&EPA
EPA/600/R-23/110 | May 2023 | wviAv.epa.gov/research
United States
Environmental
Protection Agency
Smoke Plume Air Monitoring
Analysis During Freshwater
In-Situ Oil Burn (ISB) Research
Office of Research and Development
Center for Environmental Measurement & Modeling
-------
SMOKE PLUME AIR MONITORING ANALYSIS DURING FRESHWATER
IN-SITU OIL BURN (ISB) RESEARCH
EXECUTIVE SUMMARY
This report documents the results of an interagency project to sample emissions from in situ oil
burns, comparing ground-based particulate matter measurements with measurements by aerial
unmanned aircraft system (UAS). A sensor/sampler system from the Environmental Protection
Agency (EPA), Office of Research and Development (ORD), was used to measure gas and
particulate emissions from an oil fire on water at the Army Corps of Engineers Cold Region
Research Engineering Laboratory. This sampler was placed on board a UAS and flown into the
plume of the oil burn to sample the emissions. Concurrent ground measurements were made by
EPA and the Coast Guard Strike Team using various particle and gas sampling devices for the
purpose of comparing aerial and ground measurements. As well, the EPA deployed an array of low
cost particle sensors to compare with the Strike Team's Special Monitoring of Applied Response
Technologies (SMART) protocol. Emission factors for PM2.5 calculated from the UAS
measurements were consistent with past data, showing decreasing PM2.5 values with improved oil
combustion efficiencies. Combustion efficiencies did not show a relationship with oil consumption,
verifying past results. All real time PM2.5 instruments measured considerably lower PM2.5 emissions
than the primary filter catch standard.
These instruments need to be calibrated with the plume smoke as the optical properties of the in situ
oil burn (ISB) plume particles differ greatly from that of ambient air. Deployment of a high number
of low cost sensors can provide a more comprehensive network of measurementsthan the limited
number of more expensive SMART sensors.
-------
QUALITY ASSURANCE SUMMARY
This work was conducted under the U.S. EPA Quality Assurance (QA) program to ensure data are
of known and acceptable quality to support their intended use. Surveillance of the work by the
scientific leads ensured adherence to QA processes and criteria, as well as quick and effective
resolution of any problems. The QA manager and scientific leads have determined under the QA
program that this work meets all U.S. EPA quality requirements. This project was performed
according to the approved QAPP titled "Freshwater In-Situ Burn Air Monitoring: USCG," QAPP
ID: 0033370-QP-1-0. As part of the QA system, a quality product review is done by the CEMM
QA Manager prior to management clearance.
-------
TABLE OF CONTENTS
1 INTRODUCTION 1
2 METHODS 1
3 RESULTS AND DISCUSSION 3
4 CONCLUSIONS. 13
5 ACKNOWLEDGEMENT 14
6 REFERENCES 14
LIST OF FIGURES
Figure 1. CRREL Geophysical Research Facility the location of the tank in which the in situ
burns were conducted 2
Figure 2. Photo of the GRF and UAS/Kolibri during one of the ISBs 2
Figure 3. Overhead view of CRREL and the GRF showing three co-location sites for the ground-
based monitors. The green marker is the Super Site 4
Figure 4. Locations at which PurpleAir sensors were variously positioned during thetesting 4
Figure 5. Comparison of CO and CO2 emission factors as compared to the total modified
combustion efficiency. Some burns had two filter samples, resulting in more data
points than burns 5
Figure 6. A comparison of the PM2.5 emission factors as a function of total modified combustion
efficiency 6
Figure 7. Comparison of PM2.5 emission factors collected in the near source plume and the far
source (downwind) plume 7
Figure 8. Comparison of PM2.5 emission factors from this work with previous determinations 8
Figure 9. Non-methane hydrocarbon emission factors versus total modified combustion
efficiency 8
Figure 10. A comparison of the amount of oil consumed by the fire and the total modified
combustion efficiency 9
Figure 11. Comparison of uncorrected EPA DRX DustTrak with filter-corrected SidePak. Color
indicates ISB number. Data are 2-min averaged. The red line is a linear fit line. Linear
regression coefficients are given as a and b where y = bx + a 10
Figure 12. Comparison of uncorrected EPA DRX DustTrak with uncorrected USCG SMART
