Vous êtes sur la page 1sur 7

Environ. Sci. Technol.

1997, 31, 1742-1748

Radon Entry into Buildings Driven by


Atmospheric Pressure Fluctuations
ALLEN L. ROBINSON AND
RICHARD G. SEXTRO*
Energy and Environment Division, E. O. Lawrence Berkeley
National Laboratory, Berkeley, California 94720

To examine the effects of atmospheric pressure fluctuations


on radon entry into houses, we report measurements of
soil-gas and advective radon entry made using an experimental
basement. Based on these measurements, we quantify
the contribution of atmospheric pressure fluctuations, steady
indoor-outdoor pressure differences, and molecular diffusion to the long-term radon entry rate into the
experimental basement. In the absence of a steady indooroutdoor pressure difference, atmospheric pressure
fluctuations at the study site induce a radon entry rate 1.5
times greater than that due to molecular diffusion. A
steady indoor-outdoor pressure difference reduces the
contribution of atmospheric pressure fluctuations to the longterm radon entry rate. For sustained indoor-outdoor
pressure differences with a magnitude greater than 1.5 Pa,
atmospheric pressure fluctuations have essentially no
effect on the time-averaged radon entry rate into the
experimental structure. The results of this study demonstrate
that under certain conditions, such as periods during
which indoor-outdoor pressure differences are small,
atmospheric pressure fluctuations will contribute measurably
to the total radon entry rate into a building, potentially
doubling indoor concentrations. However, in absolute terms,
atmospheric pressure fluctuations drive approximately
the same amount of entry as molecular diffusion and, therefore,
will probably not cause houses to have long-term, elevated
indoor radon concentrations.

Introduction
Advective flow of radon-laden soil gas is the dominant
transport mechanism for radon into most houses with
elevated indoor radon concentrations (1). This advective
entry is commonly associated with small (0-5 Pa) but
sustained indoor-outdoor pressure differences created by
temperature effects, wind interaction with the building shell,
and the operation of heating, ventilation, and air-conditioning
(HVAC) systems (1). In the context of radon entry into houses,
indoor-outdoor pressure differences are the pressure differences, accounting for the effects of hydrostatics, between
the ambient atmosphere at the soil surface and the indoor
air at the mouth of an opening between the building and the
soil. However, several field studies have reported higher than
expected indoor radon concentrations during periods when
these pressure differences were small (2-5). Two of these
field studies hypothesized that the elevated indoor radon
concentrations were caused by radon entry induced by
atmospheric pressure fluctuations (2, 5).
Recent theoretical (6-8) and experimental studies (9, 10)
suggest that atmospheric pressure fluctuations induce transient soil-gas and radon entry into buildings without the
* Corresponding author fax: 510-486-6658; e-mail RGSextro@lbl.gov.

1742

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997

indoor-outdoor pressure differences commonly associated


with advective radon entry. Atmospheric pressure fluctuations create an oscillating gas flow between the interior of a
building and the underlying soil through the compression
and expansion of the soil gas underneath the building
foundation. Although, over the long-term, the net gas flow
between a building and the soil caused by atmospheric
pressure fluctuations is zero, these fluctuations drive radon
entry because soil-gas radon concentrations are typically 3
orders of magnitude larger than the radon concentrations of
indoor air. Thus, falling atmospheric pressure draws more
radon into a building than is carried out by the gas flow
induced by rising atmospheric pressure, creating a net radon
entry rate. Although progress has been made on the
development of this conceptual model to describe how
atmospheric pressure fluctuations cause advective radon
entry, no experimental data indicating the relative importance
of natural changes in atmospheric pressure for driving radon
entry have ever been reported.
The goal of this paper is to determine the relative
importance of atmospheric pressure fluctuations as a mechanism for driving radon entry into an experimental basement
structure. The experiments reported here examined the
interaction of atmospheric pressure fluctuations and steady
indoor-outdoor pressure differences on the advective radon
entry rate. Based on these results, we quantify the contribution of atmospheric pressure fluctuations, steady indooroutdoor pressure differences, and molecular diffusion to the
long-term radon entry rate into the experimental structure.

