DIESEL PARTICULATE MATTER EXPOSURE ASSESSMENT
STUDY FOR THE PORTS OF LOS ANGELES AND LONG BEACH
FINAL REPORT
April 2006
i
State of California
AIR RESOURCES BOARD
DIESEL PARTICULATE MATTER EXPOSURE ASSESSMENT
STUDY FOR THE PORTS OF LOS ANGELES AND LONG BEACH
Primary Author
Pingkuan Di, Ph. D., P. E.
Contributing Staff
Anthony Servin, P. E.
Kirk Rosenkranz
Beth Schwehr
Hein Tran
Reviewed and Approved by
Robert D. Fletcher, Chief
Stationary Source Division
Daniel E. Donohoue, Chief
Emissions Assessment Branch
Peggy Taricco, Chief
Emission Inventory Branch ( PTSD)
Erik White, Manager
Technical Analysis Section
ii
Acknowledgements
Air Resources Board staff extends its appreciation to representatives of Starcrest
Consulting Group, LLC and the Ports of Los Angeles and Long Beach for providing
assistance with emissions inventory data and spatial allocation of emissions. OEHHA
provided the commentary on unquantified risk ( Appendix D).
The staff of the Air Resources Board has prepared this report. Publication does not
signify that the contents reflect the views and policies of the Air Resources Board.
iii
DIESEL PARTICULATE MATTER EXPOSURE ASSESSMENT
STUDY FOR THE PORTS OF LOS ANGELES AND LONG BEACH
TABLE OF CONTENTS
Section........................................................................................................................ ............... Page
Part I: Summary ........................................................................................................................... 1
Part II. Technical Support Document................................................................................... 15
I. INTRODUCTION.............................................................................................................. 15
A. OVERVIEW ........................................................................................................... 15
B. PURPOSE............................................................................................................. 16
C. DESCRIPTION OF THE PORTS....................................................................... 17
II. EMISSION INVENTORY DEVELOPMENT................................................................. 19
A. PORT OF LOS ANGELES.................................................................................. 19
B. PORT OF LONG BEACH ................................................................................... 22
C. IN- PORT AND OUT- OF- PORT EMISSIONS ALLOCATION........................ 23
D. EMISSION INVENTORY SUMMARY............................................................... 24
III. AIR DISPERSION MODELING..................................................................................... 26
A. AIR DISPERSION MODEL SELECTION ......................................................... 26
B. MODEL DOMAIN AND RECEPTOR NETWORK........................................... 26
C. MODEL PARAMETERS...................................................................................... 29
D. SPATIAL AND TEMPORAL ALLOCATION OF EMISSIONS ....................... 29
E. METEOROLOGICAL DATA ............................................................................... 30
IV. EXPOSURE ASSESSMENT.......................................................................................... 34
A. OEHHA GUIDELINES......................................................................................... 34
B. EXPOSURE ASSESSMENT.............................................................................. 35
C. RISK CHARACTERIZATION ............................................................................. 35
D. ESTIMATION OF NON- CANCER HEALTH .................................................... 49
E. UNQUANTIFIABLE ADVERSE HEALTH EFFECTS ..................................... 51
F. COMPARISONS WITH MONITORING RESULTS......................................... 51
G. UNCERTAINTIES AND LIMITATIONS............................................................. 54
V. SUMMARY OF FINDINGS ............................................................................................. 57
REFERENCES ................................................................................................................. 61
iv
Appendices
Appendix A: Methodologies for Developing Source Category Emission
Inventories ................................................................................................. A- 1
Appendix B: Development of Ship Auxiliary Engine Emission Factors ................. B- 1
Appendix C: Comparison of Estimated Diesel PM Cancer Risks from
Oceangoing Vessel Activity Outside of the Breakwater using
Wilmington and King Harbor Meteorological Data Sets .................... C- 1
Appendix D: Unquantifiable Health Adverse Effects ............................................... D- 1
List of Tables
TABLE 1 Estimated 2002 Diesel PM Emissions Inventory for POLA and POLB ......... 5
TABLE 2 Summary of Area Impacted by Risk Levels and
Activity Categories ( Acres) ................................................................................... 9
TABLE 3 Summary of Population Affected by Risk Levels and Activity Categories .... 9
TABLE 4 Comparisons of Predicted Potential Cancer Risks with Measurements ..... 10
TABLE 5 Estimated Diesel PM Emissions per Vessel Call and 2002 Port Calls ........ 20
TABLE 6 2002 Estimated Diesel PM Emissions for the POLA and POLB… .............. 24
TABLE 7 Emission Source Model Parameters................................................................. 29
TABLE 8 Temporal Distribution of Diesel PM Emissions at POLA and POLB............ 30
TABLE 9 Summary of Area Impacted by Risk Levels and Activity
Categories ( Acres) ............................................................................................... 38
TABLE 10 Summary of Population Affected by Risk Levels Activity Categories .......... 38
TABLE 11 Comparisons of Modeling Results with Measurements ................................. 52
List of Figures
Figure 1 Estimated Diesel PM Cancer Risk from POLA and POLB.............................. 8
Figure 2 Aerial Photos of POLA and POLB ..................................................................... 18
Figure 3 Estimated 2002 Diesel PM Emissions for POLA and POLB......................... 25
Figure 4 In- Port and Out- of- Port Distribution of POLA and POLB Diesel PM
Emissions .............................................................................................................. 25
Figure 5a Model Receptor Domain for the Ports of Los Angeles and
Long Beach ........................................................................................................... 27
Figure 5b Depiction of the Emission Source Locations ................................................... 28
Figure 6 Locations of Air Quality Measurement Sites around the Ports ..................... 31
Figure 7 Annual Wind Rose at Wilmington...................................................................... 32
v
Figure 8 Wind Speed and Stability Class Frequency Distribution at the
Wilmington Meteorological Site .......................................................................... 33
Figure 9 Estimated Diesel PM Cancer Risk from All Diesel- Fueled
Engines at POLA and POLB .............................................................................. 37
Figure 10 Estimated Diesel PM Cancer Risk from Oceangoing Vessel’s
Activity at POLA and POLB ................................................................................ 40
Figure 11 Estimated Diesel PM Cancer Risk from Ship Auxiliary
Engines’ Hotelling at POLA and POLB............................................................. 41
Figure 12 Estimated Diesel PM Cancer Risk from Commercial Harbor
Craft Vessel Activity at POLA and POLB ......................................................... 42
Figure 13 Estimated Diesel PM Cancer Risk from Cargo Handling
Equipment Activity at POLA and POLB ............................................................ 43
Figure 14 Estimated Diesel PM Cancer Risk from In- Port
Heavy- Duty Trucks at POLA and POLB........................................................... 45
Figure 15 Estimated Diesel PM Cancer Risk from In- Port Locomotive
Activity at POLA and POLB ................................................................................ 46
Figure 16 Estimated Diesel PM Cancer Risk from All In- Port Diesel
Engine Activity at POLA and POLB................................................................... 47
Figure 17 Estimated Diesel PM Cancer Risk from All Out- of- Port
Diesel Activity at POLA and POLB .................................................................... 48
Figure 18 Air Quality monitoring Stations for POLA and ARB Programs...................... 53
Figure 19 Distribution of Diesel PM Emissions by Source Categories
for POLA and POLB in 2002 .............................................................................. 57
Figure 20 Population Affected within the Model Domain by Cancer Risk
Levels and Source Categories ........................................................................... 59
Figure 21 Residential Areas Impacted within the Model Domain by
Cancer Risk Levels and Source Categories .................................................... 59
Figure B- 1 PM Emissions as a Function of Fuel Sulfur from Environmental Canada .. B- 2
Figure C- 1 Locations of Meteorological Monitoring Sites
Around the Ports.................................................................................................. C- 2
Figure C- 2 Wind Rose for King Harbor Meteorological Site ............................................. C- 3
Figure C- 3 Frequency Distributions of Wind Speed and Atmospheric
Stability for King Harbor Meteorological Site .................................................. C- 4
Figure C- 4 Comparison of Estimated Diesel PM Cancer Risks from
OGV’s Activity in the Shipping Lanes outside the
Breakwater Using Wilmington and King Harbor
Meteorological Data ............................................................................................ C- 5
1
DIESEL PARTICULATE MATTER EXPOSURE ASSESSMENT
STUDY FOR THE PORTS OF LOS ANGELES AND LONG BEACH
PART I: SUMMARY
The California Air Resources Board ( ARB or Board) conducted an exposure
assessment ( study) to evaluate the impacts from airborne particulate matter emissions
from diesel- fueled engines associated with port activities at the Ports of Los Angeles
and Long Beach ( ports) located in Southern California. The purpose of the study was to
enhance our understanding of the port- related diesel particulate matter ( PM) emission
impacts by evaluating the relative contributions of the various diesel PM emission
sources at the ports to the potential cancer risks to people living in communities near
the ports. This information will assist in the efforts underway to reduce diesel PM
emissions at the ports by helping to identify the sources that have the greatest impact
on potential cancer risks to nearby residents and by providing a tool that will allow
evaluation of the impacts of measures planned and under development that are
designed to reduce diesel PM emissions.
The study focused on the on- port property emissions from locomotives, on- road heavy-duty
trucks, and cargo handling equipment used to move containerized and bulk cargo
such as yard trucks, side- picks, rubber tire gantry cranes, and forklifts. The study also
evaluated the at- berth and over- water emissions impacts from ocean- going vessel main
and auxiliary engine emissions as well as commercial harbor craft such as passenger
ferries and tugboats. For the ocean- going vessel emissions, the study evaluated the
hotelling emissions, i. e. those emissions from vessel auxiliary engines while at berth,
separately from the maneuvering and transiting emissions. While there are locomotive
and on- road heavy- duty truck emissions associated with the movement of goods
through the ports that occur off the port boundaries, these were not evaluated in this
study. Future analyses will consider the impact of these off- port emissions.
The results from the study are presented in this report which is comprised of two parts.
Part I, “ Summary,” provides an overview and summary of the study in a less technical
and more easily understood format. Part II, “ Technical Support Document,” provides a
description of the supporting technical basis for the study and a more comprehensive
summary of the results. For simplicity, the Summary is presented in question- and-answer
format. The reader is directed to Part II fo r more detailed information.
1. What are the major elements of the study?
The major elements of the study were:
· developing a baseline ( 2002) inventory of diesel PM emissions at the two ports from
ocean going vessels ( transit, maneuvering, and hotelling), harbor craft, cargo
handling equipment, in port trucks, and in port trains ,
· estimating the ambient concentration of diesel PM downwind of the ports, and
2
· estimating the potential cancer risk levels and other non- cancer health effects
associated with the diesel PM concentrations.
2. What are the key findings from the study?
The key findings from this study are:
· Diesel PM emissions from the ports are a major contributor to diesel PM in the South
Coast Air Basin.
The combined diesel PM emissions from the ports are estimated to be about
1,760 tons per year in 2002. This represents a significant component of the
regional diesel PM emissions for the South Coast Air Basin ( SCAB) at about 21
percent of the total SCAB diesel PM emissions in 2002. Focusing only on diesel
PM emissions occurring on port property or within California Coastal Waters
( CCW) 1, the emissions from ship activities ( transiting, maneuvering, and
hotelling) account for the largest percentage of emissions at about 73 percent,
followed by commercial harbor craft vessels ( 14%), cargo handling equipment
( 10 %), in- port heavy duty trucks ( 2%), and in- port locomotives ( 1%).
· Diesel PM emissions from the ports impact a large area and the associated potential
health risks are of significant concern.
Diesel PM emissions from the ports result in elevated cancer risk levels over the
entire 20- mile by 20- mile study area. In areas near the port boundaries, potential
cancer risk levels exceed 500 in a million. As you move away from the ports, the
potential cancer risk levels decrease but continue to exceed 50 in a million for
more than 15 miles.
Primary diesel PM emissions from the ports also result in potential non- cancer
health impacts within the modeling receptor domain. The non- cancer health
effects evaluated include premature death, asthma attacks, work loss days, and
minor restricted activity days. Based on this study, average numbers of cases
per year that would be expected in the modeling area have been estimated as
follows:
Ø 29 premature deaths ( for ages 30 and older), 14 to 43 deaths as 95%
confidence interval ( CI);
Ø 750 asthma attacks, 180 to 1300 as 95% CI;
Ø 6,600 days of work loss ( for ages 18- 65), 5,600 to 7,600 as 95% CI;
Ø 35,000 minor restricted activity days ( for ages 18- 65), 28,000 to 41,000 as
95% CI.
1 In 1983, the ARB established the California Coastal Waters ( CCW) boundary based on coastal
meteorology within which pollutants released offshore would be transported onshore. The development
of the boundary was based on over 500,000 island, shipboard, and coastal observations from a variety of
records, including those from the U. S. Weather Bureau, Coast Guard, Navy, Air Force, Marine Corps, and
Army Air Force ( ARB, 1982). The CCW boundary ranges from about 25 miles off the coast at the
narrowest to just over 100 miles at the widest.
3
· “ Hotelling” emissions from ocean- going vessel auxiliary engines and emissions from
cargo handling equipment are the primary contributors to the higher pollution related
health risks near the ports.
