Illustrative Air Quality Analysis for
the Light and Medium Duty Vehicle
Multipollutant Proposed Rule
Technical Support Document (TSD)
&EPA
United States
Environmental Protection
Agency
-------
Illustrative Air Quality Analysis for
the Light and Medium Duty Vehicle
Multipollutant Proposed Rule
Technical Support Document (TSD)
This technical report does not necessarily represent final EPA decisions
or positions. It is intended to present technical analysis of issues using
data that are currently available. The purpose in the release of such
reports is to facilitate the exchange of technical information and to
inform the public of technical developments.
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
4>EPA
United States
Environmental Protection
Agency
EPA-420-D-23-002
April 2023
-------
Table of Contents
1 Introduction/Overview 1
2 Emissions Inventory Methodology 2
2.1 Emissions Inventory Sector Summary 2
2.2 The Emissions Modeling Process 4
2.3 Emissions Inventory Methodology for 2016v2-Compatible Sectors 6
2.4 2055 Emissions Inventory Methodology for the Nonroad Sector 6
2.5 2055 Emissions Inventory Methodology for Fugitive Dust 6
3 Onroad Emissions Inventory Methodology 6
3.1 Emissions Factor Table Development 7
3.2 Activity and Other Data Development 8
3.2.1 2016 Base Year Activity data 9
3.2.2 2055 Projected Activity Data 13
3.3 Onroad Emissions Modeling 23
3.3.1 Spatial Surrogates 26
3.3.2 Temporal Profiles 26
3.3.3 Chemical Speciation 27
3.3.4 Other Ancillary Files 27
4 EGU Emissions Inventory Methodology 28
4.1 Integrated Planning Model (IPM) 28
4.2 IPM Inputs 28
4.2.1 IPM Energy Demand Inputs 30
4.3 Air Quality Model-Ready EGU inventory generation 34
5 Petroleum Sector Emissions Inventory Methodology 35
5.1 Refinery Emissions 35
5.1.1 Initial Projection of Refinery Emissions to 2050/2055 35
5.1.2 Apportioning Total Refinery Emissions to Gasoline and Diesel Fuel Production... 36
5.1.3 Identifying Refinery Emissions to Adjust for Illustrative Air Quality Analysis 37
5.1.4 Illustrative Air Quality Modeling Scenarios and Associated Refined Fuel Demand37
5.1.5 Projected Change in U.S. Refinery Activity Related to Decreased Domestic Demand
38
5.1.6 Generation of Adjustment Factors 39
5.2 Crude production well and pipeline emissions 40
-------
5.2.1 Initial Projection of Crude Production Well Site and Pipeline Inventories to
2050/2055 40
5.2.2 Illustrative Air Quality Modeling Scenarios and Associated Crude Demand 40
5.2.3 Generation of Crude Production Well and Pipeline Adjustment Factors 41
5.3 Natural gas production well and pipeline emissions 41
5.3.1 Initial Projection of Natural Gas Production Well Site and Pipeline Inventories to
2050/2055 41
5.3.2 Illustrative Air Quality Modeling Scenarios and Associated Natural Gas Demand. 41
5.3.3 Generation of Natural Gas Production Well and Pipeline Adjustment Factors 42
6 Inventory Summary Tables 43
7 Air Quality Modeling Methodology 46
7.1 Air Quality Model - CMAQ 46
7.2 CMAQ Domain and Configuration 46
7.3 CMAQ Inputs 48
7.4 CMAQ Model Performance Evaluation 49
7.4.1 Monitoring Networks 51
7.4.2 Model Performance Statistics 52
7.4.3 Evaluation for 8-hour Daily Maximum Ozone 54
7.4.4 Seasonal Evaluation of PM2.5 Component Species 60
7.4.5 Seasonal Hazardous Air Pollutants Performance 126
7.4.6 Seasonal Nitrate and Sulfate Deposition Performance 127
7.5 Model Simulation Scenarios 130
8 Additional Results of Illustrative Air Quality Analysis 131
8.1 Annual 2055 Reference, LMDV Regulatory, and Onroad-Only Scenario Maps 132
8.2 Seasonal Air Toxics Maps 149
8.3 Projected Visibility in Mandatory Class I Federal Areas 169
-------
1 Introduction/Overview
The Environmental Protection Agency (EPA) is proposing a rule to build on and improve the
existing emission control program for on-highway light and medium-duty engines and vehicles
by further reducing air pollution from light and medium-duty engines across the United States.
This proposed rulemaking is formally titled "Multi-Pollutant Emissions Standards for Model
Years 2027 and Later Light-Duty and Medium-Duty Vehicles," and is more generally referred to
as the "Light Medium Duty Vehicle"(LMDV) proposed rule. The proposed rule would impact
emissions of criteria and toxic air pollutants as well as greenhouse gases (GHGs). This document
includes information related to the illustrative air quality modeling analysis done in support of
the proposed rule and focuses on impacts to ambient concentrations of criteria and toxics
pollutants.
EPA conducted an illustrative air quality modeling analysis of a regulatory scenario involving
light- and medium-duty "onroad" vehicle emission reductions and corresponding changes in
"upstream" emission sources like EGU (electric generating unit) emissions and refinery
emissions. Decisions about the emissions and other elements used in the air quality modeling
were made early in the analytical process for the proposed rulemaking. Accordingly, the air
quality analysis does not represent the proposal's regulatory scenario, nor does it reflect the
expected impacts of the Inflation Reduction Act (IRA). Based on updated power sector modeling
that incorporated expected generation mix impacts of the IRA, we are projecting the IRA will
lead to a significantly cleaner power grid; because the air quality analysis presented here does
not account for these impacts on EGU emissions, the location and magnitude of the changes in
pollutant concentrations should be considered illustrative and not viewed as Agency projections
of what we expect will be the total impact of the proposed standards. Nevertheless, the analysis
provides some insights into potential air quality impacts associated with emissions increases and
decreases from these multiple sectors.
For this analysis, emission inventories were produced, and air quality modeling was
performed, for three scenarios: a 2016 base case, a 2055 reference scenario, and a 2055 light-
and medium duty vehicle (LMDV) regulatory case. 1 The illustrative LMDV regulatory case
assumes a light- and medium-duty fleet that phased-in to reach 50 percent of new vehicle sales as
BEVs in 2030 and remained constant at about 50 percent BEVs for model years 2030-2055, for a
total national light-duty vehicle population of 48% BEVs in 2055. The regulatory case also
assumes a phase-in of gasoline particulate filters for gasoline vehicles beginning in model year
2027.
An air quality modeling platform consists of all the emissions inventories and ancillary data
files used for emissions modeling, as well as the meteorological, initial condition, and boundary
condition files needed to run the air quality model. An emissions modeling platform consists of
the emissions modeling data and techniques including the emission inventories, the ancillary data
files, and the approaches used to transform inventories for use in air quality modeling.
1 Because the regulatory proposal had not been determined at the time of this analysis, the "regulatory case"
described here is not the proposed rule, but a plausible control scenario for illustrative purposes.
1
-------
This analysis utilizes the 2016v2 emissions modeling platform,2 which includes a suite of
base year (2016) and projection year (2023, 2026, 2032) inventories, along with ancillary
emissions data, and scripts and software for preparing the emissions for air quality modeling.
The Technical Support Document (TSD) Preparation of Emissions Inventories for the 2016v2
North American Emissions Modeling Platform describes how the emission inventories for each
year of data available in the platform were developed.3
Section 2 of this document gives a summary of the emissions inventory inputs to the air
quality modeling. Section 3 of this document describes the methodology for developing onroad
mobile emission inventories, Section 4 focuses on the methodology for developing electrical
generating unit (EGU) emission inventories, and Section 5 focuses on the methodology for
developing petroleum sector emission inventories. Section 6 provides emissions summary tables.
Sections 7 and 8 provide an overview of the air quality modeling methodology and results.
2 Emissions Inventory Methodology
This section provides an overview of the emission inventories used in the air quality analysis
for the proposed rule. These inventories include point sources, nonpoint sources, onroad and
nonroad mobile sources, commercial marine vessels (CMV), locomotive and aircraft emissions,
biogenic emissions, and fires for the U.S., Canada, and Mexico. The emissions used for the 2055
control scenarios were the same as those in the 2055 reference scenario for all emissions sectors
except for onroad mobile source emissions, EGU emissions, and petroleum sector emissions
(specifically refineries, crude oil production well sites and pipelines and natural gas production
well sites and pipelines).
For this study, the 2016 emission inventories used were based on those for the 2016v2
platform except for the U.S. onroad and nonroad4 mobile sources. For the 2055 cases, the U.S.
onroad mobile sources, U.S. nonroad mobile sources were projected to year 2055 levels, while
other anthropogenic emissions sources were retained at the 2016v2 platform projected emissions
levels for the year 2032. A high-level summary of the emission inventories used is provided in
this section, while the development of the U.S. onroad mobile source emissions is described in
detail in Section 3, the development of the EGU emissions is described in Section 4, and the
development of petroleum sector emissions is described in Section 5.
2.1 Emissions Inventory Sector Summary
For the purposes of preparing the air quality model-ready emissions, emission inventories are
split into "sectors". The significance of a sector is that each sector includes a specific group of
emission sources, and those data are run through the emissions modeling system independently
2 2016v2 Emissions Modeling Platform. SMOKE inputs available from https://gaftp.epa.gov/Air/emismod/2016/v2/
3 U.S. EPA (2022) Preparation of Emissions Inventories for 2016v2 North American Emissions Modeling Platform
Technical Support Document, https://www.epa.gov/air-emissions-modeling/2016-version-2-technical-support-
document.
4 The 2016 U.S. nonroad mobile source emissions inventory in the 2016v2 platform includes emissions for Texas
and California which were developed using their own tools. For this study, those state-supplied emissions were
replaced with 2016 nonroad emissions computed with an updated version of the Motor Vehicle Emission Simulator,
MOVES3.R1.
2
-------
from the other sectors up to the point of the final merging process. The final merge process
combines the sector-specific low-level (of the vertical levels in the AQ model) gridded,
speciated, hourly emissions together to create CMAQ-ready emission inputs. While pertinent
atmospheric emissions related to the problem being studied are included in each modeling
platform, the splitting of inventories into specific sectors for emissions modeling varies by
platform. The sectors for the 2016v2 emissions modeling platform used in this study are shown
in Table 2-1. Descriptions for each sector are provided. For more detail on the data used to
develop the 2016v2 inventories and on the processing of those inventories into air quality model-
ready inputs, see the 2016v2 emissions modeling platform TSD.5
Table 2-1 Inventory sectors included in the emissions modeling platform
Inventory Sector
Sector Description
Mobile - Nonroad
Mobile sources that do not drive on roads, excluding
locomotives, aircraft, and commercial marine vessels (see
Section 2.3)
Mobile - Onroad
Onroad mobile source gasoline and diesel vehicles from moving
and non-moving vehicles that drive on roads (see Section 3)
Mobile - Category 3 Commercial Marine
Vessels
Commercial marine vessels with Category 3 engines within and
outside of U.S. waters
Mobile - Category 1 and 2 Commercial
Marine Vessels
Commercial marine vessels with Category 1 and 2 engines
within and outside of U.S. waters
Mobile - Rail
U.S. Class I line haul, Class II/III line haul, passenger, and
commuter locomotives (does not include railyards and
switchers)
Nonpoint - Fertilizer
NH3 emissions from U.S. fertilizer sources
Nonpoint - Livestock
Primarily NH3 and VOC emissions from U.S. livestock sources
Nonpoint - Area Fugitive Dust
PM emissions from paved roads, unpaved roads and airstrips,
construction, agriculture production, and mining and quarrying
in the U.S.
Nonpoint - Residential Wood Combustion
U.S. residential wood burning emissions from devices such as
fireplaces, woodstoves, pellet stoves, indoor furnaces, outdoor
burning in fire pits and chimneys
Nonpoint - Oil and Gas
Oil and gas exploration and production, both onshore and
offshore
Nonpoint - Solvents
Nonpoint VOC emissions from solvents such as cleaners,
personal care products, and adhesives.
Nonpoint - Other
All nonpoint emissions in the U.S. not included in other sectors,
including industrial processes, waste disposal, storage and
transport of chemicals and petroleum, waste disposal,
commercial cooking, and miscellaneous area sources
Point - Airports
Aircraft engines and ground support equipment at U.S. airports
Point - Electrical Generating Units
Electric generating units that provide power to the U.S. electric
grid
Point - Oil and Gas
Point sources related to the extraction and distribution of oil and
gas in the U.S.
Point - Other
All point sources in the U.S. not included in other sectors.
Includes rail yards and refineries.
5 U.S. EPA (2022) Preparation of Emissions Inventories for 2016v2 North American Emissions Modeling Platform
Technical Support Document, https://www.epa.gov/air-emissions-modeling/2016-version-2-technical-support-
document.
3
-------
Point - Fires - Agricultural
Fires due to agricultural burning in the U.S.
Point - Fires - Wild and Prescribed
Wildfires and prescribed burns in the U.S.
Point - Non-U.S. Fires
Fires within the domain but outside of the U.S.
Biogenic (beis)
Emissions from trees, shrubs, grasses, and soils within and
outside of the U.S.
Canada - Mobile - Onroad
Onroad mobile sources in Canada (see Section 2.5)
Mexico - Mobile - Onroad
Onroad mobile sources in Mexico (see Section 2.5)
Canada/Mexico - Point
Canadian and Mexican point sources
Canada/Mexico - Nonpoint and Nonroad
Canadian and Mexican nonroad sources and nonpoint sources
not included in other sectors
Canada - Agricultural Point
Canadian agricultural ammonia sources
Canada - oil and gas 2D
Canadian low-level point oil and gas sources
Canada - Nonpoint - Area Fugitive Dust
Area source fugitive dust sources in Canada
Canada - Point - Point Fugitive Dust
Point source fugitive dust sources in Canada
2.2 The Emissions Modeling Process
The CMAQ air quality model requires hourly emissions of specific gas and particle species
for the horizontal and vertical grid cells contained within the modeled region (i.e., modeling
domain). To provide emissions in the form and format required by the model, it is necessary to
"pre-process" the emissions inventories for the sectors described above. The process of
emissions modeling transforms the emissions inventories from their original temporal, pollutant,
and spatial resolution into the hourly, speciated, gridded resolution required by the air quality
model. Emissions modeling includes the chemical speciation, temporal allocation, and spatial
allocation of emissions along with final formatting of the data that will be input to the air quality
model.
Chemical speciation creates the "model species" needed by CMAQ, for a specific chemical
mechanism, from the "inventory pollutants" of the input emission inventories. These model
species are either individual chemical compounds (i.e., "explicit species") or groups of species
(i.e., "lumped species"). The chemical mechanism used for this platform is the CB6 mechanism.6
This platform generates the PM2.5 model species associated with the CMAQ Aerosol Module
version 7 (AE7). See Section 3.2 of the 2016v2 platform TSD for more information about
chemical speciation in the 2016v2 platform.
Temporal allocation is the process of distributing aggregated emissions to a finer temporal
resolution, for example converting annual emissions to hourly emissions as is required by
CMAQ. While the total annual, monthly, or daily emissions are important, the hourly timing of
the occurrence of emissions is also essential for accurately simulating ozone, PM, and other
pollutant concentrations in the atmosphere. Many emissions inventories are annual or monthly in
nature. Temporal allocation takes these aggregated emissions and distributes the emissions to the
hours of each day. This process is typically done by applying temporal profiles to the inventories
in this order: monthly, day of the week, and diurnal, with monthly and day-of-week profiles
applied only if the inventory is not already at that level of detail. See Section 3.3 of the 2016v2
6 Yarwood, G., et al. (2010) Updates to the Carbon Bond Chemical Mechanism for Version 6 (CB6). Presented at
the 9th Annual CMAS Conference, Chapel Hill, NC. Available at
https://www.cmascenter.org/conference/2010/abstracts/emery updates carbon 2010.pdf.
4
-------
platform TSD for more information about temporal allocation of emissions in the 2016v2
platform.
Spatial allocation is the process of distributing aggregated emissions to a finer spatial
resolution, as is required by CMAQ. Over 60 spatial surrogates are used to spatially allocate U.S.
county-level emissions to thel2-km grid cells used by the air quality model. See Section 3.4 of
the 2016v2 platform TSD for a description of the spatial surrogates used for allocating county-
level emissions in the 2016v2 platform.
The primary tool used to perform the emissions modeling to create the air quality model-
ready emissions was the Sparse Matrix Operator Kernel Emissions (SMOKE) modeling system,
version 4.9 (SMOKE 4.9). When preparing emissions for CMAQ, emissions for each sector are
processed separately through SMOKE. The elevated point source emissions are passed to
CMAQ directly so the model can perform plume rise based on hourly meteorological conditions,
while the low-level emissions are combined to create model-ready 2-D gridded emissions.
Gridded emissions files were created for a 36-km national grid named 36US3 and for a 12-km
national grid named 12US2, both of which include the contiguous states and parts of Canada and
Mexico as shown in Figure 2-1. This figure also shows the region covered by other grids that are
relevant to the development of emissions for this and related studies.
Figure 2-1 Air quality modeling domains
ฆWRF_36NOAM
BELD4
12US1
12US2
5
-------
2.3 Emissions Inventory Methodology for 2016v2-Compatible Sectors
Except for the onroad mobile source emissions, the emissions used for the 2016 air quality
case are consistent with those developed through the 2016v2 Platform. For the 2055 cases,
emissions for sectors other than U.S. onroad and nonroad mobile sources and emissions for
onroad mobile sources for Canada and Mexico, were developed to be consistent with the 2032
emissions developed by the Inventory Collaborative and are described in the 2016v2 Platform
TSD. Development of the 2055 nonroad emissions is described in Section 2.4. The development
of the U.S. onroad mobile source emissions for each of the cases is described below in Section 3.
Additionally, for CMV, the 2016v3 inventories, which have improved state and county
apportionment compared to 2016v2, were used. Nonpoint-solvents also used the new 2016v3
inventory as of May 2022 (not final 2016v3) and associated speciation/gridding/temporal xrefs
and profiles. For the point (non-egu) sector 2016v2 was used with the difference being that point
solvents were included. Another update that was made for this modeling included using the
corrected BEIS 3.7 (processed in April/May 2022), with CMAQ run using inline biogenics.
2.4 2055 Emissions Inventory Methodology for the Nonroad Sector
To prepare the nonroad mobile source emissions, an updated version of the Motor Vehicle
Emission Simulator (MOVES), MOVES3.R1, was run using inputs compatible with the 2016v2
platform for all states. The nonroad component of MOVES was configured to create a national
nonroad inventory for 2055. The 2055 MOVES nonroad inventory was used in all states.
2.5 2055 Emissions Inventory Methodology for Fugitive Dust
The inventory for road dust is generated using VMT7, and the total projected VMT in 2055
did not change between the reference and LMDV regulatory scenario (only the fraction of EVs
changed). Road dust inventories for 2055 were projected using 2055 VMT (see Section 3.2.2)
and are presented in Table 6-7.
3 Onroad Emissions Inventory Methodology
This section focuses on the approach and data sources used to develop gridded, hourly
emissions for the onroad mobile sector that are suitable for input to an air quality model in terms
of the format, grid resolution, and chemical species. While the emission factors used to develop
emissions for the reference and control scenarios differed, the approach and all other data
sources used to calculate emissions for both scenarios were identical.
Onroad mobile source emissions result from motorized vehicles operating on public
roadways. These include passenger cars, motorcycles, minivans, sport-utility vehicles, light-duty
trucks, heavy-duty trucks, and buses. The sources are further divided by the fuel they use,
including diesel, gasoline, E-85, electricity, and compressed natural gas (CNG) vehicles. The
sector accounts for emissions from parked vehicle processes (e.g., starts, hot soak, and extended
idle) as well as from on-network processes (i.e., from vehicles as they move along the roads).
The onroad emissions are generated using SMOKE programs that leverage MOVES-generated
7 See Section 4.2.3.1 of the 2016v2 TSD for more detail on how fugitive dust is projected.
6
-------
emission factors with county, fuel type, source type, and road type-specific activity data, along
with hourly meteorological data.
The MOVES-generated onroad emission factors were combined with activity data (e.g.,
vehicle miles traveled, vehicle population) to produce emissions within the Sparse Matrix
Operator Kernel Emissions (SMOKE) modeling system. The collection of programs that
compute the onroad mobile source emissions are known as SMOKE-MOVES. SMOKE-MOVES
uses a combination of vehicle activity data, emission factors from MOVES, meteorology data,
and temporal allocation information needed to estimate hourly onroad emissions. Additional
types of ancillary data are used for the emissions processing, such as spatial surrogates which
spatially allocate emissions to the grid used for air quality modeling.
More details on the generation of the emission factors, activity data, and on the modeling of
the emissions are in the following subsections. National onroad emission summaries for key
pollutants are provided in Section 4.
3.1 Emissions Factor Table Development
Onroad mobile source emission factors were generated for the modeled cases by running an
updated version of MOVES38 (MOVES3.R1) which incorporates the latest vehicle activity data,
newer emission rules, and changes that reflect improvements in EPA's understanding of vehicle
emissions and add features to better model electric vehicles. More information on these updates
is available in a memo to the docket.9
In addition to the updates incorporated in MOVES3.R1 that were used in all three modeling
cases, we also developed case-specific inputs. The LMDV regulatory case assumes light-duty
and medium-duty BEV sales of about 50 percent for model years 2030 and beyond, plus some
improvements to particulate matter emissions for light-duty and medium-duty vehicles as
detailed in the DRIA. Thus, case-specific BEV fractions were incorporated into each county's
fuel mix described in Section 3.2.2.5 below. Also, for the LMDV regulatory case, we reduced
the gasoline light- and medium-duty PM exhaust emission rates to account for GPF control and
changed the adjustment weight in EVPopICEAdjustLD to zero for HC and NOx to indicate no
averaging with electric vehicles. This effectively caps the light- and medium-duty internal
combustion NOx and HC emissions at the model year 2026 rate. We did not change the
adjustment weights for energy consumption since this case assumed that CO2 averaging with
electric vehicles would continue.
The emission factor tables input to SMOKE-MOVES are generated by running MOVES.
These tables differentiate emissions by process (i.e., running, start, vapor venting, etc.), fuel type,
vehicle type, road type, temperature, speed bin for rate per distance processes, hour of day, and
day of week. To generate the MOVES emission factors across the U.S., MOVES was run to
produce emission factors for a series of temperatures and speeds for a set of "representative
8 USEPA (2020) Motor Vehicle Emission Simulator: MOVES3. Office of Transportation and Air Quality. US
Environmental Protection Agency. Ann Arbor, MI. November 2020. https://www.epa.gov/moves.
9 USEPA (2023). Updates to MOVES for the Multi-Pollutant Emissions Standards for Model Years 2027 and Later
Light-Duty and Medium-Duty Vehicles, Memo to Docket, February 2023. Memorandum to Docket EPA-HQ-OAR-
2022-0829. February 2023
7
-------
counties," to which every other county in the country is mapped. The representative counties for
which emission factors are generated are selected according to their state, elevation, fuels used in
the region, vehicle age distribution, and inspection and maintenance programs. Every county in
the country is mapped to a representative county based on its similarity to the representative
county with respect to those attributes. The representative counties selected for the 2016v2
platform were retained for this analysis. More details on the methodology behind choosing
representative counties is available in the 2016v2 TSD.
Emission factors were generated by running MOVES for each representative county for two
"fuel months" - January to represent winter months and July to represent summer months -
because in some parts of the country different types of fuels are used in each season. MOVES
was run for the range of temperatures that occur in each representative county for each season.
The calculations of the temperature ranges needed for each fuel month were based on
meteorology for every county and grid cell in the continental U.S. for each hour of the year. The
SMOKE interface accounts for the sensitivity of the on-road emissions to temperature and
humidity by using the gridded hourly temperature information available from the meteorological
model outputs used for air quality modeling.
MOVES3.R1 was run using the above approach to create emission factors for each of the
three modeling cases: 2016 base year, 2055 reference, and a 2055 regulatory case. A new set of
emission factor tables were developed for this study using the same representative counties as
were used the 2016v2 platform. The county databases (CDBs) input to MOVES for 2016 were
equivalent to those used for the 2016v2 platform. To prepare the 2055 CDBs used to generate
year 2055 emissions factors, the vehicle age distributions were projected to reflect the year 2055
as were the tables representing the inspection and maintenance programs. The fuels used were
also representative of year 2055. The CDBs for each of the 2055 modeling cases incorporated the
case-specific fuel mix as detailed in Section 3.2.2.5 below.
3.2 Activity and Other Data Development
To compute onroad mobile source emissions, SMOKE selects the appropriate MOVES
emission rates for each county, hourly temperature, speed bin, and SCC (which includes the fuel
type, source type and road type), then multiplies the emission rate by the appropriate activity
data such as VMT (vehicle miles travelled), VPOP (vehicle population), SPEED/SPDIST (speed
distributions and averages), HOTELING (hours of extended idle), ONI (hours of off-network
idling), or STARTS (engine starts), to produce emissions. For each of these activity datasets, first
a national dataset was developed; this national dataset is called the "EPA default" dataset. Data
submitted by state agencies were incorporated into the activity data sets used for the study where
they were available and passed quality assurance checks.
The activity data for the 2016 base year were consistent with the activity data used in the
2016v2. Additional details on the development of activity data are available in the 2016v2 TSD.
In addition to activity data, this section also describes inputs for fuel parameters and county-
specific vehicle inspection and maintenance programs.
8
-------
3.2.1 2016 Base Year Activity data
3.2.1.1 Vehicle Miles Traveled (VMT)
EPA calculated default 2016 VMT by backcasting the 2017 NEI VMT to 2016. The 2017 NEI
Technical Support Document10 contains details on the development of the 2017 VMT. The data
backcast to 2016 were used for states that did not submit 2016 VMT data. The factors to adjust
VMT from 2017 to 2016 were based on VMT data from the FHWA county-level VM-2 reports..
For most states, EPA calculated county-road type factors based on FHWA VM-2 County data for
2017 and 2016. Separate factors were calculated by vehicle type for each MOVES road type.
Some states have a very different distribution of urban activity versus rural activity between
2017 NEI and the FHWA data, due to inconsistencies in the definition of urban versus rural. For
those states, a single county-wide projection factor based on total FHWA VMT across all road
types was applied to all VMT, independent of road type. County-total-based (instead of
county+road type) factors were used for all counties in IN, MS, MO, NM, TN, TX, and UT
because many counties had large increases in one particular road type and decreases in another
road type.
For the 2016vl platform, VMT data submitted by state and local agencies were incorporated
and used in place of EPA defaults. Note that VMT data need to be provided to SMOKE for each
county and SCC. The onroad SCCs characterize vehicles by MOVES fuel type, vehicle (aka
source) type, emissions process, and road type. Any VMT provided at a different resolution than
this were converted to a full county-SCC resolution to prepare the data for processing by
SMOKE.
A final step was performed on all state-submitted VMT. The distinction between a "passenger
car" (MOVES source type 21) versus a "passenger truck" (MOVES source type 31) versus a
"light commercial truck" (MOVES source type 32) is not always consistent between different
datasets. This distinction can have a noticeable effect on the resulting emissions, since MOVES
emission factors for passenger cars are quite different than those for passenger trucks and light
commercial trucks.
To ensure consistency in the 21/31/32 splits across the country, all state-submitted VMT for
MOVES vehicle types 21,31, and 32 (all of which are part of HPMS vehicle type 25) was
summed, and then re-split using the 21/31/32 splits from the EPA 2016v2 default VMT. VMT
for each source type as a percentage of total 21/31/32 VMT was calculated by county from the
EPA default VMT. Then, state-submitted VMT for 21/31/32 was summed and re-split according
to those percentages.
3.2.1.2 Vehicle Population (VPOP)
The EPA default VPOP dataset was based on the EPA default VMT dataset described above.
