* A vo
l ®!
PRO"^
Technical Support Document for EPA's
Updated 2028 Regional Haze Modeling for
Hawaii, Virgin Islands, and Alaska

-------
EPA-454/R-21-007
August 2021
Technical Support Document for EPA's Updated 2028 Regional Haze Modeling for Hawaii,
Virgin Islands, and Alaska
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Quality Assessment Division
Research Triangle Park, NC

-------
Table of Contents
Table of Contents	1
1.0 Background	Error! Bookmark not defined.
1.1 Introduction	2
2.0 Air Quality Modeling Platform	3
2.1	Air Quality Model Configuration	3
2.2	Meteorological Data for 2016	6
2.3	Initial and Boundary Concentrations	7
2.3.1	Hemispheric Simulation	8
2.3.2	Processing Boundaries from the Hemispheric Simulation	9
2.4	Emissions Inventories	9
2.5	Air Quality Model Evaluation	12
3.0 Projection of Future Year 2028 Visibility	13
3.1	Regional Haze Rule Requirement	14
3.2	Calculation of Visibility	15
3.2.1	2000-2018 Visibility	16
3.2.2	2028 Visibility	21
3.3	Comparison to Regional Haze "Glidepath"	24
3.4	Contribution from International & U.S. anthropogenic sources	28
4.0 References	29
Appendix A Model Performance Evaluation-Alaska	A-l
A.l Spatial Plots of Average Model Predictions on the Most Impaired Days	A-l
A.2 Time Series for 2016	A-6
A.3	Particulate Matter Composition on Clearest and Most Impaired Days in 2016 A-10
Appendix B Model Performance Evaluation - Hawaii	B-l
B.l	Spatial Plots of Average Model Predictions on the Most Impaired Days	B-l
B.2 Time Series for 2016	B-4
B.3	Particulate Matter Composition on Clearest and Most Impaired Days in 2016 B-6
Appendix C Model Performance Evaluation - Virgin Islands	C-l
C.l	Spatial Plots of Average Model Predictions on the Most Impaired Days	C-l
C.2 Time Series for 2016	C-3
C.3	Particulate Matter Composition on Clearest and Most Impaired Days in 2016 C-4
Appendix D Emissions Summary	D-5
D.l	Emissions summary table for Alaska	D-5
D.2 Emissions summary table for Hawaii	D-8
D.3 Emissions summary table for Puerto Rico/Virgin Islands	D-ll
1

-------
1.0 Background
The Regional Haze Rule (RHR) requires states to develop and submit state implementation plans
(SIP) that evaluate reasonable progress for implementation periods in approximately 10-year
increments. The next regional haze SIP is due in 2021 for the second implementation period
which ends in 2028.1 The EPA conducted visibility modeling for 2028 with the intention of
informing the regional haze SIP development process.
This modeling provides a number of outputs and metrics that may be helpful in the state
regional haze SIP planning process. These include:
1)	Projected 2028 visibility impairment on the 20% most anthropogenically impaired and
20% clearest days at each Class I area in Hawaii, Alaska, and the U.S. Virgin Islands.
2)	Estimated contribution of U.S. anthropogenic emissions to visibility impairment at each
Class I area.
Our goal is that this information, along with future collaborative work, will improve the
technical foundation of modeling used in regional haze SIP development. States should consult
with their EPA Regional Office to determine the usefulness of these model results for any
particular Class I area. Information about EPA's modeling for Class I areas in the contiguous U.S.
is provided elsewhere (U.S. Environmental Protection Agency, 2019c).
1.1 Introduction
In this technical support document (TSD) we describe the air quality modeling performed to
examine regional haze in 2028 at Class I areas in Alaska, Hawaii, and the Virgin Islands. For this
assessment, air quality modeling is used to project visibility levels at individual Class I areas
(represented by IMPROVE monitoring sites) to 2028 and to estimate U.S. anthropogenic
contribution to 2028 particulate matter (PM) concentrations and visibility. The projected 2028
PM concentrations are converted to light extinction coefficients and then to deciviews and used
to evaluate visibility progress in 2028. Hemispheric CMAQ modeling provides an estimate of
international anthropogenic contribution and a zero-out of U.S. anthropogenic emissions for
1 On January 10, 2017 (82 FR 3078), the EPA revised the Regional Flaze Rule to clarify and streamline
certain planning requirements for states. The rule also extended the deadline for second
implementation period plans by three years, to July 31, 2021. The second implementation period covers
2019 to 2028.
2

-------
each area provides an estimate of more local contribution. This information allows for a better
understanding and accounting of sources of future visibility impairment.
The remaining sections of this TSD are as follows. Section 2 describes the air quality modeling
platform and the evaluation of model predictions using measured concentrations. Section 3
defines the procedures for projecting regional haze deciview values to 2028, with comparisons
to both the "unadjusted" and "unadjusted" glidepath.
2.0 Air Quality Modeling Platform
The EPA used a 2016-based air quality modeling platform which includes emissions,
meteorology, and other inputs for 2016 as the base year for the modeling described in this TSD.
The 2016 base year emissions were projected to a future year base case scenario, 2028. The
emissions, meteorological modeling, and photochemical modeling used for this regional haze
assessment are further described below.
2.1 Air Quality Model Configuration
The photochemical model simulations performed for this Regional Haze assessment used the
Community Multiscale Air Quality Modeling (CMAQ) system version 5.3
(https://www.epa.gov/cmaq). CMAQ is a three-dimensional grid-based Eulerian air quality
model designed to simulate the formation and fate of oxidant precursors, primary and
secondary particulate matter concentrations, and deposition over regional and urban spatial
scales. Consideration of the different processes (e.g., transport and deposition) that affect
primary (directly emitted) and secondary (formed by atmospheric processes) pollutants at the
regional scale in different locations is fundamental to understanding and assessing the effects
of emissions on air quality concentrations.
Figures 2-1 and 2-2 show the geographic extent of the modeling domains that were used for air
quality modeling in this analysis. The three domains will subsequently be referred to as the AK,
HI and PR & VI domains, respectively. Domain specifications are provided in Table 2-1. All
domains are Lambert Conic projections with centers and true latitudes noted in Table 2-1. Each
nested simulation used initial and lateral boundary inflow from the coarser domain. The 27 km
domains were initialized using a hemispheric scale model simulation. The modeling domain
contains 35 vertical layers with a top at about 17,550 meters, or 50 millibars (mb). The model
simulations produce hourly air quality concentrations for each cell across the modeling domain.
3

-------
Table 2-1. Domain specifications used for each area.
Domain
Cell size
(km)
X and Y origin (km)
NX
NY
Lambert
center
Lambert true
latitudes
27AK1
27
-1,971,000.0, -1,701,000.0
146
126
-115, 63
60, 70
9AK1
9
-1,107,000.0, -1,134,000.0
312
252
-115, 63
60, 70
27HI1
27
-1,012,500.0, -1,012,500.0
75
75
-157, 21
19, 22
9HI1
9
-517,500.0, -490,500.0
100
100
-157, 21
19, 22
3HI1
3
-391,500.0, -346,500.0
225
201
-157, 21
19, 22
27PR1
27
-1,012,500.0, -1,012,500.0
75
75
-66, 18
17, 19
9PR1
9
-517,500.0, -436,500.0
100
100
-66, 18
17, 19
3PR1
3
-274,500.0, -202,500.0
150
150
-66, 18
17, 19
CMAQ requires a variety of input files that contain information pertaining to the modeling
domain and simulation period. These include gridded hourly emissions estimates,
meteorological data, and boundary concentrations. Separate emissions inventories were
prepared for the 2016 base year and the 2028 base case. All other inputs (i.e., meteorological
fields, initial concentrations, and boundary concentrations) were specified for the 2016 base
year model application and remained unchanged forthe future-year model simulations.The
CMAQ modeling scenarios were each performed using a single time segment with a 10-day ramp-up
period at the end of December 2015.
4

-------
Domain = 9AK1
DELTAXY =9 km
NX =312 and NY = 252
35 layers
Domain = 27AK1
DELTAXY = 27 km
NX = 146 and NY = 126
35 layers
27 km domain
9 km domain
1400
1200
1000
800
600
400
200
0
Figure 2-1. Maps of the 27km and 9km (insert) WRF and CMAQ modeling domains used for
regional haze modeling covering Alaska.
/ O (MH,



Figure 2-2. Maps of the WRF and CMAQ modeling domains used for regional haze modeling
covering Hawaii, Puerto Rico and the Virgin Islands
5

-------
Table 2-2 shows each of the CMAQ model runs performed for this analysis. For each of the
three modeling domains there are three model simulations: a 2016 base case, a 2028 future
base case, and a 2028 U.S. anthropogenic emissions zero-out model run.
Table 2-2. CMAQ model simulations for Alaska, Hawaii, and Puerto Rico/Virgin Islands.
Scenario Name
Description
2016fh_16j
Historical 2016 base case
2028fh_16j
Future year 2028 "on the books" scenario
2028fh_16j_zeroanth
Future year 2028 "on the books" scenario, with U.S.
anthropogenic emissions zeroed out.
2.2 Meteorological Data for 2016
Meteorological inputs for the photochemical and emissions models were generated with
version 3.9.1.1 of the Weather Research and Forecasting (WRF) model
(www2.mmm.ucar.edu/wrf/users). WRF was applied for the entire year of 2016 using 35 layers
between the surface and 50 mb with thinner layers closer to the surface to better resolve
diurnal variation in the planetary boundary layer. The Hawaii and Puerto Rico domains were
modeled using grid cells sized at 27, 9, and 3 km horizontal resolutions (Figure 2-2), the Alaska
domain was modeled at 27 and 9 km. The 27 km model domain was initialized with the
National Centers for Environmental Protection (NCEP) 0.25 degree Global Forecast System
(GFS) 0 hour analysis and 3 hour forecast (National Centers for Environmental Prediction, 2015).
The 9 and 3 km model domains, where applicable, were initialized using the coarser WRF
domain output. WRF was applied with the settings and inputs as described in Table 2-3 and
briefly summarized here. Unless otherwise noted in Table 2-3, WRF was applied with Morrison
microphysics, RRTMG radiation, Kain-Fritsch cumulus at 29 and 9 km (none at 3 km domains),
MODIS landcover, NOAH land surface model, revised MM5 Monin-Obukhov surface layer
scheme, and YSU planetary boundary layer option. Analysis nudging was applied for each model
domain using default nudging coefficients. WRF output was processed for input to CMAQ using
MCIP version 4.5 (Otte and Pleim, 2010). The MODIS landcover dataset was adjusted to change
incorrect landcover assignments of lakes to barren for Hawaii and Puerto Rico.
6

-------
Table 2-3. WRF Options - 27, 9, and 3 km Hawaii, Puerto Rico, and Alaska domains
Option
Hawaii
Puerto Rico
Alaska
WRF version
3.9.1.1
3.9.1.1
3.9.1.1
Boundary layer
YSU
YSU
MYNN Level 2.5
(bl_pbl_physics)



Surface layer
Revised MM5 Monin-
Revised MM5 Monin-
MYNN
(sf_sfclay_physics)
Obukhov scheme
Obukhov scheme

Land surface model
Noah
Noah
Noah
(sf_s u rf a ce_p hy s i cs)




Kain Fritsch - 27 and 9
Kain Fritsch - 27 and 9
Kain Fritsch - 27

km (cu_rad_feedback =
km (cu_rad_feedback =
and 9 km

.true.)
.true.)
(cu_rad_feedback
Cumulus scheme
none - 3 km
none - 3 km
= .true.)
Microphysics
Morrison
Morrison
Morrison
Longwave radiation
RRTMG
RRTMG
RRTMG
Shortwave radiation
RRTMG
RRTMG
RRTMG

GFS 0.25-degree (Oh
GFS 0.25-degree (Oh
GFS 0.25-degree

analysis and 3h forecast)
analysis and 3h forecast)
(Oh analysis and
Initialization data


3h forecast)
Horizontal grid resolution
27, 9, and 3 km
27, 9, and 3 km
27 and 9 km
Model top
50 mb
50 mb
50 mb
Number of vertical layers
35
35
35
Sea surface temperature
GFS- 27 and 9km/ NLCD-
GFS- 27 and 9km/ NLCD-
GFS- 27 and 9km
data
3km
3km

Analysis nudging
yes
yes
yes
MCIP version
4.5
4.5
4.5
Land cover data
Modis
Modis
Modis
Details of the annual 2016 meteorological model simulation and evaluation for the AK domain
are provided in a separate technical support document (U.S. Environmental Protection Agency,
2020a). Additional evaluation for the Hawaii and Puerto Rico/Virgin Islands domain are also
available elsewhere (Baker et al., 2020).
2.3 Initial and Boundary Concentrations
The lateral boundary and initial species concentrations are based on a hemispheric
modeling platform. The standard hemispheric simulation is described in detail in the
Hemispheric CMAQ 2016 Simulation TSD (U.S. Environmental Protection Agency, 2019a). The
hemispheric simulation is summarized in Section 2.3.1 and processing to boundary conditions is
summarized in Section 2.3.2.
7

-------
2.3.1 Hemispheric Simulation
The hemispheric modeling platform uses the Weather Research and Forecasting model (WRF
v3.8) meteorological model, the Sparse Matrix Operating Kernel for Emissions (SMOKE v4.5)
emissions model, and the Community Multiscale Air Quality model (CMAQ) version 5.2.1 with
the Carbon Bond mechanism (CB6r3) and the non-volatile aerosol option (AE6).
The hemispheric scale model uses a polar stereographic projection at 108 kilometer (km)
resolution to completely and continuously cover the Northern Hemisphere. The hemispheric
scale allows for long-range free tropospheric transport with 44 layers between the surface and
50 hPa (~20 km asl). The hemispheric modeling system was initiated on May 1st 2015 and run
continuously through December 31st, 2016.
The regional inventories over North America are based on the Inventory Collaborative 2016
emissions modeling platform (http://views.cira.colostate.edu/wiki/wiki/9169). which was
developed through the summer of 2019. The hemispheric modeling analysis used the 2016
"alpha release" (specifically the modeling case abbreviated 2016fe) that is publicly available
from https://www.epa.gov/air-emissions-modeling/2016-alpha-platform.
For the hemispheric emissions modeling platform, there are thirty anthropogenic sectors of
emissions including nine sectors based on the Hemispheric Transport of Air Pollution Version 2
inventory (EDGAR-HTAPv2) inventory and 15 sectors that represent emissions in China which
together comprise the anthropogenic emissions outside of North America. The international
emission inventories are synthesized from the EDGAR-HTAP v2 harmonized emission inventory
and country specific databases where updates were likely to be influential.
The EDGAR-HTAP v2 inventories were projected to represent the year 2014. Projection factors
were calculated from the Community Emissions Data System (CEDS) inventory at a country-
sector level. This allowed our inventory to evolve without the risks associated with transitioning
to a new inventory system. Especially because EDGAR-HTAP v2 is superseded for critical
counties, this was the optimal approach. Details of scaling factor development are described in
Section 2.1.5 of the 2016v7.1 Hemispheric Modeling Platform Technical Support Document
(U.S. Environmental Protection Agency, 2019a).
The China emission inventory was developed at Tsinghua University (THU) (Zhao et al., 2018).
This inventory was extensively compared to the EDGAR-HTAP v2 and EDGAR v4.3 inventories
before use. The largest differences for NOx in 2016 occurred in individual emissions sectors
rather than inventory totals. The SO2 emissions were more different than NOx emissions
between the two inventories because the THU inventory applies controls to the metal industry
that have been adopted by China.
8

