950R91034
United States
Environmental Protection
Agency
Research and Development
Air and Energy Environmental
Research Laboratory
Research Triangle Park NC 27711
April 1991
The 1991 International
Symposium on Radon
and Radon Reduction
Technology:
Volume 3. Preprints
Session V: Radon Entry Dynamics
Session VI: Radon Surveys
April 2-5,1991
Adam's Mark Hotel
Philadelphia, Pennsylvania
-------
The 1991 International
Symposium on Radon and
Radon Reduction Technology
"A New Decade of Progress"
April 2-5,1991
Adam's Mark Hotel
Philadelphia, Pennsylvania
Sponsored by:
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
and
U. S. Environmental Protection Agency
Office of Radiation Programs
and
Conference of Radiation Control Program Directors, Inc.
(CRCPD)
^ Printed on Recycled Paper
-------
The 1991 International Symposium on Radon
and Radon Reduction Technology
Opening Session
Opening Remarks Symposium Co-Chairpersons
Introduction Charles M. Hardin, CRCPD, Inc.
Welcome Edwin B. Erickson,
EPA Region III Administrator
An Overview of the NAS Report on Radon Dosimetry Jonathan Samet
New Mexico Tumor Registry
iii
-------
The 1991 International Symposium on Radon
and Radon Reduction Technology
Table of Contents
Session I: Government Programs and Policies Relating to Radon
The Need for Coordinated International Assessment of the
Radon Problem - the IAEA Approach
Friedrich Steinhausler, International Atomic Energy Agency, Austria 1-1
The European Research Program and the Commission of
European Communities
Jaak Sinnaeve, Belgium I-2
United Kingdom Programs
Michael O'Riordan, National Radiological Protection Board, UK I-3
The U.S. DOE Radon Research Program: A Different Viewpoint
Susan L. Rose, Office of Energy Research, U. S. DOE I-4
U.S. EPA Future Directions
Margo Oge, U.S. EPA, Office of Radiation Programs 1-5
Session I Posters
The State Indoor Radon Grants Program: Analysis of Results
After the First Year of Funding
Sharon Saile, U.S. EPA, Office of Radiation Programs IP-1
EPA Radon Policy and Its Effects on the Private Sector
David Saum and Marc Messing, INFILTEC ip-2
Evaluation of EPA's National Radon Contractor Proficiency Program
and Network of Regional Radon Training Centers
G. Lee Salmon, U.S. EPA, Office of Radiation Programs IP-3
State Certification Guidance
John Hoornbeek, U.S. EPA, Office of Radiation Programs IP-4
The U.S. EPA Radon Measurement Proficiency (RMP) Program
Jed Harrison, U.S. EPA, Office of Radiation Programs IP-5
iv
-------
Session II: Radon-Related Health Studies
Residential Radon Exposure and Lung Cancer in Women
Goran Pershagen, Karolinska Institute, Sweden 11-1
An Evaluation of Ecologic Studies of Indoor Radon and Lung Cancer
Christine Stidley, University of New Mexico II-2
Comparison of Radon Risk Estimates
Richard Hornung, NIOSH II-3
Lung Cancer in Rats Exposed to Radon/Radon Progeny
F. T. Cross and G. E. Dagle, Pacific Northwest Laboratory II-4
Startling Radon Risk Comparisons
JoAnne D. Martin, DMA-RADTECH, Inc II-5
Estimated Levels of Radon from Absorbed Polonium-210 in Glass
Hans Vanmarcke, Belgium H-6
Expanded and Upgraded Tests of the Linear-No Threshold Theory
for Radon-Induced Lung Cancer
Bernard L. Cohen, University of Pittsburgh II-7
Session II Panel: Risk Communication
Apathy vs. Hysteria, Science vs. Drama: What Works in Radon
Risk Communication
Peter Sandman, Rutgers University II-8
American Lung Association's Radon Public Information Program
Leyla Erk McCurdy, American Lung Association II-9
Ad Council Radon Campaign Evaluation
Mark Dickson and Dennis Wagner, U.S. EPA, Office of
Radiation Programs 11-10
Developing a Community Radon Outreach Program: A Model for
Statewide Implementation
M. Jeana Phelps, Kentucky Cabinet for Human Resources 11-11
v
-------
Session II Posters
Occupational Safety During Radon Mitigation: Field Experience and
Survey Monitoring Results
Jean-Claude F. Dehmel, S. Cohen & Associates; Peter Nowlan,
R. F. Simon Company; Eugene Fisher, U.S. EPA,
Office of Radiation Programs
Consumer Cost/Benefit Analysis of Radon Reductions in 146 Homes
Kenneth D. Wiggers, American Radon Services, Ltd HP_2
The Effect of Passive Cigarette Smoke on Working Level
Exposures in Homes
Raymond H. Johnson, Jr. and Randolph S. Kline, Key Technology, Inc.;
Eric Geiger and Augustine Rosario, Jr., Radon QC HP_3
Session III: Measurement Methods
Quality Assurance of Radon and Radon Decay Product Measurements
During Controlled Exposures
Douglas J. Van Cleef, U.S. EPA, Office of Radiation Programs 111-1
Current Status of Glass as a Retrospective Radon Monitor
Richard Lively, MN Geological Survey, and Daniel Steck,
St. John's University ||t-2
Soil Gas Measurement Technologies
Harry E. Rector, GEOMET Technologies, Inc |||-3
Results From a Pilot Study to Compare Residential Radon Concentrations
with Occupant Exposures Using Personal Monitoring
B. R. Litt, New Jersey Department of Environmental Protection,
J. M. Waldman, UMDNJ, and N. H. Harley and P. Chittaporn,
New York University Medical Center W-4
Rapid Determination of the Radon Profile in a Structure by
Measuring Ions in the Ambient Atmosphere
W. G. Buckman and H. B. Steen III, Western Kentucky University III-5
Intercomparison of Activity Size Distribution Measurements with
Manual and Automated Diffusion Batteries - Field Test
P. K. Hopke and P. Wasiolek, Clarkson University; E. O. Knutson,
K. W. Tu, and C. Gogolak, U. S. DOE; A. Cavallo and K. Gadsby,
Princeton University; D. Van Cleef, NAREL W-6
vi
-------
Influence of Radon Concentrations on the Relationship Among Radon
Measurements Within Dwellings
Judith B. Klotz, NJ State Department of Health 111-7
The Use of Indoor Radon Measurements and Geological Data in Assessing
the Radon Risk of Soil and Rock in Construction Sites in Tampere
Anne Voutilainen and llona Makelainen, Finnish Centre for
Radiation and Nuclear Safety 111-8
Session III Panel: Detection of Radon Measurement Tampering
Policy and Technical Considerations for the Development of EPA
Guidance on Radon and Real Estate
Lawrence Pratt, U. S. EPA, Office of Radiation Programs 111-9
State Property Transfer Laws Now Include Radon Gas Disclosure
Michael A. Nardi, The Nardi Group 111-10
Update of AARST Real Estate Testing Guidelines
William P. Brodhead, WPB Enterprises 111-11
Real Estate Transaction Radon Testing Interference
Dean Ritter, ABE Testing 111-12
How to Determine if Radon Measurement Firms are Providing
Accurate Readings
Herbert C. Roy and Mohammed Rahman, New Jersey Department
of Environmental Protection 111-13
Grab Sampling as a Method of Discovering Test Interference
Marvin Goldstein, Building Inspection Service, Inc 111-14
Exploring Software Device Management Routines that Ensure the Overall
Quality of Continuous Working Level and Continuous Radon Monitor
Performance in a Field Environment
Richard Tucker, Gemini Research, and Rick Holland,
Radonics, Inc 111-15
Use of Grab Samples as a Quality Assurance Tool to Enhance Overall
Radon Measurement Accuracy and Reproducibility
Brian Fimian, Radonics, Inc., and Richard Tucker, Gemini Research 111-16
vii
-------
Session III Panel: Short-Term/Long-Term Measurement
Predicting Long-Term Indoor Radon Concentrations from Short-Term
Measurements: Evaluation of a Method Involving Temperature Correction
T. Agami Reddy, A. Cavallo, K. Gadsby, and R. Socolow,
Princeton University 111-17
Correlation Between Short-Term and Long-Term Indoor Radon
Concentrations in Florida Houses
Susan E. McDonough, Southern Research Institute IH-18
Relationship Between Two-day Screening Measurements of Radon-222
and Annual Living Area Averages in Basement and
Nonbasement Houses
S. B. White, N. F. Rodman, and B. V. Alexamder, Research Triangle
Institute; J. Phillips and F. Marcinowski, U. S. EPA, Office of
Radiation Programs
The Use of Multiple Short-Term Measurements to Predict Annual Average
Radon Concentrations
Frank Marcinowski, U. S. EPA, Office of Radiation Programs 111-20
Session III Posters
Characterization of Structures Using Simultaneous Single Source
Continuous Working Level and Continuous Radon Gas Measurements
Brian Fimian and John E. McGreevy, Radonics, Inc IIIP-1
Pennsylvania Department of Environmental Resources Radon in Water
Measurement Intercom pari son
Douglas Heim and Carl Granlund, Pennsylvania Department of
Environmental Resources IUP-2
A Field Comparison of Several Types of Radon Measurement Devices
Elhannan L. Keller, Trenton State College IIIP-3
Radon and Water Vapor Co-Adsorption on Solid Adsorbents
Neguib M. Hassan, Tushar K. Ghosh, Sudarshan K. Loyalka,
and Anthony L. Hines, University of Missouri-Columbia, and
Davor Novosel, Gas Research Institute IIIP-4
Calibration of Modified Electret Ion Chamber for Passive Measurement
of Radon-222 (Thoron) in Air
P. Kotrappa and J. C. Dempsey, Rad Elec, Inc IHP-5
viii
-------
Unit Ventilator Operation and Radon Concentrations in a
Pennsylvania School
William P. Brodhead, WPB Enterprises IIIP-6
Session IV: Radon Reduction Methods
Causes of Elevated Post-Mitigation Radon Concentrations in Basement
Houses Having Extremely High Pre-Mitigation Levels
D. Bruce Henschel, AEERL; Arthur G. Scott, AMERICAN
ATCON, Inc IV-1
A Measurement and Visual Inspection Critique to Evaluate the
Quality of Sub-Slab Ventilation Systems
Richard W. Tucker, Gemini Research, Inc.; Keith S. Fimian,
Radonics, Inc IV-2
Correlation of Diagnostic Data to Mitigation System Design and
Performance as Related to Soil Pressure Manipulation
Ronald F. Simon, R. F. Simon Company IV-3
Pressure Field Extension Using a Pressure Washer
William P. Brodhead, WPB Enterprises IV-4
A Variable and Discontinuous Subslab Ventilation System and Its
Impact on Radon Mitigation
Willy V. Abeele, New Mexico Environmental Improvement Division IV-5
Natural Basement Ventilation as a Radon Mitigation Technique
A. Cavallo, K. Gadsby, and T.A. Reddy, Princeton University IV-6
Attic Pressurization - A Radon Mitigation Technique for Residential Structures
Myron R. Edelkind, Southern Mechanical IV-7
Section IV Posters
Radon Mitigation Failure Modes
William M. Yeager, Research Triangle Institute; D. Bruce Harris, AEERL;
Terry Brennan and Mike Clarkin, Camroden Associates, Inc IVP-1
Mitigation by Sub-SlabDepressurization Under Structures Founded on
Relatively Impermeable Sand
Donald A. Crawshaw and Geoffrey K. Crawshaw, Pelican
Environmental Corporation IVP-2
ix
-------
A Laboratory Test of the Effects of Various Rain Caps on Sub-Slab
Depressurization Systems
Mike Clarkin, Camroden Associates, Inc IVP-3
Analysis of the Performance of a Radon Mitigation System Based on
Charcoal Beds
P. Wasiolek, N. Montassier, P. K. Hopke, Clarkson University;
R. Abrams, RAd Systems, Inc IVP-4
Control of Radon Releases in Indoor Commercial Water Treatment
D. Bruce Harris and A. B. Craig, AEERL IVP-5
Session V: Radon Entry Dynamics
A Modeling Examination of Parameters Affecting Radon and Soil Gas
Entry Into Florida-Style Slab-on-Grade Houses
G. G. Sextro, Lawrence Berkeley Laboratory V-1
Effect of Winds in Reducing Sub-Slab Radon Concentrations Under
Houses Laid Over Gravel Beds
P. C. Owczarski, D. J. Holford, K. W. Burk, H. D. Freeman, and
G. W. Gee, Pacific Northwest Laboratory V-2
Radon Entry Into Dwellings Through Concrete Floors
K. K. Nielson and V. C. Rogers, Rogers and Associates
Engineering Corporation V-3
Radon Dynamics in Swedish Dwellings: A Status Report
Lynn M. Hubbard, National Institute of Radiation Protection, Sweden V-4
Soil Gas and Radon Entry Potentials for Slab-on-Grade Houses
Bradley H. Turk, New Mexico; David Grumm, Yanxia Li, and Stephen
D. Schery, New Mexico Institute of Mining and Technology;
D. Bruce Henschel, AEERL V-5
Direct Measurement of the Dependence of Radon Flux Through
Structure Boundaries on Differential Pressure
D. T. Kendrick and G. Harold Langner, Jr., U.S. DOE/Chem-Nuclear
Geotech, Inc V-6
Radon Resistance Under Pressure
William F. McKelvey, Versar, Inc.; Jay W. Davis, Versar A/E, Inc V-7
Recommendations to Reduce Soil Gas Radon Entry Based on an
Evaluation of Air Permeability of Concrete Blocks and Coatings
J. S. Ruppersberger, U. S. EPA, Office of Research and Development V-8
X
-------
Session V Posters
A Simple Model for Describing Radon Migration and Entry Into Houses
Ronald B. Mosley, AEERL VP-1
Effect of Non-Darcy Flow on the Operation of Sub-Slab
Depressurization Systems
R. G. Sextro, Lawrence Berkeley Laboratory VP-2
Effects of Humidity and Rainfall on Radon Levels in a Residential Dwelling
Albert Montague and William E. Belanger, U. S. EPA;
Francis J. Haughey, Rutgers University VP-3
Session VI: Radon Surveys
Factors Associated with Home Radon Concentrations in Illinois
Thomas J. Bierma and Jennifer O'Neill, Illinois State University VI-1
Radon in Federal Buildings
Michael Boyd, U. S. EPA, Office of Radiation Programs VI-2
Radon in Switzerland
H. Surbeck and H. Volkle, University Perolles; W. Zeller, Federal
Office of Public Health VI-3
A Cross-Sectional Survey of Indoor Radon Concentrations in 966 Housing
Units at the Canadian Forces Base in Winnipeg, Manitoba
D. A. Figley and J. T. Makohon, Saskatchewan Research Council VI-4
Radon Studies in British Columbia, Canada
D. R. Morley and B. G. Phillips, Ministry of Health; M. M. Ghomshei,
Orchard Geothermal Inc.; C. Van Netten, The University of
British Columbia VI-5
The State of Maine Schools Radon Project; Results
L. Grodzins, NITON Corporation; T. Bradstreet, Division of Safety
and Environmental Services, Maine; E. Moreau, Department of
Human Services, Maine VI-6
Radon in Belgium: The Actual Situation and Plans for the Future
A. Poffijn, State University of Gent VI-7
A Radiological Study of the Greek Radon Spas
P. Kritidis, Institute of Nuclear Technology - Radiation Protection VI-8
xi
-------
Session VI Posters
A Cumulative Examination of the State/EPA Radon Survey
Jeffrey Phillips, U. S. EPA, Office of Radiation Programs VIP-1
Seasonal Variation in Two-Day Screening Measurements of Radon-222
Nat F. Rodman, Barbara V. Alexander, and S. B. White, Research
Triangle Institute; Jeffrey Phillips and Frank Marcinowski,
U. S. EPA, Office of Radiations Programs VIP-2
The State of Maine School Radon Project: Protocols and Procedures of
the Testing Program
Lee Grodzins and Ethel G. Romm, NITON Corporation;
Henry E. Warren, Bureau of Public Improvement, Maine VIP-3
Results of the Nationwide Screening for Radon in DOE Buildings
Mark D. Pearson, D. T. Kendrick, and G. H. Langner, Jr., U. S. DOE/
Chem-Nuclear Geotech, Inc VIP-4
Session VII: State Programs and Policies Relating to Radon
Washington State's Innovative Grant: Community Support Radon Action
Team for Schools
Patricia A. McLachlan, Department of Health, Washington VII-1
Kentucky Innovative Grant: Radon in Schools Telecommunication Project
M. Jeana Phelps, Kentucky Cabinet for Human Resources;
Carolyn Rude-Parkins, University of Louisville VII-2
Regulation of Radon Professionals by States: the Connecticut Experience
and Policy Issues
Alan J. Siniscalchi, Zygmunt F. Dembek, Nicholas Macelletti, Laurie
Gokey, and Paul Schur, Connecticut Department of Health Services;
Susan Nichols, Connecticut Department of Consumer Protection;
Jessie Stratton, State Representative, Connecticut
General Assembly VII-3
New Jersey's Program - A Three-tiered Approach to Radon
Jill A. Lapoti, New Jersey Department of Environmental Protection VIM
Session VII Posters
Quality Assurance - The Key to Successful Radon Programs in the 1990s
Raymond H. Johnson, Jr., Key Technology, Inc VIIP-1
xii
-------
Radon in Illinois: A Status Report
Richard Allen and Melanie Hamel-Caspary, Illinois Department
of Nuclear Safety VIIP-2
Session VIII: Radon Prevention in New Construction
Long-Term Monitoring of the Effect of Soil and Construction Type on
Radon Mitigation Systems in New Houses
D. B. Harris, U. S. EPA, Office of Research and Development VIII-1
A Comparison of Indoor Radon Concentrations Between Preconstruction
and Post-Construction Mitigated Single Family Dwellings
James F. Burkhart, University of Colorado at Colorado Springs;
Douglas L. Kladder, Residential Service Network, Inc VIII-2
Radon Reduction in New Construction: Double-Barrier Approach
C. Kunz, New York State Department of Health VIII-3
Radon Control - Towards a Systems Approach
R. M. Nuess and R. J. Prill, Washington State Energy Office VIII -4
Mini Fan for SSD Radon Mitigation
David Saum, INFILTEC VIII-5
Building Radon Mitigation into Inaccessible Crawlspace New
Residential Construction
D. Bruce Harris and A. B. Craig, AEERL; Jerry Haynes, Hunt
Building Corporation VIII-6
The Effect of Subslab Aggregate Size on Pressure Field Extension
K. J. Gadsby, T. Agami Reddy, D. F. Anderson, and R. Gafgen,
Princeton University; A. B. Craig, AEERL VIII-7
Session VIII Posters
Radon Prevention in Residential New Construction: Passive Designs
That Work
C. Martin Grisham, National Radon Consulting Group VIIIP-1
Preliminary Results of HVAC System Modifications to Control Indoor
Radon Concentrations
Terry Brennan and Michael Clarkin, Camroden Associates;
Timothy M. Dyess, AEERL; William Brodhead, Buffalo Homes VIIIP-2
xiii
-------
Correlation of Soil Radon Availability Number with Indoor Radon and
Geology in Virginia and Maryland
Stephen T. Hall, Radon Control Professionals, Inc VIIIP-3
Session !X: Radon Occurrence in the Natural Environment
Combining Mitigation and Geology: Indoor Radon Reduction by
Accessing the Source
Stephen T. Hall, Radon Control Professionals, Inc ix-1
A Comparison of Radon Results to Geologic Formations for the
State of Kentucky
David McFarland, Merit Environmental Services |X-2
Geologic Radon Potential of the United States
Linda Gunderson, U. S. Geological Survey IX-3
Technological Enhancement of Radon Daughter Exposures Due to
Non-nuclear Energy Activities
Jadranka Kovac, University of Zagreb, Yugoslavia IX-4
A Site Study of Soil Characteristics and Soil Gas Radon
Richard Lively, Minnesota Geological Survey; Daniel Steck,
St. John's University IX-5
Geological Parameters in Radon Risk Assessment - A Case History
of Deliberate Exploration
Donald Carlisle and Haydar Azzouz, University of California
at Los Angeles IX-6
Session IX Posters
Geologic Evaluation of Radon Availability in New Mexico: A Progress Report
Virginia T. McLemore and John W. Hawley, New Mexico Bureau of Mines
and Mineral Resources; Ralph A. Manchego, New Mexico
Environmental Improvement Division IXP-1
Paleozoic Granites in the Southeastern United States as Sources
of Indoor Radon
Stephen T. Hall, Radon Control Professionals, Inc IXP-2
Comparison of Long-Term Radon Detectors and Their Correlations with
Bedrock Sources and Fracturing
Darioush T. Gharemani, Radon Survey Systems, Inc IXP-3
xiv
-------
Geologic Assessment of Radon-222 in McLennan County, Texas
Mary L. Podsednik, Law Engineering, Inc IXP-4
Radon Emanation from Fractal Surfaces
Thomas M. Semkow, Pravin P. Parekh, and Charles O. Kunz, New York
State Department of Health and State University of New York
at Albany; Charles D. Schwenker, New York State Department
of Health IXP-5
National Ambient Radon Study
Richard Hopper, U. S. EPA, Office of Radiation Programs IXP-6
Session X: Radon in Schools and Large Buildings
The Results of EPA's School Protocol Development Study
Anita L. Schmidt, U. S. EPA, Office of Radiation Programs X-1
Diagnostic Evaluations of Twenty-six U. S. School - EPA's School
Evaluation Program
Gene Fisher, U. S. EPA, Office of Radiation Programs X-2
Extended Heating, Ventilating and Air Conditioning Diagnostics in
Schools in Maine
Terry Brennan, Camroden Associates X-3
Mitigation Diagnostics: The Need for Understanding Both HVAC and
Geologic Effects in Schools
Stephen T. Hall, Radon Control Professionals, Inc X-4
A Comparison of Radon Mitigation Options for Crawl Space School Buildings
Bobby E. Pyle, Southern Research Institute; Kelly W. Leovic, AEERL X-5
HVAC System Complications and Controls for Radon Reduction in
School Buildings
Kelly W. Leovic, D. Bruce Harris, and Timothy M. Dyess, AEERL;
Bobby E. Pyle, Sourthe Research Institute; Tom Borak, Western
Radon Regional Training Center; David W. Saum, INFILTEC X-6
Radon Diagnosis and Mitigation of a Large Commercial Office Building
David Saum, INFILTEC X-7
New School Radon Abatement Systems
Ronald F. Simon, R. F. Simon Company X-8
Design of Radon-Resistant and Easy-to-Mitigate New School Buildings
Alfred B. Craig, Kelly W. Leovic, and D. Bruce Harris, AEERL X-9
XV
-------
Session X Posters
Design and Application of Active Soil Depressurization (ASD) Systems
in School Buildings
Kelly W. Leovic, A. B. Craig, and D. Bruce Harris, AEERL; Bobby E.
Pyle, Southern Research Institute; Kenneth Webb, Bowling Green
(KY) Public Schools XP-1
Radon in Large Buildings: Pre-Construction Soil Radon Surveys
Ralph A. Llewellyn, University of Central Florida XP-2
Radon Measurements in North Dakota Schools
Thomas H. Morth, Arlen L. Jacobson, James E. Killingbeck,
Terry D. Lindsey, and Allen L. Johnson, North Dakota State
Department of Health and Consolidated Laboratories XP-3
Major Renovation of Public Schools that Includes Radon Prevention:
A case Study of Approach, System Design and Installation, and
Problems Encountered
Thomas Meehan XP-4
The State of Maine School Radon Project: The Design Study
Henry E. Warren, Maine Bureau of Public Improvement;
Ethel G. Romm, NITON Corporation XP-5
Design for the National Schools Survey
Lisa Ratcliff, U. S. EPA, Office of Radiation Programs XP-6
xvi
-------
Session V:
Radon Entry Dynamics
-------
V-l
TITLE: A Modeling Examination of Parameters Affecting Radon and Soil
Gas Entry Into Florida-style SIab-on-grade Houses
AUTHOR:
G.G. Sextro, Lawrence Berkeley Laboratory
This paper was not received included ^in^the
you^registration packet, for a complete copy of the paper.
Entry of soil gas and radon into structures is dependent upon a number of variables
including driving pressures, the air permeability and geometry of the building substruc-
ture and any adjacent soil or gravel layers, and the radon generation rate in the soil
Residential building substructures in many areas of Florida often consist of & perimeter
stem vail constructed of hollow-core concrete blocks and an above-grade floor slab resting
on backfill, or in some cases, a monolithic concrete slab and footer. When the building is
depressurized with respect to the ambient pressure, radon-bearing soil air Mill flow
through various combinations of soil, backfill and blockwall components, entering the
house through perimeter slab-stem wall gaps or interior cracks or openings in the floor
slab.
We have examined the influence of soil, backfill and construction characteristics on
radon and soil gas entry using a two- dimensional finite difference model employing
cylindrical symmetry. At a constant building depressurbation, steady-state pressure
flow, and radon concentration fields are predicted for a soil block 10 m deep and ex-tend-
ing 10 m beyond the 7-m-diameter slab.
Under basecaae conditions, approximately 02 percent of the soil gas entry is through
the exterior section of the stem wall, 5 percent is through the interior section of the stem
wall, 2 percent through an interior slab opening and less than 1 percent through gaps
assumed to exist between the stem wall and footer or the stem wall and floor slab. In
contrast, 57 percent of the radon entry rate occurs through the interior section of the
stem wall, 22 percent through the interior slab opening, 20 percent through the exterior
section of the stem wall, and less than 0.5 percent through the gaps, Changes in backfill
permeability have significant effects on radon entry, while changes in blockwall permeabil-
ity are somewhat offset by increased flow and entry through structural gaps. These
results, along with those from other model configurations will be discussed.
-------
EFFECT OF WINDS IN REDUCING SUB-SLAB RADON CONCENTRATIONS
UNDER HOUSES LAID OVER GRAVEL BEDS
by: P. C. Owczarski
D. J. Holford
K. W. Burk
H. D. Freeman
G. W. Gee
Pacific Northwest Laboratory
ABSTRACT
Wind pressures on houses can affect the availability of radon beneath
houses that are surrounded by highly permeable materials. The Rn3D computer
code was used in a two-dimensional study of the effects of winds on radon
concentration profiles beneath a slab-on-grade house when dry gravel of
various thicknesses surrounded the outer surfaces of the slab. Four generic
soil types (sand, silt, loam, and clay) at several moisture saturations under
laid the gravel layer and house. For a typical annual distribution of wind
speeds, radon concentrations under the houses were obtained as functions of
gravel thickness and underlying soil type and saturation. Results show that
for a gravel depth of 0.1 m, radon concentrations are reduced by up to 502 fo
sand and loam, 60Z for silt, and 90Z for clay. At a depth of 0.3 m the
percent reductions are correspondingly 75, 75, 85, and 95Z.
-------
INTRODUCTION
In a recent study (1), we used the Rn3D computer code to examine the
effects of winds in reducing sub-slab radon concentrations under houses laid
over deep homogeneous dry soils with uniform radium concentrations. Only with
gravel was a significant reduction seen. This recent study led to the present
study of the significance of the radon-reducing effect of winds on houses
deliberately laid over low-radium-content gravel beds. The technical concerns
of this study were the effects of controlling parameters: wind speed, gravel
de th and underlying soil type and moisture saturation. The ultimate goal is
to^determine the practicality of using gravel beds in new home construction.
STUDY SYSTEM
To describe this study it is necessary to first discuss the study house
and then the computational tool, the Rn3D code. The properties of the soils
and annual wind speed distribution used are also discussed.
The study building consists of a two-dimensional 15-m wide slab-on-
de house (Figure 1). The soil surrounding the house and gravel bed is
homogeneous directly underlying the slab. The gravel bed is of a uniform
thickness around the edges of the slab. The gravel around the edges is
necessary to ensure a flow path to communicate wind pressures to the sub-slab
gravel.
z
Wind Driven
Flow
House
Rn Flow by Diffusion & Advection
Figure 1. Schematic diagram of the two-dimensional study house
showing gravel bed laid under and alongside the 15-m-wide slab
in uniform thickness. Wind is assumed to be perpendicular to
the side of the house.
-------
Rn3D is a variably saturated, three-dimensional version of CRACK (2,3),
a two-dimensional finite element model that simulates soil and crack pressure
gradients that are caused by atmospheric pressure variations at the ground
surface. The resulting soil gas flow drives the advective transport of radon
gas, affecting the flux density of radon across the soil/air boundary. The
soil in the model may be heterogeneous and anisotropic. Molecular diffusion
of radon is assumed to be governed by Fick's Law and occurs in both the air
and water phase. The flow of soil gas is assumed to be laminar and governed
by Darcy's Law. The water phase is static. The code can be run for one-,
two-, and three-dimensional problems, both steady-state and unsteady-state.
The initial boundary conditions used for the Rn3D code simulation are no
radon concentration gradient at large depth, zero radon flux upward into the
slab, zero radon concentration at the free soil-air surface, and a linear
pressure gradient from windward to leeward under the slab. The wind-induced
pressures ^in Pa) on the free soil surface are (1): 1. Upwind, AP = 0.422U
where U = U(1 - x^/9), U = wind speed (in m/s) above the house, and x = upwind
distance (in m) from the windward house wall. 2. Downwind, AP = -0.1507U2,
where U = U(1 - zz/9) and z = downwind distance (in m) from leeward wall.
The soil properties used in the simulation are as follows. (See
Reference 4 for methods used in calculating these properties.) Radium
activity of 110 Bq/kg was assigned to dry sand, silt, clay, and loam, and
0 Bq/kg was assumed in gravel. The following bulk dry soil densities (kg/m3)
were assumed: gravel 1275, sand 1934, loam 1494, silt 1583, and clay 1290.
The emanation coefficient, e, for each soil is assumed to be the same for each
water saturation level with the following dependence on saturation, S: 0 < S
< 0.2, e = 0.1 + 1.5S; S > 0.2, e = 0.4. This simple representation approxi-
mates the observed behavior of the emanation coefficient in many soils.
Table 1 summarizes the essential soil property inputs used in the Rn3D
runs for this study. All computations were made for soil temperatures of
20°C. The gravel size distribution used here had a geometric mean radius of
0.371 cm and geometric standard deviation of 2.41 (4).
The wind speed distribution used to determine the annual effects of
gravel bed reductions in radon concentration is listed in Table 2. These wind
properties are those of the Richland, Washington, airport.
-------
TABLE 1. SOIL PROPERTIES
Soil type
Gravel
Clay
Saturation
(2)
Porosity
0.519
Permeability
(m2)
1.9 x 10_
0
0.5131
1.0
X
10""
30
4.9
X
10-16
47.5
2.8
X
10-16
65
1.2
X
l0-!6
Silt
0
0.4026
1.5
X
10"15
30
7.6
X
10-15
65
1.8
X
10-15
Loam
0
0.4362
2.0
X
10-13
30
9.8
X
10-"
65
2.4
X
O
l
Sand
0
0.27
3.4 x 10"12
30
1.7 x 10"12
65
3.4 x 10-13
Rn diffusivity
at 20°C (m2/s)
7. 72
X
10"6
1.98
X
10"8
1.51
X
10"8
8.64
X
10-9
3.35
X
10~9
2.57
X
10'6
7.95
X
10"7
1.48
X
10"7
6.89
X
10"6
2 . 50
X
10"6
3.82
X
10"7
7.10
X
10"6
5.20
X
10"6
9.59
X
10"/
TABLE 2. WIND SPEED CLASSES AND ANNUAL PERCENTAGES
Speed class
Percent of
Range
used
in model
class
(mph)
(m/s)
(mph)
annually
1-3
0.95
2
19.8
4-7
2.62
5.5
43.0
8-12
4.77
10.0
23.3
13-18
7.39
15.5
10.6
19-24
10.26
21.5
2.7
>24
(not
modeled)
0.6
-------
RESULTS
We began this study by searching for a suitable gravel bed depth. To
determine an optimum thickness of gravel for radon venting, we examined a case
where the soil type (clay) and saturation (47.52) were held constant and the
gravel bed thickness and wind speed were varied. The Rn3D results for this
case are plotted in Figure 2. In this figure, the slope of each wind speed
curve, representing gravel-region average soil gas concentration versus gravel
bed depth, starts leveling out at a gravel depth of 0.1 m. Increments of
gravel depth at larger gravel depths do not produce as much reduction in radon
concentrations. Even at no wind, there is about a factor of three reduction
in sub-slab radon concentrations for a gravel bed of 0.1-m depth above 47.52
saturated clay.
Figure 3 shows the effects of moisture saturation and wind speed on sub-
slab radon concentration in 0.1-m-deep gravel above clay. The effect of
saturation on percent reduction is small for clay, since clay transports radon
slowly even when dry. Figure 4 for dry soils shows that clay provides the
greatest sensitivity to wind speeds. Since the air permeability and radon
diffusivity in clay are lower than those of the other soils, radon removed by
the wind cannot be replenished into the gravel region as quickly from clay.
100,000
Q 10,000
Wind Speed, m/s
0 (0 mph)
2.62 (5.5 mph)
1000
"O
4.77 (10 mph)
100
10.3 (21.5 mph)
-Q
10
0.7
0.6
0.4
0.5
0.3
0.2
0.1
0
Depth of Gravel, m
Figure 2. Effects of bed depth and wind speed in reducing
average sub-slab radon concentrations. Underlying soil is
47.52 saturated clay.
-------
10,000
1000
100
Percent Soj|
Saturation
>0 47.5%
^ 65%
' ¦ ! - '
^ 0
i 1 . i . i
6 8
Wind Speed, m/s
10
12
14
Figure 3. Effects of wind speed and saturation in reducing
average gravel-region sub-slab radon concentrations. Gravel
bed is 0.1m thick and underlying soil is clay.
100,000
g 10.000
Soil Types
Sand
Loam
Sill
-D
1000
Clay
100
Wind Speed, m/s
Figure 4. Effects of wind speed and underlying dry soil type
in reducing gravel-region sub-slab radon concentrations.
Gravel bed is 0.1 m thick.
-------
As soils become more saturated, this replenishment ability diminishes, as we
see in Figure 5. Here 652 saturated soils show a greater sensitivity to wind
speed than the previous soils.
We now look at the end results of this study where we determined the
annual average reduction in sub-slab radon that could be expected if a gravel
bed underlaid the study house. Using the wind classes of Table 2, we con-
structed Figures 6 and 7. Figure 6 shows the percent reduction that could be
expected from 0.1 m of gravel as a function of saturation. At 652 saturation,
clay reaches 902 reduction, followed by silt at 602 and sand and loam at 502
reduction. Figure 7 shows that even greater percent reductions is achieved
for 0.3 m (1 ft) beds. Here reductions achieved at 652 saturation were 95,
85, 75, and 752 for clay, silt, sand, and loam, respectively. The reader can
easily extrapolate on Figure 7 to gravel beds of depth greater than 0.3 m.
Simplifying assumptions used in producing Figures 6 and 7 were 1) con-
stant temperature (20°C) of air and soil throughout the year and 2) constant
saturation throughout year at each saturation level. Running Rn3D for all
temperature and saturation variations that would occur at a specific site were
beyond the scope of this paper.
Wind Speed, m/s
Figure 5. Effects of wind speed and underlying 652 saturated
soil type in reducing gravel-region sub-slab radon concentrations.
Gravel bed is 0.1 m thick.
-------
100
Soil Types
nl 1 1 ' 1 1— 1 i
° 20 40 ~to —i
Percent Saturation
Figure 6. Effects of 0.1 m gravel hpd or,j,
and soil type in reducing gravel-ret>inn k ®aturati°n,
concentrations for annual spectrum of wi^h ~S radon
H rura ot wind speeds.
Percent Saturation
Figure 7. Effects of gravel bed df>r>t-h
type in reducing gravel-region Sub-slah a saturated soil
for annual spectrum of wind speeds. ^ °n concentrations
-------
CONCLUSIONS
The following general conclusions can be made from the results above and
from the related information:
• Gravel beds underlying slab-on-grade houses can be useful in reducing
sub-slab radon concentrations.
• Similar achieved reductions can be expected from basement houses having
gravel beds if the gravel path length from windward to leeward sides of
the basement is about the same as the slab-on-grade house.
• Use of gravel beds in conjunction with a sub-slab barrier immediately
beneath the slab could achieve greater reductions.
• The presence of gravel beds beneath a slab or basement would make sub-
slab ventilation more effective if needed.
• The beneficial effects of winds can only be achieved if the gravel beds
are well drained. However, there are possible benefits in saturating
the gravel beds. In such a situation radon cannot be easily trans-
ported to entry paths in the house foundations.
• Tightly packed houses in residential neighborhoods might see higher
frequencies of low wind speeds than those of this study.
ACKNOWLEDGMENT
This work was supported by the U.S. Department of Energy under Contract
DE-AC06-76RL0 1830. Battelle Memorial Institute operates the Pacific
Northwest Laboratory for the U.S. Department of Energy.
The work described in this paper was not funded by the U.S. Environ-
mental Protection Agency and therefore the contents do not necessarily reflect
the views of the Agency and no official endorsement should be inferred.
REFERENCES
1. Owczarski, P. C., Holford, D. J., Freeman, H. D., Gee, G. W., and Burk,
K. W. Radon transport from the subsurfaces: the roles of certain
boundary conditions at the subsurface/environment boundaries. Paper
presented at Twenty-Ninth Symposium on Health and the Environment:
Indoor Radon and Lung Cancer: Reality or Myth? Richland, Washington.
October 16-19, 1990.
-------
Hoi ford, D. J., Schery, S. D., Wilson, J. L., and Phillips, F. H. Radon
transport in dry, cracked soil: two-dimensional finite element model.
PNL-7116, Pacific Northwest Laboratory, Richland, Washington, 1989.
Schery, S. D., Holford, D. J., Wilson, J. L., and Phillips, F. M. The
flow and diffusion of radon isotopes in fractured porous media: Part 2,
semi-infinite media. Rad. Protect. Dosimetry 2k: 191-197, 1988.
Owczarski, P. C., Holford, D. J., Freeman, H. D., and Gee, G. W.
Effects of changing water content and atmospheric pressure on radon
flux from surfaces of five soil types. Geophysical Research Letters
17: 817-820, May 1990.
-------
V-3
RADON ENTRY INTO DWELLINGS THROUGH CONCRETE FLOORS
by: K. K. Nielson and V. C. Rogers
Rogers and Associates Engineering Corporation
Salt Lake City, Utah 84110-0330
ABSTRACT
Indoor radon entry commonly is modeled as advection from pressure-driven flow through
foundation cracks and openings. However recent diffusive-advective model analyses suggest that
diffusion through intact concrete floor areas also may be significant. Diffusive radon entry is
characterized further by new measurements of the porosities, radon diffusion coefficients, and air
permeability coefficients of concrete. The measurements include concretes from typical Florida
slab-on-grade housing, and others from industrial mixes that were selected to inhibit radon movement.
Measured radon diffusion coefficients ranged from 9x1 fr6 cm2 s"1 for the industrial concretes to
0.003 cm2 s"1 for floor-slab concretes. Air permeabilities generally were below lxlO"11 cm2. The
diffusion measurements included concretes made from Type I, II, and V cements, and exhibited a
correlation with the water/cement (W/C) ratio of the concretes. The correlation had the form D =
1.8xl0'6 exp(ll.l W/C). The average diffusion coefficient measured for the floor slab concretes
(1.6x103 cm2 s'1) suggested that diffusion may account for indoor concentrations up to 2 pCi L'1 in cases
of soil-gas radon concentrations of 3,000 pCi L K Radium concentrations of 5 pCi g'1 in the concrete
similarly may contribute more than 1 pCi L'1 of indoor radon. Combined diffusion through an intact slab
and through a perimeter floor crack slightly exceeded advective radon entry for a reference house on
sandy soil, and exceeded advective entry rates by more than 100 times for the house on clayey and silty
clay loam soils.
This paper has been reviewed in accordance with the U. S. Environmental Protection Agency's
peer and administrative review policies and approved for presentation and publication.
INTRODUCTION
Indoor radon entry from soil gas has been modeled most commonly as advective transport by
pressure-driven air flow through foundation openings or cracks. The flow is caused by the typically
negative indoor pressure compared with that in the soil and the outdoor atmosphere. Recently, attention
has been directed toward the importance of diffusion as a significant mechanism for radon entry. In
-------
particular, Tanner (1) identified radon diffusion as the dominant entry mechanism when foundation soil
permeabilities are less than 7x1012 m2. Rogers and Nielson (2) also identified diffusion through concrete
floors and the contiguous soil as a significant mechanism for radon entry for many soils under typical
long-term average foundation pressure gradients. Loureiro et al. (3) have compared theoretical diffusive
and advective radon transport in soils to estimate conditions when diffusion is insignificant.
While the diffusive radon flux through concrete floors is much smaller than the advective flux
through cracks in the floor, the predominance of the intact floor area over the crack area may compensate
for the difference in fluxes. Thus, it is desirable to examine the diffusive properties of concretes used
in dwelling floors to better assess this mode of radon entry. It also is instructive to characterize the
relative importance of radon generated within the concrete to determine whether aggregates or other
concrete components may contribute significantly to indoor radon concentrations.
This paper identifies the main properties of concrete that influence radon migration into dwellings
It reports measured values of these parameters for concretes typical of Florida residential floors and also
for industrial concretes specifically selected to restrict radon migration. It then examines the relation of
the measured properties to other physical properties of the concretes. Finally, it examines the relative
importance of the concrete properties, including radium concentrations, to radon entry into dwellings
The radon entry analyses are based on an approximate indoor radon balance equation and on a complete
numerical analysis of combined diffusive-advective radon entry.
CONCRETE CHARACTERISTICS IMPORTANT TO RADON TRANSPORT
The simplified equation describing radon generation and transport in concrete (ignoring radon
adsorption) is given by (2,4):
dC
at
D^c - iS VP • vc - XC +
RpEX
P
(1)
where
C
D
K
/*
P
X
R
P
E
P
v
pore radon concentration (pCi cm"3)
radon diffusion coefficient (cm2 s"1)
permeability coefficient (cm2)
viscosity of air (Pa*s)
three dimensional gradient operator
atmospheric pressure (Pa)
radon decay constant (2.1x10"ft s'1)
radium concentration (pCi g')
bulk density (g cm"', dry)
radon emanation coefficient
porosity.
-------
The diffusive component of radon transport in equation (1) is based on Fick's law, which defines
diffusive radon flux as:
F = -DpVC
where
F = bulk radon flux (pCi cm 2 s'1).
Equations (1) and (2) show that the concrete parameters dominating radon transport are D, K,
and p. The D and K parameters are influenced in turn by the concrete free moisture content, m, and its
mean pore radius, rp. The measurements reported herein locus on these parameters. The parameters
influencing radon generation in concrete are R, E, p, and p. Of these parameters, the influences of only
R and E are examined here in the radon entry calculations with RAETRAN (4) and RAETRAD (2).
Four samples (F-l through F-4) of residential concretes typical of new Florida dwellings were
tested. These samples had a water-to-cement (W/C) ratio of about 0.6. Two additional samples (T-l and
T-2) were obtained of concretes intended to have relatively low radon transport constants (D and K).
These samples had W/C ratios slightly less than 0.4. Bulk densities and effective porosities of the T-l
and T-2 samples were measured. For the (F-l through F-4) samples, bulk densities were measured and
porosities were estimated from the bulk densities by assuming a specific gravity of 2.71. All six of the
samples were taken from separately poured test cylinders, and thus may vary in curing conditions and
finishing from their construction counterparts. The physical properties of the concretes are given in
Table 1.
TABLE 1. PHYSICAL PROPERTIES OF CONCRETE SAMPLES
Sample
ID
Concrete
Type
Water/Cement
Ratio
Density
(g cm'3)
Porosity
F-l
I
0.60
2.15
0.20
F-2
I
0.60
2.14
0.21
F-3
I
0.61
1.99
0.26
F-4
I
0.61
2.00
0.26
T-l
II
0.39
2.40
0.11
T-2
V
0.38
2.35
0.13
PERMEABILITY COEFFICIENT
Permeability coefficients were measured on samples T-l and T-2 using the transient method, as
described previously (5). The permeability coefficient of sample F-4 was measured using a 10 cm
diameter by 5 cm long concrete sample that was tightly secured in a steel sample holder. Simultaneous
measurements were then made of the air flow through the sample and the pressure drop across the
-------
sample. The permeability then was deduced using the one-dimensional form of Darcy's Law:
K _ Qml
(3)
A AP
where
0
L
A
AP
air flow rate through the sample (cm3 s"1)
sample length (cm)
sample cross-sectional area (cm2)
pressure drop across sample (Pa).
Measurements of air permeability for samples T-l, T-2, and F-4 all were less than 1x10" cm3
and did not exhibit a strong dependence on porosity or on the W/C ratios. This suggests that the mean
pore radius for the higher porosity samples may not be significantly larger than for the lower porositv
samples, but that there simply may be more pores. These permeabilities are sufficiently low that radon
advective flow through the intact bulk concrete is negligible, even for very high foundation pressure
gradients.
DIFFUSION COEFFICIENT
Radon diffusion coefficients were measured on all concrete samples using the standard procedure
for porous materials (6). For initial measurements, the samples were equilibrated with laboratory air
and had pore water contents that corresponded to the capillary retention by the concrete pores (7) In
addition, for samples T-l and T-2, diffusion measurements were made on water-saturated samples and
at lower water contents, including zero pore water. The various water contents were achieved by
vacuum-saturating the samples, with subsequent vacuum pumping to reduce water contents to the desired
levels (measured gravimetrically). The results of the diffusion measurements are shown in Figure 1
There is considerable uncertainty in the water contents of samples F-l through F-4.
The measurements at zero pore water provide estimates of the Knudsen diffusion coefficient and
the associated effective pore radius for samples T-l and T-2 (8). The total diffusion coefficient for the
sample pores combines the Knudsen diffusion coefficient and the diffusion coefficient of the pore fluid
with the relationship (7,8):
1 = _L
D " Dk
D,
c
(4)
where
Dk
Dc
Knudsen diffusion coefficient (cm2 s"1)
diffusion coefficient of the pore fluid (cm2 s'1)-
Using the prior, predictive correlation for the diffusion coefficient in soil pores to represent D
(9), and using equation (4) to estimate the total diffusion coefficient give the two curves shown in Figure
1. The curves agree with the data for T-l and T-2 to within the experimental diffusion measurement
-------
10
•2
10"
10^
v>
CM
E
c
a>
o
E
a>
o
o
c
o
CO
3
IE 10
c
O
T3
TO
£E
10J
1 1 ' 1
¦ i ¦ i 1 :
+
4.
>»
N»
_ N
S
• ; ;
• T-1 (Type II)
\\
O T-2 (Type V)
V
1 >
+ F1 - F4 (Type 1)
J
O
T-1 calc'd
— T-2 calc'd
¦ ¦ '
¦ - 1
0.0 0.2 0.4 0.6 0.8 1.0
Water Content (fraction of saturation)
RAE-103423
Figure 1. Measured radon diffusion coefficients for the six concrete samples, with
comparisons to soil-based water-dependence trends for samples T-l and
T-2.
uncertainty. If it is assumed that the mean pore radius in samples F-l through F-4 scales with the mean
pore radius in sample T-2 according to the ratio of the sample porosities, then similar diffusion coefficient
predictions can be made for samples F-l through F-4 that have similar degrees of accuracy as the curves
for samples T-l and T-2.
It is also of interest to plot the concrete diffusion coefficients for samples equilibrated with
ambient air as a function of their W/C ratios. The diffusion data, shown in Figure 2, correlate well with
the W/C ratios to yield the least-squares fitted expression:
D = 1.8X10'6 exp(ll.l W/C).
(5)
-------
V)
CM
E
o
c
0>
"5
•*-
a;
O
O
c
o
3
C
O
"O
m
tr
F1 - F4 (Type I)
¦ T-1 (Type II)
~ T-2 (Type V)
O Culot et a! (10)
1.8E-6 e*p( 11 1 W/C)
0.35 0.40 0.45 0.50 0.55 0.60 0.65
Concrete Water/Cement Ratio
RAE-103424
Figure 2. Regression of ambient-moisture radon diffusion measurements on the
water/cement ratio of the initial concrete mixture.
The correlation coefficient associated with this fit is r = 0.96. Also included in Figure 2 and in the
determination of equation (5) is the radon diffusion coefficient for concrete reported by Culot et al. (10).
Equation (5) is also an excellent predictor for the value reported by Culot, and suggests an important
dependence of the radon diffusion coefficient on the W/C ratio of the concrete.