Hotelling emissions from ocean- going vessels account for about 20 percent of
the total diesel PM emissions from the ports. These emissions are responsible
for about 34 percent of the port emissions related risk in the modeling receptor
domain based on the population- weighted average risk. These emissions
resulted in the largest area ( 2,036 acres) where the potential cancer risk levels
were greater than 200 in a million in the nearby communities. The second
highest category contributing to cancer risk levels above 200 in a million was
cargo handling equipment, which impacted a residential area of 410 acres and is
responsible for about 20 percent of the total risk in the modeling receptor domain
based on the population- weighted average risk. Reducing emissions from these
two categories will have the most dramatic effect on reducing the port emissions
related risks in nearby communities.
· Emissions from commercial harbor craft, in- port trucks, in- port rail, and ocean- going
vessels ( transit and maneuvering activities) account for about 46 percent of the port
emissions related risk in the modeling receptor domain based on the population-weighted
average risk. These emissions are an important contributor to elevated
cancer risk levels over a very large area.
Emissions from commercial harbor craft, on- port trucks, on- port rail, and ocean
going vessels ( maneuvering and transit activities) account for about 70 percent of
the total diesel PM emissions for the ports. While emissions from these source
categories do not have a major role in the near port risk le vels, they are
significant contributors to the overall elevated risk levels in the study area.
Addressing the emissions from these sources is critical if we are to significantly
reduce the exposure of a large population ( over 2 million people) to cancer risk
levels in the 50 in a million range.
3. Why is ARB concerned about Diesel PM?
Diesel engines emit a complex mixture of air pollutants, composed of gaseous and solid
material. The visible emissions in diesel exhaust are known as particulate matter or
PM, which includes carbon particles or " soot.” In 1998, ARB identified diesel PM as a
toxic air contaminant based on its potential to cause cancer, premature deaths, and
other health problems. Health risks from diesel PM are highest in areas of concentrated
emissions, such as near ports, rail yards, freeways, or warehouse distribution centers.
Exposure to diesel PM is a health hazard, particularly to children whose lungs are still
developing and the elderly who may have other serious health problems.
The health impacts of particulate matter ( PM10 and PM 2.5) have been studied in
epidemiological studies conducted in many different cities. Diesel particulate matter is a
major component of particulate matter in many cities. Diesel particulate matter is
composed of carbonaceous particles ( soot) and particles that can form from nitrogen
4
A risk assessment is a tool used
to evaluate the potential for a
chemical or pollutant to cause
cancer and other illnesses.
For cancer health effects, the risk is expressed as
the number of chances in a population of a million
people who might be expected to get cancer over a
70- year lifetime. The number may be stated as “ 10
in a million” or “ 10 chances per million”. Often
times scientific notation is used and you may see it
expressed as 1 x 10- 5. or 10- 5. Therefore, if you
have a potential cancer risk of 10 in a million, that
means if one million people were exposed to a
certain level of a pollutant or chemical there is a
chance that 10 of them may develop cancer over
their 70- year lifetime. This would be 10 new cases
of cancer above the expected rate of cancer in the
population. The expected rate of cancer for all
causes, including smoking, is about 200,000 to
250,000 chances in a million ( one in four to five
people).
oxides ( NOX) emitted by diesel engines. These studies have found an increase of one
to two percent in daily mortality associated with each 10 m g/ m3 increase in PM10
exposure. The most vulnerable subpopulations are those with preexisting respiratory or
cardiovascular disease, especially the elderly. In addition, increased hospital
admissions and morbidity from respiratory disease have been associated with
particulate matter exposure in adults and children. Particulate matter exposure is
associated with an increased risk of lung cancer in epidemiological studies.
The ARB staff has estimated that 2,000 premature deaths statewide are linked to direct
diesel PM exposure and 900 premature deaths are associated with indirect diesel PM
exposure in the year 2000 alone. Exposure to fine particulate matter, including diesel
PM 2.5, can also be linked to a number of heart and lung diseases. For example, the
ARB staff has estimated that 5,400 hospital admissions for chronic obstructive
pulmonary disease, pneumonia, cardiovascular disease, and asthma were due to
exposure to direct diesel PM 2.5 in California. An additional 2,400 admissions were
linked to exposure to indirect diesel PM ( Lloyd, 2001). There are uncertainties in these
analyses, but the non- cancer public health impacts of diesel PM exposure may
outweigh the considerable public health impacts of diesel PM as a carcinogenic
substance.
4. What are exposure and risk assessments?
Risk assessment is a yardstick useful for comparing
the potential health impacts of various sources of
air pollution. For this risk assessment, the amount
of diesel PM emitted from each source ( e. g. cruise
ships) is estimated. An air modeling computer
program uses local meteorological data ( e. g. wind speed and direction) to estimate the
annual average ground level concentrations of diesel PM in the communities around the
facility. The increased risk of developing lung cancer from exposure to a particular level
of diesel PM can be estimated using
the Office of Environmental Health
Assessment’s ( OEHHA) cancer
potency factor for diesel PM. The non-cancer
health impacts of diesel PM
exposure are possible to quantify, but
the cancer health impacts have more
commonly been used as the yardstick
with which to compare the impacts of
various diesel sources. Risk
assessment has various uncertainties
in the methodology and is therefore
deliberately designed so that risks are
not under predicted. Risk assessment
is thus best understood as a tool for
comparing risks from various sources,
usually for purposes of prioritizing risk reduction, and not as literal prediction of the
community incidence of disease from exposure.
5
In a risk assessment, risk is expressed as the number of chances in a population of a
million people who might be expected to get cancer over a 70- year lifetime. However,
for informational purposes only, the risk is sometimes reported for other exposure times,
such as a 30- year or a 9- year risk. The longer the exposure to a given air
concentration, the greater the cancer risk will be. In this report, only the 70- year lifetime
risk is presented. The exposure assessment study for the Ports of Los Angeles and
Long Beach focuses on potential cancer cases due to exposure to diesel PM emissions.
However, there is a growing body of scientific data suggesting that exposure to fine PM
results in premature death and morbidity ( illness) due to respiratory and cardiovascular
disease. The sensitive subpopulations include people with pre- existing cardiovascular
disease and respiratory disease, including asthma, particularly those who are also
elderly.
5. Where are the Port of Los Angeles and the Port of Long Beach located and
what port activities occur there?
The Ports of Los Angeles ( POLA) and Long Beach ( POLB) are located adjacent to each
other on San Pedro Bay, about 20 miles south of downtown Los Angeles. Together,
they form the third- largest port complex in the world. The primary purpose of the ports
is to move cargo on and off ocean- going ships and onto trucks or railcars. The majority
of goods are transported in containers although the ports also handle non- containerized
goods such as coke and motor vehicles. These activities involve a wide variety of
sources that contribute to diesel PM and oxides of nitrogen ( NOx) emissions such as
the ocean- going ships that participate in international trade. Other sources include
trucks, locomotives, cargo handling equipment, and harbor craft such as tug boats, crew
boats, and fishing vessels.
6. What are the diesel PM emissions from port- related activities at POLA and
POLB?
The emissions of diesel PM from port- related activities were estimated to be
approximately 965 tons per year for the POLA and 795 tons per year for the POLB in
the year 2002, or a total of 1,760 tons per year for both ports. As shown in Table 1, by
source category, ocean- going vessels, ship auxiliary engines’ hotelling, harbor craft,
cargo handling equipment, in- port heavy- duty trucks, and in- port locomotives account
for about 53, 20, 14, 10, 2, and 1 percent of the mass emissions, respectively.
Table1: Estimated 2002 Diesel PM Emissions Inventory for POLA and POLB
OGV HOTEL CHC CHE IPT IPL COMBINED
Diesel PM
Emissions
T/ Y
942 343 244 172 41 18 1760
Percent of
Total
53% 20% 14% 10% 2% 1% 100%
Note: OGV – Oceangoing vessels; HOTEL – Ship’s auxiliary engine hotelling; CHC – Commercial harbor crafts; CHE – Cargo
handling equipment; IPT – In- Port heavy- duty trucks; IPL – In- Port locomotive.
6
By source area, about 43 percent of the emissions occur on land - based port property
and over the water within the breakwater2 and the remaining ( 57 percent) occur outside
of the breakwater over water. These emissions estimates include only the emissions
that are occurring on port property and the over- water emissions from ocean- going
ships. It does not include the more regional land - based emissions from trucks and
locomotives that occur outside of the port boundaries.
The diesel PM emissions resulting from port activities have been a significant and
growing contributor to regional air pollution and community exposure to toxic air
pollutants. For example, in the South Coast Air Basin ( SCAB), the diesel PM emissions
resulting from the ports activities accounted for about 21 percent of the total SCAB
diesel PM emissions in 2002. Growth forecasts predict that trade at the POLA and
POLB will triple by 2020, resulting in a 60 percent increase in diesel PM emissions from
current levels unless further controls are enacted.
7. How were the diesel PM concentrations near the ports estimated?
ARB staff used the United States Environmental Protection Agency ( U. S. EPA)
approved computer model ( ISCST3) to estimate the annual average offsite
concentration of diesel PM resulting from the activity at the two ports. The key inputs to
the computer model were the diesel PM emissions information ( magnitude, timing, and
location), the meteorological data ( wind speed, direction, etc.), and the dispersion
coefficients ( rural or urban). Meteorological data, used as a direct input to the
dispersion model, are obtained from an air quality monitoring study conducted in
Wilmington in 2001. The meteorological observations were located about one mile from
the north boundary of the Port of Los Angeles. These data are the most recent and
most representative meteorological data for the dock areas of the Ports of Los Angeles
and Long Beach. Because the area surrounding the ports has urban characteristics,
the modeling was done using the urban dispersion coefficients.
8. How were the potential cancer risks from diesel PM estimated?
The potential cancer risks were estimated using standard risk assessment procedures
based on the annual average concentration of diesel PM predicted by the model and a
health risk factor ( referred to as a cancer potency factor) that correlates cancer risk to
the amount of diesel PM inhaled.
The methodology used to estimate the potential cancer risks is consistent with the
Tier- 1 analysis presented in OEHHA’s Air Toxics Hot Spots Program Guidance Manual
for Preparation of Health Risk Assessments ( September 2003). A Tier- 1 analysis
assumes that an individual is exposed to an annual average concentration of a pollutant
2 The breakwater protects POLA and POLB Harbor from rough seas and waves. The breakwater is about
nine miles long ( east- west) and was built in a pyramid shape with rocks from Catalina Island. The bottom
on the ocean floor is 200 feet wide and the top is only 23 feet wide. Construction of the breakwater
began in 1899 and took 50 years to complete. The breakwater is approximately 4.5 miles from the ports’
north land boundary.
7
continuously for 70 years. 3 The cancer potency factor was developed by the OEHHA
and approved by the State’s Scientific Review Panel on Toxic Air Contaminants ( SRP)
as part of the process o f identifying diesel PM emission as a toxic air contaminant
( TAC).
9. What is the estimated potential cancer risk from all sources at the ports?
Figure 1 shows the potential cancer risk isopleths for all emission sources at the two
ports superimposed on a map showing the ports and the nearby communities. The risk
contour of 100 in a million extends beyond the modeling receptor domain to the north of
the ports. The domain boundary is about 10 miles north of the port boundary. The area
with predicted cancer risk levels in excess of 100 in a million is estimated to be about
94,000 acres, which is 57 percent of the effective land area ( 163,400 acres, excluding
the port property and the water acreage) within the modeling receptor domain. The
area in which the risks are predicted to exceed 200 in a million is also very large,
covering an area of about 29,000 acres ( 18 percent of the effective land area within the
modeling receptor domain). The areas with the greatest impact have an estimated
potential cancer risk of over 500 in a million and cover about 2 percent of the effective
land area within the domain. The risk isopleths of 1000 and 1500 in a million occur on
the ports’ property and the nearby ocean surfaces, and are not considered in this study
as people do not reside in these areas.
Using the U. S. Census Bureau’s year 2000 cens us data, we estimated the population
within the isopleth boundaries. Nearly 60 percent of the 2 million people that live in the
area around the ports have predicted risks of greater than 100 in a million. The affected
population numbers for the cancer risk ranges of 100- 200, 200- 500, and over 500 have
been estimated to be about 724,000 people, 360,000 people , and 53,000 people,
respectively. The affected population numbers account for about 37, 18 and 3 percent
of the total population within the modeling receptor domain, respectively. Note that the
risk isopleth of 10 in a million is not shown in Figure 1 because it is outside of the
modeling receptor domain. Also, note that if the modeling receptor domain expands,
the impacted areas and affected popula tion would be increased.
3According to the OEHHA Guidelines, the relatively health- protective assumptions incorporated into the
Tier- 1 risk assessment make it unlikely that the risks are underestimated for the general population.
8
Figure 1
Estimated Diesel PM Cancer Risk from POLA and POLB
Notes: Wilmington Meteorological Data, Urban Dispersion Coefficients, 80th Percentile
Breathing Rate, Emission = 1,760 TPY, Modeling Receptor Domain = 20 mi x 20 mi,
Resolution = 200 m x 200 m.
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
9
10. What are the relative contributions to the potential cancer risks from the
various diesel PM emission sources at the ports?
The different emission sources are used at various locations on the ports property.
