United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-90-005
March 1990
Air
&EPA
Guidance On Applying
The Data Quality
Objectives Process For
Ambient Air Monitoring
Around Superfund Sites
(Stage III)
-------
EPA-450/4-90-005
GUIDANCE ON APPLYING THE
DATA QUALITY OBJECTIVES PROCESS
FOR AMBIENT AIR MONITORING
SUPERFUND SITES
(STAGE III)
\ Y, __ ฃ>y
( ~J , nr 'search Triangle Institute
Triangle Park, NC 27709
EPA Contract No. 68-02-4550
EPA Work Assignment Officer: Jane Leonard
Technical Support Division
EPA Project Officer: Darryl von Lehmden
Atmospheric Research And Exposure Assessment Laboratory
Office Of Air Quality Planning And Standards
Office Of Air And Radiation
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
March 1990 y,S, Envlrohme--
Region b,!
77 West .--'- '
Chicago, SL '-.'-
-------
This report has been reviewed by the Office Of Air Quality Planning And Standards, U. S.
Environmental Protection Agency, and has been approved for publication as received from the
contractor. Approval does not signify that the contents necessarily reflect the views and policies of the
Agency, neither does mention of trade names or commercial products constitute endorsement or
recommendation for use.
EPA450/4-90-005
11
-------
ACKNOWLEDGEMENTS
This document was prepared for the U. S. Environmental Protection Agency
by the Research Triangle Institute, Research Triangle Park, North Carolina.
The contributing authors were Franklin Smith, Cynthia Salmons, Michael Messner
and Richard Shores.
We hereby acknowledge our appreciation for the extensive reviews and
assistance provided by the following individuals:
Stanley Sleva, OAQPS
Thomas Pritchett, OSWER
Peter Kahn, Region I
Joseph Padgett, OAQPS
11 i
-------
PREFACE
This document, Guidance for Applying the Data Quality Objectives Process
for Ambient Air Monitoring Around Superfund Sites (Stage III), along with a
previous document (EPA-450/4-89-015, August 1989) covering Stages I and II,
provides an illustration of how the data quality objectives (DQO) process is
used to design an ambient air monitoring system that will be adequate for the
intended use of the data. The two documents combined are intended to serve as
a bridge between the Quality Assurance Management Staff's (QAMS1) DQO guidance
and an actual application of the DQO process at a Superfund site. This docu-
ment illustrates the process of developing a monitoring system for volatile
organic compounds in the ambient air to the point of submitting the design to
the decision maker for approval. If the decision maker approves the design
then a complete DQO document integrating a sampling and analysis plan, a
quality assurance project plan, and a work plan can be prepared. Specifi-
cally, these documents were written to aid the Remedial Project Managers, On-
site Coordinators, Enforcement Project Managers, and the EPA Regional and
Superfund contractor personnel responsible for ambient air sampling and analy-
sis at Superfund sites to carry out their jobs in an efficient and effective
manner.
The DQO process as outlined by QAMS consists of three stages with several
steps within each stage. In Stage I the decision maker (or ultimate data
user) takes the lead role in stating the problem and defining the decision.
The program and technical staff lead in establishing qualitative and quantita-
tive constraints in Stage II. These first two stages of the process result in
proposed DQOs with accompanying specifications and/or constraints. These DQOs
and specifications are passed to the technical staff to complete Stage III;
that is, to design a monitoring system which ensures that the DQOs will be met
in an economical manner.
iv
-------
TABLE OF CONTENTS
Section Page
Preface i v
1.0 INTRODUCTION 1
1.1 PURPOSE AND SCOPE 1
1.2 SUMMARY OF DQOs AND CONSTRAINTS 1
1.3 TECHNICAL APPROACH 3
1.4 USE AND ORGANIZATION OF THIS DOCUMENT 4
2.0 SITE DESCRIPTION 6
2.1 SITE DESCRIPTION AND RECEPTOR LOCATIONS 6
2.2 METEOROLOGICAL CONDITIONS 8
2.3 WASTE CHARACTERIZATION 11
3.0 MODELING VOC CONCENTRATIONS 16
3.1 NEED FOR MODELING DATA 16
3.2 APPROACH FOR MODELING 16
3.3 MODELED EMISSION RATES 17
3.4 MODELED COMPOUND CONCENTRATIONS 18
3.5 CONCLUSIONS BASED ON MODELING RESULTS 22
4.0 SELECTION OF MONITORING INSTRUMENTS 24
4.1 INTRODUCTION 24
4.2 SCREENING STRATEGY 25
4.2.1 Operational and Performance Requirements 25
4.2.2 Overview of Commercially Available
Instruments 26
4.2.3 Capabilities of Selected Instrument 26
4.2.4 Purchase and Maintenance Costs 28
4.3 REFINED SCREENING STRATEGY 28
4.3.1 Operational and Performance Requirements 28
4.3.2 Overview of Commercially Available
Instruments 29
4.3.3 Capabilities of Selected Instrument 29
4.3.4 Purchase and Maintenance Costs 31
4.4 QUANTITATIVE ASSESSMENT STRATEGY 31
4.4.1 Operational and Performance Requirements 31
4.4.2 Overview of Commercially Available
Instruments 32
4.4.3 Capabilities of Selected System 33
4.4.4 Purchase, Maintenance, and Analysis
Costs 34
5.0 MONITORING SYSTEM DESIGN 35
5.1 GENERAL 35
5.2 SCREENING STRATEGY 37
5.2.1 Deployment of the Total VOC-PID Monitor 37
5.2.2 Actions Based on Screening Results 37
5.2.3 Estimated Operating Manpower and Costs 38
-------
TABLE OF CONTENTS
(continued)
5.3 REFINED SCREENING STRATEGY 38
5.3.1 Deployment of the Portable GC-AID
and the Total VOC-PID 39
5.3.2 Actions Based on Refined Screening Results 39
5.3.3 Estimated Operating Manpower and Costs 39
5.4 QUANTITATIVE ASSESSMENT STRATEGY 40
5.4.1 Deployment of Sampler 41
5.4.2 Actions Based on Quantitative Assessment
Resul ts 41
5.4.3 Estimated Operating Manpower and Costs 41
5.5 CRITERIA FOR EMPLOYING A MONITORING STRATEGY 42
5.5.1 Screening Strategy 43
5.5.2 Refined Screening Strategy 43
5.5.3 Quantitative Assessment Strategy 43
5.6 CRITERIA FOR APPLYING EMISSION CONTROL ACTIONS 45
5.6.1 Screeni ng Strategy 45
5.6.2 Refined Screening Strategy 45
5.6.3 Quantitative Assessment Strategy 46
5.7 ESTIMATED TOTAL LABOR AND COST 46
6.0 ERROR ANALYSIS FOR COMPLIANCE WITH DQOs 49
7.0 REFERENCES 55
vl
-------
LIST OF FIGURES
Figure Page
1 HYPOTHETICAL SUPERFUND SITE WITH LOCATION OF BURIED
DRUMS 7
2 WIND ROSE OF WIND SPEED AND WIND DIRECTION DATA 10
3 TEMPERATURE PROFILE OF THE HYPOTHETICAL SUPERFUND
SITE 12
4 AVERAGE EMISSION RATES OVER TIME FOR AN OPEN SPILL (POOL) OF
WASTE CONTAINING VOLATILE COMPOUNDS AT CONCENTRATIONS OF
50 PERCENT EACH 19
5 AVERAGE EMISSION RATES FOR NEWLY UNEARTHED SOIL CONTAMINATED
NEAR SATURATION WITH WASTE CONTAINING VOLATILE COMPOUNDS
AT CONCENTRATIONS OF 50 PERCENT EACH 20
6 ILLUSTRATION OF THE PROBABILITY OF A FALSE POSITIVE ERROR WHEN
THE TRUE CONCENTRATION IS 50 PPB AND THE CONCENTRATION OF
CONCERN IS 100 PPB 51
7 ILLUSTRATION OF THE PROBABILITY OF A FALSE NEGATIVE ERROR WHEN
THE TRUE CONCENTRATION IS 150 PPB AND THE CONCENTRATION OF
CONCERN IS 100 PPB 52
8 ILLUSTRATION OF THE PROBABILITY OF A FALSE NEGATIVE ERROR WHEN
THE TRUE CONCENTRATION IS 200 PPB AND THE CONCENTRATION OF
CONCERN IS 100 PPB 54
LIST OF TABLES
Table Page
1 TWO-WAY FREQUENCY DISTRIBUTION OF WIND SPEED
AND DIRECTION 9
2 CHARACTERISTICS OF COMPOUNDS AT THE SITE 13
3 EXPOSURE LIMITS FOR SELECTED INDICATOR COMPOUNDS 15
4 MODELED 8-HOUR AVERAGE EMISSION RATES 17
5 MODELED 8-HOUR AVERAGE CONCENTRATIONS AT RECEPTOR 1 22
6 COMPARISON OF TOTAL VOC DETECTION TECHNIQUES 27
7 COMPARISON OF GAS CHROMATOGRAPH DETECTION TECHNIQUES 30
VII
-------
SECTION 1.0
INTRODUCTION
1.1 PURPOSE AND SCOPE
This document reflects the nature of the sampling and analytical costs
versus usable information decisions that must be made when initially develop-
ing a draft of a proposed air monitoring plan for subsequent approval by the
ultimate decision maker, the remedial project manager (RPM) and/or the on-site
coordinator (OSC). It addresses Stage III of the DQO process as a continua-
tion of the DQOs and constraints developed in Stages I and II, the results of
which were published as a separate report (!_). In the DQO process, DQOs with
quantitative and qualitative constraints are developed by the decision maker
in consultation with the program and technical staffs. The results of Stages
I and II are passed to the technical staff to design a monitoring system that
will ensure that the DQOs will be satisfied under the constraints of manpower,
instrumentation, facilities, and funds. The purpose of this document is only
to illustrate the implementation of DQOs developed in Stages I and II. This
involves the design of a data collection and analysis plan that will satisfy
the DQOs as stated in Stages I and II. To complete the DQO process this data
collection and analysis (or monitoring) design would be submitted to the deci-
sion maker for approval after which a detailed sampling and analysis plan
including elements of a quality assurance project plan and a work plan could
be developed.
This application is limited to monitoring volatile organic compounds in
the ambient air around a hypothetical Superfund site. In a real application,
consideration would be given to gaseous inorganic compounds and airborne
particulate matter.
1.2 SUMMARY OF DQOs AND CONSTRAINTS
It is desirable that a user of this document have ready access to the
preceding document that covers Stages I and II of the DQO process. For con-
venience of those readers who do not have the preceding document, however, it
-------
Is pointed out that many topics such as site description, waste characteris-
tics, and available instrumentation are discussed in more detail in this docu-
ment. Also, key points from Stages I and II such as the statement of the
decision, actions to consider, ranking of errors by their seriousness, and the
DQOs (presented as acceptable levels of false positives and false negatives)
are restated here.
statement of decision
The decision to be made is whether or not action must be taken to
protect the public from VOCs in ambient air during remedial response
activities.
actions to consider when a problem is identified
Any of the following actions may be considered singularly or in
combination with other actions:
- institute controls to lower air emissions
- halt remedial activity
- evacuate the receptor population which is at risk
- employ a more rigorous monitoring strategy
ranking of decision errors by their seriousness
- concluding there is a problem when the true concentration is well
below the level which poses a health risk is a false positive error
and the least serious.
- concluding there is no problem when the true concentration is in
the range associated with a health risk is a false negative error
and is serious.
- concluding there is no problem when the true concentration is in
the range associated with immediate acute health problems is a false
negative error and very serious.
desired DQOs expressed as acceptable probabilities of false positive
and false negative error rates at selected concentration levels
The following statements of DQOs (or desired levels of performance)
are given first as a general statement, then followed by a specific
statement for benzene as an example. The permissible exposure limit
(PEL) for benzene expressed as an 8-hour average is 1000 ppb (1 ppm).
