United States       Air and Energy Environmental   EPA/600/9-90/005C
          Environmental Protection   Research Laboratory      January 1990
          Agency         Research Triangle Park NC 27711

          Research and Development	
ve/EPA     The 1990 International
          Symposium on Radon
          and Radon Reduction
          Volume III. Preprints

          Session IV:  Radon Surveys
          Session V: Radon Entry
          Session VI:  Radon in the
           Natural Environment
          Session C-V: Radon Entry
          Session C-VI:
           Radon in the Natural
          February 19-23,1990
          Stouffer Waverly Hotel
          Atlanta, Georgia

  Session IV:
Radon Surveys


by: Alan J. Siniscalchi, M.S., M.P.H. (1), Lynne M.  Rothney,  M.P.H.  (2),
    Brian F. Toal, M.S.P.H. (1) ,  Margaret A. Thomas, M.S.  (3),
    David R. Brown, Sc.D. (1), Maria C.  van der Werff,  B.S.  (4),
    Carolyn J. Dupuy, M.S., S.M.  (1)

       (1)   State of Connecticut Department of Health Services
             Hartford, Connecticut 06106-4474
       (2)   Yale University School of Medicine
       (3)   State of Connecticut Department of Environmental Protection
       (4)   U.S. Environmental Protection Agency Region I -  work
             performed while employed with the State of Connecticut
Statewide radon measures are needed to establish comprehensive population
estimates of risk.  The Connecticut Department of Health Services has
measured indoor radon concentrations in over 5,000 living units which
represent nearly 1% of the homes statewide.  Short-term exposure data
were obtained from both lowest livable areas (i.e., basements) and lowest
living areas in 3,378 houses.  Long-term living area data were collected
in over 500 homes.  Analysis of these data has shown associations with
home construction type including a strong positive correlation with house
age.  Analysis of over 1,000 homes with energy use data did not reveal an
association between energy efficient homes and high radon levels.
Differences in air and water levels have been identified among various
geologic units.  Consistent findings among the three surveys include
geometric mean basement and living area radon levels and percent of homes
exceeding 4 pCi/L.


    The State of Connecticut Department of Health Services (DHS) began
receiving numerous telephone inquiries on radon following the discovery
of elevated indoor air radon levels in Pennsylvania (1).   Since the DHS
lacked information on radon exposure in Connecticut, plans were made to
investigate the distribution of radon in the state.  In addition to
providing information to residents concerned about radon, the Department
wanted to determine the extent of the radon problem in Connecticut.  The
risk reduction priorities of the DHS needed to be refined by under-
standing the true hazard posed by this newly discovered residential
risk.  Moreover, if elevated radon levels were identified in certain
areas of the state,  then local health agencies could focus their risk
reduction resources in these areas.  Since risk determinations require
exposure data, the DHS proceeded to assess Connecticut population
exposure to radon.

    A pilot project and three studies have been performed.  The first
study was the 1985-87 "Connecticut Radon Survey" (2).  This study was
followed by an expanded 1986-87 "EPA-Connecticut Radon Survey" of
basement radon levels in 168 of Connecticut's 169 municipalities.

    Results of the EPA-Connecticut Radon Survey were then utilized to
plan the 1987-88 "Household Testing Program" where 3,400 homeowners were
given radon testing kits in towns which were selected based on radon
potential or in towns not previously sampled.

                       THE CONNECTICUT RADON SURVEY
    The Connecticut Radon Survey was conducted between 1985 and 1987 to
assess the predictive ability of geologic, hydrologic and household
factors on well water and indoor air radon.  In the first phase, homes
were chosen  in six areas thought to be potentially high in radon levels
on the basis of the Bedrock Geological Hap of Connecticut (3), and later,
on the basis of an aeroradioactivity survey  (4).  The second phase
expanded  the survey to also include low and  intermediate radon areas.
Information  from  local health department well completion reports were
also used.   Private well water was sampled using the EPA scintillation
method from  262 homes.  The indoor air radon level was sampled with a
single alpha track device placed in the living area for three winter
months in 202 of  the homes from which water  samples were taken.  These
homes were located in 44 towns and represent sampling from 26 geologic

    The private well water radon levels ranged from 100 to 130,240
picocuries per liter (pCi/L) with a geometric mean level of 3,179 pCi/L.
Twenty-six percent (26%) of the wells had water radon levels greater than
10,000 pCi/L.  The long-term living area indoor air radon levels ranged

from 0.1 pCi/L to 25.6 pCi/L, with a geometric mean level of 1.25 pCi/L.
Eleven percent (11%) of the homes with radon had levels greater than 4
pCi/L.  Regression analysis estimated that 18% of the variation in indoor
air radon can be attributed to radon in private well water, with a water
to air ratio of 10,000 pCi/L to 1.6 pCi/L (C.I. - 0.80 to 3.40 pCi/L).
Figure 1 shows this relationship.

    Geologic, hydrologic and household parameters were analyzed for their
ability to predict radon levels.  Figures 2 and 3 show the mean water and
air radon levels respectively, compared to the bedrock classifications as
shown on the Connecticut portion (Figure 4) of a generalized bedrock
mapping scheme of New England (5).  Geology was a significant predictive
factor of radon in both private well water and indoor air with granitic
and sedimentary formations associated with higher and lower radon levels
respectively,  while the radon potential of bedrock may be generally
characterized by such data, it is possible that water radon levels alone
are sufficient to estimate the geologic radon potential.  Indoor air
radon measurements represent radon emissions not only from bedrock, but
also from surficial materials and also appear to vary significantly with
various household factors.

    Hydrologic factors, foundation type and energy efficiency were
examined.  Only the depth of unconsolidated material overlaying the
bedrock had a strong positive correlation with private well water radon
levels.  Homes with block wall foundations had higher indoor air radon
levels than those with other types of foundations.  Homes characterized
by the homeowner as being more energy efficient did not have higher
indoor air radon levels than homes characterized as being less energy
                       EPA-CONNECTICUT RADON SURVEY
    During the summer of 1986 the DHS cooperated with the U.S.
Environmental Protection Agency (EPA) to conduct a survey of basement
radon levels in Connecticut using charcoal testing devices supplied and
analyzed by the EPA Eastern Environmental Radiation Facility in
Montgomery, Alabama.  The survey design called for differing sampling
densities based upon radon potential estimations conducted by a geologist
from the Connecticut Department of Environmental Protection (DEP).   The
radon potential was estimated using geological mapping,  the Connecticut
Radon Survey results and aeroradioactivity mapping.  A target sample
number from 7 to 15 samples per town was established based on the radon
potential estimation for each of Connecticut's 169 towns.  The survey
design called for distribution of charcoal testing devices to 1600 homes,
and for placement in the lowest livable area (i.e. basement) of the house
for two days.  The devices were placed during the winter of 1986-87 by
energy auditors from CONN SAVE, a non-profit energy conservation
organization.  Homeowners were offered a radon test when they requested

an energy audit.  Testing was on a first-come first-served basis until
the target sample number was reached in each town.   At the time of kit
placement, information on housing characteristics,  air infiltration rate,
and house location were recorded by the energy auditor.  Air infiltration
rates were later estimated with a computer model commonly used by energy
conservation organizations.

    A total of 1,572 homeowners agreed to participate in the survey, with
a refusal rate of less than 1%.  One hundred and forty seven or 9.4% of
the tests, were dropped from the analyses due to improper testing or
mail-in procedures.  A total of 1,157 tests of detached home basements
were included for the detailed analysis reported here.  Nineteen percent
(19%) of the basements tested exceeded the EFA guideline of 4 pCi/L.
There were significant differences (p/ 0.05) in the percentage of homes
with radon levels greater than 4 pCi/L between major areas of the state.
These regions were established by their estimated geological potential
for radon.  Furthermore, correlation analysis demonstrated strong
associations between radon levels and geology.  Detailed analysis with
regard to specific bedrock units also showed strong associations, with
homes above granitic rock units having higher radon values than homes
overlaying sedimentary rock units.  Of the many housing characteristics
studied, the age of the house was the most predictive factor with older
housing having higher levels (see Figure 5).  Fieldstone and block
foundations were also accompanied by higher radon levels compared to
concrete and concrete mix foundations (see Figure 6).  No association was
found between radon levels and the estimated air infiltration rate.
These findings are consistent with those of the previous survey.

    Geographic Information System (CIS) analyses of indoor radon data
with digitized aeroradioactivity mapping (4) showed a strong correspon-
dence between the percentage of homes with basement levels that were
greater than or equal to 4 pCi/L and the generalized aeroradioactivity
mapping as measured in counts per second gamma radiation (see Figure 7).
    Since a review of radon distribution revealed that some elevated
 radon occurs  in all areas, the DHS issued an advisory in August 1987
 stating  that  all Connecticut homeowners should have their houses tested
 for radon.  The DHS then established a full-time Connecticut Radon

                        HOUSEHOLD TESTING PROGRAM
     In  December  of  1987  the  Connecticut Radon Program began distribution
 of free radon testing devices under  the Household Testing  Program  (HTP).
 The objectives of the HTP were:   to  provide free radon  testing devices

 and appropriate  placement instructions  to residents living in areas
 suspected  of  having high radon levels;  to obtain additional data on radon
 concentrations in selected Connecticut  municipalities; and to examine the
 ratio between basement  and living  area  radon concentrations.

     In planning  the HTF,  fifty-three cities and towns were identified by
 a  geologist employed with the  Connecticut Department of Environmental
 Protection (DEP)  based  on results  of the previous two radon surveys and
 existing information on terrestrial radiation and bedrock geology.
 Thirty-eight  of  these municipalities were selected for the HTP based on
 the  ability and  interest  of the local health departments or other local
 agencies to participate in the distribution of testing devices (see
 Figure 8).

     Each municipality was given 200 charcoal testing devices and asked to
 recruit 100 volunteer households.  One  charcoal testing device was placed
 in the basement  or other  "lowest livable area" and the second device
 placed in  the "lowest lived-in area."   All testing device analyses were
 conducted  by  the  same contract laboratory.  Three hundred and forty
 households with  living  area radon  concentrations over 4 pCi/L and/or
 basement radon concentrations  over 20 pCi/L received alpha-track devices
 for  long-term follow-up testing.

     The results of the  HTP were not different from those of the
 EPA-Connecticut  Survey  and the Connecticut Radon Survey (see Table 1).
 The  data also revealed  an apparently consistent 3:2 ratio between
 basement and  living area  radon concentrations (6).  Analysis indicated
 that the basement radon level  is strongly predictive of the upstairs or
 living area radon level (R2-0.48,  P/ 0.00001).

     The program  is now  evaluating  results of long-term (9-12 month) alpha
 track devices provided  to the  340  participants (10%) whose living area
 radon levels  exceeded 4 pCi/L  or basement levels that exceeded 20 pCi/L.
 The  program also  examined the  influence of waterborne radon on indoor air
 radon levels.  Figure 1 showed the analysis of variation between radon in
water and  indoor  air radon levels  for the Connecticut Radon Survey.  When
HTP  homes  on  private well  water were compared to homes with public water
 the  ratio  of  living area  to basement air levels was 0.75 ± 0.46 versus
0.49 ± 0.29 respectively  (mean ± SD, N - 50).  This effect was non-linear
which suggests that higher levels  of water radon are more significant
contributors  to indoor  living  area radon exposure.

    The Connecticut DHS and DEP have collected radon data on 5,036
households.  Information on housing characteristics and detailed
locations is available on nearly all of the units tested.  Household
locations have been mapped on U.S. Geological Survey topographic maps at

1:24,000 scale and digitized into the DEP computerized Geographic
Information System (CIS).  This information has proven to be extremely
useful in evaluating the radon potential of specific areas of the state.

    The CIS is being used to test correlations of radon occurance at
various  concentrations with geophysical data, specific geological
materials and modifying environmental conditions.  Statistical analyses
of radon indoor air and water data with respect to Connecticut geological
terranes and individual bedrock formations have produced geologically
stratified sampling schemes for the EPA - Connecticut Radon Survey and
the Household Testing Program.  For the EPA - Connecticut Radon Survey,
CIS analyses have shown a strong correspondence between aeroradioactivity
mapping (4) and basement radon testing (see Figure 7).  These analyses
produced statewide radon potential mapping that is being used as a tool
to increase our understanding of the locations of,homes with radon levels
above 4 pCi/L.

    The aeroradioactivity data was also utilized by earlier researchers
in conducting an epidemiological study of cancer rates in Connecticut
(7).  The weighted regression analysis by towns used by Walter, et al.
did not reveal an increased cancer rate by radiation level.  However, the
authors cited the limited statistical power of the study which provided
only a small probability of detecting a radiation effect if a two-fold
excess cancer risk existed in the higher gamma areas.  The present survey
data suggests that towns are too large a unit to detect cancer increases
from radon exposure.

    The similarity in the results for both the percentage of homes above
the guideline and geometric mean radon levels are especially notable when
one considers the methodological differences in both device (short-term
charcoal and long-term alpha track) and home selection among each of the
three studies.  For example, differences in bias of the selection among
the three studies is shown in Table 1.  The selection of households for
the Connecticut Radon Survey and the Household Testing Program tended to
be toward areas with higher radon levels (although to a lesser degree in
the HTP).   The homes tested in the EPA-Connecticut Survey were not
selected in areas that were known for high or low levels.  Thus, the
selection bias of this study was neutral.

    The results of these studies provided information to the Department
that suggested radon exposure in Connecticut was higher than expected.
Many local health departments then initiated their own surveys.  This
information was also used to plan a number of educational campaigns
designed to encourage further testing among Connecticut residents.

    The studies also provided information useful to the Department's
evaluation of its risk reduction priorities.   An analysis of the survey
data has been used to generate estimates of the risk for developing lung
cancer from radon exposure in Connecticut.   Table 2 displays the

distribution of living area radon concentrations for the three studies.
Analysis of the distribution of this data using the risk tables provided
in the National Academy of Sciences BEIR IV report (8) and smoking
information would predict 280 excess lung cancers could occur each year
in Connecticut from exposure to radon in homes.  Risk calculations using
a U.S. EPA model (9) predict similar rates.

    An extrapolation of this information yielded estimates of additional
exposures to radon that occur in schools.  Assuming a school population
of 542,000 students with 10 years of exposure, an estimated 425
additional cases of lung cancer could result from radon exposure in
schools alone (10).  If, as chronic disease surveys (11) imply, that 19%
of elementary and secondary school of students smoke, the excess cancer
cases due to radon alone would be much higher.

    This series of surveys illustrate the complexity of predicting radon
exposures from limited measures of radon.  Additional factors which are
being investigated further are: the contribution of water radon in homes
with private wells, the influence of differences in bedrock type within
small geographic areas, and the radon contribution from unconsolidated
materials, differences in housing characteristics and lifestyles of the


    The three Connecticut radon surveys show that a radon problem exists
in the state.  The studies also provided information on the differences
in risk due to variability in exposure from housing type and location.

    Information on the relationship between radon levels and geology will
continue to be developed to refine our knowledge of the locations and
conditions in Connecticut where radon may be of the highest risk.  Local
agencies can then focus educational and outreach efforts in these areas.

    The studies have been the foundation for the establishment of a
formal radon program within the DHS.  Information from these studies will
continue to guide the Program's efforts in public outreach and
educational campaigns designed to encourage testing and appropriate
mitigation within all Connecticut structures.
    With the exception of the charcoal testing devices provided by the
U.S. Environmental Protection Agency (EPA) for use in the EPA Connecticut
Radon Survey, the work described in this paper was not funded by the
EPA.  Therefore, the contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.


    The authors wish to thank Ms. Suzanne Lessard for her expert word
processing assistance in the preparation of this paper, Ms. Laurie Gokey
and Mr. Zygmunt Dembek for their review of the manuscript, and Mr. Marc
T. Rothney for the preparation of Figures 1, 2, and 3.  The authors also
express their appreciation to the staff of CONN SAVE for their assistance
in the EPA-Connecticut Radon Survey, the local health and other officials
who participated in the Household Testing Program, and all the households
who participated in these surveys.

1.  Logue, J. and Fox, J. Health hazards associated with elevated levels
    of indoor radon - Pennsylvania.  Morbidity and Mortality Weekly
    Report (MMWR).  34:657, 1985.

2.  Rothney, L.M.  Connecticut radon survey of private well water and
    indoor air:  assessing geologic, hydrologic and household
    parameters.  Connecticut Department of Health Services, Toxic Hazards
    Section, Hartford, Connecticut, 1987.  45 pp.

3.  Rogers, J. Bedrock geological map of Connecticut.   Scale 1:125,000,
    Connecticut Geological and Natural History Survey.  1985.

4.  Popenoe, P. Aeroradioactivity and generalized geologic maps of parts
    of New York, Connecticut,  Rhode Island and Massachusetts.  U.S.
    Geologic Survey Geophys, Inv. Map. GP-359, Scale 1:250,000, U.S.G.S.

5.  Olszewski, W., Jr. and Boudette, E.L. Generalized bedrock map of New
    England, Scale 1:1,000,000, New Hampshire Water Supply and Pollution
    Control Commission and U.S. EPA Region I, 1986.

6.  Toal, B.F., Dupuy, C.J., Rothney L.M.,  Siniscalchi, A.J., Brown,
    D.R., and Thomas, M.A.  Radon exposure assessment - Connecticut.
    Morbidity and Mortality Weekly Report (MMWR).  38:713, 1989.

7.  Walter, D.S., Meigs, J.W., and Heston,  J.F.  The relationship of
    cancer incidence to terrestrial radiation and population density in
    Connecticut 1935-1974.  American Journal of Epidemiology 123:1,

8.  National Research Council Committee on the Biological Effects of
    Ionizing Radiations.  Health Risks of Radon and Other Internally
    Deposited Alpha-Emitters (BIER IV).  National Academy Press,
    Washington, D.C., 1988.  602 pp.

 9.   Oge,  M.  Current ORP estimate  of  radon-induced lung cancer deaths in
     the general population.  U.S.  EPA memorandum,   August 17, 1989.

 10.  Dupuy,  C.J. and Rothney, L.M.  Radon risk in schools.  DHS
     memorandum, December 21, 1988.

 11.  Unpublished data.  Center for  Chronic Disease,  Urban/Rural Health,
     Connecticut Department of Health  Services.   Connecticut Health Check
     data  1988.







                     STATE      E3 GRANITIC GNEISS

                        n = NUMBER OF HOMES
                0-5K      5-1 OK      IO-20K    20-40K     40K
                               pCi/L (water)
Figure 1.  Connecticut Radon Survey:   geometric mean indoor air radon
levels by range of water  radon levels.

                  Sedtrrentary Besln
    Extrusive Igneous'Sedimentary Basin
                   Two-Mica Granite
                     Mafic Plutonic
                     General Gneiss
                     Granitic Gneiss
           Intermediate to Mafic Gneiss
              Stratified Metamorphic
       Carbonate Stratified Melamorphic
       Quartzite Stratified Metamorphic
                    General Mixture
                                      5000   10000  15000   20000   25000  30000
                                     pCi/L  (water)
                                                                      n - NUMBER OF HOMES
Figure  2.   Connecticut  Radon Survey:   geometric  mean  water radon  level by
generalized bedrock type.
                 Sedimentary Basin
   Extrusive Igneous Sedimentary Basin
                 Two-Mica Granite
                    Mafic Plutonic
                    General Gneiss
                    Granitic Gneiss
          Intermediate to Maftc Gneiss
             Stratified Metamorphic
     Carbonate Stratified Metamorphic
      Quartzite Stratified Metamorphic
                   General Mixture



i i i





F 	 -4 	 1
0 0.5

.5 2 2.5 3 3.5 4
                                      pCi/L (air)
Figure  3.   Connecticut  Radon Survey:   geometric  mean air  radon levels by
generalized bedrock type.


                                       GNEISSES -  GENERAL
                                             GNEISSES OF GRANITIC
                                             GNEISSES OF  INTERMEDIATE
                                             TO MAFIC COMPOSITION
                                          10. ULTRAMAFIC ROCKS


                                          11.  STRATIFIED METAMORPHIC

                                          lla.  CARBONATE STRATIFIED
                                               METAMORPHIC  ROCKS
                                          lib.  QUARTZITE STRATIFIED
                                               METAMORPHIC ROCKS

                                    12.   GENERAL MIXTURES OF ABOVE
                                         UNITS  (AREAS  NOT SHOWN)

Figure 4.   Connecticut portion,  generalized bedrock map of New England
           (modified from reference 5)

                       (2-9)    (10-24)   (25-50)    (> 50)
                          AGE  OF HOUSE (YRS)
 Figure 5.   EPA  - Connecticut Radon Survey:   geometric mean radon levels
 by age of house.
5          BLOCK
                0.0   0.3  0.5  0.8  1.0  1.3  1.5   1.8   2.0   2.3   2.5
                              PICOCURIES  PER LITER

 Figure 6.  EPA - Connecticut Radon Survey:  geometric mean radon levels
 by  type of foundation wall.

         —    Basement Air Radon vs. Aeroradioactivity
          ^         < 300        300-500       500-700       700-900
          °~        Aeroradioactivity  (counts per second  gamma)

Figure 7.  EPA - Connecticut  Radon Survey:  percent frequency  of
basements with 4 pCi/L radon  levels  by aeroradioactivity.
                                       HOUSEHOLD TESTING PROliRfltl
                                       CONNECTICUT  TOUN SELECTION
                                    H SELECTED FOR TESTING, UIMTER 1987-

                                    '//. RECOrrtNDED FOR FUTURE TESTING
Figure 8.   Household Testing Program:   municipalities recommended and
selected  for testing.

Survey device
Radon Survey
|n = 1157»)
Testing Program
In =3409!

of measurement
Lived-in area



•Number of detached houses out of 1425 total homes tested.
fBias toward geologic locations with a higher probability of finding high radon homes.
'Picocuries per liter.
fNot tested.
                             (from reference 6)
Radon Concentration (pCi/L)
   4  (geometric mean
    4  - 9.9
   10  - 19.9
   20  • 49.9
   50  - 99.9
      & 100
-1.3 pCl/L)

                  by:   Jacolyn A.  Dziuban1
                        Maureen Anderson Clifford1
                        S. B. White2
                        Jane U. Bergstein2
                        Barbara V.  Alexander2

                   1.  U.  S.  Environmental Protection Agency
                   2.  Research Triangle Institute

      This paper describes the cumulative results from 23 states that
conducted surveys with the assistance of the Environmental Protection Agency
(EPA) during the 1987, 1988 and 1989 heating seasons.  It also describes the
survey designs, provides population estimates of medians and means, and
defines the proportion of households in each state exceeding specified
exposure levels.

      The goal of these surveys was twofold:  to locate areas with elevated
radon levels, and to characterize the statewide frequency distribution of
radon screening measurements.  Each survey was designed to provide a
statistically valid comparison of radon levels in households in defined areas
within each state and for each state as a whole.  Overall, approximately
34,400 randomly-selected households provided screening measurements.
Experience gained through these surveys will be highlighted and applied to the
next series of state surveys scheduled for the winter of 1990.

      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.


                      PITTSBURGH RADON PROJECT

                                Bernard L. Cohen
                             University of Pittsburgh
                              Pittsburgh, PA 15260


Data are analyzed on measurements of radon levels in numerous U.S. homes, accompanied
by responses to questionnaires. Substantial bias reduction was accomplished by use of ques-
tionnaire responses, leaving 37,000 measurements in living areas and 33,000 in basements for
the analysis. Variables studied include level with respect to ground where measurement was
made, room type, age of house, recent weatherization actions, draftiness, location (urban-
suburban- rural), air pollution, market value of house, annual household income, educational
attainment of head of household, cigarette smoking, whether the house is rented or owner-
occupied, and geographic section of U.S. Mean radon levels are determined for each response
to questionnaire items  (correlations), and  for each pair of responses (cross correlations).
Many interesting correlations and cross correlations are found, and their explanation  and
consequences are discussed.
   From July, 1985 to January, 1988 the University of Pittsburgh Radon Project and its
successor (starting in May, 1987), The Radon Project provided measurements of radon levels
in homes almost exclusively in response to mail orders (cost - $12.00) stimulated by media
publicity. This yielded a fairly broad national coverage. Each measurement was accompanied
by a questionnaire which was filled out by the householder and returned before the result
of the measurement  was known.  The  purpose  of this paper is to present  results of the
statistical analysis of these data. Preliminary reports on this work have been published in
connection with presentations at conferences (Cohen and Gromicko 1988), but they were
based on portions of the data and incomplete analysis.
   The measurement and quality control procedures have been described previously (Cohen
and Nason 1986, Cohen 1988, Cohen and Cohen 1989). Diffusion barrier charcoal adsorption
collectors  were exposed for  seven days, after which  the quantity of  radon  collected  was
measured by 40 minute counts of gamma rays under the 295 KeV, 352 KeV, and 609 KeV
photoelectric peaks with Nal (Tl) scintillation detectors. Results are corrected for seasonal
variations to best represent annual averages (Cohen 1989). Statistical uncertainties are about
45% at 18 Bq m~3, 25% at 37 Bq m'3, and less than 10% above 120 Bq m'3.

                                  Bias Reduction

   Since our purpose is to obtain statistically meaningful data, it is important to reduce the
biases in selection of houses measured as much as possible.  Since these measurements are
purchased by the householder, the potential for bias is very great.  Bias reduction is done by
use of the questionnaires.
   One obvious source  of bias is that poor people  are less able to afford purchase of  a
measurement. To overcome this, the questionnaire includes questions about socioeconomic
status, and data can be stratified on that basis. Another source of bias is that houses with
high radon levels are much more likely to be remeasured. The questionnaire asks*  "Has the
radon level in this house  been measured previously?"  The  ratio of mean radon levels for
yes/no answers is 1.93 [1-61]; here and elsewhere in this paper, the first figure is for the main
living areas of houses, and the figure in  brackets is for basements. The expected effect is
clearly evident, and it has been eliminated by requiring a "no" answer to the question. Our
study is therefore restricted to first measurements on houses.
   Another obvious source of bias is that a person is more likely to purchase a measurement
if a nearby neighbor has a high radon level. The questionnaire therefore asks "How far away
is the closest house you know to have a high radon level". The ratios of mean radon levels for
distance D to distance 16 km or more (or don't know) are: 0=0-1.6 km,  1.88 [2.11]; D=1.6-8
km, 1.20 [1.26]; D=8-16  km, 1.03 [1.15]. Based on these results, data were eliminated if 0 is
less than 8 km.
   The remainder of this paper  deals with analysis of the 70,000 measurements described
above.  We will present  mean radon levels for various situations.  This  raises the question
of whether to use geometric or arithmetic means. A problem with the latter is that it is
heavily influenced by very large  measurements which, since  they occur  infrequently,  intro-
duces "random" fluctuations into the results. This is avoided by using geometric means. The
latter would be heavily influenced by measurements close to zero which have large statistical
uncertainty but this problem is avoided by setting a lower limit of 11 Bq in~3 (0.3 p Ci IT1);
any measurements below that were  set equal to this lower limit. Mean radon levels will be
presented, in units of the widely  accepted world mean for houses, r0, where r0 = 37 Bq m~3
(1.0 p Ci L-1).
*This question was suggested to the author by David Gur.

                              Questionnaire Items

   The questionnaires were changed frequently, but the most persistently asked questions
(names for later reference given in capital letters) were:

- LEVEL -  "Is the room where test was made (a) mostly below ground level, (b) lowest floor
     above ground, (c) second floor above ground, (d) third or higher."

- ROOM -  "Type of room where test was made (e.g., livingroom, bedroom)	"

- AGE -  "Age of house	(If not known, best guess)"

- WEATHERIZE - "How much has been done since 1975 to reduce heat loss from your
     house by weather-stripping, closing gaps under doors, sealing windows, etc.?
     	much,	little,	nothing"

- DRAFTY - "How drafty is your house compared to other houses in your area?
     	 more than average, 	 about  average, 	less
     than average"

- LOCATION - "The location of the house is (check one):
     	urban,	suburban,	rural"

- POLLUTION - "Is the air pollution in your area (check one):
     	less than average,	about average, 	worse
     than average?"

- VALUE -  "What is the approximate market value of your house?"	

- INCOME - "Annual household income?" 	

- EDUCATION - "Head of household's years of formal education beyond eighth grade (for
     example, high school graduate = 4)?" 	

- CIGARETTES - "How many cigarettes per day are smoked by all members of your house-
     hold combined?"  	
- OWN - RENT -  "Do you own or rent this house?" 	


    The analysis consists mainly of determining mean radon levels for each response to each
questionnaire item for the total file and for each geographic area, and also for each pair of
responses which we refer to as cross correlations. Since complete tables of cross correlations
would be far too voluminous for this paper, these results will mostly be presented through

   In Tables 1-3, for the total file the first entry (in parenthesis) is the number of responses
in hundreds, the second entry is  the mean radon level in units of r0, and the third entry is
the ratio of the mean radon level to the one designated l.QO. For the separate geographic
areas, the second of these entries is omitted.

Level and Room
   Table 1 gives the results for LEVEL and ROOM. It is evident that basements which are
mostly below ground have about twice the radon level of first floor living areas. Levels on
second floors are a few percent lower than on first floors. Only 4% of our measurements are
from above the first floor, and only 0.2% are from above the second floor but, while statistics
are poor,  they are consistent in  indicating substantially lower  radon levels on third floors
and above. High rise apartment dwellers are clearly grossly under-represented in our data.
   The results on ROOM types in Table  1 can be better understood with the aid of the
cross-correlation with level,  shown in  Table 4, with the last column giving the percent of
cases where the room was mostly below ground.  We see that the reason why family rooms,
bedrooms, and halls in Table 1 have higher average radon levels than living rooms, dining
rooms, and kitchens is because  they  are more frequently below ground.  This represents
another type of measurement bias: it seems probable that far less than 27% of U.S. bedrooms
are below  ground,  but that below ground bedrooms are an incentive to purchase radon
measurements. The residual room-to-room variations in Table 4 may be due to variations in
the fraction of the room that is  below ground; for example,  "mostly below ground" rooms
may average 60% below ground  for living  rooms vs.  90% for basements.  Another factor is
that below ground rooms that are  used as living areas are less likely to have cracked walls
and floors.
    There have been claims  that kitchens are not a suitable place for radon measurements.
The fact that average levels are similar in living rooms, dining rooms, and kitchens in Tables
1 and 4 is strong evidence to the contrary.
    Results for other  questionnaire items  are listed separately for living areas  in Table 2
and for basements in Table  3. The first row of these gives mean radon levels for the entire
data files - 1.68 for living  areas and  3.34 for  basements - and also shows the number of
measurements in each  geographic area.

    The results on AGE (of houses) indicate that radon levels increase during the first year.
Houses 1 - 9 y old have the highest levels, and levels decrease monotonically with increasing
age up to  80 y old where they are only about  | as high.  For houses more than 80 y old,
radon levels increase.  The  decrease out to 80 y and the increase beyond is observed for
both living areas and basements; for all geographic areas; for all degrees of weatherization,
draftiness, or pollution; for urban, suburban, or rural locations; for houses of all value above
$40,000; for occupants of all EDUCATION, and of all incomes above $15,000 per year; and
for both owner-occupied and rented houses. The only exceptions are houses valued at less
than $40,000,  or with occupant annual incomes below $15,000.  For these, radon levels seem
to increase slowly with AGE beyond 5 y.

   Data on the former are shown in Table 5. The bottom row, which gives the ratio of radon
levels for age 0-9/age 50-79, shows that this ratio increases monotonically from the lowest
to the highest value houses.  Also notable from Table 5 is the fact that expensive houses are
predominantly new while low cost  houses are predominantly old, with a steady progression
between these extremes as the value changes.

   The WEATHERIZE results in  Tables 2 and 3 yield the surprising conclusion that
tightening houses to conserve fuel has done essentially nothing to increase radon levels except
perhaps in the SC-W area. With that exception,  the most consistent finding  is that houses
where "little" has been done to weatherize have lower radon levels than those that have done
"nothing" or "much". It is difficult to imagine an explanation for that behavior.
   The ratio of mean radon levels for houses where "much"/"nothing" has been  done is
1.06 for living rooms, 1.06 for dining rooms, 1.00 for kitchens,  0.97 for family  rooms, 1.07
for bedrooms, 1.08 for halls and 0.99 for basements; there are no large differences among
the various rooms. It is 1.02 [1.06]  for houses less than  10 y  old, increasing  with age to
1.27 [1.21] for houses 50-79 y old, which seems reasonable  since new houses would not have
as many cracks,  but it drops back to 1.12 [1.10] for ages over 80 y.  It  is 1.21 [1.09]  for
urban houses, 1.07 [0.98] for suburban, and 0.98 [0.95] for  rural; perhaps weatherizing rural
houses includes measures which prevent entry of radon from the ground(?). For houses that
are more drafty than average, average, and less drafty than average, the "much"/"nothing"
ratio is respectively 1.23 [1.67], 0.98 [0.96], and 0.93 [0.84]. This means that weatherizing
increases radon levels only in exceptionally drafty houses, where  the weatherization probably
consisted mainly of sealing openings. In houses that are  not drafty, this was presumably not
done and perhaps insulation was added which might act to reduce radon entry. Of course,
the judgement of  how drafty the house is  refers  to the time of the measurement which is
after the weatherizing has been completed.
   The "much"/"nothing" ratio decreases monotonically with increasing values of the house:
1.26 [1.01] for less than $40,000; 1.14 [1.08] for $40-75,000; 1.01  [0.97] for 875-130,000; 0.93
[0.97] for $130-200,000; and 0.97 [0.88] for over $200,000; apparently weatherizing does most
to reduce air exchange with outdoors in lower cost housing. A similar pattern for the ratio
appears in the cross correlation with income:  1.15 [1.12] for incomes less  than $15,000 per
year; 1.04 [1.04] for $15-25,000; 1.07 [1.01] for $25-45,000; 0.99 [0.97] for $45-70,000; and 0.90
[0.92] for over $70,000. For owner-occupied houses, the  ratio is  0.98 [0.94] vs. 1.19 [1.14] in
rented houses.  In summary weatherizing has increased  radon levels by  15-20% in low cost
houses, occupied  by low income families, but has  actually  reduced radon levels in expensive
houses owned and occupied by high income families. Perhaps wealthier people weatherize by
improving insulation which does not affect air exchange, whereas poorer people seal cracks
which inhibits air exchange and thereby increases radon levels.
    It is interesting to  note  that the fraction of respondents who did much/little/ nothing
to weatherize is about the same for all income levels, roughly 2/2/1. One might think that
high income people would have done more because they are better able to afford the costs

of weatherizing, but on the other hand they are also better able to afford the cost of heating
   Only 20% of our respondents said they have done nothing to weatherize their houses since
1975, whereas typical estimates are that about 50% of all houses have been weatherized. This
can be explained by the fact that people who purchase radon measurements are untypically
interested in home improvement.

   The correlations with how drafty a house is are large and consistent in Tables 2 and 3,
as is expected from the fact that drafty houses have more air exchange with outdoors where
radon levels are much lower  than indoors.  The ratio of mean  radon levels in houses  less
drafty/more drafty than average is 1.56 [1.47] for our total files, and for geographic areas it
is 1.58 [1.48] in NE (-NJ),  1.47 [1.38] in NJ, 1.48 [1.34] in SE, 1.59 [1.36] in MW, and 1.57 in
SC-W. It is 1.39 in living  rooms, 1.40 in  kitchens, 1.76 in family rooms, 1.64 in bedrooms,
and  1.46 in basements, with  much of this variation explainable by statistics.  There is no
clear trend with age of the house, although there may be a peak at 10-19 y where the ratio
is 1.80 [1.68], and the effect seems to be reduced for houses over 80 y old, ratio = 1.09 [1.22].
There does seem  to be a dependence on  weatherization activities; for houses where much,
little, and nothing has been  done to weatherize,  the  ratio is respectively 1.44 [1.12], 1.55
[1.40], and 1.91 [2.23].  As suggested above, perhaps  weatherizing drafty houses consisted
largely of sealing openings, which did increase radon levels.
   For urban, suburban, and rural houses, the ratio is respectively 2.08 [1.55], 1.52 [1.41],
and  1.49 [1.53], which seems to indicate that draftiness plays a more important role in urban
homes. There is no consistent trend in  the ratio with socioeconomic factors.
   The ratio is substantially  higher for owner-occupied than  for rented houses, 1.58 [1.41]
vs.  1.16 [1.25]. Perhaps rented houses have more very small  openings which contribute in
an important way to air exchange rates without affecting draftiness.  Only 4.5% [4.0%] of
owner-occupied houses, vs. 17% [17%] of rented houses were judged to be more drafty than

   The data  in Tables 2 and 3 clearly  indicate that rural houses have substantially higher
radon levels than urban and suburban houses. One possible explanation is that rural areas
are more windy, leading to stronger chimney effects which cause radon to be sucked in from
the ground. Another suggestion^ is that urban sewer drains go into  storm sewers whereas
rural drains go into the ground, leaving an open path  for radon  to come out of the ground,
into  the house. There may be other construction differences between urban and rural houses
to explain this effect.  The result is quite important since it means that urban people, who
have more lung cancer, have lower radon exposure.
   For most  geographic areas, Tables 2 and 3 indicate that radon levels are substantially
higher in suburban than in urban houses,  but a major  exception here is the Midwest  (MW).
This last is a "fluke" resulting from the fact that  urban and suburban measurements came

from different areas; urban measurements came heavily from Columbus and Dayton, OH and
Des Moines, IA, where radon levels are high, while one third of the suburban measurements
are from the Chicago area where radon levels are comparatively low.
   Since this case represents a potential serious weakness in this type of study, we consider
it in more detail with reference to Table 6 where the geogrphic area is divided into its three
principal components by zip codes. Note that the middle zip code range has  much  higher
radon levels than the other two, and  it includes far fewer suburban cases but more urban
cases (note that what is perceived as  "urban" in Iowa is not necessarily the same as in the
Chicago area).  It is easy to calculate averages from the data given, and in doing so one finds
that, although the urban/suburban ratio (last column of Table 8) is less than unity for each
of the three components, it is substantially greater than unity when all three are combined.
   On the other hand, it should be recognized that when the whole nation is considered,
with its multitude of variations in types of urban and suburban  areas located in various
geologic settings, the correct conclusion, that suburban areas have higher radon levels, does
come through, and it also comes through in the majority of individual geographic areas.
   While the question of urban vs. suburban radon levels involves some complications, there
can be no question but that rural houses have appreciably higher radon levels than  either.
For our entire data file, the rural/suburban ratio is 1.31 in living rooms, 1.26 in dining rooms,
1.28 in kitchens, 1.29 in bedrooms, 1.34  in halls, but 1.47 in basements and 1.49 in  family
rooms. It is 1.45 for  rooms mostly below ground vs. 1.28 on first floors and 1.24 on second
floors.  This may indicate that the reason for  the difference is in the basement source.
   The rural/suburban ratio vs.  age of the house is 1.30 [1.26]  for age 0-4, 1.23 [1.39] for
age 5-9, 1.23 [1.44] for 10-19, 1.23 [1.43]  for 20-29, 1.24 [1.47] for 30-39, 1.26 [1.38] for age
40-49, 1.53 [1.48] for age 50-79, and 1.50 [1.25] for ages over 80 y. It is 1.26 [1.46] for  houses
that have done much to weatherize, and 1.27 [1.48] for those that  have done nothing. It is
1.29 [1.31] for drafty houses vs.  1.27 [1.42] in houses that are less drafty than average. We
see in Table 7 that the ratio increases with increasing value of the house; expensive rural
houses have much higher radon levels than expensive suburban houses, but there is far less
difference for low cost houses. However, there is no indication of such a trend with income:
progressing from the lowest to the highest income bracket gives rural/suburban ratios  1.24
[1.51], 1.35 [1.45], 1.30 [1.46], 1.28 [1.56], 1.28  [1.41]. The ratio is somewhat higher for rented
than for owner-occupied houses, 1.51  [1.63] vs. 1.33 [1-44].

   The results in Tables 2 and 3 indicate that there is  a clear negative correlation between
air pollution and mean radon levels.  High air pollution regions  have lower levels than low
pollution regions for each geographic section of the nation in both living areas and basements.
It is difficult to concoct  a direct cause-effect explanation; the most obvious one would be
that in polluted areas people do not open their windows as much, but that would increase
indoor radon levels, contrary to our finding. An obvious confounding relationship is that air
pollution is highest in urban areas where we have found radon levels to be low.
   The cross-correlation between LOCATION and POLLUTION is shown in Table 8.  It is
true that the great majority (71%) of rural houses are in regions of below average pollution,

while most suburban (62%) and urban (64%) houses are in areas of "average"  pollution.
Moreover 9% of urban houses, vs. only 4% of suburban and 1% of rural houses are in "above
average" pollution areas.  But still, among rural houses only, or among suburban houses only,
or among urban houses only, low pollution areas have higher radon levels than high pollution
areas by substantial factors both in living areas and in basements. One might assign part of
the explanation to varying judgements on what is high or low pollution, or what is urban or
suburban, or what is suburban or rural.  But even suburban and rural highpollution areas
have substantially lower radon  levels than low pollution urban areas. This would  seem to
imply that the negative correlation between air pollution and radon is not due only to urban-
rural effects.
   The low/high pollution ratio of mean radon levels is 1.43 in basements, 1.31 in first floors,
and 1.19 on second floors. This might be interpreted as indicating that the difference is due
to basement entry.  There is little  variation  in this ratio with room  type (other  than it is
somewhat higher in  basements and family rooms),  with age of the house, with how  much
has been done to weatherize, or with how drafty the house is.  There is also no systematic
variation with any of our socioeconomic factors, value of the house, household income, or
education of head of household. It is higher in rented than in owner-occupied  houses, 1.83
[1.53] vs. 1.23 [1.37]. In summary, radon levels are substantially higher in low pollution than
in high pollution  areas, there is evidence that this is not simply due  to urban-rural effects,
but there is no clear hint from our studies of what these other causes might be.

   The results in Tables 2 and 3 seem to indicate that  houses valued at less than  $40,000
have about 15% lower radon levels than others, and there is a tendency for expensive houses
to have lower levels than intermediate value houses. Mean radon levels in the cheapest houses
are lower than in any other value category in 4 [3] of the 5 geographic areas and are second
lowest in the remaining 1 [2] categories.  They have the lowest mean radon level in kitchens,
bedrooms,  and basements, but the most expensive houses have the lowest in living rooms,
dining rooms, family rooms and halls. Explanations for this are difficult to concoct.
   The cross correlation between value and age is shown in Table 5.  We see there  that for
living areas of houses less than 30 y old, the least expensive houses have substantially lower
radon levels than the others, but for houses more than 30 y old,  the least expensive houses
have the highest radon levels.  For basements, they have the lowest levels up to age 20, and
the second highest levels above age 50.  This switch-over can perhaps be explained by the
fact that older low-value houses were often originally constructed as expensive houses  but
their value deteriorated by aging or by  changes in the neighborhood.
*t This suggestion was offered by David Gur.

   About 10% of the "more drafty than average" houses, but only 3% of the "less drafty
than average" houses are valued at less than $40,000.  Among the more drafty houses, radon
levels are nearly constant but  declining  with increasing value of the house, but for houses
of average  or lower draftiness, radon levels are lowest for the least and most expensive,
and highest at intermediate value.  About 13% of the urban houses, but only 2.5% of the
suburban and 6% of the rural houses are valued below $40,000.
   The cross correlation between LOCATION and VALUE is given in Table 7. We see there
that lowest value houses have higher radon levels than expensive houses in urban areas, but
the former  have by far the lowest levels in rural houses. Intermediate value houses have the
highest radon levels for all locations.
   The cross correlation between VALUE and INCOME is given  in Table 9. As expected
we see that low income people live in  low value houses and high income people live in high
value houses with a steady progression between. But for each income level independently,
intermediate value houses have the highest radon levels.
   For owner-occupied  houses, radon levels are lowest for the least  and most  expensive
houses with a peak at about $75,000, but for rented houses low value houses have higher
radon levels and these levels decline steadily with value.

   Table 2 indicates that houses of lowest income people have about 15% lower radon con-
centration  than others, with no clear  trend for incomes above $15,000 per year. According
to Table  3, there is essentially  no variation of radon level with income for basements.
   The ratio of radon levels for annual incomes <$15,000/>$70,000 is 0.87 for living rooms,
0.70 for dining rooms, 0.92 for kitchens, 0.95 for family rooms, 0.69 for bedrooms, 0.87 for
halls, and 0.94 for basements.  The cross- correlation between INCOME and AGE is similar
to that between VALUE and AGE  (Table 5), with low income families having lower radon
levels in  new houses  and higher levels in old  houses, but the differences are less  sharply
defined here.
   The ratio of radon levels for <$15,000/>$70,000  is 0.83 [1.04]  for those who have done
"much" to  weatherize, 0.76 [0.93]  for those who have done "little" and 0.65 [0.85] for those
who have done "nothing".  This monotonic  trend confirms our  previous conclusion that
weatherizing increases radon exposure for poor people much more than for rich people.  This
ratio is 0.82  [0.84] for drafty houses,  0.78 [0.95] for houses of average draftiness, and 0.93
[1.02] in  "less drafty  than average" houses. Apparently the  effect  of income  is  largest  in
drafty houses, and almost disappears  in  houses that are not drafty.
   This  ratio is 0.84 [0.96] for urban, 0.76 [0.90]  for suburban,  and 0.73  [0.96] for rural
houses, and it is 0.89  [0.96] for owner-occupied houses vs. 0.72 [0.82] for rented houses.
   From Table 9 we note that  for expensive houses radon levels increase monotonically with
increasing income, but for houses valued at less than $40,000, averaging between living areas
and basements there is a slow monotonic decrease with increasing income. Low income people
living in  expensive houses and high income people living in cheap houses have substantially
lower radon levels than people living in houses matched to their income.

    The questionnaire item on head of household's years of formal education beyond eighth
 grade was apparently widely misunderstood because the majority of responses correspond
 to post-baccalaureate college education. Perhaps it was interpreted as total years of formal
 education.  Nevertheless, the data do provide some information on trends of mean radon
 levels vs.  education on a relative scale.
    There is some indication in Tables 2 and 3 of a slight trend for more educated people to
 have higher radon levels.  The ratio of mean radon levels for most/least educated is 1.03 in
 living rooms, 1.09 in dining rooms, 1.03 in kitchens, 1.15 in family rooms, 1.04 in bedrooms,
 1.05 in halls, and 1.14 in basements.
    The least educated  have the lowest radon  levels, averaging between living  areas and
 basements, for houses of all ages, although the differences are by only a few percent for houses
 over 40 y old.  They have the lowest radon levels for each response to  the WEATHERIZE
 question,  for each response to the DRAFTY question, for each response to the LOCATION
 question,  for each response to the  POLLUTION question, for each but the lowest  VALUE
 of the house response, for all but one INCOME bracket, and for owner-occupied - but not
 for rented-houses. One reason for this unusually high degree of consistency is that there are
 reasonably good statistics for all responses to the EDUCATION question.  The conclusion is
 that homes in which the head of household did not finish high school have about 10% lower
 radon levels than average.

    The results in Tables  2 and  3  clearly  indicate that households with cigarette smokers
 have substantially lower radon levels than the others, but for some strange reason it seems
 like the difference decreases with increasing number of cigarettes smoked.  It should also be
 noted  from those tables  that only 17% of all people who purchased radon measurements
 have smokers in their households, whereas 33% of American adults  are smokers.
   The ratio of mean radon levels  between non-smokers and the average of the three cate-
 gories of smokers (CIGARETTES=0/>0) is 1.13 in living rooms, 1.19 in dining rooms,  1.13
 in kitchens, 1.06 in family rooms,  1.11 in bedrooms, 1.08 in halls, and 1.12 in basements.
 Our earlier indication that this ratio is much larger in living rooms and dining rooms (Cohen
 and Gromicko 1988) proved to be misleading.
   The ratio is 1.06 [1.07] for houses of age 0-4, 1.12 [1.12] for age 5-9, 1.10  [1.10] for age
 10-19, 1.06 [1.10] for age 20-29,1.08 [1.11] for age 30-39, 1.08  [1.08] for age 40-49, 1.10 [1.07]
 for age 50-79, and 1.03 [1.12] for age over 80. Clearly there is little or no correlation with age
of the house. There is similarly little correlation with how much has  been done to weatherize
and with whether the location is urban, suburban, or rural.  For drafty houses, the ratio is
 1.30 [1.15]  whereas  for average and less drafty houses it is respectively  1.03  [1.21] and  1.11
 [1.07]. 7.7% [5.0%]  of smokers vs 4.3% [3.8%] of non-smokers reported their  houses to be
excessively drafty.
   There  does  seem to  be a correlation with value of the house and income.  The 0/>0
ratio is 1.32 [1.30] for house values below $40,000, 1.17  [1.18] for $40-75,000, 1.11 [1.01] for


$70-130,000,1.10 [1.10] for $130-200,000, and 0.95 [1.03] for > $200,000, and it is 1.22 [1.22]
for incomes below $15,000 per year vs.  1.10  [1.10] for annual incomes over $70,000. The
ratio is about the same for owner-occupied as for rented houses — 1.14 [1.13] vs. 1.16 [1.10].
   In summary, houses of cigarette smokers have about 15% lower radon levels than houses
of non-smokers, with the effect almost twice as large for low income families living in low
value houses, and almost disappearing for high income people living in high value houses.

   The results listed in Tables 2 and 3 clearly indicate  that owner-occupied houses have
higher radon levels than rented houses; the own/rent ratio is 1.28 [1.14]. This is important
because rented houses are grossly under-represented in our data base.
   Cross correlations are hindered by the small numbers of rented houses, leading to rela-
tively large statistical uncertainties.  The own/rent ratio is  1.21 for living rooms, 1.23  for
dining rooms, 1.31 for kitchens, 1.16 for family room, 1.23 for bedrooms, 1.50 for halls (with
poor statistics), and 1.14 for basements.  There is little systematic variation with age of
the house except that for houses over 50 years old basement ratios fall below  1.00. The
ratio  is 1.24 [1.12]  in  houses where "much" has been done to  weatherize, vs.  1.50 [1.36]
where "nothing" has been done. Apparently weatherized rented houses are more like owner-
occupied houses. The ratio is only 0.99 [1.07] in drafty vs.  1.35 [1.21] in  "less drafty than
average" houses.  The ratio is 1.48 [1.20] in urban, 1.28 [1.17] in suburban, and 1.13 [1.04] in
rural houses, a monotonic relationship. It is 1.08 [1-19] in areas where air pollution is below
average vs.  1.60 [1.34] where pollution is above average, a rather large difference which is at
least partly related to the urban- rural difference.
   The own/rent ratio is below  unity, 0.82 [1.06], for houses valued at less than $40,000,
but increases with value to 1.16 [0.98] for $40-75,000, 1.38 [1.14] for $75-130,000, 1.21 ]1.52]
for $130-200,000, and 1.48 [1.27] for >$200,000.  The cross correlation with income seems to
behave differently, 1.39 [1.38] for income <$15,000 per year, 1.08 [1.12] for $15-25,000, 1.36
[1.07] for $25-45,000, 1.56 [1.31] for $45-70,000, and 1.13 [1.19] for >$70,000.  It is difficult
to reconcile this difference except as statistical fluctuations. There is no systematic trend in
the own/rent ratio vs. education of head of household.
   In summary, the own-rent ratio seems to vary systematically with several factors but  the
reasons for these variations are obscure. Perhaps many of them are due to  poor statistics as
there are less than 100 rented homes in most categories.


Cohen, B.L.; Nason, R. A diffusion barrier charcoal adsorption detector for measuring radon
concentrations in indoor air. Health Phys. 50: 457-463; 1986.

Cohen, B.L.;  Gromicko,  N. Variation of radon levels in U.S. homes with various factors.
Jour. Air Polltuion Control Assn. 38: 129-134; 1988.

Cohen, B.L.; Performance characteristics of DBCA radon detectors. Had.  Protection mag-
net.  5:47-54; 1988.


   Cohen, B.L.;  Cohen, F.B.  Error prevention at a radon measurement service lab.  Rad.
   Protection Mngmt.  6:43-47; 1989.

   Cohen, B.L.; Seasonal variations of radon levels in homes. Health Phys. - submitted; 1989.
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.

Table  1: Mean radon concentration vs  level with respect to ((round and  v«  room i\pv
     for various geographic areas of the United States  -living" refers to the rooms luted
     other  than •basement1'  Geographic areas refer to up codes-  NJ  07000-0899!). M:
       0-19999. bK - 20000-39999, MW - 4000-69999. SC + VV   70000 99999  Under • I'S
     total" column, the figure in parenthesis is number of measurements (in  hundreds), the
     next number is the mean radon level in units of r0 and the third number is the ratio of
     the mean radon level to the one  marked LflQ in  the game column  in other columns.
     the second of these numbers is not included
below ground :
lit above grd :
2nd above grd :
)rd or higher •
living rood
dining roooj
fully rooei
klcehon :

(251)- 3 39-1 96 •
( 3l)-2 86-1 65 :

( S)-2 43-1 40 :
(1S7}-1 73-1.00 •
( 20) -1 (2- 94 '.
( 1)-1 04- 60
(298)-J 32-2.27 .
(106)-1 46-1.00 :
( 26)-l 40- 96 :
( 60)-1.«3-1.25 .
( 80) -2 05-1 40 .
( 12)-1.77-1.21 .
( 50) -1 46-1.00 :
NE (-NJ)-

(92) -2 09 •
( 8)-l 68

( 3)-l 41 .
(47).JJ)fl :
( 7)- 91
( 3)- 51 .