In the areas where EPA backcasted 2017 NEI VMT:
10 U.S. EPA (2021) 2017 National Emissions Inventory: January 2021 Updated Release, Technical Support
Document, https://www.epa.gov/air-emissions-inventories/2017-national-emissions-inventory-nei-technical-
support-document-tsd
9
-------
2016v2 VPOP = 2016v2 VMT * (VPOP/VMT ratio by county-SCC6).
where the ratio by county-SCC is based on 2017NEI with MOVES3.R1 fuel splits and SCC6
means the first six digits of the SCC code that include fuel type and source type but exclude the
road type and process. In the areas where we used 2016vl VMT resplit to MOVES3.R1 fuels,
2016v2 VPOP = 2016vl VPOP with two resplits: first, source types 21/31/32 were resplit
according to 2017 NEI EPA default 21/31/32 splits so that the whole country has consistent
21/31/32 splits. Next, fuels were resplit to MOVES3.R1 fuels. There are some areas where 2016
VMT was submitted but 2016 VPOP was not; those areas are using 2016vl VPOP (with
resplits). The same method was applied to the 2016 EPA default VMT to produce an EPA
default VPOP data set.
3.2.1.3 Speed Activity (SPEED/SPDIST)
In the version of SMOKE used for this analysis (SMOKE 4.7), SMOKE-MOVES was
updated to use speed distributions similarly to how they are used when running MOVES in
inventory mode. This new speed distribution file, called SPDIST, specifies the amount of time
spent in each MOVES speed bin for each county, vehicle (aka source) type, road type,
weekday/weekend, and hour of day. This file contains the same information at the same
resolution as the Speed Distribution table used by MOVES but is reformatted for SMOKE.
Using the SPDIST file results in a SMOKE emissions calculation that is more consistent with
MOVES than the previous hourly speed profile (SPDPRO) approach, because emission factors
from all speed bins can be used, rather than interpolating between the two bins surrounding the
single average speed value for each hour as is done with the SPDPRO approach.
As was the case with the previous SPDPRO approach, the SPEED inventory that includes a
single overall average speed for each county, SCC, and month, must still be read in by the
SMOKE program Smkinven. SMOKE requires the SPEED dataset to exist even when speed
distribution data are available, even though only the speed distribution data affects the selection
of emission factors. The SPEED and SPDIST datasets are carried over from 2017 NEI and are
based on a combination of the Coordinating Research Council (CRC) A-100 data and 2017 NEI
MOVES CDBs.
3.2.1.4 Hoteling Hours (HOTELING)
Hoteling hours activity is used to calculate emissions from extended idling and auxiliary
power units (APUs) for heavy duty diesel vehicles. For the 2016v2 platform, hoteling hours were
computed using a new factor identified by EPA's Office of Transportation and Air Quality as
more appropriate based on recent studies.
The method used in 2016v2 is the following:
1 Start with 2016 VMT for combination long haul trucks (i.e., MOVES source type 62)
on restricted roads, by county.
10
-------
2 Multiply the VMT by 0.007248 hours/mile.11
3 Apply parking space reductions to keep hoteling within the estimated maximum
hours by county, except for states that requested EPA do not do that (CO, ME, NJ,
NY).
Hoteling hours were adjusted down in counties for which there were more hoteling hours
assigned to the county than could be supported by the known parking spaces. To compute the
adjustment, the hoteling hours for the county were computed using the above method, and
reductions were applied directly to the 2016 hoteling hours based on known parking space
availability so that there were not more hours assigned to the county than the available parking
spaces could support if they were full every hour of every day.
A dataset of truck stop parking space availability with the total number of parking spaces per
county was used in the computation of the adjustment factors.12 This same dataset is used to
develop the spatial surrogate for hoteling emissions. Since there are 8,784 hours in the year
2016; the maximum number of possible hoteling hours in a particular county is equal to 8,784 *
the number of parking spaces in that county. Hoteling hours for each county were capped at that
theoretical maximum value for 2016 in that county unless the number of parking spaces listed
was less than 12, in which case the hours were not reduced.
For 2016v2, hoteling was calculated as:
2016v2 HOTELING = 2017NEI HOTELING * 2016v2 VMT/2017NEI VMT
This is effectively consistent with applying the 0.007248 factor directly to the 2016v2 VMT.
Then, for counties that provided 2017 hoteling but did not have vehicle type 62 restricted VMT
in 2016 - that is, counties that should have hoteling, but do not have any VMT from which to
calculate it - EPA backcast 2017 hoteling to 2016 using the FHWA-based county total 2017 to
2016 trend. Finally, the annual parking-space-based caps for hoteling hours were applied. The
same caps were used as for 2017 NEI, except recalculated for a leap year (multiplied by
366/365).
3.2.1.5 Off-Network Idling Hours (ONI)
After creating VMT inputs for SMOKE-MOVES, Off-network idle (ONI) activity data were
also needed. ONI is defined in MOVES as time during which a vehicle engine is running idle
and the vehicle is somewhere other than on the road, such as in a parking lot, a driveway, or at
the side of the road. This engine activity contributes to total mobile source emissions but does
not take place on the road network.
11 USEPA (2020). Population and Activity of Onroad Vehicles in MOVES3. EPA-420-R-20-023. Office of
Transportation and Air Quality. US Environmental Protection Agency. Ann Arbor, MI. November 2020.
https://www.epa.gov/moves/moves-technical-reports.
12 From 2016 version 1 hoteling workbook.xlsx developed based on the input dataset for the hoteling spatial
surrogate in the 2016vl platform.
11
-------
Examples of ONI activity include:
light duty passenger vehicles idling while waiting to pick up children at school or to
pick up passengers at the airport or train station,
single unit and combination trucks idling while loading or unloading cargo or making
deliveries, and
vehicles idling at drive-through restaurants.
Note that ONI does not include idling that occurs on the road, such as idling at traffic signals,
stop signs, and in trafficthese emissions are included as part of the running and crankcase
running exhaust processes on the other road types. ONI also does not include long-duration
idling by long-haul combination trucks (hoteling/extended idle), as that type of long duration
idling is accounted for in other MOVES processes.
ONI activity hours were calculated based on VMT. For each representative county, the ratio
of ONI hours to onroad VMT (on all road types) was calculated using the MOVES ONI Tool by
source type, fuel type, and month. These ratios were then multiplied by each county's total VMT
(aggregated by source type, fuel type, and month) to get hours of ONI activity.
3.2.1.6 Engine Starts (STARTS)
Onroad "start" emissions are the instantaneous exhaust emissions that occur at the engine start
(e.g., due to the fuel rich conditions in the cylinder to initiate combustion) as well as the
additional running exhaust emissions that occur because the engine and emission control systems
have not yet stabilized at the running operating temperature. Operationally, start emissions are
defined as the difference in emissions between an exhaust emissions test with an ambient
temperature start and the same test with the engine and emission control systems already at
operating temperature. As such, the units for start emission rates are instantaneous grams/start.
MOVES3.R1 uses vehicle population information to sort the vehicle population into source
bins defined by vehicle source type, fuel type (gas, diesel, etc.), regulatory class, model year and
age. The model uses default data from instrumented vehicles (or user-provided values) to
estimate the number of starts for each source bin and to allocate them among eight operating
mode bins defined by the amount of time parked ("soak time") prior to the start. Thus,
MOVES3.R1 accounts for different amounts of cooling of the engine and emission control
systems. Each source bin and operating mode has an associated g/start emission rate. Start
emissions are also adjusted to account for fuel characteristics, LD inspection and maintenance
programs, and ambient temperatures.
2016v2 STARTS = 2016v2 VMT * (2017 STARTS/ 2017 VMT by county&SCC6)
12
-------
3.2.1.7 Fuels
The 2016 scenario used MOVES3.R1 default fuels. These fuels are the same as the fuels in
MOVES3.0.1.13
3.2.2 2055 Projected Activity Data
The projected 2055 activity data are primarily based on the 2016v2 platform's projected 2032
data, updated to be consistent with the default data and algorithms in MOVES3.R1, as well as to
estimate geographic differences in fuel and age distributions. To accomplish this analysis, the
following steps were taken:
1. Calendar year 2055 CDBs were developed for each representative county, as
described in more detail later in this section. Each scenario (the reference case and the
two control cases) had its own set of CDBs.
2. MOVES3.R1 was run with each CDB to calculate detailed activity data for each
representative county.
3. The MOVES activity results for each representative county were allocated to the
individual counties represented by each representative county using the 2016v2
platform allocations.
The following sections describe how the 2055 CDBs were developed to calculate the 2055
projected activity data.
3.2.2.1 Data Used As-is from the 2016v2 Platform
The starting point for developing the 2055 CDBs was the 2016v2 platform for calendar year
2032. The following data were used as-is from the 2016v2 platform data in the 2055 CDBs:
Geography tables: State, County, Zone, and ZoneMonthHour
VMT distribution tables: MonthFraction, DayFraction, and HourFraction
Speed distribution table: AvgSpeedDistribution
Road distribution tables: RoadTypeDistribution and ZoneRoadType
Retrofit table: OnroadRetrofit
3.2.2.2 Default Data Used As-is from MOVES3.R1
National default data and algorithms in MOVES3.R1 were used for the following tables:
Some (but not all) fuels tables: FuelFormulation, FuelSupply, and FuelUsageFraction
13U.S. EPA (2021) Fuel Supply Defaults: Regional Fuels and the Fuel Wizard in MOVES3, EPA-420-R-21-006.
Office of Transportation and Air Quality. US Environmental Protection Agency. Ann Arbor, MI. March 2021.
13
-------
Starts tables: StartsPerDayPerVehicle, StartsMonthAdjust, StartsHourFraction,
Starts Age Adjustment, and StartsOpModeDistribution
Hotelling tables: HotellingHoursPerDay, HotellingMonthAdjust,
HotellingHourFraction, HotellingAgeFraction, and HotellingActivityDistribution
Off-Network Idle tables: TotalldleFraction, IdleModelYearGrouping,
IdleMonthAdjust, and IdleDayAdjust
I/M table: IMCoverage
Note that in MOVES3.R1, starts, hotelling, and off-network idle tables are optional tables,
and therefore can be empty in a CDB if the intention is to use default data. Therefore, these
tables are empty in the 2055 CDBs. However, the fuels tables and I/M table are required inputs.
Since the default database contains county (or region) specific data, the 2055 CDBs contain the
relevant subset of the default database's data. See the MOVES3.R1 technical reports14'15'16'17'
18'19for more information about how these default data were derived.
3.2.2.3 Default Data from MOVES3.R1 Allocated Using 2016v2 Platform
National default data in MOVES3.R1 were allocated to representative counties for the
following tables:
VMT table: HPMSVTypeYear
VPOP table: SourceTypeYear
VMT fractions by HPMSVTypelD and county were calculated from the 2032 VMT
projections in the 2016v2 platform and used to allocate the national default VMT projections for
2055 to the county level. Similarly, VMT fractions by sourceTypelD and county were calculated
from the 2016v2 platform to allocate the national default VPOP projections for 2055. See the
MOVES3.R1 technical report for more information about how the national default data were
derived.18
3.2.2.4 2055 Age Distributions
Each CDB has a sourceTypeAgeDistribution table. The 2055 age distributions by
representing county were primarily derived using July 1, 2020, vehicle registration data
14 U.S. EPA (2023) Exhaust Emission Rates for Light Duty Onroad Vehicles in MOVES3.R1. Office of
Transportation and Air Quality. US Environmental Protection Agency. Ann Arbor, MI. February 2023.
15 U.S. EPA (2023) Exhaust Emission Rates for Heavy Duty Onroad Vehicles in MOVES3.R1. Office of
Transportation and Air Quality. US Environmental Protection Agency. Ann Arbor, MI. February 2023.
16 U.S. EPA (2023) Emission Adjustments for Onroad Vehicles inMOVES3.Rl. Office of Transportation and Air
Quality. US Environmental Protection Agency. Ann Arbor, MI. February 2023.
17 U.S. EPA (2023) Evaporative Emissions from Onroad Vehicles in MOVES3.R1. Office of Transportation and Air
Quality. US Environmental Protection Agency. Ann Arbor, MI. February 2023.
18 U.S. EPA (2023) Population and Activity of Onroad Vehicles inMOVES3.Rl. Office of Transportation and Air
Quality. US Environmental Protection Agency. Ann Arbor, MI. February 2023.
1919 U.S. EPA (2023) Greenhouse Gas and Energy Consumption Rates for Onroad Vehicles in MOVES3.R1. Office
of Transportation and Air Quality. US Environmental Protection Agency. Ann Arbor, MI. February 2023.
14
-------
purchased from IHS Markit, vehicle stock and sales projections from the Annual Energy Outlook
(AEO) 202120, and vehicle scrappage rates presented in the Transportation Energy Data Book
(TEDB).21 The age distributions were calculated using a modified version of the age distribution
projection algorithm described in Appendix C of the Population and Activity of Onroad Vehicles
in MOVES3 ,R1 technical report. 18 The algorithm was modified to maintain differences between
counties, such that counties that had newer-than-average fleets in 2020 continue to have newer-
than-average fleets in 2055 and, similarly, counties with older fleets now have older fleets in the
future. The fundamental approach to solving this problem was to define how age distributions in
a local area are different from the national average, and then apply that difference to future years.
The following algorithm was implemented for calculating a representative county's base age
distribution:
1. Subset the 2020 registration data to get vehicle counts by source type and model year
for all counties represented by the representative county.
2. Group all model years 1990 and older together, because MOVES groups all vehicles
ages 30 and older together.
3. Calculate age fractions by source type.
4. Replace age 0 (model year 2020) fractions with the ratio of vehicle sales to stock from
AEO. This is because the July 1 registration data pull represents an incomplete year.
5. Renormalize the age distributions, retaining the age 0 fractions.
The following equations were used to project a representative county's base age distribution
one year into the future:
Population distribution for the next calendar year = Population distribution for the
current calendar year, minus vehicles scrapped in the current calendar year, plus
locally adjusted new vehicle sales in the next year
Vehicles scrapped in the current calendar year = Scrappage factor times the base
scrappage rate times the population distribution for the current calendar year
Scrappage factor = (Total number of vehicles in the current year, minus total number
of vehicles in the next year, plus locally adjusted new vehicle sales in the next year)
divided by the sum of the base scrappage rate times the current year's population
distribution. The purpose of the scrappage factor is to scale the base scrappage rate to
balance the equation accounting for the total number of vehicles in each calendar year.
For example, if the total number of vehicles remains constant from one year to the
next and vehicle sales are high, then the scrappage factor would be high as well, as
20 US Energy Information Administration (EIA), Annual Energy Outlook 2021, Washington, DC: February 2021.
https://www.eia. gov/outlooks/archive/aeo21/
21 Davis, S. and R Boundy (2022), Transportation Energy Data Book, Ed. 40, Oak Ridge National Laboratory,
ORNL/TM-2022/2376, https://tedb.orn.Lgov/wp-content/npioads/2022/03/TEDB Ed_40.pdf
15
-------
more vehicles would be scrapped to balance out the higher sales while maintaining
constant number of total vehicles.
The population distribution of the current calendar year is known; thus, the algorithm starts
with the base age distribution and then the algorithm iterates, so the output of the algorithm is the
input for the next year. The total number of vehicles in the next year and the vehicle sales in the
next year are also known, based on AEO. The base scrappage curve is also known, based on data
presented in TEDB.
The differences between local areas were accounted for by applying a local sales scaling
factor to the number of new vehicles sold in the next year in the equations above. This scaling
factor was defined as the difference between the local and the national population fractions
summed over an age range [1, j], divided by the national population fraction over the same age
range. Essentially, this is using the fraction of newer vehicles in a local fleet compared to the
national average as a surrogate for what future sales in a local area might be.
The precise age range [1, j] used was determined for each source type, chosen so that the
difference between the local average age and the projected national average age in 2055 was as
close as possible to the difference between the local and national average ages in 2020. That is,
the chosen age ranges tried to maintain the same delta in average age between the local and the
national case in the future. The chosen age ranges by source type were:
Motorcycles: [1, 7]
Passenger cars: [1, 10]
Passenger trucks: [1, 5]
Light commercial trucks: [1, 5]
Other buses: [1, 10]
Transit buses: [1, 10]
School buses: [1,8]
Refuse trucks: [1, 10]
Single unit trucks: [1,7]
Motor homes: [1, 10]
Combination short-haul trucks: [1, 10]
Note that for some counties, some source types were not present in the IHS data. In these rare
cases, the national default age distributions were assumed. Additionally, combination long-haul
trucks were assumed to have the same age distribution nationally.
16
-------
The algorithm and data described above were used to calculate SourceTypeAgeDistribution
tables for each representing county in 2055. The same age distributions were used for all
scenarios. The following figures show the resulting projected average age in 2055 by county for
the light-duty source types.
Average Age
19 8
I 18.0
I
14.0
10.0
7.7
Figure 3-1 Projected average age of passenger cars in 2055
Average Age
20.2
18.0
N
15.0
12.0
9.0
7.2
17
-------
Figure 3-2 Projected average age of passenger trucks in 2055
Figure 3-3 Projected average age of light commercial trucks in 2055
3.2.2.5 2055 Fuel Mix
The mix of the fuel types used in vehicles (or "fuel distributions") for 2055 rely on national
projections, which vary by scenario. The national projected fuel distributions for the reference
case rely on July 1, 2020, vehicle registration data purchased from IHS Markit, vehicle sales
projections from AEO2021,20 EPA's Revised 2023 and Later Model Year Light-Duty Vehicle
Greenhouse Gas Emissions Standards,22 and CARB's Advanced Clean Trucks regulation. More
information about the national projected fuel distributions for the reference case can be found in
the MOVES3.R1 technical report.18
Fuel distributions for the regulatory case assume a shift to more electric vehicles, with the
market share for electric vehicles reaching about 50 percent of light and medium-duty sales.
Additional details are available in the Draft RIA.
The goal of the representative county fuel distribution projection was to maintain differences
between counties, while simultaneously maintaining the projected national average electric
vehicle (EV) penetration rates. That is, counties with higher-than-average fractions of EVs in
2020 are assumed to have higher-than-average fractions of EVs in the future reference and
regulatory scenario. The fundamental approach to solving this problem was to define how a
22 U.S. EPA (2021). Revised 2023 and Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions Standards
(86 FR 74434, December 30, 2021)
18
-------
local area is different from the national average, and then apply that difference with the
appropriate weighting to future years.
The 2020 vehicle registration data were subset for all counties represented by a representative
county, and the local EV fraction for model year 2019 (the most recent complete model year in
the registration data) was calculated by dividing the number of EVs by the total number of
vehicles for each light-duty source type. The local EV fraction was divided by the national EV
fraction to obtain the local to national ratio for each representative county.
Since EVs are not randomly distributed throughout the country (they are more likely to appear
in areas that incentivize EVs), these ratios were weighted based on the vehicle populations in the
local area, so that the national average EV fraction was maintained. Specifically, a national
average ratio was calculated for each source type and model year, weighted by the source type
and model year VPOP in each local area; the local ratios were then divided by this national
average ratio, so that the effective national average ratio was 1.
Additionally, EPA assumes that as EVs become more prevalent, they will be less concentrated
geographically. Therefore, a maximum ratio value of 2.0 was chosen, representing that in the
future, no county will have more than 2 times the EV penetration as the national average. In the
algorithm description below, the "uncapped ratio" represents the local ratio as described above,
and the "capped ratio" is the lesser of the uncapped ratio or 2.0.
To calculate light-duty fuel distributions in each representative county, the following
algorithm was implemented for each source type and model year:
1. Calculate the local vehicle population (based on the SourceTypeYear and
SourceTypeAgeDistribution tables in the CDB).
2. Calculate the uncapped EV fraction by multiplying the uncapped ratio by the national
EV fraction.
3. Calculate the capped EV fraction by multiplying the capped ratio by the national EV
fraction.
4. If the uncapped fraction is greater than the capped fraction, the number of excess EVs
in this county is calculated by multiplying the vehicle population by the difference
between the uncapped and capped fractions.
5. If the capped fraction is greater than 1.0 (this is possible when the national average EV
fraction is greater than 50%), additional excess EVs are calculated by multiplying the
vehicle population by the difference between the capped fraction and 1.0.
6. The number of EVs in the county is calculated by multiplying the vehicle population
by the capped fraction or 1.0, whichever is less.
7. The internal combustion engine (ICE) fraction is 1 minus the EV fraction. Gasoline
counts are calculated by multiplying the ICE fraction by AEO gasoline sales divided
19
-------
by AEO (gasoline + diesel + FFV sales). Diesel and FFV counts are calculated
similarly.
8. Steps 1-7 are performed for each representative county.
9. Excess EVs are reallocated to all representative counties that have not reached their
EV penetration cap. This is done proportionally across representative counties,
weighted by the number of remaining ICE vehicles in each county. Within a
representative county, the excess EVs proportionally reduce the number of ICE
vehicles. If this reallocation would bring a county to over 100% EVs, the reallocation
step is repeated until all excess EVs have been placed in a county.
Once all excess EVs were reallocated, the light-duty fuel distributions were formatted for use
in the MOVES AVFT table and were stored in the CDBs.
Note that the heavy-duty fuel distributions were not assumed to vary geographically. The
national average fuel distributions for all heavy-duty source types were used uniformly across all
representative counties. The following figures compare the projected EV penetration rates by
county in 2055 between the reference case and the regulatory case for each light-duty source
type.
20
-------
-------
-------
EV Fraction
100%
I
75%
50%
25%
0%
Figure 3-6 Comparing light commercial truck EV penetrations in 2055
3.3 Onroad Emissions Modeling
The SMOKE-MOVES process for creating the air quality model-ready onroad mobile
emissions consists of the following steps:
1) Select the representative counties to use in the MOVES runs.
2) Determine which months will be used to represent other month's fuel characteristics.
23
-------
3) Create inputs needed only by MOVES. MOVES requires county-specific information on
vehicle populations, age distributions, speed distribution, road type distributions,
temporal profiles, inspection-maintenance programs, and presence of Low Emission
Vehicle (LEV) program for each of the representative counties.
4) Create inputs needed both by MOVES and by SMOKE, including temperatures and
activity data.
5) Run MOVES to create emission factor tables for the temperatures and speeds that exist in
each county during the modeled period.
6) Run SMOKE to apply the emission factors to activity data (VMT, VPOP, HOTELING,
STARTS, ONI) to calculate emissions based on the gridded hourly temperatures in the
meteorological data.
7) Aggregate the results to the county-SCC level for summaries and quality assurance.
The onroad emissions are processed as five components that are merged into the final onroad
sector emissions:
rate-per-distance (RPD) uses VMT as the activity data plus speed and speed profile
information to compute on-network emissions from exhaust, evaporative, permeation,
refueling, and brake and tire wear processes;
rate-per-vehicle (RPV) uses VPOP activity data to compute off-network emissions from
exhaust, evaporative, and permeation processes;
rate-per-profile (RPP) uses VPOP activity data to compute off-network emissions from
evaporative fuel vapor venting, including hot soak (immediately after a trip) and diurnal
(vehicle parked for a long period) emissions;
rate-per-start (RPS) uses START activity data to compute off-network emissions from
vehicle starts;
rate-per-hour (RPH) uses hoteling hours activity data to compute off-network emissions
for idling of long-haul trucks from extended idling and auxiliary power unit process; and
rate-per-hour-ONI (RPHO) uses off-network idling hours activity data to compute
emissions for vehicles while idling off-network, (e.g., idling in a parking lot or unloading
freight). This is a new emission calculation which was added to the CTI version of
MOVES.
As described above, MOVES3.R1 was run for three scenarios: 2016, a 2055 reference case,
and a 2055 regulatory case. Scenario specific EV fractions were developed for each
representative county. MOVES was used to compute onroad emissions in California.
24
-------
SCC descriptions for onroad emissions
SCCs in the onroad sector follow the pattern 220FVV0RPP. where:
F = MOVES fuel type (1 for gasoline, 2 for diesel, 3 for CNG, 5 for E-85, and 9 for
electric)
VV = MOVES vehicle (aka source) type, see Table 3-1
R = MOVES road type (1 for off-network, 2 for rural restricted, 3 for rural unrestricted, 4
for urban restricted, 5 for urban unrestricted)
PP = SMOKE aggregate process. In the activity data, the last two digits of the SCC are
always 00, because activity data is process independent. MOVES separately tracks over a
dozen processes, but for computational reasons it is not practical to model all of these
processes separately within SMOKE-MOVES. Instead, "aggregate" processes are used in
SMOKE. To support this, the MOVES processes are mapped to SMOKE aggregate
processes according to Table 3-2. The MOVES3.R1 model includes a process, 92 that
corresponds to emissions from off-network idling (ONI).
Table 3-1 MOVES vehicle types
MOVES Vehicle Type
Description
11
Motorcycle
21
Passenger Car
31
Passenger Truck
32
Light Commercial Truck
41
Intercity Bus
42
Transit Bus
43
School Bus
51
Refuse Truck
52
Single Unit Short-haul Truck
53
Single Unit Long-haul Truck
54
Motor Home
61
Combination Short-haul Truck
62
Combination Long-haul Truck
25
-------
Table 3-2 SMOKE-MOVES aggregate processes
MOVES Process II)
Process description
SMOKE aggregate process
01
Running Exhaust
72
02
Start Exhaust
72
09
Brakewear
40
10
Tirewear
40
11
Evap Permeation
72
12
Evap Fuel Vapor Venting
72
13
Evap Fuel Leaks
72
15
Crankcase Running Exhaust
72
16
Crankcase Start Exhaust
72
17
Crankcase Extended Idle Exhaust
53
18
Refueling Displacement Vapor Loss
62
19
Refueling Spillage Loss
62
90
Extended Idle Exhaust
53
91
Auxiliary Power Exhaust
91
92
Off-network Idle Exhaust
92
3.3.1 Spatial Surrogates
Onroad county activity data were allocated to a national 12 km grid for air quality modeling
using spatial surrogates. For all processes other than the ONI process present in the MOVES3.R1
model, the spatial surrogates used to allocate onroad activity to the national 12km grid are the
same as in the 2016v2 platform and are described in the 2016v2 platform TSD. ONI and other
off-network activity data including VPOP and STARTS were spatially allocated using the
surrogates listed in Table 3-3.
Table 3-3 Spatial surrogates for on-network idling (ONI)
Source Type
Description
Spatial Surrogate
Description
11
Motorcycle
307
NLCD All Development
21
Passenger Car
307
NLCD All Development
31
Passenger Truck
307
NLCD All Development
32
Light Commercial Truck
308
NLCD Low + Med + High
41
Other Bus (non-transit, non-
school)
258
Other Bus Terminals
42
Transit Bus
259
Transit Bus Terminals
43
School Bus
506
Education
51
Refuse Truck
306
NLCD Med + High
52
Single Unit Short-haul Truck
306
NLCD Med + High
53
Single Unit Long-haul Truck
306
NLCD Med + High
54
Motor Home
304
NLCD Open + Low
61
Combination Short-haul Truck
306
NLCD Med + High
62
Combination Long-haul Truck
306
NLCD Med + High
3.3.2 Temporal Profiles
For on-network and hoteling emissions, VMT and hoteling activity were temporally allocated
from annual or monthly values to hourly and SMOKE was run for every day of the year. The
temporal profiles for VMT and hoteling activity are the same as in the 2016vl platform and are
26
-------
described in more detail in the 2016vl platform TSD. ONI monthly activity data were
temporally allocated to hourly values using a subset of the temporal profiles that are used to
temporally allocate VMT. VMT data were temporally allocated using temporal profiles which
vary by region (e.g., county, MSA), source type, and road type. ONI activity was developed for
each county and source type, but not road type. This means ONI cannot be temporalized in
exactly the same way as VMT. Instead, a subset of the VMT temporal profiles was selected to be
applied to ONI. Only temporal profiles for unrestricted road types were chosen to be used for
ONI, since off-network idling activity is assumed to better match the temporal pattern of
unrestricted road type driving, rather than on freeways. There are also different VMT temporal
profiles for urban road types and rural road types. ONI activity has no urban or rural designation,
and so within each county, we can only apply either a rural temporal profile or an urban temporal
profile. Therefore, we used the MOVES3.R1 county classification as either an urban county or a
rural county for the purposes of choosing appropriate temporal profiles for ONI in each county.23
In urban counties, ONI activity was temporally allocated using VMT profiles for urban
unrestricted roads, and in rural counties, For rural unrestricted roads, ONI activity was
temporally allocate using VMT profiles.