The level of concern specified by the decision maker for this Super-
fund site is 0.1 PEL, or 100 ppb.
-------
false positive
- At a true average concentration of 1/2 the level of concern (1/10
PEL), the probability of a positive finding (that Is, concluding
that there 1s a problem) should be limited to 10% or less (the least
serious error).
For example, 1f the true benzene concentration Is 50 ppb for an 8-
hour average, there should be no more than a 10% chance of obtaining
a measurement of 100 ppb or more.
false negatives
- When the true average concentration Is 1.5 times the level of con-
cern, the probability of a negative finding (that is, concluding
that there is no problem) should be limited to 5% or less (the
serious error).
For example, if the true benzene concentration is 150 ppb for an 8-
hour average, there should be no more than a 5% chance of obtaining
a measurement of 100 ppb or less.
- When the true average concentration is 2 times either level of con-
cern, the probability of a negative finding should be limited to 1%
or less (the very serious error).
For example, if the true benzene concentration is 200 ppb for an 8-
hour average, there should be no more than 1% chance of obtaining a
measurement of 100 ppb or less.
1.3 TECHNICAL APPROACH
This subsection presents an overview of the approach taken in designing a
monitoring system for volatile organic compounds (VOCs) in ambient air for
this Superfund site. While an effort was made to ensure reasonableness of the
monitoring system design developed herein it is emphasized that the design
only serves to illustrate the application of Stage III of the DQO process to
this problem. Each phase of this approach is detailed in the sections that
follow. First, to design an effective and efficient monitoring system, it was
necessary to identify the compounds of interest and estimate their potential
concentration levels at the point(s) of measurement. Because there were no
applicable monitoring data on VOC concentrations in the ambient air during
site remedial activity, site records were used to determine the RCRA waste
codes represented in the buried drums. Each waste code description was used
to identify individual VOCs in the waste. These VOCs were ranked according to
-------
their vapor pressure, Henry's Law constant, and air toxicity (from Appendix A
of the Superfund Public Health Evaluation Manual). Six of the most volatile
and most toxic compounds were selected as indicator compounds.
Expected concentrations of these six compounds, relating to the distance
from the work site, were generated using models for an "extreme-case" scenario
and again for a more practical scenario. The extreme-case scenario assumed
that several drums of waste were spilled, resulting in a pool of waste that
was 7 meters by 7 meters and 1 centimeter deep; that the compound being model-
ed accounted for 50% of the waste (i.e., the concentration of the subject
compound in the waste was 50%); and that the meteorological conditions were a
temperature of 30ฐC and a wind speed of 12 m/s. For these conditions, an
emissions model was used to provide estimates of 8-hour average emission rates
for each VOC. These modeled 8-hour emission rates were used as inputs for a
Gaussian diffusion model and the resulting 8-hour average concentrations at
the distance of the closest receptor site were estimated.
This same procedure was carried out for a situation involving a volume of
soil saturated with the liquid waste. As expected, the modeled 8-hour emis-
sion rates from saturated soil and resulting estimated ambient air concentra-
tions were much less than those from the pool of waste. These modeling
results were used as a guide for designing the monitoring system.
The objectives of this monitoring system are 1) to provide on-site (that
is, near the work site), real-time or near real-time data of sufficient quali-
ty to allow for emission control actions to be implemented during the 8-hour
work day (this would ensure that concentration levels of concern are not
exceeded at receptor sites); and 2) to provide off-site measurements of 8-hour
average concentrations at receptor sites to assure the decision maker that the
DQOs were met for the subject sampling period and to provide real-time concen-
trations at receptor sites to allow timely application of emission control
actions if there is an indication that VOCs are reaching the receptor site.
1.4 USE AND ORGANIZATION OF THIS DOCUMENT
This document attempts to present the thought process that the technical
staff would follow in designing a monitoring system. To complete the DQO
process the design would be submitted to the decision maker for approval after
which the sampling and analysis plan including the quality assurance project
-------
plan and work plan could be completed. The monitoring strategies, the deci-
sion criteria for initiating emissions control actions, and the criteria for
deciding when to change strategies were based only on the information present-
ed for this hypothetical site. Thus, these monitoring strategies and decision
criteria are not to be assumed applicable to other sites without first going
through the thought process using relevant site-specific information.
This document serves as an aid to the Remedial Project Managers, Enforce-
ment Project Managers, and EPA Regional and Superfund contractor personnel
responsible for ambient air sampling and analysis at Superfund sites. Its
intent is to help these staff develop and apply site-specific DQOs for ambient
air monitoring requirements at other Superfund sites.
It would be helpful if the reader is familiar with the following:
Guidance on Applying the Data Quality Objectives Process for Ambient
Air Monitoring around Superfund Sites (Stages I and II) (1)
The QAMS guidance: Development of Data Quality Objectives,
Description of Stages I and II (2)
the four volumes of Procedures for Conducting Air Pathway Analyses for
Superfund Applications (3.4,5,6)
the two volumes by the Office of Solid Waste and Emergency Response:
Data Quality Objectives for Remedial Response Activities (7,8)
the two DQO papers prepared by OAQPS: Data Quality Objectives for the
Toxic Air Monitoring Systems and Data Quality Objectives for the Urban
Air Toxic Monitoring Program (9,101
The remainder of the document is organized as follows. Section 2.0 gives
a physical description of the Superfund site, its meteorological profile, and
its waste characteristics. Since there were no VOC monitoring data available
when the site was being disturbed during remedial activity, it was necessary
to rely on emission and diffusion models to obtain some idea of the concentra-
tion levels to expect at the receptor sites for different scenarios at the
work site. Section 3.0 discusses modeling results. Section 4.0 provides the
process for selecting VOC instrumentation for this site. Section 5.0 presents
and discusses the monitoring system design, including monitoring strategies
and decision criteria. Section 6.0 gives an error analysis which indicates
that the proposed monitoring system will satisfy the DQOs in terms of data
quality.
-------
SECTION 2.0
SITE DESCRIPTION
This section presents a more detailed description of the hypothetical
Superfund site than was given 1n the document covering Stages I and II. The
additional details Include distances from the perimeter of the waste site to
the different receptor sites, a meteorological profile for the site, and a
discussion of waste characterization. These details were required as inputs
to models for generating potential VOC concentrations downwind of the work
site during remediation.
2.1 SITE DESCRIPTION AND RECEPTOR LOCATIONS
This hypothetical Superfund site, located on the outskirts of an urban
area, consists of approximately 180 acres. The fenced-in site property is
mostly flat, covered by weeds and brushes 1 to 2 meters in height. Also,
there is a continuous row of both deciduous and nondeciduous trees ranging
from 18 to 23 meters in height, just inside the boundary fence. Figure 1 is a
scaled map of the site.
Site records indicate that approximately 20,000 drums containing waste
from local Industries were buried here from 1959 until 1980. Most of the
drums were buried in a single layer at a depth of about 2 meters. The 3-acre
area marked "A" in Figure 1 designates where the drums were buried. The drums
contain mainly spent solvents representing RCRA waste codes F001 through F006.
Some of the drums have leaked and contamination of soil and groundwater
is possible. This potential problem led to a Remedial Investigation/Feasi-
bility Study (RI/FS) for the site. The remedial action selected following the
RI/FS was to dig up the buried drums, pack them into larger drums labeled as
containing hazardous waste, and transport them to a nearby hazardous waste
facility. This remediation effort is expected to take 12 months of on-site
activity based on an 8-hour, 5-day-a-week work schedule.
During the RI/FS, drums were disturbed and an unusual odor was detected
at the house identified as receptor Rl in Figure 1. A measurement made with a
total VOC instrument indicated that VOCs had been released into the air;
therefore, monitoring will be necessary to detect any subsequent releases to
the ambient air which would pose a threat to public health during remediation.
-------
Trees and brush,
combination of
spruce, pine, oak,
and maple, 18-23 m
785m-
ซ OR1
30 mT
-H
/Site Boundary
Grave/ road, fugitive
emissions control
with water spraying
Open area
-1-2 m weeds,
brush, and grass.
Bush-hogged annually.
Designated route
for excavation work
and site visitors
1>R4 t}R5
ฐ*iPLocal
agency air
monitoring
station
Elevation at 470 m;
grade is mostly flat ฑ 6m
0 60 120 180 240
1 cm = 60 m
ฎ Canister Sampling Site
N
Figure 1. Hypothetical Superfund site with location of buried drums.
-------
The guidance given 1n Procedures for Conducting Air Pathway Analyses for
Superfund Applications, Volumes I and II will be followed in designing the
ambient air quality monitoring system for this site.
The domain of interest consists of six houses located within 184 to 714
meters of the work site's perimeter (see Figure 1). The following information
shows the bearing and distance to those receptors:
Distance from
Receptor Bearing (Degrees) Boundary (Meters)
Rl 358 184
R2 100 289
R3 93 390
R4 133 630
R5 128 660
R6 208 714
2.2 METEOROLOGICAL CONDITIONS
A State-operated air monitoring station has been operating for a number
of years outside the fenceline of this site (see Figure 1). Its location
adjacent to the site is coincidental, but historical meteorological data such
as wind speed, wind direction, rainfall, and temperature will be useful in
designing the monitoring network.
Because there are no applicable air monitoring data available, historical
meteorological data are needed for two purposes: (1) to use in an emission
rate model to assess the potential emissions, and (2) to model exposure risks,
in terms of concentration levels, that may be experienced at receptor sites
during remediation. Because the emission rates of volatile organic compounds
are dependent upon temperature and wind speed, a knowledge of meteorological
data that are representative of extreme cases is important. Table 1 repre-
sents the two-way distribution of wind speed and direction. Figure 2 is a
graphical representation (a wind rose) of these wind speed and wind direction
data. These historical meteorological data were generated at the State-
operated air monitoring station. Table 1 lists the number of measured wind
speed occurrences for each wind speed range and for each direction. For
example, there were six occurrences of winds out of the south (S) at speeds
between 3.1 and 4.4 meters per second (m/s). For modeling purposes, the data
were reviewed to obtain a wind speed value that would represent an extreme-
-------
TABLE 1. TWO-WAY FREQUENCY DISTRIBUTION
OF WIND SPEED AND DIRECTION
Speed
(m/s)
1.8 -
3.1 -
4.5 -
5.9 -
7.2 -
8.5 -
9.9 -
11.2 -
12.6 -
13.9 -
15.3 -
16.6 -
17.9 -
Direction
3.0
4.4
5.8
7.1
8.4
9.8
11.1
12.5
13.8
15.2
16.5
17.8
19.2
S
6
11
11
5
1
1
SW
1
16
16
21
8
8
2
2
2
W
2
2
5
10
9
6
5
4
4
2
1
NW
8
12
16
8
5
1
1
N NE
13
8 4
7 14
1 22
37
26
11
14
4
5
1
E
1
17
15
6
5
1
2
SE
2
7
2
5
1
2
I
4
64
78
87
63
58
37
17
22
6
3
5
1
35 76 50 51 29 138 47 19 445
-------
NW
NE
Legend
Strong
Moderate
I Weak
10% occurrence
20% occurrence
SE
Figure 2. Wind rose of wind speed and wind direction data.
10
-------
case situation. Numbers in Table 1 under the column headed "E" show that of a
total of 445 occurrences recorded, 408 (slightly over 90% of the occurrences)
were for wind speeds of 12.5 m/s or less. Thus, a value of 12 m/s was used to
represent an extreme case for modeling purposes. Figure 3 shows that site
temperatures of 30ฐC or greater are experienced only in the summer months,
with the highest temperature of 37ฐC occurring in July. For modeling, a value
of 30ฐC was used to represent an extreme case.