(104) -2 44 .
( 33).JJffi. .
( 8). .92 :
( 13)-1 14 .
( 20) -1 70 •
< 3)-l 17 .
( 14). 1 02 .

(47)- 2 12 :
{ 2)-1.62 •

( 1)-1.«4 !
(23)-UUl .
( 4). 94 .
( 2)- 70
(70). 2. 46 :
(28)-.U|Q. :
( 8)- 98 i
(12)-1.21 :
(22)-l J4 •
( 3)-l 24 :
(14). 1.07 :

(31)- 2. 02 .
( J)-1.J6 •
( 1)-1 46
( *>• 99
( 2)- .63 .

(40). 2. 38
( 4). 1. 01 •
(14). 1 28 :
(IS)- 1.44
( 2)-l 22 •
( 8)-l 08 '.

(69) -I 77 .
(ll)-l 4J :

( 1)-1 22 •
( 3)- 88 •
( 2)- 54

(72). 1.92 .
( 4). 1.01 :
(ll)-l 33
(17)-1 26
( 3)-1.29 !
( 8)-l 04

( 8)-2 52
( 4). 2 07

( 5)-l 00
( 2)-l 16
( 2)- 72

( 9). 2 30
( 2). 96
( 9)-l 17
( 0-1 25
( I)- 95
( 5)- 89

Table 2: Mean radon  levels from measurements in living areas for various responses to
     questionnaire items See caption (or Table I

SO -79











(4). 1. 61-. 84
(19). 1.47. .77
(33J-1. 22-. 64

(129J-1 76-1 02
(147M.60- 92
(63). 1.73-1. 00

(11). 1 38- IS
(67)- 2. 17-1. 33

(137). 1.69-1.00
(67)-2. 16-1731"

(89) -2. 09-1. 24
( 9). 1.64-79T"

(11). 1. 56-. 86
(61)- 1.82-1. 10
(IQH-l. 81-1.00
(36). 1. 61-. 89

( 9)-1.39-.85
(28)- 1.72-1. OS
(122)-1 64-1.00

(35). 1. 61-. 99
HB1- 1.65-1. 00
(172)- 1.76-1. 07

(12)-1.50- 87
( 34). 1. 55-. 90
(18)- 1. 62-. 94

( 6). 1.49- 1.00
HE (-NJ)
(105)- 1.74
(20) -.95
(16) -.81
(14).. 87
( 6)-. 71
(11)-. SB
(12)- .67
(37)- 87
(42)-. 84
(17). 1.00
( D-.80
(48) -1 00
(20). 1.26
(41). I ' 00
(21). 1.11
(27)- I. 11
( 2)T7?
< 4). .69
(20)-. 94
(28). 1.00
(24).. 76
(12)-. 7S
( 3)- .84
( 9)- 1.01
(26)- 1.07
(12)- 1.10
( 4). 82
( 8). .92
( S>-.96

(2) -.96
(21)-. 96
(18) -.82
(13)-. 72
( 5)-. 75
( 8)- .58
( 5)-. 88

(24).. 99
(16)-. 91
(14). 1.00

( 2)-. 92
(40). 1.00
( 7)- 1.35

( D-.69
(22). 1.00

(IS)- 1.00
(.6)-. 92

( 2). .94
(15). 1.00
(14). 1.21

( D-.74
( D-.89
(15) -.97
(34). 1.00

( 8). 1.00

( 3)- .91
( 6). .89
( 4).. 98

: (17)- 1.92

(71)- 1.75

(.7)- .72
(12)-. 96
( 7). .82
( D-.69
( 4). .61
( 2)-. 71

(25)- 1.06
(27). .98
(16). 1.00

( 2)-. 86
(11). 1.27

( 4)-. 84
(29)- 1.00

(14)- 1.12
( 2)-78T'

( 2). .71
(11). .85

( D-.74
(30)- 1.00

(10)-. 90
( 5)-. 97
(12)- .99
(40) -1.01

( 2)-. 85
( 7)- .87
( 4). 90


(67). 1.84

(.I)-. 82
(15)-. 91
(11)-. 86
( 4)-. 82
( 7)-. 71
( 4).. 81

(27)- 1.07
(28). .96
( 9). 1.00

( 2). .82
(17)- 1.30

( 9). 1.18
(3D- 1.00

( 2>.T7T

(22)- 1.00
(10)-. 91
( D-.84

( D-.74
( 6)- .95

( 7)-. 96
(10)-. 97
(11). 1 17

( 2)-. 86
( 3)-. 90
( 2)- .77

(34). 1.21
: ( 2)-U|0.

(41). 2. 06

(.2)-. 90
(14). 1.00
(10).. 94
( 6)- .81
( 4)- .91
( 2)-. 86
( D-.H
(11)-. 97

(ISM 21
(14). 1.05
( 7). 1.00

( I)-. 76
(17). 1.00
(11) -1 19

( 6). .89
(ll)-l 00

(14)- 1 26
( 9)-1.00
( D-TTTf

( 2). .81
( 9). .96
( 2)-!s8

( 2)-. 90
( 5)- 1.25
(10) -1 00
( 4).T7HT

(10)-. 91
( 3)-1.02
( 5)- 99
(19). 94

( 1M OS

<19)-1 11
• ( D-.LP£

Table 3: Mean radon levels from measurements in basements for various responses to ques-
     tionnaire items See caption for Table 1

laia than av :
•ere than av
S4SK-$70K .

(32S)-3 34
( 3)-) 23-. 77 :
(80)-4.18-1 00 •
(6D-3.74. 89 .
(55)-3 19-. 76
(46)-2 67- 64
(20)-2 72- 6) •
(37)-2 69- 64 •
(21)-3 40- 81
(116)-3 49- 99 •
(136)-3 14- 89
(49)-3 51-1.00
(ll)-2 71- 86
(178)- 3 14-1 00
(80)-3 96-1 26
(31)-3 10-1.02 •
(166)-3 05-1 0* •
(100) -3 96-1.31 :
( 9)- 2 85- 94
(1D-2.96. 83 :
(51)-3 40- 95
(8D-3 S8-JLJ20.
(7})- 3 30- 92 .
(56)-3 28- 92
( 6)- 3. 30- 96 :
(19)- 3. 34-. 97 •
(7D-3.37- 98 :
(92)-3.44-1.00 :
(56)-3.45-l 00 :
(«S)-3 00.89
(10) -3 41-1.01 .
(22j-3.38-J.jaL •
(23). 3 54-1. OS
(215)-3 42-1.01 :
(258)- 3. 42-1. 00
(34)-1.55-.90 •
(16)-3 07-. 85 .
(203) -3 37-1.14 :
( 5)-2.95-1.00 •
HE (-NJ)
(113J-3 69
( 9). 64
(26) -1 00
(19)- .92
(18)-. 81
(16)-. 69
( 7)- 66 .
(14)- 62
(U)- 72 •
(41). 94
(48)- 88
(16) -1 00
< 5)- 82
(62) -1 00
(30)- 1 21
( 9)-. 82
(61)-1 00
(36)-l 18 •
(53)-l 00
( 3). 95
( 4). .64 •
(19)- 89
(23)- 81
(19)- 76
( 2>- 86 .
( 8). .86 •
(27)-1.00 :
(M^TT .
(25). 83
(4)-. 96
(7)-l 07
(73)- 97 .
(88)-l 00
( J)> 85 .
(10)- 95
( 5)- 91
(74). 1 24
( 2)-l 00

(75)- 2. 97 •
(19)-1 00
(14)-. 84
(15)-. 64
(10)-. J4
( 4)..J4
( 8)- 48
( 4)- 71
(21)-. 87
(32)- 82
(12)-1 00 •
( 2)- 91 •
(40) -1 00 •
(12)-1 26 •
( D-.64 :
(35). 1 00 :
(14) -1.84 :
(20). 1.49 •
(26) -1 00 •
(.!)•!. 06 :
( 9)-. 82 :
(8)-l 00 •
(26)-l 33
(.»)•!. 07 :
( 2)-. 88 :
(10). 90 :
(24)-1.00 :
(17) -TUT:
(19) -.87 :
( 3)- 91 .
(6)-UBL •
(7)-1.0i .
(44)-1.00 •
(60)-1 00 •
( 2)-. 92 :
( 6)-. 89 •
( 3)-1.01 •
OS)- 1.23 •
(.7)-!. 00 :

(45)- 3 47:
(.7)-.9S :
(17)-1 00:
(10)- 1.05:
(6). .81 :
( *)-.59
( 3). 65 :
< 3). 61 :
( I)- 85
(15)-. 97 •
(17)- 90 •
(9)-l 00
( D-.94 •
(24). 1.00.
(10)-1 26
( 3)-. 76 :
(22)-l 00-
(8)-1.63 :
(20)-l 00:
( D-.75 :
(.4). .84 :
(4). 1.01 .
(1S)-1 00-
(»)•!. 13.
( 6)- 94
(.3)-!. 14
( 2)-1.07
(9)-. 97
(17)- 1.00
( 8).TW~
(6) -.86
( 2)-1.12
( 6). .99 :(.2)- 53
(26)- 1 01:( 2)-. 68
<2S)-1.00:(5M 00
(9)-~9f-( 2>T"P
( 4)- 87 •( 7)-l 12
( 2)-. 99 :(.2)-.94
( 7)-1.06:(.6)- 90
(235-1.01: (3)-. 99
(10).798~:( 2FTTS
(ll)-l.lO: (2)-. 76
( 2)-1.09:(.5).l 14
t5).1.00 :( 1)-1 00
(4).l 07 .( 9)-l 03
(58)-1.21:(7)-l 14
(61)-1^00:(9)-1 00
( 2)-. 76 :(.3)- 91
( 8)-. 87 :( 7)-1.05
( 5). 80 ( 3) -.89
(57)- .96 :(6)-l 65
( 2) -1.00.: (.2) -1^00.

  Table 4. CTOM correlation between room type and level with respect to ground  Number
       in  parenthesis is number or measurements (in hundreds), and other number 11 mean
       radon level in units or r0 Last column is the percent of cases for each room type where
       level is below ground
                  BELOW GROUND
                                     FIRST ABOVE
                                                        SECOND  ABOVE
                                                                          % BELOW GROUND
( 2)-241
( })•! 40
( 1)-1 92
( 17) -2 Sk
( 9)-3 17
( I) -2 94
(25D-3 39
(42)-l 58
(10)-1 55
U6)-l 60
(22)-l 92
(18)-1 62
( 4>-l 73
( 8)-2 43
( 3)-l 54
(.8)-! 31
( 1)-1 36
(.8)-! 83
( 7)-l 69
( 8)-l 57
4 0
3 1
Table 5. Crosi correlation between age of house and value of house L and B refer to living
     areaa  and baicments respectively  See caption for Table 4  Last column u ratio of
     radon levels for value OIO.OUO/>S20U,000  bottom row is ratio of radon levels for ace
     0-9/age 50-79
• L(.9)-l 20
B( 2)-2 81
. L( D-1.31
B( 6)-2 99
L( 1)-1 42
B( 9J-3.1S
L( 2)-l 96
B( 2)-2 58
L( D-1.84
B( D-2.47
L( 3J-1.S3
B( 3)-3.00
. L< 2)-l 59
B( 2)-3.63
L - 90
B -1.09
. $40-75.000
: L( S)-2 02
B( 2)-3.83
L( 7>-l 80
: B( 3)-3 95
• L(12)-l 98
• B( 7)-3 48
• L(ll)-l 96
B( 9)-3.65
. L(1D)-1 95
B(12)-3 06
. L( 5)-1.63
. B( 5)- 3. 18
L( 7)-l 46
B( 8). 3 14
: L( 4>-l.S5
: B( 4)- 3 87
. L -1 31
: B -1 24
. $75-130.000
. L(14)-l 96
B(10)-4 32
L(16>-2 14
• B(10)-4 48
L(25)-2 OS
B(18)-4 15
L(19)-l 77
• B<16). 3 58
L(13)-l 58
B(ll)-2 74
L( 5) -I 37
: B( 5>- 2. 82
L( 7)-l 26
• B( 8). 2 26
L( 5)-l 58
B( 4)- 3. 43
L -1 63
B -1 68
: $130-200.000
• L(18)-l 79
: B(13)-4 10
. L(15)-l 91
: B(10)-4 07
• L(23)-l 78
: B(13)-3 84
L(17)-l 47
: B(12)-2 90
• L(ll)-l 27
B( 9)-2 27
: L( 4)-l 27
. B( 4). 2 42
. L( 7)-0 91
• B( 6) -2 44
• L( 5>-l 29
: B( 4)- 3 61
: L -2 03
: B -1 67
>$200.000 :
L( 8)- .86 :
B(14)- .39 :
'. U 5)- .88 :
B( 7)- .05 .
. L( 7)- 78
BUD- il
L( 5)-l 48 .
B( 8)-2.62
L( 4)- I 49
B( 6)-2 43
L< 1)-1 36
B( 2)-2 68
L( 3)-l 07 .
B< 5) -2 37
L< 2)-l 33 .
B( 3)-2 73
L -2 05
B -1 78 .
*9 au.uuu
1 18
1 17
1 44
I 36
L 17
1 04
0 83
0 76
0 95
0 74
1 09
0 70
0 79
0 84
0 75

Table 8: Breakdown of geometric medn dud arithmetic average radon levels for location
     =  urban (U), suburban (S), and rural (Ft)  Last column u ratio of U/S for previous





UI (2)
ND (.8)



, KN
. SD





IA (9).
. HT (.5) :

( 3)


309 .
1436 •
S48 .
164 •
416 .
1231 .
: 2
: A
: 1
0 93

0 99

'. 0 79

1 18

  Table 7: Cross correlation between location and value of house  Last column u ratio of
       mean radon levels in  previous two columns  See caption (or Table 5
$75.130.000 :
< 3)
( A)
( B)

( 7). 3
( 3)
( »
( 2)
( 2)
. ( 3)-l
: ( 3)- 2
: (22)-l
(23)- 3
(41). 1 86
: (43). 3. 22
: (3D- 1
• (38). 3
(24)- 1
(39). 3
: ( 3>- 3
(19)- 2
• (16) -4
: (12) -7.
'. ( «>-2
(13) -4

                      RADON IN NORWEIGAN DWELLINGS
                             12                3
                     T.Strand  , B.M.R. Green and  E.Stranden

            National Institute of Radiation Hygiene, Osteras, Norway
           2 National Radiological Protection Board, Oxfordshire, United Kingdom
           3Radforsk A/S, Fjellhamar, Norway

     The results of a large scale survey of radon concentrations in Norwegian dwellings
are reported. Measurements of radon have been made in 7500 dwellings representing all 450
municipalities. The dwellings were selected by a random sampling procedure based on data from
the Central Bureau of Statistics. The number of measurements in each municipality is proportional
to its population. The measurements were performed by nuclear track detectors from the National
Radiological Protection Board in United Kingdom. The results will be used in an epidemiological
study on radon and lung cancer.

     From 1983 to 1986 a pilot survey of radon in Norwegian dwellings was carried out (1). In the
winter seasons measurements of radon were performed in a total of about 1500 detached houses
from 79 out of the 450 municipalities of Norway. These municipalities included about 30% of the
total population. Taking into account seasonal differences in radon concentration and that
multifamily houses were not included in the survey, the year average of radon concentration in
Norwegian dwellings was estimated to be between 80 and 100 Bq/m*. Comparing these estimates
with those from other surveys in other countries, it was concluded that the overall level of radon
in Norwegian dwellings is about the same as in Sweden and Finland (2,3). Owing to weaknesses  in
the sampling procedure (no random selection of dwellings) and because the integration time in the
measurements was only 5-7 days, it was recognised that it would be difficult to use this material in
any kind of epidemiological study on radon and lung cancer.

     In 1986, financial support for a large scale epidemiological study on radon and lung cancer
in Norwegian dwellings were given from the Norwegian Cancer Society. The study was started in
early 1987 and has been conducted as a collaboration between the National Institute of Radiation
Hygiene, the Cancer Registry of Norway and the National Radiological Protection Board in United
Kingdom. The objectives and strategy of the study have been presented in an earlier paper (4).

     In this paper, some initial results of the radon survey are reported. These data are combined
with earlier radon data and information on the building stock and indoor occupancy patterns to
give a better estimate of the average radon concentration in Norwegian dwellings. The results from
the epidemiological part of the study will be presented later.
                           MATERIAL AND METHODS

     The aim of the survey is to obtain a representative average value of the indoor radon
concentration for each of the 450 municipalities in Norway for use in an epidemiological study of
radon and lung cancer. The number of dwellings in each municipality is proportional to the
population except for the two largest cities, Oslo and Bergen, where somewhat smaller samples
were taken due to the higher population density.

     Influx of radon from the subsoil and bedrock is the main source of indoor radon in
Norwegian dwellings. For an average Norwegian detached or undetached house about 90% of the
source term is  due to the influx of radon from the ground. For blocks of fiats and other multifamily
houses the picture is somewhat different: The radon concentrations are generally much lower, and
building materials, which are not usually correlated with geology, are usually the most important
source of indoor radon. Less than 15% of the population live in flats or other multifamily houses
(5). In Oslo, die largest city, about 65% of die population are living in flats. Due to die significant
differences in radon levels between houses and flats, it was necessary to stratify die sampling
procedure.          A census data base with information on die building stock is available at die
Central Bureau of Statistics. This data base was used for die selection of dwellings. The sampling
procedure is similar to that used in a large scale survey of natural 7 radiation in  Norwegian
dwellings (6).

     Because die financial resources were limited, is was necessary to limit die number of
measurements  to about 10,000. From an epidemiological point of view it was found more
appropriate to  do one measurement in  10,000 dwellings instead of two measurements in 5,000
dwellings. From an earlier study (1), die radon concentration in bedrooms was found to be on die
average about  10% lower than  in living rooms. This difference can be explained by die fact that in
a large proportion of single family houses die bedrooms are  often better ventilated and are on a
floor above die living room. However, die higher occupancy in die bedroom compared to other
parts of die house means that die measurements in bedrooms would be  a good indicator of die
relative radon exposure of die population at die municipality level.

     The measurements were performed by NRPB type nuclear track detectors (7,8). Preparation,
calibration and analyses were performed at die NRPB in United Kingdom, while die distribution of
dosemeters to die householders were organized from die National Institute of Radiation Hygiene in
Norway. Each month, for a period of 18 mondis (from  March 1987 to October  1988), about 550
detectors were issued to householders in different municipalities. The integration time for die
measurements  was 6 mondis. In order to avoid any difficulties from long-term variations in radon
concentration in die estimates of annual average of radon concentration  for each municipality, die
detectors were spread evenly over all seasons  of die year. After six mondis, die householder was
asked to return die detector and a questionnaire to die National Institute of Radiation Hygiene.
The detectors were stored in a  very low radon atmosphere (<10 Bq/m>) for a maximum of a week
at the institute before it was sendt to die NRPB for analysis.

     The detectors were sent to the householders without first asking for their agreement to take
part in the study.  It was anticipated that this procedure would increase the percentage of selected
households taking part in the study. A variety of reasons, such as detectors returned without
completed questionnaire, changes of address, an unwillingness to participate, lost detectors and so
on, meant that the final number of valid measurements was about 7,500 out of the initial 10,000,
a success rate of 75%. There was no geographical pattern in the missing fraction of measurements.
                           RESULTS AND DISCUSSION

     In figure 1, a frequency distribution of the measurements for the whole country is shown.
The distribution is found to be log-normal. The arithmetic mean of the measurements is calculated
to be 53 Bo/nP. Owing to the fact that the sampling procedure was population based and that the
measurements in each municipality were evenly distributed in time, this mean value is a good
estimation of the annual average radon concentration in Norwegian bedrooms. On an individual
level it may be necessary to correct for the seasonal variations in radon concentration. Owing to
climatic conditions, concentrations in winter are usually somewhat higher than in summer.


              700 -


              soo -
          b    ** "

          «    300
               200 -
                               RAOON CONCENTRATION (Bq/n>)

Figure 1. Frequency distribution for radon concentration in bedrooms in Norwegian dwellings.
     The frequency distribution in figure 1 has been truncated at 200 Bq/m>. However, about 4%
of the results were above 200 Bq/m* and about 0.3% above 800 Bq/m*. The highest values were
found in die eastern pan of southern Norway. In this  region of the country there are relatively
large occurrences of precambrian to Silurian rocks like alum shales and granites. Measurements on
samples of alum shale from the upper cambrian or lower ordovician period have shown

 concentrations of radium up to 4,500 Bq/kg, which is about 100 times higher than the normal
 level in rocks and soils. In an ealier survey of radon in dwellings on typical alum shale ground (9),
 more than 75% of the  dwellings had an average radon concentration in the heating season above
 200 Bq/m>. In our survey, the lowest concentrations were found in dwellings on Caledonian ground
 with large occurrences of gneisses.

      In most detached  and undetached houses,  living room and kitchen are located on the first
 floor (ie groundlevel). However, in a large proportion of the dwellings, the bedrooms are located
 one floor higher. In such houses the radon level is assumed to be somewhat lower in the
 bedroom(s) than in the rooms on  the first floor. On the questionnairies, information on "what floor
 the main bedroom is located" were recorded. In figure 2, the average radon level is shown for
 bedrooms in  the basement, on the first floor and on the second floor. On the average, the radon
 level was found to be about 20% higher if the main bedroom was on the first floor and 240% in
 the basement relative to a bedroom on the second floor. According to census data from the Central
 Bureau of Statistics (5) we may assume that the living  room are located on the first floor in all
 detached and undetached houses in Norway. Assuming that the average radon concentration for
 the category "bedroom  on the first floor* is a representative average for the radon concentration in
 living rooms  for detached and undetached houses, it is  possible to correct the annual average of
 radon concentration in bedrooms to become a more representative estimate for the average in the
 dwelling. In these estimates it is assumed that 50% of the time indoors is spent in the bedroom,
 40% of the time in  the living room and 10% of the time in the basement.  If  the bedroom  in
 located on the basement floor, it is assumed in  the calculations that 40% of the time is spent on
 the first floor. From these calculations the average radon concentration in Norwegian dwellings was
 estimated to 60 Bq/m>. This average is somewhat lower than the estimate from the  pilot study.
 This may partly be explained by the unrandomized sampling in the pilot study and partly  by the
 fact  that the pilot measurements were performed during the heating season in detached houses
Figure 2. Average radon concentration in bedroom for different categories of dwelling according to
         on what floor the bedroom is.

      In figure 3, the arithmetic averages are calculated for different categories of dwelling. As
 illustrated the average concentration in detached and undetached houses is 30-35% higher than in
 blocks of flats. The higher level of radon in detached/undetached houses was expected owing to
 the fact that the influx of radon from the ground has been found to be the most the most
 significant source of radon in Norwegian dwellings.
 Figure 3. Average radon concentration in the bedrooms for different categories of dwellings in

     In figure 4, average radon concentration in bedrooms are calculated for different categories of
 dwellings according to year it was built. All types of dwellings are included the estimates.
 The average radon level does appear to be higher in newer dwellings. This may be attributed in
 part to the "save energy campaign" in the seventies. It is to be expected that the air exchange rate
 in newer houses will be less than in older houses.
2. 60 •
i 40"
§ 20"








                                1*01  1M1  1941   It4«   1991
                                1*10  1*40  IMS   1*50   !•««  1»7«  1MO
Figure 4. Average radon concentration in bedrooms according to the year die house was
        built Percentage of the total sample are included in the histogram for each category.


     The annual average of radon concentration in Norwegian bedrooms is calculated to be
 53 Bq/m>. The radon concentrations were found to be about 30-35 % higher for bedrooms in
 detached and undetached houses than in blocks of flats. There was no significant seasonal
 dependence in the results. The highest concentrations were found in the eastern pan of southern
 Norway. However, it is assumed that the radon level in bedrooms on the average is somewhat
 lower than in the living room. In a large proportion of Norwegian houses, the bedrooms are
 located one floor higher than the living room and kitchen. By assuming that most living rooms are
 located on the first floor, by using an average factor for the ratio between bedrooms on the second
 and first floors to estimate the level in  the living room and by assuming that that an average of
 the concentration in the bedroom and living room is representative for the dwelling, we estimate
 the average radon concentration in Norwegian dwellings to be 60 Bq/m3. This is 30-40% lower
 than the earlier estimates (1).

     This work was supported by the Norwegian Cancer Society - Landsforeningen mot
     The work described in this paper was not funded by the U.S. Environmental Protection
Agency and therefore the content do not necessarily reflect the views of die Agency and no official
endosrement should be inferred.

1. Stranden, E.   Radon in Norwegian dwellings.  In: Proceedings of the Symposium Radon and its
   Decay Products: Occurrence, Properties, and Health Effects, New York, April 13-18, 1986.
   American Chemical Society Symposium Series 331, p.70-83.
2. Swedjemark, GA and Mjones, L.  Radon and radon daughter concentrations in Swedish homes.
   Radiat.Prot.Dosim. 7 (1-4): 341-345, 1984.
3. Castren, O., Winquist, XL, Makelainen, I. and Voutilainen, A. Radon measurements in Finnish
   houses.  Radiat-ProcDosim. 7 (1-4): 333-336, 1984.
4. Stranden, E., Magnus, JC, James, A.C., Green, B.M.R. and Strand, T.   Radon and lung cancer:
   an epidemiological study in  Norway. Radiat.Prot.Dosim. 24 (1-4): 471-474, 1988.
5. Central  Bureau of Statistics. Census data. Personal Communications,  1986.
6. Strand,  T., Magnus, K. and Stranden, E.  Sampling strategy for a large scale indoor radiation
   survey - a pilot project Radiat.ProtDosim.  14 (3): 251-255, 1986.
7. Bartlett, D.T., Gflvin, OJ., Still, R., Dixon, D.W. and Miles, J.C.H. The NRPB radon
   personal dosimetry service.  Jour.Radiol.Prot. 8: 19-24, 1988.
8. Bartlett, D.T. and Bird, T.V., Technical specification of the NRPB radon personal dosemeter.
   Chilton, NRPB-r208 (London, HMSO).
9. Stranden, E. and Strand, T.  Radon in an alum shale rich Norwegian  area.
   RadiatProLDosim. 24(1-4): 367-370, 1988.

     Session V:
Radon Entry Dynamics

            A Simplified Modeling Approach and Field Verification
              of Airflow Dynamics  in SSD Radon Mitigation Systems

     Kenneth J.  Gadsby,  T.  Agami Reddy, Rajiv de Silva, and David T.Harrje

                  Center for Energy and Environmental Studies
                             Princeton University
                           Princeton, NJ 08544, USA

     The airflow  characteristics in  the  subslab area  of a building  must be
known  in  order  to provide engineering  guidance  for designing  a  subslab
depressurization (SSD) radon mitigation system. An earlier study had described
our research effort  to  model subslab flows as radial airflow through a porous
media  confined  between two  impermeable disks. A  laboratory device was  also
built  to  experimentally determine  the  aerodynamic  pressure drop  versus  flow
model  coefficients  for  a variety of  subslab  gravel materials  and to test the
validity of our modeling approach.

     This  paper will address the  issue of how  the  simplified model approach
along with the laboratory-determined pressure drop coefficients can be used as
a  rational  means  of  assessing  subslab connectivity  in  actual  houses.
Preliminary  field  verification  results  in a house  with gravel  under the
basement slab  are  presented and discussed.  The  way in  which the  simplified
modeling  approach  could  be useful  to  professional  mitigators  is  described.
Illustrative  figures  of  the pressure  field extension  map from closed-form
solutions  are  presented.  Experimental  results of the  subslab pressure  field
under  single  and dual  suction  penetrations (perimeter  and  central  locations)
are  shown.  Finally, certain practical  aspects relating to  proper engineering
design of the SSD system are addressed and initial recommendations  are made.

     This  paper  has been  reviewed  in  accordance with  the  U.S.  Environmental
Agency's peer and administrative review policies and  approved  for  presentation
and publication.


     An EPA  sponsored workshop  was  held  at Princeton University  to summarize
available  knowledge on  various  radon mitigation diagnostic  techniques  (1). The
emphasis of  the  workshop  was  on  diagnostics since  each house, housing  sub-
division,  and  region may  have  different   characteristics which would  require
that special attention be paid to system design in order to maximize  mitigation
system performance  and minimize  installation cost. This  issue was of  particular
importance since it was found that a large number of mitigated  houses still had

radon levels above the recommended guideline of 4 pCi/L. In fact a recent study
(2) found that 64% of the mitigated houses  in New Jersey, where post-mitigation
measurements  have been made,  remained  above  the recommended  radon level.
Diagnostics  are  therefore  crucial  for  providing information  relevant  and
necessary to  the  successful  design  and implementation of a  radon mitigation

     A survey of the workshop participants indicated that systems based on SSD
account  for  more than  50% of  all  installed  systems.  (Another promising
technique involves subslab pressurization.  Since the two techniques are similar
in terms  of subslab dynamics of  induced airflow,  they can both be treated in
the  same  scientific framework.) In  the pre-mitigation  diagnostics phase,  the
degree of  "connectivity"  under  the slab  as  well as  the permeability
characteristics of the subslab medium must be determined before a  suitable SSD
system can  be  designed.  Proper attention to  these aspects  will  ensure that
reasonable flows,  and hence  the  desired  degree of  depressurization,  will
prevail at  all  points  under  the slab.  Lowering the pressures at all points of
the  subslab to  values  below  those of the basement/crawlspace/living area will
then greatly reduce the flow of radon-rich  soil-gas into the building.

     Parallel with the above aspect is  the  concern that presently mitigators
tend to over-design  the  SSD  system in order  to err on the safe side. In doing
so,  more  radon (a 10-fold increase has been cited in Ref.3)  from the soil is
removed  and  vented to  the  ambient air  than  would have  occurred naturally.
There is the  need  to  try to  prevent  these  overly robust SSD mitigation systems
and  therefore  decrease exhaust  emissions of  radon and conditioned indoor air,
while simultaneously  ensuring that the indoor radon levels  do not exceed the
recommended value.

     We are currently  involved  in the formulation and verification of a rapid
diagnostic  protocol  for  subslab and wall depressurization systems designed to
control indoor  radon levels   (4).  It is hoped  that the protocol will lead not
only  to  the  ability  to  distinguish between houses that  are  hard or easy to
mitigate,  but also  to  the articulation of  a more rational and scientific
approach which  would be of  special  usefulness  to the ever-increasing body of
professional mitigators.

     Our approach  to the formulation of the diagnostics protocol  consists of:
(i)  a practical component in that  specific guidelines would be  suggested so
'that the effectiveness of the engineering design of the  radon mitigation  system
would be  enhanced,   and  (ii)  involvement of  scientific  studies at a more
fundamental  level  which would  both  lend  credence and  enable defining the
guidelines  more  rationally. This paper  will  briefly  address the scientific
component of  the protocol  and present  preliminary  validation  of the previously
suggested simplified modeling approach to  predict  the  subslab pressure-induced
flow by  a single  central suction  hole  (5).  This  paper will also  show how the
closed-form solutions  thereby obtained could be conveniently  used to  generate
figures useful to the professional  mitigator.  Aspects  relating to the  proper
engineering design of  the piping system will also be addressed.

Previous Work

     Following an  earlier  study (6),  we had forwarded arguments to support the
suggestion that subslab airflow of a residence with either gravel or soil under
the concrete slab  be  visualized as  occurring in radial streamlines terminating
at the  central  suction point (5). It was  pointed out that the Reynolds number
(Re) indicates the flow regime, where Re is defined as:
          Re -  (q/A)  (l/va) (oV4)                                      (1)

     where q -  total volumetric flow rate,
           A -  cross  sectional  area of the flow (in the case of radial flow
           through a circular bed of radius r and thickness h, the area -2wrh) ,
           va - kinematic viscosity of air,
           dy - equivalent diameter of pebbles, and
           $  - void fraction or porosity of the gravel bed.

     It  was  then  shown  that subslab  airflow  under  actual  operation of
mitigation systems  is likely to be  turbulent  if  a gravel bed is present under
the slab, and laminar when soil is present (6).

     The  core  of  any model  is the  formulation  of the  correlation structure
between  pressure  drop and Re  (or flow rate).   For laminar  flow,  Darcy's law
holds and we have (7) :

           - a  (q/A)                                    (2)

where pf -  density of the flowing fluid, and
       g -  gravitational constant.

     In  the  case of  turbulent  flows,  a model such as  the following is widely
          U/Pf g)  (dp/dx) - a  (q/A)b                                   (3)

     The  left  side is the pressure  drop  in head per unit bed  length,  and the
parameter "a"  can be loosely interpreted as  the  resistivity of the porous bed
to the  flow  of the particular fluid.  The permeability k of the porous bed is
given by:

          * - ( VS) (l/a>                                              (4)

     It was  then  shown in Ref. 5 that  the following  closed- form solutions are
obtained for the pressure drop (in units of head of water) in a homogeneous bed
with a  circular boundary and with  a single  suction hole  at  the  center of the
disk, see Fig.  1:

     For laminar flow:

                 [p(r)-pa]/(pw g)  - a (pa//>w) (q/2>rh)  In(r/r0)          (5)

     where  pv and pa are the densities  of water and air respectively.

     For turbulent flow:

      [P(r)-pa]/(pw g) - a a/Pw)  [(q/2irh)b]  [l/(l-b)J (r1^-^1^)    (6)
     The practical implications of the  parameters a and b are that, if they are
really  constant  for a  given bed  material  and can  be determined  by  actual
experiments in the field, they will serve as indices by which a mitigator will
be able  to assess how much  of the area from the suction hole he can hope to
access for a given suction pressure.

     In  order  to evaluate  the  soundness  of  the  mathematical derivation
presented  above  and  also to determine the numerical - values  of  the empirical
coefficients,  a laboratory model  consisting  of a  2.4 m  diameter circular
section and 0.15 m deep was  constructed as shown in Figs. 2 and 3. The top and
bottom impermeable disks were made from 2 cm thick plywood, and a wire mesh at
the outer  periphery  of the  disks  was  used to contain  the  gravel between the
disks. The  apparatus  allowed experiments to be conducted  with  a maximum disk
spacing (depth of gravel bed) of 9.5 cm.

     A 3.8  cm diameter  hole was drilled at the  center of the disk to serve as
the suction hole. Nine holes, whose layout is shown in  Fig. 3, were drilled on
three separate  rays of  the  top disk and fitted with  a  sleeve of 1.3 em inner
diameter FVC pipe with chamfered entrances  at  either end. Pressure measurements
at these nine holes would then yield an  accurate picture of the pressure field
over the entire bed.

     A number  of  different  experimental  runs  were performed on the  laboratory
apparatus  using two different sizes  of river-run gravel  (1.3 and  1.9 cm
diameter) .  Least square regression for both the constant "a" and exponent "b"
was performed on the observed experimental  pressure drop data  using eq.(6)
(since the  flow in the  gravel  bed was determined to be  turbulent) .  Table 1
summarizes  the  different experiments performed using the laboratory apparatus
and the values of the physical parameters obtained.

Field Verification

     The irregular boundary  conditions and the non-homogeneity in  subslab beds
that  arise in  practice  are  however  not easily  tractable with a simple
expression  such as eqs. (5) and (6), and resorting to  a  numerical  computer code
may be  the only  rigorous way to proceed in  order  to predict pressure fields
under actual  situations  (8,9).   We  shall show in  this  section that our
simplified  approach  nevertheless  has practical relevance in  that it could be
used to determine which areas under the slab have  poorer connectivity.

     The house  under  investigation (H21) has  a partial basement with a gravel
bed  under the  basement slab.  As  shown  in  Fig. 4,  the  basement  though
rectangular, is close to being square (6.45 x  7.60 m) . It has  two  sides exposed
to the ambient air above grade, while the other  two  sides are adjacent to slab-
on- grade construction.   Initially, one  suction  hole  of  0.1 m  diameter was

drilled     at roughly  the  center of  the  basement slab  to  which a  temporary
mitigation system was installed.  Though 19 holes  were  drilled  across the  slab
(Fig. 4), two of them (holes 11 and 12) were  found to be blocked beneath the
slab.  Consequently, data from only 17  holes have been used in this  study.  This
blockage was  later  found to be  due  to  the  presence of an oversized  footing for
a support column.

     Three sets  of  runs were carried out which, depending on the airflow  rate
through the single suction pipe, are termed:

          i)     28 L/s  -   High flow,
          ii)    23.4 L/s -  Hedium flow,  and
          iii)   18.1 L/s -  Low flow,

     Note that our analytical expression for the pressure field under turbulent
flow  given  by  eq.  (6)  is  strictly valid for  a  circular disk with boundary
conditions at r - ro and p - pa. We approximate the  rectangular basement  by a
circle of 3.5 m mean radius. We need  to also  include  the extra path length of
ambient air flowing down  the outer  basement wall,  going under the  footing, and
then flowing  through the subslab gravel into the suction hole. We  estimate  this
to be  about  2 m. Consequently, we  find that r0 - 5.5  m.  The thickness of the
subslab gravel bed, h, has been found to be about 5 cm.

     The  gravel under the slab,  though  river-run,  was found  to  be highly
heterogeneous in size and shape. In general, its average size was  slightly  less
than  1.3  cm.  However, we  decided to use  the  properties of  the 1.3  cm gravel
determined experimentally in the laboratory (see Table 1).

     Fig. 5 shows  the observed and calculated pressure drops for the high and
low  flow  rates.  Readings from holes 13 and  14 are lower and we suspect poorer
connectivity  to  these holes; i.e.,  some sort of blockage in this  general area.
Ve note  that the  agreement between model  and observation (Fig.  5)  is indeed
striking,  given  the  simplification in our  model  and  also  the  various
assumptions outlined above.

     Fig. 5 indicates which areas under the slab are  non-uniform. A better way
of illustrating  how well the model fares  against  actual observations is shown
in Fig.  6.  The solid line represents  the  model  predictions  while observations
are shown by discrete points. Ve note again the satisfactory predictive ability
of this  modeling approach and  also the  fact that certain holes have pressure
drop values higher than those predicted by the model.

     In  order to illustrate  the fact that our approach is sensitive  to the
selection of  type of  gravel bed,  Fig.  7 presents the  experimental observations
plotted  against model  predictions  with gravel  bed coefficients  taken to  be
those that correspond to  the 1.9  cm gravel.  Ve note the very large  differences
between model  prediction and observed  pressures over the entire  basement,
thereby suggesting  that our approach has enough  sensitivity  to  be of practical

     An alternate approach, to the one adopted here and described above, would
be not to  assume  specific  gravel bed coefficients but to determine these from
regression.  This entails using eq.(6)  along  with the  data set  of actual
observations and  determining  the  parameters k and b by regression.  Since b is
not a parameter that varies greatly (5),  we have  chosen two different values of
b  (1.6 and 1.7) to  see  what  difference  this  leads to  in terms  of the
coefficient of determination (R2) and in the values  of k.

     Regression results  are  summarized  in Table  2.  We have  performed three
trial  runs.  Trial 1  uses  all  data points while,  in trial  2,  pressure drop
observations from holes  11 and 12 (holes that are blocked) have been removed.
We note that the R2  improves dramatically,  from O.BO to 0.96.   For trial 3,
holes  9 and 10 have been  equally removed  in  order not  to bias the regression
since  these holes have high pressure drop values. We note that the R2 of trial
3 is 0.88, an improvement over that of trial  1.
     Other  than the very  high R  values found,  the  most striking feature is
that regression yields a value of k which is practically identical to  that of
the 1.3 cm gravel determined experimentally  in our laboratory  apparatus. This
suggests that  even a visual  inspection  of the porous material under the slab
can  be an indicator  good enough  for  a mitigator  to  select  a  standard bed
material  before  using the physical  properties  of  the  material get  a  sound
estimate  of  what the suction pressure ought  to be  in order to  generate  a
certain pressure  field under the  slab.  The  need to categorize commonly  found
subslab material,  deduce  their  aerodynamic pressure  drop coefficients in
laboratory  experiments,  and  then tabulate  these in handbooks seems to be an
avenue worth pursuing.

Graphical Representation

     The approach developed here will show how closed-form solutions  for the
pressure drop  in  porous  beds  can be represented  in graphical  form suitable for
professional  radon  mitigators. Let  us  illustrate our  approach  using the
simplest   case of  a  circular  porous  bed  with  radial inflow  between two
impermeable disks. From the discussion in the above section,  it would seem that
we  could  apply our  model  equally to square  basements  and  to houses with  a
partial basement.

     Eq.  (5)  is valid for  laminar  flow  which would prevail where soil is the
subslab material  (5).  It can be written as:

          Aph(r)  k - (ua/g) (/>a/pw) (1/2*)  (q/h)  In(r/r0)                (7)

where  Ap^  is  the pressure drop  in head of  water and is equal to
lp(r)-pa]/0»w 8>-

     Four curves have been plotted in Fig.  8  for four values  of
(q/h):   l.OxlO'3, 5xlO'3,  IxlO'2,  and  2xlO'2 m2/s.   Thus,  if the values of
(r/r0), (q/h),  and k are known, [Aph(r)l    can be obtained from this  figure.

     For gravel  under the slab, the  flows  will probably be turbulent and  the
pressure drop is given by eq. (6) which can be rewritten as:
     Aph(r) F- 
Pressure Drop Considerations

     There are  basically  three  different sources  of pressure  drops in  the
mitigation system:

          APtotal - APbed + APent + APpipe

     where ApDe(j - pressure drop in porous bed,
                 ** pressure drop due  to  change of direction and that
                   associated with entrance effects into the mitigation
                   pipe, and
                  ~ pressure drop in the mitigation pipe.
     The pressure drop in the subslab bed is given by equations akin to eq. (5)
or  (6) .  The  pressure  drop  at the  entrance  to  the  suction pipe  involves
accounting for the following effects: (i) change in flow direction, (ii) change
in  cross-sectional  area,  and  (iii)  entrance effects  at the  throat of the
suction  pipe.  From an  engineering  viewpoint, it  is more convenient  to  treat
these  together.   In  accordance  with  actual  practice  (10),  we propose the
following simplified empirical equation for the head loss:

           APent/Pw g - Kp [l-(Ap/Ab)]2 (Vp2/2g)                      (10)

     where Kp  is  the  dimensionless  pressure  loss  coefficient  which should not
              depend on the velocity or  the  bed  thickness,  and is a constant
              for a specific type of bed material,
           Ap is the cross -sectional area of the suction pipe,
           AD is  the  surface area of a cylinder  of diameter equal to that of
              the suction pipe and  height  equal  to the thickness of the porous
              bed, and
           Vp is the velocity of air in the suction pipe.

     If dp is the diameter of the suction pipe, then:

                - (irdp2/4) (l/ffdph)  - dp/4h                             (lla)
            Vp  - q/Ap - 4q/wdp2                                        (lib)

     Table  3  assembles  the  results of  determining the  entry pressure  loss
coefficient for three flow rates. We note  that  Kp values are exactly the same,
a gratifying result. This enables us to place a certain amount of confidence in
our model for the entrance losses.

     The pressure drop  in the piping includes losses due  to elbows,  fittings,
as well  as  straight pipe. Following Ref.  10,  losses in the  straight pipe  are
given by:
               - f a/dp) (q/A)/2g                                     (12)

     Pressure losses in  bends  and fittings are normally expressed  in  terms  of

an equivalent  pipe diameter. For example,  a 90° elbow has  the  same pressure
drop as  a  straight pipe of  length  equal  to about 25  times  the  pipe diameter

     Since  the  primary  objective  of  the  mitigation system  is to  create a
suction  pressure  under  the  slab  only,  we  can  define  a  hydrodynamic
effectiveness of the mitigation system based on  these three pressure drops:

     Hydrodynamic effectiveness - Apbed / APtotal                       (13)
     We have computed these various pressure  drop values  for house H21 in order
to get  an idea of their relative  magnitude.   The  mitigation system (with one
suction hole  only)  in house  H21  has  about 7 m  of  straight  pipe of  0.1  m
diameter and three 90° elbows.  This translates  into  a  total length of
7 + (3 x 2.5 x 0.1) - 14.5 m.

     Table 4 assembles the various pressure  drops in  the three  elements of the
mitigation system.   While  AP^ed and APent have been measured, APpipe has been
calculated from eq.  (12).   The hydrodynamic effectiveness defined by eq.  (13)
is also given.

     Table 4 shows  that APpipe is negligible compared  to APDed>  while AP  is
about  one  third  of  AP^,e

     The assistance of R. Gafgen during the experimental phase of this study is


A     cross-sectional area of flow
a     parameter representative of the resistivity to flow of the porous bed
b     pressure drop exponent for turbulent flow in gravel beds
d     diameter
dy    equivalent diameter of pebbles
F     correction factor given by eq. (8b)
f     friction factor
g     acceleration due to gravity
h     thickness of porous bed
Kp    pressure loss coefficient at entry to suction pipe
k     permeability of porous bed
L     length of pipe
Ap    pressure drop
p     pressure
pa    atmospheric pressure
q     total volume flow rate
R2    coefficient of determination of regression
Re    Reynolds number
r     radial distance from center of the suction hole
r0    outer radius of the laboratory apparatus
SEM   standard error of the mean of the regression estimate
V     air velocity
x     distance along flow
p     density
v     dynamic viscosity
4     porosity of porous bed


a      air, ambient
b      porous bed
ent    entrance
f      fluid
P      Pipe
w      water


1. D.T. Harrje  & L.M.  Hubbard,  Proceedings  of the Radon Diagnostics Workshop,
April 13-14, 1987 (EPA-600/9-89-057)  (NTIS PB89-207898), June 1989.

2. J. Wang & M.  Cahill,  Radon reduction efforts in New Jersey, paper presented
at the Annual Meeting  of the National Health Physics Society, Boston, MA, July
4-8, 1988.

3. D.C.  Sanchez, Technical issues related  to emission releases  from subslab
radon mitigation  systems,  presented  at ASCE  National Conference  on
Environmental Engineering,  Austin,  TX, July  9-12, 1989.

4. K.J. Gadsby,  L.M.  Hubbard, D.T. Harrje  &  D.C.  Sanchez,  Rapid Diagnostics:
Subslab  and  Wall  Depressurization Systems  for  Control of  Indoor  Radon,
Proceedings:    The 1988  Symposium on  Radon  and Radon Reduction Technology,
Volume 2,  EPA-600/9-89-006b (NTIS  PB89-167498), March 1989.

5. T.A.Reddy, H.E.Black III, K.J.Gadsby, D.T.Harrje  & R.G.Sextro, Modeling air
flow dynamics through a homogeneous porous bed with relevance to proper design
of  radon  mitigation  systems  using  subslab  depressurization,  PU/CEES  draft
report, Center  for Energy  and  Environmental  Studies,  Princetion University,
Jan. 1990;  also,  "Airflow dynamics under subslab depressurization:  Simplified
model approach and preliminary validation" paper presented at the Third Annual
AARST Conference, Baltimore, Oct.  16-17,  1989.

6. T.G. Matthews,  D.L. Wilson, P.K.  TerKonda,  R.J. Saultz,  G.  Goolsby, S.E.
Burns & J.W.  Haas,  Radon diagnostics:  Subslab communication and  permeability
measurements,  Proceedings:   The 1988 Symposium  on  Radon and Radon Reduction
Technology, Volume 1,  EPA-600/9-89-006a  (NTIS  PB89-167480), March 1989.

7. M. Muskat,  The Flow of Homogeneous  Fluids through Porous Media,
McGraw-Hill, 1937.

8. C.  de  0.  Loureiro,  "Simulation of the Steady-State Transport of Radon from
Soil  into  Houses with Basement under  Constant  Negative Pressure",  LBL-24378,
Lawrence Berkeley Laboratory,  Berkeley CA 1987.

9.  J.M.Barbar  &  D.E.Hintenlang, Computer  modeling of  subslab ventilation
systems  in  Florida,  paper presented  at the  34th  Annual Meeting  of  Health
Physics,  Abstract No.  TAM-E8,  Albuquerque, NM, 1989.

10.   ASHRAE,  Handbook of Fundamentals.  American Society  of Heating,
Refrigeration and Air-Conditioning Engineers,  Atlanta,  1985.

          Table 1.  Summary of laboratory experiments using river run gravel
	and the physical parameters deduced in Ref.  5	

Experiment      Diameter of      Measured  Pressure   Permeability
                particles        porosity  drop         of bed
             nominal measured             exponent    (m2)
	(m)     fin)	
0.013   0.011
0.374    1.60
           9.4 x 10
0.019   0.022
0.424    1.40
           34 x 10
   Table 2. Results of regressing experimental data using eq,(6)
            1.6    9.13 x 10'9

            1.7    7.5 x 10'9
         O.BO    With all data points


     7.1  x 10'9

     5.8  x 10'9

With data of holes 11 and 12

     10.0 x 10'9

      7.3 x 10"9
        With data of holes 9, 10, 11
        and 12 removed

  Table 3.   Determination of the pressure loss coefficient at the
            throat of the mitigation suction pipe in house H21




(cm water)
{en water)


Table 4.  Relative pressure drops in the mitieatlon system of house H21


(cm water)

(cm water)

(cm wat
8.0 x
13.0 x
17.7 x

)e Hydrodynamic
:er) effectiveness

x"  pipe

                                            . Tap to measure static pressure
 Fig. 1  Schematic of  a model  to  duplicate  flow conditions
         occurring beneath the concrete slab of a residence
         when induced by a single suction  point.  The air-
         flow is  assumed to be radial flow through a
         homogenous porous bed of circular boundary.
                                        Attachement Co measure
                                        flow and pressure

                                           Suction tube

                                           Top plywood disk

                                           Foam cover
                                           Gravel bed
                                            Bottom plywood disk
  Fig. 2   Cross-section of the  experimental  laboratory apparatus.
Fig.  3   Layout of  the test  holes to measure static
         pressures  in the  porous bed.

                                                  • 1
          • 5
                        7        .6
10 V9

                      Slab on grade
    Fig. 4  Plan of the basement slab showing the relative
            positions of the various subslab penetrations.
            The suction hole  of the temporary mitigation
            system is marked as +, while the location of
            the central and perimeter suction holes (M) of
            the final mitigation system are also shown.

2.5 -

2.0 -

1.5 -

1.0 -

0.5 -
                n Observed
                + Calculated
          0  -F-T
            I   1  3  4  8  a   7  •  •  10  11 14  IS  II  17 II II
                                                                                           DISTANCE (m)
                                     10 11 14  19  II 17 ia  It
   Fig.  5   Using coefficients for 1.3 cm grave] (b =  1.6        Fig. 6   Comparison  of observed and computed pressure
            and k = 9.4 x  10~9m2).  Data of holes 11 nnd  12             drops us!nj.', coelT !.c. ients of 1.3 cm Rr.nvcl.
            not included.                                                  Data of holes 11 and 12  not included.

                                     10  II  14  IS U  IT  11  It
Fig. 7  Using coefficients of 1.9 cm gravel  (b =  1.4
        and  k = 3.4 x 10~9 m2).


  0.1  0.2 0.3  0.4  0.5 0.6  0.7  0.8 0.9
8  Pressure drop  in  a  sand bed  with radial
   airflow between two impermeable disks.
                                     GRAVEL BED
                                     River-run gravel
                                     Small 1.3 cm

                                     Large 1.9 Cm
Fig. 9
  0.1  0.2 0.3  0.4  0.5  0.6  0.7  0.8  0.9
 Pressure drop  in a gravel bed with radial
 airflow between two  impermeable disks.  The
 correction factor F  can be determined from
 Fig. 10.
                                                                          12343678   9 10
                                                                           1234567  89 10
    0.13 •
    0.09 ^
    0.07 •
           (4) -7.5=1
                                                                            Large gravel
                                                                                   ' Snail gravel
    1234   5678   9  10
                 ro (m)
10  Correction  factor F for gravel  beds to  he
    used in  Fig. 9.

    35 -
    30 -
25 -
    15  -
                                                      Both suction
                                                      pipes open
                                              pipe only
                                            pipe only
        1   2   S  4   5   S  7   8   9   10   13  14  15   16  17  18  1»

                              HOIX NUUBER
         Fig.  11  Subslab suction pressure fields generated by the
                  mitigation system when different suction pipes
                  are used.


                      by:  Allan B. Tanner
                           U.S. Geological Survey
                           Reston, VA, 22092

     Pressure-driven flow of radon-bearing soil gas is commonly accepted as
Che usual mechanism whereby radon moves from outside house foundations to
cause elevated indoor radon concentrations.  It is less clear how radon moves
to the backfill-and-subslab zone just outside the foundation.  Fourteen houses
having elevated indoor radon concentrations were investigated by the U.S.
Environmental Protection Agency and its contractors.  The permeability of the
ground to gas flow was measured next to and several meters from each house
foundation.  For 6 of the 14 houses none of the intrinsic permeability values
exceeded 7.6xlO~12 m2, below which diffusion is likely to be the dominant mech-
anism of radon movement.  Because it can be significant in unsaturated soils
of moderate-to-low permeability, diffusion should not be ignored in consider-
ing radon movement to house foundations.


