United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/SR-96/045 August 1996
EPA Project Summary
Site-Specific Protocol for
Measuring Soil Radon
Potentials for Florida Houses
Kirk K. Nielson, Vern C. Rogers, and Rodger B. Holt
The full report describes a protocol
for site-specific measurement of radon
potentials for Florida houses that is
consistent with existing residential ra-
don protection maps. The protocol
gives further guidance on the possible
need for radon-protective house con-
struction features. In applying the test
results, the user should also consider
the relative costs of using conserva-
tive radon controls and the EPA guid-
ance on further reducing radon levels
even in the range below 4 pCi L~1.
The measurements included in the
protocol were selected from sensitivity
analyses of radon entry into the same
reference house as was used to de-
velop the radon protection maps. The
sensitivity analyses also used the same
RAdon Emanation and TRAnsport into
Dwellings (RAETRAD) model, provid-
ing a common basis to that of the maps.
The sensitivity analyses identified ra-
dium concentration, soil layer depth,
soil density, soil texture, and water table
depth as the independent parameters
dominating indoor radon. Radium con-
centration and water table depth were
most important. Soils up to 2.4 m deep
contributed to indoor radon in uniform-
radium scenarios, and soil layers about
0.6 m thick significantly affected radon
in cases of nonuniform radium distri-
butions.
A conservative upper limit for radon
potentials was defined as the 95% con-
fidence limit for radon in the reference
house, corresponding to the radon pro-
tection map definition. The number of
samples needed to represent a site was
determined from equivalent regional
map precisions to be 20 samples per
4,000 m2 (1 acre). The samples are taken
at four depth increments extending to
2.4 m. Equivalent precisions for a
smaller parcel of land can use fewer
total samples.
The site-specific protocol involves
drilling five boreholes in each 4,000 m2
parcel of land. Twenty soil samples ob-
tained from the borings are used for
radium measurements. The soil borings
are also used for density and textural
estimates unless default values are cho-
sen. A soil radon measurement detects
the potential presence of anomalous
elevated-radium materials at depth. The
site water table depths are defined from
soil survey data or site observations.
The site measurements are analyzed
by a special-purpose computer code
called RAETRAD-F to determine the
upper limit for indoor radon concentra-
tions in the reference house. Using the
same category definitions as for the
residential radon protection map, the
RAETRAD-F code determines whether
the site is in the low, intermediate, or
elevated map category for radon pro-
tection purposes.
This Project Summary was devel-
oped by EPA's National Risk Manage-
ment Research Laboratory's Air Pollu-
tion Prevention and Control Division,
Research Triangle Park, NC, to an-
nounce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
The Florida Department of Community
Affairs (DCA) has developed prescriptive
building standards to reduce radon-related
health risks. The standards require pas-
-------
sive radon barriers and active sub-slab
ventilation to reduce radon entry from soils
that can cause elevated radon levels. Re-
gions requiring these radon controls are
identified by radon protection maps that
show where soils require no special con-
trols, passive radon controls, or passive
and active radon controls. Despite the con-
venience and low cost of using the maps,
it is sometimes desirable to directly mea-
sure the radon potential category of a
site. Although not generally required, site-
specific analyses can give valuable guid-
ance on alternative planning or anoma-
lous site conditions. Alternative planning
decisions should also consider the rela-
tive costs of using conservative radon con-
trols versus testing, as well as the EPA
guidance on further reducing radon levels
even in the range below 4 pCi L~1.
The full report builds on a previous study
of methods for measuring site-specific ra-
don potentials as a consistent alternative
to the radon protection map. It examines
more closely the fundamental parameters
controlling radon potential and identifies
from model simulations the minimum mea-
surements to characterize radon poten-
tial. The measurement requirements are
used to define a field sampling and mea-
surement protocol for characterizing site
radon potentials with the appropriate level
of sampling and replication. The measure-
ments are analyzed by the RAETRAD
model to find a reference-house radon
potential that is consistent with the resi-
dential radon protection map.
Model Simulations
Model simulations of radon generation
and movement from soils into houses were
used to estimate the dependence of in-
door radon levels on different site param-
eters. The simulations focused on long-
term average values to avoid the tempo-
ral radon fluctuations from varying soil and
house conditions. Radon potentials for
open land used the same reference house
and approach as the soil radon potential
maps to remove house variables from the
definition of radon potential. The hypo-
thetical reference house was simulated
on the individual soil profiles that occur in
different regions, with radon generation
and movement from the soils into the ref-
erence house. This approach determines
the geographic soil contributions to indoor
radon without the complicating effects of
individual house properties.