-------
More details on the 2016 hemispheric CMAQ modeling are available in (U.S. Environmental
Protection Agency, 2019a) and more details on the hemispheric emissions inventories are
available elsewhere (U.S. Environmental Protection Agency, 2019b).
2.3.2 Processing Boundaries from the Hemispheric Simulation
The 108 km resolution hemispheric CMAQ predictions were used to provide one-way dynamic
boundary concentrations at one-hour intervals and an initial concentration field for the CMAQ
simulations. The hemispheric CMAQ results are spatially interpolated to lateral boundary and
initial boundary conditions.
Boundary conditions for the regional CMAQ domain require mapping hemispheric results from
the polar stereographic grid and the vertical layer structure. Both lateral and initial conditions
use nearest neighbor horizontal interpolation and vertical mass conserving interpolation. The
lateral boundaries perform the interpolation along the perimeter for each hour, while the initial
boundaries perform the interpolation for the entire domain at only specific hours. The initial
boundaries were created for 2015-12-22 at 00:00:00 UTC. These results are directly usable for
CMAQ.
2.4 Emissic sntories
CMAQ requires detailed emissions inventories containing temporally allocated (i.e., hourly)
emissions for each grid-cell in the modeling domain for a large number of chemical species that
act as primary pollutants and precursors to secondary pollutants. Annual emission inventories
for 2016 and 2028 were preprocessed into CMAQ-ready inputs using the Sparse Matrix
Operator Kernel Emissions (SMOKE) modeling system (https://www.cmascenter.org/smoke/).2
Biogenic emissions of volatile organic compounds (VOC) and nitrogen oxide (NO) were
generated using the Model of Emissions for Gases in Nature (MEGAN) version 2.0 (Guenther et
al., 2006) at 0.5 degree scale and allocated to the finer scales using relevant MODIS landcover
categories. Day-specific wildland fire emissions were based on FINN (Wiedinmyer et al., 2011)
for Puerto Rico and SmartFire2/BlueSky framework (Baker et al., 2016) for Alaska and Hawaii.
Agricultural burning emissions for Hawaii were developed from the 2016 Hazard Mapping
System (HMS) fire activity over agricultural land. Crop-specific emission factors were applied to
each daily fire to calculate emissions (Pouliot et al., 2017). Sea-salt (Gantt et al., 2015) and
2 The SMOKE output emissions case name for the 2016 base year is "2016fh_16j" and the emissions case
name for the 2028 base case is "2028fh_16j".
9

-------
halogen (Sarwar et al., 2015) emissions from the ocean were included. Lightning, wind-blown
dust, and volcanic emissions were not included.
Electric generating unit (EGU) emissions were based on state submitted data to the 2016
emissions modeling platform. Fuel use data provided by the Energy Information Administration
(EIA) was used to allocate annual EGU emissions to month when available or a state profile
based on total monthly fuel use otherwise. Monthly emissions were allocated to week and hour
of the day using default EGU temporal profiles reflective of typical energy use and load
patterns. Annual total EGU emissions in Puerto Rico were allocated to month, week, and hour
of the day based on regional fuel profiles (south Florida) from the Continuous Emissions
Modeling System (CEMS). The EGU emissions were based on 2016 values that were submitted
to the 2016 NEI. Values from the 2014 NEI were used for smaller sources that were not
submitted for 2016. Alaska provided comments on the EGU emissions from the beta platform
that were incorporated into the 2016vl inventories used for this case. The EGU inventories
were held constant at 2016 levels in the 2028 inventories.
The primary data source for non-EGU point sources is the 2016 point source National Emissions
Inventory (NEI). For point sources not updated for the 2016 point source NEI, 2014NEIv2
emissions were carried forward with additional updates provided by the States. Industrial
emissions were grown to 2028 according to factors derived from the 2019 Annual Energy
Outlook. Controls were applied to reflect relevant New Source Performance Standards (NSPS)
rules (e.g., reciprocating internal combustion engines (RICE), natural gas turbines, and process
heaters). Airport emissions for 2016 were derived from the 2017 draft National Emissions
Inventory (NEI) airport inventory, back projected to 2016 using Federal Aviation Administration
(FAA) data. Airport emissions were projected to 2028 using the FAA's Terminal Area Forecast
(TAF) data. Alaska rail emissions were developed from data maintained by the Federal Railroad
Administration (FRA) and tier fleet mix information from the Association of American Railroads
(AAR).
The onroad mobile source emissions were generated using the released version of the Motor
Vehicle Emissions Simulator (MOVES2014b). The activity data were temporally allocated based
on regional average temporal profiles from the Coordinating Research Council (CRC) A-100
data. The CRC A-100 data were available for the continental United States and did not include
AK / HI / PR / VI specifically. The A-100 data included regional average profiles, and those were
used in AK/HI (West region average) and PR/VI (South region average). Onroad and nonroad
mobile source emissions were created for 2028 using emission factors based on MOVES2014b
run for 2028, combined with activity data projected from 2016 to 2028 based on data from the
Annual Energy Outlook (AEO) 2018 and state/local-provided data, where available.
10

-------
Meteorological data from the year 2016 were used to compute the emissions for both 2016
and 2028.
Commercial Marine Vessel (CMV) emissions for ships with Category 1 and Category 2 (i.e., small
to medium-sized) engines, as well as ships with Category 3 (i.e., large) engines, were modeled
as point sources. All CMV emissions were based on AIS hourly ship data for the year 2017,
mapped to 2016 dates and adjusted to represent the year 2016 based on national adjustment
factors. CMV emissions were projected to 2028 using region-specific factors for NOx, S02, and
other pollutants. More details are available in the 2016 vl platform specification sheets
(National Emissions Inventory Collaborative, 2020) and the 2017 NEI TSD (U.S. Environmental
Protection Agency, 2020b).
The majority of nonpoint source emissions for the year 2016 were used as-is from the
2014NEIv2, except for emissions estimated using census data. Historical population data for
2016 from the US Census were used to project these emissions from the 2014NEIv2 to 2016.
Puerto Rico and Hawaii nonpoint emissions were held constant from 2014NEIv2 to 2016. Alaska
and Puerto Rico industrial emissions were grown to 2028 according to factors derived from the
2019 Annual Energy Outlook. Portions of the nonpoint sector were grown using factors based
on expected grown in human population. Controls were applied to reflect relevant New Source
Performance Standards (NSPS) rules (i.e., reciprocating internal combustion engines (RICE),
natural gas turbines, and process heaters). Nonpoint agricultural emissions, which includes
ammonia (NH3) and VOC emissions from livestock and fertilizer sources, were not included in
this assessment due to a lack of available data for 2016.
The nonpoint area fugitive dust sector consists of fugitive dust particulate matter (PM)
emissions from the 2014NEIv2 nonpoint source category. Emissions from paved roads were
projected from 2014 to 2016 based on county total vehicle miles traveled (VMT), but emissions
from all other sources, including unpaved roads, were held constant. After SMOKE processing,
road dust emissions were reduced using a gridded transport fraction file that considers the
impact of the roughness of the landscape on the emissions and further reduced at specific
hours based on snow cover and precipitation. Paved road dust was grown to 2028 levels based
on the growth in VMT from 2016 to 2028. The remainder of the sector including building
construction, road construction, agricultural dust, and road dust was held constant. The
projected emissions are reduced during modeling according to a transport fraction (newly
computed for the beta platform) and a year 2016 meteorology-based (precipitation and
snow/ice cover) adjustment as they are for the base year.
11

-------
Residential wood combustion (RWC) emissions were projected from the 2014NEIv2 values to
represent 2016 and 2028 using factors based on EPA's 2011v6.3 emissions modeling platform
and implemented into spreadsheet tools by MARAMA. Day-of-year temporalization of these
sources for Alaska is based on daily minimum temperature by county and calculated by the
SMOKE program GenTPRO, with more general profiles used for Hawaii and Puerto Rico. RWC
emissions were projected from 2014 to 2028 using the same spreadsheet tools used to create
2016 emissions. The projected emissions account for growth, retirements, and NSPS.
Point oil and gas emissions were based on the 2016 point source emissions modeling platform.
Any sources from the 2014 NEI which were not submitted for 2016 were included for Alaska at
their 2014 levels unless they were marked as shut down. Nonpoint oil and gas emissions were
estimated from the 2016 Nonpoint Oil and Gas Emission Estimation Tool developed by EPA.
State air agencies provided the 2016 oil and gas activity data to EPA. When state data is not
supplied, EPA populates the inventory with the best available data. Oil and gas emissions were
not projected to year 2028 for Alaska, Hawaii or Puerto Rico.
Annual total emissions are provided by major sector in Appendix D for each of the areas and
major pollutants relevant for regional haze.
2.5 Air Quality Model Evaluation
An operational model performance evaluation was performed for particulate matter (PM2.5
species components and coarse PM) and regional haze to examine the ability of the modeling
system to simulate 2016 measured concentrations. Model performance results are provided in
Appendix A.
The model evaluation was focused on the ability of the model to predict visibility related PM
components at Class I areas (represented by IMPROVE monitoring sites). The analysis looked at
PM species component performance at IMPROVE and other PM monitoring networks, and
performance on the 20% most impaired (and 20% clearest) days3 at individual IMPROVE sites.
This provides a comprehensive assessment of the components that make up visibility
performance.
3 The values for the 20% most impaired and clearest days are calculated according to the draft
recommended method in the draft EPA guidance document "Draft Guidance for the Second
Implementation Period of the Regional Haze Rule" posted at https://www.epa.gov/visibility/regional-
haze-guidance-technical-support-document-and-data-file.
12

-------
The measured concentrations of PM components such as sulfate and nitrate on the 20% most
impaired days at many Class I areas are extremely small. Some Class I areas have average
sulfate and nitrate observations (on the 20% most impaired days) of less than 1 |-ig/m3. This
makes it challenging to correctly model observed visibility. Assumptions regarding particular
emissions categories and boundary conditions can have a large impact on model performance.
Even when model performance appears to be accurate, it is sometimes difficult (without
further sensitivity modeling and analysis) to determine if the model is getting the right answer
for the right reasons.
Overall, the visibility performance for 2016 was generally good, with some regional exceptions.
In different parts of the country, varying PM components contribute to visibility impairment,
which also varies by season.
Appendix A contains tables and figures, including individual IMPROVE site PM species
component performance information for the 20% most impaired days. Performance issues seen
in the 2016 operational performance evaluation indicate uncertainty in the model results at
some Class I areas. However, visibility performance at most Class I areas is quite good, adding
to confidence in the future year contribution analyses and calculations. Further improvements
in emissions inputs, boundary conditions, and model chemistry may help improve model
performance in specific regions. More details about how the model compared to
measurements of chemically speciated PM2.5 and precursors are available in a separate
document for Hawaii and Puerto Rico/Virgin Islands (Baker et al., 2020).
3.0 Projection of Future Year 2028 Visibility
The PM predictions from the 2016 and 2028 CMAQ model simulations were used to project
2014-20174 IMPROVE visibility data to 2028 following the approach described in EPA's ozone,
PM2.5and regional haze modeling guidance (U.S. Environmental Protection Agency, 2018).5 The
SIP Modeling Guidance describes the recommended modeling analysis used to help set
reasonable progress goals (RPGs) that reflect the regional haze SIP's long-term strategy
containing adopted emissions control measures.
4	Based on EPA modeling guidance, a five-year average centered on the base modeling year (2014-2018)
would be the appropriate ambient data base period. However, as of September 2019, the 2018
IMPROVE data is not available. Therefore, a four-year average (2014-2017) period was used instead. The
ambient data can be updated when the final 2018 IMPROVE data becomes available.
5	The EPA's ozone, PM2.5, and regional haze modeling guidance is referred to as "the SIP Modeling
Guidance" in the remainder of this document.
13

-------
3.1 Regional Haze Rule Requirement
As required by the Regional Haze Rule (RHR) RPGs must provide for an improvement in visibility
for the 20 percent most anthropogenically impaired days relative to baseline visibility
conditions and ensure no degradation in visibility for the 20 percent clearest days relative to
baseline visibility conditions.6 The baseline for each Class I area is the average visibility (in
deciviews) for the years 2000 through 2004.7 The visibility conditions in these years are the
benchmark for the "provide for an improvement" and "no degradation" requirements. In
addition, states are required to determine the rate of improvement in visibility needed to reach
natural conditions by 2064 for the 20 percent most anthropogenically impaired days.8 A line
drawn between the end of the 2000-2004 baseline period and 2064 (dv/year) shows a uniform
rate of progress (URP) or "glidepath" between these two points. The glidepath represents a
linear or uniform rate of progress and is the amount of visibility improvement needed in each
implementation period to stay on the glidepath. The URP is a framework for consideration but
there is no rule requirement to be on or below the glidepath. An example glidepath plot is
shown in Figure 3-1.
5-year average
of 20% most
impaired days
o
cr>
c\j
Uniform rate of progress
or Glidepath
CO
5
Q>
>
O
(V
T3
O
CM
Yearly average
of 20% most
impaired days
2028
Reasonable
progress
goal
o
Natura
visibility
condition
2000
2010
2020
2030
2040
2050
2060
Year
Figure 3-1 Example Glidepath Plot.
s40 CFR 51.308(f)(3)(i).
740 CFR 51.308(f)(1) and definitions in 51.301.
8 40 CFR 51.308(f)(1).
14