     Radon (222Rn) in the ground moves by two principal mechanisms.  In res-
ponse to a gradient of radon concentration, there is a net movement of radon
atoms in the direction of lessening concentration by the process of diffusion.
In a porous medium such as soil, the effective diffusion coefficient, which
characterizes the rate of diffusion, depends primarily on the degree of liquid
saturation of the soil and secondarily on the porosity, pore sizes, adsorptive
properties of the soil grains, and absorptive properties of the liquid phase.
Diffusion can occur with or without soil-gas flow, which is caused by a
pressure gradient in the soil.  The rate of soil-gas flow is controlled mainly
by the pressure gradient and the soil's permeability to gases.  Permeability
depends strongly on pore and grain size; like the diffusion coefficient, it
decreases greatly as the fractional saturation of the soil by liquids

     It is generally accepted that where enough radon enters a house to cause
concern, entry Is usually due to pressure-driven flow of radon-bearing soil
gas (1,2).  Diffusion of radon through concrete slabs and walls is slow enough
to reduce the radon concentration greatly by decay (3).  Radon movement
through local cracks, sumps, and utility openings by soil-gas flow is favored
over diffusion if the air pressure within a house at the slab level is less
than that in the soil, and if the zone comprising the backfill and subslab
material is fairly permeable.  Both conditions often exist, particularly if
the house slab is laid on the layer of coarse aggregate required by many
building codes.  It is less clear to what degree each of the two mechanisms is
responsible for radon movement from the soil to the foundation wall or to the
underside of the slab.  Presumably because of evidence that pressure-driven
flow causes the actual radon entry, most modeling to date has assumed that
diffusion can be neglected.  I know of no study that justifies such an assump-
tion for other than very permeable soils.

     The purpose of this paper is to present the several factors most impor-
tant to diffusion and flow, to relate them to various soil types and condi-
tions, and to show that diffusion, rather than flow, is likely to be the
dominant mechanism of radon movement to the foundations of some houses having
moderately elevated indoor radon levels.

                          COMBINED DIFFUSION AND  FLOW

     The first equation for steady-state, one-dimensional radon movement
incorporating both diffusion and flow velocity was derived by Grammakov (4).
Clements (5) derived an equation for the steady-state radon flux density
crossing the Earth's surface into the atmosphere.  Multiplying the radon flux
density by the mean life of radon (the reciprocal of the decay constant)
yields the maximum amount-of radon per unit area that can be sustained by
steady migration from the source soil (6).  Dividing that amount by the
concentration of radon that would build up in the soil gas if there were no
migration yields a distance, K, which is the volume of undepleted soil gas per
unit surface area that would contain the amount of radon sustained externally.
I call M the "mean migration distance."  It is calculated as follows:
             M -  [l/(2L)][(-k/u)(dp/dx)+/(-k/u)2(dp/dx)2+4eLD],           (1)

where H is in meters, L is the radon decay constant, e is the soil porosity,
k (mz)  is the soil's intrinsic permeability to gas,  u is the viscosity of air
(1.8xlO"s Pa-s at  typical  soil temperatures),  dp/dx  (Pa/m)  is  the  pressure
gradient, and D (nr/s) is the (bulk) effective diffusion coefficient (see
references 3 and 7), equal to the interstitial effective diffusion coefficient
times the soil porosity.

     The bulk effective diffusion coefficient can be estimated by the
following equation  (8):

                        D/e - 7xl
     Typical permeabilities of soils range from 10~16 m2 for clays to
10~7 m2 for clean gravels (7).  Although the mean migration distance in clays
is only of the order of 1 cm, diffusion is the dominant mechanism, and flow
can be neglected.  For even very low pressure gradients, flow is dominant in
gravels and coarse sands.  Because of their mixed grain sizes, most soils have
permeability values between the extremes,  and the relative contributions by
flow and by diffusion depend not only on permeability and diffusion
coefficient, but also on the pressure gradient.

     Figures 1 through 4 are graphs computed from equation (1) in order to
show the relative importance of diffusion and soil-gas flow for different
values of permeability, porosity, diffusion coefficient, and pressure
gradient.  Because of interaction among the variables, the graphs should not
be used to infer the effect of changing a single variable on the flux density
of radon.
         SILT AND CLAY

         CLEAN SANDS
              -13            -12            -11
                 DDG10 PERMEABILITY (m2)
Figure 1.
Mean radon migration distance for soils of 40 percent porosity,
50 percent water saturation, and -1 Fa/m pressure gradient.  The
correlations of soil types with permeability values or ranges are
from reference (7).

     Figure 1 has been computed for a soil  porosity of 0.4, a pressure
gradient of -1 Pa/m, a bulk effective diffusion  coefficient of 4.8x10"7 mz/s,
corresponding to 50 percent water saturation,  and  the range of permeability
extending from that of a sandy clay to that of a uniform medium sand.  The
permeability ranges of the soil types are taken  from reference (7).  Figure 1
shows that the mean interstitial migration  distance is nearly independent of
permeability below several times 10'12 m2  under the specified conditions.
Soil-gas flow should contribute little to radon  migration  in poorly graded
soils containing substantial fractions of clay,  silt, or fine sand.  With
well-graded grains and larger grain sizes,  soil-gas flow is more important
than diffusion.
 W   3
-13            -12            -11
Figure 2.  Mean radon migration distance  for soil  of  SO percent water satura-
           tion, -1 Pa/m pressure gradient,  and porosities of 20,  30, 40, 50,
           and 60 percent (uppermost to lowest curves, respectively).

     Figure 2 presents a family of curves computed for a pressure  gradient of
-1 Fa/n, a diffusion coefficient corresponding to  50  percent water saturation,
and different values of porosity in the range from 20 to 60 percent.  Soil-gas

flow is significant at slightly lower values of permeability in low-porosity
soils than in high-porosity soils.
•13           -12            -11
Figure 3.  Mean radon migration distances for soils of 40 percent  porosity,  50
           percent water saturation, and pressure gradients  of -20,  -5,  -3,
           -1, -0.5, and -0.1 Pa/m (uppermost to lowest curves,  respectively).

     Figure 3 presents a family of curves computed for soils of 40 percent
porosity, a diffusion coefficient corresponding to 50 percent water  satura-
tion, and pressure gradients ranging from -20 to -0.1 Pa/m.   Pressure  differ-
ences of -5 Pa between U.S. houses and the soil are considered to  be fairly
high (9).  If a -5 Pa difference were distributed over the typical radon
diffusion length of about 1 m in soil, the resulting -5 Pa/m gradient  should
cause soil-gas transport of radon to be more important than diffusion  in soils
of permeability greater than about 10"12 m2.   Such gradients  are likely where
the radon flux converges at entry points such as cracks and utility  openings,
and the gradients decrease markedly with distance from the entry points.  At a

house site where radon is effectively gathered from  a  zone several meters from
the foundation, the gradients should be much  less  than 1 Pa/m.  Where radon
entry occurs over a broad front,  such as through porous block walls or a
distributed crack system, the one-dimensional migration regime discussed in
this paper should be suitable.
   -~»  4
-13            -12            -11
Figure 4.  Mean radon migration distance  for  -1 Pa/m pressure gradient in
           soils of 40 percent porosity and 10, 30, SO, 70, and 90 percent
           water saturation (uppermost  to lowest curves, respectively).

     Figure 4 presents a family of curves computed for a pressure gradient of
-1 Pa/m in soils of 40 percent porosity and 10, 30, 50, 70, and 90 percent
water saturation.  In the drier soils,  soil-gas flow should be significant
with permeability exceeding 10"ll m2;  in the wetter soils, soil-gas flow should
be significant at somewhat lower values.   Increasing saturation of a given
soil reduces both its radon diffusion coefficient and its gas permeability.

I am not aware of studies of the effects of increasing saturation on both
permeability and diffusion coefficient on the same material.  However, Rogers
and others (10) presented figures showing approximately 30-fold reduction of
permeability from dry to saturated state and a three order-of-magnitude reduc-
tion of diffusion coefficient over the same range.  Nazaroff and others (7)
gave data indicating a decrease of gas permeability by a factor of 50 to 100
between the dry and saturated states of loamy sand.  Because the diffusion
coefficient enters into equation (1) to the 1/2 power, the effect of satur-
ating soil on the mean migration distance in it by diffusion should be
comparable with that for soil-gas flow.

     Loureiro (11) included both diffusive and soil-gas flow mechanisms in
modeling radon entry from dry soil via a slab-to-footing gap.  He concluded
that the diffusive mechanism was dominant for soils of permeability less than
10*12 m2, and that the soil-gas flow mechanism was dominant above that value.

                                  FIELD DATA

     Two sets of field data were available that included soil permeability
measurements for houses having indoor radon measurements exceeding the lowest
action level.  The first set was obtained as part of the U.S. Environmental
Protection Agency's House Evaluation Program in work performed by Agency
personnel and contractors.  Radon in soil gas and the permeability of the
ground close to and several meters from house foundations were measured by
soil-probe methods at locations in Colorado Springs and Denver, Colorado, in
northern Virginia, in Bartow and Lakeland, Florida, and in northern New
Jersey.  Of 14 different houses having indoor radon measurements exceeding  the
148 Bq/m3 (4 pCi/L) level,  six had no measurement exceeding 7.6xlO"12 m2.   The
maximum values measured were 2.2xlO'12, 7.6xlO'12,  7.4x10'",  1.2xlO'12,
3.3xlO"12, and 6.4xlO"12 m2.  Permeabilities at the sites typically ranged to
values one to two orders of magnitude lower than the maximum.  The greatest
permeability was usually found in the backfill zone near the foundation wall.

     I investigated three homesites that were included in the House Evaluation
Program at the times of those tests.  The three houses had indoor radon
concentrations of about 800 Bq/m3 (20 pCi/L) but are not included among the
six houses discussed above because the House Evaluation Program permeability
measurements exceeded 1x10"u m2 for those houses.  By means of the procedure
described in reference (6), I measured permeabilities ranging from l.SxlO"12
to 7xlO'14 and from 2.2xlO'12 to  1.2x10'" m2 at two  of  the houses;  two  meas-
urements at different sites at the third house gave a value of 1.4x10'" m2.
At two houses in northern New Jersey, both of which had indoor radon  levels
exceeding 7 kBq/m3 (200 pCi/L),  I measured permeabilities of 7x10'" and
2xlO"10 m2, which were consistent with other observations of high permeabil-
ities associated with severe indoor radon levels  (12).


     There is little doubt that severe indoor radon levels are very likely to
be associated with highly permeable soils.  However, moderately elevated
indoor radon levels are sometimes associated with soils of low average perme-
ability.  At such sites, diffusion is probably the dominant mechanism of radon
movement in the soil, particularly beyond the disturbed zone comprising the
backfill and subslab aggregate.


     This study was supported in part by the Office of Health and Environ-
mental Research of the U.S. Department of Energy under interagency agreement
no. DE-A105-87-ER60578.  I thank R. Thomas Peake of the Office of Radiation
Programs, U.S. Environmental Protection Agency,  for making some of the House
Evaluation Program data available.

     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.

1.  Akerblom, G.,  Andersson, P., and Clavensjfl, B.   Soil gas radon--a source
    for indoor radon daughters.  Radiation Protection Dosimetry.  7: 49, 1984.

2.  Nero, A.V. and Nazaroff, W.W.  Characterising the source of radon indoors.
    Radiation Protection Dosimetry.  7: 23, 1984.

3.  Culot, M.V.J., Olson, H.G., and Schiager, K.J.   Effective diffusion
    coefficient of radon in concrete, theory and method for field
    measurements.   Health Phys.  30: 263, 1976.

4.  Grammakov, A.G.  On the influence of some factors in the spreading of
    radioactive emanations under natural conditions [in Russian].  Zhur.
    Geofiziki.  6: 123, 1936.

5.  Clements, W.E.  The effect of atmospheric pressure variation on the
    transport of Z22Rn from the soil to the atmosphere [Ph.D. dissertation].
    New Mexico Inst. Mining and Technology, Socorro, N.M.  110 p.

6.  Tanner, A.B.  A tentative protocol for measurement of radon availability
    from the ground.  Radiation Protection Dosimetry.  24: 79, 1988.

7.  Nazaroff, W.W., Hoed, B.A., and Sextro, R.G.  Soil as a source of indoor
    radon: generation, migration, and entry.  In:  W.W.  Nazaroff and A.V.
    Nero, Jr. (eds.), Radon and Its Decay Products in Indoor Air.  John Wiley,
    New York, 1988.  p. 57.

8.  Rogers, V.C.,  Nielsen, K.K., and Kalkwarf, D.R.  Radon attenuation
    handbook for uranium mill tailings cover design.  NUREG/CR-3533, U.S.
    Nuclear Regulatory Commission, Washington, D.C., 1984.  85 p.

9.  Sextro, R.G.  Oral communication, 1986.

10. Rogers, V.C.,  Nielson, K.K., and Merrell, G.B.  Radon generation,
    adsorption, absorption, and transport in porous media.  RAE-8810-1 and
    DOE/ER/60664-1, Rogers and Associates Engineering Corp., Salt Lake City,
    Utah.  48 p.

11. Loureiro, C. de 0.  Simulation of the steady-state transport of radon from
    the soil into houses with basements under constant negative pressure
    [Ph.D. dissertation).  LBL-24378, Lawrence Berkeley Laboratory, Berkeley,
    Calif.  278 p.

12. 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.  In: Proceedings of the 1988 Symposium on Radon and
    Radon Reduction Technology, Vol. 1, Symposium Oral Papers.
    EPA/600/9-89/006a.   [U.S. Environmental Protection Agency] Radian Corp.,
    Research Triangle Park, N.C., 1989.  p. 5-61.


                                Bradley H.  Turk
                           Rm.  109,  105 E.  Marcy St.
                         Santa  Fe, New Mexico    87501

                                 Jed Harrison
                     U.S.  EPA Office of Radiation Programs
                           Washington, D.C.   20460

                               Richard J. Prill
                      Washington Energy Extension Service
                          Spokane, Washington   99201

                               Richard G.  Sextro
                          Indoor Environment Program
                         Lawrence Berkeley Laboratory
                         Berkeley, California   94720


     Measurement  techniques  and  parameters that  describe the  potential for
areas of a  building substructure to have high  soil  gas  and radon entry rates
have  been   developed.   Flows   and   pressures  measured   at   test  holes  in
substructure  surfaces  while the substructure  was intentionally depressurized
were  used   in   a   highly  simplified   electrical   circuit   to   model  the
substructure/soil network.  Data from four New Jersey houses indicate that (1)
the soil was a factor of two to six times more resistant to soil gas flow than
substructure  surfaces,  (2)  concrete slab  floors,  including  perimeter gaps,
cracks,  and  other penetrations, were approximately  five  times more resistant
to soil gas  movement than hollow block walls,  and  (3)  radon entry potentials
were highest  for  slab  floors.   These Indices of entry potential may be useful
for characterizing the relative leakiness of below-grade substructure surfaces
and for determining the selection and placement of radon control systems.


     It  is  widely  accepted  that the  pressure-driven flow  of soil  air into
buildings is  the  most  frequent cause of elevated indoor radon levels. Several
studies  have  proposed  techniques   that  define  the  ability  of  soil  at  a
particular site  to supply radon  to a building structure  via  this convective
flow  (1,2,3,4,5).   In  some of these studies,  the  magnitude  of  the pressure
field extending  from a mechanically-depressurized house has  also been mapped
(6,7,8).  The pressure coupling between the house and the soil location can be
defined  as   the   ratio  of  the pressure  difference  between  the  measurement
location and the  substructure to  the depressurization  of the  substructure
relative to  outside.   A snail  value  for  the pressure coupling indicates the
existence of  pathways connecting the  house and  soil  location.  The pathways

 consist  of  openings  through  the  substructure  surfaces   (cracks,   holes,
 perimeter drains, etc.)  and regions in  the materials  around the house  that
 permit relatively unrestricted transport  of soil air (permeable  soil,  gravel
 layers, and  air  gaps).  Other work,  some  related  to  the  control of  indoor
 radon,  has attempted to  identify  soil  air pathways near substructures  and  to
 locate radon entry points in the  building envelope  (9,10,11,12).
     Various measurements of the air  permeability of  the soil  performed  during
 many of  the above  studies are  indicators of  the  mobility  of  soil  gas  in
 materials around  a  house.  The  air  permeability  can vary widely over  both
 vertical  and horizontal dimensions.  High permeability zones  were  often found
 near substructures and  were caused  by  less tightly-packed soil  in the  area
 disturbed by construction, layers of material provided  for drainage  (gravel),
 and  air  gaps below slab floors  or  near walls  -  some extended  to  the  soil
 surface - created by expansion/contraction cycles and settling (4,13).
     In this  paper,  we  describe  entry  potentials  for soil gas  and  radon
 through substructure surfaces  that  are  based  on  the  developments  from the
 research  mentioned above and on  flows,   pressures,  and radon concentrations
 measured  at  the   substructure/soil   interface.  This  approach simplistically
 considers the  below  grade  substructure  surfaces,  near-house materials, and
 nearby  soils  as  elements  of  a  "black  box" whose  aggregate  characteristics
 assist  in estimating the likelihood and  relative  magnitude  of  soil  gas and
 radon  entry.  We  do  not  include  transport  by  diffusion.  In  addition, the
 analysis  enables  comparisons of the resistance of soil and of the  below-grade
 substructure surfaces to  air flow.


     Below  grade, most  houses are  surrounded  by  a  very  complex  matrix  of
 soils,  rock,  and construction-related materials  or  structures,  each  having
 different capacities for  radon production and soil gas and radon transport.  In
 most situations,  details  of the geometry,  characteristics,  and  interactions
 between these features  and with the  substructure cannot be known.  Plus, the
 condition and construction  details of  the substructure surfaces  in contact
 with these exterior materials are not fully known.
     In   order  to  account  for  these  complexities  without knowing   their
 specifics,  a method  to  integrate  the  flow  and pressure  characteristics  is
 useful.   If  we  assume a  linear  pressure  dependence  for air flow through  soils
 and  materials around the  substructure (i.e.,  Darcy  flow), the passage of  soil
 gas  through  soil  and into a  substructure depends on the resistance of the  flow
 path through the  materials around the substructure and through the surfaces  of
 the  substructure.  Therefore, a simplified electrical resistance  analog of a
 typical basement/soil system can be created, as shown  in Figures la and  Ib,  to
 simulate  the flows,  pressure drops,  and  resistances  in  the soils, near-house
 materials, and substructure  surfaces.
     A  blower door  that  enhances depressurization   in  the  substructure  is
 represented by the battery in the circuit.  The test holes drilled through the
 floor and walls permit measurements of  air flows (current)  and pressure drops
 (voltage) at various nodes in the circuit.  While the  blower is operating, the
 system  is tested  in two  conditions:  with  the  test hole closed and  with the
 test hole open.  With the test hole sealed, good pressure coupling between the
 interior of the substructure and the exterior at a test hole location (pathway
with low  effective resistance)  may result  from nearby cracks  and openings  in
 the  substructure surfaces   or  from  more  distant  openings  through  the
 substructure  that are  connected  by  high permeability  pathways  to the  test

hole.  Good coupling does not, however, necessarily mean that large quantities
of soil gas  will  enter through the nearby  openings.   The  soil, aggregate, or
backfill material around  the  substructure must also be sufficiently permeable
so that substantial quantities of soil gas can be transported to the openings.
Therefore, with the test hole open (indicated in Figure Ib by the dotted lines
representing  the  current, IH, and resistance,  R^,  of the test hole and  flow
measurement adaptor), a high air flow rate through the test hole suggests  high
permeability   (low   effective  resistance)   in  the  materials   around   the
substructure and relatively rapid transport of soil gas to the  nearby openings
in the  substructure surfaces.  In general,  when both good  pressure coupling
and  high  flow  rates are  measured at  a test  hole,  then  it  is  likely  that
significant amounts of soil gas will enter the house around that location.
     A more  quantitative  interpretation of the data  requires analysis of  the
analogous electrical circuit.  Following are  symbols and definitions (and the
corresponding  electrical  parameters)  used  in  the derivations.  The subscript
"C" identifies the condition with the test hole closed.

QH (IH)  =  measured (corrected) flow through open test hole  and flow adaptor

QF (Jp)  "  defined  flow through cracks  and openings in below-grade substructure
          surfaces with test  hole open  (m3/s),

QT (IT)  =  defined  total flow through cracks, openings, and test hole (m3/s),

PB (VB)  -  measured pressure difference between inside of basement and
          outside, point  a to c (Fa),

PH (VH)  =  calculated pressure drop across open test hole and flow adaptor,
          point a to b  (Pa),

Ps (Vs)  -  pressure drop across soil paths between point b and outside with
          test hole  open,  PB  - PH (Pa),

RH      =  defined resistance of open test hole and flow adaptor (Pa-s/m3),

RF-EFF   =  calculated effective resistance that lumps resistances of cracks
          and openings  in substructure  surfaces  and resistances of near-
          substructure  materials  surrounding  the  open test hole (RF1, RF2»
          RF3, etc.) (Pa-s/m3),  and

Rg.gpp   - calculated effective resistance of soil paths to measurement
          point b (Rg!, R^, Rgg,  etc.), with test hole open (Pa-s/m3).
     We  assume that the  complex network of  resistances through the  soil  and
 substructure  surfaces is represented  approximately by  the  simplified circuit
 shown  to  the left  in Figure  Ib.   With the  test hole  closed,  a  resistance
 ratio,   Z   (i.e.,   the   resistance   of   the   substructure   surfaces   and
 near-substructure materials  divided by the  resistance of  the soil),  can be
 defined  as

        ~"  17      ID          D
           V sc    'TCKSC-SFF     KSC-BFF

 Using Kirchoff's rule  for circuit analysis  when the  test  hole is  open,  the
 following three independent equations are derived

      ^8 -  IHRH ~  h*S-EFt -  0 ,                                      [3]

      IF&F-F.FF ~  IH*H = 0 ,    and                                     [4]

      IT  ~ If ~ IH  =  0  '                                               [5]

 If we assume that

      RFC-CFF "  RF-SFF     ano<

      RSC-EFF "  &S-SFF

 then  we  can solve for Rg-gpy,  using equations 2 through  5  and substituting for
 analogous  air  flow and  pressure  parameters,  to  obtain the effective  soil
RF-EFP is found by  solving equation 2.  These two resistance  parameters enable
comparisons of the  resistance  to  air  flow created by the soil and substructure
surfaces .
     Combining these  resistances, we  now define the  entry potential  of  soil
gas at a location,  G  (a3/Pa-s), as the net conductance through the surrounding
soil, near- substructure materials, and substructure  surface materials from
S-SFF * Kf-tFF
      fr  - » - ;r» -  .       or                                    [7]
           K-    *
                            Qr  •                                       [8]
Thus,  larger values  of either  Rg.gpp  or Rp.Epp result  in a  smaller soil  gas
entry potential.
     Note  that  the  area  over which to  apply the  effective resistances  and
entry  potential is  not defined.  In an ideal situation,  where the  subfloor
materials  are highly permeable  (RF2  is  small) and  the substructure  surfaces
and  surrounding  soils are  homogeneous -  without  discontinuities  such  as
impermeable  barriers or  large  short  circuits, a  single  test  hole  location
would suffice to  calculate the  total  resistance  of the substructure  surfaces
and  of  the  soils.   However,  many of  these discontinuities  may exist  around
typical houses.  Consequently, more  than one measurement location is  required
to determine  the local  resistances at different locations.  Unfortunately,  in
this situation with  many  test  locations, it is difficult to know the  distance
from each  test  location over which  the  resistances are derived.  Indeed,  for
those test locations  in homogeneous  materials such as subfloor gravel layers,
identical  conditions may be measured over  a large area.  While  at the  same
house,  measurements  made at  test  locations  in  different  materials   could
represent conditions very near to the test location.

     The entry  potential  of radon,  E  (Bq/Pa-s),  may  be  defined as  the mass
transfer of  radon found near the  substructure surfaces,  C  (Bq/m3),  with the
prevailing pressure -normalized  flow of soil gas  into the building  (soil gas
entry potential):

      £ = CC •                                                       [9]

Similar  to  the  soil  gas entry  potential,  the radon entry  potential  should
indicate the likelihood that significant amounts of radon can enter a building
through an area  of the substructure surface.  Both high  soil  gas entry rates
(soil gas entry potential)  and high radon  concentrations at  the exterior of
the substructure will cause a higher radon entry rate (radon entry potential).

                            EXPERIMENTAL PROCEDURES

     At four  New Jersey houses , a blower door depressurized the substructure
by  -10   Pa  to  -37 Pa  while air  velocities  and  pressure  differences  were
measured  at  indoor  test  holes.   Measurements  were made  at   up   to  three
different  pressures  in  two   houses,   LBL13   and  LBL14C.   By  artificially
depressurizing the building, the magnitude of most parameters was increased so
that  they  were  more  easily measured and so that many environmental effects
were minimized.  All tests were conducted during June 1987.  These houses were
part of a larger research project investigating radon entry and control (14).
     In each  house,  approximately 30 test holes  (6 mm to 13  mm in diameter)
had been drilled through  substructure  slab  floors and hollow block walls, and
into the block  cavities of  these  walls (approximately 0.25 m above the floor)
for a  variety of  measurement  purposes.  Some  of  the test  holes through the
floors  penetrated  only  to the  space or gravel  layer  directly below the slab,
while at  other  holes,  probes  extended approximately  1 m into  soil  that was
compacted before the  slab was poured.   At one  location on each exterior wall,
a probe completely penetrated the block wall to the soil.
     In  separate experiments conducted while  the house  was depressurized by
the blower  door, a pressure field map was  made  of  the  pressure coupling at
some  indoor  test holes and  at  approximately 25 locations in  the soil around
the house.   The  soil  probes were  13 mm OD  and were  placed  at depths ranging
from  0.2 m to  2.2 m and distances of 0.5 m  to  3.5 m from the houses.  The
pressure coupling  data  for  the  soil probes  provides an interesting comparison
to those data measured  at the test holes in the substructure surfaces.
     Pressure differences were measured between each soil probe or exterior of
each  test hole,  VK  (test  hole  sealed),  and  the basement.  All  other test
holes and soil  probes  were kept  sealed.  The basement depressurization,  PB,
was  measured  relative  to   outside.  All  pressures  were measured  using an
electronic micromanometer with  a minimum resolvable pressure difference of 0.5
Pa and  an accuracy of 1%.
     Air velocities as  small as 0.025 m/s at the open test holes  were measured
with  a  hot wire anemometer  attached to a flow adaptor designed to mate with
the  various-sized holes  (Figure 2).  Where necessary,  flow  rates  (IH) were
corrected  for the effects  of  the  size  of  the test  hole.   The  pressure drop
across  the flow adaptor and test hole  (PH) was  estimated using the engineering
formula for  laminar flow  through a  tube
        _  m                                                             [10]
      x     nr4

where     PH - pressure drop (Pa),
          x  = length of the test hole plus flow adaptor (m),
          H  . absolute viscosity of air, 1.8 x 10~s  kg/in-s,  and
          r  = radius of the tube (in our case, we used the radius of the
               flow adaptor - 0.0045 m).

Estimated pressure drops in the test hole and flow adaptor ranged from 0.01 Pa
to 3.5  Pa.   In theory,  a more accurate value for  this pressure drop could be
determined by  direct measurement; however,  the small pressure differences are
difficult to measure in practice.
     To determine  the  radon concentration in soil gas near the substructure,
G, grab samples of soil gas  from the test holes  were collected in evacuated
alpha scintillation flasks.  The radon activity in the flasks was counted on a
portable  photomultiplier  tube  and  sealer.  Because  grab  samples were  not
always   collected   concurrently  with  measurements   of   flow  and  pressure
difference,  samples collected at  other times during  the study were  used to
compute average radon  concentrations  for the test holes.  Uncertainties  in the
radon concentrations measured with  this procedure  are  estimated to be ±  20 X.

                            RESULTS AND DISCUSSION

     A  total  of  117  measurements  were made  at  75  test  holes in  the four
houses.  The average  calculated effective  resistances for  soil   (RS-EH?) and
substructure surfaces  (1^-^) are summarized in Table  1.  The data  are grouped
according  to location  of the test hole.  There is considerable variability  in
the   resistances   among  test  holes  as  indicated   by  the   large  standard
deviations.   By examining the geometric means,  several patterns are apparent
and  statistically  significant:  1) the effective  soil resistance that is  'seen'
by the test  locations across the  slab  floors and  in block wall  cavities  is
similar, probably  because  large  surface  areas  of soil  (and  for  the  block walls
-  wall  areas exposed directly to the outside  air) are accessible  to the  test
holes;  (2)  the slab floors are approximately five times more resistant  to soil
gas  movement than the  interior surface of the porous  block  walls;  and  (3)  for
all   locations,  except  those  in   the  soil  exterior  to the   walls,   the
substructure 'sees' the soil as being a factor of 2 to  6 times more resistant
to soil gas flow than  the  substructure surfaces  and the materials  very  near to
the  substructure.   We  also find that the  entire thickness  of  a block wall  is
many times more resistant  to  soil  gas flow than only the  interior surface  of
the  block - presumably  because  of  the coatings and sealants that  are  applied
to the exterior surface for waterproofing.   These data support the view that
the   flow  of  soil  gas into  buildings depends  to  a  lesser degree   on  the
 resistance of the building surfaces  below grade  than  on the resistance  of the
 surrounding  soil   and  materials.   It  is  important  to  recognize  that  low
 resistance values can result from  low resistance  in the materials  near to the
 test locations  or from the sum of  many parallel  resistances when the  test
 location 'sees' a large area of soil and building material.
      Average soil gas  entry  potentials, G,  for each  of the  75 test holes are
 shown  on  Figures  3  through  6.  Pressure  coupling ratios  from the  pressure
 field mapping tests are also shown so that comparisons  can be made  among the
 test locations.   See  Table 3 for  a  description of the symbols used  on these
 figures. Values for G ranged from less than 0.01 x ID'9 m3/Pa-s to 3.2 x  lO'*
 mVPa-s.   These data  are  summarized in Table 2 by the same  groupings as in
 Table  1   The geometric  mean  soil  gas entry  potential  is  highest for  the
 hollow  block walls,  probably  because of the  high porosity  (lower  effective

resistance) of the block wall material  and  because  the large exterior surface
area exposed to  soil  and/or outdoor air is available  to  most entry locations
on the interior  surface via the  interconnecting network of cavities.  Similar
conclusions were reached by Garbesi and Sextro (6).
     For  the  two houses   where  G was  calculated   for  different  basement
pressures, the mean  coefficient  of variation for replicates  at  35 test holes
was  49%.   The entry potential  did  not appear to  be  biased  by level  of
     If we assume the average soil gas entry potential  (0.73 x 10'6 m3/Pa-s)
is for an effective  area of 1 m2,  then  for  a house  with 175 m2 of below grade
surface area at  a natural  depressurization of 3 Pa,   the soil gas entry rate
is predicted to  be 1.4 m3/h. This soil  gas entry rate  is  similar to  the 1 m3/h
calculated by other  researchers  (6,15,16,17),  and lower than that measured in
houses on highly permeable  soils (18).
     Data for the  radon  entry potentials,  E, at each  house  are  also shown on
Figures  3  to  6 and  are   summarized   in Table 2.  For  73  test  holes,  the
geometric  mean  radon entry potential  was highest  for the  test locations in
subslab aggregate.  Although the test holes into the block wall  cavities had a
slightly higher  soil gas entry potential, the  subslab test  holes had greater
concentrations of radon in the  soil  gas which  compensated  for  their smaller
soil gas entry potential.   Calculated values  of E  ranged from less than 0.1 x
10'3 Bq/Pa-s  to  1300  x 10'3  Bq/Pa-s.
     When reviewing  the  radon entry potential  plotted on Figures  3 to 6, we
find that the areas  of highest potential generally  coincide with the locations
where  the pipes  of  successful  subsurface  ventilation-depressurization  (SSD)
radon  control systems  were placed through  the  slabs.  For  these  houses,  a
'high'  radon entry  potential  would be considered  greater than  approximately
15   x   10'3  Bq/Pa-s.  Since   the   entry  potentials  were  calculated   after
installation  of  the  SSD  systems,   these   indices   appear  to  provide  a
quantitative  method  for  replicating   the  intuitive  approach  of  successful
mitigation  contractors.  House LBL12 is an  exception, where it was difficult
to  bring indoor radon levels  below the  target concentration  of  148  Bq/m3.
There  were areas of  high radon entry potential in  this house  that were not in
proximity  to an SSD  pipe, and may have  been  the  sources  of  inadequately
controlled radon entry (Figure 4).
     In  general,  the  radon  entry  potential  may  indicate  the preferred
locations  for SSD pipes, but  will  not provide information about  the ability of
a  specific  SSD system  to  reduce radon  entry  rates.  The  pressure  field
extension test that  uses a vacuum  cleaner or  depressurizing  blower remains the
best technique  to measure the extent  to which a  SSD system can  reverse the
natural  pressure gradient  around  a substructure,  and  therefore  control  radon
entry  (10).  Combining  results  from the  pressure   field extension test  with
identified areas of  high radon entry potential may  assist contractors  deciding
on  the placement of SSD pipes.  When the soil gas  (and radon) entry potential
is  high  for  a  particular location,  but  the pressure field extension or
connection to the vacuum  .is poor,  obstructions   or  high permeability  short
circuits  are probably blocking or intercepting the  pressure field from  the
vacuum.   The problem  then is to  provide access to  the  areas  of high  radon
entry  potential.
     The  geometric mean radon entry potential for  each of the four houses was
compared  with  the   average  indoor  radon  concentration,   measured  between
September 1 and May 1  and weighted by  the volumes  for various  zones  where
indoor radon  was measured.  From the  lowest  to  the highest  average  indoor
radon concentration;  540,  620,  650, and 660 Bq/m3, the geometric  mean  radon
entry potentials were 6.4, 10,  7.2, and 18 x 10'3  Bq/Pa-s,  respectively. Only

for  the third  house  listed (LBL13),  does  the  geometric  mean  radon entry
potential  fail  to  trend  with  increasing  indoor  radon  concentrations.  The
geometric mean radon entry potential (assumed to be for an effective area of 1
m2}  can  also  be used  in a  mass  balance equation  along with  actual  surface
areas and  structure volumes and  an assumed  pressure  difference of  3  Pa and
ventilation  rate  of  0.5  h'1   to  calculate  a  steady-state  indoor  radon
concentration  for   each  house.   The  correlation  is   poor  between  these
calculated values  (42,  39,  24,  and 103  Bq/m3,  respectively) and  the actual
radon levels.   And we see that using the radon entry potentials under-predicts
actual radon levels by factors of 6 to 30.


     That the current development and  application of  these new parameters may
still have weaknesses,  is suggested in  some  of the data  just  presented. The
electrical analog is imperfect since  it  assumes linear flow characteristics -
which may not occur either under natural depressurization at entry locations
or  under  the  greater  mechanically-induced  depressurization  at   both entry
locations and  test holes.  A  more appropriate  analogous circuit  would also
include  capacitance  to represent the  storage and discharge  of  radon that is
assumed  to occur  in spaces near the substructure  in  response to time-varying
driving forces.
     In addition, the  assumptions,  REC-EFF = &F-EFF an<* RSC-EFF °  &S-EFF>  are not
exactly valid, since the paths  for air flowing through  the soil and building
surfaces  are  different  in the  two measurement  conditions.  The  effects  of
inhomogeneities  in  soils,  near-house  materials,  and  substructure   surface
materials on the  assumption have not  been  examined.   Although  the derivation
is generally not sensitive to the  pressure  drop across the test hole and flow
adaptor  (right  hand  term  in the  numerator  of Equation  6),   the  term was
occasionally as large  as 55% of  the substructure depressurization  (PB).  This
term has  the  greatest  impact when  the pressure drop across  the  test hole is
large or when substructure surfaces are  leaky (small  Z).   Since the test hole
diameter will also  have  an  important effect  on  the  measured  flow rate, all
test holes  should be drilled to  the  same  size or corrections  to a standard
hole  size  (for this  study,  a  9 mm  dia.  hole)  should be  made.   Then entry
potentials may  be  more  accurately compared  between  test holes  and  between
     The soil gas entry  potential at a particular location is affected by all
soils, materials, and  openings in  the  below-grade surfaces around a building,
but to  a greater  degree  by  those  nearby or  connected by  a high permeability
path.  More study is required to determine  the distance from a  test hole over
which  the  soil gas entry  potential   is  applicable.   For  example,  if  RF2 in
Figure  Lb is  small  (i.e.,  highly permeable  subfloor  material),  then the
distance  is  large.  If  RF2  is  large,  then  the  entry  potential   is more
localized.  When these relationships are better understood,  the test holes can
be better  placed  to best  represent the  soil  gas entry  throughout the entire
     The time-varying  nature of  many  of the measured parameters  may create
difficulty in  using a  one-time  determination of  soil gas and  radon entry
potentials to represent   typical entry potentials for a  house.  For example,
the  radon concentrations  measured in  grab  samples  collected  from  the test
holes  under  natural   conditions  showed  large  variations.   Repeated   samples
(between two and seven) were collected from 45 test holes at  six houses over a
12-month  period  when  radon  control   systems were  not operating.  The mean

coefficient of  variation for the  radon concentration  at  each test  hole was
79%.  Concentrations at  the  same test hole varied  by up to a  factor of 1000
from one  sample to another.  It is possible  that  these large variations  in
radon concentrations were  caused by changes in wind,  radon  production in the
soil, or soil gas flow rates into the houses.   The radon concentration in grab
samples  from  individual  test   holes  collected  during  mechanically-induced
depressurization can also be considerably different from concentrations during
natural conditions.  However, when concentrations measured at 99 test holes in
seven houses during mechanically-induced depressurization were correlated with
corresponding concentrations during natural conditions, we found a correlation
coefficient  (R)  of 0.68 and  that  the means for  the two conditions  were not
significantly different.   This  suggests  that  radon  concentrations  in samples
collected during a  brief period of mechanically-induced depressurization may,
on average, be representative of concentrations under natural conditions.
     In  addition  to  variations  in  radon concentrations,  it would not  be
surprising  to  also observe  changes in  the soil gas entry potential  due  to
changes in soil permeability (for example, caused by precipitation or a moving
water table).  More study  on the time-dependent variation of these parameters
is required.


     A procedure has  been described  to  determine the  potential  for  soil gas
and  radon to enter a house  at  substructure surfaces  by convective flow from
the  soil.  The  necessary  field measurements  of flow,  pressure,  and  radon
concentration are  relatively simple and utilize commonly available equipment.
These parameters for entry potential may be useful to:  (1) identify areas in a
substructure with  the  potential for comparatively  high soil gas  entry rates;
(2)  compare  the  relative  leakiness  of  below-grade  surfaces  in  different
houses;  (3) provide approximate  measures of the resistance of the substructure
surfaces  and soils/materials around the substructure  and  provide a basis for
establishing the relative  importance of these  features  to radon entry; and (4)
identify  areas  in  a substructure with potentially  high radon entry  rates for
placement of radon  control systems.
     Initial measurements  in four houses  indicate  that the soils  surrounding
the  houses  are approximately three times more resistant  to the  transport of
soil   gas  than  the  substructure   surfaces   and   the  materials  near  the
substructure.   Soil  gas has  approximately twice  the entry potential  at the
surfaces  of hollow block walls  than at  slab  floors,  but the slab floors have
much higher  radon  entry  potentials  because of  the greater radon concentrations
below  the slabs.
     Modifications  to  the simple  electrical  resistance circuit  used in this
development   to represent  the  substructure/soil  system  may  improve  the
predictive  capability  of  the procedure.   In  addition, determination of  these
new parameters  in  additional houses and under changing  seasonal conditions is
necessary to more  fully examine their  suitability as diagnostic and  research


     We  appreciate the  insightful  comments of William Fisk. Ashok  Gadgil, and
Mark Modera;  the program support of David Sanchez at AEERL, U.S. EPA;  and the
cooperation of  the families  of  the  seven New Jersey houses.
     This work was supported by the  Assistant Secretary for Conservation and

 Renewable Energy, Office  of  Building and Community  Systems,  Building Systems
 Division; by  the Director, Office  of Energy Research,  Office of  Health and
 Environmental Research, Human Health and  Assessments Division and Pollutant
 Characterization and Safety Research Division of the U.S. Department of Energy
 (DOE) under  Contract  No.  DE-AC03-76SF00098; and by  the U.S.  Environmental
 Protection Agency  (EPA)  through Interagency Agreement DW89931876-01-0  with
 DOE.   This paper has been reviewed in accordance  with the  U.S.  EPA's  peer and
 administrative review policies and approved for  presentation and publication.


 1.  Nazaroff,  W.W.  Predicting the rate  of  radon-222 entry from soil  into the
 basement  of   a  dwelling  due   to  pressure-driven  air  flow.   Radiation
 Protection Dosimetry.  24:  199-202,  1988.

 2.  Nazaroff,  W.W.  and  Sextro,  R.G.  Technique  for  measuring the indoor  222Rn
 potential.  Environmental Science  and Technolpgy. 23:  451, 1989.

 3.  Schery,  S.D. The design of accumulators and their use in determining 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 of the U.S.  EPA Symposium on Radon and Radon Reduction Technology.
 EPA-600/9-89/006a. pp.  5-61 to 5-74.  Research Triangle Park,  N.C., 1989.

 5.  Tanner,  A.  B.   Measurement   of  radon  availability  from soil.   In:  M.A.
 Marikos  and  R.H. Hansman (eds.),  Geologic Causes  of  Natural Radionuclide
 Anomalies,  Proceedings  of  the  GEORAD Conference, St. Louis, MO., April 21-22,
 1987:  Rolla, MO., Missouri Dept.  Nat. Resources, Div. Geology and Land Survey
 Spec.  Pub. No. 4, p. 139.

 6.  Garbesi,   K.,   and  Sextro,  R.G.   Modeling   and  field   evidence   of
 pressure-driven  entry  of soil gas into a house  through  permeable below-grade
 walls.   Environmental Science  and Technology. 23:  1481-1487,  1989.

 7.  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. 21: 459,  1987.

 8.  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. In:  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,  N.Y., 1987.

 9.  Figley, D.A.  and Dumont,  R.S.  Techniques for measuring the  air  leakage
 characteristics of below grade foundation components.  In:  Proceedings of the
 82nd  Annual  Meeting  of the Air  and  Waste Management  Association. 89-79.4,
Anaheim,  CA., 1989.

10. Gadsby, K.L.,  Hubbard,  L.M.,  and Harrje, D.T.  Rapid  diagnostics:  subslab
and wall depressurization systems for control of indoor radon. la: Proceedings
of  the  U.S.   EPA  Symposium  on  Radon  and   Radon  Reduction  Technology.
EPA-600/9-89/006a. pp. 3-69 to 3-85. Research Triangle Park, N.C., 1989.

11. Matthews, T.G.,  Wilson,  D.L.,  TerKonda,  P.K.,  Saultz,  R.J.,  Goolsby,  G.,
Burns,  S.E.,  and  Haas,  J.W.  Radon  diagnostics:  subslab communication  and
permeability measurements. IQ: Proceedings of the U.S.  EPA Symposium on Radon
and Radon Reduction  Technology. EPA-600/9-89/006a.  pp.  6-45 to 6-66. Research
Triangle Park, N.C., 1989.

12.  Turk,  B.H.,  Prill,  R.J.,  Sextro,  R.G.,  Harrison,   J.   Intensive  radon
mitigation  research:  lessons  learned.  In:  Proceedings  of  the  U.S.  EPA
Symposium on Radon and Radon Reduction Technology. EPA-600/9-89/006a. pp. 6-25
to 6-44. Research Triangle Park, N.C., 1989.

13. Turk, B.H.,  Harrison,  J.,  Prill,  R.J.,  and  Sextro,  R.G.   Developing soil
gas and 222Rn entry  potentials  for substructure surfaces and  assessing 22ZRn
control diagnostic techniques.  Submitted to Health Physics. Lawrence Berkeley
Laboratory Report No. 27319, Berkeley, CA.  1989.

14.  Sextro,   R.G.,  Harrison,   J.,  Moed,  B.A.,   Revzan,  K.L.,  Turk,  B.H.,
Grimsrud, D.T.,  Nero,  A.V., Sanchez, D. C. , and  Teichman, K.Y.  An intensive
study of radon  and remedial measures in New Jersey homes.  In:  B. Seifert, H.
Esdorn,  M.  Fischer,  H.  Ruden,  and  I.   Wegner  (eds.),  Indoor  Air   '87:
Proceedings  of  the  4th International  Conference on  Indoor  Air  Quality  and
Climate.  Institute  fur  Wasser-,   Boden-  und   Lufthygiene:  Berlin.  2:  pp.
295-299, 1987.

15.  DSMA Atcon  Ltd. Review of  existing  instrumentation  and  evaluation of
possibilities for research  and development of  instrumentation  to determine
future  levels  of  radon  at a proposed  building  site.  Report  No.   INFO-0096.
Atomic Energy Board. Ottawa, Canada. 1983.

16. Fisk, W.J. and Mowris, R.J.  The impacts of balanced and exhaust  mechanical
ventilation on indoor radon.  In: B. Seifert, H.  Esdorn,  M. Fischer, H. Ruden,
and  I.  Wegner (eds.),  Indoor  Air  '87:  Proceedings of the  4th  International
Conference on  Indoor Air Quality  and Climate.   Institute  fur Wasser-, Boden-
und Lufthygiene: Berlin. 2: pp.  316-320, 1987.

17. Nazaroff,  W.W., Feustel, H.,  Nero, A.V.,  Revzan, K.L.,  Grimsrud,  D.T.,
Essling, M.A.,  and  Toohey,  R.E.   Radon transport into a  detached one-story
house with a basement. Atmospheric Environment.  19: 31, 1985.

18.  Turk,  B.H.,  Prill,  R.J., Grimsrud,  D.T.,   Moed,  B.A., and  Sextro, R.G.
Characterizing  the occurrence,  sources, and variability of  radon in Pacific
Northwest homes.  Submitted  to JAPCA.  Journal of the  Air  and Waste  Management
Association. Lawrence Berkeley  Laboratory  Report No.  LBL-26960, Berkeley,  CA.

                                              BRSEtlENT  INTERIOR
                                BLOCK  UflLLS
  Figure 1  Drawing  of  substructure during  pressure field mapping  and  basement depreaauriiation (a).
  A simplified electrical analog of the various flows,  pressure drop*, and reaiatancea during the teat
  depicted in (a)  is shown  in  (b).  The dotted line indicates the variables associated with an open
  teat hole.  A further simplification ia shown by the circuit on the left aide of (b).


           • 9


, /






                           HOT  UIRE
                            TOP   VIEU
Figure 2. Flow adaptor device used to measure the flows through the teat holes. The bottom opening
of the adaptor seals against the test hole surfaces, while flows are measured with a  hot wire
anemometer probe placed Inside the open end of the adaptor.

PFIICI>> a. 81
ocui pFmi)« a. 87
pFnu>. i.e t
PFt1>I>» 8. 88
PFHCI >• a. aa
PFH< i>- a. aa
PFN(I>* a. 88
Qimg pFn< ia>> a.ii
PF«1)« 0. 72 \
\ *
0 l^|"^ 1 1 ,U6>e.3B
PIPE E> 18 G* 1.2
	 £• 14
- J I: i3M VRC ma9\9^
pn>a.aa p|{(t,,.e ^55Lfl?F
?! I'l1* iuia I
PFIK ie>«a.sa__L_
_ v.
iuaa —
1U22(B> PFmiai-0. 19
— pFncia»a. IB E> sa \
r C* 8. 87 \
pFticie>- a. IB onst • • • •
£?5 PFn. 8.88 / OHS? »
E« a.'i ,--• • pFti« a. aa \
ORS3 x^ "BS2 OHSb
PF(1CI>« 8.91 PFIK1*. 8.84 PFMCI).
5£H! OCH2 PFn< B.ei
Fnci>« a. a? PFn•••" ^
• c- a. a •
1FB E< S'a 1F9
pFn>a.ia pFn>a.78
G> 1.3 G> l.a
E* 18 E> 8.4
PFncia>*8. ia
GI 1.1
1 IU1S«B)
/ IF n PFnciei -a. ae ifi8
I PP^iat.n «^C« 1.6 PFn(ia>«a.6S
I I-, i S ~X^E« 11 G« 1.3
1 u» i • r j^ p _ A A ,
/ E. 33 .S E' ••• /
— • ^/^ IHiltB.' FS12 • /
^^ PFn(lB)*8.2B PFnfiaiaa TO /
1-^ 1
• • •
PFIU I )• 8.91 .
8.81 *
PFn(I>« 8.84
| ( 1 |>
i —

G> 1.1
£• 1.7
PFH« ieS-8.28
c* a. a
E* 1.2
Figure 3. A site plaa of  house LBL08 that  shows the  locations  of probes  in the soil  around the
house, holes  drilled through  the  slob floors  (solid dots),  test holes  drilled into  and through
hollow block walls  (vertical and horizontal lines),  and the  pipes  for the radon  control systems.
Data for pressure coupling  ratios  during  pressure field mapping  tests (PFM),  and for  soil gas (6)
and radon (E)  entry potentials are placed  under the identification  code for the  test holes.  See
Table 3 for  more complete descriptions of  the codes and  symbols that are used.

PFM 1 ><
pFmi>« a.ra

l_ Dl_ 1 C N
OCMl* • »
pFmi>. 8.TT SSio. e.B8
OBN8 OBNB OBM. a.Tl PFH(I» OR PFtKI» a. 63 1
OBNI g) _ _ _
ppmi>. a. 68
ome 0|W8 PFnciJ- B.BI FFmi>> a.es J£?
,.„ • •pFmi>. 8.66 •PFMI»O* • • /£"•!*!"*'**
BBB / '" "
iui8iB>| ^E SSnyl-aas 'if? IpBHi >••.!• IIH | ,/
'; •;* " "* •HOLE IE* 11
iuia — j
xpFmia>>a.»a 1 •
J.B.SB j; •••• pFiui»a.aa ,US(B>
""alS' *"*3 PFmia>«a.as
/PFIH ia>-a. la
' C* 1.8
/ CRflUL PFH< !•>••. l« 1 '" *"*
JL iF7 g; {;S ' 	 ,„, BHSEHENT
• pFmia>.a W ^pTn(i>.B.8B( pTJuo-ees
C> 8.0
E* 88
PFIKia>>8. 13
	 ... C« 8.8
	 If? «• 88 \
PFH(I8>*B. 6-ne!Bl"B"M ~~*
PFM ia>*a. ia E* 8.8
O 8.8
E* 188 IU12CB)
IU6 PFt1-a. 13
1 PFTHD.a.TT 1 \\ }." I -*
'"»•»<•>/ ,F9 /
I'"!1!'"*'1* pFn«ia>>a.B9
c. D. , S- e.88
PORCH ' •• ••»
PFH(I>> 8.81
PFIKI>> a. an
% —
PFH< 1 >• 8.86 — jj
ppmi>- a. ai
PFII* 8.88
                                                             rasa  •         oasi   a)
                                                             PFR(I>*  B.BB      PFtKI). 8.61
Figure 4.  Similar to Figure 3, except for  LBL12.
                                                                                 PFIK!>« 8.81

_|_ v ) pFmi>" i.a PFHC t >• a.ai
DECK eau a) a) • PFH. a.aa •) " _PFIKI» 8.77
PFn(l8>* 8.18
PFIKM>> a.ia
GI a.ai
•> a.c
PPM 18
C« I.T
c* e.e
1 1 1 '
^',«2f S?'""'""""^
PFIU18»a.l8 PFn.B.7B
Q« 1.1 	
i> 11 ... vnc.
•F» MOLE 	
pFmi>*a.as — ^ j
BflSEMENT t- 18E<1
ma pFncia>> a .m
pFmi>.a.aa ^cla>: a —
pFmia>-8.87 e> afSa
c> a.a J E" H'8
\ IUBCB) /
\ pFti.a.ie /
>t « iuia*a /
•a.ci pFmia>>a.ia \ K"^'"*'*1 Jt—
•a.ia pFiuia>>a.ar \ |. Ji ,
pFit(ta» a
pFncaa>* e.7
c* a.aa :us
c- 11 , pFH(iai> a
// pFH> a.ai
/ pFMia>> a
/ c« a.es
/ ۥ 1.8
/ OO81
/ wPFmi>« 8.78
pFn> a.as
PFIKl>. 8.TB m
9 OCBE *
PFIKI>> a.
PFH> 8.71
PFH<1>. 8.73
• oasa
PFN(1>. 8.71
^^ 01181
^ PFml>. 8.78
^ •
/pFiKia>» a
ppn* a.ai
C* 8.81
E. s.i
/pFn> a.sa
FFncaa>. a. 38
C- 8.18
• • 38
I             I
                                                                             •PFIKI>> a. ea  • PFHCIX e.eg
                                                                                PFHCI>> B.OT
                                                                                                     PFH> B.T8
                                                                      PFIKIX B.TT
                                                 PFH( 1 >• B.TB
                    Figure S. Similar  to Figure 3, except for LBL13.