The radon simulations also used soil
and geology definitions that were consis-
tent with the prior radon map analyses.
For example, soil radium distributions were
defined to be uniform throughout the strata
in an upper layer about 2.5 m thick, and
they were also defined to be uniform
throughout a lower, geology-dominated
zone that was also about 2.5 m thick. Soil
water distributions throughout both zones
were determined from drainage properties
and water table depths.
The RAETRAD model was used to simu-
late radon generation and transport from
the soil profiles into the reference house.
RAETRAD is a steady-state, numerical-
analytical computer code that simulates
(a) radon generation and decay in soil
regions around a house and in the house
understructure (e.g., floor slab, footings);
(b) diffusive and pressure-driven advec-
tive movement of radon through the soils
and the house understructure; and (c) ra-
don accumulation in a single-chamber
house. RAETRAD's multiphase radon
source and transport equations explicitly
represent the solid soil particles, pore wa-
ter, and air-filled pore space.
The reference house corresponds to the
one used previously for radon potential
maps, giving the site-specific analyses a
common basis for estimating equivalent
radon potentials. The house is a 143-m2
(1,540-ft2) rectangular, slab-on-grade struc-
ture with the approximate characteristics
of Florida single-family dwellings. The
dominant characteristics affecting indoor
radon include the indoor air pressure, the
indoor-outdoor air exchange rate, and the
type and quality of foundation design and
construction. The house has a floating-
slab floor design, causing a shrinkage
crack about 0.5 cm wide at the slab pe-
rimeter. Other properties of the house are
listed in Table 1.
Surface and deep soils were defined for
sensitivity analyses to have a sandy tex-
ture, a density of 1.6 g cnr3, and a poros-
ity of 0.407. Soil air permeability and ra-
don diffusion coefficients were defined by
RAETRAD from porosity, water content,
and mean particle diameters using empiri-
cal correlations. Soil water contents were
defined by the distance of each 30-cm
soil layer above a 5.2-m deep water table.
Soil radon emanation fractions were de-
fined from an empirical trend with radium
concentrations for consistency with previ-
ous radon map calculations.
Radon Sensitivity Analyses
Sensitivity analyses with the RAETRAD
model determined which site parameters
strongly affect radon potentials and which
have smaller influences. Five main pa-
rameters were analyzed: soil texture,
depth, radium, density, and water table
depth. The depth parameter was applied
only to radium distributions since water
contents and their dependent parameters
already varied with depth from the water
table. For consistency with previous analy-
ses, surface and geologic soil zones com-
prised the depth parameters for the radon
simulations. Soil gas radon was also ana-
lyzed as a possible surrogate for deep-
soil radium concentration to detect poten-
tial high-radium anomalies beneath the
sampling depth range.
Sensitivity analyses determined the net
radon concentration in the reference house
while individually varying each of the five
independent parameters. Analyses first
were performed in which the entire soil
column consisted of one of the 12 Soil
Conservation Service (SCS) textural clas-
sifications. The resulting changes in cal-
culated moistures, diffusion coefficients,
and permeabilities caused decreased ra-
don concentrations for the fine-textured
soils, as shown in Figure 1.
A semi-quantitative approach helped in-
terpret the sensitivity analyses. The rela-
tive sensitivity of each parameter was es-
timated from a log-log plot of radon ver-
sus the parameter of interest. Although
the textural classifications had no direct
numeric scale for such a plot, they were
assigned integer values for comparison
purposes, starting at 1 for sand and in-
creasing to 12 for silt. The slopes of the
lines on the plot, near the parameter value
of interest, correspond to the power-curve
(y = axb) dependence of each parameter.
This approach helped rank the relative
importance of each parameter for site-
specific measurements. Figure 2 compares
the sensitivity of soil texture with that of
the other parameters.
