  TOC \o "1-3" \h \z \u    HYPERLINK \l "_Toc350119146"  I.	Table of
Contents	  PAGEREF _Toc350119146 \h  i  

  HYPERLINK \l "_Toc350119147"  II.	Table of Tables	  PAGEREF
_Toc350119147 \h  ii  

  HYPERLINK \l "_Toc350119148"  III.	Table of Figures	  PAGEREF
_Toc350119148 \h  iv  

  HYPERLINK \l "_Toc350119149"  IV.	Introduction	  PAGEREF _Toc350119149
\h  1  

  HYPERLINK \l "_Toc350119150"  V.	2002 and 2008 Emission Inventories	 
PAGEREF _Toc350119150 \h  1  

  HYPERLINK \l "_Toc350119151"  V.A.	2002 Arizona Emission Inventory
Methodology	  PAGEREF _Toc350119151 \h  3  

  HYPERLINK \l "_Toc350119152"  V.B.	2008 Arizona Emission Inventory
Methodology	  PAGEREF _Toc350119152 \h  3  

  HYPERLINK \l "_Toc350119153"  V.C.	2002 & 2008 Emission Inventory
Methodology Comparisons	  PAGEREF _Toc350119153 \h  4  

  HYPERLINK \l "_Toc350119154"  V.D.	2002 and 2008 Emissions	  PAGEREF
_Toc350119154 \h  13  

  HYPERLINK \l "_Toc350119155"  V.D.1.	Estimated 2002 and 2008 Arizona
Emissions	  PAGEREF _Toc350119155 \h  14  

  HYPERLINK \l "_Toc350119156"  V.D.2.	Summary of Major Methodological
Changes	  PAGEREF _Toc350119156 \h  23  

  HYPERLINK \l "_Toc350119157"  V.D.3.	Regional Inventory Trends for
Emissions	  PAGEREF _Toc350119157 \h  24  

  HYPERLINK \l "_Toc350119158"  VI.	IMPROVE Monitoring Data	  PAGEREF
_Toc350119158 \h  27  

  HYPERLINK \l "_Toc350119159"  VI.A.	Data Completeness Requirements	 
PAGEREF _Toc350119159 \h  27  

  HYPERLINK \l "_Toc350119160"  VI.A.1.	SIAN1 data substitution
methodology	  PAGEREF _Toc350119160 \h  28  

  HYPERLINK \l "_Toc350119161"  VI.B.	Baseline and Progress Period
Visibility	  PAGEREF _Toc350119161 \h  32  

  HYPERLINK \l "_Toc350119162"  VI.B.1.	Progress Period (2005-2009)
Visibility	  PAGEREF _Toc350119162 \h  34  

  HYPERLINK \l "_Toc350119163"  VI.B.2.	Visibility Trend Analyses	 
PAGEREF _Toc350119163 \h  37  

  HYPERLINK \l "_Toc350119164"  VI.C.	Ammonium Sulfate Analysis	 
PAGEREF _Toc350119164 \h  47  

  HYPERLINK \l "_Toc350119165"  VI.C.1.	20% Most Impaired Ammonium
Sulfate Days	  PAGEREF _Toc350119165 \h  47  

  HYPERLINK \l "_Toc350119166"  VI.C.2.	Regional Ammonium Sulfate Trends
  PAGEREF _Toc350119166 \h  49  

  HYPERLINK \l "_Toc350119167"  VI.D.	Coarse Mass Analysis	  PAGEREF
_Toc350119167 \h  51  

  HYPERLINK \l "_Toc350119168"  VI.D.1.	Worst 20% Coarse Mass Days	 
PAGEREF _Toc350119168 \h  52  

  HYPERLINK \l "_Toc350119169"  VI.D.2.	Large Point Source Locations	 
PAGEREF _Toc350119169 \h  53  

  HYPERLINK \l "_Toc350119170"  VI.E.	Method Comparison Summary	 
PAGEREF _Toc350119170 \h  56  

  HYPERLINK \l "_Toc350119171"  VII.	Reasonable Progress Goals	  PAGEREF
_Toc350119171 \h  59  

  HYPERLINK \l "_Toc350119172"  VIII.	Conclusions	  PAGEREF
_Toc350119172 \h  64  

 

Table of Tables

  TOC \h \z \c "Table"    HYPERLINK \l "_Toc350119173"  Table 1: 
Emission Inventory Methodology Comparison for 2002 and 2008.	  PAGEREF
_Toc350119173 \h  5  

  HYPERLINK \l "_Toc350119174"  Table 2:  Arizona Sulfur Dioxide
Emissions by Source Category	  PAGEREF _Toc350119174 \h  15  

  HYPERLINK \l "_Toc350119175"  Table 3:  Arizona Nitrogen Oxide
Emissions by Source Category	  PAGEREF _Toc350119175 \h  16  

  HYPERLINK \l "_Toc350119176"  Table 4:  Arizona Ammonia Emissions by
Source Category	  PAGEREF _Toc350119176 \h  17  

  HYPERLINK \l "_Toc350119177"  Table 5:  Arizona Volatile Organic
Compound Emissions by Source Category	  PAGEREF _Toc350119177 \h  18  

  HYPERLINK \l "_Toc350119178"  Table 6:  Arizona Primary Organic
Aerosol Emissions by Source Category	  PAGEREF _Toc350119178 \h  19  

  HYPERLINK \l "_Toc350119179"  Table 7:  Arizona Elemental Carbon
Emissions by Source Category	  PAGEREF _Toc350119179 \h  20  

  HYPERLINK \l "_Toc350119180"  Table 8:  Arizona Fine Particulate
Matter Emissions by Source Category	  PAGEREF _Toc350119180 \h  21  

  HYPERLINK \l "_Toc350119181"  Table 9:  Arizona Coarse Particulate
Matter Emissions by Source Category	  PAGEREF _Toc350119181 \h  22  

  HYPERLINK \l "_Toc350119182"  Table 10:  Relative Contribution of
Pollutants to Visibility Conditions on the 20% Most Impaired Days at
Arizona Class I area IMPROVE Sites for the Progress Period (2005-2009).	
 PAGEREF _Toc350119182 \h  35  

  HYPERLINK \l "_Toc350119183"  Table 11:  Relative Contribution of
Pollutants to Visibility Conditions on the 20% Least Impaired Days at
Arizona Class I area IMPROVE Sites for the Progress Period (2005-2009).	
 PAGEREF _Toc350119183 \h  36  

  HYPERLINK \l "_Toc350119184"  Table 12:  Difference in Aerosol
Extinction by Component between the Baseline Period (2000-2004) and the
Progress Period (2005-2009) on the 20% Most Impaired Days for Arizona
Class I IMPROVE Sites.	  PAGEREF _Toc350119184 \h  41  

  HYPERLINK \l "_Toc350119185"  Table 13:  Difference in Aerosol
Extinction by Component between the Baseline Period (2000-2004) and the
Progress Period (2005-2009) on the 20% Least Impaired Days for Arizona
Class I IMPROVE Sites.	  PAGEREF _Toc350119185 \h  42  

  HYPERLINK \l "_Toc350119186"  Table 14:  Statistically Significant
2000-2009 Annual Average Trends for Aerosol Extinction by Component for
Arizona Class I area IMPROVE Sites.	  PAGEREF _Toc350119186 \h  45  

  HYPERLINK \l "_Toc350119187"  Table 15:  2000-2009 Ammonium Sulfate
Visibility Extinction (mM-1) Trends, Baseline (2000-2004) vs. Progress
(2005-2009) Period Comparisons, and Altered Baseline (2000-2005) vs.
Altered Progress (2005-2009) Period Comparisons at each Arizona Class I
area IMPROVE Site for the 20% Worst Ammonium Sulfate Days.	  PAGEREF
_Toc350119187 \h  48  

  HYPERLINK \l "_Toc350119188"  Table 16:  2000-2010 Coarse Matter
Visibility Extinction (mM-1) Trends and Baseline vs. Progress Period
Comparisons at each Arizona Class I area IMPROVE Site for the 20% Worst
Coarse Matter Days.	  PAGEREF _Toc350119188 \h  53  

  HYPERLINK \l "_Toc350119189"  Table 17:  CHIR1 IMPROVE Site comparison
of Baseline and Progress Period Ammonium Sulfate Averages for All Days
and the 20% Worst Days	  PAGEREF _Toc350119189 \h  59  

  HYPERLINK \l "_Toc350119190"  Table 18:  Arizona Class I Area
Reasonable Progress Goals Comparison to Progress Period Visibility for
the 20% Worst Days.  '2018 Projected Visibility' was extrapolated based
on the rate of Visibility change between the Baseline and Progress
Period Visibilities.	  PAGEREF _Toc350119190 \h  62  

  HYPERLINK \l "_Toc350119191"  Table 19:  Alternative method for the
20% Most Impaired Days at GRCA2.  EC and POM visibility extinctions are
replaced by ten-year average for 2003 and 2009.	  PAGEREF _Toc350119191
\h  62  

  HYPERLINK \l "_Toc350119192"  Table 20:  Arizona Class I Area
Reasonable Progress Goals Adjusted Comparison to the Altered Progress
Period Visibility (2006-2010) for the 20% Worst Days.  '2018 Projected
Visibility' was extrapolated based on the rate of visibility change
between the Baseline and Progress Period Visibilities.  In this case the
Baseline period was altered to the years 2000-2005 and the Progress
Period was adjusted to the years 2006-2010.	  PAGEREF _Toc350119192 \h 
63  

  HYPERLINK \l "_Toc350119193"  Table 21:  Arizona Class I Area
Reasonable Progress Goals Comparison to Progress Period Visibility for
the 20% Best Days.  '2018 Projected Visibility' was extrapolated based
on the rate of Visibility change between the Baseline and Progress
Period Visibilities.	  PAGEREF _Toc350119193 \h  64  

 

Table of Figures

  TOC \h \z \c "Figure"    HYPERLINK \l "_Toc350119194"  Figure 1:  2002
and 2008 Sulfur Dioxide Emissions by Source Category	  PAGEREF
_Toc350119194 \h  15  

  HYPERLINK \l "_Toc350119195"  Figure 2:  2002 and 2008 Nitrogen Oxide
Emissions by Source Category	  PAGEREF _Toc350119195 \h  16  

  HYPERLINK \l "_Toc350119196"  Figure 3:  2002 and 2008 Ammonia
Emissions by Source Category	  PAGEREF _Toc350119196 \h  17  

  HYPERLINK \l "_Toc350119197"  Figure 4:  2002 and 2008 Volatile
Organic Compound Emissions by Source Category	  PAGEREF _Toc350119197 \h
 18  

  HYPERLINK \l "_Toc350119198"  Figure 5:  2002 and 2008 Primary Organic
Aerosol Emissions by Source Category	  PAGEREF _Toc350119198 \h  19  

  HYPERLINK \l "_Toc350119199"  Figure 6:  2002 and 2008 Elemental
Carbon Emissions by Source Category	  PAGEREF _Toc350119199 \h  20  

  HYPERLINK \l "_Toc350119200"  Figure 7:  2002 and 2008 Fine
Particulate Matter Emissions by Source Category	  PAGEREF _Toc350119200
\h  21  

  HYPERLINK \l "_Toc350119201"  Figure 8:  2002 and 2008 Coarse
Particulate Matter Emissions by Source Category	  PAGEREF _Toc350119201
\h  22  

  HYPERLINK \l "_Toc350119202"  Figure 9:  Regional differences in VOC
emissions between 2002 and 2008 Emission Inventories	  PAGEREF
_Toc350119202 \h  25  

  HYPERLINK \l "_Toc350119203"  Figure 10:  Regional differences in
Coarse Mass emissions between 2002 and 2008 Emission Inventories	 
PAGEREF _Toc350119203 \h  25  

  HYPERLINK \l "_Toc350119204"  Figure 11:  Regional differences in Fine
Mass emissions between 2002 and 2008 Emission Inventories	  PAGEREF
_Toc350119204 \h  26  

  HYPERLINK \l "_Toc350119205"  Figure 12:  IMPROVE SIAN1 data collected
during the 2005-2009 progress period, where original SIAN1 RHR data are
depicted in dark blue, and substituted data are depicted with separate
colors by species.	  PAGEREF _Toc350119205 \h  30  

  HYPERLINK \l "_Toc350119206"  Figure 13:  IMPROVE SIAN1 data collected
during the 2005-2009 progress period, where substituted days are
depicted with a black bar beneath the data.	  PAGEREF _Toc350119206 \h 
31  

  HYPERLINK \l "_Toc350119207"  Figure 14:  Average Extinction for
Current Progress Period (2005-2009) for the Worst (Most Impaired) and
Best (Least Impaired) Days Measured at Arizona Class I area IMPROVE
Sites.	  PAGEREF _Toc350119207 \h  37  

  HYPERLINK \l "_Toc350119208"  Figure 15:  Average Extinction for
Baseline and Progress Period Extinction for Worst (Most Impaired) Days
Measured at Arizona Class I area IMPROVE Sites.	  PAGEREF _Toc350119208
\h  43  

  HYPERLINK \l "_Toc350119209"  Figure 16:  Difference between Average
Extinction for Current Progress Period (2005-2009) and Baseline Period
(2000-2004) for the Worst (Most Impaired) Days Measured at Arizona Class
I area IMPROVE Sites.	  PAGEREF _Toc350119209 \h  43  

  HYPERLINK \l "_Toc350119210"  Figure 17:  Average Extinction for
Baseline and Progress Period Extinction for Best (Least Impaired) Days
Measured at Arizona Class I area IMPROVE Sites.	  PAGEREF _Toc350119210
\h  44  

  HYPERLINK \l "_Toc350119211"  Figure 18:  Difference between Average
Extinction for Current Progress Period (2005-2009) and Baseline Period
(2000-2004) for the Best (Least Impaired) Days Measured at Arizona Class
I area IMPROVE Sites.	  PAGEREF _Toc350119211 \h  44  

  HYPERLINK \l "_Toc350119212"  Figure 19:  Average Annual Ammonium
Sulfate Extinction (mM-1) at each Arizona Class I area IMPROVE Site for
the 20% Worst Ammonium Sulfate Days.	  PAGEREF _Toc350119212 \h  48  