OCMt •
PFn« M
                                        OCN2  •
                                        PFmi» 8.M
                                PFtl- 8.36
                                                                   • OCH3
                                                                      PFIKD-  8.91
                  PFH< !)•  8.68
PFHC !>•  8.68
PFHCIJi  8.38.
                                                                                         PFIKI)- 8.91
                                                                                          PFn-  8.98
             PFHCIX 8.TT
      PFIKI >•
lu* lu.xn, | I"* »»«•> 1

PFHC 10 )*0* SB •em 41* \ a" •« PFH( !• >• 0* 33 PFtlC
,.,. ,1 il.- •
8B>" 8.88 PFIK18}* 8.88
PFnCM»B.87 £".?•>• •••• PFh<88». 8.88 C- 8.88 "."\"" ''"
e. Big e> »•• ,_, E. 888
— , 	 IFH PFTIC18>* 8.89
PFn(18><8.33 iun».» PFH<88>' 8.18
IUIZ«B> PFnCM)- 8.18 !_!!£_' «* ';•
Pfitiimt.a SB c> '-I PFn(18>> 8.98 c* "
PFIIC88>>8ll9 f' ** PFnC88» 8.98
C' 1 T B* I * • 0J IF 1


PFHC88>- 8.82 •
E* 1.2 E* >>T PFn> 8.88 SSD PIPE
PFIt<18>- 8.8


PFn(S8>- 0.8
C- 8.91
C' 18

PFnC18>* 8.98
PFIUM)* 8.98
C* 1.6
E> 1.1

PFmi8>> 8.19
PFHC88>> 8.18
_ C- 1.8
• E> 1.8

C* 8.88
• EC 8.1

•• _ 	 ^
1 • IU6CB) AB
pFmii)- B ^^
PFn<88>* 8.B8
6« 1.8 	
£• 18

PFnt 12* 8. 98
PFIU18>> 8.98
PFHCBB)* 8.88
C. 8.81
i- 8.8
PFnciB)* e.ds
PFnC88< 8.18 	
C* 2.8 » 8)
E* Z68
IU7C8) 1
PFm,ie>-e 1
87 PFnrtaj.a.aa
\\ !;T9 if •
pFncie>-8. !•
PFnC88>>8. 13
, C- 8.8
1 E> 89
                                                                                                                                     PFnCD- 8.82
                                                                                                                                   QBE I
                                                                                                                                   PFH<1>.  9.72
                                                                                                                               ' PFMI>>  8.6T
                                                                                                                                OHt 8
                                                                                                                                PFHCI>• OR
                                                             C>  1.8
                                                             E>  1.1
                                       PFfl(38>*  8.78
                                       C< 8.88
                                       E* 8.1
                                                       PFnCI8>> 8.28
                                                       PFm88>- 8.83
                                                       C«  1.1
                                                       E>  8.6
                     Figure 6  Similar  to Figure 3.  except for LBL1AC
                                               PFm I >• tf. id
                                                                          CFn< I >• Urt

Table 1. Statistical Summary of Effective Resistances for Soils and
         Substructure Surfaces from Four Houses
Below Slab
Test Hole Location
Block Wall Hall
Cavity Exterior
Soils, Rs-EFF (1C6 Pa-s/m3)
Geom. Mean
Geom Std Dev
Arith Mean
Arith Std Dev.
Substructure Surfaces
Geom Mean
Geom Std Dev.
Arith Mean
Arith Std Dev
Table 2 Statistical
Four Houses
, RF-EFF <106 Pa"s/
0 65
4 8
3 0
Summary of Soil Gas
Below Slab
2 2
and Radon
Test Hole
Block Wall
7 8
3 6
Entry Potentials from
Soil Gas Entry Potential, G (ID'S m3/Pa-s)
Geom. Mean
Geom Std Dev
Arith Mean
Arith Std Dev
Redon Entry Potential
Geom. Mean
Geom Std. Dev.
Arith Mean
Arith Std Dev.
4 8
, E (10-3 Bq/Pa-s)
6 7

4 3
0 73
3 6

Table 3. Key to Symbols for Figures 3 to 6

I = Indoor
      W = Wall
           (B)   = Into block wall cavities
           Blank = Through wall, into soil
           Top   = Opening into block at top of wall
      F = Through floor

FS = Floating slab

0 = Outdoor
      A « 0.5 m from house
      B « 1.5 m from house
      C « 3.0 m from house
      N,E,W,S = Orientation to compass direction

1,2,3,... = Arbitrary sample location number

SSD PIPE = Subsurface depressurization pipe for radon control

BWV PIPE = Block wall ventilation pipe for radon control system

VAC. HOLE = Test hole where vacuum was placed for pressure field
            extension tests


PFM = Pressure field map coupling ratio  [•=	"""""	I
                                         \' OUTSIDE ~ r BASEMENT J
      I  = Initial test, basement depressurized to -30 Pa
      10 = Basement depressurized to -10 Pa
      30 = Subsequent tests with basement depressurized
           to approx. -30 Pa
      OR = Over range

G = Soil gas entry potential (10~6 m3/Pa-s)

E = Radon entry potential (10~3 Bq/Pa-s)



               by:  F. Stoop, F.J.  Aldenkamp, E.J.T. Loos, R.J.  de Meijer and
                    L.W. Put
                    Kernfysisch Versneller Instituut
                    Rijksuniversiteit Groningen, Zernikelaan 25, 9747 AA
                    Groningen, the Netherlands
     In situ  measurements  in a Dutch  test  dwelling of radon  exhalation of
walls, floor and soil in and under  the dwelling  are interpreted in terms of
source strengths. Continuous measurements of  radon  concentration in various
compartments  (e.g.   crawl  space  and  living  room)  of   the  same  dwelling
together  with measurements  of  leakage parameters,  temperature,  pressure
differences between  compartments  and  ground-water table are used  to obtain
information  on  the  dynamic aspects   of  the  radon  infiltration.  Source
strengths and dynamic variables are used as  input for a  multi-room model to
describe  the  variations   and   interrelations  of  radon  concentrations  in
various compartments. Despite the relatively  low  radon concentration in the
soil gas (~ 10000 Bq.nf ) and the low  permeability  of the soil, preliminary
analyses suggest pressure driven flow through the soil to the crawl space to
be an important radon source. A possible influence  of precipitation on this
flow  and   on  the  source  strength  for the  soil  of the  crawl  space  is
*) Research supported in part by  the Dutch Government and  the Commission of
   the European Communities.

                               1. INTRODUCTION
     Radon  being  a noble gas has the property  to move from its location of
birth  by   two  physical  processes:  diffusion  and  pressure   driven   flov.
Diffusion is  the  process by which radon is transported due to  concentration
differences.  Pressure   differences  cause  transport   through  e.g.   porous
materials   and  openings  in  walls.  The  importance  of   the   transport  is
determined  by the  product of current and concentration; a  low current  with a
high  concentration may  be  equally  important  as  a high current with  a low

     To   design    effective   countermeasures   against   elevated    radon
concentrations, source strengths, entry routes and  driving forces have to be
known. Source strengths  can be estimated for a dwelling by measuring in situ
the exhalation  rate of  various  components  of a  dwelling.  Entry routes and
driving  forces  may be determined from  measurements of radon concentrations
simultaneously with other relevant physical parameters.

     The dynamic  aspects have been measured in houses  in  New  Jersey  by the
groups of  Berkeley and  Princeton-Oak Ridge.  So  far, to  our  knowledge,  no
results  have  been  published  in  the   open   literature.  Also in  Finland
continuous  measurements  of  radon have been made  by Arvela tt  al.  (1). They
compare  the variation in  measured  radon  concentrations  of 33 houses with
variations  calculated with a model which relates  indoor radon concentrations
to radon entry rate, air infiltrations and meteorological  factors. In  houses
with a slab on  ground the measured seasonal  variations are often explained
almost entirely by pressure  driven  flow. Diffusion is according to ref. (1)
a  significant  source  in case of  large porous  concrete  walls  against the
soil. According to their measurements and calculations  the diurnal maximum
in the radon concentrations  occurs  in  the  morning and is caused by pressure
driven flow.

     Median radon  concentrations  in the Netherlands (2-3)  are  comparable to
values found in the Federal Republic of Germany and England. However,  due to
the absence of rock near the  surface, except  for the far south eastern tip,
the percentage  of  dwellings  with  high concentrations  is smaller. In the
absence of  the need for  immediate mitigation an investigation of radon entry
and transport  for a  "typical"  house  may  provide knowledge which  leads  to
reduction of the population dose due to exposure  to radon.

     In  the present investigation  measurements  are  carried out in a test
dwelling in Roden in  the  northern  part of the  Netherlands.  As most  Dutch
dwellings it has a crawl space  which  is the result  of  excavating  a part  of
the soil and refilling it partly with sand. The soil in the 50 cm high crawl
space is uncovered. Radon  concentrations in the  crawl  space and  the  living
room  have  been  monitored  with  time-integrating  detectors   since  1980.

Unexpected   variations   (4)  were  observed  in   the  year  average  radon
concentrations,  which are  likely due to  fluctuations in  the  ground water
level often  inundating the  crawl  space.

     The aim of  the  present investigation  is to obtain better understanding
of  the behaviour of  radon concentrations in dwellings. This goal is  thought
to  be  achieved  by  developing  a  dynamic   model   in  which  the  radon
concentrations are calculated from static source strengths and air currents.
For the source strengths  the exhalation  by walls,  floors, ceilings and soil
were measured; air currents were deduced from measured pressure differences
and measured or  estimated leak sizes. Moreover  a number of possibly relevant
parameters as ground water  level, relative humidity,  temperature, barometric
pressure, and precipitation rate were  measured and  ventilation  rates were

     In this paper we present measurements  and  calculations for two  periods
of  one week  with different  characteristics. An hypothesis  is formulated in
which rainfall plays an  important  role  in changes of values of exhalation
rate, source strength and pressure driven flow.

     The test house  is  located in Roden,  southwest of the city of Groningen
in  the  northern part of the Netherlands.  It  is located on  the  edge of the
"Drents  Plateau";   the  local  soil   consists   of  silty  fine  sand  with
intercalations  of  boulder  clay  upon  "pot  clay". The  impermeability of the
pot clay causes large and  rapid variations in  the groundwater level. Prior
to  the  installation  of  a water drainage system in February 1989 the ground-
water levels often almost  reached  the soil surface.  Such situations are not
uncommon for recently build houses in various parts of the Netherlands.

     The  dwelling  is a single  family house,  built  in  1973.  It  has  a
rectangular floor  shape with  a  wall  extending  from the crawl  space to the
roof, which divides  the  house  into a northern and southern part which can be
regarded as  reasonably  independent.  In the present  investigation  only the
southern part  is investigated.  This  part  consists of  a crawl  space  and a
high loft living room, with an "open kitchen", directly covered by the roof.
The ceiling of  the crawl space consists of prefab hollow concrete bars which
are covered by  a layer  of  a  few centimeters cement;  in the living room this
layer is  covered with  ceramic  tiles.  The outside  walls are  cavity walls
consisting of  masonry (about  10 cm thick);  the cavity has  been  filled in
1976 with a polyurethane foam  for thermal insulation. The floor of the crawl
space consists  of  uncovered  sand and  is  situated  at about  40  cm below the

surrounding soil of the yard. Around the house, at a distance of 1 to 1.5 m,
a ring type drain has been installed in February 1989 at a depth of 70-90 cm
belov  the surface.  The dividing  vail  in the  crawl space  has  tvo crawl
openings  in  addition  to  feed  throughs   for  central  heating  (water)  and
utilities. In June  1989  these openings have been closed  by  5 cm thick foam
board. All remaining openings in the wall were sealed with caulk. In October
1989 a 20 cm diameter ventilation  duct  plus  fan was installed  in order to
(de)pressurize the  southern  part of the  crawl  space.  Natural ventilation of
the  crawl  space occurs  via  ventilation  shafts  to  the outside air;  in
principle these shafts should have no connection with the cavity wall.
     Radon  exhalation  rates were  measured with  a  device developed  from a
prototype designed by Ackers  (5).  A description of  the modified version and
its collection properties is given in ref. (6). For a measurement the device
is placed  on a surface;  soft-rubber rings at  the  ends of  the two coaxial
cylinders are supposed  to fit air-tight to the surface.  For rough surfaces
caulk is applied to the surface and  the outer ring after the device has been
mounted. For measurements in  soil a 30  cm long cylinder  with a  flat and
smooth  top  surface is placed  in  the soil  on  which the device is mounted.
Prior to each measurement the device is flushed with dry nitrogen.

     The exhalation  rate,  E,  is determined from a  fit to  the  growth curve
for the radon concentration C

     C(t) =   * (1 - e'Xt).                                           (2.1)
In this equation  A is the area  covered  by the inner cylinder  and  V is the
measuring volume. The quantity X is  the  effective decay constant and is the
sum of  the  nuclear decay constant of  radon and  a  constant  associated with
leakage by diffusion from the measuring volume to its surrounding.

     Radon  concentrations  were  measured with a  radon  meter  based  on the
Lucas  cell   principle.  The  device  is  designed  for  low background,  high
efficiency and short measurement time. The efficiency for radon with respect
to radon daughters was optimized by  segmenting a  13 cm diameter, 30 cm long
cylindrical cell into 8 longitudinal sections. The cell and segment wall are
covered by ZnS(Ag) except for  the side where the cell  is  mounted onto a 13
cm diameter,  low  background photo multiplier  tube. Due  to  this geometry a
high efficiency for radon is obtained: about 50%.  The efficiency  for each of
the two a-emitting radon daughters is about 60%.

     The setup was operated in a quasi-continuous mode: a sample  is taken by
flushing  the cell  over  a  filter and  a  drying  column  for  five  minutes;
subsequently  the sample  is  counted for 25 minutes  and  the number of counts

is stored  in  the memory of a data logger.  Simultaneously the values of  the
following  parameters averaged  over  25  minutes  are  stored in  the memory:
pressure differences between  the  outside air  at  three sides of the dwelling
at  the  crawl  space  ventilation  openings,  the pressure  difference between
living room  and crawl  space,  temperature and relative  humidity  of outside
air, air in the  living  room and in  the crawl  space, temperature of the soil
near the house and in the yard and barometric pressure.

     After transfer of  the data to  the central computer the counts from  the
radon meters  were converted  to radon concentrations  after corrections  for
daughter activity and background.
                                 3. RESULTS



crawl space
soil d = 58
11 d = 48
" d = 46
inner wall
outer wall
living room
(V*26 m)
11 1.78 ± 0.08
0.88 ± 0.06
1.02 ± 0.06
2.44 ± 0.06
1.96 ± 0.04
3.40 ± 0.05
(V=208 m)
1.34 ± 0.04
0.26 ± 0.05







   water level relative to the soil surface, averaged over the measuring
 .  period
   taken identical to the value of the floor in the study

     Table  1 lists exhalation rates,  effective decay constants and  surface
 areas  for  the perimeters in the crawl space and living  room. The  exhalation
 rate for the ceiling is taken equal  to  the value  obtained for  the floor  in
 the study  because the value for  the  ceiling could not be measured directly
 due to its  curved surface.  In  the study  the prefab concrete  bars are  covered
 by a few centimeters  of  cement. In the table three measurements  are reported
 taken  at different levels of  the  ground-water  in  the crawl space. Here one
 notes  a decreasing exhalation rate with  a  decreasing thickness of the sand
 layer  between the exhalation meter and the ground-water. In the crawl  space
 the exhalation  rate  of  the  soil  is a factor of two to three smaller  than  of
 the walls  and the ceiling.  The product  of  surface area and exhalation rate
 makes  that  the  ceiling yields  the  largest  contribution to  the diffusive
 source strength  in the  crawl space. From the values  of  X   one notices the
                                                          af f
 large  value for  the  inner  wall.  This value indicates  the  roughness  of the
 surface  which made it difficult to mount  the device leak-tight.

     Fig.  1  shows from top  to  bottom four pressure differences  P -P   with
                                                                  i  cs
i=west,  south,  east and living room,  respectively,  and P   is the pressure

in the crawl space,  the deduced ventilation rate, the radon concentration  in
the crawl  space,  and the barometric  pressure  in  week 24 (June 11th -  17th)
of 1989. This  week was characterized  by absence  of wind, high temperatures
during day time and cool nights.  In  the figure one notices that during the
night time the  living room is at  a lower  pressure  than the crawl space and
the  crawl space  is  at  lower  pressure  than  the  outside   air.  This   is
attributed to   the  temperature  difference between  living room  and  outside
air. In  the absence of wind or  forced  ventilation  this  stack effect   is
considered the  driving force  for the  pressure difference  between crawl space
and living room. The ventilation rate  is calculated from  conservation of air
mass and the flow calculated from  the pressure differences  and the measured
leak of the crawl space ventilation shafts. From the figure one notices that
both ventilation rate  and  radon  concentration  in  the  crawl space  are   in
phase with each other and the pressure differences over the wall.

     Fig. 2 shows  the  time dependence of a number of parameters during week
31, 1989.  In this  week there  has  been almost  continuously  wind  from the
southwest, resulting in over  pressure at the  south  and west walls and under
pressure in the living room, all with respect to the crawl space. Heavy rain
fall occurs on  Sunday afternoon and evening as showers. No obvious effect  on
the radon  concentration is observed.  On Tuesday  and Vednesday  evening the
open fire  was  lit  causing   a  sharp  increase  in  the pressure  difference
between  living  room  and  crawl  space.   Surprisingly  no  effect  of  this
increased under  pressure is  observed on either radon  concentrations and/or
pressure difference of the  crawl  space with  the  outside world. The radon

concentration  in  the  crawl space shows the diurnal cycle similar to  the  one
in  fig.  1.  Also in the pressure differences such a cycle is noticeable.  The
difference  with fig.  1, however,  is  that  the maxima in  the  crawl space  radon
concentration  are  more  pronounced  and  that  for  Tuesday,   Wednesday  and
Thursday  they  coincide with  the  minima in the ventilation rate. However,they
are more  pronounced than  the variations in  the ventilation  rate. The  overall
trend  of  the   crawl  space  radon concentration  is  decreasing  towards  the
middle of the  week and increasing afterwards. For  the radon concentration in
the living  room no diurnal  cycle is observed;  the concentration diminishes
somewhat  on Wednesday and  Thursday.
                       4. ANALYSIS AND INTERPRETATION
     In the extremely simplified model for  the radon concentration  in a  room
which  is  ventilated with  radon-free  air  one  considers  constant  static
sources and  a constant  ventilation rate. For  the equilibrium situation one
may write  for the radon concentration C in a room with ventilation rate X
(A » X  ) and volume V:

     C = 1- I E. A ,                                                  (4.1)
         XV i  L  i

where the  summation  is  over all radon exhaling surfaces indicated by i  with
exhalation rate  E.  and  surface area  A .  Here  it is assumed that the radon

concentration in the incoming ventilation air may be neglected.

     For the  crawl  space with volume V=26  m3,  a ventilation rate  X=0.5 h"1
(based on  the average value in  fig.2e) and  E  and A  taken from table 1 one
obtains a  value of C=21 and 17  Bq.nT3  for a dry soil  (d=46 cm) and a barely
inundated  (d=0  cm)  crawl  space,  respectively.  In  the  latter  case  the
contribution from the soil has  been neglected.  In the period 1980-1987  time
averaged radon concentrations were measured  (4); the average value was 50±20
Bq.nT .  The large uncertainty in this value is  due to large fluctuations in
the radon  concentration, presumably  due to variations  in the ground-water
level (4).

     From  the  numbers   one   may   conclude   that  based  on  the  exhalation
measurements  the  radon  concentration in the  crawl space is  two  to three
times higher than expected on diffusion from the materials. This discrepancy
indicates another source of radon which,  for this crawl space, is equally or
even more important.

     A  similar  simplified  model  yields  for  the  living  room with
constant source  terms  and a constant ventilation  with a constant fraction,
a, of air from the crawl space:

     Cl, - « Cc.  * "-I  C.  * W f EiV                                  <4'2>

Substituting C  =22 Bq.m'3 (ref. 4). C  =50 Bq.nf3, C  =3 Bq.nf3, X  =0.5 and
              lr                      CB             0    ^       It
E  and A  from  table  1 one obtains o=0.4. This  value  would correspond  to a
crawl  space ventilation  rate  of  at  least  3 h"1  and hence  a discrepancy
between  the  radon contribution  calculated  from  static  sources  and  the
observed value for the crawl space of a factor of 15.

     A  dwelling   is  considered  with   N  rooms,  each   having  a  radon
concentration  C (t)  and a  volume V ; outdoor  parameters  are  indicated by

N+l. The radon concentration in room i is determined  by the total strength
of  radon  sources  in  the  room,  S. ,  and  the   transport  of  radon  by  air

currents.   If  q    denotes   the  air  current  from  room  k  into  room  i  the

following mass-balance equation for room  i can be written:


                                                               V     <4'3>
Defining X. = «-  Z  q   as the ventilation rate of room i and f  = » — ,
          1    i k=i                                            ik   i
eq. (4.3) transforms into:
Eq.  (4.4)  can be  written for  each  of the  N rooms,  giving a system  of N
coupled linear non-homogeneous  differential  equations. The set of equations
can be described  more comprehensively using vector  and matrix notation. In
principle  the  equations can be solved analytically,  in practice, however,
instabilities occur  and a numerical  solution is  preferred.  The algorithms
have been implemented in a computer programme (CARACO).

     Essential  for  the calculation are  the input parameters:  the currents
q   and the source strengths S . The currents q   have been deduced from the

pressure  difference   between   room   i   and  k,  Ap. =p -p.,   and  the  air

transparency of a barrier between the compartments i and k, T  .

      l          1  Pa
The values of T   and n are  taken  from  a previous study on the air leaks in
this test dwelling (7).  For the source term in  the  crawl  space a pressure
driven flow  term (S  ) was  added  to the static  sources (S  ) as discussed
                    pdf                                      at
in sect. A.I. This pressure  driven flow stems  from the pressure difference
between the outside and the crawl space and causes a small, Darcy-type flow,
through  the  soil  (and  the  walls  of  the  crawl  space). This  flow becomes
important due  to  the relatively  high  concentration  of radon  in  soil.  The
relative intensity of this term depends on the permeability of  the soil  (and
walls),  the  radon concentration  in the  soil  and  the  length  of  the crawl
space perimeter. As starting values were taken: for  the permeability  5.10"
ra . the value for fine sand,  and for the radon concentration in the soil gas
10  Bq.nf .

     Fig. 3 shows  the radon concentration in the  crawl  space measured during
week 24,  1989.  The  dashed curve  is  the calculated  concentration  based on
measured leakage parameter,  measured S  and  a  pressure  driven flow (PDF)
term with  a strength  40  times  the estimated  value of S     based  on the
starting parameters.  Without  the  adjustment of  S     the magnitude  of the
concentration is a factor of three too low  and the oscillations are out of
phase.  For the  radon concentration in  the living room  (not measured during
that week) a value of about 75 Bq.nf3  was assumed  at the rare moments  that
the flow was from  the living room  to the crawl space.  The results indicate
that the missing radon concentration in the crawl space may be  accounted for
by a values of  the S   which  is almost two  orders of magnitude larger  than
estimated from rather arbitrary starting parameters.

     Fig. 4  shows  radon  concentrations in  the living  room  (top)  and crawl
space  (bottom)  measured during week 31,  1989. The  dashed curves represent
the concentrations calculated  with a two-room model. First  the crawl space
concentration was  optimized  by starting from values for parameters as found
for week 24, 1989. The result  in  the bottom  part  of fig. 4 is  obtained  with
a seven  times  larger value  than  the  measured value of S  and a value for
S    equal  to  the  value  estimated from  the original  starting parameters.
With these  values  of  S  and  S     the  average concentration  in  the crawl
                        at     pdf
space is reproduced; in the living  room the concentration  is a  factor of two

too lov. The larger static source term is necessary to reproduce the dips in
the cravl  space  radon concentration at  times  of  increased ventilation; the
smaller S    to obtain the correct average values.

     At  first  sight  it  seems  rather  strange  that  the  size of  the   terms
changes in  time and one may  argue that  the infiltration of radon is still a
large puzzle. Without  trying to reduce  such a conclusion one could think of
a possible explanation. During week 31 the ground-water level reached one of
its lowest values of  the year (d=58 cm).  The exhalation of the soil in this
period, measured at one location  increases sharply (d=58 cm in table 1). The
increase  of  the exhalation  rate  is  larger  than  expected on  basis  of
difference of water level at d=48 and 46 cm.

     As a hypothesis  we propose a role  to rainfall to qualitatively account
for both the increase  in exhalation rate and the  reduction of the PDF-term.
The  concentrated precipitation did hardly change  the  ground-water   level
until the latter part  of  the week. This means that the percolating rain has
been forming a type of piston on  the soil closing of pores in air with  water
thereby  reducing the  porosity  and  permeability  of   the  soil.  The  slowly
moving  piston  blocks   the diffusion of  radon  to  the  outdoor soil  and thus
causes an increase of  the radon in  the  soil gas below the piston.  This will
enhance the exhalation of  the soil in the  crawl space,  especially near the
perimeters (the measurement of  the exhalation was made close  to the centre).
Moreover the reduction of permeability will reduce the PDF-term.  It should
be  stressed again  that  this  is only  a  hypothesis, which qualitatively
accounts for these rather large changes.
     Radon concentrations  in  crawl spaces and living  rooms  are larger than
calculated from  measured static source  strengths.  Time evolution  of radon
concentrations measured  simultaneously with  pressure  differences indicates
that pressure  driven  flow may account  for the missing  contribution  to the
concentration  as  well  in  the  crawl  space as  the  living room. This  is the
result of the  preliminary  analysis of two weeks in  the  summer of 1989 with
different weather characteristics.

     From this analysis we conclude that the effective strength of diffusive
and pressure  driven terms are not constant  in  time. It  is  proposed  as  a
hypothesis that  changes  in water  content  of the soil due  to precipitation
may temporarily increase the  radon concentration below the  wetted layer and
that the increased water content may reduce the pressure driven flow.

     Although we are still at  the  beginning  of the analysis of the data and
our hypothesis may  be  replaced by others we would be  tempted to state that

understanding radon infiltration into dwellings is not  so  much a problem of
transport in the dwelling  than it  is a problem of radon  transport of soil.
In our present  opinion  the way to better understanding is  a combination of
analysing measurements, as described in this  paper,  over  a longer period of
time to  identify possible  important parameters and  investigations of these
parameters under controlled conditions in the laboratory.

     This  work is  part  of  the  research  programme  of   the  Environmental
Radioactivity Research Group  at the KVI and has been  financed by the Dutch
Government,  as  part of  the National  Research Programme  RENA,  and  by  the
Commission of  the  European Communities, as part of  the 1985-1989 Radiation
Protection Programme.


1. Arvela, H., Voutilainen, A., Makelainen, I., Castren, 0. and Winquist, K.
   Comparison   of   predicted   and  measured   variations   of  indoor  radon
   concentration. Radiat. Prot. Dosim. 24: 231-235,  1988.

2. Put,  L.W., de Meijer, R.J.  and  Hogeweg,  B. Survey of radon concentration
   in Dutch dwellings. Sci. Total Environ. 45: 441-448, 1985.

3. Put,  L.W.,  de  Meijer,  R.J.  and  Bosniakovic,  B.F.M.   Radon   in  the
   Netherlands. In: Indoor RADON II. Proc. 2   APCA. Int. Spec. Conf. Cherry
   Hill, APCA,  Pittsburgh, U.S.A., pp. 107-119, 1987.

4. Put,  L.W., de Meijer, R.J.  Variation of time-averaged indoor and  outdoor
   radon concentrations with  time,  location  and  sampling height.   Radiat.
   Prot. Dosim. 24: 317-320, 1988.

5. Ackers,  J.G. Direct measurements  of  radon  exhalation  from surfaces.
   Radiat. Prot. Dosim. 7:  199-201,  1984.

6. Aldenkamp,  F.J.,  de  Meijer,   R.J.,  Put,  L.W.  and  Stoop,   P.   Removal
   processes  for  charged  radon decay  products, submitted  to Radiat. Prot.

7. Phaff,  J.C.,  de  Gids, W.F.   and  Knoll,  B.   Ventilatie  van gebouwen.
   Metingen  van de  luchtlekken en  voorspelling  van de  ventilatie  van een
   woning  in Roden, IMG-TNO Technical Report  C535, Sept.  1983, TNO-Delft.

— i— —i— —i— —i 	
	 1 	 1 	 1 	 1 	 i 	 i 	
Fig. 1 Pressure  differences  P -P  (Pa)  with  P   being  the  pressure at  the
                              i  as            i
west (a), south (b) and east (c) outside wall and in the living room (d) and
with P   being  the  pressure  in the crawl space. The  calculated  ventilation

rate (If1) of the cravl space  and  the  measured  radon  concentration (Bq.nf3)
and the  barometric  pressure  (hPa)  are  presented in parts (e), (f)  and  (g),
respectively. All data were collected in week 24, 1989 (June 11-17).

Fig. 2 Pressure  differences PI-PCI (Pa) with PA being the  pressure at  the
west (a),  south  (b), and east (c) outside  wall  and the living room (d),  the
calculated ventilation rate (h~ )  of  the crawl  space, part (e) the measured
radon  concentrations(Bq.m~  )  in  living   room  (f)  and   crawl  space   (g),
precipitation  rate  (mm.h~ ) in  part  (h),  and  barometric  pressure  (hPa)  in
the bottom part.  All data are collected in week 31,  1989  (July  30 - August

  Fig. 3 Radon concentration in the crawl space of  the  test dwelling  in Roden
  during  week  24,   1989   (June   11-17).  The  solid   curve   represents   the
  measurement,  the  dashed curve the  calculated  concentration  with a  dynamic
  model. Multiplication factors for the  measured  S   and  estimated  S    are 1
                                                 at                pdf
                      .    *     ^V\
 Fig. 4 Radon concentration in  the  crawl  space and living  room  of  the test
 dwelling in Roden during week 31,  1989  (July 30 - August 5). (See fig. 3).
  Multiplication factors for the measured S   and estimated S    are 7
                                           St                pdf
 and 1, respectively.

          Session VI:
Radon in the Natural Environment


               by:   V.C.  Rogers
                    K.K.  Nielson
                    Rogers and Associates  Engineering  Corporation
                    Salt  Lake City,  Utah  84110-0330

     Field measurements  were used  to benchmark  a  simple  new  predictive
correlation between soil  gas permeability and soil  grain size, moisture, and
porosity.   The  correlation  was  incorporated  with  a previous  diffusion
correlation into the new RAETRAD  code, that calculates radon generation and
two-dimensional transport in soils, and radon entry into structures.  RAETRAD
generalizes  the one-dimensional   RAETRAN model,  combining advective  and
diffusive  radon transport with  radon emanation,  decay,  absorption,  and
adsorption.  RAETRAD calculations  suggest  0.03 percent radon-entry efficiency
for slab-on-grade homes  on low-permeability soils «10 ° cm'), increasing to
0.1 percent for sandy soils.   Soil  or  fill properties  in  the first few feet
dominate  radon entry efficiency and limiting  radium concentrations  for
prescribed  indoor  radon  levels.    For  indoor  radon  concentrations  of  2
pCi/liter. sandy soils may contain only 2-3 pCi/g radium compared to 10-20
pCi/g for more clayey 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.


     Radon generation and transport in soils and its subsequent entry into
dwellings  is  a  complex process  requiring  characterization  of the  soil
conditions, meteorological conditions, and  the  structure.   Radon emanates
from radium-bearing minerals  into the soil pore space,  followed by diffusive
and advective  transport in  both  liquid  and gas  phases  into the dwelling.
entering via cracks,  sumps, porous building materials, and other routes.  The
RAECOM (Radon Attenuation Effectiveness and £over Optimization with Moisture)
(1) multlregion. one-dimension radon generation  and  transport code has been
used widely to  predict radon migration through porous media. RAETRAN (RAdon
Emanation and TRANsport)  (2) provides similar capabilities, but  also includes
Idvective transport mechanisms.  These codes  are easy to use and require very
little  input  data:  however,  because  they are  one-dimensional,  they have
limited application for radon entry into structures.

     The  mathematics  of  the  RAETRAN  code  have  been  extended  to  two
dimensions.  In addition, the pressure-driven flow equation now is solved in
the  code  instead of  externally  as required with  RAETRAN.   The  resulting
position-dependent velocities have corresponding boundary conditions to those
used for  the radon generation and  transport  calculations.   The  resulting
code,  called RAETRAD (RAdon  Emanation  and  TRAnsport into Dwellings)  (3).
retains the  general  simplicity of  operation and minimal inpUt  requirements
as  the earlier  RAECQM  and  RAETRAN  codes.    However,  it provides  a  more
detailed description of radon movement through porous materials such  as soil
and concrete and subsequent radon entry into  structures coupled to the soils.

     Key factors in the simplicity of the  RAETRAD  input  data  are the simple
correlations   for   predicting  gas  permeabilities  and  radon   diffusion
cpefficients for the porous materials.  These correlations and their  use are
discussed  in the next  section.   After that  the RAETRAD code is  briefly
described, and finally is applied to typical  Florida soils  and structures to
obtain radon entry efficiency factors and  radon  entry rates into dwellings.
and to estimate example maximum soil  radium concentrations   for  foundation
fill materials.


     Radon migration through soils  and entry into  dwellings depend strongly
on values  of radon  diffusion coefficients  and  air  permeabilities for  the
soils  and  for  the  applicable  house  construction materials.     Simple
correlations for the radon diffusion coefficient  have been developed and have
been widely  used (4-5).   The  diffusion  coefficient correlation that  is
incorporated in RAETRAD is (6-7):

           3p(l+p)d       7p(l+p)d
     D  -  	a-exp [	a m - 7m5].                            (1)
           8(2+ndg)          2+dg


     D     -  pore average radon diffusion  coefficient (cmV1)
     p     -  soil porosity
    dg    -  geometric mean particle diameter  (pm)
    m     -  fraction of moisture saturation.

Predictive correlations for gas permeability have been proposed previously
(7-8).   A recently improved permeability correlation for shallow  soils  (9)
was incorporated into RAETRAD:


        -  Pi—1  d«'3.                                               (2)

     K     -  air permeability in porous material  (cm2)
    da    -  arithmetic average particle diameter (cm).

     Equations  (1)  and (2)  reveal  that  soil  gas permeability  and radon

diffusion coefficients both can be estimated from soil moisture, porosity,
and particle diameter averages d_ and da.   In  turn,  the particle diameter
averages can be estimated  from standard soil classifications such as the 1Z
categories  used  by  the  U.S.  Soil  Conservation  Service  (SCS)  (10).
Furthermore, the appropriate soil  moisture  near a dwelling can be estimated
from the soil classification and soil matric potential (11).

     As an example of the above methodology, measurements were made of the
in-situ moisture, gas permeability,  porosity,  and  soil  particle sizes for
several soils in Florida.   Soil  samples were obtained  at depths of 60 t9 75
cm from several  locations around the state.   The data for the soils are given
in Table 1.  The  soil gas  permeabilities were  estimated from Equation (Z)
using   the   field  soil   data.      The   resulting  correlation-predicted
permeabilities and measured field  permeabilities  are also given in Table 1.
In general,  the agreement  is  within  the  experimental  uncertainties.  Two-
thirds of the predictions were within a factor of 2 of  the field-measured gas

     From the particle size and moisture information in Table  1. soil matric
potentials were estimated using the methodology, described in4Reference 11.
The estimated matric potentials ranged from 1x10*  Pa to 3.4xlOn Pa.  A matnc
potential of  5xl04  Pa was selected  as a  reasonably  conservative dry-side
average  for conditions  in  the locations  sampled  in  Florida.   Using the
5x10* Pa matric potential and soil  particle size distribution parameters from
the soil samples,  soil moisture, permeabilities,  and diffusion coefficients
were  estimated  for the broader  range of  soils  defined by  the  U.S.  Soil
Conservation Service classifications  (10).   These data are given in Table Z.
and  were used  in the  example  radon migration  and   house  entry analyses
performed by RAETRAD.
                          WITH CALCULATED VALUES
E Orlando
N Orlando
SH Orlando
SH Orlando
NE Tampa
S Tampa
Permeabl 1 1 ty
Ratio of
Cal c.Perm./


SCS Soil
Loamy Sand
Sandy Loam
Sandy Clay Loam
Sandy Clay
Clay Loam
Silt Loam
Silty Clay Loam
Silty Clay
Radon Diffusion
Soil Gas
*At 1.6 g/cm3 bulk dry density:  0.407  porosity.
Estimated from Equation 1.
'Estimated from Equation 2.

                             THE RAETRAD CODE

     RAETRAD  solves  the-two  dimensional   radon  balance and  air pressure
balance equations in  cylindrical geometry.  The two-dimensional  rate balance
equation for radon in the gas component of the soil pore space is given by:
        /d2Ca    1  dCa   dzCa\
        I— +	+ —-) - xc
        \drz     r  dr    dzz /

                 /dP dCa   dP dCa\
                 I	+	1
                 \dr dr    dz dz /

mX              dC=
            (l-m)kd     a     tit


    D.    -  radon diffusion coefficient in air,  including tortuosity
    C.    -  radon concentration in the air-filled pore space




        radial  distance from center of house
        vertical  depth from ground surface
        radon decay constant
        air-surface adsorption coefficient for radon
        bulk dry density
        fraction of moisture saturation
        pore gas permeability
        total porosity
        pore gas pressure
        radium concentration in the solid matrix
        component of emanation coefficient  that  is a direct pore  air
        source of radon                                   .      .
        equilibrium distribution  coefficient  for radium  in  sond-to-
        transfer factor of radon from pore water to pore  air
     The T   transfer factor  from  pore water to pore air is  obtained  from
combining Equation (3) with a similar rate balance equation for radon in pore
water (6).  The  derivatives  of the atmospheric and soil air  pressures are
obtained by solving the following equation using the same approach as for the
radon transport equation:
          1  dP   d'P      dP
K [— r + -- +— y]  --
     The boundary  conditions  for  Equation (4) are the  indoor  air  pressure
applied to  the  inside surface of the  dwelling floor, and the  outdoor air
pressure (typically averaging zero)  applied to the outdoor soil surface.  If
the dwelling  is  at a  negative pressure compared to the  outdoors,  then air
movement proceeds from the outdoor soil  surface downward  through the soil and
then inward and  upward towards the  structure as shown  in  Figure  1.   Radon
entry into the slab-on-grade  dwelling  in  Figure  1 is  assumed to be through
a  perimeter  crack,  such  as  may occur  between  the slab   and  foundation
                            House Area- 141 mz(1SOOsq ft)
                              Indoor Pressure. -2 4 Pa
                              Perimeter Crack Area: 4%
          Fiqure  1.   Flow  lines  and  peripheral  air  entry locations
                      for a structure on a 61-cm deep  foundation  in
                      sandy soil  (K - 1.7 x  10"' cm').

     After  the pressure  field is  determined.  RAETRAO  solves  the radon
generation  and transport  equations to  obtain  values  for  the following
parameters (3):

     1.   Radon concentration in soil air pores as a function of position.

     2.   Average   radon   concentration   under  the  dwelling   slab  (if

     3.   Diffusive, advective. and total surface radon  fluxes.

     4.   Radon entry rates  through  dwelling  floors,  walls,  and cracks in
          contact with the soil.

     5.   Air entry rates from the soil.

     6.   Radon entry efficiency factors  for the dwelling-soil  system.

The radon entry  efficiency factor is defined  as  the  average indoor radon
concentration  divided   by  the  area-weighted   average   sub-slab  radon
concentration in the soil pores.

                          APPLICATION OF RAETRAD

     The RAETRAO code was applied  to the  soils  and soil conditions given in
Table 2.  A  slab-on-grade  structure was  coupled to the soils  as shown in
Figure  1,  and it was  assumed that  radon  entered the dwelling through   a
perimeter crack  between  the  10-cm  thick concrete  slab  and a  60-cm deep
foundation footing.   The dwelling is assumed to be at  a  -2.4  Pa pressure
compared to the atmosphere.  The radon emanation  coefficient  of  the  soil is
0.25.   Other parameters used in the analyses are shown in  Figure 1.

     Radon entry efficiency factors  computed by  RAETRAD for the  dwelling on
each of the SCS soils are shown in Figure 2.  They increase with  increasing
soil permeability mainly for coarse-grained soils.    The  entry  efficiency
factor becomes less dependent on permeability for permeabilities less than
about  10   cm. because  diffusion processes  dominate  the  radon  entry rate
into the dwelling for the low-permeability soils.  For these examples,  the
entry  efficiency varies  from  about  0.025 to 0.1  percent.   The   radon entry
efficiency factors are determined  from the radon entry rates, dwelling air
volumes, and air exchange rates. The average air exchange rate can be either
input  directly or estimated from procedures published by the American  Society
of Heating.  Refrigerating and Air-Conditioning  Engineers (12).   The  present
example used an air exchange rate  of 1  hr  .

          Soil Moistures: 5x104 Pa matric potential
          Soil Densities: 1.6 g/cm3
                                              Sandy Loam

                                         Sandy Clay Loam
•  sit0"*   Loam
           10'1U       10'9        10'8        10

                 Soil Gas  Permeability  (cm*)
          Figure 2.  Radon entry efficiencies  computed  by  RAETRAD
                     for a slab-on-grade structure  (Figure 1)  on
                     uniform soils defined  in  Table 2.
     Maximum soil  radium  concentrations can also  be determined from  the
example analyses by assuming a maximum indoor radon concentration guideline.
A guideline of  2 pCi/liter  applied to the example calculations  gives  the
maximum soil radium  concentrations shown in Figure  3.   The maximum  soil
radium increases with  decreasing  soil permeability.  Calculations were  also
made of the maximum soil radium concentrations for a layer of foundation  fill
material placed over the natural  soil.   Fill  material  properties generally
obscured effects  from  the  underlying soils  when the  fill  layer  thickness
exceeded approximately 1 m.   For thinner fill  layers, high or  low  radium
contents in the underlying soil affected the acceptable radium content of the
fill material.   As shown in Figure 3. the maximum soil radium  for the  fill
material also becomes  insensitive  to the natural  soJl  conditions  for  low-
permeability fill materials  (less than about 10   cnr  permeability).  Sandy
soils permitted  only 2-3 pCi/g radium  before exceeding  the 2 pCi/liter indoor
radon concentration, while finer-grained soils could have 10-20 pCi/g due to
their lower permeabilities  and diffusion coefficients  (Figure 3).


Silt, Silty Clay, &
Silty Clay Loam
                               Clay Loam


                                        Sandy Clay

                                                 Clay Loam

                                                Sandy Loam

        Uniform Soil

        30-cm Layer on 1 pCi/g Silt Loam

        30-cm Layer on 24 pCi/g Silt Loam
1Q-10       1Q-9        10'°        10

     Soil Gas  Permeability  (cm*)
            Figure 3.  Maximum soil  radium concentrations to maintain
                       2 pCi/liter radon in a slab-on-grade structure
                       (Figure 1) on SCS soils that are uniform (solid
                       line) or used for a 30 cm fill  layer over a
                       silt-loam base soil (broken  lines).   Soil
                       properties are defined in Table 2.
       Radon entry efficiency factors and maximum soil  radium concentrations
  also vary  according  to  the  perimeter crack  width.    Figure  4  shows  the
  variation of the entry efficiency factor  with  the perimeter crack area for
  a  sandy soil.   The perimeter crack  area is expressed as a percentage of the
  total slab area, and also may be used to approximate the effects of perimeter
  utility penetrations  through  the  slab.

 "g 0.02
                         Perimeter Crack  Area (%)
          Figure 4.  Variation of radon entry efficiencies  with  the
                     size of perimeter crack for  the  slab-on-grade
                     structure (Figure 1)  on SCS  sandy  soil.
     As a benchmark for RAETRAD,  an analysis was performed for a house-soil
system in Florida  for which some field data are  available.  The indoor radon
concentration for  a 203  m  slab-on-grade  dwelling was measured  to  average
about 10 pCi/liter.
     The radium concentration  in  the top 30 cm of subslab soil is about 0.9
pCi/g. the soil  moisture is about 15 percent of saturation, and the measured
permeability  is  8 x  10    cm*.    A subslab  radon  concentration of  4.200
pCi/liter indicates the presence of a deeper soil  layer with elevated radium.
This is  represented by a  5  pCi/g soil  radium layer beneath  the  top  31  cm
layer characterized above.  A radon emanation coefficient of 0.25  is also
used  in  the  analysis.    The  RAETRAD  calculation  gives  a   subslab  radon
concentration of  4.000  pCi/liter and  an  indoor  radon  concentration  of  7
pCi/liter, for a house pressure differential of 1.0 Pa.  The estimated indoor
concentration is within  30 percent of  the  measured  value  of 10 pCi/liter.
     This work was  supported  in part by  U.S.  Department of Energy  grant
DE-FG02-88ER60664 and in part  under subcontract  IAG-RWFL933783  to  the U.S.
Environmental  Protection Agency.


1.   Rogers.  V.C..  and  Nielsen.  K.K..   "Radon  Attenuation  Handbook  for
     Uranium Mill Tailings  Cover Design," U.S. Nuclear Regulatory Commission
     report NUREG/CR-3533. April 1984.

2.   Nielsen,  K.K..  and Rogers,  V.C.,  "Radon  Generation.  Absorption  and
     Transport in  Porous Media -- The RAETRAN Model." EOS. 70. 497 (1989).

3.   Nielson, K.K.,  and Rogers, V.C.. "A Mathematical Description of  Radon
     Generation. Transport and Entry Into Structures," in preparation.

4.   Nielson.  K.K..  Rogers.  V.C..  and Gee.  G.W..  Soil  Science Society  of
     America Journal 52, 898 (1988).

5.   U.S. Nuclear Regulatory Commission  Regulatory  Guide  2.64. "Calculation
     of  Radon  Flux Attenuation by  Earthen  Uranium Mill  Tailings Covers,"
     June 1989.

6.   Rogers, V.C.,  Nielson,  K.K.,  and  Merrell, G.B..  "Radon  Generation.
     Adsorption. Absorption,  and Transport in Porous Media."  U.S. Department
     of Energy report DOE/ER/60664-1, May 1989.

7.   Rogers. V.C..  and Nielson. K.K..  "Radon Emanation  and  Transport  in
     Porous Media," in Proceedings:  The 1988 Symposium on Radon and  Radon
     Reduction Technology.  Volume 1. EPA-600/9-89/006a (NTIS PB89-167480).
     March 1989.

8.   Shepherd.   R.G..  "Correlations  of  Permeability  and  Grain  Size."
     Groundwater 27. 633-638. 1989.

9.   Rogers. V.C..  and Nielson.  K.K..   "Predictive Correlations for Air
     Permeabilities and Radon  Diffusion  Coefficients  in  Porous  Media,"  in

10.  Dunn.  I.S..  Anderson,   L.R.,  and  Kiefer.  F.W..   Fundamentals   of
     Geotechnical Analysis. New York: Wiley & Sons. 1980.

11.  Nielson,  K.K.,  and Rogers,  V.C.,  "Moisture,  Radon  Diffusion  and Air
     Permeability  Characteristics   of   SCS  Soil   Classifications."   in

12.  Fundamentals Handbook. "Ventilation and Infiltration." American Society
     of Heating.  Refrigerating  and  Air-Conditioning Engineers  (1981).

                                   L.T. Gregg
                        Atlanta Testing & Engineering, Inc.
                            11420 Johns Creek Parkway
                             Duluth, Georgia 30136

                                   Gene Coker
                  United States Environmental Protection Agency
                            345 Courtland Street, N.E.
                             Atlanta, Georgia  30365

Through a combination of geologic models and  field measurements, each of the  four
geologic  provinces of  Georgia can be  characterized  for radon concentration.   The
combinations of bedrock lithology and soil  characteristics most likely to exhibit higher
radon concentration in Georgia are granites, granodiorites, granite gneisses, pegmatites,
mylonites, carbonaceous shales, phosphates, and monazite/heavy mineral placers, coupled
with high to medium permeability soils such as gravels, sands, and uniformly-graded silts
and sandy silts.  Saprolite and surficial soil may either enhance or impede radon migration,
as may hydrogeologic characteristics and rock structures such as faults and joint/fractures.


    The four radioisotopes considered in this paper are uranium, thorium, radium, and
radon. The principal uranium isotope is U238, which decays to Ra226. Radium-226 decays
to Rn222, which decays to  polonium, bismuth and finally lead.  The principal thorium
isotope is Th230, which decays to Ra228, then  to Rn220, polonium, bismuth and finally lead.
Because of its half-life  of 3.8  days, Rn222 is  of much more concern and interest from an
environmental standpoint than Rn220 (56 seconds  half-life).

    The basic properties of radon are: it is  gaseous and therefore highly mobile, it is an
alpha and gamma emitter so its presence can be measured (usually with an  alpha-particle
measuring device), it is a daughter or decay product of radium, it is moderately soluble in
water and  its solubility decreases with increasing temperature, and it does  not readily
combine chemically with other elements.  These properties form the framework for our
understanding  of  radon generation,  occurrence and migration  in  rocks,  soils,  and

    In considering the geologic setting of Georgia and the Southern Appalachians, the
major structural and lithologic trends run from northeast to southwest. The principal rock
types in  the Cumberland Plateau  and the Valley and Ridge Provinces are limestones,
dolomites, shales, and sandstones, in the Blue Ridge and Piedmont Provinces igneous and
metamorphic rocks such as  granites,  gneisses,  and schists,  and in the Coastal Plain
limestones, sandstones, phosphates, and unconsolidated sediments.

                          URANIUM GEOCHEMISTRY

    Uranium abundance typically ranges from 2 to 5 parts per million (ppm) in granites
to as high as 9 ppm in nepheline syenites, from 0.5 to 2 ppm in the andesites and mafic
rocks and much less in the ultramafic rocks, to as high as 34 ppm in the Marcellus Shale
(an  analog of the Chattanooga Shale) and 120 to  140 ppm in the phosphates  and
phosphorites.  Thorium shows similar ranges, as high as SO ppm in granites.

    Why is uranium so widely distributed?  Uranium is polyvalent, with the three principal
ions being +4, +5, and +6.  It has a large atomic radius of 0.8 to 0.97 angstroms;  it is
highly active chemically and forms strong complexes with many ligand species; and the
hexavalent compounds are more soluble than the tetravalent compounds, the latter being
isomorphic with Ca, Th, Zr, W, Mo, and so forth.  This suite of properties results in a
complex geochemistry.

    During magmatic differentiation, uranium  does not seem to form  separate mineral
precipitates. There is isomorphic substitution in  some rock-forming minerals.  Uranium
tends to  concentrate  in late  stage crystallization in  acidic rocks  and  minerals such as
granites and felsic volcanics, and through hydrothermal action in pegmatite dikes and veins.
This is primarily due to its large atomic radius and to its affinity for late forming members
in the reaction sequence such as quartz,  potassium feldspar and muscovite. In igneous
rocks, the petrofabrics play an important  role in  uranium concentration (or  enrichment).
The petrofabrics provide microstructural control of uranium movement and deposition such
as coatings around mineral grains, and within microfractures  and along crystal cleavage
planes (which provide  a type of  porosity  for the uranium bearing waters).  In felsic
volcanics, both acidic  and alkalic, uranium may be highly dispersed  and  thus readily
leachable.  Some concentration of  uranium has been observed in ashes and tuffs and in
cross-cutting dikes and veins.

    In sedimentary rocks, uranium concentration is more dependent upon the geochemical
cycle, including  such  mechanisms as oxidation-reduction, absorption,  adsorption  by
substrates such as iron  oxide, silica, and organic material, formation of fluoride, sulfate,
phosphate, carbonate, and organic  complexes, the relative mobility of the hexavalent and
tetravalent ions, and the weathering and solution of the uranium source and the transport,
precipitation and deposition of the weathered uranium.  Figure 1 shows the processes

 controlling the occurrence of syngenetic uranium in the sedimentary environment.  Through
 various mechanisms - oxidation, reduction, adsorption, ionic substitution, and evaporation -
 uranium concentrates  in different sedimentary rocks.   Figure  2 shows  the  processes
 controlling epigenetic uranium deposition. Concentration mechanisms here are primarily
 adsorption, precipitation, evaporation, and changes in redox potential and pH.

    Figure  3 shows an  idealized sequence of the  processes  that  control  uranium
 concentration in metamorphic rocks, where recrystallization may cause a grain size increase,
 porosity reduction, and liquid-gas expulsion, resulting in uranium movement into fractures,
 shear zones,  and lower pressure zones and concentration in them.

    To summarize, the principal factors in uranium mobility and concentration are the long
 half-life, ionic size, polyvalence, mineralogy and uranium content of the source, petrofabric,
 and geochemical conditions such as the amount and rate  of circulating water, climatic
 factors, pH and Eh,  the presence or absence of complexing agents, and the  presence or
 absence of sorptive materials.

                           RADIONUCLIDE MOBILITY

    Uranium is  more mobile than  its daughter product radium.  Generally, uranium may
 be considered much more mobile in oxidizing environments than in reducing environments,
 whereas radium is most mobile in chloride-rich reducing environments.  Radium tends to
 behave chemically somewhat similar  to the alkaline earths  such as calcium  and may
 complex with sulfate, carbonate, or chloride.

