Analytical Support Group 715 Grantham Lane

New Castle, DE 19720 302-323-4542

Delaware Department of Natural Resources and Environmental Control Division of Air Quality

Marine Vessel Impact on Ambient Air Quality, Delaware City, Delaware Final Report

March 14, 2016

Page 1 of 14

Project Title: Marine Vessel Impact on Ambient Air Quality, Delaware City, Delaware Submitted by: Keith Hoffman – Environmental Scientist I Category: Short-term Point-Source Investigative Deployment Background: The citizens of Historic Delaware City have voiced concerns regarding the marine vessels that deliver crude oil to the Delaware City Refinery (DCR) and possible air toxic emissions from that process. There is also concern regarding the exhaust from these vessels while docked. It has been noted that at times these ships emit black smoke from their stacks when docked for unloading crude oil. The community is concerned their health is negatively impacted by this activity. This study investigated any identifiable contributions for certain pollutants specified below from these vessels to the local air quality of Delaware City.

Black Smoke from docked ship Ship docked at DCR offloading dock The Moveable Monitoring Platform (MMP) was sited near the DCR off-loading docks on the Delaware River. This location is adjacent to the Delaware City Refinery fence line and Washington Street. This site was also located near tourist attractions like Fort DuPont. This location also provided important exposure for the MMP to educate the public regarding air quality in general and the Air Quality Index (AQI). The aerial photo below depicts the location for the MMP in relationship to the off-loading docks:

Page 2 of 14

Location of MMP within Delaware City and location of DCR Docks

The monitoring project focused on the following parameters: • Continuous PM2.5 • Black Carbon • BTEX (benzene, ethyl benzene, toluene, m-p-o xylenes) • NO, NO2 & NOx • Ozone • SO2

• Wind Speed / Wind Direction Objectives: The study investigated whether the local ambient air concentrations of certain pollutants are impacted by marine vessels delivering crude oil. The marine vessels operate on diesel engines to provide power while at port. While these vessels run their auxiliary engines for power they have been observed to show visible signs of emissions. This project sought to determine if there was any impact from these marine vessels. Pollutants associated with diesel emissions include particulate matter and black carbon. Air toxic emissions from the crude oil off-loading may contain the BTEX compounds as a fingerprint. Additional monitoring for ozone, sulfur dioxide, and oxides of nitrogen were conducted for comparison to nearby sites. Implementation: Site Selection: A site close to the off-loading operation in Delaware City was selected. Space was available near the refinery fence line in a municipal parking lot. This space was selected due to the availability for long- term parking and the availability of electricity to power the MMP.

Delaware City Refinery Docks

Page 3 of 14

Compass quadrant views from top of MMP

Monitoring: Continuous gas analyzers were used to monitor for NO-NO2-NOx, SO2 & Ozone. Continuous Particle monitors were used for PM2.5 and Black Carbon. The Black Carbon Analyzer provides the contribution of “Diesel Emissions” to the PM2.5 concentrations detected. Wind Speed and Wind Direction were measured using an ultrasonic transducer. Data were collected at 1, 5 and 60 minute averages and reported to the Central Data Acquisition System. An automated GC was used to analyze the BTEX compounds at a 15 minute interval.

Data Analysis: All data was stored in the MMP data logger with wireless transmission to the Central Data Acquisition System. Data was analyzed weekly to determine if there were any instances when local ambient air quality is above what other monitoring stations were reporting. In particular, the Delaware City/Route 9 monitoring station includes continuous PM2.5 and SO2 monitors but is located in a different downwind direction from the loading dock which was helpful in identifying impacts on those pollutants related to loading dock activities. The urban NCore site at MLK Blvd in Wilmington, DE was also used for comparison particularly for Black Carbon, PM2.5, NO-NO2-NOx, and Ozone.

Quality Assurance: All monitors were operated as per approved SOPs and manufacturer specifications. Quality control and assessment procedures followed the project Quality Assurance Project Plan, and included at a minimum multi-point calibrations on-site at the beginning of the study period and weekly quality control checks on all gaseous analyzers. The continuous PM2.5 and Black Carbon monitors had monthly quality control checks performed as per current SOP. The auto GC used the auto calibration check function to verify data accuracy for the BTEX compounds. The EPA established method quality objectives (MQOs) for all FEM analyzers and monitors were followed.

Page 4 of 14

Timeline: Week of July 6, 2015 – MMP deployed to monitoring site August, 2015 – All Monitors Calibrated August, 2015 - January, 2016 – Data collected and checks performed in accordance with Project QAPP February, 2016 – Project shut down, MMP returned to Grantham Lane, and data analysis performed. Results: Data Summary Data capture for the study period from August 2015 through January 2016 was well above 60% for most monitors. Issues with the computer used for data acquisition and maintenance issues with the ozone and NOx monitors account for lower data capture.

