EPA/451/R-96/007
vyEPA
ited States Office of Air Quality EPA-451/R-96-007
/ironmental Protection Planning and Standards August 1996
ency Research Triangle Park, NC 27711
Air/Superfund
AIR/SUPERFUND
NATIONAL TECHNICAL
GUIDANCE STUDY SERIES
Estimation of Air Impacts for
Soil Vapor Extraction (SVE)
Systems (Revised)
-------
AIR/SUPERFUND NATIONAL TECHNICAL
GUIDANCE STUDY SERIES
ASF-22
ESTIMATION OF AIR IMPACTS
FOR SOIL VAPOR EXTRACTION
(SVE) SYSTEMS
(REVISED)
Contract No. 68-D3-0032
Work Assignment No. 11-59
Prepared for:
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
August 1996
-------
DISCLAIMER
The information in this document has been reviewed in its entirety by the U.S.
Environmental Protection Agency (EPA), and approved for publication as an EPA
document. Any mention of trade names, products, or services does not convey, and
should not be interpreted as conveying official EPA approval, endorsement, or
recommendation.
-------
CONTENTS
Disclaimer ii
Figures iv
Tables iv
Acknowledgment v
1. Introduction 1
2. Process Description 3
3. Estimation of Air Emissions 6
First-Order Estimates of Removal Rates 7
Estimation of Multicomponent System Removal Rates 9
Screening Model Computer Codes for Multicomponent
Systems 15
4. Refined Estimates of Removal Rates 19
5. Limitations and Assumptions of Models Reviewed 25
Gas Flow 25
Vapor Concentration 27
6. Estimation of Ambient Air Concentrations 29
7. Recommendations 34
Estimating Emissions During the Feasibility Study 34
Estimating Emissions During the Remedial Design/
Remedial Action 37
Dispersion Modeling to Estimate Ambient Air Impacts 37
References 39
Appendix A. Multicomponent Estimation Model (Case Example) A-1
-------
FIGURES
Number
1 Generalized process flow diagram for soil vapor extraction 4
2 Comparison of time-dependent soil concentrations of benzene,
toluene, and ethylbenzene 17
TABLES
1 Conversion Factors for 1-h Ambient Air Concentrations 30
2 Example Scenarios for SVE with No Controls Based on
Size of System 32
3 Example Scenarios for SVE with Activated Carbon Controls
Based on Size of System 32
4 Example Scenarios for SVE with Catalytic or Thermal Oxidation
Controls Based on Size of System 33
5 Vapor Flow Rate Model Variables 36
6 Soil Property Variables for Phase Partioning 37
IV
-------
ACKNOWLEDGMENT
This document was prepared for the U.S. Environmental Protection Agency by
Environmental Quality Management, Inc., Durham, North Carolina, under contract to
Pacific Environmental Services, Inc. Greg Pagett (Pacific Environmental Services, Inc.)
and David Dunbar (Environmental Quality Management, Inc.) managed the project. Craig
Mann of Environmental Quality Management, Inc. was the principal author. Patricia Flores
of the U.S. Environmental Protection Agency, Region III provided overall program direction
and served as the work assignment manager.
This document is a revision to the initial document of the same title (EPA-450/1-92-
001) dated January 1992 which was produced for the U.S. Environmental Protection
Agency by Radian Corporation of Austin, Texas.
v
-------
SECTION 1
INTRODUCTION
The U.S. Environmental Protection Agency's (EPA's) Office of Air Quality Planning
and Standards (OAQPS) and the Regional Air Offices have been given the responsibility
to evaluate air impacts from Superfund sites. An important part of this program is the
analysis of air impacts from various alternatives for cleaning up Superfund sites. Since
these analyses are frequently required for planning purposes prior to actual cleanup they
depend on estimated emissions and ambient concentrations rather than on field
measurements.
Soil Vapor Extraction (SVE) is also known as soil venting, vacuum extraction,
aeration, or in-situ volatilization. It is a widely used technique for removing volatile organic
compound (VOC) vapors from contaminated soil. This report provides several predictive
emission models ranging from first-order gross estimates requiring little site-specific data,
to refined models which calculate air flow through a porous medium as a result of the
pressure gradient created by an extraction well and the vapor-phase concentration of the
extracted air.
When evaluating the array of predictive models, a primary concern is the ability of
each model to estimate time-dependent removal rates. This is especially important for
SVE in that initial removal rates may be quite high (e.g., 500 to 600 kg/day). Typically,
removal rates decrease with time as initial soil concentrations are depleted. Time-
dependent removal rates may be critical for estimating other than long-term exposures,
and thus short-term risks to the community and to site workers.
This document does not, however, include an analysis of the selection or use of
air pollution control devices. The models reviewed are used to estimate contaminant
-------
removal rates which are also synonymous with uncontrolled emission rates. Modeling
results are therefore most useful to determine the need for air pollution controls.
The report also provides guidance on air dispersion model selection and use to
help develop different modeling scenarios and estimate ambient air impacts. Model
inputs, including source and receptor data can vary greatly. These data can have a direct
bearing on ambient air concentrations, which are important both in risk assessments and
compliance with air applicable or relevant and appropriate requirements (ARARs).
Modeling several scenarios can allow for the adjustment of equipment size, design,
control equipment, and location in a cost effective manner.
-------
SECTION 2
PROCESS DESCRIPTION
Soil vapor extraction systems involve the removal of air containing volatile
compounds from unsaturated soil. A general process schematic is shown in Figure 1.
One or more extraction wells are placed in the vadose (unsaturated) zone at the site of
the contamination. The extraction wells typically are constructed of slotted plastic pipe
placed in permeable packing such as coarse sand. The wells are connected via a
manifold that leads to an air blower or vacuum pump. The vacuum pump is used to
withdraw air from the extraction wells and is typically capable of inducing a vacuum of
about 30 in. Hg. The subsurface vacuum that is created and the resulting pressure
gradient in the soil will dictate the rate at which vapors are withdrawn and the air flow
patterns in the vadose zone.
As shown in Figure 1, air inlet wells can be used to introduce ambient air at the
edge of the contaminated area. They serve to both increase the rate that contaminant-
laden air is extracted and to control the direction of vapor migration. The air inlet wells
can be passive or air can be forced into the ground using an air blower. Only a fraction
of the extracted air comes from the air inlets, with the remainder of the air drawn from the
surface through the soil. Therefore, it is important to have an impermeable cap in place
at the top of each extraction well to prevent the direct inflow of ambient air down the well
casing.
Several design options are available for increasing the air flow rate and removal
efficiency of a SVE system. The inlet air can be heated to increase the partitioning of
subsurface contaminants into the vapor-phase. The air blower used to inject air into the
ground will heat the air to some extent due to compression; additional heat is added in
some SVE systems by also injecting steam into the ground. In some cases, the air inlet
-------
Clean Flue Gas
Clean
Water
Liquid
Treatment
Air Air
Inlet
Well
Vapor
Treatment
Stack
Pump
Extraction
Well
Soil Surface
Contaminated
Soil
Vandose Zone
Vapors
Extraction
Well
Figure 1. Generalized process flow diagram for soil vapor extraction.
-------
wells extend down into the saturated zone. The inlet air bubbles up through the water
table and this sparging action can enhance the removal of slightly soluble volatile organic
compounds.
An air/water separator is typically employed if water is pulled from the extraction
wells. The separator serves to protect the air blowers or pumps and to increase the
efficiency of any vapor treatment system that is used. The condensate from the separator
may require treatment prior to discharge. For the large-scale remediations done at
Superfund Sites, the air stream will normally require treatment to reduce its VOC content
prior to discharge into the atmosphere. Carbon adsorption, catalytic, or thermal oxidation
are the most commonly used control technologies. Carbon adsorption systems may be
present upstream or downstream of the vacuum pump. For SVE systems with small air
flow rates (i.e., 30 to 100 scfm), internal combustion engines can also be used to control
the air emissions.
-------
SECTION 3
ESTIMATION OF AIR EMISSIONS
There are several alternative approaches for estimating the emissions from a SVE
system. The best method is to directly measure the emissions from the system while it
is in full-scale or pilot-scale operation. The next best method is to estimate the emissions
using predictive equations with site-specific inputs. If site-specific inputs are not available,
conservative estimates can be made by using default values for the input parameters.
This report describes several existing models for estimating the removal rate of
volatile soil contaminants via SVE. Each model was evaluated based on theoretical
approach, data requirements, ease of use, and the relative refinement of the removal rate
estimates. Critical to the estimation of the emission rate, is a time-dependent estimate of
the vapor-phase contaminant concentrations within the soil column. For this reason, the
more refined models include equilibrium contaminant phase distribution estimates.
Each model was evaluated as to its usefulness for conducting risk evaluations
and/or ARAR compliance evaluations at each step in the remedial process. Screening
models were considered most useful during the detailed analysis of alternative remedial
technologies and were therefore assessed as to the relative refinement of their predictions
using limited site-specific and/or default data. Refined models are typically more useful
during the remedial design/remedial action (RD/RA), when more data are available. In
the case of computer models, availability and support were also considered as key
criteria.
