'om each sampled
household. During the initial household visit a questionnaire addressing socio-economic
and human behavioral variables was administered and a visual inspection of housing
structural characteristics were made and recorded, all according the operational plan.

The study was designed to select houses to be sampled based upon underlying
geologic strata (i.e. bedrock, surficial features, and soil types) and numbers of households
within townships. Every effort was made to balance the number of observations in each
cell and honor the assumptions in certain statistical tests. Detailed air photos were used to
identify households in selected areas. Selected households were contacted initially by
mail, and then by telephone, to arrange an initial visit. A short questionnaire was
administered, a radon detector placed, and a visual inspection of the structure was made
during the initial visit. A second visit four days later was arranged for the collection of the

radon detector.

33

Sample Design

This research utilized a stratified random sample into which the sample size of 190
was allocated. This sample was stratified based on overall geologic risk potential
(bedrOck, surficial features, and soil association potentials) and on the number of occupied
housing units in the county. The sample size of 190 households was allocated into the 22
townships of Lenawee County in order to reflect population housing patterns at the
township level based on the number of occupied housing units. It was most practical to
work at the township level since all county records are organized by township and section
numbers. Allocation at the township level was also more practical given the great
variation in the county of geologic strata components within the sampling fi'amework.
Table 4.1 lists the allocation of the sample size by township.

Overall geologic risk potential is expressed as a three digit code. The first digit is
bedrock risk potential, the second digit represents surficial features risk potential, and the
third digit represents soil association risk potentials. Each of these three geologic strata
was assigned a code of high (3), medium (2), and low (1) potential based on the estimated
potential to emit or transmit radon. In theory 27 possible combinations exist (3*3*3).
Nineteen of these possible combinations actually occur in Lenawee County. Of these 19
geologic risk potential classifications that occur in the county, a total of 16 had households
located within their boundaries. Therefore sampling occurred in 16 different areas of

varying potential geologic risk.

34
Table 4.1: Allocation of Sample Size By Township

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Macon (l) 423 2 4
Clinton (2) 2224 12 5
Tecmnseh (3) 3433 18 16
Franklin (4) 887 5 5
Cambridge (5) 1829 10 10
Woodstock (6) 1572 8 13
Rollin (7) 2095 11 4
Rome (8) 530 3 3
Adrian (9) 8944 47 61
Raisin (10) 1834 10 22
Ridgeway (l l) 782 4 5
Deerfield ( 12) 904 5 7
Blissfield (13) 2714 14 20
Palmyra (14) 771 4 6
Madison (15) 1624 9 13
Dover (16) 660 3 4
Hudson (17) 1498 8 6
Medina (18) 448 2 4
Seneca (19) 13 10 7 4
'Fairfield (20) 652 3 4
Ogden (21) 379 2 7
Riga (22) 519 3 7
Totals 36032 190 230

 

 

 

 

 

35

The overall purpose when allotting the sample size was to collect the best set of
representative radon measurements. The challenge of the sample allocation was to place
190 samples into the 16 areas of geologic potential with respect to household and
population distributions, while attempting to balance the cells for future statistical analysis.
Because of the great variation in the amounts of surface area each geologic potential
group occupies, and the number of homes in each of these areas, balancing the cells was
not possible. Furthermore, sampling could only occur in areas where houses exist.

Every effort was made to allocate the sample according to the assumptions of
certain statistical tests. Certain assumptions of statistical tests presuppose the sampling
area as a friction free smooth plane with equal distributions of phenomena. It is known
that the real world is not this way, and that the real world is not random. There exist
outside influences and uneven distributions of phenomena.

Sampling for this study occurred in every township, every census tract, and in the
majority of block groups. The result is a fairly representative reading of domestic radon in

Lenawee County.

Geologic Risk Classification System

This section discusses the characteristics of bedrock, surficial features, and soil
association groups of Lenawee County. The different formations within each strata were
categorized into high, medium, or low potential to emit or transmit radon. These

classifications were developed as working assumptions.

36
Following this classification of assumed high, medium, and low within each

geologic strata, a numerical value was assigned to each: 3 to high, 2 to medium, and 1 to
low. Components of each strata have a number of 3, 2, or 1, with three being highest and
one being the lowest potential for radon emission or transmission. Within this system, a
three digit code (in the order of bedrock, surficial, soil association) reflects the underlying
geologic radon potential at any given point in the county. For example the code 3,2,1
would indicate high potential bedrock, medium potential surficial geology, and low
potential soil (see Table 4.2: Key to Geologic Radon Risk Potential Codes).

The bedrock in Lenawee County is comprised of formations from the
Mississippian and Devonian Systems. Figure 4.4 shows the bedrock formation

classifications of the county.

37
TABLE 4.2: KEY TO GEOLOGIC RADON RISK POTENTIAL CODES

First digit: bedrock
Second digit: surficial features
Third digit: soil associations

These classifications were developed as working assumptions.

Components of each strata have a number of 3, 2, or 1 with three being the assumed
highest and one being the assumed lowest potential for radon emission or transmission.

First Digit: BEDROCK
Code 3: high potential risk: Antrim Shale and Sunbury Shale.

Code 2: medium potential risk: Coldwater Shale, Bedford Shale, Marshall Sandstone, and
Berea Sandstone.

Code 1: low potential risk: Dundee Limestone, Traverse Group, and Detroit River Group.
Second Digit: SURFICIAL FEATURES

Code 3: high potential risk: moraines and outwash/glacial channels.

Code 2: medium potential risk: sand lake beds.

Code 1: low potential risk: ground moraines (till plains) and clay lake beds.

Third Digit: SOIL ASSOCIATIONS

Code 3: high potential risk: associations defined as rolling to hilly, well-drained loamy
sands and sandy loams; and level to gently rolling, well-drained soils developed from
sandy loam and loamy sand overlying sand and gravel.

Code 2: medium potential risk: associations defined as gently rolling to rolling, well-
drained and imperfectly drained loams; level to nearly level, poorly drained soils developed
from loam, sandy loam, and loamy sand overlying limy sand and gravel; and level and
undulating, imperfectly and poorly drained soils developed in deltaic and lacustrine
deposits.

Code 1: low potential risk: associations defined as undulating and rolling soils developed
from limy clay loams, silty clay loams, and clays; nearly level, imperfectly and poorly
drained soils developed from clay loams, silty clay loams, and clays; and level, poorly
drained soils developed from clay loams, silty clays, and clays.

38

 

 

FIGURE 4.4: BEDROCK RISK POTENTIAL CLASSIFICATIONS

 

 

....-.- _

-

 

 

 

 

 

 

 

 

r.f. ”271,874

 

five miles
Assumed Bedrock Radon Risk Potential Codes:

3 = High: Antrim Shale and Sunbury Shale

2 = Medium: Coldwater Shale. Bedford Shale. Marshall
Sandstone. and Berea Sandstone

1 = Low: Dundee Limestone, Traverse Group, and
Detriot River Group

 

 

39

It is very diflicult to determine actual uranium content of bedrock. In order to
estimate potential risk to emit radon, working assumptions were developed as predictors
to classify bedrock formations based upon description, color, and composition (Schumann,
1993; Western Michigan University, 1981; Ruotsala, 1980).

Organic rich shales (especially black shales) possess higher-than-average amounts
of uranium, while lighter gray shales have lower uranium content (Schumann, 1993). Two
formations have been classified as having the highest potential to emit radon (potential risk
code 3). These are the highly radioactive Antrim Shale and Sunbury Shale. Antrim Shale
is a hard, dark gray to black or dark brown, pyritiferous shale that is part of the greater
Eastern Black Shale that stretches into Ohio, Indiana, and Kentucky. Antrim shale is
considered to have high potential to emit radon since uranium content is greatest in this
type of shale (Ruotsala, 1980). Sunbury Shale is similar in many respects to Antrim Shale
including elevated uranium content, and is dark gray to black or brown, bituminous, and
pyritic (Schumann, 1993; Western Michigan University, 1981).

Four formations have been classified as having medium or moderate potential to
emit radon (code 2): Coldwater Shale, Bedford Shale, Marshall Sandstone, and Berea
Sandstone. Coldwater Shale is a gray micaceous shale with a considerable sand
component. Another medium potential shale is the gray Bedford Shale. Coldwater and
Bedford Shales contain uranium but in lower amounts than Antrim and Sunbury Shale
(Schumann, I993; Ruotsala, 1980). The Marshall Sandstone is made up of sandstone,

siltstone, and shale. These sandstones are well-sorted and dominantly quartz. Berea

4O
Sandstone consists of a moderately fine-grained sandstone which grades upward and

downward into shaly, dolomitic sandstone (Western Michigan University, 1981).

The final three formations have been assigned a low potential rating for radon
emission (code 1): Dundee Limestone, Traverse Group, and Detroit River Group.

Dundee Limestone is made up of fine grain, dolomitized limestone. The Traverse Group
is light gray in color with dolomite on the top, crystal, and limestone, with a very thin layer
of shale in the middle to lower portion. Finally, the Detroit River Group is made up of a
wide variety of rock including sandstone, limestone, dolomite, anhydrite, and halite.
(Western Michigan University, 1981).

Figure 4.5 provides the surficial formation classifications of the county. The five
surficial formations within Lenawee County were also grouped into three levels of
potential radon transmission, based on content, particle size and permeability. High
potential (3) include moraines, and outwash/glacial channels; medium or moderate
potential (2) includes sand lake beds; finally the low potential group (1) is comprised of
ground moraines (till plains) and clay lake beds (Martin, 1955).

Soil association groups are shown in Figure 4.6. The eight soil association groups
were placed into the categories of high, medium, and low potential to transmit radon.
Drainage, particle size, and clay content were taken into consideration in determining over
all permeability potential of the soil association groups. Categorization was complicated
by wide variation and composition of certain association groups. The breakdown of high,
medium, and low follows: -- high potential associations (3) defined as rolling to hilly, well-

drained loamy sands and sandy loams; and level to gently rolling, well-drained soils

 

41

 

FIGURE 4.5: SURFICIAL FORMATION
RISK POTENTIAL CLASSIFICATIONS

 

 

 

 

 

 

 

 

 

 

 

 

l

 

 

r.I. ”271,874

 

five miles

Assumed Surficial Feature Radon Risk Potential Codes:
3 = High: moraines and outwash/glacial channels

2 = Medium: sand lake beds

1 =Low: ground moraines (till plains) and clay lake beds

 

 

42

 

 

FIGURE 4.6: SOIL ASSOCIATION RISK POTENTIAL CLASSIFICATIONS

 

 

 

 

_____________________.__..
r.f. “271.874

 

 

 

five miles
Assumed Soil Association Radon Risk Potential Codes:

3 = High: associations defined as rolling to hilly. well-drained loamy sands and sandy loams;
and level to gently rolling, well—drained soils developed from sandy loam and loamy sand
overlying sand and gravel

2 = Medium: associations defmed as gently rolling to rolling , well-drained and imperfectly
drained loams; level to nearly, poorly drained soils developed from loam. sandy loam, and
loamy sand overlying limy sand and gravel; and level and undulating. imperfectly and poorly
drained soils developed in deltaic and lacustrine deposits.

1 =Low: associations defined as undulating and rolling soils developed from limy clay loams,
silty clay loams, and clays; and level, poorly drained soils developed from clay loams, silty
clays, and clays.

 

 

43
developed from sandy loam and loamy sand overlying sand and gravel; -- medium

potential associations (2) defined as gently rolling to rolling, well-drained and imperfectly
drained loams; level to nearly level, poorly drained soils developed fiom loam, sandy loam,
and loamy sand overlying limy sand and gravel; and level and undulating, imperfectly and
poorly drained soils developed in deltaic and lacustrine deposits; -- low potential
associations (1) defined as undulating and rolling soils developed from limy clay loams,
silty clay loams, and clays; nearly level, imperfectly and poorly drained soils developed
from clay loams, silty clay loams, and clays; and level, poorly drained soils developed from
clay loams, silty clays, and clays (US. Department of Agriculture, 1961).

Whenever possible, this soil association information was supplemented with more
specific soil information. Additional sources of soil information were well-log records on
file at the county health department, Lenawee County Health Department sanitarians, and
on site information provided by the home owners.

A primary objective of this study was to develop a geologic risk potential based
not only on bedrock, but surficial formations and soil associations as well. To accomplish
this goal, the potential risk code maps of these three geologic strata were combined. This
combination process involved an overlay of the three geologic strata in order to determine
site specific overall geologic risk. The bedrock, surficial, and soil association files were
exported fi'om C-MAP into ArcInfo for the final overlay. This process involved two
separate overlays in order to generate the final geologic radon risk potential map. The
first sequence overlayed the bedrock file (Figure 4.4) with the surficial features file (Figure
4.5) which produced an intermediate file. The second sequence involved the overlay of

the intermediate file with the soil association file (Figure 4.6). This step completed the

44
overlay of the three geologic strata and produced 191 polygons. Finally, the risk potential

value labels for each strata were displayed in each polygon, giving potential risk in the
three digit code sequence of bedrock, surficial and soil association. This final overlay
(Figure 4.7) represents the overall geologic radon risk potential for the entire county by
individual strata.

An additional step involved the plotting of the township boundaries over the final
geologic risk potential map (Figure 4.8). This map enables the underlying geologic risk
potential to be known for each township and section. The three numbers in each polygon
represent potential radon risk as high (3), medium (2), or low (1), presented in the order
of bedrock, surficial features, and soil association.

Household Selection, Contact, and Visitation

Houses for this study were selected using a stratified random sample. Satisfying
the geologic and census stratification criteria, houses within each sampling unit were
assigned a number fi'om one to ten. A random number table was used to select the houses
from each sampling unit. Table 4.1 represents the allocation of the sample size of 190
houses into the 22 townships of Lenawee County. Sampling could only occur in areas
where households are located.

Detailed air photos of the county were used to precisely locate houses within
township sections. These air photos provided coverage of the entire county at the section
level. On the actual air photo, each property is assigned an identification number. This

property identification number was used to obtain the home owner name and address.

45

 

 

FIGURE 4.7: OVERALL GEOLOGIC RADON RISK POTENTIAL CODES

 

 

 

 

 

 

 

 

 

2‘ f r.f. ”271.874

 

 

five miles

Assumed Overall Radon Risk Potential Codes:
First Digit = Assumed Bedrock Radon Risk Potential
Second Digit = Assumed Surficial Feature Radon Risk Potential

Third Digit = Assumed Soil Association Radon Risk Potential

 

 

 

 

 

FIGURE 4.8: OVERALL GEOLOGIC RADON RISK POTENTIAL CODES
WITH TOWNSHIP BOUNDARIES AND SECTION CORNERS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L; a v A l . .
2 233 12
O O O 0 g e 9 23' + Q
s O 9
e e e e e on 2”
‘ O " p
I
* t233' ‘ ’ s 3 i
e e e O O ‘ e |
eat I t . 1‘ ’ ’
O 4
I . . . ' .5 O O O '
H2A . tfl
w V b
23! s j‘
2 O .23. 2.) Q . 3"
z , e e
3 ’ ’ r
O . e 0 12”. e e l , e e
23 e
e e s e 9 e
4 e e . 1 ,
‘21 +
. . Q 0 '9 '
| O
'7 'VV .V VV ’ 2'2
e e 9 ‘ T T
o e e 0 ‘ '
e 9 ' ' I
I - 22 '2
4 C
e + 2'
e e 4 0
’ e . , 4 9 2
s e e e
4 e e 6 i3“ 3§I
‘ AA/ 4 -5 y
. O 8‘ O O
6 o t O
211‘
22: , e e e 9
e
e 9 ’
22! 4 . O O O
. e e t t 2” an
O O O Q Q
, an O O , ,
4 O 9
. . . Q 0 T .
e 9 Ill
9 e O O 9
2m . s .1 t
2H , 2
2t
‘ O A

 

 

 

 

 

 

 

r.f. ”271.874

 

five miles

 

 

47
To further help in the household location process, a map representing the geologic risk

potential with township boundaries and section corners was displayed (Figure 4.8).