SIGNIFICANCE OF RADON DIFFUSION THROUGH CONCRETE FLOORS
The significance of indoor radon entry by diffusion through concrete floors can be estimated from
a simplified approximation of the indoor radon balance equation. The approximation assumes that all
indoor radon enters via the concrete foundation area, and that the indoor volume is uniformly diluted at
the continuous rate of \ with clean air having an insignificant radon concentration:
-------
J • A = CVXv
(6)
where
J = radon flux entering the house foundation (pCi m"2 s"1)
A = house foundation area over which radon entry occurs (m2)
C = steady-state indoor radon concentration (pCi L"1)
V = indoor volume (L)
Xv = ventilation rate of indoor volume (s1)-
For a simple slab-on-grade house geometry typical of Florida construction, equation (6) can be simplified
further by introducing the indoor height as the ratio of the house volume to its area. This leads to the
expression:
C = 10"3 J / (h X ) (7)
where
h = height of indoor volume (m)
103 = L per m3 conversion.
Based on the approximate separability of diffusive and advective radon entry into dwellings,
equation (7) can be used directly to estimate the component of the indoor radon concentration that results
from diffusion through the concrete floor slab. The diffusive radon flux through the slab was estimated
by repeated analyses with the RAETRAN code (4), in which a 10-cm slab separated an indoor radon
concentration of 2 pCi L'1 from 5 m of sandy foundation soil that had varying source strengths
corresponding to deep-soil radon concentrations of 100 pCi L'1 to 10,000 pCi L'1. Various diffusion
coefficients also were used for the slab, which had a fixed porosity of p=0.23, corresponding to the
average porosity of the Florida concrete samples (F-l to F-4). The radium concentration in the concrete
first was assumed to be 1 pCi g', and to have a radon emanation coefficient of 0.25.
The resulting radon fluxes from the slab were divided by an indoor height of 2.3 m and a
ventilation rate of 1.4xl0"4 s'1 (0.5 h'1) to estimate the diffusive component of the indoor radon
concentration (equation 7). The resulting indoor concentrations (Figure 3) start to exceed the 1 pCi L1
level for elevated soil gas radon concentrations (several thousand pCi L"1) only when concrete diffusion
coefficients exceed 3xl0'4 cm2 s'1 or higher. Using concrete diffusion coefficients similar to the 3.4xl0'4
cm2 s'1 reported by Culot et al. (10), previous estimates of radon entry by diffusion were relatively small
(2), but still were significant compared to advective radon entry. Using the higher diffusion coefficients
measured with Florida floor-slab concretes (averaging 1.6xl0'3 cm2 s"1), indoor radon concentrations of
more than 2 pCi L'1 may result from soil radon sources of about 3,000 pCi L"1. The assumption of
separability of diffusive and advective entry made a difference of less than 5% in the entry rates. Even
if a 1-cm perimeter floor crack is assumed for the house, diffusion rates through the intact part of the slab
are affected by less than 25%, mainly from altered gradients near the perimeter.
-------
100
o
— o.
O'—
~- c
? o
0} »5
c to
O k-
x ~*
C
E a
o «
Og
a> o
i§
.2 xj
5 w
DOC
o
o
TJ
C
Diffusion Coefficients
o! Concrete (cm2 s"1)
1 E-2
x-
¦'"2E-3 -»''
3 E-4
7.E-5
.1
1 E-5
Ra-226 in Concrete
- 1 pCig'1
.01
100 1000
Radon Concentration in Soil Gas
(pCi L" •)
10000
RAE-103425
Figure 3. Diffusive contributions to indoor radon concentrations for varying soil
radon sources and five different radon diffusion coefficients.
Figure ^These^sed^radon S'ab " SUmmari2ed
Tre,T ,hAVlluf s' TT 4,6 b'"l! mmai! may imroducepcfK?
radon into the house, and 20pCi g of radium in the concrete may exceed the4pCi L"' indoor level w
to the absence of significant subslab radon sources. These analyst assume a radon eL a ScTem
of 0.25, however, little is known about the radon emanation coefficients of concrete « hTS
compare w.th those of ^ aggregate and other concrete components. If radon emanation mi ,!
concrete are less than 0.25, the resu tine radon sourre strpmrthc qah j ^oemcients tor
scale accordingly. ngthS a"d ind00r radon ^^entrations will
-------
100
.01
Ra 226
in Concrete
(pCi g-1)
Concrete
D*1 6E-3 cm^ e'1
P=0.23
100 1000 10000
Radon Concentration in Soil Gas (pCi L"1)
RAE-103430
Figure 4. Diffusive contributions to indoor radon concentrations for varying soil
radon sources and eight different concrete radium concentrations.
contents of the soils were defined as the drainage limits at -0.3 bar matric potential (11), and radium
concentrations in the soils were varied to the maximum that could be tolerated for an indoor radon
concentration of 2 pCi L'1. The radium concentration in the concrete was kept constant at 0.3 pCi g"1.
The resulting radon entry rates for the diffusion and advection mechanisms are plotted in Figure
5, and illustrate the dominance of diffusive radon entry for this case. Although the diffusive and
advective entry rates have the same order of magnitude for the sandy soils, they are more divergent for
the low-permeability soils, differing by more than 2 orders of magnitude for the clay and silty clay loam
soils.
In summary, the present radon diffusion measurements on Florida floor slab concretes exceed
previous diffusion estimates for concrete, emphasizing the importance and sometimes dominance of the
diffusive mechanism for indoor radon entry. The correlation of radon diffusion coefficients with the
concrete W/C ratio suggests this ratio as a possible surrogate for estimating radon diffusion coefficients
of concretes when better data are unavailable. Parametric model analyses indicate several pCi L'1 of
-------
100
SCS SiCILo
33 pCi g''
SCS Clay
15 pCi g"1
SCS Loam
5 G pCi g"1
O 10
a
a>
+->
(0
DC
Diffusive Entry
Adveclive Entry
SCS Sand
2 5 pCi g"1
SCS SjCHo
3 G pCi g 1
C
LU
c
o
"O
03
QC
Pormeabilitios for SCS Soils at Water Contents |
Corresponding to -0 3 bar Maine Potential
Max Radium for 2 pCi L'1 Indoors (0 5 a c h )
141 m' S.O.G. House, 0.9 m Footer, 1 cm
Perimeter Crack
10
-12
10'
10"
10
10"8
10
-8
10'7
Soil Air Permeability (cm^)
10"
RAE-103427
Figure 5. Comparison of diffusive and advective radon entry rates for a slab-on-
grade house with a 1-cm perimeter floor crack.
indoor radon may result from diffusive entry from near-background soils, or from slightly-elevated
radium concentrations (10-15 pCi g1) in the concrete. J 641
ACKNOWLEDGEMENT
This work was supported in part by U.S. Department of Energy grant DE-FG02-88ER60664
U.S. Environmental Protection Agency Interagency Agreement IAG-RWFL933783.
REFERENCES
1, Tanner, A. B. The role of diffusion in radon entry into houses. Presented at The 1990
International Symposium on Radon and Radon Reduction Technology, Atlanta, GA, February 19-
23, 1990.
-------
2. Rogers, V. C. and Nielson, K. K. Benchmark and application of the RAETRAD model.
Presented at The 1990 International Symposium on Radon and Radon Reduction Technology.
Atlanta, GA, February 19-23, 1990.
3. Loureiro, C. 0., Abriola, L. M., Martin, J. E., and Sextro, R. G. Three-dimensional simulation
of radon transport into houses with basements under constant negative pressure. Environmental
Science and Technology 24: 1338-1348, 1990.
4. Rogers, V. C., Nielson, K. K., and Merrell, G. B. Radon generation, adsorption, absorption,
and transport in porous media. DOE/ER/60664-1, U. S. Department of Energy, Washington,
D. C. 1989. 48pp.
5. Shuman, R., Rogers, V. C., and Nielson, K. K. Measurements of concrete properties for
low-level waste disposal facilities. RAE-8716-3, Rogers and Associates Engineering Corporation,
Salt Lake City, Utah, 1988. 31pp.
6. Nielson, K. K. and Rogers, V. C. Comparison of radon diffusion coefficients measured by
transient-diffusion and steady-state laboratory methods. NUREG/CR-2875, U. S. Nuclear
Regulatory Commission, Washington, D. C., 1982. 30pp.
7. Nielson, K. K., Rogers, V. C., and Gee, G. W. Diffusion of radon through soils: a pore
distribution model. Soil Science Society of America Journal 48: 482-487, 1984.
8. Youngquist, G. R. Diffusion and flow of gases in porous solids. In: Flow Through Porous
Media. American Chemical Society, Washington, D. C., 1970. p. 57-69.
9. Rogers, V. C. and Nielson, K. K. Multiphase radon generation and transport in porous
materials. Health Physics, 1991 (in press).
10. Culot, M. J. V., Olson, H. G., and Schiager, K. J. Effective diffusion coefficient of radon in
concrete, theory and method for field measurements. Health Physics 30: 263-270, 1976.
11. Nielson, K. K. and Rogers, V. C. Radon Transport Properties of Soil Classes for Estimating
Indoor Radon Entry. Presented at the 29th Hanford Symposium on Health and the Environment,
Richland, WA, October 16-19, 1990.
-------
TITLE: Radon Dynamics in Swedish Dwellings: A Status Report
AUTHOR: Lynn M. Hubbard, National Institute of Radiation Protection, Sweden
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
A status report of a long term study on radon entry into Svedish
dwellings will be given. Both physical modelling and continuous
measurements of radon and other relevant parameters in real home
environments are used in the investigation. The dynamic modelling used to
study the temporal behavior of radon indoors will be discusssed, and
current results and examples of the modelling and measurements will be
presented. We will show hov changes in the model parameters can be used
to study how the radon behavior responds to these changes. Building
characteristics typical of Swedish dwellings and geological factors
typical of Svedish ground will be discussed with regard to their relevance
to radon entry, and compared to similar factors typical in the housing
stock in the United States.
-------
V-5
SOIL GAS AND RADON ENTRY POTENTIALS FOR SLAB-ON-GRADE HOUSES
by: Bradley H. Turk
105 E. Marcy St., Rm. 109
Santa Fe, NM 87501
David Grumm, Yanxia Li, and Stephen D. Schery
Physics Department
New Mexico Institute of Mining and Technology
Socorro, NM 87801
D. Bruce Henschel
AEERL
U.S. EPA
Research Triangle Park, NC 2//ll
ABSTRACT
The technique to quantify the potential for pressure-driven entry of
soil gas and radon through house surfaces in contact with the soil is evaluat-
ed for six New Mexico houses with slab-on-grade floors. Flows, pressures, and
radon concentrations were measured through test holes in these floors while
the houses were mechanically depressurized from -10 to -30 Pa. Soil and
substructure surface resistances, and soil gas and radon entry potentials were
calculated for each test location. These data support earlier work in four
basement houses in New Jersey that showed the soils surrounding a building's
substructure are many times more resistant to soil gas movement than the
substructure surfaces themselves. Locations along the perimeter of the slab
floors had soil gas entry potentials approximately 40 times greater and radon
entry potentials approximately 15 times greater than locations more central to
the slab. Mean radon entry potentials for a house were found to be a satis-
factory indicator of the average indoor radon concentrations during the
heating season. The radon entry potential data were also useful in the design
and placement of subsurface depressurization radon mitigation systems in these
houses where system locations that are acceptable to the homeowners are
limited.
OVERVIEW
For both scientific and practical reasons, it is important to be able to
characterize the convective flow of soil gas and radon through the surrounding
-------
soils, materials, and const mcl hI surf aces of individual buildings. Intorma-
t ion on this movement: can improve our understand i nj; of tin has i c phvs i cal
mechanisms at work in and around actual bui 1 di ngs. It can also provide
guidance on the selection and design of techniques to reduce radon levels
inside buildings where elevated concentrations rna y pose an excessive health
r isk.
A technique developed two years ago by one of the authors (1) quantifies
the relative leakiness of substructure surfaces in contact with tin- soil and
the resistance to soil gas movement of the soils and materials around a
building. The technique also has proven useful in guiding t lie placement of
subsurface depressurization (SSD) radon control systems. The objective of the
technique is to develop entry potentials for soil gas and radon at various
locations in the substructure surfaces assuming that the detailed characteris-
tics of the substructure surfaces and surrounding soils cannot be known
Interpretive and measurement methods are based on the procedures of rest-arch
ers investigating the radon source potential of soil (') , i /t) and the pressure
fields created in the soil around houses 0,6). Entry potentials were
originally evaluated in four New Jersey (NJ) houses with basements. in this
paper, the technique is examined in six slab-on-grade houses in New Mexico
(NM).
DEFINING SOIL GAS AND RADON ENTRY POTENTIALS
For many existing buildings it is impossible to know the specific
details of the usually complex and non-uniform structure construction and
underlying soils and materials. Consequently, a complete understanding of
soil gas and radon movement around and into a building is unattainable.
Similar to the work of others (7,8), a steady-state lumped parameter model
seeks to simplify the building/soil system by substituting a few simple
electrical circuit elements for the many detailed structure and soil features
We assume that, at each soil gas/radon entry location the building 'views' an
aggregation of the network of pathways through the local cracks, gaps, and low
and high permeability regions in the below-grade soils, materials, and con-
struction features around the substructure. Likewise, the entry location
itself is affected by nearby imperfections in the substructure surfaces. if
air flow through all of these materials is laminar, a direct current electri-
cal analog may be applied where air flow is represented by electrical current
and pressure differences by voltage drops. At each test location we create a
series circuit by substituting an effective resistor for the complex network
of resistances in the surrounding soil and another resistor for the substruc-
ture surface. The negative pressure found in the building is substituted by a
battery (see Figure 1). Other researchers have developed lumped parameter
models to more accurately represent the transient conditions of an actual
building (9), The circuits in these models include capacitors -- which would
involve longer-term, more complex experimental procedures.
To determine the values of the circuit parameters, a test hole is
drilled through a substructure surface. While the negative pressure in the
building is mechanically enhanced by a blower, the pressure difference (AP,
"voltage drop") across the sealed test hole is measured. The measured AP
depends on the resistance of the substructure surface relative to the resis-
-------
Kv-
K-KFF
II"
(a)
Kh
(b)
la
o (c)
EKF
LA
Figure 1. A simplified electrical circuit is shown that substitutes for the
resistances to soil gas flow through the soil and substructure materials.
tance of the surrounding soils. A ratio of these two resistances, Z, is
defined as
7. - ^HC = ^HC = ^TC^FC-EFF _ ^FC-EFF [ ]_ ]
where (a subscript "C" identifies the condition with the test hole closed):
PB (Vb) -
P« (V„) -
Ps (Vs) =
Qt (It) =
Rf-eff =
Rs-e
measured pressure difference (or "applied voltage") between inside
of house/substructure and outdoors, point a to c (Pa),
measured pressure (or "voltage") drop across open test hole and
flow adaptor, point a to b (Pa),
calculated pressure drop across soil paths between point b and
outdoors (c) with test hole open, PB - PH (Pa),
defined total flow ("current") through cracks, openings, and test
hole (m3/s),
calculated effective resistance that lumps resistances of cracks
and openings in substructure surfaces and resistances ,of near-
substructure materials surrounding the open test hole (Pa-s/m3),
and
calculated effective resistance of soil paths to measurement point
b with test hole open (Pa-s/m3).
Typically, a low resistance (leaky) surface will cause a smaller AP to be
measured across the sealed test hole. To determine the effective resistance
of the surrounding soils, the air flow through open test holes is measured.
Because the test hole is not a perfect short circuit -- soil gas continues to
-------
pass through nearby cracks and openings in flu subst rue1 un surf/ire -- the AP
across the test hole is measured with the test hole- open. Assuming that
RfC-EFF ^ Rf-EFF a nd Rsc-EFF Rs-EFF > t hp fl p0 T~ f () Tlfl i- Ug CM* i 'C II I t ."J TN'i ] VS i S . 1 Ilti S I ] bs t 1 t 111
ing analogous air flow and pressure parameters, the effect ivt soil resistance
is calculated from
where 0„ (I„) = measured (corrected) flow through open test lmU :md flow
adaptor (mVs). The effective resistance of the subst rur t ui i- sin face. Rf-e(t.
is found from Equation [1].
The entry potential of soil gas at a test location (. (in'/P.-.-s) . is
proportional to the soil gas flow through surrounding so, ,u,d materials and
nearby cracks and openings in the substructure surface. 1. .s defuud as a
net conductance:
The radon entry potential, E (Bq/Pa-s), is defined is t lu mass t ransfer
of radon in the soil gas near the substructure surface, C (Bq/m') , with the
pressure-normalized flow of soil gas into the building:
Thus, if soil and subsurface effective resistances arc low, the poten-
tial for soil gas to enter a building is increased. In addition, if Vrjdon
levels in the soil gas are elevated, then the potential for radon to enter a
building through an area near a test hole is also increased. The term
'potential', used to define soil gas and radon movement into buildings, does
not refer to electrical potential. Instead it is a more casual term for the
possibility, or capability, of soil gas or radon to ertter near a particular
test location.
The six New Mexico houses that were studied for this paper are locnteH
in or around Santa Fe and Albuquerque. They are part of an eight-house
research project in New Mexico investigating radon and thoron entry nnd
control. All six houses are single story and have siab-on-grade cons, nu-t i ™
Typically, the slab floors were poured over compacted existing soil altho, T'
fill material containing some gravel was used in part of one house (AIDA)
Existing soils range from slightly expansive clays to coarse sandy loam with
rock fragments. Where the slabs are exposed, hairline cracks are'of ten
visible, although most slabs are covered with carpet, linoleum or tile At
least one house (T142) appears t0 have a monolithic (downturn) slab while ,h
remaining houses have floating slabs inside of concrete block or poured
concrete stem walls. In every hoUse hut AL04, extensive cracks (up to 10 mm
wide) exist along the perimeter at the slab/stem wall boundary. These crack*
are sometimes aggravated by styrofoam insulation panels placed vertically 0
FFF + ^F-FFF
E = C,C.
DESCRIPTION OF HOUSES AND EXPERIMENTAL PROCEDURES
-------
the inside of the stem wall. Horizontal sheets of styrofoam insulation were
found directly below the slab along the perimeter walls in other houses.
Plumbing services are always routed below the floor, as well as forced air
furnace supply ducts in some houses (AL02, AL03, and AL04).
EXPERIMENTAL PROCEDURES
From five to thirteen 16 mm (5/8 in.) diameter holes were drilled
through the slab at accessible locations in each house. The holes were placed
to be accessible to the researchers, to avoid sub-floor services, and to
satisfy the aesthetic requirements of the homeowners. By comparison, approxi-
mately 30 test holes were drilled through the surfaces of unoccupied substruc-
tures in each New Jersey house. A blower door fan was used to mechanically
depressurize the houses to approximately -10 Pa and -30 Pa. At each test hole
radon grab samples of soil gas were collected first and then air flows and APs
were measured during depressurization. The enhanced depressurization mini-
mized environmental influences and increased the magnitude of the flow and AP
so they could be more easily measured.
Radon grab samples were collected through a filtered sample train into
evacuated 300 cm3 alpha scintillation flasks. Alpha activity in the cells was
counted after a 3-hour delay on a portable counter/scaler (Pylon Model AB-5).
Uncertainties with this technique are estimated to be ±20%.
Pressure differences were measured using an electronic micromanometer
(Neotronics Model MP20SR) with a minimum resolvable AP of 0.1 Pa with a
specified accuracy of 1% of full scale. All APs were measured with the house
as reference. Pressure differences were first measured across the slab at
each test hole with all holes sealed, PHc- Then the AP was measured across
the open hole with a flow adaptor in place, PH. An alternative approach would
be to calculate the AP across the open test hole using a standard engineering
formula. The flow adaptor is used to establish uniform flow and pressure
measurement conditions at each test hole and is slightly modified from the
adaptor used in the New Jersey houses (1). The adaptor is a 0.3 m (12 in.)
long metal tube with an inside diameter of approximately 10 mm (the outside
diameter is 1/2 in.). A small diameter (3 mm, 1/8 in.) static pressure tube
runs along the inside length of the adaptor to sense pressure at the end
placed into the test hole. A fitting into the side of the adaptor allows a
hot wire anemometer probe to be inserted into the air stream within the
adaptor. Flow rates less than 0.015 m/s (3 fpm) could not be reliably
measured on the Hastings Model B-22 hot wire anemometer. Calculated flow
through the adaptor could be in error by as much as ±50%.
RESULTS
Data from this study of New Mexico houses are shown in Tables 1, 3, and
4 and in Figures 2, 3, and 4. Effective resistances and entry potentials were
calculated using Equations [1] through [4j. Where measured flows and pres-
sures were less than the detection limits of the instrumentation, values of
-------
approximately half the detection limit wort substituted jnt(, , ilt calculations
The summary in Table 1 includes statistics for bot j, norma] uid
distributions. However, since cumulative probability |>|ots °&normal
Indicate that the data are' most closely approximated bv t i>, i "
i i i . 1 ,H' 1 norma 1 distri-
bution, subsequent tables only present the geometric nit-;,,, (i;M) ,
standard deviation (GSD). Aggregation „f tlx test loc;,i M'°nu't ric
- . .i r . 411 Oils OIK-
statistical grouping (as in the first grouping of Tahh j} . . ,
These summary statistics for all test locations at all |u,u't'< ur' f 'nl/s _ f'
DISCUSSION
SUBSTRUCTURE, REGIONAL, AND TEST LOCATION COMPARISONS
Table 1 summarizes the entry potential parameters from the -10 Pa
f-i „ t„ct at- the 48 NM test locations. The GSl) for all data is
deDressurizatton test at
larPe Indicating a very l.r»* ""8e l" the the Parameters. By
examining the CM for the soil ««1 «rt«. rtfnv* resrstane-ea .« «
b rL"oundLne the houses is from i to 15 times more
armarerit tViat tl~i0 soil >- , . i
- *1 as flow than slab floor surfaces. This is consistent with
resistant to soi g ^ houses (Table 2) which show that soils were between
the ear ier a a reSistant to soil gas flow than the substructure
two and six tmes ra _ entry potentials for the NM slab-on-grade floors
surfaces. ^ t^an fof NJ basement floors, possibly due to more
perLableesoils surrounding the NJ houses (although data on soil permeability
are not yet available) or to the closer proximity to the outdoor soil grade in
d ™ Q mm data have been grouped according to location of the
the NM houses. ihe Nn oau , , . , , \ • • i
s i m of the slab perimeter, and b) interior locations
teat hole' st.) witnin 1 , . ,
1 the slab perimeter. The perimeter locations have much
fnr-tber than I m Iron) int r r
^ s entry potentiais (and lower soil and surface resistances).
ISov^mitrto outdoor soil &^ade. the extensive cracking observed along the
. ancj t^e disturhed sQil and materials around the stem walls and
footTnesrprobably combine to contribute to the very high soil gas entry
t- Is at these perimeter locations. Likewise, the soil gas entry
notential for block wall cavity locations in the NJ houses is slightly
J;. also possibly due to tlrie same factors affecting the perimeter
locations in the NM siabs-o^-grade. The interior locations in the NM houses
generally had poor pressure field connection to other locations as measured
with SSDmitigation systems operating or with the sub-slab vacuum cleaner
test.
For both NM and NJ li°uses> where soil gas entry potentials are low, the
radon concentrations in the sou &as tend tQ be M &h The low soil gas entry
potentials may be indicative poor flushing of radon from Hie soil by slight
-------
soil gas movement into the buildings and by minimal diffusion over the large
distance to the soil surface. The GM soil gas radon concentrations were
higher for the NJ houses than the NM houses, suggesting a larger radon source
in the NJ soils. These higher soil gas radon concentrations (42000 Bq/m3)
resulted in the NJ houses' having a high GM radon entry potential through the
basement floors. However, the relatively high soil gas entry potential for
perimeter slab locations in the NM houses created a high radon entry potential
(GM of 15 x 10~3 Bq/Pa-s) despite the low GM radon concentration of 3000 Bq/m3.
The GM radon entry potential for interior slab locations in the NM houses was
15 times lower than for the perimeter locations, despite the higher GM radon
concentrations at the interior locations.
Table 1. Summary of Entry Potential Parameters at -10 Pa Calculated for New Mexico Slab-on-Grade Houses
Entry
Potential Parameters
Statistic
Soil
Resistance,
Rs-eff
(10s Pa-s/m3)
Surface
Resi stance,
Rf-eff
C106 Pa-s/i.i3)
Soil Gas
Entry
Potential,
G
C10"6 m3/Pa-s)
Radon
Cone.
(kBq/m3)
Radon Entry
Potential,
E
C10"3 Bq/Pa-s)
All Locations at 6 Houses
Geometric Mean
0.32
0.064
2.2
3.6
8.5
Geometric Std. Dev.
15. 1
10.5
13.0
4 . 00
13 . 4
Arithmetic Mean
5.1
1.1
13
7.7
65
Arithmetic Std. Dev
8. 75
3.02
21.5
9.00
136
Number of Locations
48
48
48
47
47
Slab Perimeter Locations
at 5 Houses
Geometric Mean
0.20
0.038
4.6
3.0
15
Geometric Std. Dev.
10.9
7.73
9.22
4 .09
11.0
Number of Locations
38
38
38
37
37
Interior Locations > 1 Meter from Perimeter at 5 Houses
Geometric Mean
6.2
0.47
0.14
7.4
1.0
Geometric Std. Dev
6. 18
12.1
6.16
2.92
9.58
Number of Locations
10
10
10
10
10
Table 2. Summary of Entry
Potential Parameters Calculated
for Four New Jersey
Houses
Statistic
Rs_eff
(106 Pa-s/m3)
®F-EFF
(106 Pa-s/m3)
G
U0"6 m3/Pa-s)
Rn Cone.
(kBq/m3)
E
(10-3 Bq/Pa-s)
Basement Slab Floor
Locations
Geometric Mean
1.2
0.65
0.5
42
23
Geometric Std. Dev.
5.0
4.8
4.8
8.09
6.7
Number of Locations
22
22
22
22
22
Basement Block Wall
Cavity Locations
Geometric Mean
0.68
0.12
1.2
6.5
7.9
Geometric Std. Dev
2.2
2.1
1.6
4.85
5.1
Number of Locations
44
44
44
42
42
-------
RADON ENTRY POTENTIALS AND INDOOR RADON LEVELS
It is expected that houses with higher average radon cut rv potentials
i-i hieher indoor radon levels. Figure 7 relates lhf.se data for the
foul NJ and six NM houses. The plot on the right of Figure ; shows the modest
r-nrrelution of the average indoor radon concentration during the winter with
the GM radon entry potential for each group of houses. The two groups do not
line because of differences in house construction, number
0T1 LOG Sciili" xj-ii
. . r tpcf holes at each house, and possibly measurement and
and. location 01 u»c
averaging techniques. The left plot in Figure 2 displays the same data but
?• A aarh proup by the median average radon concentration and median
normalized for each gr P y corre]at-ion coefficient, R, for these normal-
CM rarlon entry pOtenLLai . ll,v-
, . q jhe data indicate that, with improved techniques, radon
ized data is . ¦ practicable predictor of long-term winter radon
pntrv potentials may ut k ,¦>*.. . , , . ,
' p . ¦„ cnmP houses. In models of indoor air pollution, the indoor
roncent.rations in some ¦¦
... _f nollutants is not solely dependent on the pollutant source
roncentrat ion 01 .
,, rates are also an impoitant factor and may explain some
strength. Ventilation lo.
of the lack of correlation in these data.
Normalized 14
Indoor Radon
Concentration j 2
J m 0.23x + 0.78
R2 - 0.43
0,5 10 1.5 2.0 2.6
Horm«lii«d R«don Entry Potential
Average 400
Indoor Radon
Concentration
(Bq/m3) 300
NM Houiei
N/ Houses
\^y ¦ 64* ~ 550
R2 - 0 3?
20 30
GM Radon Entry Potential
Figure 2. Average winter indoor radon levels for fn„r mi v.
Mexico houses are related to the GM radon entry potential SU
right plot. In the left plot, the data are normalized bv the m^- ^ in the
group. Solid curves are lines of best-fit from linear rLr*! dlan for each
lines are 95% confidence curves. ® ession. Dashed
ENTRY POTENTIALS AT DIFFERENT APPLIED DEPRESSURIZATIONS
At five of the six New Mexico houses, entry potentials werP h „
with the structure depressurized to both -10 and "*0 Pa tv, dete™ned
these two levels of depressuriZatio„ arD ^ Table 3 " ,
grab samples were used to calculate radon entry potentials for both "
however, flows and APs changed, causing chan.es j pressures;
A one-sided, two-sample, n°n-parame"S 7 ^ rad°" Potential,
the soil g« entry potentials at -10 Pa « l""' c°™Jus.vely show that
- . . * ~ y are significantly higher than at
-30 Pa
31 test
cant, s
(p < 0.6). The data from a similar ° 7 i»an at
t: locations in the NJ houses also i a- mparisor> °* applied pressures ^
tatistica] difference
-------
different pressures. However, because the range of values is large for these
parameters (large GSD), the statistical tests are not conclusive. In addi-
tion, since the soil gas entry potential at -10 Pa is higher for 20 of the 26
test locations and for the GM of the group, there may be some bias in the
procedure to cause this difference. The GM soil gas entry potential at
approximately -10 Pa for the NJ test locations was 0.64 x 10"6 m3/Pa-s (GSD of
4.16) and at approximately -30 Pa was 0.47 x 10"6 m3/Pa-s. An analysis of flow
data from 22 test holes in the NJ houses suggests that flow through those
holes did not increase in direct proportion to the applied pressure. For a
general equation of the form, Q - A{&P)n, the exponent, n, was calculated to
be approximately 0.9. This result may be due to non-linear air flow through
the materials below the test hole or within the flow adaptor, and could
explain slightly lower entry potentials at higher pressures.
Table 3 Comparing Entry Potential Parameters at -10 and -30 Pa from 5 New Mexico Houses
Statistic
(10
rs-eff
6 Pas/»!)
^F-EFF
(106 Pa-s/m3)
G
(10~6 m3/Pa-s)
E
(10~3 Bq/Pa-s)
House Depr essun zed
to Approx.
-30 Pa
Geometric Mean
0 . 44
0.13
1.4
6.0
Geometric Std. Dev.
13 . 2
House Depressurized
to Approx.
-10 Pa
Geometric Mean
0 . 35
0.10
1.9
7.8
Geometric Etc). Dev.
14 . 4
15.2
12.9
14.0
Number of Locations
26
26
26
26
SEASONAL CHANGES
Thus far in the study, entry potential measurements have been conducted
in different seasons in only one house, AL03. The data for the measurements
made at seven test locations in September and January are shown in Table 4.
By inspection, it appears that the entry potentials are very similar for the
two periods, although because of the small sample size and large standard
deviation, the equality of the samples is not statistically robust
(p < 0.70). It is conceivable that seasonal changes in environmental and
structural conditions, and soil moisture could have a significant effect on
the flows, pressures, and radon concentrations measured during an entry
potential test. These additional measurements are planned for the remaining
five houses.
Table A, Comparibon of Entry Potential Parameters for Two Seasons at New Mexico House AL03
Statisti c
rs-efp
CIO6 Pa-s/m3)
rf-eff
(106 Pa-s/m3)
G
{10~6 m3/Pa-s)
Rn Cone.
(kBq/m3)
E
(10'5 Bq/Pa-s)
September 1990
Geometric Mean
6.0
2.0
0.12
4.2
0.50
Geometric Std. Dev.
6.45
9.92
7.36
2.15
8.14
January 1991
Geometric Mean
6.8
1.5
0.12
4.5
0.34
Geometric Std. Dev
7. 77
9.68
8.01
3.00
6.05
Number of Locations
7
7
7
7
7
-------
APPLYING ENTRY POTENTIALS TO SSD DESIGN
An important aspect of this work has been to apply the radon entry
potential data to the design of SSD radon control systems. The material below
the slabs in the center of these houses does not support the broad extension
of a pressure field from an SSD pipe. Therefore, many pipes would have been
required to develop an adequate pressure field beneath the entire slab. Since
almost the total occupied floor area of these houses is finished, finding
acceptable locations for SSD pipes is difficult.
Figures 3 and 4 display the entry potential data and mitigation system
location plotted on floor plans for each of the houses. Table 5 contains
descriptions of the symbols used in these figures. Because the highest radon
entry potentials generally occur at the slab perimeter, SSD systems were
designed' to depressurize the sub-floor areas with the highest entry potentials
by penetrating the exterior stem walls from outdoors. With this approach,
interior locations for pipes were avoided. By penetrating the stem wall,
pressure fields were often more easily extended along the perimeter through
the existing gaps, channels, and more permeable materials. As seen in
Table 6, the installations based on control of local areas with high radon
entry potentials have been very successful in houses AL04, SF31 , and TI42. Of
these three houses, the SSD pressure field extends through a relatively
permeable layer of soil beneath the entire slab only in SF31 .
Table 5. Description of Symbols Used jn Figures 3 and 4,
Symbol Description
Floor Test Hole Identification
IF(1,2,3. . . )
(YES)/(NO)
Rs
RN
Pressure Field Developed by SSD Mitigation System
Detected at this Test Hole
Effective Resistance of Soil (Rs.Err), 10s Pas/m3
u Eff0Ctive Resistance of Substructure Floor
-------
(a) ALO
o
HfL
•lRs =
Rr~
G =
RN-
F =
(NO)
0.6
0 3
' 1.1
300
12
*)E± (NO)
Rs= 0.06
RT- 0.03
G - 11
RN- 14
E = 5.8
I£2.(VT9)
R9 — 0.04
RP= 0.01
G -t 19
RN.- 82
E = 57
1 -PIPE
SSD
SYSTEM
IF4 (no)
Rs =
Rf~
G =
RN =
0.04
0.01
20
13
9.1
1F5 (NOi
Rg = 0.4
Rr= 0.7
G = 0.93
RN = 110
E ¦= 3 6
IO (**9)
Rs= 0.4
Rf= 0.09
G = 2.2
RN- 230
E - 19
T_
(BRICK --ON-SAND)
(b) AI.,03
lEil(NO)
Rg = 0.2
IF 10 (YI5)
PATIO
RP= 0.03
RN= 72
lEfi (YtS)
Re= 0.5
IF 13 1NQ\
RN= 36
lEfl, (MO)
imtxo) I
Ra= 19 I
FOLLOW
2-PIPE
TF7 (NO)
Rs= 22
R?= 6
G = 0.04
RN = 18
E = 0.02
SYSTEM
INITIAL
SSD PIPE
-------
(a) SF31
2-PIPE
SYSTEM
IE5 (YE3) •
Rs - 0 02
Rr^ 0.01
0 = 34
KN =• 71
E = 08
* 1E4 (w«
Rs- 0 02
R»=- 0 05
G - 16
RN= 10
E » 5.7
tt'3 (TOO
Rs-
Hy-
G
RN
E -
0 01
o oa
28
156
150
tt'2 IYM)
Kg - 0 04
£
RN
r. ¦
0 10
7 2
7ft
21
il l ttW)
Rs ¦= 0.02
Rr»
G -
RN
E -
0.04
17
230
140 •
(b) TI41
(c) TI42
INITIAL
1 PIPE
SSD SYSTEM
li
PATIO
1L4 (YES)
mi (no)
r8= 0.1
Rr= 0.04
G = 0.8
KN- 75
E - 19
0.03
0.001
31
460
520
J£1(N0)
1E2 (wo) •
20
IEa(*o)
0.02
0.005
= 43
RN- 71
E = 110
gakagi:
IFp (NO)
0.004
SD SYSTEM
• IF fl (Yts\
R8- 0.2
c' =
RN
E =
IrA (YES)
Rs- 0 000
0 002
120
12
53
VACHOLE (ris)
Rs = 0.5
0 05
19
550
38
I£i(na)
0.004
i£5 (Y18)
1E&(no)
Rs- O.Ott
\
* SSD
PIPE
1£Z(yk8)
Rs¦» 0.01
R» - 0 005
G = 66
RN- 85
E = 210
\F8 fYBSi
BRICK PATIO
Rr = 0.0 09
RN- 630
ulI (no)
0.006
Figure 4a,b,c. Similar to Figure 3, except for houses SF31, TI41, and TI42.
-------
There are various reasons for the lower effectiveness of the mitigation
systems in the other three houses. The single SSD pipe installed in TI41 has
reduced indoor radon levels approximately 40%, but is not extending the
pressure field to other important radon entry locations (IF2, IF5, and IF6).
Since SSI) mitigation systems are being evaluated before any cracks and holes
are sealed, the large perimeter crack in this house has not been sealed.
During future mitigation, this crack will be sealed to improve the extension
of the pressure field and additional SSD pipes will be installed, if neces-
sary. Both AL02 and AL03 are unique in that their indoor radon levels are
very responsive to changes in barometric pressure. A negative rate of change
(drop) in barometric pressure appears to cause an increase in indoor radon
levels, while a positive rate of change (rise) in barometric pressure causes
radon levels to decrease. From the perspective of this study, two questions
are raised by this condition: a) what is the mechanism forcing radon into the
houses, and are existing mitigation techniques appropriate, and b) should
diagnostic (entry potential) measurements and post-mitigation radon monitoring
be conducted only during periods of falling barometric pressure? In house
ALO?, indoor radon levels appear to have been reduced approximately 70% to an
average of about 80 Bq/m3. However, indoor radon levels have peaked over 520
Bq/m3 during periods of falling barometric pressure with the SSD system
operating. A similar, though more difficult, problem exists at AL03 where two
SSD systems have been installed that are only partially effective.
Although soil gas entry rates may be low at the center of slabs, in
areas of the country where soils have extremely high radon concentrations,
radon entry potentials in the center of the slabs could be quite high. In
these situations, radon control may have to extend to all locations of the
slab.
Table 6. Entr* Potential Data and Pre- and Post-Mitigation Indoor Radon Levels - New Mexico Houses
Soil Gas Ent.rv Potential (G)
Radon Entry Potential (E)
Pre-Mitigation
Post-Hitigation
at 10 Pa'
[at 10
a:
(after 11/1/90)
(after 11/1/90)
Mean
Approx.
Mean
Approx.
House
Geo. /lean Geo.
Number
Geo, Mean
Geo.
Indoor Rn
Duration
Indoor Rn
Duration
ID
(10"c m:/Pa-s) Std. Dev.
Loc.
(r Bq/Pa-s)
Std. Dev.
(Bq/n3)
(hrs)
(Bq/«J)
(hrs)
AL02
4.6 4.10
6
12
2.62
290
1250
80
360
AL03
0.30 11.6
11
0.83
8.98
260
1110
140/130'
470/290'
AL04
0.90 14.2
7
6.8
13.6
290
840
40
440
sni
1/ 1.80
3
47
4.14
360
1040
30
400
1141
2.8 14,9
7
24
21.7
440
1540
230
130
T142
14 4.38
9
42
6.10
390
1070
40
120
» Tvo different SSD mitigation svstems were evaluated.
-------
SUMMARY
Data on soil gas and radon entry potentials f rom this steely 0f ^
Mexico houses with slab-on-grade substructures support the r<-\su]/ts ^ f
earlier study of New Jersey houses with basements. Soi.ls and materi ,1 • ^
surrounding; the substructures of the houses in both studies art. m„
many times
more resistant to soil gas flow than the below-grade structure
, ouriaces and
materials. Those areas of the substructure closer to the open c
i i surr3C6
tended to have higher soil gas entry potentials. Locations aw^y from
perimeter of the slab floors in the New Mexico houses generallv , ,
t • • ¦/ diso ncid much
lower radon entry potentials. Radon mitigation designs incorp0ra( • n .
information emphasized that pipes for SSD systems in these houSps should be
located along the perimeter of the slab and only at areas of reian„ 1 < . ,
i • i j- ^J-aiively hi en
radori entry potentials. Some difficulties with this diagnostic approach ha
been encountered in two houses where indoor radon levels are vp v- ¦ 6
1 . , . . ». . • ' LY responsive
to changes m barometric pressure. Radon entry potentials; were modestl
correlated with average indoor radon levels during the heating season'
implying that with additional modifications to the technique and anilysis
radon entry potentials might be a satisfactory indicator of 1onp.term indoo
radon levels. J 1
ACKNOWLEDGEMENTS
The authors would like to acknowledge the assistance of John Hawley of
the New Mexico Bureau of Mines and Mineral Resources for his interpretation of
the soils and geology at each house site, Gregory Powell for his work on the
radon control systems, and all homeowners for their active participation in
this project.
This work is supported by the U.S. Environmental Protection Agency (EPA)
AEERL through the New Mexico Institute of Mining and Technology^ Cooperative
Agreement No. CR-816688-01-0. This paper has been reviewed in accordance with
the U.S. EPA's peer and administrative review policies and approved for
presentation and publication.
REFERENCES
1. Turk, B.H., Harrison, J., Prill, R.J., and Sextro, r .
-------
ing the radon availability of soil. .In: Proceedings of the 82nd
Annual Meeting of the Air and Waste Management Association.
89-79.2, Anaheim, CA, 1989.
4. Sextro, R.G., Nazaroff, W.W., and Turk, B.H. Spatial and temporal
variation in factors governing the radon source potential of soil.
In: Proceedings: The 1988 Symposium on Radon and Radon Reduction
Technology, Volume 1, EPA-600/9-89/006a (NTIS PB89-167480). pp.
5-61 to 5-74. Research Triangle Park, NC, 1989.
5. Nazaroff, W.W., Lewis, S.R., Doyle, S.M., Moed, B.A., and Nero,
A.V. Experiments on pollutant transport from soil into residential
basements by pressure-driven airflow. Environmental Science and
Technology. ?1: 459, 1987.
6. Sextro, R.G., Moed, B.A., Nazaroff, W.W., Revzan, K.L., and Nero,
A.V. Investigations of soil as a source of indoor radon. Chapter
2. Tn: P.K. Hopke (ed.), ACS Symposium Series No. 331, Radon and
Its Decay Products: Occurrence, Properties, and Health Effects,
pp. 10-29, American Chemical Society, New York, NY, 1987.
1. Mowris, R.J., and Fisk, W.J. Modeling the effects of exhaust ventilation
on 222Rn entry rates and indoor 222Rn concentrations. Health Physics. 54:
491-501, 1988.
8. Scott:, A.G., and Findlay, W.O., Demonstration of remedial techniques
against radon in houses on Florida phosphate lands. EPA-520/5-83-009
(NTIS PB84-156157). Montgomery, AL, 1983.
9. Personal communication with Claus Andersen, Lawrence Berkeley
Laboratory and Riso National Laboratory, Denmark, 1990.
-------
DIRECT MEASUREMENT OF THE DBPKWnBwrw OF RAnn„ „
THROUGH STRUCTURE BOUNDARIES ON PIPPKRKNTlAL~pi^jj^B
by: D. T. Kendrick and G. Harold Langner J
U.S. Department of Energy ' r'
Grand Junction Projects Office
Chem-Nuclear Geotech, Inc.
P.O. Box 14000
Grand Junction, CO 81502
ABSTRACT
A relatively simple measurement technique was developed to estimate the
fraction of radon flux, through a specific portion of a structure boundary
that is attributable to pressure riven flow. The technique consists of
directly measuring the differential pressure across a structure boundary,
while concurrently measuring the radon-flux density for a portion of the
boundary. Measurements are als0 raade of several other related parameters
(temperatures, barometric pressure, wind speed, and wind direction) . The
resulting data allow a simple regression of the measured radon-flux density
on differential pressure. The zero pressure intercept of this regression
represents zero driving force f°r Pressure-driven flow and, therefore, the
nonpressure-driven flow component of the total flux density.
This measurement technique was conducted in three houses in or near Grand
Junction, Colorado. Two of these houses are considered to have low radon
levels, while the third house exhibits somewhat elevated radon levels during
the winter. Interpretation of data acquired from measurements at these
houses indicates that, contrary t0 the commonly accepted theory, pressure-
driven flow accounts for only a small fraction of the total radon entering
these structures through the surfaces measured.
-------
INTRODUCTION
Understanding the dynamics of radon entry into even a simple structure
is not an easy task. However, two entry mechanisms are considered to be of
importance for those cases where soil gas appears to be the major source
of radon available for entry into a structure. These mechanisms are
(1) diffusion of radon through the structure boundary and (2) pressure-driven
flow of radon-laden soil gases through the structure boundary.
Diffusion accomplishes transport of radon atoms by a random-walk process
from regions of high radon concentration to regions of lower radon
concentration. The rate of diffusion depends on the difference in
concentration between the two regions and the physical parameters (including
porosity, permeability, and water saturation) of the media separating the two
regions. The potential exists in most houses for diffusive transport of
radon into the structure because the radon concentration in the soil gas is
considerably higher than the radon concentration within the structure.
Pressure-driven flow is simply the movement of molecules in a fluid
state from regions of high pressure to regions of lower pressure. The lower
levels of most houses exhibit modest pressure differences between their
interior volumes and the atmospheric and soil gases during the colder months
of the year. These pressure differences, commonly on the order of several
pascals (Pa), result from temperature differences between the structure's
interior volume and outdoors, suction produced by combustion appliances
operating within the structure, effects of wind, and changes in barometric
pressure. The resulting depressurization of the structure volume creates a
potential for flow of outside gases into the lower levels of the structure.
The rate of flow depends on the driving force between the two regions (the
differential pressure) and on the physical parameters of the media separating
the two regions.
The overwhelming majority of authors considering radon-entry dynamics
favor pressure-driven flow as the dominant radon entry mechanism (1, 2,
3, 4) . However, some authors (5, 6) point to diffusion as an important
transport mechanism.
It seems likely that both processes are acting to transport radon into
structures and that factors specific to the structure under consideration
determine the relative importance of the two mechanisms. Our understanding
of the relative contributions of these two mechanisms would be greatly
improved if there was a relatively simple method to directly measure the
pressure-driven flow component of the total radon entering a given structure.
This paper presents the progress to date in developing such a technique and
the results of field measurements employing the technique in three houses in
the Grand Junction, Colorado, area.
APPROACH
Even a relatively simple structure such as a house represents a complex
network of potential entry pathways for radon from soil gases. Moreover,
factors such as soil moisture and permeability, which may dramatically affect
both pressure-driven flow and diffusive transport of radon, may change
-------
seasonally and add further complexity. Our approach was to greatly simplify
the system under consideration. Rather than treat the structure and its
surroundings as a multicompartment system as some authors have done (5, 7),
we elected to consider only small portions of the structure boundary
separating the structure interior from the surrounding soil.
The term "structure boundary" means the concrete walls and floor slab of
the basements under study. We believe the importance of pressure-driven flow
in transporting radon across the boundary can be directly ascertained by
examining the radon-flux density from a small portion of the structure
boundary and concurrently measuring the differential pressure across this
same boundary portion. Performing a regression of the observed radon flux on
the measured differential pressure and determining the zero pressure
intercept of such a regression line represents the nonpressure-driven flow
component of the observed radon flux.
Measurements were conducted only in houses with poured concrete basements
for three reasons. First, a basement house represents a better characterized
and simpler system than either a crawl-space house or a slab-on-grade house.
Second, a basement with a floor slab and walls provides two distinctly
different structure boundaries that are bounded by soil on the exterior
surface. Third, these structure boundaries are more readily accessible in an
unfinished basement than in the finished living spaces above a crawl space or
a slab-on-grade house.
METHODS AND INSTRUMENTATION
The basic requirements of
this study were measurements of
radon flux and differential
pressure at several locations
in the basements of the studied
houses. Additionally, monitoring
a number of associated parameters
more fully describes the system
under investigation and ensures
that the measurements being made
do not adversely affect the
natural situation. These
associated measurements consist
of the radon concentrations in
the soil gas adjacent to the
structure boundary, in the
outside air, and in the
basement; various temperatures:
and barometric pressure, wind
speed, and wind direction.
Figure 1 depicts some of
these measured parameters.
[Rn]
Figure 1. Some Measured Parameters
1
[Rn] = Radon-Concentration Measurement
P = Differential-Pressure Measurement
Rn | = Radon-Flux Density Measurement
Measured Parameters
-------
RADON MEASUREMENTS
Both active and passive continuous radon monitors were employed in this
study. The two types of monitors differ principally in the means of
delivering the gas sample to the scintillation chamber and in the size of the
scintillation chamber. Both types of monitors have a zinc - sulfide - coated
scintillation chamber connected to a photomultiplier tube, which is powered by
a Ludlum model 2000 scaler. Modifications to the scaler produce a logic-
level pulse as output in addition to the standard light - emitting diode (LED)
display and printed output. This modification permits the photomultiplier
pulses counted by the scaler to be transmitted directly to the data logger.
The data logger counts the number of pulses per sampling interval and applies
a calibration equation to compute radon concentration.