Experimental Methods
Experiments were conducted in a controlled experimental
basement. In this section, we first describe the experimental
facility and the structure instrumentation. We then describe
the experimental procedures used to determine the contribution of molecular diffusion, steady indoor-outdoor pressure differences, and atmospheric pressure fluctuations to
the long-term radon entry rate.
Structure and Site Description. The measurements
reported in this study were made using a highly instrumented
experimental structure that was designed and constructed to
study soil-gas and radon entry into buildings in a wellcharacterized setting (11-13). To ensure that the results
obtained from these experiments can be applied to real
houses, this structure has several features in common with
actual houses. The structures concrete walls and floor are
15 cm thick, a common footer design was employed, and the
floor of the basement is located 1.9 m below grade. A
schematic diagram of the structure and the surrounding soil
is shown in Figure 1. The structure is a single chamber with
interior dimensions of 2.0 3.2 m and a height of 2.0 m; only
about 0.1 m of the walls extend above grade. The floor slab
rests on a 0.1 m thick, high-permeability gravel layer.
The soil at the experimental site has been extensively
characterized (12, 14, 15). Table 1 reports the measured
permeability of the gravel, backfill, and undisturbed soil at
the structure site. The backfill region, shown in Figure 1, was
excavated during construction of the structure. After construction, it was carefully refilled and compacted in an attempt
to minimize the disturbance of the native soil environment;
however, as Table 1 indicates, the backfill region has a lower
permeability than the undisturbed soil. Table 2 summarizes
measurements of the air-filled porosity, radon emanation
fraction, and radium content of the soil at the structure site.
All openings between the structure interior and the soil
are sealed except for a 3.8-cm-diameter hole in the center of
the structure floor. Although this hole is not geometrically

S0013-936X(96)00715-8 CCC: $14.00

1997 American Chemical Society

FIGURE 1. Schematic of experimental structure and flow sensor.


A CRM is a continuous radon monitor. The structure and backfill are
drawn to scale. The flow sensor and the hole in the center of the
structure floor are drawn at a scale approximately four times larger
than that of the structure drawing.

TABLE 1. Averages of Measured Soil and Gravel Permeability


at Structure Site
soil region

permeability (m2)

undisturbeda

3.0 10-11 (h); 1.8 10-11 (v)


3.5 10-12
2.0 10-8

backfillb
gravelc

a Horizontal permeability (h) based on measured permeability at


3.5-m length scale; vertical permeability (v) based on measured ratio
of vertical to horizontal permeability (14). b The average of single-point
measurements taken around the basement structure (12). c Based on
laboratory measurements using a vertical column filled with a sample
of the gravel used below the basement structure (11).

TABLE 2. Averages of Measured Air-Filled Porosity, Radon


Emanation Fraction, and Radium Content of Soil at Structure
Site
depth of
layer (m)

soil-grain radium
densitya
contentb
(kg m-3) (Bq kg-1)

0-1.6
2.8 103
1.6-2.2 2.8 103

30
30

2.2-5
2.8 103
5.0-8.5 2.8 103

30
30

air-filled
porosityc

emanation
fractionb

0.45
approximately
linear decrease
from 0.45 to 0.25
0.25
0.25 (inferred)d

0.31
0.45
0.31
0.31

a Ref 24. b Ref 15. c Based on gravimetric analysis (25) of soil cores
taken by Flexser et al. (15). d We have extrapolated the measured profile
to 8.5 m, the measured depth of the water table below the soil surface.

representative of the perimeter cracks that exist in many


houses (16), it is similar to the gaps frequently found around
plumbing and utility penetrations through a building foundation (17). These experiments require such an opening in
combination with a high-permeability subslab gravel layer
to enable atmospheric pressure fluctuations to generate soilgas velocities greater than the detection limit of our flow
sensor. Previous experiments have shown that, as long as
the opening in the structure floor does not provide significant
resistance to flow, the advective radon entry rate into the
structure only weakly depends on the geometry of this opening
due to the presence of a high-permeability subslab gravel
layer (13). Therefore, the measured radon entry rate through
the hole is representative of the flow between the structure
and the underlying soil for a wide range of opening configurations.