The sensitivity to surface soil emanat-
ing radium concentrations was found to
be highest, with a power-curve exponent
of b = 0.81 at Ra-E = 1 pCi g-1 of emanat-
ing radium. As shown in Figure 2, this
sensitivity increases at higher radium con-
centrations, approaching b = 1.0 and po-
tentially raising indoor radon levels above
20 pCi L1 as the emanating radium con-
centration approaches 5 pCi g-1. The sen-
sitivity of the deeper, geologic soil radium
level is smaller, with an exponent of b =
0.13 at Ra-E = 1 pCi g~1 of emanating
radium. Its sensitivity also increases at
higher radium concentrations but remains
lower than that for surface soils because
of the decreasing sensitivity of radon
sources that are farther from the house
foundation.
The sensitivity to soil density is moder-
ate at the low-density end of the curve in
Figure 2, but it diminishes at higher soil
densities. The strong dependence at low
densities is dominated by the emanation
-------
Table 1. Reference House Parameters Used in Radon Entry Simulations
Parameter
Footprint dimensions
Interior volume
External ventilation
Floor crack width
Floor crack location
Fill soil thickness
Indoor air pressure
Value
8.6 x 16.5 m
350 m3
0.25 tr1
0.5 cm
Slab perimeter
30 cm
-2.4 Pa
Parameter
Slab water/cement ratio
Floor slab thickness
Floor slab porosity
Slab 226Ra emanation
Exterior footing depth
Slab air permeability
Slab Rn diffusion
Value
0.55
10 cm
0.22
0.07 pdg-1
61 cm
1x10-" cm2
8x1 0-4 cm2 s-'
o
Q.
c
g
"ro
o
O
c
o
T3
ro
or
o
o
T3
^ ^
of &
Water Table: 5.2m
Soil Radium: 1.8 pCi/g (E = 0.47)
Soil Density: 1.6 g/cm3
^ ^
-------
100
o
g
^4-*
(0
o
O
c
o
T3
(0
_
O
O
T3
10
0.1
Surface Soil
Radium:
b=0.81
at RaE=1pCi/g
Geologic Soil
Radium:
b=0.13
at RaE=1pCi/g
Density:
b=0.35
at
1.6 g/cm3
0.1 1
Figure 2. Comparison of parameter sensitivities.
10
Parameter
100
1000
coefficient both showed a square-root de-
pendence (b = 0.5).
Mathematical Definition of Site-
Specific Radon Potential
The radon potential for a residential
building site is defined mathematically as
it was for the residential radon protection
map. It is the calculated radon concentra-
tion in the reference house that would not
be exceeded if the house were placed on
at least 95% of the land areas in a geo-
graphic region (map polygon). Stated more
simply, it is the 95% confidence limit of
radon concentration for the reference
house. Although site measurements give
a more representative estimate of radon
in the reference house, site variations and
measurement uncertainties define a distri-
bution of possible radon concentrations.
This distribution is defined to be log-nor-
mal, consistent with the mapped distribu-
tions, and is used to estimate the 95%
confidence limit for radon that is consis-
tent with the maps.
The equation for site-specific radon po-
tential is defined from site-specific values
of the log-normal radon distribution pa-
rameters. An annual-average radon con-
centration is first calculated for the refer-
ence house from seasonal values using
the RAETRAD algorithm with the geomet-
ric means of the measured site param-
eters. A geometric standard deviation
(GSD) is then calculated to represent the
indoor radon variability associated with the
distributions of the site-specific parameters.
The radon concentration in the reference
house at the 95% confidence limit is fi-
nally calculated from the estimated geo-
metric mean and GSD to estimate the
site-specific radon potential as:
(1)
where C95 = site-specific radon poten-
tial, at 95% confidence
limit, for residences
(pCi L1)
C = net radon concentration
gm
calculated for the refer-
ence house from geomet-
ric means of the site
parameters (pCi L1)
G = GSD of the indoor radon
distribution (dimension-
less)
1.645 = inverse of the normal
distribution integral at
95% (standard devia-
tions).
The GSD of the indoor radon distribu-
tion is estimated from a sensitivity-weighted
sum of the variances of the individual pa-
rameters that contribute to the total uncer-
tainty in the indoor radon calculation. Since
the distributions are log-normal, the com-
ponent uncertainties are GSDs, which re-
quire log-transformation for the usual qua-
dratic addition of uncertainties, giving:
G = exp
(2)
-------
where g. = GSD of parameter i
b: = power-curve dependence of
indoor radon on parameter
i.
