Microsoft Word - 20091027 - PARTNER-15 DRAFT FINAL REPORT

Partnership for AiR Transportation

Noise and Emissions Reduction

An FAA/NASA/Transport Canada-

Category

metric tons

% of off-highway

% of mobile

% of total

Aircraft

73,152

3.73%

1.25%

0.80%

Recreational Marine

Diesel

13,520

0.69%

0.23%

0.15%

Commercial Marine

(C1 & C2)

398,338

20.34%

6.78%

4.33%

Land-Based Nonroad

Diesel

755,208

38.56%

12.86%

8.21%

Commercial Marine

(C3)

105,414

5.38%

1.80%

1.15%

Small Nonroad SI

83,735

4.27%

1.43%

0.91%

Recreational Marine SI

27,661

1.41%

0.47%

0.30%

SI Recreational

Vehicles

2,411

0.12%

0.04%

0.03%

Large Nonroad SI

(>25hp)

168,424

8.60%

2.87%

1.83%

Locomotive

330,894

16.89%

5.64%

3.60%

Total Off-Highway

1,958,755

100.00%

33.36%

21.29%

Highway non-diesel

2,229,330

37.97%

24.23%

Highway Diesel

1,683,882

28.68%

18.30%

Total Highway

3,913,213

66.64%

42.53%

Total Mobile Sources

5,871,967

100.00%

63.82%

Notes:

If an area had more than type of nonattainment area (e.g., PM

2.5

and CO nonattainment areas), the nonattainment

area was selected based on the area with the largest population base.

Except for aircraft, the emission levels for categories are from the inventories developed for the 2008 Final Rule on

Emission Standards for New Nonroad Spark-Ignition Engines, Equipment, and Vessels, which is available at

http://www.epa.gov/otaq/equip-ld.htm .

c 2005 (and not 2002 as for other emission sources) is the base year for aircraft emissions.

SI means spark-ignition engine, usually gasoline-powered

Categories 1, 2, and 3 (C1, C2, and C3, respectively) are EPA categories for marine engines with displacements of

less than 5 liters per cylinder, between 5 and 30 liters per cylinder, and greater than 30 liters per cylinder,

respectively. 72 FR 15937.

While aircraft contribute to the emission inventories of all the criteria pollutants, the analysis shows that the largest

contributors to inventories are NO

, VOCs (NO

and VOCs are ozone precursors; NO

is also a secondary PM

precursor), PM

2.5

and SO

(also a secondary PM precursor). SO

emissions depend on fuel sulfur levels and overall

fuel burn. NOx and PM

2.5

emissions depend on combustor and engine technology in addition to overall fuel burn. The

contribution of aircraft emissions to the national annually-averaged ambient PM

2.5

level was estimated to be 0.01

µg/m

. On a percentage basis, the contribution is approximately 0.08% for all counties and 0.06% for counties in

nonattainment areas.

The aircraft contributions to county-level ambient PM

2.5

concentrations ranged from

approximately 0% to 0.5%. Aircraft emissions were also estimated to contribute 0.12% (0.10 parts per billion) to

average 8-hour ozone values in both attainment and NAAs. Near some urban centers aircraft emissions reduced

ozone, whereas in suburban and rural areas, aircraft emissions increased ambient ozone levels. The largest county-

level decrease was 0.6%; the largest county-level increase was 0.3%.

The air quality modeling performed for this analysis was based on the Community Multi-Scale Air Quality Model

(CMAQ) with a 36-square-kilometer grid cell coverage across the contiguous lower 48 states. (Byun, D. W. and K. L.

Schere 2006) Approximately 166 million people live within the 118 NAAs identified in Table 3.1 and of these, about

29 million live within 10 kilometers of a commercial service airport within the NAAs (based on population data for the

year 2000).

The adverse health impacts of aircraft emissions were estimated to derive almost entirely from fine ambient

particulate matter. Nationally, about 160 yearly incidences of PM-related premature mortality were estimated due to

ambient particulate matter exposure attributable to the aircraft emissions estimated for this study (from 325 airports)

(with a 90 percent confidence interval of 64 to 270 incidences). One-third of these 160 premature mortalities were

estimated to occur within the greater Southern California region, while another fourteen counties (located within NY,

NJ, IL, Northern CA, MI, TN, TX and OH) accounted for approximately 21 percent of total premature mortality. In

total, 47 counties within the United States had a measurable PM-related premature mortality risk of greater than one

premature mortality incidence associated with aircraft emissions. Other PM-related health impacts, such as chronic

bronchitis, non-fatal heart attacks, respiratory and cardiovascular illnesses were also associated with aircraft

emissions. No significant health impacts were estimated due to the changes in ambient ozone concentrations

attributable to aircraft emissions. Although the health impacts estimated for aircraft LTO emissions are important, it is

very likely

they constitute less than 0.6% of the total adverse health impacts due to poor local and regional air quality

from anthropogenic emissions sources in the United States.

Evaluation of aviation emissions and their impacts on emission inventories, air quality, and public health is difficult. As

discussed further within the text, there are several important assumptions and limitations associated with the results

of this study, including some related to emission inventory development and air quality modeling. Measurement and

modeling of aircraft PM emissions is still an emerging area, and there are data limitations and uncertainties.

3, 4

The

Note that these estimates for percent contributions to total ambient concentrations carry uncertainties due to the fact

that some emissions sources are not well-quantified in U.S. National Emissions Inventories.

Greater than 90% probability based on judgment of the authors. This convention is based on that utilized by the

Intergovernmental Panel on Climate Change (IPCC 2007), where “very likely” represents a 90 to 99% probability of

an occurrence.

The determination of fine particulate matter emissions from aircraft engines is an active area of research. Methods

to estimate primary PM emissions from aircraft are relatively immature: test data are sparse, and test methods are

still under development. ICAO and EPA do not have approved test methods or certification standards for aircraft PM

emissions. ICAO’s Committee on Aviation Environmental Protection (CAEP) has developed and approved the use of

an interim First Order Approximation (FOA3) method to estimate total PM emissions (or total fine PM emissions) from

certified aircraft engines. Subsequent to the completion of FOA3, the FOA3 methodology was modified with margins

to conservatively account for the potential effects of uncertainties that include the lack of a standard test procedure,

poor definition of volatile PM formation in the aircraft plume, and the limited amount of data available on aircraft PM

emissions. This modified methodology is known as FOA3a. FOA3a is currently the agreed upon method to estimate

total PM emissions from aircraft engines, and it has been incorporated into the latest version of the FAA Emissions

and Dispersion Modeling System (EDMS), version 5.02, June 2007. FOA3a was used in this study. FOA3a predicts

fine PM inventory levels that are approximately 5 times those predicted by FOA3. The factor of 5 difference between

use of a 36 km x 36 km grid scale for the air quality analyses is expected to underestimate health impacts, especially

those that may occur close to airport boundaries. Omitting the effect of cruise level emissions on surface air quality is

also expected to lead to underestimation of health impacts by an unknown amount. Further, analysis of only one year

of aircraft emissions data may lead to an over- or under-estimation of aircraft impacts on ambient air quality due to

year-to-year changes in meteorology. Non-aircraft airport sources were also not included (e.g. emissions of ground

service equipment and other airport sources). Finally, results are reported for one concentration-response

relationship for the health effects of ambient PM; a range of concentration-response relationships has been reported

in the literature. The net effect of these assumptions and limitations is not known.

General aviation (GA) aircraft emissions were not included in our emissions inventory since GA aircraft were

responsible for less than 1% of jet fuel use by volume in 2005. However, a separate estimate of lead emissions from

GA aircraft was made (most piston-engine powered GA aircraft operate on leaded aviation gasoline (avgas); gas

turbine powered jet engines and turboprops operate on Jet A which does not contain significant levels of lead). It is

estimated that in 2002 approximately 281 million gallons of avgas were supplied for GA use in the U.S., contributing

an estimated 563 metric tons of lead to the air, and comprising 46% of the EPA year 2002 National Emissions

Inventory (NEI) for lead.

It is expected that about 50-60% of this inventory is related to LTO and local flying

operations. The health impacts of these lead emissions were not estimated.

The contribution of aircraft emissions to poor air quality is influenced by air traffic management (ATM) inefficiencies

that result in increased fuel burn and emissions. Emissions and fuel use are a function of the amount of time spent in

each phase of aircraft operations, and delays cause longer idle and taxi times and introduce ground hold times, which

in turn, increase fuel use and ground level emissions. From among the 148 U.S. airports in air quality nonattainment

areas, 113 were selected for further study and it was estimated that delays at these airports account for

approximately 320 million gallons of annual additional fuel usage due to increased taxi times. This is approximately

1% of all jet fuel used in the U.S. during 2005 and approximately 17% of fuel use during the LTO portion of the flight

for these 113 airports. Based on these results, unimpeded taxi times would result in average LTO emissions

reductions of 22% (28,000 metric tons) for CO, 7% (5,000 metric tons) for NO

, 16% (4,000 metric tons) each for

VOCs and non-methane hydrocarbons, 17% (1,000 metric tons) for SO

, 15% (260 metric tons) for PM

2.5

, and 17%

(986,000 metric tons) for fuel. These values represent about five percent of LTO emissions in these non-attainment

areas.

While there are many strategies available to reduce emissions, including aircraft and engine technology

advancements, the relationship between taxi-out time and emissions suggests that ATM initiatives can play an

important role in reducing emissions and fuel use at U.S. airports. This study suggests that initiatives such as

airspace flow programs, schedule de-peaking, continuous descent arrivals, and new runways could offer viable

means of reducing fuel burn and emissions. The analyses of these initiatives performed for this study were not

the method used for this study and that determined by the ICAO method reflects the scientific uncertainty associated

with PM emissions rates from aircraft engines.

In particular, a fuel sulfur level of 400 parts per million (ppm) was assumed for some airports and 680 ppm was

assumed for others. Our intention was to assume 680 ppm for all airports. However, year-to-year and location-to-

location variations of fuel sulfur of this level (±200 ppm) are typical and are thus within the uncertainty of the

estimation methods.

Note that the uncertainties in the primary PM estimate (footnote 3), and the uncertainties in the SO

inventory level

(footnote 4) were found to result in changes in the health impact assessment that fall within the quoted 90%

confidence interval for yearly mortality incidences, and thus do not add a substantial amount of uncertainty to the

estimate of health impacts.

U.S. EPA, Correction to May 1, 2008 Memorandum titled, ‘Revised Airport-specific Lead Emission Estimates,’

Memorandum from Marion Hoyer, Solveig Irvine, Bryan Manning to Lead NAAQS Review Docket EPA-HQ-OAR-

2006-0735, May 14, 2008.

intended to provide representative results for all airports, but to illustrate the extent to which such ATM initiatives

reduce fuel use and emissions. In order to increase efficiency without adversely affecting safety, noise and security,

these and other operational initiatives must be implemented with consideration of the larger system and numerous

complex interdependencies. Moreover, there are no universal strategies for improving operational efficiency, and a

single technology or procedure will not reduce fuel consumption and emissions at all U.S. airports.

2 Overview of Study and Report Organization

This study was conducted to identify:

 The impact of aircraft emissions on air quality in non-attainment areas;

 Ways to promote fuel conservation measures for aviation to enhance fuel efficiency and reduce emissions;

and

 Opportunities to reduce air traffic inefficiencies that increase fuel burn and emissions.

The study considered how air traffic management inefficiencies, such as aircraft idling at airports, result in

unnecessary fuel burn and air emissions. The study also makes recommendations on ways to address these

inefficiencies without adversely affecting safety and security or increasing individual aircraft noise, and that it do so

while taking account of all aircraft emissions and the impact of those emissions on human health. The scope of the

study was limited to aircraft activities in and around airports (versus operational efficiencies at altitude and in the

enroute airspace).

The Study was conducted by the Partnership for AiR Transportation Noise and Emissions Reduction (PARTNER), an

FAA/NASA/Transport Canada-sponsored Center of Excellence. Appendix A contains the full list of study participants.

The study was conducted through the coordinated efforts of five contractors and subcontractors: CSSI Inc. (CSSI),

Metron Aviation (Metron), the Massachusetts Institute of Technology (MIT), Abt Associates, Inc. (Abt), and Computer

Sciences Corporation (CSC). Figure 2.1 shows the objectives and their relationship to the tasks undertaken in the

study.

Figure 2.1: Organization of this study

This document is the final report resulting from the study. Sections 1 and 2 contain the Executive Summary and

Study Overview, respectively. The body of the report is divided into three sections:

 Section 3 addresses the impact of aircraft emissions on air quality and public health. This section describes

the methods used to estimate emissions from aircraft operating from U.S. commercial service airports, and

includes a comparison of the resulting inventory to total emissions from anthropogenic sources. Section 3

also contains results of air quality modeling to determine how these aircraft emissions impact ambient

concentrations of criteria pollutants. Finally, results of a health impact analysis are presented to estimate

how these aircraft emissions contribute to adverse health consequences.

 Section 4 focuses on opportunities to reduce fuel burn and emissions by assessing the pool of available

benefits that may be achieved by reducing ground delays.

 Section 5 identifies four air traffic management (ATM) initiatives aimed at reducing operational inefficiencies

and examines the benefit of these initiatives for reducing fuel use and emissions. These initiatives do not

represent a complete list, but are analyzed to provide illustrative estimates of the benefits that may be

achieved by pursuing these and other initiatives.

Section 6 provides the study conclusions and recommendations.

3 The Impact of Aircraft Emissions on Nonattainment Area, Local, and Regional

Air Quality and Public Health

The Clean Air Act requires the EPA to set standards for ambient levels of pollutants that have been shown to have

negative impacts on public health and welfare (40 CFR part 50). The EPA has set standards, called National Ambient

Air Quality Standards (NAAQS), for six pollutants: ozone, particulate matter (PM), carbon monoxide (CO), nitrogen

dioxide (NO

), sulfur dioxide (SO

), and lead (Pb). Standards for these pollutants, called criteria pollutants, are set by

developing human health-based and environmentally-based criteria from scientific studies. Primary standards are set

to protect public health. Secondary standards are set to protect public welfare, including items such as crop damage

and decreased visibility. These standards set the maximum concentration of the pollutant acceptable over a variety of

averaging times dependent on the scientific literature. The averaging times vary by criteria pollutant. Areas that do

not meet primary standards are called nonattainment areas (NAAs).

An assessment of the impact of aircraft emissions on air quality in NAAs was performed in this study. As is discussed

further below, in 2005, there were a total of 118 NAAs in the US (see Table 3.1 below). Figure 3.1 shows the major

commercial service airports located in ozone, PM

2.5

, CO, PM

, NO

, and SO

NAAs.

There were 150 airports

located in these areas in 2005, of which 148 were included.

This study also directly assessed the health impacts

that result from the changes in air quality that could be attributable to aircraft operations. This section describes the

three elements necessary to complete these study goals:

 A baseline aircraft emissions inventory was developed to provide an estimate of criteria pollutants and

precursor emissions attributable to aircraft operations from U.S. commercial service airports (Section 3.1);

 Air quality modeling was performed to estimate the impacts of these emissions on ambient concentrations of

PM and ozone

(Section 3.2); and

 Health impact analyses were conducted to determine the changes in public health endpoints if aircraft

emissions at these airports were eliminated (Section 3.3).

In addition, an assessment of lead emissions from piston engine (general aviation) aircraft using aviation gasoline is

provided (Section 3.4).

Airports were identified based on airports listed in the FAA’s Voluntary Airport Low Emissions Program (VALE),

which focused on airports in CO, PM, and ozone non-attainment areas for 2005 see

http://www.faa.gov/airports_airtraffic/airports/environmental/vale/media/vale_eligible_airports.xls.

148 of these airports were used in this study; Block Island State Airport (Block Island, Rhode Island) and Lake Hood

Airport (Anchorage, Alaska) were not included due to insufficient aircraft operations data.

It is typical EPA practice to focus on PM and ozone impacts in air quality analyses due to their importance for

human health. Note that ozone and PM

2.5

nonattainment areas are more prevalent than NO

, SO

, CO, and lead

nonattainment areas. Several EPA Regulatory Impact Analyses have considered changes in ambient concentrations

of PM and ozone and resulting changes in health incidences (EPA 2005, EPA 2006).

Figure 3.1: Commercial service airports located in ozone, PM

2.5

, CO, PM

, NO

, and SO

nonattainment areas in

2005.

3.1 Creation of a Baseline Inventory

Aircraft jet engines emit carbon dioxide (CO

), water vapor, nitrogen oxides (NO

), carbon monoxide, oxides of sulfur

(SO

), unburned hydrocarbons (HC), primary fine particulate matter (PM

2.5

), and other trace compounds such as

various hazardous air pollutants (e.g., formaldehyde, acetaldehyde). Typical emission indices for these pollutants are

3200 g CO

/kg-fuel-burned, 1200 g water vapor/ kg-fuel-burned, 13 g NO

/ kg-fuel-burned, 11 g CO/ kg-fuel-burned,

1 g SO

/ kg-fuel-burned, 1 g HC/ kg-fuel-burned, and 0.06 g PM

2.5

/ kg-fuel-burned. While some health impacts are

related directly to the compounds being emitted (e.g. primary particulate matter) other health impacts result from the

contributions that these emissions make to the formation of secondary pollutants, especially ozone and secondary

ambient particulate matter. Aircraft jet engines do not emit lead, except perhaps in trace amounts, since lead is not

added to jet fuel. However, most general aviation aircraft powered by piston engines use leaded gasoline as

described in Section 3.4.

Aircraft emissions can be broken into two segments: cruise and LTO cycle. Most aircraft operating hours and

emissions take place at cruise altitudes. Depending on the pollutant involved approximately 68-91% of full flight

emissions occur during cruise operations.

However, it is aircraft emissions released in the lower layer of the

atmosphere, that are typically quantified in local and regional emission inventories. The mixing height (the region of

For domestic flights for 2004, FAA’s System for Assessing Aviation’s Global Emissions (SAGE) indicates that 91%

of fuel burn and SOx, 90% of NOx, 72% of CO, and 68% of VOC emissions occurred outside the LTO. Data on

2.5

is not available. FAA, System for Assessing Aviation’s Global Emissions, Version 1.5, Global Aviation

Emissions Inventories for 2000 through 2004, FAA-EE-2005-02, September 2005, revised March 2008, available at

http://www.faa.gov/about/office_org/headquarters_offices/aep/models/sage/

the atmosphere near the earth’s surface in which turbulent mixing occurs) varies greatly by location, time of day,

season, and synoptic meteorological pattern. For this study, we considered only emissions that occur below 3,000

feet above ground level; this is normally deemed equivalent to emissions which occur during the LTO cycle. The LTO

cycle includes idle, taxi to and from terminal gates, take-off and climb-out, and approach to the airport. To provide an

estimate of the contribution of aircraft to the total emissions inventories associated with non-natural sources, and to

provide a basis for the air quality modeling, a baseline inventory of aircraft LTO cycle emissions was created as

described below.

Airport Selection

An emissions inventory for the study was generated for 325 airports with commercial activity in the United States. Of

these 325 airports, there are 263 commercial service airports and 62 airports that are either reliever or general

aviation airports with commercial activity.

The decision to include these 325 airports was made in two phases. First,

the study participants estimated aircraft emissions from those commercial service airports located in the NAAs. The

U.S. Federal Aviation Administration Voluntary Airport Low Emissions Program (VALE)

identified 150 commercial

service airports that are located in the 2005 ozone, PM

2.5

, CO, PM

, NO

, and SO

NAAs areas as shown in Figure

3.1 and Table 3.1.

During the study, it also became apparent that aircraft emissions from upwind airports (in attainment areas) could

influence air quality in NAAs because of atmospheric chemistry and regional transport processes. While it was not

feasible to model aircraft emissions from all airports in the United States within the timeframe of this research,

emissions data were generated for an additional 177 commercial service airports to account for upwind aircraft

sources that could influence air quality in NAAs and to more fully estimate the impacts of aircraft activities. A total of

177 airports in attainment areas (those with the greatest number of operations and readily available flight operations

data) were selected for inclusion in the analysis. The 325 airports modeled for the study cover all 50 states and

approximately 95 percent of U.S. jet engine aircraft operations from June 2005 to May 2006 for which flight plans

were filed (including commercial, military, and general aviation). (These airports also represent 95% of the operations

with ICAO certified jet engines in the U.S.) The study includes 63 percent of all U.S. commercial service airports (325

of 515 airports). Figure 3.2 shows the 148 NAA airports and the additional 177 airports modeled for the study. A list of

the 325 airports and their number of aircraft operations (and LTOs) is provided in Appendix B.

FAA’s National Plan of Integrated Airport Systems (NPIAS) report at

http://www.faa.gov/airports_airtraffic/airports/planning_capacity/npias/reports/ .

http://www.faa.gov/airports_airtraffic/airports/environmental/vale/

Figure 3.2: 148 Nonattainment airports and the additional 177 modeled for the study

Data and Methods

The aircraft emissions inventory used for this study was created with the FAA Emissions and Dispersion Modeling

System (EDMS), a computer program used to estimate emissions in and around airports, and to provide dispersion

calculations around airports. EDMS was developed in the mid-1980’s (and has been regularly improved since that

time) to assess the air quality impacts of proposed airport development projects. EDMS is the program required by

the FAA for performing airport inventory and dispersion analyses for aviation.

EDMS was used to generate an emissions inventory for LTO activity for flights arriving to, and departing from, the

325 study airports during the one-year period between June 2005 and May 2006. The inventory generated includes

emissions from aircraft main engines, and also auxiliary power units (APUs). APUs are small, self-contained

generators installed on aircraft that are used to start the main engines and to provide electricity and air conditioning to

aircraft parked on the ground.

EDMS requires several data inputs. Operations data were obtained from the 2005 FAA Enhanced Traffic

Management System (ETMS)

, the Bureau of Transportation Statistics (BTS) On Time Performance Data

, and the

Air Traffic Activity Data System (ATADS).

EDMS also requires jet fuel quality data, main engine and APU

specifications, aircraft weight, and ground operating times. These data were obtained from a number of sources

More details regarding EDMS may be found at

http://www.faa.gov/about/office_org/headquarters_offices/aep/models/edms_model/.

http://www.fly.faa.gov/Products/Information/ETMS/etms.html

http://www.transtats.bts.gov/OT_Delay/OT_DelayCause1.asp

http://aspm.faa.gov/main/atads.asp

including BTS

, the BACK fleet database

, and the National Airspace System Resources (NASR)

. Figure 3.3

shows the data inputs to EDMS.

Figure 3.3: Overview of EDMS inputs

EDMS computes emissions of primary particulate matter, CO, hydrocarbons,

, and SO

for all phases of taxi

and flight based on ICAO engine emissions indices. Emissions indices are estimates of the mass of pollutant

produced per mass of fuel consumed and are measured during engine certification testing and reported in the ICAO

Engine Emissions Certification Databank.

However, ICAO does not have a primary PM aircraft engine standard or

test procedure, and, thus, PM emission indices are not reported in the ICAO Databank. To estimate total emissions of

primary particulate matter (PM), a criteria pollutant composed of a complex mixture of solid particles and liquid

droplets, EDMS relied on a research-based estimation technique to derive emissions indices from available data such

as ICAO certification smoke number,

and experimental results, as described more fully below.

Historically, primary PM emissions from aircraft have been difficult to estimate due to the lack of physical

understanding of their formation and evolution in gas turbine engines and exhaust plumes, and the difficulty in

measuring fine particles in the hot, high speed flow at the point where the exhaust exits the engine. Aircraft PM

exhaust emission data are sparse, and test methods are still under development. ICAO and EPA do not have

approved test methods or certification standards for aircraft PM emissions. ICAO’s Committee on Aviation

Bureau of Transportation Statistics, Airline On-Time Performance Data, June 2005 through May 2006, available

from http://www.transtats.bts.gov/

http://www.backaviation.com/Information_Services/

Federal Aviation Administration, National Airspace System Resources (NASR) data, 2006.

Hydrocarbons are classified as non-methane hydrocarbons (NMHC) & volatile organic compounds (VOCs). VOCs

play a role in the formation of ozone.

An error was made in the specification of the fuel sulfur level for some of the airports in this inventory such that the

aircraft SO

inventory is expected to be biased towards underestimating the contribution of aircraft by approximately a

factor of 0.8. In particular, a fuel sulfur level of 400 ppm was assumed for some airports and 680 ppm was assumed

for others. Our intention was to assume 680 ppm for all airports. However, variations of fuel sulfur of this level

(±200ppm) are typical and are thus within the uncertainty of the estimation methods.

http://www.caa.co.uk/default.aspx?catid=702&pagetype=90

Smoke number is a dimensionless measure that quantifies smoke emissions from aircraft engines. ICAO requires

smoke number testing for engine certification.

Environmental Protection (CAEP) has developed and approved the use of an interim First Order Approximation

(FOA3)

method to estimate total PM emissions (or total fine PM emissions) from certified aircraft engines.

Subsequent to the completion of FOA3, the methodology was modified by adding margins to account for the potential

effects of uncertainties that include the lack of a standard test procedure, poor definition of volatile PM formation in

the aircraft plume, and the limited amount of data available on aircraft PM exhaust emission rates. This modified

methodology is known as FOA3a. FOA3a is currently the agreed upon method to estimate PM emissions from aircraft

engines, and it has been incorporated into the version of the FAA Emissions and Dispersion Modeling System

(EDMS) that was used for this study, which was, version 5.02, June 2007. FOA3a predicts fine PM inventory levels

that are approximately 5 times those predicted by FOA3 and reflects the scientific uncertainty associated with PM

emissions rates from aircraft engines. This is discussed further in Appendix C.