Instrumentation. The atmospheric pressure was measured at 5-s intervals using a high-resolution pressure
transducer (Paroscientific Model 1015a) connected to an
outdoor omnidirectional static pressure tap located 3 m
from the structure. The response time, accuracy, and
resolution of this pressure transducer are 1 s, ( 5 Pa, and 0.1
Pa, respectively. The pressure difference between the interior
of the structure and the static pressure tap was measured at
30-s intervals using a differential pressure transducer (Validyne Model DP103).
The gas flow rate through the entry hole in the center of
the structure floor was measured at 5-s intervals using the
flow sensor shown in Figure 1. This sampling interval is much
shorter than the 2-min characteristic response time of the
soil gas at the structure site to changes in atmospheric pressure
(10). The sensor incorporates two omnidirectional hot-film
velocity transducers (TSI Model 8470) mounted in a U-shaped
tube (1.9 cm i.d., 30 cm high) and can measure the
magnitude and direction of gas flow as small as 0.15 L min-1.
Two velocity transducers are required to determine the
direction of the soil-gas flow; details of the sensor and its
calibration are described by Robinson (18). The response
time, accuracy, and resolution of the flow sensor are 2 s, 5%
of reading, and 0.02 L min-1, respectively. For the range of
flows considered in this study, the pressure drop in the flow
sensor tube varies linearly with flow rate and was measured
in the laboratory to be 0.14 Pa min L-1.
Continuous radon monitors (CRMs) were used to measure
radon concentrations in three locations. As indicated in
Figure 1, one CRM sampled air from the top of the U-shaped
flow sensor tube, another CRM sampled soil gas from the
subslab gravel layer at a location 15 cm from the entry hole
in the center of the structure floor. From these locations, gas
samples were drawn at a constant flow rate of 66 cm3 min-1,
passed through a 33 cm3 scintillation cell, and then exhausted
to the outdoors. To reduce the potential effects of 220Rn on
the measurements, the samples were drawn through an 11m-long tube, which, at these flow rates, provided a 3-min
delay before the sample gas arrived in the scintillation cell.
Counts from each CRM were recorded at 1-min intervals and
interpreted using the algorithm described by Busigin et al.
(19).
These low-detection-volume, low-flow CRMs were designed to minimize the effects of sampling on the flows into
and out of the structure and yet maximize the temporal
resolution of our measurements. The 66 cm3 min-1 CRM
sampling flow rate is much smaller than the 500-1000 cm3
min-1 flow rates in the sensor tube caused by typical
atmospheric pressure fluctuations. The high radon concentration of the soil gas underneath the structure, 105 Bq m-3,
enables us to achieve acceptable levels of statistical uncertainty despite the small size of the scintillation cell.
An unmodified CRM (scintillation cell volume of 100 cm3
and a sample flow rate of 200 cm3 min-1) was used to
measure the radon concentration of the air inside the
structure. To allow accurate sampling of the structure radon
concentration from a single location, an oscillating fan
continually mixed the structure air. The indoor radon
concentration was determined by analyzing the counts from
the structure CRM using the method described by Thomas
and Countess (20).
Measurements of Radon Entry Rate. Radon enters the
experimental structure by advection of soil gas through the
hole in the center of the structure floor and by molecular
diffusion through the structures concrete walls and floor.
The total radon entry rate into the structure, ST(t) (Bq s-1),
is

ST(t) ) SA(t) + SD

(1)

where SA(t) is the advective radon entry rate as a function of

VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

1743

time and SD is the diffusive entry rate through the concrete


walls and floor. Both atmospheric pressure fluctuations and
steady indoor-outdoor pressure differences induce advective
radon transport.
Experiments were conducted to make time series measurements of SA(t) as a function of steady indoor-outdoor
pressure difference. For each experiment, a pump and a
computer-controlled mass-flow controller was used to establish a steady indoor-outdoor pressure difference, henceforth referred to as P. In addition to the imposed P, the
pressure of the interior of the structure was allowed to
fluctuate with the natural variations in atmospheric pressure.
Two 1.25-cm-diameter holes in one wall of the access hatch
permitted the pressure in the structure to respond faster than
0.1 s to changes in atmospheric pressure. This time scale is
comparable to the measured response times of the interior
of real houses to changes in atmospheric pressure (21).
Because changes in atmospheric pressure occur simultaneously inside the structure and at the soil surface, these
changes do not affect P. For those experiments conducted
at neutral pressure conditions (P ) 0 Pa), the interior of the
structure was pressurized to offset the slight P, -0.15 Pa,
created by the thermal stack effect. This effect is created by
the 5-10 C temperature difference between the soil gas and
the air inside the structure.
We calculated SA(t) using the measured gas flow rate
through the flow sensor tube, Q(t) (m3 s-1), and the measured
radon concentration of the air inside the flow sensor tube,
C(t) (Bq m-3):

SA(t) ) Q(t)C(t)

(2)

When the magnitude of C(t) falls below the detection limit


of the CRM used to monitor the radon concentration of gas
inside the flow sensor tube, we assume that the radon
concentration of the air inside the sensor tube is equal to the
measured indoor radon concentration. This occurs when
structure air is forced out through the sensor tube and into
the soil by rising atmospheric pressure. The detection limit
associated with our measurement of C(t), typically 7500 Bq
m-3, is similar in magnitude to the measured indoor radon
concentration, but it is much smaller than the 105 Bq m-3
radon concentration of the soil gas underneath the structure
floor slab. Most of the detection limit is due to counting
statistics and to the 1-min counting interval used in the
analysis of the flow sensor CRM data.
The diffusive radon entry rate into the experimental
structure, SD, is 0.1 Bq s-1. This value was determined by
averaging radon flux measurements from various locations
on the inside surface of the structures concrete walls and
floor (22). The flux measurements were made under neutral
pressure conditions, P ) 0 Pa. Because of the small crosssectional area of the entry hole in the center of the structure
floor, we neglect the contribution of the diffusive flux through
this hole to the total radon entry rate.
For this study, we approximate the diffusive radon entry
rate into the structure as constant, independent of the
advective radon entry rate, because the difference between
the indoor radon concentration and that of the surrounding
soil gas was relatively invariant. Radon measurements made
in the subslab gravel layer indicate that the radon concentration of the soil gas underneath the structure floor slab was
essentially constant except for a small region surrounding
the entry hole. In addition, the variation in the measured
indoor radon concentration was small relative to the concentration difference across the concrete walls and floor.
Radon Entry Driven by Atmospheric Pressure Fluctuations. Since both atmospheric pressure fluctuations and P
drive advective radon entry, we must determine the contribution each of these mechanisms to the total advective radon
entry rate. We calculate the time-averaged contribution of