-------
3.1 FIRST-ORDER ESTIMATES OF REMOVAL RATES
*
The most basic techniques for estimating SVE removal rates involve either an
assumption of total contaminant mass removed over a given time period, or the product
of estimated subsurface vapor flow and vapor concentration. The former assumes a
constant removal rate and the latter a constant flow and vapor-phase concentration.
An estimate of the total emission potential for a site may be made using the
following equation:
ERfve = Mjt = (Vs C, (1) ^It d)
where EFfve = Average emission rate of component i, g/s
Mj = Total mass of i in soil, g
t = Duration of remediation, s
Vs = Volume of contaminated soil to be treated, nf
Q = Average concentration of i in soil, //g/g
1 = Constant, g/106 /JQ x 10s crrf/m3
pb = Average soil dry bulk density, g/crrf.
Equation 1 assumes complete recovery of the contaminant over the duration of the
remediation and thus constitutes a gross estimate of the average emission rate. In
addition, Equation 1 assumes that the concentration in soil is constant over time and
relies on an estimate of the time required to remove all of the contaminant. In reality, the
concentration will not remain constant and the time required for remediation cannot be
determined without further site information. Therefore, Equation 1 may best be used as
a check on the total mass lost over long time periods.
A better estimate of the removal rate can be obtained as the product of the
subsurface vapor flow induced by the extraction well and the soil vapor-phase
7
-------
concentration. Johnson et al. (1990a) estimate the maximum uncontrolled emission rate
over relatively short time periods as:
= CvJ Q (2)
where EF\max = Maximum emission rate of component i, g/s
Q,j = Vapor-phase concentration of i in the vent, g/cm3 -vapor
Q = Induced soil vapor flow rate, cm3-vapor/s.
Equation 2 estimates the maximum emission rate at the start of remediation and
requires an estimate of the soil vapor concentration and flow rate. The soil vapor
concentration may be estimated via soil gas measurements, however, this necessitates
determining the lateral and vertical soil gas profile to estimate an average in situ soil gas
concentration. Another option would be to extract headspace gas samples from bore
hole samples. This requires careful core sample handling and sample preparation to
avoid loss of analytes with low boiling points.
If soil gas measurements are not available or if data are not sufficient, a first-order
estimate of vapor concentration in soil may be made using the ideal gas law. The
maximum vapor concentration of any compound in extracted vapors is its equilibrium or
saturated vapor concentration which is calculated as:
MW;
8
where Q,|eq = Equilibrium saturation vapor concentration of component i,
g/cm3-vapor
xj = Mole fraction of i in the residual mixture, unitless
8
-------
PIV(TS) = Pure component vapor pressure of i corrected for system
• temperature, atm
= Molecular weight of i, g/mol
R = Gas constant (=82.1 cm3-atm/mol-°K)
Ts = System temperature, ° K.
Equation 3 is applicable when the residual total hydrocarbons are present as
nonaqueous phase liquids (NAPLs) or solids. At low soil concentrations, nonideal
conditions exist and Equation 3 will overpredict the vapor concentration.
With a measured or calculated vapor concentration in soil, an estimate of the
induced soil vapor flow is required to predict the maximum emission rate using Equation
2. Johnson et al. (1990a) give a range of vent vapor flow rates typically seen for SVE
systems. The range of values is given as 10 to 200 if/min per well.
Emissions calculated using Equation 2 may give reasonable estimates at the start
of venting, however, vapor concentrations decrease with time due to changes in the
composition of the residual and due to mass-transfer resistances (Johnson et al., 1990).
Because emissions calculated using Equation 2 are constant, an estimate of the time
required to reduce residual soil levels to some target concentration cannot accurately be
made.
3.2 ESTIMATION OF MULTICOMPONENT SYSTEM REMOVAL RATES
An estimate of the removal rate of an individual compound in a multicomponent
system (i.e., one volatile compound in a mixture) may be obtained from the procedures
of U.S. EPA (1994) and Johnson et al. (1990 and 1990a, b). In a single comoonent
system, the soil concentration at which the sorbed, dissolved and vapor-phases are
saturated is commonly referred to as the soil saturation concentration (Csat). The value
of Csat (mg/kg) can be estimated from U.S. EPA (1994) as:
-------
(4)
where S is the pure component solubility in water (mg/L);yob is the soil dry bulk density
(kg/L); KD is the soil/water partition coefficient (L/kg); 6W is the soil water-filled porosity
(unitless); H/ is the Henry's law constant corrected system for temperature (unitless); and
0a is the soil air-filled porosity (unitless). Concentrations greater than C^ indicate the
presence of a residual-phase (e.g., NAPL). Although not as accurate as computer codes
which solve for multicomponent thermodynamics, Equation 4 may be used as a
preliminary estimate of the soil concentration above which a residual phase is present for
individual components of a mixture.
Equilibrium vapor-phase soil concentrations during SVE can be calculated at
discreet time-steps using either Raoult's law or Henry's law depending upon whether the
soil concentration is greater than Csat. If the soil concentration for any time-step is greater
than or equal to Csat, the vapor-phase concentration at equilibrium can be calculated
using Equation 3. If the soil concentration at any time-step is less than 0^,, the vapor-
phase concentration can be calculated from Johnson et al. (1990b) as:
Pb
(H'T ea) + ew
where H/ is the dimensionless Henry's law constant corrected for system temperature;
Cs is the soil concentration at each time-step; Qa is the soil air-filled porosity; pb is the soil
dry bulk density; 0W is the soil water-filled porosity, and KD is the soil/water partition
coefficient.
The pure component vapor pressure corrected for system temperature used in
Equation 3 may be calculated using Antoine coefficients if available or may be estimated
by:
10
-------
exp
In
(6)
where PiV(TR) is the vapor pressure of i at the reference temperature (atm); TB is the
normal boiling point (°K), TR is the reference temperature (°K), Ts is the system
temperature (°K), and PB is the pressure at which the boiling point is measured (atm).
The Henry's law constant used in Equations 4 and 5 may be adjusted for system
temperature by:
AH.
V,TS
H
R
(7)
where H,- is the Henry's law constant at the desired temperature (atm-m3/mol); HR is the
Henry's law constant at the reference temperature (atm-nf/mol); AH^ is the enthalpy
of vaporization of the compound at the system temperature (cal/mol); R is the gas
constant 1.987 cal/mol-0 K; Ts is the system temperature (°K); and TR is the Henry's law
constant reference temperature (°K) (U.S. EPA, 1990). The enthalpy of vaporization at
the system temperature may be estimated from Lyman et al. (1990):
AHVT> = AH b
(8)
where AH^b is the enthalpy of vaporization at the normal boiling point (cal/mol); Tsis the
system temperature (°K); Tc is the critical temperature (°K); TB is the normal boiling point
(° K); and n is an exponential term given the value of 0.38. The dimensionless form of the
Henry's law constant corrected for temperature (H/) is simply H^/RTg where R is the gas
constant 8.21E-05 m3-atm/mol-°K and Ts is the system temperature (°K).
11
-------
With an estimate of the equilibrium vapor-phase concentration as a function of the
degree of saturation, the rertioval rate at each time-step may be calculated using Equation
2. This also requires an estimate of the induced vapor flow rate from the SVE well (Q).
Constant vapor flow rate may be estimated from Johnson et al. (1990) as:
Q =H
P.
w
- (Pajpw)2}
In (RJR,)
(9)
where H is the length of the vacuum well that is screened through the vadose zone (cm),
K, is the permeability of the soil to vapor flow (cm2), jj is the viscosity of air (1.8E-04
g/cm-s), Pw is the well absolute pressure (g/cm-s2), Patm is the average atmospheric
pressure (g/cm-s2), l\ is the vent well radius (cm), and F^ is the well radius of influence
(cm) where the pressure is equal to atmospheric pressure.
For screening purposes, the screened interval (H) is assumed to be equal to the
thickness of the contaminated zone. The removal rate (g/s) for each time-step is
calculated using Equation 2. The initial total mass of the contaminant within the radius
of influence of a single vacuum well at time-zero is estimated from the initial soil
concentration and the well radius of influence by:
2
M, = n R, H C0
(10)
where IvJ is the total initial mass of component i (g); F^ is the well radius of influence (cm);
H is the screened interval (cm); and C0 j is the initial soil concentration of i (g/cm3) at time-
zero.
For the first time-step, the initial soil concentration is compared to the value of Csat
calculated by Equation 4. Equation 3 or Equation 5 is then used to calculate the vapor-
phase concentration, as appropriate. The vapor concentration is then multiplied by the
12
-------
flow rate calculated by Equation 9 to obtain the removal rate (g/s). The removal rate is
then multiplied by the time-step interval (s) to derive the total mass removed. This mass
is subsequently subtracted from the initial mass calculated by Equation 10. The
remaining mass is converted to a concentration by dividing the remaining mass by the
volume determined from the values of F^ and H (i.e., V = n F^2 H). This concentration is
then compared to the value of Csat to determine the appropriate phase-partitioning for the
next time-step. This procedure is duplicated for each succeeding time-step until a target
soil concentration is reached.