In addition to addressing the challenges of estimating aircraft PM emissions, another area requiring investigation was

APU usage. APU usage depends on a range of factors including aircraft size, weather, and practices specific to

individual airlines and pilots. One of the most important determinants of APU usage time is the availability of ground

support equipment (e.g. preconditioned air) that can be used in place of the APU to heat or cool the cabin and

provide ground-based power to aircraft parked at the gate. While many airlines have standard operating procedures

for APU use, the ultimate decision rests with the pilot.

An APU usage survey was conducted and the results were integrated into EDMS for more accurate characterization

of APU emissions. Because of the wide range of reported usage in the survey data, low, medium, and high values

were analyzed to account for variations in aircraft size and the availability of ground support. For the study baseline

inventory, a medium level of APU usage was used to account for a wide range of ground support access at the 325

airports, seasonal conditions, and other factors that define APU usage. The range of contribution of the medium level

of APU usage to aircraft emissions below 3,000 feet is between 0% and slightly over 25%. The average is below 5%

for CO and VOCs and under 10% for NO

and SO

. For only four non-attainment areas considered in this report, the

medium level of APU usage contributes more than 1% to census area emissions (or total emissions) as estimated in

the EPA year 2002 National Emissions Inventory. A description of the APU survey methods and results can be found

in Appendix D.

Airport Air Quality Guidance Manual. Preliminary Edition 2007 (Doc 9889).

http://www.icao.int/icaonet/dcs/9889/9889_en.pdf

Before discussing the inventory results, there is one other point which requires discussion. An error was made in the specification of the fuel sulfur level for 78 of

the airports in this inventory such that the aircraft SO

inventory is expected to be biased towards underestimating the contribution of aircraft by approximately a

factor of 0.8. In particular, a fuel sulfur level of 400 ppm was assumed for some airports and 680 ppm was assumed for others. The intention was to use 680 ppm

for all airports. However, variations of fuel sulfur of this level (±200 ppm) are typical and are thus within the uncertainty of the estimation methods.

Using the above data and methods, EDMS was used to generate an emissions inventory for each of the 325 study airports. A more detailed description of EDMS,

baseline runs, data inputs, model specifications, limitations, and sources of discrepancies in the EDMS inventory are discussed in Appendix E.

Emissions Inventory Discussion

The first step in assessing the contribution of aircraft operations to NAAQS non-attainment is to develop emission inventories for the primary pollutants (NO

, SO

HC, CO, and primary PM

2.5

) for each of the NAAs.

There were a total of 118 NAAs identified for this study; each contained at least one commercial service

airport. The NAAs in the study and the commercial service airports in each area are listed in Table 3.1 (see Appendix B for the airport name that coincides with the

airport code), together with the pollutant(s) of concern. Of the 325 airports modeled, 148 commercial service airports were located in a NAA. Emissions from the

remaining airports potentially contribute to the emission concentrations in these NAAs, due to atmospheric transport of emissions.

Table 3.1: List of nonattainment areas with at least one commercial service airport, as of September 7, 2005

State

EPA

Green Book Name

Ozone

(8-Hour)

c,d,e

2.5

(V=violation)

Notes

Airport

Code

Anchorage, AK

Serious

ANC,

MRI, LHD

Fairbanks, AK

Serious

FAI

Jefferson Co, AL

Subpart 1

BHM

Colbert Co, AL

MSL

Phoenix, AZ

Subpart 1

Maintenance

Serious

PHX

Tucson, AZ

Maintenance

TUS

Mohave Co, AZ

Maintenance

IFP

Yuma, AZ

Moderate

YUM

Secondary pollutants such as ozone and secondary particulate matter are not emitted directly from aircraft engines and require air quality modeling to simulate

their formation.

State

EPA

Green Book Name

Ozone

(8-Hour)

c,d,e

2.5

(V=violation)

Notes

Airport

Code

Los Angeles South Coast Air Basin, CA

Severe 17

Serious

LAX,

SNA,

ONT,

BUR,

LGB

San Francisco-Oakland-San Jose, CA

Marginal

Maintenance

SFO,

OAK,

SJC

San Diego, CA

Subpart 1

SAN,

CRQ

Sacramento Co, CA

Serious

Moderate

SMF

Coachella Valley, CA

Serious

PSP

San Joaquin Valley, CA

Serious

Maintenance

Serious

FAT,

BFL,

MOD,

SCK,

MCE, VIS

San Bernardino Co, CA

Moderate

VCV

Ventura Co, CA

Moderate

OXR

Chico, CA

Subpart 1

Maintenance

CIC

Indian Wells, CA

Maintenance

IYK

Imperial Valley, CA

Marginal

Moderate

IPL

Denver Metro, CO

Subpt. 1 EAC

Maintenance

DEN

Colorado Springs, CO

Maintenance

COS

Aspen, CO

Maintenance

ASE

Hartford-New Britain-Middletown, CT

Moderate

Maintenance

BDL

New Haven Co, CT

Moderate

Maintenance

Moderate

HVN

Greater Connecticut, CT

Moderate

GON

Atlanta, GA

Marginal

ATL

State

EPA

Green Book Name

Ozone

(8-Hour)

c,d,e

2.5

(V=violation)

Notes

Airport

Code

Macon, GA

Subpart 1

MCN

Boise-Northern Ada Co. ID

Maintenance

BOI

Fort Hall Reservation, ID

Moderate

PIH

Chicago-Gary-Lake Counties IL-IN

Moderate

ORD,

MDW,

BLV

Marion County, IN

Subpart 1

IND

Evansville, IN

Subpart 1

EVV

Cinc.-Hamilton, OH-KY-IN

Subpart 1

CVG

Louisville, KY-IN

Subpart 1

SDF

Boston, MA

Moderate

Maintenance

BOS

Baltimore, MD

Moderate

BWI

Washington Co (Hagerstown), MD

Subpart 1 EAC

HGR

Portland, ME

Marginal

PWM

Presque Isle, ME

Maintenance

PQI

Hancock, Knox, Lincoln & Waldo

Counties, ME

Subpart 1

RKD,

BHB

Detroit-Ann Arbor, MI

Marginal

DTW

Grand Rapids, MI

Subpart 1

GRR

Flint, MI

Subpart 1

FNT

Lansing-East Lansing, MI

Subpart 1

LAN

Kalam.-Battle Creek, MI

Subpart 1

AZO

Muskegon, MI

Marginal

MKG

Minneapolis-St Paul, MN

Maintenance

MSP

Duluth, MN

Maintenance

DLH

St Louis, MO

Moderate

Maintenance

STL

Laurel Area,Yellowstone Co.

Maintenance

BIL

East Helena Area (Lewis and Clark Co.),

B,D

HLN

State

EPA

Green Book Name

Ozone

(8-Hour)

c,d,e

2.5

(V=violation)

Notes

Airport

Code

Butte, MT

Moderate

BTM

Charlotte, NC

Moderate

Maintenance

CLT

Raleigh-Durham, NC

Subpart 1

Maintenance

RDU

Greensboro-Winston Salem-High Point,

Moderate EAC

GSO

Fayetteville, NC

Subpart 1 EAC

FAY

Boston-Lawrence-Worcester (E. MA), MA

Moderate

Maintenance

MHT

Portsmouth-Dover-Rochester,NH

Moderate

PSM

New York-N. New Jersey-Long Island,

NY-NJ-CT

Moderate

Maintenance

EWR,

JFK,

LGA,

ISP, HPN

Atlantic City, NJ

Moderate

Maintenance

ACY

Trenton, NJ

Moderate

Maintenance

TTN

Albuquerque, NM

Maintenance

ABQ

Clark Co, NV

Subpart 1

Maintenance

Serious

LAS,

VGT,

HND

Washoe Co, NV

Moderate <=

12.7ppm

Serious

RNO

Buffalo-Niagara Falls, NY

Subpart 1

BUF

Albany-Schenectady-Troy, NY

Subpart 1

ALB

Rochester, NY

Subpart 1

ROC

Syracuse, NY

Maintenance

SYR

Poughkeepsie, NY

Moderate

SWF

Jamestown, NY

Subpart 1

JHW

Cuyahoga Co, OH

Moderate

Maintenance

CLE

State

EPA

Green Book Name

Ozone

(8-Hour)

c,d,e

2.5

(V=violation)

Notes

Airport

Code

Columbus, OH

Subpart 1

CMH,

LCK

Dayton-Springfield, OH

Subpart 1

DAY

Cleve.-Akron-Lorain,OH

Moderate

CAK

Lucas Co, OH

Subpart 1

TOL

Youn.-Warren-Shar.OH-PA

Subpart 1

YNG

Portland OR-Vancouver WA area

Maintenance

PDX

Medford-Ashland, OR

Maintenance

Moderate

MFR

Klamath Falls, OR

Maintenance

LMT

Phil.-Wilmington-Atl. City, PA-NJ-MD-DE

Moderate

PHL

Hazelwood, PA

Subpart 1

Maintenance

PIT

Harris.-Lebanon-Carlisle,PA

Subpart 1

MDT

Allen.-Bethl.-Easton, PA

Subpart 1

ABE

Scranton-Wilkes-Barre, PA

Subpart 1

AVP

Erie, PA

Subpart 1

ERI

State College, PA

Subpart 1

UNV

Reading, PA

Subpart 1

RDG

Pitts.-Beaver Valley, PA

Subpart 1

LBE

Johnstown, PA

Subpart 1

JST

Altoona, PA

Subpart 1

AOO

Providence (All RI), RI

Moderate

PVD,

WST,

BID

Greenville-Spartanburg-Anderson, SC

Subpart 1 EAC

GSP

Columbia, SC

Subpart 1 EAC

CAE

Memphis, TN

Marginal

Maintenance

MEM

Nashville, TN

Subpt. 1 EAC

BNA

Knoxville, TN

Subpart 1

TYS

Chattanooga, TN-GA

Subpart 1 EAC

CHA

State

EPA

Green Book Name

Ozone

(8-Hour)

c,d,e

2.5

(V=violation)

Notes

Airport

Code

Johnson City-Kingsport-Bristol, TN

Subpart 1 EAC

TRI

Dallas-Fort Worth, TX

Moderate

DFW,

DAL

Houston-Galvest.-Braz, TX

Moderate

IAH,

HOU,

EFD,

LBX

San Antonio, TX

Subpart 1 EAC

SAT

El Paso Co, TX

Moderate

ELP

Beaumont-Port Arthur, TX

Marginal

BPT

Salt Lake Co, UT

Maintenance

Moderate

SLC

Washington, DC-MD-VA

Moderate

Maintenance

IAD, DCA

Norfolk-Virginia Beach-Newport News

(HR),VA

Marginal

ORF,

PHF

Richmond-Petersburg, VA

Marginal

RIC

Roanoke, VA

Subpart 1 EAC

ROA

Seattle-Tacoma, WA

Maintenance

SEA

Spokane Co, WA

Serious

Moderate

GEG

Yakima Co, WA

Moderate

YKM

King Co, WA

Maintenance

BFI

Milwaukee, WI

Moderate

MKE

Madison, WI

MSN

Charleston, WV

Subpart 1

CRW

Huntingt.-Ashland,WV-KY

Subpart 1

HTS

Parkersb.-Marietta,WV-OH

Subpart 1

PKB

Sheridan, WY

Moderate

SHR

Notes:

State

EPA

Green Book Name

Ozone

(8-Hour)

c,d,e

2.5

(V=violation)

Notes

Airport

Code

Commercial service airports listed in the National Plan for Integrated Airport Systems (NPIAS) per §47102(7) of Title

49 USC.

An empty cell in criteria pollutant columns indicates that the airport is in attainment for that

pollutant.

Green Book Name is the name of the nonattainment area.

The 8-hr. ozone national ambient air quality standard took effect on June 15, 2005, replacing the previous 1-hr.

standard.

"Subpart 1" denotes 8-hour ozone nonattainment areas that are covered under Subpart 1, Part D, Title I of the Clean Air Act.

"Subpart 1" is considered nonattainment without a classification.

Early Action Compacts (EACs) are not a classification, but areas for which the effective date of their nonattainment

designation has been deferred because they are expected to reach or maintain attainment status by December 31,

2006.

Notes description below:

A -

Lead nonattainment or maintenance

confirmed

D -

nonattainment or

maintenance unconfirmed

B -

Lead nonattainment or maintenance not

confirmed

E -

nonattainment or maintenance confirmed

C -

nonattainment or maintenance

confirmed

F -

nonattainment or maintenance unconfirmed

The two airports that were not included in the study because of insufficient operations data are Block Island State Airport (BID) and Lake Hood

Airport (LHD).

BID is in Block Island, Rhode Island and LHD is Anchorage, Alaska.

As part of this process, a quantitative comparison of the baseline aircraft inventory to total county level emissions

inventories was performed for the primary pollutants. The county level inventories used for the aircraft inventory

comparison shown in this section were derived from EPA’s year 2002 National Emissions Inventory (NEI), a database

of criteria pollutants and their precursors. The NEI provides emissions by Federal Information Processing Standards

(FIPS) area; FIPS are generally the same as counties. An estimate of all FIPS area emissions was obtained by

aggregating NEI data from all sources including point sources (e.g. smokestacks at a factory), mobile sources (e.g.

cars) and area sources (e.g. gas stations). While the NEI does include aircraft emissions, the baseline aircraft

emission inventory for each airport in this study was based on EDMS as described above, rather than the NEI.

That

is, the baseline aircraft emissions inventory for each airport for the period June 2005 through May 2006 were used

and aircraft emissions originally within the NEI were removed. The NAA and regional inventories were built from this

county level inventory information. As presented below, the aircraft emissions inventory was then compared with total

emissions inventories (which thus included EDMS aircraft emissions rather than NEI aircraft emissions) to get a

measure of relative contributions.

Focusing first on the NAAs, Table 3.2, below, shows a distribution of the percent contribution of emissions for aircraft

in the 118 NAAs. The average value in each row reflects the average of the values for aircraft contributions in each of

the 118 NAAs.

As seen in Table 3.2 the aircraft LTO emissions at the 148 commercial service airports within the

118 NAAs are small. (Note, some of the general aviation airports and reliever airports studied were located in NAAs,

but they were not included with the below inventories for NAAs. The aircraft emissions from these airports are

estimated to be a small fraction of the aircraft emissions in NAAs compared to those from commercial service

airports. This is because commercial aircraft are generally larger than general aviation aircraft and thus burn more

fuel; emissions are proportional to fuel burn.)

Table 3.2: Contribution of U.S. aircraft LTO operations at 148 commercial service airports to emission inventories in

118 NAAs

a, b, c, d

Aircraft Emissions Inventory

VOCs

2.5

2002: Average and range as a

percentage of aircraft LTO

contributions to emission

inventories for 118 NAA with at least

one commercial service airport

0.44%

0.06% to

4.36%

0.66%

0.004% to

10.93%

0.48%

0.05% to

5.03%

0.37%

0.002% to

6.91%

0.15%

0.002%

to 2.57%

Notes:

This table presents aircraft LTO emission inventories for the 148 commercial service airports in the nonattainment

EDMS aircraft emissions were used instead of NEI aircraft emissions because the level of fidelity for modeling

aircraft in the 2001 NEI is lower than that for the inventories used for this study. In particular, NEI emissions for

commercial aircraft were generated using the default EDMS times in mode (0.7 minutes for take-off, 2.2 minutes for

climb-out, 4 minutes for approach, and 26 minutes for taxi and ground idle). Also, aircraft PM emissions in the 2001

NEI were based on several engines with PM emissions data in AP 42, which is an EPA publication of air pollutant

emissions factors (http://www.epa.gov/ttn/chief/ap42/). For the aircraft inventory comparison in this study, NEI

commercial aircraft emissions were instead replaced with aircraft emissions generated for this study using a newer

version of EDMS (version 5.02) along with actual aircraft operational data and the PM emissions estimation method

FOA3a as described in Appendix E. See Appendix J for a comparison of EDMS aircraft emissions with the 2002

NEI.

If the values were calculated as total aircraft emissions over total NAA area inventories for each pollutant the

values for each pollutant for 2002/2020 would be as follows: CO: 0.36/0.78%, NO

: 0.80/2.27%, VOCs: 0.43/0.77%,

: 0.12/0.32%, PM

2.5

: 0.16/0.24%.

areas.

If an area had more than type of nonattainment area (e.g., PM

2.5

and CO nonattainment areas), the nonattainment

area was selected based on the area with the largest population base.

Except for aircraft, the emission levels for categories are from the inventories developed for the 2008 Final Rule on

Emission Standards for New Nonroad Spark-Ignition Engines, Equipment, and Vessels, which is available at

http://www.epa.gov/otaq/equip-ld.htm .

2005 is the base year for aircraft emissions.

Looking deeper into the information, Table 3.3 and Table 3.4 show the top 25 PM

2.5

and NO

aircraft emission

inventory NAAs ranked according to the percent of inventory contributed by aircraft emissions (from commercial

service airports). Table 3.3 shows that for PM

2.5

, 9 of the areas with the greatest aircraft direct PM contributions were

also PM

2.5

NAAs in 2005. Similarly for ozone, Table 3.4 shows that 16 of the areas with the greatest aircraft NO

contributions were ozone NAAs in 2005 (as described earlier, 2002 is the base year for non-aircraft emissions, and

2005 is the base year for aircraft emissions).

Table 3.3: Top 25 NAAs according to aircraft PM

2.5

contribution

NAA Name

% of total

% of

mobile

Anchorage

2.57%

8.88%

Memphis

1.14%

4.06%

Salt Lake City

0.85%

3.99%

Las Vegas

0.68%

3.20%

Aspen

0.44%

5.20%

New York-N. New Jersey-Long Island*

0.41%

1.48%

Louisville*

0.39%

2.90%

Minneapolis-St. Paul

0.39%

1.87%

Chicago-Gary-Lake County*

0.36%

1.37%

Providence (all of RI)

0.31%

1.06%

Denver-Boulder-Greeley-Ft. Collins-Love. Area

0.31%

1.65%

Phoenix-Mesa

0.30%

1.29%

San Francisco-Bay Area

0.29%

1.23%

Charlotte-Gastonia-Rock Hill

0.29%

1.56%

Los Angeles-South Coast Air Basin*

0.27%

0.92%

Southeast Desert Modified AQMA (Riverside

County, CA - Coachella Valley, CA Area)

0.27%

0.98%

Cincinnati-Hamilton*

0.26%

2.27%

Detroit-Ann Arbor*

0.26%

1.27%

Seattle-Tacoma

0.25%

0.87%

Dallas-Fort Worth

0.23%

1.52%

Atlanta*

0.23%

1.74%

Syracuse

0.22%

1.10%

Washington DC*

0.21%

1.49%

Philadelphia-Wilmington-Trenton*

0.20%

0.85%

NAA Name

% of total

% of

mobile

Albuquerque

0.19%

1.28%

* 2005 PM

2.5

NAA according to Table 3.1.

Table 3.4: Top 25 NAAs according to aircraft NO

contribution

NAA name

% of total

% of

mobile

Anchorage

10.93%

19.63%

Aspen

4.45%

5.16%

Memphis*

3.23%

4.76%

Las Vegas*

3.06%

7.13%

Salt Lake City

2.98%

3.64%

Dallas-Fort Worth*

1.76%

2.27%

Reno

1.73%

2.07%

Phoenix-Mesa*

1.72%

1.87%

San Francisco-Bay Area*

1.57%

1.85%

Lake Tahoe Nevada (Washoe County)

1.43%

1.75%

Denver-Boulder-Greeley-Ft. Collins-Love. Area*

1.42%

2.13%

New York-N. New Jersey-Long Island*

1.40%

1.98%

Charlotte-Gastonia-Rock Hill*

1.39%

2.05%

Atlanta*

1.32%

2.19%

Albuquerque

1.27%

1.62%

Chicago-Gary-Lake County*

1.27%

1.93%

Washington DC*

1.22%

1.93%

Minneapolis-St. Paul

1.07%

1.90%

Southeast Desert Modified AQMA (Riverside

County, CA - Coachella Valley, CA Area)*

1.07%

1.28%

Seattle-Tacoma

1.03%

1.15%

Indianapolis*

1.02%

1.42%

Los Angeles-South Coast Air Basin*

1.02%

1.21%

San Diego*

0.99%

1.07%

Providence (all of RI)*

0.95%

1.19%

El Paso

0.84%

1.11%

*2005 Ozone NAA according to Table 3.1.

It is also interesting to consider these inventory contributions from other perspectives. For the 118 NAAs listed in Table 3.1, Table 3.5 shows the 25 which are the

busiest based on the total number of LTOs at all commercial service airports in that NAA. Of these, 21 of 25 were either an ozone or PM

2.5

NAA, or both, in 2005

(10 areas both ozone and PM

2.5

NAAs, 21 ozone NAAs, and 10 PM

2.5

NAAs). The airports associated with these LTOs are among the busiest in the nation. Also

for the 118 NAAs listed above, Table 3.6 shows the 25 largest NAAs by population. The population (based on population data for the year 2000) in these NAAs

represents 74 percent of those in all 118 NAAs. Many of the same busy airports are also shown in Table 3.5. Of the 25 large population centers in Table 3.6, 24

were either an ozone or PM

2.5

NAA, or both in 2005 (14 areas both ozone and PM

2.5

NAAs, 24 ozone NAAs, and 14 PM

2.5

NAAs). Both of these analyses indicate

that airports are an important emissions source in these NAAs.

Table 3.5: Aircraft emissions contribution for top 25 NAAs according to LTOs (NO

, VOCs, and PM

2.5

)

VOCs

2.5

NAA Name

2005

LTOs

% total

% mobile

% total

% mobile

% total

% mobile

Los Angeles-South Coast Air Basin

a,b

937,157

1.02%

1.21%

0.50%

0.97%

0.27%

0.92%

New York-N. New Jersey-Long

Island

a,b

930,014

1.40%

1.98%

0.42%

0.93%

0.41%

1.48%

Southeast Desert Modified AQMA

(Riverside County, CA - Coachella

Valley, CA Area)

756,196

1.07%

1.28%

0.54%

1.04%

0.27%

0.98%

Chicago-Gary-Lake County

a,b

660,721

1.27%

1.93%

0.49%

1.02%

0.36%

1.37%

Houston-Galveston-Brazoria

512,986

0.68%

1.08%

0.49%

1.25%

0.16%

0.81%

Atlanta

a,b

491,426

1.32%

2.19%

0.54%

1.10%

0.23%

1.74%

Dallas-Fort Worth

486,402

1.76%

2.27%

0.58%

1.05%

0.23%

1.52%

Las Vegas

481,057

3.06%

7.13%

1.71%

2.34%

0.68%

3.20%

San Francisco-Bay Area

469,251

1.57%

1.85%

0.63%

1.20%

0.29%

1.23%

Washington DC

a,b

393,169

1.22%

1.93%

0.57%

1.14%

0.21%

1.49%

Philadelphia-Wilmington-Trenton

a,b

380,249

0.64%

0.92%

0.35%

0.74%

0.20%

0.85%

Seattle-Tacoma

318,786

1.03%

1.15%

0.28%

0.45%

0.25%

0.87%

Phoenix-Mesa

308,259

1.72%

1.87%

0.61%

1.12%

0.30%

1.29%

Denver-Boulder-Greeley-Ft. Collins-

Love

303,065

1.42%

0.67%

0.54%

0.58%

0.31%

0.60%

Boston-Worcester-Manchester

282,139

0.78%

1.17%

0.35%

0.87%

0.11%

0.93%

Charlotte-Gastonia-Rock Hill

265,175

1.39%

2.05%

0.69%

1.70%

0.29%

1.56%

VOCs

2.5

NAA Name

2005

LTOs

% total

% mobile

% total

% mobile

% total

% mobile

Detroit-Ann Arbor

a,b

255,504

0.64%

1.06%

0.42%

0.73%

0.26%

1.27%

Minneapolis-St. Paul

254,326

1.07%

1.90%

0.59%

1.05%

0.39%

1.87%

San Joaquin Valley

a,b

249,458

0.04%

0.06%

0.10%

0.29%

0.01%

0.06%

Anchorage

248,459

10.93%

19.63%

3.89%

5.78%

2.57%

8.88%

Salt Lake City

227,358

2.98%

3.64%

1.13%

2.12%

0.85%

3.99%

San Diego

222,798

0.99%

1.07%

0.37%

0.76%

0.16%

0.63%

Cincinnati-Hamilton

a,b

220,115

0.62%

1.37%

1.45%

2.99%

0.26%

2.27%

Memphis

196,202

3.23%

4.76%

2.95%

5.93%

1.14%

4.06%

Cleveland-Akron-Lorain

a,b

184,501

0.41%

0.62%

0.33%

0.62%

0.13%

0.51%

Notes:

Ozone NAA in 2005 according to Table 3.1.

2.5

NAA in 2005 according to Table 3.1.