Active Monitors
An active continuous radon monitor was used for all radon measurements
except soil-gas measurements. This monitor uses a 7.5-centimeter-diameter
photomultiplier tube looking into a 1-liter scintillation chamber. A small
pump delivers the sample through a filter into the scintillation chamber at a
flow rate of 0.8 liter per minute (L/min). Whenever possible, an external
pump immersed in the atmosphere to be sampled supplies the air sample to the
radon monitor to minimize the possibility of contaminating the sample by a
leak in the delivery system. The sensitivity of the active continuous
monitors is approximately 220 counts per hour per picocurie per liter
[cph/(pCi/L)].
Passive Monitors
A passive continuous radon monitor was developed to measure soil-gas
radon concentration. Early attempts to use the active version of the
continuous radon monitor demonstrated how sensitive soil-gas radon
concentration and sub-slab differential pressure are to even slight
perturbations. Extraction of extremely small flow rates of soil gas resulted
in significant depletion of the sub-slab radon concentration and
depressurization of the sub-slab soil in the region of the measurements. A
closed-loop recirculation system was also tried, but the system produced
unacceptably large perturbations.
To overcome these problems, a passive in-soil radon monitor was assembled.
The probe portion of the monitor contains a Hamamatsu R-1010, 13-millimeter-
diameter photomultiplier tube that looks into a scintillation chamber made of
a 51-millimeter (mm) length of 1/2-inch copper tubing coated with zinc
sulfide. A number of small holes drilled in the wall of the chamber at the
end opposite the photomultiplier tube allow soil gases to enter. This
assembly is inserted into a 17.5-mm inside diameter (I.D.) stainless steel
tube, approximately 60 cm in length, that is used as the probe housing. The
lower end of the housing is fitted with a tapered plug. Just above the plug,
a number of small holes drilled in the wall of the probe allow soil gases to
reach the scintillation chamber. A rubber stopper seals the open end of the
probe housing against light entry and gas exchange: a high voltage-signal
cable passes through the stopper. A light seal forms at the lower end of the
-------
probe by its insertion into the soil,
measurements under a concrete slab, the
between the probe housing and concrete
A hole is drilled in the concrete for
probe is inserted, and the area
is sealed with caulking material.
The small volume of the scintillation chamber produces a relatively low
sensitivity [2.5 cph/(pCi/L)], and the instrument exhibits a good high-
voltage plateau. When used for determining soil-gas concentrations with a
counting interval of 1 hour, the coefficient of variation based on counting
statistics is within a few percent for typical values of radon concentration
in soil gas.
RADON-FLUX MEASUREMENTS
Radon-flux density measurements were made using a modification of the
flow method described in Colle et al.(8). The device consists of a metal can
that is 0.44 meter (m) in diameter and 0.16 m in height and has one 3-cm-
diameter hole in the side of the can. A small aquarium pump, mounted on the
inside of the can, permits continuous sampling of the air within the can. The
open end of the can is placed against the surface where the radon-flux
density will be measured, and the can edges in contact with the concrete are
sealed with a caulking material. The pump delivers a steady flow of
0.2 L/min from the can to a continuous radon monitor. The hole in the side
of the can permits room air to replace the volume of gas delivered to the
radon monitor. When operated at a flow rate of 0.2 L/min, there is no
measurable (( 0.1 Pa) pressure difference between the can and the room. The
ventilation rate of the can at this flow rate is approximately 0.5 hour
Figure 2 is a schematic diagram of the flux can applied to a concrete surface
[Rn]
Sample Delivered
to Radon Monitor
0.44-m Diameter Metal Can
Aquarium Pump
LLL
Concrete Slab
[Rn]
.— 3-cm Diameter Hole
[Rn] = Radon Concentration Measured
Figure 2. Schematic Diagram of the Radon-Flux Can Apparatus
Appliec^ a Surface
-------
The can acts as a partial accumulator and allows the radon concentration
to build up within the can to a concentration that is a function of the
radon flux into the can, the concentration of radon in the air backfilling
the can, and the rate of air flow through the can. A simple mass balance
approach, neglecting the radioactive decay of radon, yields the following
expression for the measured radon-flux density:
QsCs + E %>Cm " QrCr
J (
A A
where:
J = radon-flux density from the surface of interest,
Qg = flow rate of soil gases into the can,
Cg = radon concentration of the soil gases,
E = contribution of radon from direct emanation,
A = area of the can,
Qm = flow rate of sample into the continuous radon monitor,
Cm = radon concentration measured by the continuous monitor,
Qr = flow rate of room air into the can, and
Cr = radon concentration of the room air.
Because the rate of air flow out of the can is equal to the rate of air
into the can, Qm = Qg + Qr. If the assumption is made that Qg is small
compared to Qr, that is to say that Qm « Qr, then the net radon-flux density
from the surface being measured may be approximated by:
QsCs + E VCm " Cr>
j _ =
A A
Therefore, the following equation was used to calculate the measured radon-
flux density:
VCm * Cr>
j „ (3)
DIFFERENTIAL-PRESSURE MEASUREMENTS
Differential-pressure measurements were made using several Validyne
model DP 103-10 Differential Pressure Transducers coupled to Validyne model
CD-15 Sine Wave Carrier Demodulators. The CD-15 produces a ± 10-volt signal
corresponding to the + 87.2-Pa (0.350-inch full-scale pressure range of
the transducer. The stated accuracy of the transducer is ± 0.25 percent full
scale or approximately ± 0.2 Pa.
The differential pressure measurements were made by drilling a 3/8- to
l/2-inch-diameter hole in the floor slab or basement wall. A length of
-------
1/4-inch-diameter stainless steel tubing was inserted a distance of
approximately 20 cm into the soil beyond the thickness of the concrete.
Caulking material was used to fill the space between the tubing and the
concrete to form a gas seal. The open end of this tubing was then connected
to the signal-in side of the pressure transducer. In all cases, the reference
side of the pressure transducer was left open to the basement room.
DATA LOGGING
A Campbell Scientific CR-10 data logger continuously recorded the
values provided by the various sensors. A pulsccounting module provided
inputs for as many as eight radon monitors. A portable weather station
equipped with wind direction and speed sensors, as well as a barometric
pressure transducer, was coupled to the data logger to provide relevant
weather data. All temperature measurements were made using type-T
thermocouple wire. The data logger was programmed to sample each of the
input channels at 10-second intervals and tabulate and record an average
value for a 1-hour sampling period.
FIELD MEASUREMENTS
Field measurements were conducted in basements of three houses in or near
the vicinity of Grand Junction, Colorado. No known uranium mill tailings
deposits are associated with any of these houses. All three houses are
situated on soils derived from the Mancos Shale, which underlies most of the
Grand Junction area. These soils are dominated by the clay-sized fraction;
a significant portion of the clay-sized fraction consists of smectitic
clays. The smectitic clays produce large shrink-swell potentials that affect
not only the soils' engineering properties but also the permeability and
porosity characteristics.
House-1 has a full basement, approximately 22 by 43 feet, finished with
plywood paneling or paint on the wall surfaces and asphalt tile on the floor.
The basement is divided into five rooms, one of which is a bathroom. The
measurements were conducted in the largest of these rooms, measuring
12 by 28 feet, during October. This house was unoccupied during all of the
measurements. Radon-flux density measurements were made at two locations.
The first location was over an open joint in the floor slab approximately
3 mm wide; the second location was on one of the exterior walls at a height
of approximately 1 meter. The average radon concentrations for this
structure for the last 6 years are 1,8 pCi/L in the upstairs area and
2.8 pCi/L in the basement.
House-2 has a partial basement measuring 28 by 29 feet. The remaining
portion of the house is located over a crawl space. The basement has been
partially finished by the addition of two interior walls that separate the
basement into two rooms. The measurements were conducted in the smaller of
the two rooms, measuring 14 by 16 feet, during August and September. This
house was occupied during the period of the field measurements. The exterior
basement walls and floor slab are bare concrete. Radon-flux measurements were
~iade at two locations on the floor s^ab. The first location was over a crack
pproximately 0.5 mm wide; the second location was over an undisturbed
-------
portion of the floor slab. A flux can was also located on one of the
exterior walls at a height of approximately 1 meter. The average radon
concentrations measured in this house during the past 6 years are 0.9 pCi/L
in the upstairs area and 2.8 pCi/L in the basement.
House-3 has a partial basement measuring approximately 16 by 38 feet.
The remaining portion of this house rests on slab-on-grade and over crawl
space. The basement is unfinished with bare concrete floor and walls.
Radon-flux measurements were made at two locations on the basement floor slab
during the months of December and January. This house was occupied during
all of the measurements. The first location was over a 30-cm by 30-cm
opening in the floor slab that presumably was intended for a lavatory. The
hole exposes soil directly beneath the slab (no aggregate is present). The
second flux measurement was made over undisturbed concrete floor at a
distance of approximately 1 m from the closest edge of the first location.
No flux measurements were made on the walls and no long-term average radon
measurements are available for this house.
RESULTS AND DISCUSSION
This investigation was not intended as a comprehensive study of the role of
pressure-driven flow as an entry mechanism for radon into houses but as an
evaluation of the usefulness of the measurement technique. These three
houses were selected for field measurement because of availability and their
range of radon levels. Initial measurements were conducted over concrete
surfaces with structural cracks and undisturbed sections of slab. Plans are
underway to test this measurement technique on sections of the floor-wall
joints at House-2 and House-3.
HOUSE-1
There is a strong degree of correlation between the measured flux and the
differential pressure in House-1. Figure 3 shows a time plot of the
naturally existing differential pressure across the floor slab and of radon-
flux density measured directly above an open joint in the floor slab. The
data were collected during the month of October. The radon flux lags
behind the differential pressure by about 1 hour, primarily due to the
ingrowth of radon-daughter activity within the radon monitor's scintillation
chamber and the ventilation rate of the flux can. The occurrence of some
negative differential-pressure values may represent situations where the
sub-slab region is actually depressurized with respect to the basement, but
it is more likely that these values represent a slight zero shift of the
pressure transducer.
Figure 4 shows a regression of the data from the time plots of Figure 3.
The radon-flux data have been shifted to compensate for the time lag seen in
Figure 3. The correlation coefficient of 0.83 indicates a strong dependence
of variations in radon-flux density on variations in differential pressure.
More simply stated, pressure-driven flow is affecting the observed radon-flux
density, but the slope of the regression line is comparatively low. The zero
pressure intercept of 0.18 pCi'm'^'s"^ represents 80 percent of the average
measured flux. Our interpretation is that approximately 80 percent of the
radon flux transported through the measured portion of the slab occurs due to
-------
Radon-Flux Density
Differential Pressure
Figure 3.
H 0.9
0.7
0.5
H 0.3
0.1
-0-1
(B
Q.
ID
(A
o>
h.
Q_
1
C
£
n
o
36 48
Time in Hours
Time Plots of Measured Radon-Flux Density Through
Open Joint in Floor and Differential Pressure
Measured Across Floor (House-1)
0.4
0.3
0.2
Slope = 0.12
Intercept = 0.18
Correlation Coefficient = 0.83
o 0.1
0 U
i i i i i i
i . ¦ . ¦ i
-0.1
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Differential Pressure (Pa)
Figure 4. Regression of Measured Radon-Flux Density Through
Open Joint in Floor on Differential Pressure
Measured Across Floor (House-1)
-------
a nonpressure-driven flow mechanism
driven flow. An obvious choice for
mechanism is diffusion.
and only 20 percent occurs by pressure-
this nonpressure-driven flow transport
HOUSE-2
Because the measurements were made in August and September, this house
exhibited no significant depressurization of the basement relative to the
underlying soil. In an effort to simulate wintertime conditions, a small box
fan was installed in one of the basement windows to slightly depressurize the
basement. The fan was run at a constant speed so that the pressure variations
would be similar in magnitude to what would be expected during the winter.
Figures 5 and 6 present the results of this experimental setup.
Figure 5 shows the regression of radon-flux density, measured over a small
crack in the basement floor, on the differential pressure across the floor.
Figure 6 shows a similar regression for radon-flux density measured over an
undisturbed portion of the floor. Although some dependence of the measured
radon flux on differential pressure is depicted on the graphs, the scatter
about the regression line is quite large and, correspondingly, the correlation
coefficients are extremely small. The zero pressure intercepts of 0.025 and
0.017 pCi'm'^'s~* for Figures 5 and 6, respectively, represent approximately
80 percent of the average measured flux values. This indicates that
pressure-driven flow adds only about 20 percent to the average nonpressure-
driven flow component of the radon flux. The measured flux values are quite
low for these two locations and are, in fact, comparable to values given in
the literature for emanation from some typical concrete samples (8, 9) .
HOUSE-3
This house exhibits significantly higher radon levels than either of the
other two houses in this study. The average basement radon concentration
during the measurements was approximately 8 pCi/L, compared to 3.6 and
1.6 pCi/L for House-1 and House-2, respectively. The radon concentration
measured in the sub-slab soil gas averaged approximately 3,760 pCi/L during
the measurement period.
Figure 7 shows a regression of the radon-flux density on the differential
pressure across the floor. The radon flux was measured over a 30-cm by 30-cm
opening in the floor slab. As might be predicted in this situation, the
measured flux density exhibits a relatively strong dependence on differential
pressure. However, the high intercept of 13.2 pCi'm *s ^ indicates that only
approximately 20 percent of the average flux density is due to the pressure-
driven flow component. Considering the relatively large area of exposed soil,
it is not entirely unreasonable that this is the result of diffusive
transport directly from the soil.
-------
0.06 h
0.05
0.04 h
0.03
2 0.02 U
0.01 L
0 U
Slope = 0.009
Intercept = 0.024
Correlation Coefficient =
0.27
'o°o 0 < g?« 1 «*, %
OO
o
° O O oO
-0.1
0.1 0.2
0.3 0.4 0.5 0.6
Differential Pressure (Pa)
0.7 0.8
0.9
Figure 5. Regression of Measured Radon-Flux Density Through
Cracked Portion of Floor on Differential Pressure
Measured Across Floor (House-2)
Slope = 0.01
Intercept - 0.016
Correlation Coefficient - 0.34
CP O
1° °
O O Oo . _
*° oftGL ° o 3 o0o
°°o \ 0 0
O ^OO
o /T 0
o o o 0 o° 0
1 I I I I. .1 I 1-1. 1 I I I I I I I i I I-1.1 t .1 I .... I .... I .... I ..., I ... f I .. t l I
-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Differential Pressure (Pa) *
Figure 6. Regression of Measured Radon-Flux Density Through
Undisturbed Portion of Floor on Differential
Pressure Measured Across Floor (House-2)
-------
E
5
V)
c
4)
Q
C
o
¦o
<0
DC
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
L
Figure 7.
Slope = 15.0
Intercept = 13.2
Correlation Coefficient = 0.87
0.1
0.2
0.3
0.4
0.5
0.6
Differential Pressure (Pa)
0.7
0.8
Regression of Measured Radon-Flux Density Through
30-cm by 30-cm Hole in Floor on Differential
Pressure Measured Across Floor (House-3)
CONCLUSIONS
A relatively simple technique was developed to measure the pressure-
driven flow component of radon flux passing through a selected portion of a
structure boundary. The technique consists of directly measuring the radon-
flux density over a selected portion of the structure boundary, while
concurrently measuring the differential pressure across the same boundary.
Regression analysis of the resulting data allows a quantitative
interpretation of the fraction of radon flux that is attributable to
pressure-driven flow.
Certain precautions are necessary in the use of this technique. The
pressure transducers must have adequate sensitivity to measure the small
differential pressures observed and must exhibit' good zero stability to
produce meaningful regression analysis results. Additionally, the radon
concentration of the room air backfilling the flux cans must be known.
This technique was performed in the basements of three houses with low to
slightly elevated radon levels. Interpretation of the measurement results
indicates that, on the average, only 20 percent of the radon enters these
structures through the boundaries tested because of pressure-driven flow.
-------
ACKNOWLEDGMENTS
Work supported by the Exploratory Research and Development Program of the
U.S. Department of Energy Office of Environmental Restoration and Waste
Management, under DOE Contract No. DE-AC07-86ID12584.
The work described in this paper was not funded by the U.S. Environmental
Protection Agency and therefore the contents do not necessarily reflect the
views of the Agency and no official endorsement should be inferred.
REFERENCES
1. Bruno, R. C. Sources of indoor radon in houses: A review. Journal of
the Air Pollution Control Association. 33: 105, 1983.
2. Nazaroff, W. W., Moed, B. A., and Sextro, R. G. Soil as a source of indoor
radon: Generation, migration, and entry, ^n: W. W. Nazaroff and A. V.
Nero, (eds.). Radon and its decay products in indoor air. John Wiley &
Sons. New York, New York, 1988. p. 57.
3. Sextro, R. G., Nazaroff W. W., and Turk, B. H. Soil permeability and radon
concentration measurements and a technique for predicting the radon
source potential of soil. Paper presented at the 1988 Symposium on Radon
and Radon Reduction Technology, Environmental Protection Agency, Denver,
Colorado. October 17-21, 1988.
4. Nero, A. V., Gadgil, A. J., Nazaroff, W. W., and Revzan, K. L. Indoor radon
and decay products: Concentrations, causes and control strategies.
D0E/ER-0480P, U.S. Department of Energy, Office of Health and
Environmental Research, Washington D.C., 1990. 138 pp.
5. Holub, R. F., Droullard, R. F., Borak, T. B., Indret, W. C., Morse, J. G., and
Bf.xter, J. R. Radon- 222 and Rn-222 progeny concentrations measured in an
energy-efficient house equipped with a heat exchanger. Health Physics
49: 267, 1985.
6. Tanner, A. B. The role of diffusion in radon entry into houses. Paper
presented at the 1990 International Symposium on Radon and Radon
Reduction Technology, Environmental Protection Agency, Atlanta, Georgia.
February 19-23, 1990.
7. Stoop, P., Aldenkamp, F. J., Loos, E. J. T., de Meijer, R. J., and
Put, L. W. Measurements and modeling of radon infiltration into a
dwelling. Paper presented at the 1990 International Symposium on Radon
and Radon Reduction Technology, Environmental Protection Agency, Atlanta
Georgia. February 19-23, 1990.
8. Colle' R., Rubin, R. J., Knab, L. I., and Hutchinson, J. M. R., Radon
transport through and exhalation from building materials: A review and
assessment. NBS Technical Note 1139. U.S. Department of Commerce,
National Bureau of Standards, Washington, D.C., 1981. 97 pp.
-------
9. Ingersol, J. G. A survey of radionuclide contents and radon
emanation rates in building materials used in the U.S. Health Physics.
45: 363, 1983.
-------
V-
RADON RESISTANCE UNDER PRESSIJRK
By: William F. McKelvey
Versar, Inc.
2010 Cabot Boulevard West
Langhorne, PA 19047
and: Jay W. Davis
Versar A/E, Inc.
760 Horizon Drive
Grand Junction, CO 81506
ABSTRACT
The radon mitigation field has many products available for the purpose of
controlling the influx of radon gas through the cracks and joints which occur
in structural components. These products are generally classified as
caulkings, paints, membranes or cenientitious materials.
Since it is difficult to evaluate the true effectiveness of these
products in the field, an air tight laboratory chamber was designed and
constructed to evaluate each product. The chamber and test conditions were
set up to determine the resistance of each material to radon permeation
(Transport) under various pressure differentials that would be similar to
field conditions.
Each material was exposed to chamber ambient radon concentrations of
several thousand picocuries per liter with an average pressure differential
across the test material of 0.5" to 2" H20. Each material was tested as the
product would be used in the field and compared with a control for QA/QC.
-------
RADON MATERIALS TESTING CHAMBER
BACKGROUND AND PAST METHODOLOGY
The materials testing chamber (patent pending) was designed and
constructed to fulfill the need to test various materials for their ability to
resist the permeation of radon gas through the material under pressure in a
controlled environment, where both the radon gas concentration and the
pressure can be controlled.
Prior to the construction of this chamber, one of the most common methods
for testing materials for radon resistance was the "bucket test", which
entailed the use of two five gallon buckets placed end to end. The opposed
ends of the buckets were structured to remain open to each other, save for the
material to be tested, which was sealed in place between the two ends. The
radon gas generating material, or the gas itself, would be placed or pumped
into the "hot" bucket. Under atmospheric conditions, the radon would diffuse
across the test material and accumulate in the "cold" bucket. Radon gas
samples would be collected from the "hot" and "cold" buckets and compared.
The percent reduction was obtained by dividing the "cold" bucket radon
concentration by the "hot" bucket radon concentration times 100%.
Hot pCi/1 - cold pCi/1 x 100% - percent reduction in radon concentration,
hot pCi/1
With the above described method there were several constraints. They are
as follows:
1. No way of controlling the radon concentrations generated by the
source material.
2. Could not maintain a constant pressure differential across the
material under test.
3. Size of material to be tested had to be at least 13"to 14" in
diameter to span the diameter of the buckets, which limited the
types of materials that could be tested, due to the large surface
area requirement.
4. Difficult to seal and maintain a seal between the test buckets and
the material undergoing the test, especially for a 30-day test
period.
5. Only one type of material could be tested at a time.
To overcome the above listed problems, Versar designed and constructed a
radon materials testing chamber, and filed for a patent (pending). The
chamber was constructed primarily for testing materials for radon resistance,
however, the application can be used for other gases as well.
DESCRIPTION
The radon material testing chamber has six (6) major components. They
are as follows:
-------
1. Radon Test Chamber
2. Radon Source Chamber
3. Test Cells (4)
4. Pump with flow rate calibrator
5. Manometers (2)
6. Radon Gas Monitor
Radon Test Chamber (Reference Figures I & II)
The radon test chamber is constructed of 1/2 inch thick clear acrylic
with an internal volume of 12 cubic feet (339.97 liters). The top of the test
chamber is removable for access to the chamber interior. There are 18 brass
valves attached to one end of the test chamber. The valves and their
functions are as follows:
• Test Cell Supply Valve - There are four valves, numbered 1 through
4. These numbers coincide with the number of the test cells placed
inside the chamber. The valves penetrate the wall of the chamber
and allow for the connection inside the chamber from the valve to
the corresponding test cell with tubing.
• Test Cell Return Valve - There are four valves numbered 1 through 4
These numbers also coincide with the number of the test cells placed
inside the chamber. As with the supply valves, tubing connects the
valve to the corresponding test cell.
The supply and return valves provide the means to take a radon gas sample from
inside each of the test cells. The radon gas is drawn through the supply
valve, collected and/or passed through the radon monitor back into the test
cell via the return valve. This keeps the pressure inside the test cell in
equilibrium and allows for a complete exchange of the air in the cell.
• Test Cell Pressure Valves and Chamber Pressure Valve - There are
four valves, numbered 1 through 4, and as with the test cell supply
and return valves, each one is connected to a test cell. A
manometer is connected to each valve and to the signal-in port of
the manometer. The reference port of each manometer is connected to
the chamber pressure valve. This provision gives an accurate
reading of the pressure differential between the internal pressure
of each test cell and the radon chamber.
• Chamber to Atmosphere Pressure Valve - This single valve penetrates
the wall of the chamber for the purpose of measuring the pressure
differential between the chamber and the atmosphere with a
manometer. The reference port of the manometer is open to the
atmosphere and the signal-in port is connected to the chamber valve
Radon Source Chamber (Reference Figure I)
The radon source chamber is a metal air tight chamber with an internal
volume of .245 cubic feet (6.94 liters). This chamber is used to store low
level radioactive material that generates the radon gas. The chamber is
equipped with two valves; a supply and a return valve. The radon gas is
-------
Return
Valve
Gas Out
Test Cell
(Sample)
Test Material
In-Line Filters
Support Rack
Supply
Valve
Test Cell
(Control)
Control Membrane
Gas In
Pump
I— Material Testing Chamber
Radon
Source
Chamber
Radon
Source
Gas Sample i
Supply Port Is
Differential
Pressure Port
Material Sample
Gas Sample]
Return Port
Typical Test Cell
Charging the Testing Chamber (Patent Pending) Versar, Inc.
Figure I
-------
In-Line
Filter
Scintillation
Cell
Pump
Manometer A
Manometer B
Material Testing Chamber
Test Cell
(Sample)
Test Cell
(Control)
Supply Valve
Pressure Valve
Return Valve
t-Each Test Cell
Test Material
Support Rack
Control Membrane
Material Testing Chamber fTop View^
Manometer A : Measures AP between chamber and test cell.
Manometer B : Measures AP between chamber and atmosphere.
In-line Filter: Collects radon decay products, allowing only the radon gas to pass.
Scintillation Cell: Collects radon gas that has diffused through test material and accumulated in test ce
Pump: Draws radon gas from test cell through scintillation cell and returns gas back to test c<
Radon Gas Sampling Diagram (Patent Pending) Versar, Inc.
Figure II
-------
circulated from this chamber into the testing chamber and back again until the
testing chamber is charged with a predetermined level of radon gas.
Test Cells (Reference Figure I)
The test cells are constructed of either 4 or 6 inch diameter PVC,
Schedule 40 pipe and have internal volumes of 0.102 cubic feet (2.89 liters)
for the 4-inch diameter cells and 0.196 cubic feet (5.55 liters) for the 6-
inch diameter cells. Each test cell is sealed at one end with end caps of the
same material. Each end cap is equipped with three brass nipples which are
connected to the appropriate chamber valves via tubing, as described under
Radon Test Chamber. The material to be tested for radon permeation
characteristics is mounted and sealed to the open end of the cells. The cells
are placed on a loading rack inside the test chamber.
Pump (Reference Figure I)
This is a low volume air flow pump, calibrated and set to a flow rate of
2 liters/minute. The pump serves two purposes:
1. To charge the testing chamber with radon gas.
2. To maintain a predetermined pressure within the testing chamber.
Manometer (Reference Figure II)
The manometers used are electronic digital manometers (EDMs) that are
capable of measuring one thousandth of an inch of water (1/1000 inch H20)
pressure. A manometer is connected to the testing chamber, (the valve labeled
chamber to atmosphere) to measure the pressure differential between the
testing chamber internal air pressure and that of the atmosphere. EDMs are
also used to measure the pressure differential between each of the test cells
and the testing chamber.
CLASSIFICATIONS OF TESTED MATERIALS
As stated earlier, the materials testing chamber and related components
were designed to test all types of materials and to test up to three materials
simultaneously. In the radon industry these materials are generally described
as sealants. They are manufactured and applied to the building components of
a structure in a variety of ways. The five general categories of sealants are
as follows:
1. Caulks - generally these are either flowable and self-leveling or of
gun grade quality. They are used to seal cracks, control, and
expansion joints in floors and walls that are in contact with the
soil. Examples include one and two part polyurethanes, silicones,
latex, oil based, and butyl rubber caulks.
2. Cementitious materials - generally these are either premixed or in
powder form. When applied they are of a consistency suitable to be
brushed or trowelled onto floor or wall surfaces. Generally, they
are applied from 1/16" to 1/4" in thickness. These materials are
-------
usually used to seal basements against water and moisture
infiltrations. The characteristics of these products have appealed
to the radon industry.
3. Membranes - generally these are either pliable or semi-rigid. The
pliable materials come in rolls 3 to 4 feet wide and 2 to 40 mil
thicknesses. The roll material usually consists of polyethylene or
polyurethane based products. The semi-rigid sheet material is
generally manufactured in sheets of 3 to 4 feet in width and 8 to 10
feet in length. The sheet material usually consists of plastic or
layers of bituminous material laminated with polyethylenes.
4. Paints and epoxies - generally these are manufactured in a liquid
state and, when dry, form a water and moisture resistant
film. These can be brushed, rolled, or sprayed onto their intended
surfaces. Generally paints are applied at 2 to 4 mil. thicknesses
and epoxies are applied at heavier rates, up to approximately 1/16
of an inch in thickness. Epoxies, when cured, usually form a hard
and impenetrable surface.
5. Foams - are expandable polyurethanes that when exposed to the air
expand volumetrically forming closed air cells. These materials are
generally used to seal and insulate large cracks, openings and
cavities found in the various components of a building.
The radon industry has used all of the above mentioned types of sealants
to prevent or reduce the infiltration of radon into a structure. However
sealing techniques, in general, do not have a high success rate as a sole
mitigation method to prevent radon infiltration. Sealing is generally used in
conjunction with other mitigation techniques such as subslab and/or wall
depressurization for a complete remediation package.
Manufacturers became interested in how their products would perform in
radon reduction and contacted Versar for an evaluation of their product(s).
Versar's objective was to evaluate each product in a manner that simulated
field conditions and to test each material as a separate entity. For example
if testing a paint material, it would have to be bonded to a substrate
material which would not bias the test results. The radon gas would have to
pass easily through the substrate material in order to evaluate the radon
resistance of the paint. The output of the evaluation would be simply
expressed in effective percent reduction in radon gas.
OPERATION METHODOLOGY
(Reference Figure I)
The material to be tested is sealed to the open end of a test cell.
After the seal has cured, a negative pressure is applied to the test cell to
check for leaks. If no leaks are found, the test cell is placed in the
testing chamber with the control test cell. Up to three different test cells
and the control can be placed in the testing chamber at one time. The testing
chamber is then sealed.
A radon source chamber with an in-line filter is connected to the suction
side of a pump, and the positive side of the pump is connected to the chamber
-------
A radon source chamber with an in-line filter is connected to the suction
side of a pump, and the positive side of the pump is connected to the chamber
supply valve. The chamber return valve is connected to the return side of the
radon source chamber via tubing with an inline filter. This provides a closed
loop system for charging the chamber with radon gas. The in-line filters
capture the radon daughters that have accumulated in the radon source chamber.
The desired operational concentration of radon in the test chamber is
determined by a function of flow rate and time that the radon is pumped from
the source chamber.
With our radon source, pumping five minutes at 2 liters/minute provides
approximately 12,000 pCi/1 of radon in the test chamber. A radon sample is
collected and analyzed immediately after charging the test chamber to verify
the actual concentration. The chamber is then pressurized through the use of
an air pump. This places a positive pressure on the chamber side of the
material under differential test. Manometers measure the pressure across the
tested material. Pressure can be variable, but is usually held at 0.1 to 0.2
inches of water. To maintain this pressure differential, the test chamber is
maintained at approximately 6 inches of water relative to atmospheric
pressure.
Testing for each product sample is usually conducted at 24-hour intervals
for a period of five days. Testing is conducted (reference Figure II) by
connecting a radon sampling device (scintillation cells and pumps) via tubing
with an inline filter to each of the supply valves for the test chamber and
each test cell. The positive discharge side of the pump is connected to each
of the return valves of the test chamber and to each test cell. As before,
when charging the test chamber, each inline filter captures the radon
daughters that have decayed from the radon gas between test periods. Also,
sampling is conducted through a closed loop system. This prevents altering
the equilibrium of the test chamber or any of the test cells by applying
positive or negative pressures to these components. This method allows for
the determination of the available radon concentrations in the chamber and
each test cell for each test period.
The percent reduction, or the ability of the test material to resist
radon infiltration into the test cell, is found by comparing the chamber
concentration with that accumulated in each test cell for each test period.
The relative overall performance is based on the average of all the five
tests. Chamber pCi/1 - Test Cell pCi/1 x 100% - % reduction in radon
Chamber pCi/1 concentrations
MATERIAL TESTS AND RESULTS
Because most of our tests were performed under contract to our clients
and we have agreed to keep these clients confidential, the following
discussion on the products tested refers to these products generically rather
than by brand name or manufacturer. The following graphs represent the tests
and the results of the many different products that manufacturers thought
could be useful in the radon field. The vertical bars represent the radon
concentrations measured in the test chamber and the test cell for each product
sample. The bottom horizontal axis represents the elapsed time in hours
-------
between sampling. The tests and the results are both graphically depicted
and above each bar, the actual numerical value is given in pCi/1 Also the
average radon concentration over the testing period for the test chamber and
the average percent reduction for each product sample is shown.
For each of the following graphs, sample preparation procedures and the
base materials in the product, when known, are discussed. The products were
prepared for the tests as they would be used in the field and in accordance
with the manufacturer's recommendations. However, as stated earlier we
interested in evaluating the product only, not the material it may be bonded
or applied to. For example, joints were constructed by using two pieces of
acrylic, which is dense and not permeable to radon. In the field caulks used
to seal joints may be very effective, but radon could bypass the joint or
crack and diffuse through the host material that is being sealed.
CAULKS
Graph #1 represents the results of the tests conducted on a gun grade
caulk of 1 part polyurethane. Sample numbers 1, 2, and 3 represent three
different size joints. The joint aspect ratios for the samples are as
follows:
Sample #1 - 1/2" deep x 1/4" wide
Sample #2 - 1/4" deep x 1/2" wide
Sample #3 - 3/4" deep x 3/4" wide
The joints were constructed by using two pieces of acrylic, the thickness
being equal to the depth of the joint. After the joints cured, the samples
were sealed to the open end of the test cell. The data indicates that all 3
test samples and the control performed equally.
Graph #2 represents the results of the tests conducted on gun grade
caulks supplied by a different manufacturer than those of Graph # 1. Sample
#1 is a silicone base caulk and sample #2 is a siliconized acrylic latex based
caulk. Sample #1 was constructed with a joint aspect ratio of 1/2" deep x
1/4" wide and for sample #2 the joint aspect ratio was 1/4" deep x 1/4" wide
These caulks performed similarly to the polyurethane caulks.
Graph #3 represents the results of the tests conducted on gun grade
caulks and a brushable sealant that would be applied to concrete or masonrv
surfaces. These materials were supplied by the same manufacturer as those
represented in graph #2. Sample #1, a brushable sealant of unknown
derivation, was prepared by coating a filter paper and sealing it to the
end of the test cell. Sample #2 is an elastomeric copolymer based material
and sample #3 is a metallic sealant. Sample #2 was constructed with a 1ol t-
aspect ratio of 1/4" deep x 1/4" wide. Sample #3 was also constructed with a
joint aspect ratio of 1/4" deep x 1/4" wide.
Overall, regardless of the base material in the caulks, they performed
about equally in radon reduction as represented in graphs #1, #2, and #3
CEMENTITIOUS MATERIALS
Graphs #4, 5, 6, and 7 depict the test results for cement based
materials. All of the samples were formulated by the same manufacturer and
the exact content of these materials were not known by Versar. Each sample
-------
Graph #1 - Sealants; Caulks
12000
10000 -
8000
6000 -
4000
2000 -
o o ° in
0 (s. CO N
CM PI N W
U) O CO O
r-, c- o
CM CM cm co
O u to o
oo r- co oo
CVI CM W W
t- o m
-------
CO
CO
00
Graph #3 - Sealant; Caulks
12000 i
10000 H
8000 H
6000 H
4000 H
Testing Chamber
Sample #1 Avg. Red. 92.4%
Sample #2 Avg. Red. 88.3%
Sample #3 Avg. Red. 86.9%
Control Avg. Red. 95.7%
7720 pCi/l
Test Chamber
Avg. Concentration
2000 H
o
CD CO
00 C\J
Elapsed Time (hrs)
Graph #4 - Sealants; Cementitious Surfacing Materials
oo
10000 1
8000
6000 H
4000 J
2000 J
^ a) n n
m in
-------
was cast to a diameter of 4 1/2" and approximately 3/32" thickness. These
samples are surfacing materials to be applied over concrete and/or block
substrates. Sample #1 on graph #4 is a premixed material shipped from the
manufacturer. Samples #2, 3, and 4 were mixed at the testing site by applying
water.
Graph #5 represents the results on the tests of four samples of materials
that are also used as surfacing materials over concrete and block substrates.
These samples were cast to the same dimensions as stated for Graph #4. The
main difference between these materials and those of graph #4 is that
plasticizers are used as part of the material matrix.
Graph #6 represents the results of the tests conducted on three samples
of materials that are used for patching and sealing around pipes and/or water
plugs. As with the samples shown on graphs #4 and #5, these samples were also
cast to the same dimensions. Unlike the previous samples that were designed
as surfacing materials, these samples are hydraulic cements.
Graph #7 represents the tests and the test results for three samples that
are used as surfacing materials over concrete and block substrates. However,
there are differences between the materials. Sample #1 is a 50/50 mixture of
the products identified as sample #2 from graphs #3 and #4, respectively.
There was relatively no increase in performance. Sample #2 is a retest of
sample #4 from graph #5, and the results were within 3% for average reduction
of radon between the two tests. Sample #3 was a clear coat glazing material
to be applied over cement finishing products. This coating was applied over
filter paper and allowed to cure before testing. The results were not very
impressive, as can be seen from the graph. As a point of interest, the first
time this coating was sent to us the manufacturer had applied the coating to a
piece of mylar. The sample had a 99% average reduction in radon transference.
At the completion of the test, we found that the sample could not be destroyed
because of the mylar. While not specifically tested, it would seem that mylar
containing materials may have a potential use in the radon remediation
business.
MEMBRANES
Graph #8 represents the results of the tests conducted on three samples
of products used primarily to waterproof foundations. The sheets of materials
were supplied by the manufacturer with instructions to laminate the products
as they would be applied in the field. All three samples were cut to 7" in
diameter, laminated, and sealed to the test cells. The samples when laminated
were approximately 1/4" in thickness. The material was rigid and appeared to
be a bituminous-like product reinforced with fiberglass-like material. One
side of each sheet had adhesive qualities to facilitate the bonding of two
sheets. In each of the three samples the bottom sheet was joined to form a
seam. Sample #1 had no vapor barrier. Sample #2 had a cellophane like vapor
barrier and sample #3 had a 2 mil. thick vapor barrier. Observing the graph,
samples #2 and #3 performed similar to the control, but sample #1 did not
perform as well. The vapor barriers could have had some effect in the overall
performance.
Graph #9 represents the results of the tests conducted on a special
formulation that the manufacturer does not want divulged. This coating was
-------
Graph #5 - Sealants; plasticized Cement Coatings
12000 -|
10000 -
8000
6000
4000 -
2000 -
Testing Chamber
(3 Sample #1
H Sample #2
E3 Samole #3
u, Sample #4
OlW M
r\ i O .
o co
n w w
C\J " cvj
CVJ oi
in cm
Avg. Red. 56.7%
Avg. Red. 70.1%
Avg. Red. 74.1%
Avg. Red. 75.1%
8810 pCi/l
Test Chamber
Avg. Concentration
3 3
Elapsed Time (hrs)
Graph #6 - Sealants; Cementitious Waterplug Material
CO
in
G)
CO
20000 -1
15000 -
10000 -
5000 -
CO ID
co ir> o>
o o 00
en
46
Elapsed Time (hrs)
¦
Testing Chamber
a
Sample #1
Avg. Red.
71.1%
¦
Sample #2
Avg. Red.
72.2%
u
Sample #3
Avg. Red.
76.0%
~
Control
Avg. Red.
99.7%
T- 00
~- o
T- o
to if) CO
CVJ
xf
%
///'
///j;
|
%
I
88
. 10978 pCi/l
Test Chamber
Avg. Concentration
-------
Graph # 7 - Sealants; Cementitious Coating
•s:
O
a
c
o
10000 n
8000 -
6000 -
4000 -
2000 -
Testing Chamber
Sample #1
Sample #2
Sample #3
Control
112
Elapsed Time (hrs)
Avg. Red. 85.4%
Avg. Red. 71.2%
Avq. Red. 28.8%
Avg. Red. 93.2%
6386 pCi/1
Test Chamber
Avg. Concentration
Graph #8 - Sealants; Membranes
8000 -i
o
a
c
o
+3
a
*3
c
4)
O
c
6000 -
4000 -
2000 -
¦
Testing Chamber
H
Sample #1
Avg. Red.
81.0%
¦
Sample #2
Avg. Red.
92.7%
~
Sample #3
Avg. Red.
92.0%
~
Control
Avg. Red.
91.9%
CVJ
¦«fr
N O W N
to O) O)
co co co
co cvj cvj
rt (O f-
co co co
in ® cm
CJ a>
C\J CM cm
4876 pCi/l
Test Chamber
Avg. Concentration
44 68
Elapsed Time (hrs)
-------
Graph #9 - Sealant; Paints
30000 -i
25000 - "
20000
15000
10000
5000 -
0 ¦+-
Testing Chamber
Sample 1 Avg. Red. 40.0%
Sample 2 Avg. Red. 74.6%
Avg. Red. 97.7%
Control
oo m
13948 pCi/l
Test Chamber
Avg. Concentration
72 ge
Elapsed Time (hrs)
120
Graph #10 - Sealants; Foam (expandable)
12000
10000 -
8000 -
6000 -
4000
2000 -
¦ Testing Chamber
B3 Sample #1 Avg. Red 94=/
~ Control Avg. Red. 95.8%
6953 pCi/l
Test Chamber
Avg. Concentration
74
E,aPseti Time (hrs)
1 18
139
-------
prepared specifically for the purpose of sealing block (concrete masonry
units) surfaces and is to be applied with a spray gun. These samples were
prepared by cutting the face shell from a standard block. The face shells
were cut in half and each mounted and sealed to a test cell. Sample #1 is the
treated block with all exposed surfaces and edges coated with the special
formulation. Sample #2 is the untreated block. The overall performance of
this coating was not up to expectations. Limited space prevents the showing
of two additional follow-up tests that were conducted. However, the results
were similar, with only the control showing consistent results in percent
reduction of radon.
Graph #10 represents the results of tests conducted on expandable
polyurethane foam. A screening material was molded to fit over the open end
of the test cell. The screen was used as a porous backing material and the
foam was sprayed onto it. The foam expanded to a range of 1/2" to 1"
thickness. After a seven day curing time, the test was conducted.
Additionally, the test chamber was pressurized to induce a negative pressure
of 2.5 inches of water inside the test cell. This would equal or exceed the
negative pressures this material would be subjected to in a hollow core
foundation wall that was being evacuated with a wall depressurization system.
The material performed remarkably well, and it is expected that the skin
coating that forms over the surfaces exposed to air may have been a
contributing factor to its performance.
MISCELLANEOUS MATERIALS
Graphs #11, 12, and 13 represent the results of the tests conducted on a
variety of floor coverings. These are not classified as one of the five
categories under sealants. The manufacturer requested these materials be
tested, as they thought that under certain circumstances these materials, when
bonded properly to cracked concrete substrates, may have a favorable impact on
reducing radon infiltration. The manufacturer supplied the floor covering
samples. Each sample was cut, fitted, and sealed to the open end of the test
cells. Graph #11 shows the results of one at three individual tests for
carpet samples #1 and #2. The second test was a duplicate test, and the third
test was conducted using the same carpet material, but with a seamed joint.
All three tests were within a 2% maximum deviation in radon reduction for each
sample and the control.
Graphs #12 and #13 represent the results of the tests conducted on
additional carpet samples supplied by the same manufacturer. The exception
was sample #2, as represented on graph #13. This sample consisted of a sheet
vinyl material as would be used on kitchen floors. In general, the
performance of all of these samples appear to be based on how the manufacturer
bonded the carpet threads to the backing material. For example, sample #2 on
graph #12, the carpet threads were very dense (piles or loops per square inch)
and bonded in a dense polyurethane backing material. This was an expensive
industrial grade carpet and performed surprisingly well in the tests. Sample
#1 of graph #13 was an inexpensive carpet with no backing. The carpet threads
were woven into a separately constructed backing material. It is suspected
that the glues and resins may have been the biggest contributing performance
factor, or there would have been none at all.
-------
Graph #11 - Floor Coverings
14000 n
12000 -
10000
8000 -
6000
4000 -
2000 -
0 4-
U>
CD
CM
CVJ
Testing Chamber
Sample #1 Avg. Red. 66.1%
El Sample #2 Avg. Red. 87.3%
~ Control Avg. Red. 95.9%
7234 pCi/l
Test Chamber
Avg. Concentration
142
166
Elapsed Time (hrs)
Graph #12 - Floor Coverings
15000 n
=¦ 10000 -
5000 -
0 -4-
Testing Chamber
Sample #1 Avg. Red. 29.7%
Avg. Red. 96.7%
Avg. Red. 23.6%
Avg. Red. 96.3%
Sample #2
Sample #3
Control
~
8184 pCi/l
Test Chamber
Avg. Concentration
75
Elapsed Time
152
168
-------
Graph #13 - Floor Coverings
m
co
o>
¦ Testing Chamber
B3 Sample #1 Avg. Red. 15.7%
@ Sample #2 Avg. Red. 63.0%
~ Control Avg. Red. 93.9%
n oi a
m co
ID (o
C\i
-------
SUMMATION
Many commercial products are available and are being used for the purpose
of controlling radon gas infiltration into buildings. Recently, a number of
manufacturers have begun developing and marketing materials specifically for
radon entry control. These materials have been developed, marketed, and used
without real data regarding the radon resistant properties of the product.
The radon mitigators have selected these products based on physical or
chemical characteristics, or simply availability.
Obviously the physical and chemical properties of the products are
important considerations, however, these properties are clearly secondary to
the intended function (i.e., as a radon barrier). In the past, the success of
a product was based on the comparison of pre- and post-remediation test
results of the ambient air in the structure. However, it was impossible to
determine if the success or failure of the product was due to the product,
workmanship, or the application of the product.
This paper presents a testing methodology that has over come many of the
constraints associated with the previously used testing methodologies. The
equipment and testing techniques described provide mitigators with
quantitative radon permeation data on a specific product to assist them in
making material selection decisions. The methods are also useful to
manufacturers in evaluating the effectiveness of existing products for radon
mitigation applications, and in developing new products designed for such
applications.
The work described in this paper was not funded by the U.S.
Environmental Protection Agency, and therefore, the contents do not
necessarily reflect the views of the Agency and no official
endorsement should be inferred.
-------
V-8
TITLE: Recommendations to Reduce Soil Gas Radon Entry Based on an Evaluation
of Air Permeability of Concrete Blocks and Coatings
AUTHOR: J.S. Ruppersberger, EPA - Office of Research and Development
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
A variation of over 50 in concrete block air permeability
(0.63 to 35 SLPM/m2 at 3 Pa) emphasizes the need for effective
block selection guidance that could help reduce soil gas entry
and its associated risk and cost. This large variation was
found in a sampling of only 9 of the 1000 or so different commer-
cially available blocks in the U.S. Fortunately, comparison of
the results of the carefully controlled tests to simpler tests
and observations shows enough correlation to expect that simple
guidance would result in the correct selection of less permeable
block in a majority of cases. A simple field test is described
that correctly differentiated between low and high permeability
blocks. Even in the absence of the simple test, the surface ap-
pearance of the blocks have been found to correlate extremely
well with air permeability, that is, inexperienced observers cor-
rectly differentiated between low and high permeability blocks by
selecting blocks that appeared less permeable.
Six different coatings were evaluated and all were found to
be very effective (over 98%) in reducing air permeability across
the face of concrete block walls when sufficient amounts were
carefully applied. Selection of a coating is therefore in-
fluenced more by factors associated with the wall itself that are
expected to affect performance , service life and cost. Of the
coatings evaluated, the cementaceous coatings were found to be
highly effective (over 99.5%), and least expensive.
In general, if air permeability data is not available,
select the type of block that looks least permeable. If the
resulting structure is subsequently found to have indoor air
quality problems due to soil gas entry, and the mitigation plan
includes coating the block surfaces, it will generally require
less material for the relatively smooth surface and result in a
less permeable coated block. From a building system perspective,
minimi2ing soil gas entry and controlling ventilation through
other preferable channels, may be as worthwhile from comfort,
energy, and other indoor air quality considerations, as it is for
reducing risk due to radon entry.
-------
Session V:
Radon Entry Dynamics -- POSTERS
-------
A Simple Model for Describing Radon Migration and Entry int0 Houses
by: Ronald B. Mosley
U.S. Environmental Protection Agency,
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina '
ABSTRACT
An approximate analytical solution to describe radon transport
in soil having uniform properties is applied to the interaction of
a soil depressurization system with radon emission to the
atmosphere. The study addresses the question of whether soil
depressurizing mitigation systems are likely to significantly
increase the local ambient radon levels by increasing the emission
rate from the soil. While the model predicts that the operation of
a soil depressurization system usually increases the total emission
rate this increase does not appear to be significant except for
soils with high permeabilities. This is true because the decrease
in emission rate from the soil surface tends to compensate for the
increase in emission rate by the mitigation system unless the soil
permeability is quite high. For permeabilities below 2 x I0~n m2,
the increase in total emission rate of a single mitigation system
and its sphere of influence is less than 1%. Even for
permeabilities greater than 4 x 10 m the increase in total
emission rate associated with a single house and its sphere of
influence is probably not greater than 50%. A 50% increase in the
emission rate from a single mitigation system does not translate
into a 50% increase in the ambient radon level, if only 10% of the
soil surface in a community with permeability greater than 4 x lo~10
m2 is associated with operating mitigation systems, the local
ambient level might be expected to increase by about 5%.