The b: weighting factors give the correct
weighting to the GSD of each parameter
in proportion to the influence of the pa-
rameter on indoor radon concentrations.
The main parameters contributing to
uncertainty in Equations 1 and 2 are the
soil radium concentrations, the water table
depths, and the soil density and texture.
Soil air permeability and radon diffusion
coefficient also have inherent uncertain-
ties. Since the uncertainties from density,
permeability, and diffusion vary less with
site properties, they are represented in
Equation 2 by a constant, composed of
respective contributions of 0.001, 0.173,
and 0.120.
The dependence on soil radium is ap-
proximately linear (bt = 0.8 in Figure 2, but
approaches unity at higher radium levels),
and it is given a linear dependence (b: =
1) for use in Equation 2. Similarly, the
GSD for water table variations is assigned
3^=1 sensitivity, which corresponds to
water table depths slightly shallower than
1 m. Equations 1 and 2 can be written in
a combined form with these numerical val-
ues as:
+[ln(gwt)]2}
(3)
where 0.294 = combined variance from
permeability, diffusion,
and density (dimension-
less)
gRa = GSD of site radium con-
centrations (dimension-
less)
gwt = GSD of indoor radon
from varying site water
table depths (dimension-
less).
The minimum number of soil samples
needed to characterize a site is domi-
nated by the soil radium distributions, since
the other parameters have more predict-
able variations. The number of soil
samples must be sufficient to characterize
both the geometric mean and the GSD of
the soil radium distribution. The number
must also be adequate to limit the uncer-
tainty in the GSD to a value consistent
with the radium variability associated with
the radon protection map.
Soil radium concentrations in Florida are
distributed log-normally, so their logarithms
are normally distributed, and the standard
deviations of their logarithms follow a chi-
squared distribution. The minimum num-
ber of samples required to determine a
geometric mean with a given degree of
precision depends on the upper limit of
the GSD. The required number of soil
samples therefore can be expressed as a
fractional uncertainty that represents the
interval between the upper limit and the
best estimate (geometric mean) as:
where n
n-i
F'"
(4)
= number of samples
= chi-squared value for n-1
degrees of freedom and
for (1-p) confidence
= probability of exceeding
a given chi-squared value
= fractional uncertainty
required (consistent with
radon protection map).
Solving Equation 4 for n gives:
« = ! + (! +n)2 **_„_„ (5)
which must be solved iteratively because
the XP,n-\ values also depend on n. For a
4,000-m2 (1-acre) area, statewide radium
concentrations had typical uncertainties in
the 25 to 35% range. Using an uncer-
tainty limit of 25% (u = 0.25) for the site-
specific soil radium data, a 90% confi-
dence limit for the radium GSD (p = 0.1),
and solving for the 95% confidence limit
on the transformed radium data gives a
minimum value of n = 19. For practical
purposes, the minimum n is rounded to
20. Thus, at least 20 soil radium samples
must be collected and analyzed for a
4,000-m2 (1-acre) site to have similar con-
fidence in the site-specific radon potential
as in the radon protection map.
The sensitivity analyses also give infor-
mation on how the 20 soil samples should
be distributed. For example, the maximum
depth from which soil generally influences
indoor radon is about 2.4 m (8 ft), if the
soil is uniform and has the limiting sandy
texture. The sensitivity to non-uniform lay-
ers further shows that samples should be
collected at intervals not exceeding about
0.6 m (2 ft). Therefore, if 2.4-m (8-ft) bore-
holes are sampled with compositing over
0.6-m (2-ft) intervals, four samples will be
obtained from each borehole and a mini-
mum of five boreholes will be required to
represent the 4,000 m2 unit area.
For site areas smaller than 4,000 m2,
fewer soil samples can be considered.
However, fewer samples increases the
uncertainty in the upper limit for the GSD,
which also increases the calculated radon
potential compared to the value that would
be calculated if 20 samples were used. A
mathematical accommodation is made to
use fewer soil samples for sites smaller
than 4,000 m2. The approach transforms
the GSD among the actual number of
samples to an equivalent 20-sample GSD.
The transformed GSD is larger to main-
tain an equivalent upper confidence limit.
The transformation requires that the site
have a slightly lower Cgm than if 20 samples
were analyzed to compensate for the in-
creased uncertainty in satisfying the pre-
scribed C95 cut point limits.