According to the operational plan, all households were selected according to the
study criteria and then initially contacted by mail. Through word of mouth, involvement
of the media, interaction with the community, and efforts of the health department, some
people offered to participate in this study without first being oflicially contacted. The
criteria for household selection was applied to these self selected households, and if a
household fit into the over all survey design, then it was incorporated into the study and
considered for testing.

Those households selected for the study were first contacted by mail. Appendix A
and Appendix B show the Lenawee County Residential Radon Survey Letter and Survey
Information Sheet which was the initial contact letter that provided a brief introduction to
radon and to this study, and asked the home owner to participate. Because mail delivery
took as many as five days, follow up phone calls to the home owner requesting
participation were made five to seven days after mailings. If the home owner agreed to
participate, a time was scheduled for the initial home visit. Those self-selected households
that suited the sample design, were placed into the visitation schedule and tested when the
survey was being conducted in their area. In order to minimize travel time between home
visitations, when possible, appointments were scheduled with close proximity to each
other.

Households that participated in the study were visited twice. The first visit, which took

approximately 25 minutes involved placing the test kit device, administering the short

48
questionnaire (Appendix C), and completing the home inspection. At this time, arrangements

were made for the second visit for the sole purpose of collecting the radon detector.

The type of radon measurement device used in this study was a charcoal packet made
and analyzed by Air Check, Inc. in Arden, North Carolina. In each home tested, the detector,
was hung from the ceiling (by masking tape), and placed in the lowest livable level of the home
following manufacturer’s recommendations and EPA test kit placement protocol With this
type of device, the length of the test can be from three to seven days. In this study, the
measurement devices were exposed during four day periods.

The EPA (EPA, 1993) provides a number of measurement protocol steps which were
carried out during this survey. Measurements were taken in the lowest livable level of the
home. If a basement existed, the measurements were made there. If no basement existed,
testing occurred on the first floor of the dwelling. The detector was placed in a position where
it would not be disturbed during the measurement period. The measurement was not made
near drafts caused by heating, ventilating, doors, fans, and windows. Areas of excessive heat
were avoided. Furthermore, measurements were not made within 12 inches of the exterior
walls of the building - where a window or other potential opening existed in the exterior wall,
the device was placed at least three feet distant. Finally, the testing device was suspended five
feet above the floor.

To ensure quality control, a small percentage (7%) of duplicate measurements were
taken, in addition to 3% unexposed devices or blanks which were sent to the manufacturer for

analysis.

49
estionnaire

The Lenawee County Residential Radon Questionnaire (Appendix C) was designed to
gather data specific to the objectives of this study, while at the same time remaining brief for
ease in administering and answering. The questionnaire addressed household demographics,
housing structural characteristics, human risk behaviors, and radon awareness. Certain
questions gathered information on the physical condition of the house and presence of
intervening features such as sump pump, age of structure, and foundation type. Any unusual
features were also noted on this form.

Each household surveyed was assigned a unique household survey number which was
location specific. This 10 digit number consists of the last two digits of the census tract
number, the two digit block group code, the two digit township number, the two digit section
number, and the two digit household number. For example: 12-05-08-17-02 would indicate
the location of this house as census tract 12, block group 05, township 08, section 17,

household 02.

Environmental Protection Agency Training

To increase my knowledge of radon and awareness of Environmental Protection
Agency protocol for radon testing and mitigation, I underwent a professional training
course in July 1993. This comprehensive course was offered by The University of Illinois
at Chicago, School of Public Health. I have completed the EPA radon program
examination and am certified in radon measurement and Operator proficiency (U SEPA

RMP Number: 150620T). This training was undertaken to ensure firll professional

50
reliability for this study. (Appendix D: Radon Measurement Operator Proficiency

Certificate).

Standard Field Methods and Procedures

All investigative measures and procedures conducted in the field were undertaken
according to standardized and approved EPA methodologies. Test kit device handling,
placement, and retrieval were performed according to EPA protocol and test kit
manufacturer’s specifications. All dealings with the public occurred in a courteous and
professional manner. At the request of the local health department, all information
requests of the study investigator were answered according to EPA guidelines and
regulations, especially legally delicate questions involving the disclosure of measurement
results which can have implications in real estate transactions.

Once home testing was completed, residents were promptly sent measurement
results with a letter of explanation, interpretative information, and phone numbers for
further questions (See Appendix E). Printed information and advice about mitigation was
also provided as part of the follow up procedure. Radon mitigation procedures are listed

in Appendix F.

CHAPTER 5: SURVEY RESULTS BY LOCAL GOVERNMENT
ADMINISTRATIVE DISTRICTS

Presentation of Survey Results

This chapter will provide a description of the survey results. The data will be
presented in the following order: overall county summary statistics and results by
township.

Overall Summary Statistics

Two hundred thirty homes were surveyed during this study. Since the distribution
of measurement results is positively skewed (as seen in Figure 5.1), throughout this study
the geometric mean has been used instead of the arithmetic mean with one exception
(Table 5.2. which presents a comparison of results fiom this study with the Michigan
Indoor Radon Survey).

Table 5.1 presents the overall summary statistics for the sample size. The median
measurement value is 3.4 pCi/L and the geometric mean measurement value is 3.3 pCi/L.
In this study measurement readings ranged from 0.3 pCi/L to 48.4 pCi/L. The final
statistic in this table reveals that 41% of the homes surveyed had measurements which
meet or exceed the EPA action level of 4.0 pCi/L. These homes are considered “at risk”

due to the elevated measurement readings obtained from the dwellings.

5]

52
Table 5.1: Overall Summary Statistics

 

 

INumbetOf Median Cream. Standard Range NumbcrOf . mo:
Homes (pCitL) Mean Deviation (9091) Observations Exceeds EPA
Sampled. (pCifL) (pCiIL)_ ' . Actionlevel
230 3.4 3.3 6.4 0.3 - 48.4 230 41%

 

 

 

 

 

 

 

 

 

A histogram of the overall measurement results is presented in Figure 5.1. Ninety

four of the 230 measurement readings meet or exceed the EPA action level of 4.0 pCi/L.

Figure 5.1: Overall Measurement Results Histogram

 

Number Of Houses

 

Lenawee County Residential Radon Survey Measurement Results

 

.....ula.L._..A.iz....,.

¢®QON¢CDQ
NNNMMMMM

Measurement Results pCi/L

0
lb

48p

”ON ON¢¢D
FINN #Q’V‘I’

 

 

As illustrated in Table 5.1, within the distribution of measurement results, a greater

number of lower values occurred with relatively fewer high readings. Therefore this

distribution is positively skewed. Several statistics utilized in this study require a normally

53
distributed data set, therefore the variable was transformed by calculating the logarithm of

the measurements. Figure 5.1-A presents a histogram of the transformed measurement

variable, that approaches a normal distribution.

Figure 5.1-A: Overall Measurement Results Histogram
Log Linear Transformation

 

 

Lenawee County Residential Radon Survey Measurement Results:
Log Linear Transformation
16-
14»
g 12-
3 l
:0 10~
'5 a- l
a l
.o - l
s 6 ¢
2 4 l l m

2- I l l l}

o . w l J ‘
‘QQ'TQ'T‘QQ'T‘T‘Q‘QfiQfiQQEQfiQQEQfiQ‘Q'fiQ
rO-vO-fiqqqqqoooooPPv-v-v-NNNNNmmmmm

Log Transformed Measurement Results

 

 

Table 5.2 is the only table in this study that uses arithmetic mean and provides a
comparison of overall results of this study (1994), with the results fi'om the Michigan
Indoor Radon Survey (MIRS) (1988-1989) for Lenawee County itself, and for the entire
State of Michigan. Both the median and arithmetic mean measurement values were lower
in this study(3.4; 5:4 pCi/L respectively) than in the MIRS (3.9; 7.8 pCi/L respectively).
The lowest indoor radon levels measured were similar in both surveys (only a 0.2 pCi/L
difference). The highest end of the range produced different values in these two surveys.

The highest measurement value obtained in this survey is 48.4 pCi/L, compared with a

54
reading of 69.7 pCi/L in the MIRS. The percentage of homes that meet or exceed the

EPA action levels of 4.0 pCi/L from this survey is 41% compared with a larger percentage
of 54% in the MIRS.

The statewide results of the MIRS are also presented in Table 5.2. Over two
thousand homes were surveyed state wide, producing the following results: a
measurement result mean of 2.4 pCi/L, a range of measurement results from 0.5 - 162.1
pCi/L (the highest reading was measured in the Upper Peninsula); and 12% of homes

measured state wide were found to be at risk.

Table 5.2: Results Comparison Between Lenawee County Residential Radon Survey
And Michigan Indoor Radon Survey

 

' = Mm I}: 5.532;] ~ l 5399995393333

 

l t

 

 

Lenawee County 230 3.4 5.4 6.4 0.3 - 48.4 41%
Residential
Radon Survey

 

Michigan Indoor 39 3.9 7.8 N/A 0.5 - 69.7 54%
Radon Survey
Lenawee County

 

 

Michigan Indoor 2082 N/A 2.4 N/A 0.5 - 162.1 12%
Radon Survey
Entire State

 

 

 

 

 

 

 

 

Results By Townshi
Sampling occurred in all 22 townships in Lenawee County. Table 5.3 presents
measurement results by township. Median measurement results range from 1.3 pCi/L in

Palmyra Township to 11.9 pCi/L in Tecumseh Township. Geometric mean

55

Table 5.3: Results By Township

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Macon (1) 4 3.3 3.1 2.3 1.4 - 6.7 50%
Clinton (2) 5 3.7 5.1 5.6 3.5 - 16.3 40%
Tecumseh (3) 16 11.9 10.8 9.8 2.0 - 33.5 81%
Franklin (4) 5 3.8 2.9 5.1 0.7 - 13.2 40%
Cambridge (5) 10 4.0 3.2 2.6 1.0 - 8.2 50%
Woodstock (6) 13 5.0 4.0 2.8 1.1 - 10.1 62%
Rollin (7) 4 2.7 2.0 1.9 0.5 - 4.9 25%
Rome (8) 3 2.3 1.8 1.3 0.8 - 3.3 0%
Adrian (9) 61 2.8 2.8 4.3 0.3 - 28.2 34%
Raisin (10) 22 6.9 7.4 10.8 2.4 - 48.4 73%
Ridgeway (11) 5 2.4 2.8 1.9 1.6 - 6.5 20%
Deerfield (12) 7 1.8 1.9 1.4 0.8 - 5.0 14%
Blissfield (13) 20 2.5 1.9 2.2 0.3 - 9.5 25%
Palmyra (14) 6 1.3 1.5 2.6 0.7 - 7.4 17%
Madison (15) 13 3.3 4.2 7.9 0.9 - 23.5 46%
Dover (16) 4 2.2 2.2 1.5 1.1- 4.5 25%
Hudson (17) 6 3.9 3.2 1.2 1.6 - 4.7 50%
Medina (18) 4 3.7 3.8 8.1 0.9 - 18.6 50%
Seneca (l9) 4 2.7 3.2 7.5 0.9 - 16.9 25%
Fairfield (20) 4 1.4 1.0 1.1 0.3 - 2.5 0%
Ogden (21) 7 2.8 2.7 3.8 0.8 - 12.1 14%
Riga (22) 7 2.6 3.1 1.0 2.3 - 4.7 29%
All Townships 230 3.4 3.3 6.4 0.3 - 48.4 41%

 

 

 

 

 

 

 

 

 

56
measurements range from 1.0 pCi/L in Fairfield Township to 10.8 pCi/L in Tecumseh.

Five of the 22 townships have geometric mean measurements which meet or exceed the
EPA action level. The following 5 townships are considered at risk based on geometric
mean measurement readings: Clinton (5.1 pCi/L), Tecumseh (10.8 pCi/L), Woodstock
(4.0 pCi/L), Raisin (7.4 pCi/L), and Madison (4.2 pCi/L). The lowest measurement
reading obtained in this survey is 0.3 pCi/L This measurement value is found within the
lowest range of three townships (Adrian, Blissfield, and Fairfield). The highest
measurement value (48.4 pCi/L) was obtained in Raisin Township. An examination of the
column providing the percentage of homes that meet or exceed the EPA action level,
reveals that in 7 of the 22 townships, half or more of the homes exceed this level. In
Tecumseh Township, 81% of homes sampled exceed the EPA action level. This is the
highest percentage calculated for any township in Lenawee County.

Figure 5.2 shows measurement results by township. The first figure provides the
geometric mean measurement and the second figure provides the sample size, followed by
the range for each of the 22 townships of Lenawee County.

Figure 5.3 shows the distribution of houses at or exceeding the EPA action level
by township. None of the homes measured in Rome and Fairfield Townships have
measurement values of 4.0 pCi/L or greater. The three townships with the highest
percentage of homes with measurement values of 4.0 pCi/L or greater are Tecumseh
(81%), Raisin (73%), and Woodstock (62%). A population corridor runs through the

high percentage townships of Tecumseh and Raisin.

57

 

FIGURE 5.2: MEASUREMENT RESULTS
BY TOWNSHIP

 

 

 

 

 

 

 

 

CLINTON

3‘55",{55; MACON

“Riggs—trig)? CAMBRIDGE FRANKLIN (3-1) [4]

1.1 - 10.1 (3.2) [101 (2.9) [5] 1.4 - 6.7

1.0 - 8.2 0.7 -13.2
hrDGEWAY

(2.8) [5]

ROLLIN ROME ADRIAN RAISIN 1-6 - 6-5
(2.0) [4] (1.8) [3] (2.8) [61] (7.4) [22]
0.5 - 4.9 0.8 - 3.3 0.3 - 28.2 2.4 - 48.4

 

 

 

 

 

PALMYRA
:HUDSON DOVER

 

 

CHARTER 0.7 -7.4 B
1.6-4.7 1.1-4.5 (4.2) “3]
0.9 - 23.5
‘1"
MEDINA SENECA FAIRFIELD OGDEN RIGA
(3.8) [4] (3.2) [4] (1.0) [4] (2.7) 171 (3.1) [7]
0.9 - 18.5 0.9 - 16.9 0.3 - 2.5 0.8 - 12-1 2.3 - 4.7

 

 

 

 

 

 

:1

 

 

1WD“UQSHIP NAUWE

(Geometric Mean pCi/L) [Number of Observations] Six Miles
Range

 

 

 

58

 

 

FIGURE 5.3: DISTRIBUTION OF HOUSES
AT OR EXCEEDING EPA ACTION LEVEL
BY TOWNSHIP

\VWHHHHR ("AMBRIDUIC I'RANKIJN

ADRIAN

HUDSON 1g MADISON
. CHARl‘l-‘R

NH'DINN SENECA

Percent Houses That Meet

‘: :PALMVM

('I IN'H )N
NiAk'UN

ll(l‘\1\lll

RI l )G 13 \VA Y

It \I\IN

DL’L’JU’HLLL)

‘5 B'IJSSFIELD

OGDL’N R l(, l A

 

or Exceed EPA Action Level (4 pCi/L)

- 62-81
- 26-50

14-25

 

0

Six Miles

 

 

59
Research Question Addressed

How do test results from this survey compamfihfipflcfiin—msed upon the
screenings in the MiChiMment of Public Hem

Lenawee County was considered a site for this study based upon predictions made
from a state wide survey conducted by the State of Michigan in 1987. It was estimated
that 40 - 45% of homes in Lenawee County exceed the EPA action level. Findings from
this present study (41% of homes meet or exceed the EPA action level) reveal that the
risk estimates based on the MIRS are accurate in predicting the percentage of at risk
homes in Lenawee County.