  HYPERLINK \l "_Toc350119213"  Figure 20:  Magnitude of visibility
component extinctions that have increased between the baseline average
(2000-2004) and the first progress period average (2005-2009) for the
20% worst visibility days.	  PAGEREF _Toc350119213 \h  50  

  HYPERLINK \l "_Toc350119214"  Figure 21:  10-year annual average
ammonium sulfate extinction trends for 20% worst days at CIA IMPROVE
sites in the WRAP region.	  PAGEREF _Toc350119214 \h  50  

  HYPERLINK \l "_Toc350119215"  Figure 22:  10-year annual average
ammonium sulfate extinction trends for all measured days at CIA IMPROVE
sites in the WRAP region.	  PAGEREF _Toc350119215 \h  51  

  HYPERLINK \l "_Toc350119216"  Figure 23:  Average Annual Coarse Mass
Extinction (mM-1) at each Arizona Class I area IMPROVE Site for the 20%
Worst Coarse Mass Days.	  PAGEREF _Toc350119216 \h  53  

  HYPERLINK \l "_Toc350119217"  Figure 24:  Locations of Arizona Class I
areas, Class I area 50 km buffers, Class I area IMPROVE monitors, and
Large Point Source Emitters of PM10 (>100 tons/year).  IMPROVE site
values correspond to Visibility Extinction (mM-1) of Coarse Mass
averaged over the progress period (2005-2009).	  PAGEREF _Toc350119217
\h  55  

  HYPERLINK \l "_Toc350119218"  Figure 25:  Daily Aerosol Extinction
measured by the Chiricahua CHIR1 IMPROVE monitor during 2002.	  PAGEREF
_Toc350119218 \h  58  

  HYPERLINK \l "_Toc350119219"  Figure 26:  Daily Aerosol Extinction
measured by the Chiricahua CHIR1 IMPROVE monitor during 2008.	  PAGEREF
_Toc350119219 \h  58  

  HYPERLINK \l "_Toc350119220"  Figure 27:  Monthly Average Aerosol
Extinction measured by the CHIR1 IMPROVE monitor in 2002.	  PAGEREF
_Toc350119220 \h  58  

  HYPERLINK \l "_Toc350119221"  Figure 28:  Monthly Average Aerosol
Extinction measured by the CHIR1 IMPROVE monitor in 2008.	  PAGEREF
_Toc350119221 \h  58  

 Introduction

In 2011, the State of Arizona submitted a Regional Haze State
Implementation Plan (SIP) to the Environmental Protection Agency (EPA)
to fulfill the requirements of 40 CFR § 51.308.  On December 21, 2012
the EPA published within the Federal Register a partial disapproval of
Arizona’s Regional Haze State Implementation Plan.  This document was
created to address certain deficiencies identified by the EPA,
specifically 40 CFR § 51.308(d)(4)(v) which requires to submission of
the most recently available Emissions Inventory in a Regional Haze SIP. 
Furthermore, this document presents an alternative methodology, which
has been used by the EPA in previous studies, for the analysis of
visibility trends and provides supplemental information to address
visibility trends within the State of Arizona between the years of 2000
and 2009 and how these trends compare to Reasonable Progress Goals and
Uniform Rates of Progress.

2002 and 2008 Emission Inventories

The 2002 Arizona Emission Inventory (EI) was originally constructed to
fulfill the requirement of 40 CFR § 51.308(d)(4)(v), which states a
baseline year emission inventory must be contained within a State’s
implementation plan for regional haze.  On December 21, 2012 EPA
published within the Federal Register a partial disapproval of
Arizona’s Regional Haze State Implementation Plan.  One disapproved
provision was based on 40 CFR § 51.308(d)(4)(v) which further requires
a submission of a statewide emission inventory for the most recent year
of which data are available.  In order to fulfill this requirement, the
Arizona Department of Environmental Quality (ADEQ) is presenting a 2008
emission inventory calculated and compiled by the Western Regional Air
Partnership (WRAP) and Air Resource Specialists, Inc (ARS).

The 2002 and 2008 emission inventories calculate emissions of the
regional haze contributing pollutants:  Sulfur Dioxide (SO2), Nitric
Oxide (NOx), Ammonia (NH3), Volatile Organic Compounds (VOCs), Primary
Organic Aerosols (POA), Elemental Carbon (EC), Fine Particulate Matter
(Fine), and Coarse Particulate Matter (CM).  For the purpose of
consistency, these pollutants were reported for all major source
categories.  Where possible these source categories were further
partitioned into anthropogenic and natural sources. Source categories
for both anthropogenic and natural sources are listed and described
briefly here, followed by information related to the inventories.

Point Sources: These are sources that are identified by point locations,
typically because they are regulated and their locations are available
in regulatory reports. Point sources can be further subdivided into
electric generating unit (EGU) sources and non-EGU sources, particularly
in criteria inventories in which EGUs are a primary source of NOX and
SO2. Examples of non-EGU point sources include chemical manufacturers
and furniture refinishers.

Area Sources: Sources that are treated as being spread over a spatial
extent (usually a county or air district) and that are not movable (as
compared to non-road mobile and on-road mobile sources). Because it is
not possible to collect the emissions at each point of emission, they
are estimated over larger regions. Examples of stationary area sources
are residential heating and architectural coatings. Numerous sources,
such as dry cleaning facilities, may be treated either as stationary
area sources or as point sources.

On-Road Mobile Sources: Vehicular sources that travel on roadways.
Emissions are estimated as the product of emissions factors and activity
data, vehicle miles traveled (VMT). Examples of on-road mobile sources
include light-duty gasoline vehicles and heavy-duty diesel vehicles.

Off-Road Mobile Sources: Off-road mobile sources are vehicles and
engines that encompass a wide variety of equipment types that either
move under their own power or are capable of being moved from site to
site.  Examples include agricultural equipment such as tractors or
combines, locomotives and oil field equipment such as mechanical
drilling engines. 

Oil and Gas Sources: Oil and gas sources consist of a number of
different types of activities from engine sources for drill rigs and
compressor engines, to sources such as condensate tanks and fugitive gas
emissions. The variety of emissions types for sources specific to oil
and gas activity can, in some cases, overlap with mobile, area or point
sources, but these can also be extracted and treated separately.

Biogenic Emissions: Biogenic emissions are based on the activity fluxes
modeled from biogenic land use data, which characterizes the types of
vegetation that exist in particular areas. Emissions are generally
derived using modeled estimates of biogenic gas-phase pollutants from
land use information, emissions factors for different plant species, and
meteorology data.

Dust: Dust emissions may have a variety of sources that could include
anthropogenic sources, natural sources, and natural sources that may be
influenced by anthropogenic activity.  For emissions summary purposes,
dust is classified here as fugitive dust and windblown dust.  Fugitive
dust includes sources such as road dust, agricultural operations,
construction and mining operations and windblown dust from vacant lands.


Fire: Fire sources are difficult to predict and control, and may have a
mix of natural and anthropogenic influences. Natural sources include
wildland fires, while anthropogenic sources can include agricultural and
prescribed fires.

2002 Arizona Emission Inventory Methodology

The 2002 Arizona Emission Inventory calculations and results are
described in detail in the Arizona Regional Haze SIP.  Please refer to
Chapter 8 of this document for detailed calculation methodology of the
2002 Arizona Emission Inventory.

The 2002 Arizona Emission Inventory served as the baseline inventory for
the 2018 emission inventory estimation.  The methodology by which 2018
emissions were estimated from the 2002 emission inventory can be found
in detail in the Arizona Regional Haze SIP and an overview of these
methodologies are shown in Table 1. 

2008 Arizona Emission Inventory Methodology

The creation of the 2008 State of Arizona Emission Inventory was a
collection of efforts between WRAP, ARS, and ADEQ.  This inventory
represents the most complete and recent calculation of pollutant
emissions from all identified sources currently available at the time of
this document’s preparation.  Portions of this inventory were derived
from the 2008 EPA National Emissions Inventory, while other portions
were calculated using a variety of emission models and techniques. 
Source data and emission estimation methodology are discussed in the
following section and compared to those methods utilized during the
creation of the 2002 EI.

The 2002 emission inventory reported emissions based on the most readily
available and accurate source data and methods at the time of
preparation; however, many of the calculated source category emissions
methodologies and input data changed between the 2002 and 2008 emission
inventory preparations in order to enhance the accuracy of estimated
statewide emissions.  For this reason, many of the source category
emission differences between the 2002 and 2008 inventories presented in
this document should be viewed as a mixture of methodology, input data,
and actual emissions changes. Furthermore, since the Arizona 2018 EI was
estimated through the adjustment of the 2002 baseline Arizona emission
inventory, emissions from these two EIs are more readily comparable than
emissions from the 2008 EI presented here due to the aforementioned
methodology and source data differences.

2002 & 2008 Emission Inventory Methodology Comparisons

Table 1 presents the input data and methodologies utilized to calculate
each source category for the 2002 and 2008 EIs.  This table illustrates
how some source category methodologies, input data, or modeling
resolutions changed dramatically, which in some cases is reflected in
estimated emissions between the two inventories. 

Table   SEQ Table \* ARABIC  1 :  Emission Inventory Methodology
Comparison for 2002 and 2008.

Inventory Sector	2002 Baseline Inventory

(WRAP Plan02d)	2008 Emission Inventory

(WRAP WestJump08 and DEASCO3)	Comments

Point Sources	The WRAP generated point source inventories for both
actual reported 2002 (Base02b) EGU and all other point source data, and
for a 2000-2004 average of EGU point sources (Plan02d).  

Inventories were generated using hourly EPA CAMD CEM data for EGUs.
Other point sources for both Base02b and Plan02d were developed in
consultation with states by the ERG contractor.

Plan02d emissions are used here because they are consistent with what
was reported as baseline conditions for most WRAP region SIPs.

More details are available here:

 HYPERLINK "http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx"
http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx  

and 

http://www.wrapair.org/forums/ssjf/pivot.html	2008 inventories were
generated using hourly EPA CAMD CEM data for EGUs. Other point sources
are from the 2008 NEI v2.  

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx

	Baseline conditions presented here represent a 5-year average for EGUs,
while progress period conditions are represented with 2008 data.



Area Sources	Plan02d emissions inventories were developed in
consultation with Arizona RPOs by the ERG contractor.

More details are available here:

http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx	Arizona
reported area sources from the 2008 NEI v2.

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx	Note that area oil and gas sources
are reported separately.



Point Oil and Gas	State reported point oil and gas sources for 2002 are
included here in the point source inventory totals.  

More details are available here:

http://www.wrapair.org/forums/ssjf/pivot.html	Different basins are
comprised of a combination of state reported point oil and gas from the
2008 NEI v2 for some areas and updated WRAP Phase III inventories for
other areas.  

These emissions were developed separately in some cases, but are
included in the point source inventory totals (see above). 

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx	This industry has expanded and
evolved considerably since 2002. 



Area Oil and Gas	Developed using WRAP Phase II emissions methodologies. 
Emissions process estimated included:

Drill Rigs

Wellhead Compressor Engines

CBM Pump Engines

Heaters

Pneumatic Devices

Condensate and oil tanks

Dehydrators

Completion Venting

More details are available here:

 HYPERLINK "http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx"
http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx  

and 

http://www.wrapair.org/forums/ogwg/documents/2007-10_Phase_II_O&G_Final)
Report(v10-07%20rev.s).pdf	Developed using WRAP Phase III emissions
methodologies using 2008 production data. The following additional
categories were included in addition to those listed for 2002:

Lateral compressor engines

Workover rigs

Salt-water disposal engines

Artificial lift engines 

Vapor recovery units (VRUs)

Miscellaneous or exempt engines

Flaring

Fugitive emissions

Well blowdowns

Truck loading

Amine units (and gas removal)

Water tanks

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx	Note that many more source
categories were counted in 2008 than in 2002.

Other differences between 2002 Phase II O&G emissions vs. Phase
III/WestJumpAQMS 2008 O&G emissions:.

Phase III 2008 estimates included new and/or revised estimation
methodologies for each of the equipment types and processes included,
e.g., the surveys provided counts by device type (low-bleed vs.
high-bleed) and specific information on control device efficiency, among
other improvements to activity data.  Phase II did not have that
information available, since no surveys were made in Phase II.

Phase III used detailed surveys of operators in each basin to determine
activities, practices, and counts of small “area-source” equipment
not typically permitted by the state.  WestJumpAQMS then carried forward
these survey data and adjusted emissions to 2008 based on production
data and any controls added after Phase III.

Phase III/WestJumpAQMS used the high-quality and complete IHS commercial
database of O&G production data by well by basin.  This was not used in
Phase II, instead the state O&G Commission databases, which have been
improved quite a bit over time, were used.

4) Phase III used more refined methodologies to estimate emissions, 
Phase III also asked states and operators for gas composition data by
basin that greatly increased the information available about VOC
emissions rates.

On-Road 

Mobile	EPA MOBILE6 model applied by ENVIRON contractor using NMIM VMT
defaults for all counties except Maricopa, for which local inputs were
provided.

More information is available in the Emissions Method section of the
WRAP TSS documentation:

 HYPERLINK "http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx"
http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx  

and
http://www.wrapair.org/forums/ef/UMSI/0606_WRAP_Mobile_Source_EI_Final_R
eport.pdf

	EPA MOVES2010a model in inventory mode utilizing national default data
for each county and MET4MOVES for meteorological data.

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx	Differences in models contributed
to some differences in emissions reported, but other disparities are due
to a combination of VMT changes and new controls on vehicles.



Off-Road Mobile	EPA draft NONROAD2004 model version data by ENVIRON with
inputs from Arizona for vehicle population allocation and county level
locomotive emissions.

More details are available here:

 HYPERLINK "http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx"
http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx  

and
http://www.wrapair.org/forums/ef/UMSI/0606_WRAP_Mobile_Source_EI_Final_R
eport.pdf	State reported off-road mobile sources for 2008 (NEI08v2).

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx	The off-road models include both
emission factors and default county-level population and activity data.

Fugitive Dust and Road Dust	Emission Methodology based on AP-42 guidance
and CARB procedures for the following source categories:

Agricultural Operations – County specific cropland acreage provided by
the State.

Construction Operations – data obtained from the US Census Bureau and
the Department of Commerce.