-------
The RHR requires states to submit an implementation plan that evaluates and contains
measures found necessary to make reasonable progress for implementation periods in
approximately ten-year increments. The next regional haze SIP is due in July 2021, for the
implementation period which ends in 2028. Therefore, modeling was used to project visibility
to 2028 using a 2028 emissions inventory with "on-the-books" controls. The EPA Software for
Model Attainment Test- Community Edition (SMAT-CE) tool was used to calculate 2028 deciview
values on the 20% most anthropogenically impaired and 20% clearest days at each Class I Area
(IMPROVE site).9 SMAT-CE is an EPA software tool which implements the procedures in the SIP
Modeling Guidance to project visibility to a future year.10
3.2 Calculation of Visibility
The visibility calculations use the "revised" IMPROVE equation (Pitchford et al., 2007), which
has been used in most regional haze SIPs over the last 10 years. The IMPROVE equation (or
algorithm) uses PM species concentrations and relative humidity data to calculate visibility
impairment or beta extinction (bext) in units of inverse megameters (Mm4) as follows:
bext = 2.2 x fs(RH) x [Small Sulfate] + 4.8 x fi_(RH) x [Large Sulfate]
+ 2.4 x fs(RH) x [Small Nitrate] + 5.1 x fi_(RH) x [Large Nitrate]
+ 2.8 x {Small Organic Mass] + 6.1 x [Large Organic Mass]
+ 10 x [Elemental Carbon]
+ 1 x [Fine Soil]
+ 1.7 x fss(RH) x [Sea Salt]
+ 0.6 x [Coarse Mass]
+ Rayleigh Scattering (site specific)
The total sulfate, nitrate, and organic mass concentrations are each split into two fractions,
representing small and large size distributions of those components. Site-specific Rayleigh
scattering is calculated based on the elevation and annual average temperature of each
IMPROVE monitoring site. See Hand, 2006 for more details.
9	The base year (2014-2017) IMPROVE data for the 20% most impaired and 20% clearest days was
calculated based on the EPA recommended method described in "Technical Guidance for the Second
Implementation Period of the Regional Haze Rule." (December 2018).
10	SMAT-CE is available here: https://www.epa.gov/scram/photochemical-modeling-tools
15

-------
3.2.1 2000-2018 Visibility
The 2016 base year visibility on the 20% most anthropogenically impaired days and 20%
clearest days at each Class I area is estimated by using observed IMPROVE data. The 2000-
2018 average annual visibility for the 20% most anthropogenically impaired days is also
estimated for each year.
Figures 3-2 to 3-8 below display stacked bar charts detailing the composition of PM2.5 on the
20% most impaired and clearest days for light extinction (bext-1) at each IMPROVE monitoring
site for the base year 2016. The plots also depict the annual average composition of PM2.5 for
light extinction from 2000-2018. The plots below are organized by region and display the
amount of light extinction due to each species as follows: amount of total particle mass using
concentrations of coarse mass, crustal (soil), ammonium nitrate (N03), ammonium sulfate
(S04), elemental carbon (EC), organic mass carbon (OMC), and sea salt.
Alaska
Alaska has four Class I areas: Denali National Park, Tuxedni National Wildlife Refuge,
Simeonof Wilderness Area, and Bering Sea Wilderness Area. There are two IMPROVE monitors
associated with Denali National Park - the monitor designated DENA that is across the Park
Road from park headquarters and the monitor designated TRCR that is located west of
Trapper Creek, Alaska. The DENA monitor is the official monitor for Denali National Park.
Tuxedni National Wildlife Refuge is located on a fairly isolated pair of islands in Tuxedni Bay,
Cook Inlet in Southcentral Alaska. The original IMPROVE monitor, designated TUXE, for
Tuxedni National Wildlife Refuge was installed near Lake Clark National Park to represent
conditions at Tuxedni Wilderness Area. This site is on the west side of Cook Inlet,
approximately 5 miles from the Tuxedni Wilderness Area. However, in 2014 the property
owner and site operator could no longer service the site. A new site, designated KPBO (Kenai
Peninsula Borough), was establish roughly 3 miles south of the community of Ninilchik.
Simeonof Wilderness Area comprises 25,141 acres located in the Aleutian Chain, 58 miles
from the mainland. It is one of 30 islands that make up the Shumagin Group on the western
edge of the Gulf of Alaska. An IMPROVE monitor, designated SIME, in the community of Sand
Point has been deemed as representative of the wilderness area. Sand Point is approximately
60 miles northwest of the Simeonof Wilderness Area. The Bering Sea Wilderness Area is
located off the coast of Alaska about 350 miles southwest of Nome. Hall Island is at the
northern tip of the larger St Matthew Island. Due to access difficulties, there is no IMPROVE
monitor representing the Bering Sea Wilderness Area.
16

-------
Denali NR AK

2.0

1.5

1.0

0.5
£ 0.0
2

r

o

o

c
40
x

LU
30

20

10

0
20% Clearest Days (2016)
I
ll
I
—|—>—i—i—i—|—r
101	201
Julian Day
_ 20% Most Impaired Days (2016)
I
S.I 1 '
1	'	'	1	'	1	'	'	'	T	[¦
1	101	201
Julian Day
301
¦ H! li •
301
40-
30-
20-
10-
20% Most Impaired Days Sea Salt
Coarse
Crustal
EC
Organic
Amm N03
ii^gfi'M'siij
I l I I I I I I l I I I I I I I I I I
2000 2005 2010 2015
Year
Figure 3-2 Stacked bar charts detailing the composition of PM2.5 in 2016 on the 20%
clearest days (top left) and 20% most impaired days (bottom left) for light extinction at Denali
National Park. The right bar chart details the average composition for 2000-2018for the 20%
most impaired days. The plots display the amount of light extinction due to each species as
follows from bottom to top: ammonium sulfate (yellow), ammonium nitrate (red), organic mass
carbon (green), elemental carbon (black), crustal mass (grey), coarse mass (brown), and sea salt
(blue).
2.5-
2.0-
1.5-
1.0-
— 0.5
£ 0.0
60-
50-
40-
30-
20-
10-
0
Trapper Creek, AK
20% Clearest Days (2016)
'fli If
101
—¦—1—i—r
201
Julian Day
301
20% Most Impaired Days (2016)

i 11 i
"•'ll Sihil ,l , i-
60
50-
40-
30-
20-
10-
0'
20% Most Impaired Days
Sea Salt
Coarse
Crustal
EC
Organic
Amm N03
¦lii'i'Vi1'!
Hi
1
101
301
rr
2000
i i i i i i i m m i i i m
2005 2010 2015
Year
201
Julian Day
Figure 3-3 Stacked bar charts detailing the composition of PM2.5 in 2016 on the 20%
clearest days (top left) and 20% most impaired days (bottom left) for light extinction at Trapper
Creek. The right bar chart details the average composition for 2000-2018 for the 20% most
impaired days. The plots display the amount of light extinction due to each species as follows
from bottom to top: ammonium sulfate (yellow), ammonium nitrate (red), organic mass carbon
(green), elemental carbon (black), crustal mass (grey), coarse mass (brown), and sea salt (blue).
17

-------
Kenai Peninsula Borough, AK
20% Clearest Days (2016)
101
	'	'	1	r
201
Julian Day
Julian Day
301
20% Most Impaired Days (2016)
W 50-
70-
60
50
40-
30-
20-
10
20% Most Impaired Days
Sea Salt
Coarse
Crustal
EC
Organic
Amm N03
mfrVn1 U
0 1II I II I I I I I I I I I I I I II
2000 2005 2010 2015
Year
Figure 3-4 Stacked bar charts detailing the composition of PM2.5 in 2016 on the 20%
clearest days (top left) and 20% most impaired days (bottom left) for light extinction at Kenai
Peninsula Borough. The right bar chart details the average composition for 2000-2018 for the
20% most impaired days. Data up through 2014 were measured at the TUXE site, data for 2016
and after were measured at the KPBO site. The plots display the amount of light extinction due
to each species as follows from bottom to top: ammonium sulfate (yellowj, ammonium nitrate
(red), organic mass carbon (green), elemental carbon (black), crustal mass (grey), coarse mass
(brown), and sea salt (blue).
Simeonof, AK
15-
10-
5-
£ 0-
20% Clearest Days (2016)
101
—'—1—i—r
201
Julian Day
301
g100-j
60-
40-
20-
0-
20% Most Impaired Days (2016)
111 'i' i|i i ''i1 In
100-
80-
60-
40-
20-
0-
20% Most Impaired Days
Sea Salt
Coarse
Crustal
EC
Organic
Amm N03

1
101
301
I I I I I I I I I I I I I I I I I I I
2000 2005 2010 2015
Year
201
Julian Day
Figure 3-5 Stacked bar charts detailing the composition of PM2.5 in 2016 on the 20%
clearest days (top left) and 20% most impaired days (bottom left) for light extinction at
Simeonof. The right bar chart details the average composition for 2000-2018 for the 20% most
impaired days. The plots display the amount of light extinction due to each species as follows
from bottom to top: ammonium sulfate (yellow), ammonium nitrate (red), organic mass carbon
(green), elemental carbon (black), crustal mass (grey), coarse mass (brown), and sea salt (blue).
18

-------
Hawaii
Hawaii has IMPROVE monitors at Haleakala National Park (HACR1 and HALE1) and Hawaii
Volcanoes National Park (HAVOl). The HALE1 IMPROVE monitor began operation on Maui in
1990 at a site about 3.5 miles outside of Haleakala National Park. In 2007, a second IMPROVE
monitor (HACR1) was installed at a higher elevation within the park. The HACR1 site was
considered more representative of visibility conditions within Haleakala National Park and
replaced the HALE1 monitoring station in 2012. See 84 FR 14634 (April 11, 2019). The
extinction data presented below indicate a combined site record using conditions at HALE1
from 2000-2007 and 2008-2018 conditions at HACR1. This combined site record is the EPA
default and may not reflect the method of combining IMPROVE monitors representing
Haleakala National Park in future SIPs.
The identification of the 20% most impaired days for these Hawaii IMPROVE monitors is based
on a modification of the statistical approach detailed in the 2018 Technical Guidance on
Tracking Visibility Progress for the Second Implementation Period of the Regional Haze
Program. This 2018 Technical Guidance described an approach that screens out natural
episodic events with high haze levels related to wildfire (based on organic and elemental
carbon) or dust storm impacts (based on fine crustal and coarse mass) that are frequently
experienced at Class I areas in western half of the Continental U.S. Although this approach
effectively screens out natural episodic events at most Class I areas, it is insufficient for the
Class I areas in Hawaii which have visibility conditions that are often impacted by volcanic
emissions. For the two Class I areas in Hawaii, this approach was modified to also screen out
natural episodic events related to volcanic activity (based on sulfate) using the same method
used for wildfires and dust storms (episodic threshold determined by the lowest annual 95th
percentile daily extinction from 2000-2014 at each site). Extinction values from the new set of
20% most impaired days following this modified approach are shown in Figure 3-6 for
combined HALE1/HACR1 site and in Figure 3-7 for HAVOl. Note that this modified approach
didn't affect the 20% clearest days.
19

-------

3.0

2.5

2.0

1.5

1.0
,r-
0.5
E 0.0


c

g

o
c=
40
X
LLJ
30

20

10

0
20% Clearest Days (2016)
Haleakala NP, HI (minimizing volcanic impact)
40
I
I | !
101	201
Julian Day
301
20% Most Impaired Days (2016)
30-
20-
10-
20% Most Impaired Days Sea Salt
Coarse
Crustal
EC
Organic
Amm N03
ml!1!
H'l'1!1*!!!! ••,!
201
Julian Day
I I I I I I I I I I I I I I I I I I I
2000 2005 2010 2015
Year
Figure 3-6 Stacked bar charts detailing the composition of PM2.5 in 2016 on the 20%
clearest days (top left) and 20% most impaired days (bottom left) for light extinction at
Haleakala National Park. The right bar chart details the average composition for 2000-2018for
the 20% most impaired days. The plots display the amount of light extinction due to each
species as follows from bottom to top: ammonium sulfate (yellow), ammonium nitrate (red),
organic mass carbon (green), elemental carbon (black), crustal mass (grey), coarse mass
(brown), and sea salt (blue).
Hawaii Volcanoes NP, HI (minimizing volcanic impact)
20% Clearest Days (2016)
201
Julian Day
20% Most Impaired Days (2016)
x. 80
100-
80-
60-
40-
20-
20% Most Impaired Days Sea Salt
Coarse
Crustal
EC
Organic
Amm N03

201
Julian Day
I I I l I II I I I I l I I I I I l I
2000 2005 2010 2015
Year
Figure 3-7 Stacked bar charts detailing the composition of PM2.5 in 2016 on the 20%
clearest days (top left) and 20% most impaired days (bottom left) for light extinction at Hawaii
Volcanoes National Park. The right bar chart details the average composition for 2000-2018 for
the 20% most impaired days. The plots display the amount of light extinction due to each
species as follows from bottom to top: ammonium sulfate (yellow), ammonium nitrate (red),
organic mass carbon (green), elemental carbon (black), crustal mass (grey), coarse mass
(brown), and sea salt (blue).
20

-------
Virgin Islands
The Virgin Islands has one IMPROVE monitor, located at Virgin Islands National Park (VIIS1).
25-
20 -j
15-
10-E
— 5-
E 0-
C
o
o
c
Virgin Islands NR VI
20% Clearest Days (2016)
"
—|	1	1	1	1	1	r
101	201
Julian Day
301
"5 60-
LLJ	J
40 -|
20-
0-
20% Most Impaired Days (2016)
-
!lll! !i!
1
T
101
n	1	1	r
T
T
301
60-
40-
20-
0-
20% Most Impaired Days
Sea Salt
Coarse
Crustal
EC
Organic
Amm N03
¦¦ ¦ ¦ "Hi
I I I I I I I I I I I I I I I I I I
2000 2005 2010 2015
Year
201
Julian Day
Figure 3-8 Stacked bar charts detailing the composition of PM2.5 in 2016 on the 20%
clearest days (top left) and 20% most impaired days (bottom left) for light extinction at Virgin
Islands National Park. The right bar chart details the average composition for 2000-2018 for the
20% most impaired days. The plots display the amount of light extinction due to each species as
follows from bottom to top: ammonium sulfate (yellow), ammonium nitrate (red), organic mass
carbon (green), elemental carbon (black), crustal mass (grey), coarse mass (brown), and sea salt
(blue).
3.2.2 2028 Visibility
The visibility projections follow the procedures in section 5 of the SIP Modeling Guidance.
Based on the recommendation in the modeling guidance, the observed base period visibility
data is linked to the base modeling year. This is the 5-year ambient data base period
centered about the base modeling year. In this case, for a base modeling year of 2016, the
21