Table 5.2 provides a comparison of the results from this study, the Lenawee
County Residential Radon Survey, and the Michigan Indoor Radon Survey conducted by
the state The sample sizes greatly differ between these studies. Two hundred thirty
homes are surveyed in this study compared with 39 in the Michigan Indoor Radon Survey.
The present study has produced a larger, more representative sample with differing results.
The artihmetic mean measurement for this survey is 5.4 pCi/L which is lower than the 7.8
pCi/L mean from the MIRS. The arithmetic means from both studies exceed the EPA
action level. Forty one percent of homes measured in this survey compared with 54% of
homes in the MIRS meet or exceed the EPA action level

The findings of these two surveys, although slightly different, seem to indicate a
higher radon risk potential in Lenawee County than in the rest of the state. This indication

is strengthened when considering the NflRS results from the 2,082 homes surveyed

60
throughout the state the of Michigan -- a 2.4 pCi/L mean, with 12% of homes with

measurements that meet or exceed the EPA action level.

CHAPTER 6: SURVEY RESULTS BY GEOLOGY AND SOILS

Results By Geologic Risk Potential

This chapter will provided results by individual geologic strata (bedrock, surficia]
features, and soil associations) and results by the overall geologic risk codes. Bedrock
measurement results will be presented in two sections: first, bedrock results based on the
original assumed risk potential categories, and secondly results by bedrock formation afier
disaggregation.
Results by Assumed Bedrock Risk Potential

Table 6.1 provides survey results by assumed bedrock Iisk potential. The median
measurement readings for the bedrock strata are as follows: high potential, 2.5 pCi/L;
medium potential 3.7 pCi/L', and low potential 2.5 pCi/L. The highest geometric mean of
the measurement readings came from houses tested on medium potential bedrock with a
reading of 3.6 pCi/L, followed by high potential at 2.3 pCi/L, and low potential at 2.1
pCi/L. The highest individual measurement (48.4 pCi/L) is also found in the medium
potential group. The highest individual reading in the high potential category is 12.!
pCi/L and in the low potential category 5.0 pCi/L. The highest percentage of sampled
houses that meet or exceed the EPA action level were results from medium potential
bedrock with 45%. Eighteen percent and 14% of homes tested in the high and low

bedrock potential areas respectively, meet or exceed the EPA action level.

61

62
Table 6.1: Results By Assumed Bedrock Risk Potential

 

 

‘-_._'Ir

 

0.3 - 12.1

   

 

0.3 - 48.4 45%

 

 

 

 

 

 

 

 

 

 

0.8 - 5.0 14%

 

Bedrock data in Table 6.1 disproves the survey working assumptions that the
highest measurement readings would occur in bedrock areas classified as highest risk
potential. The highest overall readings actually occur in the bedrock groups identified as
assumed medium risk. Following analysis of the measurement results from assumed
bedrock potential risk categories, these groupings were disaggregated, further analyzed,
and empirically reclassified by bedrock formation. These measurement results by bedrock
formation and reclassification will be presented later in this chapter.

Results by Assumed Surficfl Features Risk Potential

The next geologic strata presented is surficial formations. Table 6.2 provides
measurement result data on surficial features. In the surficial features strata results, the
high potential category has the highest results in median (4.0 pCi/L), geometric mean (4.]
pCi/L) and percentage of homes that meet or exceed the EPA action level (51%) The low
potential risk category has the lowest results in these same categories: median, 2.5 pCi/L;

geometric mean 2.6 pCi/L; and 26% of homes meet or exceed the EPA action level.

63
Elevated readings were found in all three surficial feature potential risk categories: 33.5

pCi/L in high potential; 48.4 pCi/L in medium potential, and 30.5 pCi/L in low potential.

Table 6.2: Results By Assumed Surficial Features Risk Potential

 

 
  
 

Fmrwm w , ,
We“

...h I

 

High (3) 124 4.0 4.1 6.5 0.3 - 33.5 51%

 

Medium (2) 53 2.7 2.7 7.3 0.3 - 48.4 32%

 

 

Low (1) 53 2.5 2.6 4.3 0.3 - 30.5 26%

 

 

 

 

 

 

 

 

Results by Assumed Soil Asscmtion Risk Potential

Table 6.3 provides the measurement results for soil association risk codes which is
the third and final geologic strata. As with the surficial feature results, the highest median
(5.2 pCi/L) and geometric mean (5.6 pCi/L ) results in the soil association potential risk
group also occur in the assumed high risk group. Equally the lowest median (2.6 pCi/L)
and geometric mean (2.6 pCi/L) measurement values occur in the assumed low potential
risk group. Elevated measurement levels occurred in all three potential risk groups, the
highest value of 48.4 pCi/L was measured in a home in a medium risk potential area.
Sixty five percent of homes surveyed in the soil association high potential risk category
have measurement results that meet or exceed the EPA action level, followed by 32% of

homes in medium risk potential, and 29% of homes in the low risk potential.

64
Table 6.3: Results By Assumed Soil Association Risk Potential

 

 

 

71 5.2 5.6 7.0 0.5 - 33.5 65%

 

 

 

 

 

 

 

 

 

Medium (2) 66 2.4 2.7 7.7 0.3 - 48.4 32%
Low (1) 93 2.6 2.6 4.0 0.3 - 27.1 29%

 

 

Results by Overall Geologic Risk Potentiacl Codes

The three geologic potential risk strata (bedrock, surficial features , and soil
associations) were combined to produce overall geologic risk potential codes. The results
of these three digit overall geologic risk potential codes are presented in Table 6.4. The
first digit of the code represents assumed bedrock risk potential; the second digit assumed
surficial features risk potential, and the third digit, assumed soil association risk potential.
The number 3 is used to indicate high, 2 to indicate medium, and 1 to indicate low
assumed potential to emit or transmit radon. Risk codes 213 and 233 produced the
highest median measurement readings with 5.7 pCi/L and 5.3 pCi/L respectively. The
highest geometric mean measurement readings also occur in risk codes 213 (5.3 pCi/L)
and 233 (5.9 pCi/L). The lowest median (1.4 pCi/L) occurs in the 331 risk potential
group, and the lowest geometric mean (1.1 pCi/L) in the 322 risk potential group. These
figures are based on one and eight observations respectively. The highest geometric mean

measurement readings appear in areas of assumed medium potential bedrock (first digit)

65
Table 6.4: Results By Overall Geologic Risk Potential Codes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

33.1. . .. ., . l I .. . 1'4. . . . 1.4 . . '1‘; .. .. 14-14 . ..Oo./.o
322 8 1.5 1.1 0.9 0.3 - 2.7 0%
321 5 3.1 3.0 2.0 1.6 - 6.7 20%
312 4 2.5 2.5 0.3 2.2 - 3.0 0%
311 10 3.3 3.9 3.2 0.8 - 12.1 40%
233 56 5.3 5.9 6.9 0.5 - 33.5 68%
232 24 3.6 3.7 6.9 0.6 - 28.2 42%
231 43 2.9 2.8 5.2 0.3 - 27.1 33%
223 10 3.8 4.4 3.0 2.0 - 10.3 40%
222 23 2.6 2.9 10.6 0.4 - 48.4 39%
221 3 4.6 4.7 3.6 2.4 - 9.5 67%
213 5 5.7 5.3 11.9 0.8 - 30.5 80%
212 3 2.4 2.0 1.7 0.8 - 4.2 33%
211 28 2.3 2.1 1.6 0.3 - 7.6 21%
122 4 1.6 1.7 1.9 0.8 - 5.0 25%
111 3 2.6 2.6 0.1 2.5 - 2.6 0%

All Codes 230 3.4 3.3 6.4 0.3 - 48.4 41%

 

 

Note: key to risk codes: first digit represents bedrock risk potential; second digit, surficial
feature risk potential; and third digit soil association risk potential. 3 indicates high, 2
medium, and 1 low assumed potential for radon emission or transmission

66
and the lowest geometric mean readings appear in the areas of assumed high potential risk.

The highest measurement reading (48.4 pCi/L) was recorded in an area of 222 overall
geologic risk potential. The greatest percentages of houses that meet or exceed the EPA
action levels were measured in the following risk potential areas : 213 (80%); 233 (68%);
and 221 (67%). There are four geologic risk potential areas (331, 322, 312, and 111)
where none of the homes measured meet or exceed the EPA action level.

Results by Bedrock Formation After Disaggregation

As previously detailed in this chapter, the measurement results for the assumed
bedrock risk potential categories did not follow the expected pattern. Measurement
results from both the assumed surficial geology and soil association risk categories
produced the highest geometric mean readings in the “high” potential risk category and
the lowest geometric mean readings in the “low” risk category (see Tables 6.2 and 6.3).
Measurement results from the assumed bedrock risk potential categories revealed the
highest geometric mean in the “medium” risk category (see Table 6.1).

These bedrock results called for a reassessment of the original working
assumptions and risk potential classifications. Based on actual field findings, the original
working assumption and classification of the high potential risk category is incorrect. The
highest geometric mean was produced from measurements in the medium potential risk
category and not the high potential risk category. This finding is one of the original
intents of this research -- to measure and locate geographic variations in indoor radon

levels.

67
To firrther explore and examine these actual site measurement readings, the

assumed bedrock potential risk categories were disaggregated and re-assessed. Table 6.5
reveals the original assumed bedrock risk categories with disaggregated measurement
results. The assumed high risk potential category is made up of Antrim Shale and Sunbury
Shale. Neither of the geometric mean measurement results from these two groups (3.2
pCi/L and 1.9 pCi/L respectively) meet or exceed the EPA action level of risk (4.0 pCi/L).
The assumed medium risk potential category is made up of Coldwater Shale, Marshall
Sandstone, Berea Sandstone, and Bedford Shale. Geometric mean measurement reading
results from two of these four bedrock formations exceed the EPA action level of radon
risk (Marshall Sandstone 4.5 pCi/L; Berea Sandstone 4.1 pCi/L). Finally, the assumed
low risk potential category is made up of The Detroit River Group, Dundee Limestone,
and the Traverse Group. Geometric mean measurement results fiom these three bedrock
groups do not meet or exceed the EPA action level.

Further analysis of the nine formations of bedrock after disaggregation produced a
ranking by geometric mean from highest to lowest. These results are presented in Table
6.6. The highest geometric mean measurement result is 4.5 pCi/L for Marshall Sandstone
which is based on 10 on site measurements. Seventy percent (7/10) of these measurement
results meet or exceed the EPA action level. The lowest geometric mean among the nine
bedrock groups is found in Bedford Shale (1.5 pCi/l). No measurement results in this
group meets or exceeds the EPA Action level. From this ranking, bedrock reclassification

occurred.

68
Table 6.5: Original Assumptions Of Bedrock Risk Potential

As Against Measured Results

 

 

 

 

 

 

Antrim Shale 10 3.3 3.2 1.5 1.5 - 6.5 (3/10) 30%
Sunbury Shale 18 2.2 1.9 2.7 0.3 - 12.1 (2/18) 11%
(3) High 28 2.5 2.3 2.4 0.3 - 12.1 (5/28) 18%

Potential Totals

 

 

 

 

 

 

 

 

 

 

 

Coldwater Shale 172 3.7 3.6 7.12 0.3 - 48.4 (76/172) 44%
Marshall 10 5.1 4.5 2.7 1.0 - 10.1 (7/10) 70%
Sandstone

Berea Sandstone 7 4.5 4.1 2.5 1.1 - 9.5 (5/7) 71%

Bedford Shale 6 2.4 1.5 1.1 0.5 - 3.4 (0/0) 0
(2) Medium 195 3.7 3.6 6.8 0.3 - 48.4 (88/195) 45%
Potential Totals
Detroit River 2 3.2 2.6 2.6 1.3 - 5.0 (1/2) 50%
Group
Dundee 1 2.6 2.6 - 2.6 (0/0) 0
Limestone
Traverse Group 4 2.2 1.8 0.8 0.8 - 2.6 (0/0) 0
(1) Low Potential 7 2.5 2.1 1.4 0.8 - 5.0 (1/7) 14%
Totals

 

 

Overall Totals 230 3.4 3.3 6.4 0.3 - 48.4 (94/230) 41%

 

 

 

 

 

 

 

 

 

69

Table 6.6: Measured Results By Bedrock Formation Ranked By Mean

 

  
    

4 ' ’1 '.“‘V' '.' .'.' '.'."‘. '.". ‘1' i' r747 '. . .'.'
4 l I i." .‘ ,'.‘-'»’.".'- ‘1' i '.
. .11 ....l . i . . agyau

EPA:

 

    

Sandstone ( 1)

Marshall

10

5.1

(7/10) 70%

 

Berea Sandstone
(4)

4.5

4.1

2.5

(5/7) 71%

 

Coldwater Shale
( 2)

172

3.7

3.6

7.12

0.3 - 48.4

(76/172) 44%

 

Antrim Shale (6)

10

3.3

3.2

1.5

1.5 -6.5

(3/ 10) 30%

 

Detroit River
Group (9)

ix)

3.2

2.6

2.6

1.3 - 5.0

(1/2) 50%

 

Dundee
Limestone (8)

2.6

2.6

2.6

(0/0) 0

 

Sunbury Shale
( 3)

18

2.2

1.9

2.7

0.3 -12.1

(2/18) 11%

 

Traverse Group
( 7)

2.2

1.8

0.8

0.8 - 2.6

(0/0) 0

 

Bedford Shale
(5)

2.4

1.1

0.5 - 3.4

(0/0) 0

 

 

Overall Totals

 

230

 

3.4

 

3.3

 

6.4

 

 

0.3 - 48.4

(94/230) 41 %

 

Figure 6.1 presents the locations of disaggregated bedrock within Lenawee County

by formations. There are nine bedrock groups each assigned an individual identification

number. Lenawee County is dominated by the occurrence of Coldwater Shale (ID

Number 2).

 

70

 

 

FIGURE 6.1: DISAGGREGATED BEDROCK F ORMATIONS
WITH MEASUREMENT RESULTS

 

 

 

 

..a

mfl

I

 

‘

v

H
kw

L

 

 

 

 

 

..r. ”271.374
ID Bedrock Formation GeoMean (pCi/L) Median (pCi/L) N
five miles

1 Marshall Sandstone 4.5 5. 1 10

2 Coldwater Shale 3.6 3.7 172

3 Sunbury Shale 1 .9 2.2 18

4 Berea Sandstone 4.1 4.5 7

5 Bedford Shale 1.5 2.4 6

6 Antrim Shale 3.2 3.3 10

7 Traverse Group 1.8 2.2 4

8 Dundee Limestone 2.6 2.6 1

9 Detriot River Group 2.6 3.2 2

 

 

71
Based upon ranking of geometric mean measurement results from field findings,

bedrock risk groups were reclassified. Table 6.7 presents empirically derived bedrock risk
categories based upon measured results. The nine disaggregated bedrock groups have
been classified into two risk groups based upon geometric mean. The high risk group is
comprised of three bedrock groups with the highest geometric mean measurement results,
and the lower risk category is comprised of six bedrock groups with the lowest geometric
mean measurement results. The combined geometric mean of the empirically derived high
risk bedrock category is 3.7 pCi/L, and a median of 3.8 pCi/L, with 47% (88/ 189) of
measurement results which meet or exceed the EPA action level. The combined geometric
mean of the empirically derived lower risk bedrock category is 2.1 pCi/L, and a median of
2.4 pCi/L, with 15% (6/41) of measurement results which meet or exceed the EPA action
level While these bedrock measurement results are mediated by surficial features, soils
and housing structural characteristics, this reclassification of bedrock based upon field
findings, provides a better characterization of potential radon risk in Lenawee County.
Research Questions Addressed
What are the spatial patterns of risk to radon associated with ggilggic vflation in
L§n_awee County? i.e. where are high_and low momma] riskfl

It is known that the distribution of radon in any given geographic area is not
uniform. Identifying potential high risk areas is the first step in minimizing the potential
negative impacts of radon exposure to home residents. If public awareness in these “hot

spot” areas of potential high risk is increased, residents may be more willing to test homes.