Table 3.6: Aircraft emissions contribution for top 25 NAAs according to population (NO

, VOCs, and PM

2.5

)

VOCs

2.5

NAA Name

Year 2000

Population

% total

% mobile

% total

% mobile

% total

% mobile

New York-N. New Jersey-Long

Island

a,b

20,364,647

1.40%

1.98%

0.42%

0.93%

0.41%

1.48%

Los Angeles-South Coast Air

Basin

a,b

14,593,587

1.02%

1.21%

0.50%

0.97%

0.27%

0.92%

Chicago-Gary-Lake County

a,b

8,757,808

1.27%

1.93%

0.49%

1.02%

0.36%

1.37%

Philadelphia-Wilmington-

Trenton

a,b

7,333,475

0.64%

0.92%

0.35%

0.74%

0.20%

0.85%

San Francisco-Bay Area

6,576,113

1.57%

1.85%

0.63%

1.20%

0.29%

1.23%

Boston-Worcester-Manchester

6,230,843

0.78%

1.17%

0.35%

0.87%

0.11%

0.93%

Dallas-Fort Worth

5,030,828

1.76%

2.27%

0.58%

1.05%

0.23%

1.52%

Detroit-Ann Arbor

a,b

4,932,383

0.64%

1.06%

0.42%

0.73%

0.26%

1.27%

Houston-Galveston-Brazoria

4,669,571

0.68%

1.08%

0.49%

1.25%

0.16%

0.81%

Washington DC

a,b

4,654,618

1.22%

1.93%

0.57%

1.14%

0.21%

1.49%

VOCs

2.5

NAA Name

Year 2000

Population

% total

% mobile

% total

% mobile

% total

% mobile

Atlanta

a,b

4,231,750

1.32%

2.19%

0.54%

1.10%

0.23%

1.74%

San Joaquin Valley

a,b

3,290,618

0.04%

0.06%

0.10%

0.29%

0.01%

0.06%

Phoenix-Mesa

3,111,876

1.72%

1.87%

0.61%

1.12%

0.30%

1.29%

Cleveland-Akron-Lorain

a,b

2,945,575

0.41%

0.62%

0.33%

0.62%

0.13%

0.51%

San Diego

2,813,431

0.99%

1.07%

0.37%

0.76%

0.16%

0.63%

Minneapolis-St. Paul

2,723,925

1.07%

1.90%

0.59%

1.05%

0.39%

1.87%

Denver-Boulder-Greeley-Ft.

Collins-Loveland Area

2,715,806

1.42%

2.13%

0.54%

1.42%

0.31%

1.65%

Baltimore

2,512,431

0.82%

1.40%

0.30%

0.70%

0.14%

1.04%

Greater Connecticut (Hartford-

New Britain-Middletown Area,

New Haven County)

a,b

2,510,470

0.41%

0.52%

0.12%

0.28%

0.08%

0.41%

St. Louis

a,b

2,508,230

0.34%

0.58%

0.26%

0.57%

0.08%

0.48%

Pittsburgh-Beaver Valley

a,b

2,433,999

0.26%

0.62%

0.39%

0.80%

0.05%

0.63%

Sacramento Metro

1,978,348

0.61%

0.74%

0.28%

0.54%

0.07%

0.53%

Cincinnati-Hamilton

a,b

1,891,518

0.62%

1.37%

1.45%

2.99%

0.26%

2.27%

Milwaukee-Racine

1,839,149

0.45%

0.81%

0.33%

0.99%

0.12%

0.82%

Indianapolis

a,b

1,607,486

1.02%

1.42%

0.63%

1.17%

0.13%

1.02%

Notes:

Ozone NAA in 2005 according to Table 3.1.

2.5

NAA in 2005 according to Table 3.1.

It is also interesting to consider aircraft emissions in the context of other mobile source emission categories for the 118 NAAs. For example, Table 3.7 and Table

3.8 present NO

and PM

2.5

emissions for mobile source categories, including aircraft at the 148 commercial service airports. 2002 is the base year for non-aircraft

emissions and 2005 is the base year for aircraft emissions. Appendix J contains similar information for VOCs, CO, and SO

Table 3.7: Nonattainment area annual NO

emission levels for mobile sources(metric tons)

a,b,c,d

Source

Aircraft

73,152

Recreational Marine

Diesel

13,520

Commercial Marine (C1

& C2)

398,338

Land-Based Nonroad

Diesel

755,208

Commercial Marine

(C3)

105,414

Small Nonroad SI

83,735

Recreational Marine SI

27,661

SI Recreational

Vehicles

2,411

Large Nonroad SI

(>25hp)

168,424

Locomotive

330,894

Total Off-Highway

1,958,755

Highway non-diesel

2,229,330

Highway Diesel

1,683,882

Total Highway

3,913,213

Total Mobile Sources

5,871,967

Notes:

This table presents aircraft LTO emission inventories for the 148 commercial service airports in the nonattainment

areas.

If an area had more than type of nonattainment area (e.g., PM

2.5

and CO nonattainment areas), the nonattainment

area was selected based on the area with the largest population base.

Except for aircraft, the emission levels for categories are from the inventories developed for the 2008 Final Rule on

Emission Standards for New Nonroad Spark-Ignition Engines, Equipment, and Vessels, which is available at

http://www.epa.gov/otaq/equip-ld.htm .

2005 is the base year for aircraft emissions.

Table 3.8: Nonattainment area annual PM

2.5

emission levels for mobile sources (metric tons)

a,b,c,d

Source

2.5

Aircraft

1,948

Recreational Marine

Diesel

368

Commercial Marine

(C1 & C2)

14,342

Land-Based

Nonroad Diesel

65,572

Commercial Marine

(C3)

5,475

Source

2.5

Small Nonroad SI

14,304

Recreational Marine

6,488

SI Recreational

Vehicles

2,668

Large Nonroad SI

(>25hp)

833

Locomotive

8,301

Total Off-Highway

120,299

Highway non-diesel

28,504

Highway Diesel

42,729

Total Highway

71,233

Total Mobile

Sources

191,532

Notes:

This table presents aircraft LTO emission inventories for the 148 commercial service airports in the nonattainment

areas.

If an area had more than type of nonattainment area (e.g., PM

2.5

and CO nonattainment areas), the nonattainment

area was selected based on the area with the largest population base.

Except for aircraft, the emission levels for categories are from the inventories developed for the 2008 Final Rule on

Emission Standards for New Nonroad Spark-Ignition Engines, Equipment, and Vessels, which is available at

http://www.epa.gov/otaq/equip-ld.htm .

2005 is the base year for aircraft emissions.

As is shown in Table 3.2 above and also included in Table 3.9, aircraft operations at the 148 commercial service

airports in the 118 NAAs are a relatively small source of emissions in these areas. Finally, as presented in Table 3.9,

the study also examined contributions to national inventories for both mobile sources and total emissions for the 325

commercial service airports.

Table 3.9: Contribution of aircraft LTO operations at commercial service airports to emissions inventories

Aircraft emissions inventory

VOCs

2.5

2002: average and range as a

percentage of total emissions

inventories in 118 NAAs with at

least one commercial service airport

(118 airports)

0.44%

0.06% to

4.36%

0.66%

0.004% to

10.93%

0.48%

0.05% to

5.03%

0.37%

0.002% to

6.91%

0.15%

0.002-

2.57%

2002: average and range as a

percentage of Mobile Source

emissions inventories in 118 NAAs

with at least one commercial service

airport (118 airports)

0.54%

0.089% to

4.72%

1.04%

0.014% to

19.63%

0.98%

0.064%%

to 9.04%

2.24%

0.026% to

30.92%

0.84%

0.016%

to 8.88%

Aircraft emissions inventory

VOCs

2.5

As a percentage of EPA year 2002

National Emissions Inventory (325

airports)

0.18%

0.41%

0.23%

0.07%

0.05%

As a percentage of Mobile Source

emissions inventory in EPA year

2002 National Emissions Inventory

(325 airports)

0.22%

0.71%

0.51%

1.29%

0.53%

3.2 Impact of Aircraft Emissions on Ambient Air Quality

The results of the baseline emissions inventory comparison presented in the previous section offer a first estimate of

aviation’s influence on air quality. However, these primary pollutants are subject to atmospheric transport and

atmospheric chemistry processes that affect air quality levels downwind of primary sources. Atmospheric residence

times can extend for multiple days and it is important to consider regional scales even when assessing aircraft

emissions from distinct airport locations. Further, these processes lead to the formation of secondary pollutants such

as ozone and secondary particulate matter – the latter results from the condensation of chemical species minutes to

days after emission of the precursor emissions (predominantly NO

and SO

for aircraft). To determine the impact of

the baseline aircraft emissions inventory on air quality, a national-scale air quality simulation was performed for the

study.

Air Quality Modeling Simulation- Data and Methods

Consistent with EPA analyses such as the Clean Air Interstate Rule Regulatory Impact Analysis (EPA 2005), the air

quality modeling performed for the study included the formation, transport, and destruction of two pollutants: ozone

and fine particulate matter (PM

2.5

). These two pollutants are expected to be the dominant causes of human health

impacts associated with local air quality. To model changes in 8-hour ozone

and annual average PM

2.5

concentrations, an air quality simulation was performed using the Community Multiscale Air Quality (CMAQ) modeling

system, a three dimensional grid-based, air quality model designed to estimate the fate of ozone precursors and

primary and secondary particulate matter concentrations and their deposition over regional and urban scales (Byun

and Schere 2006). The analysis used a 36 km x 36 km grid scale that is expected to lead to an underestimation of

some local effects close to airport boundaries.

Inputs to the CMAQ modeling system include emissions estimates (from aircraft and other sources), initial/boundary

conditions, and meteorological fields. While the baseline emissions inventory from EDMS for June 2005 through May

2006 was used to estimate total emissions from aircraft (see section 3.1), emissions estimates for non-aircraft

sources were obtained from the EPA year 2001 National Emissions Inventory (NEI). The 2001 NEI, rather than the

2002 NEI, was used for the modeling of aircraft emissions impacts on air quality and human health because it was

the most carefully assessed national inventory at the time and it was readily available for air quality modeling -- it had

already been used for other rulemakings such as the proposed rule for "Control of Emissions of Air Pollution from

Locomotives and Marine Compression-Ignition Engines Less than 30 Liters per Cylinder", 72 FR 15938, April 3,

2007. For the annual PM

2.5

estimates, an entire year of meteorology was modeled. For 8-hour ozone estimates, a

five month simulation was performed to account for the summer months in which ozone concentrations peak (May

through September).

The air quality modeling methods used in this study have been used to support several regulatory actions initiated by

EPA, including the final PM

2.5

NAAQS (EPA 2006), the 8-hour Ozone NAAQS (EPA 2008), and the rule for the

"Control of Emissions of Air Pollution from Locomotives and Marine Compression-Ignition Engines Less than 30

Liters per Cylinder" (EPA 2007c). A detailed description of the air quality modeling methods used for the study can be

found in Appendix F.

Air Quality Modeling Results

For the study, three national emission scenarios were modeled with CMAQ to estimate the potential air quality

impacts of aircraft emissions: a base line scenario with all 2001 NEI emissions (including NEI aircraft emissions), an

Given in parts per billion (ppb). The averaging time for the ozone NAAQS is 8 hours.

Given in micrograms per cubic meter (µg/m

). The PM NAAQS is expressed as an annual average. There is also

a 24-hour PM

2.5

NAAQS; however, this study only considered the annual average.

EDMS aircraft emissions scenario with a full set of emissions data for non-aircraft sources obtained from the 2001

NEI plus the specific aircraft emissions generated in the EDMS baseline inventory (see Section 3.1), and another with

all aircraft emissions (both EDMS and NEI) removed. The difference in estimated pollutant concentrations between

these two simulations was used to determine the local and regional air quality impacts of the aircraft emissions. The

approach used is consistent with the EPA guidance document for modeling ozone and PM

2.5

(EPA 2007b) and is

described more fully in Appendix F.

Turning first to PM

2.5

, almost all areas experienced increases in annual average PM

2.5

concentrations due to modeled

aircraft emissions. The CMAQ simulation for PM

2.5

utilized data from 557 counties with monitoring systems for this

emission. Of the 557 counties with PM

2.5

monitoring data, 546 showed increases, 9 showed no change, and 2

showed decreases of less than 0.001 µg/m

; these decreases are expected to be within the range of model

uncertainty. On average, the modeling revealed that aircraft emissions contribute 0.01 µg/m

to overall annual

average ambient PM

2.5

levels.

 The largest impact was found in Riverside County, CA where modeled aircraft emissions increased annual

average PM

2.5

values by 0.15 µg/m

(a 0.52% increase from 28.73 to 28.88 µg/m

 San Bernardino County, CA also showed an impact greater than 0.10 µg/m

. Another 13 counties showed

an impact of at least 0.05 µg/m

and another 38 counties in the U.S. had an impact of at least 0.02 µg/m

The results of the PM modeling for NAAs and all counties appear below in Table 3.10. Figure 3.4 shows a map of the

national changes in average annual PM determined by the air quality simulation. The individual results for the 557

U.S. counties with PM

2.5

monitoring data are provided in Appendix F.

Table 3.10: Average annual PM

2.5

estimates. Results are given in µg/m

. The annual National Ambient Air Quality

Standard for PM

2.5

is 15.0 µg/m

Without

Aircraft

Emissions

(µg/m

)

With Aircraft

Emissions

(µg/m

)

Percent

Increase Due

to Aircraft

Emissions

Non-Attainment

Areas

17.75

17.76

0.06%

All Counties

12.59

12.60

0.08%

Figure 3.4: Estimated change in annual PM

2.5

concentrations (µg/m

) due to aircraft emissions.

For ozone, the analysis revealed a mix of potential benefits and disbenefits resulting from aircraft emissions. The

photochemistry associated with ozone formation is complex, depending on local quantities of NO

, VOCs, and other

ozone catalysts. Normally, increasing NO

emissions increases ozone concentrations in suburban and rural areas

where VOC sources are plentiful. Sometimes however, the addition of NO

emissions (from aircraft and other

sources) decreases ozone concentrations in urban cores, where VOC concentrations are more limited. The air quality

modeling simulation revealed areas in which the addition of aircraft emissions increased ozone as well as areas in

which decreased ozone concentrations (sometimes referred to as ozone or NO

disbenefits) were projected:

 The CMAQ simulation for ozone utilized monitoring data from 571 U.S. counties. For all of these counties,

the average change in 8-hour average ozone values was found to be an increase of 0.10 parts per billion

(ppb) due to modeled aircraft emissions. The largest increase due to aircraft emissions occurred near the

Atlanta area, a 0.6% increase from 95.9 to 96.5 ppb.

 However, there were 24 counties across the U.S. where modeled aircraft emissions caused a decrease in 8-

hour ozone values. The largest reduction was projected in Richmond County, NY, a 0.3% decrease from

96.3 to 96.0 ppb.

A summary of the results of the ozone modeling for NAAs and all counties appears below in Table 3.11. Individual

results for the 571 U.S. counties with valid ozone monitoring data are provided in Appendix F. Figure 3.5 depicts the

county-level changes in 8-hour ozone determined by the air quality simulation.

Table 3.11: Average 8-hour ozone values (ppb) with and without EDMS aircraft emissions. The National Ambient Air

Quality Standard for 8 hour ozone is 80 ppb. Based on rounding convention, values greater than or equal to 85 ppb

are considered non-attainment.

Without

Aircraft

Emissions (ppb)

With Aircraft

Emissions (ppb)

Percent

Increase Due to

Aircraft Emissions

Nonattainment Areas

91.10

91.21

0.12%

All Counties

84.85

84.95

0.12%

Figure 3.5: Estimated change in 8-hour ozone concentrations (ppb) due to aircraft emissions. Negative values

represent regions where aircraft emissions reduce levels of ozone. Positive values represent regions where the

aircraft emissions increase ozone levels.

The air quality modeling results presented above depict changes in ambient concentrations of ozone and PM that

influence attainment of National Ambient Air Quality Standards and may result in changes in public health.

The

Note that on March 27 of 2008, EPA published a rule revising the primary ozone NAAQS from 0.08 ppm to 0.075

ppm and setting the secondary ozone NAAQS to 0.075 ppm, effective May 27 of 2008. 73 FR 16436. Nonattainment

statuses of the counties assessed in this study are from 2005; the effect of the new ozone NAAQS on the

nonattainment statuses of these counties was not considered in this study.

following section describes the health impact analysis that was performed to assess the changes in public health due

to aircraft contributions to ozone and ambient fine PM concentrations.

3.3 The Impact of Aircraft Emissions on Public Health

The health impact analysis performed for this study used a methodology consistent with benefit analyses performed

by EPA for the PM NAAQS and the Ozone NAAQS (EPA 2006; EPA 2008). It should be noted that there are data

limitations and uncertainties that may affect the results by an unknown amount (in terms of both under- and over-

estimates). The use of a 36 km x 36 km grid cell size for the air quality analyses is expected to underestimate health

impacts, especially those that may occur close to airport boundaries. The omission of air quality impacts from airports

not included in this analysis is expected to lead to underestimation of aircraft-related health impacts. Omitting the

effect of cruise level emissions on surface air quality is also expected to lead to underestimation of health impacts by

an unknown amount. Further, analysis of only one year may lead to overestimation or underestimation of aircraft

impacts due to year-to-year changes in meteorology. Non-aircraft sources were also not included (e.g. emissions of

ground service equipment and other aircraft sources). Finally, we report the results for one concentration-response

relationship for the health effects of ambient PM; a range of concentration-response relationships has been reported

in the literature. The net effect of these assumptions and limitations is not known. Further research is recommended

into these areas.

EPA’s general health impact analysis framework uses the following framework (EPA 2006):

 Given baseline and post-control

emissions inventories, EPA uses photochemical air quality modeling to

estimate baseline and post-control ambient concentrations of the pollutant of concern.

 Changes in ambient concentrations of that pollutant are then combined with monitoring data to estimate

population-level potential exposure to changes in ambient concentrations.

 Changes in population exposure are then used as input to impact functions to generate changes in the

incidences of health effects, or changes in other exposure metrics are input into dose-response functions to

generate changes in welfare effects.

The results of the air quality modeling described in the previous section were used as inputs to determine changes in

human health effects across the continental United States. Consistent with EPA regulatory impact analyses such as

the Clean Air Interstate Rule (CAIR), this analysis focused on the health effects linked to two pollutants, fine ambient

particulate matter (PM

2.5

) and ambient ozone. (EPA 2005)

The air quality modeling results described in Section 3.2 were processed for use in the Environmental Benefits

Mapping and Analysis Program (BenMAP), an EPA tool that combines air pollution monitoring data, air quality

modeling data, census data, and population projections to calculate a population’s potential exposure to ambient air

pollution (Abt 2005). Appendix G contains the specific health impact functions and baseline incidence rates used in

BenMAP to perform the health impact analysis. Further information on the methodologies used for this health impact

analysis and EPA benefit analyses can be found in the PM NAAQS Regulatory Impact Analysis or the Ozone NAAQS

Regulatory Impact Analysis (EPA 2006; EPA 2008).

Note that the uncertainties in the primary PM estimate (footnote 3), and the uncertainties in the SO

inventory level

(footnote 4) were found to result in changes in the health impact assessment that fall within the quoted 90%

confidence interval for yearly mortality incidences, and thus do not add a substantial amount of uncertainty to the

estimate of health impacts.

For this study, the “baseline” inventory included EDMS aircraft emissions and 2001 NEI non-aircraft emissions, the

“post-control” inventory was that with all aircraft emissions removed.

The national results of the health impact analysis appear below in Table 3.12. The mean incidence reduction for the

continental U.S. represents the estimated change in number of yearly health incidents if all of the aircraft emissions

were to be removed.

Table 3.12: Health effects due to aircraft emissions, continental United States.

Health Effect

Yearly Baseline

Incidence

Yearly Mean Incidence Due

to Aircraft Emissions

(90% Confidence Interval)

PM-Related Endpoints:

Premature mortality

Adult, age 30 and over

Infant, age <1

2,300,000

9,000

160

(64 – 270)

(0 – 1)

Chronic bronchitis (adult, age 27 – 99)

630,000

110

(20 – 200)

Non-fatal myocardial infarction (adult, age 18 - 99)

780,000

290

(160 – 430)

Hospital admissions–respiratory (adult, age 0 – 64)

640,000

(12 – 39)

Hospital admissions–respiratory (adult, age 65 – 99)

570,000

(6 – 16)

Hospital admissions–cardiovascular

(adult, age 18 – 64)

1,400,000

(14 – 34)

Hospital admissions–cardiovascular

(adult, age 65 – 99)

2,500,000

(29 – 60)

Emergency room visits for asthma (age 0 - 17)

730,000

140

(81 – 194)

Acute bronchitis (children, age 8-12)

880,000

340

(-12 – 700)

Upper respiratory symptoms

(asthmatic children, age 9-11)

87,000,000

2,700

(860 – 4,600)

We present total baseline incidence for each health effect. Baseline incidence represents all cases of a particular

health effect in a specific population (for all causes, not just air quality), defined by the epidemiological study from

which the health effect measure is derived.

Mean incidences for the continental U.S. are rounded to the nearest whole number and to two significant figures

where applicable. These represent the estimated changes in yearly health incidences due to modeled aircraft

emissions.

Adult premature mortality based upon the Pope et al., 2002 American Cancer Society cohort study. Infant

premature mortality based upon studies by Woodruff, Grillo, and Schoendorf, 1997

Respiratory hospital admissions ages 0 – 64 for PM include admissions for chronic obstructive pulmonary disease

(COPD) and asthma.

Respiratory hospital admissions ages 65 – 99 for PM include admissions for COPD and pneumonia.

Cardiovascular admissions include cardiovascular ailments except for myocardial infarctions.

Cardiovascular admissions include cardiovascular ailments and subcategories for ischemic heart disease,

dysrhythmia and heart failure. Myocardial infarctions not included.

Health Effect

Yearly Baseline

Incidence

Yearly Mean Incidence Due

to Aircraft Emissions

(90% Confidence Interval)

Lower respiratory symptoms

(asthmatic children, age 7-14)

14,000,000

3,700

(1,800 – 5,700)

Asthma exacerbation (asthmatic children, age 6-18)

130,000,000

3,300

(370 – 9,600)

Work loss days (adults, age 18-64)

380,000,000

23,000

(20,000 – 25,000)

Minor restricted activity days (MRADs)

(adults, age 18-64)

1,400,000,000

130,000

(110,000 – 150,000)

Ozone-Related Endpoints:

Premature Mortality

(all ages)

Bell et al. (2004)

930,000

(0 – -1)

Bell et al. (2005)

1,000,000

-2

(-1 – -2)

Levy et al. (2005)

1,000,000

-2

(-2 – -2)

Meta-Analyses

Ito et al. (2005)

930,000

-2

(-1 – -2)

Hospital admissions–respiratory causes

(adults, age 65 – 99)

450,000

-3

(-5 – 0)

Hospital admissions–respiratory causes

(children, age 0 – 1)

180,000

-6

(-3 – -10)

Emergency room visits for asthma (age 0 – 99)

710,000

-4

(-12 – 0)

Minor restricted activity days (MRADs)

(adults, age 18 – 65)

570,000,000

-7,500

(-3,800 – -11,000)

School absence days (children, age 6 – 11)

3,200,000,000

-2,800

(-4,700 – -990)

The results of the BenMAP health impact analysis indicate that ambient particulate matter related to emissions of

, SO

(both gaseous precursors to secondary PM) and primary PM

2.5

causes almost all of the total aircraft-related

health impacts, including all of the mortality incidences. Approximately 160 yearly incidences of premature mortality

were estimated due to ambient particulate matter exposure attributable to aircraft emissions (with a 90% confidence

Consistent with the methodology used in the 2007 Ozone NAAQS Regulatory Impact Analysis, ozone mortality

estimates are included with the recognition that the exact magnitude of the effects estimate is subject to continuing

uncertainty. Effect estimates from Bell et al. (2004) as well as effect estimates from three meta-analyses are given.

An effect estimate of zero is also given to account for the possibility that there is no causal association between

ozone and mortality.

Respiratory hospital admissions for ozone include admissions for all respiratory causes and subcategories for

COPD and pneumonia.

Respiratory hospital admissions for acute respiratory diseases.

interval of 64-270 yearly incidences).

The adverse health effects for aircraft emissions were localized to a small

number of counties; 43% of the health impacts occurred in 10 counties, 5 of which are in southern California. The 10

counties with the highest PM-related mortality incidences due to aircraft emissions appear in Table 3.13. A list of the

twenty counties with the highest PM-related mortality incidences can be found in Appendix H.

Table 3.13: Ten counties with highest PM-related mortality incidences

Rank

County

State

Incidences

Percent of Total

Los Angeles

Orange

San Diego

San Bernardino

Cook

Riverside

Nassau

Alameda

Queens

Kings

All other counties

The results of the health impact analysis also indicated that ozone exposure related to aircraft emissions, in

comparison to PM

2.5

exposure related to aircraft emissions, produces small health impacts. This is expected due to

the small changes in ambient ozone concentrations presented in Section 3.2.