2.3 WASTE CHARACTERIZATION
As previously stated, there are no compound-specific air monitoring data
available for this site. Therefore, to gain an understanding of which VOCs
may be released to the atmosphere during remediation, it was necessary to
determine the composition of the buried waste. The first step was to review
the site records, which showed that the site contained waste codes F001
through F006. A list of the VOCs contained within each waste code was obtain-
ed from 40 CFR, Part 261 (subpart D, paragraph 261.31 of Table 1). These RCRA
waste codes and VOCs are listed in the first two columns of Table 2. As
shown, there are 30 VOCs listed for waste codes F001 through F005. Waste code
F006 contains inorganics and does not enter into this illustration.
For those VOCs in Table 2, Appendix A of the Superfund Public Health
Evaluation Manual was used to obtain each compound's vapor pressure, Henry's
Law constant, air toxicity constant for carcinogens, and air toxicity constant
for noncarcinogens (columns 4 through 7, respectively). These data were used
for two purposes. First, a compound's volatility is a function of its vapor
pressure and Henry's Law constant. Therefore, both vapor pressure and Henry's
Law constant were used as inputs to the model for determining emission rates.
Second, data from these columns (4 through 7) were used to rank the compounds
according to their volatility (vapor pressure or Henry's Law constant) and
their toxicity (carcinogen or noncarcinogen) to select six as indicator
compounds. Ideally, these six indicator compounds would be both highly toxic
and highly volatile. For example, benzene was one of the compounds selected.
Table 2 shows that benzene had the 5th highest vapor pressure, the 8th highest
Henry's Law constant, and the 5th highest air toxicity constant for noncar-
cinogens. It is obvious from the table that there are a lot of missing data;
thus, in each individual case, professional judgement should be used to add or
11
-------
0
01
o
*
&
I
40
30
20
10
0-
-10-
-* Highest Temp.
a- Lowest Temp.
Average Temp
-20
JAN FEB MAR APR MAY JUN JUL AUS SEP OCT NOV DEC
Month
Figure 3. Temperature profile of the hypothetical Superfund site.
12
-------
UJ
X
O
u.
o
o
a:
UJ
i
co
<:
w>
Z
0
Or- 3
oe oc , -,
< Z *""
Httl Z
o u *-i
O
U
C
I- Of-
>H 00
ซJH C^-,
t-IZ~ 0
X< U E
Oh- t^
r- V) ซo
Z U E
ceo i ป*
MซJ C
< 0
c
***
^
> ^
r- CO
UJ
X >*
UJf)
ceซ-^
eel? 01
O V) X
0. (A
H? UJ E
A N^
oe
UJ
CO
3
3
Z
to
HC
u>
O
z
o
3
O
ui
o
0
u
^
oc
u
oe
10 tH CM
a a a
CO S Ok S
o u. u^
CM fH CO a fH IH
a a co aaa a
i ป a i*i i i i i
u> oo i ซ ปป- ok i u> ป> a m a oป a
^ a a to CM oป fH 10 co a
iii iii i i i i
r- o> u> tH o>
CM f* r* F u) a r"ป Oi h- fป
fH fH
c
c
C ฃ. C
T> -a jc m
c x 4> x
-O 4ป I- O 4>
c o i i- c 6
x <- o o o o o
4> X O U O COO
XXX C * f) 3
o 4> 4ป ux i. ปx ja x
tU U 4) M U ซซ.I Ot-t.
Jit O U 9 V"" m
a a
i i i i i i i i i
1 ซO 1 CO 1 1 1 | 1
i a i * i i i t i
CM ซ
IH CM a
aa a
a a i a
Sf- 1 S 1 1 1 1 1
fH CM r-
fH fO H/ fH <) fH
III II 1
T 09 fH a CO CO
1 ซD r- T 1 rH CO 1 UJ
1 1 1 1 1 1 1 1 1
1 r- fH a 1 09 fH II-
ซ ป a a r>. CD
fH fH rH
f)
c
0
0
JC
X O
4> C 3 O C
v. ja u o
4> M 0 C
ex ซ
0 4ป X
C XX C
CO X4ป O
*ป X X XX 3 X
^ X X X 4) Cfi U 4)
XU4>4>4>ซ 1 X
x
^4
a
ซ i
r- i
0ป 1
u>
i i
i i
I i
CO
a
i i
a i
tH 1
fH
fH fH
aa
i i
aa
via
CMfH
CO Cf)
1 1
r-uป
a>oo
fHOI
n
fH
^
.^
U
u
^B
X
,)
i.
u e
^
4 N
C
Jป
O O
t.
4k
u
^
a
a
u.
u>
C4
*
CM CM a CM
aaa a
i i ป i i ป i i
a u> v i i oo i i
CMr-CM 1 1 fH 1 1
U>F- ff fH
CM
a
i i i i i i i i
1 t 1 1 1 fH 1 1
1 1 1 1 1 f- 1 1
|^
COIACM <>
aaa a
i i i i i i i i
r>-T ซ i i cป i i
or*. CM i i u> i i
CO CMfH U>
fH fH CM fH fH
aaa aa
IHU> a i a M i i
OBfsCO 1 BU> 1 1
CMr-ซ CMC*
mraa iHfHCM
i i i i i i
09COU> (0 O CO
CO Ot IH 00 00 T 1 1
1 1 1 1 1 1 1 1
co CD 10 oo a fH i i
fH fH
0
C
o
o o
ป c
X
x 3 a a.
x o
4> 0 C,
a c x a.
c e 4> cot,
XO 3-D ซX4>
4kt.OC.CUJZ
O Ck X 1 1
H a iปi a. ca CM CM
10
a
a
u.
^ T9 "U ^ "O
fH a a a CP
a a a s
fH ซH U> f-
CM fH Cf
aa a
* * i * i
Ok s I r- t
iO tO 1 LO 1
OCM fH
fH CM a
aa a
ป ป i * i
UJ fH 1 U> 1
CO fH 1 00 1
fHfH CM
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
.J
L,
a a aa>
oป ซ co a u>
i i i i i
eo ป>-ปซ. CM CM
f ป* afH
i i i i i
a a a a r-
Kt- r>. ^
f- t- t- r-
~m
o
41
CO 1
*E E
"i"i"c>~
E o Ojf c
^3 Cป t. U
mjc JS x
cj cj o z cj
0
a
a
u.
t.
CO
a
CO
CO
fH Ck
a
V X 01
CO 4> O
^s C
OL u
UJ XI.
o
H 0
o
c
3 ป
C U 4> t.
O O
2 f-Z*-
c a m
O 4>
U 4> C
4> - 0
C 4>
oปซป- c
4> OUJ O
> c u
TJ UJ - 0
9 x >> e 4>
C 4>
U) 0ป U
o
ck - e x
X 0
E " o 4>
o ป u
C X
U 4>
4> J3 tl c
3 O O *
W O. X Z
Q T3 I. t.
C COOL,
1 * Q.
i c e
1 f> 4> O
i a f) u
3 30,
e x xja
O Ck 4> 4>
4> UJ X X
O O O TJ
C ^ H H
O 3 4>
U - - 0 C
fH 01 tO X
ป 1 1 1 Ck U
4ป ป
c 4> 4> 4> c a
ja jo. ja Ok4>
u ~ o o
XX X C C
XXX *~
C IUUJUJ UTJ
t. c
H; HI HJ u
L -.
x x x t. E
0 0.
M ^3 ^0 ^3 *,yป Q,
c c c
0.0. Q.*> l_
CL a CL c
4> II
E E E II 4>
o o o c
ttl L, t. L. O C
H U.U.U. 03
o
Z J3 U T3
13
-------
remove compounds from the list of indicators as monitoring data become avail-
able. The six indicator compounds selected for this site are given in Table
3, along with permissible exposure limits (PELs) representing 8-hour averages
and short-term exposure limits (STELs) representing 15-minute averages.
Modeling for potential VOC concentrations at receptor sites will be limited to
these six compounds.
An additional piece of information needed in order to model emission
rates is an estimate of the VOC concentration in the waste; for example, what
percent of the waste is benzene. The estimated total VOC concentrations in
the waste drums (except F004) were taken from results of a random sample of
responses to the National Survey of Hazardous Waste Generators (conducted for
the EPA by Research Triangle Institute, 1987). Due to the lack of reported
generator data for F004, characterizations for this waste code found in the
RCRA Risk-Cost Analysis Model, Appendix A; Waste Stream Data Base (prepared
for the EPA Office of Solid Waste, by ICF Corporation, March 1984) were used
instead. The last column in Table 2 contains the estimated total VOC concen-
trations by waste code. These estimates range from 42.5% (F005) to 90%
(F002). Based on these data, a concentration of 50% for an individual com-
pound was used as an extreme-case for modeling. Note again that this exercise
was necessary only because there were no applicable monitoring data available.
14
-------
TABLE 3. EXPOSURE LIMITS FOR SELECTED INDICATOR COMPOUNDS
Compound
Permissible Exposure
Limit (PEL)
Short-Term
Exposure Limit (STEL)
Benzene
1 ppm for 8-hour time-weighted 5 ppm for 15 minutes3
average
Trichloroethylene 50 ppm for 8-hour time-weighted 200 ppm for 15 minutes*5
average
Carbon
tetrachloride
2 ppm for 8-hour time-weighted
average
1,1,2-
Trichloroethane 10 ppm for 8-hour time-weighted
average
Carbon disulfide 4 ppm for 8-hour time-weighted 12 ppm for 15 minutesb
average
Methyl ethyl ketone 200 ppm for 8-hour time-
weighted average
300 ppm for 15 minutes0
ppm - parts per million
a 29 Code of Federal Regulations 1910.1028, Subpart Z.
b Federal Register, 54, Thursday, January 19, 1989, pages 2332-2983.
indicates no short-term exposure limit.
15
-------
SECTION 3.0
MODELING VOC CONCENTRATIONS
3.1 NEED FOR MODELING DATA
There are no ambient air concentrations available for this site to repre-
sent the remedial activities. For this reason, models were used for the
single purpose of estimating potential risk to residents of the nearby houses.
Specifically, what are judged to be the two extreme-case scenarios that could
occur during remediation were modeled to provide estimates of the 8-hour aver-
age concentrations that would occur at the receptor closest to the work site.
These scenarios were provided by the technical staff responsible for designing
the monitoring network.
This modeling required the efforts of experienced modelers with necessary
information on compound characteristics, meteorological conditions, and the
scenarios. These models are reasonably complex, hence a detailed discussion
is not attempted in this document. References to documents containing detail-
ed information on the models are provided, however.
3.2 APPROACH FOR MODELING
Modeling was carried out in two steps. First, emission rates for the six
indicator compounds were modeled for two extreme-case scenarios. These model-
ed emission rates were then used as inputs to a diffusion model to predict
compound concentrations in the ambient air at the nearest receptor site.
The two scenarios selected to approximate extreme-case conditions follow:
A waste spill emptying two 55-gallon drums, resulting in a liquid pool
7 meters by 7 meters and 1 centimeter deep. Fifty percent of the
liquid waste was assumed to consist of the compound being modeled (for
example, when modeling benzene, 1t was assumed that 50% of the waste
was benzene). Meteorological conditions used were a wind speed of 12
m/s and a temperature of 30*C.
Uncovering or working in an area 7 meters by 7 meters of soil satur-
ated with liquid waste. As above, the compound being modeled was
assumed to account for 50% of the liquid waste. A wind speed of 12
m/s and a temperature of 30ฐC were used for modeling.
16
-------
3.3 MODELED EMISSION RATES
The modelers followed the guidance provided in the EPA document,
Hazardous Waste Treatment, Storage, and Disposal Facilities (TSDF) - Air
Emission Models (11). Inputs to the model Included the compound's vapor pres-
sure and Henry's Law constant (both obtained from Table 2, Section 2.0); wind
speed; wind direction; and temperature, all taken to approximate extreme-case
conditions from Section 2.0. The final equation giving the average emission
rate for the compound follows:
Q = fair V C0/t
where Q = Average emission rate over time period t, grams per
second (g/s).
fair = The fraction of the compound originally in the waste,
emitted to the air during time t.
Note: = fai-r was modeled by a complex exponential function
requiring input values on waste and compound
characteristics and atmospheric conditions.