72
Table 6.7: Empirically Derived Bedrock Risk Categories

Based On Measured Results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Marshall 10 5.1 4.5 2.7 1.0 - 10.1 (7/10) 70%
Sandstone (1)
Berea Sandstone 7 4.5 4.1 2.5 1.1 - 9.5 (5/7) 71%
(4)
Coldwater Shale 172 3.7 3.6 7.12 0.3 - 48.4 (76/172) 44%
(2)
High Risk 189 3.8 3.7 6.8 0.3 - 48.4 (88/ 189) 47%
Totals
Antrim Shale (6) 10 3.3 3.2 1.5 1.5 - 6.5 (3/10) 30%
Detroit River 2 3.2 2.6 2.6 1.3 - 5.0 (1/2) 50%
Group (9)
Dundee 1 2.6 2.6 - 2.6 (0/0) 0
Limestone (8)
Sunbury Shale 18 2.2 1.9 2.7 0.3 -12.1 (2/18) 11%
(3)
Traverse Group 4 2.2 1.8 0.8 0.8 - 2.6 (0/0) 0
(7)
Bedford Shale 6 2.4 1.5 1.1 0.5 - 3.4 (0/0) 0
(5)
Lower Risk 41 2.4 2.1 2.1 0.3 - 12.1 (6/41) 15%
Totals
Overall Totals 230 3.4 3.3 6.4 0.3 - 48.4 (94/230) 41%

 

 

73
If elevated levels of radon are present, the proper steps can be taken to reduce the level of

radon and reduce risk of exposure.

Areas of assumed high and low potential risk are represented in Figure 4.7: Overall
Geologic Radon Risk Potential Codes. These codes identify the assumed potential radon
risk of an area based on the original classification of bedrock, surficial features, and soil
associations. Considering only geologic features in assessing risk, the highest readings
would be expected in homes located on the assumed highest potential geology. The
highest potential bedrock in Lenawee County is classified with a risk code of 3, followed
by medium potential bedrock as 2; and finally low potential bedrock as l. The majority of
the county rests on what was originally classified as medium potential bedrock.

Table 6.1 provides results by assumed bedrock risk potential. The highest
measurement result geometric mean occurs in the medium potential bedrock category.
Based upon the radon literature, it was expected that the highest measurement results
would occur in areas of assumed high potential bedrock. These results are not as
expected. The geometric mean measurement result from homes on assumed high potential
bedrock is 2.3 pCi/L; while the mean measurement result from homes on assumed medium
potential bedrock is 3.6 pCi/L. These measurement outcomes are influenced by the
working assumptions of the bedrock classification system (i.e. where the categories were
drawn separating assumed medium potential risk from assumed high potential risk).

In the field, the working assumption of potential geologic risk was tested. Data
and analysis indicate that the assumptions were not correct. The highest measured results

occurred in the assumed medium risk potential category and not the high risk category as

74
expected. Bedrock in the study area was disaggregated and reassessed based upon actual

measurement readings and empirically reclassified. This reclassification resulted in the
empirically derived risk categories presented in Table 6.7. Three groups of bedrock with
the highest geometric mean measurement readings, have been reclassified into a high risk
category. The remaining six bedrock groups, with the lower geometric mean
measurement readings, have been reclassified into a lower risk category.

The location of these nine bedrock groups is indicated in Figure 6.1. This figure
with the accompanying geometric mean and median measurement readings reveals
potential radon risk based on bedrock in Lenawee County. The occurrence of Coldwater
Shale (ID number 2) dominates the bedrock of the county, occupying approximately two-
thirds of the area. Measurement results fiom the 172 houses tested on Coldwater Shale
produce a geometric mean of 3.6 pCi/L, the third highest of all nine bedrock groups, and
44% (76/172) measurements which meet or exceed the EPA action level. Other areas
classified as high risk are those of Marshall Sandstone (ID number 1; geometric mean 4.5
pCi/L, 70% [7/10] meet or exceed EPA action level) and Berea Sandstone (ID number 4;
mean 4.1 pCi/L, 71% [5/7] meet or exceed EPA action level). Residents of these three
areas could be at higher risk based upon these measurement results. The lowest areas of
potential risk based upon field findings are those of Bedford Shale (ID number 5) and the
Traverse Group (ID number 7), with geometric means of 1.5 pCi/L and 1.8 pCi/L
respectively with all measurement readings below the EPA action level.

The bedrock measurement results are mediated by and also suggest the importance

of surficial features and soil associations as radon filtering layers. These surficial features

75
and soil association strata will be examined and discussed in reference to another research

question.

Figure 5.2 provides measurement results by township in Lenawee County. Five of
the 22 townships have average geometric mean measurement readings above the EPA
action level of 4.0 pCi/L. All five of these townships are dominated either by Coldwater
Shale (ID number 2) or by Marshall Sandstone (II) number 1). These readings indicate
that residents in these five townships may be at a higher risk to indoor radon exposure. A
spatial pattern of risk emerges from the location of these five townships on the map. A
“crescent” of potential higher risk can be observed from the south west corner of the
county beginning in Medina, moving east to neighboring Seneca, then north east to
Madison, Adrian, and Raisin, heading north to Tecumseh and Clinton, followed by
movement east to Franklin, Cambridge, and finally Woodstock to complete the crescent
pattern.

Measurement results by township are presented in Table 5.3. The four townships
with the highest geometric mean measurement readings are: Tecumseh (10.8 pCi/L),
Raisin (7 .4 pCi/L), Clinton (5.1 pCi/L) and Madison (4.2 pCi/L). A large population
corridor extends from the city of Adrian, in both Adrian and Madison Townships, through
Raisin Township, into Tecumseh and Clinton Townships. Measurement results suggest
that this population corridor or cluster falls in an area of potential high radon risk. Many
of the highest measurement readings came from homes tested within these townships

(Adrian 28.2 pCi/L; Madison 23.5 pCi/L; Raisin 48.4 pCi/L; Tecumseh 33.5 pCi/L; and

76
Clinton 16.3 pCi/L). The elevated measurements and geometric mean values suggest this

area to be a “hot spot” of radon risk.

Seventeen of the 22 townships of Lenawee County have geometric mean
measurement results below the EPA action level. These lower risk townships are located
in the extreme east and south east of the county, and in the central western portion. The

township with the lowest measurement result geometric mean is Fairfield (1.0 pCi/L).

In areas of similar risk potential based upon bedrock, to what deggee do surficial features

and soil association impact the movement of Ladon?

Very often radon risk is determined by considering only bedrock in the geologic
risk. Of primary importance to this study is the impact of surficial geology and soil
associations as sub-surface filtering layers. Surficial features and soil associations are
classified into the same assumed risk categories as bedrock: high (3), medium (2), and low
(1). As with bedrock, it was also expected with these two strata that higher measurement
readings would occur in the high risk potential categories.

Chi squared analysis was conducted on each of the three geologic strata of
bedrock, surficial features, and soil associations. The chi square test of bedrock potential
risk and measurement results produced a Stuart Tau-C value of -0.080 which suggests a
relatively weak negative association between bedrock potential risk and the measurement
values. The probability value of 0.008 indicates that the association is significant.

Table 6.8 provides the frequency distributions of assumed bedrock risk potential

by measurement result. A violation of a chi-squared assumption exists in this table since

77
there are fewer than the desired five observations present in one cell. Due to the minimal

occurrence of low potential bedrock that occurs in the county, location of houses, and no
advance knowledge of actual measurement values, it is difficult to meet the minimum

requirements in all cells.

Table 6.8: Frequencies - Bedrock Risk Potential By Measurement Result

 

 

 

 

 

 

High (3) 23 5 28
Medium (2) 107 88 195
Low (1) 6 l 7

Total 136 94 230

 

 

 

 

 

 

Based on the radon literature it is expected that a greater percentage of
measurement results above 4.0 pCi/L would occur in the assumed high potential bedrock.
Of the 28 homes measured on high potential bedrock, only five have measurement results
that meet or exceed the EPA action level. The measurement results from bedrock
contradict the working assumptions of this study. As previously stated, these outcomes
could be based upon the placement of bedrock into the risk categories and also point to
the significance of surficial features and soil associations as filtering layers.

The measurement results by surficial features are presented in Table 6.2 and soil
association measurement results are presented in Table 6.3. Unlike the bedrock

measurement results, the measurement results from both of these strata follow an expected

78
pattern: the assumed high potential risk categories produced the highest median,

geometric mean, and percentage of homes that meet or exceed the EPA action level;
equally, the assumed low potential risk categories produced the lowest median (with one
exception, the low potential risk soil association median is 0.2 pCi/L greater than the
medium risk potential category), geometric mean, and percent of homes that meet or
exceed the EPA action level. i

The chi-squared test of surficial features and measurement results produced a
Stuart Tau-C value of -0.027 which suggests a relatively weak negative association
between surficial features potential risk and measurement results. The probability value of
0.008 indicates that the association is significant.

Table 6.9 presents the frequency distributions of surficial feature risk potential by
measurement result. Half of the houses (62 of 124) measured on assumed high potential
surficial features have measurement readings of 4.0 pCi/L or greater while only 28% (15
of 53) houses measured on low potential surficial features have measurement readings of
4.0 pCi/L or greater. These results reflect the expected pattern of a greater percentage of
houses in the assumed high potential risk category with measurement results that meet or
exceed the EPA action level, and a lower percentage of houses in the assumed low
potential risk category with measurement results that meet or exceed the EPA action level.

The chi-squared test of soil association risk potentials by measurement result
produced a Stuart Tau-C value of 0.282 which suggests a positive association between
soil association potential risk and measurement results. The probability value of 0.000

indicates that the association is highly significant.

Table 6.9: Frequencies -- Surficial Feature Risk Potential By Measurement

79

Result

 

 

 

 

 

 

Total 136

 

 

94

 

 

230

 

Table 6.10 presents the frequency distributions of soil association risk potential by
measurement result. Sixty five percent (46 of 71) of the measurement results from houses
in the soil association assumed high potential risk category have readings that meet or
exceed the EPA action level, while only 29% (27 of 93) of the measurement results from
houses in the soil association assumed low potential risk category have readings that meet
or exceed the EPA action level. These results like those of the surficial features, also
reflect the expected pattern of a greater percentage of houses in the assumed high
potential risk category with measurement results that meet or exceed the EPA action level,

and a lower percentage of houses in the assumed low potential risk category with

measurement results that meet or exceed the EPA action level.

 

80
Table 6.10: Frequencies - Soil Association Risk Potential By Measurement Result

 

 

 

 

 

 

Total 136 94 230

 

 

 

 

 

 

Table 6.4 presents measurement results by overall geologic risk potential codes.
These codes represent a combination of the individual strata of bedrock, surficial features,
and soil association designed to indicate the overall geologic potential radon risk.

Results suggest that people living in areas originally classified as assumed medium
potential bedrock are at an increased risk. Of the 16 geologic risk potential categories
where sampling occurs, 4 have geometric mean measurement results above the EPA
action level. All four of these geologic potential risk categories occur within medium
potential bedrock.

The highest geometric mean (5.9 pCi/L) occurs in the 233 geologic risk potential
category. This combination of assumed medium potential bedrock, high potential surficial
features, and high'potential soil associations, not only produced the highest geometric
mean measurement, but also has 68% of the 56 homes tested which meet or exceed the

EPA action level.

81
Sampling occurs in three geologic risk potential codes with high potential soil

association (233, 223, and 213). All three of these categories have geometric means
above the EPA action level of 4.0 pCi/L (233: 5.9 pCi/L; 223: 4.4 pCi/L; and 213: 5.3
pCi/L). These results suggest that assumed medium potential bedrock (which includes
Coldwater Shale, Marshall Sandstone, and Berea Sandstone) combined with assumed high
potential soil associations, produce higher overall indoor radon measurements.

While the three lowest overall geologic potential risk codes (211, 122, and 111)
produced low geometric mean readings of 2.6 pCi/L or less, two overall potential risk
codes with high potential bedrock (331 and 322) produced the lowest geometric means
14 pCi/L and 1.1 pCi/L respectively. Results based upon the geologic risk potential
codes suggest that the classification system produces more expected results in the codes
with medium and low potential bedrock. These measurement results also suggest that

surficial features and soil associations do effect overall domestic radon risk.

CHAPTER 7: SURVEY RESULTS BY HOUSING CHARACTERISTICS

Results by Housing Structural Characteristics

Table 7.1 provides survey results based on the location of the measurement device
within a household. Following EPA guidelines and measurement protocol, measurements
were taken in the lowest livable level of a home. If a basement was present the reading
was taken there. If a home did not have a basement, the measurement was obtained fi'om
the first floor. Of the 230 homes sampled in this survey, 202 are basement measurement
readings and 28 are first floor measurement readings. Basement measurement readings
are higher in median (3.7 pCi/L) and geometric mean (3.8 pCi/L) when compared with
first floor measurement of median (1 .3 pCi/L) and geometric mean (1.3 pCi/L). While low

readings of 0.3 can be found in measurement taken in the basement and on the first floor,

Table 7.1: Results By Location 01' Measurement Device

 

‘ j ‘; msmmmgc. armor)?

 

   

 

 

 

***** (pea; Am I... %
Basement .202 3.7 3.8 6.6 0.3 - 48.4 45%
First Floor 28 1.3 1.3 1.9 0.3 -8.8 14%

 

 

 

 

 

 

 

 

the highest measurement obtained in a basement is 48.4 pCi/L, while the highest

measurement collected fiom a first floor measurement is 8.8 pCi/L. Forty-five percent of

82

83
the measurement results taken in homes with basements meet or exceed the EPA action

level, while 14% of the measurement readings obtained on the first floor of homes meet or
exceed this criteria.
Measurement results by age of housing structure are presented in Table 7.2. These

results are presented in five categories. The age of housing tested ranged from one year to

180 years since time of construction. The lowest median measurement is 1.6 pCi/L and
occurs in the 50 -74 year age of structure category; while the highest median value is 4.0
pCi/L and occurs in the 25 - 49 year age of structure category. Geometric mean
measurement results range fiom 2.1 pCi/L in the 50 - 74 year age of structure category, to
4.4 pCi/L in the 25 - 49 year age of structure category. Measurement readings of 0.6
pCi/L or less occur in each category, while the highest measurement reading of 48.4 pCi/L
was obtained in the 1 - 24 year age of housing structure category. The lowest percentage
of any group that meets or exceeds the EPA action level is 12% in the 50 - 74 year age of
housing structure category, while the highest percentage of any group is 53% in the 25 -

49 year age of housing structure category.

Table 7.2: Results By Age Of Housing Structure

 

 
  
 

 

 

 

 

ya“
125 - 180 22 3.2 3.5 3.4 0.3 - 13.8 41%
75 -124 59 3.0 2.9 3.1 0.5 - 13.2 39°/o
50 - 74 17 1.6 2.1 6.5 0.6 - 28.2 12%
25 - 49 61 4.0 4.4 7.2 0.3 - 30.8 53%

 

l - 24 71 3.3 3.2 8.0 0.3 - 48.4 39%

 

 

 

 

 

 

 

 

 

84

The next set of figures presented in Table 7.3 combine the data from the previous

two tables to view results by age of housing structure and location of measurement device.

