MEMORANDUM

To:	Patricia Rowley, Lucie Audette, and Ken Davidson, USEPA

From:	Arlene Rosenbaum, Seth Hartley, Mark Bethony, Jeffrey Hoye

Date:	September 28, 2007

Re:	Estimation of diesel particulate matter concentration isopleths for
marine harbor areas and rail yards (revised)



	

BACKGROUND

At the present time, there is evidence that people living in close
proximity to major transportation sources like roads experience higher
exposure to pollutants that are directly emitted by motor vehicles. At
least one study suggests that people living near major rail terminals
are also exposed to significantly higher concentrations of directly
emitted pollutants. Likewise, another study suggests that nearby
residents of marine ports are exposed to significantly higher
concentrations of pollution, including particulate matter.

The purpose of this work assignment is to collect population data that
EPA will need to evaluate potential impacts of higher pollution
exposures on populations living near large sources of diesel exhaust,
i.e., locomotives and Class 1, 2, and 3 marine engines. The major areas
for study include (1) public and privately owned marine ports, and (2)
rail terminals or rail yards.  This study is a screening level analysis,
and as such, is an inexact tool and not appropriate for regulatory
decision-making; it is used for illustrative purposes only.

INTRODUCTION

Summarized below is the methodology used to identify and digitize diesel
particulate matter (DPM) concentration isopleths surrounding 47 US
harbor areas and 37 US rail yards. The isopleths were selected to
correspond to estimated DPM concentrations of interest. They were
estimated with air dispersion modeling for marine harbor areas, and a
process for scaling from published rail yard modeling reports for rail
yards. 

These isopleths will be used to assess the size and characteristics of
populations that are potentially exposed to various levels of DPM
concentrations resulting from emissions at the selected facilities.

This memorandum describes the technical approach used to digitize
facility boundaries and to estimate the size and shape of the
concentration isopleths.

DIGITIZATION

All footprint and buffer spatial data sets for marine harbor areas, and
rail yards are contained in personal geodatabase (.mdb) format,
including the Federal Geographic Data Committee (FGDC) metadata and the
geospatial models used to create the buffers.

Prior to calculating buffer distances for each marine harbor area and
rail yard, the facility feature data set was reprojected from geographic
coordinates (WGS 1984 datum) to Albers Equal Area Conic (in order to
obtain rectangular linear units) using the following spatial parameters:

Coordinate system:	Albers Equal Area Conic

Datum:			NAD 1983

Central Meridian:	96º West

Standard Parallel 1:	29.5º North

Standard Parallel 2:	45.5º North

Latitude of Origin:	39.5º North

Linear Unit:		U.S. Mile

ISOPLETH DISTANCES

μg/m3 (0.4 μg/m3 when converted to equivalent environmental exposure
over a 70-year lifetime) resulted in increased cancer risk.

In addition, scientific studies show ambient PM, of which DPM is an
important component, is associated with a series of adverse health
effects. These health effects are discussed in detail in the 2004 EPA
Particulate Matter Air Quality Criteria Document (PM AQCD) for PM, and
the 2005 PM Staff Paper.,,  Health effects associated with short-term
exposures (hours to days) to ambient PM include premature mortality,
increased hospital admissions, heart and lung 

diseases, increased cough, adverse lower-respiratory symptoms,
decrements in lung function and changes in heart rate rhythm and other
cardiac effects.  Studies examining populations exposed to different
levels of air pollution over a number of years, including the Harvard
Six Cities Study and the American Cancer Society Study, show
associations between long-term exposure to ambient  PM2.5 and both total
and cardio respiratory mortality.  Recent local impact studies have also
documented the health effects due to PM exposures measures on or near
roadways.  Also, a number of studies have shown associations between
residential or school outdoor concentrations of constituents of fine
particles found in vehicle exhaust and adverse respiratory outcomes such
as asthma who live near roadways.  Finally, the State of California,  in
the past two years have completed a number of studies of multiple rail
yards and marine ports finding that emissions from these facilities
contributed significantly to elevated ambient concentrations near these
sources leading to a substantial number of people being exposed to
diesel engine emissions. 

EPA believes that areas of the U.S. that have relatively higher annual
exposure levels for diesel exhaust, certainly those areas with
substantial portions of their populations exposed to annual average
exposure concentrations above 2.0 µg/m3, the lower end of the range of
occupational exposures where increased cancer risk was found, should be
aware of EPA’s scientific findings regarding elevated risks. 
Furthermore, while considering the important limitations of the science
at this time, they should use this information to compare air toxic
risks and set priorities for their programs. For the U. S. overall, the
90th percentile of mean census tract concentrations is about 1.9 ug/m3. 


Given the preliminary nature of EPA’s recommendations about DPM
concentrations exceeding 2.0 µg/m3, this analysis provides information
about the populations exposed to concentration levels above urban
background of both:

2.0 µg/m3, and

0.2 µg/m3.

MARINE HARBOR AREAS

The 47 marine harbor areas identified by EPA for this study are listed
in Table 1.

Table 1. Marine Harbor areas 

Baltimore, MD

Boston, MA

Charleston, SC

Chicago, IL

Cincinnati, OH

Cleveland, OH

Corpus Christi, TX

Detroit, MI

Duluth-Superior, MN

Freeport, TX

Gary, IN

Helena, AR

Houston, TX

Jacksonville, FL

Lake Charles, LA

Long Beach, CA

Los Angeles, CA

Louisville, KY

Miami, FL

Mobile, AL

Mount Vernon, IN

Nashville, TN

New Orleans, LA

New York, NY

Norfolk Harbor, VA

Oakland, CA

Panama City, FL

Paulsboro, NJ

Philadelphia, PA

Pittsburgh, PA

Port Arthur, TX

Port Everglades, FL

Port of Baton Rouge, LA

Port of Plaquemines, LA

Portland, ME

Portland, OR

Richmond, CA

Savannah, GA

Seattle, WA

South Louisiana, LA

St. Louis, MO

Tacoma, WA

Tampa, FL

Texas City, TX

Tulsa - Port of Catoosa, OK

Two Harbors, MN

Wilmington, NC

 

Facility Footprints

Using a combination of USDA Data Gateway (  HYPERLINK
"http://datagateway.nrcs.usda.gov"  http://datagateway.nrcs.usda.gov )
aerial imagery and U.S. Census Bureau census block boundaries, marine
harbor areas were hand-digitized and subsequently refined to exclude
areas adjacent to the marine harbor area, refineries, chemical plants,
and obviously residential areas. All data were input in geographic
coordinates using the WGS 1984 datum.

Note that the American Association of Port Authorities (  HYPERLINK
"http://www.aapa-ports.org/"  http://www.aapa-ports.org/ ) maps
generally exclude areas of marine harbor area activity that are not
under the jurisdiction of the Port Authority, such as privately-owned
piers. As noted above, these digitized footprints include areas not
owned or operated by Port Authorities.