-------
ambient IMPROVE data should be from the 2014-2018 period.11 However, since 2018
IMPROVE data was not available in the attainment test software tool, the most recent four-
year average 2014-2017 base period was used.
The 2028 future year visibility on the 20% most anthropogenically impaired days and 20%
clearest days at each Class I area is estimated by using the observed IMPROVE data (2014-
2017) and the relative percent modeled change in PM species between 2016 and 2028. The
process is described in the following six steps (see the SIP Modeling Guidance for a more
detailed description and examples).
1)	For each Class I area (IMPROVE site), estimate anthropogenic impairment on each day
using observed speciated PM2.5 data plus PM10 data (and other information) for each of
the 5 years comprising the base period (four years, 2014-2017 in this case) and rank the
days on this indicator.12 This ranking will determine the 20 percent most
anthropogenically impaired days. For each Class I area, also rank observed visibility (in
deciviews) on each day using observed speciated PM2.5 data plus PM10 data for each of
the 5 years comprising the base period. This ranking will determine the 20 percent
clearest days.
2)	For each of the 5 years comprising the base period, calculate the mean deciviews for the
20 percent most anthropogenically impaired days and 20 percent clearest days. For
each Class I area, calculate the 5 year mean deciviews for most impaired and clearest
days from the 5 year-specific values.
3)	Use an air quality model to simulate air quality with base period (2016) emissions and
future year (2028) emissions. Use the resulting information to develop site-specific
relative response factors (RRFs) for each component of PM13 identified in the "revised"
11	The baseline period for the regional haze program continues to be 2000-2004, and the uniform rate of
progress is calculated using that historical data. However, the modeled visibility projections should use
ambient data from a 5-year base period that corresponds to the modeled base year meteorological and
emissions data. Also, unlike the ozone and PM2.5 attainment tests, the ambient data averaging
calculation is a 5-year mean, where each year counts equally (unlike the 5-year weighted average values
recommended for the ozone and PM2.5 attainment test).
12	The EPA recommended methodology for determining the most anthropogenically impaired days (which
includes the explanation of how anthropogenic vs. natural daily light extinction was determined) can be
found in Technical Guidance on Tracking Visibility Progress for the Second Implementation Period of the
Regional Haze Program.
13	Relative response factors (RRFs) are calculated for sulfate, nitrate, organic carbon mass, elemental
carbon, fine soil mass, and coarse mass. Since observed sea salt is primarily from natural sources which
are not expected to be year-sensitive, and the modeled sea salt is uncertain, the sea salt RRF for all sites
is assumed to be 1.0.
22

-------
IMPROVE equation. The RRFs are an average percent change in species concentrations
based on the measured 20% most impaired and 20% clearest days from 2016 (the
calendar days from 2016 identified from the IMPROVE data above are matched by day
to the modeled days).
4)	Multiply the species-specific RRFs by the measured daily species concentration data
during the 2014-2017 base period (for each day in the measured 20% most impaired day
set and each day in the 20% clearest day set), for each site. This results in daily future
year 2028 PM species concentration data.
5)	Using the results in Step 4 and the IMPROVE algorithm, calculate the future daily
extinction coefficients for the previously identified 20 percent most impaired days and
20 percent clearest days in each of the five base years.
6)	Calculate daily deciview values (from total daily extinction) and then compute the future
year (2028) average mean deciviews for the 20 percent most impaired days and 20
percent clearest days for each year. Average the five years together to get the final
future mean deciview values for the 20 percent most impaired days and 20 percent
clearest days.
The SMAT-CE tool outputs individual year and 5-year average base year and future year
deciview values on the 20% most impaired days and 20% clearest days. Additional SMAT output
variables include the results of intermediate calculations such as species-specific extinction
values (both base and future year) and species specific RRFs (on the 20% most impaired and
clearest days).
Table 3-1 details the settings used for the SMAT runs to generate the 2028 future year
deciview projections:
Table 1-1. SMAT settings for 2028 visibility calculations
SMAT Option
Setting or File Used
IMPROVE algorithm
Use new version
Grid cells at monitor or Class 1 area centroid?
Use grid cells at monitor
Temporal adjustment at monitor
3x3
Start monitor year
2014
End monitor year
2017
Base Model year
2016
Minimum years required for a valid monitor
1
23

-------
Table 3-2 shows the base and future year deciview values on the 20% clearest and most
impaired days at each Class I area for the base model period (2014-2017) and future year
(2028).
Table 3-2. Base and future year deciview values on the 20% clearest and 20% most impaired
days at each Class I area for the base model period (2014-2017) and future year (2028)



Future
Base Year



Base Year
Year
(2014-



(2014-2017)
(2028)
2017) 20%
Future Year


20%
20%
Most
(2028) 20%
Class 1 Area
IMPROVE
Clearest
Clearest
Impaired
Most Impaired
Name
Site ID
Days (dv)
Days (dv)
Days (dv)
Days (dv)
Denali NP
TRCR1
3.34
3.32
8.99
8.95
Denali NP
DENA1
2.19
2.16
6.86
6.84
Haleakala





Crater NP
HALE1/HACR1
0.51
0.50
7.70
7.55
Hawaii





Volcanoes





NP
HAVOl
3.50
3.49
16.31
16.03
Tuxedni





National





Wildlife





Refuge
KPBOl/TUXEl
4.62
4.23
11.43
10.9
Simeonof





Wilderness





Area
SIME1
7.68
7.42
13.86
13.43
Virgin Islands





NP
VIIS1
9.90
9.7
15.45
15.14
3.3 Comparison to Regional Haze "Glidepath"
The future year 2028 deciview projections can be compared to the unadjusted visibility
"glidepath" at each Class I area, as defined above.14 The unadjusted "glidepath" represents
the amount of visibility improvement needed in each implementation period, starting from
the baseline 2000-2004 period, to stay on a linear path to natural visibility conditions by
2064. Visibility on the 20% most impaired days is compared to the relevant value of the
glidepath, in this case for a future year of 2028. Since the glidepath is a linear path between
2004 and 2064, a glidepath value (in deciviews) can be calculated for any future year, using a
14 The projected 2028 visibility level is compared to the "unadjusted" glidepath for each Class I area. In
this calculation, no adjustments have been made for impacts from international anthropogenic sources
or wildland prescribed fires.
24

-------
simple equation. The following formula was used to calculate the 2028 unadjusted glidepath
value:
Glidepath2028 = Baseline avg deciview - (((Baseline avg deciview - Natural
conditions)/60)*24)
where
Baseline avg deciview = average observed deciview value on the 20% most impaired
days for 2000-2004 (in dv)
Natural conditions= Natural conditions on the 20% most impaired days at the Class I
area (in dv)
Table 3-3 shows the 2028 glidepath values (in dv) at each Class I area, including the data
needed to calculate the glidepath (natural conditions and the 2000-2004 baseline deciview
values).15 Both "adjusted" and "unadjusted" glidepath values for 2028 are also provided. The
observed 2014-2017 values and projected 2028 values are repeated from Table 3-2.
Table 3-3 Natural conditions, 2000-2004 baseline visibility, 2028 projected visibility, and 2028
glidepath values (all in deciviews).
Class 1 Area Name
State
IMPROVE
Site ID
Observed
00-04
Baseline
20% Most
Impaired
Days(dv)
Projected
2028
Impairment
20% Most
Impaired
Days(dv)
2028
Unadjusted
Glidepath
20% Most
Impaired
Days(dv)
2028
Adjusted
Glidepath
20% Most
Impaired
Days(dv)
Natural
Conditions
20% Most
Impaired
Days (dv)
Adjusted
Natural
Conditions
20% Most
Impaired
Days (dv)
Denali NP
AK
TRCR1
9.16
8.95
8.05
8.52
6.38
7.55
Denali NP
AK
DENA1
7.06
6.84
6.15
6.47
4.79
5.60
Haleakala Crater NP
HI
HALE1/
HACR1
10.94
7.55
8.73
9.93
5.41
8.43
Hawaii Volcanoes NP
HI
HAVOl
15.6
16.03
12.01
15.06
6.62
14.26
Tuxedni National
Wildlife Refuge
AK
KPBOl/
TUXE1
10.47
10.9
9.07
10.25
6.96
9.92
Simeonof Wilderness
Area
AK
SIME1
13.67
13.43
11.6
13.35
8.49
12.86
Virgin Islands NP
VI
VIIS1
14.29
15.14
11.99
13.05
8.53
11.2
15 The values for the 20% most impaired and clearest days and natural conditions are calculated
according to the draft recommended method in the draft EPA guidance document "Draft Guidance for
the Second Implementation Period of the Regional Haze Rule" posted at
https://www.epa.gov/visibility/regional-haze-guidance-technical-support-document-and-data-file.
25

-------
The 2028 future year projected deciview values can be compared to the unadjusted glidepath
for 2028. While the RHR requires future year projected visibility impairment be compared to
the glidepath, it does not require the RPGs be on or below the glidepath. However, the rule
has different requirements depending on whether the projected value (RPG) is above or
below the glidepath.16 The RHR provides flexibility regarding adjustments of the glideslope
related to international and natural contribution. Details about the approach used to
estimate an adjusted glideslope are provided in section 3.4. Glideslopes are shown for each
of the Class I areas in Figures 3-9-1 to 3-9-3.
C\J _

Haleakala Crater National Park
o _



CO -
*••••* •
••• •
•

CD -
O
•

-
o
o
o
o
o
o
o


CM -
o -
0
o oo0o°o°°o
0
•	20% Most Impaired o 20% Clearest U.S. zero-out
o 20% Clearest A Unadjusted glidepath
•	20% Most Impaired U.S. zero-out ¦ Adjusted glidepath
2000	2010	2020	2030	2040	2050	2060

Hawaii Volcanos National Park

••i! • •«•••
•


• • —	
00°000°°00oo000000
•


8
•	20% Most Impaired
o 20% Clearest
•	20% Most Impaired U.S. zero-out
o 20% Clearest U.S. zero-out
A Unadjusted glidepath
¦ Adjusted glidepath
2000	2010	2020	2030	2040	2050	2060
Figure 3-9-1. Unadjusted (blue triangle) and adjusted (orange square) glidepath (in deciviews)
at each Class I area in Hawaii. The closed black circles represent the 20% most impaired days
and the open black circles are the 20% clearest days. The green dots represent the 2028 RPG
for the sensitivity simulation where all U.S. emissions were zeroed-out.
16 See 40 CFR 51.308(f)(3)(ii) and (iii)
26

-------
Tuxedni National Wildlife Refuge

• • • •





• •
•
°o
• Oo
o°oOo00o°o0o0o


• 20% Most Impaired
20% Clearest U.S. zero-out

o 20% Clearest *
Unadjusted glidepath

• 20% Most Impaired U.S. zero-out
Adjusted glidepath
2000	2010	2020	2030	2040	2050	2060

Simeonof Wilderness Area
0°o0o°°o0°°oo0oo00
0

•	20% Most Impaired o 20% Clearest U.S. zero-out
o 20% Clearest » Unadjusted glidepath
•	20% Most Impaired U.S. zero-out Adjusted glidepath
2000
2010
2020
2030
2040
2050
2060
t • Denali National Park (Trapper Creek monitor)
• . • •
• • •
' 		—_	-
°°o°o 00 °°0o°°

•	20% Most Impaired 20% Clearest U.S. zero-out
o 20% Clearest A Unadjusted glidepath
•	20% Most Impaired U.S. zero-out Adjusted glidepath
2000
2010
2030
2050
co —
(/) CD —
a)
">
o ^ -
0
~0
CM -
O -
Figure 3-9-2. Unadjusted (blue triangle) and adjusted (orange square) glidepath (in deciviews)
at each Class I area in Alaska. The closed black circles represent the 20% most impaired days
and the open black circles are the 20% clearest days. The green dots represent the 2028 RPG
for the sensitivity simulation where all U.S. emissions were zeroed-out.
Denali National Park (Denali monitor)
°o0o°0oooo°ooo00o0o
2000
2010
20% Most Impaired
20% Clearest
20% Most Impaired U.S. zero-out
2020
—I	
2030
	1	
2040
20% Clearest U.S. zero-out
A Unadjusted glidepath
Adjusted glidepath
—I—
2050
	1	
2060
27

-------
in _
(/)
£
CD o _
> T"
'o
0
"O
iO -
o -
Figure 3-9-3. Unadjusted (blue triangle) and adjusted (orange square) glidepath (in deciviews)
at each Class I area in the Virgin Islands. The closed black circles represent the 20% most
impaired days and the open black circles are the 20% clearest days. The green dots represent
the 2028 RPG for the sensitivity simulation where all U.S. emissions were zeroed-out.
Virgin Islands National Park
• •
• ••
oo
o
Oo
•	20% Most Impaired	o 20% Clearest U.S. zero-out
o 20% Clearest ± Unadjusted glidepath
•	20% Most Impaired U.S. zero-out ¦ Adjusted glidepath
2000	2010	2020	2030	2040	2050	2060
3.4 Contribution from International & U.S. anthropogenic sources
Visibility at Class I areas is impacted not only by natural and anthropogenic emissions from
within the U.S., but also by natural and anthropogenic international emissions. Due to the fact
that international anthropogenic emissions are beyond the control of states preparing regional
haze SIPs, the Regional Haze Rule allows states to optionally propose an adjustment of the 2064
URP endpoint to account for international anthropogenic impacts, if the adjustment has been
developed using scientifically valid data and methods.17 The URP can be adjusted by adding an
estimate of the visibility impact of international anthropogenic sources to the value of the
natural visibility conditions to get an adjusted 2064 endpoint. See the Technical Guidance on
Tracking Visibility Progress18 for more details. The regional haze rule also allows for an optional
adjustment to the URP relating to certain prescribed fires. However, since prescribed fire
activity is anticipated to be uncommon in these areas in 2028, only international anthropogenic
contribution was considered as part of this analyses.
The EPA modeling calculates estimated Class I area (IMPROVE site) contributions from
17 See 40 CFR 51.308(f)(l)(vi)
18'Technical Guidance on Tracking Visibility Progress for the Second Implementation Period of the Regional Haze
Program", December 20, 2018, available at: https://www.epa.gov/visibility/technical-guidance-tracking-visibility-
progress-second-implementation-period-regional
28