DRX DustTrak. Color indicates ISB number. Data are 2-min averaged. Linear
regression coefficients are given as a and b where y = bx + a 11
Figure 13. Uncorrected PurpleAir sensor versus filter-corrected SidePak. Color indicates ISB
111
-------
number. Data are 2-min averaged. Linear regression coefficients are given as a and b
where y = bx + a 11
Figure 14. Corrected EPA DRX DustTrak versus filter-corrected SidePak. Color indicates ISB
number. Data are 2-min averaged. Linear regression coefficients are given as a and b
where y = bx + a 12
Figure 15. Corrected PurpleAir results versus filter-corrected SidePak. Color indicates ISB
number. Data are 2-min averaged. Linear regression coefficients are given as a and b
where y = bx + a 12
LIST OF TABLES
Table 1. Meteorological conditions during the ISBs 3
Table 2. CO, CO2, and MCEt values for the seven burns 5
Table 3. PM2.5 emission factors and MCEt values 6
ACRONYMS & ABBREVIATIONS
ANS
Alaska North Slope crude oil
AVG
Average
CEMM
Center for Environmental Measurement and Modeling
CO
Carbon monoxide
C02
Carbon dioxide
CRREL
Cold Regions Research and Engineering Laboratory
EC
Elemental Carbon
EPA
Environmental Protection Agency
GRF
Geophysical Research Facility
ISB
In situ oil burn
MAX
Maximum
MCE
Modified combustion efficiency
MIN
Minimum
NMHC
non-methane hydrocarbon
OC
Organic carbon
ORD
Office of Research and Development
OTM
Other test method
PM2.5
Partiuclate matter equal to and less than 2.5 um
QA
Quality assurance
QAPP
Quality assurance project plan
RDC
Coast Guard Research and Development Center
SD
Standard deviation
iv
-------
SMART Special Monitoring of Applied Response Technologies
TC Total carbon
UAS Unmanned aircraft system
USCG United States Coast Guard
VOCs Volatilie organic compounds
v
-------
1 INTRODUCTION
Fresh water in-situ oil burns (ISBs), purposeful burns of surface oil resulting from accidental spills
and leaks, tend to be close to near shore population centers, raising exposure and visibility concerns
from the resultant smoke. Limited data are available on emissions from these burns making
decisions about whether to conduct ISBs uncertain. Current methods of measuring potential
exposure risk require placing a limited number ofdownwind samplers in the estimated direction of
the plumes. This method is documented in the SMART protocol, employed by emergency
responders including the US Coast Guard Strike Teams. The SMART protocol runs the risk of
missing the predicted plume direction due to wind direction changes resulting in limited or no data
on the plume. The protocol also requires potential personnel exposure to the plume during
placement and monitoring of the sensors. Use of remotely piloted vehicles, or unmanned aircraft
systems(UAS) carrying gas and particle sensors eliminate this uncertainty and exposure risk due to
their locational flexibility and standoff operation. UAS-based measurements on ISBs have been
extensively demonstrated in recent years [1,2]. A further limitation of the current SMART protocol
is the minimal number of instruments available for deployment due to their relatively high cost. The
current availability of low cost sensors that can be deployed in an array, shows promise for better
area coverage provided larger numbers of sensors can be used for a lower cost [3],
An ISB sampling project at the US Army Corps of Engineers Cold Regions Research Engineering
Laboratory (CRREL) was conducted in concert with the U.S EnvironmentalProtection Agency's
Office of Research and Development (ORD), the U.S. Coast Guard's Strike Team, and the U.S.
Coast Guard Research and Development Center (RDC). UAS-based sampling and ground-based
sampling were conducted to:
quantify emissions from fresh water, in-situ oil burns using a UAS,
compare emissions quantity to the combustion efficiency,
compare UAS-based aerial measurements with SMART measurements,
compare SMART measurements with low-cost particulate matter sensormeasurements
provide information for decision makers related to near shore, freshwaterenvironments
(e.g., Great Lakes).
2 METHODS
ISBs were conducted at CRREL's (GRF) tank (Figure 1), with fresh water and Alaska North Slope
crude oil (ANS). ANS was poured gently intoa 1.5 m2 pan near the center of the GRF tank, creating
a 4.17 cm slick thickness (B. Booker, CRREL, email). The oil was ignited via a propane torch. The
plume wassampled using a UAS with the ORD Kolibri sensor/sampler system (Figure 2).