As we also described in Section 3.2, due to the complex photochemistry of ozone production, reductions in NO

emissions (from aircraft and other sources) lead to both the formation and destruction of ozone, depending on the

relative quantities of NO

, VOCs, and ozone catalysts such as the OH and HO

radicals. In areas dominated by fresh

emissions of NO

, ozone catalysts are removed via the production of nitric acid, which slows the ozone formation

rate. Because NO

is generally depleted more rapidly than VOCs, this effect is usually short-lived and the emitted

can lead to ozone formation later and further downwind. The terms “NO

disbenefits” or “ozone disbenefits” refer

to the ozone increases that can result from the removal of NO

in these localized areas. According to the North

American Research Strategy for Tropospheric Ozone (NARSTO) Ozone Assessment (NARSTO, 2000), these

disbenefits are generally limited to small regions within specific urban cores (with relatively high population density)

and are surrounded by larger regions in which NO

reductions are beneficial. The ozone-related health impacts

shown in Table 3.12 are all negative (e.g. aviation emissions lead to fewer health incidences). This is because the

ozone disbenefits due to aircraft emissions occur in regions of higher population than the regions of ozone benefits

due to aircraft emissions. In addition, as discussed earlier, NO

emissions at low altitude also react in the atmosphere

to form secondary particulate matter (PM

2.5

), particularly ammonium nitrate, and contribute to regional haze. Thus, in

areas or regions with ozone disbenefits, NO

reductions will still help reduce secondary PM levels and regional haze.

Note that the uncertainties in the primary PM estimate uncertainties and the errors in the SO

inventory level were

found to result in changes in the health impact assessment that fall within the quoted 90% confidence interval for

yearly mortality incidences, and thus do not add a substantial amount of uncertainty to the estimate of health impacts.

Yearly incidences of premature mortality from PM

2.5

based on upon the Pope et al., 2002 American Cancer Society

cohort study. Incidences rounded to the nearest whole number and to two significant figures where applicable. Total

refers to total nationwide premature mortality incidences from aviation-related PM

2.5

exposure (approximately 160).

It is important to note that aircraft-related NO

emissions modeled on their own, as was done for this analysis, may

yield a different ambient ozone concentration than if NO

emission reductions are modeled in combination with other

required, planned, or future NO

emission controls. For example, California State Implementation Plan (SIP) modeling

indicates that with a combined program of national and local controls, Southern California can reach ozone

attainment by 2024 through a mixture of substantial NO

(and VOC) reductions (SCAQMD, 2007). In areas prone to

ozone disbenefits, our ability to draw conclusions about the future air quality and health impacts of a particular source

of NO

is limited because our analytical approach does not reflect yet-to-occur emission reductions in these areas.

Within a region such as Southern California, we expect that future NO

reductions from SIP-based controls will lead

to fewer ozone disbenefits than the disbenefits modeled here. More detailed information about the air quality

modeling conducted for this analysis is contained in Appendix F.

Interpreting the PM Mortality Results for Aviation

The health impacts from aircraft LTO emissions should be viewed in the context of the total health impacts of poor

local air quality to avoid misperceptions of the relative risks associated with aircraft emissions. People frequently do

not accurately perceive risks—such misperception of risk is not unique to aviation or air quality health impacts.

However, the characteristics of aviation are such that the perceived risks (e.g. of safety-related fatalities) are often

higher than the true risks; and the characteristics of local air quality health impacts are such that the perceived risks

are often lower than the true risks. This is in part because people have a strong fear of catastrophic fatal events they

cannot control, such as the crash of an airplane, and are less afraid of risks caused by events that occur over a long

period of time, such as the chronic effects of poor air quality (cf. Slovic 2002).

Although the health impacts of aviation estimated by our study are important, it is very likely

that they constitute less

than 0.6 percent of the total adverse health impacts due to poor local air quality from all sources in the United States.

A detailed analysis of the total health effects due to poor air quality in the United States was not made for this study,

but other sources and analyses suggest that the total number of yearly premature deaths due to poor air quality in the

U.S. is very likely greater than 25,000 as described below.

EPA has finalized three mobile source air quality rules that mandate cleaner fuels (gasoline and diesel) as well as

engine standards to control pollutant emissions such as direct PM and NO

. In 2000, EPA finalized the Tier 2 rule,

regulating the sulfur content in gasoline and setting vehicle and engine standards for passenger cars and trucks (EPA

1999). In 2000 and 2004, EPA finalized the Heavy Duty Diesel Rule and the Nonroad Diesel Engine Rule,

respectively (EPA 2000, EPA 2004). Each of these mobile source rules is projected to control a significant fraction of

the PM-related emissions associated with diesel and gasoline engines and fuels. It was projected that in 2030, the

Tier 2 rule will reduce NO

emissions by 3.71 million metric tons, reduce total VOC emissions by 0.73 million metric

tons, and reduce SO

emissions by 0.25 million metric tons. It was projected that in 2030, the Heavy Duty Diesel Rule

will reduce vehicle PM

emissions by 0.09 million metric tons, NO

emissions by 2.25 million metric tons, and NMHC

emissions by 0.07 million metric tons. It was also projected that in 2030, the Nonroad Diesel Engine Rule will reduce

PM emissions by 0.12 million metric tons, reduce NO

emissions by 0.67 million metric tons, reduce VOC emissions

by 0.03 million metric tons, and reduce SO

emissions by 0.34 million metric tons. By comparison, the EDMS aircraft

emissions in this study totaled 0.01 million metric tons of SO

, 0.08 million metric tons of NO

, 0.04 million metric tons

of VOC, 0.03 million metric tons of NMHC, and less than 0.01 million metric tons of primary PM. In terms of health

impacts, EPA estimated that when fully implemented, these three mobile source programs will together prevent

Greater than 90% probability based on judgment of the authors. This convention is based on that utilized by the

Intergovernmental Panel on Climate Change (IPCC 2007), where “very likely” represents a 90 to 99% probability of

an occurrence.

approximately 25,000 PM-related premature mortalities each year. The regulatory impact analyses for these rules

used a health impacts methodology similar to that utilized in this study, and thus, may be used to put the health

impacts we estimate for the commercial aircraft LTO inventory in context. (EPA 1999; 66 Fed. Reg. 5002, January

18, 2001; 69 Fed Reg. 38958, June 29, 2004).

Other studies corroborate the overall magnitude of health impacts from air pollution in the U.S. For example, Cohen

et al. (2004) estimated that in the year 2000, urban PM accounted for approximately 28,000 premature mortalities for

U.S. cities with a population of 100,000 or more. Furthermore, the Clean Air Task Force, using emissions projections

for 2010, estimates that diesel soot is responsible for approximately 21,000 annual deaths in the U.S., and power

plant emissions are responsible for approximately 24,000 annual deaths in the U.S. (Hill, 2005).

Our purpose in comparing the health impacts of aircraft LTO emissions to the larger total health impacts of poor local

air quality from all sources in the United States and elsewhere is not to dismiss these aircraft impacts as being

unimportant. Indeed, one of the challenges of improving poor local air quality is that it results from many small

sources acting in concert. Still, we provide these overall impact estimates so that the risks imposed by aircraft LTO

emissions can be understood in the context of the overall risks associated with poor local air quality.

3.4 Lead Emissions from Piston Engine Aircraft

In 1978 EPA established a National Ambient Air Quality Standard for lead of 1.5 micrograms per cubic meter, as a

maximum quarterly average as measured in total suspended particulates. Currently, there are two areas officially

designated as non-attainment for the lead NAAQS: Herculaneum in Jefferson County, Missouri and East Helena

Area portion of Lewis and Clark Counties, Montana.

The main lead emission source associated with the East

Helena Area closed in early 2001 and monitoring ceased in late 2001 so that location is not discussed here.

While commercial and military jet engine fuel contains only trace amounts of lead, tetraethyl lead is commonly added

to aviation gasoline used in piston-engine powered, general aviation aircraft. Exhaust emissions from these piston-

engine powered aircraft that operate on leaded aviation gasoline (avgas) contribute to levels of ambient lead. The

most commonly used leaded avgas contains 2.12 grams of lead per gallon of fuel. In 2002 approximately 280 million

gallons of aviation gasoline were supplied to the U.S. (DOE Energy Information Administration 2006) contributing an

estimated 565 metric tons of lead to the air and comprising 46 percent of the EPA year 2002 National Emissions

Inventory for lead. The 2002 NEI includes an analysis of the airport-specific contribution of lead for 3,410 airports

located throughout the United States (EPA 2007a). These lead emissions are allocated to each airport based on its

percentage of piston-engine operations nationwide. These operations for 2002 can be found in the Terminal Area

Forecast system, which is the official forecast of aviation activity at the Federal Aviation Administration facilities.

Airport-specific lead emissions estimates in the NEI include lead emitted during the entire flight (i.e., not limited to the

landing and take-off cycle and local operations).

At this time, this allocation method for lead emissions was used

here to account for all lead emissions associated with avgas use. Allocating lead emissions to airports from

operations outside the landing-takeoff cycle and local flying operations has a tendency to overstate the local

emissions near airports because longer duration (e.g., itinerant) flights emit lead at altitude as well as in the local

flying area near the airport.

While there are no airports in the Herculaneum NAA (the city limits of Herculaneum), there are seven registered

http://www.epa.gov/air/oaqps/greenbk/Inca.html

Lead emissions from general aviation are calculated as the product of the fuel consumed, the concentration of Pb

in the fuel and the factor 0.95 to account for an estimated 5 percent of Pb being retained in the engine and/or exhaust

system of the aircraft. The estimate of 5 percent Pb retention was derived from measurements of lead in used oil

samples and a factor for exhaust system retention from other literature.

airports within the twenty mile local flying area around Herculaneum where general aviation aircraft operate. A

proposed revision to Missouri’s SIP characterizes general aviation aircraft lead emissions as “background.” This

characterization seems appropriate since emissions from piston-powered aircraft operating on leaded aviation

gasoline are expected to contribute to ambient concentrations of lead entering the Herculaneum NAA both from

landing and take-off at local airports as well as piston-engine powered aircraft flying through the NAA. However, they

are not necessarily the cause of the non-attainment problem.

EPA conducted a review of the lead NAAQS which has included the assessment of health and welfare effects of lead

documented in the 2006 Air Quality Criteria Document for Lead (available at www.epa.gov/ncea). Integral to the

NAAQS review were decisions regarding the adequacy of the current standard for lead and whether the Agency

should retain or revise it. The final revisions to the lead NAAQS were published in the Federal Register on November

12 of 2008. 73 FR 66964. Additional information about the review is available at:

http://www.epa.gov/ttn/naaqs/standards/pb/s_pb_index.html.

4 Opportunities to Enhance Fuel Efficiency and Reduce Emissions: Benefits of

Reducing Airport Delays

Delay is often the result of the inability of the air transportation system to meet operational demands. The imbalance

between demand and the timely operation of flights can be caused by over-scheduling of the airport, maintenance

and airline operating inefficiencies, weather events, or air traffic management (ATM) programs that hold planes in a

location because of congestion or weather elsewhere. Emissions and fuel use are tied to the amount of time spent in

each phase of aircraft operations, and system delays can cause longer idle and taxi times, and in turn, increase fuel

burn and ground level emissions.

This study investigates ways that ATM inefficiencies result in unnecessary fuel burn and air emissions, caused by

factors such as aircraft idling at airports. The relationship between delay and emissions was examined to develop an

estimate of the emissions reductions and improvements in fuel burn achievable in the absence of ground delays.

Note there are numerous opportunities to reduce aircraft fuel consumption and emissions beyond those associated

with improving performance on the surface or in the vicinity of the airport (e.g. below 3,000 feet). These include

enroute operational initiatives, the use of alternative fuels, improvements in aircraft and aircraft engine design, and

policy options to promote these advances. Further research into ways to promote fuel efficiency should include an

investigation of these opportunities in addition to further assessment of operational initiatives.

4.1 The Relationship between Delay and Emissions

Emissions are related to the amount of fuel consumed during each mode of aircraft operations. For ground delays

this relationship is complicated by the fact that for some delays, airlines switch to APUs or use single-engine taxi

rather than taxiing using all engines. Due to the high uncertainty associated with predicting when an aircraft may

switch to APU power or to single-engine taxiing, full engine taxiing was modeled, even for longer delays. Therefore,

the results of this analysis provide an upper bound for the effects of delays.

The relationship between delay and emissions is influenced by various factors including the fleet mix at the airport

and the particular pollutant that is being evaluated. To provide a better understanding of these factors and to further

explore the relationship between delay and emissions, we focus on the relationship between two metrics: taxi-out

time and the mass of each pollutant emitted for specific aircraft types at three airports.

Scoping & Airport Selection

The delay and emissions analysis focused on the six-week period from November 15

through December 27

, 2005,

one of the busiest travel times of the year. While other time periods were considered, including those in which spring

storms brought delays to the system, the November-December timeframe was chosen to focus more on volume-

related congestion rather than delay that may be attributed to particular weather events—although the two are

interdependent.

While the baseline inventory described in Section 3.1 was created using both instrument flight rules (IFR) and visual

flight rules (VFR) operations, only IFR traffic was considered for the delay and emissions analysis, as VFR operations

were assumed to operate at maximum efficiency.

Of the 325 airports selected for the creation of the baseline

Instrument Flight Rules (IFR) are a set of procedures for operating aircraft where it is assumed that the pilot may

not be able to see outside the aircraft. The majority of commercial flights operate under IFR. Visual Flight Rules

(VFR) apply to flights in which it is assumed the pilot can use visual references to the ground and other aircraft. VFR

flights are mainly performed by general aviation aircraft operating in good weather conditions.

inventory, three airports were studied in more depth to evaluate the relationship between delay and emissions. These

airports were chosen because they represent a spectrum of operational delays and because there are a variety of

aircraft types operating at these airports:

• Hartsfield-Jackson Atlanta International Airport (ATL) is one of the busiest airports in the National Airspace

System (NAS), with over 480,000 annual operations, and is part of a large air traffic hub.

Almost all of

these operations are commercial service flights. An assessment of delays from November 15

through

December 27th at ATL indicated delays due to large numbers of departures during particular peak times of

operation.

• Newport News/Williamsburg International Airport (PHF) has approximately 17,000 annual IFR flights and

belongs to a small air traffic hub. This airport also serves a significant general aviation population with

almost 100,000 VFR flights each year. PHF is a relatively uncongested airport that operates well below

capacity. Congestion at other destination airports was the likely source for delayed flights departing from

PHF during the November-December study timeframe. PHF was investigated because it contributes a

relatively large percentage of emissions to Poquoson County’s emissions inventory.

• Newark Liberty International Airport (EWR) is a busy airport with approximately 225,000 operations and is a

large hub airport. Delays that occurred at EWR during the study timeframe were indicative of the departure

demand generally exceeding the available departure capacity for the airport for almost all times of operation.

Differences in emissions were complicated by the different fleet mixes at these three airports, so two aircraft types

were examined for the analysis: CRJ-200s at airports ATL and PHF, and B737s at ATL and EWR. There were not

enough B737 operations at PHF to make a meaningful comparison. Additionally, there were very few CRJ-200

operations at EWR.

The Relationship between Taxi-Out Time and Emissions

The relationship between taxi-out time and total emissions for the individual aircraft types was examined at the three

study airports. Figure 4.1 shows the relationship between taxi-out time (in minutes) and pollutants emitted (in grams

per operation normalized by the departure mass in metric tons) for Boeing 737’s at ATL. The variability in slope (most

visible for CO) is due to two elements of the aggregations: Boeing 737’s were aggregated together regardless of the

specific type of 737 and the airframes have different engines that lead to different emission rates.

According to U.S. Department of Transportation Bureau of Transportation Statistics, Airport Activity Statistics of

Certificated Air Carriers - Summary Tables - twelve months ending December 31, 2000

(http://www.bts.gov/publications/airport_activity_statistics_of_certificated_air_carriers/2000/index.html) and the BTS

Air Traffic Hubs 2007 map

(http://www.bts.gov/programs/geographic_information_services/maps/hub_maps/2007/html/map.html), an air traffic

hub is a geographic area that enplanes at least 0.05% of all enplaned passengers in the United States. A hub may

have more than one airport in it. This definition of hub should not be confused with the definition used by the airlines

in describing their “hub-and-spoke” route structures. Large air traffic hubs serve 1 percent or more of the total

enplaned passengers in all services and all operations for all communities within the 50 states, the District of

Columbia, and other U.S. areas, while medium hubs serve 0.25% to 0.99% and small hubs serve 0.05% to 0.24%.

Figure 4.1: Taxi-Out Emissions of Boeing 737s at ATL Mapped to their Corresponding Taxi-Out Time. Grams of

pollutant per operation are normalized by the mass of the aircraft in metric tons.

Similar results were produced for the aircraft types studied at EWR and PHF, each showing a similar relationship.

The relationship between delay and emissions provides a common metric for examining the effects of delays that

result from a range of sources. However, the appropriate mitigation techniques are directly tied to the particular

source of delay. Examining the patterns of delay at the three airports used for the analysis, suggests different

initiatives may be helpful for reducing emissions. Figure 4.2 through Figure 4.4 show the patterns of delay found at

each of the airports examined for the analysis. Section 5 will discuss initiatives that target different sources of delay.

Figure 4.2: Average carbon monoxide (CO) and NO

emissions per operation as function of time of day for Boeing

737 aircraft at ATL averaged over the period between November 15

and December 27

, 2005. Increased emissions

are found around 9 o’clock in the morning and between 4pm and 8pm in the evening, corresponding with increases in

taxi out times. This pattern of delay and emissions is related directly to the increases in the number of departure

operation during these times.

There were five flights that departed at 2am; one of these flights experienced a three-hour delay.

Figure 4.3: Average carbon monoxide (CO) and NO

emissions per operation as function of time of day for CRJ-200

aircraft at PHF averaged over the period between November 15

and December 27

, 2005. There is a consistent

range of taxi out times between 10 and 15 minutes with the exception of three hours of operation. At noon there was

only one operation. The delays at 8:00 PM are unlikely to be the result of congestion since the capacity at this airport

is 55 operations per hour and during these two hours of the day only 32 aircraft departed over the six-week period.

Congestion at other destinations likely delayed flights from PHF.

Figure 4.4: Average carbon monoxide (CO) and NO

emissions per operation as function of time of day from Boeing

737’s at EWR averaged over the period between November 15

and December 27

, 2005. This delay pattern is

more indicative of the departure demand generally exceeding the available departure capacity for the airport, with the

exception of the time period between 4:00 AM and 6:00 AM, where the taxi-out times are below 20 minutes and very

few flights depart relative to the rest of the day.

4.2 Potential Benefits from Reduced Ground Delays

Ideally, aircraft would leave the gate, taxi, take off, fly to their destinations, land, and taxi in without experiencing

delay. To understand the potential reductions in local emissions and fuel use for such an ideal system, the pool of

benefits achievable was estimated by comparing to a case with unimpeded taxi times.

The baseline inventory discussed in Section 3.1 was created using reported taxi times obtained from the Bureau of

Transportation Statistics (BTS). BTS provides operations data that list taxi times for air carriers carrying more than

1% of the total passengers. BTS provided data for 113 of the 148 commercial service airports in non-attainment

areas. (These 113 airports are listed in Appendix I.) Twenty-six minutes of taxi time per LTO cycle was

conservatively assumed for those airports without BTS data based on the ICAO test procedure. To measure the

effects of delays, only the 113 airports with BTS data were used in the comparison. As a basis for comparison,

unimpeded taxi times were gathered from the Aviation System Performance Metrics (ASPM) for 75 airports.

For the

remaining airports, unimpeded taxi times were calculated from the airport layout.

EDMS was used to compute total emissions and fuel consumed (see Section 3.1) and the outputs from the two

Aviation System Performance Metrics provides information on individual flight performance and airport efficiency.

See http://aspm.faa.gov/getInfo.asp.

scenarios were compared. Estimates of fuel consumed and mass of CO, hydrocarbons, NO

, SO

, and PM were

compared and the difference between the values for each scenario was used as an estimate of the reductions

possible with the absence of delay. Figure 4.5 shows fuel savings as a percentage of total fuel consumed for the LTO

portion (below 3,000 feet above ground level) of all operations.

Figure 4.6 shows the metric tons of fuel saved for

the 113 airports.

Figure 4.5: Percentage savings in LTO fuel use with the absence of ground delays at the 113 selected airports. With

fewer operations and less fuel consumed, smaller airports are able to achieve large percentage changes when

comparing the operational baseline to the no delay scenario. While at larger airports with more delay and operations,

small percentage changes in the fuel consumption result in large quantities of fuel saved.

Figure 4.6: Metric tons of fuel saved with the absence of ground delays for the 113 selected airports

The smallest potential savings for the 113 airports was approximately 23 metric tons over a year. The largest

Taxi times were adjusted for IFR flights only, but total fuel use and emissions estimates include VFR traffic as well.

VFR traffic was assumed to operate as efficiently as possible.

Metric tons of kerosene-based fuel can be converted to gallons by multiplying by 326.13.

potential savings was over 86,000 metric tons. Overall 17% of the fuel burned below 3,000 feet could be saved with

no taxi-in or taxi-out delay. This translates to 986,000 metric tons of fuel, approximately 320 million gallons per year

out of 1.8 billion gallons (6 million metric tons) of fuel burned below 3,000 feet. 320 million gallons is approximately

1% of the 25.7 billion gallons of jet fuel used in 2005.

Table 4.1 shows how these fuel reductions translate into emissions reductions. Taxi-in and taxi-out are the only

phases of the LTO cycle altered, but the percentage change in total LTO emissions is given in Table 4.1. From 260

metric tons of PM

2.5

to 28,071 metric tons of CO could be saved with no taxi-in or taxi-out delay.

A total of 42,668

metric tons of emission reductions is an overall 15% reduction in LTO emissions.

Table 4.1: Emissions reductions at selected airports with no ground delay

Pollutant

Mass Reduction (metric

tons)

Percentage

Reduction

Carbon Monoxide

28,071

22%

Non-Methane Hydrocarbons

3,978

16%

Volatile Organic Carbons

4,266

16%

4,882

1,211

17%

2.5

260

15%

Fuel

985,954

17%

Not all engines have ICAO smoke numbers (and thus, nonvolatile PM emissions could not be computed for these

engines). PM emissions from aircraft APUs were not computed.

A list of the 113 airports used in the analysis of ground delays is shown in Appendix I.

5 Ways to Promote Fuel Conservation: Initiatives Aimed at Improving Air Traffic

Efficiency

Section 4 describes the effects that ground delays can have on emissions and fuel burn. This study investigated ways

to reduce these effects by promoting greater operational efficiency. To identify methods for improving air traffic

efficiency, an examination of several surface and airspace ATM operational initiatives was conducted. This section

provides illustrative examples of the reductions in ground-level emissions and fuel consumption that can be achieved

by implementing specific ATM initiatives.

Eleven ATM initiatives were surveyed for the study, and four were chosen for modeling based on available data and

publicly available assessments of the initiatives. The initiatives surveyed for this section have broad applicability in

reducing delay throughout the system, and our estimates serve only as representative examples of the magnitude of

their effects. However, understanding the full system-wide impact of multiple, interacting initiatives in different phases

of maturity was beyond the scope of this study. Further research in this regard is recommended.

The 11 initiatives examined span a range of strategies and are described below:

 New and extended runways – New runways create capacity at congested airports and relieve delay by

serving the already existing demand for flights. Runway extensions allow larger aircraft to operate and may

allow more operations of these flights and therefore reduce delay.

 Airport Surface Detection Equipment, Model X (ASDE-X) – Airport surface-surveillance data with better

accuracy, faster update rate, and stronger reliability can improve airport safety and efficiency in all weather

conditions by giving the controllers better knowledge of aircraft locations on the ground.

 Cockpit Display of Traffic Information (CDTI) Assisted Visual Separation (CAVS) – This initiative aims to

avoid capacity loss when weather or other environmental conditions like haze or smoke force an airport to

use instrument approach operations. This is expected to allow airports to continue visual arrival rates under

poor weather conditions, and reduce the frequency and duration of instrument approach operations.

 Integrated Terminal Weather System (ITWS) – This is an ATM tool that provides air traffic managers,

controllers, and airlines with more accurate, easily understood, and immediately useable graphical weather

information and hazard alerts on a single, integrated color display. It is anticipated that, among other effects,

this will enable coordination of the movement of traffic through alternate arrival/departure routes and will

result in overall increases in capacity and reduction of delays.

 Precision Runway Monitor (PRM) – PRM consists of enhanced surveillance capabilities and procedures to

support simultaneous approaches to closely spaced parallel runways, with the goal of increasing throughput

and reducing delays.

 Departure Flow Management (DFM) and Departure Spacing Programs (DSP) – DFM and DSP provide ATM

with the capability to automate coordination of departure releases into congested airspace, with the goal of

improving efficiency and reducing delays.

 Schedule De-Peaking – This refers to measures that adjust demand for departures and arrivals at

congested airports to ensure that the demand does not exceed capacity. The objective is to reduce delays

associated with operating airports at levels at or above capacity.

 RNAV/RNP Arrivals and Departures – RNAV (Area Navigation) refers to a method of navigation that enables

aircraft to fly on more optimal flight paths within the coverage of reference navigation aids and/or within the

limits of the capability of self-contained systems (Flight Management System [FMS]- or Global Positioning

System [GPS]-based). RNP (Required Navigation Performance) refers to RNAV operations within navigation

containment and monitoring, enabling the aircraft navigation system to monitor its achieved navigation

performance within specified tolerances.

 More efficient de-icing procedures – This refers to procedures that enable de-icing activities to be performed

with less waiting time, fewer instances of repeated de-icing, etc.