-------
international anthropogenic emissions using hemispheric scale CMAQ zero-out model
simulations. The hemispheric CMAQ zero-out simulations provided an estimate of
international anthropogenic SO2 emissions to sulfate related extinction. The estimate of
international anthropogenic sulfate was added to the 2064 goal at each of these Class I areas
to provide an alternative, or "adjusted" glideslope. Other international anthropogenic
emissions were not added to the 2064 goal for several reasons: because non-linearity of
secondary organics and nitrate are difficult to interpret, because sulfate was the dominant
component of observed visibility at these Class I areas, and commercial shipping is the largest
component of the global inventory near these Class I areas.
The estimate of international anthropogenic contribution is based on 2016 emissions and may
not reflect all anticipated reductions in certain sectors such as commercial marine.
Commercial marine emissions are expected to be lower in 2028 than 2016 which means this
assumption may be over-stating international contribution to the 2064 endpoint. This
adjustment to the international contribution would likely result in a smaller increment added
to the 2064 goal and a steeper adjusted glideslope. However, the analysis is not considering
the contribution of international emissions to nitrate or primary PM2.5 components. The
inclusion of these species might increase the international contribution and increase the
increment added to the 2064 goal to some extent.
Additional information about international and U.S. anthropogenic emission contribution was
provided by model simulations where U.S. anthropogenic emissions were zeroed out. U.S.
anthropogenic sources were any point, mobile, or area source located in the U.S. or
territories. This included Class 1 and 2 commercial marine vessels but not Class 3 vessels.
4.0 References
Baker, K., Nguyen, T., Sareen, N., Henderson, B., 2020. Meteorological and air quality modeling
for Hawaii, Puerto Rico, and Virgin Islands. Atmospheric Environment, 117543.
Baker, K., Woody, M., Tonnesen, G., Hutzell, W., Pye, H., Beaver, M., Pouliot, G., Pierce, T.,
2016. Contribution of regional-scale fire events to ozone and PM 2.5 air quality estimated by
photochemical modeling approaches. Atmospheric Environment 140, 539-554.
Gantt, B., Kelly, J., Bash, J., 2015. Updating sea spray aerosol emissions in the Community
Multiscale Air Quality (CMAQ) model version 5.0. 2. Geoscientific Model Development 8, 3733-
3746.
29

-------
Guenther, A., Karl, T., Harley, P., Wiedinmyer, C., Palmer, P., Geron, C., 2006. Estimates of
global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols
from Nature).
National Centers for Environmental Prediction, 2015. NCEP GFS 0.25 Degree Global Forecast
Grids Historical Archive. Research Data Archive at the National Center for Atmospheric
Research, Computational and Information Systems Laboratory. Dataset.
https://doi.org/10.5065/D65D8PWK. Accessed October 2018.
National Emissions Inventory Collaborative, 2020. 2016 Emissions Modeling Platform. Retrieved
from http://views.cira.colostate.edu/wiki/wiki/10202.
Otte, T., Pleim, J., 2010. The Meteorology-Chemistry Interface Processor (MCIP) for the CMAQ
modeling system: updates through MCIPv3. 4.1. Geoscientific Model Development 3, 243-256.
Pitchford, M., Malm, W., Schichtel, B., Kumar, N., Lowenthal, D., Hand, J., 2007. Revised
algorithm for estimating light extinction from IMPROVE particle speciation data. Journal of the
Air & Waste Management Association 57,1326-1336.
Pouliot, G., Rao, V., McCarty, J.L., Soja, A., 2017. Development of the crop residue and
rangeland burning in the 2014 National Emissions Inventory using information from multiple
sources. Journal of the Air & Waste Management Association 67, 613-622.
Sarwar, G., Gantt, B., Schwede, D., Foley, K., Mathur, R., Saiz-Lopez, A., 2015. Impact of
enhanced ozone deposition and halogen chemistry on tropospheric ozone over the Northern
Hemisphere. Environmental science & technology 49, 9203-9211.
U.S. Environmental Protection Agency, 2018. Modeling Guidance for Demonstrating Attainment
of Air Quality Goals for Ozone, PM2.5, and Regional Haze. EPA-454/R-18-009.
https://www3.epa.gov/ttn/scram/guidance/guide/O3-PM-RH-Modeling_Guidance-2018.pdf.
U.S. Environmental Protection Agency, 2019a. 2016 Hemispheric Modeling Platform Version 1:
Implementation, Evaluation, and Attribution. Research Triangle Park, NC. U.S. Environmental
Protection Agency. U.S. EPA.
U.S. Environmental Protection Agency, 2019b. Preparation of Emissions Inventories for the
Version 7.2 2016 Hemispheric Emissions Modeling Platform. Research Triangle Park, NC. U.S.
Environmental Protection Agency. U.S. EPA.
U.S. Environmental Protection Agency, 2019c. Technical Support Document for EPA's Updated
2028 Regional Haze Modeling, https://www.epa.gov/visibility/technical-support-document-
epas-updated-2028-regional-haze-modeling.
U.S. Environmental Protection Agency, 2020a. 2016 Alaska State-wide Weather Research
Forecast (WRF) Meteorological Model Performance Evaluation. EPA-454/R-20-003.
30

-------
U.S. Environmental Protection Agency, 2020b. 2017 National Emissions Inventory (NEI)
Techincal Support Document, https://www.epa.gov/air-emissions-inventories/2017-national-
emissions-inventory-nei-technical-support-document-tsd
Wiedinmyer, C., Akagi, S., Yokelson, R.J., Emmons, L., Al-Saadi, J., Orlando, J., Soja, A., 2011. The
Fire INventory from NCAR (FINN): A high resolution global model to estimate the emissions
from open burning. Geoscientific Model Development 4, 625.
Zhao, B., Zheng, H., Wang, S., Smith, K.R., Lu, X., Aunan, K., Gu, Y., Wang, Y., Ding, D., Xing, J.,
2018. Change in household fuels dominates the decrease in PM2. 5 exposure and premature
mortality in China in 2005-2015. Proceedings of the National Academy of Sciences 115, 12401-
12406.
31

-------
Appendix A Model Performance Evaluation - Alaska
A.l. Spatial Plots of Average Model Predictions on the Mc taired Days
The plots in this section show average daily average measurements and model predictions for
the major components of total PM2.5 mass aggregated over the most impaired days in 2016 at
each of the Class I areas (left panels). The difference between the 2028 and 2016 simulations
are also shown (middle panels) and the difference between the 2028 simulation and 2028
simulation with zero U.S. anthropogenic emissions (right panels).
A-l

-------
PM2.5 organic carbon
PM2.5 organic carbon
PM2.5 organic carbon
0.010
2028 no
028 -2016

ODEN41
QTRCR
- 0.000
TRCR
- -0.005
£
" 0.2

LigC/m
PM2.5 sulfate ion
PM2.5 sulfate ion
PM2.5 sulfate ion
2028 no
028 -2016
- 0.3
notN
ODENy\1
QTRCR
- 0.00
QTRCR
QTRCR
9
X
-0.02
PM2.5 nitrate ion
PM2.5 nitrate ion
PM2.5 nitrate ion
o	. is
028-2016
2028 no
QDEN
QDENA1
QDEN
- 0.00
QTRCR
QTRCR
TRCR
-0.02
ug/m
lig/m
Figure A-l Spatial plots showing the average model predictions for PM2.5 organic carbon
(top three panels), PM2.5 sulfate ion (middle three panels), and PM2.5 nitrate ion (bottom three
panels) on the 20% worst days at Denali National Park. The left panels depict concentrations for
year 2016, the middle panels depict the differences in concentrations for year 2028 - 2016, and
the right panels depict the differences in concentrations for year 2028 with no anthropogenic
influences minus year 2028.
A-2

-------
PM2.5 organic carbon
2016
PM2.5 organic carbon
iTHCfl
PM2.5 organic carbon
2023rKi U

ThCH
PM2.5 sulfate ion
® 2828 m 1 "	-"TP

O thch

L.
Figure A-2 Spatial plots showing the average model predictions for PM2.5 organic carbon
(top three panels), PM2.5 sulfate ion (middle three panels), and PM2.5 nitrate ion (bottom three
panels) on the 20% worst days at Trapper Creek. The left panels depict concentrations for year
2016, the middle panels depict the differences in concentrations for year 2028 - 2016, and the
right panels depict the differences in concentrations for year 2028 with no anthropogenic
influences minus year 2028.
o-- 1
OTRcn-
¦¦ 04
-	Q2
-	0 2
r01
¦L Q Q
MflAir
PM2.5 sulfate ion
QD€NAi
OthcrI
¦- 0 .02
-	0 01
-	0 .00
j - -Q.Q1
*0 02
*ig/m3
PM2.5 nitrate ion
2016
OdcnAi
OthcrI
PM2.5 nitrate ion
r
-	Q.01
-	0.03
I - -Q.01
,0.02
jig/m3
PM^.b nil ml a on
I THCH
A-3

-------
PM2.5 organic carbon
PM2.5 organic carbon
PM2.5 organic carbon
i ii ¦ i
2a2a ru U.S. 3-1*1



2016
020 - 2016
- 0 035
OOLNA1
'I 'AM

THCR
0 'Al 5
PM2.5 sulfate on
PM2.5 sulfate ton
PM2.5 sulfate ion
2028 ra US antl -


028 - 2016
QQ EN.^1
qtrch
- 0 0G

ligfm'
jig/m
PM2.5 nitrate ton

2016
QDLNA1
O iflcfl
.ug/m'
0.3Q
s
0 25
,v
020
V
0.15
r %
0.10

0.05
«
£j.0Q

PM2.5 nitrate ion
028 - 2016
ODCN-V
THCH
PM2.S nitrate ion
I
.ug/m
QDLNAl
¦men!
Figure A-3 Spatial plots showing the average model predictions for PM2.5 organic carbon
(top three panels), PM2.5 sulfate ion (middle three panels), and PM2.5 nitrate ion (bottom three
panels) on the 20% worst days at Kenai Peninsula Borough.
A-4

-------
PM2.5 organic carbon
2016
0M1
QTRCRl
PM2.5 sulfate ion
QD€NAi
OthcrI
PM2.5 nitrate ion
PM2.5 organic carbon
iTHCfl
If 0X110
-	Q-005
-	0.000
U- -Q.005
-0.010
jigC/m
|- 0 .02
-	0 01
-	0 .00
j - -Q.Q1
*0 02
Jig/m3
r
-	Q.01
-	0.03
I - -0.01
,0.02
jig/m3
PM2.5 organic carbon
iTHCH
PM2.5 sulfate ton
ITHCH
I OXM
-	002
-	OjM
...
I- -0 0
ngC/m3
2016
p
-	OJ0
I - 0 6
-	Q.4
l " 02
¦L of
LigC/m
r
030
025
020
0.15
0.10
QQ5
ixg/m
PM2.5 nitrate ion
men
I,
M3''m
Figure A-4 Spatial plots showing the average model predictions for PM2.5 organic carbon
(top three panels), PM2.5 sulfate ion (middle three panels), and PM2.5 nitrate ion (bottom three
panels) on the 20% worst days at Simeonof Wilderness Area,
A-5

-------
A.2 Time Series for 2016
The plots in this section show daily average measurements and model predictions for the major
components of total PM2.5 mass at each of the Class I areas.


+ Observation
DENA1

0 Model

+



+ Observation
DENA1

0 Model



Denali NP
DENA1
PM2.5 organic carbon
+ ++ +
*
.+
+ Observation
° Model
01	02	03	04	05	06	07	08	09	10	11	12


+ Observation
DENA1

0 Model



01	02	03	04	05	06	07	08	09	10	11	12

PM2.5 sodium + chloride
+ Observation
DENA1
0 Model



01	02	03	04	05	06	07
09	10



+ Observation
DENA1


0 Model



&®w^
8
§ -
8 "
o _
Figure A-5 Time series plots for 2016 comparing model predictions (blue circles) with
IMPROVE monitor measurements (black crosses) for PM2.5 sulfate ion (top panel), PM2.5 nitrate
ion (second panel), PM2.5 organic carbon (third panel), PM2.5 elemental carbon (fourth panel),
NaCI (fifth panel), and PM coarse (bottom panel) at Denali National Park.
A-6

-------
Trapper Creek (Denali NP)
TRCR1
PM2.5 sulfate ion
-------
o
+ Observation
o Model
Tuxedni NWR
KPBOI
PM2.5 sulfate ion
CO
CD
O
01
02
03
04
05
06
07
08
09
10
11
12
+ Observation
o Model
Tuxedni NWR
KPBOI
PM2.5 nitrate ion
o
o
p
o
01
02
03
04
05
06
07
08
09
10
11
12
Tuxedni NWR
KPBOI
+
PM2.5 organic carbon
++
"h+L

&

+
+
+,
i
+ Observation
o Model
L+,
+
Tuxedni NWR
KPBOI
PM2.5 elemental carbon
+ Observation
o Model
Tuxedni NWR
KPBOI

PM2.5 sodium + chloride


+ Observation
0 Model





+
Tuxedni NWR
KPBOI
tcbal

PM coarse
+++
+ -H-
+
¦M-
-h++
i. +4
LlK
+
+ Observation
o Model
. rtch t
Figure A-7 Time series plots for 2016 comparing model predictions (blue circles) with
IMPROVE monitor measurements (black crosses) for PM2.5 sulfate ion (top panel), PM2.5 nitrate
ion (second panel), PM2.5 organic carbon (third panel), PM2.5 elemental carbon (fourth panel),
NaCI (fifth panel), and PM coarse (bottom panel) at Kenai Peninsula Borough.
A-8
-------
Simeonof WA
SIME1
PM2.5 sulfate ion
d-jj-, J-
A
+ Observation
o Model
Simeonof WA
SIME1