1744

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997

FIGURE 2. Measured gas flow rate through the 3.8-cm-diameter


hole in the center of the structure floor as a function of steady
indoor-outdoor pressure difference. The solid line is the linear
regression, the slope of which represents the inverse of the resistance
of the soil-structure system to flow.
atmospheric pressure fluctuations, SAP (Bq s-1), to the total
advective radon entry rate using

SAP ) Sh A - SSS(P)

(3)

where Sh A is the measured advective radon entry rate averaged


over a time period T:

Sh A )

1
T

S (t) dt
T

(4)

and SSS(P) is the entry rate driven by a steady indooroutdoor pressure difference (P).
For a given value of P, we determine SSS(P) using

SSS(P) ) QSS(P)CSS

(5)

where QSS(P) (m3 s-1) is the soil-gas entry rate caused by P,


and CSS is the radon concentration of the undisturbed soil
gas underneath the structure floor slab, 96 000 Bq m-3. CSS
is the measured radon concentration of the gas inside the
flow sensor when the interior of the structure is depressurized
a steady 5 Pa, a value at which atmospheric pressure
fluctuations are not observed to cause any low-concentration
indoor air to flow out of the structure and into the soil. In
addition, this value of P is not high enough to produce air
flow rates from the soil surface through the soil sufficient to
dilute the radon concentration of the soil gas next to the
structure. We calculate QSS(P) using

QSS(P) )

P
R

(6)

where R is the resistance of the soil-structure system to steadystate soil-gas entry (Pa min L-1). We determined R by
measuring the gas flow rate through the entry hole in the
center of the structure floor for a range of P. A linear
regression of these measurements, shown in Figure 2, yields
the resistance of the soil-structure system, 2.9 Pa min L-1.
The measurements shown in Figure 2 were made with the
holes in the structure access hatch sealed to prevent
atmospheric pressure fluctuations from inducing soil-gas flow.
Long-Term, Time-Averaged Radon Entry Rates. To
examine the importance of atmospheric pressure fluctuations
as a mechanism for driving radon entry, we calculated the
time-average advective radon entry rate into the experimental
structure as a function of P. Each average contained at
least 7 days of measurements. Since more than 99% of the

total power of the time rate of change of the atmospheric


pressure spectrum occurs at frequencies greater than 1 day-1
(8), this average entry rate should represent the effect of typical
atmospheric pressure fluctuations on the long-term radon
entry rate into the experimental structure.
For this analysis, data from each experiment were first
broken into 24-h time blocks. To minimize the influence of
temporal variations in the indoor-outdoor pressure difference on the radon entry rate, we discarded time blocks in
which the standard deviation of the measured indooroutdoor pressure difference exceeded 0.5 Pa. Large fluctuations in indoor-outdoor pressure typically occurred during
storms or other periods with high winds (wind speeds greater
than 5 m s-1). For the remaining time blocks, we calculated
the 24-h average advective radon entry rate and the 24-h
average indoor-outdoor pressure difference. Time blocks
with similar indoor-outdoor pressure differences, within
(0.15 Pa, were grouped together and averaged to estimate
the long-term advective radon entry rate as function of P.
The total, long-term radon entry rate is the sum of SD and the
long-term advective radon entry rate.

Results and Discussion


In this section, we first present time series measurements of
radon entry to examine the relationship between atmospheric
pressure fluctuations and the advective radon entry rate into
the experimental structure. Next, we discuss the effects of
soil-gas dilution caused by the flow of low-concentration
indoor air out of the structure and into the soil on the advective
radon entry rate. We then time-average our measurements
of radon entry to estimate the contribution of atmospheric
pressure fluctuations to the long-term radon entry rate into
the experimental structure. Finally, we discuss briefly the
implications of this work to our understanding of radon entry
into buildings.
Time Series Measurements of Radon Entry. Time series
measurements are shown in Figure 3 to illustrate the
relationship between atmospheric pressure fluctuations, the
resulting soil-gas flows, the soil-gas radon concentration, and
the advective radon entry rate into the experimental structure.
These measurements were made under neutral pressure
conditions, P ) 0 Pa. Figure 3a shows the measured
atmospheric pressure signal for a 2-h period chosen to
illustrate the dynamics of the system. Since changes in
pressure drive soil-gas flow, the time rate of change of
atmospheric pressure, calculated over a 15-s interval, is shown
in Figure 3b. The corresponding gas flow rate through the
hole in the center of the structure floor, Q(t), is shown in
Figure 3c. The measured radon concentration in the Ushaped pipe, C(t), is shown in Figure 3d. The radon entry
rate into the structure, calculated using eq 2, is shown in
Figure 3e. Summary statistics for this 2-h period, such as the
average advective radon entry rate and the volume of gas
forced into and out of the structure by the atmospheric
pressure fluctuations, are listed in Table 3.
Before examining the measurements of radon entry, it is
important to understand the relationship between changes
in atmospheric pressure and the gas flow rate between the
structure interior and the underlying soil because this flow
transports the radon into the structure. In this paper, we
only outline a couple of important aspects of this relationship,
for more details the reader is referred to Robinson et al. (8,
10). A comparison of the calculated time rate of change of
atmospheric pressure, shown in Figure 3b, and the measured
gas flow rate, shown in Figure 3c, reveals that the soil-gas
flow rate follows the time rate of change of atmospheric
pressure. Falling atmospheric pressure drives soil-gas entry
into the structure; rising atmospheric pressure forces air from
inside the structure into the soil. The larger the time rate of
change of atmospheric pressure, the larger the gas flow rate
into or out of the structure. Over the long-term, the net gas