Meteorological and oil measurements were made by CRREL.
1
-------
Figure 1. CRUEL Geophysical Research Facility the location of the tank in which the in situ burns
were conducted.
Figure 2. Photo of the GRF and llAS Kolibri during one of the ISBs.
Aerial measurements were conducted with the EPA's Kolibri which provided time- resolved CO,
CO2, and particles with aerodynamic diameter less than 2.5 jam diameter (PM2.5) as well as batch
measurements of Organic Carbon/Elemental Carbon/Total Carbon (OC/EC/TC) and PM2.5. Ground
sampling was conducted by the USCG Strike Team using the SMART protocol (USCGNo. CG-D-
-------
08-14, 2014), which consisted of TSI DustTraks for PM2.5 in three locations and several AreaRae
samplers (https://www.electrogasmonitors.com/proiect/arearae-pro/) for H2S, CO, and VOCs. Up
to eight PurpleAir PM sensors (https://www2.purpleair.com/pages/technology) were deployed by
ORD at each of the three Strike team locations and at other sites deemed to be impinged by the
plume. At one"Super Site" collocated samplers consisted of the Smart Team's DustTrak, the ORD
PurpleAir sensor, an ORD DRX DustTrak and ORD's backup Kolibri. The reader is referred to
Appendix 1, Quality Assurance Project Plan J-IO-0033370-QP-1-0, "Freshwater In-Situ Burn Air
Monitoring: USCG," for more details on the ORD measurements.
The primary data reported are emissions factors which here are in units of pollutant mass per unit
mass of oil burned. These measurements are determined using the carbon balance method as
described in OTM-48 [4],
3 RESULTS AND DISCUSSION
Eight (8), replicate burns were conducted from October 25th to October 27th, 2021. Seven (7) of
these burns were sampled for emission. The average burn duration was 15.6min, and the average
burn rate was 2.67 mm/min (B. Booker). CRREL pre-weighed the oil absorbent pads and
reweighed the used, post-burn pads to determine residual oil massand oil mass burned. Their results
showed that 90-96% (93% avg) of the oil mass burned(B. Booker).
The meteorological conditions were determined at the ISB burn site and are shown inTable 1 for the
3-day operations.
Table 1. Meteorological conditions daring the ISBs.
MAX
MIN
AVG
SD
Temperature, °C
14.0
7.3
11.6
2.2
Rel. Humidity, %
93.6
46.8
63.7
19.1
Wind Speed, km/h
10.8
0.1
6.6
3.9
Figure 3 shows the three ground, co-location sites of the emission samplers - the Super Site and
two others. Figure 4 shows the location of the PurpleAir sensors throughout the test program.
These sensors were moved around in order to accommodate the prevailing wind patterns.
-------
Jr-
There were three collocation sites. Site #1
"Super Site" collocated PurpleAir, EPA
DustTrak, EPA SidePak, and SMART
DustTrak together
CRREL Oct 25-27
7 views
Last edit was seconds ago
^ Add layer Share <) Preview
V SMART Collocation Sites
^ Individual styles
9 Site #1 f Super Site")
9 Site #2: Downwind E
Q Site #3: Upwind
Base map
Figure 3. Overhead view of CRREL and the GRF showing three co-location sites for the
ground-based monitors. The green marker is the Super Site.
MyjMaDs
Eight PurpleAir
were deployed to
14 locations over
the courseof 7
burns
| Map data ©2022 Imagery ©2022, Maxar
Figure 4. Locations at which PurpleAir sensors were variously positioned during the
testing.
-------
CO and CO2 emission factors are shown in Figure 5 and Table 2 as a function of the total modified
combustion efficiency, MCEt. MCEtis the ratio of the CO2 divided by the CO, CO2, and particulate
carbon, hence is based on both gas phase and solid phase carbon measurements, the latter derived
from analysis of the captured particles' carbon content. Typical trends are exhibited: the CO
emission factor decreases as MCEt improves (gets higher) and the opposite is observed for CO2-
the CO2 emission factor increases as the burn gets a higher MCEt. While the burns were replicates,
MCEt, ranged 9% under thosesimilar burn conditions.