Table 7.3: Results By Age Of Housing Structure

And Location Of Measurement Device

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AgeOf LocationOf ' No. Of * Range ' Exceeds
Structure Measurement . Obvs. . i (PG/L} ..EPASF
In Years Device ' ' - . ' Action"

Level '

125 - 180 basement 20 3.4 4.2 3.3 1.8 - 13.8 45%
first floor 2 1.0 0.7 0.9 0.3 - 1.6 0%

overall 22 3.2 3.5 3.4 0.3 - 13.8 41%

75 - 124 basement 54 3.5 3.2 3.1 0.5 - 13.2 43%
first floor 5 1.0 1.2 1.0 0.5 - 2.8 0%

overall 59 3.0 2.9 3.1 0.5 - 13.2 39%

50 - 74 basement 17 1.6 2.1 6.5 0.6 - 28.2 12%
first floor 0 0 0 0 0 0%

overall 17 1.6 2.1 6.5 0.6 - 28.2 12%

25 - 49 basement 52 4.4 5.0 7.5 0.9 - 30.8 56%
first floor 9 2.4 2.0 2.7 0.3 - 8.8 33%

overall 61 4.0 4.4 7.2 0.3 - 30.8 53%

l - 24 basement 59 3.6 4.0 8.5 0.3 - 48.4 46%
first floor 12 1.1 1.2 1.3 0.3 -4.6 8%

overall 71 3.3 3.2 8.0 0.3 - 48.4 39%

 

 

For each of the five housing age categories, data are listed by location of measurement

device (basement or first floor) and by overall or combined results for each of the

categories. In each of the five categories, the median and geometric mean measurement

results are higher in the basement than on the first floor. The highest individual

measurement readings in each of the five housing age categories occurred in basement

 

85
measurement readings. The percentage of homes in each category that meet or exceed the

EPA action level is greater in measurements taken in basements. The highest percentage
of houses that meet or exceed the EPA action level from first floor measurement is 33% in
the 25 - 49 housing age category.

Table 7.4 presents results by board of review housing assessment values. This
figure is an overall assessment of the worth of a property and is given in fifty percent
values. Board of review housing assessment values are divided into five categories. The
highest median measurement reading is 4.5 pCi/L in the $30,001 - $45,000 category; while
the lowest median measurement value of 2.3 pCi/L is in the $15,001 - $30,000 category.
The two highest geometric mean values of 4.6 pCi/L and 3.6 pCi/L occur in the categories

of $30,001 - $45,000, and $45,001 - $60,000 respectively.

Table 7.4: Results By Board of Review Housing Assessment Values

Assessment values are fifty percent values in $US.

 

. i . '.‘ .'.. ~ ' ' v . - p ‘ ‘ ‘ y r . V . V " I V. r r. . . . . . . V .
‘ . , , . . .. '. ‘ . ' ‘ . ... . ‘ ~ . .. I 3“ ‘7 . '. . ‘ .. . ' .' .'.' Vi.‘ ‘ , . , . '
. .. q . ~ ‘ ,‘ .‘,‘ '.

 

 

 

60.001 - 132.500 32 2.7

 

45,001 - 60,000 30 3.5 3.6 10.0 0.5 - 48.4 43%

 

30.001 - 45,000 83 4.5 4.6 6.6 0.3 - 30.8 54%

 

15,001 - 30,000 64 2.3 2.5 2.9 0.3 - 13.2 31%

 

 

 

 

 

 

 

 

 

5,600 - 15,000 21 2.8 2.1 2.2 0.3 - 9.2 29%

 

86
Measurement values of 0.5 pCi/L or less were recorded in all categories, while readings of

30 pCi/L or more occur in three categories ($60,001 - $132,500; $45,001 - $60,000; and
$30,001 - $45,000). The greatest percent of homes measured that meet or exceed the
EPA action level (54%) occur in the $30,001 - $45,000 category, while the lowest (29%)
occurs in the 5,600 - 15,000 category. This data seems to indicate that higher income
families with more expensive housing are not at a lower risk to radon.

Table 7.5 presents results by presence of sump pump. In homes with sump pumps
the measurement results in both median (3.7 pCi/L)and geometric mean (3.8 pCi/L)
values are higher than those home without sump pumps (3.1 pCi/L; and 2.9 pCi/L,
respectively). Measurement readings ranging from 0.3 pCi/L to 30.8 pCi/L or greater
occur in both categories. The highest measurement reading of 48.4 pCi/L occurs in a
home with a sump pump present. Forty six percent of homes measured with sump pumps

meet or exceed the EPA action level compared with 36% in homes without sump pumps.

Table 7.5: Results By Presence Of Sump Pump

 

 

 

 

 

 

3mm (pom : Mm e " Wu 4 Exceeds - EPA
- _;., _- : -. aw) " " ”1°“ Lev“

Sump Present 112 3.7 3.8 7.5 0.3 - 48.4 46%

Sump Absent 118 3.1 2.9 5.0 0.3 - 30.8 36%

 

 

 

 

 

 

 

 

Results by floor material type are presented in Table 7.6. There are seven
categories of material type, including a “mix” category for homes with more than one

floor material. The highest median (9.4 pCi/L) and geometric mean (8.3 pCi/L)

 

87
measurement readings occur in homes with stone as the floor material. Only two homes

measured have stone floors. The second highest median (6.5 pCi/L) and geometric mean
(5.8 pCi/L) occur in homes with dirt or earth floors. The lowest median (1.4 pCi/L) and
geometric mean (1.4 pCi/L) measurement results occur in homes with wood floors. These
are typically homes without basements thus the measurement readings were taken on the
first floor. The highest measurement result of 48.4 pCi/L occurred in a home with a
poured basement. In the same category of poured floor material, measurement results as
low as 0.3 pCi/L also occur. One hundred percent of homes measured with stone floors
(2 of 2) meet or exceed the EPA action level; 57% of earth floor homes; and 44% of

poured floor homes.

Table 7.6: Results By Floor Material

 

’ ,Floor . ‘ : Standaxdkange . MeetsOt‘
- Material .. 13mm {90113 3‘: 5W3“
H (1’5“) I? VIC} 7 - ‘ ”mm

 

 

Stone 2 9.4 8.3 6.2 5.0 - 13.8 100%

 

Poured 192 3.6 3.6 6.8 0.3 - 48.4 44%

 

Block 1 1.8 1.8 - 1.8 0%

 

Clay Brick 1 3.9 3.9 - 3.9 0%

 

Wood 20 1.4 1.4 1.4 0.3 - 5.1 15%

 

Mix 7 3.3 2.5 1.2 0.8 -4.5 14%

 

Earth 7 6.5 5.8 4.6 2.3 -13.2 57%

 

 

 

 

 

 

 

 

88

Table 7.7 presents results by presence of floor cracks. The median measurement
value is 3.6 pCi/L for homes with floor cracks and 2.9 pCi/L for homes measured without
floor cracks. The geometric mean measurement result is 3.7 pCi/L for houses with floor
cracks and 2.8 pCi/L for houses measured without floor cracks. Low and high
measurement results were obtained from houses with and without floor cracks. Forty two
percent of houses with floor cracks present and 39% percent of houses without floor

cracks meet or exceed the EPA action level.

Table 7.7: Results By Presence Of Floor Cracks

 

A; 5......   - chm mam)

 

 

Cracks Present 142 3.6

 

 

Cracks Absent 88 2.9 2.8 5.4 0.3 - 30.8 39%

 

 

 

 

 

 

 

 

The next set of data presented in this chapter are results by wall material type,
presented in Table 7.8. There are six wall material categories including a category called
“mix”. Homes with a mixture of wall material are placed in this category. The highest
median (4.7 pCi/L) and geometric mean (4.8 pCi/L) measurement readings occur in homes
with poured walls. The second highest median (4.4 pCi/L) and geometric mean (4.7
pCi/L) measurement readings occur in homes with stone walls. The lowest median (1.5
pCi/L) and geometric mean (1.7 pCi/L) measurement values occur in homes with wood
walls. Homes in this category would typically not have basements, therefore the

measurement reading would be from the first floor. Measurement results of 1.3 pCi/L or

89
lower occur in all categories. The highest recorded measurement of 48.4 pCi/L occurred

in a home with poured walls. Fifty nine percent of homes measured with stone walls meet

or exceed the EPA action level; 56% of homes with block walls; and 25% with walls of

wood.

Table 7.8: Results By Wall Material

 

 

 

 

 

 

 

 

Stone 32 4.4 4.7 3.4 1.7 -13.8 59%
Poured 45 4.7 4.8 10.1 0.3 - 48.4 56%
Block 111 3.4 3.2 5.5 0.3 - 30.8 35%
Clay Brick 10 2.8 3.0 1.3 . 1.3 - 5.5 30%
Wood 28 1.5 1.7 3.7 0.3 -18.6 25%
Mix 4 3.1 2.4 1.7 0.8 -4.7 25%

 

 

 

 

 

 

 

 

 

Results by presence of wall cracks are presented in Table 7.9. The median
measurement value is 3.6 pCi/L for those houses with wall cracks and 3.1 pCi/L for those
houses without wall cracks. The geometric mean measurement value is 3.7 pCi/L for
those houses with wall cracks and 3.0 pCi/L for those houses without wall cracks.
Measurement values as low as 0.3 pCi/L were recorded in houses with and with out wall
cracks. The highest measurement recorded in a home without wall cracks was 33.5 pCi/L
while the highest measurement result recorded in a home with wall cracks was 48.4 pCi/L.
Forty two percent of houses measured with wall cracks and 40% of houses measured

without wall cracks meet or exceed the EPA action level of 4 pCi/L.

 

90
Table 7.9: Results By Presence Of Wall Cracks

 

 

 

Cracks Present 102 3.6 3.7 6.4 0.3 - 48.4 42%

 

Cracks Absent 128 3.1 3.0 6.4 0.3 - 33.5 40%

 

 

 

 

 

 

 

 

Interaction Between Bedrock Risk Groups and Selected Housing Features

Three tables have been created in order to examine and analyze the interaction and
combined effects between bedrock and housing structural features. The following tables
utilize bedrock information based on emperically derived measurement results originally
presented in Table 6.7.

Table 7.10 presents results by bedrock risk categories and presence of basement.
In the high risk bedrock category, those homes measured with basements produce a
geometric mean of 4.1 pCi/L, and those homes without basements have a geometric mean
of 1.5 pCi/L. Fifty percent of homes with basements and 20% of homes without
basements have measurement values that meet or exceed the EPA action level. These
results fiom the high risk bedrock category are compared with the results from the lower
risk bedrock category. Homes with basements in the lower risk category produce a
geometric mean of 2.5 pCi/L and have 18% of measurement values that meet or exceed
the EPA action level. Those homes with out basements in the lower risk bedrock category
have a geometric mean of 1.1 pCi/L with no measurements that meet or exceed the EPA

action level.

91
Data presented in Table 7.10 indicates homes with basements in both bedrock risk

categories are at greater risk to increased radon levels than those without basements.
Houses in the lower risk bedrock category with basements (geometric mean 2.5 pCi/L)
appear to be at higher risk than those houses in the higher risk category without basements

(geometric mean 1.5 pCi/L).

Table 7.10: Bedrock Risk Groups and Presence of Basement

 

Bedrock Risk NumberOf
(3W ' W 9‘” ‘ muons -
Stains V '

 

 

 

High Risk 169 3.9 4.1 7.1 0.3 - 48.4 (84/169) 50%
Bedrock/Base.
Present

 

High Risk 20 1.1 1.5 2.5 0.3-8.8 (4/20) 20%
Bedrock/Base.
Absent

 

High Risk 189 3.8 3.7 6.8 0.3 - 48.4 (88/ 189) 47%
Totals

 

 

Lower Risk 33 2.5 2.5 2.2 0.5-12.1 (6/33) 18%
Bedrock/Base.
Present

 

Lower Risk 8 1.5 1.1 1.0 0.3-2.8 (0/8) 0%
Bedrock/Base.
Absent

 

Lower Risk 41 2.4 2.1 2.1 0.3 - 12.1 (6/41) 15%
Totals

 

 

 

Overall Totals 230 3.4 3.3 6.4 0.3 - 48.4 (94/230) 41 %

 

 

 

 

 

 

 

 

92
Measurement results from bedrock risk categories are compared with the presence

of sump pump in Table 7.11. Houses with sump pumps in the high bedrock risk category
have a geometric mean measurement of 4.4 pCi/L, while homes without sump pumps have
a geometric mean of 3.1 pCi/L. In the high risk bedrock category, 55% of the
measurements in homes with sump pumps and 39% of the measurements in homes
without sump pumps meet or exceed the EPA action level. These results are compared
with data from the lower risk bedrock category. In the lower risk bedrock category
homes with a sump pump present have a geometric mean of 2.3 pCi/L. Those homes with
out a sump pump have a geometric mean of 1.8 pCi/L. In this lower risk bedrock
category, 15% of homes with a sump pump and 14% of home without a sump pump have
measurements that meet or exceed the EPA action level.

A comparison of data in Table 7.11 between the two risk categories reveals that
geometric means in the high risk bedrock category are higher than the geometric means in
the lower risk category. Within each category, geometric means from measurements
taken in houses with sump pumps are higher than those readings taken in houses without
sump pumps. If bedrock risk is assumed to be similar in each risk category, the presence
or absence of a sump pump appears to effect indoor radon levels.

Table 7.12 presents results by bedrock risk categories and the presence of floor
cracks. In the high. risk bedrock category, those homes measured with floor cracks

produce a geometric mean of 4.1 pCi/L, and those homes without floor cracks have a

93
Table 7.11: Bedrock Risk Groups and Presence of Sump Pump

 

 

 

 

High Risk 85 4.5 4.4 8.3 0.3 - 48.4 (47/85) 55%
Bedrock/Sump
Present

 

High Risk 104 3.5 3.1 5.2 0.3-30.8 (41/104) 39%
Bedrock/Sump
Absent

 

High Risk 189 3.8 3.7 6.8 0.3 - 48.4 (88/189) 47%
Totals

 

 

Lower Risk 27 2.5 2.3 1.4 0.3-6.7 (4/27) 15%
Bedrock/Sump
Present

 

Lower Risk 14 2.3 1.8 3.1 0.3-12.1 (2/14) 14%
Bedrock/Sump
Absent

 

Lower Risk 41 2.4 2.1 2.1 0.3 - 12.1 (6/41) 15%
Totals

 

 

 

Overall Totals 230 3.4 3.3 6.4 0.3 - 48.4 (94/230) 41%

 

 

 

 

 

 

 

 

geometric mean of 3.1 pCi/L. Forty seven percent of homes with floor cracks and 46% of
homes without floor cracks have measurement values that meet or exceed the EPA action
level. These results fiom the high risk bedrock category are compared with the results
from the lower risk bedrock category. Homes with floor cracks in the lower risk category

produce a geometric mean of 2.4 pCi/L and have 20% of measurement values that meet or

94
exceed the EPA action level. Those homes with out floor cracks in the lower risk bedrock

category have a geometric mean of 1.7 pCi/L with 6% of measurements that meet or
exceed the EPA action level.

When data are compared between bedrock risk categories in Table 7.12, they
reveal higher geometric means in the high risk bedrock category than in the lower risk
bedrock category. Within each category, geometric means fi'om measurements taken in
houses with floor cracks are higher than those readings taken in houses without floor
cracks. If bedrock risk is assumed to be similar in each risk category, the presence or
absence of floor cracks appears to effect indoor radon levels.