FIGURE 3. Time series measurements made during neutral pressure


conditions (no steady indoor-outdoor pressure difference): (a)
measured atmospheric pressure, (b) time-rate-of-change of atmospheric pressure, (c) Q(t), measured gas flow rate, (d) C(t), measured
radon concentration in the U-shaped flow sensor, and (e) SA(t),
calculated advective radon entry rate. Negative values in Q(t) in
panel c indicate flow from the soil into the experimental structure.
The dashed line in panel e indicates the time-averaged advective
radon entry rate for this 2-h period.
flow between a building and the underlying soil driven by
atmospheric pressure fluctuations is zero; Table 3 indicates
that for the 2-h period shown in Figure 3, changes in
atmospheric pressure cause 30 L of soil-gas to enter the
structure and 32 L of indoor air to exit the structure.
Although over the long-term atmospheric pressure fluctuations do not create a net gas flow rate into the structure,
the measurements shown Figure 3 illustrate how the oscillating gas flow created by these fluctuations causes a net
radon entry rate into the structure. Figure 3 indicates that
falling atmospheric pressure draws high radon concentration
soil gas, 80 000 Bq m-3, into the structure. In contrast, rising
atmospheric pressure forces low radon concentration indoor
air, 2600 Bq m-3, back into the soil. This large concentration
difference causes substantially more radon to be transported
into the structure by falling atmospheric pressure than is
forced out by rising atmospheric pressure, creating a net radon
entry rate without a net gas flow between the structure and
the underlying soil. The time-average advective radon entry
rate for the 2-h period shown in Figure 3 is 0.23 Bq s-1, which
is more than two times the measured diffusive entry rate of
0.1 Bq s-1.

VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

1745

TABLE 3. 2-h Time-Average Values for Soil-Gas Flow and Radon Entry Measurements Shown in Figures 3-5
figure

P (Pa)

QSS
(L min-1)

vol of
gas in (L)

vol of
gas out (L)

av advective radon
entry rate (Bq s-1)

no-dilution entry
rate (Bq s-1)

indoor Rn concn
(Bq m-3)

3
4
5

0
0.9
-0.9

0
0.3
-0.3

32
15
55

30
52
19

0.23
0.02
0.58

0.43
0.21
0.74

2600
1800
4900

FIGURE 4. As in Figure 3 except the measurements were made


during a period when the interior of the structure was steadily
pressurized 0.9 Pa relative to the ambient atmosphere. The dashed
line shown in panel c indicates the gas flow rate out of the structure
caused by a 0.9 Pa steady indoor-outdoor pressure difference. The
dashed line in panel e indicates the time-averaged advective radon
entry rate for this 2-h period.

FIGURE 5. As in Figure 3 except the measurements were made


during a period when the interior of the structure was steadily
depressurized 0.9 Pa relative to the ambient atmosphere. The dashed
line shown in panel c indicates the soil-gas entry rate into the
structure caused by a -0.9 Pa steady indoor-outdoor pressure
difference. The dashed line in panel e indicates the time-averaged
advective radon entry rate for this 2-h period.