3.3.3 Chemical Speciation
Chemical speciation of onroad emissions is internal to MOVES except for brake and tire-wear
particulate matter (PM) speciation, which occurs in SMOKE. The emission factor tables from
MOVES include both unspeciated emissions totals in grams for criteria air pollutants (CAPs) and
hazardous air pollutants (HAPs), and speciated emissions totals for CB6 model species in moles
(or grams for PM). The speciation cross reference (GSREF) and speciation profile (GSPRO)
input files used by SMOKE-MOVES do not do any actual speciation. The GSREF file has no
function and only exists to prevent a SMOKE error. The GSPRO and mobile emissions process
and pollutant (MEPROC) files in SMOKE work in tandem to select which species and pollutants
to include in SMOKE outputs. The MEPROC includes all unspeciated pollutants, and the
GSPRO maps unspeciated pollutants to individual model species (e.g., brake wear PM2 5 to all
individual PM species). Model-ready emissions files will include all species in the GSPRO that
are mapped to one or more pollutants present in the MEPROC. Movesmrg reports include all of
those model species, plus all of the pollutants listed in the MEPROC.
3.3.4 Other Ancillary Files
SMOKE-MOVES requires several other types of ancillary files to prepare emissions for air
quality modeling:
Mobile county cross reference (MCXREF): Maps individual counties to representative
counties.
Mobile fuel month cross reference (MFMREF): Maps actual months to fuel months for
each representative county. May through September are mapped to the July fuel month,
and all other months to the January fuel month.
23 USEPA (2020). Population and Activity of On-road Vehicles in MOVES CTINPRM. Office of Transportation and
Air Quality. US Environmental Protection Agency. Ann Arbor, MI.
27
-------
MOVES lookup table list (MRCLIST): Lists emission factor table filenames for each
representative county.
Mobile emissions processes and pollutants (MEPROC): Lists which pollutants to include
in the SMOKE run.
Meteorological data for MOVES (METMOVES): Gridded daily minimum and maximum
temperature data. This file is created by the SMOKE program Met4moves and is used for
RatePerProfile (RPP) processing.
4 EGU Emissions Inventory Methodology
This section focuses on the approach and data sources used to develop gridded, hourly
emissions for the electrical generating unit (EGU) or "power plant" sector that are suitable for
input to an air quality model in terms of the format, grid resolution, and chemical species.
4.1 Integrated Planning Model (IPM)
The 2055 EGU emissions inventories were developed from the output of the Pre-IRA 2022
Reference Case run of the Integrated Planning Model (IPM). This version of IPM included EGU
fleet information, and rules and regulations that were final at the time the IPM version was
finalized, but not impacts due to the Inflation Reduction Act (IRA).24 IPM is a linear
programming model that accounts for variables and information such as energy demand, planned
unit retirements, and planned rules to project unit-level energy production and configurations.
4.2 IPM Inputs
The following specific rules and regulations are included in the IPM run:
The Revised Cross-State Air Pollution Rule (CSAPR) Update, a federal regulatory
measure affecting EGU emissions from 12 states to address transport under the 2008
National Ambient Air Quality Standards (NAAQS) for ozone.
The Standards of Performance for Greenhouse Gas Emissions from New, Modified,
and Reconstructed Stationary Sources: Electric Utility Generating Units through rate
limits.
The Mercury and Air Toxics Rule (MATS) finalized in 2011. MATS establishes
National Emissions Standards for Hazardous Air Pollutants (NESHAP) for the
"electric utility steam generating unit" source category.
Current and existing state regulations, including current and existing Renewable
Portfolio Standards and Clean Energy Standards as of the summer of 2021.
24 https://www.epa.gov/power-sector-modeling/pre-ira-2022-reference-case
28
-------
The latest actions EPA has taken to implement the Regional Haze Regulations and
Guidelines for Best Available Retrofit Technology (BART) Determinations Final
Rule. The regulation requires states to submit revised State Implementation Plans
(SIPs) that include (1) goals for improving visibility in Class I areas on the 20% worst
days and allowing no degradation on the 20% best days and (2) assessments and plans
for achieving Best Available Retrofit Technology (BART) emission targets for
sources placed in operation between 1962 and 1977. Since 2010, EPA has approved
SIPs or, in a few cases, put in place regional haze Federal Implementation Plans for
several states. The BART limits approved in these plans (as of summer 2020) that will
be in place for EGUs are represented in the Summer 2021 Reference Case.
California AB 32 C02 allowance price projections and the Regional Greenhouse Gas
Initiative (RGGI) rule.
Three non-air federal rules affecting EGUs: National Pollutant Discharge Elimination
System Final Regulations to Establish Requirements for Cooling Water Intake
Structures at Existing Facilities and Amend Requirements at Phase I Facilities,
Hazardous, and Solid Waste Management System; Disposal of Coal Combustion
Residuals from Electric Utilities; and the Effluent Limitation Guidelines and
Standards for the Steam Electric Power Generating Point Source Category.
IPM is run for a set of years, including 2050 and 2055. 2055 was the furthest out year in this
set of runs so in order to avoid end of timeframe issues we used the 2050 outputs and assumed
they are constant through 2055. All inputs, outputs and full documentation of EPA's IPM Pre-
IRA 2022 Reference Case and the associated NEEDS version is available on the power sector
modeling website (https://www.epa.gov/power-sector-modeling/pre-ira-2022-reference-case).
Some of the key parameters used in the IPM run are:
Demand: AEO 2021
Gas and Coal Market assumptions: updated as of December 2021
Cost and performance of fossil generation technologies: AEO 2021
Cost and performance of renewable energy generation technologies: NREL ATBG
2021 (mid-case)
Nuclear unit operational costs: AEO 2020 with some adjustments
Environmental rules and regulations (on-the-books): Revised CSAPR, MATS, BART,
CA AB 32, RGGI, various RPS and CES, non-air rules (Cooling Water Intake, ELC,
CCR), State Rules. This version does not include IRA
Financial assumptions: 2016-2020 data, reflects tax credit extensions from
Consolidated Appropriations Act of 2021
Transmission: updated data with build options
29
-------
Retrofits: carbon capture and sequestration option for CCs
Operating reserves (in select runs): Greater detail in representing interaction of load,
wind, and solar, ensuring availability of quick response of resources at higher levels of
RE penetration
Fleet: Summer 2021 reference case NEEDS
4.2.1 IPM Energy Demand Inputs
4.2.1.1 LMDVReference Case
IPM requires an electricity demand, and the default electricity demand for the version of IPM
used in this analysis is based on AEO2021, which does not include the full forecasted zero
emission vehicle (ZEV) adoption in its reference case. Relative to AEO2021, the LMDV
reference case has increased HD ZEV adoption and LD BEV adoption (to account for EPA's
Revised 2023 and Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions Standards
(LD GHG 2023-2026) final rule (86 FR 74434, December 30, 2021).25 Therefore, we developed
IPM input files specific to the demand of electric vehicles not captured by IPM's defaults, which
we call incremental demand input files.
We used the output of national MOVES3.R1 runs to develop the set of IPM incremental
demand input files for the LMDV reference scenario.26 Electricity demand was calculated using
the MOVES national modeling domain, with output by each type of day (i.e., for an average
weekday and weekend). IPM requires grid demand to be specified by day type, by each of IPM's
geographic regions, and by each hour of the day.
IPM requires grid demand to be geographically allocated by IPM region. We developed
regional allocation factors based on county-level CO2 emissions in the 2016v2 emissions
modeling platform.27'28 We used CO2 emissions as our basis for regional allocation because CO2
scales well with VMT while capturing differing fleet characteristics in different counties. IPM
includes a mapping of each county to an IPM region, which we used to aggregate county
allocation factors by IPM region.
Inputs to the IPM model include not only the anticipated electricity demand from plug-in
electric vehicles (PEVs), but also how that demand is distributed by time of day. This will
depend on when PEVs charge. We develop and apply charging profiles to reflect the share of
demand from PEV charging that we assume occurs each hour on weekdays and weekends.
25 Beardsley, Megan. 2023. "Updates to MOVES for the Mult-Pollutant Emissions Standards for Model Years 2027
and Later Light-Duty and Medium-Duty Vehicles." Memorandum to the Docket EPA-HQ-OAR-2022-0829.
26 US EPA, 2023. "Incremental Demand Input Files for the Multi-Pollutant Emissions Standards for Model Years
2027 and Later Light-Duty and Medium-Duty Vehicles." Memorandum to the Docket EPA-HQ-OAR-2022-0829.
27 The emissions modeling platform is a product of the National Emissions Inventory Collaborative consistent of
more than 245 employees of state and regional air agencies, EPA, and Federal Land Management agencies. It
includes a full suite of base year (2016) and projection year (2023 and 2028) emission inventories modeled using
EPA's full suite of emissions modeling tools, including MOVES, SMOKE, and CMAQ.
28 U.S. EPA. "2016v2 Platform". January 23, 2023. Available online: https://www.epa.gov/air-emissions-
modeling/2016v2-platform
30
-------
We source charging profiles for light-duty PEVs from the Electric Vehicle Infrastructure
Projection Tool (EVI-Pro) Lite developed by the National Renewable Energy Laboratory in
collaboration with others.29 EVI-Pro Lite allows users to generate charging profiles30 for
different scenarios based on the number31 and mix of vehicles, daily vehicle miles traveled,
ambient temperature, and availability and preference for certain charging types and charging
strategies. While full customization isn't possible in the tool, we generally tried to make
selections among the available options most consistent with our reference case where applicable,
using default selections for other variables.32 The resulting weekday and weekend charging
profiles33 are shown in Figure 4-1.
29 U.S. Department of Energy, Alternative Fuels Data Center. 2023. "Electric Vehicle Infrastructure Projection Tool
(EVI-Pro) Lite." Available at: https://afdc.energy.gov/evi-pro-lite/load-profile.
30 The tool asks users to select a city or urban area, which changes default selections for average ambient
temperature and vehicle miles traveled. Since we use the resulting profiles nationwide, we made selections (e.g.,
50ฐF) intended to reflect that.
31 We selected 30,000 PEVs (the highest default option available in the tool). However, it is important to note that
we do not use the charging profiles from EVI-Pro Lite to estimate the amount of PEV demand. Rather, we use the
profiles only to distribute our estimate of PEV demand for the Reference and Regulatory cases by hour of day.
32 We made the following selections: average daily miles traveled per vehicle: 35 miles; average ambient
temperature: 50ฐF; PEVs that are all-electric: 75% (highest available option); PEVs that are sedans: 50%; mix of
workplace charging: 20% Level 1 and 80% Level 2; access to home charging: 75%; mix of home charging: 50%
Level 1 and 50% Level 2; preference for home charging: 100%; home charging strategy: immediate - as fast as
possible; work charging strategy: immediate - as fast as possible.
33 Profiles from the EVI-Pro Lite tool are generated in 15-minute increments. Here we have aggregated to hourly
shares for use in IPM. We also normalized profiles such that the sum of hourly demand shares totals 100%.
31
-------
Weekday Weekend
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour
Figure 4-1: Charging profiles for light-duty PEV demand in the reference Case34
Heavy-duty vehicles comprise a broad spectrum of vehicle types and applications, and we
would expect charging patterns to vary accordingly. For this reason, we develop individual
charging profiles for seven vehicle categories: transit buses, school buses, other buses, refuse
trucks, single unit short-haul trucks, combination short-haul trucks, and motorhomes. We start
from data on vehicle soaks (or times when vehicles are not operating) in MOVES3.R1 for each
of the above categories. For our analysis, we considered only soak lengths that were greater than
or equal to 12 hours, using this as a proxy for when vehicles may be parked at a depot,
warehouse, or other off-shift location and may have an opportunity to charge. How long a
particular vehicle will take to charge will depend on a variety of factors including the vehicle's
daily electricity consumption and the power level of the charging equipment. The time that
charging occurs will also depend on the charging preferences of BEV owners or operators. Some
may choose to start charging as soon as the vehicle is parked, while others may delay charging to
accommodate other vehicles in a fleet, take advantage of time-of-use electricity rates, or for other
reasons. In developing national, fleetwide profiles, we made the simplifying assumption that
charging demand would be evenly distributed across the 12 hours before vehicles start daily
operation, i.e. when the soak periods end.35
As a final step, we weight the seven individual charging profiles by the relative share of
electricity demand for each vehicle category in MOVES3.R1 under the reference case. The
resulting aggregate weekday and weekend profiles are shown in Figure 4-2.
9.0%
8.0%
1 7.0%
a 6.0%
O
g 5.0%
^ 4.0%
c 3.0%
(U
ฃ 2.0%
Q.
1.0%
0.0%
34 We use light-duty charging profiles to distribute PEV demand for cars, passenger trucks, and light commercial
trucks (MOVES vehicle types 21, 31, and 32, see Table 3-1).
35 See "Heavy-duty BEV Charging Profiles.xlsx," available in the docket.
32
-------
Figure 4-2: Charging profiles for heavy-duty PEV demand in the reference case36
Finally, upstream emissions that would be incurred for fuel cell electric vehicles (FCEVs) due
to the production of hydrogen are not captured by MOVES. We made a simplifying assumption
that all hydrogen used to fuel FCEVs is produced via the electrolysis of water, and thus in this
analysis, all hydrogen production is represented as additional demand to EGUs and the emissions
are modeled using IPM. Hydrogen in the U.S. today is primarily produced via steam methane
reforming (SMR) largely in support of petroleum refining and ammonia production. New
transportation demand and economic incentives may shift how hydrogen is produced, and
electrolysis is a key mature technology for hydrogen production. The relative emissions impact
of hydrogen production via SMR versus electrolysis depends on the source of electricity
generation, and this varies significantly by region across the country. Electrolysis powered by
electricity from the grid on average in the U.S. may overestimate the upstream emissions impacts
that are attributable to HD FCEVs in the near-term.
We developed yearly scalar multipliers which were applied to MOVES FCEV energy
consumption to represent total grid demand from the hydrogen production necessary to support
the projected levels of FCEVs. First, we assumed hydrogen is produced by a series of
decentralized, grid-powered polymer electrolyte membrane (PEM) electrolyzer systems, each
with a hydrogen production capacity around 1,500 kilograms per day.37 38 Next, we assumed the
gaseous hydrogen is compressed and pre-cooled for delivery to vehicles using grid-powered
electrical equipment. Finally, we assumed a linear improvement between our estimated current
and future efficiency for hydrogen production. The linear interpolation is between current values
36 We use heavy-duty charging profiles to distribute demand for PEVs of MOVES vehicle type 41 and higher (see
Table 3-1).
37 This is based on assumptions from the Hydrogen Analysis Production (H2A) Model from the National Renewable
Energy Laboratory (NREL).
38 National Renewable Energy Laboratory (NREL). "H2A: Hydrogen Analysis Production Model: Version 3.2018".
Available online: https://www.nrel.gov/liydrogen/li2a-production-arcliive.html
33
-------
that start in 2025 and future values represented for 2055, assuming a period of diffusion for more
efficient electrolysis technology improvements to spread. The final scaling factors range from
1.748 in 2025 to 1.616 in 2055.
4.2.1.2 LMD VRegulatory Case
Similar to the LMDV reference case incremental demand input files described in Chapter
4.2.1.1, we used output from a national MOVES 3R.1 run to develop the set of IPM incremental
demand input files for the LMDV regulatory scenario.39
We use the same charging profiles for light-duty PEV demand in our regulatory case as in the
Reference case (see Figure 4-1). For heavy-duty charging profiles, we start from the same
charging profiles developed for each of the seven vehicle types, but apply weightings from
MOVES 3.R1 for the regulatory case. The resulting profiles are show in Figure 4-3.
Figure 4-3: Charging profiles for heavy-duty PEV demand in the regulatory case
4.3 Air Quality Model-Ready EGU inventory generation
The EGU emissions are calculated for the inventory using the output of the IPM model for the
forecast year. Units that are identified to have a primary fuel of landfill gas, fossil waste, non-
fossil waste, residual fuel oil, or distillate fuel oil may be missing emissions values for certain
pollutants in the generated inventory flat file. Units with missing emissions values are gapfilled
using projected base year values.
The projections are calculated using the ratio of the future year seasonal generation in the IPM
parsed file and the base year seasonal generation at each unit for each fuel type in the unit as
derived from the 2018 EIA923 tables and the 2018 NEI. New controls identified at a unit in the
IPM parsed file are accounted for with appropriate emissions reductions in the gapfill projection
values. When base year unit-level generation data cannot be obtained no gapfill value is
39 US EPA, 2023. "Incremental Demand Input Files for the Multi-Pollutant Emissions Standards for Model Years
2027 and Later Light-Duty and Medium-Duty Vehicles." Memorandum to the Docket EPA-HQ-OAR-2022-0829.
34
-------
calculated for that unit. Additionally, some units, such as landfill gas, may not be assigned a
valid SCC in the initial flat file. The SCCs for these units are updated based on the base year
SCC for the unit-fuel type. Combined cycle units produce some of their energy from process
steam that turns a steam turbine. The IPM model assigns a fraction of the total combined cycle
production to the steam turbine. When the emissions are calculated these steam units are
assigned emissions values that come from the combustion portion of the process. In the base year
NEI steam turbines are usually implicit to the total combined cycle unit. To achieve the proper
plume rise for the total combined cycle emissions, the stack parameters for the steam turbine
units are updated with the parameters from the combustion release point. Large EGUs in the
IPM-derived flat file inventory are associated with hourly CEMS data for NOx and SO2
emissions values in the base year. To maintain a temporal pattern consistent with the 2016 base
year, the NOx and SO2 values in the hourly CEMS inventories are projected to match the total
seasonal emissions values in the future years.
5 Petroleum Sector Emissions Inventory Methodology
This section focuses on the approach and data sources used to develop adjusted gridded,
hourly emissions for some of the sectors related to producing petroleum liquid fuels for mobile
sources. While the emission factors used to develop emissions for the reference and control
scenarios differed, the approach and data sources used to calculate emissions for both scenarios
were consistent.
Emission sources related to producing petroleum liquid fuels for mobile sources include
extracting, transporting, and storing crude oil, extracting, transporting, and storing natural gas,
and refining and transporting and storing finished fuels like gasoline and diesel. These sources
are described in the emissions modeling platform TSD in Section 2.1.2 (point oil and gas) and
2.2.4 (nonpoint oil and gas).40
More details on the modeling of the petroleum sector emissions are in the following
subsections and national emission summaries for key pollutants are provided in Section 6. The
docketed spreadsheet "2055 LMDV AQM petroleum adjustment factors final.xlsx" presents the
calculations described in this Section.
5.1 Refinery Emissions
5.1.1 Initial Projection of Refinery Emissions to 2050/2055
The starting point for developing refinery inventories for the illustrative air quality analysis
was the 2016v2 emissions modeling platform, which includes projection years of 2023, 2026,
and 2032.41 The 2032 refinery inventory from the 2016v2 emissions modeling platform was
40 https://www.epa.gov/air-emissions-modeling/2016v2-platform
41 https://www.epa.gov/air-emissions-modeling/2016v2-platform
35
-------
projected to 2050 using AEO2021 growth factors.42'43 We assumed no change in refinery
emissions between 2050 and 2055. The national total refinery inventory is presented in Table
5-1, and see docketed spreadsheet "2050 national refinery summary for OTAQ.xlsx".
Table 5-1 2016v2 Emissions Modeling Platform Refinery Inventory Projected to 2032 and 2055
Pollutant
Projected emissions in 2032 (tons/yr)
Projected emissions in 2055 (tons/yr)
NOx
69,330
71,525
pm25
19,919
19,514
so2
30,777
29,347
voc
64,019
58,675
5.1.2 Apportioning Total Refinery Emissions to Gasoline and Diesel Fuel Production
The projected 2055 refinery emissions inventory was apportioned to the refining of gasoline
and diesel fuel by assuming that refinery emissions are correlated with the refinery's energy
demand and accounting for the estimated refinery's fraction of input energy going towards
producing gasoline or diesel. The energy demand was calculated from modeled refinery energy
allocations.44 The original energy allocations were adjusted to ignore modeled outputs that are
not refinery products and converted from mass-based to volume-based percentages, see Table
5-2. Relative emission factors per unit of gasoline and diesel produced45 were then used to
generate the apportioned, pollutant-specific refinery inventory shown in Table 5-3.
Table 5-2 Refinery Energy Demand Percentages
Refinery Energy Demand
Gasoline
59.2
Diesel
6.1
Other
34.7
42 Specifically, a projection packet was prepared for 2032->2050 using AEO2021 (except for the categories where
we have been using AE02020 instead of AEO2021, e.g. cement) for refineries. AEO categories were mapped to
SCCs and SCC+NAICS combinations (with SCC+NAICS taking precedence if a mapping exists for the refinery
NAICS, which are 32411/324110) using the usual industrial source AEO-SCC and AEO-SCC-NAICS xrefs from
past platforms. Only refineries NAICS and SCCs which have refinery emissions were included when making the
packet, so the 2032-2050 packet is not something that can be used to project the entire ptnonipm sector. Each record
in the packet references the refineries NAICS so that it can be applied to the entire ptnonipm sector without
changing any non-refineries.
43 https://www.eia.gov/outlooks/archive/aeo21/
44 Table 1 in Wang el al. 2004. http://dx.doi.org/10.1065/lca2003.07.129
45 Wang. Michael. Elgowainv. Amgad. Lee. Uisung. Bafana. Adarsh. Bancrjcc. Sudhanva. Bcnavidcs. Pahola T..
Bobba. Pallavi. Burnham. Andrew. Cai. Hao. Gracida. Uliscs. Hawkins. Troy R.. Iyer. Rakcsh K.. Kelly. Jarod C.
Kim. Tacmin. Kingsbury. Kalhryn. Kwon. Hoyoung. Li. Yuan. Liu. Xinvu. Lu. Zifeng. On. Longwcn. Siddiquc.
Na/.ib. Sun. Pingping. Vyawaharc. Pradccp. Winjobi. Olumidc. Wu. May. Xu. Hui. Yoo. Eunji. Zaimcs. George G..
andZang, Guivan. Greenhouse gases, Regulated Emissions, and Energy use in Technologies Model ฎ (2021
Excel). Computer Software. USDOE Office of Energy Efficiency and Renewable Energy (EERE). 11 Oct. 2021.
Web. doi: 10.11578/GREET-E.\cel-2021/dc.20210902.1.
36
-------
Table 5-3 Refinery Inventory Apportioned to Gas, Diesel, and Other
Pollutant
Projected emissions in 2055 (tons/yr)
NOx
71,525
gasoline
43,595
diesel
3,984
other
23,946
pm25
19,514
gasoline
12,107
diesel
1,043
other
6,364
S02
29,347
gasoline
17,503
diesel
1,715
other
10,129
VOC
58,675
gasoline
33,454
diesel
3,423
other
21,798
5.1.3 Identifying Refinery Emissions to Adjust for Illustrative Air Quality Analysis
The refineries report from the 2016v2 emissions modeling platform was reviewed to identify
and remove from consideration any facilities that did not produce gasoline or diesel fuel for
onroad vehicles, see docketed spreadsheet "2016v2_platform_refineries_report.xlsx". The
resulting list had 114 refineries that produce onroad fuel, see docketed spreadsheet "Refineries to
Adjust LMDV and HDP3 NPRM.xlsx".
5.1.4 Illustrative Air Quality Modeling Scenarios and Associated Refined Fuel Demand
Refinery inventories were calculated for two scenarios, a reference scenario and a LMDV
regulatory scenario. Fuel demand volumes were used to adjust the original 2055 refinery
inventory (Table 5-3) and create an updated 2055 reference scenario refinery inventory (that
accounts for EPA's Revised 2023 and Later Model Year Light-Duty Vehicle Greenhouse Gas
Emissions Standards (LD GHG 2023-2026) final rule (86 FR 74434, December 30, 2021) and
additional heavy-duty ZEV adoption) and a 2055 refinery inventory that reflects the LMDV
regulatory scenario.
The onroad fuel consumed and total refined fuel supplied associated with the original refinery
inventory that is presented in Table 5-3 are from AEO202146 and are presented in Table 5-4.
Both volumes are projections for 2050 that we assume stay constant through 2055.
46 https://www.eia.gov/outlooks/aeo/data/browser/#/?id=46-AE02021®ion=0-
0&cases=ref2021&start=2019&end=2050&f=A&linechart=ref2021-dll3020a.2-46-AE02021&sourcekey=0
37
-------
Table 5-4 Fuel Volumes from AEO2021 Reference Case (billion gallons/yr)
Onroad Fuel Consumed3
Total Refined Fuel
Suppliedb
Gasoline
121.67
126.94
Diesel
36.73
54.91
a Onroad fuel consumed from Table 36 of AEO2021, with units converted from trillion BTU
b Total refined fuel supplied from Table 11 of AEO2021, with units converted from million barrels per day
The fuel demanded in 2055 by onroad vehicles (gallons of gasoline and gallons of diesel) in
the LMDV reference scenario and LMDV regulatory scenario was generated using MOVES, see
Table 5-5 and docketed spreadsheets "LMDV reference petroelumconsumption.xlsx" and
"LMDV regulatory petroelumconsumption.xlsx".
Table 5-5 2055 Onroad Fuel Demand for Illustrative Air Quality Analysis Scenarios from MOVES
(billion gallons/yr)
LMDV Reference Onroad
Fuel Demand
LMDV Regulatory
Onroad Fuel Demand
Gasoline
99.55
94.48
Diesel
38.21
37.05
There are methodological differences in how onroad fuel demand is calculated by MOVES
and by AEO. An adjustment factor to account for the difference between MOVES3.R1 and
AEO2021 reference was applied to the MOVES onroad fuel demand numbers to make them more
consistent with AEO Table 36, see Table 5-6.
Table 5-6 Factor to apply to MOVES fuel demand to make consistent with AEO fuel demand
Gasoline
0.995
Diesel
0.975
The reduction in onroad fuel demand due to the LMDV reference and LMDV regulatory
scenarios was generated separately for gasoline and for diesel by subtracting the fuel demands in
Table 5-5, once it was adjusted to account for AEO/MOVES methodological differences, from
the AEO2021 onroad fuel demand in Table 5-4.
5.1.5 Projected Change in U.S. Refinery Activity Related to Decreased Domestic Demand
It was necessary to project how the reduced onroad fuel demand associated with the LMDV
reference and LMDV regulatory scenarios would affect US refinery emissions since US refined
product demand is also satisfied by imports, not just production by US refineries. We projected
how the change in petroleum demand would affect US refinery production, averaged over the
time period 2027-2050, based on a comparison of two separate economic cases modeled by EIA
in AEO2021: the Low Economic Growth Case and the Reference Case.47 The AEO Low
Economic Growth Case estimates lower refined product demand than that of the AEO Reference
47 In this paragraph, Reference Case refers to the 2021 Annual Energy Outlook Reference Case, not the reference
case used elsewhere in this chapter to evaluate the impacts of the proposal and alternative.
38
-------
Case. Due to the reduced refined product demand, AEO estimates reduced US refinery
production and reduced imports of crude oil refined products. The two AEO cases project that,
for a volume of reduced gasoline or diesel fuel demand, 93% of the reduction is due to reduced
US refinery production while seven percent can be attributed to imports of refined products, see
attached spreadsheet "AEO 2021 Change in product demand on imports.xlsx". The reduced
domestic demand (gallons of gasoline or diesel) is multiplied by 93% to estimate the reduction in
domestically produced gasoline and diesel fuel.
5.1.6 Generation of Adjustment Factors
The reduced gallons of onroad gasoline and diesel that would be refined domestically was
subtracted from the total refined fuel supplied in Table 5-4 and used to create an adjustment
factor to be applied to the gasoline and diesel portions of the 2050 onroad refinery inventory, see
Table 5-7. The resulting emissions, associated with refining gasoline and diesel fuel only, are
presented in Table 5-8.