 Airspace Flow Program – This is a new form of ATM control activity that applies the concept of a Ground

Delay Program (GDP) to airspace regions whose capacity has been reduced due to bad weather or other

factors. The objective is to balance demand and capacity for these airspace regions, and to perform this

balancing with more specificity and less delay than was possible using GDPs.

 Continuous Descent Arrivals (CDA) – This refers to approach procedures that enable aircraft to use lower

power settings during the approach to the airport therefore reducing noise and emissions.

Choice of Metrics and Airport Selection

Airports were selected based on available radar-based flight path data for the months of April 2005 and April 2006.

Given the study’s emphasis on ground-level emissions and fuel burn, taxi time was chosen as the appropriate metric

to evaluate initiatives. Delays associated with departure taxi operations are generally longer than those associated

with arrivals; thus, taxi-out time for departing flights was selected as the primary metric for matching airports to

initiatives.

By using BTS on-time performance data, taxi-out times were extracted and analyzed for Operational Evolution

Partnership (OEP, formerly Operational Evolution Plan) airports that reside in non-attainment areas.58 Taxi-out

times were reviewed in 15-minute bins to identify potential periods of airport or terminal-area congestion. Notable

peaks in taxi-out times were identified.

Figure 5.1 shows the variation in taxi-out times for Cleveland Hopkins Airport (CLE

) for the month of April 2005.

It is important to note that ATM initiatives effect other phases of flight although the focus of this study was on taxi

times.

The OEP is a rolling ten-year plan to address capacity and delay problems through the NAS by focusing on

selected airports, with these airports changing over the years as various issues are corrected by the FAA. At the time

of this research all OEP airports except those in Florida and Honolulu were located in non-attainment areas (30 of the

35). The 35 airports included in the OEP account for about 75 percent of all passenger enplanements.

Part of a medium hub

(http://www.bts.gov/programs/geographic_information_services/maps/hub_maps/2007/html/map.html).

Figure 5.1: Taxi-out times for Cleveland Hopkins Airport (CLE) during the month of April 2005.

Plots similar to Figure 5.1 were created to determine the nature of delays at OEP airports in non-attainment areas. In

this example, we see three days containing significant delay, with taxi-out times greater than 20 minutes throughout

the day. Examining the hour axis we see consistent evening delays (around 18:00 hours or 6pm) throughout the

month.

After examining the patterns of delay, OEP non-attainment area airports were matched to FAA initiatives based on

the pattern of delay at the airport and the potential for improving operational efficiency with the particular initiative.

Multiple airports were chosen for some initiatives (based on available data) to provide a range of the benefits.

The following sections provide estimates of the potential improvements in air traffic efficiency and fuel consumption

that can be achieved with the implementation of four types of initiatives at representative airports:

 Airspace Flow Program effects at Boston Logan International Airport and O’Hare International Airport

(Section 5.1)

 Schedule De-peaking effects at Phoenix International Airport, Boston Logan International Airport,

Minneapolis St. Paul International Airport, Dulles International Airport, and Memphis International Airport

(Section 5.2)

 Continuous Descent and Arrivals (CDAs) effects at Los Angeles International Airport

(Section 5.3)

 New Runways and Runway Extensions effects at Minneapolis St. Paul International Airport (Section 5.4)

Boston Logan International and O’Hare International are each part of large hubs.

Phoenix International, Minneapolis St. Paul International, and Dulles International are each part of large hubs;

Memphis International is part of a medium hub.

Los Angeles International Airport is part of a large hub.

5.1 Airspace Flow Programs in Support of Severe Weather Avoidance Procedures

Airspace Flow Programs (AFPs) refer to programs that allow air traffic management specialists to restrict flights with

the use of defined airspace, as opposed to the use of Ground Delay Programs (GDPs). AFPs can help reduce delays

when used during severe weather events.

Before the advent of AFPs, reductions in en route capacity caused by severe weather were addressed, in part, by

using GDPs, which delay flights to and from airports on both sides of the bad-weather area, regardless of the

proximity of the flight routes to the bad weather. With the introduction of AFPs, only flights flying through the affected

area are delayed. Additionally, operators of those flights have the option of routing around the affected area, further

reducing the number of flights delayed. This type of ATM initiative promotes greater specificity in the assignment of

delay for flights attempting to depart: those not using the weather-impacted airspace will not be assigned delay under

an AFP, whereas they might have been assigned delay under a GDP. Assuming that the AFP results in fewer

delayed outbound flights, this will result in more departures and fewer taxi-out delays. Thus, while both AFPs and

GDPs necessarily result in delays in order to cope with decreased en route capacity, AFPs have the potential to

result in less widespread delays. To provide a measure of one of the benefits of implementing AFPs, this analysis

provides an estimate of the fuel and emissions savings related to the reduced impacts of severe en route weather on

taxi-out times.

Impact Estimation Method

To provide an estimate of the benefits of Airspace Flow Programs, changes in taxi-out times with the implementation

of AFPs were compared to those resulting from Ground Delay Programs. The shorter delays associated with AFPs

were used as an estimate of the benefits. Two sets of taxi times were examined: taxi-out times with and without

GDPs and taxi-out times with and without AFPs (for the year 2005 and 2006 respectively). 26 airports had readily

available data to support this analysis. This comparison indicated that, while AFPs applied to cope with severe en

route weather resulted in shorter taxi-out times than GDPs for some airports (17 airports), others experienced longer

taxi-out times with the use of AFPs (9 airports). Further analysis would be needed to understand the differences

between these groups of airports. To focus solely on the benefits of this type of initiative, 2 of the 17 airports were

selected for further examination.

Boston Logan International Airport (BOS) and O’Hare International Airport (ORD) were chosen for further analysis.

Boston Logan International Airport (BOS) experienced a 20% average increase in taxi-out times when an AFP was

implemented and a 30% increase when GDPs were implements. Similarly, Chicago O’Hare International Airport

(ORD) showed a 27% increase with AFPs and a 30% increase for GDPs.

To estimate the impacts during periods of congestion, a sample day was selected on which multiple airports

experienced increased delays as a result of severe weather (April 20, 2005). Bad weather over New York,

Pennsylvania, Ohio, and Indiana affected airspace capacity, and ORD and BOS experienced delays during the

afternoon hours.

Increased hourly taxi times for BOS and ORD are shown in Figure 5.2. BTS taxi-out data were

obtained for that day and the estimated increases in taxi-out times obtained from the above comparison were then

applied to the April 20 sample day.

Note that a GDP was implemented at ORD, but not at BOS that day, and the decreases in taxi time reflect a

different starting point.

Figure 5.2: Hourly minutes of delay at BOS (left) and ORD (right) during the afternoon of April 20, 2005 compared to

average minutes of delay for the entire month of April 2005. Bad weather brought delays resulting in longer taxi out

times during the afternoon hours.

For the period of congestion observed in the BTS taxi-out data (1:00pm-4:00pm), BOS was estimated to experience

16% shorter taxi-out times for total flights with the implementation of an AFP instead of a GDP. ORD was estimated

to experience an 11% reduction. These reduced taxi times were estimated to result in a 9% decrease in LTO fuel

burn for BOS and a 4% reduction for ORD compared to what is expected with the use of GDPs. Reductions in

emissions are shown in Table 5.1. (THC is total hydrocarbon.)

Table 5.1: Reduction in emissions and fuel burn due to the implementation of AFPs instead of GDPs at Boston

Logon and Chicago O'Hare airports.

THC

NMHC

VOCs

Fuel

BOS

13.2%

8.3%

4.3%

8.9%

9.2%

ORD

8.2%

3.6%

1.4%

3.8%

4.3%

5.2 Schedule De-Peaking

Schedule de-peaking refers to reducing the demand for departures and arrivals during specific periods in which

demand exceeds the capacity of the airport.

Reduction of demand peaks when demand is close to, or greater than

maximum capacity, can significantly affect average delays and queue sizes, as well as their variability from flight to

flight.

A range of studies was reviewed to develop a simplified means of estimating the effects of schedule de-peaking (Fan

and Odoni 2002; Zhang, Menendez et al. 2003; Le, Donohue et al. 2005; Le 2006; Levine and Gao 2007). All sources

suggest that bringing demand into alignment with capacity throughout the day can affect taxi-out times. However, the

size and dynamics of the effect depend upon airport-specific factors and the timing and extent of the de-peaking.

More thorough, nationwide analysis for specific airports was beyond the scope of this project. Instead, a conservative

estimate of the magnitude of de-peaking effects was used. The study assumed that de-peaking would reduce excess

taxi-out times (that is, times in excess of the unimpeded taxi-out time) by approximately half during times of excess

This study did not evaluate the methods of achieving schedule de-peaking but rather the theoretical gains if the

schedule was spread out across the day. Past and current initiatives to reduce schedules include voluntary efforts at

Chicago and slot control at LaGuardia Airport (LGA). There are other methods including slot auctions, peak-time

pricing, and other economic schemes to increase the cost of operating certain flights to reduce demand. However,

pricing the operations is not the sole option to reduce the schedule of operations by carriers.

demand.

Impact Estimation Method

This assumption for de-peaking effects was modeled for five airports: Phoenix International Airport (PHX), Boston

Logan International Airport (BOS), Minneapolis St. Paul International Airport (MSP), Dulles International Airport (IAD),

and Memphis International Airport (MEM). For these 5 airports, unimpeded taxi-out time ranged from 7 to 10 minutes,

and times in excess were divided by 2 to estimate the effects of de-peaking. The results of this for PHX for April 2005

appear in Figure 5.3.

Figure 5.3: Original and modified hourly taxi-out times for PHX are based on monthly average for April 2005

(estimated unimpeded time of 8 minutes)

Using the new taxi-out times to estimate the effect of schedule de-peaking, taxi-out fuel burn reductions of between

16% and 23% were found. This translates to a range of 6% to 10% reduction for total LTO fuel burn. Fuel burn

reductions for all 5 airports appear below in Table 5.2.

Table 5.2: Estimated reductions from schedule de-peaking

Carbon Monoxide

Non-Methane

Hydrocarbons

Volatile Organic

Carbons

Fuel

BOS

17.5%

9.6%

9.5%

3.6%

9.6%

10.4%

IAD

13.6%

6.3%

2.3%

6.1%

MEM

17.0%

7.4%

1.8%

6.9%

7.9%

MSP

18.0%

8.7%

3.4%

9.3%

10.1%

PHX

14.5%

6.1%

2.0%

6.1%

6.9%

5.3 Continuous Descent Arrivals

Continuous Descent Arrival procedures reduce noise and emissions by changing the approach path so that it more

closely follows a 3˚ glide slope as shown in Figure 5.4. Using the 3˚ glide slope, aircraft are able to reduce the thrust

of the engines to reduce fuel burn and lessen the noise impacts on approach. Figure 5.4 depicts the vertical

dispersion that normally occurs on approach and landing using the downwind approach at Los Angeles International

Airport (LAX).

Figure 5.4: Baseline downwind approaches at LAX from Dinges, 2007.

An estimate of the benefits of implementing CDA procedures is provided by (Dinges 2007). Dinges evaluated the

benefits for different fractions of the aircraft using CDA. The five threshold levels were 5.9% (threshold 1), 21%

(threshold 2), 42.9% (threshold 3), 67.3% (threshold 4) and 100% (all-CDA). These thresholds were chosen to

explore the space of potential benefits from CDA and illustrate the incremental gains available with varying levels of

properly equipped aircraft and conditions suitable for flying the approach. Table 5.3 shows the range of benefits from

converting to CDA paths.

Table 5.3: Emissions and fuel burn percentage reductions relative to the baseline below 3,000 feet, comparing five

levels of CDA usage to the baseline for all modeled approaches to LAX (Dinges 2007).

Percent Reduction for Each CDA Threshold Level

Category of

Reduction

5.9%

Threshold

21%

Threshold

42.9%

Threshold

67.3%

Threshold

100%

Threshold

0.2%

1.6%

3.7%

5.5%

6.8%

THC

0.1%

0.9%

2.2%

3.4%

4.5%

VOC

0.1%

0.9%

2.2%

3.4%

4.5%

1.7%

6.0%

13.1%

21.7%

28.4%

1.2%

4.5%

9.7%

15.6%

19.9%

Percent Reduction for Each CDA Threshold Level

Category of

Reduction

5.9%

Threshold

21%

Threshold

42.9%

Threshold

67.3%

Threshold

100%

Threshold

Fuel

1.2%

4.5%

9.7%

15.6%

19.9%

5.4 New Runways and Runway Extensions

New runways and runways extensions are part of the FAA Operational Evolution Partnership (Version 8, FAA 2006):

New runways and runway extensions provide very significant capacity increases for the NAS. Since

1999, ten new runways have opened at the 35 Operational Evolution Plan airports, providing these

airports with the potential to accommodate almost 1.2 million more operations annually. Currently,

there are eight runway projects (five new runways, one runway extension, and two airfield

reconfigurations) included in the OEP. All eight will be commissioned by 2010 providing these

airports with the potential to accommodate more than one million more annual operations.

Impact Estimation Method

The impact of new runways was estimated by examining Minneapolis St. Paul International Airport (MSP), an airport

in which an additional runway became operational between 2005 and 2006. The two days used for this comparison

were April 2, 2005 (before the runway was completed) and April 26, 2006 (with the new runway operational). The

period between 9:00 and 12:00 on April 2, 2005 was selected because of increased taxi out times. The post-

enhancement period (9:00-12:00 on April 24, 2006) was selected because of similar weather and similar volume, as

compared to the baseline period. These two data samples were used for estimating the benefits.

For the 2005 time period, flights had an average taxi-out time of 19 minutes; for the 2006 time period, flights had a

taxi-out time of 16 minutes, an approximate 15% improvement. By applying, the 2006 taxi-out time to the 2005 flights,

we determined how a 15% improvement would decrease the emissions of the 2005 flights. As noted in Section 4.1,

we assume that emissions are linear with taxi out time so a 15% reduction in taxi out time reduces taxi-out emissions

by 15%. However, taxi-out emissions are only a portion of the departure emissions below 3,000 feet. Table 5.4 shows

the reduction of the LTO emissions for a 15% reduction in taxi time. A summary of all initiatives is shown in XXXX.

Table 5.4: Table of percentage reduction in fuel burn and emissions achieved by applying the 2006 taxi out time to

the 2005 flights for an effective 15% reduction in taxi-out time

Pollutant/Fuel

% Change

Carbon Monoxide

12%

Hydrocarbons

VOC

Fuel

Table 5.1: Summary of emissions reductions potential from operational initiatives

Emissions Reduction

THC/NMHC

VOC

Fuel

Initiative

Min

Max

Min

Max

Min

Max

Min

Max

Min

Max

Min

Max

AFPs (BOS, ORD)

8.2%

13.2%

3.6%

8.3%

3.6%

8.3%

1.4%

4.3%

3.8%

8.9%

4.3%

9.2%

Schedule depeaking

(BOS, IAD, MEM, MSP,

PHX)

13.6%

18.0%

6.1%

9.6%

6.1%

9.5%

1.8%

3.6%

6.1%

9.6%

6.1%

10.4%

CDA (LAX)

0.2%

6.8%

0.1%

4.5%

0.1%

4.5%

1.7%

28.4%

1.2%

19.9%

1.2%

19.9%

New Runways and

Runway Extensions

(MSP)

12%

Notes:

NMHC for schedule depeaking initiative; THC for all other initiatives.

DRAFT OCTOBER 19, 2009 – DO NOT CITE OR QUOTE

6 Conclusions and Recommendations

This study analyzed aircraft LTO emissions at 325 airports with commercial activity in the U.S (includes 263

commercial service airports and 62 airports that are either reliever or general aviation airports) for operations that

occurred from June 2005 through May 2006. The flights studied represent 95 percent of the commercial jet aircraft

operations for which flight plans were filed and 95 percent of the operations with ICAO certified engines. Of the 325

airports (or the 263 commercial service airports), 148 are commercial service airports in at least one of 118 ambient

air quality NAAs (for ozone, CO, PM

2.5,

10,

or NO

) for 2005 using the criteria specified by the National

Ambient Air Quality Standards (40 CFR Part 50).

The purpose of this study was to assess the impact of aircraft operations on air quality in these NAAs. This study

found that aircraft LTO emissions during the period June 2005 through May 2006 at the 148 U.S. commercial service

airports in the 118 NAAs represented the following average percentages of the 2002 emissions inventory in these

NAAs:

0.44% of carbon monoxide (CO) emissions, 0.66% of oxides of nitrogen (NO

) emissions, 0.48% of

emissions of volatile organic compounds (VOCs), 0.37% of oxides of sulfur (SO

) emissions, and 0.15% of fine

particulate matter (PM

2.5

) emissions.

Looking more broadly, this study found that aircraft LTO emissions during the period June 2005 through May 2006 at

the 325 U.S. airports with commercial activity included in the study represented the following percentages of the total

2002 U.S. National Emissions Inventory: 0.18% of CO emissions, 0.41% of NO

emissions, 0.23% of VOCs, 0.07% of

emissions, and 0.05% of PM

2.5

emissions.

Air quality and health effects impacts from aircraft LTO operations were assessed by removing all aircraft operations

from the inventories and modeling ozone and PM concentrations and population based health impacts. Within the

capabilities of the modeling, the impacts on health from aircraft emissions were found to derive almost entirely from

fine ambient particulate matter. The dominant emissions from aircraft that contribute to ambient PM

2.5

are the

secondary PM precursor emissions, SO

and NO

, as well as direct emissions of primary PM

2.5

. SO

emissions

depend on fuel sulfur levels and overall fuel burn. NO

and PM emissions depend on combustor and engine

technology in addition to overall fuel burn. The contribution of aircraft emissions to the national annually-averaged

ambient PM

2.5

level was estimated to be 0.01µg/m

. On a percentage basis, the contribution is approximately 0.08%

for all counties and 0.06% for counties in NAAs.

The aircraft contributions to county-level ambient PM

2.5

concentrations ranged from approximately 0% to 0.5%. Aircraft emissions were also estimated to contribute 0.12%

(0.10 parts per billion) to average 8-hour ozone values in both attainment and non-attainment areas. Near some

urban centers aircraft emissions reduced ozone, whereas in suburban and rural areas, aircraft emissions increased

ambient ozone levels. The largest county-level decrease was 0.6%; the largest county-level increase was 0.3%.

The health impacts of aircraft LTO emissions were derived almost entirely from fine ambient particulate matter.

Nationally, about 160 yearly incidences of PM-related premature mortality were estimated due to ambient particulate

matter exposure attributable to aircraft emissions at the 325 airports studied (with a 90% confidence interval of 64 to

270 yearly incidences). Although the health impacts we estimate for aircraft LTO emissions are important, it is very

likely

that they constitute less than 0.6% of the total adverse health impacts due to poor air quality from

anthropogenic emissions sources in the United States. One-third of these 160 premature mortalities were estimated

to occur within the greater Southern California region, while another fourteen counties (located within NY, NJ, IL,

2005 is the base year for aircraft emissions, and 2002 is the base year for non-aircraft emissions.

Note that these estimates for percent contributions to total ambient concentrations carry uncertainties due to the

fact that some emissions sources are not well quantified in U.S. National Emissions Inventories.

Nonvolatile PM mass was not computed for non-ICAO certified aircraft engines; however, sulfates- and organics-

related PM mass were computed. No PM mass was computed for APUs.

Northern CA, MI, TN, TX and OH) accounted for approximately 21 percent of total premature mortality. In total, 47

counties within the United States had a PM-related premature mortality risk associated with aircraft emissions that

was greater than 1 incidence per year. Other health impacts, such as chronic bronchitis, non-fatal heart attacks,

respiratory and cardiovascular illnesses were also associated with aircraft emissions. No significant health impacts

were estimated due to changes in ozone concentrations attributable to aircraft emissions.

There are several important assumptions and limitations associated with the results of this study. The method used to

estimate aircraft primary PM emissions in this study (known as FOA3a) includes margins to conservatively

accommodate uncertainties in aircraft PM emissions. An error was made in the specification of the fuel sulfur level for

some of the airports in this inventory such that the aircraft SO

inventory is expected to be biased towards

underestimating the contribution of aircraft by 20 percent (i.e. the contribution of aircraft to the national SO

inventory

may be closer to 0.07%). This would have an effect on sulfate secondary PM contributions to fine PM air quality and

health effects as well. The use of a 36 km x 36 km grid scale for the air quality analyses is expected to underestimate

health impacts, especially those that may occur close to airport boundaries. Omitting the effect of cruise level

emissions on surface air quality is also expected to lead to underestimation of health impacts by an unknown amount,

especially for fine primary and secondary PM. Further, analysis of only one year may lead to overestimation or

underestimation of aircraft impacts due to year-to-year changes in meteorology. Non-aircraft sources were also not

included (e.g. emissions of ground service equipment and other aircraft sources). Finally, we report the results for

one concentration-response relationship for the health effects of ambient PM; a range of concentration-response

relationships has been reported in the literature. The net effect of these assumptions and limitations is not known.

Further research is recommended into these areas.

General aviation aircraft emissions were not included in our emissions inventory since GA aircraft are responsible for

less than 1 percent of fuel use by volume. However, a separate estimate of lead emissions from GA aircraft was

made (most piston-engine powered GA aircraft operate on leaded aviation gasoline; gas turbine powered jet engines

and turboprops operate on Jet A which does not contain significant levels of lead). We estimate that in 2002

approximately 280 million gallons of aviation gasoline were supplied for GA use in the U.S., contributing an estimated

565 metric tons of lead to the air, and comprising 46 percent of the 2002 U.S. National Emissions Inventory (NEI) for

lead. We did not estimate the health impacts of these lead emissions.

Aircraft emissions are influenced by weather, air traffic management, and other inefficiencies that compound,

resulting in increased fuel burn and emissions. During a one-year period, airport delays accounted for approximately

320 million gallons of fuel use due to increased taxi times for the 113 non-attainment airports examined in Section 4.

This is approximately 1% of all jet fuel used in the U.S. during 2005, and approximately 17% of fuel use during the

LTO portion of the flight for these 113 airports. Based on these results, unimpeded taxi times would result in average

LTO emissions reductions of 22% (28,000 metric tons) for CO, 7% (5,000 metric tons) for NO

, 16% (4,000 metric

tons) each for VOCs and non-methane hydrocarbons, 17% (1,000 metric tons) for SO

, 15% (260 metric tons) for

2.5

, and 17% (986,000 metric tons) for fuel. These values represent about five percent of LTO emissions in these

non-attainment areas.

While there are many strategies available to achieve these reductions, including technological, operational and policy

options, the relationship between taxi-out time and emissions suggests that ATM initiatives have the potential to play

an important role in increasing operational efficiency and, in turn, reducing emissions and fuel use at U.S. airports.

This study provides illustrative examples of potential reductions in fuel use and emissions that may be obtained

through initiatives such as airspace flow programs, schedule de-peaking, continuous decent arrivals, and new

runways. To increase efficiency without adversely affecting safety, noise and security, operational initiatives must be

implemented with consideration of the larger system, and the numerous, complex interdependencies that are inherent

in the system. Further, there are no universal mitigation strategies for operational efficiency, and a single technology

or procedure will not be applicable at all U.S. airports.

This study highlights some of the needs for future work in the area of aviation fuel conservation and emissions

reduction. Some of the data, methods, and modeling used for the study would benefit from further development:

The dominance of PM health effects suggests the need for more complete PM measurements from aircraft engines.

An agreed upon test method is needed and is now under development. The current analytical methods (see

Appendix B) are intended as temporary estimation methods until mass emissions data are collected for ICAO-

certified engines. PM data is also needed for APUs and non-ICAO certified engines.

As noted above, further analysis of air quality impacts of aviation emissions is required to understand the impacts

cruise level emissions, better grid resolution (near airport health effects), and year-to-year meteorological variations.

Investigation of source-specific dose response functions for health impacts may also be beneficial. It is currently not

known if primary particulate matter due to aviation has unique health impacts that differ from other emission sources.

Currently, dose-response functions used to assess impacts of particulate matter do not discriminate between PM

sources or components. Research is underway to better understand source-specific and component-specific health

impacts of particulate matter.

There are numerous ATM initiatives that can effectively reduce delays. This study estimates the benefits of only a

few; and even in these cases only illustrative cases are presented. Further study is recommended to more fully

evaluate the potential benefits of different ATM initiatives. Moreover, there are numerous opportunities to reduce

aircraft fuel consumption and emissions that are beyond the scope of this study including the use alternative fuels,

improvements in aircraft and aircraft engine design, and policy options to promote these advances. Further research

into ways to promote fuel efficiency should include an investigation of these opportunities in addition to further

assessment of operational initiatives.

To better understand the relationship between delay and ground-level emissions and fuel burn, it is necessary to

model the numerous factors that govern APU use and single engine taxi. Further analysis of variations in APU usage

by carrier, aircraft type, airport, season, and operating environment is also needed to understand engine cut-off and

APU use.

7 References

Abt (2005). Environmental Benefits Mapping and Analysis Program (BenMAP) User's Manual. Bethesda, MD, Abt

Associates, Inc. Prepared for the U.S. Environmental Protection Agency Office of Air Quality and Standards.

Byun, D. W. and K. L. Schere (2006). "Review of the Governing Equations, Computational Algorithms, and Other

Components of the Models-3 Community Multiscale Air Quality (CMAQ) Modeling System." Applied Mechanics

Reviews 59(2): 51-57.