    Since radon is a gas, the oxidation-reduction environment is  immaterial.  Radon has
 two principal components to its movement:  diffusion, which is generally thought to be a
 minor component (probably an  average of about 1 meter from its radium parent source),
 and convection, which  is the major  component.   Radon can  move  many meters by
 convection, but it has to be carried in some type of "geo-gas" that can be either a mixture
 of helium, nitrogen, methane, CO2, and so forth, or groundwater, or both.

    Once radon is liberated (or "emanates") from its radium  parent source, within rock,
 soil, or water,  it will tend to  migrate  by  diffusion  and convection  to zones  of lower
 pressure, for example vertically toward the surface.  A  high permeability soil will allow
 radon to more readily permeate upward to the surface than will  soils high in  the clay
 minerals that have resultant high porosity  but  low permeability.  Figure 4 shows  an
 idealized cross-section of the Piedmont from surface soils down through massive saprolite,
 structured saprolite,  partially weathered  rock and fresh bedrock.  Circulating meteoric
waters near the surface should more actively dissolve and transport leachable uranium than
 at greater depths, where these waters may  concentrate iron, silica, and clay substrates
which would tend to act as scavengers to concentrate uranium.  Relict structures within the

structured saprolite such as joints, fractures, faults, mineral veins, and foliation planes
would provide pathways for percolating waters and deposition of substrates and organic
material, resulting in sites for precipitation of radionuclide complexes and concentration
of uranium, thorium, and radium. These relict structures should also act as conduits for
radon migration, through diffusion and  convection, toward the surface.

                       NURE STREAM SEDIMENT DATA

    Under the NURE Program, uranium and thorium were measured in stream sediments.
The mean in Georgia stream sediments for uranium was 11.62 ppm and 56.62 ppm for
thorium. The average upper continental crustal abundances are 2.5 ppm and 10 ppm for
uranium and thorium respectively. Thus,  the ratio of the mean to  the crustal abundance
in Georgia is about 4.5 for uranium and 5.5 for thorium.

    Figure 5 shows the NURE stream sediment data for uranium (1). The outlined zones
include the  two highest concentrations  of uranium, ranging from  6.1 ppm to  426 ppm.
Note the  trends  from northeast  to  southwest.  The NURE  thorium data  in stream
sediments is almost a  perfect overlay of the uranium data, which is not too surprising if
one considers the location of the so-called monazite belt as it traverses the Carolinas and
Georgia.  However, it should be noted that  heavy minerals, as they are weathered and
eroded from host rocks,  tend to concentrate in stream  sediments  at  highly variable
transport distances. Thus uranium and thorium stream sediment concentrations are at best
an imperfect indicator of adjacent host  rock concentrations.


    We have categorized the geologic factors controlling radon occurrence in the Piedmont
Province as bedrock, saprolite, soil, groundwater, and surface processes.

    Obviously these factors must interact  with one another, but the degrees and types of
interaction are not well known, except in a few instances.  In examining bedrock, we must
consider the  lithology  and mineralogy,  the mobility of radium in either an  oxidizing or
reducing environment, the amount and  continuity of near-surface jointing and fracturing,
the proximity of major faults and shear  zones, the depth of the  water table, and the
proximity of pegmatite dikes and veins. Radon concentration and  migration in saprolite
is influenced by the lithology of the parent rock, the amount and degree of jointing and
fracturing and interconnection, the degree of water saturation, permeability and porosity,
thickness, zonation (whether the saprolite is structured or massive), and the distribution
and  extent of nanopores (pores  less than one micron in width).  In surface  and near
surface soil the principal influences are thickness, zonation (A, B, and C zones), moisture
content (8  to 15  percent has been suggested (2) as  optimum for radon  emanation),
permeability and porosity of the soil, and finally the temperature gradient from the surface,

which determines the water vapor pressure of the soil. The major groundwater influences
are the recharge area, the flow directions and flow  rates, seasonal  fluctuations and the
presence of water supply wells (both of which cause a pumping effect), and the infiltration
of surface precipitation.  Finally, there are meteorologic and topographic effects on radon
migration that can be enumerated but are not well understood. Meteorologic controls on
soil-gas transport that have been  identified  are temperature, humidity, precipitation,
barometric pressure, presence or absence of snow cover, and wind speed and  direction.
Topographic effects are primarily the varying thickness of soil cover on ridge tops and
hillsides versus that  in valleys.


    Until recently, radon occurrence in groundwater has in general been the subject of less
research than in soil  and bedrock, although there  is a fairly large body  of  empirical
groundwater data for the U.S. and some quantitative analysis of that database. There have
been  several published physical models of radon  release from  soil and rock  into
groundwater and transport through aquifers into pumping wells. Together with aquifer
lithology, the width and frequency of fractures and pores/nanopores in the aquifer are key
determinants of both radon release and transport.

    Aquifer lithology has been shown (3) to be a  useful and  relatively accurate predictor
of radon  concentrations in groundwater across  a  wide suite of rock types  in  North
Carolina:   granite,  metasediments and metavolcanics, gneisses  and schists,  mafics,
nonmarine and marine clastic sediments. The average radon concentrations measured in
each of these rock types were generally consistent with relative abundances of uranium in
these rocks.


    In recognition of the need for additional research, EPA  initiated a  groundwater
sampling program in 1988 in Georgia and Tennessee (4). In Georgia, nine  aquifer units
or rock types, with more  or  less homogeneous geologic  conditions, were selected  for
sampling. These nine units or sampling cells are somewhat representative of portions of
the four geologic provinces in Georgia.  Within each sampling cell, ten private well sample
sites were located and sampled.  The sampling protocol required continuous pumping and
measurements every five minutes of temperature,  specific conductance,  dissolved oxygen,
and  pH  until these parameters stabilized.   After purging and stabilization, the water
sampling took place at the  faucet nearest the wellhead. For each sample, a total of forty-
eight  chemical parameters and fourteen radionuclide parameters, including radon and
uranium, were analyzed by EPA analytical laboratories.

    Some preliminary results from  the EPA groundwater sampling  program in Georgia

show a range of radon in groundwater in a Piedmont granite gneiss from 3,160 to 268,500
picocuries per liter (pCi/1) with an average of nearly 82,000 pCi/1.  In the Blue Ridge, as
expected, the minimums, maximums  and average are much lower, as is the case in  the
Valley and Ridge and the Coast Plain.


    The most suspect terranes for radon occurrence in Georgia are granite, granodiorite,
granitic  gneiss,  pegmatites,  mylonites  and  other cataclastics,  carbonaceous shales,
phosphates and phosphorites, and monazite/heavy mineral placers, which are overlain by
high to medium permeability soils and which have conduits for radon migration from  the
source to the surface, such as joints, fractures, faults, bedding planes and foliation planes.

    Much research remains to be done.   The E.P.A. groundwater sampling program
currently underway will be an important contributor  to understanding radon occurrence in
groundwater in the Southeast.

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.

1.    Koch, George S., Jr., A Geochemical Atlas of Georgia, Georgia Geologic Survey,
      Geologic Atlas 3, 1988.

2.    Otton, James K., et al, Map Showing Radon Potential of Rocks and Soils in Fairfax
      County, Virginia, U.S. Geological Survey,  Miscellaneous Field Studies Map MF-
      2047, 1988.

3.    Loomis, Dana P., Aquifer Lithology As a Predictor  of Radon Concentration in
      Groundwater: Research Results in North Carolina, U.S.  E.P.A.  Conference on
      Indoor Radon, EPA 904/9-87 145, 1987.

4.    Coker, Gene  and Olive, Robert, Radionuclide Concentrations from  Waters of
      Selected Aquifers in Georgia, U.S. Environmental Protection Agency, Region IV,



                     (FLUVIAL; MARGINAL
                     IONIC SUBSTITUTION
                   *•  MARINE BLACK SHALES











               URANIUM LOSS INTO
                  - FRACTURES
                  • SHEAR ZONES

                  • LOWER PRESSURE ZONES




                                                        Fe & Mn CONCENTRATION
                                                           (U SCAVENGERS)
                 -JOINTS, FRACTURES,
                  FAULTS, ETC.

                 - POSSIBLE ZONES OF
                   CONCENTRATION OF

                                     U (Uranium!



    R. Randall Schumann1, R. Thomas Peake2, Kevin M. Schmidt3, and Douglass E. Owen1
       iU.S. Geological Survey, MS 939 Denver Federal Center, Denver, CO 80225-0046
                    2U.S. EPA, 401 M St. SW, Washington, DC 20460
                    3USGS, 345 Middlefield Rd., Menlo Park, CA 94025


       A higher percentage of homes in parts of the northern Great Plains underlain by soils
derived from continental glacial deposits have elevated indoor radon levels (greater than 4 pCi/L)
than any other area in the country. Soil-gas radon concentrations, surface radioactivity, indoor
radon levels, and soil characteristics were studied in areas underlain by glacially-derived soils in
North Dakota and Minnesota to examine the factors responsible for these elevated levels. Clay-rich
till soils in North Dakota have generally higher soil-gas radon levels, and correspondingly higher
indoor radon levels, than the sandy till soils common to west-central Minnesota.  Although the
proportions of homes with indoor radon levels greater than 4 pCi/L are similar in both areas,
relatively few homes underlain by sandy tills have screening indoor radon levels greater than
20 pCi/L, whereas a relatively large proportion of homes underlain by clayey tills have screening
indoor radon levels exceeding 20 pCi/L. The higher radon levels in North Dakota are likely due to
enhanced emanation from the  smaller grains and to relatively higher soil radium concentrations in
the clay-rich soils, whereas the generally higher permeability of the sandy till soils in Minnesota
allows soil gas to be drawn into structures from a larger source volume, increasing indoor radon
levels in these areas.

       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.


       Preliminary testing by the U.S. Environmental Protection Agency (EPA) as part of the
EPA/State Indoor Radon Survey (1) indicates that a large proportion of homes built on soils
derived from continental glacial deposits have screening indoor radon concentrations greater than
EPA's recommended action level of 4 pCi/L. Although a correlation between highly permeable
glacial deposits and elevated indoor radon levels has been documented by previous researchers,
notably in northern Europe (2,3,4), the magnitude of the problem in the northern Great Plains
States (Table 1) was unexpected because the deposits have low to moderate permeability and low
surface gamma radioactivity signatures.

      A regional-scale correlation between radon potential and surface gamma radioactivity of
areas underlain by bedrock or soils derived from underlying bedrock (5,6,7,8)  allows
preliminary predictions of radon potential to be made from equivalent uranium or radium data
calculated from gamma radioactivity, one of the best sources of which is the map of NURE aerial
radioactivity data for the conterminous United States compiled by the U.S. Geological Survey

(USGS) (9). However, the number of homes in North Dakota, Minnesota, and Iowa with
elevated indoor radon levels (Table 1) is disproportionately high compared to the generally low
aerial radiometric signature of the area.  Glacial drift derived largely from the Pierre Shale in North
Dakota and from crystalline rocks of the Canadian Shield in Minnesota generate elevated indoor
radon levels in a large number of homes in those States. Soils in the area are not necessarily highly
permeable; in fact, some of the highest indoor radon levels in North Dakota were measured in
homes in the Red River Valley along the eastern border of the State, which is underlain primarily
by silty-clay lacustrine deposits of glacial Lake Agassiz (10,11).

       As part of the EPA-USGS joint effort to identify and characterize the radon potential of the
United States, a field investigation was initiated to quantitatively identify and describe the geologic
factors responsible for anomalous indoor radon concentrations in the northern Great Plains and
Great Lakes States underlain by continental glacial deposits. This report presents preliminary
observations from field investigations in North Dakota and Minnesota.

                                  GLACIAL GEOLOGY

       Most of North Dakota and Minnesota are underlain by Wisconsin-age continental glacial
deposits, except for the southwest corner of North Dakota, which is unglaciated and underlain by
Tertiary and Upper Cretaceous sandstones and shales (Figure 1), and the southeast comer of
Minnesota, which is underlain by pre-Wisconsin glacial drift (Figure 2). In North Dakota, ice
advanced from the north and northwest in six separate glacial advances during Wisconsin time
(Figure 1).  The tills of all the ice advances are lithologically similar and are derived primarily from
Tertiary and Upper Cretaceous shales, siltstones, and sandstones that comprise the underlying
bedrock in North Dakota and southwestern Manitoba. Some of the deposits in the  northeastern
part of the State also include carbonate-rich till derived from Paleozoic limestone and dolomite in
southern Manitoba (11). Most of the tills consist of nearly equal parts sand, silt, and clay (12,13).
Lacustrine deposits of glacial Lakes Agassiz, Souris, and Devil's Lake (Figure 1) are composed
primarily of silty clays and clays, and are commonly interbedded with  tills. The unoxidized tills
are generally dark olive gray to bluish gray. Iron oxidation and accumulation of calcium carbonate
are common weathering effects (14); both were noted in nearly all of the soils sampled in the study

       Wisconsin-age glacial drift covers most of Minnesota. The drift can be classified into
deposits of four major ice lobes that advanced at different times and in different directions, from
areas with different source lithologies (Figure 2). Each lobe experienced multiple phases of ice
advance, some of which overlapped other lobes in time and space.  In  order of roughly decreasing
age, the major lobes are the Wadena, Rainy, Superior, and Des Moines (15). Drift of the Wadena
lobe is exposed only in central Minnesota (Figure 2). The Wadena lobe advanced southward from
the north and northwest, moving in a generally north-south direction in central Minnesota.
Wadena drift is dominantly gray to buff-colored, sandy, calcareous till derived from carbonate
rocks of southern Manitoba. The Rainy lobe moved from northeast to southwest.  Rainy lobe drift
covers parts of northeastern and central Minnesota and varies in both color and constituent
lithology. In the northeast the drift is derived primarily from gabbro and basalt, giving it a gray
color. Further west the drift is light gray to light brown, reflecting a dominantly granite source. In
central Minnesota, Rainy lobe drift is derived mostly from metamorphic rocks and  has a brown
color resulting from oxidation of the initially gray metamorphic rock fragments.

       The Superior lobe advanced from northeast to southwest in the eastern part of the State,
roughly parallel to the Rainy lobe. It carried generally sandy red till containing sandstone and slate
pebbles. Drift of the Des Moines lobe was primarily derived from Upper Cretaceous shales of

southern Manitoba, eastern North Dakota, and western Minnesota. Sublobes of the Des Moines
lobe moved eastward across northern Minnesota and southward to central Iowa. Till derived from
the Des Moines lobe is generally gray to buff, calcareous, silty to clayey (16). Silty and clayey
lacustrine deposits of Lake Agassiz cover much of the Red River Valley in the northwestern pan of
the State (Figure 2).


       Field sampling was conducted at 132 locations along four traverses, each 100-150 km
long, along highway rights-of-way (Figure 1,2). Sample stations were spaced about 4 km apart.
The field measurements were made during two weeks in August, 1989 during which the weather
was mostly warm and dry, so any variations in measured values due to climate or weather-related
effects are estimated to be negligible compared to variations caused by geologic factors. At each
station, soil-gas radon was sampled at 1 m depth using the method of Reimer (17) and surface
gamma radioactivity was measured with a portable gamma-ray spectrometer. The gamma-ray
spectrometer gives an estimate of the concentrations of uranium, radium, thorium, and potassium
in the upper 30 cm of soil. At alternate stations, an additional soil-gas sample was collected and
soil permeability was estimated using soil-gas probes and equipment developed by V.C. Rogers
and Associates (18), and soil profiles were examined, described, and sampled with a bucket auger.
Soil samples were collected at the surface and at 1 m at alternate stations. Laboratory analyses of
these samples were not completed at the time of this writing.

                              PRELIMINARY FIELD DATA

       Results of field sampling of soil gas, permeability, and surface radioactivity are
summarized in Table  2. Each traverse generally characterizes a specific glacial lobe or group of
similar lobes. Traverse ND-1 crosses deposits of advances 2,3, and 4. Most of traverse ND-2
characterizes deposits of advances 5 and 6, except for those samples at the eastern end of the
traverse, which were collected in sediments of glacial Lake Agassiz, and which are treated
separately from the rest of ND-2 (Figure 1; Table 2). In Minnesota, traverse MN-1  crosses
deposits of the Wadena and Rainy lobes, and traverse MN-2 primarily crosses deposits of the Des
Moines lobe (Figure 2).

       Average soil-gas radon concentrations, equivalent uranium (eU), and permeabilities for
each transect are compared in Figure 3 and Table 2.  Soil-gas radon concentrations and eU are
generally higher in North Dakota than in Minnesota,. The highest soil radon and eU
concentrations and lowest permeabilities were measured in soils derived from sediments of glacial
Lake Agassiz. A good correlation exists between eU measured at the  surface and soil radon
measured at 1 m (Figure 4a).  Average ratios of eU to soil-gas radon concentration are significantly
lower in this study than for similar ratios in unglaciated areas, such as those reported by
Gundersen (19) and Gundersen and others (6).

       Permeability appears to exhibit a weak inverse correlation with soil-gas radon concentration
(Figure 4b), although  the range of permeability values is relatively small; all the average values are
between 2xlQ-8 and 9xlQ-8 cm2, and all 39 permeability measurements fall  between  10"10 and 10"7
cm2. It was expected that the Minnesota soils, especially the sandy and silty soils derived from
Wadena and Rainy lobe deposits (traverse MN-1), would have higher permeabilities  than were
measured (U.S. Soil Conservation Service soil surveys for the area describe many of the soils as
moderately to rapidly  permeable). The lower measured values may have been due to localized wet
soil conditions, or to better development of the B horizon, which is generally a zone of

accumulation of fines and therefore less permeable than over- or underlying horizons, in the older
Wadena and Rainy lobe soils. Soils sampled along the Minnesota transects have somewhat higher
and generally more variable permeability than those sampled in North Dakota (Figure 3).


       The physical, chemical, and drainage characteristics of soils formed on glacial deposits
vary according to source bedrock type and the glacial features on which they are formed. Although
the effects of glaciation have modified the relationship between the bedrock source and the soil's
radon generation and transport characteristics, source rock lithology exerts a major control on
radon potential because it determines the initial uranium and/or radium concentrations in the glacial
drift and the type of soil (clayey or sandy, for example) that develops. Soils formed on ground or
stagnation moraine deposits, which underlie most of the study area, tend to be more poorly drained
and contain more fine-grained material than soils formed on outwash or eskers, which are
generally coarser and well drained.

       In general, soils developed from glacial deposits are moderately to highly permeable and
rapidly weathered, because crushing and grinding of the rocks by glacial action may enhance and
speed up soil weathering processes (20). Grinding of the rocks increases the radionuclide mobility
in the resulting soils by exposing the uranium and radium at grain surfaces, where they are more
easily leached and moved downward through the soil profile with other mobile ions.
Accumulations of CaCOj and iron oxides were observed below about 75 cm in most soils in the
study area. CaCOs and iron oxides form soil-grain coatings or concretions that sorb or associate
with uranium (21,22), providing a possible mechanism for uranium accumulation and enhanced
radon emanation in deeper soil horizons. The low surface radioactivity and comparatively high soil
radon concentrations of the glacial soils suggests that radionuclides have been removed from the
upper soil layers and are probably concentrated in deeper horizons. This may explain why
preliminary radon potential predictions based primarily on aerial gamma-ray data typically provide
an underestimate of the number of homes with indoor radon problems in some areas underlain by
glacial drift

       Clayey till soils, such as those underlying most of North Dakota, have high emanation
coefficients (23) and usually have low to moderate permeability, depending on the degree to which
the clays are mixed with coarser sediments. Soils formed on tills consisting of mostly coarse
material, such as the sandy tills that underlie much of Minnesota, tend to emanate less radon
because the larger grains have lower surface area-to-volume ratios, but because these soils have
generally higher permeability, radon transport distances are longer, so buildings constructed in
these materials are able to draw soil air from a larger source volume, and moderately elevated
indoor radon concentrations may be achieved from comparatively lower radioactivity soils
(24, 25). In till soils with extremely high permeability, atmospheric dilution may become
significant, so elevated indoor radon levels are less common.

       Two general classes of glacial soils can be identified from this study:  1) clay-rich soils
with lower permeability and higher emanation coefficients, and 2) coarser-grained soils with lower
emanation coefficients and higher permeability. The effect of this difference in soil characteristics
can be clearly seen by comparing the distributions of indoor radon levels in North Dakota
(Figure 5) and Minnesota (Figure 6). Although both types of soils can generate indoor radon
levels greater than 4 pCi/L, till soils with high emanation coefficients can generate a significant
number of indoor radon levels greater than 20 pCi/L (Figure 5), whereas soils with high
permeability tend to  generate few indoor radon levels greater than 20 pCi/L (Figure 6).


       A comparison of glacial geology, soil-gas radon and gamma radioactivity data, and indoor
radon data for Minnesota and North Dakota indicates that glaciation modifies the relation among
source bedrock, soils, and their resulting radon emanation and transport characteristics, but radon
potential predictions are possible if the nature of these modifications is understood.  The
preliminary observations presented in this paper suggest that two factors are important to consider
when predicting radon potential in glaciated terranes.  1) Glaciers transport and redistribute
bedrock, so a bedrock geologic map may not accurately reflect the parent material lithology of
glacially-derived soils, but knowledge of source rock lithology (or lithologi&f, as glaciers may also
mix rocks from several sources together) will aid in determining the radon emanation and transport
characteristics of the derivative tills. 2) Grinding of the rocks by glaciers reduces grain size and
therefore increases grain surface area, enhancing radon emanation by exposing more radium at
grain surfaces than in coarser-grained soils. Glacial mixing and crushing also speeds up the
weathering process, so radionuclides may be leached from shallow horizons of till soils more
rapidly than in soils developed on untransported bedrock, thus giving a surface gamma
radioactivity reading that may underestimate the uranium and radium concentrations at depth.


       This study was conducted in cooperation with, and funded in part by, the U.S.
Environmental Protection Agency under Interagency Agreement DW14933884-0. Thanks are due
to S.S. Agard and G.M. Reimer for reviewing the manuscript and to L.C.S. Gundersen for
helpful discussions.


1. Ronca-Battista, M., Moon, M., Bergsten, J., White, S.B., Holt, N., and Alexander, B.,
       Radon-222 concentrations in the United States-Results of sample surveys in five states.
       Radiation Protection Dosimetry 24:307-312,1988.

2. Akerblom, G., Andersson, P., and Clavensjo, B., Soil gas radon—A source for indoor radon
       daughters.  Radiation Protection Dosimetry 7:49-54,1984.

3. Castre'n, O., Makelainen, I., Winqvist, K., and Voutilainen, A., Indoor radon measurements in
       Finland: A status report.  la: Hopke, P.K. (ed.), Radon and Its Decay Products.
       American Chemical Society Symposium Series 331:97-103,1987.

4. Stranden, E, Radon-222 in Norwegian dwellings. In: Hopke, P.K. (ed.), Radon and Its
       Decay Products.  American Chemical Society Symposium Series 331:70-83,1987.

5. Duval, J.S., Use of aerial gamma-ray data to estimate relative amounts of radon in soil gas. In:
       Gundersen, L.C.S., and Wanty, R.B. (eds), Field studies of radon in natural rocks, soils,
       and water.  U.S. Geological Survey Bulletin, in press.

6. Gundersen, L.C.S, Reimer, G.M., Wiggs, C.R., and Rice, C.A., Map showing radon
       potential of rocks and soils in Montgomery County, Maryland. U.S. Geological Survey
       Miscellaneous Field Studies Map MF-2043, scale 1:62,500,1988.

7. Peake, R.T., and Schumann, R.R., Regional radon characterizations. la: Gundersen, L.C.S.,
       and Wanty, R.B. (eds), Field studies of radon in natural rocks, soils, and water. U.S.
       Geological Survey Bulletin, in press.

8. Schumann, R.R., and Owen, D.E., Relationships between geology, equivalent uranium
       concentration, and radon in soil gas, Fairfax County, Virginia.  U.S. Geological Survey
       Open-File Report 88-18,1988,28 p.

9. Duval, J.S., Jones, W.J., Riggle, F.R., and Pitkin, J.A., Equivalent uranium map of the
       conterminous United States. U. S. Geological Survey Open-File Report 89-478,1989.

10. Clayton, Lee, Moran, S.R., and Bluemle, J.P., Explanatory text to accompany the  geologic
       map of North Dakota.  North Dakota Geological Survey Report of Investigation 69,1980,
       93 p.

11. Moran, S.R., Arndt, M., Bluemle, J.P., Camara, M., Clayton, L., Fenton, M.M., Harris,
       K.L., Hobbs, H.C., Keatinge, R., Sackreiter, D.K., Salomon, N.L., and Teller, J.,
       Quaternary stratigraphy and history of North Dakota, southern Manitoba, and northwestern
       Minnesota. In: Mahaney, W.C. (ed.), Quaternary stratigraphy of North America.
       Stroudsburg, Pennsylvania, Dowden, Huchinson, and Ross, 1976, p. 133-158.

12. Lemke, R.W., Geology of the Souris River area, North Dakota. U.S. Geological Survey
       Professional Paper 325, 1960,138 p.

13. Winters, H.A., Geology and ground water resources of Stutsman County, North Dakota, Part
       I:  Geology. North Dakota Geological Survey Bulletin 41,1963,84 p.

 14.  Lemke, R.W., Laird, W.M., Tipton, M.J., and Lindvall, R.M., Quaternary geology of the
       northern Great Plains. In: Wright, H.E., Jr., and Frey, D.G. (eds), The Quaternary of the
       United States. Princeton, NJ, Princeton University Press, 1965, p. 15-27.

 15.  Hobbs, H.C., and Goebel, J.E., Quaternary geologic map of Minnesota. Minnesota
       Geological Survey Map S-l, scale 1:500,000,1982.

 16.  Wright, H.E., Jr., and Ruhe, R.V., Glaciation of Minnesota and Iowa. IQ:  Wright, H.E.,
       Jr., and Frey, D.G. (eds.), The Quaternary of the United  States. Princeton, NJ, Princeton
       University Press, 1965, p. 29-41.

 17.  Reimer, G.M., Simple techniques for soil-gas and water sampling for radon analysis.  In:
       Gundersen, L.C.S., and Wanty, R.B. (eds), Field studies of radon in natural rocks, soils,
       and water. U.S. Geological Survey Bulletin, in press.

 18.  Nielson, K.K., Bollenbacher, M.K., Rogers, V.C., and Woodruff, G., Users guide for the
       MK-II radon/permeability sampler. U.S. Environmental Protection Agency report, in

 19.  Gundersen, L. C. S., Anomalously high radon in shear zones. In: Osborne, M., and
       Harrison, J., symposium cochairmen, Proceedings of the  1988 Symposium on Radon and
       Radon Reduction Technology, Volume 1, oral presentations. U.S. Environmental
       Protection Agency publication EPA/600/9-89/006A, 1989, p. 5-27 to 5-44.

20. Jenny, H., The clay content of the soil as related to climatic factors, particularly temperature.
       Soil Science 40:111-128,1935.

21. Hansen, R. O., and  Stout, P.  R., Isotopic distributions of uranium and thorium in soils.  Soil
       Science 105:44-50,1968.

22.  Nash, J. T., Granger, H. C, and Adams, S. S., Geology and concepts of genesis of
       important types of uranium deposits. Economic Geology, 75th Anniversary volume: 63-
       116, 1981.

23. Grasty, R. L., The relationship of geology and gamma-ray spectrometry to radon in homes.
       EOS 70:496,1989.

24. Duval, J.S., Otton, J.K., and Jones, W.J., Estimation of radon potential in the Pacific
       Northwest using geological data. U.S. Department of Energy, Bonneville Power
       Administration report DOE/BP-1234,1989.

25. Kunz, C, Laympn,  C. A., and Parker, C., Gravelly soils and indoor radon.  In: Osborne,
       M., and Harrison, J., symposium cochairmen, Proceedings of the 1988 Symposium on
       Radon and Radon Reduction Technology, Volume 1, oral presentations. U.S.
       Environmental Protection Agency publication EPA/600/9-89/006A, 1989, p. 5-75 to 5-86.

North Dakota
                                           (data from EPA press releases, 1987-89)
See figures 1 and 2 for locations of sampling traverses. Rn=soil-gas radon concentration at 1m,
eU=equivalent uranium measured at the surface by gamma spectrometer, S.D.=sample standard
deviation, n=number of measurements.
eU (ppm)
          ND-2    LakeAgassiz    MN-1
Rn (pd/L)







 **'          V V V ^ V V
 'XUngladated A'X>v

                              20  40
60  80  100
 i    i	I
Figure 1. Map of North Dakota showing Wisconsin-age glacial deposits and locations of sampling
         traverses (ND-1 and ND-2). Numbered lines indicate maximum extent of each glacial
         advance. Modified from (14).

                                                               Glacial Lake
                                                               Oes Moines



                                 30   60 Miles
Figure 2. Map of Minnesota showing Wisconsin-age glacial deposits and locations of s
         traverses (MN-1 and MN-2). Patterned areas indicate deposits of named glacial lobes,
         as indicated in map legend.  Modified from (15).

                              EQUIVALENT URANIUM, ppm
                      2000 H
                            ND-1    ND-2  LakoAg.  MN-1    MN-2

                                   SOIL RADON,  pCi/L
                            ND-1    ND-2  LakaAg.  MN-1   MN-2
                                SOIL PERMEABILITY, cm'
                            ND-1   ND-2  LakeAg.  MN-1   MN-2
                                                               •  AVERAGE
                                                               O  minimum
                                                               +  maximum
O minimum
+ niBxbnum
                                                               • AVERAGE
                                                               O mWmum
                                                               ^ irojufflum
Figure 3. Average, minimum, and maximum equivalent uranium (eU), soil-gas radon, and
         permeability for soils in the study aiea. See Figures 1 and 2 for locations of traverses.

               7    1500 -
               z   1500 -
                                      1             2

                                     AVERAGE eU.ppm
                                  Lake Agassiz
                                2       4       6       8       10

                          AVERAGE SOIL PERMEABILITY, cm 2 x10 "*
Figure 4. a) eU versus soil radon.  Each point represents the average for one traverse.
         b) Permeability versus soil radon. Each point represents the average for one traverse.
         Line fitted visually, excluding the point representing MN-1.

                   20  40  60
                             80  100
                              I   I
                                                   10  25  50  75  100
                    20  40
                              •0  100
                                                 % QTMMr than 20 pCW.
                                             M>  o   1   5  10 25*
Figure S. Percent of homes tested in the State/EPA Indoor Radon Survey with basement screening
         indoor radon values in excess of 4 pCi/L and 20 pCi/L in North Dakota, by county.

   0   30  60
                I*   o   10 25  50 76 100
                                                                     % graatcr than 20 pCM.
                                                       30  60
                                                                  I   tv
                                                                              10  2S»
Figure 6. Percent of homes tested in the State/EPA Indoor Radon Survey with basement screening
         indoor radon values in excess of 4 pCi/L and 20 pCi/L in Minnesota, by county.


                by:  Charles Laymen and Charles Kunz
                     Wadsworth Center for Laboratories and Research
                     New York State Department of Health
                     Albany, New York 12201-0509
    Indoor radon in Onondaga County, New York is largely controlled by
bedrock and surficial geology.  At more local scales, these alone are insuf-
ficient to characterize indoor radon potential.  A detailed study of the
concentration of indoor radon, soil radium, soil-gas radon, soil and bedrock
type, permeability, and home construction practices indicates that above-
average indoor radon concentrations are associated with gravelly moraine and
glaciofluvial deposits, the radium-bearing Marcellus Shale, and high
permeability zones around the substructure of houses built into limestone


    Indoor radon*- produced by the radioactive decay of radium-226 in rocks
and soil, is the product of a complex system of interrelated variables.
Because of the inherent relationship between geology and indoor radon, geolo-
gists have been challenged in recent years with the problems of identifying
areas of potential public health risk and developing protocols for evaluating
indoor radon potential prior to new house construction.  It is recognized
that many factors affect indoor radon, but the importance of these factors at
different scales is not well understood.  New York is geologically very
diverse and the New York State Department of Health has been actively
involved in trying to better understand the geologic controls on indoor
radon.  Onondaga County in central New York is one of several counties in the
state that has received special attention.
       occurs in three forms, Rn-119, Rn-220, and Rn-222.  The longest-lived
isotope, Rn-222, is the most important with regard to public health and will
be referred to exclusively in this paper.

    For the past several years, the New York State Department of Health has
provided activated charcoal canister radon detectors to state residents at a
minimal cost.  Over 4300 measurements have been made in Onondaga County as
part of this program.  In this paper, we examine the county-wide distribution
of indoor radon and the radon problem at several local areas.  Results of
this investigation demonstrate that bedrock and surficial geology are
adequate to characterize indoor radon potential at the county-wide scale, but
at smaller scales of a few square kilometers or less, many other character-
istics need to be considered.  Particularly important is an understanding of
the variability within the major controls and how homes within the area
interface with their environment.

                               GEOLOGIC  SETTING


    Onondaga County is centrally located in New York State and at the
northeastern edge of the Finger Lakes Region (Figure 1).  The county spans
the border between two physiographic provinces,  the Erie-Ontario Lowland to
the north and the Appalachian Upland to the south.  The Erie-Ontario Lowland
encompasses the relatively low, gently undulating plain lying south of Lakes
Erie and Ontario.  Glacial scour has produced numerous ridges of till and
rock, which are surrounded by extensive and thick deposits of stratified
glaciolacustrine sand and mud (3).
    Onondaga County is underlain by Upper Silurian to Upper Devonian shales,
limestones, and minor sandstones that are exposed in east-west trending bands
and dip slightly to the south beneath progressively younger strata (1,2).
Differing resistance of the bedrock to weathering processes has produced a
scarp at the northern edge of the Appalachian Upland that rises sharply to
the south more than 300 m in elevation (Figure 1).  The scarp is comprised of
several benches formed primarily on the Onondaga and Helderberg Limestones
and is, therefore, referred to as the Onondaga or Helderberg Escarpment
    The Appalachian Upland is characterized by greater relief as a result of
fluvial dissection of the uplifted bedrock.  The northern edge of the upland
is cut by the Finger Lakes and adjacent through valleys, which are glacially
modified valleys of preglacial river drainage.  A segment of the largest
moraine in New York, the Valley Heads Moraine, crosses the southern end of
Onondaga County (3,6).  Thick sequences of gravelly outwash fill the deeply
cut through valleys of the upland and thin till covers the upland interfluves
between the valleys (3).
    The dominant soils in Onondaga County are derived from glacial deposits
containing varying amounts of limestone, shale,  and sandstone.  For the most
part these soils are deep, medium-textured, well drained to moderately well
drained, and medium to high in lime content (7).


    Before discussing the geologic controls on indoor radon along the
Onondaga Escarpment, a review of the bedrock lithologies found on the

                                                                  44° N
43° N
         Onondaga & Halderberg Limestone

      :^3 Marcvllut Formation
Figure  1.   Maps showing  the physiographic  provinces of New York
            State and  location of Onondaga  County,  and the study
            areas along the  Onondaga Escarpment.

escarpment is beneficial.  The Upper Silurian Rondout Dolostone is a poorly
exposed gray dolostone at the base of the escarpment (5).   The Rondout
Dolostone grades upward into the overlying Lower Devonian Helderberg Group,
which consists of about 45 m of limestone and generally forms the middle part
of the escarpment (4,5).  The Middle Devonian Oriskany Sandstone overlies the
Helderberg Group.  This sandstone is absent from the western part of the
county, but increases in thickness from about 1 cm southwest of Syracuse
(Figure 1) to about 5 m on the eastern edge of the county (5).  Although the
Oriskany Sandstone is relatively thin, it is notable for containing radium-
bearing, black nodules of calcium phosphate such that it is readily detected
in gamma-ray logs from wells (4,8,9) (Figure 2).  Despite this, indoor radon
has not been attributed to the nodules.  The Onondaga Limestone, which is
about 24 m thick, overlies the Oriskany Sandstone and is the most durable
rock in the escarpment, cropping out at or near the top (5).
    Immediately overlying the Onondaga Limestone are thick beds of black,
bituminous, argillaceous shale collectively known as the Marcellus Formation
(5,9).  This formation increases in thickness from about 30 m at the western
edge of the county to about 90 m at the eastern edge (9).   The Marcellus
Formation is most noted in New York State for its radioactivity and it is
consequently easily recognized in gamma-ray logs of wells (8,9,10) (Figure
2).  In general, radioactivity of the Marcellus Formation is greatest at the
base and decreases progressively upward to the top with the exception of a
1 m thick limestone bed of relatively low activity about 5 m above the base
of the formation.
    Results of whole-rock gamma spectrometry analysis of the Marcellus
Formation are given in table 1.  Similar results for the Onondaga and
Helderberg Limestones are also given in table 1 for comparison.  It is
apparent from these results that the Marcellus Formation shales are enriched
in uranium and radium compared to the underlying limestone units.  In
addition, the uranium-to-radium ratio is nearly one to one for both rock
types (taking analytical error and the geometric standard deviation into
account) suggesting that the radium is in equilibrium with the uranium and
there is no evidence of preferential removal of radium by solution.
                                                226Ra          232Th
Marcellus Formation
Geom. Mean
Geom. Std. Dev.



          Onondaga and Helderberg Limestones
               Geom. Mean          0.3            0.3            0.1
               Geom. Std. Dev.     1.8            1.6            1.8
               n                    31             32             32


                   Permit 11632
Van Patten
Permit 12148
Permit 9578
. — _ _**^«»^^^.

    Figure 2.   Gamma-ray  logs of the Marcellus Formation, Onondaga
              Limestone, and upper Helderberg Group from three wells
              in and adjacent to Onondaga County.



     Indoor radon  survey results reported in this paper are based on volun-
 teered two or  four day screening measurements primarily in the basement using
 activated charcoal canister detectors.  Measurements were made at any time of
 the  year, but  most were made during the fall, winter, and spring.  Follow-up
 measurements were made in homes with screening results greater than 20 pCi/L.


     Soil-gas samples were collected by driving a 2.5 cm diameter steel pipe
 into the soil  to bedrock or a depth of 120 cm (11).  Soil gas was withdrawn
 through perforations at the tip of the pipe and the concentration of radon
 was  measured by alpha scintillation using Lucas cells.


     Permeability  is measured using the same apparatus used to sample soil
 gas.  Soil gas is withdrawn from the ground at a fixed flow rate while the
 pressure differential required to maintain that flow is measured (11).
 Permeability can then be calculated using a standard equation (11,12).


     Soil samples were prepared for radionuclide analysis by placing approx-
 imately 100 g  of air-dried soil less than 2 mm in diameter in a sealed
 polyethylene bottle and set aside for three weeks to allow the radon and
 radon daughters to obtain equilibrium with the radium of the sample.  Approx-
 imately 100 g  of each rock sample was crushed and ground to a powder and
 similarly sealed in a bottle and set aside for three weeks.  The uranium-
 238, radium-226, and thorium-232 concentrations were determined after
 measuring the  rate of gamma emissions by spectroscopy with a Ge(Li) detector.

                           RESULTS AND DISCUSSION


    Over 4300  indoor radon screening measurements in Onondaga County yield a
mean and geometric mean of 10.1 and 4.4 pCi/L, respectively.   Examination of
 the distribution of indoor radon survey results grouped by zip code area can
be helpful in  identifying large-scale geologic controls on indoor radon (13).
 Indoor radon in the Erie-Ontario Lowland physiographic province is generally
below 4 pCi/L  (Figure 3).   This can be attributed to the low concentration of
uranium and radium in the bedrock (9) and an extensive and thick cover of
glacial lacustrine sediments (3) (Figure 3).   These sediments are generally
 fine-grained and well sorted with almost no gravel-size clasts, resulting in
moderate to low permeability for gas flow (7).  Much of the area covered by
 these sediments is poorly drained,  resulting in large swamps.  Only a few

                                        Indoor Radon Concentration
Surficial Geology
               < 2.0 pCi/L

           ^ 2.0 - 3.9 pCi/L
                                                                  4.0 • 10.0 pCi/L

                                                                  > 10.0 pCi/L
     1 Till   [V] Outwiih Qravtt

   PI Licuttrfte Sand & Mud
J2 Moraine
   Figure 3.  Maps of Onondaga County  showing indoor radon concentrations
               and surficial geology

drumlins and areas covered with gravelly loamy till remained above the ice
marginal lake levels and were not covered with glaciolacustrine sediments
(3).  In spite of the gravel component of the till, the fine-grained nature
of its matrix restricts permeability to moderate or lower.
    In the southern part of the county, including the Onondaga Escarpment,
indoor radon concentrations greater than 4 pCi/L are more common (Figure 3).
These higher concentrations are attributed to a) the radium-bearing, black
Marcellus Shale, b) building practices in the limestone bedrock, and c) the
highly permeable moraine and glaciofluvial deposits.  Almost the entire
Appalachian Upland of Onondaga County is covered by till, outwash, or moraine
deposits (3) (Figure 3).  Numerous permeability measurements in these
deposits throughout New York State show variability depending on the texture
and degree of sorting, but in general, demonstrate relatively high permea-
bility for gas flow ranging between 10"° and 10'^ nr (11).  Above-average
concentrations of indoor radon are often associated with permeable surficial
deposits, even in the absence of above-average concentrations of radium in
the soil and bedrock (11,14,15).  We attribute many of the elevated concen-
trations of indoor radon in southern and central Onondaga County to the
highly permeable nature of these glacial deposits.
    Many homes with indoor radon concentrations greater than 4 pCi/L occur
along a narrow east-west trending band across the central part of the county
corresponding to the Onondaga Escarpment.  Although some of the elevated
indoor radon concentrations can be attributed to the Marcellus shale,
detailed field investigations with laboratory analyses reveal that many homes
along the escarpment underlain by Onondaga or Helderberg Limestone also
possess elevated concentrations of indoor radon.  Hand and Banikowski (16)
attributed elevated indoor radon concentrations on, and north of, the
Onondaga Escarpment to redistribution of uranium from the Marcellus Formation
into subjacent limestones by ground water.  No direct evidence for this has
been observed.   Below, we present results from several small study areas
along the escarpment that illustrate the complex nature of the association
between topography, bedrock, erosion and transport processes, soils, and home
construction practices that occur along the Onondaga Escarpment leading to
elevated radon levels.


    This study area is located on the Marcellus Formation and forms the basis
for comparison with the other study areas located on the underlying lime-
stones (Figure 1).  Indoor radon concentrations in 19 single-family homes
located within the study area yield a mean and geometric mean of 8.5 and 6.0
pCi/L, respectively.   The soil derived from rapidly weathered Marcellus shale
is silt loam with shale fragments and is typically less than one-meter thick
and poorly drained (7).  Permeability of the soil is generally moderate to
high,  ranging from 5 x 10'12 to 1 x 10'9 m2, and in most cases, it is fairly
consistent with depth (Figure 4).  The soil contains above-average concentra-
tions of uranium and radium (Table 2).  Soil-gas radon concentrations range
up to 5300 pCi/L.  The mean soil-gas radon concentration near the foundation
of homes is about 1650 pCi/L, whereas in less disturbed soil in the yard away
from the homes it averages about 2200 pCi/L.

             Study Area 1
           Permeability (m2)
         10'13  10'"  10'11  10'
        Study Area 2
       Permeability (m*)
ID'14  ID'13   10"*  10'"
         Study Area 3
       Permeability (m2)
IP'14  IP"'3  10"2  10"'  10'10  10'9
                                          30 •
                                                  Ltowilont TU
                           Figure  4.   Permeability profiles for each study  area.

                                                 226Ra         232^
Study Area 1 (n=47)
Geom. Mean
Geom. Std. Dev.
Study Area 2 (n=14)
Ceom. Mean
Geom. Std. Dev.
Study Area 3 (n=16)
Geom. Mean
Geom. Std. Dev.










    This study area is located on the Onondaga and the Helderberg Limestones
on the escarpment downslope from Study Area 1 (Figure 1).  Indoor radon
concentrations in 39 homes have a mean and geometric mean of 12.7 and 7.4
pCi/L, respectively.  The overburden in this area is variable in thickness
and ranges in texture from gravelly loam to silty clay loam reflecting its
multiple sources (7).  Some of the overburden in this area is calcareous till
derived from limestone and shale bedrock to the north that was transported
southward over the escarpment.  Elsewhere, there is evidence that Marcellus
shale has been transported northward over some parts of the area by slope
wash and fluvial processes.  Disturbance of the overburden during home
construction has made it difficult to separate soils derived from both of
these sources.  The radionuclide content of the till is average to slightly
below average for soils (Table 2).  In general, permeability of the till is
low to moderate, ranging from 5 x 10"^ to 1 x 10"^ m ,  but permeability of
the soil in cracks between limestone blocks and buried limestone talus is
higher, approximately 5 x 10*^-^ or (Figure 4).  The radon concentration of
several soil-gas samples from undisturbed soil in these cracks and buried
talus range from about 1400 to 1700 pCi/L.  For homes built into the
limestone bedrock, permeability is slightly higher adjacent to the home in
the backfilled area around the basement than it is in the undisturbed soil


    This study area also is located downslope from the Marcellus Formation on
Onondaga and Helderberg Limestones (Figure 1).  Indoor radon concentrations
in 56 homes yield a mean and geometric mean of 46.0 and 20.8 pCi/L, respec-
tively.  The soil is shaly loam or silt loam, often less than one-meter thick
(7), and contains above-average concentrations of uranium and radium (Table
2).  Permeability of the soil is generally moderate to high, ranging from

5 x 10'12 to 5 x 10"10 m2, and in general, is fairly consistent with depth
(Figure U).  As in Study Area 2, permeability is higher in the backfilled
area adjacent to homes than in the undisturbed soil farther away.  Soil-gas
radon concentrations range from about 400 to 5000 pCi/L.  The mean soil-gas
radon concentration near the foundation of homes is about 1600 pCi/L, whereas
less disturbed soils in the yard away from the home have a mean of about 1900
    It has become increasingly clear after intensive study that several
important factors are acting together to increase the radon potential of
homes in this area.  Even though these homes are built on limestone that
possesses very little radon source potential (Table 1), the overlying soil,
which infiltrates joints and solution cavities in the limestone bedrock
provides a sufficient source of radium for radon production.  From the
texture of the soil, abundance of shale fragments, and similar radionuclide
content to the Marcellus shale and corresponding soil, it is clear that much
of the soil in this area is derived from the Marcellus Formation and not the
underlying limestone bedrock.  It is unclear if the Marcellus shale was
incompletely eroded by glaciers from the Onondaga Limestone, or if it was
transported beyond its bedrock limits and over both the Onondaga and the
Helderberg Limestones by mass movement or fluvial processes.  Inconclusive
evidence exists for both mechanisms.
    The choice of home construction practices by builders in Study Area 3
also influences indoor radon potential.  Most homes in the area are two-
story, colonial, single-family homes between 10 and 30 years old with full
basements or combination basements and crawl spaces.  Most of these homes are
heated by forced air from gas furnaces located in the basement.  Basement
walls were made with cement blocks, open at the top, and penetrated by the
utilities.  Slabs were poured up to the wall and now slab-wall separations
occur in at least 50 percent of the homes examined.  Floor drains, rather
than sump holes, were used almost exclusively, and floor cracks are common.
    Indoor radon measurements on the first floor and in the basement of homes
in Study Area 3 shows that radon concentrations on the first floor are on
average 18 percent of the concentration in the basement.  Radon concentra-
tions in crawl spaces are about 50 percent higher than in the adjacent
basement.  Radon concentrations in the block walls are about the same or
slightly higher than in the basement, and concentrations below the slab, as
indicated by grab samples from floor drains, cracks, and slab-wall separa-
tions, are about 100 percent higher than in the basement.


    Excavation of limestone bedrock is a difficult, time-consuming, and
expensive task for the home builder, particularly when drilling and blasting
are required.   Consequently, builders are forced to either build a split-
level home partially below grade, build the home with a full basement only
partially below grade,  or build combination partial basements with crawl
spaces.   Below-grade combinations seem to be more common in the limestone
bedrock terrain along the Onondaga Escarpment.  Basement and crawl space
slabs were poured directly on the underlying limestone bedrock and rubble
that was generated in the excavation process or over an intervening gravel

bed.  In either case, sub-slab communication is excellent and the sub-slab
area becomes a part of a large network of interconnected joints and solution
cavities in the limestone bedrock underlying and surrounding the house.
    Joints developed in most beds of the limestone bedrock are widely spaced,
resulting in large blocks commonly more than one meter to a side.  Large
numbers of these boulders are excavated for basement construction.  In the
absence of much overburden on the limestone bedrock along the escarpment,
backfilling around the below-grade foundation with the excavated material
produces a poorly compacted zone with void spaces.  One builder, who needed
to import top soil to cover the bedrock around some homes that were built
into the limestone bedrock, stripped top soil from its source along with the
brush that was growing on it to help prevent the top soil from infiltrating
the cracks and cavities.  In Study Area 3, soil derived from Marcellus shale
was incorporated in the backfill, adding a fine-grained, radium-bearing
component to the matrix of the gravelly limestone backfill.  The result of
these practices is the formation of a highly permeable zone around the
foundation of the homes that may even be interconnected with the network of
joints and solution cavities in the bedrock.  The potential soil-gas
collection area of these homes is so large that, even in the absence of a
substantial radon source, indoor radon concentrations can become very high.

                           SUMMARY AND CONCLUSIONS
    The radium content of bedrock and permeability of surficial materials are
the primarily factors controlling indoor radon in Onondaga County.  However,
at smaller scales, additional factors also affect indoor radon to the same
extent.  The most important of these factors is how each house interfaces
with its environment.  Along the Onondaga Escarpment, where overburden is
thin and many homes are built into limestone bedrock and surrounded by a
highly permeable zone, there is a high potential for these homes to have
elevated levels of indoor radon even though the radium content of the bedrock
is well below average.  The effectiveness with which a home draws soil gas
from this zone is dependent on many design characteristics.  Other factors
affecting indoor radon in Onondaga County include dispersal of radium-bearing
rock or soil by glacial, mass movement, and fluvial processes.  These
processes extend the effect of the radon source rock beyond the mapped limits
of its bedrock source.  Consequently, identification of radon risk based on
geologic contacts can greatly underestimate the area affected.  The
complexities of the radon system identified in this paper are not unique to
Onondaga County.  They most certainly occur elsewhere in New York State and
may be prevalent throughout much of the eastern U.S., particularly in the
Appalachian Mountain region.
    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 made.


    Funding for this research was provided by the New York State Department
of Health Radon Program and the New York State Energy Research and
Development Authority.  We are grateful for the field and laboratory
assistance of R. Mahoney, C. Parker, M. Reynolds, and D.  Gordon.

1.  Broughton, J.G., Fisher, D.W.,  Isachsen,  Y.W.  and Rickard,  L.V.   Geology
    of New York.  New York State Museum and Science Service,  Educational
    Leaflet 20, Albany, New York, 1966, 49 pp.

2.  Rickard, L.V. and Fisher, D.W.   Geologic map of New York,  Finger Lakes
    sheet.  1:250,000.  New York State Museum and Science Service,  Map and
    Chart Series No. 15.  1970.

3.  Muller, E.H. and Cadwell, D.H.   Surficial geologic map of New York,
    Finger Lakes sheet.  1:250,000.  New York State Museum and Science
    Service, Map and Chart Series No. 40.  1986.

4.  Clarke, J.M. and Luther, D.D.  Geologic map of the Tully  Quadrangle.  New
    York State Museum, Bulletin 83.  1905.  70 pp., map.

5.  Hopkins, T.C.  The geology of the Syracuse quadrangle. New York State
    Museum, Bulletin 171.  1914.  80 pp., map.

6.  Muller, E.H.  Surficial geology of the Syracuse field area.  In:  J.J.
    Prucha (ed.), Guidebook, New York State Geological Association,  36th Ann.
    Mtg., May 8-10, 1964. p. 25.

7.  Hutton, Jr., F.Z. and Rice, C.E.  Soil survey of Onondaga County, New
    York.  U.S. Department of Agriculture, Soil Conservation  Service and
    Cornell University Agricultural Experiment Station, 1977.   235 pp., maps.

8.  Vickers, R.C.  A radioactivity study of the sedimentary rocks of central
    New York State and a description of the methods and apparatus used.
    M.Sc. Thesis, Syracuse University.  Syracuse,  New York.  1951.   39 pp.

9.  Rickard, L.V.  Stratigraphy of the subsurface Lower and Middle Devonian
    of New York, Pennsylvania, Ohio and Ontario.  New York State Museum, Map
    and Chart Series, No. 39.  1989.  59 pp., maps.

10. Leventhal, J.^S., Crock, J.G. and Malcolm, M.J.  Geochemistry of trace
    elements and'uranium in Devonian shales of the Appalachian Basin.  U.S.
    Geological Survey Open-File Report 81-778.   1981.  72 pp.

11. Kunz, C.,  Laymon, C.A. and Parker, C.  Gravelly soils and indoor radon.
    In:  Proceedings of the 1988 Symposium on Radon and Radon Reduction

    Technology,   U.S.  Environmental Protection Agency,  Denver,  CO.   1988.