Table 5-7 Adjustment Factors to Apply to Apportioned 2050 Onroad Refinery Inventory
LMDV
Reference case
LMDV Regulatory
Case
Gasoline
0.84
0.80
Diesel
1.03
1.01
Table 5-8 Projected 2055 Emissions from Refineries Associated with Producing Gasoline and Diesel
Only
Scenario
NOx (tons/yr)
PM2.5 (tons/yr)
SO2 (tons/yr)
VOC (tons/yr)
LMDV
Reference
gasoline
36,527
10,144
14,666
28,030
diesel
4,084
1,069
1,757
3,509
LMDV
Regulatory
gasoline
34,909
9,695
14,016
26,789
diesel
4,006
1,049
1,724
3,441
A final adjustment factor, based on a ratio of the emissions associated with gas and diesel and
the total refinery emissions, was then calculated for each of the illustrative air quality analysis
scenarios, see Table 5-9.
Table 5-9 Adjustment Factor to Apply to 2050 Refinery Inventory
Scenario
NOx
PM2.5
SO2
VOC
LMDV Reference
0.90
0.90
0.90
0.91
LMDV Regulatory
0.88
0.88
0.88
0.89
We recognize that there is significant uncertainty in the impact reduced fuel demand has on
refinery emissions and that the refinery industry could respond in a different way to reduced
domestic demand for gasoline and diesel fuel. Because many US refineries experience lower
crude oil and natural gas prices than refineries elsewhere, they may continue their production of
39
-------
refined products, and instead import less refined product, or increase exports of US refined
products. If refineries do not decrease production in response to lower domestic demand, we
would project no emission reductions from refineries rather than the reductions associated with
the adjustment factors presented in Table 5-9. This method assumes that reduced demand would
be spread evenly across all refineries, as a scalar of emissions.
5.2 Crude production well and pipeline emissions
5.2.1 Initial Projection of Crude Production Well Site and Pipeline Inventories to 2050/2055
The estimated emission inventories for crude production wells and associated pipelines in the
2016v2 emissions modeling platform for the year 2032 are projected to the year 2050 using
AEO2021 reference case production forecast data in the year 2050 relative to that in the year
2032. The estimated crude production well and pipeline inventories were assumed to remain
constant from 2050 to 2055.
5.2.2 Illustrative Air Quality Modeling Scenarios and Associated Crude Demand
The initial 2050/2055 crude production well and pipeline inventories for 2055 needed to be
adjusted to reflect the impact of the LMDV reference and LMDV regulatory scenarios which
reduced the domestic demand for liquid fuel, see Table 5-5.48 The total reduced gallons of
refined fuel consumed for each scenario were adjusted to account for refinery efficiency, using
factors from Forman, et al (2014), to get reduced gallons of crude-equivalent finished fuel.49'50
Then the gallons of reduced crude-equivalent finished fuel were converted to gallons of reduced
crude using the energy density of crude and gasoline and diesel fuel.51
5.2.3 Projected Change in U.S. Crude Production Activity Related to Decreased Domestic
Demand
It was necessary to project how the reduced crude demand associated with the reference and
LMDV regulatory scenarios would affect US crude production well and pipeline emissions since
US crude demand is also satisfied by imports, not just domestic production. We projected how
the change in crude demand would affect US crude production, averaged over the time period
2027-2050, based on a comparison of two separate economic cases modeled by EIA in
AEO2021: the Low Economic Growth Case and the Reference Case.52 The AEO Low Economic
Growth Case estimates lower crude demand than that of the AEO Reference Case. Due to the
reduced crude demand, AEO estimates reduced US crude production and reduced imports of
crude oil. The two AEO cases project that, for a volume of reduced crude demand, 8% of the
reduction is due to reduced US crude production, see attached spreadsheet "AEO 2021 Change in
product demand on imports.xlsx". The reduced domestic demand (gallons of crude) is multiplied
by 8% to estimate the reduction in domestically produced crude.
48 The calculations are provided on docketed spreadsheet "2055 LMDV AQM petroleum adjustment factors
final.xlsx".
49 Forman et al, 2014 dx.doi.org/10.1021/es501035a
50 The conversion of crude oil to products may be more efficient that this value as this value represents the overall
refinery efficiency, not the efficiency for converting crude oil into products.
51 Energy densities came from EIA, https://www.eia.gov/energyexplained/units-and-calculators/
52 In this paragraph, Reference Case refers to the 2021 Annual Energy Outlook Reference Case, not the reference
case used elsewhere in this chapter to evaluate the impacts of the proposal and alternative.
40
-------
5.2.4 Generation of Crude Production Well and Pipeline Adjustment Factors
The reduced gallons of crude that would be domestically produced was subtracted from the
total crude produced in the AEO2021 reference case and used to create an adjustment factor to
be applied to the crude production well and pipeline inventories, see Table 5-10 and Equation 1.
Equation 1
AEO2021 reference case Bgal crude produced domestically in 2050
CrudeProdUCtion reduced crude produced domestically due to additional EV penetration
adjustment f actov AEO2021 reference case Bgal crude produced domestically in 2050
Table 5-10Adjustment Factor Applied to Crude Gas Well and Pipeline Inventories to Generate AQM
Scenarios
Scenario
Adjustment Factor
LMDV Reference
0.992
LMDV Regulatory Scenario
0.990
5.3 Natural gas production well and pipeline emissions
5.3.1 Initial Projection of Natural Gas Production Well Site and Pipeline Inventories to
2050/2055
Emission inventories for natural gas production wells and associated pipelines in the 2016v2
emissions modeling platform were projected from 2032 to 2050 using AEO2021 reference case
production forecast data. The 2050 natural gas well and pipeline emission inventories were
assumed to remain constant in the 2055 LMDV reference case.
5.3.2 Illustrative Air Quality Modeling Scenarios and Associated Natural Gas Demand
The 2050/2055 natural gas production well and pipeline inventories needed to be adjusted to
reflect the impact of the LMDV regulatory scenarios, which increased the domestic demand for
electricity, leading to more demand for natural gas.53 Natural gas use projections (trillion cubic
feet) from IPM are presented in Table 5-11, and AEO2021 reference case projections of the
percentage of produced natural gas going to EGUs, are presented in Table 5-12.
Table 5-11 IPM projections of Natural Gas Usage, trillion cubic feet, in 2050
Scenario
Natural Gas Usage (Tcf)
1 MDV Reference
11.61
I.MDY Regulatory
11.99
53 The calculations are provided on docketed spreadsheet "2055 LMDV AQM petroleum adjustment factors
final.xlsx".
41
-------
Table 5-12 Projections of Natural Gas, in trillion cubic feet, in 2050
AEO2021 Reference case, Table I t
Natural Gas (Tcf)
Total Dry Gas Production
4:
Consumption of Natural Gas by EGUs
12.13
5.3.3 Generation of Natural Gas Production Well and Pipeline Adjustment Factors
Based on the increased natural gas usage by EGUs indicated in Table 5-11, the LMDV
regulatory case has 1.03% more natural gas usage than the LMDV reference case. The AEO
projections from Table 5-12 indicate that 72% of the natural gas projected to be produced
domestically in 2050 goes towards EGUs. The growth factor applied to the LMDV reference
case natural gas well site and pipeline pump emission inventories to get the LMDV regulatory
scenario natural gas well and pipeline pump emission inventories was 1.01, see Equation 2.
Equation 2
Growth factor = (1-0.28) + (0.28*1.03)
42
-------
6 Inventory Summary Tables
This section includes summary tables of emission inventories used in the illustrative AQM
analysis and described in this document.
Table 6-1 Modeled PM2.5 Emissions Used in Illustrative AQ Modeling (tons)
PM2.5
2016 Base
Year
2055
Reference
Case
2055
LMDV
Regulatory
Case
Reference -
LMDV
Regulatory
Onroad Total (48 state)
104,005
35,737
26,833
8,904
Upstream Total (48 state)
167,795
92,358
94,533
-2,174
EGU
133,570
54,589
57,033
-2,444
Refinery
19,958
18,855
18,468
387
Crude Production Wells + Pipeline Pumps
3,393
4,824
4,814
10
Natural Gas Production Wells + Pipeline Pumps
10,875
14,091
14,217
-127
Total
271,800
128,096
121,365
6,730
Table 6-2 Modeled NOx Emissions Used in Illustrative AQ Modeling (tons)
NOx
2016 Base
Year
2055
Reference
Case
2055
LMDV
Regulatory
Case
Reference -
LMDV
Regulatory
Onroad Total (48 state)
3.444.356
729,707
683,096
46,610
Upstream Total (48 state)
2,067,563
920,948
933,078
-12,130
EGU
1,319,734
232,631
243,010 -10,379
Refinery
78,332
67,470
66,067 | 1.403
Crude Production Wells + Pipeline Pumps
161,605
221,243
220,800 | 442
Natural Gas Production Wells + Pipeline Pumps
507,891
399,604
403,201
-3,596
Total
5,511,919
1,650,655
1,616,174
34,481
Table 6-3 Modeled SO2 Emissions Used in Illustrative AQ Modeling (tons)
S02
2016 Base
Year
2055
Reference
Case
2055
LMDV
Regulatory
Case
Reference -
LMDV
Regulatory
Onroad Total (48 state)
1.342.456
4>>S,495
'J 2,534
105.961
Upstream Total (48 state)
2.415.830
1:1
-8.544
EGU
33,763
U.IK.5
-1.572
Refinery
67.853
5<>.lU<>
55.876
1.070
Crude Production Wells + Pipeline Pumps
1.229.169
1.455.550
I45:.<.^
2.911
Natural Gas Production Wells + Pipeline Pumps
1.085.046
i.:r.n:
i.::s.os(.
-10.954
Total
3.758.286
'.:<.().<,i<,
3.163.200
>7.417
43
-------
Table 6-4 Modeled VOC Emissions Used in Illustrative AQ Modeling (tons)
VOC
2016 Base
Year
2055
Reference
Case
2055
LMDV
Regulatory
Case
Reference -
LMDV
Regulatory
Onroad Total (48 state)
1.342.456
4^8,495
^>2,534
105.961
Upstream Total (48 state)
2.415.830
i:i
2.
-8.544
EGU
33,763
U.IK.5
-1.572
Refinery
(>~.X5 '
56.946
55.X~<>
1.070
Crude Production Wells + Pipeline Pumps
1.229.169
1.455.550
I45:.<.^
2.911
Natural Gas Production Wells + Pipeline Pumps
1.085.046
1.217.132
1.228.086
-10.954
Total
3.758.286
3.260.616
3.163.200
97.417
44
-------
Table 6-5 Modeled Onroad Emissions (short tons)
Pollutant
2016 Base
2055
2055
Base -
%
Reference
% Change
Year
Reference
LMDV
Reference
Change
-LMDV
Reference -
Case
Regulatory
2016 Base
Regulatory
LMDV
Case
- 2055
Regulatory
Reference
pm2.-,
104.005
35,737
26.833
68.268
-(.(."
8.904
-25ฐ/.,
Nth
3.444.356
729,707
683.096
2.714.649
-"9%
46.610
-(."
so.
:4.i>ix
" !8d
"4 12
17.638
-_1%
1 (>8
-2".,
YOC
1.342.456
498.495
392.534
84VK.I
-l.V',,
|05.'K.|
-21".,
( ()
18.986.388
5.770.997
4.4ซ.5. -(. 1
i v: 15.^1
1.305.636
-23%
Acrolein
l.'S"
226
n.
I4(>l
-84".,
50
-22%
Acetaldehyde
1 \454
4.207
v44^
>.24"
765
-18";,
Benzene
25.918
8.056
\~2I
17.862
2.336
-2Th
1,3-Butadiene
619
874
584
2.745
JO
Ethylbenzene
l').S(.5
7.880
(.454
11.985
-(.0%
l."2"
-22".,
Formaldehyde
18.286
3.433
2.945
14.853
-8 I
488
-14%
Naphthalene
2.380
323
28
2,057
-8(i",i
>5
-29%
Table 6-6 Nonroad Emissions (short tons)
Year
Pollutant
2016
2055
PM.-,
|Ol,.XO(,
56.535
NOx
1.115.733
671.502
SO'
1.467
1. i5(>
YOC
1.167.957
J(>5.o24
CO
11.384.766
15.262.152
Acrolein
2.080
5>>5
Acetaldehyde
11.177
5.531
Benzene
2'J. 1 '5
28.423
1,3-Butadiene
4.^1
4.890
Ethylbenzene
2o 44(>
17.258
Formaldehyde
28.435
1 V'l"
Naphthalene
1.944
1,413
Table 6-7 Fugitive Dust Emissions (short tons)
Year
Pollutant
2016
2055
PM.-,
880.002
916.040
45
-------
7 Air Quality Modeling Methodology
7.1 Air Quality Model - CMAQ
CMAQ is a non-proprietary computer model that simulates the formation and fate of
photochemical oxidants, primary and secondary PM concentrations, acid deposition, and air
toxics, over regional and urban spatial scales for given inputs of meteorological conditions and
emissions. CMAQ includes numerous science modules that simulate the emission, production,
decay, deposition and transport of organic and inorganic gas-phase and particle pollutants in the
atmosphere. The CMAQ model is a well-known and well-respected tool and has been used in
numerous national and international applications.54 The air quality modeling completed for the
rulemaking proposal used the 2016v2 platform with the most recent multi-pollutant CMAQ code
available at the time of air quality modeling (CMAQ version 5.3.2).55 The 2016 CMAQ runs
utilized the CB6r3 chemical mechanism (Carbon Bond with linearized halogen chemistry) for
gas-phase chemistry, and AER07 (aerosol model with non-volatile primary organic aerosol) for
aerosols. The CMAQ model is regularly peer-reviewed, CMAQ versions 5.2 and 5.3 beta were
most recently peer-reviewed in 2019 for the U.S. EPA.56
7.2 CMAQ Domain and Configuration
The CMAQ modeling analyses used a domain covering the continental United States, as
shown in Figure 7-1. This single domain covers the entire continental U.S. (CONUS) and large
portions of Canada and Mexico using 12 km x 12 km horizontal grid spacing. The 2016
simulation used a Lambert Conformal map projection centered at (-97, 40) with true latitudes at
33 and 45 degrees north. The model extends vertically from the surface to 50 millibars
(approximately 17,600 meters) using a sigma-pressure coordinate system with 35 vertical layers.
Table 7-1 provides some basic geographic information regarding the CMAQ domains and Table
7-2 provides the vertical layer structure for the CMAQ domain.
Table 7-1 Geographic elements of domains used in air quality modeling
CMAQ Modeling Configuration
Grid Resolution
12 km National Grid
Map Projection
Lambert Conformal Projection
Coordinate Center
97 deg W, 40 deg N
True Latitudes
33 deg N and 45 deg N
Dimensions
396 x 246 x 35
Vertical extent
35 Layers: Surface to 50 millibar level
(See Table 7-2)
54 More information available at: fattps://www. epa. gov/e maq.
55Model code for CMAQ v5.3.2 is available from the Community Modeling and Analysis System (CMAS) at:
http://www.cmascenter.org.
56 The Sixth External Peer Review of the Community Multiscale Air Quality (CMAQ) Modeling System. Available
online at: https://www.epa.gov/sites/production/files/2019-
08/documents/sixth_cmaq_peer_review_comment_report_6.19.19.pdf.
46
-------
Table 7-2 Vertical layer structure for CMAQ domain
Vertical
Layers
Sigma P
Pressure
(mb)
Approximate
Height (m)
35
0.0000
50.00
17,556
34
0.0500
97.50
14,780
33
0.1000
145.00
12,822
32
0.1500
192.50
11,282
31
0.2000
240.00
10,002
30
0.2500
287.50
8,901
29
0.3000
335.00
7,932
28
0.3500
382.50
7,064
27
0.4000
430.00
6,275
26
0.4500
477.50
5,553
25
0.5000
525.00
4,885
24
0.5500
572.50
4,264
23
0.6000
620.00
3,683
22
0.6500
667.50
3,136
21
0.7000
715.00
2,619
20
0.7400
753.00
2,226
19
0.7700
781.50
1,941
18
0.8000
810.00
1,665
17
0.8200
829.00
1,485
16
0.8400
848.00
1,308
15
0.8600
867.00
1,134
14
0.8800
886.00
964
13
0.9000
905.00
797
12
0.9100
914.50
714
11
0.9200
924.00
632
10
0.9300
933.50
551
9
0.9400
943.00
470
8
0.9500
952.50
390
7
0.9600
962.00
311
6
0.9700
971.50
232
5
0.9800
981.00
154
4
0.9850
985.75
115
3
0.9900
990.50
77
2
0.9950
995.25
38
47
-------
Vertical
Sigma P
Pressure
Approximate
Layers
(mb)
Height (m)
1
0.9975
997.63
19
0
1.0000
1000.00
0
Figure 7-1 Map of the CMAQ 12 km modeling domain (noted by the purple box)
7.3 CMAQ Inputs
The key inputs to the CMAQ model include emissions from anthropogenic and biogenic
sources, meteorological data, and initial and boundary conditions.
The emissions inputs are summarized in earlier sections of this document.
48
-------
The CMAQ meteorological input files were derived from simulations of the Weather
Research and Forecasting Model (WRF) version 3.8 for the entire 2016 year.57'58 The WRF
Model is a state-of-the-science mesoscale numerical weather prediction system developed for
both operational forecasting and atmospheric research applications.59 The meteorological outputs
from WRF were processed to create 12 km model-ready inputs for CMAQ using the
Meteorology-Chemistry Interface Processor (MCIP) version 4.3. These inputs included hourly
varying horizontal wind components (i.e., speed and direction), temperature, moisture, vertical
diffusion rates, and rainfall rates for each grid cell in each vertical layer.60
The boundary and initial species concentrations were provided by a northern hemispheric
CMAQ modeling platform for the year 2016.61'62 The hemispheric-scale platform uses a polar
stereographic projection at 108 km resolution to completely and continuously cover the northern
hemisphere for 2016. Meteorology is provided by WRF v3.8. Details on the emissions used for
hemispheric CMAQ can be found in the 2016 hemispheric emissions modeling platform TSD.63
The atmospheric processing (transformation and fate) was simulated by CMAQ (v5.2.1) using
the CB6r3 and the aerosol model with non-volatile primary organic carbon (AE6nvPOA). The
CMAQ model also included the on-line windblown dust emission sources (excluding agricultural
land), which are not always included in the regional platform but are important for large-scale
transport of dust.
7.4 CMAQ Model Performance Evaluation
An operational model performance evaluation for ozone, PM2.5 and its related speciated
components, specific air toxics (i.e., formaldehyde, acetaldehyde, benzene, and 1,3-butadiene),
as well as nitrate and sulfate deposition were conducted using 2016 state/local monitoring sites
data in order to estimate the ability of the CMAQ modeling system to replicate the base year
concentrations for the 12 km Continental United States domain (Section 7.2, Figure 7-1).
Included in this evaluation are statistical measures of model versus observed pairs that were
paired in space and time on a daily or weekly basis, depending on the sampling frequency of
each network (measured data). For certain time periods with missing ozone, PM2.5, air toxic, and
nitrate and sulfate deposition observations we excluded the CMAQ predictions from those time
periods in our calculations. It should be noted when pairing model and observed data that each
57 Skamarock, W.C., et al. (2008) A Description of the Advanced Research WRF Version 3.
https://opensky.ncar.edn/istandora/obiect/techiiotes:500.
58 USEPA (2019). Meteorological Model Performance for Annual 2016 Simulation WRF v3.8
https://www.epa.gov/scram/air-modeiing-reports-and-ionrnai-art.icles.
59 https://www.mmm.ucar.edu/models/wrf.
60 By un. D.W., Ching, J. K.S. (1999). Science algorithms of EPA Models-3 Community Multiscale Air Quality
(CMAQ) modeling system, EPA/6O0/R-99/O3O, Office of Research and Development. Please also see:
https://www.cniascenter.org/.
61 Henderson, B., et al. (2018) Hemispheric-CMAQ Application and Evaluation for 2016, Presented at 2019 CMAS
Conference, available https://cmascenter.Org/conference//20.l.8/slides/0850 henderson hemispheric-
cmaq application 2018.pptx.
62 Mathur, R., et al. (2017) Extending the Community Multiscale Air Quality (CMAQ) modeling system to
hemispheric scales: overview of process considerations and initial applications, Atmos. Chem. Phys., 17, 12449-
12474, https://doi.Org/10.5194/a -20.1.7.
63 USEPA (2019). Technical Support Document: Preparation of Emissions Inventories for the Version 7.1 2016
Hemispheric Emissions Modeling Platform. Qffice of Air Quality Planning and Standards.
49
-------
CMAQ concentration represents a grid-cell volume-averaged value, while the ambient network
measurements are made at specific locations.
Model performance statistics were calculated for several spatial scales and temporal periods
(statistics are defined in Section 7.4.2). Statistics were calculated for individual monitoring sites
and for each of the nine National Oceanic and Atmospheric Administration (NOAA) climate
regions of the 12-km U.S. modeling domain (Figure 7-2).64 The regions include the Northeast,
Ohio Valley, Upper Midwest, Southeast, South, Southwest, Northern Rockies, Northwest and
West65'66 as were originally identified in Karl and Koss (1984).67 The statistics for each site and
climate region were calculated by season ("winter" is defined as average of December, January,
and February; "spring" is defined as average of March, April, and May; "summer" is defined as
average of June, July, and August; and "fall" is defined as average of September, October, and
November). For 8-hour daily maximum ozone, we also calculated performance statistics by
region for the April through September ozone season.68 In addition to the performance statistics,
we prepared several graphical presentations of model performance. These graphical
presentations include regional maps which show the mean bias, mean error, normalized mean
bias and normalized mean error calculated for each season at individual monitoring sites.
64 NOAA, National Centers for Environmental Information scientists have identified nine climatically consistent
regions within the contiguous U.S., http://www.ncdc.noaa.gov/monitoring-references/maps/us-climate-regions.php.
65 The nine climate regions are defined by States where: Northeast includes CT, DE, ME, MA, MD, NH, NJ, NY,
PA, RI, and VT; Ohio Valley includes IL, IN, KY, MO, OH, TN, and WV; Upper Midwest includes IA, MI, MN,
and WI; Southeast includes AL, FL, GA, NC, SC, and VA; South includes AR, KS, LA, MS, OK, and TX;
Southwest includes AZ, CO, NM, and UT; Northern Rockies includes MT, NE, ND, SD, WY; Northwest includes
ID, OR, and WA; and West includes CA and NV.
66 Note most monitoring sites in the West region are located in California (see Figure 7-2), therefore statistics for the
West will be mostly representative of California ozone air quality.
67 Karl, T. R. and Koss, W. J., 1984: "Regional and National Monthly, Seasonal, and Annual Temperature Weighted
by Area, 1895-1983." Historical Climatology Series 4-3, National Climatic Data Center, Asheville, NC, 38 pp.
68 In calculating the ozone season statistics, we limited the data to those observed and predicted pairs with
observations that exceeded 60 ppb in order to focus on concentrations at the upper portion of the distribution of
values.
50
-------
U.S. Climate Regions
Figure 7-2 NOAA Nine Climate Regions (source: http://www.ncdc.noaa. gov/moniloring-
references/maps/us-climate-regions.php#references)
7.4.1 Monitoring Networks
The model evaluation for ozone was based upon comparisons of model predicted 8-hour daily
maximum concentrations to the corresponding ambient measurements for 2016 at monitoring
sites in the EPA Air Quality System (AQS) and the Clean Air Status and Trends Network
(CASTNet). The observed ozone data were measured and reported on an hourly basis. The PM2.5
evaluation focuses on concentrations of PM2.5 total mass and its components including sulfate
(SO4), nitrate (NO3), total nitrate (TNO3), ammonium (Nll i). elemental carbon (EC), and organic
carbon (OC) as well as wet deposition for nitrate and sulfate. The PM2.5 performance statistics
were calculated for each season (e.g., "winter" is defined as December, January, and February).
PM2.5 ambient measurements for 2016 were obtained from the following networks: Chemical
Speciation Network (CSN), Interagency Monitoring of PROtected Visual Environments
(IMPROVE), Clean Air Status and Trends Network (CASTNet), and National Acid Deposition
Program/National Trends (NADP/NTN). NADP/NTN collects and reports wet deposition
measurements as weekly average data. The pollutant species included in the evaluation for each
monitoring network are listed in Table 7-3. For PM2.5 species that are measured by more than
one network, we calculated separate sets of statistics for each network. The CSN and IMPROVE
networks provide 24-hour average concentrations on a 1 in every 3-day, or 1 in every 6-day
sampling cycle. The PM2.5 species data at CASTNet sites are weekly integrated samples. In this
analysis we use the term "urban sites" to refer to CSN sites; "suburban/rural sites" to refer to
CASTNet sites; and "rural sites" to refer to IMPROVE sites.
51
-------
Table 7-3 PM2.5 monitoring networks and pollutants species included in the CMAQ performance
evaluation
Ambient
Monitoring
Networks
Particulate
Species
Wet Deposition
Species
PM2 5 Mass
so4
N03
TN03a
EC
oc
nh4
S04
N03
IMPROVE
X
X
X
X
X
CASTNet
X
X
X
CSN
X
X
X
X
X
X
NADP
X
X
a TNO3 = (N03 + HNCb)
The air toxics evaluation focuses on specific species relevant this proposed rulemaking, i.e.,
formaldehyde, acetaldehyde, benzene, and 1,3-butadiene. Similar to the PM2.5 evaluation, the air
toxics performance statistics were calculated for each season to estimate the ability of the CMAQ
modeling system to replicate the base year concentrations for the 12 km continental U.S. domain.
Toxic measurements for 2016 were obtained from the air toxics archive,
https://www.epa.gov/amtic/amtic-air-toxics-data-ambient-monitoring-archive. While most of the
data in the archive are from the AQS database including the National Air Toxics Trends Stations
(NATTS), additional data (e.g., special studies) are included in the archive but not reported in the
AQS.
7.4.2 Model Performance Statistics
The Atmospheric Model Evaluation Tool (AMET) was used to conduct the evaluation
described in this document.69 There are various statistical metrics available and used by the
science community for model performance evaluation. For this evaluation of the 2016 CMAQ
modeling platform, we have selected the mean bias, mean error, normalized mean bias, and
normalized mean error to characterize model performance, statistics which are consistent with
the recommendations in Simon et al. (2012)70 and the draft photochemical modeling guidance.71
Mean bias (MB) is used as average of the difference (predicted - observed) divided by the
total number of replicates (n). Mean bias is given in units of ppb and is defined as:
1
MB = ~Hi(P ~ 0) , where P = predicted and O = observed concentrations
69 Appel, K.W., Gilliam, R.C., Davis, N., Zubrow, A., and Howard, S.C.: Overview of the Atmospheric Model
Evaluation Tool (AMET) vl.l for evaluating meteorological and air quality models, Environ. Modell. Softw.,26, 4,
434-443, 2011. (http://www.cmascenter.org/).
70
Simon, H., Baker, K., Phillips, S., 2012: Compilation and interpretation of photochemical model performance
statistics published between 2006 and 2012. Atmospheric Environment 61, 124-139.
71 U.S. Environmental Protection Agency (US EPA), Draft Modeling Guidance for Demonstrating Attainment of
Air Quality Goals for Ozone, PM25, and Regional Haze. December 2014, U.S. EPA, Research Triangle Park, NC,
27711.
52
-------
Mean error (ME) calculates the absolute value of the difference (predicted - observed)
divided by the total number of replicates (n). Mean error is given in units of ppb and is defined
as:
ME = iฃI|P-0|
Normalized mean bias (NMB) is used as a normalization to facilitate a range of concentration
magnitudes. This statistic averages the difference (predicted - observed) over the sum of
observed values. NMB is a useful model performance indicator because it avoids over inflating
the observed range of values, especially at low concentrations. Normalized mean bias is given in
percentage units and is defined as:
f(P-O)
NMB = ^ *100
n
X(o)
Normalized mean error (NME) is also similar to NMB, where the performance statistic is
used as a normalization of the mean error. NME calculates the absolute value of the difference
(predicted - observed) over the sum of observed values. Normalized mean error is given in
percentage units and is defined as:
i\p-o\
NME = ^ *100
n
t(o)
The "acceptability" of model performance was judged by comparing our CMAQ 2016
performance results in light of the range of performance found in recent regional ozone and
53
-------
PM2.5 model applications.72'73'74'75'76'77'78'79'80'81 These other modeling studies represent a wide
range of modeling analyses that cover various models, model configurations, domains, years
and/or episodes, chemical mechanisms, and aerosol modules. Overall, the ozone and PM2.5
model performance results for the 2016 CMAQ simulations are within the range found in other
recent peer-reviewed and regulatory applications. The model performance results, as described in
this document, demonstrate that that our applications of CMAQ using this 2016 modeling
platform provide a scientifically credible approach for assessing ozone and PM2.5 concentrations
for the purposes of this proposed rulemaking.
