2015 Clean Air Status and Trends Network Five Year Network Assessment


2015

Clean Air Status and Trends Network
Five Year Network Assessment

-------
Table of Contents

1.	Introduction	2

A.	Purpose	2

B.	CASTNET Objectives	2

C.	Network Overview	3

D.	CASTNET Partners	5

E.	CASTNET Ozone Monitoring Program	6

F.	Network Modifications for Regulatory Ozone Monitoring	9

2.	Monitoring Results	10

A.	Ambient Ozone Concentrations	10

B.	W126	11

C.	Ozone Trends	12

D.	Shelter Temperature	14

3.	Quality assurance	15

A.	Overview	15

B.	Precision	16

C.	Bias	17

D.	Accuracy	18

1.	Semi-Annual Site Visits	18

2.	Independent PE Results	19

E.	Completeness	19

4.	Precursor Measurements and Meteorology	20

A.	NOy Monitoring	20

B.	CASTNET meteorology	21

5.	Summary	22

1 | P a g e

-------
Five Year Network Assessment

1. Introduction

A. Purpose

Monitoring agencies that submit data to the U.S. Environmental Protection Agency (EPA) for regulatory
purposes are required to conduct an assessment of their air quality surveillance system once every five
years (40 CFR Part 58.10). The EPA has adapted these requirements to conduct this assessment for the
Clean Air Status and Trends Network (CASTNET). The purpose of the assessment is to determine, at a
minimum, if the network meets the monitoring agencies' objectives. The focus of this assessment is on
the CASTNET ozone (O3) monitoring program from 2012 to 2014, reflecting the 2011 network
enhancements in accordance with federal regulatory requirements. The assessment includes a review of
the network's effectiveness in reporting trends and regional concentrations of O3 and recommendations
to improve network performance as CASTNET adapts to meet agency objectives.

B. CASTNET Objectives

CASTNET is a long-term monitoring network designed to measure regionally representative
concentrations of acidic pollutants, base cations, chloride (CI ), and ambient O3. The Environmental
Protection Agency - Clean Air Markets Division (EPA), the National Park Service - Air Resources Division
(NPS), and the Bureau of Land Management - Wyoming State Office (BLM-WSO) collaboratively manage
and operate CASTNET. In addition to EPA, NPS, and BLM-WSO, numerous other participants including
Tribes, other federal agencies, States, private land owners, and universities provide network support.
CASTNET monitors provide critically important, regionally representative data used to provide air quality
trends, estimate background O3 concentrations, and evaluate air quality models in the absence of local
emissions (Cooper et al., 2012; Fiore et al., 2002; Lin et al., 2012; Rieder et al., 2013; Zhang et al., 2011;
Zoogman et al., 2014). Additionally, CASTNET data are used to evaluate the effectiveness of national and
regional emission reduction control programs, gauge compliance with National Ambient Air Quality
Standards (NAAQS), and provide input into regional air quality and total deposition models.

CASTNET currently operates 93 monitoring stations throughout the contiguous United States, Alaska,
and Canada. EPA operates 63 CASTNET monitoring stations, NPS operates 25 CASTNET stations, and
BLM-WSO operates five CASTNET stations. More than 25 years of consistent, long-term measurements
reported by CASTNET demonstrate reductions in O3, nitrogen, and sulfur concentrations throughout the
United States. Additionally, continuous trace-level gas monitoring for sulfur dioxide (SO2), nitrogen
oxide/total reactive oxides of nitrogen (NO/NOy), and carbon monoxide (CO) is ongoing at four, eight,
and three CASTNET sites, respectively.

CASTNET's three operating agencies, EPA, NPS, and BLM-WSO coordinate their resources to fulfill the
following goals:

• monitor the status and trends in regional air quality and atmospheric deposition;

• provide information on the contribution of atmospheric pollution to ecosystem conditions; and

• provide measurements for validating and improving atmospheric models

2 | P a g e

-------
Each operating agency also utilizes CASTNET to fulfill their own monitoring objectives. Specific examples
are described below.