Generally, the “footprint” specified for each harbor area was
selected to include all areas of marine-related activity for the general
waterway area, and not limited to the jurisdiction of the proximate Port
Authority. This was done to 

include all engines and equipment associated with harbor areas,
regardless of whether the equipment is operated in association with a
Port Authority facility or a nearby private facility, and

insure that the designated area for the air dispersion modeling is
consistent with the activity data on which the emission estimates were
based (see below). 

Pictures of the digitized footprints for the 47 US harbor areas are
presented in Appendix A.

Air Dispersion Modeling

The annual average DPM concentrations were determined by applying the
USEPA’s regulatory air dispersion model, AERMOD (version 07026), to
each of the port complexes with local meteorology and other relevant
parameters, and determining averages of annual concentrations over a
regular grid of receptors. Because of the large number of facilities to
be considered in this analysis, a detailed air quality simulation could
not be performed for each port, but AERMOD was applied to perform a more
limited screening analysis for DPM concentrations from harbor area
activity, as described in this section.

In order to assess some of the uncertainties in the concentration
predictions for this study, a sensitivity test of our approach,
including a comparison with published results of detailed modeling for
the Ports of Los Angeles and Long Beach, is presented in Appendix B. In
addition, a detailed discussion of the meteorological inputs and
sensitivity to those inputs is presented in Appendix C. 

Emission Source Characterization and Location Because of the large
number of facilities and lack of specific information for many of the
ports to be considered in this analysis, a detailed air quality
simulation for each harbor area could not be performed. In particular,
the precise locations of the emission releases that occur in the harbor
area of question could not be determined. Instead, each of the ports was
represented as two or more area sources, with vertices of each source
determined using the digitized “footprint” of that harbor area and
distributed the emissions uniformly (horizontally) throughout the areas.


Each digitized “footprint” for a given harbor area was translated to
specific polygons, from which coordinates of each vertex were extracted.
To incorporate vertical variation in emissions release, two area sources
were simulated with the same horizontal layout, but differing locations
vertically, for each polygon. Ocean-going and deep-draft vessel
emissions were assumed to be released from an elevated area source at
50.0 m above ground level and an initial vertical dimension of 23.0 m.
All other sources of DPM emissions, such as heavy duty vehicles
(trucks), locomotives, cargo handling equipment (CHE), and harbor craft
(H/C), are expected to show 

much less variation in actual release height. Thus, these sources were
combined into a single area source polygon with a release height of 4.4
m, and an initial vertical dimension of 2.1 m.  Table 2 shows these
values, along with the median contribution from each layer to total
emissions over all 47 ports. 

The assumption underlying the area source characterization of emission
releases is that they are equally likely to occur anywhere within the
generalized boundary of the harbor area. While this approach removes any
bias associated with allocation of source locations, it is not likely to
represent the actual emissions sources well. This could be particularly
true for the vessel categories, where the location of their activity is
largely limited to specific areas (often edges) of the harbor complex.
Although in principle this approach could lead to underestimation of
emission densities and the corresponding impacted areas by overstating
the initial horizontal dispersion, the sensitivity tests presented in
Appendix B suggest that, at least for medium sized ports, the size of
the predicted concentration isopleths are not very sensitive to the
source characterization. 

zi (m)	Median Emissions Contribution

Near Surface	4.4

 	2.1

 	34%

 

Trucks, Rail, H/C, CHE



	Ocean-Going Vessels and Hotelling	50	23	63%



Particle Deposition Dry and wet deposition processes were included in
the simulations. Particulate size distributions for DPM can vary
significantly, particularly in the sub-micron range, and are not well
characterized for marine vessels. Hence, the modeling included
AERMOD’s “Method2” logic for deposition. The deposition velocity
for Method 2 is calculated as the weighted average of the deposition
velocity for particles in the fine mode and the deposition velocity for
the coarse mode.  

Values of the fine mass fraction and mass mean diameter were taken from
an exhaustive study ICF performed for EPA on heavy duty diesel vehicle
exhaust size distributions and settling.  Table 3 shows these values.
Although derived for trucks, the size distribution is not likely to be
very different for all diesel PM sources, when aggregated and viewed
only from Method 2’s fine/coarse perspective. These values are also
consistent with AP-42 recommendations
(http://www.epa.gov/ttnchie1/ap42/). These values are not likely to
introduce significant uncertainty in the resulting calculations. 

Table 3. DPM size and settling parameters

Exhaust Emissions	Size Fraction	MMD (m)	Fine Fraction

Heavy heavy-duty diesel vehicles	PM10	0.193	94.5 %



Urban and Rural Specification In AERMOD, all sources default to rural
dispersion. However, an "urban source" option is available, which
accounts for enhanced plume dispersion during nighttime stable hours due
to the urban heat island effect. The magnitude of the urban heat island
effect is a function of the user-specified population for the urban
area. Harbor areas were characterized as urban if a majority of the
regional land use within 3 km of the port could be considered urban and
the adjacent Consolidated Metropolitan Statistical Area (CMSA) had a
population of 500,000 or more for 2006.  Of the 47 marine harbor areas,
25 were simulated as urban and 22 as rural.  Table 4 lists the
classifications and Table 5 shows the CMSA population input to AERMOD
for the areas classified as urban. The current AERMOD Implementation
Guide suggests that using census data for the Metropolitan Statistical
Area (MSA) may be appropriate for relatively isolated urban areas. 
Determination of the appropriate population to input to AERMOD is more
difficult for applications within urban corridors. The populations used
based on the CMSA may overestimate the heat island effect in some cases,
producing somewhat lower concentrations for low-level releases. 
However, the main urban effect in AERMOD is dependent on population to
the 1/4th power, so a factor of 2 differences in population would only
account for about a 20 percent difference in heat island effect.  

As discussed in Appendix C, concerns regarding accurate treatment of the
atmospheric boundary layer and surface roughness at the land-sea
boundary were important considerations in the selection of urban and
rural site characterization. In several cases, the “land use” within
3 km of the application site included a substantial amount of water,
which affects the boundary layer structure and plume dispersion, due to
the influences of lower surface roughness and lower albedo. Had all
“whole urban complexes” been characterized as urban sources, AERMOD
predictions could have differed, depending on the distribution between
low-level and elevated sources and the underlying CMSA population.
However, the assumption of rural dispersion for several areas would tend
to produce more conservative concentration estimates than the urban
option (i.e., more likely to overestimate impacts than to
underestimate).  Also, as discussed more fully below and in Appendix C,
our treatment of surface roughness is likely to lead to more
conservative estimates of DPM impacts. 

Receptor Grid The model receptor grids are defined in a polar system,
centered on one vertex of the port polygon and extending out far enough
to capture a predicted concentration of 0.2 µg/m3. Each total grid was
composed of concentric rings of 7 miles width. Inside each ring group,
were sub-rings of receptors placed every 5 degrees with radial spacing
between 250 and 500 meters. 

To capture details of the concentration profiles near the harbor area,
the innermost ring (0-7 miles) employed spacing between concentric
receptors of 250 m. However, to reduce run time, rings 2 (7-14 miles), 3
(14-21 miles), and beyond employed spacing of 500 m. Receptors within
the polar grid that would fall within the area sources (i.e., the
digitized harbor area footprint) were excluded. 