70
60
50
« 40
an
O
O
30
20
10
A
> *****
*>¦
ii""s
<
~
*
#
3080
3070
3060
3050
3040
3030
3020
3010
aj
*4
ao
ao
rs
O
U
0.820 0.840 0.860 0.880 0.900 0.920 0.940
MCEt
CO g/kg fuel
C02 g/kg fuel
Linear (CO g/kg fuel)
Linear (C02 g/kg fuel)
Figure 5. Comparison of CO and CO2 emission factors as compared to the total modified combustion
efficiency. Some burns had two filter samples, resulting in more data points than burns
Table 2. CO, CO 2, and MCEt values for the seven burns.
CO2
CO
MCEt
g/kg fuel
g/kg fuel
fraction
AVG
3044
47
0.88
SD
18
11
0.031
RSD
0.58%
24.30%
3.55%
5
-------
The PM2.5 emission factors are shown in Table 3 and Figure 6. The PM2.5 emission factor decreases
with increases in combustion efficiency, MCEt. Note that some burns had more than one PM2.5 filter
collected, resulting in more data points than the number of ISBs. A 3-fold range of PM2.5 emission
factors are observed despite similar oil burn conditions.
350.0
300.0
250.0
= 200.0
O
J*
"52 iso.o
in
c*j
| 100.0
50.0
0.0
0.830 0.840 0.850 0.860 0.870 0.880 0.890 0.900 0.910 0.920 0.930
MCEt
Figure 6. A comparison of the PM2.5 emission factors as a function of total modified
combustion efficiency.
Table 3. PM2.5 emission factors and MCEt values.
EF PM
MCEt
g/kg oil
fraction
AVG
154
0.88
SD
70
0.031
RSD
45.63%
3.55%
***..
r2 = o
.5078
«
' ' .
%
-------
Figure 6 is repeated in terms of near source and far source UAS/Kolibri measurements in Figure 7.
Downwind UAS transects of the plume returned similar emission factors as those UAS flights in the
thick part of the plume, indicating the versatility of the UAS location - the emission factors are
spatially consistent no matter where the measurements made in the plume.
350.0
300.0
250.0
0
txo
200.0
J*.
ao
1^
150.0
5
CL
100.0
50.0
0.0
PM
2.5
~
A
~
A
~
R2 = 0.51
0.820 0.840 0.860 0.880 0.900 0.920 0.940
MCEt
Downwind ~ Plume
Linear (All Burns)
Figure 7. Comparison ofPM2.s emission factors collected in the near source plume and the far
source (downwind) plume.
Comparison of PM2.5 emission factors with previous USCGRDC results at Little Sand Island,
Mobile Bay, AL [1] are shown in Figure 8. Emission factors are similar but the MCEt values are
higher. The Little Sand Island experiments were done with different oil sources and, perhaps more
importantly, different combustion conditions. In those tests, fresh oil was continuously pumped into
the fire, likely promoting a higher combustion efficiency.
7
-------
350.0
300.0
250.0
^ 200.0
60
-S£
M 150.0
in
rj
I 100.0
50.0
0.0
<
~
R2 = C
.5078
*
m
*
*
.
«
ANS 2021
Crude oil arid RMG 380
Linear (ANS 2021)
0.820 0.840 0.860 0.880 0.900 0.920 0.940 0.960 0.980 1.000
MCEt
Figure 8. Comparison of PM2.5 emission factors from this work with previous determinations.
A limited number of non-methane hydrocarbon (NMHC) measurements were made.Figure 8 shows
trends similar to CO2: decreasing NMHC emission factors with increasing MCEt.
-------
Consistent with previous observations [5] there is no apparent relationship between combustion
efficiency (MCEt) and the amount of oil consumed (Figure 9). However more tests should be
conducted to verify this finding. Particular attention should be placed on the recovery of the oily
residue, as this procedure is difficult and may lead to miscalculation of the amount of oil burned.
"D
(D
E
3
W
C
o
u
o
3?
97.00%
96.00%
95.00%
94.00%
93.00%
92.00%
91.00%
90.00%
89.00%
r2 =
0.0042
<
~
*
0.S30 0.840 0.350 0.360 0.370 0.380 0.390 0.900
MCEt
Figure 10. A comparison of the amount of oil consumed by the fire and the total modified combustion
efficiency.