Research Question Addressed

What role does housing construction and structural characteristics have in the variation of

 

Ladon mezfiurement reagljpgfl

According to the radon literature, housing structural characteristics can affect the
amount of radon in a home. Certain housing features may increase or decrease the actual
amount of indoor radon. The identification of household construction features which
promote or discourage indoor radon entry radon is vital for public awareness, mitigation,
and new construction considerations. Regardless of housing structural features, a

geologic source for radon must be present in order to have potential indoor radon risk.

95
Table 7.12: Bedrock Risk Groups and Presence of Floor Cracks

 

 

 

 

High Risk 117 3.9 4.1 7.4 0.4 - 48.4 (55/117) 47%
Bedrock/Cracks
Present

 

High Risk 72 3.7 3.1 5.8 0.3-30.8 (33/72) 46%
Bedrock/C racks
Absent

 

High Risk 189 3.8 3.7 6.8 0.3 - 48.4 (88/189) 47%
Totals

 

 

Lower Risk 25 2.5 2.4 2.5 0.5-12.1 (5/25) 20%
Bedrock/Cracks
Present

 

Lower Risk 16 2.4 1.7 1.1 0.3-4.6 (1/16) 6%
Bedrock/Cracks
Absent

 

Lower Risk 41 2.4 2.1 2.1 0.3 - 12.1 (6/41) 15%
Totals

 

 

Overall Totals 230 3.4 3.3 6.4 0.3 - 48.4 (94/230) 41%

 

 

 

 

 

 

 

 

 

Tables 7.10, 7.11, and 7.12 present data to address the combined effects of
bedrock and housing structural characteristics. Conclusions drawn from these tables

indicate that in areas of sirrrilar bedrock potential, certain housing structural features such

96
as the presence of basements, sump pumps, and floor cracks do influence the amounts of

radon in homes.

Houses with basements in the high risk bedrock category have a geometric mean
of 4.1 pCi/L compared with a geometric mean of 1.5 pCi/L for houses without basements.
In the lower risk bedrock category, homes with a basement have a geometric mean higher
than those homes with out a basement (2.5 pCi/L and 1.1 pCi/L, respectively). Within
each bedrock risk group, homes with sump pumps present have a higher geometric mean
than homes without sump pumps (4.4 pCi/L compared with 3.1 pCi/L in the higher
bedrock risk group; 2.3 pCi/L compared with 1.8 pCi/L in the lower bedrock risk group).
When examining the presence of floor cracks within each bedrock risk group, homes with
floor cracks present have a higher geometric mean than homes without floor cracks (4.1
pCi/L compared with 3.1 pCi/L in the higher bedrock risk group; 2.4 pCi/L compared
with 1.7 pCi/L in the lower bedrock risk group). These results, which are presented in
Tables 7.10, 7.11, and 7.12 respectively, identify housing structural characteristics that
seem to influence indoor radon levels.

Table 7.] presents results by location of the measurement device. According to
EPA testing protocol, the lowest livable level of a dwelling should be tested. Therefore if
a basement is present, testing should occur at this level. Of the 230 measurement readings
in the study, 202 are basement results, and 28 are first floor results.

Typically, basement readings are expected to be higher than first floor readings due
to closeness to the source, additional entry points (floors and walls), and fewer changes in

air pressure due to the opening and closing of doors. Table 7 .1 presents measurement

97
results by location of measurement device. The median and geometric mean measurement

results from basement readings are higher (3.7 pCi/L and 3.8 pCi/L respectively) than first
floor measurement results (1.3 pCi/L and 1.3 pCi/L respectively). While both categories
recorded measurements of 0.3, as expected, higher measurement readings occurred in the
basement as opposed to the first floor. The highest measurement result collected in a
basement is 48.4 pCi/L, while the highest measurement result collected on a first floor is
8.8 pCi/L. Forty five percent of basement measurement results compared with 14% of
first floor results, meet or exceed the EPA action level.

A t-test was computed to compare the means of basement measurements and first
floor measurements using the log transformed measurement values. The two means are
significantly different with a t-value of 5.657 and a probability value of 0.000. These
results suggest that in Lenawee County, human risk to radon is greater in the basement
than on the first floor of a dwelling.

Measurement results are compared with the age of the housing structure in Table
7.2. Based on the radon literature, older homes could be at less risk because they lack
insulation allowing additional air exchange and pathways for indoor radon to escape, or
more at risk because of poor maintenance and the wear of age. Newer housing can be at
higher risk due to better insulation producing a more “tightly” sealed house where there is
less air exchange. Radon entering a well sealed home may have fewer exit routes.

No clear patterns emerge in Table 7.2 when examining age of housing structure
and measurement results. The highest geometric mean is 4.4 pCi/L in the 25 - 49 year

structure category, and the lowest geometric mean is 2.1 pCi/L in the 50 - 74 year

98
structure category. The lack of an obvious pattern emerging could be due to a high

variation in structural features among houses in each category.

A simple regression with the age of the housing structure produced a coefficient
value of -0.021 and a probability value of 0.030, indicating a significant relationship.

A Board of Review assessment value was recorded for each home surveyed.
These fifty percent values were divided into five categories for analysis and are presented
in Table 7.4. Based on assumptions from the radon and other literature, it was expected .
that houses of lower value would be at higher risk and that houses of higher value would i
be at lower risk. These assumptions are primarily due to expected levels of house
maintenance. The expected pattern did not occur in the measurement reading results. Of
the five assessment value categories, the lowest geometric mean (2.1 pCi/L) occurs in the
lowest assessment value category (5,600 - 15,000). A geometric mean of 3.3 pCi/L
occurs in the highest assessment value category (60,001 - 132,500).

A simple regression with the board of review assessment values produced a
coefficient value of 0.000 and a probability value of 0.01] indicating significance.

The sump pump is a commonly recognized housing structural characteristic that
can increase radon risk. The presence of a sump pump commonly requires an opening in
the floor of the structure which provides a likely radon entry point. Homes with sump
pumps are expected to have radon readings higher than those homes without sump pumps.

Table 7.5 presents measurement results based on presence of sump pump. While
low and high measurement readings were obtained in houses with and without sump

pumps, the geometric mean of measurement results taken in dwellings where a sump is

99
present is higher (3.8 pCi/L) than the geometric mean measurement result from homes

with out sump pumps (2.9 pCi/L). F orty-six percent of homes with sump pumps present
compared with 36% of homes with sump pumps absent meet or exceed the EPA action
level. These descriptive results indicate that homes in Lenawee County with sump pumps
are potentially at higher risk for indoor radon.

A chi-squared test and a two sample t-test produced probability values indicating
that the association and variance is not significant. Table 7.13 presents the fi'equency
results of presence of sump pump by measurement result. The Kendall Tau-B value is
-0.092 indicating a weak negative association. The probability value is 0.161. The t-test
using the log transformed measurement values was slightly better with a t-value of 1.903
and a probability value of 0.058. This computation does suggest some difference between

the presence and absence of sump pumps.

Table 7.13: Frequencies - Presence Of Sump Pump By Measurement Result

 

sum» ‘ «to was m

, 4:» ~ - - .¢. 'v: :-: :-: w 4: .....

 

 

 

Sump Present 61 51 112

 

Sump Absent 75 43 118

 

 

 

 

 

 

Total 136 94 230

 

The type of housing construction materials can effect the amount of radon in a

dwelling. Information on floor and wall material was collected for this survey.

100
An f-test was conducted with housing floor material and log linear transformed

measurement values. The results reveal an association between different types of floor
material and measurement values. The f-ratio statistic is 4.011 with a probability value of
0.001 which indicates that floor material type effects measurement values significantly
within a 99% confidence interval.

Results by floor material type are presented in Table 7.6. Homes with floors of
stone and earth are expected to produce elevated radon readings because of composition
and lack of any human made barrier between natural material and the housing structure.
The highest geometric mean in the floor material type occurs in the stone category (8.3
pCi/L)., as expected. The second highest geometric mean occurs in the earth category
(5.8 pCi/L), also as expected. Measurement results in the poured concrete floor category
have a geometric mean of 3.6 pCi/L. Forty four percent of the 192 homes surveyed with
poured floors have measurement results above the EPA action level. These relatively high
readings obtained in homes with poured floors may be due to the tendency of large slabs
to crack over time with changes in temperature and as settling occurs. These floor cracks
provide potential entry points for radon. Table 7.7 reveals that 42% of houses measured
with floor cracks and 39% of houses without floor cracks meet or exceed the EPA action
level. The geometric mean measurement value is 3.7 pCi/L for houses with floor cracks
and 2.8 pCi/L for houses without floor cracks. These results indicate that on average,

houses with floor cracks are at a greater risk than houses without cracks in the floor.

101
An f-test was conducted with housing wall material means using the log linear

transformed measurement values. The f-ratio statistic is 5.179 which is highly significant
with a probability value of 0.000.

The suggested potential risk of poured concrete is even more apparent when
examining measurement results and wall material types presented in Table 7.8. The
highest geometric mean (4.8 pCi/L) occurs in the poured wall category. The second
highest geometric mean (4.7 pCi/L) occurs in the stone wall category. While
measurement readings in the poured wall category range form 0.3 - 48.4 pCi/L, 56% of
homes in this category have measurement results which meet or exceed EPA action levels.
Table 7.9 presents results by presence of wall cracks. Houses with wall cracks have a
greater median (3.6 pCi/L) and geometric mean (3.7 pCi/L) than houses with out wall
cracks (3.1 pCi/L and 3.0 pCi/L respectively).

These results indicate that there is an association between wall materials, floor
materials, structural cracks, and radon measurement results. These results also suggest
that housing structural materials of stone, earth material, and poorly maintained poured

concrete create a higher radon risk potential.

 

CHAPTER 8: HUMAN BEHAVIORS THAT AUGMEN T CANCER RISK

Intervening Human Behavior Variables

This chapter presents data gathered on intervening human variables. These data
were collected to examine and help determine overall human risk to indoor radon. These
intervening human behavioral variables have no effect on the actual measured result of
radon in a home, however, they can effect the level of risk to residents in the home when
radon is present.

Table 8.1 provides the number of households with smokers present. Fifty-eight
(or 25%) of the 230 households surveyed had at least one smoker present. One hundred

seventy two (or 75%) of the 230 households surveyed had no members who smoke.

Table 8.1: Number of Households with Smokers Present

 

 

 

 

mm Mathews
SWAE m 4- W345
At Least One 58 25%
Smoker
No Smokers 172 75%
Present
Totals 230 100%

 

 

 

 

 

Table 8.2 presents results by smoking status. In homes with one or more smokers

9

the median (2.5 pCi/L) and geometric mean (2.7 pCi/L) measurement values are lower

102

103
than households tested where no one smokes (3 .6 pCi/L and 3.6 pCi/L, respectively).

Low measurement results of 0.3 pCi/L occur in both categories. Elevated measurement

results occur in both categories as well. In homes with one or more smokers present, 33%

of measurements meet or exceed the EPA action level, while 44% of measurements in

homes with no smokers present meet or exceed the EPA action level.

Table 8.2: Results By Smoking Status

 

_ Household

ySIatusBy

mm: m ‘

m o.
1 Exceeds EPA
"Am: ; f ; AAA:

 

One Or More
Smokers

2.5

 

33%

 

No Smokers
Present

172

3.6

 

3.6

 

6.8

0.3 - 48.4

 

44%

 

 

 

 

 

Data gathered on the number of households where sleeping occurs in the basement

is presented in Table 8.3. Two hundred two of the 230 houses surveyed in this study have

Table 8.3: Number of Households with Basement Sleeping

Based on 202 houses with basements. Twenty eight of 230 homes had no basement.

 

 

Sleeping Occurs

 

No Sleeping Occurs

160

 

 

Totals

 

202

 

 

 

 

104
basements. Sleeping occurs in the basement in 42 (or 21%) of the 202 homes with

basements Sleeping does not occur in 160 (or 79%) of the 202 homes surveyed with
basements.

Table 8.4 provides measurement results by use of the basement for sleeping. The
median (4.8 pCi/L) and geometric mean (4.6 pCi/L) values in homes where sleeping does
occur in the basement are higher than the values in homes where basement sleeping does ......
not occur (3.6 pCi/L and 3.6 pCi/L respectively). Measurement values of 0.9 pCi/L or ll
less were recorded in both categories as were readings of 27.1 pCi/L or greater. Fifty five

percent of homes where sleeping occurs in the basement and 42 % of homes where no

sleeping occurs in the basement meet or exceed the EPA action level.

Table 8.4: Results By Use Of Basement For Sleeping
Based on 202 houses with basements. Twenty eight of 230 homes had no basement.

 

38mm  > waor ' Median Geom : swede: d

mu—

 

 

Yes 42 4.8 4.6 7.0 0.9 - 27.1 55%
NO 160 3.6 3.6 6.5 0.3 - 48.4 42%

 

 

 

 

 

 

 

 

 

Table 8.5 presents the number of households where basement television watching
occurs. These results are based on 202 houses that have basements. In 66 (or 33%) of
the 202 homes with basements, television viewing occurs. Television viewing does not

occur in the basement of 136 (or 67%) of the 202 homes surveyed with basements.

105

Table 8.5: Number of Households Where Basement Television Viewing Occurs
Based on 202 houses with basements. Twenty eight of 230 homes had no basement.

 

H, ...,...,p

 

 

 

Viewing Occurs

 

 

No Viewing Occurs

136

 

Totals

 

 

202

 

 

 

Results by use of basement for television viewing are presented in Table 8.6. The

median (4.0 pCi/L) and geometric mean (4.6 pCi/L) measurement results are higher in the

homes where basement television viewing does occur, as opposed to homes where no

basement television occurs (3 .3 pCi/L and 3.5 pCi/L respectively). In both categories

measurement exist of 0.9 pCi/L or lower and 27.1 pCi/L or higher. Fifty percent of homes

with basement television viewing, and 42% of homes without television viewing meet or

exceed the EPA action level.

Table 8.6: Results By Use Of Basement For Television Viewing

Based on 202 houses with basements. Twenty eight of 230 homes had no basement.

 

  
  

Assessed
Vms ‘ ' ‘~ *

*‘ze.:<sc.mzf '

 

- ......
Exceeds EPA

5...... ,,,,, (new Aster—eve:

 

Yes

4.0 4.6

0.9 - 27.1

 

No

 

 

 

3.3 3.5

 

 

0.3 - 48.4

 

 

 

 

 

1 06
Research Question Addressed

mmwmm—mmmmmfi
Wt increased swung?

This study seeks to determine overall human risk to indoor radon. Intervening
human behaviors must be included in this holistic framework. While these human
variables in no way affect the amount of radon present in a home (the measurement
result), they can increase or minimize overall indoor radon risk.

Cigarette smoking and radon have a synergistic affect. Typically, smokers are at a
higher potential risk when radon is present. The number of households with smokers
present is presented in Table 8.1. Twenty five percent (58 of 230) of homes surveyed
have at least one smoker present. Residents in these dwellings are at higher risk to radon
than the residents in homes where no smoking occurs.

Table 8.2 presents measurement results by smoking status. The geometric mean of
measurement results in homes where smoking occurs (2.7 pCi/L) is lower than the
geometric mean in homes where no one smokes (3.6 pCi/L). Thirty three percent of
homes where smoking occurs compared with 44% of homes with no smokers present
meet or exceed the EPA action level. While low and high measurement results are
recorded in houses of both categories, average radon levels in houses where smoking
occurs are lower than those homes with no smokers present.