EPA utilizes CASTNET measurements to provide air pollutant concentration data to evaluate the
effectiveness of national and regional emission reduction programs and to determine compliance with
O3 NAAQS. EPA uses these data to provide consistent, long-term measurements for determining
relationships between changes in emissions and subsequent changes in air quality, atmospheric
deposition, and ecological effects. Under Title IV of the Clean Air Act Amendments (CAAA), the Acid Rain
Program (ARP) was promulgated to reduce emissions of sulfur dioxide (SO2) and nitrogen oxides (NOx)
from electric generating units (EGUs). A critical component of the CAAA required CASTNET to assess and
track real-world environmental results as the ARP was implemented and emissions were reduced.

The NPS uses CASTNET monitoring data to assess environmental conditions and trends in O3, sulfur and
nitrogen deposition. Coupled with special studies data, this information allows the NPS to understand
how air pollutants are currently impacting park air quality and air quality related values (AQRVs). These
data help the NPS and the public understand which parks are at highest risk for impacts, and where
conditions of park air quality and AQRVs are declining or improving. Specifically, ambient measurements
of O3, NOx and SO2 concentrations, deposition, and effects on visibility, soils, waters, and plants are
critical components of periodic assessments. For example, C^and vegetation data from Sequoia and
Yosemite NPs have been used to document the concentrations at which O3 pollution causes damage to
Ponderosa pine trees.

The BLM-WSO uses CASTNET data to identify air quality concerns and evaluate air strategy
effectiveness. These data also fulfill air monitoring commitments in Resource Management Plans
(RMPs) and Records of Decisions (RODs). Lastly, CASTNET data provide necessary information to assess
existing conditions, impacts of federal actions, and long-term trends in air quality and deposition on BLM
land.

While these CASTNET monitoring objectives go beyond the scope of this assessment, they are provided
here in brief to illustrate the utility and breadth of the data generated by CASTNET. In this assessment
we provide an overview of the CASTNET monitoring program, the sponsoring agencies' objectives for
the regulatory O3 monitoring program, O3 trends and annual results, quality assurance metrics, and the
future outlook for the program.

C. Network Overview

CASTNET was established under the 1990 CAAA, expanding the National Dry Deposition Network
(NDDN), which began in 1987. NPS began its participation with CASTNET in 1994 under an agreement
with EPA. With the involvement of NPS, the network became a national, rather than a primarily eastern,
network. BLM-WSO began participation in CASTNET in late 2012, with additional coverage provided in
Wyoming. CASTNET was designed to provide measurements for determining relationships between
changes in emissions and subsequent changes in air quality, atmospheric deposition, and ecological
effects. To meet those goals, CASTNET site locations were selected in rural areas to provide regionally
representative concentrations and estimates of dry deposition fluxes. CASTNET has historically used the
Multi-Layer Model (MLM) to estimate dry deposition fluxes using measured concentrations, on-site
meteorology and site characteristics, including land use and vegetation, as input. The CASTNET filterpack

3 | P a g e

-------
measurements provide weekly concentrations of gaseous sulfur dioxide {SO2) and nitric acid (HNO3), and
particulate sulfate (S042), nitrate (NO3), ammonium (NH4+), base cations (Ca2+, K+, Mg2+, Na+), and
chloride (CI ). A single laboratory, operated under contract to the EPA, analyzes the filterpack samples
for all CASTNET sites. Figure 1 shows the locations of all CAST NET monitoring sites. Circles represent
sites operating a filterpack and a continuous O3 monitor. Continuous O3 concentrations are measured at
79 sites. Diamonds represent sites with trace gas monitors operated by EPA or NPS. Sites at Bondville, IL
(BVL130) and Great Smoky Mountains National Park - Look Rock, TN (GRS420) are NCore sites with
trace gas NO/NOy, SO2, and CO. Additionally, there are five EPA sites that operate filterpack only sites,
represented by a square on the map in Figure 1 (EPA Small Footprint Sites). These are sites without a
temperature-controlled monitoring shelter. Thirty six CASTNET sites measure hourly meteorological
parameters including all NPS sites, all BLM-WSO sites, and six EPA sites. Additional information and data
from the CASTNET monitoring program can be found on the CASTNET website at
http://www.epa.gov/castnet.