C0 = Initial concentration of the individual VOC in the
liquid waste, g/m^
V = Volume of liquid waste in the pool, m^
The 8-hour emission rates in grams per second (g/s) for the six indicator
compounds for the pool and saturated soil are given in Table 4.
TABLE 4. MODELED 8-HOUR AVERAGE EMISSION RATES
Compound
Benzene
Carbon Disulfide
Carbon Tetrachloride
Methyl Ethyl Ketone
Trichloroethylene
1,1, 2-Tri chl oroethane
Emission Rate
for Pool
(g/s)
8
8
8
8
8
8
Emission Rate for
Saturated Soil
(g/s)
1.1
0.7
0.7
0.7
0.5
0.3
17
-------
Modeling results for the waste spill pool were plotted to provide pro-
files of emission rate as a function of time over an 8-hour period for each of
the six Indicator compounds in Figure 4. Moving from left to right, the plot-
ted points for emission rates are a 15-minute average, a 1-hour average, a 4-
hour average, and, at the far right of the figure, an 8-hour average. For
example, the figure shows that carbon disulfide had the highest average emis-
sion rate of all six indicator compounds for the first 15 minutes after the
spill. (This was expected since carbon disulfide has the highest vapor pres-
sure of the indicator compounds.) The 1-hour average emission rate for carbon
disulfide was approximately 70 g/s, or about 1/4 of the 15 minute average of
275 g/s. This shows that nearly all of the carbon disulfide originally in the
waste was emitted into the atmosphere the first 15 minutes after the spill.
Average emission rates for 4 hours and 8 hours were the same for all six Indi-
cator compounds. Carbon tetrachloride, benzene, and methyl ethyl ketone also
showed high emission rates within minutes after the spill. This rapid emit-
tance of at least four of the six indicator compounds indicates that the site
safety plan should specify level B protective clothing during all drum excava-
tions. Time will not allow the crew to put on protective clothing or the
monitoring staff to start up a monitoring system after a spill has occurred.
This rapid emittance also indicates that, to capture the short-term high com-
pound concentration levels resulting from a spill, the monitoring system must
be set up and collecting a sample over the 8-hour work day. Also, a real-time
or near real-time monitoring capability should be in place at the impacted
receptor site to assess short-term exposure and to allow for timely
application of emission control actions.
Emission rate results from modeling the saturated soil scenario were
plotted similarly in Figure 5. Note that the emission rates from soil have
approximately the same profile over time as emissions from the pool. The
average 8-hour emission rates from soil are less by factors ranging from 7 to
27 than those from the pool, as seen in Table 4.
3.4 MODELED COMPOUND CONCENTRATIONS
A Gaussian dispersion model, based on the work of Turner (12), was used
to estimate 8-hour average concentrations of the six indicator compounds at
the nearest receptor (Rl). Wind rose and temperature data presented in
18
-------
Q>
C
c c.
ฃ O q>
J2 O Z ^ l^J UJ
O f< UJ 2 O O
r O
Q.CO
"ง
8-2
(O O
O CL
*>ป
o
>
4ป C
V "=
O>-^
(O *
S- 4J J=
V C= U
> O 10
< U
-------
s-
ซ0
V
IS
if
ui
i O
ฃ-
O g
10 10
O) r
c o
3 >
-
>> oป
o> c
C f- -M
4-1 4-> O
re co in
s- 10
-------
Section 2.0 suggested a combination of a wind speed of 12 m/s and a
temperature of 30ฐC for an extreme case.
The Gaussian plume equation for a continuously emitting point source with
no effective plume rise, but along the center line of the plume, for concen-
trations at ground level and directly downwind is given by
C= Q/(f <7y ffZ U)
where C 1s the ground-level air concentration, Q is the emission rate, u is
the wind speed, and ay and az are the horizontal and vertical variances of
plume spread, respectively.
In general, stability categories were identified in terms of wind speed,
cloud cover, and the intensity of solar radiation (during the day) at the site
location. Stability class C was used in this analysis in accordance with the
above-mentioned wind speed.
The variances ay and az were calculated using formulas recommended by
Briggs (1^) for urban conditions. Urban conditions were selected because of
the trees between the work site and the receptors (see Figure 2). These
formulas are valid for distances of 100 to 1000 meters from the source to the
receptor.
The simple model described above provides a quick estimate of the ground
concentrations using data for this Superfund site; however, more accurate and
detailed analyses can be provided using higher accuracy models such as those
described in EPA guidelines on air quality models (14).
The meteorological conditions of a site should be reviewed from two per-
spectives. The first is to model the concentration at the receptors of con-
cern (the closest) under an extreme-case scenario (in high wind velocity and
high ambient temperature). The second is to model the concentrations at the
receptors of concern under actual conditions. Modeling is dependent upon the
availability of sufficient meteorological data that are representative of the
site, and upon whether or not the extreme-case modeling has identified concen-
trations of concern at the receptors. Table 5 contains the modeled 8-hour
average concentrations for the six indicator compounds at the distance to the
closest receptor, based on the emission rates from the open pool of waste and
from the soil saturated with waste.
21
-------
TABLE 5. MODELED 8-HOUR AVERAGE CONCENTRATIONS AT RECEPTOR 1
Concentration
from Pool
Compound (ppb)
Benzene 40
Carbon Dlsulfide 40
Carbon Tetrachloride 40
Methyl Ethyl Ketone 40
Trlchloroethylene 40
1,1,2-Trichloroethane 40
Concentration
from
Saturated Soil
(ppb)
2
4
2
2
2
1
Level of Concern
0.1 PEL
(ppb)
100
400
200
20000
5000
1000
The results of the modeling identified benzene as the compound posing the
greatest health risk for residents at receptor Rl. For results from the pool,
the ratio of modeled concentration to the level of concern (0.1 PEL) was
largest for benzene (40/100 = 0.4) and smallest for methyl ethyl ketone
(40/20000 = 0.002). For the saturated soil scenario, the ratio for benzene
was 0.02 (2/100) and for methyl ethyl ketone, 0.00001 (2/20000).
3.5 CONCLUSIONS BASED ON MODELING RESULTS
Based on the modeled 8-hour average concentrations given in Table 5, it
was concluded that emissions from accidental spills (as large as 7 m by 7 m by
1 cm) could pose a risk to residents at the nearest receptor site (Rl).
Therefore, the operating procedure for remediation should be to clean up a
spill immediately. On the other hand, the modeled data also indicated that
emissions from contaminated soil would not pose a risk to anyone outside the
site fenceline. Furthermore, these modeled concentrations from the saturated
soil scenario were low enough to indicate that there should be no problem in
securing the site and terminating monitoring activity at the end of the work
day. Emissions from a secured site should be so low as to not pose a problem,
except maybe, during inversions. Monitoring only during the 8-hour work day
should be sufficient a great majority of the time to protect the residents at
the receptor sites, however, the monitoring system design should include pro-
visions for monitoring during non-work hours or evenings when meteorological
conditions are conducive to inversions.
22
-------
Based on the modeled emission rates given 1n Figures 4 and 5, which show
a very short lag time between a spill and resulting high emission rates (and
thus high concentrations), the site workers should wear protective clothing
while on-s1te. Also, the monitoring system designed to provide 8-hour average
concentrations at the receptor site must be operated over the entire 8-hour
work day so that the 8-hour averages will be measured accurately when acciden-
tal spills occur. In addition, some form of monitoring that provides real-
time or near real-time data should be In operation at the receptor site to
monitor for short-term exposure and to allow for timely application of
emission control actions at the work site.
23
-------
SECTION 4.0
SELECTION OF MONITORING INSTRUMENTS
4.1 INTRODUCTION
The purpose of this section 1s to work through the process of selecting
monitoring Instruments that will fulfill the requirements of the measurement
system design. Personal experience with specific Instruments and sample
collection techniques and unique features of the subject superfund site will
always be factors in the selection process and ultimately in the monitoring
system design. The selections made in this section were based on input from
individuals, with some experience with the methods, as to the capability of
the methods to produce the quality and quantity of data needed to satisfy the
stated DQOs. Since the purpose of this document is to illustrate how the DQO
process can be applied to this site, discussions of technical strengths and
weaknesses of each monitor and sampling technique are minimized so as not to
detract from the process itself. To some extent, the instrument selection
process and the measurement system design effort had to proceed in parallel
since they depend on each other. Details of the measurement system design are
given in Section 5.0. The information considered in selecting monitoring
instruments is presented in the following paragraphs.
The measurement system design consists of three monitoring strategies.
First, modeling data indicated that in the absence of open pools of liquid
waste, the probability of the residents at the receptor sites being exposed to
VOC concentration levels of concern was very small. Thus, a single screening
strategy was chosen as the first element of the monitoring system design.
This strategy was designed to provide the RPM/OSC with real-time information
on total VOC concentrations near and downwind of the work site to allow timely
initiation of emission control procedures when necessary.
The second element selected for the monitoring system design was a refin-
ed screening strategy. This strategy will be employed if the results from the
screening strategy show that a preset criterion for implementing the refined
screening strategy has been exceeded. The objective of the refined screening
strategy is to provide the RPM/OSC with real-time or near real-time concentra-
tions of the six indicator compounds near and downwind of the work site.
24
-------
Since the concentration level of concern Is different for each VOC, a know-
ledge of the Individual VOC concentrations provides the RPM/OSC with more
Information to assess the risk of not Initiating emission control actions than
is available from the screening results. Also, this refined screening strate-
gy provides total VOC concentrations at the impacted receptor site in real-
time to allow for timely initiation of emission control actions and provide
Information on short-term exposure levels.
When on-site measurements from the screening and/or refined screening
strategies exceed preset criteria, a third element of the monitoring system, a
quantitative assessment strategy will be employed to provide the RPM/OSC with
8-hour average concentrations of individual VOCs at the receptor sites for
that work day.
Procedures for selecting monitoring instrumentation for each of the three
monitoring strategies are discussed separately by strategy in the following
subsections. The selection process starts with a description of the perform-
ance and operational requirements of the monitoring instrument. This is
followed by an overview of the commercially available instruments considered
for this application. Finally, the instrument and associated apparatus sel-
ected for this application are discussed.
4.2 SCREENING STRATEGY
The primary function of the monitoring instrument for the screening stra-
tegy is to provide a continuous, real-time indication of VOC concentrations
near and downwind of the work site. There are several commercially available
VOC monitors that provide real-time data for total VOC concentrations.
4.2.1 Operational and Performance Requirements
The operational requirements and performance capabilities of an instru-
ment for this application are as follows:
Operate outside year-round.
Operate continuously over an 8-hour work day without external
electrical power.
Be portable, easily carried by one person.
Respond to all six indicator compounds.
25
-------
Have a detection limit of better than 1 ppm for isobutylene (the
calibration gas).
Be capable of recording the total VOC data over an 8-hour period.
Not respond to interferences common to Superfund sites including
methane, water, carbon dioxide, nitrogen, and oxygen.
Be capable of visual or audible alarms at preset total VOC
concentration levels.
4.2.2 Overview of Commercially Available Instruments
There are a number of manufacturers marketing total VOC monitors. Each
manufacturer usually specializes in a monitor with one of two possible detec-
tors: a flame ionization detector (FID) and a photoionization detector (PID).
Thus, the first decision in the Instrument selection process was to decide on
the appropriate detection technique. Table 6 contains a comparison of the two
techniques including sensitivity, response time, and drift over 7 hours for
the six indicator compounds. As seen in the table, the PID responds to all
six compounds with detection limits below the respective levels of concern
(0.1 PEL). The FID does not respond to carbon disulfide. Also, a total VOC
monitor equipped with a PID does not require a source of hydrogen gas as does
a monitor equipped with an FID. The FID hydrogen-burning system requires more
controls and a more complicated pneumatic system.
Once the detection technique was selected, recommendations were solicited
from individuals with experience using total VOC monitors in the field. One
highly recommended instrument that possessed all the capabilities specified in
subsection 4.2.1 was reviewed and selected for this application.