PM2.5 nitrate ion

+ Observation
0 Model

O



Simeonof WA
SIME1
+

PM2.5 organic carbon
+
-rh
+
rl%»
+
o
+
+ Observation
o Model
Simeonof WA
SIME1
PM2.5 elemental carbon
+ Observation
o Model
Simeonof WA
SIME1
+-HJ-
+ + +
±4-^+,rf
PM2.5 sodium + chloride
+
+ Observation
o Model
+ ¦ V
Simeonof WA
SIME1
-h
+
PM coarse
ft
+ Observation
o Model
i&
Figure A-8 Time series plots for 2016 comparing model predictions (blue circles) with
IMPROVE monitor measurements (black crosses) for PMzs sulfate ion (top panel), PM2.5 nitrate
ion (second panel), PM2.5 organic carbon (third panel), PM2.5 elemental carbon (fourth panel),
NaCI (fifth panel), and PM coarse (bottom panel) at Simeonof Wilderness Area.
A-9
-------
A.3 Particulate Matter Composition on Clearest and Most Impaired Days in 2016
The plots in this section show average daily average measurements and model predictions for
the major components of total PM2.5 mass aggregated over the most and least impaired days
in 2016 at each of the Class I areas.
DENA1
5 2.5
S 2.0
E 0.5
Observed	Modeled
20% Most Impaired Days
NaCI
PM Coarse
PM2.5 EC
PM2.5 0C
PM2.5 N03
Observed	Modeled
20% Clearest Days
Figure A-9 Stacked bar charts detailing the average composition of speciated particulate
matter in 2016 on the 20% most impaired days (right) and 20% clearest days (right) at Denali
National Park. The plots display concentration from bottom to top for the following: PM2.5
sulfate (yellow), PM2.5 nitrate (red), PM2.5 organic carbon (green), PM2.5 elemental carbon
(black), PM coarse (brown), and sea salt (blue).
A-10
-------
TRCR1
Observed	Modeled	Observed	Modeled
20% Most Impaired Days	20% Clearest Days
Figure A-10 Stacked bar charts detailing the average composition of speciated particulate
matter in 2016 on the 20% most impaired days (right) and 20% clearest days (right) at Trapper
Creek. The plots display concentration from bottom to top for the following: PM2.5 sulfate
(yellow), PM2.5 nitrate (red), PM2.5 organic carbon (green), PM2.5 elemental carbon (black), PM
coarse (brown), and sea salt (blue).
A-ll
-------
w
E
ro
O)
2
o
E
c
o
CD
L_
CI
<1)
O
c
O
O
KPB01
Observed	Modeled
20% Most Impaired Days
NaCl
PM Coarse
PM2.5 EC
PM2.5 0C
PM2.5 N03
Observed	Modeled
20% Clearest Days
Figure A-ll Stacked bar charts detailing the average composition of speciated particulate
matter in 2016 on the 20% most impaired days (right) and 20% clearest days (right) at Kenai
Peninsula Borough. The plots display concentration from bottom to top for the following: PM2.5
sulfate (yellow), PM2.5 nitrate (red), PM2.5 organic carbon (green), PM2.5 elemental carbon
(black), PM coarse (brown), and sea salt (blue).
A-12
-------
SIME1
Observed	Modeled	Observed	Modeled
20% Most Impaired Days	20% Clearest Days
Figure A-12 Stacked bar charts detailing the average composition of speciated particulate
matter in 2016 on the 20% most impaired days (right) and 20% clearest days (right) at Simeonof
Wilderness Area. The plots display concentration from bottom to top for the following: PM2.5
sulfate (yellow), PM2.5 nitrate (red), PM2.5 organic carbon (green), PM2.5 elemental carbon
(black), PM coarse (brown), and sea salt (blue).
A-13
-------
Appendix B Model Performance Evaluation - Hawaii
B.l Spa'M '! M "ts of Average Model Predictions on the Mc •' !r
-------
PM2.5 orqariic carbon
PM2.5 orqanic carbon
2028 - 2016
- 0 .4
\}
u-oM
ngC/m
PM2.5 sulfate ion
PM2.5 sulfate ion
ug.'m
PM2.S nitrate ion
0"
f
r 05

[• 0 6

- 02

Loo
PM2.5 nitrate ion
it
I
ugC/nn'
:S°
PM2.S orqanic carbon

2029 fKj U.S 81*1
-2028
10







K^I<£R1
$»•
2028 — 2016
- 0.4

Wf'm
ug/m
2028-2016
-	0 04
-	0 .02
-	0.00
-	-0.0

r 0-10

- 0.05

- 0.00

- -0.05

LgO.10
HflAn
PM2.5 sulfate ran
¦ 0.10
2U29 rio U S. arifi - 2029
- 0.00

PM2.5 nitrate on
¦ 0.10
2029 na U .S. crifi - 2029
: ¦ n
- 0 00
jig/m
Figure B-l Spatial plots showing the average model predictions for PM2.5 organic carbon
(top three panelsj, PM2.5 sulfate ion (middle three panels), and PM2.5 nitrate ion (bottom three
panels) on the 20% worst days at Haleakala National Park. The left panels depict concentrations
for year 2016, the middle panels depict the differences in concentrations for year 2028 - 2016,
and the right panels depict the differences in concentrations for year 2028 with no
anthropogenic influences minus year 2028.
B-2
-------
PM2.5 organic carbon
2016

&
PM2.5 oraanic carbon
ugC/m
PM2.5 sulfate ton
ZDie
©S&
PM2.5 nitrate ion
2016

IACRI

2028 — 2016

r QJ010

[¦ a .035

L O-OQO

I- -0.035

L -0.010
PM2.5 organic carbon
2028 no U.S. anil - 2028

I'

ugC/m



-0,

¦ 0.6

- a 2

La
ngmi

r Qi

0.6

- 02

L3.0
PM2.5 sulfate ion
2028-2016
PM2.5 nitrate ion
PM2.5 nitrate ion
2028 no U .S. criti - 2028
2028 - 2016
- 0 05

jig/m
ng,'m
Figure B-2 Spatial plots showing the average model predictions for PM2.5 organic carbon
(top three panels), PM2.5 sulfate ion (middle three panels), and PM2.5 nitrate ion (bottom three
panels) on the 20% worst days at Hawaii Volcanoes National Park.
PM2.5 sulfate ion
no US. ariti - 2028
B-3
-------
B.2 Time Series for 2016
The plots in this section show daily average measurements and model predictions for the major
components of total PM2.5 mass at each of the Class I areas.
Haleakala Crater NP
HACR1


PM2.5 sulfate ion
+ Observation
0 Model

+



\ t+o
I IT' TTI- (
+

i i i i i i i

01	02	03	04	05	06	07
OS	10
Haleakala Crater NP
HACR1
PM2.5 nitrate ion
02	03	04	05
07	08	09	10
Haleakala Crater NP
HACR1
PM2.5 organic carbon
+ Observation
o Model
01	02	03	04	05	06	07
I		 I
10	11	12
Haleakala Crater NP
PM2.5 elemental carbon
+ Observation
HACR1

0 Model

h

01	02	03	04	05
09	10	11	12
Haleakala Crater NP
HACR1
PM2.5 sodium + chloride

+ Observation
0 Model


O





l	I
01	02	03	04	05	06	07
09	10	11	12
Haleakala Crater NP
HACR1

K Observation
> Model
-r	 i 	1— 	1 	r"— —r——r
01	02	03	04	05	06	07
—	'I ——T1
OS	10	11	12
oo
OnO'J'o
tit-
Figure B-3 Time series plots for 2016 comparing model predictions (blue circles) with
IMPROVE monitor measurements (black crosses) for PMzs sulfate ion (top panel), PM2.5 nitrate
ion (second panel), PM2.5 organic carbon (third panel), PM2.5 elemental carbon (fourth panel),
NaCI (fifth panel), and PM coarse (bottom panel) at Haleakala National Park.
B-4
-------
o
+ Observation
o Model
Hawaii Volcanos NP_
HAV01+ _|_
PM2.5 sulfate ion
+ +
CO
+ +
CD
O
01
02
03
04
05
06
07
08
09
10
11
12
o
+ Observation
o Model
Hawaii Volcanos NP
HAVOI
PM2.5 nitrate ion
o
o
o
01
02
03
04
05
06
07
08
09
10
11
12
Hawaii Volcanos NP
HAVOI
+
-to
-£k-
k"^1 ~ i "ft1
PM2.5 organic carbon

+ Observation
o Model
Hawaii Volcanos NP
HAVOI
PM2.5 elemental carbon
+ Observation
o Model
Hawaii Volcanos NP
HAVOI
+ +
PM2.5 sodium + chloride
+
jfc +, .
+
+ Observation
o Model
+
£
+
QD
Hawaii Volcanos NP
HAVOI
PM coarse


cp°a
m
+ Observation
o Model
CP
P„ O _	°
-"to:
O CD
% ®
Figure B-4 Time series plots for 2016 comparing model predictions (blue circles) with
IMPROVE monitor measurements (black crosses) for PM2.5 sulfate ion (top panel), PM2.5 nitrate
ion (second panel), PM2.5 organic carbon (third panel), PM2.5 elemental carbon (fourth panel),
NaCI (fifth panel), and PM coarse (bottom panel) at Hawaii Volcanoes National Park.
B-5
-------
B.3 Particulate Matter Composition on Clearest and Most Impaired Days in 2016
The plots in this section show average daily average measurements and model predictions for
the major components of total PM2.5 mass aggregated over the most and least impaired days
in 2016 at each of the Class I areas.
HACR1
Observed	Modeled	Observed	Modeled
20% Most Impaired Days	20% Clearest Days
Figure B-5 Stacked bar charts detailing the average composition of speciated particulate
matter in 2016 on the 20% most impaired days (right) and 20% clearest days (right) at
Haleakala National Park. The plots display concentration from bottom to top for the following:
PM2.5 sulfate (yellow), PM2.5 nitrate (red), PM2.5 organic carbon (green), PM2.5 elemental carbon
(black), PM coarse (brown), and sea salt (blue).
B-6
-------
HAV01
NaCI
PM Coarse
Observed	Modeled	Observed	Modeled
20% Most Impaired Days	20% Clearest Days
Figure B-6 Stacked bar charts detailing the average composition of speciated particulate
matter in 2016 on the 20% most impaired days (right) and 20% clearest days (right) at Hawaii
Volcanoes National Park. The plots display concentration from bottom to top for the following:
PM2.5 sulfate (yellow), PM2.5 nitrate (red), PM2.5 organic carbon (green), PM2.5 elemental carbon
(black), PM coarse (brown), and sea salt (blue).
B-7
-------
Appendix C
Model Performance Evaluation-Virgin Islands
C.l Spain! r|.-.ts of Average Model Predictions on the I¥lc ) ii"i. ised Days
The plots in this section show average daily average measurements and model predictions for
the major components of total PM2.5 mass aggregated over the most impaired days in 2016 at
each of the Class I areas (left panels). The difference between the 2028 and 2016 simulations
are also shown (middle panels) and the difference between the 2028 simulation and 2028
simulation with zero U.S. anthropogenic emissions (right panels).
C-l
-------
PM2.5 organic carbon

r 1 0

k 08
k 0j6

k Q-4
k 0 2


PM2.5 organic carbon
PM2.5 sulfate Ion
-	1.0
-	0.8
0.6
LJ-n Q
[lQfrcr
PM2.5 nitrate ion
PM2.5 organic carbon
2Q2& no US. aifi - 2028
0 010
2323 - 2:r.rj
- ,j ,M:,
- a joaa
-0.035
PM2.5 sulfate ion
2028 - 2016

PM2^ sulfate ion
Mg'm
PM2.S nitrate ion
ng/m
2028 - 2016

r 0.1Q

- 0.Q5

- 0 .00

¦ -0 05

L *0-10
u s an* - 202a
u.gC/m
r01
SE

Jigton
PM2.S nitrate ion
ug.'m
2021 no US am - 2028
- 0 00

Figure C-l Spatial plots showing the average model predictions for PM2.5 organic carbon
(top three panels), PM2.5 sulfate ion (middle three panels), and PM2.5 nitrate ion (bottom three
panels) on the 20% worst days at Virgin Islands National Park. The left panels depict
concentrations for year 2016, the middle panels depict the differences in concentrations for year
2028 - 2016, and the right panels depict the differences in concentrations for year 2028 with no
anthropogenic influences minus year 2028.
C-2
-------
C.2 Time Series for 2016
The plots in this section show daily average measurements and model predictions for the major
components of total PM2.5 mass at each of the Class I areas.
Virgin Islands NP
VIIS1
PM2.5 sulfate ion

1	r
1	r~
~t	I-
01	02	03	04	05	06	07	08	09	10	11	12
Virgin Islands NP
VIIS1
PM2.5 nitrate ion
° ° °o°


i	r
n	r
M>
1	r~
09	10	11	12
01	02	03	04	05	06	07
Virgin Islands NP
VIIS1
PfSXO,;
O
o o
PM2.5 organic carbon

i	r
01	02	03	04	05
i 	r~
ast>
~I	l~~
10	11	12
Virgin Islands NP
VIIS1
PM2.5 elemental carbon

Virgin Islands NP
VIIS1
r%-Jit
PM2.5 sodium + chloride

o
"I	T"
~l	T"
~l	T"
02	03	04	05	06	07	08	09	10	11
Virgin Islands NP
VIIS1
+
o
o° o ¦
o
°o°/
%
oS>
o O OocP
^ o_n&°

~i	r~
02	03	04	05
1	r~
~i	r~
10	11	12
Figure C-2 Time series plots for 2016 comparing model predictions (blue circles) with
IMPROVE monitor measurements (black crosses) for PM2.5 sulfate ion (top panel), PM2.5 nitrate
ion (second panel), PM2.5 organic carbon (third panel), PM2.5 elemental carbon (fourth panel),
NaCI (fifth panel), and PM coarse (bottom panel) at Virgin Islands National Park.
C-3
-------
C.3 Particulate Matter Composition on Clearest and Most Impaired Days in 2016
The plots in this section show average daily average measurements and model predictions for
the major components of total PM2.5 mass aggregated over the most and least impaired days
in 2016 at each of the Class I areas.
VIIS1