Road Dust – data obtained from the Federal Highway Administration and
State and local datasets.

Vegetative scavenging factors were applied pre-processing at the county
level

More details are available here:

http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx	Extracted
from state reported area sources for 2008 (NEI08v2).

Vegetative scavenging factors were applied post-processing at the higher
resolution grid cell level, compared to 2002 data.

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx	Note that fugitive dust and road
dust categories were available separately in WRAP Plan02d inventories,
but are combined for summary purposes here. For the 2008 inventory,
vegetative scavenging factors were applied to the combined sources; thus
these source categories are not easily separated.



Windblown Dust	Generated using WRAP Windblown Dust Model and 2002 MM5
meteorology, at 36km grid cell resolution.

Vegetative scavenging factors applied pre-processing at the county level

More details are available here:

http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx	Generated
using WRAP Windblown Dust Model and 2008WRF meteorology, at 4km and 12km
grid cell resolution for the WRAP region.

Vegetative scavenging factors applied post-processing at the grid cell
level.

More details are available here:

http://wrapair2.org/WestJumpAQMS.aspx	Difference between 2002 and 2008
meteorology introduce factors that make judgment of progress for this
category difficult.

MM5 vs. WRF met models – different actual meteorology in each year and
increased grid cell resolution in 2008.

Higher resolution of grid cells leads to higher average wind speeds in
individual cells, which leads to increased windblown dust emissions
aggregated at the county level.

MM5 Layer 1 36 meter height winds vs. WRF average winds across lowest 3
layers spanning ~40 meter height.

Error in 2002 WBD model application treating rainfall in cm as inches.

Biogenic	BEIS3.12 with BELD3 landuse and 2002 MM5 meteorology data, at
36km grid cell resolution.

More details are available here:

http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx	MEGAN2.10
with 2008 WRF meteorology data, at 4 and 12 km grid cell resolution

More details are available here:

 HYPERLINK "http://www.wrapair2.org/emissions.aspx"
http://www.wrapair2.org/emissions.aspx#  

	Comparisons of biogenic inventories between baseline and progress years
will show large differences due to methodology, and not actual changes
in emissions. Examples of biogenic emissions input factors that may
affect differences between the BEIS3.12 and MEGAN2.10 model outputs
include:

Different meteorological years and models (2002 MM5 vs. 2008 WRF).

Higher temporal and spatial variability of land cover and other
environmental input factors.

Improved emissions factors based on better sources of data (e.g.,
satellites and field studies).

A model comparison study between BEIS3.12 and MEGAN2.10 was performed by
WRAP and can be found at: 
http://wrapair2.org/pdf/Memo_9_Biogenics_May9_2012_Final.pdf

Fires	WRAP Phase III fire inventory

More details are available here:

http://vista.cira.colostate.edu/TSS/Results/Emissions.aspx	Current
summaries use interim WESTJUMP08 fire data currently based on satellite
fire data for 2008. This inventory does not separate anthropogenic from
natural fire.

DEASCO3 fire summaries will include separate reporting of anthropogenic
and natural fires.

More details are available here:

http://www.wrapfets.org/deasco3.cfm	Baseline conditions are represented
with a 5-year average of fire activity at the same locations and
occurrence dates as actually occurred in 2002, while progress period
conditions are represented with actual 2008 data.

Comparisons between these inventories are complicated by the variable
and sporadic nature of wildfires.



2002 and 2008 Emissions

The 2008 Arizona statewide EI was originally created by WRAP and ARS to
fulfill the requirements of 40 CFR § 51.308(g)(4) which states
individual States must address the change in the emissions of pollutants
which contribute to visibility impairment every 5 years for all sources
and activities.  ADEQ is currently submitting WRAP and ARS’ version of
the 2008 EI to fulfill the requirement of 40 CFR § 51.308(d)(4)(v). 
While the WRAP and ARS 2008 emission inventory provides the most
consistent inventory available in relation to the baseline 2002 State of
Arizona EI, Table 1 illustrates the differences which occurred during
the calculation of these inventories.  Inferences related to emission
changes between 2002 and 2008 should not be made for many of the source
sectors due to these emission changes likely resulting from model
resolution, methodology, and input data enhancements.  Despite this
concern, general observations regarding emission differences between the
two inventories are listed below.

For sulfur dioxide, all categories except area sources exhibited lower
emissions in the 2008 inventory as compared to the 2002 inventory, with
the largest difference reported in point sources.

For nitrogen dioxide, all source categories except area sources
exhibited lower emissions in the 2008 inventory as compared to the 2002
inventory, with the largest difference reported for mobile sources.

Ammonia emissions remained relatively similar between inventories, with
2008 area sources showing slightly higher estimated emissions than 2002
and 2008 on-road mobile sources showing lower estimated emissions.

2008 EI volatile organic compound emissions were much lower than 2002 in
biogenic sector, due to enhancements in biogenic inventory methodology.

Primary organic aerosol emissions from fire were higher in 2002 than
2008. Note that current year inventories represent only snapshots of
fire emissions for the year 2008.

Elemental carbon showed large decreases in fire emissions, but on-road
mobile emissions were higher in the 2008 inventory than the 2002
inventory.

Fine particulate matter (crustal) and coarse mass were much larger for
windblown, fugitive and road dust sectors of the 2008 EI as compared to
the 2002 EI. The increase in windblown dust is thought to be due in part
to enhancements in dust inventory methodology. The 2008 EI was also
slightly higher for area and point sources for the crustal components of
fine particulate matter as compared to the 2002 EI.

Estimated 2002 and 2008 Arizona Emissions

Table 2 and Figure 1 present 2002 and 2008 estimated SO2 emissions by
source category. Tables 3 and Figure 2 present data for NOx, and
subsequent Tables and Figures (Tables 4 through 9 and Figures 3 through
8) present NH3, VOCs, POA, EC, Fines and CM emissions for the years of
2002 and 2008.  Source categories are qualified (highlighted) in Tables
where methodology, input data, or modeling resolution enhancements are
believed to significantly affect emission differences between the 2002
and 2008 EIs.

Table   SEQ Table \* ARABIC  2 :  Arizona Sulfur Dioxide Emissions by
Source Category

Source Category	Sulfur Dioxide Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	94,716	79,015	-15,700

Area	2,677	3,678	1,001

On-Road Mobile	2,715	812	-1,904

Off-Road Mobile	4,223	673	-3,550

Area Oil and Gas	0	0	0

Fugitive and Road Dust	0	0	0

Anthropogenic Fire*



	Total Anthropogenic	104,330	84,177	-20,153 (-19%)

Natural Sources

Natural Fire*	4,559	607	-3,952

Biogenic	0	0	0

Wind Blown Dust	0	0	0

Total Natural	4,559	607	-3,952 (-87%)

All Sources

Total Emissions	108,890	84,784	-24,105 (-22%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  1 :  2002 and 2008 Sulfur Dioxide
Emissions by Source Category

Table   SEQ Table \* ARABIC  3 :  Arizona Nitrogen Oxide Emissions by
Source Category

Source Category	Nitrogen Oxides Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	69,968	60,759	-9,209

Area	9,049	39,403	30,354

On-Road Mobile	178,009	137,555	-40,453

Off-Road Mobile	66,414	33,857	-32,557

Area Oil and Gas	17	0	-17

Fugitive and Road Dust	0	0	0

Anthropogenic Fire*



	Total Anthropogenic	323,458	271,575	-51,882 (-16%)

Natural Sources

Natural Fire*	17,218	3,513	-13,704

Biogenic	27,664	15,256	-12,408

Wind Blown Dust	0	0	0

Total Natural	44,881	18,769	-26,112 (-58%)

All Sources

Total Emissions	368,339	290,344	-77,995 (-21%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  2 :  2002 and 2008 Nitrogen Oxide
Emissions by Source Category

Table   SEQ Table \* ARABIC  4 :  Arizona Ammonia Emissions by Source
Category

Source Category	Ammonia Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	531	971	440

Area	32,713	34,878	2,165

On-Road Mobile	5,035	2,377	-2,658

Off-Road Mobile	48	40	-8

Area Oil and Gas	0	0	0

Fugitive and Road Dust	0	0	0

Anthropogenic Fire*



	Total Anthropogenic	38,326	38,265	-61 (0%)

Natural Sources

Natural Fire*	3,878	0	-3,878

Biogenic	0	0	0

Wind Blown Dust	0	0	0

Total Natural	3,878	0	-3,878 (-100%)

All Sources

Total Emissions	42,203	38,265	-3,939 (-9%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  3 :  2002 and 2008 Ammonia Emissions by
Source Category

Table   SEQ Table \* ARABIC  5 :  Arizona Volatile Organic Compound
Emissions by Source Category

Source Category	Volatile Organic Compound Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	5,464	3,489	-1,975

Area	102,918	100,256	-2,661

On-Road Mobile	110,424	54,589	-55,834

Off-Road Mobile	56,901	42,297	-14,604

Area Oil and Gas	46	12	-34

Fugitive and Road Dust	0	0	0

Anthropogenic Fire*



	Total Anthropogenic	275,753	200,644	-75,109 (-27%)

Natural Sources

Natural Fire*	37,232	4,989	-32,243

Biogenic	1,576,698	686,255	-890,443

Wind Blown Dust	0	0	0

Total Natural	1,613,930	691,243	-922,686 (-57%)

All Sources

Total Emissions	1,889,682	891,887	-997,795 (-53%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  4 :  2002 and 2008 Volatile Organic
Compound Emissions by Source Category

Table   SEQ Table \* ARABIC  6 :  Arizona Primary Organic Aerosol
Emissions by Source Category

Source Category	Primary Organic Aerosol Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	276	410	134

Area	4,728	6,445	1,718

On-Road Mobile	1,583	2,666	1,083

Off-Road Mobile	2,006	1,383	-624

Area Oil and Gas	0	0	0

Fugitive and Road Dust	535	1,393	858

Anthropogenic Fire*



	Total Anthropogenic	9,128	12,298	3,169 (35%)

Natural Sources

Natural Fire*	48,625	5,669	-42,957

Biogenic	0	0	0

Wind Blown Dust	0	0	0

Total Natural	48,625	5,669	-42,957 (-88%)

All Sources

Total Emissions	57,754	17,966	-39,787 (-69%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  5 :  2002 and 2008 Primary Organic
Aerosol Emissions by Source Category

Table   SEQ Table \* ARABIC  7 :  Arizona Elemental Carbon Emissions by
Source Category

Source Category	Elemental Carbon Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	26	283	257

Area	449	1,337	889

On-Road Mobile	1,761	5,559	3,798

Off-Road Mobile	2,752	1,813	-940

Area Oil and Gas	0	0	0

Fugitive and Road Dust	39	47	8

Anthropogenic Fire*



	Total Anthropogenic	5,027	9,039	4,012 (80%)

Natural Sources

Natural Fire*	9,719	412	-9,307

Biogenic	0	0	0

Wind Blown Dust	0	0	0

Total Natural	9,719	412	-9,307 (-96%)

All Sources

Total Emissions	14,745	9,450	-5,295 (-36%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  6 :  2002 and 2008 Elemental Carbon
Emissions by Source Category

Table   SEQ Table \* ARABIC  8 :  Arizona Fine Particulate Matter
Emissions by Source Category

Source Category	Fine Particulate Matter Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	632	4,434	3,801

Area	4,223	7,906	3,684

On-Road Mobile	1,080	511	-569

Off-Road Mobile	0	97	97

Area Oil and Gas	0	0	0

Fugitive and Road Dust	10,072	24,592	14,520

Anthropogenic Fire*



	Total Anthropogenic	16,007	37,540	21,533 (>100%)

Natural Sources

Natural Fire*	3,945	1,938	-2,006

Biogenic	0	0	0

Wind Blown Dust	6,422	9,307	2,885

Total Natural	10,367	11,246	879 (8%)

All Sources

Total Emissions	26,373	48,785	22,412 (85%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  7 :  2002 and 2008 Fine Particulate
Matter Emissions by Source Category

Table   SEQ Table \* ARABIC  9 :  Arizona Coarse Particulate Matter
Emissions by Source Category

Source Category	Coarse Particulate Matter Emissions (tons/year)

	2002

(Plan02d)	2008

(WestJump2008)	Difference

(Percent Change)

Anthropogenic Sources

Point	8,473	5,260	-3,214

Area	1,384	2,389	1,005

On-Road Mobile	1,004	5,597	4,593

Off-Road Mobile	0	162	162

Area Oil and Gas	0	0	0

Fugitive and Road Dust	79,316	141,117	61,801

Anthropogenic Fire*



	Total Anthropogenic	90,178	154,525	64,348 (71%)

Natural Sources

Natural Fire*	10,125	1,692	-8,433

Biogenic	0	0	0

Wind Blown Dust	57,796	83,765	25,969

Total Natural	67,921	85,457	17,536 (26%)

All Sources

Total Emissions	158,099	239,983	81,884 (52%)

*Natural fire totals for the 2008 inventory include both anthropogenic
and natural sources. Updated data distinguishing these sources are
expected.

Figure   SEQ Figure \* ARABIC  8 :  2002 and 2008 Coarse Particulate
Matter Emissions by Source Category

Summary of Major Methodological Changes

The Arizona Department of Administration (ADOA) provides publicly
available records of State and County estimated populations for each
year dating to 1980.  For the years of 2002 and 2008, ADOA estimates
Arizona state-wide populations to be 5,470,720 and 6,629,455
respectively.  This is an increase in state-wide population of 21.2%. 
An increase in population this significant will undoubtedly lead to
pollutant emission changes for a number of source categories; however,
the extreme degree to which certain pollutants change for given source
categories indicate that population increases are not solely responsible
for emission changes between the 2002 and 2008 EIs.  Below is a list of
methodology, input data, and model resolution changes which are believed
to significantly contribute to emission differences between the 2002 and
2008 EIs.  This list describes the possible changes which could affect
all qualified data from Tables 2-9.

ADEQ has reviewed emission estimates to understand the drastic changes
in Area Source SOx and NOx emissions between the 2002 and 2008 EIs. 
This review indicated that these changes are due to a mixture
methodological changes and data completion issues.  Therefore, ADEQ
believes a more accurate indicator of NOx and SOx changes are the
IMPROVE data.