The calculation of site-specific radon
potentials from site measurements uses a
specialized version of the RAETRAD com-
puter code. The computer code, called
RAETRAD-F, first calculates the geomet-
ric means and GSDs of the measured soil
radium concentrations, and determines the
seasonal water table distribution as it was
defined for the radon protection maps. It
then compares the measured soil radon
concentration with concentrations calcu-
lated from the radium measurements and
uses the larger of the two to extrapolate
the deep-soil radium concentrations
throughout the 2.4 - 5.0 m range.
The RAETRAD-F code next computes
best estimates of indoor radon concentra-
tions for the different seasonal water table
conditions using the geometric means of
the radium measurements. After determin-
ing the geometric mean annual radon con-
centration, Cgm, from the seasonal values,
the variations among seasonal conditions
and among the radium concentrations are
used in Equation 3 to estimate C95. If less
than five boreholes were used, the code
also modifies gRa. The resulting radon pro-
tection category is determined by compar-
ing C95 to the 4.0- and 8.3-pCi L1 cut
points used in the radon protection map
to determine whether the site is in the
low, intermediate, or elevated radon pro-
tection category.
Site-Specific Radon Potential
Measurement Protocol
A protocol has been developed for mea-
suring the radon protection category of a
building site. The protocol is an alterna-
tive to the radon protection map and must
only be performed with standard methods
after any site recontouring that could af-
fect the water table or the distribution of
soils. The protocol applies to land areas
-------
of 4,000 m2 (1 acre) or smaller. Larger
areas must be divided into < 4,000 m2
parcels for using this protocol.
Site Sampling and
Measurements
Site soil sampling must use five bore-
holes spread over the site at potential
building locations. Sites smaller than 4,000
m2 require at least one borehole for every
potential residential building location. If the
site is an individual lot, as few as one
sampling borehole may be used if it is
supplemented with two additional soil
samples at least 10 m away from the
borehole and from each other, and from
the 0-61 cm depth interval (representing
horizontal and vertical variations). Soils
shall be collected from each borehole to
represent the 0-61, 61 - 122, 122- 183,
and 183 - 244 cm depth intervals. The
borehole samples and any supplementary
samples shall be used for measurement
of density and for textural classification.
The remaining material from each depth
interval shall be composited for individual
measurements of radium concentration.
The concentration of radon in the soil gas
shall be measured at or near one or more
borehole sites. Water table depth must be
determined on the site or on nearby prop-
erty.
Concentrations of 226Ra shall be mea-
sured from gastight sealed, equilibrated
aliquots of individual samples using a cali-
brated gamma-ray spectrometer. The
spectrometer shall be calibrated by analy-
ses of standard reference materials and
blanks in identical configurations. The con-
centrations of 226Ra shall be reported indi-
vidually in picocuries per gram on a dry
mass basis.
The in-situ soil density shall be mea-
sured by the drive cylinder method, or by
other methods approved by the DCA, us-
ing calibrated equipment. Soil density shall
be reported in grams per cubic centimeter
(dry mass basis) as 20 individual mea-
surements, as four layer means, or as an
overall site mean. If the in-situ soil density
is not measured at the site, a default value
of 1.5 g cnr3 shall be used in the analyses
for computing the site radon protection
category.
Soil texture classification shall use labo-
ratory or field methods to select one of
the 12 SCS classes: sand, loamy sand,
sandy loam, sandy clay loam, sandy clay,
loam, clay loam, silty loam, clay, silty clay
loam, silty clay, or silt. The classifications
should be reported for all 20 samples, or
for homogeneous sites, for the four layer
classes, or for the entire site. Because of
the conservative results obtained with the
sand classification and the prevalence of
sandy soils throughout Florida, soil tex-
ture classifications need not be performed
if a default classification of "sand" is used
in the analyses.
Soil gas 222Rn concentrations shall be
sampled by drawing soil gas from a driven
tube and measuring the 222Rn concentra-
tion with a calibrated device. The soil ra-
don measurements must be made near
one or more borehole locations at 1.2 m
or deeper. To avoid radon perturbation,
the gas samples shall be collected before
soil boring, or afterward if they are about
2 - 3 m away from the bore holes. If the
water table is shallower than 1.2 m at the
time of soil gas sampling, soil radon mea-
surements may not be required, if (a) the
minimum water table depth is less than
0.6 m and (b) there is no evidence that
the water table would be below 1.2 m
during an alternative season.