During the transition from four to three phases (i.e., when the residual phase is
depleted), the dissolved-phase concentration will not be at the solubility limit, as assumed
in Equation 5, but will be at a reduced concentration such that the activity coefficients of
the dissolved and residual phases are in local equilibrium as a function of the mole
fraction. Therefore during the transition, Equation 5 will tend to overpredict the vapor-
phase concentration and thus the emission rate. Once the soil concentration drops below
the value of Csat, the vapor-phase concentration predicted at each time-step is compared
to that predicted at the previous time-step. If the vapor-phase concentration predicted
by Equation 5 is higher than that predicted for the previous time-step, the vapor-phase
concentration is multiplied by the mole fraction.
Appendix A contains a case example spreadsheet printout of the multicomponent
estimation procedures for 10 contaminants. The spreadsheet is constructed in
MICROSOFT EXCEL Version 5.0. Enclosed with this document is a diskette containing
the EXCEL spreadsheet. In addition, this file has also been converted to a LOTUS 1-2-3
spreadsheet.
The spreadsheet may be used with the original 10 chemicals or new chemicals
may be substituted. If analysis of more than 10 chemicals is needed, the spreadsheet
must be altered to accommodate the additional compounds. If less than 10 chemicals
are to be analyzed, the initial soil concentrations of the excess listed chemicals must be
set to zero. To execute the analysis, the data entry sheet is filled-in first. The next step
is to simply copy down row 14 starting at column AF (the column containing the time-
steps for the first chemical) until the soil concentration remaining for each chemical is less
13
-------
than the target soil concentration. Emission calculations for each chemical cannot be
done separately because the computed mole fractions are based on the relative weight
fractions of all chemicals.
Once the analysis is complete, the average emission rate of each chemical
(ERJ over the operating time of the SVE system may be numerically estimated using the
trapezoidal rule:
= 7 [
Af .
o \ 0 1 2 •• • /)_^ p| »/
where r is the averaging period (s), At is the constant time-step interval (s), EF\,i1i2 n are
the emission rates at the initial and each succeeding time-step (g/s), and n is the number
of time-steps. To ensure that a mass balance violation does not occur, the cumulative
mass lost of each chemical (ERi x r) must be compared with the initial mass in the soil
at time-zero. If the cumulative mass lost is greater than the initial mass, the analysis is
rerun with a shorter time-step interval.
The multicomponent estimation model, as well as the screening model computer
codes described in the next section, require chemical properties for the contaminants to
be analyzed. The following is a list of references which may be consulted to attain these
chemical properties:
CHEMDAT8 Data Base of Compound Chemical and Physical Properties. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards
Technology Transfer Network (TTN), CHIEF Bulletin Board, Research Triangle
Park, North Carolina.
Cf?C Handbook of Chemistry and Physics. CRC Press, Boca Raton, Florida.
1995.
Perry, R. H., and C. H. Chilton, eds. Chemical Engineering Handbook. 5th Ed.
McGraw-Hill, New York, New York. 1973.
14
-------
Soil Screening Guidance: User's Guide. U.S. Environmental Protection Agency,
Office of Solid Waste and Emergency Response, Washington, D.C. OSWER
Publication 9355.4-23. April 1996.
Soil Screening Guidance: Technical Background Document. U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response, Washington,
D.C. EPA/540/R95/128.
Suoerfund Chemical Data Matrix and User's Guide. U.S. Environmental
Protection Agency, Office of Emergency and Remedial Response, Washington,
D.C. EPA/540/R-96-028 and EPA/540/R-96/029.
3.3 SCREENING MODEL COMPUTER CODES FOR MULTICOMPONENT SYSTEMS
The work of Johnson et al. (1990 and 1990a) are available as computer codes in
two separate programs. The Hyperventilate code (Johnson, 1991) is available as public
domain software free of change from the U.S. EPA. The Hyperventilate application
manual (U.S. EPA, 1993) is a software guidance manual for evaluating the feasibility of
using SVE at specific sites. Hyperventilate operates in the APPLE MACINTOSH or the
IBM PC-compatible MICROSOFT WINDOWS environment with OBJECT PLUS.
Hyperventilate is an educational and decision support tool consisting of several
"stacks" of graphical reference "cards" within a hypertext framework. Based on user
inputs and/or a range of default model parameters, Hyperventilate enables the user to
decide whether SVE is practical, produce preliminary SVE system designs, and project
system costs. Of particular interest to risk evaluations of remedial alternatives,
Hyperventilate can estimate maximum removal rates. The analytical mass transfer
equations used by the code are based on Raoult's law, and assume the presence of
NAPL in the vadose zone. Therefore, it is not appropriate to apply Hyperventilate at sites
where NAPL is absent because Henry's law rather than Raoult's law describes
volatilization from the dissolved-phase. The program does not compute time-dependent
removal rates. Program outputs include: 1) extraction well flow rate, 2) maximum
estimated contaminant removal rates, 3) desired versus maximum removal rate, 4) soil
permeability estimates, 5) relative removal efficiency based on diffusion-limited vapor
transport, and 6) removal rates based on vapor flow outside of the contaminated zone.
15
-------
VENTING is a program for estimating hydrocarbon recovery from SVE and is
based on the VENTING fortran code of Johnson et al. (1990). VENTING is a commercial
code available from Environmental Systems and Technologies, Inc. (ES&T, 1994). At the
v
time of publication, the cost of the VENTING software was $395. The code uses a
MICROSOFT DOS based text menu interface for data input and provides ASCII output of
the results in the form of phase partitioning and mass versus time at a series of user-
specified time-steps. Graphical output is a series of plots for total mass and individual
component mass versus time. VENTING calculates contaminant mass in the subsurface
(averaged over the entire specified contaminated zone) over time during extraction from
a single well or a multiwell system based on a user-defined SVE flow rate. The code will
also calculate flow rate using Equation 9. Soil permeability may be specified by the user
or calculated from air pumping tests.
VENTING partitions contaminant mass at each time-step following the procedures
outlined in Johnson et al. (1990). The total initial mass of each contaminant in a mixture
at each time-step is numerical calculated using a variable time weighted scheme. Using
these procedures, the program is able to partition contaminants in four and three-phase
systems. VENTING also incorporates the diffusion-limited solution of Johnson et al.
(1990) by calculating a venting efficiency factor which accounts for vapor flow through
uncontaminated soils. Finally, the code also includes a biodegradation scheme whereby
the rate constant is calculated as a function of the amount of oxygen pulled through the
contaminated zone and a default theoretical oxygen requirement for all contaminant
species in dissolved-phase.
The multicomponent estimation procedures described in Section 3.2 were
compared to the results of employing the VENTING model for an example mixture of
benzene, toluene, and ethylbenzene. Figure 2 shows the results of this comparison.
Overall, the comparison indicates similar results for the same vapor flow rate. Major
differences may be attributed to the fact that the VENTING model repartitions the soil
concentration at each time-step, and thus calculates mole fractions based only on the
mass fractions in the residual phase. The multicomponent estimation procedures do not
repartition at each time-step but compute mole fractions as a function of the total mass
16
-------
UJ
o:
to
UJ
o
o
o
_J
o
CO
(0
0)
F
o
0)
N
c
d>
X)
0)
•o
c
TO
0>
3
0)
I
0>
(A
C
s
c
O
o
o
o
'o
(/>
••-f
0>
•a
c
o
a.
a>
i
o
c
o
(Q
Q.
E
O
O
G>
IT
o
6>|/6iu ''01103 n
17
-------
fractions. In addition, the VENTING model accounts for the presence or absence of a
residual phase as a function of phase saturation for the mixture of contaminants; the
multicomponent estimation procedures estimate phase saturation for individual
components rather than for the mixture. Finally, the VENTING model calculates Henry's
law constants while the estimation procedures use published Henry's law constants and
approximate four to three phase transition by assuming that activity coefficients are a
function of the total mass mole fractions.
18
-------
SECTION 4
REFINED ESTIMATES OF REMOVAL RATES
Refined models that are applicable to SVE systems can be categorized into 1) air
flow models and 2) compositional flow and transport models. Air flow models generally
are used to predict vapor flow rates and pressure distributions in the vadose zone due
to the induced pressure differential of a single or multiple extraction wells. Compositional
flow and transport models are also used to simulate SVE-induced air flow but included
algorithms to estimate contaminant phase partitioning and multicomponent transport in
the vadose zone. Because the main objective of modeling for risk evaluation is to
determine SVE removal rates (i.e., uncontrolled emission rates), only compositional flow
and transport models provide the vapor composition data needed to estimate air pathway
exposure point concentrations.
Refined models are most useful for risk evaluation purposes during the RD/RA
when more detailed site data are available. Compositional flow and transport models
attempt to simulate pressure distributions, vapor distributions and vapor flow in complex
two or three-dimensional domains and require detailed soil and contaminant distribution
data. As a result, the predicted removal rate and time required to deplete site
contaminants to target levels can be more rigorously estimated.
Recently, a review was performed of the mathematical models for evaluating SVE
systems (U.S. EPA, 1995). The following discussion regarding refined compositional flow
and transport models is taken from that review.
Compositional flow and transport models can be used to
estimate both the subsurface airflow regime and the transport
and removal of contaminants via SVE. These models can
also be used to estimate post-remedial soil concentrations in
order to determine if cleanup goals can be met. Flow and
19
-------
transport modeling can be used to generate estimates for
most aspects of SVE design, ranging from well placement to
contaminant extraction rates. However, the level of complexity
of the modeling and data requirements are also substantially
greater than that associated with screening and air flow
models.