Time series measurements are shown in Figures 4 and 5


to illustrate the effect of atmospheric pressure fluctuations
in combination with a non-zero P on the advective radon
entry rate into the experimental structure. The results shown
in Figure 4 are from a period during which the interior of the
structure was pressurized, P ) +0.9 Pa; the measurements
shown in Figure 5 were made during a period when the interior
of the structure was depressurized, P ) -0.9 Pa. Summary
statistics for each of these 2-h periods are also listed in Table
3.
Again, before examining the measurements of radon entry
shown in Figures 4 and 5, it is helpful to understand how
atmospheric pressure fluctuations and P interact to determine the gas flow rate between the structure and the
underlying soil. Since the equations that govern soil-gas flow

are linear, we can separate the measured gas flow rate into
a component caused by P and a component induced by
atmospheric pressure fluctuations. The former creates a dc
offset in the gas flow rate into the structure. To calculate the
magnitude of this offset, we use the measured value of P
and eq 6. For example, the 0.9 Pa indoor-outdoor pressure
difference measured during the 2-h period shown in Figure
4 creates a 0.32 L min-1 offset in the soil-gas flow rate. The
calculated offsets in the gas flow rate created by P are
indicated by the dashed lines in Figures 4c and 5c. The timevarying flows due to the atmospheric pressure fluctuations
are then added to this offset.
The measurements shown in Figure 4 indicate that
atmospheric pressure fluctuations drive advective radon entry
into the structure even when the interior of the structure is

1746

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997

slightly pressurized above the ambient atmosphere. Sustained periods of rapidly falling atmospheric pressure overcome the outward gas flow rate created by the slightly positive
values of P and draw soil gas into the structure. For the 2-h
period shown in Figure 4, atmospheric pressure fluctuations
cause a time-averaged advective radon entry rate of 0.02 Bq
s-1, which is 20% of the measured diffusive entry rate.
If the interior of the structure is slightly depressurized
relative to the ambient atmosphere, atmospheric pressure
fluctuations also increase the advective radon entry rate into
the structure. In effect, the atmospheric pressure fluctuations
increase the rate at which soil gas is drawn into the structure
in comparison to a period during which the interior of the
structure is only steadily depressurized. The average advective
radon entry rate for the 2-h period shown in Figure 5 is 0.58
Bq s-1, which is 20% greater than the radon entry rate driven
by a -0.9 Pa steady indoor-outdoor pressure difference.
Soil-Gas Dilution. In this section, we examine the effects
of soil-gas dilution caused by the outflow of low radon
concentration indoor air into the soil. This dilution complicates the relationship between the soil-gas entry rate and
the corresponding advective radon entry rate. A comparison
of the two spikes in the radon entry rate immediately before
hour 139 in Figure 3e illustrates this complex relationship.
Although these spikes are caused by equivalent soil-gas entry
rates, 1 L min-1, the second spike in the radon entry rate
is much smaller than the first due to the dilution of the soil
gas underneath the structure. If there was no dilution of the
soil gas underneath the structure, then these two spikes in
the radon entry rate should have essentially the same
magnitude.
The overall effect of soil-gas dilution is to reduce the
contribution of atmospheric pressure fluctuations to the timeaveraged, advective radon entry rate. To quantify this
reduction, we calculate a hypothetical no-dilution radon
entry rate for each of the 2-h periods shown in Figures 3-5.
The no-dilution entry rate is the product of the rate at which
soil-gas is drawn unidirectionally into the structure and the
undisturbed radon concentration of the soil gas underneath
the structure floor slab, CSS. Comparing the no-dilution entry
rate to the actual, time-averaged radon entry rate listed in
Table 3 reveals that soil-gas dilution reduces the timeaveraged advective radon entry rate by a factor of 2 under
neutral pressure conditions. Pressurizing the interior of the
structure dramatically increases the effects of dilution;
depressurizing the interior of the structure reduces the effects
of dilution.
The reduction in the advective radon entry rate caused by
soil-gas dilution depends on both the history of the gas flow
into and out of the structure and the rate at which the radon
concentration underneath the structure is recharged. The
gas flow depends on both P and the atmospheric pressure
fluctuations. The recharge rate of the soil gas depends on
both the production of radon, through the decay of 226Ra in
the soil, and the transport of radon through the soil pore
space. Since gas flow, radon transport, and radon generation
all depend strongly on the soil properties, the effects of soilgas dilution on the advective radon entry rate could be
significantly different for houses located in soils with very
different properties than those of the soil at the structure site.
Long-Term Radon Entry Rate. To examine the effects of
atmospheric pressure fluctuations on radon entry for a wide
range of environmental conditions, Figure 6 presents the
measured long-term radon entry rate into the experimental
structure as a function of P. Each data point shown in Figure
6 was determined by time averaging at least 7 days of radon
entry measurements, as previously described. Short time
series, such as those shown in Figures 3-5, may not accurately
indicate the contribution of atmospheric pressure fluctuations
to the long-term radon entry rate because occasional, high-