This paper has been reviewed in accordance with the U.S.
Environmental protection Agency's peer and administrative review
policies and approved for presentation and publication.
-------
INTRODUCTION
A large number of radon mitigation systems have been installed
in U. S. houses during the past few years. One of the more
effective mitigation technigues uses active soil depressurization
to reverse the direction of flow of soil gas through the building
substructure and to flush high concentrations of radon from beneath
the building. In many instances, high radon concentrations have
been measured in the exhaust of the mitigation system. In some
instances, the high radon concentrations combined with the
relatively high flow rates generated by the mitigation system have
given rise to concern for the safety of individuals exposed to
these emissions. Some of these exhausts are released under or near
decks where the occupants of the house could receive appreciable
exposures. In other instances, the exhausts are released near
ground level where children may be prone to play. Even where the
exhausts are extended to the eaves of the houses, there is concern
that downwash may result in occupant exposure. The EPA has had
inquiries from individuals about potential exposures from the
exhausts of a neighbor's mitigation system.
Most of these situations relate to the concern about potential
exposure to a few individuals as a result of increased
concentrations at very specific locations. The local concentration
near a mitigation system exhaust might become quite high if the
mechanisms for dispersing the radon in the atmosphere, such as air
movement, were inhibited by physical obstructions or, temporarily,
by a temperature inversion. Concentrations near a mitigation
system exhaust will depend strongly on the dispersion processes.
The present paper will not address the question of dispersion of
the radon in the atmosphere. It will be assumed that radon emitted
to the atmosphere is effectively dispersed in such a way that only
the average radon concentration increases.
A related question that is frequently asked is whether
communities with hundreds or thousands of mitigation systems may
actually increase the ambient radon level resulting in increased
average exposure for the members of the community. Some have taken
the issue further by coining the term "mining" of radon to describe
the high emission rate from mitigation systems. The purpose of the
present document is to explore the effects of active soil
depressurization systems on the total radon emissions to the
atmosphere.
MODEL DESCRIPTION
In order to explore the influence of soil depressurization
systems on total radon emissions to the atmosphere, the interaction
of a model house with gas transport in the soil will be described.
For convenience, and to avoid lengthy numerical analysis,
simplifying assumptions will be used to obtain analytical solutions
to the transport equations.
A number of sources in the literature (1-4) emphasize that
pressure driven flow of soil gas into the house is the dominant
-------
c?e in most houses. Consequently, the present
radon entry process in ceptUal arrangement illustrated in
formulation will use the con pg a basement house that is
figure 1. This fig ture difference, wind effects, or
depressurized (by a P ^ fche ambient air. The resulting
mechanical appliances; driving force to cause air to flow
pressure difference S®^J®S , enter the basement through openings
downward through the soil . t entry route in many basements is
in the substructure. The In some houses, the dominant entry
the perimeter wall/floor that CQnnects directly to a sump
route is a perimeter dra interior of the basement. Both these
that, in turn, is open to _ as flow into an isolated
situations will £pth below the surface,
cylinder buried at th net effect of the mitigation system
in order to ^cesses must be considered. While
on radon emission, two p system exhausts are frequently
increased emissions about the effects of the system on
th^emissioris from the surface of the soil surrounding the house.
• • radon from the surface of the soil occurs by
Emission of rad through the soil gas contained in the
molecular diffusion o flows down through the soil
pores of the sel1; the soil gas is diluted and the
the radon concen*f*jL is modified. Consequently, radon emission
concentration ^raaien ix is reduced by increased flow through
from the sur^f. ?*tion system. The net effect on radon emission
the house and .* difference in the increased emissions
?hiough tlfe m""gatio^ syltem and the decreased emissions from the
surface of the soil. ough the house and/or the mitigation
Emission , usinq expressions developed in reference 1.
system will be comp house Qr mitigation system is given
The emission rate uuuuyn
by
, (l)
Em=j cvda,
where Em is the emission rate (Bq/s) , C is the local radon activity
concentration (Bq/mJ) in the soil gas, and v (m/s) is the velocity
of the soil gas. The integration is taken over the surface of th
cylinder which represents the entry route into the house or th
mitigation system. The activity concentration at the surface
the cylinder (1) is given by ce of
cw - |
(2)
(-A.eu^2ln(2h/b)
1-exp^—^~lc\p
i-4>cot4>lN
sin24>
-------
where
C(4>)
~
cylinder,
G
particles,
X
exp =
e
h
In
b
k
Pc "
atmosphere.
= the radon activity concentration (Bq/m3) at angle ,
= the polar angle measured from the vertical axis of the
= the emanation rate (Bq/m3/s) of radon from soil
= the radon decay constant (s-1) ,
= the exponential function,
= the soil porosity,
= the depth of the basement (m),
= the natural logarithm,
= the radius of the cylinder (m),
= the permeability of the soil, and
= the pressure (Pa) in the cylinder relative to
Equation (2) does not apply when 4> = 0 because of a local
singularity. The gas velocity at the surface of the cylinder is
given by
v = I 1 \ (3>
jbfiln (2h/b) \ h-bcos^J'
where v is the velocity, p. is the kinematic viscosity of the soil
gas, and the other parameters are as previously described. It is
a relatively straight forward matter to numerically evaluate
equation (1) using equations (2) and (3). While equation (3)
represents a rigorous result, equation (2) is a rigorous solution
of the transport equations only when the contributions from
diffusion are negligible. While this approximation is adequate for
relatively large values of soil permeability, the permeability in
many localities is not that large.
Perhaps the simplest way to estimate the local effect of
diffusion on the entry of radon into the buried cylinder would be
to compute the gradient of the activity concentration at the
surface of the cylinder and integrate the resulting diffusive flux
over the surface of the cylinder. This result would not be
rigorously correct because the concentration gradient would not be
self-consistent. That is, the influence of the diffusive process
on the concentration gradient would not be accurately reflected.
However, since the calculation is easy to do, it seems worthwhile
to incorporate this approximation.
The normal gradient of the concentration evaluated at the
surface of the cylinder is given by
-------
vcnS i-4»c°t
(4)
where V is the gradient operator. The diffusive flux is given by
Jd=~DeVCn , (5)
where is the diffusive flux and De is the effective diffusion
coefficient. This flux is to be integrated over the surface of the
cylinder,
Ed=j~DeVcnda .
(6)
- « , orma-t-ion (1). represents the change in
and the result added \he hoUse or mitigation system due to
the emission rat e from the^n
diffusion at the sur surface of the soil, the flow will be
in ^ region near ^ifln ^ the s<)il surfac^ the
nearly vef*lc^* ated as if it were one dimensional for purposes
problem activity t^Hicentration and migration. With these
assumptions^ t^ transport elation bscom.s
D.
d2C _ v dC
dy2 e dy
kC+G = 0
(7)
Eguation (7) has the solution
/
1 -exp
I
N
v
2 SDe ¦ De 2 GDe
y
(8)
Since we consider only a narrow range of y near the surface, y = o
-------
the velocity can be considered to be independent of y. It has been
assumed that C(0) = o. The diffusive flux at the surface of the
soil is given by
(9)
The total emission rate of radon from the surface of the soil would
be obtained by integrating the flux (equation 9) over all the
surface. Since the surface is very large, it can be seen that the
integral would yield a very large number. The influence of the
mitigation system on the emission rate would be contained in the
relatively small differences in very large numbers. Since this
approach would require extreme accuracy in evaluating the
integrals, the following rationale is adapted.
It is known from reference 1 that the velocity at the surface
of the soil decreases approximately as the reciprocal of the square
of the distance from the house. Consequently, the practical
influence of the house and mitigation system on radon emission from
the soil is limited to quite finite distances. Let us call the
area of the soil surface for which the house and mitigation system
influence the emission rate the house's "sphere of influence."
This sphere of influence will be characterized by its maximum
distance from the house. Reference 1 shows that the fraction of
total flow occurring within a distance s of the house is given by
where s is the distance from the house, Q(s) is the flow through
j area between the house and the boundary located at distance s,
and Qt is the total flow. The sphere of influence could be
defined, for instance, as the area within a distance, s, of the
house through which 95% of the total flow passes. with this
definition, a basement 2 m deep would have a sphere of influence
f , ie.s Wlthl*\ about 25 m of the house. in the present
calculations, this convention will be adapted so that the
integrations over the soil surface will extend to 25 m. The
emission rate from the surface becomes
(10)
Eg = LfjDdx,
(11)
-------
where Es is the J^^ence^L isT th^ length ofhthe°cylinder?
and 5°i|ethedistance from the house along the surface. The total
emission rate (%> of radon into the atmosphere is given by
(12)
Er=E^E^Es.
w X-l ¦*¦ *-*
j i^fomret the above developments, some
In order to ap^y and xnte P &
is described. It
idealizations are 5eq^i"5? *n4- wall's, floor, and common joints are
is assumed that the flow except that the sump has been left
tightly sealed against air flow, exc p^ ^ ^ bage Qf
open. There is a compl connected to the sump which was
the house. This dra water collected by the drain tiles. This
designed to remove t efficient radon entry path into the
description depicts an frequently at lower pressure than
basement. Since tte taMMnt the drain tiles are
the atmosphere, the S^P tyJjs results in convective flow of air
also at reduced " to the tile and ultimately into
from the surface thro 9 circumstances, radon entry into the
the basement. Under rninnutina the radon entry into the drain
house can be computed y P heart of the method for
tile. These assumptions were « ^ ^ reference
computing radon entry Jladon entry is now extended to describe
* This idealization forwith the soil. Suppose
the interaction of a?led in this model house consists of a
the mitigation systeni 1 ^ with the mitigation fan and
collection pipe - In this case, the mitigation system
exhaust located on the reof arKJ therefore is coupled
becomes an integral P t _ jt is easy to imagine that the
directly to the dr"J] ^de to simulate the natural conditions
mitigation system couldlb t the fan speed until the pressure
of radon entry simply by ad} relative to atmosphere is the same as
difference in the drain tiles ent. It is reasonable to
when the mitigation system rate from the exhaust under these
assume that the "don mission «>tejrr^^ entry rate lnto ^
conditions would be th <=Vstem was absent. Herein lies the
basement when the *^iga_ ^ for deScribing the interaction of
essence of the present app -oil
the mitigation system^wxth tio'n of the mitigation system can
It is assumed that the p natural entry process. This
be treated as an the mathematical formulation by the
extension is reflected ;L" _ssumes that the radon entry rate into
^hfd^tn'tfl-Tnd Jns^ntly out the system exhaust) would be
-------
the same as if the depressurization had occurred by reducing the
pressure in the basement (for instance, with a blower door).
While the idealized model described here is a very specific
case in which radon entry into a buried tube can be used to
describe radon entry into a particular type of basement, it is
believed that the formulation can be applied for other basement
construction details. For instance, if there is no drain tile, but
there is a perimeter crack at the wall/floor joint, it is believed
that entry through the crack can be simulated by flow into an
appropriately sized cylinder. In reference 1, it is shown that the
model is not very sensitive to the diameter of the cylinder.
The total emission rate in equation (12) was evaluated for a
model basement house having 144 m^ of floor area in contact with
soil. The length of the perimeter drain tile around the basement
is 50 m. Values of the parameters used to characterize the soil
properties are listed in Table 1. Figure 2 shows the normalized
total emission rate as a function of pressure for four values of
soil permeability. The emission rates have been normalized to the
values at zero pressure. The emission rate from the exhaust is
taken to be zero when the pressure is zero. Consequently, all the
emission is from the surface of the soil when the pressure is zero.
This total emission rate represents all the emissions from the
mitigation system and the soil within the assumed sphere of
influence of the house. Although the calculations were done for
rather modest values of radium content in the soil (emanation
rate), the normalized values are independent of the source
strength. Of course, these calculations assume that the radium is
uniformly distributed in the soil and that the transport properties
of the soil are uniform. From Figure 2 it can be seen that the
mitigation system produces very little increase in the total
emission rate when the soil permeability is less than about 2 x 10"
11 m2. For a permeability of 7.8 x lO--11 m2 the increase in total
emissions is about 5% at 10 Pa and 30% at 40 Pa. Even at the
rather high permeability of 3.9 x 10~10 m2, the increase is about
38% at 10 Pa and 245% at 40 Pa. There probably are few localities
with permeabilities higher than 4 x 10"10 m2.
Figure 3 shows normalized emission rates as a function of
pressure for the individual emission sources, the soil surface, and
the mitigation system. Note that, while the emission rate from the
mitigation system starts at zero for zero pressure and increases
with increasing pressure difference, the emission rate at the
surface of the soil starts at a maximum at zero pressure and
decreases with increasing pressure. This is a direct reflection of
the influence of the air flow at the soil surface on the
concentration gradient. Note that the two phenomena approximately
compensate each other. In fact the total emission rate increases
by less that 1% over the pressure range to 40 Pa, while the
individual rates change by about 10%.
Figure 4 shows normalized emission rate as a function of
pressure for the individual rates as well as for the total emission
rate. These curves represent a soil permeability of 1.6 x 10"11 m2.
While the individual emission rates vary by about 25% over the
pressure range, the total emission rate varies by less than 4%.
This result indicates that the changes in the two emission
-------
processes almost ^ ^on rates are equal at about 26 Pa.
in figure 5 the two changed by about 50%, the total
While the individual r;at: d by about 18%. However, as the
emission rate has only i gation system emissions begin to be
permeability i««ea®®8'ath® Ability It 3.9 x 10"^ m2, the
more important. At a * are somewhat more dramatic as seen
variations in the emissi , rates are equal at about 6 Pa, while
in figure 6. , The nearly linear over the pressure
the total emission rate ^ surface of the soil at 40 Pa is
range. The emission rate r value> Remember that this effect
only about 20% of inle1s sphere of influence. In fact most of
only applies over th . OCCUrs near the house.
the decrease m ^^^"meability of 2 x 10~9 m2, the emissions are
At the very hig:h p®™iticfation system above pressures of about
totally dominated by the mxx. ^ hQ the factor of 6 increase in
10 Pa as shown in f1(^re ™itiaation system is very impressive, it
emission rate due to th First of an, few soils have such high
probably is not realisy • mitigation system could not maintain
permeabilities. Sec°™*^; flow rates that would result from such
such high pressures at t ^ permeabilities of 3.9 x 10 10 m2 and
high permeabilities, r respectively, a pressure difference
2 x lO"9 m2 in f igures 6 and 7,not be maintained by the
greater than 10 Pa pro ^ in the total emission rates would
mitigation system. Tn ^ 3Q0% fQr permeabilities of 3.9 x
then be limited to> respectively.
10'10 m2 and 2 x 10 J® ' .£ t particular increase applies only
It roust be of influence. These emissions win
to a given house and rts P ai^ gince the sphere Cf influence
be mixed with all the an . is HKely to cover only a small
of installed mitigatiaoil surface, the increase of the ambient
fraction of the t°tal soil surt ^ 'smaller fraction than those
radon levei will incre ^ tion systems. For instance, if
associated with f the soil in a given community is
only 10% of the surf a systemsf then the ambient radon
associated with ^ltig ase by oniy 10% of the increase of the
concentrations would in iQn system. Even for the high
average individ"al above this would translate into the range
permeabilities discussed above, this e Qf increased
% - 30% increase m be compared with typical natural
ambient radon levels sho that Qccur from one location to
variations of 300% (io
another
CONCLUSIONS
—
radon entry into many basement houses
It has been argued racjon entry into a buried cylinder
can be estimated by ^^imeter drain tiles or wall/floor cracks,
which simulates either pe^"1 . this model can be extended to
It has been further argued^iSuriZation system with radon
simulate the interact!):>n _dvantage of this simulation is that
in the soil. The PrlIV y . obtained to describe the radon
analytical solutions can addresses the question of whether
migration and entry.
-------
soil depressurization systems significantly increase the ambient
radon concentrations. While the model predicts that the total
emission rate from the house and its sphere of influence is almost
always increased by the operation of a soil depressurization
system, the increase is not significant (less than 1% increase per
mitigation system) for soil permeabilities below about 2 x 10-11 m2.
As the permeability increases so does the total emission rate.
However, the pressure difference that the mitigation system can
sustain decreases as the permeability increases. Consequently, the
maximum increase in the total emission from a house and its sphere
of influence due to the operation of a depressurization system is
probably not more than 50%. In cases of unusually high
permeability (2 x 10~9 m2) the increase per house could be 300%.
Even the higher rates of increased emissions would lead to average
ambient levels smaller than typical indoor radon levels.
-------
REFERENCES
1. Mosley, R.B. 1990 A
SjS^iS oXaSHnaiS. Environment. 10/15 - 19/90, Richland,
WA.
iqae Analytical and numerical models for
2. Mowns, R-J- 8 * exhaust ventilation on radon entry in
estimating the ef:feet of exnaius ^ ^ _ 22Q61. M>s< ThesiS/
houses with basements or craw P CA
Lawrence Berkeley Laboratory, Berkeley,
^ M 1qqr Predicting the rate of radon-222 entry
Iron SInto tke basest Of a dwelling due to pressure-driven
air flow. Pro*- Dosim--
n lAQi Sources of indoor radon in houses: a review.
4. Bruno, R.c auuiv
JAPCA 33: 105 -109.
** u w and R.G. Sextro. 1989. Techniques of measuring
?he indolr°"^Rn'source potential of soil. Environ. Sci. Technpl^
23: 451 - 458.
-------
TABLE 1. Values of the parameters used to perform the calculations
in this paper
Parameter Baseline value Range of variation
k 1.6 X 10"11 m2 1 x 10~15 - 2 x 10~9 m2
Pc 4 Pa 0 - 40 Pa
De 2.0 x 10"6 m2/s 1 x 10~6 - 4 x 10"6 m2/s
1.7 x 10~5 kg/m/s
e 0.5
G 0.0334 Bq/m3/s
X 2.11 x 10~6
b 0.0508 m
h 2.0m
L 50. m
-------
M <4-4
CfJ
Qfi 0)
r
c
-P
(0
0^
5
o
i-H
IW
It
a)
•P
cn
c
•H
3
0
-C
W
e
Q)
¦P
01
>i
0)
O
W
a)
w
3
O
x:
0)
£!
Eh
(U
3
o>
•H
El,
-------
o
a
i
£
¦t
5
1
"tl
>s
n
c
5
01
52
E
T3
g
w
•
M
E
-5.
2
sr
O
>
CM
5=
qpm
1
•
i
S
*55
n
i
~T~
Oi
I
CO
~T~
o
~r~
CO
o
—r~
in
~i r~
K1 CM
• •
o o
1
o
. in
m
_ o
ro
-in
CM
o
Q-
K o
CM
. in
. o
- in
*- o
>-((N
07
dJ p"'
3 *
W CM
w .
W
¦p a)
(0 o
3
o
w
c
o
c
o
g (A
Q) U)
•H
T3 E
a) a)
(H ^
a«
E'S
0 -h
2-o
••5
QJ §
CT 0)
•H .Com
fa +J g
ejDy uo|ss}LU3 pazipuiJON
-------
u
0
<4-1
a
1
t/i
¦P
t/i
•H
•rH
tH
s
•r-l
a) xt
ifl
T>
a)
a)
e
N
u
•H
0)
r-1
a
(3
e
>
u
3
u
tr 3
o
pL, IM
r
I
—r-
tf)
ts|
CN
CM
CM
uo|S8|tu3 P©zhdlujon
-------
¦0
*i 1 r
^ 4-1
o
>1
¦p
0)
3
0) -H
01 £)
d) (0
u
a
o
a)
g
Q)
a
C J3
2^
•H -h
¦P >
u
§M
o
A3 I)
B®
£5
273
M c
c «>
5»
03 «
W p
¦H £
o
U1
c
0
E
0)
T3
3 -H
tr'd vo
•H c •
fa'H rl
X
-------
_o
~r ^
8
§
S
I
i
n
>»
w
O
o>
TJ
§
0
M
3
-§>
3
£
T
T
m
ro
o
KJ
m
cs
CN
- m
. o
- IT)
00 to t CM
, « • »
o O O o
®py uoj8S|Oi3 p«*llDtuJ0N
1-4 4-1
o o
0) #
U -H
3 rH
cn -H
V) Si
0) n
J-i m
a g
Q)
O a
c jC
¦H 5
-U s
u s
§ M
2 «
<0 °
Ss
10 2
JUS
2 "o
w c
c
o
-H w
(/] <1>
to P
•H >-l
e
Q)
TJ
0)
N
3
O
W
c
o
w
fl3 0)
g
o
tZ
eCB
X
3 -H
CT"0 co
¦H C •
fci *h
-------
§
M 1
0) 4-1
M -H
3 r~H
W -rH
U) XI
Q) rtj
U a)
a 0
u
a)
3
Cr> t3
•H C •
fc'Hn
9JDfcJ UO|SS]LU3 P©Z||DUJJON
-------
o
V)
<0
•0
¦4—
o
E
8"
I
t
~T~
<0
~r~
m
_r_
ro
§
•+~
n
>»
V)
O
o>
~r~
CM
8
I
i
I
g
8
1
-»n
to
. 1_ o
K)
. in
CM
a
Q_
CM
-»n
_ o
m
o
o
Q)
3
Ui
W
(U
a
O
"w
0
•P
XI
(0
a)
S
Vh
0)
a
c xx
•2 5
¦p s
o s
§-
4-i
a
•p
o
(0
w
<0
W
a)
•p
n
u
c
•H «
Sg
•H >-•
•H
x
3 -H
CT>T3 O
•H G .
k< -H C\)
©|Dy UOJ8S1UJ3 p»Zl|DUJJO|s|
-------
VP-2
TITLE: Effect of non-Darcy Flow on the Operation of Sub-slab
Depressurization Systems
AUTHOR: R.G. Sextro, Lawrence Berkeley Laboratory
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
Sub-slab depressurization (SSD) systems are widely used to limit radon entry into
residential and non-residential buildings. Satisfactory performance of these systems
depends upon reversing the norma! pressure gradient across the building substructure at
all major soil gas entry locations. Since many of these entry points are at the periphery
of the floor slab, adequate extension of the depressurization field to these locations from
the SSD suction point(s) is essential, In principal, a high permeability medium (e.g.
gravel) below the slab enhances this communication. Under some conditions, the flows
through this region are characterized by high Reynolds numbers, where the soil gas velo-
city is no longer a linear function of the pressure gradient.
This non-Darcy behavior may have important consequences for the design and applica-
tion of SSD systems because it can, in principal, adversely affect the extension of the
depressurization field in the sub-slab region.
We have constructed a horizontal flow column for measuring the air flow rates at
different applied pressures in gravels collected at various building sites in the Spokane,
\VA, vicinity,
These data are then used to evaluate the coefficients in the non-Darcy equations used to
describe this flow. This equation has, in turn, been used in a fully three-dimensional finite
difference model to simulate the pressure-field extension and the flows of soil gas and
radon in the sub-slab gravel layer and surrounding soil due to SSD system operation.
We first present here results of the flow column measurements and the non-Darcy
flow equations for various kinds of gravel. We also present predictions from the numeri-
cal model for SSD system effectiveness for these different gravels, using geometrical
specifications of an actual house.
-------
FEE: 07 '91 11 = 40 USEF'H SUPERFUNE' OF I PEG J
Effects of Humidity and Rainfall on Radon Levels
in a Residential Dwelling
by: Albert Montague, Ph.D., P.E.
William E. Belanger, P.E.
U.S. EPA, Region III
Philadelphia, PA, 19107
Francis J. Haughey, Ph.D.
Rutgers University
New Brunswick, N.J. 08903
Although the effects of precipitation and barometric pressure are not well
defined with regard to radon entry into buildings, evidence continues to be collected
showing that there appears to be some type of relationship between these factors.
Nazaroff showed that on one day of heavy rain with a moderate drop in barometric
pressure the radon concentration in a dwelling rose to a level more than five times
higher than the average (Nazaroff, 1983). Another study from which this report draws
showed a strikingly similar relationship of radon with rainfall, barometric pressure and
radon levels, (Montague, 1990). It showed that the average radon level in the dwelling
increased to over eight times its normal concentration during a heavy rainfall event. A
drop in barometric pressure also occurred, but the minimum occurred after the radon
peak, in at least one instance. Surface soils were not frozen during the rainfall event.
Other factors not previously reported are the focus of this report. These appear to be
P.£
VP-3
-------
FEE 07 '91 11: 40 USEPh SUPERFUND OFC REG 3 ':
contributing to the complexity of the radon transport and entry problem.
A residential dwelling located in a rural area near Media, Pennsylvania, was
selected for a radon study that covered a five-month period from November 1988 to
March 1989. Meteorological and radon data were collected continuously during this
period. The home, a two-story colonial wood frame structure with a full basement, was
five-years old at the time of the study. It was selected because prior short-term radon
observations showed levels in the basement to range between 15 to 20 pCi/L., and
because the homeowner was amenable to having instrumentation installation in the
house.
The house was constructed on the south side of a hill, near the 250-foot contour
line, with a southerly orientation that is unobstructed by trees. On a sunny day solar
exposure was continuous. However, the area behind the house is shielded from the
direct rays of the sun because of the dwelling's shadow. This setting decreases the
probability of freezing surface soil except for the area directly behind the dwelling
during extended periods of sub-zero temperature. The hill, with a maximum elevation
of approximately 375 feet above sea level, shields the dwelling from northerly winds and
has a surface incline of about 10 percent.
The steepness of the incline affects the rate and extent of water runoff during
periods of precipitation and snowmelt. The topography simultaneously affects the rate,
-------
FEB 07 '91 11:41 USEF'h SUPERFLIND OFC REG 3 P_
and quantity of surface water that can percolate into the soil. This ultimately affects
subsurface radon kinetics (EPA 1987). The subsurface geology has the Glenarm Series,
metamorphic formations mainly of schist and phyllite with some marble, and quartzite
of Peters Creek, Wissahickon, Cockeysville, and Setters Formations (DOA, 1963). The
local soil has been classified as the Glenelg Series, which is a moderately deep, well-
drained soil of uplands. The surface layer is a silt loam. The subsoil is also silt loam,
but it contains a little more clay than the surface layer. In some locations there are flat
channery fragments, nearly two inches across, in the surface layer. Slightly larger
fragments are in the subsoil. Beneath the subsoil there is a strong brown to reddish-
brown loam that contains many bright fragments of mica.
The dwelling is conventional from both a design, and structural standpoint. It
contains two fireplaces, power (exhaust) vents in the bathrooms, and an electric clothes
dryer verted to the outdoors. When placed into operation each of these items
discharge indoor air to the outside air. The forced air heating and air-conditioning
system uses a conventional gas-fired furnace and heat pump. The inclusion of a heat
pump makes the heating system more efficient; however, it is not the usual residential
heating system configuration. The heat pump is automatically shut off when the outside
temperature falls below 40 degrees Fahrenheit. It is this point that the gas-fired
furnace automatically begins to supply heat to the house. The gas furnace was in
operation for most of the study period.
-------
All of the meteorological readings were continuously recorded on strip charts,
and collected at the end of each week, except for the anemometer strip chart readings,
which were retrieved once a month. Basement radon levels were recorded on hard
copy printout every three hours, and collected at the end of each week. All of the
meteorological values were manually transformed from analog to digital form, by
visually estimating the average value that best represented the strip chart values for
each three-hour interval. Lotus 1-2-3 was used to conduct the analysis, because it has
the ability to process the appropriate amount of data, conduct statistical analyses, and
permits the user to generate graphs with relative ease.
Examination of all the radon data revealed a number of sharp "spikes" in radon
concentration, during which the radon concentration went to several times its "normal"
concentration. A selective retrieval was performed, to examine potential causes for
these events. Four separate events, each having rainfall amounts that exceeded one (1)
inch, were identified within the five-month period of study. It should be noted that the
four events that were selected all occurred when the dwelling was operating under the
same experimental mode; specifically, when the furnace was drawing inside (house) air
for combustion. The absence at these events during operation on outside air is not
thought by the authors to be related to the outside air supply, since the events
appeared to be related to rainfall. With the four events identified and their time
frames established, data retrievals were made for relative humidity, basement radon,
outside temperature, barometric pressure, and the wind speed and direction dwelling's
differential air pressure (vacuum). (Figures 4 to 7). Wind speed was found to have no
-------
located inside the dwelling envelope. There is a two inch gap around the outside of the
chimney which leads to the attic. This combustion system provides a conduit for the
continuous release of indoor air to the outdoors, and keeps the house under a slight
vacuum (~ .02"wc) throughout the heating season. Several factors responsible for this
phenomenon include: the buoyancy of the dwelling's indoor (warm) air relative to
outside (cool) air, outside wind-speed (velocity) that creates a pressure drop at the
mouth (top) of the chimney commonly known as the "Bernoulli effect"; and the
combustion of fossil fuels, e.g., gas or oil, which increases the temperature differential
between the combustion exhaust gases flowing through the furnace, and up the chimney,
with the surrounding ambient (outside) air, when the furnace is operating (Figure 1).
The furnace was not found to produce a measurable negative pressure in the house,
while fireplace operation caused a marked difference.
The original study focused on the effects of depressuriza-tion in a dwelling,
differential air pressure, and the kinetics of radon entry. This study focused on the
operation of a residential fossil fuel central heating appliance that is located within the
living space. The furnace combustion system was completely isolated from the air
within the dwelling envelope during predetermined intervals and returned to its normal
configuration for the remainder of the time as a control. This was accomplished by
mechanically providing only outside air to the appliance (Figure 2). The investigators
took care to assure that no unsafe condition would result from the outside air supply to
the furnace. The device used has been tested and approved by Underwriters
Laboratories. Provisions were also incorporated in the modified furnace combustion air
-------
FEB QT 'yi 11-4Z USEF'ft SUF'ERFUND OFC REG 3
supply system, to allow for switching easily from internal house air.
A five-month cold season monitoring period was selected to minimize the effect of
perturbations associated with the prolonged opening of windows and doors, to minimize
seasonal variations, and to collect a sufficient amount of data to obtain a valid result.
Within the period of study, a two-week "outside air" followed by a two-week "inside air"
interval was selected. This cycle was repeated throughout the entire investigation.
Several calibrated radon monitors were used in all floors at the house in the
original study; however, only the Pylon model AB-5 with passive call that was placed in
the basement will be the monitor of interest in this report. It was programed to
provide three-hour radon averages or eight readings each day. Data generated by the
Pylon monitor was collected eveiy two weeks to coincide with intervals selected to
switch the furnace system from indoor air to outdoor air.
Since the primary driving force moving radon into the dwelling is the soil-gas
pressure relative to the interior air pressure. A methodology developed by Belanger
was selected to gather meaningful indoor/outdoor air pressure data (Figure 3). The
approach requires that two outdoor air pressure measurements be made. A differential
outdoor air pressure reading is obtained between the upwind-side, and the downwind-
side of the dwelling. Second, a differential air pressure reading is also simultaneously
obtained between the interior and exterior of the dwelling. The second reading is taken
only when the differential air pressure in the first observation indicates very close to
-------
FEB 07 '91 11 = 43 USEF'h SUPERFUND OFC REG 3
zero. This approach removes the effect of wind on the differential pressure value
observed between the inside and outside wall of the dwelling.
Two Magnahelic gauges were used to measure differential air pressure. One
Magnahelic having a zero center, i.e., -.25 to 0 to .25 inches water column (0 to 62
Pascals) was used to measure the air pressure difference caused by winds blowing
across the exterior of the dwelling. The second gauge, with .25 inches water column
full scale, was used to measure the interior building pressure relative to the air pressure
outside the dwelling. Air diffusers normally used in a domestic aquarium were attached
to the tip of the outside probes, to prevent wind velocity (pitot) effects on the exposed
end of the pressure tube. With the diffusers connected to the tubing pitot effects
disappeared.
A tee fitting was inserted in one of the lines and connected to the pressure port
of the 0 to .25 inch pressure Magnahelic. The vacuum port was left open to the
interior building air pressure so a vacuum in the building would give a positive
deflection on the meter. The Magnahelics were then read twice each day, at
approximately 7:00am in the morning and again at 7:00pm in the evening.
A meteorological station was set up at an adjacent dwelling, approximately 200
feet from the test house. It continuously measured wind speed and direction, outside
temperature, humidity, and barometric pressure.
-------
FEE: 07 '91 11 = 44 USEPA SUPERFUHD OFC REG 3
systematic relationship with these radon perks. It was excluded from subsequent
analysis because the relationship with the other parameters was quite striking.
Visual examination of the graphed data presented in Figure 4, Event 1, provides
some valuable insight into various actions that occurred during the "Event," First, it
shows that it was preceded by a rise in relative humidity, which also represents
concurrent rain period. Second, barometric pressure readings, which were multiplied by
a constant so that they could also be plotted on the same graph, show an inverse
relationship with respect to the dwelling's basement radon levels and relative humidity.
The latter observation is well-known, but may not imply cause-and-effect. Last, the
static values representing the dwelling's differential air pressure (vacuum) decreased,
i.e., became less negative while the basement radon level surged upward, and then
increased again as the event proceeded until basement radon levels returned to their
normal values.
The fact that the differential pressure in the dwelling decreased and then started
to recover increase in negativity during the course of the event, can be explained and
attributed to the buoyancy of the dwelling's indoor air, its temperature and and relative
humidity which remained almost constant throughout the event. However, the
differential density between the inside air decreased relative to the outside air, because
the outside air became less dense as its relative humidity increased. As the event
progressed the relative humidity of the outside air declined, its density increased,
causing the buoyancy of the dwelling air, hence the dwelling's vacuum, to increase
-------
FEE: 07 '91 11:44 USEF'ft SUPERFUND OFC REG 3
relative to the ambient air outside the dwelling. This decrease in the vacuum appears
to be related in time to the measured outdoor humidity rather than the radon
concentration.
This last observation--a declining vacuum in the dwelling- was unexpected, since
it clearly conflicts with conventional wisdom. The radon spike, was presumed to be
driven by the vacuum in the house. This is apparently not the case. The reduced
differential air pressure (vacuum) in the dwelling should have produced a lower radon
level in the basement. Instead an opposite outcome occurred, clearly indicating the
some other overriding factor was causing the radon concentration in the dwelling to
increase.
It is important to note and understand that throughout the study period there were
numerous instances when the barometric pressure was low and it did not rain.
However, these occurrences were never accompanied with radon spikes. Radon spikes
were observed only during and after a rainfall event when the barometric pressure was
low and the surface soil was not frozen. Furthermore, these spikes appeared to follow
the rain event when the event was significant, i.e., precipitation was greater than 0.3
inches. Occasionally, the barometric pressure appeared to follow the radon spike. In
light of these observations it is reasonable to conclude that rainfall, and not barometric
pressure is the cause of the radon spikes. In addition, the height of the radon spike
correlates with the amount of rainfall; 8 relationship that is presented in Figure 8.
-------
FEB 07 '91 11:45 USEF'hi SUPERFUND OFC REG 3 P.
Further examination of the three remaining events yielded similar findings. In
light of these observations, it now appears that the rain percolating into the subsoil
plays a far greater role in the dynamics of radon gas transport from the surrounding
soil and into a dwelling that was previously believed. As surface water percolates
vertically downward through the soil, it displaces the gas in the soil pores. This soil gas
is compressed and forced laterally into the unsaturated soil pores immediately below the
foot print of the dwelling since these soil pores are protected from the rain and
associated percolate because of the dwelling's cover. This process, in effect, produces a
gas surge of radon-222 spike that is pushed, rather than pulled into the dwelling during
the rainfall event. However, additional studies should be done to corroborate this
theory.
-------
FEE 0, 'W1 11:45 USEPh SUPEPFIimd nrr REu
p. 10
Reference
DOA (Department of Agriculture). May 1963. Soil Survey Series 1959, No. 19.
Government Printing Office, Washington, DC.,
pp78*79.
EPA (Environmental Protection Agency). September 1987. Radon Reference Manual,
EPA/520/1 •87-20, Government Printing OFfice, Washington, DC. pp2-l to 3-8.
Montague, A., W.E. Belanger, and F.J. Haughey, 1990. "An Analysis of the Parameters
Influencing Radon Variations in a House.: In 1990 International Symposium orj
Measurements of Toxic and Related Air Pollutants. U.S. Environmental Protection
Agency/Air & Waste Management Association, Raleigh, North Carolina.
Montague, A., 1990b. An Evaluation of the Quantitative Effects on Radon Gas frnm
the Modification of a Home Heating and Air Conditioning Systems. PhD Thesis,
Rutgers University, New Brunswick, NJ.
Nazaroff, W.W. and S.M. Doyle, Radon Entry Into Houses Having a Crawl Space,
Lawrence Berkeley laboratory, Report LBL-16637, Berkelely, CA, 1983.
-------
Chimneys
Figure i. Sources of Negative Pressure
-------
BKFORE
Figure 2. Furnace Modification
-------
u
QJ
-Q
-O
Z>
CO
-------
-------
-------
-------
-------
-------
Session VI:
Radon Surveys
-------
vi-l
FACTORS ASSOCIATED WITH HOME RADON CONCENTRATIONS IN ILLINOIS
by: Thomas J. Bicrma, MBA, Ph.D. and Jennifer O'Neill
Environmental Health Program
Department of Health Sciences
Illinois State University
Normal, Illinois 61761
ABSTRACT
The Illinois Department of Nuclear Safety has performed short term
alpha-track radon testing in over 3000 homes throughout Illinois. Data were
also collected on a wide array of household characteristics and test condi-
tions. Analysis of these data revealed a number of interesting patterns.
Contrary to many investigations, the highest average concentrations were
not obtained in homes monitored in winter. Though a number of models were
explored in the analysis, few could explain more than 10% of the variation in
radon concentrations and no model could explain more than 20%. This suggests
that other factors, such as local geology, may be largely responsible for
inter-house variations.
Among houses with a crawlspace elevated radon was associated with having
an indoor entrance and not ventilating or insulating the crawlspacc. In
houses with basements, the foundation construction materials were important
explanatory variables. Surprisingly, common entry routes, such as sump pits
cracks, drains, and exposed earth, were not associated with radon concentra-
tion. Also contrary to other studies, energy efficiency was positively asso-
ciated with radon concentration.
Results suggest that factors governing radon concentrations in the
Midwest arc poorly understood.
-------
INTRODUCTION
The primary factors governing the entrance of radon into the home arc
well known: 1) the soil radon gas concentration and soil permeability, 2) the
existence of entry routes between soil and home interior, and 3) a pressure
difference between home and soil to provide a driving force for radon entry.
However, specific conditions that influence these factors arc not well under-
stood, and other factors, such as the distribution of radon in the home and
the infiltration of outside air, can influence the concentration of radon in
various parts of a home.
Though a number of studies have been performed to examine the relation-
ships between local conditions and indoor radon concentrations, considerable
uncertainty remains. In addition, relatively little work has been conducted
on factors influencing home radon concentrations in the Midwest.
This study was performed using data collected on approximately 3,000
homes in Illinois that were tested for radon as a part of the Illinois Depart-
ment of Nuclear Safety's Radon Program. In addition to radon testing, data
were collected on a number of household and monitoring conditions. The pur-
pose of this analysis is to build a predictive model for indoor radon concen-
trations in Illinois. Such a model could not only help identify homes with a
higher risk of elevated radon, but could also assist in directing mitigation
efforts.
A PRIORI MODEL SPECIFICATION
Data on a number of household and monitoring variables were collected
during monitoring (sec the following section for details). An a priori model
was developed to guide analysis of these variables and their relationships to
indoor radon. The model is also useful in identifying gaps in the array of
variables included on the questionnaire.
Tabic 1 presents the household and monitoring variables evaluated in
this analysis. Indicated in the table are expected relationships between each
variable and monitored radon concentration. These expectations arc based upon
relationships reported by others in the literature and on the authors' own
field experience. Table 1 lists the basic references used in creating the
table. A detailed discussion of the current literature and a priori model is
not presented here due to space limitations. The reader is referred to re-
views such as references (1-3). However, a few comments should be made to
clarify subsequent modeling.
Air infiltration may be associated with an increase or decrease home
radon concentrations. Increases may occur if infiltration occurs due to house
dcprcssurization, which can increase infiltration of soil gas. Similarly, the
mixing of air within a house can increase or decrease monitored radon depend-
ing upon the location of the monitor. For example, basement radon levels may
decrease and first floor radon levels may increase as basement air is distrib-
-------
utcd through a forccd-air heating system. Therefore, variables related to air
infiltration and air distribution may demonstrate either a positive or nega-
tive relationship to radon in modeling results.
No information on local geology, weather conditions, or other radon
sources (other than basement wall construction material and exterior brick)
was collected. Thus, the contribution of these factors to radon concentra-
tions cannot be assessed with this data.
METHODS
In the Fall of 1986 the Illinois Department of Nuclear Safety began a
systematic home radon testing program in Illinois. Testing was performed on a
county-by-county basis using local assistance from county health departments,
Cooperative Extension Service personnel, or other organizations. The program
continued through the Spring of 1990, though no testing was performed from the
Summer of 1988 to the Fall of 1989.
An alpha-track detector was typically placed in the lowest livable area
of the home. After approximately one month of exposure detectors were re-
turned to the contract laboratory (Tech/Ops Landaur, Inc.) for evaluation.
In addition to placement of detectors, surveyors completed a question-
naire with the assistance of a hcad-of-household. The questionnaire included
a number of questions on house characteristics and on the household occupants.
Though houses were not selected randomly, selection procedures, typical-
lv involving identification of houses from county highway maps, were intended
to avoid any systematic bias. Due to the complex interaction of factors
influencing radon concentrations, such selection processes may be relatively
free of bias (1). At least 20 houses were tested in each county. The number
of houses tested in a county increased with increasing population. The 3,021
monitored houses analyzed in this study were drawn from 73 of 102 Illinois
counties. These 73 counties contain approximately 88% of the Illinois popula-
tion.
All data were evaluated for distributional characteristics and coding
errors or ambiguities. For many quantitative variables demonstrating non-
normal distributions, transformations were performed to enhance normality. In
most cases, natural logarithm transformations were sufficient. For other
non-normal quantitative variables, however, no useful distributions could be
derived from transformation. Some variables demonstrated bi- or tri-modal
distributions. Such variables were typically transformed to artificial cate-
gorical variables using cither theoretically or empirically based cut-points.
All final variables and their coding schemes are presented in Table 1. For
quantitative variables, scatter plots were assessed for evidence of hetcroski-
dasticity and non-linearity. No problems were identified.
Multiple linear regression was used to explore models of monitored radon
concentration. A scries of theory and non-theory based models were explored.
These arc explained in more detail in the following section.
-------
RESULTS
OVERALL
Radon concentrations were approximately log-normally distributed with a
geometric mean of 2.84 pCi/1 and a standard geometric deviation of 2.30. All
subsequent modeling and statistical analyses used the natural log transform of
radon concentration.
It was anticipated, prior to analysis, that substructure type, monitor
location, and season of monitoring would be the primary explanatory variables
across all types of homes, but that the effects of monitor location and season
of monitoring could vary by substructure type. Table 2 presents the ANOVA
results for radon concentration across the four basic substructure types
encountered. Results confirm that substructure is a critical explanatory
variable, with crawlspace/no basement producing the lowest concentrations and
basement/crawlspace combination producing the highest concentration.
Some of the difference in means, however, may be due to monitor location
rather than a direct effect of substructure type. A small percentage of
houses with basements were monitored on the first floor, presumably because
the basements were not considered "livable". Table 2 presents ANOVA results
using first floor measurements only. Though significant differences are still
noted, slab-on-grade, rather than basement/crawlspace combination represents
the substructure associated with the highest average monitored concentrations
No analysis was conducted on the joint effect of monitor location and
substructure type since interaction is an artifact of the survey method: a
monitor was placed in the basement only if the home had a basement. The joint
effect of substructure and season are presented in Table 3. Both independent
variables are significantly associated with the dependent variable, and there
appears to be an interaction effect. In basement homes without a crawlspace,
fall monitoring produced the highest concentrations. For slab-on-grade homes,
fall monitoring produced the lowest concentrations.
Due to the apparent interaction between substructure type and monitoring
location and season, as well as probable interactions with other explanatory
variables, subsequent analyses were conducted separately for each of the four
substructure types.
Before leaving Table 3, however, it is important to note that, contrary
to many research findings, winter was not the season of highest radon concen-
trations . Summer produced the highest or second highest concentrations in
each substructure type.
A number of regression models were explored for each substructure type.
These are presented in Tables 4 through 7. Model A in all tables represents
the results from a series of simple linear regressions using the variable
listed as the only independent variable. Interpretation of statistical sig-
nificance under multiple comparisons is an obvious problem. Consistency with
theory and the findings of other studies, as well as consistency across sub-
structure types, should be used in evaluating the results. This point will be
explored further in the discussion section.
-------
Models B through D reflect models built upon theory and the results of
other models. In Model B, only the monitor location variables were used. In
Model C included the addition of monitoring location. Model D represents the
best attempt at a complete, theory-based model. Beginning with a base of
Model C, additional variables were added for which a good theoretic foundation
exists, and which demonstrated results consistent with theory in Models A and
E (explained below). Model D should be considered the best model for predic-
tion in Illinois houses.
Models E through C represent non-theoretical modeling. Model E includes
all variables simultaneously in the model. Because many of the independent
variables are correlated, multicolincarity is a significant problem in Model
E. However, by comparing the regression coefficients and p-valucs to those
obtained in other models, one can use Model E to identify those variable that
are relatively sensitive or relatively insensitive to the inclusion of other
variables in the model. Model F uses Model C plus a forward selection proce-
dure, and Model G used a stepwise selection procedure without any forced
variables. Variables included through such non-theoretic methods should only
be considered curious possibilities as true explanatory factors for indoor
radon. Considerable support from other studies would be needed to enhance the
reliability of such models.
The most immediate observation from the regression analyses is the
relatively low explanatory power of the models. The percentage of variation
in radon explained by the models was generally under 10%. Only for homes with
both a basement and crawl space docs theoretical modeling achieve an r-square
greater than 0.2. (Note: although Model E can produce a greater R square,
non-theoretic models arc likely to have far less predictive power than explan-
atory power.)
Specific variables of interest include monitoring location and season.
For homes with basements, knowing whether the monitor had been located in the
basement or not was an important predictor of monitored radon in all models.
Knowing whether the monitor was located in a first-floor bedroom or elsewhere
on the first was not, generally, a good explanatory variable. This changed in
the non-theoretic models for homes with basement/ no crawlspacc, possibly due
to the addition of the energy efficiency or central air conditioning varia-
bles. For homes without basements, bedroom location was important only for
homes with a crawlspacc, and then, only when monitoring season was included in
the model.
Monitoring season was important in nearly all models though the effects
were not consistent across substructure types (supporting the previous ANOVA
results). Only for slab-on-grade homes, where sample size was small, did the
season variables generally have p-valucs greater than 0.1 and not produce a
significant r-square change from Model B to Model C.
BASEMENT/CRAWLSPACE SUBSTRUCTURE
For homes with a bascmcnt/crawlspacc combination, an entrance from the
interior of the home (most likely the basement) to the crawlspacc produced a
consistent increase in monitored radon. Because the dependent variable is
logarithmic, coefficients represent the multiplicative effects of an independ-
ent variable on the radon concentration. The coefficient of 0.26 in Model D
indicates that a crawlspacc entrance increases home radon about 30%. Similar-
-------
ly, venting a crawlspacc demonstrated a consistent decrease in radon, with
Model. D indicating approximately 20% higher radon levels for homes with un
vented crawlspaccs. Coefficients for crawlspacc insulation and the presence
of exposed earth in the crawlspacc were both in the correct direction but did
not demonstrate consistent statistical significance.
Basement foundation construction material also appeared to be important.
Construction materials, from lowest to highest in their apparent contribution
to home radon levels arc poured concrete, block, "other material", stone and
mortar, and brick. From Model D, brick is associated with an average increase
in radon levels of greater than 50% over poured concrete. (It is interesting
to note that brick reverses signs in Model E. This is apparently an artifact
due to multicollincarity.)
Energy efficiency demonstrated a consistent positive relationship with
radon. The greater the occupant-assessed energy efficiency, the greater the
radon. Perhaps for similar reasons, the number of people in the house was
negatively related to radon. Both variables may be related to the amount of
air exchange and the size of the house.
Among the most significant findings is that the standard entry routes in
a basement (cracks, sump pit, exposed earth, etc.) were not important explana
tory variables. Some even had coefficients with a sign opposite that predict-
ed by theory. Also contrary to theory, the presence of a woodstove or fire-
place had a negative coefficient and large p-value.