Human risk to indoor radon is based not only on exposure dose, but also on
duration of exposure. Spending time in the basement for various activities can increase
length of exposure and create additional risk. This survey gathers information on the use

of the basement for sleeping and watching television.

107
Table 8.3 presents information on the number of homes where basement sleeping

occurs. Basement sleeping occurs in 21% (42 of 202) of homes with basements. Table
8.4 provides measurement results based on basement sleeping status. The geometric mean
measurement result for homes where basement sleeping occurs is higher (4.6 pCi/L) than
the geometric mean measurement result for homes where no basement sleeping occurs
(3.6 pCi/L). Measurement readings in 55% of homes where basement sleeping occurs and
42% of homes where no basement sleeping occurs meet or exceed the EPA action levels.

These results suggest that on average, residents who sleep in the basements of
homes in Lenawee County may be at an increased risk to radon due not only to longer
durations of exposure, but also higher average radon measurements.

Basement use for television viewing is also examined. Table 8.5 presents the
number of households where basement television viewing occurs. Basement television
viewing occurs in 33% (66 of 202) of households surveyed in Lenawee County.
Measurement results by use of basement for television viewing are provided in Table 8.6.
As observed with the results in household basement sleeping status, the geometric mean
measurement results are higher (4.6 pCi/L) in homes where basement television occurs
than the geometric mean measurement result (3.5 pCi/L) in homes where no television
basement viewing occurs. Half of the homes where basement television viewing occurs
and 42% of the homes where no television viewing occurs have measurement readings
that meet or exceed the EPA action level.

While high and low measurement readings were obtained in both television

viewing categories, these data suggest that on average, residents who watch television in

 

108
the basement are at greater risk to radon based not only on additional length of exposure,

but also on higher average measurement readings.
It must again be noted that these intervening human behavioral variables in no way
effect the amount of radon in a home. Measuring the prominence of certain human

behaviors is necessary to assess overall human risk to radon in the domestic environment.

 

CHAPTER 9: SUMMARY AND CONCLUSIONS

Summary and Conclusions

This study examined overall human risk to indoor radon exposure in Lenawee
County, Michigan, utilizing a holistic approach following the disease ecology tradition to
go beyond the common radon risk estimates based solely on bedrock. While it is beyond
the parameters of this study to examine disease outcomes specifically, of primary interest
are the major factors and variables that contribute to overall risk. Overall radon risk was
divided into three main categories: geologic risk potential (including bedrock, surficial
features, and soil associations), housing structural characteristics, and intervening human
behavior variables.

Assumed geologic risk potential categories were established for each of the three
geologic strata, individually and collectively. These categories were established as a set of
predictors. Data were collected from a survey of 230 households in the county. A
questionnaire was administered, a visual inspection was made and recorded, and an actual
radon measurement was obtained in each home.

These data were analyzed to examine primarily the following: the spatial variations
in the levels of radon risk in the county; the role of surficial features and soil associations
as filtering layers that effect the movement of radon; variations of radon risk based on

housing structural characteristics; and the prominence of certain intervening human

109

110
variables that affect overall human risk to radon. Data analysis procedures were

implemented to explore the impacts of the these topics of concern within the framework of
the study.

Findings from this study indicate a higher than average radon risk for residents of
Lenawee County. F orty-one percent of the 230 houses surveyed in this study have
measurement results that meet or exceed the EPA action level of 4.0 pCi/L. This
percentage of at-risk housing is relatively high when compared with the state-wide survey
results which revealed 12% of the 2082 homes measured with results that met or exceeded
the EPA action level. Overall summary statistics from this study revealed a median of 3.4
pCi/L, a geometric mean measurement value of 3.3 pCi/L and a range of 0.3 - 48.4 pCi/L.

Results from this survey also indicate that great variations exist in the spatial
patterns of risk to radon in Lenawee County. Certain “hot spots” (areas of higher
measurement results) and areas of lower measurement results were identified. A major
“hot spot” area of higher risk is located in a population corridor which runs between the
cities of Adrian and Tecumseh. Of particular risk are homes located on Marshall
Sandstone, Berea Sandstone, and Coldwater Shale (geometric mean measurement values
of 4.5, 4.1, 3.6 pCi/L, respectively, and median measurement values of 5. 1, 4.5, 3.7 pCi/L,
respectively).

The working assumptions of potential geologic risk, established as predictors,
were empirically tested in this study. Data indicated that these original estimated geologic
radon risk potential categories were accurate in classifying only two of the three geologic
components. The results from the bedrock strata contradicted the working assumptions of

this study. The highest measurement readings did not occur in homes located in areas of

111
bedrock assumed and classified as “high” risk potential. Instead, the highest measurement

results occurred in areas classified as assumed “medium” risk potential bedrock. This
finding is one of the original intents of this research -- to measure and locate geographic
variations in indoor radon levels.

To further explore and examine these actual site measurement readings, the
assumed bedrock potential risk categories were disaggregated and re-assessed. This
bedrock disaggregation allowed the nine bedrock formations to be ranked according to
test measurement means. Based upon this ranking of mean measurement results from field
findings, bedrock formations were reclassified into two risk groups: a high risk group of
formations with the three highest geometric mean measurement values, made up of,
Marshall Sandstone, Berea Sandstone, and Coldwater Shale; and a lower risk group of
formations with six lowest geometric mean measurements, made up of Antrim Shale,
Detroit River Group, Sunbury Shale, Dundee Lime Stone, Bedford Shale, and Traverse
Group. The overall geometric mean is 3.7 pCi/L, and the median is 3.8 pCi/L, for the
reclassified high risk group, and the overall geometric mean is 2.1 pCi/L, and the median is
2.4 pCi/L, for the reclassified lower risk group. While these bedrock measurement results
are mediated by surficial features, soils and housing structural characteristics, this
reclassification of bedrock based upon household radon measurements , provides a better
characterization of potential radon risk in Lenawee County.

Bedrock type is an important variable in assessing risk to radon, however this
variable alone cannot provide accurate overall risk. Results from this study also indicate
the importance of considering surficial features and soil associations as filtering layers

when assessing radon risk. The measurement results in the surficial and soil association

 

112
strata supported the working assumptions of this study. Of houses measured, the greatest

percentage with measurement readings that met or exceeded the EPA action level,
occurred in the high potential risk category, and the lowest percentage occur in the low
potential risk category, as assumed.

Housing structural characteristics also can affect the amount of radon present in a
home. The survey data indicated the presence of certain structural features such as
basements, sump pumps, floor and wall cracks, as well as the type of housing material can
impact the amount of radon present in a home.

The combined effect of bedrock type and housing structural characteristics were
also examined in this study. Conclusions drawn from analysis indicate that in areas of
similar bedrock potential, certain housing structural features such as the presence of
basements, sump pumps, and floor cracks do influence indoor radon levels.

The prominence of the most common human variables which exacerbate risk were
examined in the county. Data on smoking status and amount of time spent in the
basement provides the final component in assessing overall radon risk. Twenty five
percent of the 230 households surveyed had at least one smoker present. Basement
sleeping occurred in 21% of the 202 homes surveyed with basements. The geometric
mean measurement result of these 42 houses was 4.6 pCi/L, which indicated a higher risk
for these residents in Lenawee County. In the 66 households where television viewing
occurred in the basement, the mean measurement reading was 4.6 pCi/L. Occupants of
these homes are also potentially at a higher rate of risk. The prevalence of these human
intervening variables, while having no impact on the actual amount of radon in a home,

serve to better characterize domestic radon risk in the county.

 

113

Host-Agent-Environment Interaction Model Revisit—ed

The disease ecology tradition of medical geography utilizes a holistic frame work
and recognizes risk as multi-causal and multi-factored. The nature of the Host-Agent -
Environment Interaction Model addresses these complexities. In this study, predicted or
expected results of bedrock geology were not entirely accurate in light of the empirically
observed data. These occurrences do not necessarily indicate that the assumptions were
wrong but rather reveal other influences at work and illustrate the complexity of risk
associated with domestic radon exposure. Figure 9.1 shows this model in a revisited form.
The convergence of the three major components of this study is represented in the area
shaded black. Ideally, if all variables could be controlled, this “center” could be the focus
of analysis. In revisiting this model, 11 have emphasized the interaction not only of all three
components (black shaded area) but the interaction between any given two components
(gray shaded areas). For instance, the original bedrock risk assumptions may be accurate
in assessing the relationship between radon (the agent) and the environment (the housing
structure), but human (host) actions may alter the housing environment producing

measurement results that are not expected.

ll

 

 

HOST-AGENT-ENVIRONMENT MODEL REVISITED

FIGURE 9.1

RADON

HUMAN

m
m
m
m

 

HOUSING

 

 

115

mama] Components Analysia

Principal components analysis (PCA) was conducted to reduce the number of
variables into components relative to the Host-Agent-Environment Interaction Model.
Ten variables were used and four factors were identified based on the eigenvalues and
skree test. A natural break occurred between the fourth and fifth eigenvalues (1.247
and 0.870 respectively). The rotated varimax loadings were interpreted. Factor one had
high loadings for variables: location of device, wall material type, and floor material type.
Factor two had high loadings for variables: overall geologic potential and bedrock
potential. Factor three had high loadings for variables: surficial feature potential and soil
association potential. Factor four had high loadings for variables: housing assessment
value, age of household and floor drains. Together these four factors account for over
70% of the variation within the data.

In addition to reducing the number of study variables, principal components
analysis was completed to reveal patterns, interrelationships, and structure in the data.
The factors relate to the host-agent-environment interaction model: factor one, housing
structure and materials, and factor four, housing quality, can be associated with the
environment component of the model; factor two and factor three are geologic factors that
can be associated with the agent component of the model. The human host, if made aware
of these other factors can take necessary precautions to minimize risk.

This study provided valuable community outreach to residents of Lenawee County.
Through this study, services and information were provided at no cost, to raise public

awareness, promote understanding, and potentially improve quality of life.

116
The presence of indoor radon is a complex and multi-causal phenomenon. This

study has examined some of the most important variables associated with radon risk,
within a holistic framework, in order to assess overall human risk to indoor radon in
Lenawee County. While findings indicate certain variables that lead to higher and lower
risk, the EPA’s recommendation of home testing must be stressed. This study has
identified and analyzed certain features that can help predict human risk to radon. Testing,

however, is the only way to be certain of the actual amount of radon present in a house.

Recommendations

A primary goal of this study is to reduce human risk to indoor radon by identifying
overall factors of risk and spatial variations in Lenawee County. Creating public
awareness is a key step in reducing the threat of radon. Measurement results from this
survey were given to each homeowner with an interpretation of the reading, EPA
approved literature, and contact telephone numbers to enable any questions to be
answered.

All survey findings have been made available to state and local health authorities.
It is recommended that the overall findings of this study be made available to the public
and that steps be taken to lower potential radon risk. A campaign to inform residents of
potential high risk radon areas should be undertaken, as well as distribution of information
identifying housing structural characteristics that may place the occupants of a home at
greater risk. Human behaviors, such as smoking, that place residents at additional risk,

should be made known to the public. In homes where basement measurement levels meet

l 17
or exceed the EPA action level, occupants should be encouraged to minimize activities

there.

Radon testing should continue to be promoted within Lenawee and other counties
in Michigan in order to firrther assess and characterize risk. Home owners in high risk
areas should especially be urged to test homes following standard EPA procedures. Test
kits can be made available to the public at reduced rates through the health department. A
radon hot line should be set up to field radon risk related questions and to distribute
information on radon mitigation. EA

Guidelines for reducing the threat of radon in the construction of new homes
should be made available to all contractors. The public should be aware that reducing risk
to radon can most often be fairly easy and inexpensive.

Finally, since measurement results contradicted the original working assumptions
of bedrock risk classifications in this study, additional data could be gathered on bedrock
formations (perhaps actual samples). More detailed geologic information could lead to a

better understanding of radon risk in Lenawee County.

BIBLIOGRAPHY

American Lung Association. Air Polmm in Your Homa? Report No. 1001. American
Lung Association, June, 1990.

American Medical Association Council on Scientific Affairs. "Health Effects of
Radon Exposure: Report of the Council on Scientific Affairs, American Medical
Association." Archives of Internal Medicine, Vol 151, April 1991: 674-677.

Block. G. Matanoski, G. Seltser, R MitchelLT. "Cancer Morbidity and Mortality in
Phosphate Workers." Cancer Research, Vol. 49, December 1988: 7298-7303.

Cavallo, A. Gadsby, K. Reddy, TA. “The Use of Computed Radon Entry Rates to
Understand Radon Concentration in Buildings.” Health Physics, Vol. 66, No. 2,
February 1994: 178-184.

Christofides, S. and Christodoulides, G. "Airborne Rn-222 Concentration In Cypriot
Houses." Health Physics, Vol 64, No. 4, April 1993: 392-396.

Clark, WAV. and Hosking, P.L. Statistical Methods for Geoggphers. John WIley and
Sons: New York, 1986.

Clifi‘, KD. Green, B.M.R. Lomas, PR. “Domestic Radon Remedies.” Radiation
Protection Dosimgy, Vol. 45, No. 1/4, 1992: 599-601.

Crameri, H. Brunner, R "Indoor Rn Levels in Different Geological Areas of
Switzerland.” Health Physics, Vol. 57, No. 1, July 1989: 29-38.

Cross, F. Palmer, R. Filippi, R. et al "Carcinogenic Effects Of Radon Daughters,
UraniumOre Dust and Cigarette Smoke in Beagle Dogs." Health Physics,
Vol. 42, 1982: 33-52.

Demers, Raymond. "Overview of Radon, Lead, and Asbestos Exposure. " American

Family Pmadm Vol. 44, No. 5, Supplement, November 1991: 518-528, 55S-
618.

118

119

Earickson, Robert J. Greenberg, Michael R Lewis, Nancy D. Meade, Melinda S.
Taylor, S. Martin. “Medical Geography”. Geogpaply In America Gaile, Gary L
and Willmott, Cort T. (eds). Merrill Publishing Company: Columbus, 1989.

Environmental Protection Agency with the Midwest Universities Radon
Consortium. _R_adon Measurement ngrators Proficiency Course Manual.
University of Minnesota: 1993.

Environmental Protection Agency. Home Buyer’s apd Seller’s nge to Radon.
EPA/402-R-93 -003. United States Environmental Protections Agency, Air and
Radiation 6604]: Washington 1993.

Ail-sun

Environmental Protection Agency. Reducing Radon in Structures. EPA/680-D-90-020.
United States Environmental Protection Agency, Air and Radiation: Washington,
1990

‘ “no"!
,3... ~

Environmental Protection Agency. Radon Reduction Methods. United States
Environmental Protection Agency, Office of Research and Development:
Washington, 1989.

Environmental Protection Agency. Applications of Radon Reduction Methods.
EPA/625/5-88/024. United States Environmental Protection Agency, Office of
Research and Development: Washington, 1988.

Environmental Protection Agency. Mon Reduction Tecflques For Detached Houses.
EPA/625/1-88/019. United States Environmental Protection Agency, Office
of Research and Development: Washington, 1988.

Field, RW. Kross, B.C. Vust, L.J. “Radon Testing Behavior In A Sample of Individuals
With High Home Radon Screening Measurements.” Risk Analysis, vol. 13, no. 4,
August 1993: 441-447.

Gundersen, Linda C. S. "Hidden Hazards of Radon - Scanning the Country for
Problem Locations." Earth, vol. 1, no. 6, November 1992: 54-61.

Gustafson, Tracy L. EPISTAT Statistical Package for the IBM Personal Computer.
Version 3.5.. User Guide. Richardson, Texas, 1989.