A compositional flow and transport model should be used
when it is necessary to attempt to estimate both the air flow
regime and the transport of contaminants in the subsurface.
The transport portion of the model allows more robust
analysis of contaminant removal rates than simplified
screening models, and provides estimates of contaminant
levels in soil over the operating life of the SVE system.
Compositional flow and transport models can be applied to
estimate the concentration of a specified contaminant in the
vadose zone for any given point in space and time. Thus, it
may be advantageous to develop a transport model to
examine mass removal trends given sufficient site
characterization data.
Simulating vapor transport increases both the complexity and
uncertainty associated with a modeling analysis. In addition
to the parameters which affect air flow, transport simulation
also requires characterization of the contaminant distribution
and the contaminant media, and fluid properties which affect
transport but not flow. Characterizing contaminant
distribution, transport properties, and transport processes is
typically subject to significantly greater uncertainty than
characterizing flow properties and processes.
Results from a transport model consist of contaminant
concentrations throughout the model region over time in soil
and soil vapor. As such, transport models can be used to
predict or evaluate contaminant removal rates and residual
concentrations. These results can be used to assess the
overall efficacy of an SVE system and identify problem areas
where mass removal rates are inadequate. Areas where
simulated contaminant concentrations remain at unacceptable
levels for the duration of the planned SVE operation may
require the installation of additional extraction or injection
wells. Contaminant removal rates can be used to develop a
general estimate of cleanup time. A determination of residual
concentrations can be useful both in predicting possible
20
-------
cleanup levels and determining which, if any, semivolatile
compounds may remain in place after SVE has been
completed. Such an analysis requires characterization of the
existing contaminant distribution.
Two commercially available computer codes for refined modeling were reviewed
for this report. The VENT2D model, Benson et al. (1993) simulates steady-state air flow
and transient vapor-phase transport of multicomponent mixtures. Any number of vapor
extraction or injection wells can be simulated, and grid-variable soil permeability, and initial
contaminant distribution can be specified. An upper confining layer can be included to
simulate the effects of surface boundary condition effects. The equilibrium partitioning
distribution of contaminant in residual, sorbed, dissolved, and vapor-phases is calculated,
as are the effects of contaminant retardation. The transport of soil moisture is simulated,
and permeability is a time-and space-dependant function of soil moisture and NAPL
saturation.
With the VENT2D model, Benson et al. (1993) take a convenient approach to the
solution of the adjective-dispersive equation by neglecting mechanical dispersion in favor
of diffusion. This allows a simpler finite difference numerical formulation to solve the
transport equation, thus enabling transport modeling of multicomponent mixtures with
retardation. Equilibrium phase partitioning is assumed and the presence of NAPL is
indicated when all non-NAPL phases are saturated. Sorption to soil solids of individual
constituents is calculated based on octanol/water partition coefficients and the soil
fraction of organic carbon.
The model solves the steady-state equation of air flow based on the principle of
conservation of mass. The concept of relative permeability in which effective permeability
varies as a function of changing air-filled porosity is incorporated in VENT2D. Permeability
increases with decreasing soil moisture and NAPL saturation due to drying induced by
SVE. In order to save computational time, the permeability distribution is only
recalculated when 10 percent or more of the active cells' permeability values have
changed by 25 percent or more.
21
-------
VENT2D operates in the DOS environment and does not provide any input or
output interface or menu.'. The code author suggests that an effective approach to
developing a simulation is to modify one of the example input files provided to suit the
user's needs. A utility program is included that allows conversion of SURFER grid files
to ASCII text so that site concentration contour maps can be used as initial concentration
input to VENT2D. The code includes several composition files, including fresh and
weathered gasoline, and common solvents.
The outer boundary of the model domain is set by default as follows: no-flow at the
bottom of the domain (water table); and constant, atmospheric pressure and zero
concentration elsewhere around the model perimeter. A surface confining layer is
allowed, whereby the user can control vertical leakage by varying the vertical permeability
of the confining layer. Extraction wells are represented by constant-flow cells. The
pressure at an extraction well can be determined based on the pressure distribution
output from the model. An additional feature of VENT2D allows placement of zero-
concentration nodes at the model surface to simulate contaminant diffusion to the
atmosphere.
VENT2D produces a wide range of output data, including distributions for pressure,
permeability, air flow specific discharge (analogous to air flow velocity), soil moisture
content, total VOC and individual component concentrations in soil and soil gas, and
NAPL. A utility program included with the software will convert any of the parameter
maps in the output file to a SURFER file format for contouring (U.S. EPA, 1995).
VENT2D allows two-dimensional (areal or cross-sectional) analysis of air flow and
multicomponent transport with multiple extraction wells. A three-dimensional extension
of this code called VENT3D has also become available recently. Some of the new
features incorporated within VENT3D include the ability for individual layers to be given
unique values of thickness, anisotropy, porosity, moisture content, organic carbon
content, and distributions of permeability and contamination. At the time of publication
of this report, VENT2D or VENT3D is available from the author at a price of $495.
Recently, a new commercial model has become available with many of the same
features as the VENT2D model but which offers a menu-driven operating environment.
22
-------
AIRFLOW/SVE (Guiguer et a!., 1995) simulates soil vapor flow in an axisymmetric
cylindrical coordinate system whereby the vacuum extraction well is located at the center
of a theoretical cylinder. The model boundaries are the ground water table, the ground
surface, and the well radius of influence. The coordinate system results in a vertical axis
representing elevation above the water table, and a horizontal axis representing the radial
distance from the extraction well. The model simulates gas flow and vapor transport in
the radial and vertical directions. All soil properties and water saturation are assumed to
be homogeneous in the radial direction. Multiphase flow and transport are approximated
using a finite difference solution to the governing partial differential equations.
AIRFLOW/SVE features afully menu-driven operating system. From user-specified
data, the model generates a finite difference grid of the SVE environment. The user is
then transferred to a computer-aided design (CAD) environment where the grid can be
altered and model parameters manipulated. After execution of the computational analysis,
the results can be visualized and output to the screen, printer, plotter or to a graphics file.
Output includes pressure contours, velocity vectors, flow path lines, isoconcentration
contours, and NAPL distribution contours, as well as concentration, mass, and removal
rate versus time. At the time of publication of this report, AIRFLOW/SVE is available from
Waterloo Hydrogeologic Software at a price of $1,195.
Air flow and compositional flow and transport models generally incorporate more
assumptions and require more data input than simplified screening models. The user
must be able to select and apply increasingly complex boundary and initial conditions
effectively. As the complexity of the model increases, so does the complexity of the
results and their potential uncertainty. Simulation results of a compositional flow and
transport model for a multicomponent mixture depend on numerous properties of each
component. In order to effectively interpret the results of the model, the modeler must
understand the physics, chemistry, and mathematics of the model application (U.S. EPA,
1995).
To evaluate short-term risk during the remedial action and, if applicable, during the
five-year review, perhaps the best use of refined modeling is to estimate the tine for
source depletion and the vacuum well vapor constituent distributions as a function of time.
23
-------
This may best be accomplished by modeling the vacuum well removal rates at the start
of SVE system operation to.predict the decrease in component residual mass in the soil.
At periodic intervals after the start of remediation, the SVE system can be shut down and
the soil contaminants allowed to re-equilibrate. Upon restart of the system, measured
contaminant concentrations and field measurements of pressure distributions are used
to remodel the removal rates on a contaminant-specific basis. In this manner, the model
may be calibrated to field conditions and afford more accurate predictions of vapor flow
rate, total mass removed as a function of time, and the removal rates of individual
contaminants. Calibration may be achieved by applying field pressure distribution data
to the model variables affecting vapor flow rate, and by altering contaminant input
variables (e.g., initial concentration) so that the calculated vapor-phase concentrations
match measured values.
24
-------
SECTION 5
LIMITATIONS AND ASSUMPTIONS OF MODELS REVIEWED
Three screening models were reviewed for this report: 1) Hyperventilate,
2) VENTING, and 3) a multicomponent estimation model. All three models are based on
the work of Johnson et al. (1990 and 1990a). As such, the limitations and assumptions
of the analytical vapor flow rate solution are identical. The three models vary, however,
in the procedures used to estimate vapor-phase concentrations and distributions of
contaminants.
5.1 GAS FLOW
The three screening models solve for steady-state radial flow in a one-dimensional
confined porous medium using Equation 9. This solution is normally used to calculate
a range of predicted vapor flow rates based on soil vapor permeability and well pressure.
Equation 9 predicts the rate of vapor flow through an ideal isotropic soil column in that
actual flow is assumed to be unimpeded due to diffusion mass transfer limitations
between a contaminated soil layer and a gas flow field which passes parallel to but not
through the contaminated layer (e.g., for a lens of floating product). In addition, the
solution to vapor flow does not consider air flow variations due to impedances such as
unconnected pore spaces, variations in pore size and volumetric water content, or
variations in lateral and vertical permeability (leaky or confining layers).
The Hyperventilate and VENTING models attempt to account for nonideal flow with
a user-specified or calculated venting efficiency factor. If part of the total gas flow does
not pass through the contaminated soil, an efficiency factor (0) will arise such that if the
screened interval (H) is specified as the total unsaturated zone thickness, the factor will
be equal to the ratio of contaminated zone thickness to total unsaturated zone thickness.