FIGURE 6. (a) Measured long-term radon entry rate as a function


of steady indoor-outdoor pressure difference (P). Each data point
indicates the average of at least 7 days of measurements. (b) Ratio
of the long-term radon entry rate driven by atmospheric pressure
fluctuations to that due to the combined contribution of molecular
diffusion and P. A ratio of 100% indicates that atmospheric pressure
fluctuations drive the same amount of radon entry as both diffusion
and steady indoor-outdoor outdoor pressure differences, i.e., such
fluctuations increase the long-term radon entry rate by a factor of
2. Negative P indicates that the interior of the structure is
depressurized relative to the ambient atmosphere. The contribution
of steady depressurization of the structure interior to the radon entry
rate is estimated based on the resistance of the soil-structure system,
as described in the text. Diffusive entry was measured under neutral
pressure conditions and is assumed to be independent of indooroutdoor pressure difference. Only diffusion and atmospheric pressure
fluctuations drive radon entry when P is greater than or equal to
zero. Lines are smooth curves through the data and are intended
only for visual guidance. The vertical bars indicate the measurement
uncertainties.
frequency oscillations in atmospheric pressure can generate
large but intermittent soil-gas and radon entry rates.
Figure 6a identifies the contribution made by atmospheric
pressure fluctuations, P, and molecular diffusion to the total,
long-term radon entry rate. Figure 6b shows the ratio of the
long-term radon entry rate driven by atmospheric pressure
fluctuations to that due to the combined contribution of
molecular diffusion and P. The contribution of molecular
diffusion is 0.1 Bq s-1 and is assumed to be independent of
the effects of soil-gas flow, as previously described. The
contribution of P to the long-term radon entry rate is defined
by eq 5. As Figure 6a indicates, when the interior of the
structure is pressurized above the ambient atmosphere only
diffusion and atmospheric pressure fluctuations produce
radon entry. When the interior of the structure is depressurized, the resulting P, the atmospheric pressure fluctuations, and diffusion all contribute to the long-term radon
entry rate.
Figure 6 demonstrates that, in the presence of small values
of P, atmospheric pressure fluctuations measurably con-

VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

1747

tribute to the long-term radon entry rate into the experimental


structure. When P ) 0 Pa, the total radon entry rate into
the structure is 0.25 Bq s-1, of which 0.1 Bq s-1 is due to
diffusion and the remaining 0.15 Bq s-1 is caused by
atmospheric pressure fluctuations. This represents a 150%
increase in long-term radon entry due to atmospheric pressure
fluctuations, as shown in Figure 6b. Small indoor-outdoor
pressure differences commonly occur during the summer
when indoor-outdoor temperature differences are small and
winds are light.
Steady indoor-outdoor pressure differences reduce the
contribution of atmospheric pressure fluctuations to the longterm radon entry rate. When the magnitude of P is greater
than 1.5 Pa, atmospheric pressure fluctuations have a
negligible effect on the long-term radon entry rate into the
experimental structure. In real houses, indoor-outdoor
pressure differences exceeding this value are common during
the winter heating season or on windy days.
Figure 6a indicates that as the value of P becomes more
negative the radon entry rate increases. For P values beyond
-1.0 Pa, advective entry caused by this underpressure
overwhelms the contribution of molecular diffusion and
atmospheric pressure fluctuations to the long-term radon
entry rate. This finding is consistent with the existing
conceptual model for radon entry into houses which states
that advective radon entry caused by steady indoor-outdoor
pressure differences is the dominant transport mechanism
of radon into houses with elevated indoor concentrations
(1).
Implications. This study demonstrates that atmospheric
pressure fluctuations can induce measurable advective radon
entry rates when the magnitude of the indoor-outdoor
pressure differences are less than 1.5 Pa. Therefore, analyses
that only consider the effects of molecular diffusion and steady
indoor-outdoor pressure differences will underpredict the
total radon entry rate into a building. In addition, the results
of this study have implications on the transport of other soilgas contaminants into buildings.
The effects of atmospheric pressure fluctuations on longterm radon entry in actual houses will depend on the soil
properties and the building structural configuration. This
study only examined the combination of soil properties listed
in Tables 1 and 2 and the structural configuration of our
experimental basement. Several theoretical studies have
considered the effects of atmospheric pressure fluctuations
for a wider range of building and soil types (7, 8).
Ultimately, the importance of atmospheric pressure
fluctuations as a mechanism for driving radon entry must be
judged in terms of its impact on the indoor radon concentration. In relative terms, the results of this study suggest that,
under certain conditions, atmospheric pressure fluctuations
can significantly increase indoor concentrations. For example, noting that steady-state indoor radon concentrations
vary linearly with the radon entry rate when the building
ventilation rate is constant (23), Figure 6a indicates that
atmospheric pressure fluctuations will more than double
indoor concentrations in experimental structure when P )
0 Pa. Consequently, radon entry driven by atmospheric
pressure fluctuations is a plausible explanation for the higher
than expected indoor radon concentration observed when
indoor-outdoor pressure differences were small (2-5).
However, in absolute terms, the results of this study and the
previous theoretical analyses (7, 8) suggest that atmospheric
pressure fluctuations drive approximately the same amount
of entry as molecular diffusion; thus, the entry rates produced
by these fluctuations will not lead to long-term, elevated
indoor radon concentrations.

by the Director, Office of Energy Research, Office of Health


and Environmental Research, Environmental Sciences Division, and by the Assistant Secretary for Energy Efficiency and
Renewable Energy, Office of Building Technology of the U.S.
Department of Energy (DOE), under Contract DE-AC03763SF00098.