35^
CO
E
30 "i

-
E

2
25 -E
D)
-
0
-
0
20
^E

a
15-
0
-
2
c~
10-E
a
-
0

£Z
O
5^
O
—

0-
Observed	Modeled
20% Most Impaired Days
35 t
30 ~
25 ~E
20
15-
10-E
5^
0-
NaCI
PM Coarse
PM2.5 EC
PM2.5 OC
PM2.5 N03
T
T
Observed	Modeled
20% Clearest Days
Figure C-3 Stacked bar charts detailing the average composition of speciated particulate
matter in 2016 on the 20% most impaired days (right) and 20% clearest days (right) at Virgin
Islands National Park. The plots display concentration from bottom to top for the following:
PM2.5 sulfate (yellow), PM2.5 nitrate (red), PM2.5 organic carbon (green), PM2.5 elemental carbon
(black), PM coarse (brown), and sea salt (blue).
C-4
-------
Appendix D Emissions Summary
D.l Emissions i try table for Alaska
This table shows annual total (tpy) emissions for the 9 km Alaska modeling domain.
Grid
State
Sector
Species
2016 annual
emissions (tpy)
2028 annual
emissions (tpy)
9AK1
Non-US SECA C3
cmv_clc2_9akl
CO
1.23
1.24
9AK1
Non-US SECA C3
cmv_clc2_9akl
NH3
0
0
9AK1
Non-US SECA C3
cmv_clc2_9akl
NOX
8.38
4.64
9AK1
Non-US SECA C3
cmv_clc2_9akl
PM2_5
0.22
0.13
9AK1
Non-US SECA C3
cmv_clc2_9akl
SO 2
0.05
0.02
9AK1
Non-US SECA C3
cmv_clc2_9akl
VOC_INV
0.31
0.16
9AK1
Non-US SECA C3
cmv_c3_9akl
CO
376.81
538.62
9AK1
Non-US SECA C3
cmv_c3_9akl
NH3
6.2
4.56
9AK1
Non-US SECA C3
cmv_c3_9akl
NOX
4062.34
5824.12
9AK1
Non-US SECA C3
cmv_c3_9akl
PM2_5
357.04
262.32
9AK1
Non-US SECA C3
cmv_c3_9akl
SO 2
2793.74
742
9AK1
Non-US SECA C3
cmv_c3_9akl
VOC_INV
183.68
261.26
9AK1
Offshore to EEZ
cmv_clc2_9akl
CO
357.19
358.27
9AK1
Offshore to EEZ
cmv_clc2_9akl
NH3
1.14
0.65
9AK1
Offshore to EEZ
cmv_clc2_9akl
NOX
2350.5
1302.18
9AK1
Offshore to EEZ
cmv_clc2_9akl
PM2_5
59.22
33.52
9AK1
Offshore to EEZ
cmv_clc2_9akl
SO 2
3.31
1.13
9AK1
Offshore to EEZ
cmv_clc2_9akl
VOC_INV
79.77
42.12
9AK1
Offshore to EEZ
cmv_c3_9akl
CO
3666.08
5210.37
9AK1
Offshore to EEZ
cmv_c3_9akl
NH3
57.54
43.62
9AK1
Offshore to EEZ
cmv_c3_9akl
NOX
39986.75
53896.96
9AK1
Offshore to EEZ
cmv_c3_9akl
PM2_5
2989.71
2266.16
9AK1
Offshore to EEZ
cmv_c3_9akl
SO 2
23219.04
6645.79
9AK1
Offshore to EEZ
cmv c3 9akl
VOC INV
1714.89
2437.43
D-5
-------
Grid
State
Sector
Species
2016 annual
emissions (tpy)
2028 annual
emissions (tpy)
9AK1
Alaska
afdust_adj
PM2 5
1053.73
1062.96
9AK1
Alaska
ag
NH3
108.93
119.49
9AK1
Alaska
ag
VOCJNV
8.71
9.56
9AK1
Alaska
airports
CO
13478.15
14914.81
9AK1
Alaska
airports
NOX
4417.41
4370.87
9AK1
Alaska
airports
PM2 5
270.59
257.09
9AK1
Alaska
airports
S02
575.9
597.75
9AK1
Alaska
airports
VOC INV
2007.75
1945.41
9AK1
Alaska
cmv clc2 9akl
CO
598.36
600.15
9AK1
Alaska
cmv clc2 9akl
NH3
1.94
1.1
9AK1
Alaska
cmv_clc2_9akl
NOX
3966.48
2197.43
9AK1
Alaska
cmv clc2 9akl
PM2 5
100.97
57.15
9AK1
Alaska
cmv clc2 9akl
S02
7.69
2.62
9AK1
Alaska
cmv clc2 9akl
VOC INV
136.17
71.9
9AK1
Alaska
cmv_c3_9akl
CO
644.27
907.96
9AK1
Alaska
cmv c3 9akl
NH3
2.57
3.15
9AK1
Alaska
cmv_c3_9akl
NOX
6250.89
6092.58
9AK1
Alaska
cmv c3 9akl
PM2 5
133.28
163.41
9AK1
Alaska
cmv_c3_9akl
S02
517.09
434.32
9AK1
Alaska
cmv c3 9akl
VOC INV
282.96
398.38
9AK1
Alaska
nonpt
CO
28955.95
29241.97
9AK1
Alaska
nonpt
NH3
563.76
650
9AK1
Alaska
nonpt
NOX
6306.51
6725.3
9AK1
Alaska
nonpt
PM2 5
2500.06
2518.3
9AK1
Alaska
nonpt
S02
1510.39
1523.71
9AK1
Alaska
nonpt
VOC INV
8223.59
8043.48
9AK1
Alaska
nonroad
CO
34125.88
30034.58
9AK1
Alaska
nonroad
NH3
6.2
6.54
9AK1
Alaska
nonroad
NOX
2579.85
1722.26
9AK1
Alaska
nonroad
PM2 5
358.22
200.99
9AK1
Alaska
nonroad
S02
7.42
4.08
9AK1
Alaska
nonroad
VOCJNV
8599.75
5297.39
9AK1
Alaska
np_oilgas
CO
2943.75
2917.36
9AK1
Alaska
np_oilgas
NH3
0
0
9AK1
Alaska
np_oilgas
NOX
2095.26
2019.19
9AK1
Alaska
np_oilgas
PM2 5
35.87
32.9
9AK1
Alaska
np_oilgas
S02
44.17
39.81
9AK1
Alaska
np_oilgas
VOC INV
25280.67
24912.9
9AK1
Alaska
onroad_nonconus
CO
60100.75
30960.75
D-6
-------
.44 1
.96
5.9
.33
.01
.09
05
.42
,66
,44
,96
.71
77 |
,55 |
76
07
.19
D.9
,05 |
,16 1
_|
B.3 1
.81
,87
,95
,58
.83
.49
3.5
.82
08 '
15
77
,06
17
,02
64
,47
,04
.74
.84
.24
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
onroad_nonconus
onroad_nonconus
onroad_nonconus
onroad_nonconus
onroad nonconus
pt_°ilgas
NH3
NOX
PM2_5
S02
VOCJNV
CO
pt_°ilgas
pt_°ilgas
pt_°ilgas
pt_°ilgas
NH3
NOX
PM2 5
S02
152.64
11977.35
488.55
32.55
8228.21
10184.09
0.05
40683.42
503.66
1657.44
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
pt_°ilgas
ptegu
ptegu
ptegu
ptegu
ptegu
ptegu
ptfire
ptfire
ptfire
VOCJNV
CO
NH3
NOX
PM2_5
S02
VOCJNV
CO
NH3
NOX
1692.96
2444.71
1.77
7792.55
239.76
1304.07
307.19
3165510.9
51691.05
29644.16
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
ptfire
ptfire
ptfire
ptnonipm
ptnonipm
ptnonipm
ptnonipm
ptnonipm
ptnonipm
rail
rail
rail
rail
PM2_5
S02
VOC INV
CO
NH3
NOX
PM2_5
S02
VOCJNV
CO
NH3
NOX
PM2 5
262648.3
19645.81
743059.87
2561.58
48.16
7291.09
478.06
1394.23
800.1
47.54
0.15
386.37
10.94
0.17
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
rail
rail
rwc
rwc
S02
VOC INV
CO
NH3
rwc
rwc
rwc
rwc
NOX
PM2_5
S02
VOC INV
16.92
5072.5
33.89
90.09
711.98
15.54
820.28
D-7
-------
D.2 Emissions summary table for Hawaii
This table shows annual total (tpy) emissions for the 3 km Hawaii modeling domain.
Grid
State
Sector
Species
2016 annual
emissions (tpy)
2028 annual
emissions (tpy)
3HI1
Non-US SECAC3
cmv_clc2_3hil
CO
0.1
0.1
3HI1
Non-US SECAC3
cmv_clc2_3hil
NH3
0
0
3HI1
Non-US SECAC3
cmv_clc2_3hil
NOX
0.64
0.36
3HI1
Non-US SECAC3
cmv_clc2_3hil
PM2_5
0.02
0.01
3HI1
Non-US SECAC3
cmv_clc2_3hil
S02
0
0
3HI1
Non-US SECAC3
cmv_clc2_3hil
VOCJNV
0.02
0.01
3HI1
Non-US SECAC3
cmv_c3_3hil
CO
3.2
4.56
3HI1
Non-US SECAC3
cmv_c3_3hil
NH3
0.06
0.09
3HI1
Non-US SECAC3
cmv_c3_3hil
NOX
34.59
49.32
3HI1
Non-US SECAC3
cmv_c3_3hil
PM2_5
3.11
4.43
3HI1
Non-US SECAC3
cmv_c3_3hil
S02
23.71
4.83
3HI1
Non-US SECAC3
cmv_c3_3hil
VOCJNV
1.68
2.4
3HI1
Offshore to EEZ
cmv_clc2_3hil
CO
180.8
181.34
3HI1
Offshore to EEZ
cmv_clc2_3hil
NH3
0.58
0.33
3HI1
Offshore to EEZ
cmv_clc2_3hil
NOX
1204.24
667.15
3HI1
Offshore to EEZ
cmv_clc2_3hil
PM2_5
30.33
17.17
3HI1
Offshore to EEZ
cmv_clc2_3hil
S02
0.87
0.3
3HI1
Offshore to EEZ
cmv_clc2_3hil
VOCJNV
42.55
22.46
3HI1
Offshore to EEZ
cmv_c3_3hil
CO
383.59
559.51
3HI1
Offshore to EEZ
cmv_c3_3hil
NH3
1.12
1.63
3HI1
Offshore to EEZ
cmv_c3_3hil
NOX
3812.24
2517.1
3HI1
Offshore to EEZ
cmv_c3_3hil
PM2_5
58.02
84.63
3HI1
Offshore to EEZ
cmv_c3_3hil
S02
147.32
214.89
3HI1
Offshore to EEZ
cmv_c3_3hil
VOCJNV
174.82
254.99
D-8
-------
Grid
State
Sector
Species
2016 annual
emissions (tpy)
2028 annual
emissions (tpy)
3HI1
Hawaii
afdust_adj
PM2_5
3764.3
3808.93
3HI1
Hawaii
ag
NH3
1495.69
1535.9
3HI1
Hawaii
ag
VOCJNV
119.66
122.87
3HI1
Hawaii
airports
CO
10079.7
12764.45
3HI1
Hawaii
airports
NOX
3500.91
4223.8
3HI1
Hawaii
airports
PM2_5
150.57
165.29
3HI1
Hawaii
airports
SO 2
524.55
662.42
3HI1
Hawaii
airports
VOC_INV
1091.71
1223.5
3HI1
Hawaii
cmv_clc2_3hil
CO
244.54
245.28
3HI1
Hawaii
cmv_clc2_3hil
NH3
0.79
0.45
3HI1
Hawaii
cmv_clc2_3hil
NOX
1621.4
898.26
3HI1
Hawaii
cmv_clc2_3hil
PM2_5
41.02
23.22
3HI1
Hawaii
cmv_clc2_3hil
SO 2
2.19
0.75
3HI1
Hawaii
cmv_clc2_3hil
VOC_INV
55.26
29.18
3HI1
Hawaii
cmv_c3_3hil
CO
306.63
447.25
3HI1
Hawaii
cmv_c3_3hil
NH3
0.87
1.27
3HI1
Hawaii
cmv_c3_3hil
NOX
2309.82
1525.09
3HI1
Hawaii
cmv_c3_3hil
PM2_5
45.23
65.98
3HI1
Hawaii
cmv_c3_3hil
SO 2
94.4
137.7
3HI1
Hawaii
cmv_c3_3hil
VOCJNV
198.63
289.72
3HI1
Hawaii
nonpt
CO
10019.69
10019.69
3HI1
Hawaii
nonpt
NH3
53.29
53.29
3HI1
Hawaii
nonpt
NOX
396.78
396.78
3HI1
Hawaii
nonpt
PM2_5
1498.37
1498.37
3HI1
Hawaii
nonpt
SO 2
89.33
89.33
3HI1
Hawaii
nonpt
VOC_INV
14675.31
14051.12
3HI1
Hawaii
nonroad
CO
47219.26
49901.07
3HI1
Hawaii
nonroad
NH3
6.81
8.1
3HI1
Hawaii
nonroad
NOX
3440.76
2085.28
3HI1
Hawaii
nonroad
PM2_5
320.43
198.08
3HI1
Hawaii
nonroad
SO 2
8.03
5.59
3HI1
Hawaii
nonroad
VOC_INV
4423.67
3020.56
3HI1
Hawaii
onroad_nonconus
CO
82599.39
43003.13
3HI1
Hawaii
onroad_nonconus
NH3
315.68
271.01
3HI1
Hawaii
onroad_nonconus
NOX
10384.35
3220.21
3HI1
Hawaii
onroad_nonconus
PM2_5
310.4
167.17
3HI1
Hawaii
onroad_nonconus
SO 2
63.17
34.15
3HI1
Hawaii
onroad_nonconus
VOC_INV
8953.71
3959.28
3HI1
Hawaii
ptagfire
CO
1079.98
1079.98
D-9
-------
3HI1
Hawaii
ptagfire
NH3
390.38
390.38
3HI1
Hawaii
ptagfire
NOX
55.3
55.3
3HI1
Hawaii
ptagfire
PM2_5
81.58
81.58
3HI1
Hawaii
ptagfire
SO 2
30.09
30.09
3HI1
Hawaii
ptagfire
VOCJNV
82.68
82.68
3HI1
Hawaii
ptegu
CO
1599.16
1375.28
3HI1
Hawaii
ptegu
NH3
170.9
146.97
3HI1
Hawaii
ptegu
NOX
17520.17
15067.35
3HI1
Hawaii
ptegu
PM2_5
1374.39
1181.97
3HI1
Hawaii
ptegu
SO 2
18003.18
15482.73
3HI1
Hawaii
ptegu
VOCJNV
162.82
140.03
3HI1
Hawaii
ptfire_nonconus
CO
57641.9
57641.9
3HI1
Hawaii
ptfire_nonconus
NH3
836.78
836.78
3HI1
Hawaii
ptfire_nonconus
NOX
3373.68
3373.68
3HI1
Hawaii
ptfire_nonconus
PM2_5
5752.35
5752.35
3HI1
Hawaii
ptfire_nonconus
SO 2
257.9
257.9
3HI1
Hawaii
ptfire_nonconus
VOCJNV
19201.73
19201.73
3HI1
Hawaii
ptnonipm
CO
3993.81
4794.87
3HI1
Hawaii
ptnonipm
NH3
67.33
70.45
3HI1
Hawaii
ptnonipm
NOX
2716.33
2810.61
3HI1
Hawaii
ptnonipm
PM2_5
495.15
541.74
3HI1
Hawaii
ptnonipm
SO 2
913.17
883.25
3HI1
Hawaii
ptnonipm
VOCJNV
2917.09
2860.38
3HI1
Hawaii
rwc
CO
3700.83
3704.72
3HI1
Hawaii
rwc
NH3
26.97
27.14
3HI1
Hawaii
rwc
NOX
66.94
71.64
3HI1
Hawaii
rwc
PM2_5
512.77
504.04
3HI1
Hawaii
rwc
SO 2
8.82
9.31
3HI1
Hawaii
rwc
VOCJNV
599.47
572.4
D-10
-------
D.3 Emissions summary table for Puerto Rico/Virgin Islands
This table shows annual total (tpy) emissions for the 3 and 9 km Puerto Rico/Virgin Islands
modeling domains.