25
-------
This accounts for the dilution effects of "clean" air flowing parallel to but not through the
contamination.
To account for diffusion-limited mass transfer, Hyperventilate and VENTING employ
the procedure of Johnson et al. (1990) to calculate a diffusion-limiting efficiency factor as:
1 =
2DEn Rl In (R/RW)
3 H2 kv (Patm - Pw)
(12)
where D^ is the effective diffusion coefficient (cnf/s) and f\ is the radius of the
contaminated soil zone, F\. < f\, (cm). The effective diffusion coefficient is given as
DE = DA(0a333/rf), where DA is the diffusion coefficient in air (cirP/s), 0a is the soil air-
filled porosity (unitless) and n is the total soil porosity (unitless). The diffusion-limiting
efficiency factor may be used to account for nonideal conditions within the soil column,
such as a floating product layer at the top of the water table or a pocket of contamination
within a clay lens of lower permeability. The factor relates venting efficiency as limited by
diffusion from subsurface locations where the gas flow field passes parallel to, but not
through the contaminated layer. In the event that multiple factors such as fractional swept
volume and diffusion limitations occur simultaneously, the net efficiency is the product of
each efficiency factor.
Calculation of the efficiency factors described above may also be applied to the
multicomponent estimation model described herein such that
ER, = (1 - 0)nQC,7 (13)
where EF^ is the uncontrolled emission rate (removal rate) of component i (g/s); 0 is the
fraction of air flow through uncontaminated soil; rj is the fractional efficiency due to
diffusion-limited mass transfer (Equation 12); Q is the vapor flow rate calculated by
26
-------
Equation 9 (cm3/s); and Qveq is the equilibrium vapor concentration (g/cm3).
Alternatively, the user may specify efficiency factors based on observation, experience,
field measurements, or professional judgement.
5.2 VAPOR CONCENTRATION
The three screening models and the refined models reviewed for this report use
multiphase partitioning of homogeneously contaminated soil to derive the vapor-phase
concentration. The Hyperventilate code is applicable only when a residual phase is
present and calculates the maximum vapor-phase concentration, and thus removal rate,
by employing Raoult's law. Therefore, Hyperventilate cannot estimate removal rates as
a function of time. The VENTING code, VENT2D and VENT3D, and AIRFLOW/SVE
employ a molar mass-balance approach whereby the vapor-phase concentration is
derived as a function of the degree of phase saturation. These programs are capable of
equilibrating four and three-phase systems and can predict time-dependent removal rates.
The multicomponent estimation model can also estimate four and three-phase equilibrium
but is less accurate than the molar mass balance approach.
Most vapor transport models developed to date assume local equilibrium between
phases to simplify calculation of chemical mass distribution, movement, and interphase
transfer.
The local equilibrium assumption presumes: 1) the rate of chemical movement
through any phase in the vadose zone is slow compared to the rate of mass transfer
between phases, 2) chemical concentrations in all phases are therefore in thermo-
dynamic equilibrium, 3) mass transfer is reversible, 4) phase equilibrium is independent
of the presence of other phases, and 5) linear partitioning occurs which does not vary
with chemical concentration (U.S. EPA, 1995).
The soil water partition coefficient, KD, is a valid representation of the partitioning
between dissolved and sorbed-phases only if partitioning is fast compared to vapor flow,
reversible and linear. Although this is true for many organic compounds, it is not true for
others. Karickhoff et al. (1979) investigated the sorption and desorption of organic
27
-------
compounds and found that a very rapid component of sorption was followed by a slower
component visualized as diffusive transport to the interior sorption sites that were
inaccessible to the fluid (U.S. EPA, 1995). Sorption, therefore, appears to be a
combination of adsorption to the surface of the sorbent as well as adsorption within the
micropores of the soil particles. In combination with nonlinear desorption and
confinement of contaminant in isolated layers (e.g., clays), this phenomenon may be
responsible for much of the reported discrepancies between simulated removal rates and
observed removal rates.
Typically, the rate of contaminant removal usually declines rapidly in the initial stage
of remediation as the more volatile constituents are removed by advection. This initial
stage is followed by a transitional stage between advective and diffusive mass transfer.
Finally, a third stage marks the diffusion-limiting mass transfer which is characterized by
the relatively slow process of contaminant diffusion from dead-end pores and stagnation
zones, evaporation from the dissolved-phase, and desorption from the sorbed-phase.
The mass balance approach to local phase equilibrium used by the computer models
reviewed for this report and the approach used by the multicomponent estimation model
do not account for nonlinear desorption nor for retarded desorption from the interiors of
the sorbing clays and organic matter particles.
This combination of rate-limiting factors, including fractional vapor flow through the
contaminated zone, diffusion-dominated mass transfer, and nonideal phase equilibria, may
combine to produce a "tailing" effect whereby the actual time required to reduce the total
contaminant mass in the soil may be significantly longer than predicted values.
28
-------
SECTION 6
ESTIMATION OF AMBIENT AIR CONCENTRATIONS
Estimates of short-term or long-term ambient air concentrations may be obtained
by using site-specific release parameters with the EPA SCREENS Model (Version 96043,
EPA 1996), TSCREEN Model (Version 94133, EPA 1994), or the Industrial Source
Complex (ISC3) Short-Term Model (Version 96113, EPA 1996).
The EPA OAQPS has developed an electronic bulletin board network to facilitate
the exchange of information and technology associated with air pollution control. This
network, entitled the OAQPS Technology Transfer Network (TTN), is comprised of
individual bulletin boards that provide information on OAQPS organization, emission
measurement methods, regulatory air quality models, emission estimation methods, Clean
Air Act Amendments, training courses, and control technology methods. The dispersion
model codes and user's guides referred to in this document, are all available on the TTN
in the bulletin board entitled SCRAM, short for Support Center for Regulatory Air Models.
Procedures for downloading these codes and documents are also detailed in the SCRAM
bulletin board. The TTN may be accessed at the telephone number (919) 541-5742 for
users with 1200 or 2400 bps modems, or at the telephone number (919) 541-1447 for
users with a 9600 bps modem. The communications software should be configured with
the following parameter settings: 8 data bits, 1 stop bit, and no (N) parity. The SCRAM
is also available on the Internet via the world wide web through EPA's main v/ebsite
(http://www.epa.gov/scram001).
Depending upon time, resources, and modeling experience, one may use these
dispersion models as an evaluation tool to quickly assess ambient air concentrations.
Depending upon the contaminants involved, air concentrations can then be used to
perform risk evaluations or ARAR compliance evaluations. The ISC3ST and TSCREEN
29
-------
models can handle a variety of averaging times such as 1-h, 3-h, 8-h, 24-h, and annual.
The SCREENS model only computes ambient concentrations for a 1-h averaging period.
For compounds that have chronic toxicity values or long-term regulatory standards,
ISC3ST and TSCREEN provide an annual ambient air concentration. If SCREENS is used,
one must adjust the hourly concentration by applying a conversion factor of 0.08 to
convert to an annual averaging period. The conversion factors for estimating an annual,
24-h, 8-h, and 3-h ambient air concentration are presented below in Table 1.
TABLE 1. CONVERSION FACTORS FOR 1-HOUR AMBIENT AIR CONCENTRATIONS
Convert to
3-h
8-h
24-h
Annual
Conversion factor
0.9
0.7
0.4
0.08
If the SVE site is located in an area with receptors located far enough away (critical
distances vary according to contaminants involved, emission rates, and equipment
parameters), the SCREENS or TSCREEN models may satisfy the necessary modeling
requirements. If receptors are located nearby or if stringent regulatory standards are to
be met, the ISC3ST model using actual meteorological data is recommended. Other
advantages include more flexible averaging periods and more accurate handling of
building downwash and complex terrain.
These air dispersion models can also be used to help assess design parameters
for the SVE system as well as address the effects of control options, if needed. An SVE
system is modeled as a point source in most cases, and the stack parameters used in
the air dispersion models will be the same. Data required are as follows:
Emission rate (g/s)
Stack height (m)
Stack inside diameter (m)
Stack gas exit velocity (m/s)
30
-------
0 Stack gas exit temperature (° K).
Additional data required ar6 use of either the urban or rural coefficients of dispersion,
receptor coordinates or distances, and meteorological data (ambient temperature for the
screening models and a meteorological data file for the ISC3ST model). If the SVE
system will be operated for less than one year, the ISC3ST model includes an option (i.e.,
EMISFACT) that allows the user to zero emissions for any month (multiplies the emission
rate by a factor), which will affect the annual ambient air concentrations accordingly. In
addition, if the SVE system is shut down during the day or evening to allow the vapors
to accumulate, the EMISFACT allows the user to zero emissions for any hour of a 24-h
day, which allows for more accurate 24-h concentrations.
To help assess the effects of SVE size and/or evaluate the risks to receptors at
the site, Tables 2 through 4 give typical SVE parameters according to control options.
These scenarios were developed based on a review of the existing literature and
discussions with researchers in the SVE area.