Figure 1 illustrates the concept of the concentric rings used to create
the receptor grids for harbor areas ranging from relatively simple
(Wilmington, NC, 1 ring) to relatively complicated (South Louisiana, LA,
3 rings). In both cases, additional rings beyond those shown here were
added to capture the tails of the concentration distributions. Figure 1
also illustrates the scales of the modeling domains. 

Each receptor was placed at standard breathing height, 1.8 m. No terrain
effects were included in any dispersion calculations. 

Meteorology Each of the 47 harbor area was simulated for 3 years using
meteorological data processed with the AERMET (version 06341)
preprocessor. For each harbor area, three surface and two upper air
sites were initially selected based on proximity. The data sets that
were relatively complete for the 2004 to 2006 time period and that best
represented conditions for the harbor area were then selected. Table 6
shows each harbor of interest and its associated best-matched surface
and upper air observation site. 

Hourly surface observations were then extracted from the NOAA/NCDC
database Quality Controlled Local Climatological Data and converted to
SAMSON format for processing within AERMET. Upper air data was collected
from the FSL/NCDC Radiosonde Data Archive in the FSL format. All
available data from the 2004 to 2006 period, inclusive, was collected
and processed with the AERMET preprocessor. 

Surface Characteristics Surface characteristics were determined and
applied for modeling based generally on EPA guidance. However, we
deviated from the guidance with respect to the location from which the
surface characteristics of the study area were derived. That is, surface
characteristics were determined based on land use within a 3 km diameter
area around the harbor area, not around the meteorological observation
site. The reasons for following this approach were as follows. 

The areas of interest always occur at a land/sea interface, where sharp
gradients can occur in meteorological parameters, particularly those of
the boundary layer, and, 

onsite meteorological measurements were not available. 

Although this approach introduces some uncertainties into the model
predictions, since the observed wind speed is intricately linked to the
adjacent surface roughness, we believe this is the best method for
characterizing the boundary layer parameters at the harbor, in the
absence of on site meteorological observations.  

Further details of the methodology used to process meteorological data
and its potential impact on DPM concentration predictions are presented
in Appendix C. 



Table 4. Urban/Rural Characterization

Emitting Area Considered Urban 	Baltimore MD

Boston MA 

Cincinnati OH

Cleveland OH

Detroit MI

Houston TX

Long Beach CA

Los Angeles CA

Louisville KY

Miami FL, 

Nashville TN

New Orleans LA

New York NY

Oakland CA

Paulsboro NJ

Philadelphia PA

Pittsburgh PA

Port Everglades FL

Portland ME

Portland OR

Richmond CA

Seattle WA

St. Louis MO

Tacoma WA

Tampa FL

Emitting Area Considered Rural	Chicago IL

Charleston SC

Corpus Christi TX

Duluth-Superior MN 

Freeport TX 

Gary IN 

Helena AR

Jacksonville FL

Lake Charles LA

Mobile AL

Mount Vernon IN 

Norfolk Harbor VA

Panama City FL 

Port Arthur TX

Port of Baton Rouge LA

Port of Plaquemines LA 

Savannah GA

South Louisiana LA 

Texas City TX

Tulsa-Port of Catoosa OK

Two Harbors MN

Wilmington NC





Table 5. CMSA Population for Areas Classified Urban

Harbor Area	2006 CMSA Population

Mobile, AL	573,319

Long Beach, CA	17,775,984

Los Angeles, CA	17,775,984

Oakland, CA	7,228,948

Richmond, CA	7,228,948

Miami, FL	5,463,857

Port Everglades, FL	5,463,857

Tampa, FL	2,697,731

Louisville, KY	1,222,216

Lake Charles, LA	223,734

New Orleans, LA	1,069,428

Boston, MA	7,465,634

Baltimore, MD	8,211,213

Portland, ME	621,219

Detroit, MI	5,410,014

Duluth-Superior, MN	274,244

St. Louis, MO	2,858,549

Paulsboro, NJ	6,382,714

New York, NY	21,976,224

Cincinnati OH	2,147,617

Cleveland, OH	2,917,801

Portland, OR	2,137,565

Philadelphia, PA	6,382,714

Pittsburgh, PA	2,462,571

Nashville, TN	1,533,406

Houston, TX	5,641,077

Port Arthur, TX	379,640

Texas City, TX	5,641,077

Seattle, WA	3,876,211

Tacoma, WA	3,876,211





Figure 1 Source outlines and receptor grids at (a) Wilmington, NC, and
(b) South Louisiana, LA. Spacings are in meters. Concentric rings of
receptor sets are color coded. 





Table 6. Selected surface and upper-air pairs for each harbor area.



Harbor Area	Surface Station	Upper-Air Station

Name	Lat	Long	Station Name	WBAN	Lat	Long	Dist (km)	Station Name	WBAN	Lat
Long	Dist (km)

Mobile AL	30.718	-88.051	 MOBILE AL           	13838	30.633	-88.067	9.5	
SLIDELL                 	53813	30.33	-89.82	174.7

Helena AR	34.409	-90.624	 WEST MEMPHIS AR     	53959	35.133	-90.233	88.1
 N LITTLE ROCK           	3952	34.83	-92.27	158.3

Long Beach CA	33.763	-118.219	 LONG BEACH CA       	23129	33.833
-118.167	9.2	MIRAMAR NAS   	3190	32.87	-117.15	140.1

Los Angeles CA	33.746	-118.266	 LONG BEACH CA       	23129	33.833
-118.167	13.3	MIRAMAR NAS   	3190	32.87	-117.15	142

Oakland CA	37.802	-122.311	 OAKLAND CA          	23230	37.717	-122.217
12.6	 OAKLAND  INT AP         	23230	37.75	-122.22	9.9

Richmond CA	37.907	-122.369	 OAKLAND CA          	23230	37.717	-122.217
25	 OAKLAND  INT AP         	23230	37.75	-122.22	21.8

Jacksonville FL	30.402	-81.531	 JACKSONVILLE FL     	53860	30.333
-81.517	7.7	 JACKSONVILLE            	13889	30.43	-81.7	16.6

Miami FL	25.771	-80.163	 MIAMI FL            	12839	25.817	-80.3	14.6	
MIAMI/FL INTL UNIV      	92803	25.75	-80.38	21.8

Panama City FL	30.156	-84.201	 TALLAHASSEE FL      	93805	30.383	-84.35
29	 TALLAHASEE              	93805	30.45	-84.3	34

Port Everglades FL	26.091	-80.126	 FORT LAUDERDALE FL  	12849	26.067
-80.15	3.7	 MIAMI/FL INTL UNIV      	92803	25.75	-80.38	45.7

Tampa FL	27.935	-82.442	 TAMPA FL            	12842	27.967	-82.533	9.6	
TAMPA BAY/ RUSKIN        	12842	27.7	-82.4	26.4

Savannah GA	32.121	-81.141	 SAVANNAH GA         	3822	32.117	-81.2	5.6	
CHARLESTON              	13880	32.9	-80.03	135.9