7.4.3 Evaluation for 8-hour Daily Maximum Ozone
The 8-hour ozone model performance bias and error statistics for each climate region, for
each season defined above and for each monitor network (AQS and CASTNet) are provided in
Table 7-4. As indicated by the statistics in Table 7-4, bias and error for 8-hour daily maximum
ozone are relatively low in each climate region. Spatial plots of the mean bias and error as well
as the normalized mean bias and error for individual monitors are shown in Figure 7-3 through
Figure 7-6. The statistics shown in these figures were calculated over the ozone season using
data pairs on days with observed 8-hour ozone of > 60 ppb. Figure 7-3 shows MB for 8-hour
ozone > 60 ppb during the ozone season in the range of ฑ15 ppb at the majority of ozone AQS
and CASTNet measurement sites. At both AQS and CASTNet sites, NMB is within the range of
72 National Research Council (NRC), 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations, Washington, DC: National Academies Press.
73 Appel, K.W., Roselle, S.J., Gilliam, R.C., and Pleim, J.E, 2010: Sensitivity of the Community Multiscale Air
Quality (CMAQ) model v4.7 results for the eastern United States to MM5 and WRF meteorological drivers.
Geoscientific Model Development, 3, 169-188.
74 Foley, K.M., Roselle, S.J., Appel, K.W., Bhave, P.V., Pleim, J.E., Otte, T.L., Mathur, R., Sarwar, G., Young,
J.O., Gilliam, R.C., Nolte, C.G., Kelly, J.T., Gilliland, A.B., and Bash, J.O., 2010: Incremental testing of the
Community multiscale air quality (CMAQ) modeling system version 4.7. Geoscientific Model Development, 3, 205-
226.
75 Hogrefe, G., Civeroio, K.L., Hao, W., Ku, J-Y., Zalewsky, E.E., and Sistla, G., Rethinking the Assessment of
Photochemical Modeling Systems in Air Quality Planning Applications. Air & Waste Management Assoc.,
58:1086-1099, 2008.
76 Phillips, S., K. Wang, C. Jang, N. Possiel, M. Strum, T. Fox, 2007. Evaluation of 2002 Multi-pollutant Platform:
Air Toxics, Ozone, and Particulate Matter, 7th Annual CMAS Conference, Chapel Hill, NC, October 6-8, 2008.
(http://www.cmascenter.org/conference/2008/agenda.cfm).
77 Simon, H., Baker, K.R., and Phillips, S., 2012. Compilation and interpretation of photochemical model
performance statistics published between 2006 and 2012. Atmospheric Environment 61, 124-139.
http://dx.doi.Org/10.1016/i.atmosenv.2012.07.012.
78 Tesche, T.W., Morris, R., Tonnesen, G., McNally, D., Boylan, J., Brewer, P., 2006. CMAQ/CAMx annual 2002
performance evaluation over the eastern United States. Atmospheric Environment 40, 4906-4919.
79 U.S. Environmental Protection Agency; Technical Support Document for the Final Clean Air Interstate Rule: Air
Quality Modeling; Office of Air Quality Planning and Standards; RTP, NC; March 2005 (CAIR Docket OAR-2005-
0053-2149).
80 U.S. Environmental Protection Agency, Proposal to Designate an Emissions Control Area for Nitrogen Oxides,
Sulfur Oxides, and Particulate Matter: Technical Support Document. EPA-420-R-007, 329pp., 2009.
(http://www.era.gov/otaq/regs/nonroad/marine/ci/420r09007.pdf).
81 U.S. Environmental Protection Agency, 2010, Renewable Fuel Standard Program (RFS2) Regulatory Impact
Analysis. EPA-420-R-10-006. February 2010. Sections 3.4.2.1.2 and 3.4.3.3. Docket EPA-HQ-OAR-2009-0472-
11332.
54
-------
ฑ30 percent (Figure 7-5). Mean error for 8-hour maximum ozone > 60 ppb, as seen from Figure
7-4, is 20 ppb or less at most of the sites across the modeling domain.
Table 7-4 Daily Maximum 8-hour Ozone Performance Statistics by Climate Region, by Season, and by
Monitoring Network for the 2016 CMAQ Model Simulation
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ppb)
(ppb)
(%)
(%)
Winter
11,462
-2.5
4.7
-7.7
14.7
AQS
Spring
15,692
-6.7
7.7
-15.0
17.3
Summer
16,686
-0.9
6.2
-1.9
13.7
Northeast
Fall
13,780
-0.1
4.7
-0.4
13.7
Winter
1,283
-3.2
4.8
-9.4
13.9
CASTNet
Spring
1,336
-7.2
8.0
-16.0
17.7
Summer
1,315
-1.7
5.8
-3.9
13.6
Fall
1,306
0.0
4.7
0.1
13.8
Winter
4,178
-0.5
4.4
-1.7
14.7
AQS
Spring
15,498
-4.5
6.5
9.9
14.2
Summer
20,495
0.1
6.2
0.3
13.7
Ohio Valley
Fall
14,025
0.1
4.9
0.3
12.6
Winter
1,574
-1.1
4.3
-3.4
13.3
CASTNet
Spring
1,600
-5.4
6.9
-11.6
14.9
Summer
1,551
-0.7
5.8
-1.6
13.4
Fall
1,528
-1.8
5.1
-4.5
12.8
Winter
1,719
-1.1
4.5
-3.5
14.5
AQS
Spring
6,892
-6.5
7.7
-14.4
17.3
Summer
9,742
-1.6
5.9
-3.7
14.0
Upper
Fall
6,050
1.4
4.3
4.3
13.5
Midwest
Winter
435
-2.2
4.5
-6.7
13.4
CASTNet
Spring
434
-7.9
8.5
-17.6
18.9
Summer
412
-3.5
5.8
-8.6
13.9
Fall
426
-0.4
4.2
-1.3
13.2
Southeast
AQS
Winter
7,128
-3.7
5.4
-10.3
15.1
55
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ppb)
(ppb)
(%)
(%)
Spring
14,569
-5.4
6.9
-11.6
14.9
Summer
15,845
0.6
5.3
1.5
13.3
Fall
12,583
-0.9
4.7
-2.3
11.6
Winter
887
-4.1
5.4
-11.1
14.4
CASTNet
Spring
947
-7.1
8.0
-14.9
16.6
Summer
926
-0.3
5.1
-0.8
13.1
Fall
928
-2.4
5.3
-5.7
12.8
Winter
11,432
-3.1
5.5
-9.1
16.5
AQS
Spring
13,093
-3.7
7.0
-8.5
15.9
Summer
12,829
0.2
6.0
0.6
15.5
South
Fall
12,443
-1.3
5.0
-3.4
12.7
Winter
523
-3.4
5.2
-9.3
14.4
CASTNet
Spring
532
-5.0
7.2
-11.0
15.9
Summer
508
-1.6
6.5
-4.2
16.7
Fall
528
-1.3
4.4
-3.5
11.4
Winter
9,990
-4.4
6.2
-11.2
15.8
AQS
Spring
11,381
-7.8
8.5
-15.2
16.6
Summer
12,027
-6.3
7.8
-11.7
14.5
Southwest
Fall
11,097
-1.7
4.4
-4.2
10.7
Winter
757
-7.0
7.3
-15.6
16.3
CASTNet
Spring
810
00
00
1
8.5
-16.8
17.5
Summer
812
-6.4
7.3
-12.0
13.7
Fall
791
-3.1
4.3
-7.0
9.9
Winter
4,719
-2.6
5.1
-7.0
13.6
AQS
Spring
4,975
-5.8
6.9
-13.2
15.8
Northern
Rockies
Summer
5,054
-3.9
5.7
-8.2
12.4
Fall
4,876
-0.1
4.4
-0.2
12.9
CASTNet
Winter
667
-3.8
6.2
-9.7
15.7
Spring
696
-7.6
8.2
-16.5
17.7
56
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ppb)
(ppb)
(%)
(%)
Summer
693
-5.1
6.2
-10.5
12.8
Fall
605
-1.4
4.9
-3.6
13.2
Winter
677
-3.3
6.1
-10.2
18.7
AQS
Spring
1,288
-6.5
8.2
-16.1
20.4
Summer
2,444
-0.9
6.4
-2.4
16.9
Northwest
Fall
1,236
1.0
5.3
3.1
16.8
Winter
-
-
-
-
-
CASTNet
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Fall
-
-
-
-
-
Winter
14,539
-3.9
6.1
-11.4
17.6
AQS
Spring
17,191
-7.9
8.6
-17.3
18.6
Summer
18,132
-6.5
9.0
-12.3
17.0
West
Fall
16,211
-4.6
6.8
-10.6
15.8
Winter
506
-3.7
5.3
-9.3
13.5
CASTNet
Spring
519
-8.3
8.6
-17.3
17.8
Summer
526
-10.9
11.6
-18.1
19.1
Fall
530
-5.6
6.6
-11.9
14.2
57
-------
03_8hrmax MB (ppb) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for 20160401 to 20160930
CASTNET Daily AQS Daily
Figure 7-3 Mean Bias (ppb) of 8-hour daily maximum ozone greater than 60 ppb over the period
April-September 2016 at AQS and CASTNet monitoring sites in the modeling domain
for 20160401 to 20160930
03 8hrmax ME
a CASTNET Daily AQS Daily
Figure 7-4 Mean Error (ppb) of 8-hour daily maximum ozone greater than 60 ppb over the period
April-September 2016 at AQS and CASTNet monitoring sites in the modeling domain
58
-------
Figure 7-5 Normalized Mean Bias (%) of 8-hour daily maximum ozone greater than 60 ppb over the
period April-September AQS and CASTNet 2016 at monitoring sites in the modeling domain
Figure 7-6 Normalized Mean Error (%) of 8-hour daily maximum ozone greater than 60 ppb over the
period April-September AQS and CASTNet 2016 at monitoring sites in the modeling domain
59
-------
7.4.4 Seasonal Evaluation of PM2.5 Component Species
The evaluation of 2016 model predictions for PM2.5 covers the performance for the individual
PM2.5 component species (i.e., sulfate, nitrate, organic carbon, elemental carbon, and
ammonium). Performance results are provided for each PM2.5 species. As indicated above, for
each species we present tabular summaries of bias and error statistics by climate region for each
season. These statistics are based on the set of observed-predicted pairs of data for the particular
quarter at monitoring sites within the nine NOAA climate regions. Separate statistics are
provided for each monitoring network, as applicable for the particular species measured. For
sulfate and nitrate we also provide a more refined temporal and spatial analysis of model
performance that includes spatial maps which show the mean bias and error and the normalized
mean bias and error by site, aggregated by season.
7.4.4.1 Seasonal Evaluation for Sulfate
The model performance bias and error statistics for sulfate for each climate region and each
season by monitor network are provided in Table 7-5. Spatial plots of the normalized mean bias
and error by season for individual monitors are shown in Figure 7-7 through Figure 7-22.
Table 7-5 Sulfate Performance Statistics by Climate Region, by Season, and by Monitoring Network
for the 2016 CMAQ Model Simulation
Climate
Region
Monitor
Network
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Winter
431
-0.1
0.2
-7.0
31.4
IMPROVE
Spring
477
0.0
0.2
12.5
39.2
Summer
486
-0.1
0.2
-9.8
30.5
Fall
456
0.0
0.2
-1.0
29.7
Winter
716
0.1
0.5
5.9
45.9
Northeast
CSN
Spring
768
0.1
0.3
00
00
36.7
Summer
782
-0.2
0.2
-21.2
23.2
Fall
736
0.1
0.3
13.3
39.0
Winter
221
-0.2
0.2
-23.4
24.9
CASTNet
Spring
242
-0.2
0.2
-17.3
20.3
Summer
252
-0.2
0.2
-21.2
23.2
Fall
242
-0.1
0.2
-16.9
21.4
Winter
220
-0.2
0.3
-18.0
31.2
Ohio Valley
IMPROVE
Spring
244
-0.2
0.3
-17.8
28.5
Summer
239
-0.4
0.5
-27.3
36.6
Fall
227
-0.3
0.4
-17.4
27.4
60
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Winter
546
-0.2
0.3
-14.4
35.3
CSN
Spring
562
0.0
0.4
2.5
34.6
Summer
553
-0.1
0.5
-6.9
30.5
Fall
541
0.0
0.4
0.2
31.5
Winter
212
-0.4
0.4
-29.9
31.0
CASTNet
Spring
228
-0.3
0.3
-23.4
25.1
Summer
224
-0.4
0.4
-24.7
27.1
Fall
226
-0.3
0.3
-23.8
24.3
Winter
200
-0.1
0.2
-6.5
27.7
IMPROVE
Spring
208
0.0
0.2
-0.4
30.5
Summer
210
-0.1
0.2
-12.0
30.0
Fall
215
0.0
0.2
-0.4
35.3
Winter
326
0.1
0.3
8.3
34.7
Upper
Midwest
CSN
Spring
354
0.2
0.4
39.2
22.9
Summer
314
0.0
0.3
4.6
33.0
Fall
310
0.2
0.4
33.9
49.4
Winter
59
-0.2
0.3
-22.4
25.9
CASTNet
Spring
63
-0.1
0.1
-7.9
13.7
Summer
63
-0.1
0.1
-12.5
17.7
Fall
57
-0.1
0.1
-13.3
18.5
Winter
342
-0.1
0.3
-10.2
-6.0
IMPROVE
Spring
379
-0.3
0.4
-21.7
30.7
Summer
394
-0.4
0.5
-35.1
40.8
Fall
366
-0.2
0.3
-16.3
27.2
Southeast
Winter
512
0.1
0.3
11.4
36.5
CSN
Spring
551
0.0
0.4
-3.9
31.3
Summer
523
-0.2
0.4
-21.1
34.3
Fall
505
0.0
0.3
1.3
26.4
CASTNet
Winter
150
-0.3
0.4
-29.9
32.2
61
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Spring
164
-0.5
0.5
-33.6
34.3
Summer
164
-0.6
0.6
-40.9
41.2
Fall
154
-0.4
0.4
-29.8
31.1
Winter
240
0.0
0.3
3.7
33.8
IMPROVE
Spring
273
-0.2
0.4
-16.0
37.6
Summer
252
-0.6
0.7
-44.1
48.1
Fall
264
-0.2
0.4
-17.7
32.5
Winter
326
0.0
0.4
3.9
41.4
South
CSN
Spring
351
-0.3
0.7
-21.9
45.4
Summer
336
-0.6
0.7
-36.8
46.4
Fall
329
-0.2
0.5
-13.7
35.0
Winter
92
-0.3
0.3
-26.3
27.7
CASTNet
Spring
102
-0.5
0.5
-33.0
34.2
Summer
96
-0.8
0.8
-48.7
48.9
Fall
102
-0.4
0.4
-28.9
29.9
Winter
910
0.1
0.2
55.6
82.1
IMPROVE
Spring
991
0.2
0.3
59.5
69.9
Summer
985
-0.3
0.3
-39.4
50.3
Fall
962
-0.1
0.2
-12.2
43.4
Winter
246
-0.1
0.4
-9.3
71.2
Southwest
CSN
Spring
255
0.3
0.3
68.1
75.0
Summer
250
-0.3
0.4
-38.2
49.4
Fall
255
0.3
0.3
60.4
68.5
Winter
101
0.1
0.1
37.5
60.4
CASTNet
Spring
115
0.2
0.2
40.9
44.9
Summer
114
-0.2
0.2
-35.9
40.9
Fall
115
-0.1
0.2
-17.6
36.0
Northern
Rockies
IMPROVE
Winter
542
0.1
0.2
31.3
65.7
Spring
573
0.1
0.2
34.5
53.1
62
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Summer
603
0.0
0.2
10.1
42.0
Fall
574
0.1
0.2
17.5
48.5
Winter
139
0.3
0.1
19.3
52.2
CSN
Spring
151
0.1
0.2
21.5
45.7
Summer
153
0.0
0.2
3.0
37.0
Fall
136
0.1
0.2
15.5
41.2
Winter
126
0.0
0.1
-10.0
38.8
CASTNet
Spring
139
0.1
0.1
12.1
26.9
Summer
138
-0.1
0.1
-11.8
23.8
Fall
129
0.0
0.1
-2.5
28.1
Winter
427
0.1
0.1
76.4
97.9
IMPROVE
Spring
505
0.2
0.2
60.1
69.6
Summer
519
0.0
0.2
12.1
51.1
Fall
499
0.1
0.2
34.7
70.5
Winter
156
0.3
0.4
>100
>100
Northwest
CSN
Spring
146
0.3
0.4
85.8
89.9
Summer
166
0.1
0.3
15.3
52.6
Fall
161
0.3
0.3
74.2
95.8
Winter
-
-
-
-
-
CASTNet
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Fall
-
-
-
-
-
Winter
565
0.2
0.2
79.0
>100
IMPROVE
Spring
608
0.1
0.3
24.1
57.3
Summer
603
-0.2
0.3
-31.2
48.3
West
Fall
576
0.0
0.2
-6.9
48.5
Winter
340
0.1
0.3
23.2
66.5
CSN
Spring
352
0.0
0.4
-5.5
48.5
Summer
349
-0.7
0.8
-50.6
55.6
63
-------
Climate
Region
Monitor
Network
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Fall
330
-0.2
0.4
-23.5
45.6
CASTNet
Winter
69
0.1
0.2
31.2
65.4
Spring
73
-0.1
0.2
-11.9
37.7
Summer
75
-0.5
0.5
-49.4
51.9
Fall
76
-0.2
0.3
-30.1
42.6
for December to February 2016
IMPROVE ฑ CSN ฆ CASTNET Weekly
Figure 7-7 Mean Bias (ug/m3) of sulfate during winter 2016 at monitoring sites in the modeling
domain
64
-------
S04 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE a CSN ฆ CASTNET Weekly
Figure 7-8 Mean Error (ug/m3) of sulfate during winter 2016 at monitoring sites in the modeling
domain
S04 NMB (%) for run CMAQ_2016fm cb6r3_ae7nvpoa_12US2 for March to May 2016
IMPROVE a CSN ฆ CASTNET Weekly
Figure 7-9 Normalized Mean Bias (%) of sulfate during winter 2016 at monitoring sites in the
modeling domain
65
-------
> 100
90
80
70
60
50
40
30
20
10
0
Figure 7-10 Normalized Mean Error (%) of sulfate during winter 2016 at monitoring sites in the
modeling domain
units - uglirS
coverage limit = 75%
>2
1.5
1
0.5
0
-0.5
-1
-1.5
< -2
IMPROVE a CSN ฆ CASTNET Weekly
Figure 7-11 Mean Bias (ug/m3) of sulfate during spring 2016 at monitoring sites in the modeling
domain
66
S04 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
IMPROVE a CSN ฆ CASTNET Weekly
S04MB
for March to May 2016
-------
S04
2016
units - ug/m3
coverage limit = 75%
1.2
1
0.8
0.6
0.4
0.2
0
CASTNET Weekly
Figure 7-12 Mean Error (ug/m3) of sulfate during spring 2016 at monitoring sites in the modeling
domain
units - %
coverage limit = 75%
IMPROVE * CSN ฆ CASTNET Weekly
Figure 7-13 Normalized Mean Bias (%) of sulfate during spring 2016 at monitoring sites in the
modeling domain
S04 NMB (%) for run CMAQ 2016fm cb6r3 ae7nvpoa 12US2 for March to May 2016
67
-------
S04 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
coverage limit = 75%
100
IMPROVE a CSN
CASTNET Weekly
Figure 7-14 Normalized Mean Error (%) of sulfate during spring 2016 at monitoring sites in the
modeling domain
units - ug/m3
coverage limit = 76%
IMPROVE * CSN ฆ CASTNET Weekly
Figure 7-15 Mean Bias (ug/m3) of sulfate during summer 2016 at monitoring sites in the modeling
domain
68
2016
-------
S04 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
CASTNET Weekly
Figure 7-16 Mean Error (ug/m3) of sulfate during summer 2016 at monitoring sites in the modeling
domain
units -
%
coverag
a limit = 75%
1
>100
80
_
60
40
-
-
-20
-
-40
-
-60
-
-80
1
< -100
IMPROVE * CSN ฆ CASTNET Weekly
Figure 7-17 Normalized Mean Bias (%) of sulfate during summer 2016 at monitoring sites in the
modeling domain
S04 NMB (%) for run CMAQ 2016fm cb6r3 ae7nvpoa 12US2 for June to August 2016
69
-------
S04 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
coverage limit = 75%
100
IMPROVE a CSN
CASTNET Weekly
Figure 7-18 Normalized Mean Error (%) of sulfate during summer 2016 at monitoring sites in the
modeling domain
units - ug/m3
coverage limit = 75%
B
>2
1.5
1
0.5
0
-0.5
-1
-1.5
< -2
IMPROVE
CSN
CASTNET Weekly
Figure 7-19 Mean Bias (ug/m3) of sulfate during fall 2016 at monitoring sites in the modeling domain
70
S04MB
September to November 2016
-------
S04ME
for September to November 2016
units - ug/m3
coverage limit = 75%
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
CASTNET Weekly
Figure 7-20 Mean Error (ug/m3) of sulfate during fall 2016 at monitoring sites in the modeling domain
units - %
coverage limit = 75%
IMPROVE * CSN ฆ CASTNET Weekly
Figure 7-21 Normalized Mean Bias (%) of sulfate during fall 2016 at monitoring sites in the modeling
domain
S04 NMB (%) for run
for September to November 2016
71
-------
to November 2016
S04 NME
units _ %
coverage limit = 75%
I> 100
90
1 80
70
60
IMPROVE * CSN ฆ CASTNET Weekly
Figure 7-22 Normalized Mean Error (%) of sulfate during fall 2016 at monitoring sites in the
modeling domain
7.4.4.2 Seasonal Evaluation for Nitrate
The model performance bias and error statistics for nitrate for each climate region and each
season are provided in Table 7-6. This table includes statistics for particulate nitrate as measured
at CSN and IMPROVE sites and total nitrate (TNO< NOjrHNOi) as measured at CASTNet
sites. Spatial plots of the mean bias and error as well as normalized mean bias and error by
season for individual monitors are shown in Figure 7-23 through Figure 7-54.
Table 7-6 Nitrate and Total Nitrate Performance Statistics by Climate Region, by Season, and by
Monitoring Network for the 2016 CMAQ Model Simulation
Climate
Region
Monitor
Network
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Winter
431
0.7
0.8
>100
>100
IMPROVE
Spring
477
0.0
0.2
4.4
69.0
(N03)
Summer
486
0.0
0.2
18.9
>100
Northeast
Fall
456
0.1
0.2
33.6
90.8
Winter
715
1.0
1.2
58.1
72.1
CSN
Spring
770
0.1
0.5
13.9
58.4
(N03)
Summer
778
-0.1
0.2
-40.6
67.0
Fall
737
0.2
0.4
26.7
68.5
72
-------
Climate
Region
Monitor
Network
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
CASTNet
(TN03)
Winter
221
0.3
0.4
21.4
28.4
Spring
242
-0.1
0.3
-10.9
27.4
Summer
252
0.0
0.3
4.9
29.8
Fall
242
0.1
0.3
6.4
31.9
Ohio Valley
IMPROVE
(N03)
Winter
220
-0.2
0.7
-13.8
54.0
Spring
244
-0.2
0.3
-47.7
62.9
Summer
239
0.0
0.2
-23.8
87.4
Fall
227
-0.2
0.3
-37.4
68.2
CSN
(N03)
Winter
543
0.1
1.0
2.2
43.3
Spring
562
0.0
0.5
-0.4
64.9
Summer
552
0.0
0.3
-3.4
82.6
Fall
538
0.0
0.5
3.0
60.7
CASTNet
(TNO3)
Winter
212
-0.3
0.6
-12.1
22.4
Spring
228
-0.3
0.5
-20.3
27.3
Summer
224
0.1
0.4
9.9
31.2
Fall
226
0.0
0.6
-1.6
34.3
Upper
Midwest
IMPROVE
(N03)
Winter
200
-0.3
0.7
-20.2
48.1
Spring
208
-0.2
0.3
-43.2
60.3
Summer
210
0.0
0.1
-2.9
82.3
Fall
215
-0.1
0.3
-37.1
68.7
CSN
(N03)
Winter
326
0.1
1.0
2.9
38.5
Spring
354
0.0
0.6
0.1
58.3
Summer
313
0.0
0.3
1.4
88.2
Fall
307
0.0
0.4
2.0
57.4
CASTNet
(TNO3)
Winter
59
-0.4
0.6
-17.1
24.1
Spring
63
-0.2
0.4
-15.0
30.9
Summer
63
0.0
0.3
4.3
30.3
Fall
57
-0.1
0.4
-12.3
32.0
Southeast
IMPROVE
Winter
342
0.2
0.4
36.0
78.1
73
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
(NOs)
Spring
379
-0.1
0.2
-39.7
70.6
Summer
394
0.0
0.1
-18.7
73.0
Fall
366
-0.1
0.2
-20.2
67.8
Winter
573
0.7
0.8
>100
>100
CSN
Spring
643
-0.1
0.2
-19.8
70.0
(NOs)
Summer
608
-0.1
0.2
-23.2
79.9
Fall
560
0.1
0.2
19.5
80.8
Winter
150
0.0
0.5
1.5
36.6
CASTNet
Spring
164
-0.5
0.5
43.8
-35.2
(TNO3)
Summer
164
-0.2
0.4
-21.7
37.7
Fall
154
-0.2
0.5
-13.8
39.6
Winter
240
-0.1
0.5
-14.2
58.4
IMPROVE
Spring
273
-0.1
0.2
-42.1
70.2
(N03)
Summer
252
-0.1
0.2
-65.7
82.5
Fall
264
-0.1
0.2
-48.0
67.9
Winter
326
0.1
0.5
9.2
58.5
South
CSN
Spring
349
-0.1
0.2
-35.8
70.0
(NOs)
Summer
335
-0.1
0.2
-29.1
79.3
Fall
330
0.0
0.2
-14.5
76.6
Winter
92
-0.4
0.5
-21.4
30.9
CASTNet
Spring
102
-0.4
0.4
-37.5
38.4
(TNO3)
Summer
96
-0.4
0.5
-37.6
42.3
Fall
102
-0.2
0.3
-18.4
34.0
Winter
910
-0.2
0.2
-55.5
78.5
IMPROVE
Spring
991
-0.1
0.1
-60.2
82.4
Southwest
(N03)
Summer
985
-0.1
0.1
-93.4
95.2
Fall
962
-0.1
0.1
-72.8
84.0
CSN
Winter
272
-0.1
0.5
-13.2
53.3
(N03)
Spring
255
-0.2
0.3
-53.4
66.0
74
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Summer
250
-0.2
0.3
-75.7
97.4
Fall
257
-0.3
0.4
-55.1
79.5
Winter
101
-0.2
0.3
-33.6
46.9
CASTNet
Spring
115
-0.1
0.2
-29.2
37.3
(TN03)
Summer
114
-0.3
0.3
-39.8
42.5
Fall
115
-0.1
0.2
-12.9
30.2
Winter
542
-0.1
0.3
-41.9
69.6
IMPROVE
Spring
573
-0.1
0.1
-46.6
74.5
(NOs)
Summer
603
-0.1
0.1
-78.4
85.8
Fall
574
0.0
0.1
-29.8
80.3
Winter
139
-0.2
0.7
-13.6
54.7
Northern
Rockies
CSN
Spring
151
-0.2
0.3
-33.8
56.3
(N03)
Summer
153
-0.1
0.1
-43.8
76.9
Fall
135
0.0
0.2
-6.8
65.6
Winter
126
-0.3
0.3
-42.4
49.7
CASTNet
Spring
139
-0.1
0.2
32.0
36.2
(TNCb)
Summer
138
-0.2
0.2
-35.6
36.0
Fall
129
-0.1
0.1
-24.7
33.0
Winter
427
-0.1
0.3
-27.0
94.6
IMPROVE
Spring
505
0.1
0.2
49.3
>100
(N03)
Summer
519
0.1
0.3
90.5
>100
Fall
499
0.0
0.2
26.2
>100
Winter
157
-0.2
1.0
-14.4
87.3
Northwest
CSN
Spring
166
1.3
1.3
>100
>100
(N03)
Summer
153
1.2
1.2
>100
>100
Fall
161
0.6
0.8
>100
>100
CASTNet
Winter
-
-
-
-
-
(TNO3)
Spring
-
-
-
-
-
Summer
-
-
-
-
-
75
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Fall
-
-
-
-
-
Winter
565
-0.2
0.3
-36.7
64.0
IMPROVE
Spring
608
-0.1
0.2
-34.4
57.7
(NOs)
Summer
603
-0.2
0.3
-55.5
81.3
Fall
576
-0.2
0.3
-53.5
73.0
Winter
341
-2.0
2.1
-59.8
64.7
West
CSN
Spring
352
-0.9
1.0
-57.1
63.0
(NOs)
Summer
348
-0.7
0.8
-55.1
65.4
Fall
332
-1.3
1.5
-66.3
75.0
Winter
69
-0.3
0.4
-44.2
52.1
CASTNet
Spring
73
-0.5
0.5
-49.0
49.7
(TNO3)
Summer
75
-0.9
0.9
-54.4
54.6
Fall
76
-0.6
0.6
-48.4
51.3
76
-------
N03 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
3 = ug.rn3
coverage limit - 75%
1.5
1
0.5
~
s
ฆ0.5
IMPROVE a CSN
Figure 7-23 Mean Bias (ug/m3) for nitrate during winter 2016 at monitoring sites in the modeling
domain
5 = ug.'m3
coverage limit - 75%
IMPROVE a CSN
Figure 7-24 Mean Error (ug/m3) for nitrate during winter 2016 at monitoring sites in the modeling
domain
NQ3 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
77
-------
TN03 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
CASTNET Weekly
Figure 7-25 Mean Bias (ug/m3) for total nitrate during winter 2016 at monitoring sites in the modeling
domain
TN03 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
CASTNET Weekly
Figure 7-26 Mean Error (ug/m3) for total nitrate during winter 2016 at monitoring sites in the
modeling domain
78
-------
N03 NMB (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
coverage limit - 75%
IMPROVE * CSN
Figure 7-27 Normalized Mean Bias (%) for nitrate during winter 2016 at monitoring sites in the
modeling domain
coverage limit - 75%
>100
90
80
70
60
50
40
30
20
10
0
Figure 7-28 Normalized Mean Error (%) for nitrate during winter 2016 at monitoring sites in the
modeling domain
IMPROVE a CSN
79
-------
TN03NMB
12US2 for December to February 2016
units =
%
covera
je limit = 75%
J
> 100
H
80
-
60
40
ฆ
20
J
-20
-
-40
.