Four air emission control options are typically employed for SVE systems: 1) no
controls, 2) activated carbon control systems, 3) catalytic oxidation control systems, and
4) thermal oxidation control systems. Thermal and catalytic oxidation emission control
systems can reduce emissions by 90 to 95 percent. Activated carbon systems have
control efficiencies that are compound specific. The reader is referred to U.S. EPA 1989
and 1992 for more information on control options for soil vapor extraction technologies.
31
-------
TABLE 2. EXAMPLE SCENARIOS FOR SVE WITH NO CONTROLS BASED ON SIZE OF
SYSTEM
Parameter
Exhaust Gas Flow Rate
Exhaust Gas Velocity
Exit Gas Temperature
Stack Height
Stack Diameter
Units
nf/min
cfm
m/sec
°C
m
m
Scenario
Very Small
1.4
50
3.0
50
3.0
0.10
Small
14
500
7.4
50
4.6
0.20
Medium
85
3,000
12.5
50
7.6
0.38
Large
425
15,000
14.29
50
9.1
0.46
aAssume three adjacent stacks each handling 5,000 cfm. The flow is split to lower the
velocity of the exiting gas to typical design levels to minimize corrosion of the stack.
TABLE 3. EXAMPLE SCENARIOS FOR SVE WITH ACTIVATED CARBON CONTROLS
BASED ON SIZE OF SYSTEM
Parameter
Exhaust Gas Flow Rate
Exhaust Gas Velocity
Exit Gas Temperature
Stack Height
Stack Diameter
Units
m3/min
cfm
m/sec
°C
m
m
Scenario
Very Small
1.3
46
2.8
25
3.0
0.10
Small
13
461
6.9
25
4.6
0.20
Medium
78
2,770
11.5
25
7.6
0.38
Large
392
13,800
13.1a
25
9.1
0.46
'Assume three adjacent stacks each handling 4,600 cfm. The flow is split to lower the
velocity of the exiting gas to typical design levels to minimize corrosion of the stack.
32
-------
TABLE 4. EXAMPLE SCENARIOS FOR SVE WITH CATALYTIC OR THERMAL
OXIDATION .CONTROLS BASED ON SIZE OF SYSTEM
Parameter
Exhaust Gas Flow Rate
Exhaust Gas Velocity
Exit Gas Temperature
Stack Height
Stack Diameter
Units
mP/min
cfm
m/sec
°C
m
m
Scenario
Very Small
2.6
92
5.5
320
3.0
0.10
Small
26
918
13.8
320
4.6
0.20
Medium
156
5,510
22.9
320
7.6
0.38
Large
780
27,500
26. 1a
320
9.1
0.46
aAssume three adjacent stacks each handling 9,200 cfm. The flow is split to lower the
velocity of the exiting gas to typical design levels to minimize corrosion of the stack.
33
-------
SECTION 7
RECOMMENDATIONS
The following procedures are recommended for estimating uncontrolled emissions
from SVE systems for the purposes of conducting an air pathway evaluation. The
recommendations are categorized as to when the evaluation is performed within the
remedial process. When applied, these procedures will provide theoretical predictions of
removal rates of individual contaminants based on increasingly rigorous estimates of
vapor flow in the vadose zone and contaminant phase partitioning.
7.1 ESTIMATING EMISSIONS DURING THE FEASIBILITY STUDY
During the feasibility study (FS), a range of remedial alternatives is identified, if
necessary, and each alternative is evaluated with respect to effectiveness,
implementability, and cost. This process typically involves development of alternatives
and screening of technologies. Those alternatives that are clearly unfavorable relative to
other alternatives in terms of effectiveness or implementability, or that are grossly
excessive in cost are dropped from consideration after the screening. During the detailed
analysis of alternatives, more site-specific data are typically available to refine the
screening procedures to evaluate the potential risks to the community and to workers.
Screening-level models are therefore more applicable during the FS in making preliminary
assessments of SVE feasibility at contaminated sites. Screening models can be applied
with very limited site data, and thus can be used early in the remedial process.
The Hyperventilate model is perhaps best used during the initial identification and
screening of remedial alternatives. This model is designed primarily to assess the
feasibility of SVE and to grossly estimate costs. Hyperventilate is capable of estimating
subsurface vapor flow under ideal conditions and capable of estimating the maximum
34
-------
uncontrolled emissions of individual contaminants. Hyperventilate does not compute
time-dependent emissions;and is only applicable at sites were NAPL is present.
Hyperventilate does not predict the time required to deplete initial soil concentrations to
a target concentration. If NAPL is present at the site, and if the soil target concentration
is above the soil saturation concentration, Hyperventilate can be used to estimate
maximum emissions for a user-specified remediation time interval.
The VENTING model is a better choice for estimating 1) time-dependent emissions,
2) time required for initial soil contaminant depletion, and 3) residual contaminant soil
concentrations as a function of time. VENTING may be employed at sites where NAPL
is not present or when target soil cleanup concentrations are below the soil saturation
concentration. VENTING uses the same subsurface vapor flow algorithms as the
Hyperventilate code but is capable of four and three-phase contaminant partitioning.
Finally, the multicomponent estimation model described in this document and
detailed in Appendix A may be used if the VENTING commercial code is not available.
This model uses the same radial flow algorithms as the VENTING code but simplifies
phase partitioning.
The Hyperventilate model directs the user through a step-by-step approach for
estimating SVE feasibility and maximum contaminant removal rates. Therefore, the user
is referred to the model user's manual (U.S. EPA, 1993) for specific guidance. The
following guidance is applicable for estimating time-dependent uncontrolled emissions
from SVE systems using either the VENTING model or the multicomponent estimation
model.
Application of the VENTING model or the multicomponent estimation model
involves estimating a range of predicted emissions based on a typical range of values for
the most sensitive model parameters. The most sensitive variables in calculating the
subsurface flow rate are soil permeability to vapor flow, vacuum well pressure, and soil
temperature.
Table 5 gives the flow rate model variables, the practicle range of variable values,
and the default variable values from U.S. EPA (1993 and 1995).
35
-------
TABLE 5. VAPOR FLOW RATE MODEL VARIABLES
Variable
Permeability (kj
- Medium sand
- Fine sand
- Silty sand
- Clayey silts
Well pressure (Pw)
Radius of influence (Rj)
Well radius (RJ
Soil temperature (Ts)
Screened interval
thickness
Practical range
0.01 to 100 Darcys
5 - 200 in. H20
3- 12m
5.08 - 60.96 cm
0-30°C
0.3 - 6 m
Default value(s)
10-100 Darcys
1-10 Darcys
0.1-1 Darcys
0.01 - 0.1 Darcys
20 - 120 in. HjO
12m
10.16 cm
11°C
Set H equal to thickness
of contamination
Vapor flow rates should be calculated using the range of default values unless site-
specific data are available.
Contaminant phase distributions are calculated by both models based on the initial
soil concentration (total volume) and soil properties. Table 6 gives the soil properties
required for phase partitioning and the default values for subsurface soils from U.S. EPA
(1994) if site-specific data are not available.
Once a range of vapor flow rates have been calculated, the subsurface equilibrium
vapor concentration of each contaminant is estimated for a series of time-steps. If site-
specific information is available, an estimate of the fraction of subsurface vapor flow
through uncontaminated soil (0) can be made. In addition, a fractional system efficiency
^7) can be calculated using Equation 12 to account for diffusion-limited mass transfer, if
applicable.
36
-------
TABLE 6. SOIL PROPERTY VARIABLES FOR PHASE PARTITIONING
Variable •
Soil dry bulk density (ob)
Soil particle density (os)
Total soil porosity (n)
Soil moisture content (w)
Soil water-filled porosity (0W)
Soil air-filled porosity (0a)
Soil organic carbon (LJ
Default value
1.5 g/crrf
2.65 g/cm3
0.434 cm3 /cm3
0.20 g/g
0.300 cm3 /cm3
0.134 cm3 /cm3
0.002 g/g
7.2 ESTIMATING EMISSION DURING THE RD/RA
Because more data are generally available during the remedial design, a
compositional flow and transport model may be more useful for estimating uncontrolled
emission rates. VENT2D, VENT3D, or AIRFLOW/SVE allow for solution of radial flow in
stratified soils for multiwell systems and attempt to compensate subsurface vapor flow due
to nonconfined boundaries and variations in site permeability.
Perhaps the best use of these types of models is to perform calibrated estimates
of removal rates and vapor concentrations throughout the life of the remediation. This
may best be accomplished by periodically recalibrating the model with measured
pressure, soil contaminant concentration, and vapor concentration distribution data. After
recalibration, the models may be rerun to examine unexpected variations in system
performance, to re-evaluate the time required to complete remediation, or the extent to
which SVE can attain the target soil concentrations.
7.3 DISPERSION MODELING TO ESTIMATE AMBIENT AIR IMPACTS
As in estimating emissions during the feasibility study, screening-level models are
used to make preliminary assessments of ambient air impacts from the site. Depending
upon site location with respect to receptors, contaminants involved, and/or equipment
37
-------
design and controls, a screening model may be all that is needed to conservatively
predict short- or long-term, ambient air concentrations. The SCREEN3 or TSCREEN
models offer quick results for a first look at ambient air impacts and in some cases may
be all that is needed to access the risks associated with a SVE unit. If refined modeling
is desired or needed to predict more accurate ambient air concentrations, the ISC3ST
model is recommended to provide short-term or annual average impacts. The major
advantage of the ISC3ST model is the use of actual meteorological data as well as
options to zero out emissions (e.g., EMISFACT) for any hour, month, or season the SVE
system is not operated.