Chicago IL	41.736	-87.536	 CHICAGO IL          	14819	41.783	-87.75	18.5
 LINCOLN-LOGAN COUNTY AP 	4833	40.15	-89.33	230.9

Gary IN	41.617	-87.344	 CHICAGO IL          	4879	41.533	-87.533	18.3	
LINCOLN-LOGAN COUNTY AP 	4833	40.15	-89.33	232.1

Mount Vernon IN	37.93	-87.874	 HENDERSON KY        	53886	37.817	-87.683
21	 NASHVILLE               	13897	36.25	-86.57	219.1

Louisville KY	38.276	-85.789	 LOUISVILLE KY       	93821	38.183	-85.733
11.4	 WILMINGTON              	13841	39.42	-83.82	213.9

Lake Charles LA	30.214	-93.254	 LAKE CHARLES LA     	3937	30.117	-93.233
11	 LAKE CHARLES            	3937	30.12	-93.22	11

New Orleans LA	29.914	-90.101	 NEW ORLEANS LA      	12958	29.833	-90.033
11.1	 SLIDELL                 	53813	30.33	-89.82	53.6

Port of Baton Rouge LA	30.438	-91.205	 BATON ROUGE LA      	13970	30.533
-91.15	11.9	 SLIDELL                 	53813	30.33	-89.82	133.5

Port of Plaquemines LA	29.469	-89.687	 BOOTHVILLE LA       	12884	29.333
-89.417	30.2	 SLIDELL                 	53813	30.33	-89.82	96.5

South Louisiana LA	30.048	-90.814	 NEW ORLEANS LA      	12916	29.983
-90.25	54.8	 SLIDELL                 	53813	30.33	-89.82	100.7

Boston MA	42.34	-71.018	 BOSTON MA           	14739	42.367	-71.017	2.9	
CHATHAM                 	14684	41.67	-69.97	114

Baltimore MD	39.248	-76.533	 BALTIMORE MD        	93721	39.167	-76.683
15.8	 STERLING(WASH DULLES)   	93734	38.98	-77.47	86.1

Portland ME	43.655	-70.25	 PORTLAND ME         	14764	43.633	-70.3	4.7	
GRAY                    	54762	43.89	-70.25	26.1

Detroit MI	42.305	-83.092	 DETROIT MI          	14822	42.417	-83.017
13.8	 DETROIT/PONTIAC         	4830	42.7	-83.47	53.8

Duluth-Superior MN	46.767	-92.106	 DULUTH MN           	4919	46.717
-92.033	7.9	 INTERNATIONAL FALLS     	14918	48.57	-93.38	222.7

Two Harbors MN	47.018	-91.676	 TWO HARBORS MN      	4979	47.05	-91.75
6.6	 INTERNATIONAL FALLS     	14918	48.57	-93.38	215.6

St. Louis MO	38.583	-90.218	 CAHOKIA/ST.LOUIS IL 	3960	38.567	-90.167
4.8	 LINCOLN-LOGAN COUNTY AP 	4833	40.15	-89.33	190.6

Wilmington NC	34.196	-77.952	 WILMINGTON NC       	13748	34.267	-77.917
8.5	 MOREHEAD CITY/NEWPORT   	93768	34.7	-76.8	120

Paulsboro NJ	39.907	-75.126	 PHILADELPHIA PA     	13739	39.867	-75.233
10.2	 BROOKHAVEN              	94703	40.87	-72.87	220.4

New York NY	40.684	-74.009	 NEW YORK NY         	94728	40.783	-73.967
11.6	 BROOKHAVEN              	94703	40.87	-72.87	98.4

Cincinnati OH	39.097	-84.532	 CINCINNATI OH       	93812	39.1	-84.417	10
 WILMINGTON              	13841	39.42	-83.82	71.2

Cleveland OH	41.47	-81.675	 CLEVELAND OH        	4853	41.517	-81.683	5.3
 PITTSBURGH/MOON TOWNSHIP	94823	40.53	-80.23	159.4

Tulsa - Port of Catoosa OK	36.233	-95.74	 TULSA OK            	13968
36.2	-95.883	13.4	 LAMONT                  	974646	36.68	-97.47	163.1

Portland OR	45.56	-122.703	 VANCOUVER WA        	94298	45.617	-122.667
6.9	 SALEM                   	24232	44.92	-123.02	75.3

Philadelphia PA	39.893	-75.173	 PHILADELPHIA PA     	13739	39.867
-75.233	5.9	 WALLOPS ISLAND          	93739	37.93	-75.48	219.7

Pittsburgh PA	40.436	-80.009	 PITTSBURGH PA       	14762	40.35	-79.917
12.3	 PITTSBURGH/MOON TOWNSHIP	94823	40.53	-80.23	21.5

Charleston SC	32.833	-79.885	 CHARLESTON SC       	13880	32.9	-80.033
15.8	 CHARLESTON              	13880	32.9	-80.03	15.5

Nashville TN	36.186	-86.788	 NASHVILLE TN        	13897	36.117	-86.683
12.2	 NASHVILLE               	13897	36.25	-86.57	20.9

Corpus Christi TX	27.817	-97.404	 CORPUS CHRISTI TX   	12924	27.767
-97.517	12.4	 CORPUS CHRISTI          	12924	27.77	-97.5	10.8

Freeport TX	28.895	-95.364	 ANGLETON/LAKE JACKSON TX	12976	29.117
-95.467	26.5	 CORPUS CHRISTI          	12924	27.77	-97.5	242.8

Houston TX	29.608	-95.015	 HOUSTON TX          	12975	29.517	-95.233
23.5	 LAKE CHARLES            	3937	30.12	-93.22	182.8

Port Arthur TX	29.859	-93.944	 BEAUMONT/PORT ARTHUR TX	12917	29.95
-94.017	12.3	 LAKE CHARLES            	3937	30.12	-93.22	75.6

Texas City TX	29.372	-94.904	 GALVESTON TX        	12923	29.267	-94.867
12.2	 LAKE CHARLES            	3937	30.12	-93.22	183.3

Norfolk Harbor VA	36.92	-76.322	 NORFOLK VA          	13737	36.9	-76.183
12.5	 WALLOPS ISLAND          	93739	37.93	-75.48	135

Seattle WA	47.579	-122.356	 SEATTLE WA          	24234	47.533	-122.3	6.6
 QUILLAYUTE              	94240	47.95	-124.55	169.8

Tacoma WA	47.262	-122.392	 TACOMA WA           	94274	47.267	-122.583
14.5	 QUILLAYUTE              	94240	47.95	-124.55	180.1



Emissions All dispersion modeling was performed with unit emission
factors (1 g/s/m2) for all polygon area sources. Resulting concentration
predictions were then scaled to reflect the actual emission rate
estimates for each harbor area, as described below. 

Temporal variations were included in emission strengths in the modeling.
These variations were taken to be both seasonal and annual (SEASHR),
based on values from the EMS-HAP model.  These factors are disaggregated
in the EMS-HAP model database based on SCC codes. The SCC codes track
the various types of equipment operating at ports reasonably well, so
the variation was mapped to four offroad equipment categories included
here: Truck, Rail, CHE, and combined H/C and Vessels. Within each of
these categories, the variation in the SEASHR emission factors was
negligible. However, there was some variation between categories. 