Parameter Wind

Speed V Wind Dir V PM2.5 BC 880 Ozone NO2 NOX

SO2 Trace Level

Unit Measured mph Deg µg/m3(L) µg/m3(S) ppm ppb ppb ppb

Average Value 3.4 290 6.61 0.583 0.02 5.7 7.1 0.5

% Data Capture 85 86 79 89 65 76 77 83

Data summary table for particulate monitors, meteorological data, and gas analyzers

Pollutant Directionality (Wind Speed/Direction)

Wind speed and direction data were collected by the MMP and compared to the Rt9 site located approximately ¾ of a mile away to the West and the MLK site in Wilmington, DE approximately 11.2 miles to the North. During the study period the West North Westerly direction occurred 57% of the time. The Rt9 site had a more South Westerly frequent directionality. The MLK site had a similar frequency profile with slightly more frequent in the North Westerly directionality. The DCR docks are located directly North of the MMP.

MMP, MLK, and Rt9 Wind Roses

Page 5 of 14

Wind direction data for 5 out of 8 of the directions during the study period was less than 10% per direction. With 75% of the data coming from three directions, there was insufficient data to determine significant pollutant source directionality.

Frequency of Wind Direction for the MMP

Map of Delaware City Refinery overlaid with Wind Direction Frequencies from the MMP location

PM2.5 PM2.5 data was collected via continuous SHARP monitor for the duration of the study and compared to SHARPs at the MLK and Rt9 Monitoring Stations. During this period the MLK site was shut down for about a month for station upgrades and the SHARP at Rt9 went offline at the beginning of November due to maintenance issues. When daily averages are compared, the MMP data correlates well with both Rt9 (R2 = 0.89) and MLK (R2 = 0.84) data. Results suggest the PM2.5 data collected tracks with regional variation. No exceedances of the daily National Ambient Air Quality Standards (NAAQS) were observed.

170 117

694

123 280 271

1363

819

0

200

400

600

800

1000

1200

1400

1600

Freq

uenc

y

Wind Direction

MMPWind DirectionFrequency

DCR Docks

MMP

DCR

Rt9 Site

Page 6 of 14

Daily average PM2.5 concentrations compared between the MMP, Rt9 and MLK Monitoring Stations

Correlation between daily PM 2.5 average concentrations at the MMP, MLK, and Rt9

Black Carbon

Aethalometer data on black carbon from the MMP was compared to the MLK aethalometer and found to not be very well correlated (R2 = 0.31). No further MLK data was available due to station upgrade and maintenance issues with the MLK aethalometer. Adjustments were made to the aethalometer in December and January to address a similar maintenance issue being investigated with the MLK unit. Due to the adjustments during the study period, no significant conclusions could be reached relating to correlation with other pollutants. When marine vessels were observed either docking, docked, or pulling away they were noted in the MMP E-log. Marine vessel activity was highlighted for approximately 24 hour periods around the time activity was observed. No spikes were observed during vessel activity periods and no vessel activity was recorded for January. Diurnal hourly averages were calculated for the study period in the Central Data Acquisition System and indicate a typical mobile source pattern with a sharp morning peak and broad evening peak.

PM2.5 Daily NAAQS (35 µg/m3)

0

5

10

15

20

25

30

35

40

PM2.

5 (µg

/m3 (

L))

Daily Average Concentration

DailyNAAQS(ug/m3)MMP

Rt9

MLK

y = 1.2807x - 0.0413 R² = 0.893

y = 1.3363x + 1.2844 R² = 0.8351

0

5

10

15

20

25

30

35

0 10 20 30

MLK

/Rt9

PM

2.5 (µg

/m3 (

L))

MMP PM2.5 (µg/m3(L))

Rt9

MLK

Page 7 of 14

Aethalometer data correlation between the MMP and MLK

Observed marine vessel activities compared to hourly average concentrations (no observed marine vessel activity in January)

Diurnal patterns for MMP BC880

Continuous Gas Analyzers Ozone:

Ozone data was compared with the design value site for ozone at Brandywine Creek State Park and additionally Lums Pond State Park for the study period and found to be well correlated (R2 = 0.71 and 0.81 respectively). No ozone exceedances for the 8 hour NAAQS were detected for the study period and data are consistent with regional ozone patterns.

y = 0.8221x + 0.4048 R² = 0.3121

0

1

2

3

4

5

6

7

8

0 1 2 3 4 5

MLK

(µg/

m3 (

S))

MMP (µg/m3(S))

MMP BC 880 Comparison

BC 880

0123456789

10

BC 8

80 m

g/m

3 (S

)

Hourly Average Concentrations

MarineVesselActivity

MMP BC880

Page 8 of 14

MMP 8hr. average ozone concentrations compared between Brandywine Creek State Park and Lums Pond State Park

MMP 8hr. ozone concentration correlation between Brandywine Creek State Park and Lums Pond State Park

SO2: SO2 data was compared with MLK and Rt9 sites for the study period. Levels of SO2 were well below the 1-hour NAAQs. There was insufficient wind direction data from most directions to provide a statistically significant comparison of directionality.