CAFE (2005). Service Contract for Carrying out Cost-Benefit Analysis of Air Quality Related Issues, in particular in

the Clean Air for Europe (CAFE) Programme: Volume 2: Health Impact Assessment, AEA Technology Environment.

Cohen et al., (2004), “Chapter 17: Urban Air Pollution,” in Comparative Quantification of Health Risks

, World Health

Organization, Volume 2, http://www.who.int/publications/cra/chapters/volume2/part2/en/index.html

CSSI (2007). Emissions and Dispersion Modeling System (EDMS) User's Manual. Washington DC, Prepared for the

Federal Aviation Administration Office of Environment and Energy, FAA-AEE-07-01.

Dinges, E. (2007). Determining the Environmental Benefits of Implementing Continuous Descent Approach

Procedures. ATM Seminar, Barcelona, Spain.

DOE Energy Information Administration (2006). Fuel production volume data obtained from

http://tonto.eia.doe.gov/dnav/pet/hist/mgaupus1A.htm, [accessed November 2006].

EPA (1999). Final Regulatory Impact Analysis: Tier 2/Gasoline Sulfur Rule, U.S. Environmental Protection Agency

Office of Air and Radiation, EPA 420-R-99-023.

EPA (2000). Final Regulatory Impact Analysis: Clean Diesel Trucks, Buses, and Fuel: Heavy-Duty Engine and

Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements, U.S. Environmental Protection Agency

Office of Air and Radiation, http://www.epa.gov/otaq/highway-diesel/regs/2007-heavy-duty-highway.htm, [accessed

September 2007].

EPA (2004). Final Regulatory Impact Analysis: Control of Emissions from Nonroad Diesel Engines, U.S.

Environmental Protection Agency Office of Air and Radiation, http://www.epa.gov/nonroad-

diesel/2004fr/420r04007.pdf, [accessed September 2007].

EPA (2005). Regulatory Impact Analysis for the Final Clean Air Interstate Rule, U.S. Environmental Protection

Agency Office of Air and Radiation, EPA-452/R-05-002.

EPA (2006). 2006 National Ambient Air Quality Standards for Particle Pollution - Regulatory Impact Analysis (RIA).

Research Triangle Park, NC, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency.

EPA (2007a) Airport-specific emissions of lead from combustion of leaded aviation gasoline.

http://www.epa.gov/ttn/chief/net/2002inventory.html, [accessed January 2008].

EPA (2007b). Guidance on the Use of Models and Other Analyses for Demonstrating Attainment of Air Quality Goals

for Ozone, PM2.5, and Regional Haze, U.S. Environmental Protection Agency, EPA-454/B-07-002.

EPA (2007c). Technical Support Document for the Proposed Locomotive-Marine Rule: Air Quality Modeling.

Research Triangle Park, NC, U.S. Environmental Protection Agency Office of Air Quality Planning and Standards.

EPA (2008). Final Ozone NAAQS Regulatory Impact Analysis, Office of Air Quality Planning and Standards, U.S.

Environmental Protection Agency.

FAA (2006). Airport Congestion, ARPT-1: New Extended Runways,

http://www.faa.gov/programs/oep/v8/smart%20sheets/arpt-1%20new%20and%20extended%20runways.htm.

[accessed May 2006].

Fan, T. and A. Odoni (2002). "A Practical Perspective on Airport Demand Management." Air Traffic Control Quarterly

10(3): 285-306.

66 Fed. Reg. 5002 (January 18, 2001). “Control of Air Pollution from New Motor Vehicles: Heavy-Duty Engine and

Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements.”

69 Fed. Reg. 38958 (June 29, 2004). “Control of Emissions of Air Pollution from Nonroad Diesel Engines and Fuel.”

72 Fed. Reg. 15938 (April 3, 2007). “Control of Emissions of Air Pollution from Locomotives and Marine

Compression-Ignition Engines Less than 30 Liters per Cylinder.”

Hill, L. B., (2005). “An Analysis of Diesel Air Pollution and Public Health in America,” v1.3, June 2005, Clean Air Task

Force, http://www.catf.us/publications/reports/Diesel_in_America_Technical_Paper.pdf

IPCC (2007). Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the

Fourth Assessment Report of the Intergovernmental Panel on Climate Change, M.L. Parry, O.F. Canziani, J.P.

Palutikof, P.J. van der Linden and C.E. Hanson, Eds., Cambridge University Press, Cambridge, UK, 976pp.

Le, L. (2006). Demand Management at Congested Airports: How Far Are We from Utopia? Ph. D. Dissertation.

Center for Air Transportation System Research

, George Mason University.

Le, L., G. L. Donohue, et al. (2005). "Auction-Based Slot Allocation for Traffic Demand Management at Hartsfield

International Airport: A Case Study." Journal of the Transportation Research Board

1888: 50-58.

Levine, B. and H. Gao (2007). Aircraft Taxi-Out Emissions at Congested Airports and the Implication for Aviation

Emissions Reduction in the United States. TRB 2007 Annual Meeting, Washington, DC.

NARSTO Synthesis Team (2000). An Assessment of Tropospheric Ozone Pollution: A North American Perspective.

SCAQMD (2007). Final 2007 Air Quality Management Plan. Available at:

http://www.aqmd.gov/aqmp/07aqmp/index.html. Accessed November 8, 2007.

Slovic, P. (2002). Perception of Risk Posed by Extreme Events. Risk Management Strategies in an Uncertain World

Palisades, NY.

World Bank (2007). Cost of Pollution in China: Economic Estimates of Physical Damages, The World Bank, State

Environmental Protection Administration, P.R. China.

Zhang, Y., M. Menendez, et al. (2003). Analysis of De-peaking Strategies Implemented by American Airlines,

NEXTOR Technical Paper.

Appendix A Study Participants

Zachariah Adelman

Research Associate

University of North Carolina at Chapel Hill

Institute for the Environment

Gayle Ratliff

Research Engineer

Massachusetts Institute of Technology

Department of Aeronautics and Astronautics

Dr. Saravanan Arunachalam

Research Associate Professor

University of North Carolina at Chapel Hill

Institute for the Environment

Christopher Sequeira

Graduate Research Assistant

Massachusetts Institute of Technology

Department of Aeronautics and

Astronautics/Engineering Systems Division

Dr. Bok Haeng Baek

Research Associate

University of North Carolina at Chapel Hill

Institute for the Environment

Dr. Terence Thompson

Director of Business and Strategic Development

Metron AviationTheodore Thrasher

Director of Simulation, Modeling, and Analysis

CSSI, Inc.

Michael Graham

Environmental Analyst

Metron Aviation

Dr. Roger Wayson

National Expert on Emissions and Modeling

John Volpe National Transportation Systems

Center

Environmental Measurement and Modeling

Division

Dr. Adel Hanna

Research Professor

University of North Carolina at Chapel Hill

Institute for the Environment

Tyler White

Environmental Analyst

Metron Aviation

Dr. Donald McCubbin

Abt Associates, Inc.

Computer Sciences Corporation (CSC)

Andrew Holland

Research Associate

University of North Carolina at Chapel Hill

Institute for the Environment

Melissa Ohsfeldt

Analyst

CSSI, Inc

REALab

Appendix B Study Airports

The study analyzed 325 airports with commercial activity (commercial service, reliever, and general aviation airports)

in the U.S for operations that occurred from June 2005 through May 2006. These airports and their IFR and VFR

operations (and LTOs) for this time period are shown below.