A single set of SEASHR emission factors for each harbor area were
determined by combining the values from each of the four offroad
categories, weighted by the relative emissions from each of these
categories. Although this method determined a unique set of SEASHR
factors for each harbor area, it eliminated the variation from the
individual sources categories. Given the uncertainty in applying these
generalized temporal profiles to specific harbor areas, we believe this
aggregation is likely the best approach to incorporate temporal
variation in the modeling. 

Figure 2 shows these profiles for three ports. In each case, there are
96 emission scaling factors, representing factors for the 24 hours of
each day in each of the four seasons: winter, spring, summer, and
autumn, sequentially. 

Actual emission rates were derived from an internal EPA analysis of
harbor area emissions.  The emission rates used as inputs into the
analysis are not official estimates and are believed to underestimate
overall emissions at the port areas.  Work is currently underway to
improve these inventories and EPA will revisit this assessment as better
information is developed for local port areas.  

PM10 was used as a surrogate for DPM emissions at each of the 40 ports
that were included in both projects. Although DPM emissions were not
explicitly calculated for the Harbor Area Emissions Inventory, the PM10
emissions exclusively represent exhaust sources, essentially all of
which are diesel fueled (see caveats below). 

For the 7 port areas included in the present analysis, but not in the
Harbor Area Emissions Inventory (Corpus Christi, TX, Gary, IN, Helena,
AR, Mount Vernon, IN, Norfolk Harbor, VA, Tulsa, Port of Catoosa, OK,
and Two Harbors, MN), total annual DPM emissions for 2005 were
determined through a multilinear regression of emissions against cargo
tonnage in four conveyance categories (liquid, bulk, container, and
other) from the 84 other harbor areas included in the Harbor Area
Emissions Inventory. This analysis produced an adjusted coefficient of
determination (R2) of 61%, and thus is believed to provide satisfactory
emissions estimates for these ports. 



Figure 2. SEASHR temporal emission scaling factors for South Louisiana,
LA, Los Angeles, CA, and Wilmington, NC.

There are several caveats associated with the use of the Harbor Area
Emissions Inventory. The most important ones are as follows.

The values are not official estimates and have not been reviewed by the
individual harbor representatives and other stakeholders.  It is likely
that they underestimate overall emissions at the individual ports;
therefore work is currently underway to improve these inventories.

A small portion of the equipment, which was assumed for this analysis to
be fueled by diesel, is alternatively fueled. For example, some deep
draft vessels are fueled by gas or steam turbines, and some CHE can be
operated on liquefied natural gas or other alternative fuels. Hence, the
assumption that all the estimated PM10 is diesel could lead to a slight
overestimate of the total DPM emissions at each port complex. 

Some comparisons have been made between the Harbor Area Emissions
Inventory values and the published values for the few harbors that have
conducted detailed inventories (e.g., Ports of Los Angeles, Long Beach,
Oakland, and Portland OR). Overall, the agreement is fairly good.
However, there are technical differences that cause discrepancies. For
example, the maximum cruising distance from shore that is tracked for
emissions is shorter in the Harbor Area Emissions Inventory than in the
published inventories (i.e., 25 nautical miles vs. 250 nautical miles),
typically resulting in lower total vessel emissions. Truck and rail
emissions in the Harbor Area Emissions Inventory calculations also are
typically different from published inventories, again due to boundary
definitions.  Finally, the Harbor Area Emissions Inventory includes rail


emissions for transport to the MSA boundary, and truck emissions for
transport either to the MSA boundary or the point of
onloading/offloading, whichever is smaller.

On the other hand preliminary estimates show the Harbor Area Emissions
Inventory estimates for Portland, OR, for example, to be higher than
published values. This, too, is due to mismatching, as the published
values include only the four terminals under the jurisdiction of the
Port Authority, and the Harbor Area Emissions Inventory attempts to
include all activity in the general waterway. As explained, the area
source boundaries were designed for this modeling study to be roughly
consistent with the boundary definitions used in the Harbor Area
Emissions Inventory.

Emission estimates for the Harbor Area Emissions Inventory were believed
to represent the best estimates for current (year 2005) activities in
these harbor areas. However, uncertainties in emissions are likely to be
the largest contributor to uncertainty in the predicted DPM
concentrations in nearby communities, larger than the contribution from
uncertainties in modeling parameters discussed above. 

Resulting Concentrations. Each harbor area was simulated for three years
and annual average concentrations were computed at each receptor. The
three annual average concentrations at each receptor were then averaged
together. Each receptor point was then translated from local coordinates
to latitude and longitude and digitized over the port “footprint”,
to determine contours of annual average concentration at each of the
threshold levels. Figure 3 shows an example of the digitized facility
boundaries and the 2.0 µg/m3 and 0.2 µg/m3 isopleths for Baltimore
Harbor.  Pictures of the digitized facility boundaries with the
concentration isopleths for all 47 US harbor areas are presented in
Appendix D.



Figure 3. Concentration Isopleths for Baltimore Harbor.

The area encompassed by each of the isopleths at each port (excluding
the port footprint) is presented in Table 7.

Table 7. Areas Enclosed by Concentration Isopleths at Marine Harbor
areas. 

Marine Harbor area	Area 

(acres, excluding water)

	Facility	Isopleth Concentrations (µg/m3)



2.0	0.2

Baltimore, MD	985	2,052	48,430

Boston, MA	1,132	182	13,953

Charleston, SC	346	1,154	27,805

Chicago, IL	6,206	0	4,207

Cincinnati, OH	1,101	0	4,430

Cleveland, OH	648	0	1,755

Corpus Christi, TX	568	643	6,074

Detroit, MI	1,923	0	575

Duluth-Superior, MN	2,922	0	2,401

Freeport, TX	9,271	0	6,332

Gary, IN	4,969	0	8,511

Helena, AR	599	2,421	39,058

Houston, TX	26,181	0	40,435

Jacksonville, FL	3,454	0	8,783

Lake Charles, LA	1,767	100	21,731

Long Beach, CA	5,052	1,190	62,343

Los Angeles, CA	5,140	950	45,492

Louisville, KY	85	98	1,344

Miami, FL	568	3,113	127,270

Mobile, AL	3,241	127	20,233

Mount Vernon, IN	753	881	22,700

Nashville, TN	503	0	688

New Orleans, LA	440	5,566	90,854

New York, NY	9,065	4,195	216,430

Norfolk Harbor, VA	1,827	667	30,871

Oakland, CA	2,030	766	29,338

Panama City, FL	88	315	3,629

Paulsboro, NJ	874	0	5,336

Philadelphia, PA	4,018	0	23,044

Pittsburgh, PA	197	0	7,504

Port Arthur, TX	249	196	5,131

Port Everglades, FL	1,096	0	6,005

Portland ME	428	87	6,394

Portland OR	3,545	0	30,062

Port of Baton Rouge, LA	3,274	261	32,747

Port of Plaquemines, LA	236	1,191	24,356

Richmond, CA	573	0	2,899

Savannah, GA	2,512	1,325	57,826

Seattle, WA	2,383	3,747	99,853

South Louisiana, LA	13,223	0	23,188

St. Louis, MO	860	483	7,061

Tacoma, WA	3,249	1,206	55,100

Tampa, FL	1,635	660	10,905

Texas City, TX	1,226	11	4,885

Tulsa - Port of Catoosa, OK	603	283	7,496

Two Harbors, MN	396	286	5,003

Wilmington, NC	558	71	6,068



RAIL YARDS

Thirty-seven rail yards identified by EPA were selected for analysis.
The facilities addressed are listed in Table 8. 