There arc a number of interesting curiosities in the regressions of
houses with bascmcnt/crawlspacc combinations. In Model A having room air
conditioning or an electric space heater were associated with lower radon
concentrations. These did not retain their significance in Model E and may
reflect the effects of energy efficiency. Another interesting point is the
importance of a brick exterior appearing only in the non- theoretic models.
Having an interior entrance to the basement was associated with higher radon
levels in Model A, though this association lessened dramatically in Model E.
A test for an interaction effect between basement monitor location and base-
ment entry was negative (results not presented here).
BASEMENT/NO CRAWLSPACE SUBSTRUCTURE
For basement homes without a crawlspacc, a similar pattern appears.
Basement construction materials, in the order of their apparent contribution
to radon, arc: poured concrete, block, brick, and stone and mortar. This
order, and magnitude of effect, arc roughly consistent with the findings of
bascmcnt/crawlspacc homes.
Standard basement entry routes were, again, not significant, though the
presence of cracks had a marginal p-value and correctly signed coefficient in
Model A. The presence of a woodstove or fireplace also was not an important
explanatory factor.
An interaction term between basement monitor location and B_FINISH
(using the basement as a bedroom or living area) was found to be significant
in a separate test. The inclusion of this term in Model D resulted in a
dramatic change in the effect of B_FINISH. Together, these variables would
appear to have a substantial effect on radon. For a home with a basement used
-------
as a bedroom or living area, basement levels averaged nearly 5 times higher
than in homes without finished basements.
Energy efficiency demonstrated the same relationship with radon as in
basement/crawl space homes.
Among the curiosities, the presence of central air conditioning or
central forced-air heat, which may be inversely related to energy efficiency
demonstrated significance in Model A. Brick exterior was negatively related
to radon, the opposite effect of that in bascment/crawlspacc homes.
CRAWLSPACE/NO BASEMENT SUBSTRUCTURE
In homes with a crawlspacc but no basement, having an entrance to the
craw]space from the home interior was an important predictor of radon m all
models Insulation and ventilation of crawlspacc were directionally consist-
ent with theory, though p-values were marginal.
A number of curious associations were found. Having an attic fan was
consistently an important explanatory variable, though in the opposite direc-
tion generally expected from theory. Other variable,,, ^uch ao ha\mg hot
liter heat gravity feed furnace, and "other" fuel type, had low or marginally
low p values in at least one model. Having a gravity feed furnace, for exam-
ple, produced an average 757. increase in home radon levels. Explanations for
such associations arc unclear.
SLAB-ON GRADE SUBSTRUCTURE
For slab-on-grade homes, age of house and energy efficiency (which may
hr related to age of house) were consistently important explanatory factors,
indicating that the older and less energy-efficient the home, the lower the
radon.
A curiosity was the very strong inverse relationship between central air
conditioning and radon. Having central air was associated with an average
decrease in radon of about 50%. Having a fireplace or voodstovc was associat-
ed with a decrease in radon, though p-valucs were relatively large.
s
Reeau-e of the relatively snail sa-plc size for slab-on-grade homo
caution should be used in assessing the i.portancc of varrables basod on p.
value alone.
DISCUSSION
* li.ifntinnq should be recognized when drawing conclusion"
A number of mcthods used to select homes for testing produce
from this stud>. rigorous, pscudo-random selection meth
greater likelihood of bras than^ore^rrgor ^ ^ conslstcncy ^
ods. Though reasonabl - p personnel and home occupants is likely to
data collection, the use of local per.ox
introduce some error.
conducted by independent local agencies as time
Because saraP^| between local factors (such as housing or geology)
permitted, acacia (such as weather) may have been introduced. In
and time-variant factors
a
-------
addition, lack of local geological data means that correlations between gcolo
gy and household conditions cannot be identified. Thus, some results may be
due to the confounding effects of geology and time-variant factors.
Finally, due to the high degree of correlation between variables, model
ing is prone to error. Model results should be considered suggestive, not
confirmatory evidence of the actual underlying relationships.
It appears that the vast majority of radon variation is due to factors
other than the common household factors considered in this analysis. Such
factors may include differences in regional or local geology, housing con-
struction types, weather conditions at the time of monitoring, or household
factors not considered (such as the existence of thermal bypasses or vented
appliances). This docs not indicate, however, that the factors considered in
this survey are not important determinants of radon (for example having a
brick foundation may increase radon readings an average of 50%) but only that
other factors appear to be more important in explaining the variation between
houses. Household factors may not be consistent in their effects. This
supports policy recommendations that all homes be tested for radon, despite
the apparent presence or absence of known risk factors.
Location of the monitor in the basement produced, as expected, a signif
icant increase in radon concentration. Location of the monitor in the bed
room, as opposed to elsewhere on the first floor, was generally unimportant.
For research primarily intended to evaluate determinants of home radon conccn
trations, first floor monitoring in all homes is desirable to allow direct
comparisons across all house types.
Season of monitoring was an important explanatory factor. However, the
relationship between season and home radon concentration was not consistent
with other investigations nor was it consistent across housing substructures.
The finding that summer radon concentrations could be as high or higher than
other seasons suggests that current guidelines for winter monitoring be re-
evaluated. However, since these data reflect the measurement of different
homes in different seasons, they should be interpreted with caution. Addition-
al research on seasonal effects in the Midwest and elsewhere in the country
are needed.
Basement foundation construction material was a relatively consistent
predictor of radon. Brick and stone foundations were consistently higher than
poured concrete, even after adjustment for age of house. Block foundations
were slightly elevated compared to poured concrete. This suggests opportuni-
ties for low cost mitigative strategies if foundation walls arc accessible and
if a suitable radon barrier can be found.
An entrance to a crawlspacc was consistently associated with increased
radon. Insulation, exposed earth, and lack of ventilation in the crawlspacc
were also associated with increased radon, though less consistently. These
findings suggest that the common wcathcrization practices of scaling crawl-
spacc entries, insulating floors, and installing vapor barriers not only save
energy, but may reduce radon (through vapor barriers may need to be sealed and
vented). Limiting crawlspacc ventilation as an energy-saving measure, howev-
er, docs not appear advisable.
The work described in this paper was not funded by the U.S. Environmen-
-------
, o the contents do not necessarily reflect
and_thcrcforc^th^,^orscmcnt shouid be interred.
Protection agency ana ^ - -
views of the Agency and no official endorsement
REFERENCES
, v nroducts in indoor air: an overview. In:
Nero, A. P^don and its decay P ^ ^ Dccay Products in
Nazaroff and A. V IScro !1= ^
Indoor Air. John Wiley and Son,, N
.u oriein of radon indoors- building a pre-
scxtro, r. "»'•
dictivc capability. ali y
U ^ n R Application of radon reduction methods.
Moscly, R-, and HcnSChf ' ^nnmcntal Protection Agency. Research Trian-
EPA/625/5-88/024, U.S. Environment
ale Park, NC, 1988.
R Soil an a source of indoor radon:
Nazoroff. "ocd- B" and,SC"t°y In: W.W. Nazar°££ and A.V. Nero
generation, "'f"10": J"pr0(jucts inTndoor Air. John Wiley and Sons.
(rd ) Radon and its ueca>
Ncv York, 1988. p. 57 .
j 0 Moed, B.. and Scxtro. R. Character
Turk, B., Prill, R.. Crim^^;cS" and variability of radon in Pacific
Izing rTto « I"*' tot- AO: 498, 1990.
Northwest hornet.
• p The impacts of balanced and exhaust mechanical
Fisk, W. and Mowris. R. ^e f ^ Scifcrt, H.Esdorn, M. Fischer, H.
ventilation on indoor rado ( j titutc for Water, Soil and Air
Rudcn, J.Wagner (ed > Indoor Air
Hygiene, Berlin, 1987. p. 316.
• vn *3 Variation of radon levels in U.S. homes
Cohen, B., and CromlCkT '.lir p0il. Cont. Assoc. 38:129.1988.
with various factors.
R Correlation of lung dose with Rn conccntra-
Bicrma, T., and Toohcy K. ccntrati.on and daughter surface deposi-
tion, potential alVha_ g Rcalth Physics, 57:429, 1989.
tioti: a Monte Carlo analysis
rn i/ Rcvzan K and Nero, A. Invc^tlgci-
Scxtro, R., Mood, B. , Nazaro ' radon. ' in J ' P. Hopke (ed.). Radon and
tions of soil as a SOQccurrcncc, Properties, and Health Effects. Amen-
Its Decay Products 1987. P- 10.
can Chemical Society, N.Y., W
10.
11.
can ujiejiiJ- j .
Borak, T., Woodruff, B., and Toohcy, R. A survey of winter, summer and
annual average Rn 222 concentrations in family dwellings. Health Phys-
ics, 57:465, 1989.
Arnold, L. A scale model study of the effects of meteorological, soil,
and house parameters on soil gas pressures. Health Physics, 58:559.
1 rtnr\
-------
TABLE 1. VARIABLE NAMES, EXPLANATIONS, CODING, AND THEORETICAL RELATIONSHIP
TO INDOOR RADON CONCENTRATION
Variable name
Cnhpt-rnrf nrpd
U U WO Wi U V/ «. Ui. C
BMTCRL
BMTNOC
CRLNOB
SLAB
Season
WINTER
SPRING
SUMMER
FALL
Monitor
LQCBNT
LOCBDR
OTHER
N_PEOPLE
LNAGEHS
N_SM0KER
ADJJLAB
PC_BRICK
ENER EPF
Description
Coding3
Relationship to
indoor radon"|C
CENT AC
ROOM AC
FAN ATIC
Basement and crawlspace
Basement but no crawlspace
Crawlspace bat no basement
Slab-on-grade only
Monitoring done during the winter
Monitoring done during the spring
Monitoring done during the summer
Monitoring done during the fall
Monitor located in the basement
Monitor located in the bedroom
Monitor located somewhere else on
first floor
How nany people will be living in the
house for the next month?
What is the age of the house?
Is there at least one smoker in the house?
Does the house have any attached asphalt or
concrete slabs(attached garage, carport slab,
patio, driveway, etc.)?
What percent of the outside of the house is
covered with brick?
Sua of occupant ratings on five energy
efficiency characteristic. Each
characteristic rated on scale of 1 to 5.
Do you use central air-conditioning during
warm weather?
Do you use a rooi air_conditioner during
warm weather?
Do you use a whole house and/or attic fan
during warm weather?
count
natural log
of years
OH. >01=1
A scale of
5=least to
25=aost
energy
efficient
¦H
+ /"
v-
+/•
-------
TABLE 1. VARIABLE NAMES, EXPLANATIONS, CODING, AND THEORETICAL RELATIONSHIP
TO INDOOR RADON CONCENTRATION (CONTINUEDj
Variable name
Hpopr ¦! nf i nn
utui ik/vivn
Coding3
Relationship to
indoor rador,"'c
CENT FA
ft BASE
CENT
ELEC SPA
Do you use a forced air central heating system
during cold weather?
Dc you use a hot water baseboard or radiator
system during cold weather?
Do you use a gravity flow central heating system
during cold weather?
Do you use an electric space heater during
cold weather?
+/-
A /.
' I
X /.
• /
WOOD Do you use a fireplace or wood burning stove
during cold weather?
KEROHEAT Do you use a kerosene space heater during cold
weather?
GAS STOV Do you use a gas stove during cold weather?
HEAT OTHER Do you use some other kind of heating system
during cold weather that was not mentioned on
this questionnaire? If so, specify.
B_FINISH Is all or a portion of the basement frequently
used as a bedroom or living area?
Foundation The outside basement walls are primarily composed
3L0CK Concrete or cinder block
CRET Poured concrete
MORT Stone and mortar
BRK Brick
OTHER Other and tile
ENT BASE Can you enter the basement from inside
the house?
X /.
•I
+
+
7
+ /*
CRACKS
Does the basement floor(or sub-surface
floor in a split-level hose) have large
cracks or holes?
DRAINS
SUMP
Does the basement floorfor sub_surface
floor in a split-level home) have drains?
Does the basement floor(cr sub-surface
floor in a split-level home) have sump
pumps?
-------
TABLE 1. VARIABLE NAMES, EXPLANATIONS, CODING, AND THEORETICAL RELATIONSHIP
TO INDOOR RADON CONCENTRATION (CONTINUED)
Variable name
Description Coding3
Relationship tc
indoor radon",c
EARTH
Does the basement floorjor sub-surface
floor in a split-level home) have exposed
pavf
Cu. vii .
+
CSP ENTR
Car, you enter the crawl space frou inside the
hoase[from basement, for example)?
CSP_INSL
Is the floor above the crawl space insulated?
-
CSP_VENT
Will the crawl space be vented to the outside
during the monitoring period?
-
CSP_EXP
Does the crawl space have exposed earth?
+
a. All Yes/No variables codes as 0-no, 1-yes
b. ++ - strongly and positively related
+ - positively related
- - negatively related
-- = strongly and negatively related
+/- - both positive and negative relationships nay be expected
? - too little information to predict a relationship
c. Based upon references (1] through (11)
d. These variables were used to define datasets and do not appear explicitly in later tables.
-------
TABLE 2: ANALYSIS OF VARIANCE RESULTS FOR RADON CONCENTRATION BY SUBSTRUCTURE TYPE.
Monitor
Geometric mean
Geosetric
location
Substructure type
concentration [pCi/1)
Standard Dev.
N
All(a)
All types
2.84
2.30
3,021
Baseaent and crawlspace
3.56
2.21
752
Crawlspace, no basement
1.67
2.11
572
Baseisent, no crawlspace
3.16
2.21
1,548
Slab-on-grade only
2.32
2.33
134
All types
1 70
i. i > j
2,14
asi
WW
Basement and crawlspace
1.95
1.88
78
Crawlspace, no basement
1.67
2.11
572
Basement, no crawlspace
1.81
2,07
69
Slab-on-grade only
2.32
2.33
134
(a) p-value - <.001 for at least two means differing by substructure type.
(b) p-value - <.001 for at least two means differing by substructure type.
TABLE 3: GEOMETRIC MEAN CONCENTRATION IN pCi/I (AND SAMPLE SIZE)
BY SEASON OF MONITORING AND SUBSTRUCTURE TYPE3
BASEMENT BASEMENT NO BASEMENT NO BASEMENT
CRAWLSPACE NO CRAWLSPACE CRAWLSPACE NO CRAWLSPACE
SEASON
WINTER 3.74 (509) 3.22 (113?) 1.68 (354) 2.27 (97)
SPRING 2.34 (122] 2.51 (218) 1.39 (103) 2.83 (17)
SUMMER 4.76 (65) 3.74 (118) 1,67 (45) 2.34 (14)
FALL 3.94 (56) 3.78 ( 75) 2.03 (70) 1.70 (6)
a. p-values = <.001 for both main effects and interaction
-------
TABLE 4: MULTIPLE REGRESSION RESULTS FOR MODELS OF HOME RADON CONCENTRATION:
HOMES BUILT OVER BASEMENT AND CRAKLSPACE
VARIABLE MODEL A MODEL B MODEL C MODEL D MODEL E MODEL F MODEL G
INTERCEPT
-
.62447
.90441
.63720
.13636
-.06271
.42381
-
'' (I11
\ N,Uii
(<•01)
(<•01)
(<•01)
(.79)
(.20)
WINTER
-.13391
-.11985
-.13353
-.19473
-.17511
(.34)
(.38)
(.34)
(.14)
(.19)
SPRING
-.60418
-.54158
-.62871
-.66264
-.48379
-.64939
[<•01)
(<•01)
(<•01)
(<-oi)
{<•01)
(<.01)
SUMMER
.15321
.10257
.05199
-.03172
-.08679
t it\
i"41
(.82)
(.91)
(.94)
(.84)
LOCBKT
.66483
.66483
.57401
.55822
.58134
.63876
.57936
(<-oi)
(<•01)
(<¦01)
(<•01)
(<.01)
(<-oi)
(<.01)
LOCBDR
-.02230
-.02230
-.08867
-.22949
-.11436
-.11565
(.91)
(.91)
(.65)
(.24)
(.55)
(.54)
N_PEOPLE
-.05905
-.05277
-.04495
(.03)
(.05)
(.07)
ADJ_SLAB
.10375
.08877
(.17)
(.26)
NJMOKER
.09009
.10630
(.21)
(.13)
LNAGEHS
-.04625
-.05477
-.06157
(.18)
(•20)
(.13)
EHERJFF
.02528
.01958
.02655
.02516
.02513
(.01)
(.04)
(<.01)
(<.01)
(<.01)
PCJRICK
-.07020
.50320
.34598
(.45)
(<.01)
(<.01)
CENTAC
.09944
.08401
(.17)
(-33)
ROOHAC
-.14919
-.06130
(.06)
(.49)
FAN_ATIC
.05429
.10389
(.60)
(.30)
CENT_FA
.05843
-.03392
(.52)
(.81)
WBASE
-.05531
.01585
(.62)
(.92)
CENT_GF
-.30802
-.27069
(.19)
(.26)
ELEC_SPA
-.22500
-.02960
(.06)
(.80)
HOOD
-.03622
-.10073
(.63)
(.16)
KEROHEAT
-.09649
.00570
(.47)
(.96)
GAS_STOV
-.31293
-.24238
(.14)
(.24)
HEAT_OTHER
.12606
.31080
(.35)
(.03)
-------
TABLE 4: MULTIPLE REGRESSION RESULTS FOR MODELS OF HOKE RADON CONCENTRATION:
HOMES BUILT OVER BASEMENT AND CRAHLSPACE (CONTINUE)
VARIABLE MODEL A MODEL B MODEL C MODEL D MODEL E MODEL F MODEL G
BLOCK
.09408
.10501
.13430
(.27)
(.23)
(.10)
MOST
.13457
.41139
.39434
(.32)
(<•01)
(<•01)
BRK
.33532
-.13263
.34835
.50465
(<.01)
(.15)
(<•01)
(<•01)
OTHERBMT
.18767
.32299
.34548
(.21)
(.03)
(•02)
BJINISH
.01927
.02686
(.81)
(•75)
ENTJASE
.26429
.05568
(.02)
(.61)
CRACKS
-.05275
-.03644
(.47)
(.61)
DRAINS
-.04133
-.11459
(.66)
(.20)
SUMP
-.03177
-.00855
(.66)
(.26)
EARTH
.06191
.08333
(.45)
(.32)
CSPJNTR
.30078
.21091
.23927
.22295
.25269
(<.01)
(.03)
(.01)
(.01)
(<•01)
CSP_IHSL
-.03825
-.06228
(.62)
(.40)
CSPJENT
-.13419
-.20053
-.18782
-.18901
-.19699
(.07)
(<•01)
(.01)
(.01)
(.01)
CSP_EXP
.06678
.09223
.19421
.20200
.14370
(.47)
(.34)
(.03)
(.02)
(.13)
Multiple R2 .07481 .12360 .25315 .19353 .18892 .21718
(p-value) (<.01) (<.01) {<.01] (<.01} (<.01) (c.oi)
R2-Ctiange .04879 .12955
(p-value) (<.01) (<.01)
-------
TABLE 5: MULTIPLE REGRESSION RESULTS FOR MODELS OF HOME RADON CONCENTRATION:
HOMES BUILT OVER BASEMENT BUT NO CRAMLSPACE
VARIABLE MODEL A MODEL B MODEL C MODEL D MODEL E MODEL F MODEL G
INTERCEPT
-
.68446
.96252
.73018
.84825
.50784
.78389
-
(<•01)
(<.01)
(.02)
(<.01)
(.02)
(<.01)
WINTER
-.18722
-.16864
-.20245
-.17015
-.17590
1 ni
I "-J i
(.17)
(.10)
(.16)
(.M)
SPRING
-.46281
-.41092
-.42518
-.40259
-.23534
-.41147
// mi
i -'"i
(<.01)
(<•01)
(<•01)
Ml)
(<.01)
SUMMER
-.09921
-.09921
-.11711
-.07166
-.09674
(.76)
(.76)
(.71)
(.82)
(.76)
LOCBMT
.47646
.47646
.39752
.40905
.39413
.57138
.39319
1/ mi
p.»'i
(<¦01)
(•01)
(<.01)
(.01)
MU
(<.01)
LOCBDR
-.29852
-.29852
-.33525
-.47366
-.43520
-.47317
-.36979
(.19)
(.19)
(.14)
(.04)
(.05)
(.04)
i in
i i
H_PEOPLE
.02780
t
.03166
(¦08)
ADJJLAB
.07140
(.16)
.11599
(.04)
N_SMOKER
.02038
(.67)
-. 00685
(.89)
LNAGEHS
-.02615
(.29)
-.08970
(<.01)
-.08629
Ml)
-.09404
Ml)
ENERJFF
.01519
.02134
.01692
.01666
.01699
(.02)
(<.01)
(.01)
(.02)
(¦01)
PC_BRICK
-.11062
-.07839
(.04)
(.15)
CENT_AC
-.11208
-.12129
-.12048
-.12346
(.02)
(.05)
(.02)
(.02)
ROOM_AC
.04593
-.08926
(.43)
(.21)
FANJTIC
-.11001
-.09317
(.13)
(.20)
CENT_FA
-.11838
-.13650
-.11611
(.05)
(.24)
(.04)
W_BASE
.10301
-.03680
( 1 R \
(.77)
CENT_GF
.01615
-.14266
(.92)
(.47)
ELECJPA
-.02207
-.02826
(.30)
(.75)
WOOD
.02871
(.57)
.01815
(.72).
KEROHEAT
.11347
(.29)
.12184
(.25)
GASJTOV
-.01028
.02939
(.94)
(.83)
HEAT_OTHER
.06506
(.52)
-.03535
(.77)
BLOCK
.21535
.28573
.27904
.26975
.27563
f/ (111
Ml)
Ml)
Ml)
{<.01)
-------
TABLE 5: MULTIPLE REGRESSION RESULTS FOR MODELS OF HOKE RADON CONCENTRATION:
HOMES BUILT OVER BASEMENT BUT NO CRAWLSPACE (CONTINUED)
VARIABLE MODEL A MODEL B MODEL C MODEL D MODEL E MODEL F MODEL G
MORT
.35438
.46092
.48845
.48387
(<.01)
(<.oi)
(<¦01)
t / ni \
i
BRK
.26990
.38522
.40233
A 0290
.49618
(<.01)
(<.01)
(<.01)
(<•01)
(<.01)
OTHERBMT
-.04626
.07880
.39319
(.72)
I kr\
i.jjI
(<•01)
B_FINISH
-.06937
-.10759
-.11224
.08526
(.15)
(.03)
(.02)
i ti \
\ 1
LOCFIN (a)
.64436
i n \
I # a fc I
ENTJASE
.02046
-.01569
-.74116
(.85)
(.89)
(.08)
CRACKS
.08293
.05840
(.09)
(.25)
DRAINS
.04636
.06675
(.52)
(.35)
SUMP
-.02555
-.04301
(.59)
(-39)
EARTH
.09408
.13934
.15700
(.31)
(.15)
(.10)
Multiple R^ .02426 .03894 .10581 .08507 .08036 .09169
(p-value) (<.01) (<.01) (<.01) (<.01) (<.oi) (<.01)
R^-Change .01468 .06687
(p-value) (<•01) (<. 01)
-------
TABLE 6: MULTIPLE REGRESSION RESULTS FOR MODELS OF HOME RADON CONCENTRATION:
HOMES BUILT OVER CRAWLSPACE BUT NO BASEMENT
VARIABLE
MODEL A MODEL B
MODEL C
MODEL D
MODEL E
MODEL F
MODEL G
INTERCEPT
.43901
.63841
.28924
.60853
.50219
.74458
(<-oi)
(<•01)
(.42)
(<.0i)
(<.01)
i / nn
WINTER
-.16455
-.19250
-.23470
-.26874
-.29359
(.14)
(.09)
(.07)
(.02)
i nn
l"4!
SPRING
-.38879
-.42165
-.40197
-.44044
-.19195
-.50885
(<.01)
(<-oi)
(-01)
(<•01)
(.03)
(<-oi)
SUMMER
-.21478
-.27490
-.32321
-.41428
-.39123
(.24)
(.14)
(.11)
(.03)
(.04)
LOCBMT
LOCBDR
.11316 .11316
.14463
.15581
.17344
.15547
(•13) (.13)
(.05)
(.04)
(.02)
(.04)
N_PEOPLE
.00389
.01128
1 OQ\
I • 1
(.70)
ADJJLAB
.04341
.10461
/ KB \
I • ¦"J 1
(.20)
NJMOKER
.11543
.06467
(.11)
(•38)
LNAGEKS
.04566
.03790
(.24)
(.41)
ENERJFF
.00355
.01355
(.72)
(.21)
PC_BRICK
.02862
.03722
(.77)
(.71)
CENT_AC
-.02534
.06893
(.72)
(.51)
ROOM_AC
.01281
.02986
(.87)
(.77)
FAN_ATIC
-.29695
-.25446
-.24573
-.26775
(<.01)
(.02)
(.02)
(.01)
CENTJA
-.05006
-.17104
(.52)
(.22)
W_BASE
.30619
.22883
.36088
.32856
(.03)
(.21)
(.01)
(.02)
CENT_GF
-.57371
-.56661
(.07)
(.09)
PTPf CDS
uuuU u k n
-.16027
-.11494
(.18)
(.35)
WOOD
-.07852
-.06947
(.34)
(.40)
KEROHEAT
-.06352
-.00899
(.71)
(.96)
GAS_STOV
.19141
.07700
(.17)
(.64)
HEAT_OTHER
-.19467
-.24163
(.09)
(.12)
CSPJNTR
.22661
.28046
.26974
.21505
.28106
(<•01)
(<.01)
(<.01)
(.01)
(<.01)
-------
TABLE 6: MULTIPLE REGRESSION RESULTS FOR MODELS OF HOME RADON CONCENTRATION:
HOMES BUILT OVER CRAWLSPACE BUT NO BASEMENT (CONTINUED)
VARIABLE MODEL A MODEL B MODEL C MODEL D MODEL E MODEL F MODEL G
CSP INSL -.12095 -.11359 -.10899
(.12) (.16) (.14)
CSP VENT -.08019 -.13048 -.12354
(.26) (.08) (.09)
CSP EXP .03655 .01987
(.69) (.83)
Multiple"? " ".00492 702936 7U3Q6 .07685 . 0 5 571~ "06398"
(p-value) (.13) (.01) (<.01) (<. C1) (.00) (<.01)
R^-Change .02444 .08370
(p-value) (.01) (.01)
-------
TABLE 7: MULTIPLE REGRESSION RESULTS FOR MODELS OF HOME RADON CONCENTRATION:
HOMES BUILT OVER SLAB-ON-GRADE ONLY
VARIABLE MODEL A MODEL B MODEL C MODEL D MODEL E MODEL F MODEL G
INTERCEPT
-
.72577
.30640
-1.88722
.47208
1.07722
-1.47057
-
(<.01)
(.44)
(.08)
(.23)
(<.oi)
(.05)
WINTER
.44310
.43118
.55750
.49918
.40731
(.27)
(.28)
(.18)
(21)
(.30)
SPRING
.70936
.63447
.79930
.63153
.63846
(.11)
(.16)
(.09)
(.15)
(.15)
SUMMER
.45747
.38257
.59130
.56041
.36715
(.32)
(.41)
(.22)
(.23)
(.42)
LOCBDR
.24511
.24511
.21398
.11732
.19420
.19473
(.14)
(.14)
(.21)
(.52)
(.25)
(.24)
N_PEOPLE
-.06949
-.04225
(.31)
(.59)
N_SMOKER
.25388
.26591
(.13)
(.13)
LNAGEHS
.13552
.23415
.23172
(.14)
(•03)
(.02)
ENERJFF
.03208
.08430
.05745
(.17)
(<.01)
(.03)
PCJRICK
-.09390
-.23546
(.65)
(.28)
CENTAC
-.39525
-.54079
-.39441
-.39525
(.02)
(.01)
(.02)
(.02)
ROOMAC
.10428
-.41429
(.60)
(.13)
FANJTIC
.26593
.57891
(.43)
(.10)
CEHTFA
-.16207
.20092
(.39)
(.54)
»_BASE
.13815
.38537
(.63)
(.31)
CENT_GF
.11750
.32170
(.82)
(.58)
ELECJPA
-.03449
.13368
(.92)
(.71)
HOOD
-.21622
-.25058
(.24)
(.19)
KEROHEAT
.73459
.91013
(.24)
(.17)
GASSTOV
-.20941
.20125
(.64)
(.70)
HEAT_OTHER
.09399
.01911
(.72)
(.95)
Multiple
.03009
.03912
.24870
.08740
.05156
.10304
(p-value)
(.14)
(.37)
(.10)
(.08)
(.02)
(.07)
R^-Chanqe
.01904
.20958
(p-value)
(.55)
(.09)
-------
TITLE: Radon in Federal Buildings
AUTHOR: Michael Boyd, EPA - Office of Radiation Programs
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
SHCtirtn*en°t9. «"ui«d to* p££k,
SweMMnt?of radon in a rapres.ntativ. .ampling of the buildings
£h«t thev own. Several agencies chose to survey all of their
bSildi£« The results of these surveys have been submitted to the
Fnvtronmental Protection Agency (EPA) where they are being compiled
for a report to Congress. Not all agencies completed their studies
in time for the first report to Congress and some agencies
w ™*rtial reports because the scope of their studies
Baveral to complete th. U.S. Army's study,
r!?^h include as many as a million test results when it is
etiolated) This paper presents a summary of the results that have
been received to date and describes briefly the testing strategies
that were employed.
-------
RADON IN SWITZERLAND
by : H.Surbeck and H.Volkle
Federal Office of Public Health
c/o Physics Institute, University
Perolles
CH-1700 Fribourg, Switzerland
and
W.Zeller
Federal Office of Public Health
Bollwerk 27
CH-3001 Bern, Switzerland
ABSTRACT
Based on measurements in nearly 1600 homes, representing 0.15% of the
housing stock, we estimate that the Swiss live on the average in rooms
with a radon concentration of 80 Bq/m3 and that 5% of them are exposed to
concentrations exceeding 200 Bq/m3.
Radon research in Switzerland started nearly a decade ago and shows
that building materials and household water use present no serious radon
problems, the soil being the main radon source. The highest values are
found in homes on highly permeable building grounds (Karst terrains,
rockslides).
We discuss the results of the radon surveys and explain how we try to
get a representative exposure estimate from biased data. We also present
geological aspects of the radon situation in our country and outline the
policy for the new decade that will see surveys concentrated on the search
for hot spots.
Several mitigation techniques have been tested successfully but few
homeowners are interested to take remedial actions. There is no great
public concern on radon in Switzerland; radon is natural.
-------
INTRODUCTION
parlv as in 1908 Gockel (1) reported on radon ("Radiumemanation")
Ze -in Switzerland. He already knew that the radon concen-
tration"1 in the soil gas depends on various geological factors, meteorolo-
gical conditions like wind speed and on the soil moisture content !
well seven decades later one started to realize that exposure to radon
mav present a serious health problem and small scale radon surveys were
carried out in Switzerland in the early 1980s.
Harmed by high values (up to 5 kBq/m3 in living rooms) found in homes
a ritv in the Western Swiss Jura Mountains (2,3) a task force was set
in a city situation in Switzerland. This eventually has led to
^nationwide 5-year research program (RAPROS) that started in 1987.
rt took some time to correct the then widely accepted but unproven
to" like • "high radon concentrations are mainly due to building
!'• "aranitic bedrock shows a high uranium concentration and
"ereJore ^J in the Alps have high radon levels", "there can't be high
radon concentrations in homes on Jurassic limestone .
Building materials and domestic water use showed to be a negligible
Hon source in Switzerland (4,5), the main source being the soil.
£ hanrpd"*Ra have been found in various soils not of granitic origin and
Ennancea ,Bon na/ka drv weight) has been measured in a soil
to
need 226Ra have been rouna in vanuus
the highest activity (880 Bq/kg dry weight) has been measured in
covering Jurassic limestone.
We show the general radon situation in Switzerland and how we try
gain representative exposure estimates from biased data. Geological
aspects of the radon problem are discussed. Mitigation techniques tested
in Swiss homes are presented and the policy for the new decade is
outlined. This policy is characterized by a concentrated search for radon
hot-spots.
GENERAL RADON SITUATION
FREQUENCY DISTRIBUTION AND AVERAGE RADON EXPOSURE
The frequency distribution of the radon concentrations in about 5000
rooms, corresponding to nearly 1600 buildings, representing 0.15% of the
-------
Swiss residential housing stock, is shown in figure 1. The radon levels
have been determined by exposing passive (etched-track) detectors for at
least two months. In general two detectors are placed per building, one in
the basement and one in an inhabited room at or above ground floor. We ask
people to use the rooms as usual in order to get radon concentrations
under realistic conditions. About 80% of the measurements have been
carried out during the winter. Few homes are represented by both summer-
and winter-values. These measurements show that summer levels are on the
average 1/3 lower than the winter levels.
The raw data in figure 1 are not representative for the radon exposure
of the Swiss for the following reasons :
1) Mainly in early surveys single family homes are overrepresented.
2) Certain regions are overrepresented due to particular
research programs like the search for radon sources in
the Jura Mountains (6) or because of the initiative of local
authorities.
3) Most measurements have been carried out during the winter
and thus don't give the annual mean.
To correct for bias 1) we sort the room data into building classes like
single family homes, blocks of flats, farms and "others". For multistory
buildings different stories (up to the forth floor) form separate classes.
For every class the number of radon values falling into a concentration
interval (subclass) is then multiplied by the percentage of the population
living in the respective class (1980 census data). Summing up the weighted
subclass contents over all classes leads to the new frequency
distribution. This first weighting is carried out for every canton (State
of the Swiss Confederation) or in the case of small cantons for a group of
cantons.
To correct for bias 2) the numbers in the subclasses of each canton
are multiplied by the percentage of the Swiss population living in this
canton. This frequency distribution having an arithmetic mean of 80 Bq/m3
is more representative for the radon concentration to which the Swiss are
exposed in their homes than the arithmetic mean of 140 Bq/m3 from the raw
data in figure 1.
The mean of 80 Bq/m3 still lacks the correction for bias 3) and for the
fact that few people stay at home 24 hours a day. We estimate that these
two factors lower the above 80 Bq/m3 to an annual mean of about 70 Bq/m3.
The Swiss map in figure 2 shows the geographical terms used.
REGIONAL DISTRIBUTION
In figure 3 we show the regional distribution of the 1540 buildings
that have at least one inhabited room at or above ground floor measured.
-------
„ 99.1
99
- 9ft
" #•
u
e
3 70
O uninhabited room at or beiow
ground Moor
1449 v«lu*«
•9
0
0
9 O
a
s 1
<¦>:
99
49
39
29
19
ft
1
9.ft
9.1
0 irifvab1**^ roof* at
or above ground floor
9997 v*lu««
% 19 29 ft9 199 299 ft99 1999 2999 *900 19990
Rn-222 conctntration C Bq/ri3 J
Figure 1. Frequency dis*r 1 huT4on of radon neasurenent results
CITV OF C.
ME8TERN
SWISS JURA
MOUNTAINS
SMI 88
COL DU
MARCHAIRUZ
+ plateau
4
UPPER RHINE UALLEV
ALPS'
100 kn
Lakes
Figure 2. Swiss nap. The dashed lines roughly represent the axis of
the respective Biographic unit.
-------
©
©
0
O
©
.fraction of
buildinfi
..ith cone •
\ / > 200 »Q/A*
0
©
1M - ft»0
/
Number ©# buildunj*
M«*fUP«d p*r #4® k**
Figure 3. Rtgionil distribution of radon ntaiurtntnt results
-------
1 -s
Number of buildings
per 840
( in thousands 1
< 1
Figure 4. Fraction of buildings Measured
-------
We consider radon levels exceeding 200 Bq/m3 in this type of rooms as an
indicator for a possible radon problem. The fraction of homes with at
least one room exceeding this level is shown in this figure for each of
the regions. The division into regions is the one used for the 1:50000-
scale maps. Each rectangle measures 24 km times 35 km (840 km2).
There are at least two regions in Switzerland with clearly enhanced
radon concentrations : the Jura Mountains in the west and the Upper Rhine
Valley in the east. Geological aspects of the radon problem in these two
regions are discussed below. Enhanced levels are present in the south-
eastern part of Switzerland too. The Swiss Plateau where most of the Swiss
live is essentially free from radon problems.
The more than 1500 homes measured so far represent 0.15 % of the
residential housing stock in Switzerland. This may be sufficient to
calculate a Swiss average but as can be seen from figure 4 many regions
are not well represented. We don't really know what "well represented"
means. What percentage of homes has to be measured per 840 km2 unit until
one can declare it as "affected" or "safe" ? A hint comes from a recent
survey in the southern part of Switzerland (Ticino) that nearly doubled
the number of homes measured in this region. From a comparison of the
frequency distributions of the radon values before and after this survey
we conclude that a representative sample has to contain at least 1 % of
the residential buildings. Another hint comes from the now best covered
region ( 3.5 % of the 6000 homes are sampled) where we have been measuring
for more than 8 years. The frequency distribution changed slightly over
the years and now has become quite stable. We don't expect any surprise
from further measurements. We therefore recommend to sample 1 to 3 % of
the housing stock before any region can be declared as safe or affected.
GEOLOGICAL ASPECTS OF THE RADON SITUATION IN SWITZERLAND
There are mainly three factors that determine the radon risk of a
building ground : 1) 226Ra activity concentration in the soil, 2) Fraction
of the 222Rn produced that is available for transport (emanation) and 3)
Gas permeability of the soil.
We will show the range of values found in Switzerland for these three
factors and discuss geological aspects of two high risk sites.
RA-226 ACTIVITIES IN ROCKS AND SOILS
Uranium data for Swiss rocks, taken from a recent compilation by
Scharli (7) are shown in figure 5. The term "Uranium" used by Scharli is
somewhat misleading for the quantity measured has been the 2a2Rn daughter
concentration. He neglects any disequilibrium in the 23aU series down to
the 222Rn daughters. We therefore call his "Uranium" values "sRa taking a
conversion factor of 12.3 Bq/kg per ppm U. From this figure it is obvious
-------
n ieoe
m m w m ^
=t
IllUUIilOillllllit
bbbbVbbb
III III III
III u u
U u u
~- b b
ZZttttffiZ 00
UOOOCUZO
>Nhh(9XIi|h
w z o o
0 0
£ N
O Z K
U tf *
IGNEOUS
ROCKS
METAMORPHIC
« ieee
H
S 6®0
>
« i® =1
nmvu
m o 5 m m *
n\\i i
J C M M
SEDIMENTARY
ROCKS
i tl 3s
NUMBER OF VALUES
maximum
}
CONTAINS
mean I- cs x of
THE VALUES
SOILS-
Figure S. Ra-22G activities in Swiss rocks and soils
-------
that "granite" is not synonymous with "high activity".
Activities in Swiss soils are shown in the figures 5 and 6. The 226Ra
concentrations in figure 5 have been determined by high resolution gamma
spectrometry on dried soil samples. The data given in figure 6 are from in
situ gamma spectrometry measurements (8). Contrary to the laboratory
measurements the poor statistics for in situ measurements exclude a
precise determination of the 23su or the 23*Th concentrations. The 23SU
contribution to the 186 keV 226Ra line has thus to be calculated assuming
perfect equilibrium down the 238u series. This leads to an underestimation
of the 226Ra activity in soils with a 23°Th ( and thus 226Ra) excess. A
23°Th excess is present in Jura Mountains soils. When comparing activities
in figures 5 and 6 one has also to take into account that the laboratory
data are for dried samples whereas the in situ values refer to the
undisturbed wet soil.
The complex nature of the Swiss geology and the important impact that
Quaternary had on our country makes it very difficult to find any
correlation between the regional activity distribution in figure 6 and
geological or tectonic maps. Soils in many parts of Switzerland are not
derived from the underlying bedrock. The most striking example is found in
the Western Swiss Jura Mountains where 22SRa activities of up to 880 Bq/kg
dry weight are present in soils covering Jurassic or Cretaceous limestone
having only about 20 Bq/kg of 226Ra.
A peculiarity of these soils is that 23°Th and 226Ra are largely in
excess of 23BU (determined quantity is 23*U), the latter being present in
"normal" quantities (30-50 Bq/kg dry weight). There is still no
explanation for this widespread anomaly. The watch industry, being very
prominent in the Jura Mountains has used large quantities of radium-
activated luminous paint but we can hardly blame them for this
"contamination". The 226Ra and it's natural precursor 23°Th are nearly at
equilibrium even in soil samples taken close to a former radium processing
workshop. In samples of luminous paint from this workshop the 23°Th
activity is orders of magnitude lower than the 22*Ra activity.
A hint for the origin of the enhanced activities may come from the
regional distribution of the 226Ra activity and it's dependence on the
altitude. There is a general trend for higher activities towards the
southwest (the main wind direction). Enhanced ( > 100 Bq/kg dry weight)
22SRa activities are abundant at high altitudes (figure 7) and no Ra
anomaly has been found so far below about 900 m above sea level. This
altitude roughly corresponds to the upper ice margin of the Rhone glacier
during the latest glacial period. These two observation are consistent
with the idea by Pochon (9) of an aeolian origin of an important part of
the Jura Mountains' soils.
RADON EMANATION
The few emanation measurements on Swiss Plateau soils (mainly glacial
till) show that for these soils about 30% of the radon produced can escape
-------
LAKES
Figure 6. Ra-226 and Ac-228 in Swiss soil*.
Dtterninvd by in situ ganna ip*ctronetry
¥
JS
91
•n
*
3
71
L
¦0
91
X
\
a
a
id
N
N
I
if
ee
1008
800
B00
0 western slope
400 —
200 -
800 900 1000 1100 1200 1300 1400 1500
Altitude of sampling point C neters above sea level ]
Figure 7. Soil sartples Col du Narchairuz
-------
to the pore space and is thus available for transport. This fraction is
far higher (about 70%) for "high radium" soil samples from the Jura
Mountains. An example is shown in figure 8. The 226Ra values given in this
figure have been determined using the 186 keV gamma line. The 23Su
contribution to this line is calculated from the measured 23,1Th activity
assuming equilibrium between 23eU and 234Th and taking the "universal"
23SU/23°U activity ratio of 0.046.
GAS PERMEABILITY OF THE SOIL
Our soil gas sampling apparatus (figure 9 and (6)) allows for the
simultaneous measurement of the gas permeability. Values from 10_iam2 to
>10~1om2 i.e. variations of at least four orders of magnitude have been
found in Swiss soils. The highest values show badly consolidated
rocks1ides.
High radon concentrations in the soil gas only present a radon risk if
the permeability of the soil is sufficiently large to allow for an
efficient radon transport to the foundation of a house. Even "normal"
radon levels in the soil gas may lead to a considerable radon risk if the
gas permeability of the soil is very high.
Therefore the product of the radon concentration in the soil gas times
the soil's gas permeability may be a better measure of the radon risk than
the bare radon concentration. This "radon availability" is plotted in
figure 10 for several regions in Switzerland. The envelopes have been
generously drawn around the respective data sets. Individual data points
are not shown in this figure. In the regions TI and FR we could not find
homes with high indoor radon concentrations whereas the regions RA and SI
are characterized by high indoor levels. Measurements in the region SI
include gas samples taken in unconsolidated rocks. Despite the large
scatter of the data points there is some evidence that building grounds
with radon availabilities larger than lO'^Bq/m to 10~6Bq/m present a radon
risk.
GEOLOGICAL ASPECTS OF TWO HIGH RISK REGIONS
Western Jura Mountains, a Limestone Karst Region
As can be seen from the figures 5 and 6 226 Ra concentrations in the
soils of the Western Jura Mountains are on the average well above the
values for soils from the Swiss Plateau. At first sight this seems to
correspond well to the high percentage of increased indoor radon levels
found in this region (figure 3). But a closer look at the houses with high
concentrations shows that they are built directly onto the bare limestone
bedrock. The contact with the soil is limited to a less than 30 cm high
zone round the walls. More important than the soil aasRa concentration
seems to be the fact that most of the homes affected in the city of C. are
close to karst features like caves and sinkholes. In the basement of one
-------
¥
£
91
¦ H
Oi
3
71
L
¦o
at
x
\
v
a
a
¥
¥
u
«
500
400 H
300 H
200
100
t i I L.
I J 1 L
t I 1 L_
-1 L
L.
I
t
O Ra-226
^ Rn—222 daughters
sanplt sealed with epoxy
F igure
r
20
Tine [ days 1
sanple dried, stored in leaky container
8. Radon buildup in a Jura Mountains soil sanple
"free"
Rn-222
"fi xed"
Rn-222
H ) DRILLED HOLE
DIA. 7 CM
P : INFLATABLE
PACKER
S : SOIL
MATER
TRAP
leee
AEROSOL
FILTER
AEROSOL
FILTER
VALUE .. .—¦ T0 LUCAS-
R CELL
Inn)..iiT^ C18C ccn)
GH
PG
T
GAS ENTRANCE HOLES
CYLINDRICAL PROTECTION GRID
BRASS TUBE, I.D. 12 nn
Figure 9. Soil gas sanpling equipnent. To deternine soil perneability
we neasure the tine it takes to punp 1000 ccn at constant
pressure into *h» borehole. Constant pressure is produced
by thp punp piston's own weight.
-------
c
0
•*4
¥
*
L
¥
e
*
u
c
0
g
N
N
N
1
e
g
see
100
50
\ 10
u
tt 5
* O
e.i
1.1.1 I Mil .
Paranettr i
Radon-availability I Bq/n]
1 i 1111'| 11 ii| 1 i 1111'| i ni|—1 i 1 i 'i'| i i ii)—r
1 5 10 38 186 588 1888 2888
Gas perneabi 1 itsi of the soil I 18'
-13 _2
M 1
Figure 18. Radon availability in Swiss soils
PERCOLATING UATER
HOUSE
BEDROCK
(KARST)
RADON DEGASSING FROM PERCOLATING UATER
Figure 11. Proposed radon transport in a karst suiten
-------
of the buildings there is even a visible connection to the karst system.
The radon concentrations in caves below this city are very high (up to
40 kBq/m3 (10,11)). In combination with the high gas permeability this
karst system represents a very powerful radon source. Even small
connections to this source are sufficient to supply large quantities of
radon to the basement of a house.
There remains to explain the high radon levels in the air of the karst
system. The Jurassic limestone contains only about 20 Bq 226Ra/kg, by far
not enough to sustain 30 to 40 kBq/m3 222Rn in the cave air. We have
therefore proposed (12) that percolating water is transporting large
quantities of radon from the (high 226Ra) soil to the caves (figure 11).
Rockslides in the Alps
There are many villages in the Swiss Alps built onto badly consolida-
ted rockslide debris. In the Upper Rhine Valley these rockslides contain
"Verrucano". In Switzerland the term "Verrucano" means an old clastic
sediment frequently showing enhanced 22SRa concentrations. 226Ra values in
soils from the Upper Rhine Valley are shown in figure 5. The combination
of relatively high 22SRa activities with the extremely high gas permeabi-
lities in these rockslides seems to be the reason for high indoor radon
concentrations. Contrary to the situation in the Jura Mountains the homes
are built onto the "high 226Ra" material.
MITIGATION
Remedial actions are still in a test phase. Homeowners willing to
participate in pilot projects have been offered a substantial financial
support by the Federal Office of Public Health that also plans and
supervises the work.
Pilot projects carried out so far have shown that passive methods like
sealing floors are insufficient. Combining sealing with subfloor suction
has lead to the successful mitigation of several homes at still reasonable
costs.
The most dramatic reduction (to nearly outdoor radon levels) has been
achieved by an air conditioning system that allows for the control of air
flow and pressure in the basement. A heat exchanger keeps the energy
consumption low. This installation is for research only. It is too
expensive for a general use but a scaled down version may give comparable
results at reasonable costs.
In a high risk area a future homeowner could be convinced to install a
subslab suction system. We hope that he will have considerably lower radon
levels in his new home than his neighbour living with the highest radon
-------
concentration ever measured in Switzerland (45 kBq/m3 in winter).
in Switzerland the mitigation technique has not yet passed to the
private sector. There is no "radon business" in our country for there is
no real public concern about radon. This may be partly due to the lack of
limits or recommendations for indoor radon levels but distinctly more
important is the general feeling that something natural like radon can't
be harmful.
THE NEW DECADE
In the 1980s we have gathered enough data to make a reasonable
estimate for the average radon exposure of the Swiss. This average will
change only slightly even if we could double the number of buildings
measured. What we need more is to find the homes with extreme values,
homes really worth remedial actions. Therefore any new survey will
concentrate on the search for high risk regions. This search will be
guided by the knowledge gained on the correlation between geology and
radon concentration. A survey started in November 1990 in the eastern part
of Switzerland has already been planned according to this new concept.