The water table depth shall be speci-
fied, as for the radon protection map, as
the minimum water table depth (in centi-
meters) and duration (in months). These
data may be obtained from the STATSGO
data base, from county soil survey data,
or from measurements during at least four
seasons at 3-month intervals. The shal-
lowest of these shall be defined as the
minimum water table depth, and a mini-
mum duration of 3 months shall be de-
fined unless a longer time is indicated by
the measurements.
Data Analysis
The site-specific measurements shall be
assembled and analyzed using the
RAETRAD-F computer code. The
RAETRAD-F code automatically imple-
ments the equations in this report with the
reference house specifications in a way
that corresponds to the calculations per-
formed for the residential radon protection
map. The user is prompted to enter the
site identification information and individual
site measurement data. The code then
computes the appropriate statistical pa-
rameters for the radium measurements,
the annual water table distribution from
the water table data, the soil moistures
from the texture and density data, and all
other parameters required for the annual
average indoor radon distribution for the
reference house. From this distribution, a
C95 value is computed and compared to
the 4.0- and 8.3-pCi L1 cut points, and the
site is designated to have either low, in-
termediate, or elevated radon potential.
The printed output from the RAETRAD-F
code includes the user-specified input pa-
rameters, the calculated C95 value, and
the site radon potential designation.
Summary and Conclusions
The site-specific protocol and basis de-
scribed in the full report can be used to
measure and interpret features at a site to
give supplementary guidance on the pos-
sible need for radon-protective construction
features. Although site-specific tests can
help in making more informed decisions,
the user should also consider the relative
costs of using conservative radon controls
and the EPA guidance on further reducing
radon levels even below 4 pCi L1.
The site-specific protocol is defined from
sensitivity analyses of radon entry into a
reference house as determined by the
RAETRAD model. By using the same
model, reference house, and other pa-
rameters used for the residential radon
protection maps, the present protocol has
the same basis as the maps. The sensi-
tivity analyses identified radium concen-
tration, soil layer depth, soil density, soil
texture, and water table depth as the in-
dependent parameters dominating indoor
radon. Of these, radium and water table
depth were most important. Soils up to
2.4 m deep contributed to indoor radon in
uniform-soil scenarios, and soil layers ex-
ceeding approximately 0.6 m thickness
significantly affected radon levels in cases
of non-uniform radium.
A conservative upper limit for radon po-
tential, defined as the 95% confidence
limit for radon in the reference house, was
defined mathematically to correspond to
the radon protection map. The number of
samples needed to represent a site was
determined from the equivalent regional
precision attained by the radon protection
map. Based on the sensitivity of the con-
tributing soil depth and layer thickness,
the site-specific measurements were
shown to require 20 samples for every
4,000 m2 (1 acre). These samples are
distributed throughout four depth incre-
ments that extend to a depth of 2.4 m. An
equivalent precision is computed for a
smaller parcel of land using fewer samples.
The site-specific protocol involves drill-
ing five boreholes in each 4,000 m2 parcel
of land. Twenty soil samples obtained from
the borings are used for radium measure-
ments. The soil borings are also used for
density and textural estimates unless de-
fault values are chosen. A soil radon mea-
surement detects the potential presence
of anomalous elevated-radium materials
at depth. The site water table depths are
defined from soil survey data or site ob-
servations. The site measurements are
-------
analyzed by a special-purpose computer same category definitions as for the resi- map category for radon protection pur-
code called RAETRAD-F to determine the dential radon protection map, the poses.
upper limit for indoor radon concentra- RAETRAD-F code determines whether the
tions in the reference house. Using the site is in the low, intermediate, or elevated
-------
Kirk K. Me/son, Vern C. Rogers, and Rodger B. Holt are with Rogers and Associates
Engineering Corp., Salt Lake City, UT 84110-0330.
David C. Sanchez is the EPA Project Officer (see below).
The complete report, entitled "Site-Specific Protocol for Measuring Soil Radon Potentials
for Florida Houses," (Order No. PB96-175260; Cost: $21.50, subjectto change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use $300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
EPA/600/SR-96/045
------- |