Table 8. Selected Rail Yards

Yard Name	Location	Railroad

Argentine	Kansas City, KS 	BNSF

Avon Yard	Indianapolis, IN 	CSXT

Bailey Yard	North Platte, NE 	UP

Barr Yard	Chicago, IL 	CSXT

Barstow Yard	Barstow, CA 	BNSF

Bellevue Yard	Bellevue, OH 	NS

Bensenville Yard	Bensenville, IL 	CP

Blue Island Yard	Blue Island, IL 	IHB

Boyles Yard	Birmingham, AL 	CSXT

Buckeye Yard	Columbus, OH 	CSXT

Clearing Yard	Chicago, IL 	BRC

Conway Yard	Conway, PA 	UP

Corwith Yard	Chicago, IL 	NS

DeButts Yard	Chattanooga, TN 	BNSF

Frontier Yard	Buffalo, NY 	NS

Frontier Yard Intermodal Terminal	Buffalo, NY 	CSXT

Galesburg Yard	Galesburg, IL 	CSXT

Hinkle Yard	Hermiston, OR 	BNSF

Inman Yard	Atlanta, GA 	UP

J.R. Davis Yard	Roseville, CA 	BSNF

Jenks Shop	North Little Rock, AR 	NS

Locomotive Maintenance Facility	Alliance, NE 	UP

Locomotive Repair Facility	Topeka, KS 	UP

Madison Yard	East St. Louis, MO 	BNSF

Moncrief Yard	Jacksonville, FL 	BNSF

Philadelphia Railyard	Philadelphia, PA 	TRRA

Pig's Eye Yard	Minneapolis, MN 	CSXT

Proviso Yard	Chicago, IL 	CSXT

Queensgate Yard	Cincinnati, OH 	CP

Radnor Yard	Nashville, TN 	UP

Rice Yard	Waycross, GA 	CSXT

Schiller Park Yard	Schiller Park, IL 	CSXT

Selkirk Yard	Selkirk, NY	CSXT

Shaffers Crossing	Roanoke, VA 	CP

Spencer Yard	Linwood, NC 	CSXT

Stanley/Walbridge Yard	Toledo, OH 	NS

West Colton Yard	West Colton, CA 	BSNF



Facility Footprints

Selected rail yards were identified using aerial imagery from USDA Data
Gateway (  HYPERLINK "http://datagateway.nrcs.usda.gov" 
http://datagateway.nrcs.usda.gov ). Rail yard boundaries and
configurations were confirmed using North American Rail Yards (Michael
Rhodes, MBI, October 2003, ISBN 0-7603-1578-7).

Pictures of the digitized footprints for the 37 US rail yards are
presented in Appendix E.

Estimating Isopleth Areas and Distances

We were not able to identify emissions data for the selected rail yards,
nor were we able to identify activity data from which emissions
estimates could be made, therefore, air dispersion modeling could not be
performed for the rail yards in order to estimate the size and location
of the isopleths for the various concentration levels of interest.

In the absence of such information, a very approximate method for making
these estimates was developed.

In order to estimate the isopleth sizes a method was developed to
extrapolate the area encompassed by the isopleths for each selected rail
yard from corresponding estimates for rail yards for which detailed air
dispersion modeling had been performed. 

Because the shape of the isopleths could not be estimated from air
dispersion modeling based on local meteorological conditions, it was
assumed that the isopleths shapes were similar to the those of the rail
yard shapes, i.e. that the isopleth boundary was everywhere equidistant
from the rail yard boundary.  This implies that the wind is equally
likely to come from any direction, and that the wind direction and wind
speed are independent. These assumptions introduce a substantial level
of uncertainty into the predictions.

As a result of the uncertainties, estimates are presented only for the
ranges of area and distance for each rail yard isopleths, rather than
point estimates.

 

Several modeling studies of rail yards have recently been reported by
the California Air Resources Board as follows.

“Roseville Rail Yard Study” (CARB, 2004)

“Draft Health Risk Assessment for the Four Commerce Railyards”
(CARB, 2007)

“Draft Health Risk Assessment for the BSNF Railways (BSNF) Commerce
Eastern Railyard” (CARB, 2007)

“Draft Health Risk Assessment for the BSNF Railways Hobart Railyard”
(CARB, 2007)

“Draft Health Risk Assessment for the BSNF Railways Shelia Mechanical
Railyard” (CARB, 2007)

“Draft Health Risk Assessment for the BSNF Railways Watson Railyard”
(CARB, 2007)

“Draft Health Risk Assessment for the Union Pacific Railroad Commerce
Railyard” (CARB, 2007)

“Draft Health Risk Assessment for the Union Pacific Railroad Los
Angeles Transportation Center Railyard” (CARB, 2007)

“Draft Health Risk Assessment for the Union Pacific Railroad Mira Loma
Auto Facility Railyard” (CARB, 2007)

“Draft Health Risk Assessment for the Union Pacific Railroad Stockton
Railyard” (CARB, 2007)

“Draft Health Risk Assessment for the Union Pacific Railroad Mira
Watson Railyard” (CARB, 2007)

The scaling metric selected for this analysis was the
isopleth-to-railyard area ratio, where the isopleths area excludes the
railyard area. Therefore, for each study this ratio was calculated for
the isopleths concentrations reported in the listed studies. The results
are presented in Table 9.

Table 9. Isopleth-to-Railyard Area Ratios for Selected Railyards

Railyard	Area (acres)	

Isopleth Concentration (µg/m3)



0.03	0.08	0.17	0.33	0.83	1.67	3.33

UP Roseville	950	50

	2

0

	BNSF Hobart	264	194	69	31	13	4	1	0

UP Commerce	202	86	28	12	4	1	0

	UP LATC	120	79	29	12	5



	UP Stockton	300	23	8	3	1



	UP Mira Loma	274	15	4	1





BNSF Richmond	150	16	6	2





BNSF Stockton	100	33	10	4	1



	BNSF Commerce Eastern	44	110	35	14	6



	BNSF Sheila	63	51	10





	BNSF Watson	37	81	21	9	3



	

The ratios were selected from UP Mira Loma as the low end of the range
and from BSNF Hobart as the high end of the range, since these had the
smallest and largest isopleths-to-railyard area ratios, respectively.