The real time PM2.5 instruments used in this work use optical measurements to calculate particle
size distribution by mass. Since optical measurements depend on the characteristics of the particles
in question, calibration of the instruments to an integrated filter mass measurement is required. In
this work, the UAS/Kolibri held both the optical SidePak as well as a filter-based, PM2.5 particle
measurement. The filter mass was a primary standard whose value was then used to correct, or
calibrate, the corresponding SidePak optical data. The filter-mass-corrected SidePak data were then
compared to the EPA DRX DustTrak results in Figure 11. This figure reports 2-min average
concentrations thatcorrespond to the time periods for data reported by the PurpleAir sensors. The
PurpleAir sensors report 2-min averaged concentrations so for each burn of about 10 minutes
duration there are five datapoints of the same color. Where color points are lacking indicates that
the burn plume didn't impinge the sensors or too clustered to distinguish.
These results indicate that the uncorrected EPA DRX DustTrak reported PM2.5 concentrations that
were potentially a factor of three lower than the filter-corrected SidePak values. This underscores
the need to calibrate the instruments with appropriatecorrection factors particular to the particle
-------
optical properties.
10 x10
E
D)
CL
X
ac
a
2
LU
10 x10
FILTER-CORRECTED SIDEPAK PM2 5 (|jg m )
Figure 11. Comparison of uncorrected EPA DRX DustTrakwith filter-corrected SidePcik. Color
indicates ISB number. Data are 2-min averaged. The red line is a linear fit line. Linear regression
coefficients are given as a and b where y = bx + a.
The EPA DRX DustTrak was not alone in its underreporting of PM2.5 concentration values - the
uncorrected SMART DustTrak agreed well with its counterpart, Figure 12. Both instruments under-
reported actual concentrations by about 3-fold (Figure 1 l).The uncorrected PurpleAir data were
even further off, approximately 8-fold lower (Figure 13).
10
-------
E
CD
X
en
Q
£
LU
4000 n
3000
2000-
1000-
Coefficient values ± one standard deviation
a = 9.49 ± 39
b =1.09 ±0.05
RZ =0.89
Spearman R = 0.92 (rank correlated)
| Site #1 "Super Site"|
4000
SMART DRX PM2 5 (|jg m" )
Figure 12. Comparison of uncorrected EPA DRX DustTrak with uncorrected USCG SMART DRX
DustTrak. Color indicates ISB number. Data are 2-min averaged. Linear regression coefficients are
given as a and b where y = hx + a.
10x10 -
5
£L
a;
<
LU
_j
a.
QC
3
CL
6
4
2-
f
5
,-2
Coefficient values ± one standard deviation
a =11.045 ±5.2
b =0.12 ±0.003
R2 =0.93
Spearman R = 0.25 (rank correlated)
1:1 line
Linear fit
Site#1 "Super Site°|
10x10
FILTER-CORRECTED SIDEPAK PM2 5 (pg m )
Figure 13. Uncorrected PurpleAir sensor versus filter-corrected SidePak. Color indicates ISB
number. Data are 2-min averaged. Linear regression coefficients are given as a and b where y
+ a.
bx
11
-------
Correction of the EPA DRX DustTrak data with the filter-corrected SidePak data shows good
agreement (R2=0.92), Figure 14. Similarly, correction of the PurpleAir data with thefilter-corrected
SidePak data (Figure 15) are encouraging, with R2 = 0.93 and a slope approaching 1 (b = 0.9).
FILTER-CORRECTED SIDEPAK PM2 5 (jjg m"3)
Figure 14. Corrected EPA DRX DustTrak versus filter-corrected SidePak. Color indicates ISB number.
Data are 2-min averaged. Linear regression coefficients are given as a and b where y = bx + a.
E
CO
Q_
<
LU
_l
CL
cr
=)
CL
Q
LU
I
0
LU
01
o:
o
o
10x10
Coefficient values ± one standard deviation
a = 64.5 ± 38.8
b = 0.908 ± 0.02
R2 =0.93
Spearman R = 0.25 (rank correlated)
10x10
FILTER-CORRECTED SIDEPAK PM2 5 (pg m 3)
Figure 15. Corrected PurpleAir results versus filter-corrected SidePak. Color indicates ISB
number. Data are 2-min averaged. Linear regression coefficients are given as a and b where y
bx
a.
12
-------
4 CONCLUSIONS
Emission factors from ISB were determined using a multicopter UAS with the Kolibri
sensor/sampler system. Emission factors were similar to previous freshwater burns, albeit
with different oil sources and combustion conditions. The remotely piloted UAS was easily
positioned within the plume for sampling emissions. Both downwind and near source
sample collection had no discernible effect on the emission factor values indicating that the
carbon balance method of determining emission factors determined like values regardless
of being in a dilute or concentrated portion of the plume.