Biogenic emission differences for NOx and VOCs are primarily due to
methodology, source data, and modeling resolution enhancements between
2002 and 2008.

Ammonia emission differences for On-road Mobile are primarily due to a
switch from the MOBILE6 model to the MOVES model.  The 2008 EPA NEI
Technical Support Document (TSD) reported a 54% decrease in highway
vehicle NH3 for 2008.

VOC emission differences for On-road Mobile are primarily due to a
switch from the MOBILE6 model to the MOVES model.

On-road Elemental Carbon (EC) and Coarse Particulate Matter (CM)
emission differences are primarily due to the switch between MOBILE6 and
MOVES (which estimates higher PM emissions).

Reported Point Source Fines emissions exhibit a dramatic increase
between 2002 and 2008, while CM decreases between 2002 and 2008.  In
theory, these two pollutants should track fairly closely to one another.
 ADEQ internal review revealed that many, if not most, sources within
the State of Arizona were not reporting PM2.5 prior to 2006 which likely
explains the drastic change in Fines emissions between the 2002 and 2008
EIs.

Area source Fines emission differences are partially due to NEI changes.
 Calculation methodology changes resulted in an overall increase in
Agricultural Tilling and Livestock emissions of 67% for the 2008 NEI.

Fugitive and Road Dust Fines and CM emission differences are primarily
due to NEI changes.  Calculation methodology changes resulted in an
overall increase in Paved Road Dust emission of 128% for the 2008 NEI.

Windblown Dust Fines and CM emission differences are primarily due to
the WRAP Windblown Dust (WBD) Model enhancing meteorological inputs and
model resolution between the 2002 and 2008 emissions calculations.

Regional Inventory Trends for Emissions

Most of the emission difference qualifying statements ADEQ presents in
Section V.D.2. are attributable to changes in input data origination or
calculation methodologies for emission estimations by sector.  Since
WRAP and ARS created statewide emission inventories for all of the
western US using similar methodology, it is reasonable to believe that
these qualifying statements would hold true for all of the compiled
emission inventories.  In this section ADEQ presents WRAP and ARS
produced figures of differences in 2002 and 2008 emission inventories by
state for 3 different pollutants to determine if the previously
mentioned qualifiers hold true.  These graphs split State emissions into
emitting source categories to identify trends for each of the calculated
or reported sources.

VOC emissions by State are presented in Figure 9.  This figure easily
illustrates qualifiers #2 and #4 from Section V.D.2.  The most evident
trend in this figure is the drastic decrease in Biogenic emissions for
each State.  These decreases are extreme and ubiquitous throughout the
region.  This trend supports qualifier #2, that Biogenic VOC emission
differences are primarily due to enhancements in calculation
methodology.  In addition, On-road Mobile emissions show reasonably
large decreases for each State.  While EPA reports a decrease of
national VMT by 0.8% for 2005-2008, it is unlikely that this small
decrease in VMT would be seen in every state.  The state of Arizona
showed a 21.2% population increase between 2002 and 2008, which would
likely result in a substantial VMT increase.  Therefore, it is
reasonable to believe that qualifier #4, that VOC decreases are likely
due to a switch from the MOBILE6 model to the MOVES model, is true.

Figure 10 presents State emissions for CM for the western US.  The
trends within this Figure are more regionally based, rather than
characteristic of the entire western US.  Coarse Mass emissions are due
to physical disturbance of an area of land by anthropogenic activities
(e.g. construction, driving on unpaved roadways, etc), natural
activities (e.g. animal movement or burrowing), or a mixture of the
natural and anthropogenic activities (e.g. wind suspension of dust from
a cleared area).  While the activity which creates the emissions may
change, the magnitude of emissions created are going to be primarily
dependant on the local environment.  Meteorology, soil

Figure   SEQ Figure \* ARABIC  9 :  Regional differences in VOC
emissions between 2002 and 2008 Emission Inventories

Figure   SEQ Figure \* ARABIC  10 :  Regional differences in Coarse Mass
emissions between 2002 and 2008 Emission Inventories

characteristics, and vegetation coverage are going to play a large role
in the magnitude of emissions produced from a certain area.  Therefore,
when examining Figure 10, it is important to group the States which have
a similar local environment.  Arizona, Southern California, Nevada, New
Mexico, and Utah comprise the southwestern US which is characterized by
its arid nature, in turn leading to sparse vegetation coverage.  When
examining Figure 10 for these five states, it is evident that local
environmental factors play a large role in how Windblown Dust emissions
differed between the 2002 and 2008 EIs.  All five southwestern US States
exhibit similar emission differences for Windblown Dust and Fugitive and
Road Dust.  Southwestern US State Windblown Dust emissions are likely to
be more affected by WRAP WBD model resolution increases and the
decreased precipitation, as reported in Table 1, than surrounding States
due to higher local wind speeds increasing dust suspension into the
atmosphere from dry, unvegetated soils.  ADEQ believes that Road Dust is
primarily responsible for the emission changes seen in the Fugitive and
Road Dust category.  The 2008 NEI reports that road dust emissions
increased by 128% over the previous NEI for the US.  Since the 2008 NEI
v1.5 was used for this source category, it is believed that this is the
reason for the difference between 2002 and 2008 emissions for the
combined categories of Fugitive and Road Dust.  Furthermore, the aridity
of the southwestern US would likely result in road dust  

Figure   SEQ Figure \* ARABIC  11 :  Regional differences in Fine Mass
emissions between 2002 and 2008 Emission Inventories

calculation disparities being maximized in this region, when compared to
other regions of the US.  Fines (Figure 11) show similar regional trends
for Windblown Dust and Fugitive and Road Dust for the arid southwestern
US as was reported for CM, further supporting the theory that
particulate matter emission differences are at least partially due to
calculation methodology changes.  While not proving qualifying statement
#7, the lack of a regional trend for Fines originating from Point
sources provides credence to this point.

IMPROVE Monitoring Data

As discussed in Section 3, comparisons between the 2002 and 2008 State
of Arizona EIs are problematic due to source data and methodology
changes.  Therefore, ADEQ has determined that IMPROVE monitoring data
are more appropriate surrogates for assessing visibility change due to
emission increases or decreases within the State of Arizona. 
Comparisons between the baseline (2000-2004) and progress (2005-2009)
periods are presented in this document in order to address EPA’s
assessment within the Federal Register that the State of Arizona
Reasonable Progress Goals (RPGs) are not acceptable for  reaching the
2064 natural visibility standards.  In this section we present IMPROVE
data comparisons between the baseline and progress periods, which is
discussed in terms of how this progress compares to RPGs in Section VII.

Data Completeness Requirements

The following information was gathered directly from ARS and describes
IMPROVE data completeness for the State of Arizona.  Furthermore, it
outlines the steps and methods utilized to gap fill missing data sets.

Progress for the Regional Haze Rule (RHR) is determined using 5-year
average visibility conditions. EPA’s guidance for tracking Regional
Haze progress includes data completeness requirements designed to ensure
that calculated averages include enough data to sufficiently represent
each daily, annual and 5-year period. The guidance specifies that the
2000-2004 baseline period and each subsequent 5-year average progress
period meet the following conditions:

Individual samples must contain all species required for the calculation
of light extinction (amm. sulfate, amm. nitrate, POM, EC, soil, coarse
mass, and sea salt)

Calendar seasons must contain at least 50% of all possible daily samples

Calendar years must contain at least 75% of all possible daily samples

Calendar years must not contain more than 10 consecutive missing daily
samples

The 5-year baseline and each 5-year progress period averages must
contain at least 3 complete years of data

RHR guidance specifies that if a 5-year period has less than three
complete years of data, then estimates should be prepared through
consultation with EPA/OAQPS.  For the state of Arizona, the 2005-2009
progress period did not have complete data available for one site. The
SIAN1 site, the Sierra Ancha Wilderness Area, did not meet RHR data
completeness criteria for the years 2006, 2007 and 2008, which
interrupted the requirement for 3 complete years required for a 5-year
average.  Substitution methodology was consistent with methodology
previously applied to the 2000-2004 baseline period for seven WRAP
sites. 

The data substitution methods include estimating missing species from
other on-site measurements and appropriately scaling data collected at a
nearby site which demonstrated favorable long-term comparisons. Only
years deemed incomplete under RHR guidance were candidates for
additional data substitutions, which included for the SIAN1 the years
2006, 2007 and 2008. Years deemed complete were not changed, although
there may have been missing samples during those years.  Substitution
methodology is described in detail below.

SIAN1 data substitution methodology

The first substitution method applied uses organic hydrogen as a
surrogate for organic carbon (OC) and elemental carbon (EC), which are
collected on the IMPROVE C module. Hydrogen (H) is measured on the A
module filter, and is assumed to be primarily associated with organic
carbon and inorganic compounds such as ammonium sulfate. Therefore, OC
can be estimated using the historical comparison between estimated
organic H and OC. Organic H is estimated by subtracting the portion of H
that is assumed to be associated with the inorganic compounds from the
total H (Org_H = H – 0.25*S). Linear regression statistics were used
to correlate all organic H and OC mass collected at the SIAN1 site
during the 2005-2009 period, and regression statistics were applied to
organic H to estimate OC on days where organic H was available, but OC
was not. OC and EC correlations for the period were then used to
calculate EC from OC. Regression statistics for these substitutions were
calculated and applied quarterly to account for seasonal variations.

Because the carbon data substitution methods were not sufficient to
complete the required years, a second method was applied that involved
scaling data from the closest neighboring IMPROVE site, TONT1. This site
had previously been determined to have favorable long-term comparisons
and similar regional characteristics for substitutions performed on the
2000-2004 baseline period, when the SIAN1 site was selected, in
consultation with the state of Arizona, as a donor site for TONT1.
Species specific correlations between SIAN1 and TONT1 of mass data
collected during the 2005-2009 period were calculated quarterly, and
applied to adjust TONT1 data to apply to missing SIAN1 days.

Figure 12 presents bar charts showing daily SIAN1 extinction data,
including substituted data, for the 2005-2009 progress period years.
Original RHR data in blue and substituted data by species in the
standard IMPROVE colors. Substituted days are also indicated by a black
bar underneath the day, and the red line indicates the threshold above
which days are counted in the 20% worst days for that year. Note that
some of the missing extinction days had partial data available and only
individual species missing in a given sample were substituted. Figure 13
presents similar bar charts indicating speciation of all data, with days
in which all or part of the day was substituted indicated by a black bar
underneath the day. Note that very few of the substituted days were
counted among the 20% worst days for the substituted years. All
summaries for the SIAN1 site in this progress report support document
include these substituted data, and substituted data and detailed
methodology information will also be made available on the WRAP TSS
website ( HYPERLINK "http://vista.cira.colostate.edu/tss/"
http://vista.cira.colostate.edu/tss/ ).

Figure   SEQ Figure \* ARABIC  12 :  IMPROVE SIAN1 data collected during
the 2005-2009 progress period, where original SIAN1 RHR data are
depicted in dark blue, and substituted data are depicted with separate
colors by species.

Figure   SEQ Figure \* ARABIC  13 :  IMPROVE SIAN1 data collected during
the 2005-2009 progress period, where substituted days are depicted with
a black bar beneath the data.

Baseline and Progress Period Visibility

This section summarizes IMPROVE monitoring and emission data comparing
the 2000-2004 baseline period to the current 2005-2009 progress period
for the state of Arizona, in line with regulatory requirements for
periodic progress (CFR 51.308(g)(3)(iii)).  Furthermore, a more robust
10 year trend analysis is presented to determine how an alternate
methods of visibility trend analysis may affect the conclusions drawn
from IMPROVE monitoring data analysis.

Arizona has 12 mandatory Federal Class I areas with associated IMPROVE
monitors. The basic premise of the RHR is to ensure that visibility on
the 20% worst days continues to improve at each Federal CIA, and that
visibility on the 20% best days does not get worse, as measured in units
of deciviews (dv) calculated from data collected at representative
Interagency Monitoring of Protected Visual Environments (IMPROVE)
monitoring sites. In addition to presenting the results of EPA
standardized 20% worst and 20% best days comparisons (RHR method), ADEQ
submits an alternative method of assessing visibility changes at State
of Arizona IMPROVE monitoring sites between the years of 2000 and 2009. 
This method utilizes Theil statistics (Theil method) to calculate an
annual trend for the 10 year period of interest and is an EPA accepted
method for annual pollutant trend analysis.  More description of Theil
method is given in Section VI.B.4.

Some of the major implications from monitoring and summaries presented
in this section are listed below, and more detailed information is
provided in the monitoring and data summaries sub-sections.

For RHR method analysis of the 20% best days, the 5-year average
deciview metric decreased at all Arizona sites, except GRCA2 which saw
no change.  The Theil method showed the same results.

For RHR method analysis of the 20% worst days, the 5-year average
deciview metric decreased at most sites, but increased at the GRCA2 and
IKBA1 sites.  The Theil method showed similar results, except no
statistically significant increasing trends.  GRCA2 and PEFO1 showed no
change when analyzed using the Theil method.

All sites experienced visibility extinction decreases of ammonium
nitrate using the RHR method. The Thiel method showed no significant
increasing trends and four sites with significant decreasing trends. 
Central and northern Arizona sites showed statistically significant
decreasing annual average trends in ammonium nitrate using the Theil
method.

 

RHR method analysis showed that ammonium sulfate increased at most
sites, though 20% worst days for ammonium sulfate showed no increases
using the Theil method.  Observations regarding ammonium sulfate were as
follows:

Ammonium sulfate RHR method increases were primarily due to higher than
average ammonium sulfate measured in 2005.

When selecting the 20% worst days for ammonium sulfate alone, no sites
showed increases when using the Theil method, and five sites showed
statistically significant decreasing trend.

RHR method analysis exhibited decreases in visibility extinction from
particulate organic mass at all sites except GRCA2 and IKBA1.  Theil
method analysis showed statistically significant particulate organic
mass decreases at four Arizona sites and no statistically significant
increasing trends at any of the other sites.