The next step was to interpolate and extrapolate the ratios to the
isopleth concentration levels of interest for this study: 2.0 µg/m3 and
0.2 µg/m3. Figures 4 and 5 are scatter plots of the ratios for Mira
Loma and Hobart Railyards, respectively, using logarithmic scales for
both the ratios and the isopleth concentrations. Inspection of these
plots suggests that the data is approximately linear in logs, so that
log-log interpolation and extrapolation is appropriate. The resulting
interpolated and extrapolated ratios are presented in Table 10.



Figure 4. Scatter Plot of Isopleth Area vs. Isopleth Concentration for
UP Mira Loma Rail Yard.

Figure 5. Scatter Plot of Isopleth Area vs. Isopleth Concentration for
BSNF Hobart Rail Yard.



Table 10. Estimated Isopleth-to-Railyard Area Ratios

Railyard	

Isopleth Concentration (µg/m3)

	2.0	0.2

UP Mira Loma	0	0.42

BSNF Hobart	0.85	25



The resulting isopleth areas, excluding the railyard areas, and
approximate associated distances from the railyard boundaries are
presented in Tables 11 and 12, respectively.

Pictures of the digitized facility boundaries with the concentration
isopleths for the 37 US rail yards are presented in Appendix F.



Table 11. Estimated Ranges of Isopleth Areas (Acres) for Selected
Railyards (railyard areas excluded).

Yard Name	Location	Area (acres)





Isopleth Concentrations (ug/m3)



	2.0	0.2



	Low	High	Low	High

Argentine	Kansas City, KS 	575	0	488	239	14,362

Avon Yard	Indianapolis, IN 	354	0	300	147	8,840

Bailey Yard	North Platte, NE 	1530	0	1,299	635	38,240

Barr Yard	Chicago, IL 	234	0	198	97	5,843

Barstow Yard	Barstow, CA 	870	0	738	361	21,734

Bellevue Yard	Bellevue, OH 	531	0	450	220	13,263

Bensenville Yard	Bensenville, IL 	385	0	327	160	9,630

Blue Island Yard	Blue Island, IL 	212	0	180	88	5,308

Boyles Yard	Birmingham, AL 	87	0	74	36	2,181

Buckeye Yard	Columbus, OH 	390	0	331	162	9,740

Clearing Yard	Chicago, IL 	1179	0	1,000	489	29,456

Conway Yard	Conway, PA 	496	0	421	206	12,385

Corwith Yard	Chicago, IL 	194	0	165	81	4,850

DeButts Yard	Chattanooga, TN 	399	0	339	166	9,975

Frontier Yard	Buffalo, NY 	205	0	174	85	5,122

Frontier Yard Intermodal Terminal	Buffalo, NY 	224	0	190	93	5,588

Galesburg Yard	Galesburg, IL 	820	0	696	341	20,500

Hinkle Yard	Hermiston, OR 	779	0	661	323	19,469

Inman Yard	Atlanta, GA 	732	0	621	304	18,290

J.R. Davis Yard	Roseville, CA 	742	21	 21	 273	 273

Jenks Shop	North Little Rock, AR 	381	0	323	158	9,515

Locomotive Maintenance Facility	Alliance, NE 	348	0	296	145	8,703

Locomotive Repair Facility	Topeka, KS 	158	0	134	66	3,948

Madison Yard	East St. Louis, MO 	127	0	107	53	3,164

Moncrief Yard	Jacksonville, FL 	257	0	218	107	6,413

Philadelphia Yard	Philadelphia, PA 	310	0	263	129	7,755

Pig's Eye Yard	Minneapolis, MN 	302	0	257	126	7,558

Proviso Yard	Chicago, IL 	643	0	546	267	16,063

Queensgate Yard	Cincinnati, OH 	466	0	396	194	11,650

Radnor Yard	Nashville, TN 	433	0	367	180	10,811

Rice Yard	Waycross, GA 	688	0	584	286	17,191

Schiller Park Yard	Schiller Park, IL 	86	0	73	36	2,156

Selkirk Yard	Selkirk, NY	1,179	0	1,001	490	29,471

Shaffers Crossing	Roanoke, VA 	270	0	229	112	6,752

Spencer Yard	Linwood, NC 	336	0	285	139	8,390

Stanley/Walbridge Yard	Toledo, OH 	383	0	326	159	9,584

West Colton Yard	West Colton, CA 	432	0	367	179	10,794



Table 12. Estimated Ranges of Isopleth Distances (Miles) from Railyard
Boundary for Selected Railyards.

Yard Name	Location	Area (acres)





Isopleth Concentrations (ug/m3)