12. DSMA, Atcon, Ltd.   Review of existing instrumentation and evaluation of
    possibilities for research and development of instrumentation to
    determine future levels of radon at a proposed building site.  Report for
    the Atomic Energy Control Board, Ottawa,  Ont.   1983.

13. Mose, D.G.,  Chrosniak, C.E. and Mushrush, G.W.   State-size radon hazard
    maps based on zip code compilations.  In:  Abstracts  with Programs,
    Geological Society of America.  1989.  p. A143.

14. Grace, J.D.   Radon anomalies in southern Michigan.  In:  Abstracts with
    Programs, Geological Society of America.   1989.  p. A144.

15. Smith, G.W., Mapes, R.H., Hinkel, R.J. and Darr, R.L.  Radon hazards
    associated with glacial deposits in Ohio.  In:   Abstracts with Programs,
    Geological Society of America.  1989.  p. A144.

16. Hand, B.M. and Banikowski, J.E.  Radon in Onondaga County, New York:
    Paleohydrogeology and redistribution of uranium in Paleozoic sedimentary
    rocks.  Geology  16: 775, 1988.


            by:   James K.  Otton
                  Joseph S. Duval
                  U.  S. G.  S.
                  Reston, VA  22092

      Indoor radon data for some townships in the Pacific Northwest
(Washington, Oregon, Idaho, western Montana and western Wyoming) have been
compared with aerial gamma-ray data which show the radium content of surface
materials.  Surface radium measurements provide a first-order estimate of the
average levels of indoor radon where soils have low to moderate intrinsic
permeability.  Areas with significantly higher average indoor radon are almost
all characterized by soils that have higher permeabilities, based on county
soil descriptions.  The permeability effect is greatest in the dry areas (less
than 50 cm annual precipitation) in the eastern part of the study area.  In
the wet Puget lowland, elevated indoor radon levels occur only in houses on
soils with extremely high permeability.  Some of the areas above the general
trend are also characterized by steep slopes.

         Session C-V:
Radon Entry Dynamics—POSTERS


                  by:   David E. Hintenlang, Ph.D.
                        Department of Nuclear Engineering Sciences
                        University of Florida
                        Gainesville, FL  32611

                        Richard A. Furman
                        School of Building Construction
                        University of Florida
                        Gainesville, FL  32611

      Soils having  permeabilities  in the range  of  10-"-10-18 m2  have sometimes
proven to  be  difficult subjects for the successful  implementation of radon-
mitigating  sub-slab suction  systems.    The characteristics  of  soils having
permeabilities  in this range  have  been  studied and  a model  developed that
describes the pressure fields and airflows that may  be  expected due  to  sub-slab
suction points.  The model has been computerized to permit its use as a design
tool for a variety of slab and  sub-slab  suction system  configurations.  Methods
of using  this  model to help optimize  sub-slab suction system  design  and the
effectiveness  of  mitigation will  be presented.    Pressure-field  and  airflow
predictions that have been made using this technique, as well as the resulting
decreases  in  indoor radon concentrations, will  be  compared with measurements
made on the University of Florida Test Slabs and for  whole-house installations.

Funding Statement.   Research sponsored  by the  Environmental Protection Agency
(CR 814-925-010)  and the State  of  Florida Board of  Regents  State University
System Radon Research Program.   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.

      Sub-slab depressurization systems have proven to be an effective means of
reducing indoor concentrations of radon-222 gas  in  single-family houses.  Their
effectiveness  has  been demonstrated  in the  northeast portion of  the United
States,  and these systems are now being implemented  in other areas of  the country

that  may  utilize  different  construction techniques  and  materials.    The
performance of sub-slab depressurization  systems  is  strongly dependent on the
characteristics of specific  construction  features  and,in  particular,  the sub-
slab medium.   In order to adapt  these  systems to the conditions  that may be
encountered in  different geographical  areas,  it  is  imperative to develop a
precise understanding of the operation of these systems and their interactions
with the house/soil system in which they are located.

      The development of a complete understanding of sub-slab depressurization
systems is also necessary for the practical aspects of optimizing the design and
installation procedures for these systems.  Some considerations encountered when
optimizing these  mitigation systems are the minimization of the number of suction
points, accounting  for the  limited  availability  of locations  to  install  the
suction points,  and  the optimization of  ventilation pipe and  mitigation  fan

      In order to address some  of these issues, a mathematical/computer model
that simulates the  performance of sub-slab depressurization  systems  has been
developed and applied to buildings of slab-on-grade construction  type, including
slab-on-stemwall  and  monolithic  poured  slabs.   These types of construction
represent the dominant construction types in Florida.  This area also utilizes
sandy, sub-slab  fill materials  almost  exclusively.   The mathematical/computer
model is, therefore, specifically designed to accommodate these conditions but
may  be  applied  equally well in any region  where the sub-slab materials  are
homogeneous and  the  air velocities  in  the sub-slab materials  induced by  the
ventilation system are sufficiently low to avoid turbulent flow.

      Sub-slab  depressurization systems  contribute  to  reduced  indoor  radon
concentrations through two mechanisms:  1)  they may provide sufficient airflow
and dilution of the soil gas immediately underneath a slab so that the radon gas
concentrations at that point are significantly reduced, or 2) they may eliminate
or reverse the direction  of  pressure-driven flow of radon containing soil gas
by  reversing  the  sub-slab  pressure  differentials.    Source   strengths  are
frequently strong enough, and  the airflow  rates  small enough  (for sub-slab
materials having air permeabilities less than about 5 x 10"" m2)-  that mechanism
2) is frequently dominant.  This results in the necessity of ensuring that any
installed  sub-slab  depressurization  system  can,   in  fact,  depressurize  a
significant area under a slab, relative to the above-slab pressure.

      Airflow through porous media, such as the fill encountered  below a slab,
is governed by several fundamental  laws (1).  These are the  conservation of mass
for a fluid moving through porous media,

                                V(p-V) = 0

where p is the fluid density, V is the fluid velocity and Darcy's Law,

                                V  = - S  vp                             (2)

where k is the air permeability of  the  porous medium, y is the fluid viscosity,
and y.p is the applied pressure gradient.  The simultaneous solutions of these
equations for appropriate boundary  conditions describe the airflow and pressure
fields developed by a sub-slab depressurization system.

      Substituting Darcy's Law into the equation of mass conservation, assuming
that the fluid (air) is incompressible, yields:

                             V'  (K V-P) = 0                             (3)

which must then be solved for appropriate boundary conditions.

      In order to obtain solutions for a wide variety of boundary and sub-slab
conditions in three dimensions, a computer solution utilizing finite difference
methods was developed.  The methodology of this approach involves dividing the
sub-slab volume into a number of small discrete volumes, referred to as cells.
Each cell  is  permitted to communicate  (have  a flow of air)  with its nearest
neighbor cells.  The program iterates a series of steps where the flow of soil
gas from  cell to  cell  is governed  by equation  (3).   The  iterative process
continues until steady-state  conditions  are  achieved;  i.e.,  the flow into the
sub-slab volume is equal to the flow out of the sub-slab volume.

      Boundary  conditions  that  govern  the  generic  operation  of  sub-slab
ventilation systems are determined by the presence of sources and sinks for the
air current.  In practical sub-slab depressurization  systems, air must be drawn
into the sub-slab  volume  to replace  the air  that  is  forced out of that volume
by the suction fan.  The precise sources of this replacement air vary according
to the physical conditions unique to that site.  For slab-on-grade construction,
there are several common sources for this replacement air. Cinderblock stemwalls
frequently provide a dominant  source,  since they have permeabilities equal to,
or larger than, those of the soils of interest and allow outdoor air to  be pulled
into the sub-slab volume.  Where the block stemwalls are well sealed, air must
be pulled into the soil around the building from the outdoor air.  This air is
then transported below the stemwalls and footings to the sub-slab volume.  Cracks
in and around  the concrete slab also allow indoor air to be transported to the
sub-slab volume.  The primary sink for  sub-slab depressurization systems is the
suction pressure  applied to the sub-slab  materials, either in the  form of a
suction pit or an extended suction scheme such as a drainage mat.

      The boundary conditions  that characterize each of these sources or sinks
may be specified as a pressure measured relative to atmospheric pressure.  The
pressures at boundaries that permit the infiltration of outdoor air are therefore
equal to atmospheric  pressure,  and the pressures  at the suction points simply
become the applied suction pressures.

      The implementation  of this technique for computer solutions requires the
complete specification of all boundary conditions. These correspond to physical
quantities  such as slab dimensions,  construction type and geometry, the sizes
and  locations of  cracks,  and the number, location,  and size  of the suction
points.  Following the specification of these parameters, the finite difference

methods  discussed  previously  are  utilized  to  obtain a  three-dimensional
simulation of the pressure and velocity fields present in the sub-slab volume.

      Although  the  calculations  are  performed in full three dimensions, it  is
usually most effective to present the data in the  form of a two-dimensional slice
(vertical or horizontal) in  which the pressure or velocity field contours may
be  easily  evaluated.   For the  purposes of  designing and  evaluating sub-slab
depressurization systems for radon mitigation,  the pressure fields of primary
interest are those in the horizontal  plane immediately below the concrete slab.
The pressure fields generated for this location also lend  themselves well  to
experimental measurements for verification of the computer simulations.

      Experimental evaluations of the pressure fields in the plane immediately
below  the  concrete slab  have been  made  on a  number of different  slabs for
comparison with the computer predictions.   These tests have been performed both
at houses where active sub-slab depressurization systems have  been  installed and
at the University of Florida Test Slab Site. The Test Slab Site was explicitly
built to  investigate  the  performance of and changes  associated with sub-slab
depressurization systems under a variety of conditions (2).  The University of
Florida Test Slab site consists  of four  test slabs built with the slab-in-stem-
wall construction technique using standard construction practices.   Each slab
has dimensions of 24 ft x 48 ft (7.3 m x 14.6 m)  and represents a typical Florida
house  construction technique  through  the completion  of  the  slab,  with  no
additional building shell.  The measurements reported here are  obtained from Test
Slab #2 which is constructed with 24 in  (0.6 m)  of sand fill material above the
natural soil, which is also sandy  in nature (Figure 1).  This  test slab utilizes
a polyethylene curtain placed along the  stem wall  for  half the perimeter of the
slab.  This  effectively  seals the block walls  from air  infiltration  for that
portion of the slab.

      The site  provides an  excellent facility for mapping the  pressure field
created by a variety of different suction point configurations  and comparing them
with  the  pressure fields generated from  the  computer  simulation for  those
experimental conditions.

      The Test Slab has a series  of  22  3/4-in.  (0.02  m) holes drilled through
the concrete slab in a regular pattern across the slab (Figure 1).  These holes
serve  as  the  measurement  points  for  pressure  field  mapping  and  velocity
measurements.  A suction pit in the form of a 12 in.   (0.3 m)  radius hemisphere
was created below the  suction point penetration  and was not backfilled with any
material in this series of experiments.

      The experimental procedure for pressure field mapping consisted of applying
a  suction  pressure of  500  Pa  to  a central  suction point  while all  other
penetrations remained sealed.   Suction  pressure was  created  by  an  industrial
vacuum cleaner fitted  to a regulating valve.  Pressure at the suction point was
monitored throughout the  experiment using a Neotronics  micromanometer,  Model
MP20SR.   Airflow out  of the  suction point was calculated from a measurement of


1 2 3
4.5 6
19 y*J

. * •
10* 11 12
V 21
13 U 15
16 17 18

•— B

Figure 1.  University of Florida Test Slab Configuration,   (a)  A hori-
           zontal cut-away of the footing, stemwall, slab,  and  fill
           construction,  (b)  A vertical view of the test  slab illus-
           trating the plumbing rough-in (A), stemwall curtain  (B) stem-
           wall (c), suction point  (SP), and test points.

the velocity of  the  air flowing through  a  pipe from the suction  point.   Air
velocity was measured with  a Kurz  Model 440 air velocity meter.   The induced
pressure was sequentially measured at each of the test points by inserting the
measurement hose of a micromanometer (Neotronics, Model MP20SR)  through the test
hole and sealing it from the above slab atmosphere.  All pressures were measured
and reported relative to atmospheric  pressure.   A commercial software package
(Surfer, Golden Software) was used to interpolate between data points, using a
Kriging  algorithm,  and  to  develop  contour lines  of constant pressure.   The
empirical pressure fields developed  for the central suction point configuration
are illustrated in Figure 2a.

      Pressure field extension measurements, total  system  airflow, and indoor
radon measurements were also made on  four research houses.   These houses were
initially  evaluated  using  EPA  Diagnostic  Protocols(3).     Based  upon  the
information obtained during  these visits, sub-slab depressurization systems were
designed for  each of  these houses  utilizing   the  computer  simulation.   The
parameters describing the operation and performance  for  each of these systems
are measured and compared to the predictions.  The airflows are predicted from
measurements of the soil permeability and applied suction pressure, along with
the  geometric  dimensions  of  the  specific  house  slab  and  compared  to  the
measurements made at each suction point  which are shown in Table 1.  All but one
of  these  sub-slab depressurization  systems consist  of  two  suction  points
connected to a single fan.

      The  system  is  also  configured  so  that  the suction  points  may  be
individually valved off.   By valving one suction  point  off  and measuring the
pressure induced at that point by  the  other suction point,  it is possible to
obtain a measure of the pressure field  that may be compared with the predicted
pressure field.   The  pressure field due  to  single  suction  point operation as
predicted by the computer model for  House 789  is illustrated in Figure 3 with
the measurement of the pressure induced at the opposite suction point posted at
that location.   The reduction of  indoor radon  concentration  by the designed
pressure fields is listed in Table 1.

                            RESULTS AND DISCUSSION
      The  pressure fields  generated under  the test  slabs are  simulated by
utilizing the computer model configured to the same boundary conditions as the
test slab.  This configuration permits the comparison of  the model results with
experimental results for two different sets of boundary conditions:  sealed and
unsealed  stemwalls.    The  boundary conditions  utilized for  the  model  that
correlate with the test slab are:  1) no  interaction of  the sub-slab environment
with the outdoor atmosphere along the outer perimeter of  the slab to a depth of
1 m, corresponding to the case of the polyethylene-curtained stemwalls, and 2)
differential pressure equal  to atmospheric pressure  at the outer perimeter of
the slab,  corresponding to unsealed stemwalls.  The  test slab pressure-field
contours  developed from the computer  simulation,  Figure 2b, may be directly
compared with the contours developed from the experimental  test points, Figure
2a.  The simulated contours coincide with the experimental data over the entire

Ro *
Po -
Qexp -
kd *
R0(cm) P0(Pa)
46 385
40.7 388
46 387
46 388
46 378
42 325
46 241
Qexp(cfm) Qpred(cfm) kd(m2) kc(m2) [Rn]Before [Rn]After
3.6 4.
2.2 3.
3.0 4.
3.2 4.
2.0 4.
20.4 4.
2.6 2.
3 e.Oxlo"11 S.oxio"11 12 1.9
9 6.0X10"11 3.5X10"11
4 6.0X10"11 4.2X10"11 17.5 1.4
4 6.0X10'11 4.2X10"11
5 6.5X10'11 3.0X10"11 18.9 7.1
9 6.5X10"11 	
7 e.oxio'11 e.oxio'11 8.9 2.0
The hemispherical suction pit radius.
The empirically measured
The empirically measured
Values are ± 0.3 cfm (1.
The flow rate predicted
The air permeability of
«*_.!..._._ «*•.**. -L. ^ S\ ft.
pressure at the suction point. Values are ± 3 Pa.
flow rate from
4xlO~4m3/s) .
the suction point.
by the computer model calculations.
the sub-slab soils determined by diagnostic communication tests.
      The corrected air permeability that permits the computer model calculations to reproduce Qexp.

                Indoor radon concentrations averaged over several weeks  of  lived-in conditions,
Berore/Arter    before and after the installation of  sub-slab depressurization systems.

      0  2  4  6  8  10  12 14 16  18 20 22  24 26 28 JO 32 34 36 38 40 42 44  46 48
                             Slab  Length,  ft.
      0  2  4  6  8 10 12 14  16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48
                              Slab Length, ft.
    Figure  2.   Test Slab Pressure Fields.  The vertical  and  horizontal
                axes are the slab dimensions in feet, and the right half
                of  the slab incorporates the stemwall curtain,   (a)
                Measured pressure field contours  (Pa) from an applied
                suction pressure of -500 Pa, plotted for  pressures
                greater than -60 Pa.  (b) Pressure  field  predicted by
                the computer simulation of the experimental conditions.

slab, indicating that the model accurately predicts the pressure fields for both
sealed and unsealed stemwall conditions.

      Both the experimental results and those from the computer model demonstrate
how polyethylene-curtained stemwalls extend the pressure field immediately under
the slab.  Figure 2  demonstrates  that,  although sealing the stemwall does not
dramatically change the pressure-field extension along the long axis of the slab,
it does permit the pressure field to reach the edges of the slab across the short
axis.   In  this  example,  if a  pressure  differential of 4 Pa were required to
neutralize the natural depressurization of a house,  the sealed stemwall could
account  for  an  additional 4%   of the  slab  area being maintained  below this
differential  pressure.   These  results  demonstrate  how sealing  stemwalls from
airflow  extends the  pressure field and that the effect of  stemwall sealing will
be most  evident for small pressures,  on the order  of a few Pascals.   These
pressure field extensions can  dramatically reduce radon entry by encompassing
major entry routes such as the slab/footing joint.

      The test slab experiments provide verification of the operation of the sub-
slab computer model under well-characterized conditions. Additional experiments
and tests  were  performed using houses  participating  in a U.S.  EPA mitigation
research project.  Unlike the  test slabs, the  sub-slab nature  of these houses
could only be inferred from the  results of  diagnostic tests performed at each
of the houses.  These houses,  located in Alachua and Marion  Counties, Florida,
provide data  on  the  range of applicability for the sub-slab computer model in
Florida  houses,  provide  the   opportunity  for  whole  house  verification,  and
demonstrate  the  facility of using the  computer simulations as an  aid  in the
design  of  sub-slab  depressurization  systems.    The  initial  data  for these
simulations were obtained from diagnostic visits  in the fall of 1988.   Once a
particular design was chosen and  installed, experimental measurements were made
for comparison with  the simulation predictions.

      The pressure at the suction point may  be predicted by developing a sub-
slab system  performance  curve, applied pressure vs.  airflow,  and matching it
against  the  fan performance  curve.   The sub-slab system performance curve is
developed by running  computer simulations of the sub-slab environment for  several
applied  pressures.    The  fan  performance  curve  is  obtained  from the  fan
manufacturer  and verified empirically.   The intersection of these two curves
describes  the pressure and airflow where the entire sub-slab depressurization
system will  operate,  assuming  that the  system  is designed such that there are
negligible pressure drops from other sources.  Since  each of the installations
examined here utilizes a  single fan,  when  two suction  points  are utilized,
slightly different pressures may  be measured at each  of the  suction  points due
to several variables  such as differences in suction pit size, length of pipe run,
number of elbows, and airflow rates.

      While  the predicted  pressure  may be  corrected to   account  for these
variables, for  the  purpose  of evaluating the  sub-slab computer model and the
predicted flowrate,  it is most straightforward to perform a simulation utilizing
the empirically determined applied pressure.  Predicted airflow rates are listed
in Table 1 along with the measured flow rates.  There  is agreement between all

of these measurements except for suction point  892-2.  The extraordinarily high
flow  rate  here may  be  attributed to  a crack  in  the cement  floor  that runs
directly into the suction point.

      All of these flow  rates  were predicted utilizing effective permeability
values  obtained from  the  diagnostic  sub-slab  communication  testing.   From
measurements of the applied pressure and airflow rates, we can  use the computer
model to calculate an updated value of the permeability.   It  is  therefore likely
that  some of the variations between the predicted and experimentally measured
permeabilities  may  be  attributed to  some variation  in  the  sub-slab  soil
permeability from the fall of 1988 to the installation time.

      Since these  houses are  occupied,  it is  not  possible to  obtain a high
resolution map of the pressure fields created underneath the slab.  As previously
discussed, it is, however, possible to obtain a simple measure of the pressure
field extension in the houses that utilize multiple suction points. The pressure
field induced under such a configuration for one house is demonstrated in Figure
3, along with the predicted pressure field contours from the single suction point
operation.  These  figures indicate that the predicted pressure  fields are in
agreement with the measured  pressure point along the axis connecting the suction
points.  The pressure fields  predicted by simultaneous operation of both suction
points are also illustrated  in  Figure  3.   Since the magnitude of the pressure
fields from single-point  operation and  test slab data is correctly predicted,
it is reasonable that the pressure fields created from the simultaneous operation
of two suction points are also correctly predicted.

      The  effectiveness  of  the  designed  pressure  fields  in overcoming  the
pressure-driven flow of radon-222 into these  houses is demonstrated  by  the
reduction of the indoor radon concentrations shown in Table 1.   One house, House
892, continues to have  elevated radon concentrations,  although  they are reduced
from  the original concentrations.   It  is most  likely that this is caused by a
significant reduction in the pressure field produced at suction point #2 by the
crack adjacent to  the  suction point as previously  discussed.   It  is expected
that, when this crack is sealed, the pressure fields will extend to the designed
contours and the indoor radon concentrations will be reduced to levels comparable
to  that  achieved through  the  use of  sub-slab depressurization in  the other
research houses.

      The  preceding  results indicate  that the  computer  model  that  has  been
developed accurately matches experimental  conditions. The  usefulness of such
a model lies in its ability to make useful predictions.  These may include the
use of the computer  model as a  design  tool to increase the  efficiency of sub-
slab depressurization systems,  or as a  means of predicting the  behavior of sub-
slab  depressurization  systems  under  conditions  that  are  time-consuming,
expensive, or otherwise difficult  to measure experimentally.

      The ability of  this  computer model to simulate the operation of a sub-slab
depressurization system prior to installation makes it a valuable design tool.
For sub-slab depressurization  systems to  be  acceptable  to homeowners,  it is
necessary to install the suction points in inconspicuous locations.   These are
not usually the optimum locations for the suction points. By  performing computer

      0  2  5  7 10 12 15  17 20 22  25 27 30 32  35 37 40 42  45 47 50 52  55 57 60
                              Slab Length, ft.

      0  2  5  7  10  12 15 17 20 22 25 27  30 32 35 37  40 42 45 47  50 52 55 57 60
                               Slab Length,  ft.


   Figure  3.   Research House 789 Pressure  Fields  (a)  Predicted contours

               for pressures greater  than -60  Pa from a single suction

               point.  (b)  Predicted contours for pressures greater than

               -60 Pa for the simultaneous  operation of both suction


simulations, it  is  possible  to evaluate the performance  of  a sub-slab system
utilizing the discrete  locations  that are available for  use.   Utilizing this
methodology, it  is  possible  to design a sub-slab depressurization system that
provides for the optimum number and location of suction points, that maximizes
system efficiency, and that minimizes the visual impact to the homeowner.

      Predictions  of  the  physical  processes  that  occur  during  sub-slab
depressurization are also possible  through  these  modeling techniques.   One of
the most important aspects of these modeling techniques  concerns the development
of pressure fields  under conditions of homogeneous permeability.   Under these
conditions  the  magnitude  of the  pressure  fields established by  the sub-slab
depressurization system is  independent of the permeability.   For a given set of
boundary conditions, the pressure  fields will always be the same, regardless of
the soil permeability.  In the case of non-homogeneous soil permeability, the
pressure fields  created will  be  modified  by  the inhomogeneities,  but  these
changes will be  a function of  the ratios of the different permeabilities, and
are still independent of the permeability magnitude.

      Permeability  does, however,  affect  soil  gas velocity and flow.   From
Darcy's Law, for a given pressure  field gradient,  the greater the permeability,
the greater the  flow rate.   Since the pressure  field  gradients are determined
solely by the boundary conditions and are the same regardless of permeability,
it follows  that changes in  the soil  permeability  must  affect the  flow rate
accordingly.  Both of these effects are predicted  by the computer model and may
also be derived analytically for simple geometries.

      An interesting application of the computer model  is  the simulation of the
effects  of cracks  and  other  openings  in a  concrete  slab on  a  sub-slab
depressurization system.  In order to model this case we have assumed that the
crack, or opening, is large enough and that the air velocity moving through it
is small enough  (small soil permeability) so that the pressure drop across the
opening can be neglected.  The  opening,  therefore, permits atmospheric pressure
to be maintained under the  slab immediately  below  the opening.  This becomes an
additional boundary condition for this problem.

      The incorporation of this boundary condition leads to  two effects on the
physical conditions present beneath the slab.  The induced pressure in the sub-
slab volume surrounding the penetration will be reduced and the flow through the
sub-slab volume, and out of the suction point, will be  increased.  The increased
flow, predicted  by the computer model, will provide an  indirect reduction in the
applied pressure by  changing the sub-slab performance curve in such a manner that
it intersects the fan performance curve at a higher flow rate, and therefore a
smaller applied pressure.   In most practical cases the  increase in flow rate is
small enough that these effects are negligible.

      The computer model predicts that the area under the slab affected by the
penetration will be determined by its proximity to the suction point.   Figure
4  illustrates the pressure fields  produced in  two situations for  a constant
applied pressure and where only the position  of the penetration has been changed.
This  demonstrates  that a penetration closer to  the  suction point  affects a
smaller slab area than one farther away.  The penetration near a suction point

     0  2  4  6  8  10  12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42  44 46 48

                               Slab Length,  ft.
     0  2  4  6  8  10  12 14 16  18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48
                               Slab  Length,  ft.
    Figure 4.  Effects of slab  penetrations on the pressure  field:   (a)
               A 12 x 36 in.  (0.3  x 0.9 m) penetration close to the
               suction point,   (b)   A 12 x 36 in.  (0.3 x  0.9 m) pene-
               tration far  from the suction point.

nay decrease  the  pressure  field below that required  to  overcome the negative
house pressure only in the  immediate vicinity  of  the penetration.  The pressure
field quickly recovers in this case and develops nearly unperturbed except for
that small area.  When the  penetration is  farther from the suction point, where
the pressure gradients are  smaller,  a much larger slab area is affected.  Under
these conditions, the lower limit of the pressure fields are greatly perturbed,
and can greatly affect the area under the slab where sub-slab depressurization
is effective.

      These results  indicate  that,  unless cracks near a  suction point are so
extensive that the applied pressure cannot be maintained, it is not worth a large
expenditure of effort to seal them.   Conversely, cracks  near the perimeter of
the pressure field should be sealed if the sub-slab depressurization system is
required to maintain its effectiveness at distances  beyond that from the suction

      A  computer   simulation  that  describes  the   operation   of  sub-slab
depressurization systems has been developed by modeling the physical processes
that  occur  when  this  mitigation  technique  is  applied  to  slab-on-grade
construction.  The model has been verified by extensive pressure field mapping
at  the  University  of  Florida  Test   Slab   Site  and  has  demonstrated  its
effectiveness in  the  design of sub-slab depressurization  systems  for several
radon mitigation research/demonstration houses. The physical processes predicted
by the model make it a useful tool for the design of these systems, and permit
it to make valuable predictions concerning  the performance  of these systems for
a variety of physical conditions.  These simulations permit  accurate evaluations
of various sub-slab depressurization configurations  that can be performed without
the  time  and  expense  associated  with the  construction  of  elaborate  test
facilities and extensive instrumentation.

1.    Bear, J.  The Fundamental Fluid Transport  Equations  in Porous Media,  in:
      Dynamics of Fluids in Porous Media.  Dover Publications, New York, 1988.

2.    Furman,  R.A.  and Hintenlang,  D.E.   Sub-Slab Pressure  Field  Extension
      Studies  on  Four Test   Slabs  Typical  of  Florida   Construction.    In1.
      Proceedings  of  the  1990  International  Symposium  on  Radon and  Radon
      Reduction Technology, Atlanta, Georgia, to be published.

3.    Harris,  D.B.,  Henschel,  D.B.,  Sanchez,  D.C.  and  Witter,  K.W.   Field
      Measurements in the EPA/AEERL Radon R&D Program.  In:  Proceedings: The 1988
      Symposium on Radon and  Radon  Reduction Technology,  Volume 2, EPA-600/9-
      89-0066 (NTIS PB86-167498),  March 1989.

                  by:   Terry Brennan
                        Camroden Associates
                        Oriskany, NY    13440

      Some Investigators use a vacuum cleaner to temporarily induce a low air
pressure field beneath concrete slabs.  This test is performed to gain insight
into the prospects for soil depressurization as a radon control method.  A
protocol is suggested that can provide a basis for deciding whether sub-slab
depressurization or pressurization is likely to succeed or fail.   Enough data
can be collected to select a fan with appropriate performance characteristics.
Examples are drawn from field work to illustrate the collection and
interpretation of the data.

Lacking X-ray vision, a vacuum suction test can be done to explore
strength and size that a low air pressure field can be extended beneath a
floor slab. How far the pressure field is extended and at what strength
depends two things:

  • how resistant to airflow is the material immediately beneath the slab
  • how tight the surrounding soil and covering foundation are

The results of this test can yield insight into both of these important
variables. First a vacuum cleaner is used to apply suction to a hole through
the slab, called FA by convention. The amount of suction created by the
vacuum pulling on the 'system * beneath the slab is measured in the
vacuum nozzle (FA), at a hole 10 inchs away (FB, again by convention) and
at other test holes (FC, FD etc.) several feet away (usually 10 to 30 feet). If
the material beneath the slab is loose (say stone pebbles) and  foundation
and surrounding soil are fairly resistant to airflow, then a low pressure can
be induced a great distance from the suction point. If there is loose
material beneath the slab and there are large airleaks through the
foundation  or the surrounding soil or bedrock, then a weak pressure field
will be induced and will extend only as far as the leaks (eg. French  drain or
shattered bedrock).  If there is fine but porous material beneath the  slab
then a strong vacuum will be induced at the suction point (FA) but  it will be
defeated by the resistance to airflow of the fine material and not extend
very far from the suction point. See Figure 7 for a conceptual model and
illustration of these ideas.
           Figure 1 - The Vacuum Suction Test
                                    Pressure Differential Gauges

Figure 2 - Measure the vacuum in the closed nozzle (zero airflow) and
open nozzle (maximum airflow). This establishes a rough guide to the
amount of air the vacuum can pull from under the slab, which can be
used with Figure 3 to select the fan type. Alternately the airflow could
actually be measured.
                                                Cover the vacuum nozzle
                                                with your hand. The airflow
                                                thru the vac is zero. The
                                                amount of suction measured
                                                in the nozzle is the most the
                                                vacuum can pull. For a
                                                Eureka Mighty Mite with a
                                                clean bag this is about 40
                                                inchs WC.
                                                Measure the suction in the
                                                nozzle with the vacuum
                                                nozzle in open air. This is the
                                                lowest amount of suction the
                                                vacuum will pull because it
                                                is the minimum resistance to
                                                airflow it can have. For a
                                                Eureka Mighty Mite with a
                                                clean bag this is about 5
                                                inchs WC.
    Figure 3 - Select Fan Type
Suction at
Fan Type
>1.5 Inchs WC
In Line Centrifugal

          Figure 4 - Fan and "System " Curves
          In Line
200 T  /S**1*1^
                         Stone Pebbles v
                                   Radial Fan
                                          W*  I
                                             Vortex Fan
                                           *— Fine sand
                                              or silt
                                   4      5
                 AP (Inchs WC)
The four "system" curves are measured field data that appear
"typical" for the conditions listed. As with all things radon it is
likely that there are many exceptions to these "typical" curves.
Notice that for different sub-slab conditions the in line
centrifugal fans normally used for soil depressurization systems
is probably not appropriate. Both the coarse sand and fine sand
and silt curves!ntersect the in line centrifugal fan curve below
the manufacturers suggested minimum airflow.

The power consumed by an air handler depends on how much air it
is moving. This table shows the  measured Wattage range of the
fans whose curves are shown in the graph.

               In Line Centrifugal 63-83 Watts
                Radial 87-245 Watts
                Vortex 67-132 Watts

Figure 5 - Select Number of Suction Points
Suction at
| FC, F D etc.
Probable Sub
Slab Condition
Number of
Suction Points
Low-Medium   Good everywhere      (stone pebbles)
               Good around
               slab edge
                     (sand with
                     drain tile)
                    1 anywhere

                    1 anywhere
                    at slab edge
   High        Drops off quickly
               (6 to 10 feet)
Medium-High   Drops off quickly
               (10 to  15 feet)
                     (clay or silt)

                     (coarse sand)
                    1/300 square feet

                    1/600 square feet
Drops off quickly
(8 to 15 feet) or
large leaks under
the slab through
the foundation or
underlying soil or
(Shattered shale
or limestone,
glacial outwash or
riverbed gravels)
beneath the slab
or more suction
points or
seal foundation
leaks if they are
large (french
drain, sump hole)
              Figure 6 - Select Pipe Diameter
               Suction at
             Pipe Diameter
                 High     Four inch
               Medium    Three Inch
                  Low    Two inch legs, three inch headers

   Figure 7 - A Conceptual Model for Soil Depressurization
xxx/x/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxx xxx t * t t f t t xRs . 3' t t t t t f
VVVVVVVVVVVVVN >>>>>/> .VVVVVVVV*>>>>>» »-W •.'.'•.'»»»?
                      'f 't 'f 'f 'f 'f ~f 't 'f ~f ~f "f ~f 't't't ~f "f "f 't't't't't't 't't't t t f
\x\\\\\\\\\\\\\\\\\s\\\\\\\\ \ \ \ \,v v>»>>»>>>>>>>^^^
If the material beneath the slab has very little resistance to airflow
compared to that of the surrounding soil or bedrock (Rs,a) and that of
the leaks through the foundation to the sub slab air (Rb.a), then the
amount of air and vacuum induced by the sub slab fan is a function of
these two variables and the fan performance curve.If the material
beneath the slab is fine enough so that the resistance to airflow
through a few feet of it is comparable to the other two (Rb.s and Rs.a)
then this model is too simple. The material beneath the slab would
have to be divided up into numerous "bits" and the resistance between
each bit the and the surrounding "bits", outside air and basement air
would need to be accounted for.This is the case when fine sands, silts
and clays are found directly beneath the slab and it has important
implications for both radon entry and the prevention of radon entry by
soil depressurization.


                     by:  Benny Majborn
                          Rise National Laboratory
                          DK-4000 Roskilde, Denmark

     Seasonal variations of indoor radon concentrations have been stud-
ied in a cluster of 10 single-family houses. Eight of the houses are
of a similar construction with slab-on-grade foundations. The remaining
two nouses have different substructures, one of them having a crawl
space, and the other having partly a basement and partly a crawl space.
A "normal" seasonal variation of the radon concentration with a maximum
in winter and a minimum in summer was observed in most of the houses. In
these houses the variation showed a strong correlation with the indoor-
outdoor temperature difference on a 2-month basis. However, deviating
seasonal variations were observed in some of the houses, notably in the
two houses having different substructures.

     A re-examination of the data obtained in a previous study indicates
that winter/summer ratios of indoor radon concentrations in Danish houses
depend on the house substructure. The mean winter/summer ratios were
about 2.1 for houses with slab-on-grade foundations, 1.5 for houses having
a basement, and 1.0 for houses with a crawl space (geometric mean values).
However, a study with more houses in each substructure category will be
needed to show whether or not the indicated differences are generally valid
for Danish houses.

     In this paper some results on seasonal variations of indoor radon
concentrations in Danish single-family houses are reported and discussed.
The results have been obtained in two different studies. One of these is
an extension of an investigation of factors influencing indoor radon
concentrations carried out in 1986-87 (1). The investigation comprised 16
houses built on the same site, and included integrating radon measurements

in the living-room and in a bedroom of all the houses on a two-month
basis throughout the two years. These measurements have been extended
through 1988 in ten of the houses. The other study is a pilot investiga-
tion of natural radiation in Danish houses carried out in 1983-84  (2). In
that work integrating radon measurements were made in 70 single-family
houses and 12 apartments during three months in winter and three months
in summer. The data on seasonal variations obtained in the pilot study
have been re-examined after obtaining additional information on the sub-
structure of the houses.

     Knowledge of temporal variations of indoor radon concentrations and
the variability of such variations is needed for a proper appraisal of
the significance of results obtained from short-term measurements and for
estimating the uncertainties involved in attempts to predict annual aver-
ages from measurements covering shorter time periods (3,4). Knowledge of
temporal variations and the identification of causes of such variations
can also contribute to improve our understanding of the mechanisms of
radon entry into houses (5,6).

     Integrating radon measurements were made with passive closed dose-
meters equipped with CR-39 track detectors. The dosemeter design and its
properties have been described elsewhere (2,1). The performance of the
dosemeter was tested in CEC radon dosimetry intercomparisons in 1984 (8),
1987 (9), and 1989, in all cases with good results.
                               HOUSE TYPES

     The 10 houses in which radon was measured on a two-month basis dur-
ing 1986-88 are situated next to Ris0 National Laboratory. The site of
the houses forms an area of about 150 m by 300 m, and the soil below the
houses is mainly moraine clay. The houses are one-storey single-family
houses built of bricks. Eight of the houses have slab-on-grade foundations
and differ only in length, having one bedroom more or less. The remaining
2 houses have different substructures, one having a crawl space and the
other having partly a basement and partly a crawl space. A special feature
is a district-heating duct which is extended into each house, where it
proceeds circumferentially along the outer walls as an integral part of
the foundation.

     The 70 single-family houses included in the pilot study in 1983-84
are distributed throughout Denmark. They were not chosen to be repre-
sentative for the Danish building stock as the purpose of the in-
vestigation was primarily to establish and test measurement procedures.1
^A representative survey was conducted in 1985-86 (10), but in that
work two six-month integration periods were used, and the seasonal
variation was not measured in each single house.

Out of the 70 single-family houses 25 had a basement, but no further
differentiation of the nouses according to substructure was made at the
time when the study was carried out. Recently, additional information on
the type of substructure has been acquired for most of the houses, so
that these may be grouped into the categories: 1) full basement, 2) base-
ment + crawl space, 3) basement + slab-on-grade, 4} crawl space, and 5)
no basement or crawl space, i.e. mostly slab-on-grade houses.

      Figure 1 shows the seasonal variations during 1986, 1987 and 1988
of the mean radon concentrations in seven of the Rise houses. The figure
shows the geometric mean value of the radon concentrations for each 2-
month period for all 14 rooms (7 living-rooms and 7 bedrooms). The geo-
metric standard deviations range from 1.6 to 2.1 for the 18 measurement
periods.  In the remaining 3 houses, including the two houses having a
different substructure, there was not a pronounced seasonal variation
              _ 300
               8 100
Figure 1.  Seasonal variations during 1966, 1987 and 1988 of 1) the
           geometric mean value of the radon concentrations in 7 houses,
           and 2) the average outdoor temperature.

of the radon concentration, except that the house with a crawl  space
appeared to snow a summer maximum in 1986. Figure 1 also shows  the
average outdoor temperature (2-month averages) at the nearby meteorologi-
cal station at Rise National Laboratory. The correlation coefficients
(linear regression) between outdoor temperature and indoor radon (geome-
tric mean value for 14 rooms) were found to be -0.97 for 1986,  -0.98 for
1987, and -0.91 for 1988. As the variations of the average indoor tempe-
rature are small compared with those of the outdoor temperature, the
results indicate a strong correlation on a 2-month basis between the
average indoor-outdoor temperature difference and the average indoor
radon concentration in this group of houses.

     In the pilot study integrating radon measurements were  made in a 3-
month winter (1 Dec. 1983 to 29 Feb. 1984) and a 3-month summer period
(22 May to 13 Aug. 1984). Figure 2 is a plot of the summer results  for
              •E  100
              § 300

                    0      100    200    300     &00
                  RADON CONCENTRATION (Bq/m3) WINTER

Figure 2.  Summer versus winter  radon  concentrations in 38 houses with-
           out basement or crawl space. The straight lines represent a
           slope = 1.

 each of the houses without a basement or crawl  space (i.e. mostly slab-
 on-grade houses)  versus the winter results for  the same houses.  Similarly,
 figure  3 shows  the summer versus winter results for the houses having a
 basement.  In Table 1  the winter and summer radon concentrations  and the
 winter/summer ratios  (geometric mean values and geometric standard devi-
 ations)  are given for 1)  houses with a basement, 2}  houses with  a crawl
 space,  and 3) houses  without a basement or crawl space. The  houses with
 a  full  basement or basement + crawl space have  been grouped  together,
 because the mean  radon concentrations and winter/summer ratios did not
 differ  significantly  between the two groups. Three houses having a base-
 ment +  slab-on-grade  foundation had mean radon  concentrations  in the
                        0   100  200  300  400
                   RADON CONCENTRATION (Bq/knJ) WINTER

Figure 3.  Summer versus winter radon concentrations in houses having
           a full basement (circles), basement + crawl space (squares),
           or basement + slab-on-grade foundation (triangles). The
           straight lines represent a slope = 1.

House type       Number        Radon concentration
   Roan            of          Winter        Summer
                 houses       GM    GSD     GM    GSD
                            (Bq/m3)       (Bq/m3)
                              GM    GSD
Basement*          16

Crawl space         3

No basement or
crawl space        38
1.46   1.7
1.57   1.7
1.32   1.6
* Include 11 houses with full basement and 5 houses with basement + crawl
living-rooms and bedrooms that were close to the mean values found for
the slab-on-grade houses. The results from these 3 houses have not been
included in Table 1. There are only 3 houses with a crawl space in this
sample. All 3 houses had winter/summer ratios close to unity in both
rooms (range 0.7 - 1.2). The average radon concentrations were about 20,
70, and 120 Bq/m3 in the 3 houses.

     Within each category of houses the mean winter/summer ratios are
not different between living-rooms and bedrooms (98% of the living-rooms
and 79% of the bedrooms were on the first floor). For the 16 houses hav-
ing a basement or basement + crawl space the lower mean winter/summer
ratio found for basement rooms (1.3) is not significantly different from
that found for living-rooms and bedrooms (1.5). The differences between
the mean winter/summer ratios found for living-rooms and bedrooms in 1)
houses without basement or crawl space (2.1), 2) houses with basement
(1.5), and 3) houses with crawl space (1.0) are significant at the 95%

                        DISCUSSION AND CONCLUSIONS

     Seasonal variations of  indoor radon concentrations may depend on
a number of factors,  including geological factors, climate, house cha-
racteristics and living habits. Figure 1 shows an example of the in-
fluence of climate. The figure shows the seasonal variations during
3 years of the mean radon concentrations in a group of houses with slab-
on-grade foundations. The variations show a strong correlation with the
indoor-outdoor temperature difference on a 2-month basis. A major dif-
ference between the 3 years was an unusually cold January-February in
1986 and 1987 and an  unusually mild January-February  in 1988. This dif-
ference is clearly reflected in the observed radon concentrations.

     The results.reported in this paper on winter/summer ratios indicate
that houses with slab-on-grade foundations have higher mean winter/summer
ratios than houses with a basement or crawl space. Similar observations
have been reported in a Finnish study (11), where measured seasonal
variations of indoor radon concentrations were compared with model cal-
culations. Data from the USA suggest that winter/summer ratios tend to be
greater for the first floor than for the basement of houses (4). The same
tendency was observed in the present work. It should be noted, however,
that the results reported in this paper are based on a limited number
of houses. A study with more nouses in each substructure category will
be needed to show whether or not the indicated differences are generally
valid for Danish houses.

             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.

     This paper is partly based on two studies that were partly funded by
the Commission of the European Communities. The author would like to
thank S.P. Nielsen for valuable discussions.


1.    Majborn, B., Serensen, A., Nielsen, S.P., and Better-Jensen, L.
      An investigation of factors influencing indoor radon concentra-
      tions. Riso-M-2689, Ris0 National Laboratory, Roskilde, Denmark,
      1988. 58 pp.

2.    Sorensen, A., B0tter-Jensen, L., Majborn, B., and Nielsen, S.P.
      A pilot investigation of natural radiation in Danish houses. Rise
      -M-2483, Risa National Laboratory, Roskilde, Denmark, 1985. 40 pp.

3.    Swedjemark, G.A. Radon and radon daughters indoors, problems in
      the determination of the annual average. SSI Scientific Report
      84-11-12, National Institute of Radiation Protection, Stockholm,
      Sweden, 1984. 47 pp.

4.    Bierma, T.J., Croke, K.G., and Swartzman, D. Accuracy and preci-
      sion of hone radon monitoring and the effectiveness of EPA moni-
      toring guidelines. JAPCA 39: 953, 1989.

5.    Nazaroff, W.W., Moed, B.A., and Sextro, R.G. Soil as a source of
      indoor radon: Generation, migration and entry. In; W.W. Nazaroff
      and A.V. Nero, Jr. (ed.), Radon and its decay products in indoor
      air. J. Wiley & Sons, New York, ISBN 0-417-62810-7, 1988, p. 57.

6.    Arvela, H., and Winqvist, K. Influence of source type and air
      exchange on variations of indoor radon concentration. STUK-A51,
      Finnish Centre for Radiation and Nuclear Safety, Helsinki, Finland,
      1986. 32 pp.

7.    Majborn, B. Measurements of radon in dwellings with CR-39 track
      detectors. Nuclear Tracks 12: 763, 1986.

8.    Miles, J.C.H., and Sinnaeve, J. Results of the second CEC inter-
      comparison of active and passive dosemeters for the measurement of
      radon and radon decay products. EUR 10403 EN, Commission of the
      European Communities, Brussels, Belgium, 1986. 64 pp.

9.    Miles, J.C.H., and Sinnaeve, J. Results of the third CEC inter-
      conparison of active and passive detectors for the measurement
      of radon and radon decay products. EUR 11882 EN, Commission of
      the European Ccmrounities, Brussels, Belgium, 1988. 80 pp.

10.   Ulbak, K., Stenum, B., S0rensen, A., Majborn, B., Better-Jensen,
      L., and Nielsen, S.P. Results from the Danish indoor radiation
      survey, Radiat. Prot. Dosim. 24: 401, 1988.

11.   Arvela, H., \foutilainen, A., Makelainen, I, Castren, p., and
      Winqvist, K. Comparison of predicted and measured variations of
      indoor radon concentration. Radiat. Prot. Dosim. 24:231, 1988.

                  by:   Craig P. Wray, P.Eng. and
                        Grenville K. Yuill, Ph.D., P.Eng.
                        G.K. Yuill and Associates Ltd.
                        Winnipeg, Manitoba   R3T 2C4

      A microcomputer program has been developed so systematic radon and radon
progeny  control  techniques  of optimal  effectiveness can  be  formulated  for
designing or  retrofitting  houses.  An existing  multizone airflow/contaminant
dispersal analysis  computer program was modified  by adding a  model  of radon
progeny plate-out on indoor surfaces.

      Three simulation  exercises  of  radon  and radon  progeny  levels   in  the
rooms of a hypothetical  house  for two different  HVAC systems were carried out
using the new program.   Comparisons of these  parameters  were  made with those
for the same house with only natural ventilation.  The simulations showed that
significant differences  in  average radon  levels  and total  EEDCs  can occur
between  rooms of  a house.    These differences demonstrate the  need  for  a
multizone model  and indicate  caution should  be used  in applying  radon  and
radon progeny level measurements  taken  in one  room to  any  other room  in  a
house.   The  simulation  exercises  also  showed that  radon  and  radon  progeny
levels in  a  house  strongly  depend  on the  type  of ventilation  system  in  the

       In a large fraction of Canadian houses, radon concentrations Indoors are
 significantly   higher   than  those   outdoors.     These  elevated  levels  can
 compromise the  health  of occupants who are exposed to them over long periods
 of  time.   Although several mitigation  techniques have been  used  recently to
 reduce  these  concentrations   to  acceptable  levels,   the  reasons  for  their
 effectiveness  are  not  well  understood, and no  systematic approach  has been
 established for implementing an appropriate procedure  in each situation.

      The  behavior of  radon   and  its  progeny  in  houses  is  too complex  to
 predict without simulation.  Although field  monitoring could  be used to study
 radon and radon progeny transport, this approach is prohibitively expensive if
 even  a  few common combinations of house geometry, ventilation system types,
 and soil characteristics are to be examined.

      As a step towards  the improvement of existing mitigation techniques and
 the  formulation of  mitigation  strategies,  it  is desirable  to  synthesize
 previous  research  about  radon   and radon  progeny   transport  with  current
 multizone  modelling  techniques  to  produce  a  model   that  can  predict  the
 accumulation of these pollutants  in  a house on a  time-varying  basis.   In the
 work  described here,   an  existing   multizone   airflow/pollutant   analysis
 microcomputer  program  was modified.    Parametric  studies using the  modified
 program  (CONAIR 89-2)  can  be carried  out  in  the  future to  gain   a  better
 understanding  of  radon transport,  so  systematic radon  and  radon  progeny
 control  techniques of  optimal effectiveness  can  be  developed  for new  and
 existing houses.

                   CONAIR 89-1:  AN EXISTING SIMULATION TOOL
      CONAIR  89-1  is  an  existing  multizone  airflow/contaminant  dispersal
network  analysis  microcomputer program  that has  most,  but  not  all,  of  the
capabilities required for  the  work described here.  It was  developed by G.K.
Yuill and Associates  Ltd.  using two computer programs (AIRNET  and  CONTAM  87)
obtained from the U.S.  National  Bureau of Standards as its  core  (1).  CONAIR
89-1  begins by  modelling  airflows between  zones  and  between  indoors  and
outdoors in a multizone  building.   It  simulates wind and  stack  effects  on
envelope leaks, friction  in  ducts,  fan characteristics and  two-way buoyancy-
driven flow through large openings such as doors.   Calculations are done using
hourly  mass-balances  on  a  macroscopic  basis  assuming   well-mixed  zones.
Recirculation  within   a   single  room   is   not   simulated.     Steady-state
calculations are  done for each set  of hourly operating conditions supplied,
such as weather data,  fan  settings,  or damper settings.  The  program ignores
subhourly pressure transients, because their time  constants  are typically  too
short to be of significance in this kind of calculation.

      CONAIR 89-1  also  has the capability to model  pollutant concentrations.
This part of the program uses the airflows determined in the first part, along
with source strengths and  reaction  and  decay  rates  for the various pollutants
present, to predict  the pollutant  concentrations  in each zone.   The program
does a  steady-state  analysis,  but  can  also  do  a dynamic  analysis  to predict
time-varying concentrations.

                        RADON DAUGHTER PLATE-OUT MODEL
      CONAIR 89-1 assumes all  pollutants are  airborne  and  thus are subject to
transport in ventilation airflows.   However,  this  assumption is inappropriate
for  some  pollutants  such  as  radon  progeny  (218Po, 214Pb,  and  214Bi),  which
plate-out on  surfaces within  a  house  such as walls or furniture.   With the
exception of  218Po,  once these progeny are deposited on these surfaces,  they
do not  produce significant  numbers  of progeny that return to the  air where
they can be transported from room to room  or  from  indoors  to outdoors.   Thus,
surface-deposited 214Pb and  214Bi  remain  trapped  in  the  particular room in
which  they  plated-out.   The  concentrations  of  these two surface-deposited
radon  daughters  are  of no   significance from a health standpoint therefore,
because they  cannot be inhaled  once  they  plate-out.   However,    214Pb  formed
from decaying  surface-deposited  218Po  can  return  to the air  through  a  recoil
mechanism,  so  CONAIR  89-1   needed  upgrading  to  model surface-deposition of
     with recoil.
      CONAIR 89-1  uses  air and pollutant mass-balances to  determine  the mass
transport rate of  up  to eight different species through  simple  flow elements
with  or without  filters.   The  flow through  an  element  is  assumed  to  be
instantaneous and well-mixed.  Axley  (2) defines  the transport of a pollutant
species a  between  two nodes by flow  element  equations from node  i  to node j
due to an air mass flow rate we(t)  as  follows:

     M = W(t)°
       Since  CONAIR 89-1 assumes  all  species are airborne,  equations 1 and  2
 state  that  all  species  are transported  into  element  e  where they  can  be
 removed  to some degree by a  filter.   Thus,  to account for  surface-deposition
 of  a  particular  species at  a  node,  it is  necessary  to  eliminate  the  mass
 transport  of that  species into  element e.  This  blocking of airborne  transport
 in  CONAIR 89-2 was  achieved  by defining a  surface-deposition  coefficient «s
 that affects the species mass transport  rates as  follows:



       For  airborne transport  of a particular species, as  is set to  unity,  so
 it  has no  effect on  the mass  transport rate  of that species.  For a particular
 species  that is surface-deposited, xs  is  set to zero,  so the  mass  transport
 by  ventilation of  that species  through all flow elements connecting every  node
 is  eliminated.

       The  nature   of  each   pollutant  (airborne  or  surface-deposited)   is
 identified   in  the  CONAIR  89-2   contaminant  input  file  by  specifying   the
 pollutant  type  as  either type "A" (airborne, «s  =  1.0)  or type "S"  (surface-
 deposited, "s = 0.0).