Etched-track detectors are placed in villages on high permeability grounds
(rockslides, karst, clean coarse gravel with low lying water table,
important fault zones) and/or close to known or suspected uranium
mineralizations.
The new decade will also see recommendations or even limits for safe
indoor radon concentrations and an increased engagement and responsibility
of local authorities. The federal government will concentrate on research,
the search for high concentrations, scientific support and quality
control.
We will plead for sensible radon concentration limits. Radon is only
one of the many carcinogenic substances present in our environment.
The work described in this paper was not funded
by the U.S. Environmental Protection Agency and
therefore the contents do not necessarily reflect
the views of the Agency and no official endorsement
should be inferred.
REFERENCES
1. Gockel, A., Ueber den Gehalt der Bodenluft an radioaktiver
Emanation, Phys.Zeitschr., 9 (1908) 304
-------
2. Lauffenburger, T. and Auf der Maur, A., The Concentration
of Radon in a Town where Radium-Activated Paints were
used. In: Proceedings of the 6 th Int. Congress of IRPA,
Berlin (West), May 1984
3. Surbeck, H. The Search for Radon Sources, a Multi-
disciplinary Task.
Rad.Prot.Dosim.,24,1/4 (1988) 431-434.
4. Buchli, R., Burkart, W., Correlation among the terrestrial
gamma radiation, the indoor air Rn-222 and the tap water
Rn-222 in Switzerland. Health Phys. 57, (1989) 753-759.
5. Schuler, Ch., Crameri, R., Burkart, W., Assessment of the
indoor Rn contribution of Swiss building materials.
Health Phys. (1991), in press.
6. Surbeck, H. and Piller, G. A Closer Look at the Natural
Radioactivity in Soils. In : Proceedings of The 1988
Symposium on Radon and Radon Reduction Technology,
Denver CO, October 17-21, 1988, Environmental Protection
Agency Report EPA-600/9-89-006, Research Triangle Park, 1989
7. Scharli, U., Geothermische Detailkartierung (1:100000) in der
zentralen Nordschweiz, mit besonderer Beriicksichtigung
petrophysikalischer Parameter,
Thesis, ETH-Diss.Nr.8941, Zurich 1989
8. Murith, C., Volkle, H., Surbeck, H., Ribordy, L., in situ
gamma spectrometry in Switzerland using a portable gamma
ray spectrometer. In : Feldt, W. (ed.) Proc. XVth Regional
Cong, of IRPA, Visby, Gotland,Sweden, 10-14 Sept. 1989, 389-394
Fachverband fur Strahlenschutz, ISSN 1013-4506
9. Pochon, M. Origine et evolution des sols du Haut-Jura
suisse. Memoires de la Societe Helvetique des Sciences
Naturelles. Vol.XC. 1976
10. Piller, G. and Surbeck, H., Radon and Karst,
In : Feldt, W. (ed.) Proc. XVth Regional Cong, of IRPA, Visby,
Gotland,Sweden,10-14 Sept. 1989,p.15-20
Fachverband fur Strahlenschutz, ISSN 1013-4506
11. Rybach, L., Medici, F., Surbeck, H., Geological aspects
of radon exposure in Switzerland, In: Proc. Colloque Int. sur
la Geochimie des Gazes, Mons, Belgium, October 6-13, 1990,
in press
12. Surbeck, H., Medici, F., Rn-222 transport from soil to karst
caves by percolating water. In : Proc. of the 22nd Congress
of the IAH, Lausanne, Switzerland, August 27 - September 1,
1990, in press
-------
VI-4
A CROSS-SECTIONAL SURVEY OF INDOOR RADON CONCENTRATIONS
IN 966 HOUSING UNITS AT THE
CANADIAN FORCES BASE IN WINNIPEG. MANITOBA
by: D.A. Figley & J.T. Makohon
Saskatchewan Research Council
Building Science Division
Saskatoon, SK S7N 0W9
CANADA
Acknowledgements: The authors would like to acknowledge the assistance of
Mr. D. Harvey-Smith, Project Manager, Department of National
Defence for his valued assistance and comments during the
course of this study.
ABSTRACT
This paper summarizes the results of a cross-sectional survey of indoor radon
concentrations in a total group of 966 housing units at the Canadian Forces Base (CFB)
in Winnipeg, Manitoba. The major objective of the study was to characterize the
distribution of indoor radon levels in the housing group as the first step in
developing a radon control strategy. Subsequent investigations on sub-groups of these
houses (not reported here) were conducted to examine the building factors associated
with the indoor concentrations and the efficacy of post-construction control measures.
Measurements were obtained from 670 of the 966 housing units (69% participation).
The study group was composed of large numbers of nominally identical housing units of
several different building styles. The two-day average measurements were taken using
charcoal canisters during extremely cold weather, -28°C to -35°C. A short
questionnaire administered to the occupants by the field workers who installed and
removed the canisters recorded basic data on occupant activities and building factors.
For the entire group, the geometric mean concentration was 112 Bq/m3 (3.0 pCi/L),
approximately twice as high as the geometric mean obtained by an earlier summertime
study of 563 Winnipeg houses. Data was subgrouped based on geographic location within
the city, and the subgroup geometric mean concentrations varied between 25 and 206
Bq/m3 (0.7 and 5.6 pCi/L). Individual house measurements ranged from <10 Bq/m3 to
>5400 Bq/m3 (<0.3 pCi/L to >146.0 pCi/L). No building or occupant factors were
initially identified as being associated with the variation in levels.
-------
INTRODUCTION
In the fall of 1989, the Department of National Defence (DND) retained the authors
to design and conduct a study to investigate the indoor radon levels in the residences
occupied by DND personnel at the Canadian Forces Base (CFB) Winnipeg, MB. These
residences included housing units (PMQ's) owned by DND, bulk leased (BL) housing units
rented by DND, and barracks units (BU). Radon levels were also surveyed in 46 areas
of the officers' and non-commissioned members' messes and other occupied areas on both
north and south areas of CFE Winnipeg.
Radon has been identified as a naturally occurring pollutant that is broadl
distributed throughout Canada. In 1977 to 1980, Health and Welfare Canada conducted
a study to survey indoor radon levels in 14000 homes in 19 cities across Canada (1)
These data are frequently referred to in discussions regarding the radon situation in
Canada and are used to rank cities with respect to their radon risk potential. In thi
study, Winnipeg was identified as the Canadian city having the highest geometric mean
indoor radon level (57 Bq/m3) based on a sample of 563 houses. For the purposes of
this paper, the conversion 37 Bq/m3 — 1 pCi/L can be used.
Many studies of indoor radon levels have been conducted and while a more complete
understanding of the factors that influence indoor levels is emerging, at present the
only reliable method of estimating the radon concentration in a specific building is
to measure it (2).
The study design included three parts to be conducted consecutively:
Part 1. Cross-Sectional Survey of Indoor Radon Concentrations.
The focus of this part of the project was to provide an overview screening of the
radon concentrations occurring in the homes. The data would also provide a statistical
database for future studies. DND requested that all of the occupants of both the owned
and leased housing units be given the opportunity to participate in the study. in an
attempt to obtain the highest indoor readings, measurements were taken in the lowest
levels of the houses during calm, cold weather.
Part 2. Detailed Engineering Study of a Selected Sub-group of Houses.
This work focused on identifying the building factors that influence indoor radon
concentrations and provided information for the development of mitigation techniques
It included a more intensive study on a sub-group of approximately 40 houses identified
in the part 1 work as having the highest and lowest indoor radon concentrations
Part 3. Mitigation Study on a Small Group of Houses.
The focus of this work was to select five houses with high indoor radon
concentrations, make building modifications and evaluate the impact of th
modifications on the indoor radon concentrations.
All houses having part 1 screening levels >150 Bq/m3 had alpha track monit
installed for the period from October 90 to March 91. s
-------
This paper deals with the results of the cross-sectional survey (part 1) of the
study.
For the purposes of data presentation and discussion, the current Manitoba
Government interim guidelines (3) are referred to in the report. In summary, these
are:
1) If the screening measurement is about 150 Bq/m3 or lower and was taken during
cold weather with the house closed up, there is little chance that the home
will have an annual average concentration greater than 150 Bq/m3. Follow-up
measurements are probably not required.
2) If the screening measurement is about 150 - 800 Bq/m3, consider performing
follow-up measurements.
3) If the screening measurement is about 800 Bq/m3 or greater, perform long term
(minimum three months) measurement as soon as possible.
OBJECTIVES
Part 1 of the study had two major objectives:
1) Measurement of indoor radon levels in all CFB Winnipeg residences (PMQ, BL
and BU) and selected other buildings to give DND an accurate assessment of
the current indoor radon levels. These data would be used to determine if
additional measurements or mitigation work were required to ensure indoor
radon levels were maintained below levels established by DND.
2) Characterization of the distribution of radon levels and analysis of the
levels in conjunction with selected building and occupant factors. The
analysis would identify factors that are statistically associated with the
radon concentration and will be used in subsequent phases of work.
STUDY DESIGN
The initial phase of the study was a cross-sectional survey to measure the two day
average concentration in (nominally) all of the 966 residential units potentially
inhabited by base personnel. CFB Winnipeg engineering staff also prepared a list of
17 buildings to be monitored. A total of 46 monitors were placed in various locations
in the lowest levels of these buildings.
Prior to conducting the study, the base command prepared an information package
containing basic information about radon and a brief overview of the proposed study
which was mailed to all occupants of homes in the study. Only homes that were occupied
during the test period were monitored since gaining access to homes where the occupants
were not present was not permitted. Participation in the study was at the discretion
of the occupant.
-------
The study consisted of an initial home visit to install the radon monitor and a
follow-up visit 48 hours later to remove the monitor and complete (with the homeowner's
assistance) a short questionnaire concerning the building construction and occupant
activities. If the homeowner indicated that they would not be home when the monitor
was to be removed, they were carefully instructed as to the protocol for repackaging
the monitor. The homeowner would leave the monitor in the mailbox for pick-up
Twenty-three temporary contract employees were used over a four day period to
install and remove the monitors. They were all paid on an hourly basis and instructed
to take as much time as necessary to complete each house visit (average 15 minute }
On the day before the monitoring began, a two hour seminar was held to train all of
the personnel assisting with the study. A phone-in help line was manned at all times
so that workers could phone in for assistance. Only five calls requesting minor
information were received during the study.
The field work was conducted from 10-14 December, 1989. During the test period
the weather was clear and relatively calm with the outdoor air temperature varvi '
between -28°C to -35°C. '-ymg
Additional monitors and questionnaires were available at the base engineerin
office for persons who phoned in to say they were in the city but would not be homf
when the visits were being made. These people were invited to come to the engineering
office to pick up the materials for self administration. These data are not included
in this report.
For this survey, the sample population and the target population were identical
since all residences were included in the survey. Considerations as to sample size
representativeness of sample and estimation of the distribution of indoor radon levels
are eliminated in a total sampling program. This is an important point in desienin
radon research projects since the nature of radon concentration distributions varieS
widely depending upon local circumstances.
The following potential biases may affect the study, however, they are not-
considered significant in the analysis. ' noc
Although all of the residences occupied by base personnel were included in th
survey population, some houses were not monitored. For most cases, the reason for not
being included was that the occupants were not home at the time the house was initiaM
visited (between the hours 8:00 to 21:00 Monday or Tuesday). Several attempts we/
made at various times of the day over the two day period. 6
The non-participants may bias the selection of the data group towards residence
where a co-operative individual was home, however, there does not appear to be a S
systematic reason why this would affect the validity or interpretation of the stnd^
results. The demographic and building data obtained from the questionnaire woulH
correctly account for these occupant differences. Of the 966 potential residence
670 measurements were obtained. es,
The questionnaire contained primarily descriptive and quantitative question
concerning the building and the occupant activities during the two day monitor!^
period, a section for general homeowner comments was also included. g
-------
METHODOLOGY
The indoor radon concentration was measured using RADPAC TM activated charcoal
canisters. The nominal exposure time suggested by the supplier was two days, however,
as long as the exposure time was accurately recorded, exposures in the range of 45 to
72 hours were acceptable.
The canisters were installed approximately 0.6 m above the floor in the lowest
level of the residence, centrally located away from drafts and in accordance with the
manufacturer's instructions.
All of the canisters were received by the supplier for analysis within 48 hours
of removal from the house.
The monitor supplier was listed as a registered participant in both the US EPA and
Health and Welfare Canada quality assurance programs. In 10 locations, duplicate
canisters were installed as an internal check of the measurements.
QUESTIONNAIRE DESIGN
The questionnaire was designed to be either self-administered or filled in by the
survey employee with assistance from the homeowner. It consisted of 36 questions
requiring;
1) a yes or no response about building characteristics.
2) basic physical information about the building such as the number of windows,
main floor area or type of space heating system.
3) selection of a ranked descriptor to rate the condition of the foundation walls
and floor.
4) estimation of hours spent doing specific household activities or frequency
of door/window openings.
The purpose of the questionnaire was to obtain information on building and
occupant factors that would influence the indoor radon concentration either directly
or indirectly.
ANALYSIS
The survey yield for the entire housing group was 670 measurements from a total
population of 966 (69%), Since the geographic location was considered to be an
important factor associated with the indoor radon concentration, the data were sub-
grouped 1 to 8 (somewhat arbitrarily) based on location.
-------
on
of
j • fv,o init-ial analysis were combinations of streets based
The sub-groups used in the ini g rou*ings useA< the smallest yield was 62%
a common geographic area areas were considered to be adequately sampled,
all possible houses so all ot tne area
Ko need to eroup the houses into different
Information from the jj°™naire information can not be directly assigned a
categories. Most of e q ^ ^ & mathematical analysis, but will be useful
tt»t can be grouped together on the basis of some com.cn
characteristics and compared with respect to other
RESULTS
manaeer should be contacted for information on the detailed
The authors or project manager snoui.
survey results.
. „f ali nf ^e indoor concentration data is presented in
A frequency distribution oaarithm of the concentration in Figure la. All
Figure 1 and replotted ^th g The distribution in Figure la follows the log-
logarithms are taken to the Da ^ geomeCric mean (GM) rather than the arithmetic
normal distribution and theret , tendency 0f the data. For the entire group, the
mean is used to describe th in some cases, a surrogate grouping
to note that the S^ouping ba ^ or ventilation system type. The
based on other building d and maintained by DND and bulk leased housing
leasing company.
t-h scope of this report) the data for the individual
In future analysis (beyon e further sub-grouped based on the general
1) original construction as built in the 1940'..
2) replacement of windows and doors with more modern units.
-------
3) replacement of windows and doors along with reinsulation and siding of the
exterior walls above grade.
Other possible analysis include examination of the effect of house style,
heating/ventilating system type, foundation type and foundation condition.
DISCUSSION
The initial screening survey indicated that there is a wide variation in radon
levels in the buildings occupied by CFB Winnipeg personnel.
Overall, the indoor radon concentrations are much higher than the 57 Bq/m3
geometric mean level obtained in the Cross-Canada study by Health and Welfare Canada.
To a large extent, this may be related to the test conditions under which the
measurements were taken. The Cross-Canada study used short term (<10 minute)
measurements conducted in the summer. For this study, two day averages during sealed
house conditions in very cold weather were taken. While detailed modelling is beyond
the scope of this study, several building science principles support these results:
1) The cold outdoor temperatures would result in high sustained negative
pressures at the lower level of the buildings. This would maximize the
pressure potential driving radon into the buildings.
2) Although the high negative pressures should result in an increase in the air
exchange rate for the houses, a concerted effort on the part of the homeowners
to keep all windows and doors closed (as compared to summer when children are
home from school and window/door opening may provide the only cooling
ventilation) may have offset the pressure effect and resulted in lower overall
outdoor air exchange rates. Many of the homeowners reported taking special
care to keep their homes "sealed up" during the winter to minimize drafts and
reduce heating costs.
The groups with the lowest geometric mean indoor radon levels (25-30 Bq/m3) were
the south and north base buildings - groups 7 & 8 and the two storey six/eight family
units in group 5 located adjacent to the north base. All of these buildings had hot
water heating systems and no mechanical ventilation systems.
All of the other groups were single or double family residences. While a detailed
analysis is not provided, there is a general tendency for the north base areas to have
higher geometric mean indoor radon levels (group 1 - 206.5 Bq/m3, group 2 - 173.4
Bq/m3, group 4 - 147.0 Bq/m3) than the south base areas (group 3 - 110.2 Bq/m3, group
6 - 156 Bq/m3).
Table 2 lists the values for the replicate measurement tests. For the ten
locations, two charcoal monitors were placed side by side and exposed for the same time
period. In nine cases, the agreement was within a maximum range of 16.7% and typically
much smaller. For the test at Location E, the monitor results varied by a factor of
six. There are no procedural differences that would account for this anomalous result.
Using a paired t-test analysis, the differences between the measured values (excluding
Location E) were not significant at the 5% level of significance.
-------
The work described in this paper was not funded by the
U.S. Environmental Protection Agency and therefore the
contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
Although this project was funded by the Canadian Department of National Defens
DND does not endorse the products or techniques used by the authors.
REFERENCES
1. Letourneau, E.G., McGregor, R.G., Walker, W.B., "Design and Interpretation of Lar
Surveys for Indoor Exposure to Radon Daughters", Radiation Protection Dosimetr^
Vol.7, No.1-4. y'
2. "Radon Reduction Techniques for Detached Houses - Technical Guidance, Seco h
Edition", United States Environmental Protection Agency, Washington, DC., Jamjar*1
1988. " y*
3. "RADON - An Interim Guide for Manitoba Homeowners", Manitoba Energy and Min
Information Center, Winnipeg, MB., 1989. es
4. Health and Welfare Canada, "A Radon Guideline for the Department of National H
and Welfare". ealth
-------
TABLE 1. INDOOR RADON CONCENTRATION DATA FOR GEOGRAPHICAL SUB-GROUPS
GROUP
NO. OF
HOMES
NO. OF
MEAS.
GEO. MEAN
Rn (Bq/ms)
ARITHMETIC
MEAN (Bq/ms)
NO. >
150 (Bq/m9)
NO. >
800 (Bq/m3)
1
214
145
206.5
319.1
88
10
2
180
111
173.4
206.3
68
1
3
243
181
108.2
183.0
47
6
4
106
65
147.0
200.8
34
1
5
105
77
25.7
29.2
0
0
6
118
91
152.1
266.7
40
4
7
-
35
27.2
47.0
2
0
8
-
11
27.8
35.0
0
0
-------
2. COMPARISON OF REPLICATE MEASUREMENTS
Location
RADPAC
Bq/m3
% Diff.
Location A
125.8
125.8
0
Location B
88.8
88.8
0
Location C
103.6
99.9
3.6
Location D
55.5
66.6
16.7
Location E
140.6
806.6
82.6
Location F
388.5
388.5
0
Location G
299.7
299.7
0
Location H
629.0
629.0
0
Location I
510.6
503.2
1.4
Location J
92.5
88.8
4.0
-------
n
v
in
U-l
o
¦P m
c e
a)
E
CD
3
in
•
O
z
o
H
a
to
h-l
06
H
s/j
>
0
Z
U
1
ft
Cd
g
o
M
In
- o
o o o
r*- id m
o o
•s}- to
o
CM
oooooooooo
'-ocnoor^uDin^fOCM
A0N3n03dJ
-------
o o o o o
io m ^ cm
to
£_
cr
CD
Z
0
1
<
cr
uj
a
z
o
a
z
o
a
<
a;
o
O
§
co
tH
WJ
I
d
<
h.
O
§
M
B
CO
I
<
rH
(4
A0N3nO3dJ
-------
pecu jaiawuad
-------
BM3QN JSTUDIES^USL BEmUH„Q3LJUMiyA^ CAMD6
By: D.R. Morley1, M.M. Ghomshei2, C. Van Netten3 and B.G. Phillips1
1 - Province of British Columbia. Ministry of Health, Radiation Protection
Service, 200 - 307 West Broadway, Vancouver, B.C. V5Y 1P9, Canada.
2 - Orchard Geothermal Inc. 500 - 342 Water Street, Vancouver, B.C. V6B 1B6,
Canada.
3 - Faculty of Medicine, The University of British Columbia, 5804 Fairview
Avenue, Vancouver, B.C. V6T 1V5, Canada.
ABSTRACT
Three radon studies, involving 150 background gamma measurements and
long-term alpha track tests in a total of 400 homes have been conducted in
three geologically diatinct areas of the province of British Columbia. A
positive correlation between the background gamma radiation and the measured
radon level can be depicted only at the regional scale.
In the Coa3tal Area where the terrestrial gamma radiation is low, no
homos were found to exceed 4 pCi/1 on the main floor. In the Kootenays, where
the background gamma is relatively high, considerably higher radon levels were
encountered. The highest radon levels were encountered in the West Kootenay
where the terrestrial gamma level is comparatively higher than Ea3t. Kootenay.
In West Kootenay about 45% of the homes demonstrate radon levels above 4 pCi/1
and 7% above 20 pCi/1 on the main floor. In the same area more than 60% of
the homes demonstrate radon levels above 4 pCi/1 in the basement. In the East
Kootenay, where the radon Levels were found to be lower than in West Kootenay,
the terrestrial gamma levels are generally lower.
In some areas, the age of the house and the combustion air supply system
seem to correlate with the radon level. In West Kootenay, where the highest
radon levels were encountered, a positive correlation was observed between the
radon level in the basement and the age of the house. The correlation was
negative for the main floor (possibly due to air circulation between main
floor and the basement in the new houses). tn both Kootenay areas, the homes
with combustion air supply from outside demonstrated reduced radon levels on
the main floor (moat probably due to increased inside pressure).
-------
INTRODUCTION
Surveys of radon gas in homes have been carried out in Canada since
1977(1). These early studies, although they covered over 14,000 homes, were
based upon single grab measurements normally in the basement of the house
during summer and therefore gave poor indications of the annual exposure(2)
and the correlation between annual exposure and construction parameters. The
British Columbia (B.C.) Ministry of Health began making terrestrial gamma ray
measurements in 1980 using both Thermoluminescent Dosimeters (T.L.D.s) in 25
locations(3), and a portable high pressure ionization chamber (Reuter Stoke
RSS-111) in 150 locations. The areas of higher gamma activity generally
corresponded with rock structures where uranium is likely to occur(4). We
found the province could be divided into three gamma background areas. The
coastal region of the province has very low gamma background, a moderate or
normal gamma background regions that is located in the interior of the
province: and an elevated gamma radiation area of the province which is
scattered about the interior and associated with areas favourable for uranium
deposits. Figure 1 shows the province of British Columbia and the three areas
where our^ long -term radon surveys were carried out during 1988 - 89.
/
YUKON
ALASKA
BRITISH
ALBERTA
COLUMBIA
200 km
• Castletaar
ancouver
MONTANA
WASHINGTON
Figure 1 - British Columbia with the locations of 3 radon surveys
-------
The first long term radon 3tudy was carried out in the town of
Castlegar, the West Kootenay region of the province. A previous radon grab
a ample study indicated elevated radon levels in many basements there(5). Th<=»
region has an elevated gamma background and there ha3 been uranium exploration
in the area. The second study was carried out in the East Kootenay
(Cranbrook) region of the province. Moderate gamma radiation levels had been
detected in the region. However ,iuat south of the region, moderately elevated
radon levels had been measured in Montana!6). Hie third study was conducted
in the Greater Vancouver Region of the coastal British Columbia low gamma
background region. In a previous grab aample atudy(l) by Health and Welfare
Canada in this area, only low radon concentrations were found.
This paper will compare the data obtained in these three long-term radon
surveys, the terrestrial gamma surveys, other geological and construction data
available. This is the first step in developing a good potential radon risk
model for homes in this province.
METHODOLOGY
RADON SURVEYS
The first survey was conducted in Ca3tlegar, in the We3t Kootenay area
of B.C. The home3 are located on glacial terraces created by the Columbia"
River. The soil is dry gravelly and permeable. The monitors were installed
in July of 1987 and removed in March of 1988. This period was representative
of the observed annual weather pattern. 74 homes were monitored (73 homes
returned monitors). All but one home had an upstairs and a downstairs (or
basement) monitor. The monitors were mounted 4 -- 7 feet above the ground
away from drafts and placed in an area where the family commonly resided.
Measurement of the terrestrial gamma were made at each house, usually outside
on the front- yard.
The second survey was conducted in the East Kootenays where 157 of the
160 monitors were recovered. The monitors were placed one per house. In this
study, the owner decided if the monitor should be placed in the basement or
upstairs. The monitors were again placed 4-7 feet above the ground, away
from draft and in a central living area. They were installed in .January 1988
and removed in July 1988. It was our observation that this period represented
an average annual weather conditions. Most of the monitors were placed in
Cranbrook, the principal community in the area but some were placed in the
nearby communities of Fernie, lnvermere, Kimberley, Creston and Golden. These
communities are located in valley bottoms at the foot of the Rocky Mountains
The third set of 140 monitors was distributed in Greater Vancouver area
135 monitors were recovered. Although the terrestrial gamma background is
consistently low the geology varies from the Fraser River delta (rich farming
land) to the North Shore mountains and includes bed rock and glacial out wash
One monitor was placed in the main living area of each home 4 - 7 feet above
the floor and away l'rom drafts and corners. No terrestrial gamma measurements
were made in Greater Vancouver in this study 3ince previous studies had
detected no significant difference from one home to another in this area. The
radon monitors were placed in January 1988 and removed in August 1988. This
period was observed as representative of the average annual weather pattern
Similar weather patterns have been observed in previous years by Ghomshei et
-------
a 1 (8") < lur i. ng thei v rad'..n studies -
At the time of monitor placement, a questionnaire was completed by the
surveyor. Information was coLleot-ed on the location of the monitor(s), house
construction, age uf home, home occupancy, basement or slab construction,
possible radon pathways, heating and ventilating systems and the geological
environment, .Surveyors were instructed on required procedures prior to going
out so that, survey consistency would be maintained.
ANALYTICAL PROCEDURES
Terradex alpha tract radon detectors were obtained from Landauer Inc.
They were type DRN and had a detection limit of about (.4 pCi/1) month.
Terrestrial gamma measurements were made using a Reuter--Stokes RSS-111
Environmental Radiation Monitor. The monitor's gamma ray response extends
from .060 MeV to above 8 MeV. Correction for cosmic rays was made by
recording the barometric pressure and subtracting the corresponding cosmic ray
component, as specified in the operators manual for the Renter Stokes
i nstrument.
RESULTS AND DISCUSSION
TBRRESTRIAL GAMMA MEASUREMENT
TerrestriaL gamma radiation levels were determined in 150 areas of the
province (2 to 100 mea3urement3/area). The province (Figure 1) can be divided
into 3 regions of terrestional radiation intensity. The first or low
background area is the coastal strip composed of the two tectonic belts which
were most recently rafted into North America to build the province. The
second, moderate terrestrial radiation area, composes much of the interior
area of the province. Within this interior area are large areas of high
terrestrial background. These areas correspond to the areas identified by the
British Columbia Ministry of Mines a3 being favourable environments for
uranium deposits*7).
There is a good correlation between average terrestrial gamma background
.and average radon concentration (see Figure 2). This however does not. follow
through to the individual homes. There wa3 no correlation between the
individual homes terrestrial gamma intensity and the radon levels found in the
basement or upstairs. The elevated terrestrial gamma was not the only
indicator of potential radon problems in communities. In the Castlegar, West
Kootenay area the 30il was dry, gravelly, and permeable while in the East
Kootenay area a number of communities (Cronbrook and Cre3t.on) were underlain
with cloy which appears to retard radon migration.
-------
Radon Cone, in pCi/l
10
Castlegar
8
6
4
2
Vancouver
0
60
80
40
20
0
Gamma ray intensity in mR/y
Figure 2 ~ Main Floor Radon Level3 aa a Function of
Terrestrial Gamma Radiation
CAiJTLKGAR, WKST KOOTKNAY ARKA
Ai] but one of the homes surveyed in Castlegar had two level3. The land
wa.i strongly sloped and the lower levels were sunk into the ground on at least
o sides of the home. Most of the second level was located above ground. The
average main radon level was 6.5 pCi/l and the average basement level was 10.6
pOi/1. A low pass filter waa applied to smooth out some of the fluctuation in
t.he data (figure 3). can be 3een froED figure 3. about 45% of the homes
demonstrated radon levels above 4 pCi/l and 7% were above 20 pCi/l on the main
floor. More than 60% of the homes had radon levels above 4 pCi/l in the
basement. Approximately 15% of the homes had higher radon concentrations
upstairs * In these ho®63 *"re3h air entering at the basement level was
probably "diluting the radon in the vicinity of the monitor. The age of the
home had a marked infl^ence on the radon concentration (Figure 4). Older
homes can be character i-zed as having poorly constructed basement foundations
with a doorway seal in# them olt trom a leaky upstairs. New homes, although
they have better constructed banemG'nts, have open stairways, an occupied
basement, -/central h and a better sealed housing envelop.
Although'less radon enters the newer building, it gets distributed over both
floors and is retained there. There was no direct correlation between
upstairs and downstairs radon levels If a home was supplied with make up
r-omhuation air the averse radon level was downstairs 10.8 +/- 10.5 and 5.4
. / H o w-i/l upstairs- no corobxistion air was supplied the average radon
h-vel"downstairs was 1<>-2 V- 12.0 pCi/l and upstairs was 8.2 +/- 10.5 pCi/l.
A!th uah the evidence i3 not 9tl>ong ^ appears that the combustion air supply
^ ;"ducfthrCti^ preaaure in the h°me reducin® in«Hration.PP
-------
] 00
80
>>
a
c
CD
3
a*
aj
S-
cu
>
10 15
5
30 35
20 25
40 45
50
Radon in pCi/1
Figure 3 - DistriV^ition of Main Floor and Basement Radon Values i Cast.legar, B.C.
14
12
10
a
6
4
2
0
RADON CONC. IN pCI/l
www
I
1
1
_ Jj
""Ml!
1
—^
m
M
¦
0-5 YRS 6-10 YRS 11-20 YRS 21-30 YRS
AGE OF HOME
30* YRS
¦i MAIN FLOOR EMS BASEMENT
Figure 4 -- Radon in Ca3tlegar Homes as Function of Age
-------
KAST KOOTKNAV DATA (CRANBBOOK AND VICINITY)
The East Knntenav area had, unlike the West Kootenay a uide vartetv of
rhe Hd3t iv . thft ,iverai,e ra,-lon values tound m the region.
housing types. U - - reiativeiy constant throughout the region. It is
Terrestrial gamma < — Creaton and Cranbrook are underlain with clay. This
interesting to »°« ^,Ore=tun^ ^ ^ ^
rrZtn,rt°M ^The other nearby communities are located on rocky «d coarse
(.ondtrucLea. r^rmeable These communities have higher radon
1->vpi" CNoXc]ear trend was detected when comparing the age of the housea with
levels- -miis mav be because both upstairs and downstairs
radon concent a - ' ^ ^ home. Combustion air intake however
meaaurementa were nut conducted >n ea^ ^ u
does apjear to h,,ve |once„tration „as 1.66 ~/- 2.90 vCUl on the main
supplied the r g the basement. It combustion air was not
^°°i ¦ pH the average'radon, concentration was 2.29 . / -2.7b pOi/L on the main
nPPr ald'> 29 +/- 3 36 pCi/1 in the basement. 10% of the main floor radon
Levels exceeded 4 pci/l and 1% exceeded 20 pCi/1.
-------
Table 1 Radon Concentrations in the Ka3t Kootenay Area
(Cranbrook and Vicinity)
Number Average Cone. Standard
Area of Homea in pCi/1 Deviation
All Survey Sites
Main Floor 90 ].7 2.8
Basement- 67 2.0 1.9
Fernie
Main Floor .3 3.2
Basement 7 1.7
Basement 10 3.4
Creston
Main Floor 14 j ^
Basement 1 'g
2.9
.6
Cranbrook
Main Floor 47 .9 7
Basement 41 1.7 ^ 7
Kimberley
Main Floor 15 2.1 2 3
c.
Invermere
Main Floor 6 7.1 8.5
Basement 4 1.6 .9
Golden
Main Floor 4 1.7 1.8
Basement 5 3.1 1.7
1.0
GREATER VANCOUVER AREA
Radon concentrations in the Vancouver region were very low (average of
0.49 +/- 0.23 pCi/1 and ranged between .2 and 1.6 pCi/1 in the 135 homes
measured. There is no significant difference from one area of the city to
another despite a large variation in geological environments. The three
factors that, are common to the city which may explain the low radon levels are
the low terrestrial gamma, an abundance of "hard pan" clay that deters radon
infiltration and higher than average rainfall.
CONCLUDING REMARKS
Terrestrial gamma measurements can be used in British Columbia to
predict a community radon potential. They however cannot be used to predict
an individual home's concentration of radon. Soil structure particularly
permeability appears to have a marked impact on radon potential (9) The
©rovince can be divided into three radon risk areas. There is a wide coastal
-------
3t.rip where: the risk of radon exposure is very low compared to other areas of
North America. However the majority of the province lies in an interior belt
where the radon risk is typical of other areas of North America. There are
patches within that belt where both the terrestrial gamma and radon risk are
relatively high. Further study is required to delineate these areas and all
homeowners in these areas should make it a priority to test their homes.
Modern changes to house construction have been reported to increase radon
concentration in the living area(lO). Although we have also seen this in the
Castlegar study, some construction techniques such as supplying combustion
area and well built basements are tending to mitigate this trend. There is
some concern that, annual variations in household radon levels may also have to
be investigated(10). At this time, fifteen additional regional radon surveys
are being carried out. These additional surveys should allow us to more
accurately delineate the geological, meterological, and construction
parameters that impact on radon levels in the province of British Columbia.
ACKNOWLEDGEMENTS
This work was supported by the East Kootenay and West Kootenay Health
Unit.3 and the British Columbia Centre for Disease Control.
The work described in this paper was not funded by the U.S.
Environmental Protection Agency and therefore the contents do not necessarily
reflect the views of the Agency and no official endorsement should be
inferred.
-------
RKFKIIKNCKS
1. Letourneau, E.G. et al. Design and interpretation of large surveys for
indoor exposure to radon daughters, Radiation Protection Dosimetry,
Vol.7, No.l - 4, p.303 308 Nuclear Technology Publishing.
2. Oswald. Richard A. The need for long-term radon measurements,
Environmental Radon Update, p.1-4, August 1986.
3. Morley and Green. Background radiation levels in British Colximbia 1980
-1983. province of British Columbia Ministry of Health publication,
1984.
4. Bates et al. Royal Commission of Inquiry into Uranium Mining, Province
of British Columbia, October 30, 1980.
b. McGregor, R.G. Background levels of radon and radon daughters in homes
in the Castlegar Trail area of British Columbia, Radiation Protection
Bureau, Health and Welfare Canada, Ottawa, November 1978.
6. Report No.10. Radon monitoring results from B.P.A's residential
weatherization program, U.S. Department of Energy, January 1989.
7. British Columbia, Ministry of Energy, Mine3 and Petroleum Resources.
Submission to the Royal Commission on Uranium Mining. Pages 16, 45,
1979.
8. Ghomshei, M.M. and Slawson, W.F. Secular variations of radon in
metropolitan Vancouver, British Columbia, Canada. Paper presented at
1990 International Symposium on Radon and Radon Reduction Technology at
Atlanta, GA. February, 1990.
9. Niel3on, K.K. and Rogers, V.C. Radon transport properties of soil
classes for estimating indoor radon entry. Paper presented at
Conference on Indoor Radon and Lung Cancer, October 15-19, 1990.
Richland, Washington, U.S.A.
10. Ergent, G.W., Kathren, R.L., Cross, F.T. The effect of home
weatherization on indoor radon concentration. Paper presented at
Conference on Indoor Radon and Lung Cancer. October 15-19, 1990.
Richland, Washington, U.S.A.
-------
VI-6
The State of Maine Schools Radon Project: Results
by: L. Grodzins, Professor of Physics
Massachusetts Institute of Technology
and Founder & Chairman of the Board of
NITON Corporation, Bedford, MA 01730
T. Bradstreet, Director of Information and Education
Division of Safety and Environmental Services
Augusta, ME 04333
E. Moreau, Manager of Indoor Air Quality Program
Division of Health Engineering
Department of Human Services
Augusta, ME 04333
ABSTRACT
A comprehensive study has been made of the radon concentrations in every frequently occupied
room on or below grade in every public school in the State of Maine. 32% of the 653 school
buildings covered in this report had at least one room with a radon level exceeding the EPA
guideline of 4 picoCuries of radon per Liter of air (4 pCi/L). 8.7% of the 13,353 rooms had a
radon level > 4 pCi/L; 1.9% of the rooms had radon concentrations > 10 pCi/L; 0.7% of the
rooms had radon concentrations
> 20 pCi/L. The radon concentrations were not distributed uniformly among the schools' a
building tended to have a radon problem or it was essentially free of radon. The radon concen-
trations were not uniformly distributed throughout the state. The schools in the counties contiguous
to New Hampshire were far more likely to have a serious radon problem than were schools in the
central part of the state, especially along the coast. And we note a strong correlation between the
geographical results of this state-wide school survey and the previous state-wide results of radon in
homes.
-------
INTRODUCTION
This is a report on the findings of the comprehensive survey of radon carried out in 1990 in
more than 13,000 classrooms in more than 650 school buildings in the state of Maine. Overall,
33% of the school buildings had radon levels exceeding the EPA action level of 4 pCi/L. Schools
with elevated radon values were not, however, uniformly located in the state. The western counties
tended to have considerably higher radon levels than elsewhere. Some schools in these counties
had mean concentrations exceeding 4 pCi/L and several had mean concentrations greater than 20
pCi/L. Most striking is the strong correlation between the radon levels in the schools of a county
with the radon levels in the homes in that county.
This study is unique in several ways. It is, to our knowledge, the first complete state-wide
school survey in which every regularly occupied room on or below grade in every school was
measured; nearly 50% of all school rooms in Maine were tested. It is the first to survey all the
schools using a single radon detection method analyzed by a single company. Without this unified
approach, the present study would not have been practical. It is the first survey in which the
placement and retrieval of the radon detectors were carried out by the school custodians, with all
scientific and technical decisions handled in advance by the testing firm, NITON Corporation; the
98.5% success rate of this procedure has important economic implications for future surveys.
The design of the study is described elsewhere in this meeting (1). So, too, are the protocols
and procedures of the testing program (2). For completeness, a summary follows. The next
section presents the results, first in overview, then in greater detail. The last section correlates the
school results with other information, particularly the radon survey of homes in the state and
summarizes our conclusions.
PROCEDURES
The radon tests were carried out using NITON's patented liquid scintillation charcoal detectors.
These small, 1" diameter by 2" long, detectors contain a cartridge holding about 1.5 grams of
charcoal mixed with desiccant. For each school, NITON made up individual packages containing
the test vials, data sheets, and a copy of the school floor plans marked with locations for placing
the test vials. Most important, the package included a set of simple, comprehensive, step-by-step,
check-off instructions.
Every regularly occupied room on or below grade was tested over a week-end under closed
building conditions. The air-handling systems were generally operated continuously. School
custodians set out and retrieved the tests and returned them to the NITON Laboratory in
Massachusetts, using next-day UPS service. This protocol worked exceptionally well even for
remote one-room school houses, including those on islands off the coast and those in Indian
reservations; only 1.5% of the rooms had to be resurveyed because of faulty procedures.
The NITON LS vials were set out on Friday afternoons, retrieved Monday morning, generally
arrived at the laboratory in Massachusetts and were counted in the automated LS counters on
Tuesday. The NITON "2-day" diffusion barrier is most sensitive to the last 48 hours of testing so
the first evening of the test effectively established the base line of closed conditions.
All test vials were counted for 5 minutes each. Prompt return of the test vials meant that the
radon decayed by only 25% to 35% between the time the vials were closed and the time they were
-------
counted. As a consequence, the 5-minute count in the automated scintillation counter resulted in a
standard deviation a <5 % at a concentration level of 4 pCi/L; a -10% at 1 pCi/L; and a -20%
at 0.4 pCi/L. All vials with radon concentrations exceeding 3 pCi/L were rerun for 20 minutes
each ( c < 2%). Results, quoted to the nearest 0.1 pCi/L, were sent to the State of Maine the
following day.
RESULTS
QUALITY OF THE DATA
The results from NITON vials were compared with themselves and with independent tests.
Over the course of the study, side-by-side tests were run in a total of 33 buildings. The results
were excellent. The mean of the absolute differences between the side-by-sides was 0.2 ±0.16
pCi/L. The mean of the absolute differences for results exceeding 2 pCi/L was 0.25 ± 0.17 pCi/L
Only 2% of the absolute differences were as great as 0.6 pCi/L; none were higher. The precision
of the results was <5% at 4 pCi/L, about the same as the statistical uncertainty of the initial liquid
scintillation test.
Quality control of the NITON vials is carried out routinely with in-house radon standards, and
checked periodically using independent radon quality control laboratories. We followed three'
additional procedures to establish the quality control for the Maine survey: 1) The NITON liquid
scintillation vials were specially tested at the Environmental Measurements Laboratory of the
Department of Energy; 2) 100 NITON LS detectors, in pairs, were compared with 50 Charcoal
Canisters; i.e. 3 detectors (2 NITON LS and 1 CC) were run side-by-side. The Charcoal Canisters
were tested by the State of Maine Indoor Air Quality group. Most of the compared results were
within 0.1 pCi/L. Two comparison tests differed widely: In one test, the LS values were 3.3 and
3.0 pCi/L, the CC result was 1.2 pCi/L; in another, the LS values were 5.5 and 5.3 pCi/L while
the CC result was 3.0 pCi/L.
3) 30 NITON LS detectors were compared by the State of Maine with continuous monitors.
Half the tests lasted 8 hours, half lasted 16 hours; NITON detectors are calibrated from 8 to 72
hours. The mean radon concentrations ranged from 4.4 pCi/L to 67 pCi/L, with short-term
variations ranging from 0.6 pCi/L to 74.5 pCi/L, according to the continuous monitor. The mean
of the 30 NITON results was 10% higher than the mean of the means of the 30 results from the
continuous monitor. These comparisons give considerable confidence in the results for the
individual schools and for the overall survey.
STATE-WIDE RESULTS
School Rooms
The results for 13,353 school rooms are presented in Table I and Figures 1 and 2. The
frequency distribution of radon in the school rooms of Maine had a most probable value of
<1 pCi/L, a median value of 1.1 pCi/L and a geometric mean of 1.05 pCi/L. These values are
not much different from those obtained by NITON in a survey of 5,000 school rooms in
Massachusetts. 8.7% of the school rooms in Maine had radon values of 4 pCi/L and above (the
corresponding number in Massachusetts was 6%); 1.9% of the Maine school rooms had radon
values of 10 pCi/L and above (Massachusetts was 1.1%); 0.7% of the Maine school rooms had
radon values of 20 pCi/L and above (Massachusetts also had 0.7%).
-------
TABLE 1. THE FREQUENCY DISTRIBUTION OF RADON IN SCHOOL ROOMS
AND SCHOOL BUILDINGS IN MAINE
Concentration, pCi/L
Rooms with >
Concentration
Column 2 as % of
13,353 Rooms
Buildings with 1 or
more rooms >
Concentration
Column 4 as % of
653 Buildings
0.4
11,550
86.5
644
98.6
1.0
7,322
54.8
587
89.9
1.5
4,884
36.6
520
79.6
2.0
3,365
25.2
445
68.1
2.5
2,429
18.2
373
57.1
3.0
1,828
13.7
308
47.2
3.5
1,420
10.6
248
38.0
4.0
1,164
8.7
213
32.6
5.0
808
6.1
167
25.6
6.0
601
4.5
126
19.3
7.0
454
3.4
93
14.2
8.0
370
2.8
77
11.8
9.0
312
2.3
70
10.7
10.0
255
1.9
54
8.3
11.0
222
1.7
49
7.5
12.0
197
1.5
44
6.7
13.0
172
1.3
38
5.8
14.0
155
1.2
35
5.4
15.0
144
1.1
33
5.1
16.0
131
.9
31
4.7
17.0
118
.8
28
4.3
18.0
109
.8
24
3.7
19.0
102
.7
21
3.2
20.0
98
.7
21
3.2
25.0
68
.5
14
2.1
30.0
48
.35
11
1.7
40.0
16
.11
7
1.1
50.0
2
.01
2
.3
-------
The distribution is plotted in Figure 2 as a lognormal probability graph. The data exhibit the
fact, now familiar in most radon surveys, that the radon distributions follow a normal probability
distribution up to about 4 to 5 pCi/L. The probability of observing elevated radon concentrations is
higher than would be predicted on the basis of a normal distribution. This study found twice as
many school rooms with radon concentrations above 10 pCi/L, and ten times as many above 20
pCi/L, as would be inferred from a normal distribution.
.5
5
6
X
£
o
X
u
V3
<*«
0
o
£
c
Ci
V
u
ZJ
8.7% of School Rooms
1.9% of School Rooms
0.7%
of S
phool Rooms
0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-99-10 10-15
Radon Concentration, pCi/L
15-20
>20
Figure 1. The distribution of radon in the school rooms of Maine
N B The uncertainties in these statistical values and those given later in the paper, are due
, 7' tL nnrertaintv in the accuracy of the test results, which we assume to be -10%
onTe ba"s s of NITON'S overall accuracy in several EPA Quality Assurance rounds. A 10%
uncertain y in the absolute accuracy results in a corresponding 10% uncertamty to the m«i,an
arithmetic and geometric means, as well as to the percentage of school rooms exceed.ng lO pCl/L
and 20 pCi/L The percentage of school rooms above 4 pCt/L depends more strongly on the
ana l\) pv,wi-.\ h | , -r~enness 0f the distnbution at 4 pCi/L. For example, if
aCKU ? fl^Hhrations have a systematic error such that all results are 10% too high (and
'Hen only 7% of the schcx,! rooms are
-------
above 4 pCi/L. If NITON's values are 10% low then 10.5 % of the school rooms are above 4
pCi/L. The sensitivity of the results to the absolute accuracy in the tests is a compelling reason
why surveys should be carried out using a single method and, wherever possible, by a single
group using the same calibration standards. In practice it is very difficult to accurately compare
surveys carried out by different methods or by different laboratories.
The distribution is plotted in Figure 2 as a lognormal probability graph. The data exhibit the
fact, now familiar in most radon surveys, that the radon distributions follow a normal probability
distribution up to about 4 to 5 pCi/L. The probability of observing elevated radon concentrations is
higher than would be predicted on the basis of a normal distribution. This study found twice as
many school rooms with radon concentrations above 10 pCi/L, and ten times as many above 20
pCi/L, as would be inferred from a normal distribution.
100
Frequency Distribution of Radon Concentrations
in School Rooms of M*ine
Geometric: Me^n: 1.05 ^Ci/L
10
i
i
S 10
20 >0
01
SO
70 SO 90 »S
I
Percentage of school rooms having less than given concentration (%)
Figure 2. Lognormal plot of the distribution of radon in the school rooms of Maine.
-------
School Buildings
The data for the 653 school buildings examined in this survey are presented in Table I
Columns 4 and 5 give, as a function of radon concentration shown in column 1, the number and
percentage of school buildings that have at least one room with a concentration greater than that
level. 32.6% of the buildings had at least one room with a radon level o> 4 pCi/L. Stated the
other way, 67.4% of the buildings had no room with a radon concentration exceeding the EPA
action level. 8.3% of the buildings had at least one room with a concentration of 10 pCi/L or
greater, and 3.2% of the buildings had at least one room with a radon concentration of 20 pCi/L
These state-wide percentages tell an incomplete story since both the geographic location and the
size of the school are critical variables.
School Size
Table II shows the distribution of the number of rooms per school building. The typical school
in Maine has fewer than 20 rooms, 151 schools have fewer than 10 rooms. The third row of Table
II gives the number of school buildings that have at least one room with > 4 pCi/L of radon. The
bottom row of the table gives the percentages of buildings that have at least one elevated radon
reading. The percentages vary from 21% to 50% but, within statistical uncertainties, the
percentages are essentially constant. This is a most suiprising finding since one would expect a
priori, that the larger the school the greater the probability of finding an elevated radon level. '
TABLE II: SOME DISTRIBUTIONS IN SCHOC
)L BUIL
DINGS IN MAIN
5
Number of Rooms per Building
<10
10-19
20-29
30-39
40-49
> 50
Total number of buildings
151
184
141
90
40
42
Total number of buildings with at
least one room > 4 pCi/L
40
65
41
36
20
9
% of "High-Radon" Buildings
27%
35%
29%
40%
50%
21%
The explanation is that radon is not randomly distributed in the schools. School rooms in a
building that has a common architecture and air-handling system, show very similar radon '
concentrations. To emphasize the lack of randomness, consider the larger buildings with more
than 50 rooms. If radon were distributed randomly we would find 8.7% of the rooms of each
school with elevated radon concentrations. The probability that a school with 50 rooms would
have no elevated radon room is (0.913)50 = 1.1%; the actual probability is 79%.