4.2.3 Capabilities of Selected Instrument
The instrument chosen for this task was a total VOC-PID monitor with the
following capabilities:
Designed to operate outside in the extremes of weather.
Designed to operate continuously throughout the 8-hour work day with
the use of an additional battery.
Is portable, easily carried by one person, and easy to operate and
maintain.
26
-------
V)
Ul
cr
z
Ul
1-
z
o
M
H
0
fH C
> o
fH
ฃ.
U
U
c
X
X
4)
9
o
1.
o
*ป
JJ
u
"t
h-
L.
O
4>
U
o
C
o
ฐ_JJ
0
c-o
o **
^ %
1.
m a
o ti
^
o
m
N
C
0
4)
o
.c
D.
^
c u
0 0
U JC
u
v^ t
1.
*
1-
c
N
C
9
ง* 0
M ซ
N CO * 4>
O
c
o
ปป
.a N
o. .
a. u c
0
19 M
S> K
CO ซ ป
E
ซ
U.
E U
a.
a M
K
T"* CO ^
a
a. u
^
SJ
(B K
M CO ป
^
L.
x I- j:
K
> A ^^
c
*> O -P
a. *
M
C C
a: o
w
E U
a
a. * K
a
M M i-l
E
a. u
a
K
s e
IH CM fH
E u
a.
a. N K
^
W CM t-l
| |
ee. i i
Z i i
E
a u
a.
X
s a
IH M IH
E U
a
a ซ K
s
E /^
X 1-
4ป t- -C
'> r-
ป~ w s>^
4> C
" O 4*
a. o-
c ซ
i-
w at o
c
0
4}
u
4>
9
P
t
O
C
o
a
^
^
0
c
5
i
e
la
E
jC
0
.a
0
4>
0
o
C
0
c a.
0 2
i.
c
o o
a. c
n
II
at
z
27
-------
Responds to all six indicator compounds at concentrations below the
respective levels of concern.
Shows a detection Hm1t of 0.1 ppm for Isobutylene (the calibration
gas).
Incorporates advanced microprocessor technology for real-time digital
or graphic data assessment and built-in data logging capability for
storing data, including concentration with time and location.
Does not respond to methane, water, carbon dioxide, nitrogen, and
oxygen.
Programmed to sound an alarm at predetermined total VOC
concentrations.
4.2.4 Purchase and Maintenance Costs
The cost of the selected VOC monitor was $4500, the supplementary field
kit was $600, and the battery was $300. Spare parts needed for continuous
operation included a lamp ($215), an extra battery ($300), and 5 inlet filters
($25). A calibration source (compressed gas cylinder) of isobutylene cost
$500.
4.3 REFINED SCREENING STRATEGY
The refined screening strategy provides the RPM/OSC with near real-time
concentrations of the six high risk indicator compounds near the work site and
total VOC concentrations at the impacted receptor site. Also, the refined
screening strategy will provide information on the appropriateness of the six
high risk compounds selected as indicators. This strategy will provide tenta-
tive identification of unknown compounds released from the work site, serve to
determine when to implement the quantitative assessment strategy, and provide
guidance on when to submit the quantitative assessment sample for analysis by
GC/MS to identify unknown compounds. This information is used to assess the
seriousness of the emissions from the work site.
4.3.1 Operational and Performance Requirements
The instrument for providing total VOC concentrations at the receptor is
a total VOC-PID. It was selected for the same reasons discussed in the
screening strategy (Section 4.2).
28
-------
The operational requirements and performance capabilities of an Instru-
ment for providing near real-time concentrations of at least the six Indicator
compounds near the work site are as follows:
Operate outside year-round.
Operate continuously over an 8-hour work day without external
electrical power.
Be portable, easily carried by one person.
Respond to all six indicator compounds.
Have a detection limit of better than 1 ppm for benzene.
Be capable of recording the concentrations over an 8-hour work day.
Have a precision, expressed as a relative standard deviation, of 20%
or better for each of the six indicator compounds at their respective
levels of concern, and have a negligible bias.
4.3.2 Overview of Commercially Available Instruments
There are a number of manufacturers marketing portable GCs which are
equipped with one or more detectors. The four available detection techniques
are argon ionizaton detector (AID), photoionization detector (PID), flame
ionization detector (FID), and electron capture detector (ECD). Table 7 pro-
vides sensitivity estimates interpreted as detection limits, analysis cycles,
and precision estimates in reference to the six indicator compounds. As seen
in the table, the PID and the AID respond to all six indicator compounds with
detection limits below the respective levels of concern (0.1 PEL). The AID
and PID provide similar responses for four of the six compounds with the AID
having superior sensitivity for the two remaining compounds. At least one of
the six compounds result in minimal or no response using the FID and ECD
detection techniques.
4.3.3 Capabilities of Selected Instrument
The detection technique chosen for this application is the AID because of
its ability to respond to all six indicators. Once the detection technique
was selected, recommendations were solicited from individuals with experience
using the portable GC-AID in the field. One GC-AID monitor possessing all the
required capabilities specified in subsection 4.3.1 was reviewed and selected
29
-------
(O
UI
3
a-
M
Z
w
UI
z
o
M
t-
X
JC ฅ
*>
2 X
*>
UI
c
A
X
4>
CM
ซ O
rH C.
ซ O
rH
JC
U
L
ป-
X
JC
4>
O
l_
o
JC
o
L
1-
c -o
o
J0ซ-
1.
3
U *
Q
o
C 1.
o o
JO
I. ^
II U
U
1.
4>
1-
C
N
C
ffl
JO
Q.
j& a. ja e
a a
as a E
rH K
S 1 S 0ฃ S S
rH 10 CM Z rH rH
^D
a
JO Q.JO
jo a a c
a as a
as E
S rH rH K
s s s s
CM rH W rH rH
J3
a
a
.a c
(X .O G JO
a as a E
arH a K
s s s
rH rH V CM rH rH
JO
jo a c
a.jป a
a a E
as K
s ae. 10 s s
rH U> Z rH rH rH
-0 E
JO Q. Q. C
a a a.o
a a E
s s a K
s s s s s
CM rH rH rH rH rH
JO E C
a.jo a
a a a E
a X
s so: s s
rH rH rH Z rH rH
X X X X
+> 4ป 4> 43
. ._
> > > > o
ซ- ._ ซ >l
4> *> 4> 4ป U
n n ซ >i t> o
c c c c
O > w
ซl M M X
O
ao a a
H M M l_) C 1.
_l
c
o
+>
II
L A
p _l
C UI
a.
o
c n
O
<_> s
M
C
o
4>
~
"0
C
o
u
v
1
IH-
^
o
c
3
-
o
4ป
0
C
v
c
II
u
t.
3
O
L.
9
JC
4>
O
o
e
4>
II
o
A
C.
L.
3
4>
U
<ป-
3
C
E
E
o
*-
o
*>
E
4>
ซ
U
4> II
._
E ซ ซl
(1 4ป
_l C
0 E
a
L. ซl
3 ซ
ซ ซ. C
O 0
a o
X C 4>
UI U
II
*>
o:
JO Z Q
ซ
0 6 t
'I
L
ซ
a. 4>
o
z
30
-------
for this application. The Instrument chosen was a portable GC-AID monitor
with the following features:
Designed to operate outside In the extremes of weather.
Designed to operate on batteries and continue throughout the 8-hour
work day.
Is portable, easily carried by one person, and easy to operate and
maintain.
Responds to all six Indicator compounds at concentrations below the
respective levels of concern.
Shows a detection limit of better than 0.1 ppm for benzene.
Incorporates advanced microprocessor technology for near real-time
data output. Data can be retrieved by either reviewing on a computer
screen or connecting the GC-AID to a printer. The computer program
provides peak identification of up to at least 9 peaks, calibration
information, and concentrations of the 9 peaks.
4.3.4 Purchase and Maintenance Costs
The cost of the selected portable GC-AID, equipped with a modular oven
and an external battery, was $20,500. A calibration source and a carrier gas
source (two compressed gas cylinders: calibration source @ $3000 and carrier
gas at $100) are needed for field operation. Spare parts needed for
continuous operation include fittings and columns ($500) and inlet filters
($25).
4.4 QUANTITATIVE ASSESSMENT STRATEGY
The purpose of the quantitative assessment strategy is to document the
ambient air 8-hour average concentrations that occurred at the receptors
during the work day. Note that the total VOC-PID monitor described in the
refined screening strategy provides real-time total VOC concentrations at the
receptor site.
4.4.1 Operational and Performance Requirements
The operational requirements and performance capabilities of an
instrument for this application include the following:
It must collect a representative 8-hour sample year-round.
It must provide data of sufficient quality (precision and bias) for
meeting the DQOs.
31
-------
It must provide concentration results within 2 to 3 weeks of sampling.
It must provide speciation of all VOC compounds identified in the
liquid waste.
4.4.2 Overview of Commercially Available Instruments
Sample collection can be conducted using either a sorbent or a stainless
steel canister. Summa canisters and adsorbent tubes have both been success-
fully used for ambient air monitoring at superfund sites. Both methods have
their strengths, their applications, and their problems. Experienced users of
both methods believe that they are capable of producing data of sufficient
quality to satisfy the DQOs. For actual application, the selection would
probably be based on the personal preference of the user. Adsorbent tube
based sampling programs appear to be less tolerant of a lack of experience by
the sampling crew and sampling program designer than a canister based sampling
program. For this illustration of the DQO process, in this document, canister
samplers serve as the selected method of collecting 8-hour VOC samples. (It
is realized that there are differences of opinion on the stability of methyl
ethyl ketone and carbon disulfide in summa canisters.)
Canister sampling is conducted by one of two methods. One method is
where the canister fills with ambient air against a flow restrictor, regulated
by the vacuum within the canister. In other words, the canister begins and
ends sampling under vacuum. In the laboratory, the sample is pumped from the
canister for analysis. The second method is where the canister is filled by
using a pump. Once again, there are flow restrictors in-line to maintain a
constant flow over the sample period (8 hours). Thus, the canister begins
sampling under a vacuum and ends sampling under pressure. In the laboratory,
the pressure within the canister is used to deliver a sample to the GC.
Recent sampling programs within the EPA have demonstrated this second method
to be susceptible to contamination. Canister samplers are available from
several manufacturers and can be ordered in either of the configurations
described above.
To provide data of sufficient quality to satisfy the DQO, a laboratory GC
is required for analysis of the canister samples. A GC with one or more of
three detectors or a GC with mass spectrometry will provide the quality of
data necessary. The three detectors generally available are PID, FID, and
ECD; many times more than one detector would be operated on the same GC (note
32
-------
the AID has lower detection limits but may not, at this time, be available In
most laboratories). The detection techniques are listed in Table 7. As seen
1n the table, the PID responds to all six indicator compounds with detection
limits below the respective levels of concern. For this application, the low
detection limits are important since most of the VOC concentrations will be
below the levels of concern. At least one of the six compounds provide
minimal or no response using the FID or ECD detection techniques.
Since both systems GC/MS and GC-PID with confirmation by FID will report-
ably produce data of sufficient quality to satisfy the DQOs and have similar
costs the selection is based on convenience. Because of a local laboratory
equipped with and experienced in the use of GC-PID/FID this becomes the method
of choice for this application.
4.4.3 Capabilities of Selected System
The selected system will use evacuated canisters to collect an ambient
air sample and use GC-PID with confirmation by FID to analyze the ambient air
sample. Ten percent of the samples will be subjected to GC/MS for qualitative
confirmation. The primary criterion for subjecting a sample to analysis by
GC/MS is an indication by the on-site portable GC-AID monitor that unknown
compounds were released from the work site during that work day. These GC/MS
analyses will be performed by a laboratory that does not use a perma-pure
drier in the GC/MS sampling line because of the potential sample losses of
methyl ethyl ketone and carbon disulfide. The canister will begin and end
sampling under a vacuum (not using a sample pump) to minimize the potential of
contamination. The canister analysis will be conducted by a laboratory with
demonstrated experience. Demonstrated experience was documented by partici-
pating successfully in the EPA's audit cylinder repository program. Prior to
initiating the remediation activities, the laboratory chosen to analyze the
canisters will demonstrate capabilities for analysis of at least the six
indicator compounds.