Figure 1. Map of CASTNET sites (May 2015). EPA-sponsored sites are green, NPS-sponsored sites are
orange and BLM-sponsored sites are blue. CASTNET sites with filter pack and ozone monitoring are
circles. CASTNET small footprint, filterpack-only sites are squares. Sites that also report trace gas data
are represented by triangles (NCore) and diamonds.

YOS404

PIN4^
%	SEK430

RED004

YEL408 SHE604
BA#01

y

To

cr

»

14^6

IUF603

NEC602
:6#^WNC429

DIN431 #
I ROM4Q6ROM206

jvhj/ ^
MEV405

KIC003
Kikl84

ASH 135

H0W191
*A416

UVL124
ANA115

SAL133

W-1105
HWF187 P #WST109
1	H$iC001

CTH110CAT17j|

mn3

KEF1^PSU106 -WSP144

BVL13° OXF122DCP^
m ®

A

ALH157 VIN140

^ ^ MCK231 MCKl'&l
_A_	t

ARE1^

OAK 172 •

*LRL117 A BEL116
CDR119 SHf«-ykBWR139
) PA^07 ~

VPI12CI PED108
,,	CKT136# •

. MWC426
CDZ1* _SPD111PNF126

ESP1#GRSij|0	CND125 BFT142

* covjfcgpowoos ^

SND152

EPA
NPS
BLM

EPA Small Footprint
EPA Co-located Pair A NPS NCore

EVE419

EPA Trace-level Gas
NPS Trace-level Gas

NPS/EPA Co-located Pair with EPA Trace-level Gas
EPA NCore

4 | P a g e

-------
D. CASTNET Partners

In addition to EPA, NPS, and BLM, numerous other participants including Tribes, other federal agencies,
States, private land owners, and universities provide network support (Table 1). CASTNET partners may
provide local operational support, space for shelters and equipment, or scientific expertise. The EPA
contractor, AMEC Foster Wheeler (AMEC), manages the day-to-day operations for the EPA-sponsored
sites while the NPS and BLM contractor, Air Resource Specialists, Inc. (ARS), manages the operations for
the remaining sites.

Table 1. CASTNET Program Partners

Program Partners

Federal

State/Local/Tribal

University

Allegheny National Forest (NF)

Alabama-Coushatta Tribe of Texas

Auburn University Alabama Agricultural

Apalachicola NF

Cedar Creek State Park WV Division of

Experiment Station

Environment Canada

Natural Resources

Cornell University, Ecology & Evolutionary

Gunnison NF

Cherokee Nation

Biology

Holly Springs NF

Cumberland St. Forest VA

KS State University (KSU) Division of

Hubbard Brook Experimental Forest

Department of Forestry

Biology/Konza Prairie Long-term ecological

Medicine Bow-Routt NF

Deer Creek State Park OH Dept. of

research (LTER)

Monongahela NF

Natural Resources (DNR)

Miami University Institute for the

Nantahala NF

Edgar Evans State Park TN Dept. of

Environment & Sustainability

National Park Service (NPS)

Environmental Conservation (DEC)

Ouachita Baptist University School of

United States Army Engineering

Laurel Hill State Park PA Dept. of

Natural Sciences

District/Louisville

Conservation & Natural Resources

Pennsylvania State University (PSU) Fruit

US Department of Agriculture (USDA)

(DCNR)

Research & Extension Center

Natural Resources Conservation

ME Dept. of Environmental

PSU Department of Meteorology

Service (NRCS)

Protection (ME DEP)

Proctor Maple Research Center (UVM)

USDA Agricultural Research Center

Maurice K. Goddard State Park (PA

Purdue University Department of

(ARS)

DCNR)

Agronomy

USDA Southern Research Station

New York DEC

State University of NY (SUNY) ESF

Coweeta Hydrological Lab

ND Department of Health

Adirondack Ecology Center

US Fish & Wildlife Service

NYS Energy Research & Development

Texas A&M Agrilife Research & Extension

USDA- Forest Service (FS) Timber &

Authority

Center

Watershed Lab

Rocky Mountain Biological Laboratory

University of GA, College of Agriculture &

USDA-FS Forestry Sciences Laboratory

Santee Sioux Tribe of Nebraska

Environmental Science

USDA-FS Rocky Mountain Research

St. Johns River Water Management

University of IL, Illinois State Water Survey

Station

District

University of Maine Plant, Soil &

USDA-FS Toecane District

SD Dept. of Environmental & Natural

Environmental Science

US Department of Interior (DOI)-

Resources (SD DENR)