	2.0	0.2



	Low	High	Low	High

Argentine	Kansas City, KS 	575	0	0.10	0.05	1.70

Avon Yard	Indianapolis, IN 	354	0	0.08	0.04	1.38

Bailey Yard	North Platte, NE 	1530	0	0.16	0.08	2.82

Barr Yard	Chicago, IL 	234	0	0.06	0.03	1.10

Barstow Yard	Barstow, CA 	870	0	0.13	0.06	2.18

Bellevue Yard	Bellevue, OH 	531	0	0.10	0.05	1.70

Bensenville Yard	Bensenville, IL 	385	0	0.08	0.04	1.42

Blue Island Yard	Blue Island, IL 	212	0	0.06	0.03	1.08

Boyles Yard	Birmingham, AL 	87	0	0.04	0.02	0.70

Buckeye Yard	Columbus, OH 	390	0	0.07	0.03	1.31

Clearing Yard	Chicago, IL 	1179	0	0.15	0.08	2.57

Conway Yard	Conway, PA 	496	0	0.08	0.04	1.51

Corwith Yard	Chicago, IL 	194	0	0.08	0.04	1.16

DeButts Yard	Chattanooga, TN 	399	0	0.07	0.04	1.37

Frontier Yard	Buffalo, NY 	205	0	0.08	0.04	1.16

Frontier Yard Intermodal Terminal	Buffalo, NY 	224	0	0.07	0.04	1.14

Galesburg Yard	Galesburg, IL 	820	0	0.11	0.05	2.05

Hinkle Yard	Hermiston, OR 	779	0	0.11	0.06	1.98

Inman Yard	Atlanta, GA 	732	0	0.10	0.05	2.04

J.R. Davis Yard	Roseville, CA 	742	0.004	0.004	0.40	0.40

Jenks Shop	North Little Rock, AR 	381	0	0.08	0.04	1.42

Locomotive Maintenance Facility	Alliance, NE 	348	0	0.09	0.05	1.46

Locomotive Repair Facility	Topeka, KS 	158	0	0.07	0.03	1.03

Madison Yard	East St. Louis, MO 	127	0	0.06	0.03	0.88

Moncrief Yard	Jacksonville, FL 	257	0	0.08	0.04	1.27

Philadelphia Yard	Philadelphia, PA 	310	0	0.09	0.05	1.39

Pig's Eye Yard	Minneapolis, MN 	302	0	0.09	0.04	1.38

Proviso Yard	Chicago, IL 	643	0	0.12	0.06	1.96

Queensgate Yard	Cincinnati, OH 	466	0	0.08	0.04	1.50

Radnor Yard	Nashville, TN 	433	0	0.08	0.04	1.47

Rice Yard	Waycross, GA 	688	0	0.10	0.05	1.85

Schiller Park Yard	Schiller Park, IL 	86	0	0.04	0.02	0.71

Selkirk Yard	Selkirk, NY	1,179	0	0.11	0.06	2.21

Shaffers Crossing	Roanoke, VA 	270	0	0.06	0.03	1.09

Spencer Yard	Linwood, NC 	336	0	0.07	0.04	1.28

Stanley/Walbridge Yard	Toledo, OH 	383	0	0.07	0.03	1.34

West Colton Yard	West Colton, CA 	432	0	0.07	0.04	1.37



 Mobile Source Air Toxics Rule (Control of Hazardous Air Pollutants from
Mobile Sources; 72 FR 8428, February 26, 2007) Regulatory Impact
Analysis Chapter 3, “Air Quality and Resulting Health and Welfare
Effects of Air Pollution from Mobile Sources.” This document is
available electronically at:    HYPERLINK
"http://www.epa.gov/otaq/regs/toxics/420r07002.pdf" 
http://www.epa.gov/otaq/regs/toxics/420r07002.pdf 

 State of California Air Resources Board.  Roseville Rail Yard Study.
Stationary Source Division, October 14, 2004. This document is available
electronically at:   HYPERLINK
"http://www.arb.ca.gov/diesel/documents/rrstudy.htm" 
http://www.arb.ca.gov/diesel/documents/rrstudy.htm    

 State of California Air Resources Board.  Diesel Particulate Matter
Exposure Assessment Study for the Ports of Los Angeles and Long Beach,
April 2006.  This document is available electronically at: 
http://www.arb.ca.gov/regact/marine2005/portstudy0406.pdf

 U.S. EPA. 2002. Health Assessment Document for Diesel Exhaust.
EPA/600/8-90/057F. May 2002.  

 U.S.EPA (1996) Air Quality Criteria for Particulate Matter, EPA
600-P-95-001aF, EPA 600-P-95-001bF. This document is available in Docket
EPA-HQ-OAR.

 U.S. EPA (2004) Air Quality Criteria for Particulate Matter (Oct 2004),
Volume I Document No. EPA600/P-99/002aF and Volume II Document No.
EPA600/P-99/002bF. This document is available in Docket EPA-HQ-OAR.  

 U.S. EPA (2005) Review of the National Ambient Air Quality Standard for
Particulate Matter: Policy Assessment of Scientific and Technical
Information, OAQPS Staff Paper.  EPA-452/R-05-005. This document is
available in Docket EPA-HQ-OAR.

 USEPA. 2001. National-Scale Air Toxics Assessment for 1996: DRAFT for
EPA Science Advisory Board Review: January 18, 2001. Office of Air
Quality Planning and Standards. EPA-453/R-01-003, Section 5.4.

 US EPA. 2006. National-Scale Air Toxics Assessment for 1999.  This
material is available electronically at
http://www.epa.gov/ttn/atw/nata1999/.

 Time, data availability, and financial constraints did not allow a
full, regulatory application of AERMOD for all 47 harbor areas of
interest. Consequently, there are uncertainties in the analysis that
could affect the results. Although quantitatively addressing each of
these is infeasible for this analysis, their potential impacts are
discussed qualitatively throughout this section and in Appendices B and
C. 

 Individual release heights were derived from those reported in CARB.
2006. Diesel Particulate Matter Exposure Assessment Study for The Ports
of Los Angeles and Long Beach, Final Report, Table 7. These values were
aggregated based on the median contribution to emissions from each
source category. The average initial vertical dimension was determined
by averaging the individual elements in quadrature, in the typical
method for standard deviations. (  HYPERLINK
"http://www.arb.ca.gov/ports/marinevess/documents/portstudy0406.pdf" 
http://www.arb.ca.gov/ports/marinevess/documents/portstudy0406.pdf )
Note that the values imply a factor of 2.15 for computing initial
vertical dispersion, which follows EPA guidance for low-level releases,
but not for elevated sources, for which the recommended factor is 4.3.
However, these values were chosen for consistency with the CARB
analysis. The resulting initial vertical plume spread is somewhat larger
than EPA guidance suggests for elevated sources. This could lead to the
present analysis to have somewhat larger impacts close to the source,
but lower impacts at greater distances than if EPA guidance had been
followed.

 “Method 2 is used when the particle size distribution is not well
known and when a small fraction (less than 10% of the mass) is in
particles with a diameter of 10 μm or larger” from Addendum, User's
Guide For the AMS/EPA Regulatory Model – AERMOD, USEPA 454/B-03-001,
September 2004.

 Task 1: Recommendation on Deposition Modeling in the Near Mobile Source
Environment, EPA, Contract 68-C-01-164, Work Assignment No. 3-3,
Technical Memorandum prepared for Chad Bailey, USEPA OTAQ by Edward Carr
and Seth Hartley, December 9, 2004. 

 http://www.census.gov/population/www/estimates/CBSA-est2006-annual.html

 AERMOD Implementation Guide. September 27, 2005. Available on the EPA
SCRAM website
(http://www.epa.gov/scram001/7thconf/aermod/aermod_implmtn_guide.pdf)

 http://www5.ncdc.noaa.gov/ulcd

 http://raob.fsl.noaa.gov/Raob_Software.html

 EPA, 2005. AERMOD IMPLEMENTATION GUIDE, September 27, 2005. Available
at www.epa.gov/scram001/7thconf/aermod/aermod_implmtn_guide.pdf

 EPA 2007, Air Quality Modeling Data Files for Local Populations
Impacted Analysis, December 2007. Materials can be found in Docket
EPA-HQ-OAR-2003-0190.

hì

C

D

h

i

v

w

g

h

i

v

w

x

€

‚

‹

£

°

¹

ÿ

hÆ

hÆ

hA

hA

kd 

kdÅ

摧᏿t欀╤

摧᏿t欀핤

 hÆ

 

  hƒ

  hƒ

hƒ

  hƒ

  hƒ

hƒ

Ff¬

H

H

 hÆ

€&

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

	\

”ÿð

	\

'š

'@

	\

'š

'@

 EPA 2004, User's Guide for the Emissions Modeling System for Hazardous
Air Pollutants (EMS-HAP) Version 3.0, EPA-454/B-03-006. 

 Emissions data for Corpus Christi, TX and Norfolk Harbor, VA are
included in the Harbor Area Emissions Inventory, but were inadvertently
omitted from the data base used for this analysis.

 http://www.arb.ca.gov/railyard/hra/hra.htm

 The values for this rail yard are interpolated and extrapolated from
the isopleths-to-railyard area ratios derived from data reported in
“Roseville Rail Yard Study” (CARB, 2004). 

 The values for this rail yard are interpolated and extrapolated from
the isopleths-to-railyard area ratios derived from data reported in
“Roseville Rail Yard Study” (CARB, 2004).

Page   PAGE  3  of   NUMPAGES  26 

Estimation of concentration isopleths for marine harbor areas and rail
yards

ICF International	September 28, 2007

  www. icfi.com

ICF International	September 28, 2007

  www. icfi.com