The sampling system selected for this application will provide the
following:
A representative 8-hour sample
A precision of 20% RSD (relative standard deviation) with no overall
bias to satisfy the DQOs
33
-------
Speclation of all compounds expected to be emitted from the Superfund
site.
4.4.4 Purchase, Maintenance, and Analysis Costs
The cost of the selected canister sampler was $5000 and each canister
cost $500. Based upon using four canister samplers and thirty-four canisters,
the total purchase price would be $37,000. Spare parts should include tubing,
ferrels, and valves used to connect the canisters to the samplers. Spare
parts costs are estimated to be $500.
Based on 130 canisters being analyzed at $500 each. (See Section 5.4.3
for a breakdown).
34
-------
SECTION 5.0
MONITORING SYSTEM DESIGN
5.1 GENERAL
The purpose of this section is to design a monitoring system for VOCs in
ambient air during the remediation process at this hypothetical Superfund
site. One function of this monitoring system is to provide the RPM and/or on-
site coordinator (OSC) with data of sufficient quantity and quality to allow
for timely initiation of emission control actions. These emission control
actions should preclude exposing the subject population to VOC concentrations
greater than the levels of concern. This function of the monitoring system
will be partially accomplished by a screening strategy that employs a total
VOC-PID monitor at the work site to provide real-time data on emissions from
the site. In addition, if the screening strategy results show that total VOC
concentrations exceed a preset level, then a refined screening strategy is
initiated. The refined screening strategy employs two instruments, a portable
GC-AID near the work site to provide near real-time concentration values for
the six high risk indicator compounds and a total VOC-PID at the receptor site
to provide real-time total VOC concentrations. This limited speciation by the
portable GC-AID will reveal the presence or absence of the high risk com-
pounds. The presence of one or more of these compounds at preset concentra-
tion levels will be reason to alert the RPM/OSC to initiate emission control
actions, and, under certain conditions, to shut down the remediation activity.
The total VOC concentrations measured at the receptor site will also be used
to alert the RPM/OSC if predetermined concentrations are exceeded.
Another important function of the monitoring system is to generate data
necessary to preclude the RPM/OSC from unnecessarily slowing or stopping the
remediation. This function is fulfilled by using the refined screening
strategy. As part of the refined screening strategy the portable GC-AID is
deployed to provide compound-specific data on the high risk compounds. For
example, if the total VOC-PID monitor at the work site indicates a VOC concen-
tration of 100 ppm, the RPM/OSC may be inclined to stop the remediation
activity 1f, for example, the portable GC-AID indicates that the total 100 ppm
is benzene, which has a PEL of 1 ppm. Conversely, the RPM/OSC would not be as
35
-------
concerned 1f the total 100 ppm was, for example, methyl ethyl ketone (which
has a PEL of 200 ppm). Thus, this function of the monitoring system could be
very Important In the overall efficiency of the remediation program. Also,
results from the portable GC-AID can be used to provide tentative identifica-
tion of unknown compounds by comparing the observed relative retention time of
an unknown peak against relative retention times that had been obtained for
the more volatile, non-Indicator compounds in Table 2. The other part of the
refined screening strategy Involves deploying a total VOC-PID monitor at the
nearest impacted receptor site. This total VOC-PID monitor at the receptor
site will be used to inform the RPM/OSC that VOCs emitted at the work site are
reaching the receptor site.
A third function of the monitoring system is to provide information of
sufficient quantity and quality to assure the RPM/OSC, decision maker, and
residents at the receptor sites that the DQOs were met. This function is
fulfilled by the quantitative assessment strategy consisting of an evacuated
canister that collects an 8-hour sample at the subject receptor site for sub-
sequent analysis by laboratory GC-PID with FID confirmation. The precision,
bias, and speciatlon capabilities of this procedure indicate that the DQOs
will be satisfied for the compounds of interest for this Superfund site.
There are differences of opinions on the stability of methyl ethyl ketone and
carbon disulfide in canisters. For this illustration it is assumed that they
are stable. Recent evaluation tests indicate that at least methyl ethyl
ketone can be successfully collected in canisters.
A special characteristic of this site that has influenced the monitoring
system design is that the fenceline around the site is lined with a row of
full-grown trees, invalidating the customary procedure of monitoring at the
fenceline. This, plus the closeness of the receptor sites to the fenceline,
means that for a sample to be representative of the ambient air at the recep-
tor, the sample must be collected at or adjacent to the receptor.
Based on the modeling data and to fulfill the three functions described
above, three different monitoring strategies are being employed for this
project:
Screening at the work site for total VOC concentrations in real-time.
Refined screening at the work site for near real-time concentrations
of the six indicator compounds and at the receptor site for real-time
concentrations of total VOCs.
36
-------
Quantitative assessment at the receptor site for integrated 8-hour
averages of Individual VOCs.
These monitoring strategies will be used together to guide the on-site person-
nel in applying emission control actions and to document off-site concentra-
tions at the receptor sites.
5.2 SCREENING STRATEGY
Screening Is the least costly monitoring strategy, but provides the least
amount of Information. It does provide real-time data, however, allowing for
a quick response to a problem should one occur. Screening is to be employed
at the beginning of the project and/or work day and will continue until emis-
sions from the work site result in a measured VOC concentration that exceeds a
preset level. The monitoring system selected for screening is a total VOC-PID
monitor. This monitor will serve dual roles. It will provide health and
safety data and screening data for the potential off-site migration of VOCs
emitted from the work site. This system's capabilities are described in
Section 4.0. The following subsections discuss this monitoring strategy.
5.2.1 Deployment of the Total VOC-PID Monitor
For the screening strategy, when not in conflict with health and safety
monitoring, the total VOC-PID monitor will be positioned on a stable tripod 30
meters downwind of the work site and operated for the 8-hour work day. The
monitor will be repositioned at least once an hour if necessary to remain in
the plume center directly downwind of the work site. (As needed or when oper-
ating in the refined screening level, the total VOC-PID monitor can be used to
evaluate upwind concentrations and/or to help the on-site personnel locate the
exact emission source at the work site.)
Based on the modeling data, the probability of experiencing VOC concen-
tration levels of concern at any of the receptor sites is very low. Thus,
this simple screening strategy may prove to be the only strategy required
during the remediation program.
5.2.2 Actions Based on Screening Results
Screening results are available in real-time and are used to alert the
on-site personnel to potential problems. Specifically, results from the total
VOC-PID monitor based on preset criteria are used to:
37
-------
Alert the RPM/OSC that there is an emission source, allowing the
source to be located and controlled 1f necessary;
Alert the RPM/OSC to check that proper protective clothing 1s being
worn;
Direct the monitoring crew to Initiate refined screening; and
Direct the RPM/OSC to Initiate emission control actions.
5.2.3 Estimated Operating Manpower and Costs
Operation and maintenance of the total VOC-PID monitor is estimated to
require 1 FTE. This will consist of calibrating and deploying the monitor at
the beginning of each work day. The monitor can then operate unattended until
its alarm is activated, except for those times when repositioning is necessary
to keep the monitor 1n the plume centerllne and downwind of the work site.
The monitor is recalibrated at the end of the work day.
Operating costs, including cost of purchase, replacement parts, and cali-
bration standards as provided by the manufacturer and gas supplier, are as
follows:
Purchase price $4,500
Supplementary field kit $ 600
Supplementary battery $ 300
Extra battery (replacement) $ 300
Lamp $ 215
Inlet filter $ 25
Calibration standard $ 500
TOTAL $6,440
5.3 REFINED SCREENING STRATEGY
Refined screening is employed when a potential problem is indicated by
the total VOC-PID response. Refined screening is more costly than screening,
but 1t provides some compound-specific information, allowing the RPM/OSC to
better evaluate the seriousness of the problem. The portable GC-AID monitor
selected for refined screening can be programmed to identify and quantify at
least nine compounds. The nine programmed compounds can be changed if
38
-------
unexpected compounds posing high risk to the resident population are identi-
fied during the clean-up process. Initially, the portable GC-AID monitor will
be programmed to identify and quantify responses for the six indicator com-
pounds. The total VOC-PID monitor for this strategy is in addition to the one
used in the screening strategy. It provides the RPM/OSC real-time data on
total VOC concentration at the Impacted receptor site.
5.3.1 Deployment of the Portable GC-AID and the Total VOC-PID
Refined screening is conducted with the portable GC-AID monitor position-
ed 30 meters downwind of the work site. The instrument must be calibrated and
ready for operation at any time during the 8-hour work day. It should be
capable of being placed on-site and generating data within 15 minutes. The
portable GC-AID monitor will be repositioned as necessary, but at least once
an hour (while refined screening is in effect) to stay in the plume center-
line, directly downwind of the work site. An alternative approach to be
considered is to set-up the portable GC-AID in the on-site laboratory and take
syringe samples in the field for analysis 1n the laboratory.
The total VOC-PID will be located at the receptor site with the highest
probability of being impacted should a spill occur. This monitor will be used
during the work day and after work hours on evenings when meteorological con-
dictions indicate that there could be an inversion.
5.3.2 Actions Based on Refined Screening Results
Concentration values for up to nine compounds are provided by the port-
able GC-AID approximately once every 10 minutes. These compound-specific data
help the RPM/OSC know when to take action and the proper action to take.
Results from the portable GC-AID monitor based on preset criteria are
used to:
Notify the monitoring crew to go to the quantitative assessment
strategy, or
Notify the RPM/OSC to initiate emission control actions.
5.3.3 Estimated Operating Manpower and Costs
Operation and maintenance of the portable GC-AID monitor will require 3/4
FTE. Based on the modeling data, frequent use of the portable GC-AID monitor
39
-------
is not anticipated. To be able to respond in a timely manner to work-site
situations requiring the refined screening strategy, however, the GC-AID
monitor will have to be set up and calibrated at the beginning of the work
day. It should be placed in a shelter, on a cart, and in a location where it
can be easily placed on-site and generating data within 15 minutes when need-
ed. At the end of the work day, the GC-AID monitor will be returned to the
on-site laboratory and maintained as necessary to be ready the next day.
Operating costs, including cost of purchase, replacement parts, and cali-
bration standards as provided by the manufacturer and gas supplier, are as
follows:
Purchase price of portable GC-AID
equipped with a modular oven, and
external battery $20,500
Spare parts and filters 525
Calibration standard and carrier $ 3,100
gas for portable GC-AID
Purchase price of total VOC-PID $ 6,440
(See 5.2.3)
TOTAL $30,565
5.4 QUANTITATIVE ASSESSMENT STRATEGY
Quantitative assessment strategy provides more compound-specific inform-
ation than does refined screening. Results from the quantitative assessment
strategy are directly applicable to the receptor site. The monitoring system
selected for this level of monitoring includes the use of an evacuated
canister to collect the sample and a laboratory GC-PID with FID confirmation
for analysis. Also, at least 10% of the canister samples analyzed will be
subjected to GC/MS for qualitative confirmation. Specifically, the downwind
canister sample(s) collected on days that the portable GC-AID results indicate
that unknown compounds were released from the work site will be subjected to
GC/MS analysis to identify the unknown compounds.
5.4.1 Deployment of Sampler
For quantitative assessment, one or two receptor areas identified as
having the greatest probability of being impacted (i.e., being downwind of the
work site) will be instrumented with an evacuated canister sampling system.