-60
-80
<-100
CASTNET Weekly
Figure 7-29 Normalized Mean Bias (%) for total nitrate during winter 2016 at monitoring sites in the
modeling domain
TN03 NME (%) lor run
12US2 for December to February 2016
CASTNET Weekly
Figure 7-30 Normalized Mean Error (%) for total nitrate during winter 2016 at monitoring sites in the
modeling domain
80
-------
N03 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
IMPROVE a CSN
Figure 7-31 Mean Bias (ug/in3) for nitrate during spring 2016 at monitoring sites in the modeling
domain
N03 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
IMPROVE a CSN
Figure 7-32 Mean Error (ug/m3) for nitrate during spring 2016 at monitoring sites in the modeling
domain
81
-------
TN03 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
CASTNET Weekly
Figure 7-33 Mean Bias (ug/m3) for total nitrate during spring 2016 at monitoring sites in the modeling
domain
TN03 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
CASTNET Weekly
Figure 7-34 Mean Error (ug/m3) for total nitrate during spring 2016 at monitoring sites in the
modeling domain
82
-------
N03
2016
coverage limit = 75%
IMPROVE * GSN
Figure 7-35 Normalized Mean Bias (%) for nitrate during spring 2016 at monitoring sites in the
modeling domain
units - %
coverage limit = 75%
-
> 100
90
_
80
70
60
-
50
-
40
ฆ
30
20
1
,0
L
0
IMPROVE * CSN
Figure 7-36 Normalized Mean Error (%) for nitrate during spring 2016 at monitoring sites in the
modeling domain
NQ3 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
83
-------
TN03 NMB (%) for run
for March to May 2016
units = %
coverage limit = 75%
F
> 100
ฆ
80
60
40
ฆ
20
J
-20
-
-40
.
-60
-80
<-100
CASTNET Weekly
Figure 7-37 Normalized Mean Bias (%) for total nitrate during spring 2016 at monitoring sites in the
modeling domain
TN03 NME
2016
> 100
80
70
60
50
40
30
20
10
CASTNET Weekly
Figure 7-38 Normalized Mean Error (%) for total nitrate during spring 2016 at monitoring sites in the
modeling domain
84
-------
N03
for June to August 2016
units - ug:'in3
coverage limit = 75%
IMPROVE * CSN
Figure 7-39 Mean Bias (ug/rn3) for nitrate during summer 2016 at monitoring sites in the modeling
domain
a - ug.'m3
coverage limit = 75%
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
IMPROVE a CSN
Figure 7-40 Mean Error (ug/m3) for nitrate during summer 2016 at monitoring sites in the modeling
domain
2016
N03ME
85
-------
TN03 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
CASTNET Weekly
Figure 7-41 Mean Bias (ug/m3) for total nitrate during summer 2016 at monitoring sites in the
modeling domain
TN03 ME
-------
N03 NMB (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
** iv "
TV / -T-V.
. -V< -BSQi
- * i r
^-4 tp^v
IMPROVE * CSN
Figure 7-43 Normalized Mean Bias (%) for nitrate during summer 2016 at monitoring sites in the
modeling domain
units - %
coverage limit = 75%
IMPROVE a CSN
Figure 7-44 Normalized Mean Error (%) for nitrate during summer 2016 at monitoring sites in the
modeling domain
N03NME
for June to August 2016
87
-------
TN03 NMB (%) for run
12US2 for June to August 2016
units = %
coverage limit = 75%
I> 100
80
60
40
CASTNET Weekly
Figure 7-45 Normalized Mean Bias (%) for total nitrate during summer 2016 at monitoring sites in the
modeling domain
TN03 NME
12US2 for June to August 2016
> 100
90
80
70
60
50
40
30
20
10
0
CASTNET Weekly
Figure 7-46 Normalized Mean Error (%) for total nitrate during summer 2016 at monitoring sites in
the modeling domain
88
-------
N03 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
IMPROVE a CSN
Figure 7-47 Mean Bias (ug/m3) for nitrate during fall 2016 at monitoring sites in the modeling domain
N03 ME (ug/m3) for run CMAQ_2016fm_cb6r3__ae7nvpoa_12US2 for September to November 2016
IMPROVE a CSN
Figure 7-48 Mean Error (ug/m3) for nitrate during fall 2016 at monitoring sites in the modeling
domain
89
-------
TN03 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
CASTNET Weekly
Figure 7-49 Mean Bias (ug/m3) for total nitrate during fall 2016 at monitoring sites in the modeling
domain
TN03 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
CASTNET Weekly
Figure 7-50 Mean Error (ug/m3) for total nitrate during fall 2016 at monitoring sites in the modeling
domain
90
-------
N03 NMB (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
IMPROVE a CSN
Figure 7-51 Normalized Mean Bias (%) for nitrate during fall 2016 at monitoring sites in the modeling
domain
units - %
coverage limit = 75%
IMPROVE * CSN
Figure 7-52 Normalized Mean Error (%) for nitrate during fall 2016 at monitoring sites in the
modeling domain
NQ3 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
91
-------
TN03 NMB {%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
CASTNET Weekly
Figure 7-53 Normalized Mean Bias (%) for total nitrate during fall 2016 at monitoring sites in the
modeling domain
units = %
coverage limit = 75%
> 100
90
80
70
60
50
40
30
20
10
CASTNET Weekly
Figure 7-54 Normalized Mean Error (%) for total nitrate during fall 2016 at monitoring sites in the
modeling domain
TNQ3 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
92
-------
7.4.4.3 Seasonal Ammonium Performance
The model performance bias and error statistics for ammonium for each climate region and
season are provided in Table 7-7. Spatial plots of the mean bias and error as well as normalized
mean bias and error by season for individual monitors are shown in Figure 7-55 through Figure
7-70.
Table 7-7 Ammonium Performance Statistics by Climate Region, by Season, and by Monitoring
Network for the 2016 CMAQ Model Simulation
Climate
Region
Monitor
Network
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Winter
718
0.6
0.6
>100
>100
CSN
Spring
770
0.2
0.3
81.0
>100
Summer
782
0.0
0.1
11.6
62.0
Northeast
Fall
737
0.2
0.3
77.4
>100
Winter
221
0.1
0.1
11.6
27.9
CASTNet
Spring
242
-0.1
0.1
-22.1
30.4
Summer
252
-0.1
0.1
-30.4
31.7
Fall
242
-0.1
0.1
-20.3
32.8
Winter
547
0.2
0.5
24.6
60.1
CSN
Spring
562
0.1
0.3
38.9
79.7
Summer
554
0.1
0.2
21.4
64.2
Ohio Valley
Fall
541
0.1
0.3
16.5
67.1
Winter
212
-0.2
0.2
-22.0
28.5
CASTNet
Spring
228
-0.2
0.2
-35.3
38.6
Summer
224
-0.1
0.2
-25.8
31.0
Fall
226
-0.2
0.2
-30.8
35.1
Winter
326
0.3
0.5
42.0
64.3
CSN
Spring
354
0.2
0.3
39.0
76.9
Upper
Midwest
Summer
314
0.1
0.2
58.8
87.1
Fall
310
0.2
0.3
82.4
>100
Winter
59
-0.2
0.2
-23.3
29.0
CASTNet
Spring
63
-0.1
0.2
-13.1
35.0
Summer
63
-0.1
0.1
-24.1
27.9
93
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Fall
57
-0.2
0.1
-30.0
36.3
Winter
513
0.3
0.4
92.7
>100
CSN
Spring
551
-0.1
0.2
-20.6
58.1
Summer
524
-0.1
0.2
-29.6
65.1
Southeast
Fall
503
0.0
0.2
-1.9
66.9
Winter
150
0.0
0.1
-4.6
27.5
CASTNet
Spring
164
-0.2
0.2
-45.7
46.6
Summer
164
-0.2
0.2
-46.7
46.9
Fall
154
-0.1
0.1
-34.7
38.3
Winter
327
0.1
0.3
38.4
82.2
CSN
Spring
351
-0.1
0.3
-39.2
76.1
Summer
336
-0.1
0.2
-35.3
83.3
South
Fall
331
-0.1
0.2
-18.6
60.6
Winter
92
-0.1
0.2
-19.4
35.3
CASTNet
Spring
102
-0.2
0.2
-46.9
51.1
Summer
96
-0.2
0.2
-50.3
52.2
Fall
102
-0.1
0.2
-35.2
39.8
Winter
247
-0.4
0.6
-60.9
83.7
CSN
Spring
255
0.0
0.1
-20.3
>100
Summer
250
-0.1
0.1
-59.3
>100
Southwest
Fall
260
-0.1
0.2
-43.8
>100
Winter
101
-0.1
0.1
-43.1
57.7
CASTNet
Spring
115
-0.0
0.1
-34.3
46.9
Summer
114
-0.1
0.1
-62.1
61.2
Fall
115
-0.1
0.1
-48.2
51.7
Winter
143
0.2
0.3
78.8
>100
Northern
Rockies
CSN
Spring
151
0.1
0.1
67.4
>100
Summer
153
0.1
0.1
>100
>100
Fall
139
0.1
0.1
>100
>100
94
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Winter
126
-0.1
0.1
-50.9
55.4
CASTNet
Spring
139
-0.1
0.1
-44.9
50.4
Summer
138
-0.1
0.1
-57.0
57.1
Fall
129
-0.1
0.1
-46.9
51.5
Winter
157
0.1
0.3
25.8
>100
CSN
Spring
161
0.2
0.2
>100
>100
Summer
166
0.2
0.3
>100
>100
Northwest
Fall
161
0.1
0.2
>100
>100
Winter
-
-
-
-
-
CASTNet
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Fall
-
-
-
-
-
Winter
341
-0.4
0.6
-51.0
74.0
CSN
Spring
352
-0.3
0.4
-58.8
79.1
Summer
349
-0.3
0.3
-74.8
81.6
West
Fall
332
-0.3
0.4
-64.1
82.7
Winter
69
-0.1
0.1
-38.9
58.3
CASTNet
Spring
73
-0.1
0.1
-59.8
61.5
Summer
75
-0.3
0.3
-84.0
84.0
Fall
76
-0.1
0.1
-60.9
63.3
95
-------
NH4 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
CSN a CASTNET Weekly
Figure 7-55 Mean Bias (ug/m3) of ammonium during winter 2016 at monitoring sites in the modeling
domain
NH4 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
CSN a CASTNET Weekly
Figure 7-56 Mean Error (ug/m3) of ammonium during winter 2016 at monitoring sites in the modeling
domain
96
-------
NH4 NMB (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
units-%
coverage limit = 75%
CSN * CASTNET Weekly
Figure 7-57 Normalized Mean Bias (%) of ammonium during winter 2016 at monitoring sites in the
modeling domain
units-
%
coverage limit = 75%
-
>100
90
_
80
70
60
-
50
.
,0
ฆ
30
20
10
U
0
CSN * CASTNET Weekly
Figure 7-58 Normalized Mean Error (%) of ammonium during winter 2016 at monitoring sites in the
modeling domain
NH4 NME (%) for run CMAQ 2016frn_cb6r3_ae7nvpoa 12US2 for December to February 2016
97
-------
NH4 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
CSN a CASTNET Weekly
Figure 7-59 Mean Bias (ug/m3) of ammonium during spring 2016 at monitoring sites in the modeling
domain
NH4 ME (ug/m3) for run CMAQ 2G16fm _cb6r3_ae7nvpoa 12US2 for March to May 2016
CSN * CASTNET Weekly
Figure 7-60 Mean Error (ug/m3) of ammonium during spring 2016 at monitoring sites in the modeling
domain
98
-------
NH4 NMB (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
units - %
coverage limit = 75%
CSN * CASTNET Weekly
Figure 7-61 Normalized Mean Bias (%) of ammonium during spring 2016 at monitoring sites in the
modeling domain
NH4 NME (%) for run CMAQ 2Q16fm_cb6r3 ae7nvpoa_12US2 for March to May 2016
CSN * CASTNET Weekly
Figure 7-62 Normalized Mean Error (%) of ammonium during spring 2016 at monitoring sites in the
modeling domain
99
-------
NH4 MB (ug/m3) for run CMAQj2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
CSN a CASTNET Weekly
Figure 7-63 Mean Bias (ug/m3) of ammonium during summer 2016 at monitoring sites in the modeling
domain
NH4 ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
CSN a CASTNET Weekly
Figure 7-64 Mean Error (ug/m3) of ammonium during summer 2016 at monitoring sites in the
modeling domain
100
-------
NH4 NMB (%) for run CMAQ__2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
coverage limit = 75%
CSN * CASTNET Weekly
Figure 7-65 Normalized Mean Bias (%) of ammonium during summer 2016 at monitoring sites in the
modeling domain
NH4 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for June to August 2016
CSN a CASTNET Weekly
Figure 7-66 Normalized Mean Error (%) of ammonium during summer 2016 at monitoring sites in the
modeling domain
101
-------
NH4 MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
units -
ug/m3
covera
e limit = 75%
-
>2
_
1,
ai
0.5
-
-0.5
.
:
<-2
CSN
CASTNET Weekly
Figure 7-67 Mean Bias (ug/m3) of ammonium during fall 2016 at monitoring sites in the modeling
domain
units - ug/m3
coverage limit = 75%
CSN * CASTNET Weekly
Figure 7-68 Mean Error (ug/m3) of ammonium during fall 2016 at monitoring sites in the modeling
domain
NH4 ME (ug/m3) for run CMAQ_2Q16fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
102
-------
NH4 NMB (%) for run CMAQ_2Q16fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
coverage limit = 75%
>100
80
60
40
20
0
-20
-40
-60
-80
<-100
CSN
CASTNET Weekly
Figure 7-69 Normalized Mean Bias (%) of ammonium during fall 2016 at monitoring sites in the
modeling domain
NH4 NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
CSN a CASTNET Weekly
Figure 7-70 Normalized Mean Error (%) of ammonium during fall 2016 at monitoring sites in the
modeling domain
103
-------
7.4.4.4 Seasonal Elemental Carbon Performance
The model performance bias and error statistics for elemental carbon for each of the nine
climate regions and each season are provided in Table 7-8. The statistics show clear over
prediction at urban and rural sites in most climate regions. Spatial plots of the mean bias and
error as well as normalized mean bias and error by season for individual monitors are shown in
Figure 7-71 through Figure 7-86. In the Northwest, issues in the ambient data when compared to
model predictions were found and thus removed from the performance analysis.
Table 7-8 Elemental Carbon Performance Statistics by Climate Region, by Season, and by Monitoring
Network for the 2016 CMAQ Model Simulation
Climate
Region
Monitor
Network
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Winter
429
0.1
0.1
50.0
63.9
IMPROVE
Spring
478
0.0
0.1
20.4
45.0
Summer
479
0.0
0.1
6.1
38.0
Northeast
Fall
456
0.0
0.1
14.3
43.2
Winter
722
0.1
0.4
18.1
53.7
CSN
Spring
785
0.0
0.2
-4.7
42.9
Summer
788
-0.1
0.2
-13.2
38.8
Fall
780
0.1
0.3
12.3
47.5
Winter
217
0.0
0.1
67.4
10.4
IMPROVE
Spring
242
0.0
0.1
-15.4
44.2
Summer
241
-0.1
0.1
-24.8
32.3
Ohio Valley
Fall
232
-0.1
0.1
-22.5
-21.9
Winter
535
0.1
0.2
56.5
14.9
CSN
Spring
571
-0.1
0.2
-16.1
38.0
Summer
532
-0.1
0.2
-20.6
37.3
Fall
535
-0.1
0.2
-9.0
33.9
Winter
222
0.1
0.1
40.3
55.3
IMPROVE
Spring
239
0.0
0.1
-13.8
43.9
Upper
Midwest
Summer
236
0.0
0.1
-20.6
41.7
Fall
243
0.0
0.1
-7.7
41.5
CSN
Winter
334
0.2
0.2
52.0
71.9
Spring
347
0.0
0.2
-1.3
47.0
104
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Summer
332
0.0
0.2
-8.2
44.5
Fall
338
0.0
0.2
6.1
46.4
Winter
398
0.0
0.1
-5.7
48.5
IMPROVE
Spring
446
-0.2
0.2
-42.2
55.7
Summer
442
-0.1
0.1
-24.1
45.2
Southeast
Fall
422
-0.1
0.1
-27.5
38.5
Winter
436
0.0
0.2
-4.1
40.9
CSN
Spring
478
-0.1
0.2
-24.3
41.7
Summer
445
0.0
0.2
-9.4
47.6
Fall
430
-0.1
0.3
-19.4
40.4
Winter
240
0.0
0.1
-1.5
40.5
IMPROVE
Spring
272
0.0
0.1
-7.9
50.1
Summer
242
0.0
0.0
-21.5
38.9
South
Fall
262
-0.1
0.1
-27.0
38.3
Winter
272
0.0
0.2
-6.5
40.1
CSN
Spring
297
-0.1
0.2
-15.5
37.6
Summer
251
0.0
0.2
-4.3
50.7
Fall
238
0.0
0.2
-2.8
44.0
Winter
890
-0.1
0.1
-34.1
58.2
IMPROVE
Spring
981
0.0
0.1
5.4
67.4
Summer
962
0.0
0.1
-16.9
55.3
Southwest
Fall
945
0.0
0.1
-22.5
57.7
Winter
228
0.0
0.4
-1.5
41.0
CSN
Spring
254
0.1
0.1
43.9
58.5
Summer
237
0.1
0.1
22.2
46.7
Fall
240
0.1
0.3
13.7
47.4
Northern
Rockies
Winter
557
0.0
0.0
12.8
70.3
IMPROVE
Spring
594
0.0
0.0
-24.7
63.0
Summer
616
0.0
0.1
30.8
79.4
105
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Fall
585
0.0
0.1
-13.9
60.0
Winter
124
0.0
0.3
0.6
100.0
CSN
Spring
145
0.0
0.1
-15.7
54.8
Summer
161
0.0
0.1
-18.2
43.1
Fall
146
0.0
0.2
-18.6
65.9
Winter
-
-
-
-
-
IMPROVE
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Northwest
Fall
-
-
-
-
-
Winter
-
-
-
-
-
CSN
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Fall
-
-
-
-
-
Winter
540
0.0
0.1
-14.3
62.2
IMPROVE
Spring
600
0.0
0.1
26.5
69.0
Summer
601
0.0
0.1
-24.5
61.5
West
Fall
565
0.0
0.1
4.9
59.7
Winter
294
0.2
0.2
45.5
58.4
CSN
Spring
293
0.2
0.2
42.9
56.2
Summer
267
0.1
0.2
29.0
46.3
Fall
277
0.2
0.3
51.3
32.6
106
-------
EC MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE a CSN
Figure 7-71 Mean Bias (ug/m3) of elemental carbon during winter 2016 at monitoring sites in the
modeling domain
EC ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE * CSN
Figure 7-72 Mean Error (ug/m3) of elemental carbon during winter 2016 at monitoring sites in the
modeling domain
107
-------
EC NMB {%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE a CSN
Figure 7-73 Normalized Mean Bias (%) of elemental carbon during winter 2016 at monitoring sites in
the modeling domain
EC NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE * CSN
Figure 7- 74 Normalized Mean Error (%) of elemental carbon during winter 2016 at monitoring sites
in the modeling domain
108
-------
EC MB
2016
units - ug:'in3
coverage limit = 75%
IMPROVE * CSN
Figure 7-75 Mean Bias (ug/rn3) of elemental carbon during spring 2016 at monitoring sites in the
modeling domain
EC ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
I
^ J \ -
V
rr~y^St ฆ.
1.8
1.6
IMPROVE a CSN
Figure 7-76 Mean Error (ug/m3) of elemental carbon during spring 2016 at monitoring sites in the
modeling domain
109
-------
EC NMB (%) for run
2016
coverage limit = 75%
> 100
80
| 60
40
20
|o
-20
-40
-60
-80
<-100
IMPROVE a CSN
Figure 7-77 Normalized Mean Bias (%) of elemental carbon during spring 2016 at monitoring sites in
the modeling domain
coverage limit = 75%
- 100
90
80
70
60
50
40
30
20
10
'o
IMPROVE a CSN
Figure 7-78 Normalized Mean Error (%) of elemental carbon during spring 2016 at monitoring sites
in the modeling domain
EC
2016
110
-------
EC MB
2016
5 - ug.''in3
coverage limit = 75%
1.5
0.5
-0.5
-1.5
: -2
IMPROVE a CSN
Figure 7-79 Mean Bias (ug/m3) of elemental carbon during summer 2016 at monitoring sites in the
modeling domain
units - ug,''m3
coverage limit = 75%
IMPROVE a CSN
Figure 7-80 Mean Error (ug/m3) of elemental carbon during summer 2016 at monitoring sites in the
modeling domain
EC ME
2016
111
-------
EC NMB
coverage limit = 76%
IMPROVE * CSN
Figure 7-81 Normalized Mean Bias (%) of elemental carbon during summer 2016 at monitoring sites
in the modeling domain
units - %
coverage limit = 75%
I> 100
90
80
IMPROVE a CSN
Figure 7-82 Normalized Mean Error (%) of elemental carbon during summer 2016 at monitoring sites
in the modeling domain
EC
for June to August 2016
112
-------
EC MB
to November 2016
units - ug:'in3
coverage limit = 75%
IMPROVE * CSN
Figure 7-83 Mean Bias (ug/rn3) of elemental carbon during fall 2016 at monitoring sites in the
modeling domain
units - ug.''m3
coverage limit = 75%
IMPROVE a CSN
Figure 7-84 Mean Error (ug/m3) of elemental carbon during fall 2016 at monitoring sites in the
modeling domain
EC
for September to November 2016
113
-------
EC NMB
for September to November 2016
units - %
coverage limit = 76%
pp> 100
W 80
60
40
IMPROVE a CSN
Figure 7-85 Normalized Mean Bias (%) of elemental carbon during fall 2016 at monitoring sites in the
modeling domain
coverage limit = 75%
> 100
90
80
70
60
50
40
30
20
10
IMPROVE a CSN
Figure 7-86 Normalized Mean Error (%) of elemental carbon during fall 2016 at monitoring sites in
the modeling domain
EC NME
for September to November 2016
114
-------
7.4.4.5 Seasonal Organic Carbon Performance
The model performance bias and error statistics for organic carbon for each climate region
and season are provided in Table 7-9. The statistics in this table indicate a tendency for the
modeling platform to over predict observed organic carbon concentrations during most seasons
and climate regions except in the Northern Rockies and the Western U.S. Spatial plots of the
mean bias and error as well as normalized mean bias and error by season for individual monitors
are shown in Figure 7-87 through Figure 7-102.