University of MD Department of

Bureau of Land Management

Vermont DEC

Atmospheric & Oceanic Science

White Mountain NF

Washington Crossing State Park (NJ

University of Michigan School of Natural



DEP)

Resources





University of NC Institute of Marine





Sciences





VA Tech Department of Plant Pathology,





Physiology & Weed Science

5 | P a g e

-------
E. CASTNET Ozone Monitoring Program

CASTNET operates 93 monitoring sites throughout the US and Canada and 79 of those sites measure
ground-level, continuous O3 following the regulations in the 40 Code of Federal Regulations (CFR) Part
58 and EPA's "Quality Assurance Handbook for Air Pollution Measurement Systems: Volume II: Ambient
Air Quality Monitoring Program" (US EPA, 2013a). CASTNET O3 monitors are located in 37 States with at
least one CASTNET O3 monitor in each of the ten EPA Regions. Three CASTNET O3 sites are located on
Tribal Lands including Santee Sioux, NE (SAN189), Cherokee Nation, OK (CHE185), and Alabama-
Coushatta, TX (ALC188).

Figure 2 shows the location of CASTNET monitoring sites reporting regulatory O3 data to EPA's Air
Quality System (AQS) database. These sites meet the siting criteria described in 40 CFR Part 58 Appendix
E. Most CASTNET ozone monitoring sites are used for regulatory purposes; however, the EPA-sponsored
Rocky Mountain National Park, CO (ROM206) and Mackville Collocated, KY (MCK231) sites are operated
for the purpose of network quality assurance (QA) and are designated as 'NAAQS Excluded' within AQS.
The Howland, ME (HOW191) site does not meet the siting criteria requirements in 40 CFR Part 58
Appendix E and is operated as a special study site (see Figure 1). Ozone data from HOW191 are not
reported to AQS. Additional information about CASTNET siting criteria can be found in the Quality
Assurance Project Plan version 8.2 at http://iava.epa.gov/castnet/documents.do (AMEC, 2014b).

6 | P a g e

-------
Figure 2. CASTNET Regulatory Ozone Monitors (May 2015)

LAV410

YOS404

PIN4U
0 SEK430

GRB411

VOY413

THR422

YEL408
# BAS601

^ NEC602

^PND165 WNC4#^

CNT169

DIN431 •
# ^ROM406

CAN407 «™161

MEV405

GRC474

JOT403

PET427

SAN189

STK138

HOX148
A UVL124

• •

ANA115

SAL133

HWF1^

ASH135

ACA416
WST109

CTH110 ABT147

• •

MIM113
• KEF1^PSU106 ^WSP144
** ARE1®

#BVL13
-------
Figure 3. (a) CASTNET monitoring site

¦ - 7.\>

Pine dale, WY(PND165)

8 | P a g e

-------
Figure 3. (b) Ozone instrumentation, PC, and data logger inside a CASTNETshelter

Palo Duro, TX (PAL190)

F. Network Modifications for Regulatory Ozone Monitoring

The National Park Service established their regulatory O3 monitoring program prior to 1990, The
transition to regulatory status for EPA-sponsored CASTNET sites was completed in 2011, when all O3
analyzers were replaced with Thermo Scientific™ Model 49i's, Thermo Scientific™ Model 49i's with
onboard O3 generators were installed as on-site transfer standards, and the QA/QC procedures were
implemented to comply with the requirements in 40 CFR Part 58. CASTNET uses the monitoring quality
objectives from the Quality Assurance Handbook for Air Pollution Measurement Systems, Volume i,
Appendix D (US EPA, 2013a) to ensure that the highest quality data are being submitted to EPA's Air
Quality System (AQS) database. The upgrades to the CASTNET O3 program have improved the overall
quality of data, reliability of the analyzers, and comparability of the data with other regulatory
monitoring networks (e.g., State and Local Air Monitoring Sites (SLAMS)). On-going improvements to site
equipment and infrastructure are posted to the individual CASTNET site information pages
(http://iava.epa.gov/castnet/epa jsp/sites.jsp).