Thus, contributions of these emission sources to exposures in the nearby
neighborhoods are different. As shown in Tables 2 and 3, the emissions from cargo
handling equipment and on- port heavy- duty trucks resulted in areas within the nearby
communities having risk levels exceeding 500 in a million while the highest risk levels
associated with the other categories were between 200 and 500 in a million. Within the
modeling receptor domain, ship hotelling emissions and cargo handling equipment
impacted the largest areas and affected more people than the other sources of
emissions when considering the risk levels greater than 100 in a million. When
considering risk levels greater than 10 in a million, all the port sources, other than in-port
heavy- duty trucks and locomotives, had similar impacts, affecting at least 119,000
acres and at least 1.4 million people. By source location, the impacts resulting from the
in- port emissions ( within the breakwater) are much larger than those resulting from the
out- of- port emissions ( outside the breakwater), although the emission magnitude of the
former is less than the latter ( 750 TPY vs 1010 TPY). Quantitatively, within the
modeling domain, the population- weighted risk resulting from the in- port emissions is
about 4.5 times greater than the risk resulting from the over water out- of- port
emissions.
Table 2: Summary of Area Impacted by Risk Levels and Activity Categories
( Acres)
Risk Level OGV HOTEL CHC CHE IPT IPL COMBINED
Risk > 500 0 0 0 50 50 0 2,500
Risk > 200 110 2,036 20 410 160 40 29,000
Risk > 100 227 12,700 750 4,100 376 160 94,000
Risk > 10 163,435 160,470 125,250 119,000 29,750 11,240 163,435
Table 3: Summary of Population Affected by Risk Levels and Activity Categories
( Number of People)
Risk Level OGV HOTEL CHC CHE IPT IPL COMBINED
Risk > 500 0 0 0 3,200 205 0 53,000
Risk > 200 18 46,020 5,000 11,100 1,780 680 411,200
Risk > 100 1,810 221,567 22,960 82,000 8,270 4,330 1,135,000
Risk > 10 1,977,760 1,949,850 1,516,515 1,444,000 422,910 213,430 1,977,770
Notes:
1. OGV – Oceangoing vessels; HOTEL – Ship’s auxiliary engine hotelling; CHC – Commercial harbor crafts; CHE – Cargo
handling equipment; IPT – In- Port heavy- duty trucks ; IPL – In- Port locomotive.
2. The model receptor domain of 20- mile x 20- mile with urban dispersion coefficients with a receptor resolution of 200m x 200m
was used. The effective receptor modeling domain ( excluding the port properties and the ocean water) is estimated to be about
255 square miles; The calculations here are ONLY based on the effective modeling receptor domain.
3. The 80th percentile breathing rate for adults over 70- year lifetime was assumed,
4. Meteorological data from Wilmington ( 2001) was used for POLA and POLB.
5. The risks within both ports and over the ocean water were excluded for calculations of average risks and affected areas .
6. The estimated population in this Table is ONLY based on the modeling receptor domain using the U. S. Census Bureau’s year
2000 census data.
7. If the modeling receptor domain expands, the population and area affected would be increased.
8. The combined column provides the population affected and area impacted for the cumulative impacts from all the emission
sources. The individual impacts are not additive since the combined impacts are greater than the sum of the individual
sources. For example, cargo handling equipment and commercial harbor craft emissions may impact the same location and
population. While individually the impacts may result in cancer risk levels between 100 and 200 in a million, when you combine
the impacts, the resulting risks could be greater than 200 in a million.
10
11. How do the results compare to the monitoring programs and the SCAQMD
MATES- II study
For comparison purposes, the ARB staff compared the study results to two monitoring
programs conducted by the POLA ( POLA, 2005) and ARB ( ARB, 2002) and to the
South Coast Air Quality Management District ( SCAQMD)’ s second Multiple Air Toxic
Exposure Study ( MATES- II ( SCAQMD, 2000).
The POLA is currently conducting an air quality monitoring program within the Port and
in the nearby communities to estimate the ambient levels of diesel PM in proximity to
the Port that are due to Port operational activities. For the comparison, the measured
elemental carbon ( EC) is used as the surrogate of diesel PM and it is assumed that the
ratio of EC with diesel PM is 0.5. Table 4 shows the potential cancer risks based on the
modeling results compared to those derived from the half year’s monitoring results
conducted during the period February 1 through August 5, 2005 at Wilmington
community and San Pedro monitoring stations. The computer modeling performs
adequately in simulating the measured diesel PM risks at the two locations.
The ARB conducted an air monitoring program in Wilmington from May 2001 to July
2002 as part of the Children’s Environmental Health Program. The derived potential
diesel PM cancer risks at two sites - Wilmington Park Elementary School and Hawaiian
Elementary School are also listed in Table 4 and compared with the predicted risks. It
is shown that the predicted results are favorably comparable with the monitored results
at the two sites.
Table 4. Comparisons of predicted potential cancer risks with measurements
( unit in cases per million)
Location Port of L. A.
monitoring results
ARB SB 25
monitoring results
Model prediction
Wilmington Community 585 N/ A 600
San Pedro 533 N/ A 500
Wilmington School N/ A 450 470
Hawaiian School N/ A 710 650
Note:
1. The ratio of elemental carbon ( EC) with diesel PM has been reported to be 0.375 to 0.75 by
literature. A ratio of 0.5 is used in this calculation;
2. For POLA’s monitoring program, the measured EC 24- hr average concentrations over the
half year from February 9 to August 5, 2005 are reported;
3. For ARB SB 25 Wilmington monitoring study, about 71% of the samples collected were below
the detection limit of 1 ug EC/ m3. It is assumed that all measurements below the limit are
arbitrarily assumed to be 0.5 ugEC/ m3;
4. For the detailed monitoring programs and results, please check POLA and ARB’s web sites.
The ARB staff also compared the study results to the SCAQMD’s MATES- II. The
MATES- II study indicated that the modeled potential risk in the grid cell containing the
Wilmington air quality monitoring station was 1,187 potential cancer cases per million
due to diesel PM emissions from port activities, freeways, and other sources of diesel
PM. This Wilmington grid cell is approximately 2 miles north of the Ports. Our study
11
shows a risk level of about 450 cases in a million in the same general vicinity. In the
nearby residential areas within one mile from port boundaries, risk levels ( from diesel
PM emissions as well as other toxics) ranged from 1000 to 1500 cases in a million
based on the MATES- II study. Our study shows a risk range of 500 to 1000 cases in a
million from the Ports’ diesel PM emissions. The differences can be attributed to
different modeling configurations. For example, MATES- II used the Urban Airshed
Model ( UAM) model, a grid based model with 2 km grid cells, while our study used the
ISCST3 model, a Gaussian plume model. In addition, MATES- II simulated diesel PM
from all sources ( e. g., port activities and freeway emissions) for the 1998 base year
while our study was limited to diesel PM from port activities for the year 2002. Also the
MATES- II study released ocean- going emissions near ground level ( within the first
horizontal layer of the UAM). Our study released ocean going emissions at 50 meters
above " ground" ( sea level) which will result in greater dispersion of emissions. 4
12. What are the uncertainties associated with risk assessments?
The estimated diesel PM concentrations and risk levels produced by a risk assessment
are based on a number of assumptions. Many of the assumptions are designed to be
health protective so that potential risks to individuals are not underestimated.
Therefore, the actual risk calculated by a risk assessment is intentionally designed to
avoid underprediction. There are also many uncertainties in the health values used in
the risk assessment. Some of the factors that affect the uncertainty are discussed
below.
When available, as is the case with diesel PM, scientists use studies of people exposed
at work to estimate risk from environmental exposures. There can be a wider range of
responses in the general public than in the workers in the epidemiology study used to
determine the cancer potency factor. Also, the actual worker exposures to diesel PM
were based on limited monitoring data and were mostly derived based on estimates of
emissions and duration of exposure. Different epidemiological studies suggest
somewhat different levels of risk. When the State’s Scientific Review Panel ( SRP) 5
identified diesel PM as a toxic air contaminant, they endorsed a range of inhalation
cancer potency factors ( 1.3 x 10 – 4 to 2.4 x 10 – 3 ( μg/ m3) – 1) and a risk factor of 3x10 - 4
( μg/ m3)- 1, as a reasonable point estimate of the unit risk. From the unit risk factor an
inhalation cancer potency factor of 1.1 ( mg/ kg- day)- 1 may be calculated.
As mentioned above, there is no direct measurement technique for diesel PM. This
analysis used an air dispersion modeling to estimate the concentrations to which the
public is exposed. The air dispersion models are based on the state - of- the- science
4 The higher release point was used because the average ship stack height is about 43 m tall. When the
emissions are released from the top of a ship’s exhaust stack, there is a plume rise that occurs which was
estimated to average to be about 7 meters. This results in an average release height of 50 meters.
5 The Scientific Review Panel ( SRP/ Panel) is charged with evaluating the risk assessments of substances
proposed for identification as toxic air contaminants by the Air Resources Board ( ARB) and the
Department of Pesticide Regulation ( DPR). In carrying out this responsibility, the SRP reviews the
exposure and health assessment reports and underlying scientific data upon which the reports are based,
which are prepared by the ARB, DPR, and the Office of Environmental Health Hazard Assessment
( OEHHA) pursuant to the sections 39660- 39661 of the Health and safety Code and sections 14022.
12
formulations which have uncertainties. Three air dispersion models – ISC, AERMOD,
and CALPUFF, could be used in this study. As stated above, the primary propose of
this study was to prioritize emission sources/ categories from the Ports operation which
are to be regulated. ISC was used in this study because of its fewer requirements for
the model inputs. Although AERMOD or CALPUFF may predict somewhat different
impacts in the nearby communities, we believe that the conclusions d rawn from this
study, especially the ranking of the emission sources/ categories, may not be altered.
This is because that each model assumes that the concentration is linearly proportional
to emission rate, thus, the relative contributions or prioritization scheme of each
emission source/ category to the total impacts in the nearby communities would not be
affected.
The model inputs included emission rates, release parameters, meteorological
conditions, and dispersion coefficients. Each of the model inputs has an uncertainty of
their own. In addition, a relative small model domain of 20 mi x 20 mi was used in this
study because of the ISC model’s limitation. In reality, the impacts of diesel PM from
the Ports in the nearby communities could exceed the domain. Fully impacts of diesel
PM from the Ports could be addressed using the long range transport model –
CALPUFF in future time.
13. What are the non- cancer health endpoints associated with exposures to
Diesel PM from port operations?
A substantial number of epidemiologic studies have found a strong association between
exposure to ambient particulate matter ( PM) and adverse health effects ( CARB, 2002).
As part of this study, ARB staff conducted an analysis of the potential non- cancer health
impacts associated with exposures to the model- predicted ambient levels of directly
emitted diesel PM ( primary diesel PM) within the modeling domain. The non- cancer
health effects evaluated include premature death, asthma attacks, work loss days, and
minor restricted activity days.
ARB staff assessed the potential non- cancer health impacts associated with exposures
to the model- predicted ambient levels of directly emitted diesel PM ( primary diesel PM)
within each 200 meter by 200 meter grid cell within the modeling domain. The
populations within each grid cell were determined from U. S. Census Bureau year 2000
census data. Using the methodology peer- reviewed and published in the Staff Report:
Public Hearing to Consider Amendments to the Ambient Air Quality Standards for
Particulate Matter and Sulfates, ( PM Staff Report) ( CARB, 2002), we calculated the
number of annual cases of death and other health effects associated with exposure to
the PM concentration modeled for each of the grid cells and then calculated the to tals
over the entire modeling area. Based on our analysis, it is estimated that the exposures
to the directly emitted diesel PM from on- port operations within the modeling domain
result in approximately 29 premature deaths for the 2 million people exposed per year.
In addition, these exposures are predicted to result in 750 asthma attacks, 6,600 work
loss days, and approximately 35,000 minor restricted activity days. In each case, the
values presented represent the mean value in cases per year for the health end point
listed.
13
These estimates are based on a well- established methodology for calculating changes
in health endpoints due to changes in air pollution levels. However, since the estimates
apply to a limited modeling domain ( 20 miles by 20 miles), the affected population is
small, and hence the overall estimated health impacts are smaller than estimates made
on a statewide basis. In addition, to the extent that only a subset of health outcomes is
considered here, the estimates should be considered an underestimate of the total
public health impact.
In this study, we also did not consider the diesel PM emissions of on- road heavy- duty
trucks and locomotives related to port activities that occur off- port boundary within the
SCAB ( regional emissions). We estimate the off- port regional diesel PM emissions to
be about 206 TPY for the both ports, or 10 percent of the total port- related emissions
( 206 TPY vs 1,970 TPY). These regional emissions are distributed throughout the
SCAB and may result in localized health impacts to people who are live near freeways
and railroad corridors within the SCAB. These health impacts will be evaluated in future
studies.
14. Are there other studies planned that will evaluate the impacts of port-related
diesel PM emissions?
As mentioned above, during 1998 - 1999, the South Coast Air Quality Management
District ( SCAQMD) conducted the second Multiple Air Toxics Exposure Study
( MATES- II) to determine the Basin- wide risks associated with major airborne
carcinogens, including diesel PM. Currently, SCAQMD is conducting MATES- III to
assess current air toxics levels within the Air Basin using updated emission inventories,
refined modeling methodologies, and improved assumptions. MATES- III will
incorporate all air toxic emission sources, e. g., stationary, on- road, and off- road mobile
sources, and all air toxics, e. g., diesel PM, 1,3- butadiene, benzene, chromium, etc. In
addition, ARB is conducting a neighborhood assessment study for Wilmington, which is
nearby the ports. This study is a part of ARB’s Neighborhood Assessment Program.