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

ANC

TED STEVENS

ANCHORAGE INTL

ANCHORAGE

311,729

155,865

FAI

FAIRBANKS INTL

FAIRBANKS

109,190

54,595

JNU

JUNEAU INTL

JUNEAU

12,875

6,438

KTN

KETCHIKAN INTL

KETCHIKAN

8,218

4,109

MRI

MERRILL FIELD

ANCHORAGE

185,188

92,594

SIT

SITKA ROCKY GUTIERREZ

SITKA

3,807

1,904

BFM

MOBILE DOWNTOWN

MOBILE

9,372

4,686

BHM

BIRMINGHAM INTL

JEFFERSON

142,275

71,138

HSV

HUNTSVILLE INTL

MADISON

36,868

18,434

MGM

MONTGOMERY RGNL

MONTGOMERY

17,143

8,572

MOB

MOBILE RGNL

MOBILE

23,176

11,588

MSL

NORTHWEST ALABAMA

RGNL

COLBERT

44,380

22,190

FSM

FORT SMITH RGNL

SEBASTIAN

14,676

7,338

LIT

ADAMS FIELD

PULASKI

65,507

32,754

ROG

ROGERS MUNICIPAL-

CARTER

BENTON

6,807

3,404

XNA

NORTHWEST ARKANSAS

RGNL

BENTON

35,054

17,527

IFP

LAUGHLIN/BULLHEAD INTL

MOHAVE

27,994

13,997

PHX

PHOENIX SKY HARBOR

INTL

MARICOPA

616,517

308,259

SDL

SCOTTSDALE

MARICOPA

40,000

20,000

TUS

TUCSON INTL

PIMA

279,103

139,552

YUM

YUMA INTL

YUMA

174,259

87,130

APC

NAPA COUNTY

NAPA

12,020

6,010

BFL

MEADOWS FIELD

KERN

87,613

43,807

BUR

BURBANK-GLENDALE-

PASADE

LOS ANGELES

190,447

95,224

CIC

CHICO MUNI

BUTTE

42,849

21,425

CRQ

MC CLELLAN-PALOMAR

SAN DIEGO

199,877

99,939

FAT

FRESNO YOSEMITE INTL

FRESNO

150,309

75,155

IPL

IMPERIAL COUNTY

IMPERIAL

73,054

36,527

IYK

INYOKERN

KERN

40,567

20,284

LAX

LOS ANGELES INTL

LOS ANGELES

664,609

332,305

LGB

LONG BEACH/ DAUGHERTY

LOS ANGELES

351,408

175,704

MCE

MERCED MUNICIPAL/

MACREADY FIELD

MERCED

27,972

13,986

MHR

SACRAMENTO MATHER

SACRAMENTO

20,396

10,198

MOD

MODESTO CITY

STANISLAUS

75,379

37,690

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

MRY

MONTEREY PENINSULA

MONTEREY

43,020

21,510

OAK

METROPOLITAN OAKLAND

INTL

ALAMEDA

340,174

170,087

ONT

ONTARIO INTL

SAN BERNARDINO

139,930

69,965

OXR

OXNARD

VENTURA

94,653

47,327

PSP

PALM SPRINGS INTL

RIVERSIDE

92,722

46,361

SAN

SAN DIEGO INTL-

LINDBERG

SAN DIEGO

245,719

122,860

SBA

SANTA BARBARA MUNI

SANTA BARBARA

69,657

34,829

SCK

STOCKTON

METROPOLITAN

SAN JOAQUIN

83,298

41,649

SFO

SAN FRANCISCO INTL

SAN MATEO

376,966

188,483

SJC

NORMAN Y. MINETA SAN

JOSE INTL

SANTA CLARA

221,361

110,681

SMF

SACRAMENTO INTL

SACRAMENTO

180,203

90,102

SMO

SANTA MONICA MUNI

LOS ANGELES

32,647

16,324

SNA

JOHN WAYNE AIRPORT

ORANGE

361,921

180,961

SUU

TRAVIS AFB

SOLANO

1,091

546

VCV

SOUTHERN CALIFORNIA

LOGISTICS

SAN BERNARDINO

73,276

36,638

VIS

VISALIA MUNI

TULARE

33,777

16,889

VNY

VAN NUYS

LOS ANGELES

31,642

15,821

APA

CENTENNIAL

ARAPAHOE

47,961

23,981

ASE

ASPEN-PITKIN CO/ SARDY

FIELD

PITKIN

43,939

21,970

BJC

ROCKY MOUNTAIN

METROPOLITAN

JEFFERSON

10,995

5,498

COS

CITY OF COLORADO

SPRING

EL PASO

155,740

77,870

DEN

DENVER INTL

DENVER

606,129

303,065

EGE

EAGLE COUNTY RGNL

EAGLE

20,701

10,351

GJT

WALKER FIELD

MESA

23,049

11,525

MTJ

MONTROSE RGNL

MONTROSE

13,601

6,801

TEX

TELLURIDE RGNL

SAN MIGUEL

10,879

5,440

BDL

BRADLEY INTL

HARTFORD

151,685

75,843

GON

GROTON-NEW LONDON

NEW LONDON

56,706

28,353

HVN

TWEED-NEW HAVEN

NEW HAVEN

62,430

31,215

OXC

WATERBURY-OXFORD

NEW HAVEN

6,954

3,477

DCA

RONALD REAGAN

WASHINGTON

ARLINGTON

290,998

145,499

IAD

WASHINGTON DULLES

INTL

LOUDOUN

495,340

247,670

ILG

NEW CASTLE COUNTY

NEW CASTLE

15,548

7,774

APF

NAPLES MUNI

COLLIER

15,711

7,856

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

BCT

BOCA RATON

PALM BEACH

10,162

5,081

DAB

DAYTONA BEACH INTL

VOLUSIA

15,269

7,635

DTS

DESTIN-FORT WALTON

BEACH

OKALOOSA

11,419

5,710

EYW

KEY WEST INTL

MONROE

7,481

3,741

FLL

FORT LAUDERDALE/

HOLLYWOOOD

BROWARD

187,730

93,865

FXE

FORT LAUDERDALE

EXECUTIVE

BROWARD

10,059

5,030

GNV

GAINESVILLE RGNL

ALACHUA

17,322

8,661

JAX

JACKSONVILLE INTL

DUVAL

132,554

66,277

MCO

ORLANDO INTL

ORANGE

311,475

155,738

MIA

MIAMI INTL

MIAMI-DADE

160,937

80,469

MLB

MELBOURNE INTL

BREVARD

8,428

4,214

NPA

PENSACOLA NAS/

SHERMAN FIELD

ESCAMBIA

986

493

OPF

OPA LOCKA

MIAMI-DADE

4,756

2,378

ORL

EXECUTIVE

ORANGE

9,483

4,742

PBI

PALM BEACH INTL

PALM BEACH

115,880

57,940

PFN

PANAMA CITY-BAY CO INTL

BAY

16,173

8,087

PIE

ST PETERSBURG-

CLEARWATE

PINELLAS

13,021

6,511

PNS

PENSACOLA RGNL

ESCAMBIA

38,375

19,188

RSW

SOUTHWEST FLORIDA

INTL

LEE

66,810

33,405

SFB

ORLANDO SANFORD

SEMINOLE

5,685

2,843

SRQ

SARASOTA/BRADENTON

INTL

SARASOTA

31,752

15,876

TLH

TALLAHASSEE RGNL

LEON

31,442

15,721

TPA

TAMPA INTL

HILLSBOROUGH

171,621

85,811

VPS

EGLIN AFB

OKALOOSA

13,585

6,793

ABY

SOUTHWEST GEORGIA

RGNL

DOUGHERTY

9,266

4,633

AGS

AUGUSTA RGNL

RICHMOND

20,284

10,142

ATL

HARTSFIELD INTL

FULTON

982,852

491,426

FTY

FULTON COUNTY AIRPORT

FULTON

25,708

12,854

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

LZU

GWINNETT COUNTY -

BRISCOE FIELD

GWINNETT

10,309

5,155

MCN

MIDDLE GEORGIA RGNL

BIBB

27,074

13,537

PDK

DEKALB-PEACHTREE

DE KALB

48,484

24,242

RYY

COBB COUNTY/

COBB

8,364

4,182

SAV

SAVANNAH/HILTON HEAD

INTL

CHATHAM

48,867

24,434

HNL

HONOLULU INTL

HONOLULU

97,849

48,925

ITO

HILO INTL

HAWAII

14,216

7,108

KOA

KONA INTL AT KEAHOLE

HAWAII

20,401

10,201

LIH

LIHUE

KAUAI

17,381

8,691

OGG

KAHULUI

MAUI

52,376

26,188

CID

THE EASTERN IOWA

LINN

53,207

26,604

DSM

DES MOINES INTL

POLK

68,129

34,065

BOI

BOISE/GOWEN FIELD

ADA

171,910

85,955

IDA

IDAHO FALLS RGNL

BONNEVILLE

19,294

9,647

PIH

POCATELLO RGNL

POWER

44,705

22,353

SUN

FRIEDMAN MEMORIAL

BLAINE

23,422

11,711

BLV

SCOTT AFB/ MIDAMERICA

ST CLAIR

28,832

14,416

BMI

CENTRAL IL RGNL

MC LEAN

23,261

11,631

CMI

UNIVERSITY OF ILLINOIS

CHAMPAIGN

24,819

12,410

CPS

ST LOUIS DOWNTOWN

ST CLAIR

14,281

7,141

DPA

DUPAGE

DU PAGE

12,804

6,402

MDW

CHICAGO MIDWAY INTL

COOK

300,110

150,055

MLI

QUAD CITY INTL

ROCK ISLAND

45,378

22,689

ORD

CHICAGO O'HARE INTL

COOK

1,021,331

510,666

PIA

GREATER PEORIA RGNL

PEORIA

26,380

13,190

PWK

PALWAUKEE MUNI

COOK

25,597

12,799

RFD

GREATER ROCKFORD

WINNEBAGO

16,744

8,372

SPI

CAPITAL

SANGAMON

16,331

8,166

UGN

WAUKEGAN RGNL

LAKE

5,666

2,833

EVV

EVANSVILLE RGNL

VANDERBURGH

66,915

33,458

FWA

FORT WAYNE INTL

ALLEN

77,748

38,874

IND

INDIANAPOLIS INTL

MARION

225,106

112,553

SBN

SOUTH BEND RGNL

ST JOSEPH

61,758

30,879

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

ICT

WICHITA MID-

CONTINENTAL

SEDGWICK

46,156

23,078

IXD

NEW CENTURY

AIRCENTER

JOHNSON

3,182

1,591

SLN

SALINA MUNI

SALINE

10,497

5,249

CVG

CINCINNATI/NORTHERN

KENTUCKY INTL

BOONE

440,229

220,115

LEX

BLUE GRASS

FAYETTE

47,031

23,516

SDF

LOUISVILLE INTL-

STANDIFORD FIELD

JEFFERSON

180,463

90,232

AEX

ALEXANDRIA INTL

RAPIDES

17,013

8,507

BTR

BATON ROUGE

METROPOLITAN

EAST BATON

ROUGE

110,373

55,187

LFT

LAFAYETTE RGNL

LAFAYETTE

26,262

13,131

MLU

MONROE RGNL

OUACHITA

15,523

7,762

MSY

LOUIS ARMSTRONG NEW

ORLEANS

JEFFERSON

100,185

50,093

SHV

SHREVEPORT RGNL

CADDO

43,429

21,715

ACK

NANTUCKET MEMORIAL

NANTUCKET

153,631

76,816

BED

LAURENCE G HANSCOM

FIELD

MIDDLESEX

170,107

85,054

BOS

GENERAL EDWARD

LAWRENCE

SUFFOLK

428,546

214,273

HYA

BARNSTABLE MUNI

BARNSTABLE

120,155

60,078

MVY

MARTHAS VINEYARD

DUKES

52,133

26,067

ADW

ANDREWS AFB

PRINCE GEORGES

10,263

5,132

BWI

BALTIMORE-WASHINGTON

INTL

ANNE ARUNDEL

311,503

155,752

HGR

HAGERSTOWN RGNL

WASHINGTON

50,658

25,329

BGR

BANGOR INTL

PENOBSCOT

33,927

16,964

BHB

HANCOCK COUNTY-BAR

HARBOR

HANCOCK

42,154

21,077

PQI

NORTHERN MAINE RGNL

AROOSTOOK

7,346

3,673

PWM

PORTLAND INTL JETPORT

CUMBERLAND

78,671

39,336

RKD

KNOX COUNTY RGNL

KNOX

55,497

27,749

AZO

KALAMAZOO/BATTLE

CREEK INTL

KALAMAZOO

80,503

40,252

BIV

TULIP CITY

ALLEGAN

3,886

1,943

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

BTL

W K KELLOGG

CALHOUN

5,803

2,902

DET

DETROIT CITY

WAYNE

7,612

3,806

DTW

DETROIT METROPOLITAN

WAYNE COUNTY

WAYNE

511,008

255,504

FNT

BISHOP INTL

GENESEE

113,863

56,932

GRR

GERALD R. FORD INTL

KENT

115,354

57,677

LAN

CAPITAL CITY

CLINTON

82,792

41,396

MBS

MBS INTL

SAGINAW

19,228

9,614

MKG

MUSKEGON COUNTY

MUSKEGON

48,286

24,143

PTK

OAKLAND COUNTY INTL

OAKLAND

30,586

15,293

TVC

CHERRY CAPITAL

GRAND TRAVERSE

18,129

9,065

YIP

WILLOW RUN

WAYNE

12,050

6,025

DLH

DULUTH INTL

ST LOUIS

66,709

33,355

MSP

MINNEAPOLIS-ST PAUL

INTL

HENNEPIN

508,651

254,326

RST

ROCHESTER INTL

OLMSTED

18,910

9,455

STP

ST PAUL DOWNTOWN

HOLMAN

RAMSEY

15,841

7,921

MCI

KANSAS CITY INTL

PLATTE

231,832

115,916

MKC

CHARLES B. WHEELER

DOWN

CLAY

11,517

5,759

SGF

SPRINGFIELD-BRANSON

RGNL

GREENE

38,022

19,011

STL

LAMBERT-ST LOUIS INTL

ST LOUIS CITY

294,159

147,080

SUS

SPIRIT OF ST LOUIS

ST LOUIS

20,277

10,139

GPT

GULFPORT-BILOXI INTL

HARRISON

22,775

11,388

JAN

JACKSON INTL

RANKIN

40,968

20,484

BIL

BILLINGS LOGAN INTL

YELLOWSTONE

102,361

51,181

BTM

BERT MOONEY

SILVER BOW

19,369

9,685

BZN

GALLATIN FIELD

GALLATIN

24,875

12,438

GTF

GREAT FALLS INTL

CASCADE

26,926

13,463

HLN

HELENA RGNL

LEWIS AND CLARK

55,581

27,791

MSO

MISSOULA INTL

MISSOULA

28,702

14,351

AVL

ASHEVILLE RGNL

BUNCOMBE

38,545

19,273

CLT

CHARLOTTE/DOUGLAS

INTL

MECKLENBURG

530,350

265,175

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

FAY

FAYETTEVILLE

REGIONAL/GRANNIS FIELD

CUMBERLAND

49,500

24,750

GSO

PIEDMONT TRIAD INTL

GUILFORD

122,384

61,192

ILM

WILMINGTON INTL

NEW HANOVER

41,803

20,902

INT

SMITH REYNOLDS

FORSYTH

7,959

3,980

JQF

CONCORD RGNL

CABARRUS

17,080

8,540

RDU

RALEIGH-DURHAM INTL

WAKE

243,212

121,606

BIS

BISMARCK MUNI

BURLEIGH

14,108

7,054

FAR

HECTOR INTL

CASS

15,754

7,877

GFK

GRAND FORKS INTL

GRAND FORKS

9,393

4,697

LBF

NORTH PLATTE RGNL

LINCOLN

4,330

2,165

LNK

LINCOLN MUNI

LANCASTER

24,535

12,268

OMA

EPPLEY AIRFIELD

DOUGLAS

84,548

42,274

MHT

MANCHESTER

HILLSBOROUGH

98,436

49,218

PSM

PEASE INTL TRADEPORT

ROCKINGHAM

37,296

18,648

ACY

ATLANTIC CITY INTL

ATLANTIC

124,343

62,172

EWR

NEWARK LIBERTY INTL

ESSEX

452,350

226,175

MMU

MORRISTOWN MUNI

MORRIS

35,331

17,666

TEB

TETERBORO

BERGEN

154,674

77,337

TTN

TRENTON MERCER

MERCER

96,253

48,127

WRI

MC GUIRE AFB

BURLINGTON

1,840

920

ABQ

ALBUQUERQUE INTL

SUNPORT

BERNALILLO

197,525

98,763

SAF

SANTA FE MUNI

SANTA FE

12,480

6,240

HND

HENDERSON

CLARK

74,149

37,075

LAS

MC CARRAN INTL

CLARK

654,117

327,059

RNO

RENO/TAHOE INTL

WASHOE

155,785

77,893

TNX

TALLAHASSEE RGNL

LEON

7,810

3,905

VGT

NORTH LAS VEGAS

CLARK

233,847

116,924

ALB

ALBANY INTL

ALBANY

113,233

56,617

BGM

BINGHAMTON RGNL

BROOME

23,472

11,736

BUF

BUFFALO NIAGARA INTL

ERIE

128,363

64,182

ELM

ELMIRA/CORNING RGNL

CHEMUNG

21,645

10,823

FRG

REPUBLIC

SUFFOLK

20,909

10,455

HPN

WESTCHESTER COUNTY

WESTCHESTER

189,600

94,800

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

ISP

LONG ISLAND MAC

ARTHUR

SUFFOLK

181,621

90,811

ITH

ITHACA TOMPKINS RGNL

TOMPKINS

16,015

8,008

JFK

JOHN F KENNEDY INTL

QUEENS

369,410

184,705

JHW

CHAUTAUQUA

COUNTY/JAMES

CHAUTAUQUA

20,813

10,407

LGA

LA GUARDIA

QUEENS

415,786

207,893

ROC

GREATER ROCHESTER

INTL

MONROE

140,653

70,327

SWF

STEWART INTL

ORANGE

92,577

46,289

SYR

SYRACUSE HANCOCK INTL

ONONDAGA

117,747

58,874

BKL

BURKE LAKEFRONT

CUYAHOGA

22,694

11,347

CAK

AKRON-CANTON RGNL

SUMMIT

110,365

55,183

CGF

CUYAHOGA COUNTY

CUYAHOGA

11,129

5,565

CLE

CLEVELAND-HOPKINS INTL

CUYAHOGA

258,636

129,318

CMH

PORT COLUMBUS INTL

FRANKLIN

198,084

99,042

DAY

JAMES M COX DAYTON

INTL

MONTGOMERY

117,960

58,980

ILN

AIRBORNE AIRPARK

CLINTON

44,748

22,374

LCK

RICKENBACKER INTL

FRANKLIN

38,476

19,238

LUK

CINCINNATI MUNI AIRPORT

HAMILTON

33,963

16,982

OSU

OHIO STATE UNIVERSITY

FRANKLIN

11,217

5,609

TOL

TOLEDO EXPRESS

LUCAS

66,174

33,087

YNG

YOUNGSTOWN-WARREN

RGNL

TRUMBULL

78,202

39,101

OKC

WILL ROGERS WORLD

OKLAHOMA

96,843

48,422

PWA

WILEY POST

OKLAHOMA

8,423

4,212

TUL

TULSA INTL

TULSA

64,293

32,147

EUG

MAHLON SWEET FIELD

LANE

40,428

20,214

HIO

PORTLAND-HILLSBORO

WASHINGTON

8,575

4,288

LMT

KLAMATH FALLS

KLAMATH

48,729

24,365

MFR

ROGUE VALLEY INTL

JACKSON

61,595

30,798

PDX

PORTLAND INTL

MULTNOMAH

260,005

130,003

ABE

LEHIGH VALLEY INTL

LEHIGH

120,564

60,282

AGC

ALLEGHENY COUNTY

ALLEGHENY

24,825

12,413

AOO

ALTOONA-BLAIR COUNTY

BLAIR

27,260

13,630

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

AVP

WILKES-BARRE/

SCRANTON INTL

LUZERNE

74,034

37,017

ERI

ERIE INTL/TOM RIDGE

FIELD

ERIE

48,659

24,330

JST

JOHN MURTHA

JOHNSTOWN

CAMBRIA

53,085

26,543

LBE

ARNOLD PALMER RGNL

WESTMORELAND

42,541

21,271

MDT

HARRISBURG INTL

DAUPHIN

69,276

34,638

PHL

PHILADELPHIA INTL

PHILADELPHIA

539,901

269,951

PIT

PITTSBURGH INTL

ALLEGHENY

260,027

130,014

PNE

NORTHEAST

PHILADELPHIA

15,173

7,587

RDG

READING RGNL/

BERKS

124,509

62,255

UNV

UNIVERSITY PARK

CENTRE

64,416

32,208

PVD

THEODORE FRANCIS

GREEN

KENT

112,454

56,227

WST

WESTERLY STATE

WASHINGTON

14,704

7,352

CAE

COLUMBIA METROPOLITAN

LEXINGTON

104,926

52,463

CHS

CHARLESTON AFB/INTL

CHARLESTON

83,563

41,782

GSP

GREENVILLE-

SPARTANBURG

GREENVILLE

60,933

30,467

HXD

HILTON HEAD

BEAUFORT

14,476

7,238

MYR

MYRTLE BEACH INTL

HORRY

37,695

18,848

FSD

JOE FOSS FIELD

MINNEHAHA

31,690

15,845

RAP

RAPID CITY RGNL

PENNINGTON

14,898

7,449

BNA

NASHVILLE INTL

DAVIDSON

217,774

108,887

CHA

LOVELL FIELD

HAMILTON

83,321

41,661

MEM

MEMPHIS INTL

SHELBY

392,403

196,202

TRI

TRI-CITIES RGNL

SULLIVAN

76,282

38,141

TYS

MC GHEE TYSON

BLOUNT

130,699

65,350

ABI

ABILENE RGNL

TAYLOR

13,354

6,677

ADS

ADDISON

DALLAS

17,868

8,934

AFW

FORT WORTH ALLIANCE

TARRANT

9,975

4,988

AMA

AMARILLO INTL

POTTER

24,407

12,204

AUS

AUSTIN-BERGSTROM INTL

TRAVIS

138,050

69,025

BPT

SOUTHEAST TEXAS RGNL

JEFFERSON

63,014

31,507

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

BRO

BROWNSVILLE/SOUTH

PADRE

CAMERON

6,954

3,477

CRP

CORPUS CHRISTI INTL

NUECES

24,344

12,172

DAL

DALLAS LOVE FIELD

DALLAS

235,981

117,991

DFW

DALLAS/FORT WORTH INTL

TARRANT

736,822

368,411

EFD

ELLINGTON FIELD

HARRIS

135,087

67,544

ELP

EL PASO INTL

EL PASO

101,701

50,851

FTW

FORT WORTH MEACHAM

INTL

TARRANT

12,445

6,223

GRK

ROBERT GRAY AAF

BELL

13,856

6,928

HOU

WILLIAM P HOBBY

HARRIS

238,555

119,278

HRL

VALLEY INTL

CAMERON

21,353

10,677

IAH

GEORGE BUSH

INTERCONTINENTAL

HARRIS

589,437

294,719

LBB

LUBBOCK INTL

LUBBOCK

42,677

21,339

LBX

BRAZORIA COUNTY

BRAZORIA

62,893

31,447

LRD

LAREDO INTL

WEBB

18,417

9,209

MAF

MIDLAND INTL

MIDLAND

32,509

16,255

MFE

MC ALLEN MILLER INTL

HIDALGO

15,937

7,969

SAT

SAN ANTONIO INTL

BEXAR

211,356

105,678

SGR

SUGAR LAND RGNL

FORT BEND

6,768

3,384

SLC

SALT LAKE CITY INTL

SALT LAKE

454,715

227,358

CHO

CHARLOTTESVILLE-

ALBEMAR

ALBEMARLE

33,511

16,756

HEF

MANASSAS RGNL

PRINCE WILLIAM

14,969

7,485

ORF

NORFOLK INTL

NORFOLK

123,329

61,665

PHF

NEWPORT NEWS/

WILLIAMSBURG

NEWPORT NEWS

228,525

114,263

RIC

RICHMOND INTL

HENRICO

125,583

62,792

ROA

ROANOKE RGNL/

WOODRUM FIELD

ROANOKE

85,338

42,669

BTV

BURLINGTON INTL

CHITTENDEN

62,602

31,301

BFI

BOEING FIELD/

KING

290,752

145,376

GEG

SPOKANE INTL

SPOKANE

99,770

49,885

PSC

TRI-CITIES

FRANKLIN

34,108

17,054

SEA

SEATTLE-TACOMA INTL

KING

346,820

173,410

Airport

Code

Airport Name

County

State

Total

operations

IFR + VFR

Total LTOs

IFR+VFR

YKM

YAKIMA AIR TERMINAL/

MCALLISTER FIELD

YAKIMA

48,383

24,192

ATW

OUTAGAMIE COUNTY

RGNL

OUTAGAMIE

36,715

18,358

CWA

CENTRAL WISCONSIN

MARATHON

13,693

6,847

EAU

CHIPPEWA VALLEY RGNL

CHIPPEWA

6,173

3,087

GRB

AUSTIN STRAUBEL INTL

BROWN

35,034

17,517

LSE

LA CROSSE MUNI

LA CROSSE

16,159

8,080

MKE

GENERAL MITCHELL INTL

MILWAUKEE

215,367

107,684

MSN

DANE COUNTY RGNL

DANE

114,833

57,417

CRW

YEAGER

KANAWHA

78,583

39,292

HTS

TRI-STATE/MILTON

WAYNE

34,878

17,439

LWB

GREENBRIER VALLEY

GREENBRIER

10,984

5,492

PKB

WOOD COUNTY AIRPORT

WOOD

41,544

20,772

CPR

NATRONA COUNTY INTL

NATRONA

20,278

10,139

JAC

JACKSON HOLE

TETON

22,391

11,196

SHR

SHERIDAN COUNTY

SHERIDAN

31,360

15,680

TOTAL

34,044,499

17,022,250

Notes:

Operations = departures and arrivals.

LTOs = operations divided by 2.

Appendix C PM Methodology Discussion Paper

Prepared by: John Kinsey (EPA-NRMRL) and Roger L. Wayson (Volpe)

MISSION STATEMENT

On April 11, 12, and 13, 2007, John Kinsey (EPA ORD) and Roger Wayson (FAA Volpe) were empowered to develop

a total PM methodology from commercial aircraft engines for purposes of this study only. The developed

methodology is meant to reflect current scientific understanding of aircraft PM measurements and include reasonable

margins to accommodate uncertainties. The methodology should be developed to meet the requirements of CMAQ

modeling - thereby, providing speciated estimates of (1) black carbon and volatile PM estimates from (2) sulfate and

(3) organic emissions.

After a technically sound consensus was reached on the PM method and by close of business on April 13, they were

expected to document the PM method (and the assumptions made) to the extent needed for other EPA and FAA

people involved in this study to understand and apply the methodology to this study (this paper is the aforementioned

documentation). Unless something is clearly wrong, the EPA and FAA agreed to move forward with their

recommended PM methodology.

BACKGROUND

The estimation of particulate matter (PM) from aircraft is in its infancy with data being sparse and the test methods

are still being refined.

There is an immediate need to estimate PM for airport planning and regulatory requirements,

hence the development of the First Order Approximation (FOA). The FOA is only for estimation of PM emissions from

jet turbine aircraft in the vicinity of airports. FOA 1.0

included only the non-volatile fraction of the PM emissions and

is based on the ICAO smoke number (SN). Scaling the volatile and non-volatile components was included in FOA

2.0

to make it more complete.

However, a more in-depth procedure was needed to improve the fidelity of the approximation and better estimate the

volatile fraction, resulting in further methodology development in FOA3. This methodology utilizes the ICAO SN to

estimate the non-volatile component. The volatile component was estimated by breaking down the total volatile

emissions into the most important components: sulphur, organics, and lubrication oil. Nitrates were not considered to

be an important contributor based on available information.

This paper shows the formulation of each component for FOA 3.0 (FOA3) as developed by ICAO WG3 and then

includes the changes made for the purposes of this study, which is utilizing the CMAQ model for air quality modeling.

The modified version of FOA3 created for the purposes of this study is referred to as FOA 3.0a (FOA3a).

OVERALL FORMULATION OF FOA3

The FOA 3.0 breakdown by component led to a new general form of:

PMvols = F(Fuel Sulfur Content) + F(Fuel Organics) + F(Lubrication Oil) [1]

SAE E-31 Position Paper on Particle Matter Measurements

Wayson, R.L., G. Fleming, B. Kim, A Review of Literature on Particulate Matter Emissions from Aircraft, DTS-34-

FA22A-LR1, Federal Aviation Administration, Office of Environment and Energy, Washington, D.C. 20591,

December, 2003.

CAEP WP, A First Order Approximation (FOA) for Particulate Matter, Prepared by WG2, TG4.

PMnvols = SN v. Mass Relationship [2]

TOTAL PM = PMvols + PMnvols [3]

INDIVIDUAL COMPONENTS

Non-volatiles (soot)

The FOA 3.0 assumptions made were:

As proven by multiple researchers, SN correlates to non-volatile PM mass emissions.

Average air-to-fuel ratios (AFR) per power setting

can be assumed for all commercial turbine jet aircraft as shown in

Table C.1 using input from manufacturers.

Error in SN measurement by different researchers could be as great as ± 3 in extreme conditions. The actual

measurements of the pollutants with different analyzers also have errors. However, a review of the standard

deviations of the measurement error reported for APEX1 show that the values are far less than the SN possible error.

As such, allowing the SN to change by a value of ± 3 form upper and lower bounds to the estimate.

A difference in the trends for SN and mass occur for those SNs ≤ 30 and those > 30. Most modern engines have SNs

< 30 but older engines remain in the fleet and some method is necessary to allow prediction of these engines. As

such, there must be a correlation for SN to mass for each of the four ICAO engine certification power settings as well

as below and above a SN of 30, resulting in the use of eight equations.

The methodology is based on the available mass data at this time and is related to the smoke number (SN) so that

emissions from the majority of jet turbine engines for commercial aircraft in the fleet can be approximated by using

the ICAO emissions databank.

For the estimation of mass emissions for SNs less than 30, a correlation was used for measurement data developed

by Dr. Hurley at Qinetiq in the United Kingdom. In-situ data from testing from DLR and the University of Missouri,

Rolla were used for verification.

Table C.1: Assumed Average Air-to-Fuel Ratios by Power Setting

Power Setting

AFR

7% (idle)

106

30% (approach)

85% (climbout)

100% (takeoff)

The analysis of these data, based on mass per volume of exhaust, yielded an equation to predict the concentration

index (CI) as compared to the SN as follows:

Eyers, C., CAEP/WG3/AEMTG/WP5, Improving the First Order Approximation (FOA) for Characterizing

Particulate Matter Emissions from Aircraft Engines, Alternative Emissions Methodology Task Group (AEMTG)

Meeting, Rio De Janeiro, Brazil.

[4]

Where: CI = concentration index (mg/M

)

SN = smoke number ≤ 30

For SNs > 30 a different approach was utilized. In this case data from DLR in Germany as well as Hurley were used

in the analysis.

[5]

Where: SN = smoke number > 30

Final calculation of the non-volatile estimation of PM is based on two other derivations. The first is the calculation of

the exhaust volume based on the AFR. This term is needed as a multiplier times the concentration index to allow an

emission index directly tied to fuel usage as is customary. While details are presented in the working paper by

Eyers

, the reduced equation is:

[6]

Where: Q = core exhaust volume (M

)

AFR = modal air-to-fuel mass ratio

If the SN is measured with bypass air, the bypass ratio, β, will be used as a multiplier to estimate the exhaust volume.

This would result in the form:

[7]

From this, the non-volatile PM EI for non-volatiles may be calculated from:

non-vol

= Q (CI) [8]

Where: EI

non-vol

= emission Index (mg/kg fuel)

CI = emission concentration index (mg/M

)

It is of note that upper limits were evaluated to provide a maximum bound to the predicted non-volatile EI and not

necessarily as useable values. This was done by increasing the SN by a value of 3.

The equations that allow these conservative values are:

[9]

Where: SN = smoke number ≤ 30

[10]

Where: SN = smoke number > 30

One other problem exists. The ICAO database does not always contain complete SN information. A procedure was

used based on dividing aircraft into groups by combustor design and using the trends of each group to fill in needed

Eyers, C., CAEP/WG3/AEMTG/WP5, Improving the First Order Approximation (FOA) for Characterizing

Particulate Matter Emissions from Aircraft Engines, Alternative Emissions Methodology Task Group (AEMTG)

Meeting, Rio De Janeiro, Brazil.

SNs.

Use of this method allows modal calculations and prediction of the non-volatile EIs for the four defined

modes for most engines listed in the ICAO database. The term most is used since some reported SNs are zero which

result in extremely low EI values.

MODIFICATIONS FOR NON-VOLATILE COMPONENT

Two conservative approaches were reviewed: (1) the use of certification smoke numbers presented in the ICAO data

bank plus 3 smoke numbers to bound the upper limit that could occur in smoke number measurement (Equation 9

and 10) or (2) adding a factor for bypass flow using the best estimate approach (Equation 7). Approach 1 was

eliminated because the addition of 3 to a certification smoke number was meant to form an upper bound and not

based on real conditions. For the purposes of this study, it was agreed to multiply the flow rate by the quantity (1+

bypass ratio). This approach was used for all engines, whether they are mixed flow turbofan engines or not. However,

it is recognized that the bypass ratio multiplication factor is only appropriate for engines where the core and bypass

flow are mixed prior to the engine exit (a small fraction of the existing in service engines). For engines where the core

and bypass flow are mixed externally, use of this multiplication factor conservatively increases the value of the non-

volatile primary PM component by as much as 9.40 using the ICAO bypass ratios.

Sulfur Component

The FOA3 assumptions made were:

Sulfur emissions are primarily a function of fuel sulfur since no other major source of sulfur exits.

Most sulfur results in gaseous emissions of SO

but some is converted from fuel sulfur to sulfuric acid (H

). The

total conversion requires a certain amount of residence time in the atmosphere and the sulfuric acid is being depleted

at the same time by other atmospheric components. Sulfates would dominant PM found on an ambient air monitoring

filter and a molecular weight of 96 for SO

was assumed.

Sulfur contents of fuels change from location to location and should remain a variable during the estimation process.

Default values can be defined, however, based on published values.

Conversion efficiencies also change from location to location but can be estimated and default values can be

defined.

These assumptions resulted in the form shown by Equation 11:

[11]

Where: EI

PMvols – FSC

= EI for volatile fraction due to sulfur compounds emitted (mg/kg of fuel)

W John Calvert, W.J., Revisions to Smoke Number Data in Emissions Databank, QinetiQ, Gas

Turbine Technologies, 23 February 2006.

Coordinating Research Council, Inc., Handbook of Aviation Fuel Properties, Third Edition, CRC Report No. 635,

Alpharetta, GA., 2004.

Schumann, U., F. Arnolod, R. Busen, J. Curtius, B. Karcher, A. Kiendler, A. Petzold, H. Schroder, and K.H.

Wohlfrom (2002). Influence of fuels sulfur on the composition of aircraft exhaust plumes: The experiments SULFUR

1-7, Jour. of Geophysical Research, 107:D15, 4247.

FSC = fuel sulfur content (% by weight)

ε = S

to S

conversion rate as a fraction

out

= 96 for sulfates in exhaust

= 32 for sulfur

MODIFICATIONS FOR SULFATES:

Discussions for this study were based on three topics: fuel sulfur content, conversion efficiency, and final product.

The typical value for fuel sulfur content listed in the Handbook of Aviation Fuel Properties, which is 0.068%

mass

(680

ppmm), was selected. Conversion of gaseous sulfur species, primarily SO

, occur creating particulate matter. While

much more is involved, the gas-to-particle conversion process can be simply described by the following major

chemical reactions:

Of note is that sulfuric acid (H

) is hydroscopic and will combine readily with atmospheric moisture resulting in a

hydrated compound. Aircraft engine literature indicates that as low as one molecule of water per two of sulfuric acid

or as much as two molecules of water per molecule of sulfuric acid could occur resulting in a heavier compound.

77,78

Assuming a simple conversion efficiency for this complex set of reactions, several literature references were reviewed

and an upper limit value of 5% was selected

79,80

. After discussion with the CMAQ modelling team, it was decided that

the final product should not include hydration of H

since this is done as part of the CMAQ simulation process and

that a molecular weight of 98 should be used as a modification of the term MW

out

in Equation 11.

Fuel Organic Emissions

The FOA3 assumptions made for PM fuel organics were:

Gas phase total hydrocarbons (HC) EIs are directly related to PM fuel organic emissions. That is, if unburned HC

emissions increase, so do the overall PM organic emissions in a related fashion.

Dakhel, P.M., S.P. Lukachko, I.A. Waitz, , R.C. Miake-Lye, and R.C. Brown (2005). Post-Combustion

Evolution Of Soot Properties In An Aircraft Engine, Proc. Of GT2005, ASME Turbo Expo 2005: Power for

Land, Sea and Air, Reno-Tahoe, NV., June 6-9.

Arnold, F., T.H. Stilp, R. Busen, and U. Schumann (1998). Jet engine exhaust chemiion measurements

implications for gaseous SO

and H

, Atmospheric Environment, 32:18, 3073-3077.

Sorokin, A., E. Katragkou, F. Arnold, R. Busen, and U. Schumann (2004). Gaseous SO

and H

in the exhaust

of an aircraft gas turbine engine: measurements by CIMS and implications for fuel sulphur conversion to sulfur (VI)

and conversion of SO

to H

, Atmospheric Environment, 38, 449-456.

Schumann, U., F. Arnold, R. Busen, J. Curtius, B. Karcher, A. Kiendler, A. Petzold, H. Schlager, F. Schroder, and

K.H. Wohlfrom (2002). Influence of fuel sulfur on the composition of aircraft exhaust plumes: The experiments of

SULFUR 1-7, Jour. of Geophysical Research, 107:D15, 4247.

Fuel PM organic emissions can be formed as a coating on non-volatile PM or due to condensation from the gas

phase. This process is not well understood at this time and although these emissions are included, there is no

separate calculation process.

Measurement data separating the organic fraction from the overall PM emissions from in-service engines are very

limited. Information from APEX1 would seem to be the most reliable at this time. However, only one engine (CFM56-

2-C1) is included and it is assumed that the trends shown in Figure D.1 are consistent for all commercial jet turbine

engines in the ICAO database. As such, ICAO certification EIs for hydrocarbons can be related to the PM fuel organic

emissions.

The data used is for a probe 30 meters behind the aircraft. It is assumed that in this distance volatile organic PM

emissions are representative of those in the atmospheric in the vicinity of airports since other data is not available.

The overall estimation problem is multi-faceted & many details are not well known. As such, the organics

methodology for PM fuel organics must be simplistic at this time.

Figure C.1: Trends from APEX 1 for CFM56-2-C1 engine

The resulting “non S component” was derived by subtracting the “sulfates” from the “volatile contribution” except for

the power settings of 85 and 100%. At these power settings, the values dropped below that shown as “organics”

measured by a different instrument. In an attempt to not under-predict, the values of the “organics” curve shown in

Figure D.1 for 85 and 100% power settings were used directly. This resulted in Equation 12 with all modes defined for

the “non S component.”

[12]

Where: PMvol

fuel organic

= volatile PM emissions of organics (mg/kg fuel)

Non_S_Component = a constant ratio based on the trends shown in Figure D.1.

HC(CFM56)

= ICAO emission index for hydrocarbons for the CFM56 engine

HC(Engine)

= specific ICAO emission index for hydrocarbons for the engine of

concern

MODIFICATIONS ON FUEL ORGANICS:

The CFM56 scaling method was reviewed and it was decided that a true mass balance represented a more

consistent approach across the entire power spectrum. It was also agreed that a margin of conservatism should be

added to the resulting values from the mass balance approach. This required modifications in two steps.

Step 1: The measured volatile component derived from APEX1 data was used and adjusted for the sulfur component

(shown as “sulfates” in Figure D.1). In this approach, a single set of measurements was used to avoid conflicting data

from different measurement techniques. This resulted in the curve shown as the “non S component” no longer being

adjusted for the 85 and 100% power setting as was done in the FOA 3.0 approach described previously. Instead, the

resulting curve used is simply the curve listed as the “volatile contribution” in Figure D.1 is subtracted off the values of

the “sulfates” at each engine power setting so that sulfur is not counted twice. Also, to be conservative, it is assumed

that 100% of the resulting “volatile component” curve are semi-volatile and in the particle phase.

Step 2: To ensure an even more conservative method, the APEX1 data set was further analyzed to determine total

volatile PM. Again using the APEX1 data for the base fuel condition, the ratio of sulfur to organics was determined

from reported measurements and this ratio used to subtract out the sulfate contribution from the total volatile PM.

This resulted in a volatile PM component that did not include sulfur. These results are reported in Table C.2.

Table C.2: Derived “Non_S_Component values by mode [mg/kg fuel]

Mode

Volatile Contribution

Sulfates

Derived Non_S_Component

Idle

13.2

1.9

11.3

Approach

5.7

1.2

4.5

Climbout

4.2

1.3

2.9

Takeoff

2.9

1.7

1.2

The standard deviation of the individual data points for this derived volatile component, without sulfates, was then

computed (see Table D.3) and added to the new derived “non S component”. This new, more conservative, ”non-S

component” was used in Equation 12 to calculate the EI for PM organics.

This is shown in equation form as:

(Total PM – Non-volatile PM)(1-(sulfate/organics)) = PM

non-S organics

Standard deviation(PM

non-S vol

) + non S component (Figure D.1) =

Modified non S component (to be used in Equation 12)

Table C.3: Computed standard deviations for the volatile PM component

Mode

Std. Dev.

[mg/ kg fuel]

Idle

Approach

Climbout

Takeoff

Lubrication Oil

Emissions of lubrication oil are not well documented in the literature. As such, an approximation method for this

component was not included in the FOA 3.0.

DECISION ON LUBRICATION OIL:

Data was extremely scarce and multiple engineering judgments had to be made based on data supplied by an engine

manufacturer. Lubrication oil use increases with engine wear until a critical value of about 0.3 quarts per hour occurs.

At this time, the engine is removed from service for substantial reworking and maintenance. Based on an assumption

that about 0.1 of the value used for overhaul standards represents nominal operating consumption, it was determined

that approximately 0.03 quarts per hour of lubrication oil are lost. Since venting is the primary release and tends to

occur at the higher power settings, a ratio of the time in takeoff (0.7 minutes) and climb-out (2.2 minutes) modes were

used and it was found that 0.00145 quarts could be emitted during these operations in the vicinity of airports. Using a

specific gravity of 1.0035 reported for Mobil Jet Oil II (density = 1,003.5 kg/m

or 949.7 grams/quart)

, it was found

that approximately 1.4 grams of lubrication oil volatile organic PM could be released per landing and takeoff

operation (LTO). This value is added to the volatile PM contribution from fuel organics to determine the total organic

volatile component for input into the CMAQ model. Sulfur volatile emissions are handled separately in this method

and this is also required by the CMAQ model.