An overall visibility deciview increase at the GRCA2 site was seen when
analyzing the data using the RHR method. Two visibility components
contributing to this increase were particulate organic mass and
elemental carbon.  In June 2009, the GRCA2 was in close proximity to 3
simultaneous, lightening induced wildfires.  Observations regarding two
visibility components associated with wildfire emissions are given
below.  These results indicate GRCA2 visibility deciview changes were at
least partially due to the 2009 wildfires:

Elemental carbon showed a fairly large increase in visibility extinction
using the RHR method; however, annual average elemental carbon
measurements did not show increasing trends using the Theil method. 

GRCA2 showed an increase in visibility extinction using the RHR method
for particulate organic matter; however, annual average elemental carbon
measurements did not show increasing trends using the Theil method. 

The overall visibility deciview increase at IKBA1 was affected by high
measurements in 2005.  POM and ammonium sulfate are the primary
contributing visibility components to the overall increasing deciview
trend.  These two visibility components are discussed below:

Particulate organic mass showed a large increase visibility extinction
using the RHR method, but did not show an increasing trend using the
Theil method.  Particulate matter increases were strongly controlled by
a large wildfire in July of 2005.

Ammonium sulfate showed a large increase in visibility extinction using
the RHR method, but did not show an increasing trend using the Theil
method.  This large increase in ammonium sulfate using the RHR method
was a regional trend (discussed later).

Progress Period (2005-2009) Visibility

This section addresses the regulatory question, what are the current
visibility conditions for the most impaired and least impaired days (40
CFR 51.308 (g)(3)(i))? RHR guidance specifies that 5-year averages be
calculated over successive 5-year periods, i.e. 2000-2004, 2005-2009,
2010-2014, etc. (EPA 2003). Current visibility conditions are
represented here as the most recent successive 5-year average period
available, or the 2005-2009 period average, although the most recent
IMPROVE monitoring data currently available includes 2010.

Tables 10 and 11 present the calculated deciview values for each site,
along with the percent contribution to extinction from each aerosol
component for the 20% most impaired and 20% least impaired days for each
of the Federal CIA IMPROVE monitors in Arizona. Figure 14 presents
5-year average extinction for the current progress period for both 20%
most impaired and 20% least impaired days.

Specific observations for the current visibility conditions on the 20%
most impaired days are as follows:

The largest contributors to aerosol extinction on the 20% most impaired
days at Arizona sites were particulate organic mass, ammonium sulfate
and coarse mass.

The highest aerosol extinction (15.2 dv) was measured at the SYCA1 site,
where particulate organic mass was the largest contributor to aerosol
extinction, followed by coarse mass. The lowest aerosol extinction (11.8
dv) was measured at the BALD1 site.

Specific observations for the current visibility conditions on the 20%
least impaired days are as follows:

Rayleigh, or the background visibility impairment due to atmospheric
gases in clean air, was the largest contributor to light extinction at
all sites for the 20% least impaired days. Average extinction for the
least impaired visibility days at the Arizona sites ranged between 2.2
deciview (GRCA2) and 8.0 deciview (SAWE1).

For all Arizona sites except SIAN1 and SAWE1, ammonium sulfate was the
largest contributor to aerosol extinction for the 20% least impaired
days.

At the SIAN1 site, particulate organic mass was the largest contributor
to aerosol extinction for the best days, followed by ammonium sulfate.
At the SAWE1 site, coarse mass was the largest contributor, followed by
ammonium sulfate.

Table   SEQ Table \* ARABIC  10 :  Relative Contribution of Pollutants
to Visibility Conditions on the 20% Most Impaired Days at Arizona Class
I area IMPROVE Sites for the Progress Period (2005-2009).

Site	Deciviews (dv)	Percent Contribution by Component

(% of Mm-1) and Rank*



Ammonium Sulfate	Ammonium Nitrate	Particulate Organic Mass	Elemental
Carbon	Soil	Coarse Mass	Sea Salt

BALD1	11.8	25% (2)	4% (6)	42% (1)	8% (4)	6% (5)	16% (3)	0% (7)

CHIR1	12.2	36% (1)	5% (5)	16% (3)	5% (6)	10% (4)	27% (2)	1% (7)

GRCA2	12.0	22% (2)	7% (5)	41% (1)	11% (4)	6% (6)	12% (3)	0% (7)

IKBA1	13.4	26% (2)	8% (5)	29% (1)	8% (6)	8% (4)	21% (3)	1% (7)

PEFO1	13.0	23% (2)	5% (6)	31% (1)	11% (4)	8% (5)	21% (3)	1% (7)

SAGU1	13.6	25% (2)	9% (5)	18% (3)	8% (6)	11% (4)	28% (1)	1% (7)

SAWE1	14.9	21% (2)	11% (5)	16% (3)	8% (6)	13% (4)	31% (1)	1% (7)

SIAN1	13.0	25% (2)	6% (6)	33% (1)	9% (4)	8% (5)	19% (3)	1% (7)

SYCA1	15.2	15% (4)	4% (6)	29% (1)	9% (5)	15% (3)	28% (2)	0% (7)

TONT1	13.8	28% (1)	8% (5)	21% (3)	7% (6)	9% (4)	26% (2)	1% (7)

*Highest contribution per site is highlighted in bold.

Table   SEQ Table \* ARABIC  11 :  Relative Contribution of Pollutants
to Visibility Conditions on the 20% Least Impaired Days at Arizona Class
I area IMPROVE Sites for the Progress Period (2005-2009).

Site	Deciviews (dv)	Percent Contribution by Component

(% of Mm-1) and Rank*



Ammonium Sulfate	Ammonium Nitrate	Particulate Organic Mass	Elemental
Carbon	Soil	Coarse Mass	Sea Salt

BALD1	2.9	35% (1)	7% (5)	26% (2)	13% (4)	5% (6)	13% (3)	1% (7)

CHIR1	4.4	38% (1)	7% (5)	17% (3)	10% (4)	6% (6)	21% (2)	1% (7)

GRCA2	2.2	45% (1)	13% (4)	15% (2)	9% (5)	4% (6)	14% (3)	1% (7)

IKBA1	5.1	29% (1)	10% (5)	28% (2)	12% (4)	5% (6)	14% (3)	1% (7)

PEFO1	4.6	31% (1)	9% (5)	21% (2)	19% (3)	6% (6)	14% (4)	0% (7)

SAGU1	6.7	28% (1)	8% (6)	20% (3)	12% (4)	8% (5)	21% (2)	2% (7)

SAWE1	8.0	24% (2)	8% (6)	18% (3)	11% (4)	10% (5)	26% (1)	2% (7)

SIAN1	5.3	27% (2)	7% (5)	32% (1)	17% (3)	5% (6)	13% (4)	1% (7)

SYCA1	5.1	27% (1)	10% (5)	23% (2)	17% (3)	7% (6)	15% (4)	1% (7)

TONT1	5.7	33% (1)	9% (5)	23% (2)	12% (4)	6% (6)	16% (3)	1% (7)

*Highest contribution per site is highlighted in bold.

Figure   SEQ Figure \* ARABIC  14 :  Average Extinction for Current
Progress Period (2005-2009) for the Worst (Most Impaired) and Best
(Least Impaired) Days Measured at Arizona Class I area IMPROVE Sites.

Visibility Trend Analyses

This section addresses the regulatory question, what is the difference
between current visibility conditions for the most impaired and least
impaired days and baseline visibility conditions (40 CFR 51.308
(g)(3)(ii))? Baseline visibility conditions are the basis against which
improvements in worst day visibility, and lack of degradation for the
best day visibility, are judged. Included here are comparisons between
the 5-year average baseline conditions (2000-2004) and the current
progress period extinction (2005-2009).  ADEQ refers to this method as
the RHR method within this document.

ADEQ further presents an alternative analysis for visibility trend
analysis for the 2000-2009 period.  Alternative methodology is presented
to better understand how anamalous years may have affected visibility
changes as measured at Arizona Class I Areas.  The Theil method was
chosen to characterize visibility trends, as this method has been
generally accepted by EPA in previous trend analyses, most notably in
previously prepared EPA National Air Quality and Emissions Trends
Reports.  While the RHR method is the important metric for RHR
regulatory purposes, trend statistics (e.g. the Theil Method) may be of
value to understand and address visibility impairment issues for
planning purposes.

Theil Trend

Ten-year visibility trends were analyzed for the State or Arizona in
order to better understand how anamalous years may have affected
visibility changes as measured at Arizona Class I Areas.  The Theil
method was chosen to characterize visibility trends, as this method has
been generally accepted by EPA in previous trend analyses, most notably
in previously prepared EPA National Air Quality and Emissions Trends
Reports.  

EPA described the statistical method as follows:

“The Theil test is a nonparametric statistical test that can be used
instead of regression-based methods for discerning a monotonic trend. It
examines whether the concentration from year to year tends to increase
or decrease consistently, making it a test of monotonicity. This test is
not concerned with the magnitude of the year-to-year differences. The
null hypothesis is that there is no monotonic trend in the data. 

The first step in the test is to examine all possible [n(n-1)/2] pairs
of data points from a given monitor, where n = 8, 9, 10, or 11. Next, a
count is taken of all the pairs that show an increasing or decreasing
trend. The null hypothesis will be rejected and the test results will
indicate a significant monotonic increasing (or decreasing) trend if
this count of the data point pairs is greater than (or less than) a
certain critical value. A large positive value indicates a positive
trend, and a large negative value indicates a negative trend.  

The Theil test was applied for two reasons. First, it is appropriate
when the errors from a linear regression are not normally, or close to
normally, distributed. The data here may not meet the normality
assumption.  Second, this test was recommended to EPA for determining
whether an area has a significant trend.  Therefore, this test is used
in EPA’s annual Trends Reports.”

Annual trends reported here were calculated by ARS for the years
2000-2009, with a trend defined as the slope derived using Thiel method.
 Trend statistics are useful in analyzing changes in air quality data,
because these statistics can show the overall tendency of measurements
over long periods of time to increase or decrease, while minimizing the
effects of the year-to-year fluctuations which are common in air quality
data.  The significance of the trend is represented using p-values
calculated using Mann-Kendall trend statistics.  Determining a
significance level helps to distinguish random variability in data from
a tendency to increase or decrease over time, where lower p-values
indicate higher confidence levels in the computed slopes.  In some
cases, trends may show decreasing tendencies where the difference
between the 5-year averages do not. In these cases, the 5-year average
is the important metric for RHR regulatory purposes, but trend
statistics may be of value to understand and address visibility
impairment issues for planning purposes.

2000-2009 Visibility Trend Analyses Results

This section presents visibility progress between 2000-2009 through the
use of the two previously discussed methodologies:  1) the RHR method
and 2) the Theil method.

Table 12 presents the differences between the 2000-2004 baseline period
average and the 2005-2009 progress period average for each site in
Arizona for the 20% most impaired days by use of the RHR method, and
Table 13 presents similar data for the least impaired days. Averages
that increased are depicted in red text and averages that decreased in
blue. Figure 15 presents the 5-year average extinction for the baseline
and current progress period averages for 20% most impaired days and
Figure 16 presents the differences in averages by component, with
increases represented above the zero line and decreases below the zero
line. Figures 17 and 18 present similar plots for the 20% least impaired
days using the RHR method.

10 year trends for individual visibility extinction components using the
Theil method are presented in Table 14.  Only averages with p-value
statistics less than 0.15 (85% confidence level) are presented in Table
14, with increasing slopes in red and decreasing slopes in blue. The
regional haze regulations refer specifically to changes in extinction
for the 20% most impaired and least impaired days, but trend statistics
are also presented in Table 14 for an average of all sampled days.
Selection of the most impaired and least impaired days can vary
seasonally from year to year, so in some cases the annual average of all
sampled days may better represent actual aerosol component trends over
time.

Some general observations regarding changes in visibility impairment at
sites in Arizona are as follows:

For the 20% most impaired days, the RHR method exhibited deciview metric
increases between the 2000-2004 and 2005-2009 periods at the GRCA2 and
IKBA1 sites and decreases at all other Arizona sites.  Theil method
analysis showed no sites with significantly increasing deciview metric
trends between 2000-2009 and significantly decreasing deciview metric
trends at all sites except BALD1, GRCA2, IKBA1, PEFO1, and SYCA1. 
Notable differences for individual components extinctions (mM-1) on the
20% most impaired days were as follows:

All sites except GRCA2 and IKBA1 measured decreases in particulate
organic mass using the RHR method.  No sites showed significantly
increasing trends using the Theil method and 4 sites showed
significantly decreasing trends.

The RHR method analysis of ammonium sulfate showed increased extinction
at all Arizona sites except SAGU1 and SAWE1, with the largest increases
in ammonium sulfate were measured at the CHIR1, IKBA1 and TONT1 sites.
In contrast, no statistically significant increasing annual trends in
ammonium sulfate were measured using the Theil method. Decreasing annual
ammonium sulfate trends on the order of about 0.1 Mm-1/year were
measured at the BALD1, CHIR1, SAGU1 and SAWE1 sites. Anomalously high
ammonium sulfate occurred in 2005 at most Arizona sites, which
influenced the increases noted using the RHR method.

RHR method analysis of ammonium nitrate extinction showed decreases at
all Arizona sites for the 20% most impaired days. Analysis of all
measured days showed no increasing trends, and decreasing trends on the
order of 0.1 Mm-1/year at the IKBA1, SAGU1, SAWE1, SIAN1 and TONT1
sites.

RHR method analysis of coarse mass revealed increasing extinction values
at BALD1, SAGU1, SYCA1, and TONT1.  However, only the BALD1 site showed
a statistically significant increasing trend for coarse mass for all
measured days on the order of approximately 0.1 Mm-1/year. 

Soil progress and baseline average differences decreased for five sites
using the RHR method while measuring highest at the PEFO1, BALD1, and
TONT1 sites for the 20% most impaired days. Theil method analysis showed
increasing trends at only 2 Arizona sites for the 20% most impaired days
(BALD1 and PEFO1) while SYCA1 showed a significantly decreasing trend.

Increases in deciview at the GRCA2 site using the RHR method were mostly
due to increases in ammonium sulfate and elemental carbon and the lack
of a decreasing particulate organic mass extinction which occurred at
most other Arizona Class 1 Areas. Higher progress period measurements at
GRCA2 were influenced by large events between June and August of 2009. 
These increases were partially offset by decreases in ammonium nitrate
and coarse mass.  This site did not show significantly increasing
ammonium sulfate trends using the Theil method. 