Haines, Forest PhD. Professor, Department of Geology, Adrian College, Adrian
Michigan. Interview and consultation, October 13, 1993.

Hofmann, W. Menache, M. Graham, R. "Radon Progeny Dosimetry In The Rat
Lung." Health Physics, Vol. 64, No. 3, March 1993: 279-290.

120

Johnston, RJ. Multivariate Statistical Analysis in Geoggaphy. John Wiley & Sons Inc:
New Yorlc 1991.

Klotz, J., Petix, J ., Zagraniski, R. "Mortality of a Residential Cohort Exposed to
Radon From Industrial Contaminated Soil." American Journal of Epidemiology,
Vol. 129, no. 6 (1989): 1179-1186.

Learmonth, Andrew. Diseag Ecology. Basil Blackwell, Inc: New York, 1988.

Lees, REM. Steele, R Roberts, JH. "A Case Control Study of Lung Cancer
Relative To Domestic Radon Exposure." Interntional Journal of Epidemiology,
vol. 16, 1987: 7-12.

Lounsbury, John F. and Aldrich, Frank T. Introduction to Geomphic Field Methods
and Techniques. Second Edition. Charles E. Merrill Publishing Company:
Columbus, 1986.

 

Martin, Helen M. “Map of the Surface Formations of the Southern Penisula of
Michigan.” Department of Conservation, Geological Survey Division, Publication
49, 1955.

Meade, Melinda Florin, John and Gesler, Wilbert. MedicalgGeoggpm. The Guilford
Press: New York, 1988.

Mettlin, Curtis. "Descriptive and Analytic Epidemiology: Bridges to Cancer Control. "
Cancer, vol. 62 (1988): 1680-1687.

Michigan Department of Public Health, Bureau of Environmental and Occupational
Health, Division of Radiological Health. Indoor Radon In Michigan: ResidentiaL

Survgg 8mm. RH-947/A, May 1990.

Michigan Department of Public Health, Bureau of Environmental and Occupational
Health, Division of Radiological Health. Igloor Radon In Michigan: Report To
The Governor. RH-947, December 1988.

National Safety Council. Reporting:on Rfiadon: A Joumalist's Guide to Coveringghe
fltion's Second-Leading Cause of Lung Cancer. National Safety Council,
Environmental Health Center Division: Washington, 1989.

Nazaroff, W., Nero, A. Mon and Its Decay Prod_ucts In Indoor Air. John Wiley &
Sons, New Yorlg 1988.

Neuberger, J. , Frost, F ., Gerald, K. “Residential Radon Exposure and Lung Cancer:
Evidence of an Inverse Association in Washington State.” Journal of
Environmental Health, vol. 55, no. 3 (1992): 23-25.

121

Otton, James K. The Geology of Radon. US Department of the Interior/ US.
Geological Survey: Denver, 1992.

Paul, Birnal Kanti. "Approaches To Medical Geography: An Historical
Perspective. " Social Science a_nd Medicine. vol. 20, no. 4 (1985): 399-409.

Pyle, Gerald F. Applied Medical Geoggaphy. John Wiley & Sons: New York, 1979.

Pershagen, G, Akerblom, G., Axelson, O. et a1. “Residential Radon Exposure and Lung
Cancer in Sweden.” New England Journal of Medicine. vol. 330, no. 3, January
20, 1994: 159-164.

Radford, Edward P. "Potential Health Effects of Indoor Radon Exposure."
Environmental Health Perspectives, vol 62 (1985): 281-287.

Rannou, Alain. "Radon In Dwellings: Physical and Health Properties." Endeavour,
New Series, Vol. 14, No. l, 1990: 34-39.

Reimer, GM. and Gundersen, L.C.S. "A Direct Correlation Among Indoor Rn, Soil
Gas Rn, and Geology in the Reading Prong Near Boyertown, Pennsylvania."
Health Physics, Vol. 57, No. 1, July 1989: 155-160.

Ruotsala, Albert P. Mineralogy Of Antrim Shale. Michigan. FE-2346-79. United States
Department Of Energy: Washington, 1980.

Samet, Jonathan M. "The Epidemiology of Lung Cancer." Chest, Vol. 103, No. 1,
Supplement, January 1993: 205-298

Schumann, R. R. Geolgg'm Radon Potenpial of EPA Region 5: Illinois, Indiana, Michigar;
Minnesota, Ohio, and Wisconsin. OF 93-0292-E. United States Geological
Survey: Denver, 1993.

Schumann, R. R. "Geologic Radon Potential of the Glaciated Upper Midwest." m
1992 lntemational Smsium on Radon and Radon Reduction Technology,
Additional Papers Volume, VIII-3, EPA, September 1992.

Stidley, CA. and Samet, J .M. “A Review of Ecologic Studies of Lung Cancer and Indoor
Radon.” Health Physics, Vol. 65, No. 3, September 1993: 234-251.

Svensson, C ., Eklund, G., Peshagen, G. "Indoor Exposure To Radon From the
Ground and Bronchial Cancer in Women." lntemational Archifves of

Occupational and Environmental Health, Vol. 59, 1987: 123-131.

122

Svensson, C., Pershagen, G., Klominek, J. "Lung Cancer in Women and Type of
Dwelling in Relation to Radon Exposure." Cancer Research, vol. 49 (April 1,
1989): 1861-1865.

United States Department of Agriculture, Soil Conservation Service. Soil Survey:
Lenawee Coung Mchigan. US Department of Agriculture, Soil Conservation
Service, in c00peration with Michigan Agricultural Experiment Station:
Washington, series 1947, No. 10, issued August 1961.

 

Western Michigan University. Hydrogeology For Underground Injection Con_trol In
Michigan: Part 1. US. Environmental Protection Agency, Underground Injection
Control Program. Department of Geology, Western Michigan University:
Kalamazoo, MI, 1981.

 

Xiang-Zhen, X. Lubin, J.H. Jun-Yau, L. Li-Fen,Y. Sheng, L. Lari, Y. Jian-Zhang, W.
Blot, W. "A Cohort Study in Southern China of Tin Miners Exposed To Radon
and Radon Decay Products." Health Physics, Vol. 64, No. 2, February 1993:
120-131.

123
APPENDIX A: LENAWEE COUNTY RESIDENTIAL RADON SURVEY

Date

Dear Household Resident,

There is growing interest in radon as a potential threat to human health. Radon is a
radioactive gas which is colorless and odoriess. It occurs naturally in soil and rocks and can
enter a home through entry points in foundation floors, walls, and other structural features.

In 1988, the Michigan Department of Public Health conducted a state wide household
radon survey. The results from this screening were used to make estimates of radon risk for
each county in the state. Lenawee County is estimated to have a higher potential for radon
risk.

A study is being conducted through the Michigan Department of Public Health and the
Lenawee County Health Department to learn more about radon in Lenawee County. Your
home, along with approximately 350 other houses throughout the county, has been selected for
this survey.

It is very simple and safe to test for radon. Through your participation in this survey, your
home will be tested for radon at no cost to you, and the results made available within a few
weeks. All individual results will be kept strictly confidential.

The enclosed pamphlet "Reducing Radon Risks" and the Survey Information Sheet
provide additional information. Please review these materials at your earliest convenience.

In the near future I will be telephoning you to discuss your participation in the Lenawee
County Residential Radon Survey. This study will contribute vital information to our
knowledge of radon levels in the County.

Thank you for your time and consideration. I look forward to speaking with you on the
phone in the near firture. Ifyou have firrther questions about the study please call (517) 353-
9940. You may also contact the Lenawee County Health Department at 264-5202.

Sincerely,

Tarek A. Joseph

Study Coordinator
Department of Geography
Michigan State University

124

APPENDIX B: LENAWEE COUNTY RESIDENTIAL RADON SURVEY
Survey Information Sheet

How was my household selected?

A stratified random sample was used to select households for this survey. A number of criteria
were used in the selection process including location of the home in the county, the type of
underlying geological features such as bedrock, surficial geology, and soil, and variables from
the 1990 census.

What will I need to do?

Participation is simple. In the near future you will receive a phone call to arrange a brief initial
visit to your home (approximately 15 minutes). During that visit, a harmless radon test device
will be placed in your home and you will be asked a few short questions about your home, how
much time you spend in it, and other related topics. During the duration of the test, the test
device should remain untouched, and occupants should go about their normal activities. The
only necessary condition is that windows remain closed, and opening of doors is limited to
regular entry and exit. The detector, which requires no power or maintenance, will need to
remain in your home from 3-5 days, at which time it will be picked up and sent to a laboratory
to be analyzed.

What is the radon testingalevice?

The device used to measure indoor radon is a small envelope filled with harmless activated
charcoal. This "packet", which is about the size of a standard business envelope, will be placed
in the lowest livable level of your home.

Is therefi any cost involved to me?

No. The test and survey will be conducted fiee of charge. There will be absolutely no cost to
the home occupants.

Who will visa my home?

Tarek A. Joseph, a doctoral student at Michigan State University, and coordinator of this study
will visit your home.

Who willsee the results of the measurements made in my home?

The radon measurement result from your home will be mailed to you. The data fi'om all homes
in the study will be combined in statistical summaries which will be presented in reports by the
county and state government agencies, Michigan State University, and the survey team.
Neither your name nor your street address will be revealed in any presentation of the
summarized information.

 

 

125

Why should I participate?

You will be contributing to a vital research project that can improve the quality of life in
Lenawee County. You will hopefully also find the experience interesting and valuable. A
direct benefit will be a free measurement of the level of radon in your home.

126

APPENDIX C: LENAWEE COUNTY RESIDENTIAL
RADON SURVEY QUESTIONNAIRE

HH Survey No. Test Kit Serial Number

 

Geologic Risk: (B)_(Sf)_(Sa)_ Measurement Result

 

additional soil information,
soil type/well log record:

 

Placement date: / /
Retrieval date: / /

Location of device: Rating Value:

 

 

1. How many people live in this house:

2. Number and age of males No. Age Sm. females No. Age Sm.
1 _ () 1 _ ()
2 ) 2
3 ) 3
4 ) 4
5 ) 5

AAA/K
vvvv

_(
_(
_(
_(

3. How many of these people smoke cigarettes:

 

4. Who are the smokers: check ofi‘ above

5. How long have you lived in this house:
(when did you come here)

6. Type of house: 1 story() 2 story( ) split( ) ranch()

7. Housing material: brick( ) aluminum( ) wood( ) mixed b( ) a( )w( )
8. When was this house built:

9. Do you rent() own() lease() this house?

10. Do you have a basement here: yes( ) no( )

l 1. Does anyone sleep( ) or watch TV( ) in the basement

127
12. Number of beds( ) or couches( ) in basement:

13. Where is your TV: 1 2 3
14. Watch in day( ) or evenings ( )
15. Do you go out to work: yes() no()

16. Do you speak Spanish: yes( ) no() formerly() parents( )
(do you listen to Hispanic radio)

17. Race/ethnicity of homeowner:

 

18. How did you first hear of radon: TV( ) radio() paper() never( )
word of mouth( ) this survey( )

19. Education completed by home owner: 1-8 (_) 9-12 (_) college( ) higher( )
(What is your school grade level)

20. Furnace: yes() no() oil( ) gas() electric()

21. Basement well: yes( ) no( ) depth

22. Sump pump: yes no

23. Basement damp: yes() no() minimal() moderate() extreme()

24. Foundation: poured() block() gravel() other none()

25. Basement cracks: floor yes() no() minor() major()
walls YES() 000 minor() minor()
N() S() E() W()
26. Floor material: ; wall material:
27. walkout: yes no
28. fireplace: yes no
29. previously tested: yes, measurement result no
Other notes and features:

128

APPENDIX D: EPA RADON MEASUREMENT PROFICIENCY
TRAINING CERTIFICATE

ASEQ’S
The University of Illinois at Chicago
School of Public Health

MIDWEST ENVIRONMENTAL AND INDUSTRIAL HEALTH TRAINING CENTER

mmmumr Tar-k 9- Joseph
ass-ss—rssr

Has Attended the Continuing Education Course
RQDDN MEASUREMENT DPERRTOR

DROP I C I ENCY

“a.” 07/15/93

him
COURSE Df-lTE: 07/14/93 To 37/15/93

111an snow It £9201“. mecca-{.3 ._/-){’fl._

M
"DESI “11551113 III)! Will! Acting DIIOCIO'
Outreach and Con-wing Education

...... QM

Schoolofpubicm

 

129
APPENDIX E: RESULTS/FOLLOW UP LETTER

LENAWEE COUNTY RESIDENTIAL RADON SURVEY
MEASUREMENT RESULTS

HH'

Serial Number:
Dear Survey Participant,

The result of the short-term radon test recently taken in your home
is: pCi/L.

To assist you with understanding this reading, I have enclosed the “Measurement
Results Interpretation Guide.”

All measurements are reported in units of picocuries per liter of air (pCi/L). This
represents the average amount of radon in a liter of air over the measurement period. This
measurement was taken during the “closed house” winter heating season and provides a worst
case scenario for the amount of radon in your home. The Environmental Protection Agency
radon “action level” is 4.0 pCi/L for occupied levels of the home. Ifthe measurement result
from your home meets or exceeds 4.0 pCi/L, the EPA recommends taking further steps. The
enclosed information outlines EPA recommendations.

Short term tests are generally referred to as radon screening measurements. These
tests provide initial information as to whether or not your home may have a potential indoor

radon problem.

If you have any questions please call the State of Michigan Indoor Radon Ofiice at
1-800-723-6642, or the Lenawee County Health Department at 264-5218.

Thank you for participating in the Lenawee County Residentail Radon Survey.

Sincerely,

Tarek A. Joseph
Study Coordinator

 

130
APPENDIX F: RADON REDUCTION METHODS
Radon reduction methods (mitigation techniques) are divided into four main categories:
Sealing of Entry Routes, House Ventilation, House Pressure Control, and Soil Gas
Suction
Sealing of Entry Routes

This method involves the sealing radon entry routes to prevent radon fi'om entering a
home.

Sealing should include: all cracks in floors and walls, untrapped floor drains, sump casings,
tops of hollow blocks, pipe openings, and exposed soil.

This sealing method can be a relatively economical method for radon mitigation.

House Ventilation

Amounts of indoor radon can be reduced by drawing in outside air thereby increasing
house ventilation or air exchange.

House ventilation can be increased by opening lower-level windows, doors, and vents, and
by using fans to move outdoor air into a house.

This method is relatively inexpensive.

House Pressure Control

These radon reduction methods involve steps to minimize or eliminate house
depressurization. Lower air pressure in a house “attracts” radon from the surrounding
soil. If the air pressure in a house can be controlled, radon entry can be reduced.

The use of household appliances contributes to house depressurization. It is
recommended to open a window or increase ventilation near appliances when in use.
Furnaces and water heaters can also cause house depressurization. Installing a system for
supplying outdoor air will help reduce house depressurization. Another reduction method
is to seal ff cold air return registers that are in the basement or a crawl space.

These are simple, low-cost, corrective steps.

Soil Gas Suction

The objective of soil gas suction is to reduce the air pressure under a home to a pressure
lower than that in the house.

Subslab suction involves insertion of pipes through the slab into roak or soil. A fan draws
siol gas from under the slab to the outside air. Block wall suction removes radon from the
uncapped openings of hollow blocks before it can enter a house. This is done by creating
a zone of lower pressure that reverses the direction of soil gas flow.

 

These soil gas suction methods are more expensive, but highly effective in preventing the
entry of radon into homes.

EPA. Radon Reduction Techniques for Detached Houses: 1988: EPA. Radon Reduction Methods. 1989; Cliff. Green. and Lamas. I992.