The objective is to estimate health risks in Wilmington and surrounding areas. Like
MATES- III, this project will consider all emission sources and all air toxic contaminants.
15. What activities are underway to reduce risks?
There are many efforts currently underway to reduce exposures to diesel PM. POLA
and POLB have instituted voluntary programs to reduce diesel PM emissions from port
operations including installation of diesel oxidation catalysts on yard equipment, funding
the incremental costs of cleaner fuels, cold- ironing of ocean- going ships and providing
monetary support to the Gateway Cities truck fleet modernization program. In addition,
efforts at the State and local level to implement the Diesel Risk Reduction Plan and to
fulfill commitments in the State Implementation Plan will also reduce emissions. For
example, the new off- road engine standards adopted by ARB and the U. S. EPA will
reduce emissions from new off- road engines by over 95% compared to uncontrolled
levels. In the fall of 2005, ARB has considered two measures to reduce emissions from
sources of diesel emissions at ports. One measure will require reductions from cargo
handling equipment and the other from ship auxiliary engines. To ensure continued
emission declines in the face of the expected growth, ARB is leading an effort to
14
develop a Port and Intermodal Goods Movement Comprehensive Emission Reduction
Plan that will build upon current efforts and define the additional strategies needed to
reduce public health impacts from port and related activities. This effort is part of
Governor Schwarzenegger’s Goods Movement Action Plan, a plan that reflects the
Governor’s desire to improve the movement of goods in California at the same time we
work to improve air quality and protect public health.
15
PART II: TECHNICAL SUPPORT DOCUMENT
I. INTRODUCTION
Emissions from port- related goods movement are a significant and growing contributor
to community air pollution. In communities with significant goods movement activity,
such as communities located adjacent to California maritime ports, a particular concern
is exposure to diesel particulate matter ( diesel PM). This pollutant poses a lung cancer
hazard for humans and causes non- cancer respiratory and cardiovascular effects that
increase the risk of premature death ( ARB, 1998a). The particles are readily inhaled
because of their small size and can effectively reach the lowest airways of the lung.
Many of the adsorbed compounds are known or suspected mutagens and carcinogens.
( ARB, 2002)
To better understand the impacts from port activities, Air Resources Board ( ARB) staff
conducted an exposure assessment study of diesel PM emissions from port- related
activities at the Ports of Los Angeles and Long Beach ( ports) located in Southern
California. This part provides the technical details on the exposure assessment. The
reader is directed to Part I, Summary, for a less technical discussion of the study.
A. Overview
Risk assessment is a complex process that requires the analysis of many variables to
model real- world situations. Three steps were taken to perform the exposure
assessment for the ports:
· developing a diesel PM emissions inventory that reflects the amount of diesel PM
released annually from port- related activities;
· conducting air dispersion modeling to estimate the ambient concentration of diesel
PM that results from these emissions; and
· estimating the potential cancer risk from the modeled exposures.
The following chapters provide a description of each element of the exposure
assessment. Specifically, the following information is provided:
· the methodology used to develop the port- related diesel PM emissions;
· a summary of the estimated diesel PM emissions inventory for the ports;
· a discussion on the air dispersion modeling conducted to estimate ambient
concentrations of diesel PM;
· the results of the air dispersion modeling;
· an estimate of the potential impacts ( potential cancer risks) to nearby residences
due to exposure to ambient concentrations of diesel PM from port- related activities
at the ports; and
· a comparison between the risk impacts from the various emission sources at the
ports.
16
B. Purpose
In the South Coast Air Basin ( SCAB), diesel PM emissions from port- related activities
are a significant and growing contributor to regional air pollution and community
exposures to toxic air pollutants. For example, in the SCAB, the diesel PM emissions
resulting from the movement of goods through the Ports of Los Angeles ( POLA) and the
Port of Long Beach ( POLB) accounted for about 21 percent of the total SCAB diesel PM
emissions in 2002. Growth forecasts predict that trade at POLA and POLB will triple by
2020, resulting in a 60 percent increase in diesel PM emissions from current levels
unless further controls are enacted. POLA and POLB operate in close proximity to
several communities including San Pedro, Long Beach, and Wilmington. These nearby
communities face potentially higher health risks from the port- generated diesel PM
emissions.
There are many efforts currently underway to reduce exposures to diesel PM. POLA
and POLB have instituted voluntary programs to reduce diesel PM emissions from port
operations including installation of diesel oxidation catalysts on yard equipment, funding
the incremental costs of cleaner fuels, cold- ironing of ocean- going ships, and providing
monetary support to the Gateway Cities truck fleet modernization program. In addition,
efforts at the State and local level to implement the ARB Diesel Risk Reduction Plan
and to fulfill commitments in the State Implementation Plan will also reduce emissions.
New off- road engine standards adopted by ARB and the United States Environmental
Protection Agency ( U. S. EPA) will reduce emissions from new off- road engines by over
95% compared to uncontrolled levels. In the fall of 2005, ARB has considered two
measures to reduce emissions from port sources. One measure will require reductions
from cargo handling equipment and the other from ship auxiliary engines. To ensure
continued emission declines in the face of the expected growth, ARB is leading an effort
to develop a Port and Intermodal Goods Movement Comprehensive Emission
Reduction Plan that will build upon current efforts and define the additional strategies
needed to reduce public health impacts from port and related activities.
The purpose of this exposure assessment study is to enhance our understanding of the
port- related diesel PM emission impacts on communities near POLA and POLB and to
assist in the evaluation of control measures under development or planned. Because
the emission sources are located at various locations on the port property, the
contributions of these emission sources to nearby neighborhoods will be different. Both
the location of the emissions and the magnitude need to be taken into consideration
when determining the degree of health risks to people who are living around the ports.
To summarize, the purpose of the exposure assessment is to:
§ investigate the impacts of the various port emission sources on nearby
neighborhoods;
§ identify the most significant emission source( s);
§ prioritize possible mitigation measures to control diesel PM emissions based on the
relative magnitude of health risks; and
§ assist in evaluating the impacts of measures developed to reduce emissions.
17
C. Description of the Ports
POLA and POLB are located adjacent to each other on the San Pedro Bay, about 20
miles south of downtown Los Angeles. The ports are directly adjacent to the
communities of Long Beach, San Pedro, and Wilmington. The ports are primarily
container ports, moving goods into and out of California in containers. However, they
also handle non- containerized goods such as coke and automobiles. While the majority
of the goods movement occurs during the day, the ports do operate 24 hours a day, 7
days a week, and 365 days a year. The ports are the first and second busiest seaports
in the Western United States. POLA encompasses 7500 acres, 43 miles of waterfront
and features 26 cargo terminals. These terminals handle nearly 150 million metric
revenue tons of cargo annually. In 2004, the POLA moved in 7.4 million TEUs 1, which
was a new national container record. POLB covers about 3000 acres of land. In 2004,
tonnage through POLB was 73.6 million metric tons, and about 5.8 million TEUs moved
through the Port. Combined, POLA and POLB are the world’s third- busiest port
complex, after Hong Kong and Singapore.
1 The TEU is the international standard measure used to describe containers. A 20- foot container =
1 TEU.
18
2a. Port of Los Angeles ( Courtesy of POLA, http:// www. portoflosangeles. org)
2b. Port of Long Beach ( Courtesy of POLB, http:// www. polb. com)
Figure 2: Aerial Photos of POLA and POLB
19
II. EMISSION INVENTORY DEVELOPMENT
Air dispersion models require emission inputs that properly characterize source- specific
emissions for diesel PM from various activities in the ports. The port- related activities
are categorized as: ocean- going vessels, auxiliary engine hotelling, commercial harbor
craft, cargo handling equipment, railroad locomotives, and heavy- duty trucks. POLA
and POLB recently hired Starcrest Consulting Group, LLC ( Starcrest) to develop
detailed emission inventories for all emission sources for POLA and three sources
( cargo handling equipment, in- port locomotives, and in- port heavy- duty trucks) for the
POLB. At the request of the ports, Starcrest used 2001 as the base year for POLA and
2002 as the base year for POLB. For this exposure assessment study, 2002 was
chosen as the baseline year for both ports. In this chapter, we briefly describe how we
projected the 2001 POLA emission inventory to 2002 and how we developed the 2002
emissions inventory for ocean- going ships, auxiliary engine hotelling and commercial
harbor craft for POLB. The basic methodologies used in the emission inventory
development are briefly described in Appendix A.
A. Port of Los Angeles
As stated above, Starcrest prepared an emission inventory for all emission sources at
the POLA using 2001 as the baseline year. ( Starcrest, 2004a) The inventory utilizes an
activity- based approach and focuses on emissions of diesel PM for all significant
sources operating in the Port. In addition to in- port activities, emissions from railroad
locomotives and on- road trucks transporting port cargo were also estimated based on
the activity that occurs outside the Port, but within the South Coast Air Basin
boundaries. Only in- port emissions and over water emissions from ocean- going ships
and harbor craft were evaluated in this exposure assessment. Our methodology for
projecting the 2001 POLA inventory to 2002 is presented below.
Ocean- going Vessels
For 2001, Starcrest estimated emissions from ship cruising ( includes transiting and
maneuvering) and hotelling activities. To estimate the 2002 POLA emissions, ARB staff
assumed that the emissions per vessel call would be the same in 2001 and 2002.
Emissions per vessel call were calculated from the emissions per vessel call ( expressed
in emissions/ call number) for each ocean- going vessel ( OGV) type ( i. e. auto carrier,
bulk, container, cruise, general cargo, reefer, RoRo, tanker) reported in the 2001 POLA
emission inventory data. Emissions per vessel call were estimated for each activity
( transiting, maneuvering, hotelling). ARB staff then estimated the emissions for each
OGV type in 2002 by multiplying the emissions per call in 2001 by the number of vessel
calls for each of the corresponding OGV types in 2002, that is:
20
xCNPOLA i
CNPOLA i
EPOLA i
EPOLA i , 2002,
, 2001,
, 2001,
, 2002, = ( 1)
where EPOLA, 2002, i is the estimated emissions of OGV type i ( i = 1, 10) in 2002, E POLA, 2001, i
is the emission of OGV type i at POLA for 2001 ( known), CNPOLA, 2001, i and CNPOLA, 2002, i
are the vessel call numbers from POLA in 2001 and 2002 for OGV type i, respectively.
Table 5 provides a summary of the estimated emissions per vessel call and the actual
vessel call numbers for each port in 2002.
Table 5: Estimated Diesel PM Emissions per Vessel Call and 2002 Port Calls
POLA 2001 Diesel PM Emissions
Vessel Type Per Vessel Call ( T/ Y- CALL)
Transit* Auxiliary –
Transit
Auxiliary –
Hotelling
POLA 2002
Vessel
Calls
POLB 2002
Vessel
Calls
Auto 0.0904 0.0055 0.011 154 109
Bulk 0.0887 0.0039 0.0374 86 453
Container 0.2019 0.0109 0.0581 1,673 1,304
Cruise 0.2675 0.065 0.0975 257 36
General Cargo 0.0807 0.0047 0.0234 158 126
Miscellaneous 0.0875 0 0.1143 3 207
Other Tug 0.0353 0 0 70 51
Tanker 0.0942 0.0058 0.0986 341 546
* Transit includes both transiting and maneuvering emissions. Vessel call estimates provided by
POLA and POLB.
Adjustments to the hotelling emissions were also made based on additional data
obtained subsequent to release of the Starcrest inventories. Specifically, corrections
were made to the emission factor for auxiliary engines running on heavy fuel oil ( HFO).
In addition, the assumption on the percentage of engines running on HFO and marine
distillate was modified to reflect new data obtained in an ARB survey conducted in
2004. ( ARB, 2004) With respect to the emission factor, for ship auxiliary engines,
Starcrest utilized a single diesel PM emission factor of 0.3 g/ kW- hr in calculating
auxiliary engine emissions, regardless of diesel fuel type. Based on a review of
published emissions data, the emission factor for HFO should be much higher. In U. S.
EPA’s 2002 “ Commercial Marine Emission Inventory Development” report prepared by
ENVIRON International Corporation, an emission factor of 1.74 g/ kW- hr is reported for
engines running on HFO with a 3% sulfur content. ( Environ, 2002) ARB staff adjusted
this emission factor to 1.5 g/ kW- hr based on the average sulfur content of HFO reported
as being used in the 2004 ARB survey and retained the 0.3 g/ KW- hr factor for auxiliary
21
engines operating on marine distillate. 2 ( See Appendix B.) Starcrest also assumed that
50% of the auxiliary engines were operating on HFO and 50% on marine distillate.
ARB’s survey results established that 75 percent of the auxiliary engines use HFO and
25 percent use marine distillate. These two modifications resulted in increasing the
hotelling emissions by a factor of 4 over the estimates that would have resulted from
growing the Starcrest values to 2002 based on the number of ship calls.
Cargo Handling Equipment
To project the emissions inventory for cargo handling equipment from 2001 to 2002, we
estimated the annual growth factors by interpolating between the 2001 baseline year
and the reported 2005 emissions developed for the No Net Increase ( NNI) Task Force
Project. We assumed linear growth between 2001 and 2005. The emissions for cargo
handling equipment developed for the NNI project for 2005 reflect both the impacts from
adopted control measures and any growth that has occurred in activity. This resulted in
a net annual average growth rate of about 4.5%.