Increases in deciview at the IKBA1 site were mostly due to increased
ammonium sulfate and particulate organic mass measurements. Higher
progress period measurements at IKBA1 were influenced by large events in
July 2005.  These increases were partially offset by decreases in
ammonium nitrate and soil.  This site did not show significantly
increasing ammonium sulfate trends using the Theil method.

For the 20% least impaired days, the RHR method exhibited decreasing
deciview metrics at all sites except GRCA2, where the measured deciview
average remained the same. Notable differences for individual component
averages on the 20% least impaired days were as follows:

The largest decreases were due to particulate organic mass, which
decreased at all sites except IKBA1 using the RHR method.  Theil method
analysis revealed significantly decreasing trends at 7 of the sites.

Ammonium sulfate decreased at most sites, but increased slightly at the
GRCA2, SAGU1 and SYCA1 sites using the RHR method.  Theil methodology
revealed no statistically significant increasing site trends and 3 

Table   SEQ Table \* ARABIC  12 :  Difference in Aerosol Extinction by
Component between the Baseline Period (2000-2004) and the Progress
Period (2005-2009) on the 20% Most Impaired Days for Arizona Class I
IMPROVE Sites.

Site	Deciview (dv)	Change in Extinction by Component (Mm-1)*

	2000-2004

Baseline

Period	2005-2009

Progress

Period	Change in dv*	Amm.

Sulfate	Amm.

Nitrate	POM	EC	Soil	CM	Sea

Salt

BALD1	11.8	11.8	0.0	+0.3	-0.1	-2.1	-0.7	+0.4	+1.3	+0.1

CHIR1	13.4	12.2	-1.2	+1.0	-0.1	-3.2	-0.5	-0.3	-1.9	+0.2

GRCA2	11.7	12.0	+0.3	+0.5	-0.4	+0.1	+0.5	+0.1	-0.3	0.0

IKBA1	13.3	13.4	+0.1	+1.0	-1.2	+0.7	0.0	-0.3	0.0	+0.1

PEFO1	13.2	13.0	-0.2	+0.5	-0.3	-1.4	+0.5	+0.6	-1.0	+0.1

SAGU1	14.8	13.6	-1.2	-0.1	-3.2	-4.1	-0.9	-0.1	+1.2	+0.2

SAWE1	16.2	14.9	-1.3	-0.7	-2.3	-1.9	-0.5	-1.4	-2.2	+0.2

SIAN1	13.7	13.0	-0.7	+0.7	-0.3	-2.5	+0.1	+0.1	-0.6	+0.2

SYCA1	15.3	15.2	-0.1	+0.7	-0.7	-0.5	+0.4	-1.0	+1.4	0.0

TONT1	14.2	13.8	-0.4	+1.3	-0.5	-3.5	-0.6	+0.4	+0.5	+0.2

*Change is calculated as progress period average minus baseline period
average. Values in red indicate increases in extinction, values in blue
indicate decreases.

sites experienced statistically significant decreases in ammonium
sulfate trends.

Ammonium nitrate decreased at all but the GRCA2 site using the RHR
method and 4 of the decreasing sites were found to be statistically
significant using the Theil method.

Table   SEQ Table \* ARABIC  13 :  Difference in Aerosol Extinction by
Component between the Baseline Period (2000-2004) and the Progress
Period (2005-2009) on the 20% Least Impaired Days for Arizona Class I
IMPROVE Sites.

Site	Deciview (dv)	Change in Extinction by Component (Mm-1)*

	2000-2004

Baseline

Period	2005-2009

Progress

Period	Change in dv*	Amm.

Sulfate	Amm.

Nitrate	POM	EC	Soil	CM	Sea

Salt

BALD1	3.0	2.9	-0.1	-0.1	-0.1	-0.1	0.0	0.0	+0.1	0.0

CHIR1	4.9	4.4	-0.5	-0.2	-0.1	-0.5	-0.1	0.0	0.0	0.0

GRCA2	2.2	2.2	0.0	+0.1	0.0	-0.1	0.0	0.0	0.0	0.0

IKBA1	5.4	5.1	-0.3	-0.3	-0.2	+0.1	0.0	-0.1	-0.1	+0.1

PEFO1	5.0	4.6	-0.4	-0.1	-0.2	-0.4	0.0	+0.1	0.0	0.0

SAGU1	6.9	6.7	-0.2	+0.1	-0.2	-0.2	-0.1	-0.3	+0.3	+0.1

SAWE1	8.6	8.0	-0.6	-0.2	-0.1	-0.5	-0.4	-0.3	+0.2	+0.2

SIAN1	6.2	5.3	-0.9	-0.3	-0.4	-0.7	-0.1	0.0	0.0	0.0

SYCA1	5.6	5.1	-0.5	+0.1	-0.1	-0.6	-0.2	-0.1	+0.1	0.0

TONT1	6.5	5.7	-0.8	-0.2	-0.2	-0.5	-0.2	-0.1	-0.2	+0.1

*Change is calculated as progress period average minus baseline period
average. Values in red indicate increases in extinction, values in blue
indicate decreases.

Figure   SEQ Figure \* ARABIC  15 :  Average Extinction for Baseline and
Progress Period Extinction for Worst (Most Impaired) Days Measured at
Arizona Class I area IMPROVE Sites.

Figure   SEQ Figure \* ARABIC  16 :  Difference between Average
Extinction for Current Progress Period (2005-2009) and Baseline Period
(2000-2004) for the Worst (Most Impaired) Days Measured at Arizona Class
I area IMPROVE Sites.

Figure   SEQ Figure \* ARABIC  17 :  Average Extinction for Baseline and
Progress Period Extinction for Best (Least Impaired) Days Measured at
Arizona Class I area IMPROVE Sites.

Figure   SEQ Figure \* ARABIC  18 :  Difference between Average
Extinction for Current Progress Period (2005-2009) and Baseline Period
(2000-2004) for the Best (Least Impaired) Days Measured at Arizona Class
I area IMPROVE Sites.

Table   SEQ Table \* ARABIC  14 :  Statistically Significant 2000-2009
Annual Average Trends for Aerosol Extinction by Component for Arizona
Class I area IMPROVE Sites.

Site	Group	Annual Trend* (Mm-1/year)



Site Total

(dv)	Ammonium Sulfate	Ammonium Nitrate	Particulate Organic Mass
Elemental Carbon	Soil	Coarse Mass	Sea Salt

BALD1

	20% Best	--	--	0.0	--	0.0	--	0.0	0.0

	20% Worst	--	-0.2	--	--	--	0.1	0.3	0.0

	All Days	--	-0.1	0.0	--	--	--	0.1	0.0

CHIR1

	20% Best	-0.1	0.0	0.0	-0.1	0.0	--	0.0	0.0

	20% Worst	-0.3	--	--	-0.7	-0.1	--	--	0.0

	All Days	-0.2	-0.1	0.0	-0.2	-0.1	--	-0.1	0.0

GRCA2

	20% Best	--	--	--	--	0.0	--	--	0.0

	20% Worst	--	--	-0.1	--	--	--	--	--

	All Days	--	--	0.0	--	--	--	--	--

IKBA1

	20% Best	-0.2	-0.1	-0.1	0.0	0.0	0.0	--	0.0

	20% Worst	--	--	--	--	0.0	--	--	0.0

	All Days	--	--	-0.1	--	0.0	--	--	0.0

PEFO1

	20% Best	-0.1	--	0.0	-0.1	--	--	--	0.0

	20% Worst	--	--	--	--	--	0.1	--	0.0

	All Days	-0.1	--	0.0	--	--	0.0	0.1	0.0

SAGU1

	20% Best	-0.2	--	-0.1	-0.1	--	--	--	--

	20% Worst	-0.3	-0.4	-0.5	-0.6	-0.3	--	--	0.1

	All Days	-0.2	-0.1	-0.1	-0.2	-0.1	--	--	0.0

SAWE1

	20% Best	-0.2	0.0	0.0	-0.1	-0.1	-0.1	--	0.0

	20% Worst	-0.3	-0.3	-0.6	-0.5	--	--	--	0.0

	All Days	-0.2	-0.1	-0.1	-0.3	-0.1	--	--	0.0

SIAN1

	20% Best	-0.2	-0.1	-0.1	-0.1	0.0	--	--	0.0

	20% Worst	-0.2	--	--	--	--	--	--	0.0

	All Days	-0.2	--	-0.1	-0.4	-0.1	--	--	0.0

SYCA1

	20% Best	-0.1	--	--	-0.1	--	--	--	0.0

	20% Worst	--	--	--	--	0.1	-0.3	--	--

	All Days	-0.1	--	0.0	--	--	-0.1	--	--

TONT1

	20% Best	-0.2	-0.1	-0.1	-0.1	-0.1	--	-0.1	0.0

	20% Worst	-0.2	--	-0.1	-0.8	-0.2	--	--	0.1

	All Days	-0.1	--	-0.1	-0.2	-0.1	--	--	0.0

*(--) Indicates statistically insignificant trend (<85% confidence
level). 

Ammonium Sulfate Analysis

Several analyses have been presented in this document to examine how
Ammonium sulfate extinction has changed within the 2000-2009 period at
IMPROVE monitoring sites within the State of Arizona.  When examining
the State as a whole, ammonium sulfate has shown increases at the
IMPROVE sites between the baseline period and progress period averages
for the 20% most impaired days (Section VI.B.2.), while showing no
significant trends or decreasing Theil statistic trends for the 20% most
impaired, least impaired, and all days (Section VI.B.4.ii.). 
Furthermore, ADEQ has shown that ammonium sulfate has decreased at all
IMPROVE monitors for the most recent five year progress period for all
site for the 20% most impaired days and for all sites except two for the
20% least impaired days.

20% Most Impaired Ammonium Sulfate Days

In this section ADEQ presents an alternate analysis performed by ARS in
which the 20% most impaired ammonium sulfate days were isolated,
averaged annually and then averaged for the baseline and progress
periods.  This analysis was performed in order to better understand how
the worst 20% visibility days for a particular pollutant change between
the baseline and progress periods rather than examining the 20% worst
visibility days for all combined pollutants.  The combination analysis
required by the RHR can cause seasonal shifts in the days chosen within
the baseline and progress periods which in turn can miss seasonal highs
for certain pollutant classes.  A Theil statistics trend analysis is
also performed for each monitor on the annually averaged 20% most
impaired days for the period of 2000-2009.  Figure 19 presents the
annual average of the 20% most impaired days.  Extinction decreases
between 2000 and 2004.  The years of 2005 and 2007 show exceptionally
high ammonium sulfate averages, followed by consecutive decreasing years
between 2007 and 2009.  When baseline and progress period extinction 

Figure   SEQ Figure \* ARABIC  19 :  Average Annual Ammonium Sulfate
Extinction (mM-1) at each Arizona Class I area IMPROVE Site for the 20%
Worst Ammonium Sulfate Days.

Table   SEQ Table \* ARABIC  15 :  2000-2009 Ammonium Sulfate Visibility
Extinction (mM-1) Trends, Baseline (2000-2004) vs. Progress (2005-2009)
Period Comparisons, and Altered Baseline (2000-2005) vs. Altered
Progress (2005-2009) Period Comparisons at each Arizona Class I area
IMPROVE Site for the 20% Worst Ammonium Sulfate Days.

SiteCode	Slope	p-value	Baseline (2000-2004)	Period 1 (2005-2009)
Difference	Altered Baseline (2000-2005)	Altered Period 1 (2006-2009)
Difference

BALD1	-0.18	0.08	7.52	7.84	0.32	8.15	7.13	-1.02

CHIR1	-0.15	0.14	10.33	10.51	0.18	10.55	10.22	-0.32

GRCA2	-0.05	0.24	6.39	7.12	0.73	6.87	6.70	-0.17

IKBA1	-0.09	0.36	8.16	8.73	0.57	8.48	8.47	-0.02

PEFO1	-0.15	0.03	8.16	8.31	0.15	8.49	7.86	-0.64

SAGU1	-0.29	0.13	9.54	9.58	0.05	10.26	8.87	-1.39

SAWE1	-0.33	0.09	10.05	10.00	-0.05	10.58	9.45	-1.13

SIAN1	-0.07	0.30	7.81	8.71	0.90	8.45	8.14	-0.31

SYCA1	-0.04	0.43	7.30	8.24	0.94	7.99	7.62	-0.37

TONT1	0.00	0.50	8.75	10.18	1.43	9.46	9.65	0.19



averages are compared, all IMPROVE sites show increasing ammonium
sulfate extinctions except SAWE1.  However, when 2000-2009 worst 20%
ammonium sulfate days trends are analyzed, no sites show increasing
trends.  BALD1, CHIR1, PEFO1, SAGU1, and SAWE1 all exhibit statistically
significant decreasing ammonium sulfate extinction trends.  The extreme
differences are strongly influenced by ammonium sulfate concentrations
measured in 2005.  Since this year is more like a midpoint, the data
have a more neutral effect using the Theil method.  To illustrate the
effect of the 2005 year, ADEQ presents the results of an analysis in
Table 15 where the RHR method is altered to include the year 2005 in the
baseline period rather than the progress period.  This altered RHR
method resulted in reduced ammonium sulfate extinction values between
the altered progress period (2006-2009) when compared to the altered
baseline period (2000-2005) for all sites except TONT1 for the 20% most
impaired ammonium sulfate days.  This illustrates the strong effect that
one outlier year can have in the RHR methodology.

Regional Ammonium Sulfate Trends

While the State of Arizona SIP only addresses the pollutant emissions
and progress goals for areas within Arizona’s State boundary, it is
important to analyze regional trends in pollutants in order to better
understand which phenomena are more representative of State issues and
which extend beyond State boundaries to surrounding areas.  This type of
analysis allows for a better understanding of which emission increases
are locally based in origin and which may be more representative of a
regional trend and thus may be due to some uncontrollable external
factor (e.g. NOx emissions originating from a point source located
within another State or Country, PM emission increases which are seen in
regional trends and thus may be related to environmental factors, etc.).
 In this section we analyze regional maps of IMPROVE monitor aerosol
extinction changes between baseline and progress periods in order to
determine if previously identified State of Arizona ammonium sulfate
trends may be regional phenomena.