Acknowledgments

ES960715V

The authors thank J. Daisey, K. Garbesi, and P. Price for


reviewing a draft of this manuscript. This work was supported

1748

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997

Literature Cited
(1) Nazaroff, W. W. Rev. Geophys.1992, 30, 137.
(2) Hernandez, T. L.; Ring, J. W.; Sachs, H. M. Health Phys. 1984,
46, 440.
(3) Holub, R. F.; Droullard, R. F.; Borak, T. B.; Inkret, W. C.; Morse,
J. G.; Baxter, J. F. Health Phys. 1985, 49, 267.
(4) Turk, B. H.; Prill, R. J.; Sextro, R. G.; Harrison, J. In The proceedings
of the 1988 International Symposium on Radon and Radon
Reduction Technology; U.S. Environmental Protection Agency:
Research Triangle Park, NC, 1989; pp 6-25-6-44.
(5) Hintenlang, D. E.; Al-Ahmady, K. K. Indoor Air 1992, 2, 208.
(6) Narasimhan, T. N.; Tsang, Y. W.; Holman, H. Y. Geophys. Res.
Lett. 1990, 17, 821.
(7) Tsang, Y. W.; Narasimhan, T. N. J. Geophys. Res. 1992, 97, 9161.
(8) Robinson, A. L.; Sextro, R. G.; Riley, W. J. Atmos. Environ. 1997,
31, 1487.
(9) Robinson, A. L.; Sextro, R. G. Geophys. Res. Lett. 1995, 22, 1929.
(10) Robinson, A. L.; Sextro, R. G.; Fisk, W. J. Atmos. Environ. 1997,
31, 1477.
(11) Fisk, W. J.; Modera, M. P.; Sextro, R. G.; Garbesi, K.; Wollenberg,
H. A.; Narasimhan, T. N.; Nuzum, T.; Tsang, Y. W. Report LBNL31864. Lawrence Berkeley National Laboratory: Berkeley, CA,
1992.
(12) Garbesi, K.; Sextro, R. G.; Fisk, W. J.; Modera, M. P.; Revzan, K.
L. Environ. Sci. Technol. 1993, 27, 466.
(13) Robinson, A. L.; Sextro, R. G. Health Phys. 1995, 69, 367.
(14) Garbesi, K.; Sextro, R. G.; Robinson, A. L.; Wooley, J. D.; Owens,
J. A.; Nazaroff, W. W. Water Resour. Res. 1996, 32, 547.
(15) Flexser, S.; Wollenberg, H. A.; Smith, A. R. Environ. Geol. 1993,
22, 162.
(16) Scott, A. G. In Radon and Its Decay Products in Indoor Air;
Nazaroff, W. W., Nero, A. V., Eds.; John Wiley and Sons: New
York, 1988; pp 407-433.
(17) Nazaroff, W. W.; Lewis, S. R.; Doyle, S. M.; Moed, B. A.; Nero,
A. V. Environ. Sci. Technol. 1987, 21, 459.
(18) Robinson, A. L. Ph.D. Dissertation, University of California at
Berkeley, and Report LBNL-38843; Lawrence Berkeley National
Laboratory: Barkeley, CA, 1996.
(19) Busigin, A.; van der Vooren, A. W.; Phillips, C. R. Health Phys.
1979, 37, 659.
(20) Thomas, J. W.; Countess, R. J. Health Phys. 1979, 36, 734.
(21) Allen, C. Technical Note AIC-13. Air Infiltration Center: Bracknell,
Berkshire, Great Britain, 1984.
(22) Garbesi, K. Ph.D. Dissertation, University of California at Berkeley,
and Report LBL-34244; Lawrence Berkeley National Laboratory: Berkeley, CA, 1993.
(23) Nazaroff, W. W.; Moed, B. A.; Sextro, R. G. In Radon and Its
Decay Products in Indoor Air; Nazaroff, W. W., Nero, A. V., Eds.;
John Wiley and Sons: New York, 1988; pp 57-112.
(24) Brimhall, G. H.; Lewis, C. J. Differential element transport in the
soil profile at the Ben Lomond small structure radon site: A
geochemical mass balance study; Technical Report; University
of California at Berkeley, Department of Geology and Geophysics;
Berkeley, CA, 1992.
(25) Danielson, R. E.; Sutherland, P. L. In Methods of soil analysis,
part 1. Physical and mineralogical methods; Black, C. A., Ed.;
American Society of Agronomy: Madison, WI, 1986; pp 443450.

Received for review August 19, 1996. Revised manuscript


received January 23, 1997. Accepted February 3, 1997.X

Abstract published in Advance ACS Abstracts, April 1, 1997.

Vous aimerez peut-être aussi