The "one-room" school houses do not follow a random pattern either, though one expects
the basis of the state-wide data, to find about 30% of the buildings with a radon problem The lack
of randomness is demonstrated in Table III, which breaks out the data of the third row of Table II
to give, as a function of the size of the school, the number of buildings that have a given
percentage of the school rooms with > 4 pCi/L. Thus, of the schools that have at least one radon
problem, there were 5 large school buildings (>50 rooms) in which fewer than 10% of the rooms
had elevated radon. There was, however, one large school building in which more than 80% of
the rooms had an elevated radon problem.
-------
TABLE III: THE NUMBER OF BUILDINGS AS A FUNCTION OF THE NUMBER OF
ROOMS AND THE PERCENTAGE OF ROOMS WITH ELEVATED RADON LEVELS
Number of Rooms per Building
<10
10-19
20-29
30-39
40-49
> 50
<10% of the rooms
0
14
20
19
9
5
10% - 39% of the rooms
22
37
12
11
7
1
40% - 59% of the rooms
1
5
5
3
0
1
60% - 79% of the rooms
5
6
3
2
3
1
>80% of the rooms
12
3
1
1
1
1
Total number of buildings
40
65
41
36
20
9
A number of school buildings (bottom row of Table III) were saturated with radon. There is
negligible probability that any of these saturations could occur by chance. Every one of the 48
rooms in one school building was far above the EPA guidelines; the median value was 25 pCi/L.
All of the buildings in the lower part of the table support the general observation that high radon
levels tend to cluster; there are relatively few buildings that have isolated rooms with elevated radon
levels.
The school buildings in the second row of Table III have, typically, only 1 or 2 high-radon
level rooms. Unfortunately, the odd high radon value can be very high indeed. For example, in
one school of 23 rooms, having a median radon level of 1 pCi/L, there was one classroom with
27 pCi/L; in a 4-room school house where 3 of the rooms were under 4 pCi/L, there was one
room with 38 pCi/L. In the next section we examine a few of the radon distributions in individual
school buildings.
Results of Individual Schools
The present study involved more than 650 school buildings and NITON has surveyed more
than 500 other school buildings during the past two years. The buildings have different ages,
architectures, geological sites, air-handling systems, etc. From the mitigator's point of view, each
building is unique. From the radon surveyor's point of view, there are definite patterns of radon
distributions that can be useful guides to understanding the origins of the radon problem.
Figure 3 shows three radon distributions found in schools in Maine. Distribution A is a "radon-
free" building; the median value is well below 1 pCi/L and no concentration is greater than 2 pCi/L.
There is no correlation between the radon levels and the room location. The most probable value is
similar to the radon level found outdoors.
Distribution C is a "radon-infested" building in which the radon levels are about the same in
every room. The distribution is nearly Gaussian; the mean of 24.9 pCi/L is same as the median
value.
Distribution B is very similar to C, though the mean and the median are both below 4 pCi/L.
Every room in the school has a potential radon problem as the levels fluctuate with changes in the
weather and the air-handling.
-------
Radon Concentration, pCi/L
Figure 3. Radon distributions in three schools in Maine.
Figure 4 shows three distributions observed in school buildings of Maine. These are rather
typical of the broad distributions that almost always show a correlation between the radon
concentration and the room location. Contiguous rooms show similar radon values; changes in
concentration take place over several to many rooms..
-------
25
2ft
IS
10
S
2 : 2
5 S 2 2 2
s :
s
Radon Concentration, pCi/L
Figure 4. Radon distributions in three schools in Maine
The most probable radon concentration of distribution D is below 1 pCi/L and most of the
rooms are radon-free. Nevertheless, a few rooms have values well above the EPA guidelines; it is
our experience that these rooms are generally localized to the same area of the school.
Distribution E is also sharply peaked at a low radon value, but the median value is close to
2 pCi/L, indicating a radon problem. We again anticipate a strong correlation between the radon
concentration and the geography of the room.
Distribution F has no reading above 3.5 pCi/L. But the median value of 2.5 is high. This
building should be carefully monitored over time since it is likely that there will be periods during
the year when the radon levels will rise by at least a factor of two and most of the rooms will have
concentrations exceeding EPA guidelines.
Only distributions similar to those of A in Figure 3 can give us reasonable assurance of a school
without a real or potential radon problem. The assurance is not, however, a guarantee. We have
several examples of schools in this survey in which there is one or at most two elevated radon
concentrations in an otherwise radon-free school. To find such rooms, one must survey every
room on or below grade.
-------
Survey Results by Maine County
Table 4 presents the results by county. The last 3 columns give the results of the Maine survey
of radon in homes.
TABLE 4: RADON IN THE SCHOOLS AND HOMES OF MAINE, BY COUNTY
County
Total
School
Rooms
Median
pCi/L
Maximum
pCi/L
% >4
pCi/L
Total
Houses
Maximum
pCi/L
% >4
pCi/L
Androscoggin
1,065
1.3
21.8
5.34
39
17.7
18
Aroostock
1,493
1.5
18.7
11.25
95
25.2
43
Cumberland
2,222
1.4
59.2
16.42
120
82.7
41
Franklin
345
.5
15.8
6.66
19
9.7
16
Hancock
654
1.2
21.4
5.5
49
19.4
27
Kennebec
1,062
1
43.5
5.83
57
19.4
26
Knox
151
.6
4.5
.66
24
9.7
29
Lincoln
116
.6
13.8
4.31
12
5.9
8
Oxford
580
1.5
37.7
14.47
37
30.2
51
Penobscot
1,783
.7
17.9
3.64
72
5.7
17
Piscataquis
384
.6
5.7
0.26
37
22.5
32
Sagadahoc
499
.9
5.3
1.4
32
8
19
Somerset
746
.8
26.4
4.28
27
5.8
19
Waldo
223
1
7.6
6.27
26
13
19
Washington
539
1
12.6
1.66
36
12.2
14
York
1,491
1.4
41.6
15.75
73
33
38
Total Tests
13,353
755
The counties vary widely in population and, therefore, in the number of schools and school
rooms. There are large variances in the measures of radon concentration in the homes of several of
the counties, particularly, Lincoln, Knox and Waldo. The school measurements are much more
secure. Even the smallest county had more than 100 school tests and the median value is measured
to an uncertainty of less than 0.2 pCi/L.
Table 4 gives three indicators of the radon concentration in the schools of the different counties
The maximum radon value, column 4, can be a statistical outrider and is not a useful measure of '
the radon problem in the county. The percentage of rooms that exceed 4 pCi/L, column 5, is a
much more useful indicator since it focuses on that part of the distribution which demands'action
The median radon values, column 3, while not giving the full description that would be obtained'
from the geometrical moments of each distribution, does give an easily understood measure of the
-------
radon problem and is, in our view, the best single number to quote. If half the rooms have a radon
concentration below 0.7 pCi/L, which is the case in 5 counties, one can be quite sure that fewer
than 10% of the school rooms will have elevated values. On the other hand, if half the rooms have
a radon concentration greater than 1.4 pCi/L, which is the case in 4 counties, one can be quite sure
that more than 10% of the rooms will have concentrations greater than 4 pCi/L.
The four counties with the highest radon concentrations in the schools also have the four
highest radon concentrations found in homes. The correlation between the percentage of school
rooms in a county that are > 4 pCi/L (column 5) and the median radon concentration found in the
county (column 3) is r2 = 0.58. The correlation between the values in column 5 and the highest
radon value found in any school room in the county (column 4) is r2 = 0.65. There is little
correlation between the median and the maximum values in a county.
Figure 5 shows, by county, the percentage of school rooms and homes that exceed the EPA
action level. The values for homes have been divided by a factor of 3. (These home readings are
generally basement readings. The first floor level in New England is, as a rule of thumb, about
one-third of the basement level.)
There is a striking correspondence between the results for schools and those for homes in the
same counties. There are several exceptions to the obvious correlation, but most have large
uncertainties in the individual values due to the small sample sizes.
~ Percent of School Buildings >= 4 pCI/L
Percent of Homes >= 4 pCi/LI (+3)
M 11! i M s i t i I M *
3 w £ w i
Counties in Maine
Figure 5. Radon Concentrations in School Rooms and Homes by County in Maine
-------
SAGADAHOC
ANDROSCOGIN
Median Levels of Radon
in Counties in Maine. In pCi/L
-------
The final figure shows a map of Maine divided into its 16 counties. The median radon
concentration for each county is given in pCi/L. The map has been shaded to show the strengths
of the radon values; the darker the shading the higher the median value. Radon is obviously
correlated with the geography of Maine. The four southwestern counties, York, Cumberland,
Androscogin and Oxford have uniformly high mean radon values. There is then a band of
moderate radon concentrations extending from Kennebec through Waldo, Hancock and
Washington. The central coastal counties of Sagadahoc, Lincoln and Knox have low radon
concentrations, as do the Maine-woods counties of Franklin, Somerset and Piscataquis.
Aroostook county, with towns and schools bordering New Brunswick, Canada, is another area of
high radon concentration. As we noted above, the areas bordering New Hampshire and New
Brunswick have the highest radon concentrations in both school rooms and homes.
SUMMARY
The principal aim of this comprehensive survey of radon in the schools of Maine was to find
those schools that should be mitigated immediately, as well as those schools that have potential
problems that must be monitored over time. That aim has been well met. A second aim was to
obtain a data base of the radon concentrations in every school, which would serve as the bench
mark and guide for spot checks and sample surveys that might be conducted in future years either
as part of a general "due diligence" program or because of changes in construction or air handling.
That aim, too, has been well met. A third aim was to find generalities and correlations that might
aid in understanding the radon problems in the state. This paper presents the sum of those
findings.
The frequency distribution of radon in school rooms in Maine is similar to that found in surveys
of school rooms elsewhere; for example, in Massachusetts. The distribution follows a lognormal
curve up to radon concentrations of about 5 pCi/L. The occurrence rate at the higher concentra-
tions is greater than would be predicted by the normal curve.
8.7% of the school rooms tested over weekends under closed building conditions were found to
have radon concentrations greater than the EPA action level of 4 pCi/L; 1.9% of the rooms were
above 10 pCi/L; 0.7% were > 20 pCi/L. The elevated radon concentrations were not randomly
distributed. A school that had one room with > 4 pCi/L, had, on the average, 5 such rooms. The
odds that a school building had at least one room with > 4 pCi/L, was about one in three. The
odds were almost independent of the size of school.
The patterns of radon distributions in the schools can be conveniently divided into three broad
groups: 1) those radon-free schools with median (or geometric mean) values well below 1 pCi/L,
no concentration greater than 2 pCi/L, and no correlation between the radon values and the position
of the school room; 2) those radon-infested schools with median and mean values above ~3
pCi/L, and with a majority of the values exceeding the EPA guidelines; 3) schools that have
median (or geometric mean) values in the 1-2 pCi/L range, have broad distributions (large standard
deviations of the geometric means), and generally a strong correlation between the position of the
room in the building and its radon concentration. The first group is the only one with a strong
probability of being radon-free under all circumstances of weather and air handling; the second
group must be mitigated early; the last group encompasses a wide variety of situations with few
common denominators other than the need for close examination and monitoring.
The radon concentrations showed a strong correlation with geography. The median radon
concentrations in the western counties — Oxford, York, Cumberland and Androscogin — and the
-------
northernmost county, Aroostock, were more than twice the values in the coastal counties of Knox
and Lincoln and the Maine woods counties of Franklin, Somerset, Piscataquis and Penobscot.
The counties with highest (lowest) radon concentration in the schools generally had the highest
(lowest) average concentration in the homes. This strong correlation between radon levels in
school rooms and homes in the same geographical area implies that the underlying geological
factors are the determinants for the average or median radon concentrations in county-size areas.
We anticipate that this is a general conclusion that will be observed throughout the country.
The correlation between radon levels in homes and in schools is strengthened by comparing the
Maine results with those obtained by NITON Corporation tests of more than 5,000 school rooms
in Massachusetts. The EPA state-wide studies of homes found that 25% of the "lowest-livable"
rooms in Massachusetts had values > 4 pCi/L, compared to 30% for Maine. Massachusetts has
only 6% of its school rooms exceeding the EPA guideline, compared to 8.7% for Maine. Thus
there are fewer school rooms with high readings in Massachusetts to about the same degree that
there are fewer homes with high radon readings.
We also consider it worth noting that the percentage of school rooms with > 4 pCi/L of radon is,
for both Maine and Massachusetts, about one-third the percentage of homes with basements with
radon concentrations > 4 pCi/L. This correlation implies that the average radon values in schools
is not much different from the radon concentrations found on the first floors (the living areas) of
homes in the same geographical area.
ACKNOWLEDGEMENTS
L.G. would like to take this opportunity to thank Anne McGuineas and Christopher Collins of
the staff of NITON Corporation for their indispensable help with the paper generally, and the
generation and understanding of the data, specifically.
The work described in this paper was not funded by the U.S.
Environmental Protection Agency and therefore the contents do
not necessarily reflect the views of the Agency and no official
endorsement should be inferred.
REFERENCES
1. Warren, H.E. and Romm E.G. The State of Maine School Radon Project: The Design
Study. EPA 1991 International Symposium on Radon and Radon Reduction Technnlnov
Philadelphia, PA. April 2-5, 1991.
2. Grodzins, L., Warren, H.E., and Romm, E.G. The State of Maine School Radon Project-
Protocols and Procedures. EPA 1991 International Symposium on Radon and Radon
Reduction Technology, Philadelphia, PA. April 2-5, 1991.
3. Grodzins, L. Radon in Schools in Massachusetts. EPA 1990 International Symposium on
Radon and Radon Reduction Technology, Atlanta, GA. February 19-23, 1990.
-------
IV-7
THE EFFECT OF SUBSLAB AGGREGATE SIZE ON PRESSURE FIELD EXTENSION
by: K.J. Gadsby, T. Agami Reddy,
D.F. Anderson, and R. Gafgen
Center for Energy and Environmental Studies
Princeton University
Princeton, NJ 08544
and
Alfred B. Craig
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
Four sizes of commercially available crushed blue gravel
aggregate (3/8, 1/2, 3/4, and 1 in. nominal diameter) were tested
in a laboratory apparatus designed to experimentally determine the
aerodynamic pressure drop coefficients of porous material such as
soil and gravel. Permeability values for crushed stone of 1/2 and
3/4 in. nominal diameter were found to be 10-20 times higher than
those reported in a previous study for river-run gravel of the same
nominal diameter. Pressure field extension of this aggregate, when
suction is applied to a single central suction hole (a practice
widely used for mitigating buildings with elevated radon levels by
the subslab depressurization technique), is also generated based on
a disc flow model previously studied by the authors. Application
of the disc flow model to residences with both basements and slab-
on-grade construction is also described. Theoretical computations
indicat-e that the permeability of soil around the foundation walls
and periphery of the residence is a more crucial parameter
affecting the pressure field extension than the permeability of the
subslab gravel bed. The laboratory studies will be field verified
in new construction of schools and houses in the future.
This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.1
'This work was funded by the U.S. Environmental Protection
Agency under Cooperative Agreement No. CR-817013.
-------
PROBLEM STATEMENT
Subslab air flow dynamics provide important diagnostic
information for designing optimal radon mitigation systems based on
the subslab depressurization (SSD) technique. An earlier study
(Refs. 1 and 2) showed that subslab air flow induced by a central
suction point can be mathematically treated as radial air flow
through a porous bed contained between two impermeable discs.
Subsequently, it was suggested that subslab material commonly found
under residential buildings be categorized and tested in the
laboratory in order to deduce their aerodynamic pressure drop
coefficients. This would then permit the pressure field extensions
to be inferred which would be of practical and realistic importance
to radon mitigators. To this end, a laboratory apparatus was
designed and built which is described fully in Ref. 3.
The scope of this study was limited to four aggregate mixes —
No. 600, 601, 602, and 603 — provided by Vulcan Materials from
their Manassas Quarry. The physical specifications provided by
Vulcan Materials are given in Table l. The aggregate is crushed
blue gravel, a material commonly used as subslab material for
residential and commercial housing.
This paper first reports aerodynamic flow coefficients of the
four aggregate mixes obtained ^ from Princeton's laboratory
apparatus. Subsequently, it describes how the disc flow model can
be applied to residences with both basement and slab-on-grade
construction. Finally, it shows generated pressure field
extension plots for each aggregate in the framework of the above
model and discusses the practical implications of these plots in
the design of SSD systems.
EXPERIMENTAL RESULTS
The laboratory apparatus, described fully in Ref. (3), is
shown in Fig. 1. It consists of a straight length of 8 in.* pvc
pipe approximately 5 ft long to the bottom of which a 1/4 in.
aluminum removable screen is fitted. The screen, which slides in
and out along a groove machined into the PVC pipe, is perforated
with a pattern of 1/4 in. diameter holes to permit air flow. The
porous material is loaded into the test column from the top after
which another sleeve, containing flow straighteners and the air
supply inlet tube, is placed on top of the column. Air is supplied
to the system from the top and several pressure taps provide
information on the pressure drop in the porous bed as air flows
down through the porous bed in the test column. The total air flow
rate is accurately measured. The experimental procedure involves
* For readers more familiar with metric units, l in. ¦ 2.54 cm and
1 ft « 30.48 cm.
-------
measuring the pressure drop (AP) through a known and predetermined
bed length (AL) for a specific value of total flow rate (q).
Several such tests are carried out over a range of q values for
each sample of aggregate and for different samples of the same
aggregate. A least-square regression finally provides estimates of
the permeability (k) and the flow exponent (b) of the aggregate.
Table 2 assembles results of the porosity experiments, three
different samples of each aggregate tested at least twice. This
involved choosing a volume of the aggregate material and then
finding the porosity by measuring the volume of water needed to
completely saturate each sample (Ref. 4). The values of porosity
(0) specified by the supplier and those obtained from the tests are
shown in Table 2. Note that, in general, standard deviation values
are less than 2% of the mean value, thereby indicating
reproducibility. As for the mean values of porosity, note that
those of aggregates No. 602 and 603 are close (within 10%) to those
quoted by the suppliers while those of No. 601 and 602 deviate by
as much as 20% from the quoted values. This relatively large
difference in porosity values has not been explained.
Table 3 presents the various experiments performed as well as
the values of k and b obtained by regression. The third column
presents values of (d*/#) which are needed to estimate the Reynolds
number which gives an indication of the nature of the flow regime
(whether laminar or turbulent) (Ref. 4). The value of gravel
nominal size was assumed to be the diameter and was taken from
the supplier's specifications (simply the mean value), while the
corresponding values of porosity 0 were taken from the experiments
(Table 2). Two or three samples of each aggregate mix were tested
and 10-20 runs were performed for each mix. This duplication was
to ensure that experimental results obtained were representative
and robust.
The range of air flow rates at which the experiments were
performed for each aggregate and the corresponding range of
resulting Reynolds numbers are also shown in Table 3. Generally,
the flow is turbulent when Reynolds numbers are greater than about
10 (Ref. 4). In the experiments, the Reynolds numbers were higher
than 10 but less than 50, a range of Reynold numbers which is
expected for SSD air flow in gravel beds of residential buildings
(Refs. l and 2). A further advantage of the Reynolds number range
chosen is that the statistical determination of the parameters k
and b is likely to be more accurate. Other than for aggregate No.
603, the R2 values of the regression model are very good (R2 > 0.9) .
This is an indirect indication that the experimental design was
satisfactory. The relatively lower R2 values for mix No. 603, which
is the largest gravel, could be a result of the fact that, in order
to keep Reynolds numbers low, the pressure drop was measured close
to the sensitivity of the instruments (which was 0.25 Pa or 0.001
in. wc).
-------
The mean and the coefficient of variation (CV) in percentage
values for k and b are also given in Table 3. Permeability values
of mix No. 600 are around 10"7 m2, gradually increasing to about 7
x 107 m3 for mix No. 603. It is worth pointing out that mixes No.
601 and 602 (1/2 and 3/4 in., respectively) have k values which are
10-20 times larger than those of river-run gravel of the same
nominal diameter, as reported in a previous study (Refs. 1 and 2).
A possible explanation for this is that river-run gravel beds have
lower porosity; i.e., they tend to pack more closely (Refs. 1 and
2? This important finding suggests that crushed aggregate is more
suitable as a subslab material fill than river-run gravel for
buildinqs to be mitigated using the SSD system since the former is
likely to have a larger pressure field extension from the suction
hole The flow exponent b does not vary too much with mix. The
relatively higher value for mix No. 602 does seem surprising since
both No. 601 and 603 have lower values for b A possible reason
?or the low b value for mix No. 603 is that it had relatively more
fines in the aggregate. CV values for k are low (less than 3%),
while those forb are high, as much as 12% for mix No. 603.
PRESSURE FIELD EXTENSION
This section gives results of computing the pressure field
extension under the slab when the four crushed gravel sizes are
used as subslab fill material, using the disc flow model described
in Ref. 1. Consequently, the following conditions are assumed:
only one suction hole is used,
the suction hole is located at the center of the slab,
the slab is circular, and
the edges of the slab communicate uniformly with the
ambient air.
Figure 2 illustrates how flow conditions in an actual house
with a basement can be visualized in the framework of the disc
model. The square basement is approximated as a circle of radius
r0 while the extra flow path H through the soil around the house is
accounted for by effectively increasing the circle radius to R.
Thus the flow is assumed to occur between two impermeable discs,
the upper end being the underside of the basement slab and the
lower end being the soil beneath the gravel bed. Since the
material under the slab is gravel while the extra flow path around
the sides of the house is through soil of different permeability
than that of the gravel, this model was modified to represent a
disc made up of two materials of different permeabilities: soil
between radii r0 and R, and gravel between and the central
mitigation pipe. As pointed out in Ref. 1, the flow regime would
likely be turbulent through the gravel bed and laminar through the
soil.
-------
Consequently "the total pressure drop AP (in head of water)
across the two-material bed is equal to the pressure drop through
the gravel plus the pressure drop through the soil. From equations
derived in Refs. 1 and 2:
Note that the pressure drop given by Equation (1) is not the
entire pressure drop to be overcome by the mitigation fan. The
pressure drop due to entrance effects into the mitigation pipe and
that due to the straight pipe, fittings, and bends could account
for as much as 50* of the entire pressure drop in the mitigation
system.
Two types of construction were studied: slab-on-grade and
basement houses. In the framework of the disc model, the
difference between the two is solely in terms of the extra flow
path through the soil. Basement houses will tend to have'higher H
values (see Fig. 1) than slab-on-grade houses; consequently the (R-
r0) value will be correspondingly different. The following values,
deemed representative, have been chosen in all calculations that
follow:
slab-on-grade: R ¦ r0 + 1 a
basement house: R « r0 + 3 m
where
g
p
q
h
b
r0
r,
K
R
permeability of the gravel bed,
kinematic viscosity,
gravity
density,
total air flow rate,
thickness of the gravel bed,
flow exponent,
radius of the basement
radius of the suction pipe,
permeability of the soil surrounding the house,
total radius of flow (equal to that of the
basement and of the extra flow path through
the soil).
Also selected were r, » 5 cm (i.e., a 4 in. diameter suction
-------
. . ^ »nd a value of h = 0.05 m (which is
pipe for the mitigation, y }orted by -tests in an actual house
based on experience an PPion 1 cannot be directly used to
described in Ref. 1)- pressure for different values of
compute the required su ^.otai air flow rate q is not an
basement radius r since. the practical criterion,
independent variable^ ^ actual siab (i.e., up to a
which is that the pre ambient pressure P, by an amount
radius r0) should be lowe ^ house arising from natural
larger than the . effect, heating, ventilation, and air
causes (e.g., wind, depressurization). A realistic range
conditioning system (HyA ) P (Ref. 5). Consequently, the
for this depressurization 16 3-10 Pa 1*°*^ drop #<£b1
flow rate q should be outer portion of the disc containing
ro?"Sret"l«we0erRrSandn £ T^TS. «<»* i. la.inar in this
region, q is computed from:
where AP is the prespecified .iniaui. depressurization below
all points of the slab expressed in head of water.
The value of q is then used in Equation (1) to calculate the
The vaiue 01 ^ t different values of slab radius rc.
required suction Pr . , value of k, soil permeability, greatly
Note that the ]?ume th pressure field extension. Hence
influences q and conse^ently^theprj^ ^ ^ ^
sensitivity of the pressure field extension on the type of boundary
fill material (Ref. 6):
(i) k, = 10-8 mJ, corresponding to sand and gravel mixtures,
(ii) k, ¦ 10,0 corresponding to fine sands.
¦ • H,, > value of 10 Pa has been selected as the
requiredpressure'drop through the soil between radii R and r„ due
to reasons discussed above.
«• i t oresent the computed suction pressures for
Figures 3 and 4 present ^ ^ ^ ^ tMt0d
cases (i) and (li), * P ab-on-grade type construction. Note
and for both b^ment and slab^onj ^JTP ^ ^
that differences in c<#lly mixes Ho. 602 and 603, are much
subslab bed material, P^ resulting fr0m soil type selection,
less important t there iS an order of magnitude difference
Figures 3 and 4 ®*V°* between the two construction types; it is
in "7""Y0"er%e "eability .oil. This fact highlights
lower for case (11) , P low perBeability soil around the
the importance of ha v.i g phery of the building. There is an
foundation walls and the perxpneiy
(2)
-------
important difference in suction pressures between basement and
slab-on-grade construction for case (i) , but this is not so for the
tighter soil (fig. 4). This is because most of the pressure drop
occurs across the outer ring of soil, with practically no drop
through the gravel bed itself. Thus theoretically, for fine soils
(case ii), the pressure field extension is very large (radius
greater than 50 m) . However, in practice, looseness in packing of
the soil, short-circuiting of flow paths, or small holes or cracks
in the slab will drastically decrease this value.
Finally, Fig. 5 presents the total air flow rate for case (i)
for slab-on-grade and basement construction. These flows are
independent of the gravel size since they have been computed from
Equation (2) following the condition that the pressure drop in the
outer ring consisting of soil is equal to a prespecified minimum
amount taken as 10 Pa in this study. Note that the air flow rates
between the slab-on-grade and basement cases differs by almost a
factor of three.
FUTURE WORK
This study was undertaken to experimentally determine the
aerodynamic flow coefficients (permeability and flow exponent) of
four crushed aggregate mixes of commercially available stone. The
sizes, ranging from 3/8 to 1 in. nominal diameter, are gravel sizes
often used as subslab fill for residences and large buildings. An
important finding of this study is that permeability values of 1/2
and 3/4 in. crushed gravel were 10-20 times higher that those
reported in a previous study for river-run gravel of the same
nominal diameter. Field validation of the computed pressure field
extension using laboratory test results is important and such tests
are currently being planned in newly constructed schools,
commerical buildings, and houses. The results presented here
should be used with caution until such time that the apparatus and
experimental design are more fully validated.
ACKNOWLEDGEMENTS
D. Harrje and R. de Silva contributed generously during the
design and construction of the laboratory column. Useful
discussions with A. Cavallo and R. Sextro are also acknowledged.
-------
NOMENCLATURE
b
flow exponent
dv
equivalent diameter of gravel
cv
coefficient of variation
g
acceleration due to gravity
h
thickness of the porous bed
k
permeability of gravel
AL
length of porous bed in flow direction
AP
pressure drop
P
pressure
q
total volume air flow rate
R3
coefficient of determination of regression
Re
Reynolds number
R
total radius of air flow path
r0
effective radius of basement
r,
radius of suction pipe
P
density
V
kinematic viscosity
-------
REFERENCES
1. T.A. Reddy, K.J. Gadsby, H.E. Black III, D.T. Harrje, and R.G.
Sextro, "Simplified Modeling of Air flow Dynamics in SSD Radon
Mitigation Systems for Residences with Gravel Beds," report
submitted to U.S. EPA, February 1990.
2. K.J. Gadsby, T.A. Reddy, R. de Silva, and D.T. Harrje, "A
Simplified Modeling Approach and Field Verification of Airflow
Dynamics in SSD Mitigation Systems," presented at the 1990
International Symposium on Radon and Radon Reduction
Technology, February 19-24, Atlanta, 1990.
3. T. A. Reddy, K.J. Gadsby, D.F. Anderson, and R. Gafgen,
"Experimental Laboratory Tests on Vulcan Crushed Aggregate,"
Report submitted to AEERL, U.S. EPA by the Center for Energy
and Environmental Studies, Princeton University, in January
1991.
4. M. Muskat, The Flow of Homogeneous Fluids Through Porous
Media. McGraw-Hill, 1937.
5. EPA Manual: Reducing Radon in Structures, 2nd Edition, United
States Environmental Protection Agency, Office of Radiation
Programs, Washington, D.C., 1989.
6. W.W. Nazaroff, B.A. Moed, and R.G. Sextro, "Soil as a Source
of Indoor Radon: Generation, Migration, and Entry," Chapter 2,
Radon and Its Decay Products in Indoor Air. W.W. Nazaroff and
A.V. Nero (Eds.), John Wiley & Sons, 1988.
-------
u
~J
m
H
TJ 3
o o
> >
in
±j
» .* a
8^ c
3 4)
J CO D
CN
m
e
o
£
-H
o
¦
tl
c°
«
u
H •
3 0
> Z
n
ST
in
v©
•
•
•
•
O
00
IP
GO
»H
o
IT)
r>
00
n
CM
O
o
o
r*
CN
m
CD
o
o
M
r*
0)
©
c
iH
0
U
u
u
•
•
•
•
u
c
c
c
«*4
•H
•
c
00
*»
PO
r>
H
©
H
CM
ro
o
o
O
O
V0
vX)
-------
TABLE 2. COMPARATIVE RESULTS OF POROSITY TESTS
Porosity
Vulcan Nominal From Present
No. Size (in.) Supplier Tests
600
3/8
0.493
0.403 (0.006)*
601
1/2
0. 506
0.419 (0.011)
602
3/4
0.475
0.436 (0.015)
603
1
0.486
0.452 (0.009)
~Values in parentheses are standard deviation values.
-------
r>
W
.J
CD
<
+J
c
0)
I*
X
Id
I »
<8 >*0
41 4J H
f!2 X,
Q)
0, XJ ^
c
0
-rH
ffi
D ^
0) Q)
b."0
p1 0 ~
41 X «
cc
I
w
0) CD
U CD C
01 J2 O
X) o
£ +>
p *w m
Z 0 >
H-f tt
C "U ©
^ u
0) 0 0)
cr c a
C > g
•
EMC
0 -H
Z W —
c
«
u
*¦4 •
9 0
> Z
o
^ ~
• r*
Ch
o
o
CN
I
IT)
m in
»—I •
• n
9>
&
O
ro
O
CN
GO
CN
m
o
(N
ff\
^ •
(N O
CN
CN
CN
CO
CN
I
r-
ro
rH
o
\C
*
n
CN
o
v£>
O
CM
*
VC
a
rn
a
^ —
~ in
•
**
•
CN
• CN
^ •
cn
•
vO
»
m •
O CN
O
n
r-i
00 «H
m
wo ^
«*
CD
V£)
CD
in
n
o
n
i
CN
VO
0
u
0
W«
0)
•o
u
A3
*0
c
m
4J
w
a>
3
H
10
>
c
m
a
2C
0)
Li
a>
x:
CD
0)
P
«0
>
c
o
4J
to
u
10
>
4J
c
0)
4)
0
u
0)
u
«
CD
01
B
4)
£
+>
c o
0) o
hi *H
4
a x
c ~
c
«
5 E
9
«H 01
« £
> +>
r>
O
-------
Figure !. Sketch of the flow apparatus. I i?t of components is attached
-------
Key for Figure 1
(1) Main Column - Schedule 40, PVC, 8 in. Pipe
(2) Test Rig Panel
(3) Support Blocks
(4) Test Material Cavity
(5) Lower Sleeve
(6) Movable 1/4 in. Aluminum Screen
(7) Upper Sleeve
(8) Collection Bin
(9) Assembly Bolts
10) Short Pipe Section
11) End Cap
12) "O" Ring
13) Flow Straighteners
14) Pressure Taps
15) Urethan Tubing
16) Tubing Union
17) Miniature Solenoid Valves
18) Solenoid Panel
19) Valve Selector Switch
20) 12 Volt DC Power Supply
21) Manifold
22) Manifold Panel
23) Transducer Selector Switches
24) Pressure Gauges
25) Interchangeable Connector Tubing
26) Gauge and Flowmeter Panel
27) High Pressure Transducer
28) Low Pressure Transducer
29) DVM Connector
30) Digital Voltmeter (DVM)
31) Wiring Harness
32) Air Supply
33) Shutoff Valve
34) Flow Control Valve
35) Flowmeters
36) Tubing from Flowmeter to Column
-------
Mitigation
/pipe
as e
Figure 2. Disc node! of a SSD mitigation system in a basement house.
-------
SLA3-0NI-GRADE
280-i
D-|—i—i—i—i—|—i—i—i—i—]—i—i—i—i—|——i—1—i—'—| i i—i—i—
0 10 20 30 <3 50
SLAB RADUS (m)
EASEMENT
o
0.
c:
3
m
in
(j
£
V600
V601
V602
V603
BASEMENT RADIUS (m)
50
Figure 3.
Pressure field extension for case (i) : - 10"'m' for
the four gravel mixes tested. Soil flow path length is
1 m for slab-on-grade and 3 m for basement houses.
-------
SLAB-ON-GRADE
o
CL
in
tn
LU
2
C
H-
y
U1
o
LlJ
&
1M-
112-
11-
108-
J05-
KH-
102-
V600
V601
V602
V603
n==F=r—i—i—i—i—i—i—i—i—i—i—r
10 20 30
SLAB RADIUS (m)
KL5-
BASEMENT
o
CL
00
If)
hi
101-
103-
102-
V600
V601
BASEMENT RADIUS (m)
50
Figure 4.Pressure field extension for case (ii): k, « 10'" m' for
the four gravel mixes tested. Soil flow path length is
1 m for slab-on-grade and 3 m for basement houses.
-------
SLAB-ON-GRADE
70-
6DH
50H
UJ
30H
ioH
c
"'"S
SLAB RADIUS (m)
T
T
T
T
T
BASEMENT
0—) 1—i 1—i 1—i 1—i~~i I i i i—i—| i—r—i 1—| r
0 10 20 30 40
BASEMENT RADIUS (m)
figure 5. Total air flow rates for case (i): k. - 10- mJ, computed from
-------
VI-8
A RADIOLOGICAL STUDY OF THE GREEK RADON SPAS
by: P. Kritidis
Environmental Radioactivity Laboratory
Institute of Nuclear Technology - Radiation Protection
15310 Aghia Paraskevi, Athens, Greece
ABSTRACT
A number of balneological units located in four regions of Greece and
using thermal spring waters of high 222Rn concentrations (0.1-6 MBq m-3) have
been investigated. The concentrations of 222Rn and 226Ra in the water used, as
well as those of the short-lived decay products of 222Rn (RnD) in the indoor
air have been determined. The annual doses to the personnel and the patients
have been evaluated. The results are discussed in the frame of the
Justification and ALARA principles. The main problem concerns the absence of
scientific statements relating the benefit from the procedures applied (or
certain part of it) to the exposure to 222Rn and RnD.
Keywords: natural radioactivity, radon spas, balneotherapy, inhalation doses
INTRODUCTION
The water is known to be, generally, a minor source of indoor radon. This
is not always the case in the balneological (water-physiotherapy) units using
thermal spring waters. The balneotherapy is an ancient (Greek, Roman) practice
which has survived until our days in many parts of Europe. After the discovery
of radioactivity, it was found that certain spa waters are characterised by
high concentrations of 222Rn. In Greece, the first studies of this kind were
carried out by Pertesis during the 20s and 30s (1-3). In four regions of
thermal springs the concentrations of 222Rn in the water exceeded 100 kBq m~3
(originally expressed in the old Mache units, where 1 Mache * 13.5 kBq m~3).
These regions are shown in Fig.l. The maximum radon concentrations were
measured in the Artemis and Apollon springs of the Ikaria island and equaled
10 and 7.5 MBq m-3 respectively. The major radon spas have been investigated
later by other authors (A,5) and the early values of Pertesis have been,
generally, confirmed. Nevertheless, we could not find, during 1983, any study
with radiation protection goals, i.e. study considering the critical pathway
of man's exposure - the inhalation of RnD related to the radon released from
the thermal waters. During the period 1984-88 we have visited, repeatedly the
balneological premises in the 4 spa regions mentioned, measuring the
concentrations of 222Rn and 226R® in the waters used, the concentrations of
RnD in the air of the main and auxiliary indoor spaces, collecting information
about the existing practices, occupancy factors etc. The results of these
studies, together with the related dose estimations, are given in the present
paper.
-------
THE GREEK RADON SPAS: GENERAL DATA
The major Greek radon spas are located in four regions (Fig.l): Ikaria
island in the Eastern Aegean Sea (9 springs, 3 units), Kamena Vourla (1 common
source and 4 units) and Edypsos (5 springs and 5 units) in the Northern
Evoikos gulf and Loutraki in the Eastern Korinthiakos gulf (5 springs and 5
units).
In all cases the spas are located right on the coast and, also, in the
mild climate central latitudes of Greece. This allowed their establishing,
throughout the centuries, as attractive places of the so-called "therapeutical
tourism", where certain physiotherapy practices are combined with usual by-sea
vacation, swimming, fishing etc. In most cases, the high temperature of the
water allows its use for bathing procedures, typically in separate individual
baths, but also in common pools (Kamena Vourla). In one case the water has to
be warmed before use. The water is collected, usually, in some basic
reservoirs and directed further to the points of use. This results in certain
losses of radon gas, of the order of 10-30 %.
During the typical individual bath procedure the patient enters a tub
filled with warm spa water (or mixture with cold spa water) and spends about
20 min in it. The total time spent in the bath, including undressing and
dressing, is about 30 min. The personnel has to prepare the bath (clean and
fill the tub with new water) and often, also, to help the patients before
and/or during the procedure. In most cases there are at least 5 baths per
member of the personnel and their in-the-bath occupancy exceeds 50%.
There are various architecture designs of bath premises. In some cases
the baths are isolated, from the ventilation point of view, from the common
spaces (corridors, halls), but communicate one with another through big holes
(0.5 to 4 m) in the top parts of the common walls. This construction leads to
higher ventilation of the bath, but also to dependence of the average radon
concentration in the bath air on the current working occupancy of the whole
premise. In other cases the baths are totally isolated one from another, but
there is some air exchange with the common spaces. This leads to higher
average concentrations of radon in the bath air, but also in the air of the
common spaces, especially under "stack effect" conditions in the building.
The significant dispersion of the water during the tub filling, its
enhanced temperature and the duration of the procedure lead to the release of
practically 100X of the radon in the air. The use of 1 m3 spa water per
procedure in a bath of 25 m3 air volume will result in a transfer coefficient
of 0.04 between the water and air concentrations of radon under the absence of
ventilation. In the extreme case of the 6 MBq m-3 Apollon spring water it
would correspond to average concentration of 222Rn in "the bath air equal to
240 kBq m~3.
The drinking therapy is another use of certain spa waters. In this case
the patients exposure to inhaled radon is significantly lower, with the
exception of cases like LU3, where certain patients combine the drinking with
a sort of inhalation therapy, by staying 10-15 min inside the hall built
around the twin thermal spring. In all cases, the intake of 226Ra is a pathway
n be considered as well. The water consumed per season is about 20 1.
-------
INSTRUMENTS AND METHODS
The concentrations of 222Rn and 226Ra in the spa waters have been
determined by use of the total alpha-counting method (Lucas cell instrument)
in a closed loop air circulation system (6). The glass sampler is designed to
be used for radon extraction as well, which avoids any sample transfer and the
related losses. The concentration of 222Rn is measured soon after sampling to
ensure maximum counting statistics. Temperature and volume corrections are
applied. After the "in situ" determination of 222Rn, the 0.25 1 water sample
is fully de-emanated and sealed for about 30 days, to allow the radioactive
equilibrium between 226Ra and 222Rn. Then a second measurement of 222Rn is
made to determine 226Ra. The LLD (2o of bckg) of the method is 25 Bq m-3,
which is quite adequate for 222Rn, and acceptable for 226Ra (where it
corresponds to 6.3 pSv per year for a standard 0.8 n)3 water consumption and to
0.16 (aSv for the 20 1 average patient consumption of spa water per season.
The concentrations of RnD have been determined separately by use of an
express variant of the 3-interval total alpha-counting filter method (6). The
time of air sampling is 1 min and the counting intervals - (1-5), (6-10) and
(11-15) min after the end of sampling. The sampling flow rate is 100 1 min-i
and the active filter area - 7 cm2. The LLD (2o of bckg) is 30 Bq m~3
equilibrium equivalent concentration (EEC) of 222Rn - a value adequate for the
levels observed in the radon spas studied. It was also possible to determine a
"quasi-equilibrium factor", as the ratio of EEC to the concentration of 2i8Po.
RESULTS AND DISCUSSION
The concentrations of 222Rn and 226Ra in the waters examined, as well as
the range of concentrations of RnD measured in the air of the premises using
these waters (if any) are given in Table 1. Note that the water sampling has
been done at the point of use (if any) and not at the source spring.
Therefore, in most cases the values for 222Rn are slightly lower than those in
the source point. The waters of the Ikaria springs 13-17 are not used for any
curative procedures indoors. The concentrations of RnD in the outdoor air
close to these springs are measurable, but insignificant from the radiological
point of view, so they are not given in Table 1. Note also, that only the
waters of the units L2 and L4 are used for drinking therapy.
The balneological premises in Kamena Vourla use a common water reservoir,
while the other premises have separate reservoirs, in certain cases - even
two, one for hot water and a second, where the water is cooled before use.
The maximum concentrations of RnD in air have been measured, in all
cases, at the place of the water use (bath, pool etc.). The minimum values
have been measured at the auxiliary spaces used by the personnel and/or the
patients (during waiting). The occupancy factors of the various spaces for the
patients and the personnel have been estimated both by observations and from
the information provided by the staff.
The "maximum transfer factor" is the ratio of the maximum RnD
concentration measured in air to the 222Rn concentration in the water used. It
can be seen that these factors vary significantly from premise to premise by
-------
more than two orders of magnitude: minimum value 0.00032, maximum value 0.085
and average value 0.012. One must note that both extreme values have been
measured in the air of personal bathrooms. It is also interesting, that the
highest values of RnD in air (K3, K4) are related to water concentrations of
222Rn significantly lower that the maximum observed (II, 12). We do not
consider here the case I1R, which is a reservoir of radon water, not visited
by patients. It is interesting, nevertheless, to note that certain
individuals, attracted by the idea of the "primary curative source", have been
seen to apply a sort of "private inhalation therapy" in the entry of the
reservoir, inhaling several minutes its 120 kBq in-3 RnD air!
TABLE 1.
CONCENTRATIONS OF NATURAL RADIONUCLIDES
MEASURED
IN THE WATER i
AND THE
AIR OF VARIOUS
GREEK RADON SPAS
(Bq m-3).
i
Region,
premise
222Rn in
226Ra in
RnD in
Maximum i
water
water
air
transfer factor i
LI
450000
18-400
i
0.0009 j
L2
175000
103
20-200
0.0011 i
Loutraki
L3
170000
150-3500
0.021 |
L4
140000
55
180-440
0.0031 i
L5
90000
15-40
0.0004 i
Kamena
K1
850000
1500
20-3700
i
0.0044 i
Vourla
K2
ii
it
90-1400
0.0016 |
K3
H
ii
110-15000
0.018 j
K4
II
ii
11-18000
0.021 j
Ikaria
11
5700000
950 1300-6900
i
0.0012 i
I1R
5700000
950
120000
0.021 i
12
5700000
1200
500-1800
0.00032 i
13
3000000
200
-
|
14
625000
360
-
j
15
480000
200
-
16
270000
3400
-
~ j
17
200000
4700
-
_ j
18
160000
3500
70
0.00044 i
19
<100
3400
<20
i
Edypsos
El
200000
3400
150-1750
0.0088 i
E2
72000
1600
200
0.0028 i
E3
10000
1100
400-850
0.085 i
E4
9300
2700
30-130
0.014 j
E5
1000
5000
<20
i
l
—— i
The differences in the transfer factors observed reflect, mainly, the
variability of ventilation rates, but also the differences in the water
temperatures and the "typical" dispersion of the water during its use This
variability leads to insignificant correlation between the water and air
concentrations of radon (CC=0.2). It is interesting to note, that an weak
negative correlation (CC=-0.3) is observed between the concentrations of 226Ra
-------
and 222Rn in the waters examined. The discussion of this point is beyond the
scopes of the present study.
The estimations of the annual effective dose equivalents (EDE) for the
personnel and the patients of the radon therapy centers examined are given in
Table 2. The occupancy of the personnel is 5 months per season and, in most
cases, 50?, of the working time is supposed to be spent in the areas of highest
radon concentrations in air. The patients are advised, typically, for 20 bath
procedures of 30 min each per season. The intake of radon water for the
drinking therapy patients is, typically, 20 1 per season.
""
TABLE 2.
ESTIMATIONS OF ANNUAL EFFECTIVE DOSE EQUIVALENTS FOR THE
PERSONNEL
AND THE PATIENTS OF
VARIOUS GREEK RADON
SPAS (mSv).
Region,
premise
Personnel
Patients
inhalation
inhalation
ingestion
1
»
of RnD
of RnD
Of 226Ra
LI
1.9
0.05
_
L2
0.9
0.01
0.001
Loutraki
L3
16
0.4
-
L4
4.0
0.03
0.0005
L5
0.2
0.005
-
Kamena
K1
17
0.4
(0.01)+
Vourla
K2
6.6
0.15
It
K3
70
1.7
II
K4
85
2.0
II
Ikaria
11
32
0.8
(0.006)
i
I1R
-
(1-3)*
-
I
12
8.5
0.2
(0.008)
i
18
0.35
0.01
(0.02)
19
0.1
0.002
(0.02)
Edypsos
El
8.2
0.2
(0.02)
E2
1.0
0.02
(0.01)
E3
4.0
0.09
(0.01)
E4
0.6
0.015
(0.015)
i
i
E5
0.1
0.002
(0.03)
* Based on 3 min per day "private inhalation therapy" (see text).
+ The values in parentheses are hypothetical. These waters are not used for
drinking therapy.
It can be seen, that the estimated yearly EDE for the personnel exceed,
in 2 cases, the current 50 mSv a~* dose limit, in 5 cases - 30% of this limit
(controlled area conditions) and in other 4 cases lay between 10% and 30% of
this limit (supervised area conditions). If we apply the new 0.04 per Sv fatal
cancer risk coefficient recommended by ICRP (7), then risks of the order of
10% could be attributed to the personnel of the units K3 and K4 after 30 years
of work and risks between 1% and 5% - in other 6 cases.
-------
The dose estimations for the patients vary 3 orders of magnitude, with
maximum values not exceeding 2 mSv a~i. These doses are mainly due to the
inhalation of RnD, while the ingestion of water (where applied) has an
insignificant contribution.
CONCLUSIONS
Two different radiation protection problems arise from the results
presented above.
1. The members of the personnel of certain radon therapy premises are
exposed to doses which not only exceed the 102, and/or 30X of the current dose
limit, but, in 2 cases, the limit itself. Nevertheless, these working areas
are not classified as supervised or controlled and no radiation protection
measures are applied. We have, in all cases, a violation of the Justification
and ALARA principles and in 2 cases - also of the Dose Limitation principle.
2. The patients are exposed to doses similar to those reported for other
cases of medical use of ionising radiations. Nevertheless, we could not find
any scientific material dealing with the dose-benefit relation of the radon
therapy procedures and, therefor, no risk-to-benefit analysis and no
optimisation can be applied.
It seems necessary to draw further the attention of the international
radiation protection community to the problem of the radon balneology
practice, in order to achieve, gradually, the conformity of this practice
(whatever it could mean) with the basic radiation protection principles.
REFERENCES
1. Pertesis, M. About the radioactivity of the Kamena Vourla spas. Geological
Service of Greece: No 16, Athens, 1926.
2. Pertesis, M. The Greek mineral springs. Geological Service of Greece: No 24
Athens, 1937.
3. Pertesis, M. About the radioactive hot springs of Ikaria island.
In: Proceedings of the Academy of Athens No 14, 1939.
4. Ioakimoglou, G., Georgalas, G. and Fokas, E. Analyses of the therapeutical
spas of Kamena Vourla. In: Radiosprings Kamena Vourla, Athens, 1940.