The estimation of the lubrication oil emissions in equation form is:

Nominal consumption = 0.3 quarts/hr * 0.1 = 0.03 quarts/hr

Emissions per LTO = 0.03 quarts/hr * 1 hour/60 min * 2.9 min/LTO = 0.00145 quarts/LTO

Emissions (grams/LTO) = 0.00145 quarts/LTO * 949.7 grams/quart

≈ 1.4 grams of volatile PM from lubrication oil per LTO

RESULTING EQUATIONS FOR CMAQ IMPLEMENTATION

The inclusion of the modifications results in a different set of application equations. The terms of these equations are

as previously defined unless noted. The equations for the method used in this study are:

Overall Equations:

PMvols = F(Fuel Sulfur Content) + F(Fuel Organics) + F(Lubrication Oil Organics) [1a]

1,003.5 kg/m3 * 1 m3/1,056.7 quarts * 1,000 grams/1 kg = 949.7 grams/quart

PMnvols = SN v. Mass Relationship = Q (CI) [2a]

TOTAL PM = PMvols + PMnvols [3a]

Detailed Equations:

(for SN ≤ 30) [4a]

(for SN > 30) [5a]

Equation 6 is no longer used in the method employed in this study.

[7a]

non-vol

= Q (CI) [8a]

Equations 9 and 10 are no longer used in the method employed in this study.

[11a]

Where: FSC = 0.00068 (typical mass fraction)

ε = 0.05 (conservative fractional conversion)

out

= 98

[12a]

Where: “Non_S_Component” is now the revised term and is the derived modal “Non_S_Component” (Table C.2)

with the modal standard deviation added (Table D.3).

lube oil

= 1.4 grams/LTO [13a]

lube oil

= Lubrication oil emission index per LTO cycle [g/engine-LTO]

To predict the total PM the procedure is:

Total PM EI w/o lubrication oil = (Equation 4a or 5a * Equation 7a) + Equation 11a + Equation 12a

The resulting EIs must then be multiplied by time in mode, fuel use by mode, and number of engines. Lubrication oil

emissions are then added to each aircraft LTO cycle per engine (number of engines * number of LTOs * 1.4) and

accounts for emissions separately using Equation 13a.

Lubrication oil may also be used as a typical EI with units of mg/kg fuel and applied in the climbout and takeoff

modes. While the mass over an LTO will stay constant at 1.4 grams per LTO for all aircraft engine types, the value of

the EI will vary dependent upon fuel use for a particular engine. This is necessary because of the units for EIs, mass

per kilogram of fuel used. To apply lubrication oil volatile PM emissions in this way, the following is required.

 Determine the fuel use rate in kg/s from the ICAO data bank for the engine of concern.

 Multiply the modal fuel usage rate by the time in mode (132 seconds for climbout and 42 seconds for

takeoff). This is the total fuel used in the vicinity of the airport during these two modes for the selected

engine.

 Divide the volatile PM from lubrication oil in each of the two modes by the total fuel use in each mode. The

volatile PM for each mode is 1060 mg during the climbout mode and 340 mg during the takeoff mode. This

final number has the units of mg/kg fuel as required.

An example of this application is included in the implementation section of this paper.

IMPLEMENTATION

The sum of the calculation for the volatile PM (sulfates, lubrication oil and organics) and the non-volatiles (soot) then

provides an overall total EI for the PM emitted from jet turbine aircraft. The largest uncertainties are associated with

the prediction of the volatile PM emissions; sulfur, fuel organics and lubrication oil emissions. Sulfur is better

understood than the other two. These uncertainties can only be resolved by carefully planned measurements and

further analysis. In sum, it is the opinion of the authors, that the FOA3.0a sufficiently serves the purpose of predicting

the LTO emissions for use in CMAQ for this study.

The derived EI values for this study were compared to those of FOA3.0 for four engines often used in the fleet. The

results are shown in Figure C.2 through Figure C.4. It should be noted that lubrication oil PM EIs were developed

using the method described in the last section. The details of the EI derivation for lubrication oil follows.

At the present time lubrication oil is estimated as 1.4 grams / 2.9 minutes which is the time the engines are in the

higher power settings in the vicinity of an airport (climbout and takeoff modes). Following the procedure in the last

section of this paper the following steps were performed.

Step 1: The mass was divided into two fractions for lubrication oil.

Climbout mode = (2.2 min / 2.9 min) * 1.4 grams = 1.062 or ≈ 1060 mg

Takeoff mode = (0.7 min / 2.9 min) * 1.4 grams = 0.338 or ≈ 340 mg

The fuel usage rates were determined from the ICAO Emissions Databank for each mode. These are shown in Table

C.4.

Table C.4: ICAO fuel use rates for three engines evaluated. [kg/s]

Mode

CFM56-3

RB211-535E4-B

PW4158

Climbout

0.878

1.65

2.004

Takeoff

1.056

2.08

2.481

Step 2: The fuel consumed for the time in mode were computed and are shown in Table D.5.

Table C.5: Total fuel use for climbout and takeoff modes [kg fuel]

Mode

CFM56-3

RB211-535E4-B

PW4158

Climbout

115.9

217.8

264.5

Takeoff

44.4

87.4

104.0

Step 3: The PM volatile mass from lubrication oil emissions for each of the two modes was divided by the fuel

consumed in each mode and the final results are shown in Table C.6.

Table C.6: Lubrication oil EIs for climbout and takeoff for selected engines. [mg/kg fuel]

Mode

CFM56-3

RB211-535E4-B

PW4158

Climbout

Takeoff

These values were included in the overall EIs which are shown in Figure C.2 through Figure D.5.

Figure C.2: Comparison of FOA3.0a to FOA 3.0 for the PW4158 engine

Figure C.3: Comparison of FOA3.0a method to FOA 3.0 for the CFM56-3B-2 engine.

Figure C.4: Comparison of FOA3.0a method to FOA 3.0 for the RB211-535E4 engine.

Figure C.5: Comparison of FOA3.0a method to FOA 3.0 for the GE90-77B engine.

RECOMMENDATIONS

1. For the purposes of this study only, the FOA3a method should be adopted as the current technique to

estimate PM emissions from jet turbine aircraft in the vicinity of airports for CMAQ modeling.

2. Separate from this study, efforts should continue to improve the FOA until it can be replaced by

measurement data.

Appendix D Data Collection and Analysis of Aircraft Auxiliary Power Unit Usage

Prepared by Metron Aviation, Inc.

Background

As discussed in the body of the report, a part of the overall study approach required the collection of usage data for

auxiliary power units (APUs). An APU is a relatively small self-contained generator used in aircraft to start the main

engines, usually with compressed air. In addition, they provide electrical power and compressed air to operate the

aircraft’s instruments, lights, ventilation, and other equipment (typically while the aircraft is parked at the gate). In

many aircraft, the APU can also provide electrical power for the aircraft while in the air. In most cases, the APU is

APUs are routinely used throughout the time an aircraft is on the ground. APU usage is determined by individual

airlines and varies with aircraft type and several other factors. For arrivals, some airlines will start the APU when the

aircraft is on approach. It will stay on during the entire taxi-in phase to ensure its availability if the engines need to be

restarted. Other airlines may operate the APUs during taxi-in if they are using reduced power or a single engine.

During the departure phase of a flight, the APU is used to start the main engine. Some airlines will keep the APU

operating during taxi-out as a backup. In addition, when an aircraft is expected to temporarily park away from the

gate, the APU will be used during the taxi-out phase of flight.

Factors Affecting APU Usage

APU use varies with aircraft type, airline, and airport. Aircraft size has an influence on the time it takes to service and

load the aircraft, and thus influences the time that the APU is utilized. For a given aircraft type, the specific APU used

will vary between airlines depending on the equipment onboard the aircraft.. For a particular airline, the APU unit may

be used differently at two different airports. Factors such as availability of ground-based power units and airport

environment, both climatologically and procedurally, affect the usage of APUs.

The availability of a ground-based power unit affects APU usage in several ways. If a pilot knows a ground-based unit

exists at the gate, the APU may remain off during taxi-in time with the understanding that the ground-based unit will

power the aircraft at the gate. Even when a ground-based unit is available at the gate, the airline may decide to start

the APU during flight preparations.

With regard to airport location, a flight at an airport that is located in a warmer or colder climate will often need to use

the APU longer than one operating at an airport in a more temperate location. In addition, APU usage generally

increases during the summer and winter months due to increased need for cooling or heating.

There are at least four operational phases to consider when discussing APU use:

 Departure Preparations: If ground-based support is available, APUs may be turned on just prior to pushing

back from the gate, or, if no ground support is available, the APUs may be started to help prepare the cabin

for passengers or cargo.

 Departure Taxi: Once the aircraft leaves the gate the carrier may have a standard operating procedure to

taxi on fewer than all of the engines. If the engines are not producing the needed power to maintain the

cabin environment, the APU may be used as a supplement.

 Arrival Taxi: When the aircraft lands and taxis to the gate the APU again may be used to supplement power

depending on the use of the aircraft’s engines.

 Arrival at the Gate: If power and conditioned air are available at the airport’s gate, the APU might remain on

until the aircraft is properly connected to the ground source. If no ground support is available, the APU may

be shut off or remain operating, depending on when the aircraft will be used next or for maintenance

purposes.

APUs also have varying power settings, and therefore differences in resulting emissions per unit of operating time.

Method and Results

When computing emissions associated with flight operations, the FAA Emission and Dispersion Modeling System

(EDMS) incorporates estimated APU usage times as part of the calculation. If the user cannot provide more detailed

information, EDMS Version 4.11 provides a default APU operation time of 26 minutes per aircraft landing/take-off

cycle (LTO), independent of any other factors. In EDMS Version 5.0, certain improvements have been made. In

EDMS Version 5.0, APU times are now allocated to arrivals and departures separately to allow for analysis without

looking at the entire LTO.

As an initial step toward providing additional information from which to estimate APU usage for this study, APU usage

data was collected in a limited, informal fashion from several airlines. We discussed patterns of usage, dependencies

on the factors discussed above, and the availability of carrier statistics. In addition to background information from

several airlines, quantitative data was provided by three airlines. This quantitative data can be characterized as

follows:

 Airline A – Partial data for four wide-body types and one narrow body type, covering 4-6 months of

operation, but no information on numbers of aircraft or airports sampled. The range of usage for wide-body

aircraft during the period was from 1 to 2.3 hours/flight, and 0.9 to 1.4 hours/flight for narrow-body aircraft.

Some variation in seasonal use was apparent, with the monthly averages for all aircraft sampled ranging

from about 1.1 to 2.0 hours/flight between the lowest-use month and the highest.

 Airline B – One year of data for airframes of a single narrow-body type, with the number of airframes

sampled each month ranging from 54 to 78. The number of airports serviced was not captured, but was

probably substantial. Some variation in seasonal use was apparent, with the monthly averages for all aircraft

sampled ranging from about 0.9 to 1.1 hours/flight between the lowest-use month and the highest. Wide

variation in usage between airframes was observed, with the yearly average ranging from about 0.3 to 3.4

hours/flight.

 Airline C – Average usage times per flight for one narrow-body aircraft and one wide-body aircraft. The

amount of data used to develop these averages was not specified.

As contact was made with various carriers it became clear that collection and analysis of APU usage was at different

levels of detail and maturity for each airline. Data has not been captured in a consistent fashion and is dependent on

ease of availability and on the carrier’s internal needs. Furthermore, although many carriers have standard operating

procedures for when and how to use APUs, the ultimate decision rests with the pilot.

Collection of such data is challenging for two reasons. Some airlines believe the data to be proprietary and are

reluctant to distribute it. In addition, APU usage data is evidently not trivial to record, and is consequently not

recorded by airlines on a regular and systematic basis. Due to these challenges, APU times collected in this initial

effort do not distinguish between APU usage during taxi and APU usage at the gate.

However, it should be noted that several airlines contacted were currently performing APU studies themselves to

100

determine how to reduce APU time. As the price of jet fuel continues to rise, it is expected that more airlines will study

APU usage and aim to improve efficiency. More systematic data may then become available.

Once the available data was assembled, the aircraft types represented were aggregated into two classes: wide-body

and narrow-body jet. In addition, the wide range of the available data was represented by three values of APU usage

per LTO cycle in each class: low, moderate, and high. The usage estimates derived from the available data are as

follows:

Table D.1: APU use per LTO cycle (minutes)

Narrow Body

Wide Body

Low

Moderate

High

Low

Moderate

High

130

163

The available data is not sufficiently specific to draw strong conclusions, but an interim approach might be to use the

lower values to represent situations in which aircraft have access to ground support, while the upper values could

represent situations where ground support is not available. Additional judicious use of these values might represent

differences in seasonal use and airport climatic conditions.

Next Steps

To better estimate the usage of APUs at airports, more data and supporting analysis is needed. With the assistance

of appropriate trade organizations, additional carriers should be contacted to increase the sample size, as well as the

level of detail. It would appear that some carriers are modifying operating practices in this area, and an understanding

of trends in these changes should be developed. In addition, airport-oriented data collection could be undertaken to

determine the availability of ground-based units, the average time planes are parked somewhere other than at the

gate, meteorological conditions through the year, etc. From these types of data, more accurate estimates of APU

usage under different conditions, as well as sensitivities to other factors, could be derived.

Effects of Auxiliary Power Units

The baseline inventory described in Section 3.1 provided the basis for the NEI comparison, the air quality modeling,

and health impact analysis. This inventory was created assuming a medium level of APU usage. An assessment of

the impacts of APUs on LTO emissions was performed, requiring two additional inventories with different APU

assumptions. In addition, for evaluation purposes in regard to the 148 airports in non-attainment areas, a total of

three emissions inventories were created using the high, medium, and low APU times. These APU inventories were

then compared to total aircraft LTO emissions.

Under the low APU usage scenario, the greatest percentage that APUs contributed to total aircraft emissions at an

airport was under 10% for CO and between 15 and 20% for NO

and SO

. For the high APU usage scenario, the

percentages increased to over 15% for CO and over 30% for NO

and SO

. However, investigating the airports where

APU emissions were a high percentage of total LTO emissions revealed that these airports served a higher

percentage of business jet operations. For certain small business jets with small taxi times, an hour of APU time (the

upper value) can produce enough SO

emissions to account for more than 30% of LTO emissions. This analysis

likely overstates the contribution of APU emissions since it may not be realistic to assume that a business jet will

spend an hour with the APU operating during an LTO when there is limited loading and unloading of passengers.

101

Additionally, an inventory of all 325 airports with VFR and IFR traffic was created using the medium level of APU

usage; the range of contribution of the medium level of APU usage to aircraft emissions below 3,000 feet is between

0% and slightly over 25%, as shown in Figure E.1.

The average is below 5% for CO and VOCs and under 10% for

and SO

. For only four non-attainment areas considered in this report, the medium level of APU usage

contributes over 1% to census area emissions (or total emissions) as estimated in the 2002 National Emissions

Inventory.

Figure D.1: Range of the percentage of aircraft emissions due to APU at 325 airports studied

In airports with a high volume of operations, the effect of APUs is overshadowed by the emissions from the main

engines. However, in areas with fewer operations with less delay, APU emissions play a greater role.

Using data that were generated for Section 4.2, the effects of APU usage were evaluated in a no ground delay

scenario. If the aircraft experienced no delay and the APU usage remained the same (currently there is no extra APU

usage assumed for periods of delay), then at medium levels of usage APUs would result in more than 15% of the

aircraft emissions for CO, greater than 25% for NO

and greater than 30% for SO

. As the system is driven to less

ground delay, APUs may play a greater role in aircraft emissions below the mixing height.

It is possible for airports to have aircraft that do not have APUs.

102

Appendix E Emissions and Dispersion Modeling System (EDMS) Baseline

Aircraft Emissions Inventory

A baseline emissions inventory for all aircraft arriving to and departing from the 325 study airports was generated

using aircraft operations data from the most current FAA Enhanced Traffic Management System (ETMS)

data for

the period between June 2005 and May 2006, providing one year of operations for each airport. The operations data

was used as input to the FAA Emissions and Dispersion Modeling System (EDMS

), version 5.02 An older version

of EDMS was used to generate aircraft emissions inventories for the 2001 EPA National Emissions Inventory; PM

emissions factors for this version of EDMS were based on data for several engines in AP 42, which is an EPA

compilation of air pollutant emissions factors,

In contrast, version 5.02 of EDMS contains the FOA3a method for

estimating PM emissions from aviation (described in Appendix C), and actual aircraft operational data was used as

an input to EDMS version 5.02 to generate aviation emissions estimates for this study. Rather than assuming a

particular national mix of engines and airframes, data on specific engine-airframe combinations were used.

Additionally, modeled operations were based solely on the data available and were not averaged across months to

give annual estimates of emissions. Thus, the aviation emissions data generated by EDMS 5.02 was of a higher

fidelity than the aviation emissions data in the 2001 NEI. For this reason, the EDMS emissions inventory was used for

this study.

General information on Instrument Flight Rules (IFR) flights was gathered from ETMS.

ETMS provides the flight

number, the origin and destination airport for the flight, and a generic aircraft type. The generic aircraft type is not

suitable for modeling emissions; specific airframe and engine combinations are required. The Bureau of

Transportation Statistics (BTS) On-Time Performance Database

was used to match flight numbers to aircraft

registration numbers (tail number), in order to match each flight to a specific aircraft type. Over 12.5 million operations

were generated by combining these two sources.

Registration information for the aircraft was obtained from the commercially-available BACK fleet database

or the

FAA’s aircraft registration database.

These databases were used to determine the engine models installed on

individual aircraft based on the tail number. The BTS data also provides aircraft pushback, wheels up, touchdown,

and gate arrival times. This allowed outbound and inbound taxi times to be calculated for input into EDMS. Since not

all flights appear in the BTS data, flights not reported in BTS were assumed to have taxi times equal to the average of

the reporting flights at the airport performing a similar operation during the same hour.

The data gathered through ETMS and BTS provided only a portion of the operational profile (IFR traffic). Visual Flight

Rules (VFR) traffic operations were estimated by subtracting IFR operations from the total operations for the airport

as listed in the Air Traffic Activity Data System (ATADS). The fleet mix of VFR aircraft was estimated from typical

aircraft categories based at each airport.

Aircraft operations were aggregated by airframe, engine and takeoff weight to ease the computational requirements

http://www.fly.faa.gov/Products/Information/ETMS/etms.html

http://www.faa.gov/about/office_org/headquarters_offices/aep/models/edms_model/

http://www.epa.gov/ttn/chief/ap42/

IFR traffic refers to aircraft that operate using an internal mechanism to show visually or aurally the attitude, altitude

or operation of the aircraft. These flights include electronic devices for automatically controlling the aircraft in flight.

The majority of commercial flights operate under IFR. VFR traffic refers to flights in which the pilot has responsibility

for maintaining separation distances visually. VFR flights are mainly performed by general aviation traffic operating

small aircraft.

http://www.transtats.bts.gov/OT_Delay/OT_DelayCause1.asp

http://www.backaviation.com/Information_Services/

Federal Aviation Administration Registry Database, Fall 2006, available from http://registry.faa.gov/.

103

of EDMS. The taxi in and out times were averaged across those operations at an airport level by engine and airframe

type. These averages were computed by month. If sufficient engine and airframe data did not exist, default averages

for the airport were used; if airport defaults did not exist, ICAO default taxi times were used. To compute an upper

bound on aircraft emissions during taxi, all operations were assumed to taxi in and out using all engines for the entire

estimated taxi time.

The FAA registration database and the National Airspace System Resources (NASR)

were used as additional data

sources to help determine VFR operations at airports in nonattainment areas, and the operational profile was fed into

EDMS. Inventories were generated for CO, hydrocarbons, NO

, and SO

for all phases of taxi and flight based on

International Civil Aviation Organization (ICAO) engine emissions indices—estimates of the mass of pollutant

produced per mass of fuel consumed as contained in the International Civil Aviation Organization (ICAO) Engine

Emissions Certification Databank.

To estimate total emissions of particulate matter (PM), a criteria pollutant

composed of a complex mixture of solid particles and liquid droplets, EDMS must rely on research-based estimation

techniques integrated into EDMS (see Appendix C). Emissions were then aggregated by month and mode for use in

the air quality analysis.

Figure E.1: Overview of the generation of the baseline inventory

Inventory Limitations and Sources of Discrepancies

Several generalizations, estimations and approximations were made in creating the baseline inventory that served as

Carriers frequently use single engine taxi going to and from terminal gates. Additionally, pilots often shut off main

engines and switch to APUs during long delays. The circumstances of single engine taxi use and APU use during

extended delays could not be adequately defined for consistent, realistic modeling across the variety of carriers,

airports and weather conditions.

Federal Aviation Administration, National Airspace System Resources (NASR) data, 2006.

http://www.caa.co.uk/default.aspx?catid=702&pagetype=90

104

the basis for the air quality modeling and health impact analysis. These discrepancies were mitigated when possible,

but some remain as discussed in this section.

Taxi Times

When available, exact taxi times from BTS data were used. If taxi times were not listed, the average taxi time for the

departure/arrival hour at the origin/destination was used. If there were no BTS flights during that hour, the average for

the year was used. If annual BTS information was not available for the airport, the ICAO standard time of 19 minutes

for taxi-out and 7 minutes for taxi-in was assumed.

Additionally, taxi times were assumed to consist of full engine taxi regardless of the type of aircraft or the length of the

taxi time. Anecdotally, it is known that aircraft often taxi-out on one engine and use APUs instead of main engines

during long delays, but we chose to create a conservative estimate due to the uncertainty associated with the exact

timing of how and when the aircraft may switch to APU or a single engine taxi.

APUs

The APU survey provided information about the range of APU use (see Appendix D). However, the survey was

centered on commercial carriers, not business jets. While, commercial aircraft have longer boarding and

disembarkment times than business jets, the APU assumptions were applied uniformly to both types of aircraft.

For departing flights, anticipated delays may prompt pilots to shut off main engines and run APUs to conserve fuel.

Although airlines have individual operating procedures, the ultimate decision rests with the pilot, making modeling

very difficult. The estimates of APU usage do not account for the fact that pilots may turn off main engines and use

the APU during periods of long delay.

Default Engines

Engines were matched to air frames based on tail number. However, for some flights, there was no BTS information

to provide tail numbers. In addition, some tail numbers did not match specific information in Campbell-Hill, BACK or

FAA registration databases. For these aircraft, the EDMS default engine, the most commonly occurring engine for

that air frame in the US was used.

International Flights

International flights are not listed in the BTS data set. This limits the specific information available about these flights

and requires a greater number of default values for inputs. Default values are particularly problematic as international

flights tend to operate heavy aircraft with higher fuel burn. ETMS was used to obtain information on international

flights. Because ETMS does not contain taxi data, international flights were assigned airport-level default taxi times

when possible. For airports that did not have default taxi times, the ICAO default taxi/idle time was used. Accurately

portraying these flights with the correct engines and taxi times is required to more correctly estimate total emissions

at international airports.

Particulate Matter Emissions Inventory

The measurement methodology for PM for jet turbine aircraft is still being developed and data are sparse.

Measurement and modeling of aircraft PM emissions is still an emerging area and there are data limitations and

uncertainties.

93,94

A small data set (APEX-1

) not used for development of the PM model was used as a

The determination of fine particulate matter emissions from aircraft engines is an active area of research. Methods

to estimate primary PM emissions from aircraft are relatively immature: test data are sparse, and test methods are

105

comparison to estimate non-volatile confidence limits. Additionally, limits on measurement errors of the independent

variable for non-volatile estimation (based on the reported smoke number) were evaluated as well to determine upper

and lower bounds of the estimation technique. For the non-volatiles, no direct comparison to measured data was

possible due to a lack of data.

The PM emissions inventory contains two known errors: The primary PM inventories for 78 of the 325 study airports

were generated using a fuel sulfur emissions index of 0.8 g/kg-fuel burned (corresponding to a fuel sulfur

concentration of 400 ppm), versus a value of 1.36 g/kg-fuel burned (corresponding to a fuel sulfur concentration of

680 ppm which is more representative of the current jet fuel supply). The higher fuel sulfur emissions index was used

to generate the results in Sections 4 and 5; however, time and resources were not available to repeat the air quality

and health effects modeling. The error in the sulfur specification impacted both the volatile component of the primary

PM emissions, and the secondary PM precursor emissions. By analyzing the changes in the inventories we estimate

that this led to an underestimation of the health effects of approximately 10%. However, this underestimation is

approximately offset by the conservatively-biased assumptions in the primary PM inventory estimation method

(FOA3a) such that the net effect is that the health effects shown in the body of the report are not biased high or low.

The second problem that occurred was an incorrect factor used for the fuel organics portion of the volatile PM

component. (PM emissions include volatile and non-volatile components – see Appendix C.) This was extensively

evaluated and found to cause an approximate 3% error. This error is less than the expected uncertainties of the

model and calculations show that no changes in the conclusions would occur.

still under development. ICAO and EPA do not have approved test methods or certification standards for aircraft PM

emissions. ICAO’s Committee on Aviation Environmental Protection has developed and approved the use of an

interim First Order Approximation (FOA3) method to estimate total PM emissions (or total fine PM emissions) from