       In the  Jacobi  model (3),  the sources of radon  in  the indoor air are  the
 entry  of  radon  in  soil  gas infiltrating into the basement through foundation
 penetrations  and in  the  outdoor air  infiltrating through above-grade leaks in
 the building  envelope.  The removal of radon and the removal and production of
 radon  progeny  in  the  Jacobi  model  by   radioactive   decay,  attachment,
 deposition,  and recoil  can  be  described   by  a  set of   linear  first-order
 differential  equations for each room.   CONAIR 89-1  already has  the capability
 to  model  these  processes (2),  with  the exception of  surface-deposition that
 involves recoil.   Therefore,  it was  only necessary  to define the differential
 equations  and to  select a  value for  each  of  the rate  constants  in  these
 equations for each room.

      The Jacobi model  can be  simplified  for the  purposes of  this  project.
There  is no need to  determine the  concentrations  of surface-deposited 214Pb or
 214Bi,  because these species do not  pose a  health risk once  they  are removed
 from the  air by the plate-out process.   The probability they will produce
recoiling atoms once deposited  is negligible.   This  leaves eight  species to
consider: airborne-unattached ^Rn  (URn2),  airborne-unattached 218Po (UPo8),
airborne-unattached  21«Pb (UPb4),   airborne-unattached  214Bi  (UBi4),  airborne-
attached 218Po (APo8), airborne-attached 214Pb (APb4),  airborne-attached 214Bi
 (ABi4), and surface-deposited 218Po (SPo8).

      The  differential  equations  describing  the  radon  and  radon  progeny
kinetics indoors (excluding ventilation) for any given room are:








=  -A0[URn2]                                                  (5)

=  Ao [URn2] - (Xa + A3 + A,) [UPo8]                               (6)

=  A, [UP08] - (Xa + A3 + A2) [UPb4]

   + Pi A1 [APo8]  + Po A1 [SPo8]                                 (7)

=  A2 [UPb4] - (\a + A3 + A3) [UBi4]                                (8)

=  Aa[UPo8]-(\i+AS)[APo8]                                   (9)

=  (1-Pi)Ai[APo8]+Aa[UPb4]-(x3+A2)[APb4]                 (10)

=  A2 [APb4] + Aa [UBi4] - (\3 + A3) [ABi4]                         (11)

=  A3 [UP08] + A3  [AP08] -Ai [SP08]                             (12)

     radioactive decay constant of 222Rn,  h-1.

     radioactive decay constant of 218Po,  h-1.

     radioactive decay constant of 214Pb,  Ir1.

     radioactive decay constant of 214Bi,  h-1.

     attachment  rate  of  free species onto aerosols, h-1.

     surface-deposition  rate of unattached species, h-1.

     surface-deposition  rate of species attached to
     aerosols, h~1.

     recoil probability  of 214Pb from aerosols,  d'less.

     recoil probability  of 214Pb from surfaces  in  the  room,

     concentration  of species X in the room, nuclei/gair.
      Based on the eight differential  equations  listed  here,  a  set  of five  key
parameters  that  characterizes  radon  and  radon  progeny  kinetics  must   be
specified in CONAIR 89-2 for each  room.   These parameters  are:   the attachment
rate, the unattached species deposition rate, the attached species deposition
rate,  the  recoil  probability  of   214Pb  from  aerosols,  and   the   recoil
probability of 214Pb from room surfaces.

       The  radioactive transmutation coefficients  (A0, AI, A 2, and AS) used  in
 the  Jacobi model  are not  unique  to each  room.    Instead,  they are  physical
 constants  related  to the  isotope half -lives.

       Porstendorfer (4,5),  Bruno  (6),  and Nazaroff and  Nero (7) report  that
 the  attachment  rate of radon progeny onto aerosols can vary  widely  indoors  (4
 to  2000  h'1).   The  attachment rate  strongly  depends  on   the  aerosol  size
 distribution  and aerosol  concentration.   Maximum attachment rates  correspond
 to  particle  sizes  in the  range  of 0.1  to 0.2  fjm  (7)  and  to high aerosol
 concentrations.   These  concentrations vary  from  2  to  500,000  particles/cm3
 depending  on  the amount of  cooking and  smoking and  on  the effectiveness  of air
 filtration systems,   if  any are  used.    Cooking  and  smoking  can  increase
 aerosol concentrations  indoors by  two  to  three orders of magnitude  (7), which
 in  some  cases can  result  in significant  variations  in  attachment rates  from
 room  to room  in  a house.

       An  attachment rate  of  50 h-1 was used indoors.   This  rate was  selected
 from  values listed in the  literature  for clean  air  (approximately 10,000  to
 20,000 aerosol  particles/cm3).  Unfortunately,  reported attachment rates  tend
 to  list  only the  corresponding aerosol  concentration,  but  not the  particle
 size  distributions,  so  it  is difficult  to  determine  a  typical  value  for
 typical conditions.   The  attachment rate  used here was assumed to be  constant
 everywhere in the  house for  simplicity in demonstrating the new models.    An
 attachment rate  of 40 fr1 was used outdoors based  on  Jacobi  (3).  This lower
 rate  recognizes  that  although the  aerosol  concentrations outdoors usually are
 higher than  those  indoors  when there is  no cooking  or smoking  (7),  these
 activities  generally would  be expected to  result  in  higher average aerosol
 concentrations indoors.

       CONAIR  89-2,  which  implements   the  modified   flow  element  equations
 described  earlier,  can  model  the surface-deposition  of radon-daughters  by
 specifying  the  unattached   and  attached  species  deposition   rates.    The
 deposition  rate  of  a particular  unattached species  in a   specific  room  is
 dependent  on  the area of exposed surfaces  in that room,  on  the volume of that
 room,  and on the deposition velocity of the particular species as follows:

            Ad     =    surface deposition rate of unattached
                        species in a given room, h'1.

            i/d     =    average deposition velocity of free (unattached)
                        species in the room, m/h.

            S      =    total area  of enclosure surfaces in  the  room exposed
                        to air  (including walls, doors, windows,  floors,  and
                        ceilings,  but   excluding   other  surfaces   such   as
                        furniture), m2.

             F     =     correction factor  to account  for additional  exposed
                         area of  furniture  or  other  surfaces  in  the  room,

             V     =     volume  of air in the room,  m3.

       This rate expression assumes all  surfaces  exposed to air in  a given room
 are equally effective  in removing any  radon daughter through  the use  of  an
 average deposition velocity that is the same for all radon progeny.   Although
 an analogous equation can  be applied to attached  species,  this was  not done
 because the deposition  velocities of  attached species are  usually  much smaller
 than  those  for unattached  species.   Nazaroff  and  Nero  (7)  report  that  the
 ratio  of unattached  to  attached deposition  velocities is approximately 100.

       Deposition velocities are strongly influenced  by air motion.   Equation
 13 assumes there  is  sufficient  air motion  in each room  so the radon  progeny
 concentration  in  the room  is uniform.   However,  most deposition  velocities
 found  in the literature  are not accompanied  by quantitative descriptions  of
 the air movement conditions during  the deposition velocity experiments.   Bruno
 (6) reports a  deposition  velocity for  unattached  218Po in still  air of 0.54
 m/h, which  is not typical  of real  houses.   He  also  reports that  the deposition
 velocity for unattached 218Po in  rooms with low  ventilation is 7 m/h  and with
 moderate ventilation  is  22  m/h.   Bigu (8) lists  values  of  2-19 m/h for a 26 m3
 chamber equipped  with  a circulating  fan.    The higher value corresponds  to
 operation of the  fan, while the  lower  value corresponds  to  periods when  the
 fan was off.  The circulation rate of the fan is not provided by Bigu.   Scott
 (9) reports unattached deposition  velocities  in the range of 3.6 to 18  m/h,
 with the higher values corresponding  to  higher ventilation rates.

       Unattached deposition velocities  of 8 m/h in rooms  and 20 m/h  in  ducts
 were assumed.   The value  for rooms was  selected  as  an average value (7).- The
 value  for ducting was  selected  in recognition  of their  high airflow rates.
 This  value  may  be   an  underestimate,   since it  is  based  on  maximum  room
 deposition  velocities,  but  there does not appear  to  be  any   literature
 available on radon progeny  deposition rates  in ducting.

       Room  surface   areas  and   volumes   as  required  for equation  13  were
 estimated from  building plans.  A correction factor of 2.5 (10) was  applied to
 increase  the surface area  exposed to air  in each  room,  except  in  hallways
 where  furnishings tend  to be sparse.   A factor  of 1.0 was used in hallways.
 Porstendo'rfer  (10) does  not quantify room or  furnishing surface  areas,  so
 these  factors  are somewhat  arbitrary.     The deposition  rate  for  attached
 progeny  in  a given room was assumed to  be  1/100 of the unattached deposition
 rate in that room.

      The unattached  deposition velocity outdoors was assumed to be 17 m/h so
 the outdoor  equilibrium  fraction  would  be  0.56  (7).    Therefore,  the outdoor
 value  has  little direct   basis  on  real   conditions.     It   is  assumed  all
deposition  rates  described  here  are  constant  over  time  for  simplicity in
demonstrating the new models.

       The  literature  reviewed  here  does  not  differentiate  between  radon
 progeny  when specifying deposition  rates or attachment  rates,  so the  Jacobi
 model  described here assumes the  same  attachment rate or the same  deposition
 rate  can be  applied to each of  the  three radon progeny considered.   Research
 should be carried  out  in a  future  project  to determine the  impact of  this
 assumption on CONAIR's predictions of radon  progeny  concentrations.

       When  atoms of attached or surface-deposited 218Po  decay,  a fraction  of
 the 214Pb atoms  produced can become  detached and move into the room  air.   The
 recoil probability  describes  the chance this process will occur for  any given
 atom.  The maximum  probability  is  1.0.   In this case, all  of these 214Pb atoms
 are released into the room air.  The recoil  probability depends on whether the
 218Po  is attached to an aerosol  or to a surface in the room.

       For  218Po  attached  to aerosols,  a  range  of 214Pb  recoil  probabilities
 from  0.4 to  0.83 has been  reported (11, 12).  The higher recoil probabilities
 are   expected   to  correspond  with  smaller  aerosol   sizes  (12).     Field
 investigations  to determine this  recoil  probability in  the  real  house  were
 beyond the scope of this project,  so a  typical  indoor value of 0.5 (3,  5)  was
 assumed  for  the  case considered  here.

      A  different  recoil  probability  of  0.83 was  used outdoors   for 214Pb
 produced  by 218Po  attached  to  aerosols.   This  value  was  selected  so  the
 outdoor equilibrium  fraction would be 0.56 (7).

       For  surface-deposited  218Po,  21«Pb  recoil probabilities   tend  to  be
 lower.  Due  to the presence of a laminar boundary layer on indoor surfaces,  it
 is unlikely  all  recoiling 214Pb atoms will escape into the room air.   Instead,
 a  significant  fraction (as much as  50%,  (6))  will   escape  into  the  boundary
 layer,  from which  they will  redeposit  on  the surface  they  escaped  from.
 Typically, this  recoil probability is estimated to be 0.25 (7, 13).    For  lack
 of any other evidence,  this typical  value has  been  used  indoors and  outdoors
 in the case  considered here.
                      EXAMPLE ANALYSES USING CONAIR 89-2


      Three  simulation  exercises  were  carried  out  using  CONAIR 89-2.   The
purpose of these exercises was to demonstrate the capabilities of the modified
program through  analyses of  the  radon  levels  and  EEDCs  in  the rooms  of a
hypothetical house for two different  HVAC  systems,  and through comparisons of
these parameters with those for  the  same house with only natural ventilation.
Hypothetical   soil   properties   and   radon   progeny   transport   phenomena
characteristics used in these simulations were based on ranges of values cited
in  the  literature  and  not  on  field trials,  because the  purpose of these
simulations did not justify the  expense  of an Intensive measurement effort to
determine these values.


      A hypothetical  house  based on a  real  single-storey house located  in  a
suburban  area  of  Winnipeg,   Manitoba  was  simulated.    The  house  has  an
Equivalent Leakage  Area (ELA) of  1540  cm*.   The  total  surface area  indoors
exposed to  air  (subject  to  radon  progeny plate-out)  is 1769  m2,  including
furnishings.  The house was divided into the following eight zones:  basement;
kitchen/dining  room/living  room; hallway  joining  living room,  bathroom,  and
bedrooms; bathroom  (sink  area);  bathroom  (tub area); master  bedroom;  bedroom
2-  and  bedroom  3.    The  kitchen,  living  room,  and  dining  room  were  lumped
together as one zone because there are no  significant flow resistances between
these regions.

      Using  the  total  ELA  of  the  house and  assumptions  of leakage  area
distributions,  inputs  were  developed  for CONAIR  89-2  to   characterize  the
magnitude and  location  of unintentional leaks in the building  envelope.   The
only  leaks  considered  between   zones  were  interior   doorways,  which  were
simulated as if the doors were wide open.

      The soil  surrounding  the hypothetical  house was a wet  sandy silt,  with
coarse sand  backfill  around the  walls and beneath the floor  slab.  The total
ELA of  the  combination of soil  leakage and  the  5 mm wide crack at the wall-
floor interface was 4 cm2 or 0.26% of the total  ELA of the house.

      The hypothetical  house  is  heated  electrically with baseboard heaters and
has no  mechanical  ventilation  system.   All  windows and exterior doors  were
simulated in  their  closed position, so the only  source  of  ventilation in the
house  was natural  infiltration  and exfiltration  driven by wind  and stack
effects through  leaks in the  house envelope.


      A second hypothetical house was simulated with characteristics identical
to  those  described  for the naturally-ventilated  house,  but  with two changes.
The baseboard  heating system was replaced with  a central electric forced-air
furnace,  and  a balanced heat  recovery ventilator (HRV)  system was added.   The
supply airflow rates  from the furnace to each zone were  sized to meet the same
design  heating load as the baseboard heaters.   Return  airflows for each zone
were  also specified.    Indoor air was  exhausted  continuously through the HRV
from  the  kitchen/dining room/living room  area at a  rate of 30.6 L/s and from
the bathroom tub area at  a  rate  of  14.2 L/s.

      An  extra  zone  was simulated  in  this  case to represent the estimated
combined  volume of  the  added ducting and  furnace (5 m3).  The  surface  area of
these components exposed to the air was estimated to be 60 m2.  The plate-put
and  attachment  rates  for  the  combined   ducting  and furnace  were described
earlier,  along with appropriate  214Pb recoil  probabilities.


       A third hypothetical house was  simulated with  characteristics  identical
 to those described  for  the naturally-ventilated  house, but with  two  changes.
 An LAFSW system with damper-controlled air inlets on  the windows in the living
 room and in  the  three bedrooms (14) was  added,  along with a central  exhaust
 fan in  the basement.   The central  exhaust  fan was  connected through  short
 pieces of  ducting  to the kitchen/dining  room/living  room  area  and  to  the
 bathroom tub area.   Exhaust airflow rates  identical  to those of  the HRV  were

       Each  air inlet had an ELA of  23.6 cm2 with the  damper in  the fully-open
 position.   These  inlets  were controlled on  a diurnal  schedule  corresponding to
 normal  residential activity.    From  8 a.m.  to 11 p.m.,  the air inlets  in  the
 living room were  fully  open,  while those  in  the bedrooms were fully  closed.
 From 11  p.m. to 8 a.m.  when people  would  occupy  the  bedrooms, the air  inlets
 in the bedrooms were fully open,  while those in the  living  room were  fully

      The  three  simulations  were carried  out  using  hourly  Winnipeg weather
 data  from  a four-day period in August 1981.   In  all  cases,  every indoor zone
 had a constant  temperature  of 21° C.  The average outdoor dry-bulb temperature
 during this period was 20.3° C.  The  average wind speed was  3.3  km/h.   Real
 wind direction  data were  not  available,  so the wind was simulated by rotating
 it around  the house on a  59 hour cycle.

      The   infiltration   of  radon   in  outdoor  air  can  be  a  significant
 contribution to typical indoor  levels, even though it is negligible at higher
 indoor levels.   Radon concentrations in outdoor air  are  usually in the range
 of 0.1 to  0.4  pCi/L (7).    A  typical  value of 0.2  pCi/L  (15)  was  assumed for
 the cases  simulated in this project.   A  constant  equilibrium fraction of 0.56
 was assumed outdoors (7).  The total  EEDC  outdoors based  on this equilibrium
 fraction was 0.11 pCi/L.


      Tables 1  through  3  summarize the  CONAIR  89-2 predictions of  radon  and
 radon progeny  concentrations  by listing  zonal  and whole-house  average  radon
 levels,   unattached  EEDCs,  attached  EEDCs,  total  EEDCs,   and  equilibrium
 fractions  for  the  three  cases considered  here.    Whole-house  averages  were
calculated as the  sum of  the  zonal  averages, weighted by the  fraction  of  the
total  volume of the house in  each zone.   To  understand how  the  two different
ventilation systems affect radon and radon progeny concentrations in different
areas  of  the  house,  consider the  basement,  living  room  and  one  bedroom

(Bedroom 2) separately.  Bedroom 2 was selected for illustration purposes,  but
either of the other two bedrooms could have been used instead,  since all  three
bedrooms have similar radon and radon progeny levels.

      A particularly interesting set of comparisons is that of the living room
to basement average  radon  concentration  and total  EEDC ratios.   These ratios
provide an  insight  into the inaccuracies  that  can result from  measuring  the
radon level or total EEDC  in one  region of the  house such  as  the basement  and
attempting to apply  the same results to other  regions of the house,  such  as
the living room.

      For the naturally-ventilated house the average radon level  in the living
room was only 50% of that in the basement,  while the average total EEDC in  the
living room was 48% of  that  in  the  basement.   The  average  radon  level  and  the
average total EEDC in the  living  room of the HRV-ventilated  house were closer
to those  in the  basement  (ratios  of  68%  and  71% respectively) than  in  the
naturally-ventilated house.  These reduced differences between the living room
and basement  averages were  due to the effects of  air  mixing caused  by  the
furnace air-handling  system in the house  with the  HRV.   Table  2  shows this
mixing  effect  in  particular  through the  lack of  variation  in main  floor
average zonal radon  level  and  total EEDC predictions.   For the house equipped
with the LAFSW  system,  the average radon  level and  the average  total  EEDC  in
the living room were further from those in the basement (ratios of 37% and  37%
each)  than   in   the   naturally-ventilated  house.     The   LAFSW  system  was
intentionally  designed  to  have   this effect.    By  providing  outdoor  air
specifically  only to  occupied regions  that require  ventilation,  the  LAFSW
system  saves  the energy an  HRV system wastes  by  not ventilating  the entire
house when all of the house is not occupied.

      These  observations  indicate  that  significant  differences in  average
radon levels  and  total  EEDCs could occur  between  basements  and  living rooms,
so caution  should be  exercised  in applying measured  radon  level  or total EEDC
data from one region to another.

      The  advantage  of  the LAFSW  system   in  providing  local ventilation  is
emphasized  by the average  radon levels and total  EEDCs in the  bedrooms.   In
bedroom 2 of  the  LAFSW-ventilated house, the average radon level was slightly
below that  outdoors  (0.20  pCi/L)  because  of  the  combined effects  of direct
ventilation with only outdoor air and the radioactive decay process.  The HRV-
ventilated  bedroom had  a lower average radon level  (0.34  pCi/L)  than  that  in
the naturally-ventilated bedroom  (0.71 pCi/L), but it was higher than that for
the LAFSW-ventilated bedroom due  to  mixing with  other  regions  of  the house
caused by the furnace  air-handling  system.   Bedroom 2 of the LAFSW-ventilated
house had  an average  total EEDC  of 0.08  pCi/L,  which  was  lower  than that
outdoors (0.11  pCi/L)  due  to the combined  effects of direct  ventilation with
only outdoor air, radioactive decay, and plate-out in the bedroom.  As for the
average radon level,  the HRV system lowered the average  total  EEDC in bedroom
2 slightly  to 0.14 pCi/L compared with 0.26 pCi/L  in the naturally-ventilated
house.  However, due to the  effects of mixing  with other regions of the house
caused by the furnace  air-handling  system,  the HRV-ventilated house could  not
achieve a reduction  in  bedroom average total EEDC  to below the outdoor level.

Since people normally spend a large fraction of their time at home sleeping In
their bedrooms, these results  Indicate  that  local  ventilation provided by the
LAFSW system in these bedrooms  could  significantly reduce the total long-term
exposure of occupants to radon progeny.

      CONAIR  89-2,  which  is  based  on  an  existing  computer  program,  was
developed  so it  could simulate  the accumulation  of  radon  and  its  progeny
within houses.  A radon progeny plate-out model was implemented in the program
and  a  set of  typical coefficients  describing  radon  progeny  kinetics  was
suggested  for a hypothetical house.

      Three  simulations of different ventilation strategies were  carried out
using CONAIR 89-2.   Average radon levels and total EEDCs  in  several  rooms of
the naturally-ventilated  hypothetical house with  baseboard heat  were compared
with  those of a  similar  house equipped with a  furnace/HRV  system  and with
those  in  one having  a baseboard-heat/LAFSW  system.    There were  significant
differences  in  average radon  levels  and total  EEDCs between  rooms of  the
house, which demonstrated the importance of using a multizone model instead of
a single-zone model to predict radon  and radon  progeny levels  indoors.   These
differences  also  indicate caution should be  exercised  in  extrapolating  radon
and radon  progeny level measurements  taken in one room to  any  other room in a

      The  simulations  predicted   that   the   baseboard-heat/LAFSW  system  can
control  radon  levels  in   the  bedrooms  better  than  the  furnace/HRV  system.
Since  the  average radon  levels   and  total  EEDCs  in  the  bedrooms  ventilated
using the  LAFSW  system were below those outdoors,  and since  people normally
spend a  large  fraction of their  time at home sleeping in  these bedrooms,  it
appears  that the LAFSW  system could  significantly reduce the  risk of lung
      This  project  was  carried out  with  the  support  of National  Research
Council Canada,  Energy,  Mines and Resources  Canada,  the Canada  Mortgage  and
Housing  Corporation,   and  Manitoba  Hydro.   We  gratefully  acknowledge  the
contributions of George Walton of the U.S. National  Institute of Standards  and
Technology and Jim  Axley of the Massachusetts  Institute of Technology.   The
use  of  their  programs  (AIRNET  and  CONTAM  87)   in  CONAIR  89-2  greatly
facilitated program  development.   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.

1.    Yuill, G.K.  and Wray,  C.P.  A microcomputer  program for evaluating the
      performance  of buildings  equipped  with  ventilator  window/baseboard
      heating systems.  In:   Proceedings of the 1989 Annual SESCI Conference,
      Penticton, 1989.

2.    Axley, J.   Progress toward  a  general analytical  method  for predicting
      indoor air pollution in  buildings  -  Indoor air quality modelling: Phase
      III  report.   U.S.  National Bureau of Standards to  U.S.  EPA,  NBSIR 88-
      3814, July 1988.

3.    Jacobi, W.   Activity and  potential  a-energy of  222radon-  and  ^radon-
      daughters in different air atmospheres.   Health Physics.   22(May): 441-
      450, 1972.

4.    PorstendSrfer,  0.,   Wicke,  A.,   and  Schraub,   A.    The   influence  of
      exhalation, ventilation  and deposition processes upon the concentration
      of radon  (^Radon), thoron (^Radon) and their  decay  products in room
      air.  Health Physics.  34(May): 465-473,  1978.

5.    Porstendorfer,  J.    Indoor radon  exposure in  the Federal  Republic  of
      Germany.    In:    Proceedings  of  the  Second  Air  Pollution  Control
      Association Specialty Conference:  Indoor Radon II, 1987.  pp. 57-67.

6.    Bruno, R.C.   Sources of indoor radon  in  houses: A  review.   Journal  of
      the Air Pollution Control Association.  33:(2, February): 105-109, 1983.

7.    Nazaroff, W.W.  and Nero, A.V. Jr.  Radon and its decay products indoors.
      New York:  John Wiley and Sons, 1988.

8.    Bigu, J.   Radon daughter  and thoron  daughter deposition  velocity and
      unattached  fractions  under  laboratory-controlled  conditions  and  in
      underground uranium mines.   Journal of Aerosol Science.   16(2): 157-165,

9.    Scott, A.G.   Radon  daughter  deposition velocities estimated from field
      measurements.  Health Physics.   45(2): 481-485, 1983.

10.   Porstendorfer,  J.   Behavior  of radon  daughter products  in  indoor air.
      Radiation Protection Dosimetry.  7(1-4):107-113, 1984.

11.   Kruger,  J. and Not hi ing, J.F.   A comparison  of the attachment  of the
      decay products of  radon-220  and  radon-222  to monodisperse  aerosols.
      Journal  of Aerosol  Science.  10:  571, 1979.

12.   Mercer,  T.T.   The effect of particle size on the escape of recoiling RaB
      atoms from particulate  surfaces.   Health  Physics.  31: 173-175, 1976.

13.   Shimo, M., Asano, Y., Hayash1, K.,  and Ikebe,  Y.   On  some properties of
      ^Rn  short-lived decay products  in air.  Health Physics.   48(1):  75-86,

14.   Yin 11, G.K.  and  Comeau, G.M.   Demonstration and  performance testing of
      the  laminar  airflow  super  window -  humidity controlled  air  inlet  -
      baseboard  heating  (LAFSW/HCAI/BH)  system   in  a  Winnipeg  house.    in:
      Proceedings of  the 1989 Annual SESCI Conference, Penticton, 1989.

15.   Bodansky,  D.,  Robkin,  M.A.,  and Stadler,   D.R.    Indoor  radon  and  its
      hazards.   Seattle:   University of Washington Press,  1989.
                                       TABLE 1
                              &maivy of Averago Radon Lmla. EEDCa.
                               ana Equibrun Fractions Natural Valuation
US 07
19 SI

446 48
Living Room
Blfroom 3
Both (Sink)
Both CTub)
Waiter Bldroom
Bedroom 2
*»( Cone

1 13


                        A DATA ACQUISITION SYSTEM FOR
R.P. Sieber                             D.A.  Flgley
Graduate student                        Research officer
Department of Mechanical Engineering    Prairie Regional Station
University of Saskatchewan              Institute for Research in Construction
Saskatoon, Sk.,  Canada, S7N OWO         National Research Council of Canada
                                        Saskatoon, Sk.,  Canada,  S7N OW9

R.W. Besant                             G.J.  Schoenau
Head                                    Professor
Department of Mechanical Engineering    Department of Mechanical Engineering
University of Saskatchewan              University of Saskatchewan
Saskatoon, Sk.,  Canada, S7N OWO         Saskatoon, Sk.,  Canada,  S7N OWO


      This paper describes the development of a micro-computer based Radon Data
Acquisition System  (RDAS)  for monitoring the entry  and  distribution of radon
gas in houses.  The information obtained will be valuable for the evaluation of
control measures used  to reduce radon concentrations to acceptable levels.

      There  are many  proposed control  measures  for  reducing  radon levels.
However,  these  methods are not always  reliable since they do  not accurately
account for uncontrolled  interacting factors.   Some  of  these  factors include
interzonal and building air leakage,  ventilation rate, pressure and temperature
gradients, radon concentrations in soil gas,  and environmental factors such as
wind  and  relative  humidity.    The  RDAS simultaneously  measures  pressure,
temperature,  and radon gas  concentration at  several  points.   In addition, the
RDAS characterizes the interzonal air exchange rate using tracer gas techniques.
Radon is measured with  semiconductor  sensors capable of measuring concentrations
continually over a period  of time; the output is transmitted to the data logger.
Careful design  of an experimental protocol will account for other uncontrolled


       The  presence  of elevated radon  gas concentrations  in residential housing
 can lead to  serious health  problems for the occupants  if they are exposed over
 an extended  period  of  time.    Some  control  measures   for  minimizing radon
 concentration levels include increased ventilation, sub  slab depressurization
 using  externally vented  fans,  and sealing potential  leakage  passages  (1-2).
 Typically,  evaluation of  these  control  measures  have  been  limited to  a
 before/after  air sample  analysis and  are unlikely  to  account  for temporal
 variations  in the  environment or building  factors.   Conclusions  drawn from
 studies  of this nature can contain  substantial errors.   The  purpose of this
 paper  is to outline briefly the experimental variables  that must be measured to
 conduct detailed building science based studies on radon  and  to describe  a data
 acquisition system  designed to  continuously monitor these variables.

       The Institute for Research in Construction (IRC)  of the National Research
 Council  of Canada  (NRC)  is developing a  data acquisition  system,  the RDAS,
 capable  of  continuous measurement  of experimental  parameters necessary  to
 evaluate the  effectiveness  of radon  control measures.   The parameters include
 ventilation rate, indoor/outdoor  pressure difference at various surfaces above
 and below  grade,  indoor and outdoor air temperature,  and radon  gas and radon
 daughter concentrations in the air and soil gas.  An important component of this
 research activity  is  the development  of  analytical models  that  describe  the
 functional relationships  between  the various  parameters.   These  models will
 provide a broader understanding of the dynamic processes involved, but they must
 be  validated with well characterized experimental data to be reliable.

       The RDAS will  allow  accurate evaluation of radon characterization in case-
 control  (paired) studies  and assist  in  the  development of relationships based
 on building science  principles.  This will provide valuable information to guide
 development of design and construction techniques for buildings.

                          BUILDING SCIENCE PRINCIPLES

      There are many potential  sources of radon gas in buildings.   One  of  the
most important  is the soil  surrounding the foundation.   To  accurately  assess
 radon  in buildings,  various  parameters  such  as  the  indoor  concentration,
ventilation rate, and transportation  of radon  gas  into  the  space need  to  be
quantified.  Considering a well mixed, single zone chamber with an outdoor  air
supply, indoor and outdoor radon sources and internal space conditioning (Figure
 1), a simple mass balance model yields the  equation:

       V C
             Figure  1. Single zone steady-state mass balance  model
                            Ci - C0 + S - R      (1)
                                     K  • V

where:      Ct - indoor  radon gas concentration (pCi/m3)
            C0 - outdoor radon gas concentration (pCi/m3)
            S  - indoor  radon gas source strength (pCi/s)
            R  - radon gas removal rate  (pCi/s)
            K  - ventilation efficiency
            V  - outdoor air exchange rate (m3/s)

      Equation 1 is a simplified expression that can be used  to  identify  the
major parameters that must  be  considered  and  the impact that changes in  the
parameters will  have  on  the resulting  indoor radon gas  concentration.    In
practice,  temporal and spacial  variations  in these  factors  must be  accurately
monitored to  avoid  errors in  analysis.   Other potential interacting  factors
including soil porosity  and  relative  humidity  may  also  effect  the radon
concentration  level.    In multi-zoned  buildings,  data  will  be  required  on
conditions within individual zones and on communication  among  zones which  add
to the complexity of the data acquisition system requirements.

                               SYSTEM COMPONENTS

       A schematic diagram of the RDAS  is shown in Figure 2.
                            Sciemetrica InsL
                            System 200
              Toshiba 3200 AT
              Portable, using
              Custom software
                      Figure 2. Schematic diagram of RDAS
      The system  includes  a  central MS  DOS  portable  computer and a Sciemetrics
System  200  data  acquisition module.   The  System  200  is  a modular,  general
purpose,  measurement  system suited  to numerous applications   including  data
acquisition and process  control.   Included  in the RDAS  are  two  model 220 relay
cards, a model  231 analog/digital converter, and two model  252  expansion card
(32 channels each).  Communication between the modules and the personal computer
is accomplished with  the System 200 I/O module  802 interface card.   Customized
software is  added to set up channel sampling  frequency, feedback control and data
processing and storage.

      The menu driven software allows the RDAS  operator flexibility in setting
up individual  experiments.   The  number of  transducers,  gains  or  calibration
coefficients  and  scan  rate can  be  input  from  the  keyboard to  suit  the
experimental requirements.

      The techniques to measure pressure difference across the building envelope
surfaces, air  ventilation and infiltration in the various  internal  zones,
internal and external temperature distribution, and radon gas concentrations are
presented in more detail below.


       Below grade  air  infiltration  due to  lower pressures  in basement  air
 compared to surrounding soil gas pressures has been  identified as  the primary
 cause of high  radon gas levels  in houses (1,3).   Potential entry  points  of
 airborne radon gas  include  floor and wall cracks, floor drains, and line cracks
 where the  slab and wall intersect  (Figure  3).   Methods  to reduce  radon  gas
 infiltration into basements are to eliminate or reduce the  flow of soil gas into
 the basement  by  sealing  entry  points or  by  eliminating  the  air  pressure
 differences which cause air infiltration into the basement foundation.
                   Figure 3. Potential soil gas entry sites
The rate of airflow through an opening such as a crack or a hole is given by:

                              Q - C(AP)n        (2)

where:      Q - airflow rate (m3/s)
            C - flow coefficient (m3/s-Pan)
            P - pressure difference across opening (Pa)
            n - flow exponent (between 0.5 - 1.0)

      The total pressure difference across the opening is the sum of the pressure
gradients due  to  wind, stack, changes  in atmospheric pressure,  and pressure
differences created by the mechanical system.  In the case  of radon gas in soil,
equation 2  can be  coupled  with the  radon  concentration in  the soil  gas  to
determine the radon source strength due to airflow as:

where:      Sp - radon source strength due to airflow  (pCi/s)
            R  - radon concentration in  the  soil  gas  (pCi/m3)

      For a foundation,  the  flows  and pressure  fields  are coupled in a complex
network  as  shown in Figure 4.   The  flow coefficient is  more complex than the
case of  a simple  crack  or hole.  The overall flow resistance is  a combination
of  the  flow resistance of the  foundation opening,  Rt, and the adjacent soil,
       Psg (SOIL GAS)
                   Figure 4.  Soil gas and  resistor  network
      For below grade components,  the airflow  resistance  of the soil can cause
large time shifts between the atmospheric pressure and the soil gas pressure on
the outside of the  below grade building envelope.  It is necessary to measure
the  pressure  difference  across  individual building  components  in order  to
calculate component specific airflows.

      The movement  of air and  airborne  pollutants within a  building is driven
by pressure gradients.  Although  the RDAS estimates air movement by tracer gas
techniques, the pressure regime inside and outside the building must be measured
independently for inputs  to the analytical  models.

      The RDAS can  record the  output from  10  individual  pressure transducers.
Modus T10 differential pressure transducers with  0-5 VDC  outputs were selected
for the system.  The range of  the  transducers  is  matched  to the expected range
in pressure to minimize measurement errors.   Maximum error is specified as ± 2%
of full scale capacity.


      The total ventilation rate of a building is a combination of infiltration
and mechanical ventilation.  While techniques exist for estimating or calculating
these components, they may not be sufficiently accurate for research purposes.
Further, the  distribution  of the ventilation will directly  affect  the indoor
pollutant concentrations.

      Ventilation rates are determined through the use of tracer gas techniques.
The tracer gas decay rate is an exponential time function related to ventilation.
The tracer decay method consists of an initial injection of tracer into the space
followed by recording  tracer  gas  concentrations  as a function of  time.  For a
well mixed zone with no sources  or sinks, the tracer  gas concentration  is given
by the equation:

                  c(t) - c0 •  exp[-(q/V)t]       (4)

where:      c -  tracer concentration
            c0= initial tracer concentration
            V =  space volume  (m3)
            q -  outdoor air exchange rate (m3/hr)
            t -  time (hr)

      Knowing  the effective  zone volume and the  current and  initial tracer
concentrations enables the  infiltration rate to be determined.  The air change
rate  (air  changes/hour)  is the  ratio of the air exchange rate,  q,  and space
volume, V.

      A four zone Air  Change  per  Hour Measuring Apparatus (ACHMA) was  designed
to  evaluate  ventilation and  air  movement  within and between zones.   N20 was
selected as  the tracer gas for  estimating ventilation  rates.  N20  is easy to
measure and  suitable for  small  and medium sized buildings.   To  estimate air
leakage characteristics between zones additional tracer gases (C02 and SF6) can
be  incorporated  into the system.

The requirements for the ACHMA are:
      1) Set  tracer gas maximum  and minimum zone concentrations (0-100 ppm).
      2) Discharge N20 tracer gas into one or a combination  of four well mixed
      3) Sample  in one or  a combination of four zones.
      4) Return  sample to  zone or purge outside.

      A  small amount  of N20  is  injected into  the  zone and  the  decay in gas
concentration is measured as a function of time.  Once the tracer  concentration
has decayed below the  minimum allowed concentration, additional tracer can be
injected  into the zone.   This allows for  an extended testing period.  Proper
mixing of air within the zone is essential for unbiased sampling since  equation
4  assumes  perfect mixing within the  zone.   The time period selected for the
solution  of  equation 4 must  be  carefully  considered.  Initially, following  a
tracer gas injection, the tracer will not be well mixed.  During a  test, changes
in  the air exchange rate can occur which will confuse the analysis  if  the change

is not recorded within the time step. A shorten time step requires  a higher rate
of sampling and increases the load on the sampling system. A zone  sampling rate
of five minutes was selected as the  best compromise between system requirements
and accuracy of estimation of air exchange rate.

      A schematic of  the ACHMA is shown in Figure 5.   The ACHMA consists of a
network of solenoid valves, tubing, an air pump,  gas  analyzer,  and tracer gas
source tank equipped with a pressure regulator.    The  N20 analyzer is a Beckman
Model  865  infrared gas analyzer with  a  span of  0  to 100 ±  1  ppra (0-100mVOC
output}.  Rotameters are used to monitor and control flow since only approximate
flow rates are required.

N20 Detector
                  Figure  5.  Four  zone  ACH measuring  apparatus

      Temperature differences  across the skin  of  a building often  cause air
infiltration while  internal  temperature differences may or may  not  cause air
circulation.    Temperatures  are  measured  with  Type  T  (copper-constantan)
thermocouples.  The original data logger was modified to include an isothermal
junction for the thermocouple connections.  This improved the original accuracy
from ± 2*C  to ±  0.1°C.   The  isothermal  Junction can accommodate 24 individual
thermocouple inputs.


      In general, radon concentrations  can be determined by measuring the time-
average value,  where  the  average radon  concentration  is  calculated  over  a
specified period of time, or by  instantaneous  measurement,  otherwise known as
grab sampling.   Time-averaging does not  identify transient  variations .while
manual grab sampling becomes time consuming and laborious when sampling a large
population.   Due to various interacting  factors,  radon levels  can fluctuate
substantially over an  extended period  of time (Figure 6).   The  RDAS includes
radon meters capable of continuous "real time" measurement. This allows analysis
of both the transient and steady  state  response associated with various control
measures or disturbances.
             RADON. mWL
24   36  48
                   60   72   84
                   TIME. HOURS

Figure 6. Temporal variations of radon levels
                                                       108  120   132  144
      While many previous studies on radon measurement focused on the house as
a whole, information is needed at suspected radon gas entry points and to examine
the spacial distribution of radon within the building.   Spacial variations of
radon gas within a zone  or between zones are often large enough to warrant the
use of several detectors. The RDAS incorporates  10 radon meters into  its design.

      The radon meters are designed by Thomson & Nielsen Electronics Ltd. They
produce  a  time integrated measurement similar  to  that  of other radon meters,
however, the time constant is relatively small (five minutes)  compared to other
radon detectors.   Although the radon data  is  not actually instantaneous, the
integrated values are  acceptable since radon levels are not expected to change

significantly over the five minute interval.

      Semiconductor sensors are used  to produce a voltage signal proportional
to  the  radon concentration,  or  alpha activity.   An integral pump  draws air
through the an intake filter, into a chamber, then out through an exhaust filter.
The intake filter absorbs radon daughters which are present.  As the radon gas
continues through the chamber, the gas  decays  and the daughters  are deposited
on the second filter.   With both  filters in place,  radon gas  is detected.  With
no intake filter, only the daughters are  detected.  Alpha activity  on the filters
is counted and processed into a zero  to  five volt  DC output and recorded on the
data logger.   Each  meter has its own power source and  pump for  air sampling.
The accuracy of the radon meters are within ± 10% of the true value.
                              FUTURE APPLICATIONS
      The RDAS is being installed in an experimental house to examine the effect
of various ventilation system operating modes  on the distribution of pollutants
within the house.   Initial  studies will focus on  line source  and point source
pollutants originating  at the  foundation.  Tracer gas sources will be used to
simulate radon entering through cracks and holes in the foundation.  Subsequent
studies will investigate the effect of building envelope and mechanical system
operation modifications on the radon entry and distribution characteristics.

      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.
      Special  thanks are  extended  to D.H.  Guenter  of  IRC,  Saskatoon  for
assistance in  the  design and  construction  of the RDAS,  to M.E.  Lux  of IRC.
Saskatoon for  valuable  discussion, and to  Canada Mortgage and Housing Corp.
(CMHC) for their financial assistance in the form of a scholarship.

1. Radon  Reduction  Techniques  for  Detached  Houses  -  Technical  Guidance,
   EPA/625/5-86/019, United  States  Environmental Protection  Agency,  Research
   Triangle Park, NC.,  June,  1986.

2. Figley,  D.A.,  Dumont,  R.S.   Techniques  For  Measuring  The  Air  Leakage
   Characteristics Of Below  Grade Foundation Components. In:   Proceedings of
   the 82nd Annual Meeting of  the Air and Waste Management Association,  Anaheim,
   Ca., June, 1989.

3. Indoor Air Quality Environmental Information Handbook: Radon, DOE/PE/72013-
   2, United States Department of Energy, Washington, D.C. February, 1985. pp.

4. Figley, D.A., Dumont, R.S.  Radon  in Houses  -  A Building Science Approach,
   Accepted for presentation at the Proceedings of  the 5th Conference on Building
   Science and Technology, February, 1990.

5. Dumont, R.S.,  Figley,  D.A.    Control of  Radon in  Houses.  In:    Canadian
   Building Digest 247, February, 1988.

6. Determining Air Leakage  Rate  By Tracer Dilution. In:  Annual  Book of ASTM
   Standards, Vol 04.07, E741-83, November, 1983.

7. ASHRAE Handbook 1989  Fundamentals, American Society of Heating, Refrigeration
   and Air Conditioning Engineers, Atlanta, Ga.,  1989.   pp.  23-1  -  23-10.

8. Discussions  with  M.E.  Lux.,  Research  officer,  Prairie  Regional  Station,
   Institute for Research in Construction,  National Research Council of Canada,
   Saskatoon, Sk., Canada, S7N OW9

              Session C-VI:
Radon in the Natural Environment—POSTERS

                    AREAS IN TWENTY-FIVE STATES
                          R. Thomas Peake

               U.  S. Environmental Protection Agency

     A  preliminary  radon potential  nap  of 25  states has been
prepared shoving areas of high radon potential.  The data used to
create this map include:

1)   National Uranium Resource Evaluation (NURE) aerial radiometric
data.  NURE data  in glaciated and non-glaciated  areas have to be
interpreted differently because  the properties  of  the surficial
material can be as  important as the  radium concentration.

2)   State/EPA Indoor  Radon  Survey Data.  Indoor  radon data  for 25
states  have  been  analyzed  by house construction by  county and

     Both sets of data have  been compared to geology to produce a
county scale map.   Areas of high radon potential include, but are
not limited to, the Upper Midwest, the Rocky Mountain region and
portions of the  Applachian  Mountains.   This map will  be used in
EPA's efforts  to identify and characterize  high radon potential
areas for  school  and workplace  surveys,  and  for  building code
      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.

                     BRITISH COLUMBIA.  CANADA

               By: M.M. Ghomshei1  and  W.F.  Slawson2

1. Orchard Geothermal Inc. 401-134 Abbott Street,  Vancouver, B.C.
V6B 2K4 (Canada)

2. Department of  Geophysics  and Astronomy,  University of British
Columbia,  Vancouver, B.C. V6T 1W5 (Canada)

     Sampling  of radon  within  the  soil  from three    sites  in
metropolitan   Vancouver   is  reported.  Alpha   track  bi-weekly
measurements during  a  period of 4 years  show secular variations
with a period  of 8-15  months.  There  are low-radon and high-radon
episodes enduring several  months to a  year.  Average radon level
during the high-radon episodes reaches  5-10 times that of the low-
radon  periods.   During high-radon  episodes  the high-frequency
variations show very  high amplitudes. After  filtering of the high-
frequency fluctuations, the data from different sites demonstrate
remarkably  similar  trends.  It  is  suggested  that  along  with
hydrogeological events, stress relaxation in rocks, earthquake, and
magma  emplacement  may  contribute to the   sources  of  secular
variations of  radon. Because of long-term variations, radon level
in urban areas should  be monitored on  a continuous  basis.  Single
measurements, even those integrating radiation over a period of few
months,  may sample  a  low-radon  episode,  and  provide  a  false
assurance, or  occur  during  a high-radon episode  and give a false


     Temporal  variation of  atmospheric  radon  is  a  known  fact
especially  to  those   involved   in   uranium,     and  geothermal
exploration/   hydrogeology,   and  environmental   science   (e.g.
1/2,3/4,5,6). These variations are controlled by: 1- factors which
affect the  release  of radon from  the  source geological material
(e.g. radium-bearing mineral) and 2- factors which affect the radon
transport systems. The  emanation  and  diffusion  of  radon from the
radium-bearing mineral to a porous media containing a fluid phase
(e.g. water,  steam,  or other gases)  is mainly  controlled by the
volume,  shape and structure of  the  porosity  (e.g. 7). Continuously
varying factors such as temperature and the pressure of the fluid
secondary phase have  some  control  on  the  emanation process  (e.g.
8). Co-seismic activities in ground may have a major control on the
release of  radon to  atmosphere.  This control  is  mainly through
producing microfractures which speed  the  process of diffusion of
radon from the source rock into the highly mobile  fluid phase.

     Since the discovery of hazardous radon levels  in the urban
areas (e.g.  9), the  prime concern of the environmental researchers
have  been the  spatial  variations of radon.  Study  of  temporal
variations have comparatively been neglected/ resulting in a lack
of a long-term data base.  It  is  often wrongly  assumed that alpha
track "single-shot"  surveys  with an exposure time  of few months
would average  out  the temporal fluctuations. Besides the common
high-frequency (diurnal  and weekly or monthly)  fluctuations, the
low-frequency  (seasonal  or  secular) variations   demonstrate high
amplitudes  which  should  be   taken   into  account  in  the  risk
assessment calculations.

     The present  work  introduces data on the significance of short-
term and  long-term temporal variations. Possible causes of these
variations are discussed. The data were collected during the period
of 1977 - 1981 and  are being reported now  in view of the recent
wide-spread  interests  in   radiation  protection  and  earthquake



     A number of  sites within  the  Greater Vancouver   (British
Columbia,  Canada) area were selected for long-term monitoring  (Fig.
1). The selection was made  to include a variety of geological and
hydro-geological environments. Site YVR was located in the Fraser
River delta,  lying over several hundred meters of Recent Sediments.
The UBC site  was  located above about  90  meters of Glacio-marine

Pig. 1. Location map, Greater Vancouver Area.

deposits. Several  meters of  glacial  outwash overlying  the Late
Cretaceous Coast Range granodiorite made  the host for site  CGN.
Only UBC and CGN were operated  over the more than 4 year period.
We report here  the sites UBC,  CGN  and YVR. Data  from two other
sites were not considered reliable due to frequent flooding.


     Alpha-track measurement system (10,11,9)  was  employed  for all
the sites. Detectors were placed in 60 cm deep cased holes and were
exposed for approximately two week periods  (Fig. 2).  The plastic
track detector  (from Terradex)  was placed at the  closed end of a
cup. The open  end was covered  with  a thoron  filter.  The  filter
consisting of a single  layer of  "Gladwrap" delays  the diffusion of
gases into the cup by almost a day, eliminating  all  the thoron and
a small fraction of  radon.  The  data  is reported as counted alpha
particles  per  square   meter of  detector  per  second.  Average
background  (blank) signal    was determined by another study  and
did  not  exceed 2 -± 1.5 counts nf2 s"1.    The  linear correlation
between track density  (cm"2) and dose  (pCiL^Day)  are reported by
Fleischer and coworkers  (9).
                                   HOLE COVER

• -
_ •'' .

^--PLASTIC CUP /"' ' .<
      Figure 2. The installation of a TRACK-ETCH detector.

                      RESULTS AND DISCUSSION

     The results are presented in the Figure 3. A low pass  filter
was applied to  smooth the high-frequency fluctuations and  better
highlight the longer-term effects.

                                     RADON HMANATON FROM SOIL
                                              . mm (con sast
       40  -
       M  -
       ao  -
       a  -
       10  -
        6  -
                                                    1. II
                                     RADON DONATION PROM SOIL
                                                ran (UK an>
                                     RADON EMANATION PROM SOIL
                                       TiNCOUVn. SOUTH (YVB SIB)
                                                  oar j. oat
                                                  —    ""'•""
                                MAX. Affi THMP. AYS. INTHBVAL « IS DATS
                                       nncoDTB, na (me am)
                                                 nur t. lor?
Fig.   3

     The spectrum of data appears to contain two main frequencies:
1- variations with a period of several weeks  (few data points) . 2-
 Variations  with a longer period of several months to a year.  A
third longer-term variation can be depicted from the overall rising
trend visible in all the sites.  The 4 year data collection period
does not seem sufficient to detect any periodicity in this trend.
The short-period variations are controlled  to  some  extent by the
transient atmospheric perturbations such as precipitation, pressure
and temperature (e.g. 6,4,3) . Our data did not show any significant
correlation with atmospheric parameters.  It  is possible that the
bi-weekly time-averaged measurements have filtered the fluctuations
related to relatively  rapid atmospheric  perturbations.  As to the
low-frequency   variations, most  of the  data  (first  1000  days)
demonstrate  a   periodicity  close  to  1  year  which  might  be
tentatively correlated with seasonal temperature  patterns.  This
correlation is not however consistent over the entire data spectrum
(Fig. 3). The variations induced by the atmospheric events may be
overshadowed by more vigorous changes  induced  by non-atmospheric
dynamic factors such as stress relaxation in the bedrock, related
degassing, and  some fluid mobilization.  It  should  also be noted
that the combination of the meteorological changes which  affect the
soil  parameters is  not  necessarily of  yearly  frequency.  Radon
episodes,  therefore,   may  not  coincide  with  yearly  temperature
seasons. To avoid confusion, we suggest the term episode  (instead
of season) to be employed for these  long-term variations of radon.
According  to  our data the  frequency  of the high-  and low-radon
episodes is not necessarily associated with the calendar year.

     An interesting observation  can be made on  the  amplitude of
the short-term and long-term variations.  It  seems the amplitude of
the short-term variations is positively correlated with that of the
long-term  variations.  The  short-term fluctuations  demonstrate a
higher amplitude during the  high-radon episodes  (crest  of the long-
term variations). This  observation strongly suggests that the long-
term  background  enhancement  is   the   major   source  of  radon
fluctuations.  The  short-term  atmospheric  factors  do not  have a
causal  relation with radon emanation.  These factors  act only as
transient  mobilizers.  The  mobilization  is naturally proportional
to the existing free radon  in the rock-soil  system. By  free radon
we  mean the  radon  which has been  released from  the  rock  to a
secondary  fluid media  and consequently more dynamic environment.

     The geological factors which may contribute  to the background
enhancement  are  either  of  a  pulse  nature  such  as  seismic
activities,  or  of  events  of  longer  duration   such  as  stress
relaxation  in   the rock   formation   due  to   plate  adjustment
(dilatancy-diffusion).  Some relatively  fast  stress  relaxations
precede  brittle deformations  and  can be used as  an earthquake
precursor.  Fast  relaxation  may  induce  extremely  high  radon

background in the roclc-soil system.  The degassing of this radon to
the  atmosphere  can  cause  "radon  storms"   or  "impulsive  radon
emanation" (12).  On our data the relatively sharp increase in radon
in 1981 may be related to this type of phenomenon.

     A second observation suggesting a lithospheric origin for the
long-term  variation  is  the  similarity  of  data-spectra  from
different sites. The general  rising trend is remarkable on all 3
sites. Other  shorter-period variations   appear also  to  reflect
regional behaviour. The  coincidence of  the  "episodic" variations
from the three sites is obvious. Although these variations may be
to some  extent related  to long-term meteorological  cycles,  the
remarkable similarities  of the data spectra  are suggestive  of a
common subsistent lithospheric component.

                      CONCLUDING REMARKS

     The data  presented in this  work  suggest that  the temporal
variations  of  radon  are  composed  of   high- and  low-frequency
components. The high-frequency  variations may reflect the effect
of  the  transient  atmospheric  perturbations   on  the  dynamic
properties of  soil and its  contained  fluids. These variations
demonstrate large amplitude when they are superimposed on the crest
of the low-frequency variations.  The low-frequency variations are
the main source  of overall radon enhancement at  regional scale.
These  variations  may be  related to  a  combination  of long-term
meteorological  cycles   and  crustal  events.   Therefore,   only
measurement  spanning  several  years constitutes valid base-line


     Data collection project  was  supported  by grants to WFS from
the Geophysics Division of  the DEMR. C.Y.  King of  USGS (Menlo
Park), and J.  Gingrich of  Terradex Corp. were very  helpful and
supportive during the period of 1979 - 1981.

     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.


1. Malmqvist,  L.,  Isaksson, M. and Kristiansson, K. Radon migration
through soil and bedrock. Geoexploration, 26: 135,  1989.

2.  Kruger,   P. and  Semprini,  L.  Radon as in-situ   tracer  in
geothermal reservoirs.  Electric  Power  Research  Institute  Inc.

(EPRI),  Report No. AP-5315, 1987.