Table 7-9 Organic Carbon Performance Statistics by Climate Region, by Season, and by Monitoring
Network for the 2016 CMAQ Model Simulation
Climate
Region
Monitor
Network
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Winter
427
1.0
1.0
>100
>100
IMPROVE
Spring
477
0.4
0.5
54.2
68.7
Summer
482
0.0
0.5
3.1
37.6
Northeast
Fall
459
0.5
0.6
52.5
70.5
Winter
722
1.5
1.7
84.4
95.2
CSN
Spring
785
0.7
0.9
43.7
58.8
Summer
788
0.1
0.7
5.5
33.9
Fall
780
0.9
1.1
46.6
60.9
Winter
217
0.8
1.1
87.4
>100
IMPROVE
Spring
242
0.6
0.9
50.7
82.6
Summer
242
0.0
0.5
-0.1
34.4
Ohio Valley
Fall
232
0.2
0.9
13.5
51.5
Winter
535
0.9
1.1
56.5
70.2
CSN
Spring
571
0.3
0.7
19.9
46.2
Summer
531
0.0
0.6
1.8
32.7
Fall
532
0.2
0.9
8.4
37.0
Winter
226
0.6
0.7
>100
>100
IMPROVE
Spring
238
0.2
0.6
21.2
70.8
Upper
Midwest
Summer
237
-0.2
0.4
-20.0
37.7
Fall
243
0.1
0.4
15.3
46.6
CSN
Winter
333
1.4
1.5
>100
>100
Spring
347
0.6
1.0
38.9
68.8
115
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Summer
331
0.1
0.6
5.9
36.9
Fall
337
0.5
0.8
35.9
52.2
Winter
398
0.6
1.0
47.3
82.1
IMPROVE
Spring
447
-4.5
5.5
-72.7
87.9
Summer
455
-0.1
0.7
-7.5
43.5
Southeast
Fall
423
0.0
0.9
-1.6
45.8
Winter
436
0.8
1.1
39.9
56.2
CSN
Spring
478
0.4
0.8
19.3
41.4
Summer
445
0.4
0.7
21.6
38.9
Fall
430
0.2
1.2
6.0
42.4
Winter
239
0.4
0.6
51.8
71.7
IMPROVE
Spring
272
0.1
0.7
9.4
62.2
Summer
250
-0.1
0.5
-12.3
44.6
South
Fall
264
0.0
0.5
-1.0
45.2
Winter
272
0.6
1.2
29.6
62.3
CSN
Spring
297
0.3
0.8
22.5
52.9
Summer
251
0.4
0.8
21.4
56.0
Fall
237
0.6
1.1
27.4
55.3
Winter
881
-0.2
0.4
-25.3
58.0
IMPROVE
Spring
981
0.0
0.2
3.1
55.0
Summer
978
-0.2
0.5
-24.6
55.2
Southwest
Fall
964
-0.1
0.4
-11.0
58.8
Winter
228
0.6
1.6
23.3
63.3
CSN
Spring
254
0.4
0.7
42.3
66.3
Summer
237
0.0
0.5
-2.9
37.6
Fall
240
0.4
1.0
23.0
60.4
Northern
Rockies
Winter
549
0.1
0.2
16.9
68.8
IMPROVE
Spring
590
-0.1
0.4
-22.7
63.0
Summer
631
-0.2
0.7
-16.7
57.5
116
-------
Climate
Monitor
Season
No. of
MB
ME
NMB
NME
Region
Network
Obs
(ug/m3)
(ug/m3)
(%)
(%)
Fall
600
-0.1
0.4
-22.0
60.4
Winter
140
0.1
0.9
5.6
93.9
CSN
Spring
145
-0.1
0.5
-12.8
54.8
Summer
161
-0.6
0.7
-39.8
47.0
Fall
146
-0.2
0.5
-23.4
51.9
Winter
-
-
-
-
-
IMPROVE
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Northwest
Fall
-
-
-
-
-
Winter
-
-
-
-
-
CSN
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Fall
-
-
-
-
-
Winter
552
-0.1
0.3
-13.8
50.6
IMPROVE
Spring
599
-0.1
0.3
-14.6
45.9
Summer
608
-0.4
0.9
-21.1
52.2
West
Fall
574
-0.1
0.5
-8.5
49.6
Winter
285
0.4
1.7
9.7
46.7
CSN
Spring
294
0.3
0.7
20.2
42.9
Summer
289
-0.2
0.9
-9.7
35.7
Fall
277
0.3
1.2
10.9
42.5
117
-------
OC MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE * CSN
Figure 7-87 Mean Bias (ug/m3) of organic carbon during winter 2016 at monitoring sites in the
modeling domain
IMPROVE * CSN
Figure 7-88 Mean Error (ug/m3) of organic carbon during winter 2016 at monitoring sites in the
modeling domain
118
-------
OC NMB (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE * CSN
Figure 7-89 Normalized Mean Bias (%) of organic carbon during winter 2016 at monitoring sites in
the modeling domain
OC NME (%) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for December to February 2016
IMPROVE a CSN
Figure 7-90 Normalized Mean Error (%) of organic carbon during winter 2016 at monitoring sites in
the modeling domain
119
-------
OC MB
2016
coverage limit = 75%
IMPROVE * CSN
Figure 7-91 Mean Bias (ug/in3) of organic carbon during spring 2016 at monitoring sites in the
modeling domain
V
OC ME (ug/m3) for run CMAQ 2016fm_cb6r3_ae7nvpoa_12US2 for March to May 2016
I
4 3. a
s i, . .*%
ซ I \ y'l %
l-#. 4 /
v-1 ( i * \ \ >
IMPROVE a CSN
Figure 7-92 Mean Error (ug/in3) of organic carbon during spring 2016 at monitoring sites in the
modeling domain
120
-------
OC NMB
2016
units - %
coverage limit = 75%
I> 100
an
- 60
- 40
IMPROVE A CSN
Figure 7-93 Normalized Mean Bias (%) of organic carbon during spring 2016 at monitoring sites in
the modeling domain
coverage limit = 75%
- 100
40
30
20
10
IMPROVE a CSN
Figure 7-94 Normalized Mean Error (%) of organic carbon during spring 2016 at monitoring sites in
the modeling domain
oc
2016
121
-------
OC MB
coverage limit = 75%
IMPROVE * CSN
Figure 7-95 Mean Bias (ug/m3) of organic carbon during summer 2016 at monitoring sites in the
modeling domain
units - ug.'m3
coverage limit = 75%
IMPROVE a CSN
Figure 7-96 Mean Error (ug/m3) of organic carbon during summer 2016 at monitoring sites in the
modeling domain
OC ME
for June to August 2016
122
-------
oc
for June to August 2016
units - %
coverage limit = 76%
IMPROVE a CSN
Figure 7-97 Normalized Mean Bias (%) of organic carbon during summer 2016 at monitoring sites in
the modeling domain
units - %
coverage limit = 75%
IMPROVE * CSN
Figure 7-98 Normalized Mean Error (%) of organic carbon during summer 2016 at monitoring sites in
the modeling domain
for June to August 2016
123
-------
OC MB (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
IMPROVE a CSN
Figure 7-99 Mean Bias (ug/in3) of organic carbon during fall 2016 at monitoring sites in the modeling
domain
OC ME (ug/m3) for run CMAQ_2016fm_cb6r3_ae7nvpoa_12US2 for September to November 2016
IMPROVE a CSN
Figure 7-100 Mean Error (ug/m3) of organic carbon during fall 2016 at monitoring sites in the
modeling domain
124
-------
OCNMB
for September to November 2016
coverage limit = 75%
IMPROVE * GSN
Figure 7-101 Normalized Mean Bias (%) of organic carbon during fall 2016 at monitoring sites in the
modeling domain
units - %
coverage limit = 75%
I> 100
90
80
IMPROVE a CSN
Figure 7-102 Normalized Mean Error (%) of organic carbon during fall 2016 at monitoring sites in the
modeling domain
OC NME
to November 2016
125
-------
7.4.5 Seasonal Hazardous Air Pollutants Performance
A seasonal operational model performance evaluation for specific hazardous air pollutants
(i.e., formaldehyde, acetaldehyde, benzene, and 1,3-butadiene) was conducted in order to
estimate the ability of the CMAQ modeling system to replicate the base year concentrations for
the 12 km Continental United States domain. The seasonal model performance results for the 12
km modeling domain are presented below in Table 7-10. Toxic measurements included in the
evaluation were taken from the 2016 air toxics archive, https://www.epa.gov/amtic/amtic-air-
toxics-data-ambient-monitoring-archive. While most of the data in the archive are from the AQS
database including the National Air Toxics Trends Stations (NATTS), additional data (e.g.,
special studies) are included in the archive but not reported in the AQS. Similar to PM2.5 and
ozone, the evaluation principally consists of statistical assessments of model versus observed
pairs that were paired in time and space on daily basis.
Model predictions of annual formaldehyde, acetaldehyde, benzene, and 1,3 butadiene showed
relatively small to moderate bias and error percentages when compared to observations. Model
performance for HAPs is not as good as model performance for ozone and PM2.5. Technical
issues in the HAPs data consist of (1) uncertainties in monitoring methods; (2) limited
measurements in time/space to characterize ambient concentrations ("local in nature"); (3)
ambient data below method detection limit (MDL); (4) commensurability issues between
measurements and model predictions; (5) emissions and science uncertainty issues may also
affect model performance; and (6) limited data for estimating intercontinental transport that
effects the estimation of boundary conditions (i.e., boundary estimates for some species are much
higher than predicted values inside the domain).
As with the national, annual PM2.5 and ozone CMAQ modeling, the "acceptability" of model
performance was judged by comparing our CMAQ 2016 performance results to the limited
performance found in recent regional multi-pollutant model applications.82'83 Overall, the mean
bias and error (MB and ME), as well as the normalized mean bias and error (NMB and NME)
statistics shown below in Table 7-10 indicate that CMAQ-predicted 2016 toxics (i.e., observation
vs. model predictions) are within the range of recent regional modeling applications.
Table 7-10 Hazardous Air Toxics Performance Statistics by Season for the 2016 CMAQ Model
Simulation
Air Toxic Species
Season
No. of
Obs.
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Formaldehyde
Winter
1,814
-1.0
1.1
-58.4
63.1
Spring
1,914
-1.3
1.3
-59.4
61.8
82 Phillips, S., K. Wang, C. Jang, N. Possiel, M. Strum, T. Fox, 2007: Evaluation of 2002 Multi-pollutant Platform:
Air Toxics, Ozone, and Particulate Matter, 7th Annual CMAS Conference, Chapel Hill, NC, October 6-8, 2008.
Wesson, K., N. Fann, M. Morris, T. Fox, and B. Hubbell 2010: A Multi-pollutant, Risk-based Approach to the
Air Quality Management: Case Study for Detroit, Atmospheric Pollution Research, 1 (4) (2010), pp. 296-
304, 10.5094/APR.2010.037.
126
-------
Summer
2,318
-1.5
1.6
-47.0
50.4
Fall
1,886
-1.1
1.2
-48.0
54.4
Acetaldehyde
Winter
1,818
-0.4
0.4
-51.1
57.4
Spring
1,920
-0.5
0.5
-54.1
58.4
Summer
2,316
-0.2
0.5
-27.8
47.9
Fall
1,870
-0.4
0.5
-39.1
51.4
Benzene
Winter
3,406
-0.1
0.4
-11.9
42.6
Spring
3,968
-0.2
0.3
-25.8
47.2
Summer
5,249
0.0
0.2
-11.2
54.9
Fall
3,858
-0.2
0.4
-21.9
47.9
1,3-Butadiene
Winter
2,791
-0.1
0.2
-71.5
87.4
Spring
2,926
-0.1
0.1
-72.9
89.5
Summer
2,785
-0.1
0.1
-70.5
88.8
Fall
2,629
-0.1
0.1
-73.0
88.7
7.4.6 Seasonal Nitrate and Sulfate Deposition Performance
Seasonal nitrate and sulfate wet deposition performance statistics for the 12 km Continental
U.S. domain are provided in Table 7-11 and Table 7-12. The model predictions for seasonal
nitrate deposition generally show under predictions for the continental U.S. NADP sites (NMB
values range from -1.4% to -78.0%). Sulfate deposition performance shows similar under
predictions (NMB values range from -10.3% to 78.3%). The errors for both annual nitrate and
sulfate are relatively moderate with most values ranging from 33% to 88% which reflect scatter
in the model predictions versus observation comparison.
Table 7-11 Nitrate Wet Deposition Performance Statistics by Climate Region, by Season, and by
Monitoring Network for the 2016 CMAQ Model Simulation
Climate
Region
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Northeast
Winter
600
-0.1
0.1
-37.8
53.7
Spring
649
0.0
0.1
-7.4
44.7
Summer
681
0.0
0.1
-20.2
51.2
Fall
679
0.0
0.1
6.5
51.2
Ohio Valley
Winter
297
0.0
0.1
0.9
52.0
127
-------
Climate
Region
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Spring
300
0.0
0.1
-0.8
33.5
Summer
309
-0.1
0.1
-25.9
50.1
Fall
288
0.0
0.1
11.9
53.8
Upper
Midwest
Winter
275
0.0
0.1
-36.9
63.5
Spring
277
0.0
0.1
-23.4
46.1
Summer
292
-0.1
0.1
-28.7
45.5
Fall
301
0.0
0.1
-12.4
47.2
Southeast
Winter
359
0.0
0.0
0.3
52.4
Spring
376
0.0
0.1
-9.7
46.0
Summer
413
-0.1
0.1
-30.8
50.5
Fall
385
0.0
0.0
-11.5
61.0
South
Winter
236
0.0
0.0
15.5
57.8
Spring
263
0.0
0.1
-9.2
45.6
Summer
281
-0.1
0.1
-36.0
54.6
Fall
280
0.0
0.0
-15.7
54.7
Southwest
Winter
300
0.0
0.0
-78.0
82.5
Spring
322
0.0
0.1
-68.8
80.4
Summer
292
0.0
0.1
-38.7
55.9
Fall
334
0.0
0.0
-47.5
72.8
Northern
Rockies
Winter
216
0.0
0.0
-66.9
87.5
Spring
251
0.0
0.0
-47.9
57.3
Summer
226
0.0
0.1
-37.8
50.0
Fall
237
0.0
0.0
-35.1
63.7
Northwest
Winter
121
0.0
0.0
-1.4
51.2
Spring
141
0.0
0.0
-6.4
58.7
Summer
138
0.0
0.0
-4.4
71.7
Fall
145
0.0
0.0
19.3
64.5
West
Winter
151
0.0
0.0
-28.0
57.5
128
-------
Climate
Region
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Spring
151
0.0
0.0
13.6
81.7
Summer
161
0.0
0.0
-81.7
92.7
Fall
160
0.0
0.0
-13.9
76.6
Table 7-12 Sulfate Wet Deposition Performance Statistics by Climate Region, by Season, and by
Monitoring Network for the 2016 CMAQ Model Simulation
Climate
Region
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Northeast
Winter
600
0.0
0.1
-40.9
56.8
Spring
649
0.0
0.0
-16.6
45.0
Summer
681
0.0
0.1
-11.4
57.1
Fall
679
0.0
0.1
-14.9
54.0
Ohio Valley
Winter
297
0.0
0.1
-25.8
50.3
Spring
300
0.0
0.1
-13.3
34.9
Summer
309
0.0
0.1
-17.8
50.2
Fall
288
0.0
0.0
-10.3
51.6
Upper
Midwest
Winter
275
0.0
0.0
-37.8
58.3
Spring
277
0.0
0.0
-28.5
49.3
Summer
292
0.0
0.1
-21.6
49.0
Fall
301
0.0
0.0
-32.9
52.6
Southeast
Winter
359
0.0
0.1
-25.1
51.0
Spring
376
0.0
0.1
-25.5
53.6
Summer
413
0.0
0.1
-25.0
53.4
Fall
385
0.0
0.0
-17.7
63.8
South
Winter
236
0.0
0.0
-15.3
48.4
Spring
263
-0.1
0.1
-40.3
53.7
Summer
281
-0.1
0.1
-41.0
63.2
Fall
280
-0.0
0.0
-36.5
60.7
Southwest
Winter
300
0.0
0.0
-78.2
84.5
Spring
322
0.0
0.0
-65.6
79.8
129
-------
Climate
Region
Season
No. of
Obs
MB
(ug/m3)
ME
(ug/m3)
NMB
(%)
NME
(%)
Summer
292
0.0
0.0
-31.4
59.2
Fall
334
0.0
0.0
-63.2
75.0
Northern
Rockies
Winter
216
0.0
0.0
-71.2
86.3
Spring
251
0.0
0.0
-50.7
58.7
Summer
226
0.0
0.0
-28.4
53.2
Fall
237
0.0
0.0
-46.8
64.6
Northwest
Winter
121
0.0
0.0
33.8
69.7
Spring
141
0.0
0.0
13.4
63.4
Summer
138
0.0
0.0
32.8
95.8
Fall
145
0.0
0.1
46.2
92.3
West
Winter
151
0.0
0.0
65.4
>100
Spring
151
0.0
0.0
42.0
>100
Summer
161
0.0
0.0
-78.3
92.3
Fall
160
0.0
0.0
7.7
91.4
7.5 Model Simulation Scenarios
As part of our analysis for this rulemaking, the CMAQ modeling system was used to calculate
annual PM2.5 concentrations, 8-hour maximum average ozone season concentrations, annual
NO2, SO2, and CO concentrations, annual and seasonal (summer and winter) air toxics
concentrations, and annual nitrogen and sulfur deposition for each of the following emissions
scenarios:
2016 base year
2055 reference
2055 light and medium duty regulatory scenario
We use the predictions from the CMAQ model in a relative sense by combining the 2016
base-year predictions with predictions from each future-year scenario and applying these
modeled ratios to ambient air quality observations to estimate 8-hour ozone concentrations
during the ozone season (May - Sept), daily and annual PM2.5 concentrations, and visibility
impairment for each of the 2055 scenarios. The ambient air quality observations are average
conditions, on a site-by-site basis, for a period centered around the model base year (i.e., 2014-
2018).
130
-------
The projected annual PM2.5 concentrations were calculated using the Speciated Modeled
Attainment Test (SMAT) approach that utilizes a Federal Reference Method (FRM) mass
construction methodology which results in reduced nitrates (relative to the amount measured by
routine speciation networks), higher mass associated with sulfates (reflecting water included in
FRM measurements), and a measure of organic carbonaceous mass that is derived from the
difference between measured PM2.5 and its non-carbon components. This characterization of
PM2.5 mass also reflects crustal material and other minor constituents. The resulting
characterization provides a complete mass balance. It does not have any unknown mass that is
sometimes presented as the difference between measured PM2.5 mass and the characterized
chemical components derived from routine speciation measurements. However, the assumption
that all mass difference is organic carbon has not been validated in many areas of the U.S. The
SMAT methodology uses the following PM2.5 species components: sulfates, nitrates, ammonium,
organic carbon mass, elemental carbon, crustal, water, and blank mass (a fixed value of 0.5
|ig/m3). More complete details of the SMAT procedures can be found in the report "Procedures
for Estimating Future PM2.5 Values for the CAIR Final Rule by Application of the (Revised)
Speciated Modeled Attainment Test (SMAT)." For this analysis, several datasets and techniques
were updated. These changes are fully described within the technical support document for the
Final Transport Rule AQM TSD.
Additionally, we conducted an analysis to compare the absolute differences between the
future year reference and regulatory scenario for annual and seasonal acetaldehyde, benzene,
formaldehyde, and naphthalene, as well as annual NO2, SO2, CO, and nitrate/sulfate deposition.
These data were not compared in a relative sense due to the limited observational data available.
8 Additional Results of Illustrative Air Quality Analysis
EPA conducted an illustrative air quality modeling analysis of a regulatory scenario involving
light- and medium-duty "onroad" vehicle emission reductions and corresponding changes in
"upstream" emission sources like EGU (electric generating unit) emissions and refinery
emissions. Decisions about the emissions and other elements used in the air quality modeling
were made early in the analytical process for the proposed rulemaking. Accordingly, the air
quality analysis does not represent the proposal's regulatory scenario, nor does it reflect the
expected impacts of the Inflation Reduction Act (IRA). Based on updated power sector modeling
that incorporated expected generation mix impacts of the IRA (presented in Chapter 5), we are
projecting the IRA will lead to a significantly cleaner power grid; because the air quality analysis
presented here does not account for these impacts on EGU emissions, the location and magnitude
of the changes in pollutant concentrations should be considered illustrative and not viewed as
Agency projections of what we expect will be the total impact of the proposed standards.
Nevertheless, the analysis provides some insights into potential air quality impacts associated
with emissions increases and decreases from these multiple sectors.
Given the considerable uncertainty associated with the upstream emissions inventory (see
Sections 4 and 5), we also modeled a sensitivity case that examined only the air quality impacts
of the onroad emissions changes from the LMDV regulatory scenario. This "onroad-only"
sensitivity case assumed no change in emissions from upstream sources and is based on the
onroad emission inventories described in Section 3.
131
-------
The draft RIA includes maps that present the impact of the LMDV regulatory scenario on
projected ozone, PM2.5, NO2, SO2, CO, and air toxics concentrations, and projected nitrogen
deposition. In this TSD we present annual reference and LMDV regulatory scenario maps for
CO, NO2, SO2, air toxics, and nitrogen deposition as well as seasonal difference maps for air
toxics and visibility levels at Mandatory Class I Federal Areas.
8.1 Annual 2055 Reference, LMDV Regulatory, and Onroad-Only Scenario Maps
The following section presents maps of ambient concentrations of PM2.5, ozone, CO, NO2,
SO2, acetaldehyde, benzene, formaldehyde and naphthalene and total nitrogen deposition in the
2055 reference case and the 2055 LMDV regulatory scenario and the 2055 onroad-only scenario.
PM D24HourMean: 2016 Annual, 2055fm ref
Figure 8-1 Projected Illustrative Annual Average PM2.5 Concentrations in 2055 Reference Case
(ug/m3)
132
-------
PM D24HourMean: 2016 Annual, 2055fm Idlb
Figure 8-2 Projected Illustrative Annual Average PM2.5 Concentrations in 2055 LMDV Regulatory
Scenario (ug/m3)
PM D24HourMean: 2016 Annual, 2055fm Idlb onronly
Figure 8-3 Projected Illustrative Annual Average PM2.5 Concentrations in 2055 Onroad-Only Scenario
(ug/m3)
133
-------
03 D8HourMax: 2016 April-September, 2055fm ref
>55
50
45
40
>
35 S
30
.
25
20
<15
tax: 71.8516 Min: 21.7,738 \
Figure 8-4 Projected Illustrative Ozone Season (Apr-Sept) 8-hour Maximum Average Ozone
Concentrations in 2055 Reference case (ppb)
03 D8HourMax:2016 April-September, 2055fm Id lb
tax: 71.0062 M
Figure 8-5 Projected Illustrative Ozone Season (Apr-Sept) 8-hour Maximum Average Ozone
Concentrations in 2055 LMDV Regulatory Scenario (ppb)
134
-------
Figure 8-6 Projected Illustrative Ozone Season (Apr-Sept) 8-hour Maximum Average Ozone
Concentrations in 2055 Onroad-Only Scenario (ppb)
CO: 2016 Annual, 2055fm ref
Figure 8-7 Projected Illustrative Annual Average CO Concentrations in 2055 Reference Case (ppb)
135
-------
CO: 2016 Annual, 2055fm Idlb
Figure 8-8 Projected Illustrative Annual Average CO Concentrations in 2055 LMDV Regulatory
Scenario (ppb)
C 2016 Annua1, 2055fm Idlb onronly
Figure 8-9 Projected Illustrative Annual Average CO Concentrations in 2055 Onroad-Only Scenario
(PPb)
136
-------
N02: 2016 Annual. 2055fm ref
Figure 8-10 Projected Illustrative Annual Average NO2 Concentrations in 2055 Reference Case (ppb)
N02: 2016 Annual. 2055fm Idlb
Figure 8-11 Projected Illustrative Annual Average NO2 Concentrations in 2055 LMDV Regulatory
Scenario (ppb)
137
-------
N02: 2016 Annual, 2055fm Id lb onronly
Figure 8-12 Projected Illustrative Annual Average NO2 Concentrations in 2055 Onroad-Only Scenario
(PPb)
S02: 2016 Annual, 2055fm ref
Figure 8-13 Projected Illustrative Annual Average SO2 Concentrations in 2055 Reference Case (ppb)
138
-------
S02: 2016 Annual, 2055fm Idlb
Figure 8-14 Projected Illustrative Annual Average SO2 Concentrations in 2055 LMDV Regulatory
Scenario (ppb)
S02: 2016 Annual, 2055fm Idlb onronly
Figure 8-15 Projected Illustrative Annual Average SO2 Concentrations in 2055 Onroad-Only Scenario
(ppb)
139
-------
ALD2 UGM3: 2016 Annual, 2055fm ref
'
I
>1.80
1.60
1.40
1.20
ro
1.00 -ง
oi
3
0.80
0.60
0.40
<0.20
Figure 8-16 Projected Illustrative Annual Average Acetaldehyde Concentrations in 2055 Reference
Case (ug/m3)
ALD2 UGM3: 2016 Annual, 2055fm Idlb
Figure 8-17 Projected Illustrative Annual Average Acetaldehyde Concentrations in 2055 LMDV
Regulatory Scenario (ug/m3)
140
-------
ALD2 UGM3: 20X6 Annual, 2055fm Idlb onronly
I
I
>1.80
1.60
1.40
1.20
<
1.00
0.80
0.60
0.40
<0.20
oi
Figure 8-18 Projected Illustrative Annual Average Acetaldehyde Concentrations in 2055 Onroad-Only
Scenario (ug/m3)
BENZENE: 20X6 Annual, 2055fm ref
I
.
>0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
<0.05
Figure 8-19 Projected Illustrative Annual Average Benzene Concentrations in 2055 Reference Case
(ug/m3)
141
-------
>0.45
0.40
0.35
0.30
0.25 |
Cl
0.20
0.15
0.10
<0.05
Figure 8-20 Projected Illustrative Annual Average Benzene Concentrations in 2055 LMDV Regulatory
Scenario (ug/m3)
>0.45
0.40
0.35
0.30
0.25 |
CL
0.20
0.15
0.10
<0.05
Figure 8-21 Projected Illustrative Annual Average Benzene Concentrations in 2055 Onroad-Only
Scenario (ug/m3)
142
BENZENE: 2016 Annual, 2055fm Idlta
-------
FORM: 2016 Annual, 2055fm ref
i
>1.80
1.60
1.40
1.20
1.00
0.80
0.60
0.40
<0.20
Figure 8-22 Projected Illustrative Annual Average Formaldehyde Concentrations in 2055 Reference
Case (ug/m3)
FORM: 2016 Annual, 2055fm Idlb
Figure 8-23 Projected Illustrative Annual Average Formaldehyde Concentrations in 2055 LMDV
Regulatory Scenario (ug/m3)
143
-------
I
>1.80
1.60
1.40
1.20
1.00 |
CL
0.80
0.60
0.40
<0.20
Figure 8-24 Projected Illustrative Annual Average Formaldehyde Concentrations in 2055 Onroad-
Only Scenario (ug/m3)
>0.090
0.080
0.070
0.060
m
0.050 -I
D1
0.040
0.030
0.020
<0.010
Figure 8-25 Projected Illustrative Annual Average Naphthalene Concentrations in 2055 Reference
Case (ug/m3)
144
-------
NAPHTHALENE: 2016 Annual, 2055fm Idlb
Max: 0.5814 Min: 0.0 '
T-
P>0.090
ฎ 0.080
0.070
0.060
0.050
.
D1
3
0.040
0.030
0.020
<0.010
Figure 8-26 Projected Illustrative Annual Average Naphthalene Concentrations in 2055 LMDV
Regulatory Scenario (ug/in3)
I
I
>0.090
0.080
0.070
0.060
0.050
0.040
0.030
0.020
<0.010
Figure 8-27 Projected Illustrative Annual Average Naphthalene Concentrations in 2055 Onroad-Only
Scenario (ug/rn3)
NAPHTHALENE: 2016 AnnuaI. 2055fm Idlb onronly
Max: 0.5814 Min: 0.0
145
-------
TD N TOT: 2016 Annual, 2055fm ref
Figure 8-28 Projected Illustrative Annual Nitrogen Deposition in 2055 Reference Case (kg N/ha)
TD N TOT: 2016 Annual, 2055fm Idlb
Figure 8-29 Projected Illustrative Annual Nitrogen Deposition in 2055 LMDV Regulatory Scenario (kg
N/ha)
146
-------
TD N TOT: 2016 Annual, 2055fm Idlb onronly
Figure 8-30 Projected Illustrative Annual Nitrogen Deposition in 2055 Onroad-Only Scenario (kg
N/ha)
TD S TOT: 2016 Annual, 2055fm ref
Figure 8-31 Projected Illustrative Annual Sulfur Deposition in 2055 Reference Case (kg S/ha)
147
-------
TD S TOT: 2016 Annual, 2055fm Idlb
Figure 8-32 Projected Illustrative Annual Sulfur Deposition in 2055 LMDV Regulatory Scenario (kg
S/ha)
TD S TOT: 2016 Annual, 2055fm Idlb onronly
'
I
>4.5
4.0
3.5
3.0
ra
2.5 ^
crt
2.0
1.5
1.0
<0.5
Figure 8-33 Projected Illustrative Annual Sulfur Deposition in 2055 Onroad-Only Scenario (kg S/ha)
148
-------
8.2 Seasonal Air Toxics Maps
The following section presents maps of January and July monthly average ambient
concentrations of acetaldehyde, benzene, formaldehyde and naphthalene in the 2055 reference
case and the 2055 LIVEDV regulatory scenario. Also presented are maps of January and July
monthly average changes in ambient concentrations of acetaldehyde, benzene, formaldehyde and
naphthalene in 2055.