Figure 20 shows only those aerosol extinction values which have
increased for the 20% most impaired days between the baseline
(2000-2004) and progress (2005-2009) periods for all IMPROVE monitors in
the western United States.  Figure 20 shows fairly ubiquitous and
substantial increases in ammonium sulfate across the State of Arizona,
State of New Mexico, western Texas, and south-central Colorado.  The
ammonium sulfate increases seen in this figure are obviously regional in
extent; however, it is difficult to determine an origination point. 
These increases could be due to a singular or combination 

Figure   SEQ Figure \* ARABIC  20 :  Magnitude of visibility component
extinctions that have increased between the baseline average (2000-2004)
and the first progress period average (2005-2009) for the 20% worst
visibility days.

Figure   SEQ Figure \* ARABIC  21 :  10-year annual average ammonium
sulfate extinction trends for 20% worst days at CIA IMPROVE sites in the
WRAP region.

Figure   SEQ Figure \* ARABIC  22 :  10-year annual average ammonium
sulfate extinction trends for all measured days at CIA IMPROVE sites in
the WRAP region.

of point or area sources within the above mentioned States and/or
Mexico.  

While average changes in ammonium sulfate visibility extinction
increased regionally between baseline and progress periods, Theil method
statistical analysis of ammonium sulfate extinction trends within the
south-western US for both the 20% most impaired days (Figure 21) and all
days (Figure 22) found that there was either: 1) no statistically
significant trends at IMPROVE monitors within the four corners region
(i.e. Arizona, Utah, Colorado, and New Mexico) or 2) the 10 year annual
average ammonium sulfate extinction trends at these IMPROVE monitors
exhibited statistically significant decreases.  Similar to what was
previously reported for the State of Arizona, regional Theil method
trends disagree with the RHR method of a 5 year average comparison of
the 20% most impaired days between baseline and progress periods. 
Furthermore, this agreement between State of Arizona and south-western
US regional trends may indicate that 2005 and 2007 were outlier years
for ammonium sulfate extinction within the entire 4 corners region and
the RHR method does not reflect more recent visibility extinction
improvements for this aerosol.

Coarse Mass Analysis

Coarse Particulate matter is generally recognized as having origination
sources which are locally based.  In this section, Coarse Mass (CM) is
analyzed more closely in order to gain more insight into trends for this
particular pollutant seen between the baseline and progress period for
all IMPROVE sites around the State.  This section presents the results
of an alternate approach in selecting the 20% most impaired day
analysis, similar to the trend performed for ammonium sulfate in Section
VI.C.  Furthermore, a qualitative analysis of IMPROVE site location in
relation to major PM10 emitting point sources is performed in order to
determine if an evident pattern exists between point source location and
IMPROVE monitor location.

Worst 20% Coarse Mass Days

Several analyses have been presented in this document to examine how
Coarse Mass extinction has changed within the 2000-2009 period at
IMPROVE monitoring sites within the State of Arizona.  When examining
the State as a whole, coarse mass has shown no discernable spatial
trends at the IMPROVE sites between the baseline period and progress
period averages for the 20% most impaired days (Section VI.B.2.). 
Furthermore, Theil statistic trends for the 20% most impaired days only
resulted in the BALD1 site exhibiting the only statistically significant
trend between 2000-2009, where an increase has been noted (Section
VI.B.4.ii.).  

In this section ADEQ presents an alternate analysis performed by ARS in
which the 20% most impaired coarse mass days were isolated, averaged
annually and then averaged for the baseline and progress periods.  This
analysis was performed in order to better understand how the worst 20%
visibility days for a particular pollutant change between the baseline
and progress periods rather than examining the 20% worst visibility days
for all combined pollutants.  The combination analysis required by the
RHR can cause seasonal shifts in the days chosen within the baseline and
progress periods which in turn can miss seasonal highs for certain
pollutant classes.  A Theil statistics trend analysis is also performed
for each monitor on the annually averaged 20% most impaired days for the
period of 2000-2010.  This trend analysis was extended past prior
analyses (2000-2009) to include 2010 since this year was shown to be
fairly unique in how coarse mass responded across Arizona.  As shown in
Figure 23, coarse mass is generally low in the years of 2008 and 2009,
but in the year 2010 coarse mass extinction dramatically increases at
some sites while continuing to decrease at others on the 20% most
impaired days.  When comparing the baseline period to the progress
period for the 20% worst coarse mass visibility days (Table 16) all
monitors except BALD1, PEFO1, and TONT1 recorded decreased extinction. 
Furthermore, the CHIR1, SAWI1, SAGU1, and SIAN1 monitors exhibited
drastic decreases in CM extinction for the 20% most impaired CM days. 
Theil statistics over the 11 year period showed decreasing trends at all
sites except 2; however, only CHIR1 and SIAN1 showed statistically
significant decreases, while BALD1 and PEFO exhibited statistically
significant increasing trend for CM on the 20% most impaired CM days.

Figure   SEQ Figure \* ARABIC  23 :  Average Annual Coarse Mass
Extinction (mM-1) at each Arizona Class I area IMPROVE Site for the 20%
Worst Coarse Mass Days.

Table   SEQ Table \* ARABIC  16 :  2000-2010 Coarse Matter Visibility
Extinction (mM-1) Trends and Baseline vs. Progress Period Comparisons at
each Arizona Class I area IMPROVE Site for the 20% Worst Coarse Matter
Days.

SiteCode	11-year trend

(2000-2010)	p-value	Baseline

(2000-2004)	Period 1

(2005-2009)	Period

Difference

BALD1	0.29	0.15	11.66	13.12	1.46

CHIR1	-2.73	0.02	33.92	25.47	-8.45

GRCA2	-0.38	0.24	11.38	9.04	-2.34

IKBA1	-1.10	0.19	23.27	21.24	-2.03

PEFO1	1.12	0.06	17.42	19.80	2.38

SAGU1	-1.26	0.38	25.88	18.83	-7.05

SAWE1	-0.56	0.31	40.47	20.49	-19.98

SIAN1	-1.56	0.01	22.97	12.76	-10.21

SYCA1	-1.04	0.24	26.36	24.77	-1.59

TONT1	-0.32	0.50	21.08	24.47	3.38



Large Point Source Locations

Previous analyses of coarse mass extinction between 2000 and 2009/2010
have shown fairly mixed results; however, there is evidence which
suggests coarse mass emissions originate from areas close to the
individual IMPROVE monitors.  Table 16 presents baseline vs. progress
period differences which generally show decreasing trends across the
state, but it is difficult to discern regional trends from monitors
within close proximity of one another.  IKBA1, SIAN1, and TONT1 are
three monitors which are centrally located within the state within
relatively close proximity; however, these three show a respective small
decrease, large decrease, and small increase in CM extinction for the
20% most impaired CM days.  Similarly, SAWE1 and SAGU1 are the two
monitors of closest proximity within the State, but these two monitors
show drastically different CM extinction baseline period averages. 
Therefore, ADEQ qualitatively examined the location of all NEI reported
major PM10 emitting sources within the State against the location of
individual IMPROVE monitors to determine if there is an evident trend. 
This analysis could prove to answer questions regarding whether locally
driven point source emissions are related to disparities in regional
coarse matter trends discussed above.

Figure   SEQ Figure \* ARABIC  24 :  Locations of Arizona Class I areas,
Class I area 50 km buffers, Class I area IMPROVE monitors, and Large
Point Source Emitters of PM10 (>100 tons/year).  IMPROVE site values
correspond to Visibility Extinction (mM-1) of Coarse Mass averaged over
the progress period (2005-2009).

Figure 24 presents the location and the 2008 annual emissions for each
point source in Arizona which emitted >100 tons.  The figure also
presents the progress period (2005-2009) average extinction (mM-1) for
CM for the 20% most impaired days.  This map indicates that the 20% most
impaired days may be impacted by local PM10 sources at some monitors,
while other sites show little to no effect on CM extinction from large
PM10 sources.  TONT1, SAWE1, and SAGU1 are all monitors which show
relatively high CM extinction values over the progress period and are
located relatively close to several large PM10 emitting point sources. 
However, SYCA1 is a monitor which recorded high extinction values for
the 20% most impaired days over the progress period and it is only
located near one large source and this is actually a relatively small
PM10 emitter in comparison with other large sources on the map.  Also,
PEFO1 and BALD1 are located near very large PM10 emitting sources, yet
have some of the lowest CM extinction values for the 20% most impaired
days recorded over the progress period.  Overall, it is difficult to
discern a visual relationship between large PM10 point sources and
IMPROVE monitor CM extinctions for the 20% most impaired days in the
State of Arizona.  A finer scale EI around each monitor site may be
needed in order to better understand individual site trends for CM
extinction.

Method Comparison Summary

The basic premise of the RHR is to ensure that visibility on the 20%
worst days continues to improve, and that visibility on the 20% best
days does not get worse, as measured in units of deciviews (dv)
calculated using data measured at IMPROVE monitoring sites. Progress is
measured in discreet 5-year average increments, beginning with the
2000-2004 baseline average, and proceeding with each subsequent 5-year
average (e.g. 2005-2009, 2010-2014, etc.). Some of the more subtle, but
important, considerations for RHR calculations using IMPROVE data
measurements are described below.

The discretization of visibility deciviews into 5 year averages can
result in anomalous years, experiencing extreme events, which can
identify longterm trends in visibility change which due not truly exist,
specifically for the 20% worst day comparisons.  As an example, this is
evident in data presented in this document for the State of Arizona
GRCA2 site.  This site experienced high 2005 and 2007 ammonium sulfate
extinction values (a regional trend) and 2009 fire induced elevations in
elemental carbon for the 20% most impaired days resulting in overall
elevated visibility progress period deciview averages for the site.  In
contrast, when the Theil statistical method was utilized over the 10
year period to analyze extinction trends for overall deciviews at the
site and extinction trends for individual visibility impairment
components, no significant increases were seen.  In this case, anomalous
years for individual visibility component extinctions, due to extreme
events, presented visibility degradation using the RHR method, while a
more standardized statistical trend method (the Theil method) showed no
significantly increasing visibility degradation trends. This is merely
one case where ADEQ, WRAP, and ARS have shown that outlier data can
significantly alter the data progress interpretations when using the RHR
method when compared to other standardized statistical trends which
better account for anomalous years.

Furthermore, to determine the 5-year average of the 20% best and worst
days, the highest and lowest 20% of days for each complete year are
first selected and averaged on an annual basis, with a 5-year average
calculated from these annual averages. The selection of the 20% best and
worst days may be significantly influenced by large episodic events, and
as such, may not represent the same time period from year to year.  This
selection of days may affect the averages for individual species in ways
that are independent from actual increases or decreases from one 5-year
period to the next.

Visibility impairment is the result of the cumulative effect of several
different particle pollutant types. Many of these pollutants have
consistent seasonal patterns. For example, ammonium nitrate is
temperature sensitive, with formation often favored during colder winter
months, while ammonium sulfate formation may be favored during warmer
summer months. Other pollutants, such as particulate organic mass, may
be impacted by large and variable episodic events such as wildland
fires. Variable occurrence of large episodic events may cause high
extinction measurements that will drive selection of 20% worst days to
coincide with the episodic events, effectively introducing the
possibility that the worst days occur at different times each year.

As an illustration of events driving the selection of the worst days,
consider daily average aerosol extinction calculated from IMPROVE data
at the CHIR1 site in Arizona. Figures 25 and 26 present daily aerosol
extinction measurements for 2002 and 2008 at CHIR1, with the 20% worst
days represented by an orange box with an “x” below the day. For
2002, large wildfire events in June and July contributed to high POM
measurements, resulting in more of the worst days selected during this
period.  In 2008, more of the worst days were selected in August and
October.

As an illustration of the seasonal patterns of individual compounds,
consider the monthly averages of aerosol extinction calculated from
IMPROVE data at the CHIR1 site. Figure 27 presents monthly average
aerosol pollution for CHIR1 measured during 2002, and Figure 28 presents
monthly averages in 2008.  For both years, plots show that ammonium
sulfate is highest between July and September. The monthly plots also
show the higher POM that coincided with wildfire events in 2002, which
affected the selection of more of the worst days between May and July in
2002, and more worst days in August and October in 2008. The seasonal
patterns of ammonium sulfate mean that even if annual ammonium sulfate
stayed the same, worst days in May and July will have higher ammonium
sulfate than worst days that occur between August and October.

For this case, Table 17 presents the annual averages of ammonium sulfate
for both the 20% worst days and all measured days. For these years, the
annual average of ammonium sulfate extinction decreases, while the 20%
worst day average actually increased.



Figure   SEQ Figure \* ARABIC  25 :  Daily Aerosol Extinction measured
by the Chiricahua CHIR1 IMPROVE monitor during 2002.



Figure   SEQ Figure \* ARABIC  26 :  Daily Aerosol Extinction measured
by the Chiricahua CHIR1 IMPROVE monitor during 2008.



Figure   SEQ Figure \* ARABIC  27 :  Monthly Average Aerosol Extinction
measured by the CHIR1 IMPROVE monitor in 2002.



Figure   SEQ Figure \* ARABIC  28 :  Monthly Average Aerosol Extinction
measured by the CHIR1 IMPROVE monitor in 2008.

Table   SEQ Table \* ARABIC  17 :  CHIR1 IMPROVE Site comparison of
Baseline and Progress Period Ammonium Sulfate Averages for All Days and
the 20% Worst Days

Year	All Days

Amm. Sulfate

Average (Mm-1)	20% Worst Days

Amm. Sulfate

Average (Mm-1)

2002	5.3	7.8

2008	4.9	9.0

Difference	-0.4 Mm-1	+2.2 Mm-1



Within this document ADEQ, WRAP, and ARS have presented several
different methods of analyzing IMPROVE data in order to best understand
the trends which are occurring at each of the IMPROVE sites between the
years of 2000-2009.  These methods are slight alterations in how the
analyzed data is chosen (i.e. choosing the 20% most/least impaired days
particular to a certain pollutant rather than for the entire suite of
pollutants) and how the variability inevitably seen with pollution data
can be analyzed to account for the effects of large individual events
which may skew overall pollution trends (i.e. the utilization of the
Theil statistical trend analysis).  ADEQ contests that these analyses
represent logical methods which are comparable to and in some ways
improve upon the standardized methodology required by the RHR. 
Furthermore, these methods are not drastically different from previous
EPA methodology, and in the case of Theil statistical comparison,
actually represent previously utilized EPA methodology for similar
comparisons.