40
-------
Meteorological information will be used to identify the site(s). In areas
where two or more receptors are located, a sampling site will be selected so
as to be representative of all these clustered receptors. Also, an upwind or
parallel site will be selected and instrumented. On a predetermined schedule
a co-located sampler will be placed at the receptor site most likely to be
impacted as part of the QA program. The samplers must be in place and operat-
ing over the 8-hour work day. Results from the portable GC-AID and/or total
VOC-PID, combined with meteorological data, will be used to determine if the
collected sample will be forwarded to the off-site laboratory for analysis.
5.4.2 Actions Based on Quantitative Assessment Results
Results from analysis of quantitative assessment samples are not avail-
able until about two weeks after sample collection. Thus, the data will be
used to develop a data base documenting VOC concentration levels experienced
at one or more of the receptor sites for various work-site situations. For
example, values for the ratio of the concentrations of an indicator compound
at the work site and the receptor site for known meteorological and work-site
conditions (for example, multiple accidental spills) will be calculated. This
information will assist the RPM/OSC in making decisions about the need for
emission control actions under similar future work site conditions.
Results of quantitative assessment strategy based on preset criteria will
be used to:
Notify the RPM/OSC that remediation procedures must be changed to
reduce emissions if levels of concern are regularly approached or
frequently exceeded at one or more receptor sites.
Notify the RPM/OSC that certain receptor sites must be evacuated
before continuing remediation if the levels of concern are regularly
exceeded in spite of attempted emission control procedures.
Provide the RPM/OSC with accurate measures of VOC concentration levels
experienced at receptor sites.
5.4.3 Estimated Operating Manpower and Costs
Operation and maintenance of the sample collection activities, and data
processing and analysis for this monitoring strategy are estimated to be 1 and
1/4 FTEs. Based on the modeling results, it is anticipated that this level of
monitoring will not be required for any extended period of time, except pos-
sibly for the two nearest receptor sites. During the remediation program, if
41
-------
results from the quantitative assessment samples Indicate that VOC concentra-
tions do not approach the levels of concern, even on days when spills occurred
and the meteorological conditions were extreme, then a decision will be made
to discontinue the quantitative assessment strategy. For costing purposes, it
is estimated that a maximum of 130 quantitative assessment samples will be
analyzed during the remediation program. This includes downwind, upwind, co-
located, trip blank, and batch blank (cleaned canisters) samples. At least
initially, on-site manpower is required to deploy the three or four samplers
before each work day, collect samplers after the work day, and deliver the
samples to the nearby commercial laboratory for analysis. It is expected that
a great deal of time will be required to review, evaluate, and prepare the
data in a form that can easily be interpreted by the RPM/OSC.
The costs associated with the quantitative assessment strategy as
provided by equipment manufacturers and a commercial analytical laboratory
are as follows:
Purchase price of canister
samplers (4 at $5,000 each) $ 20,000
Purchase of 34 canisters at
$500 each $ 25,000
Canister analysis by GC-PID/FID
at $500 each (130 assumed) $ 65,000
Canister analysis by GC/MS at
$1,000 each (5 assumed) $ 5,000
TOTAL $115,000
5.5 CRITERIA FOR EMPLOYING A MONITORING STRATEGY
The criteria for moving from one strategy to another is based on measured
VOC concentrations. In the absence of VOC ambient air measurement data, the
criteria are purposely set to error on the side of safety. These criteria
will be reevaluated and changed if necessary as monitoring data become
available. The criteria are discussed for each strategy in the following
subsections.
42
-------
5.5.1 Screening Strategy
The screening strategy 1s to be employed at the beginning of each work
day, unless experience has shown that the remedial activity will result in
emission levels that trigger the need for refined screening. The criterion
for moving up to the refined screening strategy is any time the total VOC
response exceeds 5 ppm for 10 consecutive minutes. This is also the safety
criterion for changing from level C to level B dress for on-site workers.
Moving from the refined screening strategy back to the screening strategy
is accomplished by simply discontinuing use of the portable GC-AID. The
criterion for this move 1s when the total VOC-PID response has been below 5
ppm for two consecutive hours after the Incident that triggered the need for
refined screening.
The rationale for the criterion of 5 ppm VOC for 10 consecutive minutes
(in addition to it being the safety criterion) follows: benzene, identified
as posing the highest risk from the list of compounds in the burled waste, has
a PEL of 1 ppm. If, for example, the refined screening results showed the 5
ppm VOC measurement to be 5 ppm benzene, then the RPM/OSC will be notified to
initiate emission control actions. Likewise, if the 5 ppm VOC was shown to be
5 ppm methyl ethyl ketone, which has a PEL of 200 ppm, the remediation
activity could proceed unimpeded.
5.5.2 Refined Screening Strategy
The refined screening strategy is employed whenever the total VOC-PID
response has exceeded 5 ppm for more than 10 consecutive minutes, or at the
beginning of the work day if experience indicates a high probability that
initiation of remediation will result in screening results greater than 5 ppm.
5.5.3 Quantitative Assessment Strategy
The quantitative assessment strategy is implemented in the event that
results from either the on-site portable GC-AID or the total VOC-PID monitor
located at the receptor site exceed preset values.
First, results from the portable GC-AID are used to calculate an
equivalent exposure (Em) value for the six indicator compounds as follows:
Em = (Ci+PELi) + (C2*PEL2) + (C3*PEL3) + (C4*PEL4) + (C5*PEL5) + (C6*PEL6)
43
-------
where C 1s the measured concentration of each Indicator compound and the PEL
is the respective permissible exposure limit.
For example, suppose the portable GC-AID monitor gave the following
results: benzene, 0.4 ppm; trichlorethylene, 5 ppm; carbon tetrachlorfde, 0.5
ppm; 1,1,2-trichloroethane, 1 ppm; carbon dlsulfide, 1 ppm; and methyl ethyl
ketone, 50 ppm. Combining these concentrations with the respective PELs given
in Table 3, Section 2.0 yields the following results:
Em = (0.4*1) + (5*50) + (0.5*2) + (1*10) + (1*4) + (50*200)
Em = 0.4 + 0.1 + 0.25 + 0.1 + 0.25 + 0.25
Em = 1.35
The quantitative assessment strategy is then initiated (that is, one or
both of the downwind canister samples (plus the upwind, co-located and trip
blank samples) 1s forwarded to the laboratory for analysis at the end of the
8-hour work day) if the calculated Em is greater than 1 for 30 minutes or more
during the work day.
A second criterion, independent of the portable GC-AID result, is based
on total VOC measurements at the receptor site. The quantitative assessment
strategy is initiated if the refined screening strategy results show total VOC
concentrations greater than 0.5 ppm for 30 minutes or more during the work
day.
These criteria are subjective, however; compound concentrations giving an
equivalent exposure value greater than 1 occurring 30 meters downwind of the
work site may result in measurable levels at the impacted receptor site(s).
Also, a total VOC concentration of 0.5 ppm, corrected for background levels,
signal the need for application of emissions control actions at the work site.
The reason for having two criteria, one based on total VOC and one on
compound-specific results, is that the total VOC concentration may represent
compounds other than the six indicator compounds. The quantitative assessment
strategy would identify and quantify all VOCs present at or above detection
limit concentrations at the receptor site.
44
-------
5.6 CRITERIA FOR APPLYING EMISSION CONTROL ACTIONS
5.6.1 Screening Strategy
Criteria for taking actions based on screening strategy results follow:
If the total VOC-PID response 1s 5 ppm or greater for 10 successive
minutes, alert the RPM/OSC that there 1s an emission source so that
the source may be located and controlled as necessary.
If the total VOC-PID response 1s 5 ppm or greater for 10 successive
minutes, alert the RPM/OSC to check that proper protective clothing is
being worn.
If the total VOC-PID response 1s 200 ppm or greater for 10 successive
minutes, alert the RPM/OSC that emissions must be reduced within the
next 30 minutes or halt the remediation activity. (This applies only
if the monitoring staff for some reason has been unable to initiate
the refined screening strategy in this time period.)
The rationale for the criterion of 5 ppm total VOC is that benzene could
account for the major portion of the VOC measurement. Benzene has a short-
term exposure limit (STEL) of 5 ppm; thus, the on-site workers would be
alerted to wear the proper protective clothing. Also, a potential 5 ppm con-
centration of benzene 30 meters downwind of the work site may result in a
concentration near the level of concern (100 ppb) at one of the receptor
sites.
The rationale for the criterion of 200 ppm for 10 successive minutes is
that on-site VOC concentrations at this level will probably result in measur-
able VOC concentrations off-site.
5.6.2 Refined Screening Strategy
The compound-specific information provided by the portable GC-AID monitor
is used to guard against allowing one or more of the high risk compounds to
reach the receptor site at concentration levels near one tenth of their
respective PELs.
There are two criteria for intiating emissions control actions based on
refined screening data. The two should be evaluated separately, that is
action must be taken should either one be exceeded. The criterion for
emission control actions follows:
The RPM/OSC is notified that emissions must be reduced within 30
minutes or remediation will have to be halted if the portable GC-AID
monitor response results in a calculated Em greater than 1 for 10
successive minutes (see the example calculation in subsection 5.5.3).
45
-------
The RPM/OSC Is notified that emissions must be reduced within 30
minutes or remediation will have to be halted.if the total VOC-PID
monitor at the receptor site shows a concentration of 0.5 ppm or
greater for 10 successive minutes.
The rationale for the criterion of Em ฃ 1 for 10 successive minutes is
that one or more of the six indicator compounds could be present at concentra-
tions near their respective PEL which could result in receptor site concentra-
tions near the levels of concern. A total VOC concentration of 0.5 ppm at the
receptor site is one tenth of the work site criterion of 5 ppm for moving from
level C to level B protective clothing. This is the same rational used for
establishing the level of concern at one tenth of the PEL.
5.6.3 Quantitative Assessment Strategy
Quantitative assessment results directly estimate the health risks
experienced by residents at the receptor sites. Thus, these results are used
by the RPM/OSC to assess when the remediation process being used needs to be
changed so that emissions are reduced.
Criteria for taking action based on quantitative assessment strategy
results follow:
The RPM/OSC is notified that remediation procedures must be changed to
reduce emissions if any compound identified routinely approaches or
exceeds 0.1 PEL, or that the remediation activity must be halted.
The RPM/OSC is notified that a certain receptor site must be evacuated
during the 8-hour work day when meteorological information indicates
that it will be in the plume's path and if results have shown one or
more compounds to exceed 0.2 PEL under similar meteorological
conditions. Otherwise, the remediation activity must be halted.
These above criteria of 0.1 PEL and 0.2 PEL are used here because they
are the concentration levels of concern (0.1 PEL) and the concentration level
at which the decision maker stated that the monitoring system should be such
that the probability of a false negative is no greater than 1 percent.
5.7 ESTIMATED TOTAL LABOR AND COST
In Stage 1 of the DQO the following resources were made available for
this ambient air monitoring effort:
A mobile laboratory to be placed on-site for the duration of the
cleanup operation.
46
-------
Equipment necessary for cleaning the sumrna canisters in the on-site
laboratory.
Two monitoring technicians to be dedicated to this air monitoring
effort (i.e., 2 FTEs).
The monitoring effort is estimated to require a total work force of 3
FTEs. Thus, one FTE will be required 1n addition to the two provided by the
ESD. It is Important to not over-commit the on-site workers because concen-
trated monitoring activities could occur at any time and the monitoring staff
needs to be prepared. The three monitoring personnel will probably have to
stagger their shifts in order to fully service the three monitoring systems
before and after the work day. That is, the two total VOC-PID monitors and
the portable GC-AID must be calibrated before and after each work day. Also,
the canisters must be deployed and the samplers set to take a sample spanning
the 8-hour work day. Costs have been estimated for each monitoring strategy
and are based on the assumption that the refined screening and quantitative
assessment strategies will be required Infrequently.
Individual costs estimates listed below are only for illustrating the
process. The individual item costs are believed to be reasonable but they do
not represent actual quotes from manufacturers or contractors nor is the
listing presented as a comprehensive list.