Deployment of an array of ground-based samplers accompanied the UAS/Kolibri
measurements. A UAS-mounted optical particle sampler was corrected using a co-located
PM2.5 impactor whose mass value provided a correction factor, specific to the optical
properties of ISB smoke. When this correction was applied to the ground-based EPA DRX
DustTrak and EPA PurpleAir sensors and compared to the primary standard, there was
good inter-agreement (R2 > 0.9). However, the raw, uncorrected data from the DustTraks
were three times lower than the actual concentrations, emphasizing the criticality of
calibrating the optical measurements with the actual source particulate matter mass
(collected using an integrated filter capture method).
Up to eight PurpleAir PM2.5 sensors were used, three of which were deployed in common
locations with the DustTraks and up to five others were positioned at other upwind and
downwind locations. The uncorrected PurpleAir sensors under-reported concentrations
compared to the primary standard even more than the DustTraks, by about a factor of eight.
However their results, like the DRX DustTrak, when calibrated with the appropriate factor,
had an R2 >0.9 when compared with the filter-corrected SidePak (the primary standard for
this test program).
The mobility of the UAS/Kolibri system assures data collection without concern for wind
direction and ground-based sensor placement. The spatial freedom allows the operators to
sample upwind from the source, presenting considerable personnel safety advantages over
ground-based system placement. The responsiveness of the UAS also allows adjustments to
be made for wind shifts as well as for thermally lofted plumes which may later cool and
descend. UAS systems, however, do have wind speed and precipitation limits on their
flights.
Deployment of ground-based systems requires astute judgement regarding downwind
placement of sensors. Arrayed sensors, numbers of which are promoted by lower cost, have
a greater chance of capturing ISB smoke plumes. These systems can also be networked
together, even working in concert with UAS, to provide real time data over a wide area for
the on-scene coordinator. These real time data would prove a more accuratepicture for
health hazards and plume direction during emergency responses.
13
-------
5 ACKNOWLEDGEMENT
This experiment was in response to an identified need from USCG District Nine (D9) and the
Coast Guard Office of Marine Environmental Response. The USCG Research and Development
Center (RDC) partnered with the Environmental Protection Agency (EPA) Office of Research
and Development (ORD) and the United States Army Corp of Engineers (USACE) Engineering
Research and Development Center's (ERDC) Cold Regions Research and Engineering
Laboratory (CRREL) to carry out In-situ burning (ISB) tests with synoptic measurements to
characterize freshwater ISB plumes. RDC also coordinated with the National Strike Force (NSF)
to use their ground-based monitoring equipment while EPA used a gas monitoring system
mounted on a small, unmanned aircraft system (sUAS) to characterize the smoke plume during
each In-situ burning (ISB) test. The measurement of air quality dynamics through remote air
monitoring was identified by the team as a research priority.
6 REFERENCES
1. Gullett, B., Aurell, J. 2020. Emission Characterization of Large Scale in-situ oil
Burns of Persistent Crude and Heavy Oil in Fresh-water with and without
Marsh-land Vegetation for U.S. Coast Guard Research and Development
Center.
2. Gullett, B., Aurell, J. 2022. Alaska In Situ Burn Plume Modeling Project.
QAPP ID: J-IO-0033671-QP-1-0
3. South Coast AQMD. http://www.aqmd.gov/aq-spec/evaluations/summary-pm.
4. U. S. EPA OTM-48. Emission Factor Determination by the Carbon Balance
Method. 2022. https://www.epa.gov/emc/emc-other-test-methods Accessed
October 11, 2022.
5. Aurell, J., et al. Analysis of emissions and residue from methods to improve
efficiency ofat-sea, in situ oil spill burns, Marine Poll. Bull. 173 (2021),
113016, https://doi.Org/10.1016/i.marpolbul.2021.l 13016.
APPENDICES
Appendix 1. Quality Assurance Project Plan: "Freshwater In-Situ Burn Air
Monitoring:USCG," J-IO-0033370-QP-1-0, October, 2021.
14
-------
&EPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
Recycled/Recyclable Printed on paper that contains a minimum of
50% postconsumer fiber content processed chlorine free
------- |