3.  Ghomshei,  M.M.  Concentrations  des  radioelements  naturels
(!'uranium,  le thorium  et le potassium) et  evolution des magmas
(exemple de quatre series volcaniques) .  Doctoral Thesis/ University
of Paris - XI, 1983. 201 pp.

4. Fleischer, R.L.  and Mogro-Campero, A. Radon enhancements in the
Earth:  Evidence  for  intermittent  upflows.  Geophysical  Research
Letters,  5:  361, 1979.

5. Pearson,  J.E.  Natural environmental radioactivity from radon
222. Public Health Service Publications (U.S. Government), 1967.

6.  Pearson,   J.  E.  and  Moses,   H.   Atmospheric    radon  222
concentration  variations with  height  and  time.  J.  of  Applied
Meteorology,  5: 175, 1965.

7. Tanner, A.B.  Radon  migration in the ground:  A  supplementary
review: United States Geological Survey Open-File,  Report 78-1050/
1978. 62 pp.

8. Satomi, K. and Kruger, P.  Radon  emanation  mechanism from finely
ground  rocks.  Proceedings  of  Fourth  International  Symposium on
Water-Rock Interactions, Misasa, Japan, August, 1983.

9. Fleischer, R.L.,  Girard,  W.R., Mogro-Campero, A.,  Turner, L.G.,
Alter,  H.W.  and  Gingrich,  J.F.  Dosimetry  of environmental radon:
Methods  and  theory for  low-dose integrated  measurements. Health
Physics, 39:  957, 1980.

10. King, C.Y. Episodic radon changes in subsurface soil gas along
active   faults  and  possible   relation  to  earthquakes.  J.  of
Geophysical  Research, 85: 3065,  1980.

11. Fleischer, R.L.  Price,  P.B., and Walker, R.M. Nuclear Tracks
in Solids: Principles and Applications: University  of California
Press,  Berkeley, California, 1975.  605 pp.

12. King, C.Y. Impulsive  radon  emanation on  a creeping segment of
the San  Andreas Fault,  California. Pageoph., 122:  1.

                    OF NORTH CENTRAL TENNESSEE
                        by:  Paul D. Collar
                          G.F.O. Box 4424
                        San Juan, PR  00936


                          Albert E. Ogden
            Center for the Management, Utilization, and
                   Protection of Water Resources
                Tennessee Technological University
                       Cookeville, TN  38505

      Radon concentrations were measured in soils, caves, and
homes of north central Tennessee.  Soil radon concentration ranged
from 72.5 to 349.2 picoCuries/Liter (pCi/L) and had a mean of
130.5 pCi/L; the maximum was measured within a soil developed on
the uraniferous Chattanooga Shale.  Cave radon ranged from 7.2
pCi/L to 386 pCi/L.  Caves within Ordovician limestones had a mean
radon concentration of 129.1 pCi/L while caves in Mississippian
limestones had a mean radon concentration of 30.1 pCi/L.  Home
radon concentration ranged from 0.2 pCi/L to 37.6 pCi/L and had a
mean of 2.02 pCi/L; twenty-three percent of the home radon
concentrations exceeded 4 pCi/L.

      Homes built on Fort Payne limestone had the highest
geometric mean (2.91 pCi/L) of lithologically classified samples;
homes built on the Warsaw limestone had the lowest mean (1.58
pCi/L).  A weak inverse correlation between radon concentration
and distance above the Chattanooga Shale was indicated by a
Pearson correlation coefficient of -0.165.  The confidence the
relation was not coincidental, however, was only 81.4 percent.
Crawl space homes and basement homes were found to have
statistically similar mean radon concentrations.  The depth of
home excavation varied directly with radon concentration at a
confidence of 99.9 % using the Pearson product-moment correlation.


       Radon concentrations were measured in 9 soils,  22  caves,  and 98 homes of
north central Tennessee.  The radon concentration of  soils  and  caves were
analyzed in order to characterize the local geologic  environment for its
potential to introduce radon to houses located  in the field area.   Radon
levels in homes were correlated with several geological  and home construction
factors in an attempt to determine if geological and  engineering controls on
radon's distribution in the indoor environment  could  be  quantified for
incorporation into mitigative home construction strategies.

                                   FIELD AREA

         Physiographically, the field area encompasses terrain from both the
Outer Central Basin and Eastern Highland Rim Provinces of north central
Tennessee (Figure 1).  A simplified northwest-southeast  trending cross-
section is provided in Figure 2 and shows that  rocks  cropping out  in the field
                 Wlllfrn vilify

             Cotllll Pill"
ol Ttnnfssti Wvtr

Mississippi Hivtr Vallty
                                                                    Unaka Mount.m.
     Mississippi EtKbaymcnt
                             Nathvill. Dome
                                              Appalachian Foldb.lt
         Figure  1.   Physiography of central Tennessee showing  location
                     of study area (modified after (1)).
area include Ordovician and Mississippian carbonates and the Devonian
Chattanooga Shale.   Table 1 summarizes the insoluble percentages  and the
phosphate content associated with each of the major rock units exposed  in the
field area.  The differing solubilities of the various lithologies  in the
field area strongly  govern the physiography of central Tennessee  (2).   The
Central Basin  is a rolling lowland which has resulted from the greater
solubility of  the Ordovician carbonates relative to the siliceous,
Mississippian  Fort Payne and Warsaw limestones.   Soils develop to  a maximum
thickness of about 20  feet in the Central Basin (3); for the most part,
however, soils do not  exceed six feet in depth.
Formations contain a high percentage of silica.
                          The Fort Payne and Warsaw
                          In the Fort Payne, silica
occurs predominantly  as  thick-bedded cherts; in the Warsaw silica occurs
mostly as silt and  sand.   As a result, these units are erosionally resistant


                                         ;i,'i t'  'X C«tfc»y«-
        Figure 2.

          Table 1,
Cross-section within study area showing the principal

Percentage of insoluble constituents and phosphate
in each of the major carbonate lithologies of the
field area (modified after (5) and (6).
Rock  Unit


St. Louis


Fort  Payne



           1.8 to  7.3



(as P2O5)





and constitute a plateau surface known as the Highland Rim Province.  Soil
thicknesses often exceed 50 feet in the Highland Rim and have been reported up
to 100 feet (4).  The Mississippian lithologies above the Warsaw Formation are
for the most part relatively pure and highly subject to dissolution.  The St.
Louis Limestone forms a sinkhole plain on the easternmost flank of the
Highland Rim Plateau surface.  A steep escarpment abuts the sinkhole plain and
rises to the Cumberland Plateau.  This escarpment is underlain by the
Monteagle, Hartselle, and Bangor Formations.  Despite the fact that the Warsaw
Limestone contains abundant silica, several caves are present in this unit as
a result of the vertical continuity of cave systems originating in the St.
Louis Limestone.  The Fort Payne Formation, although abundantly fractured and
highly porous relative to adjacent lithologies, is characterized by small
pocket caves.  The caves analyzed for radon in the present survey were all in
either the Ordovician carbonates or the Mississippian Warsaw or Monteagle

      The Chattanooga Shale has been shown (7 and 8) to contain relatively
high quantities of uranium throughout most of its geographic extent.  Whole
rock uranium analyses  of the organic-rich Gassaway Member of the Chattanooga
Shale ranged from 13 to 97 microgram per gram (ug/g) and averaged 55 ug/g (8).
The uranium content of the Chattanooga Shale is therefore significantly higher
than most granites and comparable to highly uraniferous granites.  The
Chattanooga Shale's potential as a source of environmentally threatening radon
is probably minimized, however, by the limited thickness of the rock unit.
Throughout central Tennessee, the Chattanooga Shale averages about 30 feet in
thickness, and the uranium-rich Gassaway Member averages only 15 feet in
thickness (9).

                            RESULTS AND DISCUSSION


Soil Radon Survey

      Track-etch radon measurements in nine soil associations yielded a range
from 78 to 345 picoCuries per liter (pCi/L), a geometric mean of 130.5 pCi/L,
and an arithmetic average of 153.6 pCi/L (Table 2).  The limited number of
samples prohibitted the clear definition of the type of data distribution, but
the data appear to be lognormally distributed.  The two lowest soil radon
measurements were taken in floodplains of creeks during the wet months of
December and January, when the water table was near the land surface.  These
lower measurements may reflect the inhibitting effect of a nearby water table
on radon's diffusive migration through soil.  Nevertheless, the geometric mean
and average of all soil radon measurements are both higher than the national
average soil radon concentration of 100 pCi/L (10), suggesting that north
central Tennessee soils may introduce slightly more radon to area homes than
do soils on the national average.

Cave Radon Survey

      Radon analyses in 22 regional caves were made with carbon canisters over
a 4 to 5 day exposure period.  Elementary statistics are summarized in Table

Table  2.   Summary of  elementary  statistics  of variously
             classified  radon data  sets.
                                                    i«°    #°    «*    ^     £
                                                   >      +    v*x  y    i*
                                                       A°     -*v    oK      «
                                                  2.54   125.74X   23.47X
     Warsaw                42     1.58     2.46
     Fort Payne             25     2.91     3.53
     Chattanooga            10      1.7     2.03
     Ordovician             21     2.31     4.24
                        2.8   177.22X   21.43X
                       1.84   63.23X   32.00X
                       1.98   116.47X   10.00X
                        2.8   121.21X   19.05X
     Crawl space
     Concrete slab
64     1.87     2.61      2.4  128.34X  20.97%
34     2.31     3.87     2.78  120.35X  30.56X
Very Low
Below Average
Above Average
Very High


162. 35X
     130.5    153.6

2.  Caves in Ordovician rocks had a significantly higher geometric mean (129.1
pCi/L) than did caves in Mississippian rocks (30.1 pCi/L).  This is thought to
reflect mineralogical differences between the two limestones.  The
Mississippian Monteagle Formation is a relatively pure limestone (Table 1)
with an average of 0.03 percent phosphate (as P205); similarly, the underlying
Warsaw Formation contains an average of 0.033 percent phosphate and the Fort
Payne an average of only 0.017 percent phosphate.  The Ordovician Leipers and
Catheys Formations, on the other hand, have phosphate-rich horizons with up to
20 percent phosphate by weight (5).  The average phosphate concentration in
the Catheys and Leipers formations is 0.179 percent.  Insoluble constituents
average 7.48 percent within the Catheys and Leipers (Table 1).  Unlike the
overlying siliceous Fort Payne and Warsaw limestones, however, the majority of
the insoluble portion of the Ordovician strata consists of  interbedded shales
and disseminated clays.

      The lattices of calcium phosphate minerals have been shown to readily
accomodate uranium as an isomorphic substituent (11). Uranium is also commonly
adsorbed onto the surfaces of clay minerals.  The relatively higher proportion
of clay minerals and phosphate in the Ordovician rocks is therefore likely to
provide a suitable environment for the regional incorporation of uranium into
the Ordovician rock matrix.  Although some of the uranium may have been
introduced into the sedimentary basin by the deposition of clays, it is also
possible that ground water mobilization of uranium from the Chattanooga Shale
has resulted in the concentration of uranium in phosphatic horizons below the
Chattanooga Shale.


      Basements and crawl spaces were tested for radon with carbon canisters
in homes built upon the Catheys, Leipers, Chattanooga, Fort Payne, and Warsaw
Formations.  Measurements ranged from 0.1 to 37.6 pCi/L.  The data were
lognormally distributed, with a geometric mean of 2.02 pCi/L (Table 2).
Twenty three percent of the measurements were above the U.S. Environmental
Protection Agency recommended maximum of 4 pCi/L (12).  The mean of 2.02 pCi/L
was higher than the national geometric mean of 0.89 pCi/L (13).  However, the
maximum of 37.6 pCi/L was one or two orders of magnitude below the extremes
measured  in other parts of the United States.  The 37.6 pCi/L maximum of the
present study was measured in a home constructed within 100 feet of the cave
passage possessing the highest cave radon measurement (386 pCi/L) made in this
study.  Radon was shown to enter the home via a fracture system connecting the
cave and the substructure of the house (14), demonstrating that near surface
cave systems can pose a potential avenue of radon entry to nearby houses.

      Indoor radon measurements were classified according to underlying
bedrock; the data from each lithological class was found to possess a
lognormal distribution.  The elementary statistics describing these classes
are given in Table 2.  Homes built on the Fort Payne Formation possessed the
highest geometric mean (2.9 pCi/L) while homes built on the Warsaw Formation
possessed the lowest geometric mean (1.6 pCi/L).  These were the only two
populations which proved to be statistically separable at an alpha of 0.05
with the Student's t-test.

      Despite the fact that the Chattanooga Shale has a uranium content
significantly greater than the overlying and underlying limestone lithologies,
homes built on the Shale did not have anomalously high radon concentrations.
The low observed measurements (Table 2) may be due to the very limited 2 to
2.5 feet soil thickness developed on  the Chattanooga.  A thin soil can
minimize soil radon concentrations and the radon introduced to overlying homes
in several ways.  At soil depths less than 3 feet, only a portion of the
diffusion length of radon is utilized, thereby limiting radon production.  A
thin mantle of soil is furthermore subject to atmospheric influences to a
greater extent than is a thick soil.  Consequently, wind and temperature
fluctuations are potentially more effective in removing radon from the soil.
Precipitation is more likely to produce a saturated zone near the land surface
within a thin soil, thereby decreasing the effective diffusivity of the soil.

      Besides the soil thickness, another factor which may explain the low
radon content of Chattanooga homes is the location of radium ions within the
soil matrix.  The emanation coefficient of a given soil is strongly controlled
by the location of radium ions in the solid matrix (15).  If radium is
primarily present as adsorbed ions on the exterior of clay grains, then the
emanating fraction of the total volume of radon produced by the decay of
radium is likely to be high.  If, however, radium is bound in individual
grains a distance greater than a few microns from interstitial pore space,
then the radon that actually escapes the grains and enters the pore space is
likely to be significantly lower and result in a lower overall emanation
coefficient.   The proportion of clays in Chattanooga-derived soils was found
to be noticeably lower than that characterizing soils weathered from carbonate
lithologies.  Correspondingly, rock fragments were more abundant within
Chattanooga soils than within any of the other soils augered.  The relative
abundance of rock fragments and paucity of clay in the Chattanooga soil,
combined with the low radon measurements in homes, may suggest that radium is
bound within the interior of rock fragments and not predominantly attached to
the exterior of clay surfaces by adsorption.

      Based on a radon survey of 1733 homes in Tennessee and the knowledge
that the Chattanooga Shale was the most highly uraniferous stratum exposed
statewide, Tennessee Department of Health and Environment (16) characterized
the surface outcrop of the Chattanooga Shale and downslope areas in Tennessee
with its highest "radon occurrence potential" ranking.  In order to further
assess the effects of the Chattanooga Shale in promoting high radon
concentrations in adjacent homes, the stratigraphic distance above and beneath
the Chattanooga Shale was calculated for each of the homes tested during the
present survey using geologic maps, topographic maps, structure contour maps,
and well logs.  The distance from the Chattanooga Shale was plotted against
the logarithm of the radon concentration for the two sets of homes (those
beneath the shale and those above the shale).  Logarithmic transformation of
the raw data was necessary in order to permit the parametric statistical
comparison of the two variables by calculation of Pearson correlation

      As Figure 3 shows, no relation was manifested between distance beneath
the shale and radon concentration, suggesting that downslope drift of
weathered Chattanooga Shale residuum does not appear to be a significant

control on radon distribution in Central Basin soils adjacent to the
Chattanooga Shale outcrop.  The plot of radon concentration versus distance
above the Chattanooga Shale is shown in Figure 4 and appears to manifest a
weak inverse correlation between the two variables.  A mathematical
correlation of the two variables yielded a Pearson coefficient of -0.165 at an
alpha of 0.186, implying only an 81.4 percent  probability that the relation
is not coincidental.  The weak dependence of radon concentrations on distance
above the shale appears to be lost altogether beyond about 180 feet above the
Shale.  The degree of home weatherization is identified in Figure A and shows
that highly insulated homes plot generally in the uppermost portion of the
field whereas poorly insulated homes plot generally in the lowermost portion
of the field.  These generalizations are not all-encompassing, however,
suggesting that radon distribution is influenced by other factors as well.

      The strength of the correlation is insufficient to conclude that the
location of the Chattanooga Shale has a predictable effect on the radon
concentration within overlying lithologies.  Nevertheless, the weak trend
cannot be dismissed outright as coincidence either.  What the distribution of
the data appear to indicate is that radon distribution in the subsurface and
indoor environment is an extremely complex function of a large number of
variables.  As a result, the analysis of one or two factors which may
partially control its distribution cannot be expected to yield perfect data
fits with low covariance.  The apparent relationship between radon
concentration of homes and distance above the Chattanooga Shale—if it is in
fact real and not coincidental—may be the result of at least three
fundamentally separate processes.  These are briefly summarized below.

      1.  The inverse relationship could indicate that radon is produced
      within the Chattanooga Shale and migrates through the subsurface
      to overlying homes.  Homes more removed from the Shale would
      therefore be expected to have a lower radon concentration.  One
      problem with this interpretation is that the Chattanooga Shale
      forms the lower boundary of an aquifer developed within the Fort
      Payne and Warsaw.  Consequently, radon produced in the Chattanooga
      Shale would have to travel a tortuous subsurface path through
      ground water and ground air, all within the time constraint
      imposed by its limited decay period.

      2.  The upward diffusion of uranium throughout geologic time as a
      result of ground water reworking could have promoted higher
      uranium concentrations within overlying rock.  In such a case,
      uranium content would be expected to be progressively reduced with
      increasing distance above the Shale.  Under this circumstance,
      radon would be derived from the decay of radium grains contained
      within Fort Payne soils and not the underlying Chattanooga itself.

      3.  Finally, the apparent reduction in average radon concentration
      moving away from the Chattanooga Shale could reflect a change in
      the source or quantity of clays introduced into the sedimentary
      basin and be completely independent of the uranium content of the
      underlying shale.  This possibility would imply that radon is






I I I I I I Lk I I
A A ^
A A A -
A A A A .

A *
I i i i i i i i
100    150   200   250   300   350   400   450
Figure 3.  Log(radon content) plotted against  stratigraphic
          distance beneath the  Chattanooga  Shale.




i i i i i i i
0 * ^
- A o ° ° ** OD
A A O D ° n
°0 ° D
f£> O
_A i«w O
" O AvnuK Q
_ D mm
1 1 1 1 In 1 1
200    240    280   320
Figure 4.  Log (radon content) plotted against stratigraphic
          distance above the Chattanooga Shale for populations
          classified according to degree of home weather izat ion.

      derived from clays within the soils underlying homes and not from
      distant subsurface sources.


      Although geological environment is expected to provide an important
control on the magnitude of radon's localized availability, home construction
characteristics have been shown (17) to play a strong role in determining the
extent to which radon is admitted into a structure and concentrated there.
The effects of three construction factors were considered in describing the
radon distribution in this study.  These factors were:  1) home type,  2)
degree of home weatherization, and 3) depth of basement excavation.

Home Type

      Homes surveyed were of two types:  homes with basements and concrete
foundations, and homes with crawl spaces and earthen foundations.  Both data
sets exhibitted lognonnal distributions, and Table 2 summarizes simple
statistics calculated for the two populations.  Homes with basements had a
larger geometric mean (2.3 pCi/L) than homes with crawl spaces (1.9 pCi/L),
but the difference was not statistically significant with the t-test at an
alpha of 0.10 (evaluated with log-transformed data).  This implies that the
greater air exchange characteristic of the crawl space homes studied is
sufficient to overcome their heightened susceptibility to radon entry
resulting exclusively from the absence of a concrete foundation radon
diffusion barrier.  Correspondingly, the data strongly indicate that the
presence of a concrete foundation does not ensure that homes with excavated
basements are immune from the effects of radon entry.

Degree of Home Weatherization

      Figure 5 shows that the degree of home weatherization is a factor which
influences the magnitude of indoor radon concentrations.  Highly weatherized
homes had higher geometric means than homes which were poorly insulated.  The
coefficients of variation indicate that well insulated homes similarly had the
most homogeneous radon concentrations, suggesting that air exchange promoted
by "leaky" houses had a diagnostic effect on the variability of measured radon

Depth of Home Excavation

      Depth of excavation was determined onsite for each of the homes tested.
This parameter was correlated with log-transformed radon concentration using
the Pearson product-moment correlation, yielding a coefficient of 0.51 at an
alpha of 0.001.  The relationship is depicted in Figure 6 in which a linear
regression is plotted.  Excavation depth was similarly correlated  with
log(radon) for each lithology.   None of the lithologically classified data
sets provided statistically significant correlations at an alpha level of 0.1
or lower except for the Fort Payne data set, which yielded a Pearson
coefficient of 0.47 at an alpha of 0.012.  A linear regression of the Fort
Payne data is also shown in Figure 6.  Comparison of the two trends indicate
that the minimum radon concentrations of Fort Payne homes is higher than the

   Figure 5.   Plot showing the geometric mean and coefficient of
              variation of populations  classified on the basis of
              degree of home weatherization.
                        LOG(RADON)=0.056(DEPTH)+6.3 IS
                                               DATA SET
                                              O SJISMS5..«.
        0  1  23456789  10  11  12  13  1415 16 17 18
                      DEPTH OF EXCAVATION (FT)
   Figure 6.  Log(radon content) plotted against depth of home
             excavation showing overall and Fort Payne populations.

minimum concentrations of homes built within other lithologies, that is, the
y-intercept of the Fort Payne regression has a higher magnitude.  Conversely,
the overall data set manifests the greater slope, indicating that the
probability of higher radon concentrations at greater excavation depths is
higher among lithologies other than the Fort Payne.  The regression equations
loosely represent data with high covariance.  The error in their prediction of
radon concentration based on excavation depth was generally less than 15
percent but as high as 30 percent.  The regression functions suggest that on
the average, 4 pCi/L radon will be exceeded in the overall data set at
excavation depths greater than 8.2 feet and among Fort Payne homes at
excavation depths greater than 10.1 feet.

      The close correspondence between depth of excavation and radon
concentration in homes built on the Fort Payne is thought to reflect the fact
that the radon source lies within the soil itself and not in the Chattanooga
Shale.    An additional indication that the source of radon is within the soil
among the Fort Payne homes is the relatively low coefficient of variation for
the Fort Payne population (63 percent) compared to the other lithologic
populations, which ranged from 116 percent to 177 percent.  If radon were
derived from the Chattanooga Shale and transported along solutionally enlarged
fracture systems within the Fort Payne, radon concentrations in the Fort Payne
soil would be expected to be variable with high radon concentrations above
fracture zones and low radon concentrations above non-fractured rock.  The
fact that the coefficient of variation is the lowest among Fort Payne samples
argues that radon is not transported along fracture systems but is derived
from radium grains within the soil itself.


      Radon concentrations in north central Tennessee homes and soils are
slightly higher than indoor radon concentrations on a national average.
However, radon is present in quantities much lower than in areas of the United
States considered by the Environmental Protection Agency and equivalent state
agencies to be radon "hot spots".

      In the homes surveyed, radon was shown to vary as a function of many
different factors related to both geology and home construction.  Homes built
upon soil derived from the Fort Payne Formation had the highest mean radon
concentration, and homes built upon the overlying Warsaw Formation had the
lowest radon concentration.  There was an apparent inverse correlation between
distance above the Chattanooga Shale and radon concentration; the confidence
that this relation was not coincidental, however, was only 81.4 percent.
Homes possessing crawl spaces were shown to have a geometric mean
statistically similar to homes having basements with concrete foundations.
Depth of home excavation and log(radon) manifested a direct correlation with a
confidence of 99.9 % that the relation was not coincidental.  A linear
regression predicted that the U.S. Environmental Protection Agency recommended
maximum of 4 pCi/L would be exceeded on the average at excavation depths
greater than 8.2 feet for the overall data set and 10.1 feet within Fort Payne

       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

       The authors wish to acknowledge the Center for Management,
 Utilization, and Protection of Water Resources at Tennessee Technological
 University for funding of the research summarized herein.   Acknowledgments are
 due Brad Neff, who collected most of the cave radon measurements.
 Additionally,  we would like to thank Joe Troester and Gregg Hileman of the
 U.S.G.S.  for the technical reviews each provided.

 1.   Miller,  R.A.,  1979,  The Geologic  History of  Tennessee:   Dept.  of  Cons.,
     Div.  of  Geol.,  Bulletin 74,  59  p.

 2.   Reesman, A.L.,  and Godfrey,  A.E.,  1972,  Chemical Erosion and Denudation
     Rates in Middle Tennessee:   Tenn.  Dept.  of Cons. Div. Water Resources
     Series 4,  35 p.

 3.   Burchett,  C.R., and  Moore, G.K.,  1971, Water Resources  in the  Upper
     Stones River Basin,  Central  Tennessee:   Tenn. Dept. Cons. Div. Water
     Resources  series 8,  62  p.

 4.   Moore, G.K., and Wilson, J.M.,  1972, Water Resources of  the Center Hill
     Lake  Region, Tennessee:  Dept. of Cons. Div.  of Water Resources Series 9,
     77 p.

 5.   Hershey, R.E. and Maher, S.W.,  1985, Limestone and dolomite resources of
     Tennessee:  Tenn. Div.  of Geol. Bull. //  65,  252 p.

 6.   Milici, R.C., Briggs, G., Knox, L.M., Sitterly, P.O., and Statler, A.T.,
     1979, The Mississippian and  Pennsylvanian Systems in the United
     States--Tennessee:  U.S. Geological Survey Professional Paper  110-G.

 7.   Glover, L., 1959, Stratigraphy and uranium content of the Chattanooga
     Shale  in northeastern Alabama, northwestern Georgia, and eastern
     Tennessee:  U.S. Geological Survey Bulletin  1087-E, pp. 133-168.

8.   Leimer, H.W., and Matthews, R.D., 1981, Chattanooga Shale of the Eastern
     Highland Rim, Tennessee, and methods of sampling; Synthetic Fuel from Oil
     Shale II, symposium and field trip, Institute of Gas Technology.

9.   Conant, L.C. and Swanson, V.E.,  1961, Chattanooga Shale and related rocks
     of central Tennessee; U.S.  Geological Survey Professional Paper 357.

10.  Brookins, D.G., 1988, The indoor radon problem:  studies in the
     Albuquerque, New Mexico area:  Environ. Geol.  Water Sci., v. 12, n. 3,
     pp.  187-196.

11.  Rogers, J.J.W. and Adams, J.A.S., 1969:  Uranium; IN: Handbook of
     Geochemistry, Springer, Berlin, Chapter 92, 1962.

12.  U.S. Environmental Protection Agency, 1986-a,  A Citizen's Guide to Radon:
     USEPA/OPA-86-004, 14 p.

13.  Nero, A.V., 1988, Radon and its decay products in indoor air:   an
     overview:  in:  Radon and its Decay Products in Indoor Air, Nazaroff W.W.
     and Nero, A.V. eds, Wiley-Interscience; 518 p.

14.  Collar, P.D. and Ogden, A.E., 1989, Radon in homes, soils, and caves of
     north central Tennessee and implications for the home construction
     industry:  Third Interdisciplinary Conference on Sinkholes and
     Environmental Problems in Karst Terranes; St.  Petersburg, Florida, Oct.
     14,  1989.

15.  Nazaroff, W.W., Hoed, B.A., and Sextro, R.G.,  1988, Soil as a source of
     indoor radon:  generation, migration and entry:  in:  Radon and its Decay
     Products in Indoor Air, Nazaroff W.W. and Nero, A.V. eds,
     Wiley-Interscience; 518 p.

16.  Tennessee Department of Health and Environment, 1987 Summary Report of
     the Tennessee Radon Survey.

17.  U.S. Environmental Protection Agency, 1987, Radon Reduction in New
     Construction:  an Interim Guide:  EPA/OPA-87-009.



                      Linda C. S.  Gundersen

                     U.  S. Geological Survey

  A comparison of two soQ radon sampling techniques, die Reimer grab sampling
technique (RG) and die EPA flow through grab sample technique (EFT), reveals a
strong control of grain size and sorting on the soil radon measured. Moisture
produces secondary effects on sampling, sometimes determining whether a sample
can be obtained at all. Emanation, however is the ultimate control on the amount of
radon sampled and h appears to be influenced most by the abundance of metal-
oxides and the siting of uranium and radium in and around grains.
  In well sorted, medium to coarse grained sands the two sampling techniques obtain
similar radon concentrations, usually within  10% of each other.  In poorly sorted
sands, especially clayey sands, the RG obtains higher concentrations than the EFT.
Under high moisture conditions including saturated conditions, the EFT cannot
obtain a radon sample while the RG will obtain a sample approximately 50% of the
time.  Permeability measured at all of these sites has no apparent correlation with the
radon concentration.
  Equivalent uranium from gamma spectrometry at the surface and uranium and
radium measured chemically in the soil have  been compared with the soil radon
concentrations. Although a generally positive correlation exists, anomalous amounts
of soil radon with respect to its parent radionuclides are found in a variety of
geologic settings. The common factor causing these anomalous concentrations is the
occurrence of metal oxides, particularly iron, and the sorption or precipitation of
uranium and radium in association with the metal oxides, dramatically enhancing the
emanation in the rocks and 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.



     Arthur W. Rose, Department of Geosciences
     Edward J. Ciolkosz, Department of Agronomy
     John W. Washington, Department of Geosciences
             Pennsylvania State University
             University Park, PA 16802


     Radon in houses depends partly on Rn concentration in soil
gas, which is affected by water in a variety of ways.  In this
paper, temperature-dependent effects of Rn partitioning between
changing proportions of air and water in pore space are shown to
be capable of causing variations in Rn concentration up to 5-
fold.  A map of regional soil moisture and temperature regimes,
in combination with the moisture/temperature effects on Rn,
suggests that soil gas Rn will be most elevated by moisture
effects in eastern U.S. and other regions with the Udic soil
moisture regime.  Large seasonal differences are also predicted
for soils of the Udic moisture regime, with soils of the Frigid
temperature regime having higher Rn in summer than winter.
Aridic (dry) soils show negligible effects.   Clay-rich soils
will generally be most strongly affected by moisture variations.


     The main source of radon (Rn) in houses is now generally
accepted Co be soil gas entering through the foundation, driven
by small pressure gradients between the outside and inside of the
house (1).   The total flux of Rn into a house via soil gas is a
function of two factors, the flow rate of soil gas into the
house, and the concentration of radon in the soil gas.   The
first factor, the flow rate of soil gas, varies with time owing
to changes in the pressure gradient with changing conditions
inside and outside the house, and temporal changes in air-filled
porosity due to moisture changes.  In addition, the flow rate
varies from one house to another because of differences in
construction details, as well as differences in permeability of
soil and backfill along the path of air flow into the house.
The second factor, the Rn concentration in the entering soil gas,
depends largely on the radium (Ra) content, moisture and physical
properties of the soil and backfill around the house.   In
general, increased Rn in soil gas is expected to lead to
increased flux of Rn into the house, given equal influx rate of
soil air.

     The intent of this paper is to discuss some effects of soil
moisture and temperature variations on the Rn content of soil
air.  The main emphasis is on effects involved in Rn equilibrium
between air and water in the soil.   This phenomenon has an
important influence on absolute soil gas Rn concentrations as
well as the nature of changes with time in the Rn concentration
available to enter a house.   The same phenomena are also
important considerations in the interpretation of soil gas Rn
surveys to predict Rn hazard.


     A large number of processes involving water have been
proposed to affect the Rn content of soil gas.   The following
processes involving liquid water (but not ice) can be
distinguished (2, 3, 4, 5, 6): (1) The sealing effect of a water-
saturated zone at the surface or within the soil profile; (2)
inhibition of diffusion in water-saturated pores because of much
slower diffusion in water than air; (3) the flushing effect of
water percolating through the soil and either pushing air ahead
of it or transporting dissolved radon; (4) changes in volume of
the soil due to changes in moisture, leading to cracking and
swelling; (5) for relatively dry soils, an increase in Rn
emanation coefficient with increasing moisture; (6) for radon
transport into houses, the effect of the water table in limiting
the depth of air flow; (7) temperature-dependent partition of Rn
between water and air in soil pore space (2).   Host of these
processes must be considered theoretical, because unambiguous

field evidence demonstrating one cause of Rn variability and
eliminating all other causes is scarce.

     In most soil profiles, the Rn concentration decreases toward
the surface because of diffusion from the relatively high
concentrations (200-5000 pCi/1) generated in the soil compared to
negligible concentration in open air (0.2 pCi/1).    In soils with
dominantly air-filled pores, this zone of markedly varying
concentrations ranges from a few tens of centimeters to several
meters in thickness, depending on the effective diffusion
coefficient and air-filled porosity of the soil.   However, the
Rn diffusion coefficient in water is smaller by a factor of about
10* than the diffusion coefficient in air (2).    Because the
concentration gradient due to diffusion depends on the square
root of the diffusion coefficient, the zone of changing
concentration in water- saturated soil is condensed to about 1%
of its thickness in soil with air-filled pores.

     One result of this phenomenon is that a rainfall event
intense enough to saturate the surface soil can create a seal for
Rn, leading to high Rn levels beneath the water- saturated zone.
Several half lives of Rn are required for the full increase to
occur.    A low permeability zone within the soil can have a
similar effect.

     As a soil is progressively wetted, the pores fill with water
in the order of smallest to largest because of surface tension
effects. Because of the low diffusion rate of Rn in water,
transport of Rn in water-filled micropores is very slow.   Thus,
saturated micropores may effectively cause a dual pore system: a
small proportion of air-filled pores separated by water-filled
micropores and having high Rn concentration, and a separate set
of connected macropores in which Rn concentration is controlled
by diffusion in air, and flushing effects discussed below.  This
condition may be further complicated during intense rainfall
events when the network of connected macropores may be the
preferred pathway for rapid water movement through the soil.

     Rainfall or other precipitation leading to percolation of
water through the soil can displace Rn-rich air upwards.  Some Rn
is expected to dissolve from the soil air into the downward
percolating water, resulting in additional Rn transport downward
with the water.   These processes can lead to a temporary
depletion of Rn in the zone above the percolating water.

     In a soil, Rn atoms are generated by radioactive decay of Ra
atoms occurring in soil grains.   The proportion of Rn atoms that
escape into the pore space is termed the emanation coefficient.
Emanation coefficients in very dry soils are markedly lower than
in moist soils (7).  Typical values in soils with at least 5% of
the pore space filled by water are 0.15 to 0.35, with about 0.2 a

common value.   Although this effect has been clearly
demonstrated in experiments, it appears that few soils are dry
enough to have significantly decreased emanation coefficients.

     In most soils, the pore space is filled with a mixture of
air and water.   These phases are expected to be in close enough
contact that Rn will redistribute itself between air and water by
local diffusion between pores.    At depths below the zone of
significant diffusive loss to the atmosphere, when the soil air
and water have remained long enough for emanated Rn to reach
radioactive equilibrium, the concentration of Rn per unit volume
of air-filled pore space (CR )  is given by the following relation

           CRn -

where CRfl is the concentration of Ra per unit mass of soil
particles, E is the emanation coefficient, D is the dry bulk soil
density, P is the volume fraction of total pore space, F is the
volume fraction of total pore space occupied by water, and K is
the partition coefficient of Rn between unit volumes of water and
air (CR  water/cRn air)-   Tne  v*l"e of K depends on temperature
(T), and'ranges from 0.54 at 0°C to 0.23 at 25°C (6).

     The effects of air-water partition may be isolated from
equation (1) and expressed as Q, the factor by which variations
in moisture and temperature increase Rn concentration in air-
filled pore space:

                   Q -     1	                 (2)

     Given this relation,  Figure 1 shows the effects of changes
in F and T on Rn concentrations of soil air,  given fixed values
of CRa,  E, D and P.   CR_ is seen to vary by a factor of 5
between F-0, T-0°C and F-1.0, T-30°C.    Variations of a
significant fraction of this range can occur at a single site at
different times, and most of the above range occurs within soils
of the United States.


     Soil moisture and temperature are very important in the
growth of crops, so soils are classified partly in terms of these
properties (8).   Based on mean annual soil temperature (MAST),
soils of temperate regions are  classified as Frigid (MAST-0-8°C),
Mesic (MAST-8-15°C),  Thermic (MAST-15-22°C) or Hyperthermic
(MAST>22°C).   Soil moisture regimes are basically defined by
groundwater level and proportion of the year in which moisture is
held at tensions of 1500 kPa or less,  the maximum tension at
which common crops can grow.   Most U. S.  soils are classified as

Aridic, Ustic or Udic in order of increasing length of the
growing season with moist soil.   In addition,  the Xeric soil
moisture regime occurs in areas of Mediterranean climate, i. e.,
moist cool winters and hot dry summers, in contrast to moist
summers and dry winters of the Ustic regime.

     Figure 2 gives a generalized distribution of soil
temperature and soil moisture regimes for the U.S., compiled from
references 9-12.  In the eastern and central U. S., soil
temperature regimes follow a simple pattern of Frigid, Mesic,
Thermic and Hyperthermic from north to south.   East of about
97°W Long., soils are relatively moist (Udic),  and westward to
the Rocky Mountains they are Ustic.   In the western U. S.,
Frigid regimes extend southward along the Rocky Mountains, and
Aridic regimes are common in the southwest.   Regions of Xeric
soil moisture regime are common in California and nearby areas.

     Based on Figures 1 and 2, soils with identical values of
CRa, E, D and P would be expected to increase in CRn in the
sequence Aridic-Ustic-Udic eastward across the U. S, and to
increase in the order Frigid-Mesic-Thermic-Hyperthermic
southward.   Table 1 lists a few sites for which the relevant
parameters have been measured or estimated.   For most sites in
regions of Udic moisture regime, the observed Rn (especially the
maximum observed Rn) exceeds the values calculated for dry soil
by factors of 1.5 to 5, as expected for soils of Udic moisture
regime.   For site 14-80, where measurements of moisture and
temperature have been made, the values calculated from equation
(1) match the measured values within measurement error (6).   The
sole exception, site 14-82, is strongly affected by water
saturation effects for much of the year.    In contrast, at the
site at Socorro, NM, in an Aridic moisture regime, the calculated
and observed Rn agree closely.   The moisture content of this
soil is reported as 4 wt.%, equivalent to F-0.17, a reasonable
value for the aridic regime.   For the site at Denver, CO, the
observed Rn levels for moist seasons distinctly exceed the values
calculated for dry conditions, and during dry seasons the soil
cracks to allow air penetration to depth.   These data are
consistent with a significant moisture effect.

     Extension of the relations discussed above leads to the
hypothesis that radon in houses should vary according to the
moisture and temperature regimes, if not obscured by variations
in CRa, E, D and P, and flux of soil air.  This hypothesis was
tested using statewide averages of Rn in homes from a compilation
of 175,000 measurements by The Radon Project of Pittsburgh, PA,
as compiled by Dr. Bernard Cohen of the University of Pittsburgh.

     The data on houses do not fit the hypothesis.   In eastern
and central U. S., the highest values are in IA, SD, and several
other northern plains states with an Ustic moisture regime and

Frigid or Mesic  thermal regime, rather than in states with a
Udic-Thermic regime.   Values  tend to be low in most states of
southeastern U.  S. with a Thermic temperature regime.   Likely
explanations of  the failure of the hypothesis are regional
differences in Ra and U, lower indoor-outdoor temperature
differences in the southeast leading to lower pressure gradients,
regional differences in home construction (specifically the use
of slab-on-grade construction  in the southeast vs. basements in
northern states), and error from grouping by state. Also, in
areas of high soil moisture, the air permeability may be greatly
decreased by blockage of small pores by water.

     Local differences in drainage and moisture are probably also
of significant importance in soil gas Rn.  Many relatively flat
valley areas are floored by alluvium or soil that is water
saturated at the surface, to form the Aquic moisture regime.
Other poorly drained areas may have a thin zone of moist soil
above the water  table.  Radon content of soil gas is expected to
be relatively high in such areas, but air permeability may be so
low that little Rn transport is possible.


     Soils literature also allows estimates of seasonal change in
moisture and temperature, leading to estimates of seasonal
changes in Rn concentrations.   Tables 2, 3, and 4 give these

     The soil moisture states and soil moisture tension ranges
given in Table 2 were developed by the USDA Soil Conservation
Service (14).  The data given in Table 2 for the sandy loam and
silty clay loam was calculated as follows:   A bulk density of
1.60 g/cm  (sandy loam) and 1.45 g/cm  (silty clay loam) was used
to calculate (assuming a particle density of 2.65 g/cm ) a total
pore space of 40% for the sandy loam and 45% for the silty clay
loam.   Soil moisture content based on data from Petersen et al.
(15) and Ciolkosz and Dobos (16) for 1500 and 33 KPa tensions was
used to interpolate the values given in Table 2 using the soil
moisture tension/moisture content relations given by Thome and
Thorne (17).  These data were used to compute the percent pore
space filled with water (% soil moisture x bulk density/total
pore space).  Table 3 gives estimated winter and summer soil
moisture regimes for the upper 1.5 meters of soils of the United
States.   These data are based on numerous soil moisture state
evaluations (18,  19)  for various soils of the United States.   The
soil temperatures given in Table 4 were calculated from mean
monthly air isotherms (20) using the method of van Wambeke (21)
for the various soil temperature/moisture regions given in Figure
3.  This method may give conservative values for bare soil areas
(22),  therefore the soil temperatures may have a greater range
than given for the arid area.

     These data have been used Co derive the estimates of Q shown
in Table 5.   Several conclusions of interest may be drawn from
this data:
     1. The largest seasonal effects of moisture and temperature
are expected in the Udic moisture regime.    In the warmer parts
of Udic regions, winter Rn values may be more than double the
summer values.
     2. Summer Rn values higher than winter values are predicted
for the Frigid-Udic regime, and are possible in the Mesic-Udic
regime.   This behavior has been recognized at several sites in
central Pennsylvania (14-80, 14-81, 14-82, 14-83), which is on
the cold edge of the Hesic thermal regime.
     3. Summer Rn exceeding winter Rn is possible in the Ustic
regime, but in general, large changes are not expected.
     4. In the aridic regime, moisture is low enough that
seasonal effects are very small, but typically winter levels will
slightly exceed summer values.
     5. In the Xeric regime, winter Rn values are expected to
markedly exceed summer values.
     6. Seasonal effects are expected to be much larger in clay-
rich soils than in sandy soils, because of the generally higher
moisture saturation of clay-rich soils.


     Although many effects of water on Rn in soil gas are
proposed, the effect of varying proportion of pore space occupied
by water appears to be among the largest and most universal.
Temperature also affects soil gas Rn because of its effect in
varying the partition of Rn between water and air.  The process
affects moist soils at all depths, but can be evaluated best for
depths where diffusion toward the surface is negligible.

     In the soils literature, major regional differences in soil
moisture and temperature are well documented, and have been
compiled as a map.  The meager data on Rn in soil gas agrees with
the proposed moisture temperature effects.  However, within
houses the effects of soil moisture and temperature are masked by
other Rn-determining factors (regional differences in Ra, house
construction, house pressurization, etc.) so that a clear
regional effect of the soil moisture and temperature is not
obvious in a large data set on Rn in houses.

     Seasonal variability in soil gas Rn has been predicted from
estimated temporal soil moisture and temperature variations.
Large seasonal variations are predicted for soils of the Udic
moisture regime, characteristic of the eastern U.S.  For the
northeastern U.S., from New England through Minnesota, summer
soil gas radon is predicted to exceed winter values, as found at
several sites in Pennsylvania.  Contrastingly, in soils of the

Xeric moisture  regime,  distinctly higher values  are predicted  in
the winter  than the  summer.  Clay-rich soils are more  susceptible
to moisture-induced  elevated Rn  than are sandy soils.  However,
local variations  of  moisture caused by variability in  drainage
are undoubtedly also of considerable importance.

     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.


     The authors  acknowledge the financial support of  the U.S.
Department of Energy (Grant DE-FG02-87ER60577).


1. Environmental  Protection Agency. A citizen's guide  to
   radon. USEPA,  Doc. OPA-86-004, 1986. 13 pp.
2. Tanner, A.B. Radon migration  in the ground: A review. In:
   J.A.S. Adams and W.M. Lowder  (ed.),  The Natural Radiation
   Environment. University of Chicago Press.  1963. p.   161.
3. Tanner, A.B. Radon migration  in the ground: A supplementary
   review. I_n:  T.F. Gesell and W.M. Lowder (ed.), The Natural
   Radiation Environment III. DOE Symposium Series,  CONF-780422.
   1980. p. 5.
4. Schery, S.D., Gaeddert, D.H.,  and Wilkening,  M.H.  Factors
   affecting exhalation of radon from a gravelly sandy loam.  J.
   Geophys. Res. 89:  7299, 1984.
5. Schumann, R.R., Owen, D.E.,  and Asher-Bolinder, S.  Weather
   factors affecting soil-gas radon concentrations at a single
   site in the  semiarid western U.S..  In:  Proceedings of the  1988
   E.P.A. Symposium on Radon and Radon Reduction Technology.  EPA-
   600/9-89/006B, U.S.  Environmental Protection  Agency,
   Cincinnati, Ohio,  1989. p. 3.1.
6. Washington, J.W. and Rose, A.W.  Regional and  temporal
   relations of radon in soil gas to soil  temperature and
   moisture. Geophys. Res. Lett,  in press:  1990.
7. Nielson, K.K., Rogers,  V.C.,  Mauch,  M.L.,  Hartley,  J.N.,  and
   Freeman, H.D. Radon emanation characteristics  of  uranium mill
   tailings. I_n: Symposium on Uranium Mill  Tailings  Management.
   Colorado State University, Fort  Collins,  CO,  1982.  p.  355.
8. Soil Survey Staff. Soil Taxonomy.  Agricultural Handbook No.
   436,  Soil Conservation Service,  U.S.  Department of
   Agriculture,  1975. 754 pp.
9. Soil Survey Staff. Soil areas  of the midwest  region. Soil
   Conservation Service, U.S. Department  of Agriculture.  Midwest
   Nat.  Tech. Center, Lincoln,  NE.  1972.  1  page  map.
10.  Soil Survey Staff.  Soil moisture regime and  soil  temperature
    map. Soil Conservation Service,  U.S.  Department  of

    Agriculture.  South Nat.  Tech.  Center,  Fort Worth,  TX.  1975.  1
    page map.
11.  Soil Survey Staff. Composite map of soil  moisture  and
    temperature as presented by individual states.  Soil
    Conservation Service,  U.S.  Department  of  Agriculture.  West
    Nat. Tech.  Center, Portland, OR. Draft of the  WRRC-50
    committee  project. Map and  text.
12.  Smith,  H.  Soil temperature  regimes of  the northeastern United
    States.  Soil Conservation Service, U.S.  Department of
    Agriculture.  Northeast Nat. Tech.  Center, Chester, FA. 1984.
    1 page  map.
13.  Schumann,  R.R., and Owen, D.E.  Relationships between geology,
    equivalent uranium concentration,  and  radon in soil gas,
    Fairfax County, Virginia. Open-File Report 88-18,  Geological
    Survey,  U.S.  Department  of  the  Interior,  Denver,  Colorado,
    1988. 27 pp.
14.  Quandt,  L.A.  and Grossman,  R.B.  Soil water states  for  select
    soils in the northeast region.  Soil Conservation  Service,
    U.S. Department of Agriculture.  Northeast Nat.  Tech. Center,
    Chester, PA.  1983. 51  pp.
15.  Petersen,  G.W., Cunningham, R.L.,  Hatelski, R.P.  Moisture
    characteristics of Pennsylvania soils: I. Moisture tension as
    related to texture. Soil Sci.  Soc. Am. Proc. 32:  271,  1968.
16.  Ciolkosz,  E.J., and Dobos,  R.R.  Penn State Soil
    Characterization Laboratory Data Base. Department  of
    Agronomy,  Penn State University. University Park,  PA.
17.  Thorne,  D.W.  and Thorne, M.D.  Soil, water, and crop
    production. AVI Pub. Co., Westport, CT.  1979.  353  pp.
18.  Soil Survey Staff. Proceedings  of the  1981 national technical
    work-planning conference of the cooperative soil  survey.  Soil
    Conservation Service,  U.S.  Department  of  Agriculture,
    Washington, D.C., 1982.  p.  19.
19.  Soil Survey Staff. Proceedings  of the  1983 national technical
    work-planning conference of the cooperative soil  survey.  Soil
    Conservation Service,  U.S.  Department  of  Agriculture,
    Washington, D.C., 1983.  p.  114.
20.  Environmental Data Service. Climatic atlas of  the  United
    States.  Superintendent of documents, Environmental Science
    Services Administration, U.S.  Department  of Commerce.
    Washington, D.C. 1968.
21.  van Wambeke,  A. Calculated  soil moisture  and temperature
    regimes of South America. SMSS  Tech. Monogr. No.  2, Soil
    Conservation Service,  U.S.  Department  of  Agriculture,
    Washington, D.C. 1981.
22.  Murtha,  G.G.  and Williams,  J.  Measurement, prediction, and
    interpretation of soil temperature for use in  soil taxonomy:
    Tropical Australian experience.  Geoderma. 37:  189, 1986.

                0.0       0.2      0.4      0.6
                            Fraction Saturation (F)
       Figure 1: Values of  Q  (factor by  which soil gas Rn  is
                  increased)  as  a function of temperature  (T°C) and
                        fraction  moisture  saturation (F)
                                                                    F- Frigid
                                                                    M- M..IC
                                                                    T • Thtrmle
                                                                    H • Hyp.rth.rmlc

                                                                    u • Uitlc
                                                                    0 -Arldlc
                                                                    X • Xerlc
      Figure  2:  Generalized distribution  of  soil moisture
                 temperature regimes  in the U.S.

Table 1.  Comparison of calculated and observed radon in soil gas.
Site* CRab(pCi/g)
14-80 (PA)
14-81 (PA)
14-82 (PA)
14-83 (PA)
6-10 (PA)
NC-1 (NC)
Denver .
Co, VA1









P Calc(dry)d






obs . mean
51 Oh


obs , max



Moisture^ Reference



This study
This study
This study
This study
This study
This study



a.  Sites 14-80 through 14-83 are in Centre Co., PA; site 6-10 is in Berks Co.. PA; and NC-1  is  at
    Justice, NC.
b.  Radium or uranium in soil.
c.  Measured value, or assumed as 0.2 if in parentheses.
d.  Calculated from equation 1.
e.  Mean value for 1-2 years by alpha-track detector at sites of this study.
f.  Maximum reported value for alpha track detectors exposed 2-3 months, or single reported value.
g.  Moisture regime:  ud,  udic; u, ustic; a, arldic
h.  Alpha track detector submerged part of year.
i.  Based on samples with 1.5 to 2.5 ppm eU on plot of soil gas Rn vs eU (Figure 7).  The observed  "mean"
    is the approximate median.

Table  2.   Soil moisture  states, percent moisture,  and percent saturated  pore  space for  a
           typical  sandy  loam  and silty clay  loan  soil.
Soil Moisture
State ~~ Tension~KPay
Very Dry (VD) 100, 000*-10 . 000
Slightly Dry (SD) 10.000-1,500
Slightly Moist (SM) 1.500-200
Moderately Moist (MM) 200-33
Very Moist (VM) 33-1
Wet < 1



Pore space


Silty Clay Loam
Pore space
Moisture* Saturated**

0-2 0-6
2-12 6-39

12-18 39-58
18-23 58-74
23-31 74-99
*  Air dry  (50% relative humidity); oven dry = 1,000,000 KPa

+  by weight

•ft by volume
Table 3.  Estimated average soil moisture states for the soil temperature  regimes  of  the
          United States.  See Table 2 for abbreviations.

Soil Moisture States
Winter Summer

Table 4.  Approximate mean January and July air and soil temperature  (degrees C at 51  cm
          depth) for the various soil temperature-moisture regime  areas  of the United
          States.  Mean annual soil temperature:  Frigid 0-8°C; Masic B-15°C; Thermic 15-
          22°C;  and Hyperthermic > 22°C mean.
Soil Temperature Regime
Air and Soil Temperature
Hffier. thermic


































                                                                           Eastern States





    Table 5.  Estimated winter and summer values of Q for soil moisture-temperature regimes.
                              Winter (Jan.)
                                        Summer  (July)
    Regime*    T°C
Midpoint   T"C
 Midpoint  Season
A. Silty
B. Sandy
Clay Loam
.74-. 99
.74-. 99
. 74- . 99
.74-. 99
.58-. 74
.58-. 74
.58-. 74
.39-. 58
.39-. 58
.39-. 58
.06-. 39
.74-. 99
.74-. 99
.58-. 74

.74-. 99
.74-. 99
.74-. 99
.74-. 99



.58-. 74
.58-. 74
.39-. 58
.39-. 58
.39-. 58
.39-. 58
.39-. 58
.39-. 58
.06-. 39
.06-. 39
.06-. 39
.06-. 39
.58-. 74
.06-. 39

.58-. 74
.58-. 74
.39-. 58
.39-. 58



    a.  Moisture-temperature regime.  Temperature regimes:  F, Frigid; M. Mesic; T, Thermic; H, Hyperthermic;
        Moisture regimes:  ud.  udic; u,  ustic; a. aridic; x, xeric.
    b.  Season with highest predicted Rn:  S, summer; W, winter; O, subequal.