5. Gioni-Stavropoulou, G. Hydrological and hydrogeological studies.
Publication of the Institute of Geological and Mining Research No 39, 1983.
6. Kritidis, P. and Angelou, P. Investigation of the concentration of 222Rn
and its daughter products in the Loutraki spas. Greek Atomic Energy
Commission report DEMO 84/7G, 1984.
7. ICRP/90/G-01, Recommendations of the Commission - 1990. Draft, Feb. 1990.
-------
BULShRIh
JUGOSLAVIA
J jvx^\N\\\vN\
^^3
VA\\r
•»V0\ \_sis
ALBANIA
^v:>
•jWAV'
A S V . V\\
V\V\%
«VWi*i
TURKEY
t-;
. |V«> s s s
L\1 * «•- ¦. J
Ionian '¦¦VA
*»* K.Vourla^Oi® I "
«t i?w
as
Sea
# 4 D
J5
t> ,«$
as- „~*'
-S
0*
&
<$¦
Cretian Sea
Mediterranean Sea
4
Fig.1. Locations of the regions of the major Greek radon spas.
-------
Session VI:
Radon Surveys - POSTERS
-------
VIP-1
TITLE: A Cumulative Examination of the State/EPA Radon Survey
AUTHOR: Jeffrey Phillips, EPA - Office of Radiation Programs
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
In an effort to identify the distribution of Radon
"hroughout the country on a statewide basis, a cooperative
;3tate/EPA Radon Survey commenced in 1987. To date, thirty four
states have participated in the program, creating an extensive
characterization of potential Radon distribution. The Survey has
provided valuable information to states attempting to assess this
environmental hazard and deal with it.
This paper focuses on 1) describing the results of the survey
data, 2) the process by which the results were obtained, and
3) an overview of the survey's potential utilization and
implications.
-------
VIP-2
SEASONAL VARIATION IN 2-DAY SCREENING MEASUREMENTS OF 222RN
by: Nat F. Rodman, Barbara V. Alexander, and S.B. White
Research Triangle Institute,
Research Triangle Park, N.C. 27709
Jeff Phillips and Frank Marcinowski
U.S. Environmental Protection Agency
Office of Radiation Programs
Washington, D.C. 20460
ABSTRACT
222
This study examines Rn data from a group of 234 houses 1n which each
level of every house was tested with a 1-year alpha track detector and the
lowest livable level was tested four times, once during each of four seasons,
with a 2-day charcoal canister. This study focuses on 1) how the seasonal
variation in 2-day screening measurements affects the decision to take further
action, when based on a single 2-day canister test, 2) how season affects
estimates of the annual living area average, and 3) how the average of four
seasonal canister readings compares with the 1-year alpha track measurement
taken on the same floor.
This paper has been reviewed 1n accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
-------
INTRODUCTION
Short-term screening tests for radon are used to determine if action
should be taken to reduce radon levels or if additional testing (usually of
one year duration) 1s needed to more accurately characterize health risks to
radon exposures. Radon concentrations vary seasonally (1) and so the result
of the short-term screening test will be, in part, a function of the season in
which the test is made.
The U.S. Environmental Protection Agency provides assistance to states in
conducting surveys of indoor radon. A subsample of houses in states beginnina
their indoor radon surveys during the early part of 1989 were the first to use
charcoal canisters 1n each of four seasons along with 1-year alpha track
detectors (ATDs). These data were used to determine 1) how season affects a
decision to take further action, 2) how season affects estimates of the annual
living area average (ALAA), and 3) how the average of four seasonal canister
measurements compares with the 1-year ATD measurement on the same floor.
METHODS
Homeowners were given alpha track detectors for a 1-year deployment and
four open-faced charcoal canisters over the course of 12 months. One ATD was
placed on each livable level up to a maximum of 4 levels, except for single
story homes which received two ATDs. One charcoal canister was exposed on the
lowest livable level during each of the four seasons. The homeowners were
instructed to use the same location for all four canisters.
For this analysis, seasons are defined as:
Winter: December 1 - March 31
Spring: April 1 - June 30
Summer: July 1 - September 30
Fall: October 1 - November 30.
Houses were required to have: 1) one canister measurement within each season
2) all canister measurements separated from each other by at least 30 days 3)
floor codes available for at least three of the four canister measurements'
(I.e., one missing floor code was permitted), 4) all available canister floor
codes must agree, and 5) canister floor codes must match the lowest ATD floor
code.
All houses were on a permanent foundation and had at least one floor at
or below ground level. Analyses were performed separately for basement and
nonbasement houses. A basement house 1s defined as any house where the lowest
livable level has at least one wall built against earth. All other houses in
the study were defined to be nonbasement houses.
-------
RESULTS AND DISCUSSION
DISTRIBUTION OF SCREENING MEASUREMENTS BY SEASON
A total of 234 houses met the above criteria, 162 basement houses and 72
nonbasement houses. These houses were located in Iowa, Maine, Ohio, Vermont,
and West Virginia. Table 1 displays the arithmetic mean, standard deviation,
geometric mean, and geometric standard deviation in each of the four seasons,
by floor and overall. As reflected by the parameter estimates in Table 1, the
distribution of measurements are essentially the same in all seasons except
summer. Summer measurements tend to be lower than the other season
measurements; the geometric mean was about 15% lower for basement measurements
and about 50% lower for first floor measurements 1n nonbasement houses.
Closed-house conditions may not have been maintained during the summer.
VARIATION IN SEASONAL SCREENING MEASUREMENTS
A coefficient of variation (CV) was calculated for each of the 234 houses
(CV = standard deviation of the four seasonal measurements divided by the mean
of the four measurements, expressed as a percentage). The results are
summarized 1n Table 2. As noted 1n Table 2, 12 houses were excluded because
the mean of the four measurements was less than 0.5 pC1/L and CVs become
unstable whenever the denominator of the CV statistic approaches zero.
The CVs appear to be about the same regardless of the concentration
level, e.g., houses with means of the four seasonal measurements near 2 pCi/L
exhibit about the same percentage variation as houses with means of the four
measurements near 15 pCi/L. The average CV for all 222 houses was 40%.
SEASONAL AFFECT ON DECISIONS FOR FURTHER ACTION
Using 4 pC1/L as the point at which some further action 1s taken,
classifying each seasonal measurement as S4 pC1/L or >4 pCi/L results in 16
possible patterns of outcome. Two of these patterns give clear decisions:
when all four measurements are £4 pC1/L or all four measurements are >4 pCI/L.
The other 14 patterns give conflicting decisions and are tabulated in Table 3.
Of 234 houses, 66 (28%) had conflicting decisions. Thirty-two (48%) of the 66
houses fell Into pattern #5 where the winter measurement Indicated taking
action but the other 3 measurements Indicated no action was necessary.
Thirteen (20%) of the 66 houses fell into pattern #3 where the fall, winter,
and spring measurements indicated taking further action but the summer
measurement did not. In 57 (86%) of the 66 houses, the winter and summer
measurements disagreed; however, 1n 5 of these, summer indicated taking action
while winter Indicated no action was necessary. These 5 houses were all
basement houses, 3 1n Iowa and 2 1n Maine; all were 2-story (plus basement)
except one Iowa house which was 1-story (plus basement).
SEASONAL AFFECT ON ESTIMATING ALAA
The following model (2) was used to estimate the relationship between 2-
day screening tests and ALAA (annual living area average):
-------
>
o
r— Z
ro •
U.
U
O
lo oo o un
in ^ c\j
ro »-H m o
c\j ro «h ro
UD CO O
• • • •
•3- *3- c\j co
+-»
CO
s- a.
c
CD CD
o
E m
CO
& i
3 >
CD
0O r—
CO
3
t-
3
O
Ll.
M—
CD
O
CTJ C
C 3
-C
•r- ^
u
S- 1
03
a. •
LU
CO t-
o.
C
<
•t-
¦a
CD
4->
00
•
CD
S~
h-
S- ftf
1
0)
c •
CO
•»- u
3
3 CD
o
O
n:
t-
o
M-
+
5-
co
CD
u
+-»
• r-
fl3
A3
Q.
¦+->
>>
M— +->
u
O C
A3
3
<0 m
u a>
(—
O
Z 3 00
O 0J
I h-
o
a>
a> to
a. 3
>. o
O I— MN
IT) «3" CO C\J
Q
Z Q Z OO
< OO CJ3 O
e
0)
<0
CQ
esj
c
0)
e
a>
i/>
e
a>
CO
CM C7>
• • • •
CSJ CSJ «—I CSJ
00 ro oo «-•
• • • •
LO CSJ CO
co co rx ro
• • • •
^ CSJ CO
_ Q O
s o z: <-o zqsoo
< (/) i3 O C 00 CU CJ
0)
>
a>
uo
<1)
TJ
c
T3
o
t-
•r—
nj
C 4->
X3
aj +->
0)
B
U T3
•r—
u o
+J TO
*r-» "p—
a> s_
t- 4-
E +->
XI T3
c
e e
•t- ra
o o
i- +j
aj a)
O)
m ii
n ii
o
^ GO CD U
*
-------
>
o
£
3
£
X
fd
• r—
>
"O
o
c
»—t
e
s-
£
o
•r—
M-
c
•r—
r—
z:
0)
>
a>
_i
to
>
c
CJ
o
•r-
<+-
4-»
o
<0 CO
s- c
c
-M O
o
C LO
• r—
a) ro
•M
u a>
<0
c
•r—
o
>
o u
QJ
3
a
c o
O U-
T3
-a
i-
f0 M—
<0
a: o
"CJ
C
to jC
<0
3 U
4->
l/> *0
<~>
U UJ
o
> c
-H
> -a
>
o o
o
W +J
to
c
C 0)
• r—
z:
4-> CO
QJ
•r— CO
S- 3
to
> :r
to
3
H-
O
o
3:
+¦>
#
c
o
0)
z
•r—
o
o
o
a.
c
•
o
CSJ
• r»
+->
0)
1—
t-
.Q
ro
c
H—
O)
u
c
o
o
c
f0
a>
csJcvjomiOK
co m »-h r>*
fO hs N N rs N
coo>o^tnrsin
O CO CSJ o\ O CO
«-H CSJ t-H
N o
CSJ rH i-H pH iH CSJ
O »-t U3 N tN
• •••••
LT> CM LT) >—I CTl O
««¦ *r ro ro n *r
tnco*J-otsoo
f-v «3" CO «3" ¦—I
O O
o o o • •
— • • • o 10
CSJ CM *3" v£> «-H «-i
o
in
in I I I I I
#
CD lO fH rH ^ r-H rH
V O CSJ tD O A
CSJ
LO
CO
o>
CSJ
iO
CT>
cr»
o>
m
CSJ
CSJ
CSJ
I?
u
0>
>
o
¦o
O) 4->
¦o c
•r- a>
> *r-
•f" u
-O
4-
<+-
O
-a 4-
C 0)
>> >*>
-Q -M
10
•r-
•f—
"O r—
00
0) -r-
-r- .Q
r— *o
'io
Q-'I-
c
•r- U
ra
.— >
• r—
e —
A3
+->
c c
M-
0> CO
to
T3
3 0)
a)
o >
13
-fiZ *r-
3
+->
F
c fa
u
0) r-
X
> CD
(U
•r- i.
O)
5
c
O
_l
O
—^
CO
• r—
+j a>
o
CZ
a.
a>
ui a>
V c
o
a) e
c
e
u
i- 10
+J l-
3 a> •
nj +j
O T3 O)
1- c
M- «r- 1/)
-t-> v
> 3
c o
0) o o
0) c
JZ i- SI
u o
+j a.
c o
o
H- C iO
u 5
O Or
o
•r- +J
C r—
C +J
IO
o <0 t-
« +J
•i- J- o
E t>
4-> -M 4-
(0 c
jz to
•r-
> u >
•r- o
0) cu
s
TJ O ^
U
•O <0 -f-
3 +J
•a u
CM XI
¦M
•— l/l
-e >>h-
i— C
1— -Q O
C -r-
i-H
CM
-------
Table 3. Number of Houses With and Patterns of Occurences of
Conflicting Decisions Based on
Four Seasonal Measurements of
Indoor Radon in Each of 234 Houses
Season
Pattern
Winter
Spring
Summer
Fall
No. Houses
1
_
+
+
+
1
2
+
-
+
+
5
3
+
+
-
+
13
4
+
+
+
-
1
5
+
-
_
_
32
6
-
+
-
-
1
7
-
-
+
-
4
8
-
-
-
+
1
9
_
_
+
+
0
10
-
+
-
+
1
11
-
+
+
-
0
12
+
+
-
-
3
13
+
-
+
-
0
14
+
-
-
+
4
66
+ = measurement is greater than 4 pCi/L
- = measurement is equal to or less than 4 pCi/L
-------
r—
<0
Lu
-o
c
CD
CD C
O
5-
0) *3
> O)
C/0
t-
A3 -O
C
U >
•r- h-
>
e
• r— QJ
o
—J
tO
3
3T
CO
3
C >.
C -Q
<
o>
co
e
C +->
a> c
cu a>
Q.
3t E
00
+j q;
i
3K
• ca
¦>
i-
0)
¦M
0)
s
ns
i-
TS
a.
X oo
ro O
io •
r~- cm «—i ro
f en • co oo
ii
O-
o o
X -—.
"3" LD
IT) O
f-» CM
MOWOO
• • Lf> r»v
ii
O-
o o
on ro
ld o
CM 00
Nror-iMO
«d- • • «=r co
o o
X ro
CO O
LT) •
+ -X
II
O-
r-1 «3"
UO .(MOl
lO * s—- • •
O O
II
o-
X
O CO
r*s* i-H
rocvj old
VO • • LO LO
VO 'H • •
o o
II
o-
ro r>-
cm >—l cm oo
*3- • • I—
O o
X
r». o
00 •
CM
Lf) t—i CT> VO
*£> r-» • cm oo
V0 • ¦ . .
o o
II
X >£>
CTi o
•
+ <«—»
CM
ro *—• co O
*£> CTi • CM CTv
l£> . ,
O O
II
0-
c
c
o
o
• r—
•n*
+->
4J
O
•r—
(/)
¦r—
r— +J
c
iSl
•r-
i— +->
3
+-»
f0 <"0
iA S-
-------
where
Yi
*1
Yj1/2 = (a + bXi)1/2 +
ALAA in the ith house, calculated as the mean of ATD measurements,
one from each floor, except for single story houses which had two
measurements taken on the same floor.
^ y*|
2-day charcoal canister measurement in the i house taken in the
basement for basement homes and on the first floor in nonbasement
homes,
a,b = parameters to be estimated, and
X L
e< = random error for the i house, assumed to be normally distributed
2
with mean 0 and variance a .
Eight different models were determined, one for basement houses and one for
nonbasement houses in each of the four seasons. Of the original 234 houses,
21 were deleted from this analysis for the following reasons: 15 due to
missing ATD measurements and 6 due to ATD exposure being less than 325 days.
All remaining houses had ATDs exposed for >325 days and <395 days and all ATD
start dates, within the same house, were within 30 days of each other. The
results are summarized in Table 4. Major conclusions are:
a) The prediction equation, residual error and correlation for basement
houses are basically the same for all seasons. For example, given a
2-day measurement of 4 pCi/L in each season, the predicted annual
11vina area average for the four seasons varies from 3.4 to 3.8
pCi/L.
b) The prediction equation for nonbasement houses is basically the same
for all seasons. For example, given a 2-day measurement of 4 pCi/L
in each season, the predicted annual living area average for the
four seasons varies from 4.0 to 5.4 pCi/L - the summertime equation
gave the highest predicted value.
c) In each of the four seasons, the prediction equation for basement
houses differs significantly from the prediction equation for
nonbasement houses. For example, the coefficient of X for basement
houses varies from 0.53 to 0.63 for the four seasons, whereas the
coefficient of X for nonbasement houses varies from 0.82 to 1.01 for
the four season.
d) The intercepts in all equations shown in Table 4 are significantly
greater than zero, that is, the fitted equation does not pass
through the origin. This would suggest that the ratio model Y/X=b
(or Y=bX) is not an appropriate model for analyzing long-term versus
short-term relationships since it assumes an equation that passes
through the origin.
-------
THE MEAN OF FOUR SEASONAL MEASUREMENTS VS 1-YEAR ATD MEASUREMENTS
Of the original 234 houses with four seasonal screening measurements 217
had 1-year ATD measurements taken on the same floor, with an exposure period
>325 days and <395 days. These houses were used to compare the mean of 4
seasonal measurements with a corresponding 1-year ATD measurement on the same
floor. The results are shown in Table 5.
The 1-year ATD measurements tend to be higher than the mean of the four
canister measurements, an average of .82 pCi/L higher in basement houses and
.77 pCi/L in nonbasement houses. The geometric mean of 1-year ATDs is 14%
higher than for short-term seasonal measurements in basement houses and about
67% higher than for short term measurements taken on the first floor in
nonbasement houses. Both are statistically significant at p=0.01.
CONCLUSIONS
In this sample of 234 houses, 66(28%) had conflicting results 1n one or
more seasons indicating the need to take further action. Thirty-two (48%) of
these 66 had their winter measurement >4 pC1/L while the spring, summer, and
fall measurements were *4 pCi/L; thirteen (20%) of the 66 had fall, winter,
and spring measurements >4 pCi/L while the summer measurement was £4 pCi/L.
Currently, EPA recommends additional testing 1f a screening measurement
exceeds 4 pC1/L. Also, EPA recommends mitigation if an annual measurement
exceeds 4 pC1/L. Houses with a true concentration of around 4 pCi/L will have
the largest chance for error. With a CV of 40%, the standard deviation of
canister measurements for concentrations of 3, 4, and 5 pCi/L will be 1.2,
1.6, and 2.0 respectively.
The prediction equation for estimating ALAA from 2-day screening
measurements does not differ by season. However the equations do differ by
house type, basement versus nonbasement, with the nonbasement houses having
the larger coefficient. The intercepts for all equations are significantly
greater than zero.
The 1-year ATD measurement tended to be higher than the mean of the four
canister measurements. Quality control conducted on blank ATDs indicated a
positive bias due to leaky bags used for storage and mailing.
REFERENCES
1. Ronca-Battista, M. and Magno, P. A comparison of the variability of
different techniques and sampling periods for measuring Rn and its
decay products. Health Physics 55:801-807;1988.
2. White, S.B., Clayton. C.A., and Alexander, B.V. A statistical analysis:
predicting annual Rn concentrations from 2-day screening tests. Paper
presented at The 1990 International Symposium on Radon and Radon
Reduction Technology, Atlanta, Georgia. February 19-23, 1990.
-------
Table 5. Comparison of Annual Concentrations Based
on 1-Year Alpha Track Measurements and Four
Seasonal Charcoal Canister Measurements
Instrument
No. of Four Canister
Type of House Houses Parameter* ATD Measurements
Basement
151
AM
6.52
5.70
SD
5.43
4.48
GM
4.70
4.11
GSD
2.44
2.38
Nonbasement
66
AM
2.74
1.97
SD
2.75
2.40
GM
2.02
1.21
GSD
2.05
2.61
* AM = arithmetic mean
SD = standard deviation
GM = geometric mean
6SD = geometric standard deviation.
-------
VIP-3
THE STATE OF MAINE SCHOOL RADON PROJECT:
PROTOCOLS AND PROCEDURES OF THE TESTING PROGRAM
by: Lee Grodzins, Professor of Physics
Massachusetts Institute of Technology
Founder and Chairman of the Board
NITON Corporation, Bedford, MA 01730
Henry E. Warren, Director
Division of Safety & Environmental Services
Bureau of Public Improvement
Augusta, ME 04333
Ethel G. Romm, President
Chief Executive Officer
NITON Corporation
Bedford MA 01730
A comprehensive radon test was done in every public school in the state of Maine by
NITON Corporation using its liquid scintillation vials. For each school, NITON made up
individual packages containing instruction sheets, test vials, data sheets, and a copy of school
floor plans marked with locations for placing the test vials. Every occupied room on or below
grade was tested over a week-end under closed-building conditions. Quality control procedures
compared NITON vials with independent tests. School personnel set out and retrieved the
tests and returned them to the NITON Laboratory using next day UPS service, a procedure that
worked for even remote one-room school houses, schools on islands, and in Indian reservations.
We will discuss the detailed procedures and close communication that resulted in the successful
and reliable testing of some 14,000 rooms; fewer than 200 tests had to be rerun due to faulty
procedures or late returns.
-------
INTRODUCTION
A comprehensive radon test was done in every public school in the state of Maine by NITON
Corporation using its liquid scintillation vials, with school personnel placing and harvesting the
tests.
The type of test chosen (light-weight, short-term charcoal liquid scintillation) and the use of
non-professionals greatly reduced the costs of testing and made it possible to follow the best EPA
guidelines for testing for radon in schools.
A short-term (week-end) screening test was done in (a) every, (b) frequently occupied (c) room
(d) on or below ground level, under closed building conditions. This follows the recommenda-
tions of EPA's Interim Report in Radon Measurements In Schools (1). The few rooms over crawl
spaces were also done. In addition, Maine uses many trailers for classrooms, some with rather
permanent skirts; after the first few dozen showed no radon in their rooms, no further testing was
done in trailers.
Tests were placed and retrieved by local school personnel. Using school personnel at each site
to do this task so lowered the costs that it became possible for Maine to test every school room in
the state within one year. It was necessary only to design, implement, and manage a protocol that
would make it possible for non-professionals to do the work reliably. A detailed description of the
procedures follows. In general, it follows the outline of another presentation: The Design Study
(2).
IMPLEMENTING THE PROTOCOLS
A school poses far fewer problems of judgment in placing radon tests than a home or
commercial building. The conditions for which professionals require training and much experience
are largely missing in schools. For example, every classroom or school office has a desk, which
is exactly the height EPA calls for, 30". Questions about avoiding open sumps, foundation walls,
stone fireplaces, and such, never come up. No one is interested in cheating—quite the contrary—
so no steps need be taken to prevent or reveal tampering.
The problems of using non-professionals, while onerous, are almost entirely logistical—getting
each stage done, and done on time. The protocol removed the technical burden from the test placer
by having NITON make all the scientific and technical decisions.
NITON had already had substantial experience in helping public schools to test. Although its
products are used almost exclusively by professionals, the company had devised a low-cost system
for a number of Massachusetts public school systems. Having very little money, they would not
have been able to test if they had needed to use professional testers. The program NITON
developed for those Massachusetts schools non-professionals was further refined for Maine.
OPERATIONAL MONITORING
RECEIVING THE PLANS
Roy Nesbitt, Director of School Facilities, sent out a "Form-1" to the superintendent of every
Administrative School Unit. It requested a set of ground floor plan sketches plus information: The
name of each school in the unit, its principal, and the number of "Ground floor Instruction
-------
Spaces." These forms and plans were sent directly to NITON, where each one was marked up to
show where a radon test should be placed. Problems at this stage:
• Many schools did not return their Form-1. New ones were sent with a letter and in many
cases, the schools were called by NITON. If they still did not respond, Nesbitt called. In a few
cases, no form was ever received and the information was taken over the phone.
• The instructions on the Form had described the rooms to be tested as "Instructional Spaces,"
although every office, library, gym, etc. on the ground also had to be counted. Some schools
figured this out and sent expanded figures. The estimates were rarely accurate, nor were they
expected to be. They served to get the schools to think about the radon testing. The count was
ultimately determined by NITON from the plans. (Not surprisingly, in the final results, radon had
made no distinction between classrooms and offices and high levels were found in both types of
areas.)
• When schools sent in Form 1, many did not include plans. They were each called,
sometimes more than once. Eventually, some 95% of the plans were received by NITON.
READING THE PLANS
The floor plans of each building were marked to show which rooms required a radon test, and
which ones should have the side-by-side QA/QC test, either a Maine 4" canister or a NITON vial,
or both. Problems at this stage:
• The usefulness of the plans NITON received varied greatly. Some plans were simple hand
sketches that were very clear; others were formally drafted asbestos schematics that were not.
When plans were unclear, NITON called the school. A. McGuineas, NITON Vice President, and
E. Romm, with many years of experience in construction, found someone knowledgeable about
the building and they walked through it over the phone. About 30% of the schools needed to be
called.
• A few schools had no plans. The same procedure was followed.
SCHEDULING THE TESTS
Even with professionals, schedules need to be set up and monitored. With non-professionals,
schedules had to be more detailed, and people need to be monitored closely and continuously;
In each package of tests to a school (see section below on packaging and shipping of tests) a
schedule was included that gave the test placer different date options for testing. This was done
both to give a deadline, which tends to focus people, and to assure that NITON did not receive
thousands of tests all in one day, which, together with its normal business would have violated its
strict QA, which calls for analyzing every test on the day it arrives. (NTTON's Standard Deviation
(a) at 4 pCi/L is less than 5%, twice as good as required by the EPA.)
The easy-to-follow instructions called for setting out tests on Friday afternoon and h^es^ng
them early Monday morning, shipping them immediately UPS 2nd Day. More than 98-5;® of the
schools finally followed instructions to the letter. There were, however, difficulties. P1"0 enis at
this stage:
• Package did not get to the proper person, although addressed properly. When follovvec* UP
by phone, the packages were all found and tests proceeded.
-------
• Contact person or test placer ignored the dates and never started the tests. When followed up
by phone, new schedule dates were given and tests were done.
MANAGING NON-PROFESSIONALS
SCHOOL-SPECIFIC PACKAGING OF TESTS
School-specific packages were assembled. This was possible because NITON liquid
scintillation vials are small and light, weighing but 3-1/4 pounds per 100 (in contrast to 4" canisters
that, depending on the numbers of small boxes used, weigh from 40 to 60 pounds per 100 and
cannot easily be bagged.)
Each building's detectors and supplies were packed separately in bag(s) labelled with the name
of the school. Inside each bag were the marked-up floor plan of the school, the number of vials
required, and specially written instruction, data and radon information sheets. Extra NITON vials
and Maine 4" charcoal canisters were added where required for QA/QC.
In addition, there were bright, easy-to-spot NITON place mats that say, "Radon test is in
progress," on which to set the tests. Assembling and checking these school-specific packages was
labor-intensive, but was critical to the success of the program.
TESTING
Maine schools were tested from late Friday afternoon to early Monday morning. This time
period assures that outside windows are closed and all doors are closed most of the time.
Moreover, no overtime is required of school custodians, and no special arrangements need to be
made for access to the building. (NITON detectors are calibrated from 24 hours to 72 hours for
screening.)
There were surprisingly few problems at this stage, due to two factors:
1. School-specific packs and clear instructions. Every school was packaged individually and
everything was spelled out. While it is true that if anything can go wrong, it will, there was not
much that could go wrong. The stages of the testing were carefully described, limited, and
monitored.
2. An 800 number. Prominently listed on the instruction sheet was the NITON 800 number.
It is company policy to spend time with callers no matter what their questions are. Maine people
were not shy about calling. When they had doubts, they called, and things went more smoothly
because of it. Problems at this stage:
• The largest number of calls came in to ask questions about doing the QC tests. They were
particularly complicated to do because the test retriever had to mail the 4" canister to an address
different from the NITON vials. All labels and return boxes were provided, and NITON rewrote
instructions many times, but the concept was evidently difficult. Because, however, there was
such close personal contact, the work was very well done and the results gratifying (3).
• Handwriting on data forms was sometimes nearly undecipherable, but this is not unique to
Maine folk. There seems to be no remedy.
• Occasionally, parts of the data form were unfilled. A follow-up call answered most
questions.
-------
• The question sometimes unanswered was in connection with the heating and ventilating
systems. Protocol called for them to be on continuous cycle, to replicate as closely as possible
occupied conditions. Schools with facilities managers or custodians answered these questions.
Smaller ones were sometimes at a loss: They could not find the controls, they did not how to
change the cycles, they changed them and weren't sure they were right.
In the winter, if the protocol was not followed, a system on a set-back cycle in Maine is likely
to give a too-high reading, since the radon builds up without ventilation of any kind. That is, the
screening test is likely to give a false high. In the summer, the systems are shut off, and their
influence cannot be discovered. In comparison testing of identical Massachusetts school rooms in
winter and summer, NITON found slightly higher readings in the summer (4).
Two schools goofed. Of 186 tests that had to be redone, 126 were from one school who
exposed vials for six days instead of a week-end (then promptly returned them to the lab). The
other 60 null tests were from a second school.
DATA REPORTING
QUALITY ASSURANCE/QUALITY CONTROL
To test the tests, two procedures were used. Side-by-side NITON vials were set out in some
150 designated rooms. In addition, (50) 4" charcoal canisters (75 gr) were supplied by the State
of Maine and analyzed in the Maine Radon Laboratory. Some 30 NITON vials were checked with
an electronic monitor. Preliminary analysis shows the degree of agreement was extremely
satisfactory (3).
The NITON QA/QC protocol was in force at all times: All tests are analyzed the day they are
received. At 4 pCi/L, the Standard Deviation (a) is less than 5%; at 1 pCi/L, SD (o) is 10%.
All first round tests showing more than 3 pCi/L are recounted—Standard Deviation (a) = 2% at
3 pCi/L.
USING DATA TO DETERMINE ROOMS TO RETEST AND/OR MITIGATE
Some areas of Maine were known to have high radon. When the first high readings in those
area schools showed up, the EPA was called in and mitigation begun.
The decision was made to retest every room over 3 pCi/L, the placement and retrieval to be
done by Maine officials. A cost-effective way to do this is to use the NITON vial to ascertain day
and night readings. The NITON vial is calibrated to 8 hours for this purpose, and is extremely
sensitive as well as accurate at low levels, since the liquid scintillation counter counts 100% of
5 decaying particles. Thus, a reading may be taken during the school day, when the building is
occupied, and another in the same room at night when the systems are set back. Preliminary data
corroborate the screening tests.
REPORTING THE DATA TO THE STATE OF MAINE
Results were sent to Warren within two business days of the arrival of the tests at the lab.
Results greater than 3 pCi/L were highlighted. All graphs and tables produced by NITON will be
sent upon completion of the work. In addition, NITON is making available the original laboratory
data on discs for further analysis.
-------
PUBLIC RELATIONS
The wise decision was made to tell the public of results as they were learned. When well
framed, even high readings can be disclosed without alarming the public. The wisdom of this
policy was demonstrated in the towns around Sebago Lake. Where in several schools the radon
was in excess of occupational levels in uranium mines, yet there was no hue and cry to close the
schools, as there has been in areas where high results are kept secret for too long.
ACKNOWLEDGEMENTS
E.R. wishes to acknowledge the cheerful, problem-solving, unflappable assistance of Christopher
Collins, Vicki Grzybinski, Lori Tamaro, and Anne McGuineas, all of NITON Corporation, in
converting ideas into the actual carrying out of the protocols and procedures of the Maine program.
The work described in this paper was not funded by the U.S.
Environmental Protection Agency and therefore the contents do
not necessarily reflect the views of the Agency and no official
endorsement should be inferred.
REFERENCES
1. EPA Radon Measurements in Schools: An Interim Report. U.S. Environmental Protection
Agency, Office of Radiation Programs, Washington, DC. EPA 520/1-89-010. Sect. VI.D. 1
2. Warren, H.E. and Romm E.G. The State of Maine School Radon Project: The Design
Study. EPA 1991 International Symposium on Radon and Radon Reduction Technology
Philadelphia, PA. April 2-5, 1991.
3. Grodzins, L., Bradstreet, T., Moreau, E. The State of Maine School Radon Project: Results.
EPA 1991 International Symposium on Radon and Radon Reduction Technology,
Philadelphia, PA. April 2-5, 1991.
4. Grodzins, L. Radon in Schools in Massachusetts. EPA 1990 International Symposium on
Radon and Radon Reduction Technology, Atlanta, GA. February 19-23, 1990.
-------
VIP-4
RESULTS OF THE NATIONWIDE SCREENING FOR RADON IN DOE BUILDINGS1
by: Mark D. Pearson, D. T. Kendrick, and G. H. Langner, Jr.
U.S. Department of Energy Grand Junction Projects Office
Chem-Nuclear Geotech, Inc.
P.O. Box 14000, Grand Junction, CO 81502-5504.
ABSTRACT
The U.S. Department of Energy (DOE) conducted a nationwide screening of
its buildings during the 1989-90 winter season in response to the Indoor Radon
Abatement Act of 1988. Three-month radon measurements using alpha-track radon
monitors were made in approximately 3,100 DOE buildings distributed among
74 administrative units, the goal of which was to identify occupied buildings
at DOE sites with elevated radon concentrations. Radon-in-water samples were
also obtained from approximately 120 non-public water sources distributed
among 22 DOE sites.
The screening measurements identified 86 of 3,100 buildings with winter
season radon concentrations exceeding 4 picocuries per liter (pCi-L"^-). The
geometric mean of the entire set of alpha-track measurements was 0.63 pCi-L"-'-,
the arithmetic mean was 0.91 pCi'L"^, the geometric standard deviation was
1.89, and the highest reported measurement was 148.5 pCi'L"^. Results of the
measurements in DOE buildings were considerably lower than those reported by
Nero and others for U.S. residences (1). This difference may be due to the
comparison of winter measurements versus Nero's annual average measurements or
may be the result of higher ventilation rates in commercial buildings and
other factors related to construction techniques, building use, or geographic
distribution of DOE sites.
1-Work supported by the U.S. Department of Energy Office of Administration and
Human Resource Management under DOE Contract No. DE-AC07-86ID12584.
-------
INTRODUCTION
The U.S. Department of Energy (DOE) Indoor Radon Study was conducted in
response to Public Law 100-551, the Indoor Radon Abatement Act, enacted by
Congress on October 28, 1988. This law required each Federal department or
agency to conduct a study to determine the extent of radon contamination in
its buildings and to report the results to Congress by October 1, 1990.
Management responsibility for the DOE Indoor Radon Study was assigned to
the DOE Office of Management and Administration (DOE-MA)2. DOE-MA selected
the DOE Grand Junction Projects Office (GJPO) and its contractor, UNC
Geotech3, to implement the study. Work scope and milestones for the study
were set by DOE-MA.
The DOE Indoor Radon Study was designed and implemented to conform with
the recommended protocols and guidance of the U.S. Environmental Protection
Agency (EPA) (2) for alpha-track screening measurements in buildings with
heating, ventilating, and air conditioning (HVAC) systems operating in a
normal manner.
The goal of the radon survey was identification of occupied buildings
with elevated radon concentrations at DOE sites, and identification of DOE
sites at which elevated radon concentrations might be suspected. Because
limited resources and a limited number of monitors were available for this
study, the choice was made to sample a large number of buildings at a low
density of monitors instead of intensively sampling a few buildings.
METHOD
The Radtrak™ alpha-track radon monitor^ was selected for the DOE Indoor
Radon Study. Of the 6,000 monitors purchased for the study, approximately
5,700 monitors (including 500 monitors used for duplicate measurements)
were deployed in the field at DOE sites and the remaining 300 monitors were
used as exposed and unexposed controls. Radon measurements were performed
during the winter season from mid-November 1989 to mid-February 1990. These
3-month measurements were performed under assumed normal building
operating conditions.
Given a fixed number of monitors and a fixed number of DOE sites and
buildings, the challenge was to determine the allocation of monitors that
minimized the probability of false negatives, i.e., an incorrect conclusion
that all buildings at a given DOE site were below a 4-picocuries-per-liter
(pCi-L"-*-) action level if in fact some were above 4 pCi-L"-*-, The primary
2Later renamed the Office of Administration and Human Resource Management
(DOE-AD) as a result of reorganization.
-*UNC Geotech was acquired by Chem-Nuclear Environmental Services, Inc. in
1990 and was renamed Chem-Nuclear Geotech, Inc.
^Tech/Ops Landauer, 3 Science Road, Glenwood, Illinois, 60425.
-------
assumption made in determining this allocation was that the distribution of
radon concentrations among DOE buildings was on the average similar to that
published by Nero et al. (1), even though the Nero distribution was based on
residences with typically lower building ventilation rates than those in
commercial buildings. Using this distribution as a guide, and also assuming
that a single radon monitor could be placed in a location in a building where
the radon concentration was maximized, an algorithm was created for allocating
the 5,700 field monitors among Ik administrative units that contained a total
of 10,000 buildings. The number of radon monitors allocated to each DOE site
is shown in Figure 1, as is the geographic distribution of those sites.
Allocation of monitors to specific buildings at a DOE site was left to
the discretion of the site Point of Contact for the study. In most cases,
these site contacts were personnel trained in health physics, many of whom had
knowledge and experience in performing radon monitoring. Site contacts were
advised to sample as many occupied buildings as possible with the number of
monitors provided to them, while giving higher priority to buildings with
greater potential radon concentrations and more occupants. Guidance was also
provided for placing monitors within a building to sample areas representative
of the highest radon levels likely to be found in the occupiable portions of
the building, such as in basements, interior rooms on the lowest occupied
level of the buildings, rooms isolated from a central HVAC system, and rooms
at or near structural joints (adjacent slabs or building additions). Monitors
were ultimately placed in approximately 3,100 DOE buildings.
The geographic distribution of the radon monitors across the country was
by no means random. Due to the concentration of DOE facilities in certain
regions, a preponderance of monitors were located in South Carolina,
Tennessee, New Mexico, Idaho, Washington, and the San Francisco, California,
area.
Water samples were obtained from non-public water supplies identified by
the site Points of Contact. Approximately 120 water sources distributed among
22 sites were sampled in duplicate. Water samples were collected in 4-ounce
glass bottles with Teflon™-lined caps following the guidance of the EPA's
recommended test procedures for radon in drinking water (3). The samples were
analyzed at the GJPO for radon concentration by a liquid scintillation
counting technique.
Instructional materials were provided to each site for training and
general information. A set of Implementing Instructions gave step-by-step
instructions for selecting buildings for inclusion in the survey and for the
placement of the alpha-track radon monitors. Water sampling instructions were
provided to those sites requiring radon-in-water measurements. Copies of the
DOE Indoor Radon Study Fact Sheet were distributed to each site. These fact
sheets contained information about the study and described radon health risks
and sources of radon. Additionally, a placard was placed with each monitor
that identified the site contact person and the study project manager.
-------
IPA
(HD CDIF
92
Col
toUa
[CPP
ICUP
PP1
NPRW
BNl
153
'A CA
LEHF
136
Bttila
lound
136
SNL
J84
'•rn«id I*
• 69
GJPOj
KCP
SUCI
'APA CO
NPRC
MO ETEC
44
lOSl 269
Pantex"
160
(GD
ISO
CTJ
SKO
393.
SHLI
WIPP
.<0
Legend
Acronym Sit* Nam*
AMES
Ames Laboratory
ANL
Argonne National Laboratory - East
ANIWEST
Argonne National Laboratory -Wast
AOS
Albuquarque Office Sita
Alaska Power
Alaska Power Administration
BNL
Brookhaven National Laboratory
BPA
Bonneville Powar Administration
Batu
Balaa Aocalarator Facility
BeMa
Battle Atomic Energy Laboratory
CEBAF
Continuous Election Baam Accelerator FaclHty
CTA
Albuquerque Office Site, Central Training Academy
Colonle
Colonla Intarlm Storage Sita (Fomerly Utilized
Sites Remedial Action Protect)
ETEC
Energy Technology Engineering Center
FNAL
Femtl National Accelerator Laboratory
Femald
Femald Feed Materials Production Center
GJPO
Grand Junction Protects Office
ICPP
Idaho National Engineering Laboratory -
Idaho Chemical Processing Plant
INEL
Idaho National Engineering Laboratory - EG&G
ITRI
Inhalation Toxicology Research Institute
KCP
Kansas City Plant
KEHC
Kaiser Engineers Hantord Company
Kasselitng
Knolls Atomic Energy Laboratory - Kesselring
Knolls
Knolls Atomic Energy Laboratory - Knolls
LANL
Los Alamos National Laboratory
LBL
Lawrence BerVtley Laboratory
LEHR
Laboratory tor Energy-Related Health Research
LLNL
Lawrence Uvormore National Laboratory
LREH
Laboratory tor Radloblology and Environmental Health
METC
Morgantown Energy Technology Center
MHDCDIF
MHO Component Development Integration Facility
MSU
MSU Plant Research Laboratory
Mound
Mound Facility
NPHW
Naval Petroleum Reserves (Wyoming, Colorado, and Utah)
NORL
Notre Dame Radiation Laboratory
NPRC
Naval Petroleum Reserves (California)
NRFID
Idaho Naval Reactor Facility
NVO
Nevada Operations office
Acronym Sit* Nam*
ORAU
Oak Ridge Associated Universities
ORQD
Oak Ridge Gaseous Diffusion Plant
ORNL
Oak Ridge National Laboratory
PETC
Pittsburgh Energy Technology Center
PNL
Pacific Northwest Laboratory
PPPL
Princeton Plasma Physics Laboratory
PaducahGDP
Paducah Gaaeoue Dlffualon Plant
Pantex
Pantex Plant
Pinellas
Plneliea Plant
Portsmouth GDP
Portsmouth Gaseous Diffusion Plant
RFP
Rocky Flats Plant
RoisAv.
Albuquerque Office Site - Ross Aviation
SCMP
Idaho National Engineering Laboratory -
Specific Manufacturing Capabilities Protect
SERI
Solar Energy Research Institute
SLAC
Stanford Linear Accelerator Center/
Stanford Synchrotron Radiation Laboratory
SNL
Sandla National Laboratories (Albuquerque Site)
SNLKTF
Sandia National Laboratories (Kauai Site)
SNLLS
Sandla National Laboratories (Uvermore Site)
SPR
Strategic Petroleum Reservee
SRO
Savannah River Operations Office
SWPA
Southwestern Power Administration
TSO
Albuquerque Offloe Site, Transportation Safety Division
Tonopah
Sandia National Laboratories (Tonopah Site)
UTSI
University ol Tennessee Space Institute
W. Valley
Weal Valley
WAPA
Western Area Power Administration
WAPA AZ
Western Area Power Administration - Phoenix
WAPACA
Western Area Power Administration - Sacramento
WAPA CO
Western Area Power Administration - Mommas
WAPAMT
Western Area Power Administration - Fori Peck
WAPA NO
Western Area Power Administration - Blsmardt
WAPA SO
Western Area Power Administration - Huron
WAPAWY
Western Area Power Administration - Loveland
WHC
Westtnghouse Hantord Company
WIPP
Waste Isolation Pilot Plant
Windsor
Knolls Atomic Energy Laboratory - Windsor
Y-12
Y-12 Plant
Figure 1. Participating DOE Sites and Distribution of Monitors
-------
QUALITY ASSURANCE AND QUALITY CONTROL
A Quality Assurance Program Plan was prepared that described quality-
control exposures for alpha-track radon monitors and radon-in-water sample
analyses. A number of surveillances were performed to verify laboratory
compliance with procedures for handling alpha-track monitors and to verify
compliance by site contacts for placement of the monitors.
Exposure of 252 alpha-track monitors in six groups in the GJPO radon
calibration chambers resulted in an average ratio of vendor-reported exposure
to chamber reference exposure equal to 0.89. The average coefficient of
variation for these six control groups, each consisting of 42 monitors, was
9.4 percent. The average coefficient of variation for 470 field duplicate
monitors was 13 percent. All unexposed controls, or blanks, reported results
below the vendor's detection limit of 30 pCi'd'L"^.
RESULTS AND DISCUSSION
Information for 107 alpha-track monitors placed in the field was lost for
one of the following reasons: 16 monitors were vandalized or damaged,
76 monitors were physically misplaced and not returned by site contacts, and
15 monitors were incorrectly labeled by the vendor or the placement data
sheets contained incorrect information that invalidated the data. Another
1-24 monitors were not placed by site contacts for a variety of reasons. Data
were, therefore, successfully returned from 5,469 of the 5,700 monitors
initially distributed.
Figure 2 shows the frequency distribution of the alpha-track radon
measurement results, and Figure 3 is a histogram of the natural log of the
measurement results. The geometric mean (GM) and geometric standard deviation
(GSD) of these 3-month winter measurement data are 0.63 pCi-L"-'- and 1.89
pCi L"-*-, respectively, and the arithmetic mean (AM) is 0.91 pCi'L"-*-. In a
compilation of annual average radon measurements made in U.S. homes, Nero et
al. (1) calculated a GM of 0.96, a GSD of 2.84, and an AM of 1.66 pCiL"^.
Eighty-six of the 3,100 measured buildings, or 2.8 percent, were found to
exceed a 4 pCi'L*-'- action level. The detection limit reported by the alpha-
track vendor was 30 pCi-d'L'-'-, which corresponds to a concentration of
0.3 pCi•L"1 for the 90-day exposure period recorded by the majority of
monitors in this study. This means that monitors exposed to radon levels
ranging from 0 to 0.3 pCi-L"1 all reported a value of 0.3 pCi-L"1. More than
14 percent of the DOE building measurements were below 0.3 pCi-L'i.
The DOE buildings measured in this study exhibited lower radon levels
than the levels measured in U.S. homes by Nero. If the distribution of year-
long average radon concentrations in DOE buildings was similar to Nero's
distribution of annual average measurements, the 3-month winter screening
measurements should show higher levels than measured by Nero. The
-------
25
Class Widlh = 0.1 pCi/L
20
15
ra 10
" 5
/Is J
Geometric Mean » 0 63
Geometric Standard Deviation * 1 89
Arithmetic Mean ¦ 0.91
0.25
TTTTTTTtTtttttttttt-
1.05 2.05 3.05
Radon Concentration (pCi/L)
'TTTTTI
>4
Figure 2.
Histogram of the measurement results from the DOE Indoor
Radon Study plotted as a percentage of the total number
of measurements. The class width shown is 0.1 pCi/L. which
represents the actual resolution available from the data.
The arthmetic mean and geometric mean are given in pCi/L.
30
25
20
15
10
Figure 3.
-r—r
Class Width - 0.29 ln(pCi/L)
f -T T -r—r
-1.204 -0.440 1.116 2.276
Natural Log ot Radon Concentration
I—i—i—i—i—r~ r
3.436
4.596
Histogram of the natural log of the measurement results from
from the DOE Indoor Radon Study plotted as a percentage of
the total number of measurements. The class width shown is
0.29 natural log of the radon concentration in pCi/L. This
distribution shows a marked positive skewness.
-------
measurements in this study demonstrate, in fact, just the opposite trend. A
number of factors may contribute to this difference, including the typically
higher ventilation rates in commercial buildings and other factors related to
construction techniques and building use. Also, the DOE buildings were not
distributed geographically in a random fashion. A large fraction of the
DOE building population is situated in regions with generally low radon
concentrations, such as the coastal plain of South Carolina, eastern
Washington, and the San Francisco Bay area. It also may be possible that the
winter season is not the season of highest expected radon concentration for
some of these regions.
The data were expected to fit a lognormal distribution, but as Figure 3
indicates, the natural log-transformed data are apparently skewed to the right
instead of exhibiting a bell-shaped Gaussian distribution. An important
consideration in this skewness is the left censoring of the data in the
distribution. Left censoring means there are no concentration values reported
below some detection limit, in this case 0.3 pCi'L"^.
Left censoring of the data is of concern in this case because the
censor point, 0.3 pCi'L"^, intersects the data distribution close enough to
the mean to distort the observed distribution. The observed skewing of the
distribution to the right may be the result of this left censoring. Left
censoring of the data also shifts the observed mean of the data to a
higher value.
Duplicate water samples were taken from 120 water-supply locations at
22 sites. The highest radon-in-water measurement was 1,460 pCi'L"''".
Approximately two-thirds of the measurements, 82 of 120, were less than
200 pCi-L"l. There was no observed correlation between any of the elevated
radon-in-air measurements from alpha-track monitors and the radon-in-water
measurements.
The work described in this paper was not funded by the
U.S. Environmental Protection Agency and therefore the
contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
-------
REFERENCES
1. Nero, A. V. , Schwehr, M. B., Nazaroff, W. W., and Revzan, K. L.
Distribution of airborne Rn-222 concentrations in U.S. homes. LBL Report
No. 18274, EEB-Vent 84-33, Lawrence Berkeley Laboratory, Berkeley,
California, 1984.
2. U.S. Environmental Protection Agency. Indoor radon and radon decay
product measurement protocols. EPA 520-1/89009, Office of Radiation
Programs, Washington, D.C., 1989.
3. U.S. Environmental Protection Agency. Two test procedures for radon
in drinking water, interlaboratory collaborative study. EPA/600/2-87/082,
Environmental Monitoring Systems Laboratory, Office of Research and
Development, Las Vegas, Nevada, 1989.
4U.S. GOVERNMENT PRINTING OFFICE; 1991-5« 8-1 8 7/2 0566
------- |