With time-dependent uncontrolled emissions and an estimate of the time to deplete
the initial soil concentrations to target levels, atmospheric dispersion modeling can be
used to estimate exposure point concentrations from which potential risks can be
assessed. With these data, a decision can be made as to the need for air pollution
control equipment. If controls are required, an estimate can then be made of controlled
emission rates and the relative risks of remediation via soil vapor extraction.
38
-------
REFERENCES
Benson, D. A., D. Huntley, and D. C. Johnson. 1993. Modeling Vapor Extraction and
General Transport in the Presence of NAPL Mixtures and Nonideal Conditions. Ground
Water, 31 (3):437-445.
Environmental Systems and Technologies, Inc. (ES&T). 1994. VENTING: A program for
estimating hydrocarbon recovery from soil vacuum extraction systems, Version 3.01.
ES&T, Inc., Blacksburg, Virginia.
Guiger, N., T. Franz, and J. Zaidel. 1995. AIRFLOW/SVE, Axisymmetric Vapor Flow and
Transport Model, Version 1. Waterloo Hydrogeologic Software, Waterloo, Ontario,
Canada.
Johnson, P. C.( M. W. Kemblowski, and J. D. Colthart. 1990. Quantitative analysis for
the cleanup of hydrocarbon-contaminated soils by in situ soil venting. Ground Water,
28(3) :413-429.
Johnson, P. C., C. C. Stanely, M. W. Kemblowski, D. L Beyers, and J. D. Colthart.
1990a. A practical approach to the design, operation and monitoring of in situ soil-venting
systems. Ground Water Monitoring Review, Spring, p. 159-178.
Johnson, P. C., M. B. Hertz, and D. L. Beyers. 1990b. Estimates for hydrocarbon vapor
emissions resulting from service station remediations and buried gasoline-contaminated
soils, In Petroleum Contaminated Soils, Vol. 3. Lewis Publishers. Chelsea, Michigan.
Johnson, P. C. 1991. Hyperventilate User's Manual. Shell Oil Company.
Karickoff, S. W., D. S. Brown, and T. A. Scott. 1979. Sorption of Hydrophobic Pollutants
on Natural Sediments. Water Resources, 13:241-248.
Lyman, W. J., W. F. Reehl, and D. H. Rosenblatt. 1990. Handbook of Chemical Property
Estimation Methods. McGraw-Hill, New York.
U.S. Environmental Protection Agency. 1988. Screening Procedures for Estimating the
Air Quality Impact of Stationary Sources - Draft for Public Comment. EPA-450/4-88-010.
Research Triangle Park, North Carolina.
39
-------
U.S. Environmental Protection Agency. 1989. Soil Vapor Extraction VOC Control
Technology Assessment. EPA-450/4-89-017.
U.S. Environmental Protection Agency. 1990. Air/Superfund National Technical Guidance
Study Series, Air Stripper Design Manual. EPA-450/1-90-003.
U.S. Environmental Protection Agency. 1992. Control of Air Emissions from Superfund
Sites. EPA/625/R-92/012.
U.S. Environmental Protection Agency. 1993. Decision-Support Software for Soil Vapor
Extraction Technology Application: Hyperventilate. EPA/600/R-93/028.
U.S. Environmental Protection Agency. 1994. Technical Background Document for Soil
Screening Guidance, Review Draft. EPA/540/R-94-102.
U.S. Environmental Protection Agency. 1994a. User's Guide to TSCREEN. EPA-454/B-
94-023.
U.S. Environmental Protection Agency. 1995. Review of Mathematical Modeling for
Evaluating Soil Vapor Extraction Systems. EPA/540/R-95-513.
U.S. Environmental Protection Agency. 1995a. User's Guide for the Industrial Source
Complex (ISC3) Dispersion Models. EPA-454/B-95-003a.
U.S. Environmental Protection Agency. 1995b. SCREENS Model User's Guide. EPA-
454/B-95-004.
U.S. Environmental Protection Agency. 1995c. Screening Procedures for Estimating the
Air Quality Impact of Stationary Sources, Revised. EPA-450/R-92-019. Research Triangle
Park, North Carolina.
40
-------
APPENDIX A
MULTICOMPONENT ESTIMATION MODEL (CASE EXAMPLE)
A-1
-------
X
E
-j
*
-
X
o
u.
Ill
0
o
m
8
LU
LU
X
CO
LU
Q
£
CO
1
P
ENTER Chemical P
,
m
1
1
•5
Ul
I
to
cJ
1
1
P
I
|
I
%
8.
5
Henr/s law
5
1
"ni
1
1
i
^
_
3
i
E
the normal
constant
i
c
o
JtS
I
3
8
olecular
s
CO
*
jf
8
i
•5
c
1
8.
t?
'S.
boiling point,
CN
•&
to
coefficient,
S"
C
£
0)
fr
cV
"
o-
h?
|
i
-
-
03
1
0
1
!
i
t.
!
i
£
E
I
"o
"E
S.
1
E.
1
S
B
I
1
6
^
CN
CO
O
O
IO
o
CN
I
.70E-02
CD
T-
CM
353.24
CN
s
5.55E-03
CN
in
o>
5
P
00
Benzene
CO
?
S
in
R
3
.70E-02
3)
5
CO
i
6 64E-03
o
CO
>
8
?
Toluene
;
9
S
10
0
I
.50E-02
o
CN
0
CO
i
7.88E-03
0
1
en
CO
CO
cb
o
Ethyl benzene
m
T5
in
0
I
i
in
o
5
[S
8
IO
OO
7.34E-03
00
?
CO
t-
co
co"
o
t
CO
S
in
0
I
20-306'
o
CN
s
«
?
1.03E-02
S
CD
O
O
o
CO
Tnchtoroethylene
h;
?
s
in
0
3
20-302'
0
8
i
S
1.84E-02
o>
•
«
S
s
Tetrachloroethylene
00
9
8
in'
0
1
.04E-01
0
8
CN
CO
i
3.67E-03
«
.
h-
O)
CO
0)
Chloroform
0)
•5
o
o
in
o
CN
1
20-308'
o
CO
8
s
CN
3.04E-02
co
S
CO
en
N.
CO
CO
m
Carbon Tetrachloride
o
CN
CO
O
O
to
o
CN
I
.04E-01
o
o
8
356.60
I
9.79E-04
8
.
O
CN
IT
CO
CO
O)
CO
O)
1,2-Dichloroethane
CN
9
in
0
1
20-308'
o
o
s
N
S
r>
K.
1.72E-02
8
CO
o
CO
CO
CN
CO
CO
1,1,1 -Tnchloroethane
CN
CN
CO
CN
(0
CD
8.
E
a.
'5
w
tr
UJ
r-
UJ
CN
CO
CO
CO
CO
in
CN
CO
CN
1
B
CO
I
I
verage
fr
1
CN
thickness of
o
1
o
soil moisture
i
1
CO
CN
C
1
f
8
1
C
0)
"tf
CD
CN
d
Jt
s
-
H"
O
CO
E
"E
1
n
T
e
1
CO
CN
CO
CO
CN
O
O
CD
CM
O
0
-
in
co
CO
Conditions
UJ
1
J5
g
CO
ENTER Default or C
in
CO
o>
T3
o
ra
(D
(D
CO
X.
.
CD
CO
s
1
Consider diffusion-
I
h-
co
i
_C
I
x:
G
1
1
0)
t;
1
creened
CO
1
CO
CO
a
**
i
1
i
uncontaminated
'
radius of ^
1
1
"S
0)
•t
n:
"r
1
8
i
I
i
(Enter YES, NO,
'8
•5
influence,
8f
13
Q.
I
ckness,
£
C
o
SJ
or User Defined
*
»
-
.'
Q?
I
0?
5
i
F
id
2.
Fractional Efficiency)
I
I
1
&
1
l"
l"
CN
CO
CN
IO
O
i
o
o
CO
o
o
CN
34.8084
s
§
CO
o
5
in
a = default value
CO
f
TB
jh uncontaminat
vapor flow throu;
enter fraction of
0°
A
X
^
(Q
1
o"
O
X
Q
II
5j
o:
ntaminatlon (i.e., R
real extent of cor
sfer across the a
mlted mass tran
§
(O
S
c = entering YES assum
CO
S
•«
be entered as a frac
fficiency should
•g
1
fraction
c
C/1
*:
CT
C
'C
CO
o
ass at time-zero.
t
Is
i
1
a
2
«
c?
Q
i
j necessary whe
J5
E
d = shorter time-step mt
o
in
1
1
£
X
6?
1
1
i
lyorallofthefol
necessary for ar
S
?
e = longer time-step inte
T-
in
transfer.
X
1
1
TJ
c3
15
c
in
0}
c
extraction well va
1
"6
V)
3) low permeability
CN
in
-------
>
i-
M
tc
a
a.
0
8
co
LATION
3
O
U
g
UJ
8
a
^
UJ
X
o
1
CO
^
in
CO
"o
CO
s.
$
8.