In addition, the emissions for cargo handling equipment were further modified to reflect
emission inventory adjustments that ARB staff have developed to support a 2005 rule -
making for cargo handling equipment. These adjustments result in about a 34%
decrease in the emissions from cargo handling equipment for the year 2002. The main
inventory changes to the OFFROAD model methodology used to estimate emissions
from cargo handling equipment include: ( 1) revising zero hour emission factors, and
( 2) revising equipment useful life, based on the data provided in a 2004 ARB Cargo
Handling Equipment Survey ( ARB, 2004). The zero hour emission factors are revised
by calculating composite emission factors based on the percentages of off- road,
on- road, and retrofitted equipment. Because on- road and retrofitted engines generally
have lower emission factors than off- road engines, these revisions resulted in lower
zero hour emission factors. The useful life of the equipment is used to calculate the rate
that the emissions increase over the life of the equipment. The 2004 ARB CHE Survey
results showed that CHE equipment useful lives are significantly longer than the useful
lives used in the OFFROAD Model. Since the deterioration rate is calculated as a
percentage of the zero hour emissions divided by the useful life, the revised
deterioration rates are lower than the original deterioration rates used in the OFFROAD
Model. Because both the zero hour emission factor and the deterioration rate are lower
than those used in OFFROAD Model, the resultant emissions for cargo handling
equipment are lower than those previously predicted by the OFFROAD Model for use in
the 2001 POLA emission inventory.
2 In July 2002, the European Commission published, “ Quantification of Emission from Ships Associated
with Ship Movements between Ports in the European Community” ( Entec Report). The Entec report
recommended an emission factor of 0.8 g/ kW- hr for auxiliary engines operating on HFO. ARB staff
believes this emission factor would result in an underestimation of diesel PM emissions. Applying U. S.
EPA’s methodology to estimate emissions of sulfate PM from diesel- fueled engines to an auxiliary engine
operating on 2.5% sulfur HFO would generate 0.8g/ kW- hr of sulfate PM alone. Because there are many
other components of PM such as ash and semi- volatile compounds, the 0.8 g/ kW- hr emission factor
appears to only account for the sulfate PM that is generated.
22
Harbor Craft, In- Port Heavy- duty Trucks, and In- Port Locomotives
To project the emissions inventory for commercial harbor craft, in- port trucks, and in-port
locomotives from 2001 to 2002, we estimated the annual growth factors by
interpolating between the 2001 baseline year and the reported 2005 emissions
developed for the No Net Increase ( NNI) Task Force Project. We assumed linear
growth between 2001 and 2005 for each source category. The emissions of each
category developed for the NNI project for 2005 reflect both the impacts from adopted
control measures and any growth that has occurred in activity. The resulted net annual
average growth rates are 0.0, - 6.0, and 11.0 percent for commercial harbor craft, in- port
heavy- duty trucks, and in- port locomotives, respectively.
B. Port of Long Beach
For POLB, Starcrest developed emission inventories for three categories: cargo
handling equipment, in- port locomotives, and in- port heavy- duty vehicles using 2002 as
the base year. The methodologies used in estimating emissions for these categories
are similar to those used in estimating corresponding emission inventories for the
POLA. To complete the emission inventories for POLB, ARB staff used the
methodologies described below to estimate the emissions for ocean- going vessels
( transiting, maneuvering, and hotelling) and commercial harbor craft vessels.
Ocean- going Vessels
To estimate emissions from ocean- going vessels for POLB, ARB staff assumed that the
emissions per vessel call for each OGV type in POLB in 2002 is the same as that for the
corresponding OGV type from POLA in 2001 ( see Table 5 ). The emissions for each
OGV type calling on POLB in 2002 are estimated by multiplying the emissions per call
by the number of vessel calls for the corresponding OGV type at POLB in 2002, that is:
xCNPOLB i
CNPOLA i
EPOLA i
EPOLB i , 2002,
, 2001,
, 2001,
, 2002, = ( 5)
where EPOLB, 2002, i is the estimated emission of OGV type i at POLB for 2002, E POLA, 2001, I
is the emission of OGV type i at POLA for 2001 ( known), CNPOLA, 2001, i and CNPOLB, 2002, i
are the call numbers from POLA in 2001 and from POLB in 2002 for OGV type i,
respectively.
Cargo Handling Equipment
Consistent with the approach used to adjust the POLA cargo handling equipment
emissions inventory, POLB 2002 cargo handling equipment inventory was decreased by
34 percent to reflect the inventory updates to the methodology used to estimate
23
emissions from cargo handling equipment. ( See discussion provided under A. Port of
Los Angeles.)
Harbor Craft
To estimate emissions from harbor craft vessels operating at POLB, ARB staff used the
estimates of emissions from harbor craft vessels from ARB’s 2004 commercial harbor
craft emission inventory. These emission estimates were based on information on
vessels registered ( California Department of Fish and Game), permitted ( California
Public Utilities Commission), or documented ( U. S. Coast Guard) with a “ home port”
listed as “ Long Beach.” These vessels registered as “ Long Beach” were then allocated
to the nine categories ( commercial fishing, charter fishing, ferries/ excursion, crew and
supply, pilot, tugs, tows, work boats, and others) using the harbor craft vessel
composition developed in ARB’s 2003 Commercial Harbor Craft Survey ( released in
2004). The emissions of each category for POLB in 2004 were estimated using the
emission density ( emission/ per vehicle per category) multiplied by the corresponding
vessel number in each category, that is:
( )
, 2004 ( , ) , 2004 ( , )
, 2004 ,
, 2004
2
1
9
1
x NPOLB i j
Nstatewide i j
Estatewide i j
EPOLB
i j ÷ ÷ ÷
ø
ö
ç ç ç
è
æ
= å å
= =
( 6)
where EPOLB, 2004 is the estimated emissions for all harbor craft vessels at POLB for
2004, Estatewide, 2004( i, j) is the estimated emission for engine type i and harbor craft
vessel type j in the statewide for 2004, N statewide, 2004 ( i, j) and NPOLB, 2004 ( i, j) are the
numbers of harbor craft type j for engine type i in the statewide and in POLB for 2004
respectively, i is the index for engine type ( propulsion and auxiliary), and j is the index
for harbor vessel type ( j = 1 to 9, defined above).
Consistent with the growth projections developed for the NNI project, it was assumed no
growth in harbor craft emissions between 2001 and 2005. Based on this assumption,
we assumed that for POLB, the total emissions of harbor craft vessels in the 2002
baseline year are equal to that in 2004 as calculated above.
C. In- Port and Out- of- Port Emissions Allocation
The emissions of different source categories are distributed at various locations in the
ports and over the offshore ocean water surfaces. To investigate spatial effects of
emission sources on the nearby neighborhoods, the total emissions of the two ports are
spatially allocated into two broad areas: in- port and out- of- port. In- port refers to the
area inside the breakwater of the ports, which is approximately 5 miles from the
shoreline; the out- of- port refers to the ocean water surface beyond the breakwater,
extending up to 50 miles from the ports. The land - based emissions resulting from
heavy- duty truck and locomotive activities outside of the Port boundaries are not
included in the “ out- port” for this modeling analysis.
24
D. Emission Inventory Summary
Emission estimates by source category for POLA and POLB in 2002 are summarized in
Figure 3 and Table 6. As can be seen, for both ports, OGVs ( transit and maneuvering)
are the biggest contributor to the combined total emissions. The next highest emission
source is the hotelling of ship’s auxiliary engines at berth, followed by commercial
harbor craft. Cargo handling equipment is the fourth largest, in- port trucks fifth, and in-port
locomotives are last. Based on the total combined emissions for the two ports,
OGV accounts for about 53 percent, hotelling accounts for 20 percent, harbor craft
accounts for 14 percent, cargo handling equipment accounts for 10 percent, in- port
truck accounts for 2 percent, and in- port locomotive accounts for 1 percent. The
emissions from POLA comprise about 55 percent of the total emissions from the two
ports. The in- port and out- of- port emissions for both ports are presented in Figure 4.
The in- port emissions comprise about 43 percent of the total emissions in the ports, and
the remaining 57 percent occurs in over water area outside the breakwater. By source
category, only OGVs and commercial harbor craft have emissions generated outside
the breakwater. OGV comprises about 90 percent of the total out- of- port emissions,
while commercial harbor craft accounts for the remaining 10 percent.
Table 6: 2002 Estimated Diesel PM Emissions for the POLA and POLB
Note: OGV – Oceangoing vessels; HOTEL – Ship’s auxiliary engine hotelling; CHC – Commercial harbor crafts; CHE – Cargo
handling equipment; IPT – In- Port heavy- duty trucks; IPL – In- Port locomotive.
Diesel PM Emissions Tons per Year
Source Category
OGV HOTEL CHC CHE IPT IPL
POLA 515 165 178 78 18 11
POLB 427 178 66 94 23 7
Combined 942 343 244 172 41 18
25
515
165 178
78
18 11
965
427
178
66
94
23 7
795
942
343
244
172
41 18
1760
0
200
400
600
800
1000
1200
1400
1600
1800
2000
OGVs HOTEL CHC CHE In- Port
Truck
In- Port
Loco
Total
Diesel PM Emissions ( TPY)
POLA
POLB
Combined
Figure 3: Estimated 2002 Diesel PM Emissions for POLA and POLB
Notes: OGV = Ocean- going Vessels; Hotel = Ship Auxiliary Engine Hotelling; CHC = Commercial Harbor Craft; CHE
= Cargo Handling Equipment; In- Port Loco = In- Port Locomotives
38
343
137
172
41
18
749
904
0
107
0 0 0
1011
942
343
244
172
41
18
1760
0
200
400
600
800
1000
1200
1400
1600
1800
2000
OGVs HOTEL CHC CHE In- Port
Truck
In- Port
Loco
Total
Diesel PM Emissions ( TPY)
IN- Port
Out- port
Combined
Figure 4: In- Port and Out- of- Port Distribution of POLA and POLB Diesel
PM Emissions
Notes: OGV = Ocean- going Vessels; Hotel = Ship Auxiliary Engine Hotelling; CHC = Commercial Harbor Craft; CHE
= Cargo Handling Equipment; In- Port Loco = In- Port Locomotives
26
III. AIR DISPERSION MODELING
In this chapter, we describe the air dispersion modeling performed to estimate the
downwind dispersion of diesel PM exhaust emissions resulting from the activities at
POLA and POLB. A description of the air quality modeling parameters, including air
dispersion model selection, modeling domain, emission source distribution/ allocation,
model parameters, meteorological data selection, and model receptor network, is
provided.
A. Air Dispersion Model Selection
Air quality models are often used to simulate atmospheric processes for applications
where the spatial scale is in the tens of meters to the tens of kilometers. Selection of air
dispersion models depends on many factors, such as, characteristics of emission
sources ( point, area, volume, or line), the type of terrain ( flat or complex) at the
emission source locations, and source receptor relationships. For this study, ARB staff
selected the U. S. EPA Industrial Source Complex Model Short Term Version 3
( ISCST3, Version 02035) to simulate impacts at nearby receptors due to diesel PM
emissions. The ISCST3 model is a micro- scale, steady- state Gaussian plume
dispersion model applicable for estimating impacts from a wide variety of emission
release patterns ( point, area, line, and volume) such as those found at the ports for
distances up to about 50 kilometers. The model may be used to predict annual average
concentrations. ISCST3 is also able to simulate the dispersion of emissions generated
from multiple sources and accommodate both continuous and intermittent sources in flat
and complex terrain. ARB staff has successfully used ISCST3 model to assess public
heath risk impacts of diesel PM emitted from the Roseville Railyard on nearby
residential areas.
B. Model Domain and Receptor Network
The modeling receptor domain ( study area) spans a 20 x 20 mile area as shown in
Figure 5a. The domain includes both the ports, the ocean surrounding the ports, and
nearby residential areas which have a population of about 2 million residents. Diesel
PM emissions are released within the modeling receptor domain as well as beyond the
receptor network for ocean- going vessels ( see Figure 5b). The land- based portion of
the modeling receptor domain, excluding the property of the ports, comprises about
65 percent of the modeling domain. A Cartesian grid receptor network ( 160 x 160 grids)
with 200 m x 200 m resolution is used in this study. This network is convenient to
identify the emission sources within the ports with respect to the receptors in the nearby
residential areas. Since the exposure assessment was not designed to identify hot
spots, a finer grid receptor network was not used. While receptors within the ports were
included in the network, the risks from these on- site receptors were excluded from the
final risk analyses. The elevation of each receptor within the modeling domain was
determined from the United States Geological Service topographic data.