Ozone 8hr NAAQS (0.075 ppm)

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08O

zone

(ppm

)

8hr. Average Concentration

8hr.NAAQS(ppm)

MMP

LPSP

BCSP

y = 0.8961x + 0.0002 R² = 0.8073

y = 0.8438x + 0.0087 R² = 0.7086 0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

LPSP

/BCS

P 8h

r Ozo

ne C

once

ntra

tion

(ppm

)

MMP 8hr Ozone Concentration (ppm)

LPSP

BCSP

Linear (LPSP)

Linear (BCSP)

Page 9 of 14

SOs hourly average concentrations compared to MLK and Rt9

NO/NO2/NOx:

NO/NO2/NOx data was compared to MLK with an average of 7ppb NOx at the MMP and 22ppb NOx at MLK the three fold difference reflects the urban contribution to the MLK site. NO2 levels for Delaware City were well below the 1-hour NAAQS. There is no NAAQS for NOx, however when plotted against observed marine vessel activities for an approximate 24hour period the corresponding spikes in NOx are not consistent in either concentration or duration. While there is insufficient wind data to draw statistically significant conclusions regarding source direction, when NOx data is averaged by wind direction the highest average concentrations trended from a North Easterly direction. This direction does not appear to include the docks and may be the result of mobile or point sources further away in the North Easterly direction.

Average hourly concentrations of NO2

SO2 Hourly NAAQS (75 ppb)

0

10

20

30

40

50

60

70

80

SO2 C

once

ntra

tion

(ppb

)

Average Hourly Concentration

1 Hr.NAAQS(ppb)

MMP

Rt9

MLK

NO2 Hourly NAAQS (100 ppb)

0

20

40

60

80

100

120

NO

2 Con

cent

ratio

n (p

pb)

Average Hourly Concentration

1 Hr.NAAQS(ppb)

NO2(ppb)

Page 10 of 14

Average Hourly NOx at the MMP and MLK

Average hourly concentrations of NOx compared to observed marine vessel activity (no activity observed in January)

Average NOx concentrations compared to wind direction

0

50

100

150

200

NO

x Con

cent

ratio

n (p

pb)

Average Hourly Concentration

MMP

MLK

020406080

100120140

NO

x Con

cent

ratio

n (p

pb)

Average Hourly Concentation

Marine VesselObserved Activity

NOx(ppb)

0.00

5.00

10.00

15.000

20

40

60

80

100

120

140

160180

200

220

240

260

280

300

320

340

Avg NOx Concentration (ppb) by Wind Direction

Page 11 of 14

Overlay of average NOx concentrations compared to wind direction for Delaware City

Overlay of average NOx concentrations compared to wind direction for the region

0.00

5.00

10.00

15.000

20

40

60

80

100

120

140

160180

200

220

240

260

280

300

320

340

Avg NOx Concentration (ppb) by Wind Direction

0.002.004.006.008.00

10.0012.0014.0016.00

020

40

60

80

100

120

140

160180

200

220

240

260

280

300

320

340

Avg NOx Concentration (ppb) by Wind Direction

Page 12 of 14

Automated Gas Chromatograph This study represents the first field deployment of the Synspec GC-955 automated gas chromatograph (Auto GC) configured for detection of benzene, toluene, ethylbenzene, and xylenes. Initial efforts were focused on effective calibration, quality control procedures and documentation; some technical problems required assistance from the manufacturer to resolve. The result is that verifiable valid hourly data were not collected until mid-November 2015. Gaps in the data reflect time periods when the sampler was not operational. Data capture for the entire study period from August 2015 through January 2016 for most compounds was 42%. Results for O-Xylene showed poor quality control checks throughout the study, and all data were ruled invalid.

Parameter Benzene Toluene Ethyl

Benzene M,P-Xylene O-Xylene

Unit Measured ppb ppb ppb ppb ppb

Average Value 0.638 0.352 0.023 0.253 Not applicable*

% Data Capture 42 42 42 42 0* Data summary table for Auto GC

*Data ruled invalid after final review

Auto GC Hourly averages in ppb

All concentrations for all parameters were below 5.0 ppb, and were not correlated with marine vessel docking events. Problems with o-xylene calibrations resulted in data invalidation for most of the study period, but other compounds were validated as indicated in the previous graph. While benzene is a component of diesel exhaust, it is also emitted from the Delaware City Refinery Complex as well as gasoline engines, so it is not a direct indicator of diesel emission impact. There is fairly good correlation (R2 = 0.59) between benzene and toluene, which is typical for VOC ambient air measurements since both compounds generally have similar sources. Since there are no other continuous VOC monitors operating at any other sites, no correlation could be done with other locations.