2016
2028


2016
2028




annual
annual


annual
annual




emission
emission


emissions
emissions
Grid
State
Sector
Species
s (tpy)
s (tpy)
Grid
(tpy)
(tpy)
3PR1
Non-US SECA C3
cmv_clc2_3prl
CO
150.86
151.32
9PR1
286.84
287.7
3PR1
Non-US SECA C3
cmv_clc2_3prl
NH3
0.5
0.28
9PR1
0.93
0.53
3PR1
Non-US SECA C3
cmv_clc2_3prl
NOX
998.65
553.25
9PR1
1875.83
1039.21
3PR1
Non-US SECA C3
cmv_clc2_3prl
PM2_5
25.94
14.68
9PR1
48.52
27.46
3PR1
Non-US SECA C3
cmv_clc2_3prl
S02
3.58
1.22
9PR1
7.09
2.42
3PR1
Non-US SECA C3
cmv_clc2_3prl
VOCJNV
34.95
18.46
9PR1
64.94
34.29
3PR1
Non-US SECA C3
cmv_c3_3prl
CO
624.45
890.35 !
9PR1
7721.46
11009.4
3PR1
Non-US SECA C3
cmv_c3_3prl
NH3
12.8
18.25
9PR1
119.75
170.74
3PR1
Non-US SECA C3
cmv_c3_3prl
NOX
7283.57
10385.14
9PR1
88718.6
126491.11
3PR1
Non-US SECA C3
cmv_c3_3prl
PM2_5
664.98
948.15

9PR1
7681.77
10952.96
3PR1
Non-US SECA C3
cmv_c3_3prl
S02
5436.66
1107.39

9PR1
63277.74
12888.49
3PR1
Non-US SECA C3
cmv_c3_3prl
VOC INV
285.24
406.71
9PR1
3572.1
5093.31
3PR1
Offshore to EEZ
cmv_clc2_3prl
CO
115.97
116.32
9PR1
136.28
136.69
3PR1
Offshore to EEZ
cmv_clc2_3prl
NH3
0.38
0.22
9PR1
0.45
0.25
3PR1
Offshore to EEZ
cmv_clc2_3prl
NOX
751.84
416.52
9PR1
880.73
487.93
3PR1
Offshore to EEZ
cmv_clc2_3prl
PM2_5
19.82
11.22
9PR1
23.16
13.11
3PR1
Offshore to EEZ
cmv_clc2_3prl
S02
3.14
1.07
9PR1
3.64
1.24
3PR1
Offshore to EEZ
cmv_clc2_3prl
VOCJNV
26.69
14.09
9PR1
30.89
16.31
3PR1
Offshore to EEZ
cmv_c3_3prl
CO
1311.09
1788.41
9PR1
1922.89
2622.98
3PR1
Offshore to EEZ
cmv_c3_3prl
NH3
19.97
27.24

9PR1
31.23
42.6
3PR1
Offshore to EEZ
cmv_c3_3prl
NOX
14488.85
12691.72

9PR1
21587.73
18910.14
3PR1
Offshore to EEZ
cmv_c3_3prl
PM2 5
1037.59
1415.36
9PR1
1622.51
2213.23
3PR1
Offshore to EEZ
cmv_c3_3prl
S02
8295.21
11315.27
9PR1
13087.73
17852.67
3PR1
Offshore to EEZ
cmv_c3_3prl
VOCJNV
615.91
840.14
9PR1
893.47
1218.76
D-ll
-------




2016
2028


2016
2028
Grid
State
Sector
Species
annual
emissions
(tpy)
annual
emissions
(tpy)

Grid
annual
emissions
(tpy)
annual
emissions
(tpy)
3PR1
Puerto Rico
afdust_adj
PM2_5
329.02
353.11

9PR1
329.02
353.11
3PR1
Puerto Rico
cmv_clc2_3prl
CO
181.1
181.64
9PR1
184.47
185.02
3PR1
Puerto Rico
cmv_clc2_3prl
NH3
0.59
0.33
9PR1
0.6
0.34
3PR1
Puerto Rico
cmv_clc2_3prl
NOX
1201.67
665.73
9PR1
1224.69
678.48
3PR1
Puerto Rico
cmv_clc2_3prl
PM2_5
30.69
17.37
9PR1
31.32
17.73
3PR1
Puerto Rico
cmv_clc2_3prl
S02
2.23
0 76
9PR1
2.41
0.82
3PR1
Puerto Rico
cmv_clc2_3prl
VOCJNV
41.8
22.07

9PR1
42.71
22.55
3PR1
Puerto Rico
cmv_c3_3prl
CO
241.8
329.83

9PR1
258.31
352.36
3PR1
Puerto Rico
cmv_c3_3prl
NH3
0.78
1.07
9PR1
0.84
1.15
3PR1
Puerto Rico
cmv_c3_3prl
NOX
2095.94
1835.97
9PR1
2218.17
1943.04
3PR1
Puerto Rico
cmv_c3_3prl
PM2_5
40.59
55.37
9PR1
43.76
59.69
3PR1
Puerto Rico
cmv_c3_3prl
S02
129.48
176.63

9PR1
139.05
189.67
3PR1
Puerto Rico
cmv_c3_3prl
VOCJNV
133.11
181.57

9PR1
142.46
194.33
3PR1
Puerto Rico
nonpt
CO
18201.9
18203.75
9PR1
18201.9
18203.75
3PR1
Puerto Rico
nonpt
NH3
75.86
75.86
9PR1
75.86
75.86
3PR1
Puerto Rico
nonpt
NOX
865.51
894.84
9PR1
865.51
894.84
3PR1
Puerto Rico
nonpt
PM2_5
2694.82
2695.49
9PR1
2694.82
2695.49
3PR1
Puerto Rico
nonpt
S02
188.84
189.66
9PR1
188.84
189.66
3PR1
Puerto Rico
nonpt
VOCJNV
28265.96
28272.42
9PR1
28265.96
28272.42
3PR1
Puerto Rico
nonroad
CO
122296.1
140808.4
9PR1
122296.05
140808.37
3PR1
Puerto Rico
nonroad
NH3
14.19
17.81
9PR1
14.19
17.81
3PR1
Puerto Rico
nonroad
NOX
6367.25
4384.58

9PR1
6367.25
4384.58
3PR1
Puerto Rico
nonroad
PM2_5
761.51
576.92

9PR1
761.51
576.92
3PR1
Puerto Rico
nonroad
S02
17.42
12.29
9PR1
17.42
12.29
3PR1
Puerto Rico
nonroad
VOCJNV
10985.53
9126.3
9PR1
10985.53
9126.3
3PR1
Puerto Rico
onroad_nonconus
CO
103859
103859
9PR1
103858.97
103858.97
3PR1
Puerto Rico
onroad_nonconus
NH3
404.28
404.28
9PR1
404.28
404.28
3PR1
Puerto Rico
onroad_nonconus
NOX
9974.92
9974.92
9PR1
9974.92
9974.92
3PR1
Puerto Rico
onroad_nonconus
PM2_5
346.89
346.89
9PR1
346.89
346.89
3PR1
Puerto Rico
onroad_nonconus
S02
85.88
85.88
9PR1
85.88
85.88
3PR1
Puerto Rico
onroad_nonconus
VOCJNV
9199.22
9199.22
9PR1
9199.22
9199.22
3PR1
Puerto Rico
ptegu
CO
2842.89
2842.89

9PR1
2842.89
2842.89
3PR1
Puerto Rico
ptegu
NH3
0
0

9PR1
0
0
3PR1
Puerto Rico
ptegu
NOX
18479.36
18479.36
9PR1
18479.36
18479.36
3PR1
Puerto Rico
ptegu
PM2_5
1140.9
1140.9

9PR1
1140.9
1140.9
D-12
-------
3PR1
Puerto Rico
ptegu
S02
24553.14
24553.14

9PR1
24553.14
24553.14 '
3PR1
Puerto Rico
ptegu
VOCJNV
221.92
22L.92
9PR1
221.92
221.92
3PR1
i Puerto Rico j
ptfire_nonconus
CO
14.47
14.47
9PR1
14.47
14.47
3PR1
| Puerto Rico j
ptfire_nonconus
NH3
0.29
0.29
9PR1
0.29
0.29
3PR1
Puerto Rico
ptfire_nonconus
NOX
0.4
0.4
9PR1
0.4
0.4
3PR1
| Puerto Rico |
ptfire_nonconus
PM2_5
1.78
1.78
9PR1
1.78
1.78
3PR1
| Puerto Rico |
ptfire_nonconus
S02
0.13
0.13
9PR1
0.13
0.13
3PR1
Puerto Rico
ptfire_nonconus
VOC_INV
4.05
4.05
9PR1
4.05
4.05
3PR1
| Puerto Rico |
ptnonipm
CO
487.71
462.38
9PR1
487.71
462.38
3PR1
Puerto Rico
ptnonipm
NH3
172.42
172.42

9PR1
172.42
172.42 |
3PR1
Puerto Rico '
ptnonipm
NOX
1720.61
1658.09

9PR1
1720.61
1658.09
3PR1
Puerto Rico
ptnonipm
PM2_5
77.01
67.98
9PR1
77.01
67.98
3PR1
i Puerto Rico j
ptnonipm
S02
1362.95
1089.96

9PR1
1362.95
1089.96
3PR1
| Puerto Rico |
ptnonipm
VOCJNV
247.21
244.38

9PR1
247.21
244.38
D-13
-------






2016
2028


2016
2028
Grid
State
Sector
Species
annual
emissions
annual
emissions

Grid
annual
emissions
annual
emissions






(tpy)
(tpy)


(tpy)
(tpy)
3PR1
Virgin
Is
ands
afdust_adj
PM2_5
271.75
279.14

9PR1
271.75
279.14
3PR1
Virgin
Is
ands
cmv_clc2_3prl
CO
181.89
182.44
9PR1
175.84
176.36
3PR1
Virgin
Is
ands
cmv_clc2_3prl
NH3
0.63
0.36
9PR1
0.61
0.34
3PR1
Virgin
Is
ands
cmv_clc2_3prl
NOX
1243.97
689.16
9PR1
1203.94
666.98
3PR1
Virgin
Is
ands
cmv_clc2_3prl
PM2_5
32.63
18.47
9PR1
31.62
17.9
3PR1
Virgin
Is
ands
cmv_clc2_3prl
S02
3.46
1.18
9PR1
3.41
1.16
3PR1
Virgin
Is
ands
cmv_clc2_3prl
VOCJNV
47.91
25.3
9PR1
46.56
24.58
3PR1
Virgin
Is
ands
cmv_c3_3prl
CO
217.81
297.11
9PR1
216.52
295.35
3PR1
Virgin
Is
ands
cmv_c3_3prl
NH3
0.69
0.94
9PR1
0.69
0.95
3PR1
Virgin
Is
ands
cmv_c3_3prl
NOX
1816.28
1591

9PR1
1814.14
1589.13
3PR1
Virgin
Is
ands
cmv_c3_3prl
PM2_5
35.97
49.06

9PR1
36.08
49.21
3PR1
Virgin
Is
ands
cmv_c3_3prl
S02
85.03
115.99
9PR1
86.66
118.21
3PR1
Virgin
Is
ands
cmv_c3_3prl
VOCJNV
121.87
166.25
9PR1
120.22
163.99
3PR1
Virgin
Is
ands
nonpt
CO
478.41
483.11
9PR1
478.41
483.11
3PR1
Virgin
Is
ands
nonpt
NH3
3.31
3.39

9PR1
3.31
3.39
3PR1
Virgin
Is
ands
nonpt
NOX
59.37
62.31

9PR1
59.37
62.31
3PR1
Virgin
Is
ands
nonpt
PM2_5
163.63
166.06
9PR1
163.63
166.06
3PR1
Virgin
Is
ands
nonpt
S02
13.87
13.31
9PR1
13.87
13.31
3PR1
Virgin
Is
ands
nonpt
VOCJNV
902.59
902.79
9PR1
902.59
902.79
3PR1
Virgin
Is
ands
nonroad
CO
4315.93
4760.68
9PR1
4315.93
4760.68
3PR1
Virgin
Is
ands
nonroad
NH3
0.66
0.8
9PR1
0.66
0.8
3PR1
Virgin
Is
ands
nonroad
NOX
328.41
197.08
9PR1
328.41
197.08
3PR1
Virgin
Is
ands
nonroad
PM2_5
33.14
19.99
9PR1
33.14
19.99
3PR1
Virgin
Is
ands
nonroad
S02
0.76
0 55
9PR1
0.76
0.55
3PR1
Virgin
Is
ands
nonroad
VOCJNV
471.29
321.78

9PR1
471.29
321.78
3PR1
Virgin
Is
ands
onroad_nonconus
CO
3092.04
3092.04

9PR1
3092.04
3092.04
3PR1
Virgin
Is
ands
onroad_nonconus
NH3
14.53
14.53
9PR1
14.53
14.53
3PR1
Virgin
Is
ands
onroad_nonconus
NOX
442.65
442.65
9PR1
442.65
442.65
3PR1
Virgin
Is
ands
onroad nonconus
PM2_5
16.08
16.08
9PR1
16.08
16.08
3PR1
Virgin
Is
ands
onroad_nonconus
S02
2.97
2.97
9PR1
2.97
2.97
3PR1
Virgin
Is
ands
onroad_nonconus
VOCJNV
351.8
351.8
9PR1
351.8
351.8
D-14
-------
United States	Office of Air Quality Planning and Standards	Publication No. EPA-454/R-21-007
Environmental Protection	Air Quality Assessment Division	August 2021
Agency	Research Triangle Park, NC
-------