Project Planning
DQO Package $75,000
- Sampling and analysis plan
- Quality assurance project plan
- Summary work plan
Sub Total $75,000
Manpower
One technician (1 FTE) $130,000
(estimated @ 2,500/week for 52 weeks
this includes per deim, overtime, etc.)
$130,000
47
-------
Screening Strategy Costs
purchase price of total VOC-PID monitor $ 4,500
supplementary field kit 600
supplementary battery 300
replacement battery 300
replacement lamp 215
Inlet filters (5) 25
calibration standard 500
Refined Screening Strategy Costs
Sub-Total $ 6,440
purchase price of portable GC-AID
monitor $20,500
spare parts and filters 525
calibration standard and carrier 3,100
gas for portable GC-AID
purchase price of total VOC-PID 6,440
monitor and needed supplies
Sub-Total $30,565
Quantitative Assessment Strategy costs
purchase price of canister sampler (4) $ 20,000
purchase price of 34 canisters 25,000
canister analysis by GC- PID/FID 65,000
$500 per sample (130)
canister analysis by GC/MS at $1,000 5,000
each (5 assumed)
Sub-Total $115,000
GRAND TOTAL $357,005
This total estimated cost greatly exceeds the $50,000 that the decision
maker has reported as being available for this effort. Some of the monitors
or other equipment such as canisters may be available from the ESD. If so,
this could reduce the costs but probably not nearly to the level of available
funds. If all the monitors and supplies must be purchased, the possibility of
obtaining the total $357,000 will be investigated.
48
-------
SECTION 6.0
ERROR ANALYSIS FOR COMPLIANCE WITH DQOs
The purpose of this section 1s to show that the monitoring system design
will satisfy the stated DQOs. The DQOs specified by the decision maker and
presented in subsection 1.2 of this document are compound-specific and apply
to the compound concentrations occurring at the receptor sites.
One aspect of satisfying the DQOs 1s that a valid sample is collected for
the receptor site experiencing compound concentrations near the levels of
concern. The criteria for locating the sampling systems at the beginning of
each work day are based on meteorological data; the two receptor site areas
most likely to be impacted by the work-site plume, should a spill occur, are
instrumented each day. Based on the locations of the receptor areas (see
Figure 1), the likelihood of not instrumenting the area that will be impacted
most by the work-site plume is very small. Also, the decision to submit the
collected sample for analysis 1s dependent on results from screening or refin-
ed screening that indicate on-site concentrations above predetermined values.
Since the work site 1s a below- ground-level source with no effective plume
rise, it appears highly unlikely that concentration levels of concern will
occur at receptor sites and not first be detected by the on-site screening or
refined screening procedures. Thus, the decision to submit the canister
sample for analysis will be a correct one.
The evacuated canister/laboratory GC-PID system selected for quantitative
assessment strategy was evaluated for its potential to satisfy the DQOs. It
should be pointed out here that during conduct of the project, adequate QC/QA
procedures must be carried out to assure the decision maker and/or RPM that
the measurement system is producing data of acceptable quality.
Salient features of this measurement system that influence data quality
include, but may not be limited to, the following:
The canister sampler is located at or near the receptor of interest;
that is, there is no error due to spatial variability.
The canister sampler collects the 8-hour sample at a constant flow
rate; that is, there is no error in the 8-hour average due to temporal
variability.
The method is compound-specific for the compounds of concern.
49
-------
The scientists experienced with the method provided an overall preci-
sion estimate for sample collection and analysis, expressed as a
relative standard deviation (RSD) of 20%.
The system is reported to have no overall bias across the compounds of
interest.
Compound-specific data quality objectives were established at three
concentrations: 0.5, 1.5, and 2.0 times the level of concern (the level of
concern 1s 1/10 of the permissible exposure limit). An error analysis was
performed for one compound, benzene. The permissible exposure limit (PEL) for
benzene is 1000 ppb; thus, the level of concern is 100 ppb. The detection
limit for benzene 1s 1 ppb, well below the level of concern.
Following are the data quality objectives and the results of the error
analysis, using benzene as an example:
1. "At a true average concentration of 0.5 times the level of concern,
the probability of a positive finding should be limited to less than
10%."
For benzene, 0.5 times the level of concern is 50 ppb. At this con-
centration, the standard deviation is expected to be 20% of 50 ppb,
or 10 ppb. Figure 6 shows a normal distribution having a mean of 50
and a standard deviation of 10. Virtually all of the area beneath
the curve is to the left of 100, showing that the probability of
finding a measured value above 100 ppb is practically zero when the
true concentration is 50 ppb. This is a far better (smaller) error
rate than the objective, 10%. Thus, this DQO should be satisfied.
2. "When the true average concentration is 1.5 times the level of con-
cern, the probability of a negative finding should be limited to less
than 5%."
For benzene, 1.5 times the level of concern is 150 ppb. At this con-
centration, the standard deviation is expected to be 20% of 150 ppb,
or 30 ppb. Figure 7 shows a normal distribution having a mean of 150
and standard deviation of 30. The hatched area below the curve and
to the left of 100 ppb is the probability of a negative finding when
the true concentration is 150 ppb. The figure shows that the prob-
ability of finding a measured value below 100 ppb is very near our 5%
objective. In fact, the probability is 4.8%. Thus, the DQO should
be satisfied as long as the precision, expressed as the relative
standard deviation, is not much greater than 20%.
50
-------
0
in
U
c
0
u
1
till
i ! i r
i i i i I i i I i
1
-
Q
0
r-|
* Q
e in
c Q:
9
Q .\*
0
X (M
4)
^ 0
JO E
0
0 Z
L
Q.
L h
0
00
Q
(0
-J 0
0
(M
c
0
0
L
4J
C
0
0
c
0
u
L
3
I
0
II
I
o> c
01 5
M t.
O 4)
O. U
V O
M U
ซ*-.
(O
M Q.
-- Q.
*- O
JQ m
Q. C
O
O C
ซl
C U J3
oca.
00.
*j o
C
(.
4>
Q
0
ro
0
(M
0
Q
0
t-
3
D)
tZ
51
-------
0
in
o
a
0
in
t\i
o
in
0
in
c
o
-r(
8ซ
2
L
4ป
C
0
U
c
0
u
TJ
0 J
H D
0
0
a
i
o
ป*-
v
- o
CX
CX
ซ> *j ^*
M 0> J-
jr a>
o
u
ซ- p c
tiEซ
3 4J *J
r 4> (_
M 4>
C
a> o
41 C O
o>
X
v
ouop
52
-------
3. "When the true average concentration is 2 times the level of concern,
the probability of a negative finding should be limited to less than
IX."
For benzene, 2 times the level of concern is 200 ppb. At this con-
centration, the standard deviation is expected to be 20% of 200 ppb,
or 40 ppb. Figure 8 shows a normal distribution having a mean of 200
and a standard deviation of 40. The area under the curve and to the
left of 100 ppb is the probability of a negative finding, given that
the true concentration is 200 ppb. The figure shows that the proba-
bility of finding a measured value below 100 ppb is very small. This
probability is 0.62%, which is better than the acceptable probability
of 1%.
Because the same precision (20% RSD) or better is estimated for each of
the compounds of concern, identical results would be realized for each of the
other compounds if subjected to the above error analysis. These results show
that a precision of 20% for sampling and analysis represents the limit of
imprecision that will satisfy the DQOs. If the precision of the monitoring
system is in fact 20% RSD then extreme care must be exercised to identify and
eliminate or quantify and adjust for biases. In this project a negative bias
would increase the probability of a false negative error, meaning that an
exposure of the public to concentrations in excess of the level of concern
could go undetected.
53
-------
0
0
(M
0
0
X
V
ฐ S S
o " ฃ
0
L
4>
C
8
0
C
0
U
o
o
o
o
1
L
3
0
0
V
0) 4-
O.
H- C
O 0> O
U O
gs-
Z 5^42
ฃ 3 C
M J- J-
V C
r -C O
I
O)
X-q.fsi.JOp
54
-------
SECTION 7.0
REFERENCES
1. Research Triangle Institute. Guidance on Applying the Data Quality
Objectives Process for Ambient A1r Monitoring Around Superfund Sites
(Stages I and II).EPA-450/4-89-015, August 1989.
2. U.S. Environmental Protection Agency. Development of Data Quality
Objectives, Description of Stages I and It.Quality Assurance Management
Staff, July 16, 1986.
3. U.S. Environmental Protection Agency. Procedures for Conducting A1r
Pathway Analyses for Superfund Applications, Volume I, Application of Air
Pathway Analyses for Superfund, December 1988 (draft).
4. U.S. Environmental Protection Agency. Procedures for Conducting Air
Pathway Analyses for Superfund Applications, Volume II, Procedures for
Developing Baseline Emissions from Landfills and LagoonsTEPA-450/1-89-
002, January 1989.
5. U.S. Environmental Protection Agency. Procedures for Conducting Air
Pathway Analyses for Superfund Applications, Volume III, Procedures for
Estimating Air Emissions Impacts from Remedial Activities at NPL Sites.
EPA-450/1-89-003, January 1989.
6. U.S. Environmental Protection Agency. Procedures for Conducting Air
Pathway Analyses for Superfund Application, Volume IV, Procedures for
Dispersion Modeling and Air Monitoring for Superfund Air Pathway Analyses,
f
December 1988 (draft).
7. U.S. Environmental Protection Agency. Data Quality Objectives for
Remedial Response Activities, Volume I, Development Process.EPA 540/G-
87/003, March 1987.
8. U.S. Environmental Protection Agency. Data Quality Objectives for
Remedial Response Activities, Volume II. RI/FS Activities at a SiTe with
Contaminated Soils and Ground Water. EPA 540/G-87/004, March 1987.
9. U.S. Environmental Protection Agency. Data Quality Objectives for the
Toxic Air Monitoring Program (Stages I and II).Office of Air Quality
Planning and Standards, December 1987.
10. U.S. Environmental Protection Agency. Data Quality Objectives for the
Urban Air Toxic Monitoring Program (Stages I and II). Office of Air
Quality Planning and Standards, June 6, 1988.
11. U.S. Environmental Protection Agency. Hazardous Waste Treatment, Storage,
and Disposal Facilities (TSDF) - Air Emission Models, EPA-450/7-87-026.
December 1987.
12. Turner, D.B. Workbook of Atmospheric Dispersion Estimates, U.S. Depart-
ment of Health, Education and Welfare, and Public Health Service, Public-
ation No. 999, April 26, 1970.
55
-------
13. Briggs, G.A. Diffusion Estimation for Small Emissions. 1973 Annual
Report, Air Resources Atmospheric Turbulence and Diffusion Laboratory,
Environmental Research Laboratory, Report ATDL-108, USDOC-NOAA, 1973.
14. U.S. Environmental Protection Agency. Guideline on Air Quality Models,
EPA-450/2-78-027R, 1978.
15. Holdren, M.W., D.L. Smith, and R.N Smith. Comparison of Ambient Air
Sampling Techniques for Volatile Organic Compounds.EPA/600/S4-85/067,
January 1986.
16. Berkley, Richard E. Evaluation of Photovac 10S50 Portable Photoionization
Gas Chromatograph for Analysis of Toxic Organic Pollutants in Ambient ATrT
EPA/600/S4-86/041, March 1987.
17. Ressl, Robert A., and Thomas C. Ponder, Jr. Field Experience with Four
Portable VOC Monitors. EPA/600/S4-85/012, March 1985.
18. Fulcher, James N., Jr., Alex R. Gholson, and R.K.M. Jayanty. Evaluation
of Portable Personal Phptoionization Gas Chromatography (Laboratory
Evaluation).Prepared by Research Triangle Institute under EPA Contract
No. 68-02-4544, 1989.
19. Fulcher, James N., Jr., Alex R. Gholson, and R.K.M. Jayanty. Evaluation
of Portable Personal Phptoionization Gas Chromatography (Field
Evaluation).Prepared by Research Triangle Institute under EPA Contract
No. 68-02-4544, 1989.
56
-------
U.S. Envl;
T>-it-'r->
------- |