0.01.02.03.04.05.06.07.08.09.0

10.0

(ppb

)

Benzene Toluene Ethyl Benzene M-P Xylene

Page 13 of 14

Correlation between Toluene and Benzene data

To determine association with other parameters and possibly similar sources, correlations were evaluated between Benzene and NOx, Benzene and PM2.5, and Benzene and Black Carbon monitored during the same time period. There was no significant correlation between Benzene and NOx or PM2.5, but Black Carbon was more complex, with some time periods appearing to correlate well while others did not. Due to operational problems and multiple adjustments made to the Black Carbon aethalometer during the study period, no significant conclusions could be reached relating to correlation with Benzene. There was insufficient data to evaluate any directionality with the Benzene concentrations. Because of the relatively short time period with valid data, 57% of the hourly wind directions were from the West-Northwest. This means there were insufficient hours with wind from other directions to generate a statistically significant comparison.

Discussions: The objective of this study was to determine if marine vessel activities at the DCR docks may be impacting the residents of Delaware City. In order to facilitate the investigation a shipping schedule was requested from the DCR, which was denied. Recorded observations of marine vessel activity were made when the site was visited.

There were several maintenance issues which lead to loss of data during the study. A month of data for ozone was lost due to analyzer failure and repairs. Periodic computer failures lead to gaps of several days in the data until a technician could reboot the computer. This was the first deployment for the Auto GC which had to be sent back to the manufacturer after problems with calibration were encountered. Further calibration issues resulted in a re-configuration of the analyzer in the MMP so that the calibration gas cylinder could be closer to the machine in order to avoid the gas line acting as an unintended GC column. Finally, the NOx analyzer failed to pass its checks towards the middle of January and was taken offline.

Despite these issues we were able to obtain validated data from mid-August though the end of January for most parameters. A Northerly wind was the direction of interest for our study. The wind most frequently occurred from the West-North-Westerly direction and very infrequently from the direction of interest for the duration of the study. The wind pattern during the study period is very similar to the MLK site 11.2 miles to the North, while the nearby Rt9 site had wind most frequently from the Southwesterly direction. Daily average PM2.5 concentrations were found to correlate well with data from the MLK and Rt9 monitoring sites and average

y = 1.6325x - 0.0801 R² = 0.5933

0.00.51.01.52.02.53.03.54.04.5

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Tolu

ene

(ppb

)

Benzene (ppb)

Page 14 of 14

concentrations compared to wind direction indicated multiple sources though not in the direction of interest. Black carbon which serves as an indicator for diesel emissions did not appear to show elevated response when marine vessel activity was observed. The diurnal patterns for black carbon were found to be consistent with regional patterns indicating mobile sources. Data for NOx, another diesel indicator, did not correlate consistently with observed marine vessel activities. However NOx did show a trend from a North Easterly directionality inconclusively suggesting more distant impacts. The auto GC data indicated all parameters were below 5.0 ppb and not correlated with observed marine vessel activity. The NO2, SO2, and Ozone data were all well below NAAQS levels and compare well to regional monitoring data.

Conclusions:

Data collected for the period of study is consistently representative of regional pollution patterns and no significant contributions from the direction of the DCR docks were observed. Wind direction for the duration of the study was infrequent from the direction of interest (North) for the DCR docks. There was insufficient wind data from additional directions to make statistically significant conclusions regarding pollutant source directionality. PM2.5 concentrations tracked well with nearby monitors suggesting a regional influence. The average concentrations for PM2.5 and black carbon showed poor directionality when compared to wind direction, reinforcing the suggestion of a regional influence. No NAAQS exceedances were reported for any of the continuous gas analyzers and they tracked well with permanent monitors in the region. The NOx concentrations which are considered a good indicator of diesel emissions were found to trend directionally, though not in the direction of interest and there is insufficient wind data to provide a statistically significant conclusion as to pollutant source direction. The continuous GC data for benzene, toluene, ethylbenzene, m-xylene, p-xylene and o-xylene did not indicate increased concentrations during marine vessel docking events.

References: 1. EPA Proposal to Designate an Emission Control Area for Nitrogen Oxides, Sulfur Oxides and Particulate Matter. EPA-420-R-09-007 April 2009 http://www.epa.gov/nonroad/marine/ci/420r09007.pdf 2. Air Emissions from Marine Vessels, Maine Department of Environmental Protection Bureau of Air Quality, January 15, 2005 http://www.maine.gov/dep/water/wd/vessel/airemissionsreport.pdf