27
Figure 5a. Modeling Receptor Domain for the Ports of Los Angeles and
Long Beach
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
Long Beach
Los Angeles Harbor
Harbor
28
Figure 5b. Depiction of the Emission Source Locations ( On the electronic
version of the document, the following color codes are used to designate emission sources:
Magenta = OGV+ CHC, Dark Brown = CHE, Yellow = IPT, Blue = IPL, Red = Hotelling)
378000 380500 383000 385500 388000 390500 393000 395500 398000
Easting ( m)
3732000
3734000
3736000
3738000
3740000
3742000
3744000
3746000
Northing ( m)
280000 300000 320000 340000 360000 380000 400000 420000 440000
Easting ( m)
3640000
3660000
3680000
3700000
3720000
3740000
3760000
Northing ( m)
OGV+ CHC Shipping Lanes
Ports
29
C. Model Parameters
The emission sources in the ports are characterized as area sources except for ship
hotelling, which is modeled as individual point sources. Model parameters for area
sources include emission rate/ strength, release height, lengths of X and Y sides of
rectangular areas or vertices for polygons, and initial vertical ( s zo) dimensions of the
area source plume. Model parameters for point sources include emission rate, stack
height, stack diameter, stack exhaust temperature, and stack exhaust exit velocity.
The OGV emissions are simulated as area sources. Starcrest provided the coordinates
to establish links. The link widths in the ports and in the shipping lanes over the ocean
water surface are assumed to be 160 m and 800 m, respectively. Commercial harbor
craft emissions are simulated similar to the OGVs. The links are identical to those of
OGVs. Cargo handling equipment emissions are simulated as area sources with the
polygon features of the dispersion model. Locomotive emissions are also simulated as
area sources. The links were established based on the nodes provided by Starcrest
and/ or estimated by ARB staff. Each link width is assumed to be 20 m. The terminal
and off- terminal heavy- duty trucks are simulated similar to the railroad locomotives,
except that the link width is assumed to be 35 m ( three lanes in each direction + 3
meters wake width on each side). As mentioned previously, the hotelling emissions
from ship auxiliary engines are simulated as individual point sources at the berths.
Because stack information was not available for individual engines, the average stack
height data ( 43 meters) provided in the Starcrest inventory report was applied to all
hotelling engines. The modeling parameters for each of the emission source categories
are summarized in Table 7.
Table 7: Emission Source Model Parameters
Model Parameter OGVs CHC CHE RAIL TRUCK HOTEL
Release Height ( m) 50 6 2.4 – 3.9 5 4
Link Width ( m) - - - 20 35
Link Width in Ports ( m) 160 160 - - -
Link Width in Shipping
Lane ( m)
800 800 - - -
s zo ( m) 23.26 2.79 1.1 – 1.8 2.33 1.86
H = 43 m
T = 618 K
V = 16 m/ s
D = 0.5 m
Note: OGV = Ocean- going vessels, CHC = commercial harbor craft, CHE = cargo handling equipment, H = release
height, T = exhaust temperature, V = exhaust exit velocity, and D = stack diameter.
D. Spatial and Temporal Allocation of Emissions
Starcrest provided spatial emission allocation for all source categories at POLA and for
three source categories - cargo handling equipment, In- port locomotives, and In- port
trucks at POLB. ARB staff used GIS mapping to allocate the emissions for POLB
OGVs, hotelling, and commercial harbor craft based on the descriptions provided by
Starcrest. ARB staff temporally allocated all the emission sources at both ports based
on discussions with terminal operators a nd locomotive representatives. The
30
assumptions for the temporal distribution of the emissions are listed in Table 8. The
ARB staff assumed that the temporal distribution of the emissions is the same for both
ports.
Table 8: Temporal Distribution of Diesel PM Emissions at POLA and POLB
Category Time Period Activity Distribution Hours Per Day
Ocean- Going Vessel 4 am – 8 pm
8 pm – 4 am
80%
20%
16
8
Hotelling midnight - midnight 100% 24
Harbor Craft 6 am – 6 pm
6 pm – 6 am
80%
20%
12
12
Cargo Handling 8 am – 5 pm
5 pm – 3 am
3 am – 8 am
80%
15%
5%
9
10
5
Trucks 6 am – 6 pm
6 pm – 6 am
80%
20%
12
12
Locomotives midnight - midnight 100% 24
E. Meteorological Data
Meteorological data are selected on the basis of spatial and temporal
representativeness. There are two available meteorological measurement sites around
the ports: Wilmington and North Long Beach1 ( see Figure 6). The Wilmington site is
about one mile away from the ports and the measurements were collected in 2001. The
North Long Beach site is about four miles away from the ports where data are archived
for 1981. The South Long Beach site in Figure 6 is an air quality monitoring site where
meteorological data are not archived. Normally five years of the latest consecutive
meteorological data are preferred by U. S. EPA for long term dispersion analyses.
However, one year of data are acceptable if the data are site specific according to
U. S. EPA. Therefore, ARB staff believe the Wilmington data to be the better data with
respect to spatial and temporal representativeness.
The meteorological data from the Wilmington site includes hourly wind direction, wind
speed, and atmospheric temperature. Atmospheric stability, rural mixing height, and
urban mixing height are developed following the U. S. EPA guidance. Figure 7 presents
the wind rose and Figure 8 provides the wind and stability class frequency distributions
for the meteorological conditions at the Wilmington site. Based on the yearly statistics,
the annual average wind speed at Wilmington is 1.8 m/ s with the predominant wind
directions from the northwest ( about 22 percent of the time) and from the south ( about
1 The King Harbor meteorological monitoring station is located about 10 miles northwest of the ports on
the ocean- side. To determine if diesel PM emissions transported on the ocean- side would be better
simulated using King Harbor meteorological data we conducted a sensitivi ty study ( detailed in Appendix
C) and found that there is not a significant difference between using Wilmington and using King Harbor
meteorological data sets based on the population- weighted risks in the modeling domain.
31
14 percent of the time). For the ISCST3 air quality model, urban dispersion coefficients
are used because the area at the impacted receptors is comprised of industrial,
commercial and compact residential land uses.
Figure 6: Locations of Surface Meteorological Measurement Sites around the
Ports
32
Figure 7: Wind Rose for the Period 1/ 1/ 01 to 12/ 31/ 01 at the
Wilimington Meteorological Site
33
Figure 8: Wind Speed and Stability Class Frequency Distribution at Wilmington
Meteorological Site.
34
IV. EXPOSURE ASSESSMENT
In this chapter, we briefly describe the OEHHA guidelines on health hazard risk
assessment and how we used the guidelines to characterize potential cancer risks
associated with exposure to diesel exhaust from the ports. We also present preliminary
air dispersion modeling results for the ports.
A. OEHHA Guidelines
The Air Toxics Hot Spots Program Risk Assessment Guidelines: The Air Toxics Hot
Spots Program Guidance Manual for Preparation of Health Risk Assessments ( OEHHA
guidelines, 2002a) outlines a tiered approach to risk assessment, providing risk
assessors with flexibility and allowing for consideration of site - specific differences.
Tier- 1 is a standard point- estimate approach that uses a combination of the average
and high- end point- estimates. This approach will be used in this risk assessment.
The OEHHA guidelines recommend that all health hazard risk assessments present a
Tier- 1 evaluation for the Hot Spots Program, even if other approaches are also
presented. For Tier- 1, OEHHA provides two values for breathing rate, one representing
an average and another representing a defined high- end value. The average and high-end
of point- estimates are defined in terms of the probability distribution of values for
that variate. The mean ( 65th percentile) represents the average values for point-estimates
and the 95th percentile represents the high- end point- estimates from the
distributions identified in the OEHHA guidelines. In 2004, ARB recommended the
interim use of the 80th percentile value ( the midpoint value of the 65th and 95th percentile
breathing rate) as the minimum value for risk management decisions at residential
receptors for the breathing pathway. The 80th percentile corresponds to a breathing
rate of 302 Liters/ Kilogram- day ( 302 L/ Kg- day). This risk assessment will use the
302 L/ Kg- day value and will assume that the receptors will be exposed for 24 hours per
day for 70 years. If a receptor is exposed for a shorter amount of time to the annual
average concentration of diesel PM the cancer risk will be proportionately less.
The relationship between a given level of exposure to diesel PM and the cancer risk is
estimated by using the diesel PM cancer potency factor. A description of how the diesel
cancer potency factor was derived can be found in the Proposed Identification of Diesel
Exhaust as a Toxic Air Contaminant ( ARB, 1998) and a shorter description can be
found in the Air Toxics Hot Spot Program Risk Assessment Guidelines, Part II,
Technical Support Document for Describing Available Cancer Potency Factors ( OEHHA
2002b). The use of the diesel unit risk factor for assessing cancer risk is described in
the OEHHA Guidelines. The potential cancer risk is estimated by multiplying the
inhalation dose by the cancer pote ncy factor ( CPF) of diesel PM ( 1.1 ( mg/ kg- d)- 1).
35
B. Exposure Assessment
A number of variables can have significant impacts on exposure. These include
emission estimates, meteorological conditions, and exposure duration of residents. The
emissions affect the risk levels linearly; as emissions increase, so does the risk.
Meteorological conditions can have a large impact on the resultant ambient
concentration of a toxic air pollutant with higher concentrations found along the
predominant wind direction. Key variables in human exposure are a person’s proximity
to the emission plume, how long he or she breathes the emissions ( exposure duration),
the person’s breathing rate, and body weight. The longer the duration of exposure, the
greater the potential risk.
C. Risk Characterization
Risk characterization is defined as the process of obtaining a quantitative estimate of
risk, including a discussion of its uncertainty. The risk characterization process
integrates the results of air dispersion modeling and relevant toxicity data ( e. g., diesel
PM cancer potency factor) to estimate potential cancer or noncancer health effects
associated with contaminant exposure. It is important to note that no background or
ambient diesel PM concentrations are incorporated into the risk quantification. The risk
assessment only considers the cancer risk by the inhalation pathway because the risk
contributions by other pathways of exposure are known to be negligible relative to the
inhalation pathway and difficult to quantify.
As stated in Chapter III, the modeling receptor domain of 20 mi x 20 mi with a grid
resolution of 200 m x 200 m was used in the modeling exercise. The effective land area
( excluding the Port property and the over water region) is about 255 square miles. The
population within the modeling receptor domain is about 2 million based on the
U. S. Census Bureau’s year 2000 census data. The risk numbers, impacted areas, and
affected population presented below are based on the effective land area within the
modeling domain; that is, the risk, the area, and the population within the ports property
and over the ocean surface are excluded from this analysis. Note that if the modeling
domain expands, the risks, impacted areas, and affected population presented in this
analysis would be changed.
Risk Characterization for All Emission Sources
Figure 9 shows the risk isopleths for all diesel PM emission sources from POLA and
POLB superimposed on a map that covers the ports and the nearby communities. The
risk contour of 100 in a million exceeds the modeling receptor domain in the north
direction of the ports, which is about 10 miles away from the ports boundary. The area
with predicted cancer risk levels in excess of 100 in a million within the modeling
receptor domain is estimated to be about 94,000 acres, which is 57 percent of the
effective land area within the modeling receptor domain ( see Table 9). The area in
which the risks are predicted to exceed 200 in a million is also very large, covering an
area of about 29,000 acres ( 18 percent of the effective land area within the modeling
36
receptor domain). The areas with the greatest impact have an estimated potential
cancer risk of over 500 in a million, which cover about 2 percent of the effective land
area within the domain. The risk isopleths of 1000 and 1500 in a million occur on port
property and the nearby ocean surfaces, which is not included in this study because
people do not reside in these areas.
Using the U. S. Census Bureau’s year 2000 census data, we estimated the population
within the isopleth boundaries. As shown in Table 10, the affected population numbers
for the risk ranges of 100- 200, 200- 500, and over 500 have been estimated to be about
724,000, 360,000, and 53,000, which account for 37, 18 and 3 percent of the total
population within the modeling domain, respectively. In other words, nearly 60 percent
of 2 million people live in the area around the ports that has predicted risks of greater
than 100 in a million. Note that the risk isopleth of 10 in a million is not shown in
Figure 9 because it exceeds the modeling receptor domain. Spatially, the emission
sources are located at various locations on port property and the outside of the
breakwater, thus the contributions of these emission sources to the nearby
neighborhoods would be different. Below, we discuss the contributions from the various
sources at the ports to the community risks.
37
Figure 9. Estimated Diesel PM Cancer Risk from All Diesel- Fueled Engines
at POLA and POLB ( Wilmington Meteorological Data, Urban
Dispersion Coefficients, 80th Percentile Breathing Rate, Total Emissions = 1,760
TPY, Modeling Receptor Domain = 20 mi x 20 mi, Resolution = 200 m x 200 m)
Risk Characterization for Individual Emission Sources
The different emission sources are used at various locations on the ports property in the
harbor and over ocean beyond the breakwater. Thus, the contributions of these
emission sources to exposures in the nearby neighborhoods are different. As shown in
Tables 9 and 10, the emissions from cargo handling equipment and on- port trucks
resulted in areas within the nearby communities having risk levels exceeding 500 in a
million while the highest risk levels associated with the other categories were between
200 and 500 in a million. Within the model domain, ship hotelling emissions and cargo
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
38
handling equipment impacted the largest areas and affected more people than the other
sources of emissions when considering the risk levels greater than 100 in a million.
When considering risk levels greater than 10 in a million, all the port sources, other than
in- port trucks and locomotives had similar impacts, affecting at least 119,000 acres and
at least 1.4 million people. By source location, the impacts resulting from the in- port
emissions ( within breakwater) are much larger than those resulting from the out- port
emissions ( outside of breakwater), although the emission magnitude of the former is
less than the latter ( 750 TPY vs 1010 TPY). Quantitatively, within the modeling receptor
domain, the population- weighted risk resulting from the in- port emissions is about
4.5 times of that resulting from the over water out- of- port emissions.