>1.80
1.60
1.40
1.20
m
1.00
oi
=j
0.80
0.60
0.40
<0.20
Figure 8-34 Projected Illustrative January Average Acetaldehyde Concentrations in 2055 Reference
Case (ug/m3)
ALD2 UGM3: 2016 January, 2055fm ref
149
-------
>1.80
1.60
1.40
1.20
ro
1.00 Je
oi
13
0.80
0.60
0.40
<0.20
Figure 8-35 Projected Illustrative January Average Acetaldehyde Concentrations in 2055 LMDV
Regulatory Scenario
>1.80
1.60
1.40
1.20
m
1.00 -I
Ol
3
0.80
0.60
0.40
<0.20
Figure 8-36 Projected Illustrative January Average Acetaldehyde Concentrations in 2055 Onroad-
Only Scenario
ALD2 UGM3: 2016 January, 2055fm Idlb
150
-------
ALD2 UGM3: 2016 July, 2055fm ref
Figure 8-37 Projected Illustrative July Average Acetaldehyde Concentrations in 2055 Reference Case
ALD2 UGM3: 2016 lulv. 2055fm Idlb
>1.80
1.60
1.20
m
1.00 -I
rj
0.80
<0.20
Max: 14.3272 Min:
Figure 8-38 Projected Illustrative July Average Acetaldehyde Concentrations in 2055 LMDV
Regulatory Scenario
151
-------
ALD2 UGM3: 2016 July, 2055fm Idlb onronly
I
s
>1.80
1.60
1.40
1.20
<
1.00
0.80
0.60
0.40
<0.20
cn
rj
Figure 8-39 Projected Illustrative July Average Acetaldehyde Concentrations in 2055 Onroad-Only
Scenario
>0.45
0.40
0.35
0.30
0.25 !
Q_
0.20
0.15
0.10
<0.05
BENZENE: 2016 January, 2055fm ref
Figure 8-40 Projected Illustrative January Average Benzene Concentrations in 2055 Reference Case
152
-------
BENZENE: 2016 January, 2055fm Idlb
I
I
>0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
<0.05
Figure 8-41 Projected Illustrative January Average Benzene Concentrations in 2055 LMDV
Regulatory Scenario
BENZENE: 20X6 January, 2055fm Idlb onronly
1
I
>0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
<0.05
Figure 8-42 Projected Illustrative January Average Benzene Concentrations in 2055 Onroad-Qnly
Scenario
153
-------
BENZENE: 2016 July, 2055fm ref
I
.
>0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
<0.05
Figure 8-43 Projected Illustrative July Average Benzene Concentrations in 2055 Reference Case
BENZENE: 2016 July, 2055fm Idlb
Figure 8-44 Projected Illustrative July Average Benzene Concentrations in 2055 LMDV Regulatory
Scenario
154
-------
BENZENE: 2016 July, 2055fm Idlb onronly
Figure 8-45 Projected Illustrative July Average Benzene Concentrations in 2055 Onroad-Only
Scenario
>1.80
1.60
1.40
1.20
1.00 |
Q.
0.80
0.60
0.40
<0.20
Figure 8-46 Projected Illustrative January Average Formaldehyde Concentrations in 2055 Reference
Case
155
-------
Figure 8-47 Projected Illustrative January Average Formaldehyde Concentrations in 2055 LMDV
Regulatory Scenario
>1.80
1.60
1.40
1.20
1.00 |
Q.
0.80
0.60
0.40
<0.20
FORM: 2016 January, 2055fm Id I b onronly
Figure 8-48 Projected Illustrative January Average Formaldehyde Concentrations in 2055 Onroad-
Only Scenario
156
-------
FORM: 2016 July, 2055fm ref
'
.
>1.80
1.60
1.40
1.20
1.00 1
Q.
0.80
0.60
0.40
<0.20
Figure 8-49 Projected Illustrative July Average Formaldehyde Concentrations in 2055 Reference Case
FORM: 2016 July, 2055fm Idlb
>1.80
1.60
1.40
1.20
1.00 |
Q.
0.80
0.60
0.40
<0.20
Figure 8-50 Projected Illustrative July Average Formaldehyde Concentrations in 2055 LMDV
Regulatory Scenario
157
-------
FORM: 2016 July, 2055fm Idlb onronly
\
>1.80
1.60
1.40
1.20
1.00 |
CL
0.80
0.60
0.40
<0.20
Figure 8-51 Projected Illustrative July Average Formaldehyde Concentrations in 2055 Gnroad-Only
Scenario
NAPHTHALENE; 2016 January, 2055fm ref
Max: 0.3886 Min: 0.0 ' '
.
>0.090
0.080
0.070
0.060
0.050
0.040
0.030
0.020
<0.010
m
E
ra
ZS
Figure 8-52 Projected Illustrative January Average Naphthalene Concentrations in 2055 Reference
Case
158
-------
NAPHTHALENE: 2016 January, 2055fm Idlb
I >0.090
0.080
0.070
0.060
0.050
0.040
0.030
lo.020
ฎ<0.010
Max: 0.3886 Mir: 0.0
Figure 8-53 Projected Illustrative January Average Naphthalene Concentrations in 2055 LMDV
Regulatory Scenario
NAPHTHALENE: 2016 January, 2055fm Idlb onronly
i
>0.090
0.080
0.070
0.060
0.050
0.040
0.030
0.020
<0.010
Max: 0.3886 Min: 0.0
Figure 8-54 Projected Illustrative January Average Naphthalene Concentrations in 2055 Onroad-Only
Scenario
159
-------
NAPHTHALENE: 2016 July, 2055fm ref
[;
4 ฆ
% #
4-
PฎHT\
* ฆ' ฆ;
ฆ *
' ^
? ฆ
ฆ ?ฆ
Vr .
j \
A
- 3 ^ฆฆ ฆ " p
" ฆ.' * ฆT. Y
ฆ '0 ! : j
j
I >0.090
0.080
0.070
0.060
0.050
0.040
_ 0.030
Ho.020
ฎ<0.010
A.
Figure 8-55 Projected Illustrative July Average Naphthalene Concentrations in 2055 Reference Case
NAPHTHALENE: 2016 July, 2055fm Idlb
sc.
*tji
*
s
^ ฆ
ฆ
-.v, ^
i ^ r
i' *
/
*
+_
Max: 3.2149 Min: 0.0
j X
- - -
, _ Ui ซ'
j ' r
* - - '-v.jp>
* *, ;r V-3 \
v"\ iปฃ
U J
f \ /ฆ
" 1 A-' -V%:
"v4
; $
ฆ j ฆ i ฆฆ .
r "v- *
ฆป ฆฆ*ฆฆ j
ฆ.
V
V . 4
m
>
V*
I
>0.090
0.080
0.070
0.060
0.050
0.040
0.030
0.020
<0.010
Figure 8-56 Projected Illustrative July Average Naphthalene Concentrations in 2055 LMDV
Regulatory Scenario
160
-------
.
>0.090
0.080
0.070
0.060
m
0.050 -I
cn
3
0.040
0.030
0.020
<0.010
Figure 8-57 Projected Illustrative July Average Naphthalene Concentrations in 2055 Onroad-Only
Scenario
ALD2 UGM3: 2016 January, 2055fm Idlb - 2055fm ref
I
>3.11e-03
2.22e-03
1.33e-03
4.44e-04 m
0.00e+00
-4.44e-04 D
-1.33e-03
-2.22e-03
I <-3.11e-03
Figure 8-58 Projected Illustrative Changes in Average Acetaldehyde Concentrations in January 2055
due to LMDV Regulatory Scenario
161
-------
ALD2 UGM3: 2016 July, 2055fm Idlb - 2055fm ref
>3.11e-03
2.22e-03
1.33e-03
4.44e-04 m
0.00e+00 ฆ!=
-4.44e-04 3
-1.33e-03
-2.22e-03
<-3.11e-03
Figure 8-59 Projected Illustrative Changes in Average Acetaldehyde Concentrations in July 2055 due
to LMDV Regulatory Scenario
ALD2 UGM3: 2016 January, 2055fm Idlb onronly - 2055fm ref
} V Y]-
Max:-0.0 Min:-0.0548^;, >
I
>3.11e-03
2.22e-03
1.33e-03
4.44e-04 ,
0.00e+00 i
-4.44e-04
-1.33e-03
I-2.22e-03
<-3.11e-03
Figure 8-60 Projected Illustrative Changes in Average Acetaldehyde Concentrations in January 2055
from "Onroad Only" Emissions Changes
162
-------
ALD2 UGM3: 2016 July, 2055fm Idlb onronly - 2055fm ref
>3.11e-03
2.22e-03
1.33e-03
4.44e-04 m
0.00e+00
-4.44e-04 3
-1.33e-03
-2.22e-03
<-3.11e-03
Figure 8-61 Projected Illustrative Changes in Average Acetaldehyde Concentrations in July 2055 from
"Onroad Only" Emissions Changes
BENZENE: 2016 January, 2055fm Idlb - 2055fm ref
I
I
>7.78e-04
5.56e-04
3.33e-04
l.lle-04
0.00e+00
-l.lle-04
-3.33e-04
-5.56e-04
<-7.78e-04
Figure 8-62 Projected Illustrative Changes in Average Benzene Concentrations in January 2055 due to
LMDV Regulatory Scenario
163
-------
>7.78e-04
5.56e-04
3.33e-04
l.lle-04
>
0.00e+00 -g_
-l.lle-04
-3.33e-04
-5.56e-04
<-7.78e-04
Figure 8-63 Projected Illustrative Changes in Average Benzene Concentrations in July 2055 due to
LMDV Regulatory Scenario
>7.78e-04
5.56e-04
3.33e-04
l.lle-04
0.00e+00
-l.lle-04
-3.33e-04
-5.56e-04
<-7.78e-04
Figure 8-64 Projected Illustrative Changes in Average Benzene Concentrations in January 2055 from
"Onroad Only" Emissions Changes
BENZENE: 2016 July, 2055fm Idlb - 2055fm ref
164
-------
BENZENE: 2016 July. 2055fm Idlb onronly - 2055fm ref
i tf y I
j - x
Max: 0.0 Min: -0.0165 . ?
>7.78e-04
5.56e-04
3.33e-04
l.lle-04
>
0.00e+00
-l.lle-04 ^
-3.33e-04
-5.56e-04
<-7.78e-04
Figure 8-65 Projected Illustrative Changes in Average Benzene Concentrations in July 2055 from
"Onroad Only" Emissions Changes
>3.11e-03
2.22e-03
1.33e-03
4.44e-04
0.00e+00
-4.44e-04
-1.33e-03
-2.22e-03
<-3.11e-03
Figure 8-66 Projected Illustrative Changes in Average Formaldehyde Concentrations in January 2055
due to LMDV Regulatory Scenario
165
-------
FORM: 2016 July, 2055fm Id I b - 2055fm ref
>3.11e-03
2.22e-03
1.33e-03
4.44e-04
>
0.00e+00
-4.44e-04 ^
-1.33e-03
-2.22e-03
<-3.11e-03
Figure 8-67 Projected Illustrative Changes in Average Formaldehyde Concentrations in July 2055 due
to LMDV Regulatory Scenario
>3.11e-03
2.22e-03
1.33e-03
4.44e-04
0.00e+00
-4.44e-04
-1.33e-03
-2.22e-03
<-3.11e-03
FORM: 2016 January, 2055fm Idlb onronly - 2055fm ref
Figure 8-68 Projected Illustrative Changes in Average Formaldehyde Concentrations in January 2055
from "Onroad Only" Emissions Changes
166
-------
FORM: 2016 July, 2055fm Id I b onronly - 2055fm ref
>3,lle-03
2.22e-03
1.33e-03
4.44e-04
>
0.00e+00
-4.44e-04 ^
-1.33e-03
-2.22e-03
<-3.11e-03
Figure 8-69 Projected Illustrative Changes in Average Formaldehyde Concentrations in July 2055
from "Onroad Only" Emissions Changes
NAPHTHALENE: 2016 January, 2055fm Idlb - 2055fm ref
B>7.78e-05
^5.56e-05
1
3.33e-05
0.00e+00 f
Ol
-l.lle-05 3
-3.33e-05
-5.56e-05
<-7.78e-05
Figure 8-70 Projected Illustrative Changes in Average Naphthalene Concentrations in January 2055
due to LMDV Regulatory Scenario
167
-------
NAPHTHALENE: 2016 July, 2055fm Idlb - 2055fm ref
I
>7.78e-05
5.56e-05
3.33e-05
l.lle-05 m
0.00e+00 ^
O)
-l.lle-05 D
-3.33e-05
-5.56e-05
I <-7.78e-05
Figure 8-71 Projected Illustrative Changes in Average Naphthalene Concentrations in July 2055 due
to LMDV Regulatory Scenario
NAPHTHALENE: 2016 January, 2055fm Idlb onronly - 2055fm ref
I
I
>7.78e-05
5.56e-05
3.33e-05
l.lle-05 f
0.00e+00 J
-l.lle-05
-3.33e-05
-5.56e-05
<-7.78e-05
Figure 8-72 Projected Illustrative Changes in Average Naphthalene Concentrations in January 2055
from "Onroad Only" Emissions Changes
168
-------
I>7.78e-05
5.56e-05
3.33e-05
l.lle-05 m
0.00e+00 ^
cn
-l.lle-05 D
I-3.33e-05
-5.56e-05
<-7.78e-05
Figure 8-73 Projected Illustrative Changes in Average Naphthalene Concentrations in July 2055 from
"Onroad Only" Emissions Changes
8.3 Projected Visibility in Mandatory Class I Federal Areas
,^5. ,
NAPHTHALENE: 2016 July, 2055fm Id I b onronly - 2055fm ref
T
2055
2055
Qnroad-
2016
2055
LMDV
Only
Baseline
Reference
Regulatory
Scenario
Visibility
Visibility
Scenario
Visibility
Natural
(dv) on
(dv) on
Visibility
(dv) on
Background
20%
20%
(dv) on
20%
(dv) on 20%
Most
Most
20% Most
Most
Most
Impaired
Impaired
Impaired
Impaired
Impaired
Class I Area Name
State
Days
Days
Days
Days
Days
Sipsey Wilderness
Alabama
19.03
15.54
15.49
15.47
9.62
Chiricahua NM
Arizona
9.41
8.86
8.86
8.84
4.93
Chiricahua Wilderness
Arizona
9.41
8.86
8.86
8.84
4.93
Galiuro Wilderness
Arizona
9.41
8.86
8.86
8.84
4.93
Grand Canyon NP
Arizona
6.87
6.45
6.44
6.44
4.16
Mazatzal Wilderness
Arizona
9.47
9.03
9.01
9.01
5.22
Mount Baldy Wilderness
Arizona
7.29
6.95
6.96
6.94
4.18
Petrified Forest NP
Arizona
8.16
7.57
7.59
7.55
4.21
Pine Mountain Wilderness
Arizona
9.47
9.03
9.01
9.01
5.22
Saguaro NM
Arizona
10.75
10.26
10.23
10.23
5.14
Superstition Wilderness
Arizona
10.45
9.97
9.96
9.95
5.14
169
-------
2055
2055
Onroad-
2016
2055
LMDV
Only
Baseline
Reference
Regulatory
Scenario
Visibility
(dv) on
20%
Visibility
(dv) on
20%
Scenario
Visibility
(dv) on
Visibility
(dv) on
20%
Natural
Background
(dv) on 20%
Most
Most
20% Most
Most
Most
Class I Area Name
State
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Sycamore Canyon Wilderness
Arizona
11.96
11.62
11.6
11.6
4.68
Caney Creek Wilderness
Arkansas
18.29
14.49
14.53
14.45
9.54
Upper Buffalo Wilderness
Arkansas
17.95
14.76
14.76
14.71
9.41
Agua Tibia Wilderness
California
16.34
15.41
15.32
15.32
7.66
Ansel Adams Wilderness (Minarets)
California
10.98
10.39
10.36
10.36
6.06
Caribou Wilderness
California
10.23
9.74
9.73
9.73
6.1
Cucamonga Wilderness
California
13.19
11.99
11.8
11.8
6.12
Desolation Wilderness
California
9.31
8.9
8.88
8.88
4.91
Dome Land Wilderness
California
15.14
14.37
14.35
14.35
6.19
Emigrant Wilderness
California
11.57
11.15
11.13
11.13
6.29
Hoover Wilderness
California
7.65
7.35
7.34
7.34
4.9
John Muir Wilderness
California
10.98
10.39
10.36
10.36
6.06
Joshua Tree NM
California
12.87
12.18
12.14
12.14
6.09
Kaiser Wilderness
California
10.98
10.39
10.36
10.36
6.06
Kings Canyon NP
California
18.43
17.52
17.48
17.48
6.29
Lassen Volcanic NP
California
10.23
9.74
9.73
9.73
6.1
Lava Beds NM
California
9.67
9.33
9.32
9.32
6.18
Mokelumne Wilderness
California
9.31
8.9
8.88
8.88
4.91
Pinnacles NM
California
14.1
13.49
13.46
13.45
6.94
Redwood NP
California
12.65
12.4
12.4
12.4
8.59
San Gabriel Wilderness
California
13.19
11.99
11.8
11.8
6.12
San Gorgonio Wilderness
California
14.45
12.96
12.83
12.82
6.2
San Jacinto Wilderness
California
14.45
12.96
12.83
12.82
6.2
San Rafael Wilderness
California
14.11
13.29
13.26
13.26
6.8
Sequoia NP
California
18.43
17.52
17.48
17.48
6.29
South Warner Wilderness
California
9.67
9.33
9.32
9.32
6.18
Thousand Lakes Wilderness
California
10.23
9.74
9.73
9.73
6.1
Ventana Wilderness
California
14.1
13.49
13.46
13.45
6.94
Yo Semite NP
California
11.57
11.15
11.13
11.13
6.29
Black Canyon of the Gunnison NM
Colorado
6.55
6.17
6.18
6.16
3.97
Eagles Nest Wilderness
Colorado
4.98
4.57
4.56
4.56
3.02
Flat Tops Wilderness
Colorado
4.98
4.57
4.56
4.56
3.02
Great Sand Dunes NM
Colorado
8.02
7.54
7.53
7.53
4.45
La Garita Wilderness
Colorado
6.55
6.17
6.18
6.16
3.97
170
-------
2055
2055
Onroad-
2016
2055
LMDV
Only
Baseline
Reference
Regulatory
Scenario
Visibility
(dv) on
20%
Visibility
(dv) on
20%
Scenario
Visibility
(dv) on
Visibility
(dv) on
20%
Natural
Background
(dv) on 20%
Most
Most
20% Most
Most
Most
Class I Area Name
State
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Maroon Bells-Snowmass Wilderness
Colorado
4.98
4.57
4.56
4.56
3.02
Mesa Verde NP
Colorado
6.51
5.88
5.87
5.86
4.2
Mount Zirkel Wilderness
Colorado
5.47
4.97
4.95
4.95
3.16
Rawah Wilderness
Colorado
5.47
4.97
4.95
4.95
3.16
Rocky Mountain NP
Colorado
8.41
7.58
7.54
7.53
4.94
Weminuche Wilderness
Colorado
6.55
6.17
6.18
6.16
3.97
West Elk Wilderness
Colorado
4.98
4.57
4.56
4.56
3.02
Chassahowitzka
Florida
17.41
15.69
15.67
15.66
9.03
Everglades NP
Florida
14.9
14.16
14.15
14.15
8.33
St. Marks
Florida
17.39
15.44
15.44
15.42
9.13
Cohutta Wilderness
Georgia
17.37
14.07
14.04
14.02
9.88
Okefenokee
Georgia
17.39
15.93
15.91
15.91
9.45
Wolf Island
Georgia
17.39
15.93
15.91
15.91
9.45
Craters of the Moon NM
Idaho
8.5
7.83
7.79
7.78
4.97
Sawtooth Wilderness
Idaho
8.61
8.34
8.33
8.33
4.7
Selway-Bitterroot Wilderness
Idaho
8.37
8.13
8.12
8.12
5.45
Mammoth Cave NP
Kentucky
21.02
16.78
16.67
16.68
9.8
Breton
Louisiana
18.97
17.4
17.4
17.38
9.23
Acadia NP
Maine
14.54
13.36
13.32
13.31
10.39
Moosehorn
Maine
13.32
12.49
12.46
12.46
9.98
Roosevelt Campobello International Park
Maine
13.32
12.49
12.46
12.46
9.98
Isle Royale NP
Michigan
15.54
14.37
14.34
14.32
10.17
Seney
Michigan
17.57
15.75
15.68
15.67
11.11
Boundary Waters Canoe Area
Minnesota
13.96
12.83
12.81
12.79
9.09
Voyageurs NP
Minnesota
14.18
13.18
13.18
13.15
9.37
Hercules-Glades Wilderness
Missouri
18.72
15.5
15.55
15.45
9.3
Mingo
Missouri
20.13
16.74
16.64
16.64
9.18
Anaconda-Pintler Wilderness
Montana
8.37
8.13
8.12
8.12
5.45
Bob Marshall Wilderness
Montana
10.06
9.84
9.83
9.83
5.53
Cabinet Mountains Wilderness
Montana
9.87
9.6
9.59
9.59
5.64
Gates of the Mountains Wilderness
Montana
7.47
7.33
7.33
7.32
4.53
Glacier NP
Montana
13.77
13.36
13.33
13.33
6.9
Medicine Lake
Montana
15.3
15.42
15.43
15.4
5.95
Mission Mountains Wilderness
Montana
10.06
9.84
9.83
9.83
5.53
171
-------
2055
2055
Onroad-
2016
2055
LMDV
Only
Baseline
Reference
Regulatory
Scenario
Visibility
(dv) on
20%
Visibility
(dv) on
20%
Scenario
Visibility
(dv) on
Visibility
(dv) on
20%
Natural
Background
(dv) on 20%
Most
Most
20% Most
Most
Most
Class I Area Name
State
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Red Rock Lakes
Montana
7.52
7.2
7.2
7.19
3.97
Scapegoat Wilderness
Montana
10.06
9.84
9.83
9.83
5.53
UL Bend
Montana
10.93
11.02
11.01
11.01
5.87
Jarbidge Wilderness
Nevada
7.97
7.78
7.77
7.77
5.23
Great Gulf Wilderness
New Hampshire
13.07
11.57
11.54
11.53
9.78
Presidential Range-Dry River Wilderness
New Hampshire
13.07
11.57
11.54
11.53
9.78
Brigantine
New Jersey
19.31
16.84
16.74
16.71
10.68
Bandelier NM
New Mexico
8.44
7.8
7.79
7.77
4.59
Bosque del Apache
New Mexico
10.47
9.71
9.71
9.69
5.39
Carlsbad Caverns NP
New Mexico
12.64
12.66
12.66
12.65
4.83
Gila Wilderness
New Mexico
7.58
7.16
7.19
7.15
4.2
Pecos Wilderness
New Mexico
5.95
5.4
5.42
5.38
3.5
Salt Creek
New Mexico
14.97
14.69
14.67
14.66
5.49
San Pedro Parks Wilderness
New Mexico
6.43
5.9
5.94
5.89
3.33
Wheeler Peak Wilderness
New Mexico
5.95
5.4
5.42
5.38
3.5
White Mountain Wilderness
New Mexico
9.95
9.65
9.65
9.64
4.89
Linville Gorge Wilderness
North Carolina
16.42
12.83
12.82
12.79
9.7
Shining Rock Wilderness
North Carolina
15.49
12.09
12.08
12.06
10.25
Swanquarter
North Carolina
16.3
14.01
13.93
13.92
10.01
Lostwood
North Dakota
16.18
16.33
16.3
16.3
5.87
Theodore Roosevelt NP
North Dakota
14.06
13.7
13.66
13.67
5.94
Wichita Mountains
Oklahoma
18.12
15.68
15.66
15.62
6.92
Crater Lake NP
Oregon
7.98
7.71
7.71
7.71
5.16
Diamond Peak Wilderness
Oregon
7.98
7.71
7.71
7.71
5.16
Eagle Cap Wilderness
Oregon
11.19
10.33
10.31
10.31
6.58
Gearhart Mountain Wilderness
Oregon
7.98
7.71
7.71
7.71
5.16
Hells Canyon Wilderness
Oregon
12.33
11.57
11.53
11.53
6.57
Kalmiopsis Wilderness
Oregon
11.97
11.56
11.55
11.55
7.78
Mount Hood Wilderness
Oregon
9.27
8.84
8.83
8.83
6.59
Mount Jefferson Wilderness
Oregon
11.28
10.9
10.89
10.89
7.3
Mount Washington Wilderness
Oregon
11.28
10.9
10.89
10.89
7.3
Mountain Lakes Wilderness
Oregon
7.98
7.71
7.71
7.71
5.16
Strawberry Mountain Wilderness
Oregon
11.19
10.33
10.31
10.31
6.58
Three Sisters Wilderness
Oregon
11.28
10.9
10.89
10.89
7.3
172
-------
2055
2055
Onroad-
2016
2055
LMDV
Only
Baseline
Reference
Regulatory
Scenario
Visibility
(dv) on
20%
Visibility
(dv) on
20%
Scenario
Visibility
(dv) on
Visibility
(dv) on
20%
Natural
Background
(dv) on 20%
Most
Most
20% Most
Most
Most
Class I Area Name
State
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Impaired
Days
Cape Romain
South Carolina
17.67
15.8
15.8
15.77
9.78
Badlands NP
South Dakota
12.33
11.87
11.85
11.85
6.09
Wind Cave NP
South Dakota
10.53
9.94
9.93
9.93
5.64
Great Smoky Mountains NP
Tennessee
17.21
13.85
13.82
13.8
10.05
Joyce-Kilmer-Slickrock Wilderness
Tennessee
17.21
13.85
13.82
13.8
10.05
Big Bend NP
Texas
14.06
13.48
13.5
13.48
5.33
Guadalupe Mountains NP
Texas
12.64
12.66
12.66
12.65
4.83
Arches NP
Utah
6.76
5.97
5.94
5.94
4.13
Bryce Canyon NP
Utah
6.6
6.11
6.1
6.1
4.08
Canyonlands NP
Utah
6.76
5.97
5.94
5.94
4.13
Capitol Reef NP
Utah
7.18
6.65
6.64
6.63
4
ZionNP
Utah
8.76
8.38
8.36
8.36
5.18
Lye Brook Wilderness
Vermont
14.75
12.86
12.78
12.77
10.24
James River Face Wilderness
Virginia
17.89
14.25
14.16
14.16
9.47
Shenandoah NP
Virginia
17.07
12.85
12.76
12.77
9.52
Alpine Lake Wilderness
Washington
12.74
11.83
11.77
11.75
7.27
Glacier Peak Wilderness
Washington
9.98
9.59
9.57
9.57
6.89
Goat Rocks Wilderness
Washington
7.98
7.66
7.65
7.65
6.14
Mount Adams Wilderness
Washington
7.98
7.66
7.65
7.65
6.14
Mount Rainier NP
Washington
12.66
12.15
12.14
12.12
7.66
North Cascades NP
Washington
9.98
9.59
9.57
9.57
6.89
Olympic NP
Washington
11.9
11.73
11.72
11.72
6.9
Pasayten Wilderness
Washington
9.46
9.06
9.05
9.05
5.96
Dolly Sods Wilderness
West Virginia
17.65
13.39
13.31
13.34
8.92
Otter Creek Wilderness
West Virginia
17.65
13.39
13.31
13.34
8.92
Bridger Wilderness
Wyoming
6.77
6.41
6.4
6.4
3.92
Fitzpatrick Wilderness
Wyoming
6.77
6.41
6.4
6.4
3.92
Grand Teton NP
Wyoming
7.52
7.2
7.2
7.19
3.97
North Absaroka Wilderness
Wyoming
7.17
6.85
6.85
6.84
4.55
Teton Wilderness
Wyoming
7.52
7.2
7.2
7.19
3.97
Washakie Wilderness
Wyoming
7.17
6.85
6.85
6.84
4.55
Yellowstone NP
Wyoming
7.52
7.2
7.2
7.19
3.97
a The level of visibility impairment in an area is based on the light-extinction coefficient and a unitless visibility
index, called a "deciview", which is used in the valuation of visibility. The deciview metric provides a scale for
173
-------
perceived visual changes over the entire range of conditions, from clear to hazy. Under many scenic conditions, the
average person can generally perceive a change of one deciview. The higher the deciview value, the worse the
visibility. Thus, an improvement invisibility is a decrease in deciview value.
174
------- |