•S
8.
to
I
f
5
I
0
"ra
.c
LU
o
1
g
g
ressure
a
pressure
^
•w
8
•a
O
c
o
8.
5
CO
o
1
!
To
£
•-
v
1-
f3
o>
J
CC
L
i
f
i'
O
^
!
§
"
S.
I
•»
S
^
f
E
£
,_^
1
~
^
CN
N
in
CM
M
0
if)
0
0
01253
c!
o
9
in
CO
CN
(D
CO
en
£
I
CO
31
CM
QO
h.
CM
O
S
o
0
in
o>
CO
o
d
T
CN
CO
o
CO
9
III
g
CO
00
o
CM
CD
0)
12
9
O
CO
9
III
CO
CO
CO
£
n
1
I
LU
£
•x.
O)
8
o
r—
g
o
o
d
CO
o
s
III
CO
CO
CO
s
n
1
t
CD
V
CO
D
O
0
S
o
i
o
d
in
CS
d
s
o
?
Ill
in
O)
0
1
5
Oi
•^
3
0
o
s
h-
o
o
h-
?
d
1
o
s
III
h~
in
fn
S
1
1-
^
O
o
CM
s«
CO
2
o
o
d
0
o
s
III
a>
In
m
m
1
a
CM
CN
r>
S
J
0
o
§
o
0.1618
CM
R
O
CO
9
in
o>
CO
CO
s
f-.
ra
hloroeth
o
K
^-'
CN
CN
CT
CM
CN
CD
CM
CM
S
OJ
0
CO
CN
CO
CO
CO
CO
m
CO
(D
CO
CO
CO
CO
CD
CO
o
,
CM
CO
,
in
CD
^
CO
§
S
5
ir
•
CO
u_
O
CM
-------
N
1
ed
O
0?
-------
i
j
^
,
"*
5
5
S
„.
o
(0
z
o
N CALCULA1
O
(0
(0
IU
or
0)
c
m
CO
i
i
D
Hi
/•)
5
i
c
1
1
1
UJ
•B
(0
(0
1
s
E
a
0)
o
"o
W
1
contaminant
_.
S
I
li
8
g
£
H:
8
8
1
8
£'
n*
UJ
*d
0
0
1
(-
p
1
•i?
s
1
~
f
I
1
1
S
I
TJ
O
T-
v
in
c|
UJ
1
(N
?
UJ
in
CO
T-
O
4 97E-03
CN
Si
1.0121E+07
m
CN
CO
in
5
§
5
(SI
!?
S
CO
2
o
4.75E-03
1
o
o
I-
_l
a
CALCULATED J
f
I-
CO
i
NATION REMAIN
8
H
0>
z
LESS THA
COMPOUND IS
8
UJ
rr
a
8
o
0)
rCTIVE TARGE!
Q.
to
UJ
o:
CN
5
_)
IED IN COL
NATION SPECIF
fe
y
CN
CN
S
S
8
(0
CN
h-
CN
CO
CN
O>
CN
O
CO
CO
s
CO
CO
CO
in
CO
CD
CO
CO
CO
CO
0)
CO
0
?
CN
CO
,
in
*
,
CO
0>
s
in
CN
tn
u.
O
-------
3
S
3
£
S
L
§
Z
^
O
u
in
2
g
u
o
z
o
w
tf)
•g
UJ
g
CE
c
T
,2
CO
TT
|
•g
p
1
1
IO
emission
.2
1
Tj>
e
in
UJ
•5
1
CO
•
1
a
5
1
'6
CO
'o
cy)
^
S
%
^
I
?
&
8
|
c
<
J
1
^
CD
0?
UJ
t
UJ
?d
*
0°
0°
S
F
p
o>
P
1
^^
S
S
1
1
"
a
§-
•S
o
^
1.71E+04I
9
UJ
.
?
UJ
CO
O)
S
o'
I
10"
¥
LU
O
IO
O
CN
¥
UJ
«
9
Ul
.
?
°
o>
o
4.99E-03
CN
O)
O)
¥
c^
CD
0
10
CN
CO
1.41E+05I
9
UJ
*-
CO
^
UJ
CN
CD
O>
8
O
492E-03
CN
O)
fe
UJ
ft
^
10
CD
^
CO
0)
S
CN
CM
CN
a
CN
CN
JN
CM
8
83
8
S
%
CO
CO
m
CO
S
r-
CO
00
CO
o>
CO
o
S
CN
s?
3
£
^
!?
"
™
S
E
S
u.
O
in
-------
1
9-91
8
S
-------
s
2
a
a
o
m
u.
m
in
m
s
8
to
O
K-
nnoivo N
o
CO
co
UJ
a:
11
X
E
CO
^
I
D
1
B
I
in
emission |
S
y>
"
1
t
to
mass of
CD
1
*
^
g
0
'6
CO
"o
CO
contaminant
r-
I
1
k
r
§
g
1
_
°
1
S
OO
Q?
UJ
0?
r
<3
0°
0
s
F
O)
P
1
I
--
^l
f
i
1
fn
|
—
S
•o
o
T-
4.14E+03|
9
UJ
?
in
ft
8
o
UJ
g
in
o
in
1 0179E+07
in
fN
o
CM
3.33E+04|
.93E-01
?
CM
rsi
o
o
UJ
g
in
s
OJ
1.0175E+07
in
fN
CM
CO
3.44E+04|
9
in
eo
CM
o
o
111
CO
O)
*
CM
0)
1.0141E+07
in
fN
*
u>
to
i-
CO
0>
0
v-
CM
CO
*
in
to
r-
co
CN
g>
0
5
CM
CO
CO
CO
CO
IO
CO
CO
CO
CO
CO
CO
o
,
CM
CO
?
in
%
t£
9
§
g
in
CM
in
ft
f-
-------
g
&
S
!?S
9
§
-------
ft
0?
UJ
g
a
*
0
CO
b
(0
, z
:|
1
§
w
w
UJ
3
8
c^c?cf
8
S
+04|
g
-------
CO
UJ
ted
a
i.
§
¥
-------
o
8
O
1
$
99
+05]
R
-------
8
g
y*
3
l_
O
8
g
8
co
7
O
3
§
g
z
o
co
CO
S
UJ
a:
1
q>
2
o
9
fN
CO
,
1
1
1
1
in
c
o
1
emission
•—
LU
mass of ,
(O
.
1
8.
m
6
^
'5
CO
-=
CO
:ontammant
1
8
1
i
&
S
R
(i
8
1
S
CO
0?
LU
•j
^f
0
o
•s.
0}
i=
.
P
,-t;
B
i
I
"E"
o
I
1
"oi
^
?
§
f
0
T-
42E+04
*
2.04E+00
42E-05
CN
T
O
9
LU
8
m
tn
1 0179E+07
in
r\
o
CN
79E-H35
CO
2.20E+00
60E-05
CN
J
O
9
LU
S
v
CO
rn
*
1 0135E+07
in
fN
CN
CO
SO+36/
CO
1
CN
59E-05
CM
•»
O
m
9
UJ
0)
"
CN
ff
^
9 7552E+06
in
fN
v
in
-------
§
u.
Q
Ul
Q
Q
o
Q
01
0
§
8
3
0
CO
o
H-
CALCUL/
z
0
CO
CO
Ul
o
o:
i
n
.c
o
fi
^_-
CO
^r
1
Uncontrolled
I
in
emission
emission
stimated
UJ
mass of
CD
1
1
a
5
0)
0
f
'o
CO
1
Dontaminant
r~
1
£
I
8
O
C
o
t3
*=
8
6
1
8
CO
of
UJ
a?
U)
»j
*
O
0
s
0)
CO
P
I
I
I
^^
b
t.
1
1
§
f
0
£1
9.55E+04
4.42E+00
5.23E-05
in
s
o
CO
9
UJ
o
o
in
o
8
in
1 0179E+07
in
o
m
m
CM
in
u_
O
CO
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
1. REPORT NO.
EPA-451/R-96-007
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Air/Superfund National Technical Guidance Study Series,
Estimation of Air Impacts for Soil Vapor Extraction (SVE)
Systems (Revised)
5. REPORT DATE
August 1996
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Craig S. Mann
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Quality Management, Inc.
Cedar Terrace Office Park, Suite 250
3325 Durham-Chapel Hill Boulevard
Durham, North Carolina 27707-2646
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-D3-0032
Work Assignment No. 11-59
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Final
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The purpose of this document is to provide both screening level and refined models for estimating
the uncontrolled emissions from soil vapor extraction systems and resulting ambient air
concentrations. The document references several public domain and commercially available
emission models. In addition, the document is accompanied by a PC computer diskette
containing a screening emission model for multiple contaminants. This model is written in the
form of a spreadsheet.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pathway Analysis
Air Pollution
Superfund
Air Pathway Analysis
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (ReT. 4-77) PREVIOUS EDITION IS OBSOLETE
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
f
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g.. date of issue, date of
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doc. etc.). List author's affiliation if it differs from the performing organi-
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Insert appropriate code.
15. SUPPLEMENTARY NpTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with. Translation of. Presented at conference of.
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authori/.cd terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI HELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Held/Group assignments that will follow
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." C'ite any availability to
the public, with address and price.
19. &20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-] (Rev. 4-77) (Reverse)
------- |