Table 9: Summary of Area Impacted by Risk Levels and Activity Categories
( Acres)
Risk Level OGV HOTEL CHC CHE IPT IPL COMBINED
Risk > 500 0 0 0 50 50 0 2,500
Risk > 200 110 2,036 20 410 160 40 29,000
Risk > 100 227 12,700 750 4,100 376 160 94,000
Risk > 10 163,435 160,470 125,250 119,000 29,750 11,240 163,435
Table 10: Summary of Population Affected by Risk Levels and Activity
Categories ( Number of People)
Risk Level OGV HOTEL CHC CHE IPT IPL COMBINED
Risk > 500 0 0 0 3,200 205 0 53,000
Risk > 200 18 46,020 5,000 11,100 1,780 680 411,200
Risk > 100 1,810 221,567 22,960 82,000 8,270 4,330 1,135,000
Risk > 10 1,977,760 1,949,850 1,516,515 1,444,000 422,910 213,430 1,977,770
Notes:
1. OGV – Ocean- going vessels; HOTEL – Ship’s auxiliary engine hotellng; CHC – Commercial harbor
crafts; CHE – Cargo handling equipment; IPT – In- Port trucks; IPL – In- Port locomotive.
2. The model receptor domain of 20 mile x 20 mile for urban dispersion coefficients with a grid resolution
of 200m x 200m was used. The effective modeling receptor domain ( excluding the port properties and
the ocean water) is estimated to be about 255 square miles. The calculations here are ONLY based
on the effective modeling receptor domain.
3. The 80th percentile breathing rate for adults over 70- year lifetime was assumed.
4. Meteorological data from Wilmington ( 2001) are used for POLA and POLB.
5. The risks within both ports and over the ocean water were excluded for calculations of average risks
and affected areas.
6. The estimated population in this Table is ONLY based on the modeling receptor domain using the
U. S. Census Bureau’s year 2000 census data.
7. If the modeling receptor domain expands, the numbers of population and area affected would be
increased.
8. The combined column provides the population affected and area impacted for the cumulative impacts
from all the emission sources. The individual impacts are not additives since the combined impacts
are greater than the sum of the individual sources. For example, cargo handling equipment and
commercial harbor craft emissions may impact the same location and population. While individually
the impacts may result in cancer risk levels between 100 and 200 in a million, when you combine the
impacts, the resulting risks could be greater than 200 in a million.
39
Below, we provide additional discussion on each of the contributions of each of the
emission source categories and present the predicted risk isopleths for individual
sources.
Ocean- Going Vessels
Figure 10 presents the predicted risk isopleths for the diesel PM emissions from the
OGVs ( transiting and maneuvering emissions only). The area impacted by these
emissions is very large ( has a large footprint) and many of the risk isopleths extend
beyond the boundaries o f the modeling receptor domain. The area within the modeling
domain in which the cancer risks are predicted to be greater than 100 in a million is
small, covering an area of about 227 acres with a population size of 1,800. The
potential cancer risk levels between 50 to 100 in a million are located in nearby areas
north of the ports. All areas within the modeling receptor domain are predicted to have
an estimated potential cancer risk of over 10 in a million. From the point of view of the
emission magnitude, OGVs contributed about half of the total emissions ( 940 of 1,760
TPY). This disproportional phenomenon can be attributed to the fact that the diesel PM
emissions from OGVs are distributed over a very wide area and most of these
emissions ( about 96 percent) are emitted from the offshore shipping lanes which begin
approximately 5 miles beyond the port breakwater and extend to about 50 miles away
from the ports. In other words, only a small portion of the transiting and maneuvering
emissions ( about 4 percent) are emitted in the ports. In addition, the vessels have an
average physical stack height of 43 meters above the water surface ( final plume rise
modeled as 50 m), resulting in diluted plumes over a wide area.
40
Figure 10. Estimated Diesel PM Cancer Risk from Ocean- Going Vessel’s
Activity at POLA and POLB ( Wilmington Meteorological Data,
Urban Dispersion Coefficients, 80th Percentile Breathing Rate, Emission =
942 TPY, Modeling Domain = 20 mi x 20 mi, Resolution = 200 m x 200 m)
Hotelling
The emissions from ship auxiliary engines’ hotelling resulted in a significant risk impact
to the nearby communities. As shown in Figure 11, the potential cancer risk level
ranges from 50 to 200 in a million. The area in which the risks are predicted to exceed
100 in a million has been estimated to be about 12,700 acres with a population of
221,600. Hotelling emissions from auxiliary engines result in cancer risk levels over
10 in a million in about 98 percent of the effective modeling domain. Compared to the
OGVs, the emission from the auxiliary engines hotelling is approximately 36 percent of
the OGVs ( 343 TPY vs 942 TPY), but the predicted population- weighted average risk
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
41
from the hotelling is about 1.5 times of that from the OGVs. This is not surprising
because the emissions from hotelling activities are located within the ports, which are
close to nearby communities.
Figure 11. Estimated Diesel PM Cancer Risk from Ship Auxiliary Engines’
Hotelling at POLA and POLB ( Wilmington Meteorological Data,
Urban Dispersion Coefficients, 80th Percentile Breathing Rate, Emission = 343
TPY, Modeling Domain = 20 mi x 20 mi, Resolution = 200 m x 200 m)
Commercial Harbor Craft
The emissions from commercial harbor craft resulted in a moderate risk level in the
nearby communities around the ports ( Figure 12). The area in which the risks are
predicted to exceed 100 in a million has been estimated to be about 750 acres with a
population of 23,000. Overall, about 77 percent of the effective modeling receptor
domain have estimated cancer risk levels of over 10 in a million due to emissions from
commercial harbor craft.
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
42
Figure 12. Estimated Diesel PM Cancer Risk from Commercial Harbor Craft
Vessel Activity at POLA and POLB ( Wilmington Meteorological Data,
Urban Dispersion Coefficients, 80th Percentile Breathing Rate, Emission = 244
TPY, Modeling Domain = 20 mi x 20 mi, Resolution = 200 m x 200 m)
Cargo Handling Equipment
The ground- based activities of cargo handling equipment generated an estimated
emission of about 172 TPY, which accounts for about 10 percent of the total emissions
inventory for the ports. The emissions resulted in significant risk impacts on the nearby
residential areas. As shown in Figure 13, the area in which the risks are predicted to
exceed 100 in a million has been estimated to be about 4,100 acres with a population of
82,000. For the highest risk level of over 500 in a million, the impacted areas have
been estimated to be about 50 acres and about 3,200 people living around the ports are
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
43
exposed to the risk level. Overall, about 73 percent of the effective modeling receptor
domain has an estimated risk level of over 10 in a million and about 73 percent of
2 million people who are living in the domain are exposed to the risk level. From
Figure 13, we can see that the finger- like isopleth jutting to the north exists. This is
caused by sources located within the narrow finger- like port property that contribute
about 17 TPY of emissions to the downwind direction area ( north). Based on the
population- weighted spatial average risk, the emission sources from cargo handling
equipment are the second biggest contributor to the nearby communities.
Figure 13. Estimated Diesel PM Cancer Risk from Cargo Handling
Equipment Activity at POLA and POLB ( Wilmington Met
Data, Urban Dispersion Coefficients, 80th Percentile Breathing Rate,
Emission = 172 TPY, Modeling Domain = 20 mi x 20 mi, Resolution =
200 m x 200 m)
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
44
In- Port Trucks and Locomotives
Compared with other emission sources, the emissions from in- port heavy- duty trucks
and locomotives are relatively small, accounting for about 3 percent of the emissions
inventory. These ground- based emissions resulted in localized health risk impacts. As
shown in Figures 14 and 15, the higher risk level of 100 to 200 in a million occurs on
port property. The exposure risk level to the nearby residents is relatively small. For in-port
heavy- duty trucks, about 18 percent of the effective modeling domain has an
estimated risk level of over 10 in a million, affecting about 21 percent of the residents
within the model domain. Similarly, for in- port locomotives, about 7 percent of the
effective modeling receptor domain has an estimated risk level of over 10 in a million,
affecting about 11 percent of the residents. It is important to note that there are
emissions of heavy- duty trucks and locomotives that are released beyond the
boundaries of the ports and impact residents living along freeways, rail yards and rail
corridors, and distribution centers. The impacts from these emissions ( e. g., freeway
diesel PM) are not included in this analysis.
In this study, we did not consider the diesel PM emissions of on- road heavy- duty trucks
and locomotives related to port activities that occur off- port boundary within the SCAB
( regional emissions). We estimated the off- port regional diesel PM emissions to be
about 206 TPY for the both ports, or 10 percent of the total port- related emissions ( 206
TPY vs 1,970 TPY). These regional emissions are distributed throughout the SCAB
and may result in localized health impacts to people who are live near freeways and
railroad corridors within the SCAB. These health impacts will be evaluated in future
studies.
45
Figure 14. Estimated Diesel PM Cancer Risk from In- Port Heavy Duty
Trucks at POLA and POLB ( Wilmington, Meteorological Data,
Urban Dispersion Coefficients, 80th Percentile Breathing Rate,
Emission = 41 TPY, Modeling Domain = 20 mi x 20 mi, Resolution
= 200 m x 200 m)
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
46
Figure 15. Estimated Diesel PM Cancer Risk from In- Port Locomotive
Activity at POLA and POLB ( Wilmington, Meteorological
Data, Urban Dispersion Coefficients, 80th Percentile Breathing Rate,
Emission = 18 TPY, Modeling Domain = 20 mi x 20 mi, Resolution = 200
m x 200 m)
In- Port vs Out- of- Port Emissions
As mentioned previously, a comparison between the impacts from in- port, i. e., those
emissions that occur on port land- based property and within the breakwater zone, and
the out- of- port, i. e., those emissions from oceangoing ships and harbor craft that occur
beyond the breakwater, was made. Although the in- port activities generate fewer
emissions than the out- of- port activities ( 750 TPY vs 1010 TPY), the in- port emissions
resulted in much higher health risk level in the nearby communities than the out- of- port
emissions ( see Figures 16 and 17). Quantitatively, based on the population- weighted
average cancer risk within the modeling domain, the potential cancer risk level resulting
from the in- port activities is about 4.5 times of that resulting from the out- of- port
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
47
activities. Possible reasons have been explained above. That is, there are greater
distances between the out- of- port emission sources and the receptors in the nearby
communities. This analysis identifies the emission sources within the ports as the most
significant to health risk to the nearby communities.
Figure 16. Estimated Diesel PM Cancer Risk from All In - Port Diesel Engine
Activity at POLA and POLB ( Wilmington, Meteorological Data,
Urban Dispersion Coefficients, 80th Percentile Breathing Rate, Emission = 750
TPY, Modeling Domain = 20 mi x 20 mi, Resolution = 200 m x 200 m)
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
48
Figure 17. Estimated Diesel PM Cancer Risk from All Out- of- Port Diesel
Activity at POLA and POLB ( Wilmington, Meteorological
Data, Urban Dispersion Coefficients, 80th Percentile Breathing Rate,
Emission = 1010 TPY, Modeling Domain = 20 mi x 20 mi, Resolution =
200 m x 200 m)
380000 385000 390000 395000 400000 405000
Easting ( m)
3725000
3730000
3735000
3740000
3745000
3750000
Northing ( m)
0 1 2
miles
49
D. Estimation of Non- cancer Health Endpoints
A substantial number of epidemiologic studies have found a strong association between
exposure to ambient particulate matter ( PM) and adverse health effects ( CARB, 2002).
As part of this study, ARB staff conducted an analysis o f the potential non- cancer health
impacts associated with exposures to the model- predicted ambient levels of directly
emitted diesel PM ( primary diesel PM) within the modeling domain. The non- cancer
health effects evaluated include premature death, asthma attacks, work loss days, and
minor restricted activity days.
Ambient levels of directly emitted diesel PM were predicted for 200 meter by 200 meter
grid cells within the modeling domain, and the populations within each grid cell were
determined from U. S. Census Bureau year 2000 census data. Using the methodology
peer- reviewed and published in the Staff Report: Public Hearing to Consider
Amendments to the Ambient Air Quality Standards for Particulate Matter and Sulfates,
( PM Staff Report) ( CARB, 2002), we calculated the number of annual cases of death
and other health effects associated with exposure to the PM concentration modeled for
each of the grid cells. The totals over the entire modeling area were then calculated.
For each grid cell, each health effect was estimated based on concentration- response
functions derived from published epidemiological studies relating changes in ambient
concentrations to changes in health endpoints, the population affected, and the baseline
incidence rates. The selection of the concentration- response functions was based on
the latest epidemiologic literature, as described in the PM Staff Report ( CARB, 2002)
and in Lloyd and Cackette ( 2001).
Based on our analysis, we estimate that the average number of cases per year that
would be expected in the modeling area is as follows:
· 29 premature deaths ( for ages 30 and older), 14 to 43 deaths as 95% confidence
interval ( CI);
· 750 asthma attacks, 180 to 1300 as 95% CI;
· 6,600 days of work loss ( for ages 18- 65), 5,600 to 7,600 as 95% CI;
· 35,000 minor restricted activity days ( for ages 18- 65), 28,000 to 41,000 as 95% CI.