Reasonable Progress Goals

EPA proposed to disapprove Arizona’s Reasonable Progress Goals (RPGs)
for 2018 based on the reasoning that they did not feel ADEQ demonstrated
that the goals constituted reasonable progress.  In this section ADEQ
presents Arizona’s progress towards reaching the previously presented
RPGs and Uniform Rates of Progress (URPs) as interpreted through IMPROVE
monitor data.  ADEQ chose to present IMPROVE data trends, as opposed to
surrogate measures such as Emission Inventory trends, as monitoring data
most accurately measure visibility changes within a region.  However,
ADEQ also provides analysis within this section relating trends seen at
the IMPROVE monitors to those noted within the emission inventories
where appropriate.  Finally, ADEQ compares State-wide extinction trends
for individual visibility impairment components to regional trends.

This section compares the rate of progress between the baseline and
progress periods towards the goal of natural visibility at each of the
Arizona IMPROVE monitors and how that rate compares to RPGs and URPs for
the 20% most impaired and least impaired days.  An alternate analysis of
RP is also included which illustrates the effect that one single year
has on the original results.  Furthermore, an additional analysis is
provided which show how specific fire events can have a large impact on
the baseline versus progress period comparison.

present the baseline visibility, progress period visibility, URP
visibility for 2018 (not included in Table 19), and the 2018 RPGs for
each of the IMPROVE monitor sites for the 20% most impaired days and the
20% least impaired days, respectively.  The Tables further present 2018
projected visibility based on the visibility rate of change between the
baseline period and progress periods.  The 2018 projected visibility was
calculated for each IMPROVE monitoring site using the following
equation:

 

where:

PV = 2018 projected visibility (dV)

BP = Average baseline period visibility (dV)

PP = Average progress period visibility (dV)

This equation assumes a linear rate of progress between the baseline and
progress period that can be extrapolated to 2018, that the average
baseline period visibility is the visibility for the midpoint year of
the baseline period (2002), and that the average progress period
visibility is the visibility for the midpoint year of the progress
period (2007).  The 2018 projected visibility values can be utilized in
two ways: 1) comparison with ADEQs previously calculated RPGs, or 2)
comparison with ADEQs previously calculated URPs.  This method is a
rather simplistic method but is believed to be more representative of
actual progress compared to utilization of a highly uncertain EI.

Table 18 presents the projected visibility for each IMPROVE site as
compared to the ADEQs RPGs and URPs for the 20% most impaired days.  Six
monitoring locations (shown in blue) are expected to surpass ADEQ’s
RPGs for 2018.  Furthermore, CHIR1, SAWE1, SAGU1, and SIAN1 are
projected to surpass the URPs calculated for these sites for 2018. 
While BALD1 and SYCA1 are expected to experience visibility improvements
by 2018, these improvements are not expected to meet the RPGs calculated
by ADEQ.  Two sites are projected to experience visibility degradation
by 2018 when compared to 2002; these sites are GRCA2 and IKBA1.

Visibility degradation at GRCA2 and IKBA1 are most accurately explained
through large, singular wildfire events which skew RHR method results
for the 20% most impaired days.  Within this document ADEQ has shown
evidence of the effect individual events at an IMPROVE monitoring
location can have in misrepresenting visibility trends when using the
RHR method.  Table 19 supplements previously overviewed data to show
this issue more clearly.  Table 19 presents an analysis where two
years’ (2003 and 2009) visibility extinction data are adjusted for EC
and POM to 10 year averages in order to reduce the effects of wildfires
located near an IMPROVE monitor.  In the year of 2003, total extinction
for the GRCA2 monitor was recalculated using the 10 year average
extinction values for EC (2.7 mM-1) and POM (10.7 mM-1).  This was
repeated for the year of 2009.  Both years, 2003 and 2009, experienced
large wildfire events near the GRCA2 monitor and this substitution
method was utilized in an attempt to reduce the effects of these
wildfire events on the overall trends of the RHR method.  The table
shows that without substitution, total visibility extinction increases
by 0.5 mM-1 using the RHR method, while EC and POM extinction
normalization for the year of 2003 caused this degradation to increase
to 2.9 mM-1, and 2009 extinction normalization caused the trend to
reverse with total visibility improvement on the order of 2.3 mM-1 at
GRCA2.  This exercise expresses the degree to which one large event can
skew visibility trends for the 20% most impaired days.

  In this comparison, all IMPROVE monitor sites are not just meeting
Arizona’s previously set RPG values, but are also exceeding URPs by
2018.  Again, this analysis shows the limitations of the RHR methodology
as one year near the mid-point has a strong influence on the overall
trends.

Table 21 presents the projected visibility for each IMPROVE site as
compared to the ADEQs RPGs for the 20% least impaired days.  None of the
sites are projected to experience visibility degradation on the 20%
least impaired days.  Furthermore, all sites except GRCA2 are projected
to surpass 2018 RPGs for the 20% least impaired days.  

Table   SEQ Table \* ARABIC  18 :  Arizona Class I Area Reasonable
Progress Goals Comparison to Progress Period Visibility for the 20%
Worst Days.  '2018 Projected Visibility' was extrapolated based on the
rate of Visibility change between the Baseline and Progress Period
Visibilities.

Reasonable Progress Goals for 20% Worst Days for Arizona Class I Areas

Arizona Class I Area	Site ID	Baseline (dV)	Progress (dV)	URP based 2018
visibility	2018 RPG (dV)	2018 Projected visibility

Chiricahua NM, Chiricahua W, Galiuro W	CHIR1	13.4	12.2	12.0	13.4	9.6

Grand Canyon NP	GRCA2	11.7	12.0	10.6	11.1	12.7

Mazatzal W, Pine Mountain W	IKBA1	13.4	13.4	11.8	12.8	13.4

Mount Baldy W	BALD1	11.9	11.8	10.5	11.5	11.6

Petrified NP	PEFO1	13.2	13.0	11.6	12.9	12.6

Saguaro NP - West Unit	SAWE1	16.2	14.9	13.9	16.0	12.0

Saguaro NP - East Unit	SAGU1	14.8	13.6	12.9	14.8	11.0

Sierra Ancha W	SIAN1	13.7	13.0	12.0	13.2	11.5

Superstition W	TONT1	14.2	13.8	12.4	13.9	12.9

Sycamore Canyon W	SYCA1	15.3	15.2	13.3	15.0	15.1



Table   SEQ Table \* ARABIC  19 :  Alternative method for the 20% Most
Impaired Days at GRCA2.  EC and POM visibility extinctions are replaced
by ten-year average for 2003 and 2009.

Alternative RHR Method Analysis for the 20% Worst Days  at GRCA2

 	Total Extinction (Mm-1)

Year Adjusted	Baseline	Adjusted Baseline	Progress	Adjusted Progress
Standard Change	Adjusted Change

2003	34.6	32.2	35.1	--	0.5	2.9

2009	34.6	--	35.1	32.3	0.5	-2.3



Table   SEQ Table \* ARABIC  20 :  Arizona Class I Area Reasonable
Progress Goals Adjusted Comparison to the Altered Progress Period
Visibility (2006-2010) for the 20% Worst Days.  '2018 Projected
Visibility' was extrapolated based on the rate of visibility change
between the Baseline and Progress Period Visibilities.  In this case the
Baseline period was altered to the years 2000-2005 and the Progress
Period was adjusted to the years 2006-2010.

Adjusted Reasonable Progress Goals for 20% Worst Days for Arizona Class
I Areas 

Arizona Class I Area	Site ID	2000-2005 Baseline (dV)	2006-2010 Progress
(dV)	URP based 2018 visibility	2018 RPG (dV)	2018 Projected visibility

Chiricahua NM, Chiricahua W, Galiuro W	CHIR1	13.3	11.8	12.0	13.4	8.5

Grand Canyon NP	GRCA2	11.8	11.4	10.6	11.1	10.5

Mazatzal W, Pine Mountain W	IKBA1	13.6	12.6	11.8	12.8	10.4

Mount Baldy W	BALD1	11.9	11.1	10.5	11.5	9.3

Petrified NP	PEFO1	13.3	12.5	11.6	12.9	10.7

Saguaro NP - West Unit*	SAWE1	16.0	14.8	13.9	16.0	12.2

Saguaro NP - East Unit	SAGU1	14.7	13.3	12.9	14.8	10.2

Sierra Ancha W	SIAN1	13.9	12.3	12.0	13.2	8.8

Superstition W	TONT1	14.2	13.3	12.4	13.9	11.3

Sycamore Canyon W	SYCA1	15.5	14.7	13.3	15.0	12.9

*2010 data was not included for this unit do to uncertainty of data's
accuracy.

Table   SEQ Table \* ARABIC  21 :  Arizona Class I Area Reasonable
Progress Goals Comparison to Progress Period Visibility for the 20% Best
Days.  '2018 Projected Visibility' was extrapolated based on the rate of
Visibility change between the Baseline and Progress Period Visibilities.

Reasonable Progress Goals for 20% Best Days for Arizona Class I Areas

Arizona Class I Area	Site ID	Baseline (dV)	Progress (dV)	2018 RPG (dV)
2018 Projected visibility

Chiricahua NM, Chiricahua W, Galiuro W	CHIR1	4.9	4.4	4.9	3.3

Grand Canyon NP	GRCA2	2.2	2.2	2.1	2.2

Mazatzal W, Pine Mountain W	IKBA1	5.4	5.1	5.2	4.4

Mount Baldy W	BALD1	3.0	2.9	2.9	2.7

Petrified NP	PEFO1	5.0	4.6	4.7	3.7

Saguaro NP - West Unit	SAWE1	8.6	8.0	8.3	6.7

Saguaro NP - East Unit	SAGU1	6.9	6.7	7.0	6.3

Sierra Ancha W	SIAN1	6.2	5.3	5.9	3.3

Superstition W	TONT1	6.5	5.7	6.2	3.9

Sycamore Canyon W	SYCA1	5.6	5.1	5.5	4.0



Conclusions

This document fills the required EPA Regional Haze SIP deficiency for
the State of Arizona regarding the submission of a complete and recent
emission inventory.  In this document ADEQ presents a 2008 Emission
Inventory which is comparable to the 2002 EI in some Source Categories
for a variety of pollutants.  Where this inventory is not reliably
comparable to the 2002 State of Arizona EI, ADEQ has provided an
overview of the methodology, input data, and model resolution
enhancements which have changed between the 2002 and 2008 inventory
preparations.

ADEQ has also included a review of IMPROVE monitor data between the
years of 2000 and 2009.  This review presented standardized 20% best and
worst visibility day comparisons between the baseline and progress
periods as well as Theil statistical trend analysis as an alternative
approach for understanding 10 year trends.  Visibility aerosol
extinction indicates that ammonium nitrate, organic mass, and elemental
carbon extinctions are improving within almost all Arizona Class 1
areas.  Fine Soil and Coarse Mass extinction values seem dependant of
the local environment surrounding the Class 1 Areas and show no
discernable increasing or decreasing spatial trends across the State. 
Anomalously high years (2005 and 2007) for ammonium sulfate extinction
revealed increasing ammonium sulfate visibility extinction between the
baseline and progress periods; however, decreasing trends in ammonium
sulfate in previous and more recent years resulted in Theil statistics
which either showed no statistically significant visibility extinction
increases or statistically significant visibility decreases across the
State.  Furthermore, similar trends for ammonium sulfate were noted for
the four corners region.

 Environmental Protection Agency (EPA).  Federal Register Volume 77, No.
246. Dec. 21, 2012.

 Environmental Protection Agency (EPA). Code of Federal Regulations
Title 40 Volume 2 Section 51.  2011.

 Environmental Protection Agency (EPA).  Federal Register Volume 77, No.
246. Dec. 21, 2012.

 “Arizona State Implementation Plan:  Regional Haze Under Section 308
of the Federal Regional Haze Rule”  Arizona Department of
Environmental Quality (ADEQ) Air Quality Division, 2011.

 Arizona Department of Administration (ADOA).
http://www.workforce.az.gov/pubs/demography/Estimates1980_2009With2000Ce
nsusWithNotes.xls

 EPA. 2012.  2008 National Emissions Inventory v. 2 Technical Support
Document.    HYPERLINK
"http://www.epa.gov/ttn/chief/net/2008neiv2/2008_neiv2_tsd_draft.pdf" 
http://www.epa.gov/ttn/chief/net/2008neiv2/2008_neiv2_tsd_draft.pdf 

 EPA. 2012.  2008 National Emissions Inventory v. 2 Technical Support
Document.    HYPERLINK
"http://www.epa.gov/ttn/chief/net/2008neiv2/2008_neiv2_tsd_draft.pdf" 
http://www.epa.gov/ttn/chief/net/2008neiv2/2008_neiv2_tsd_draft.pdf 

 Environmental Protection Agency (EPA).  Federal Register Volume 77, No.
246. Dec. 21, 2012.

 EPA. 2003.  Guidance for Tracking Progress Under the Regional Haze
Rule.  

 EPA. Trends in Monitored Concentrations of Carbon Monoxide.  National
Air Quality and Emissions Trends Report, 2003.  

 EPA.  Trends in Monitored Concentrations of Carbon Monoxide.  National
Air Quality and Emissions Trends Report, 2003.

 EPA.  Trends in Monitored Concentrations of Carbon Monoxide.  National
Air Quality and Emissions Trends Report, 2003.

 ibid  

 Environmental Protection Agency (EPA).  Federal Register Volume 77, No.
246. Dec. 21, 2012.

 PAGE   

  PAGE  \* roman  i 

  PAGE  \* Arabic  65 

Could we add a title page (Arizona Regional Haze Supplementary Technical
Support Document)?  Also, I might suggest that the TOC and list of
tables and figures sections don’t be given a roman numeral. Maybe
Introduction would start as Section I.

Should this be updated to say that EC and POM values for 2009 were
replaced with average values? Could also add a note here to see RP
section for details (table 19?).

Should we mention the alternate analysis done in the RP section also?

Update

restate

may want to include a few sentences explaining the alternative analysis
conducted.