Several assumptions were used in our estimation. They involve the selection and
applicability of the concentration- response functions to California data, exposure
estimation, subpopulation estimation, baseline incidence rates, and the threshold.
These are briefly described below.
· Premature death calculations were based on the concentration- response function
of Krewski et al. ( 2000). The ARB staff assumed that concentration- response
function for premature mortality in the model domain is comparable to that in the
Krewski study. It is know that the composition of PM can vary by region, and not
all constituents of PM have the same health effects. However, numerous
studies have shown that the mortality effects of PM in California are comparable
to those found in other locations in the United States, justifying our use of
50
Krewski et al’s results. Also, the U. S. EPA has been using Krewski’s study for its
regulatory impact analyses since 2000. For other health endpoints, the selection
of the concentration- response functions was based on the most recent and
relevant scientific literature. Details are in CARB’s PM Staff Report ( CARB,
2002).
· The ARB staff assumed the model- predicted exposure estimates could be
applied to the entire population within each modeling grid. That is, the entire
population within each modeling grid of 200 m x 200 m was assumed to be
exposed uniformly to modeled concentration. This assumption is typical of this
type of estimation.
· The ARB staff assumed the grid cell population had similar age distributions as
the county in which it was located. The subpopulation used for each health
endpoint was calculated by multiplying the all- age population for each grid cell by
the county- specific ratio of the subpopulation used for the endpoint over the all-age
population. For example, mortality estimates were based on subpopulations
age 30 or more estimated from ratios of people over 30 over the entire
population, specific for each county. These estimates were needed because
information on the particular subpopulation in each modeling grid was not
available.
· The ARB staff assumed the baseline incidence rates were uniform across each
modeling grid, a nd in many cases across each county. This assumption is
consistent with methods used by the U. S. EPA for its regulatory impact
assessment. The incidence rates match those used by U. S. EPA.
· Another assumption pertains to the threshold, the lowest level a t which health
impacts can be assessed. There is some evidence that the PM effect coefficient
may be larger at lower levels of PM and smaller at higher levels. However, we
assumed no threshold in our calculations. That is, the effects can be estimated
down to zero.
It should be noted that because the estimates apply to a limited modeling domain
( 20 miles by 20 miles), the affected population is small, and hence the overall estimated
health impacts are smaller than estimates made on a statewide basis. In addition, to
the extent that only a subset of health outcomes is considered here, the estimates
should be considered an under- estimate of the total public health impact.
51
E. Unquantifiable Adverse Health Effects
In this analysis, we did not qua ntify all possible health adverse effects associated diesel
PM emitted from Ports. For example, the effects of diesel PM on infant mortality,
premature births, and low birth weight are not presented. Appendix D provides a brief
overview of potential health effects of diesel PM not captured in the quantitative risk
assessment and non- cancer health evaluation.
F. Comparison with Monitoring Results
In this section, we compare the potential cancer risks from this modeling study to the
diesel PM risks based on ambient monitoring results from the Port of Los Angeles’s
( POLA) monitoring conducted during the period February 9 through August 5, 2005, and
from ARB’s 2002 Wilmington monitoring data collected as part of the Children Health
Study. We also compare this study results with the South Coast Air Quality
Management District ( SCAQMD)’ s second Multiple Air Toxic Exposure Study
( MATES- II). ( SCAQMD, 2000)
Comparison with POLA’s Monitoring Results
The POLA is currently conducting an air quality monitoring program on Port property
and in the nearby communities. The primary objective of this monitoring is to estimate
the ambient levels of diesel PM in proximity to the Port that are due to operational
activities at the Port .
There are four monitoring stations deployed within the Port and in the nearby
communities ( see Figure 18). The Wilmington community station is the primary
monitoring station located about one mile north of the Port boundary. Due to its
proximity to Port operations and the prevalence of onshore wind flows, this station has
the potential to experience elevated ambient diesel PM impacts from Port emissions.
The San Pedro station is located within the San Pedro community, on the Liberty Hill
Plaza Building. This location is near the western edge of Port emission sources and
adjacent to residential areas in San Pedro. The other two stations – Coast Boundary
Station and Source- Dominated Station, are located within the Port property.
Each monitoring station measures PM10, and PM2.5. The PM samples are analyzed for
elemental carbon ( EC), a component of air pollution that has been used as indicator of
diesel PM. The monitoring stations collect samples over specific 24- hour periods in
three- day intervals over a 12- month period. In its latest update, the POLA has released
the measured EC 24- hr average concentrations for the period from February 9 to
August 5, 2005. To estimate the concentration of diesel PM based on the monitored
concentrations of EC, ARB staff used an EC to diesel PM ratio of 0.5. The ratio of EC
to diesel PM has been reported to be 0.375 to 0.75 in literature. ( Shi, et al., 2000,
Pierson, et al., 1983, and Hildemann et al., 1991)
52
Table 11 shows the potential cancer risks based on the modeling results compared to
those calculated using the monitored EC concentrations for the Wilmington and San
Pedro monitoring sites. It can be seen that there is excellent agreement between the
predicted cancer risk levels based on modeling and the cancer risk levels based on the
monitoring data at both monitoring sites. A comparison was not made for the other two
monitoring sites because they are located within the port property. Any risks within the
port property are not reported in this study because of issues associated with the
proximity of the emission sources and on port receptors.
Comparison with ARB’s Wilmington Monitoring Results
The ARB conducted air monitoring in Wilmington from May 2001 to July 2002 as part of
the Children’s Environmental Health Program. Two monitoring sites were chosen –
Wilmington and Hawaiian. Wilmington site is located near Wilmington Park Elementary
School and the Hawaiian site is located at Hawaiian Elementary School ( also see
Figure 18 for locations). The ambient levels of EC were monitored at the two sites, but
about 70 percent of the samples collected were below the detection limit of 1.0 μg
EC/ m3. It is assumed that all measurements below the limit are arbitrarily assumed to
be 0.5 μg EC/ m3. The monitoring results are summarized and compared with our
predicted results ( see Table 11). It also can be seen that the predicted results compare
favorably with the monitored results at the two sites.
Table 11. Comparison of the predicted potential cancer risks with measurements
conducted by POLA and ARB ( cases per million)
Location Port of L. A.
monitoring results
ARB SB 25
monitoring results
Model prediction
Wilmington Community 585 N/ A 600
San Pedro 533 N/ A 500
Wilmington School N/ A 450 470
Hawaiian School N/ A 710 650
Note:
1. The ratio of elemental carbon ( EC) with diesel PM has been reported to be 0.375 to 0.75 by
literature. A ratio of 0.5 is used in this calculation;
2. For POLA’s monitoring program, the measured EC 24- hr average concentrations over the
half year from February 9 to August 5, 2005 are reported;
3. For ARB SB 25 Wilmington monitoring study, about 70% of the samples collected were
below the detection limit of 1 ug EC/ m3. It is assumed that all measurements below the limit
are arbitrarily assumed to be 0.5 ugEC/ m3;
4. For the detailed monitoring programs and results, please check POLA and ARB’s web sites.
53
Figure 18. Air Quality Monitoring Stations for the POLA and ARB Programs
( Courtesy of POLA, http:// www. portoflosangeles. org)
Comparison with the SCAQMD MATES- II Study
We also compared the modeling results to the SCAQMD’s second MATES- II study.
The MATES- II study indicated the modeled potential risk in the grid cell containing the
Wilmington air quality monitoring station is 1,187 potential cancer cases per million due
to diesel PM emissions from port activities, freeways, and other sources of diesel PM.
Wilmington
School Site
Hawaiian
School
54
This Wilmington grid cell is approximately 2 miles north of the ports. Our modeling
study shows a risk level of about 450 cases in a million in the same general vicinity. In
the nearby residential areas within one mile from port boundaries, cancer risk levels
( from diesel PM emissions as well as other toxics) ranged from 1000 to 1500 cases in a
million based on the MATES- II study. Our study shows a cancer risk range of
500 to 1000 cases in a million from diesel PM emissions. The differences can be
attributed to different modeling configurations. For example, MATES- II used the Urban
Airshed Model ( UAM) model, a grid based model with 2 km grid cells, while our study
used the ISCST3 model, a Gaussian plume model. In addition, MATES- II simulated
diesel PM from all sources ( e. g., port activities and freeway emissions) for the 1998
base year while our study was limited to diesel PM from port activities for the year 2002.
Also the MATES- II study released ocean- going emissions near ground level ( within the
first horizontal layer of the UAM). Our study released ocean going emissions at 50
meters above " ground" ( sea level) which will result in greater dispersion of emissions.
G. Uncertainty and Limitations
Risk assessment is a complex process which requires the integration of many variables
and assumptions. Due to these variables and assumptions, there are uncertainties and
limitations with the results. Generally, the assumptions are designed to be health
protective so that the estimates of risks to individuals are not underestimated. Below is
a discussion of uncertainty associated with the key elements used in a risk assessment.
These key elements are the heath risk values, the air dispersion modeling used to
predict diesel PM concentrations, and the model input parameters.
Uncertainty Associated with Health Values
Scientists often use animal studies to predict how a chemical affects humans in the
development of health values that are then used in a risk assessment. Scientists
cannot be sure that humans will respond exactly the same way as animals do to a
chemical. Also, animals used in these studies are often given very high doses of a
chemical to produce negative health effects. These doses are much higher than what
people are actually exposed to in the environment. When available, as is the case with
diesel PM, scientists use studies of people exposed at work to develop health values to
estimate potential cancer risk from environmental exposures. This can introduce
uncertainty in the potential risk estimated for the general public because there is a wide
range of responses among all individuals, and there can be a wider range of responses
in the general public than in the workers in an epidemiology study. In addition, for
diesel PM, the actual worker exposures to diesel PM were based on limited monitoring
data and were mostly derived based on estimates of emissions a nd duration of
exposure. Different epidemiological studies also suggest somewhat different levels of
risk. When the Scientific Review Panel ( SRP) identified diesel PM as a toxic air
contaminant, they endorsed a range of inhalation cancer potency factors ( 1.3 x 10 – 4 to
2.4 x 10 – 3 ( μg/ m3) – 1) and a risk factor of 3x10 - 4 ( μg/ m3)- 1, as a reasonable estimate of
55
the unit risk. From the unit risk factor an inhalation cancer potency factor of 1.1 ( mg/ kg-day)-
1 may be calculated.
Uncertainty Associated with Air Dispersion Modeling
As mentioned previously, there is no direct measurement technique for diesel PM. This
analysis used air dispersion modeling to estimate the concentrations to which the public
is exposed. While air dispersion models are based on the state - of- the- art formulations,
there are uncertainties associated with the models. The primary purpose of this study
was to prioritize emission sources/ categories from the Ports operation which are to be
regulated. The U. S. EPA Industrial Source Complex – Short Term ( ISCST) model was
selected for use in this study because of our experience using this model and it was the
U. S. EPA’s preferred air dispersion model at the time this analysis was performed.
Uncertainty Associated with the Model Inputs and Domain
The model inputs include emission rates, emission release parameters, meteorological
conditions, and dispersion coefficients. Each of the model inputs has uncertainty
associated with it. Among these inputs, emission rates and meteorological conditions
have the greatest affect on modeling results. The emission rate for each source was
calculated from the emission inventory. The emission inventory has several sources of
uncertainty including: emission factors, equipment population and age, equipment
activity, load factors, and fuel type and quality, The uncertainties in the emission
inventory can lead to over predictions or under predictions in the modeling results. To
minimize uncertainty, we relied on the most current information available.
Meteorological conditions can play a key role in predicted pollutant concentrations.
These meteorological parameters include wind speed, wind direction, atmospheric
stability, and ambient temperature. For this modeling study, we used wind data from
the Wilmington site. We assumed that this wind data was applicable over the entire
study area ( 400 square miles). This is a conservative ( health protective) assumption
and will tend to over predict the impact of emissions somewhat, particularly for
emissions released offshore.
Another critical meteorological condition that can affect pollutant concentration is the
mixing height. The greater the mixing height, the greater the volume of air is available
to dilute the pollution concentration. For this modeling study, we assumed an average
annual mixing height of about 700 meters. This value compares favorably with
U. S. Navy mixing height measurements at Point Mugu and San Nicholas Island.
( Lee Eddington, 2006)
As stated previously, a model domain of 20 miles x 20 miles was used in this study
because of the ISCST model’s limitation. In reality, the impacts of diesel PM from the
Ports in the nearby communities exceed the model domain. Based on some
preliminary modeling estimates, we believe that an additional six million people outside
the modeling study area are exposed to an annual ave rage diesel PM concentration of
56
about 0.08 μg/ m3. Additional study using a long range transport model may be
conducted to better address the full impacts outside of this model domain.
Unquantified Adverse Effects
It is not possible to quantify all possible adverse health effects associated with diesel
PM emitted from Ports. This is because peer- reviewed methodologies to quantify all of
the health effects do not currently exist. Appendix D provides a brief overview of
potential health effects from port- related emissions not captured in the quantitative risk
assessment and non- cancer health evaluation.
57
V. SUMMARY OF FINDINGS
The study evaluated the diesel PM emissions on a mass basis and with respect to what
impacts those emissions have on potential cancer risks in communities near the ports.
With respect to the mass emissions, the combined diesel PM emission from both ports
is estimated to be about 1 ,760 tons per year in 2002. This represents a significant
component of the regional diesel PM emissions for the South Coast Air B