9 | P a g e

-------
Prior to being used for regulatory monitoring, zero, span, and precision (ZSP) checks of the O3 analyzer
at EPA-sponsored sites were performed every week; now all sites perform ZSP checks daily. In addition
to the daily ZSP QC checks, technicians perform semi-annual system checks at each CASTNET site.

During these semi-annual visits, technicians audit the on-site analyzer, reverify the on-site transfer
standard, calibrate the on-site analyzer to the traveling transfer standard (Level 2) as needed, and verify
the responses of the data logger and shelter temperature probe with NIST-traceable standards. All on-
site O3 transfer standards at CASTNET sites are NIST-traceable at Level 3. Audit results are used to
perform the final validation on the hourly O3 data and validated data are submitted to the sponsoring
agency.

Prior to 2011, all CASTNET sites were visited once every other year by an independent auditor (Audit
Agency) to verify equipment was working properly and data were consistent across the network. As
required by 40 CFR Part 58 Appendix A, an annual Performance Evaluation (PE) is now conducted at
each CASTNET Ossite by an independent Audit Agency. For most CASTNET sites the Audit Agency is
Environmental Engineering & Measurement Services (EE&MS); however, some States act as an Audit
Agency and perform PEs at CASTNET sites.

The validated hourly O3 concentrations are submitted monthly to AQS by the sponsoring agency's
contractor. Additionally, the daily 1-point precision checks are submitted quarterly to AQS for each site.
PE results are submitted to AQS routinely by the designated Audit Agency. A subset of the CASTNET
partners act as the principal quality assurance organizations (PQAOs) - a unique role where States and
Tribes collect and own O3 data at CASTNET sites. This subset includes Acadia National Park, ME (ACA416)
submitted by Maine Department of Environmental Protection, Wind Cave National Park, SD (WNC429)
submitted by South Dakota Department of Environment and Natural Resources, Cherokee Nation, OK
(CHE185) submitted by Cherokee Nation Clean Air Program, and Theodore Roosevelt National
Monument, ND (THR422) submitted by North Dakota Department of Health.

2. Monitoring Results

A. Ambient Ozone Concentrations

CASTNET data provides an assessment tool for quantifying the improvements in air quality due to
regional and national emission reduction programs (e.g., the NOx Budget Trading Program, Clean Air
Interstate Rule, and Cross State Air Pollution Rule).

Ozone concentrations from CASTNET are used to gauge compliance with the primary NAAQS. Design
values are used to designate and classify nonattainment areas, as well as to assess progress towards
meeting the NAAQS. The design values are based on the 3-year average of the fourth highest daily
maximum 8-hour average. Figure 4 depicts the 2012-2014 fourth highest daily maximum 8-hour O3
average for all sites that met the completeness criteria (40 CFR Part 50, Appendix I). Ozone
concentrations are not included (shown as dots with no value) if the 3-year average was not available
because of incomplete data. In this map, exceptional events are not excluded because AQS data flags for
exceptional events that occurred in 2014 are not required to be submitted until July 1, 2015 (40 CFR Part
50.14). In 2012-2014, three sites exceeded the primary O3 standard of 75 ppb.

10 | P a g e

-------
Figure 4. Map of 2012-2014 fourth highest daily maximum 8-hour ozone average

«53

j V\

f / AfS

fFtff ft 'iia

i \ j j £ , \ j>*j

¦' :4;#'b3 hK
i <- i -71 ) .' • ,=?;

x —{ ¦ \

^ \n\«hri»£3

fe"69- .>•

Site riot pictured:
DEN417, AK 54

Concentration

| 80
75
70
65
60
55
52

B. W126

CASTNET also provides a unique dataset for evaluating the secondary NAAQS, which protect against
vegetation-related effects and other deleterious impacts to public welfare. The secondary O3 NAAQS is
currently set equal to the primary NAAQS. While the secondary NAAQS is currently set equal to the
primary NAAQS, the W126 index is often used to relate vegetation loses, such as reduced crop yield,
foliar injury, and decreased biomass accumulation, with O3 exposure. The W126 index is a cumulative
metric that sums weighted hourly O3 concentrations during the O3 season. The W126 is reported as the
maximum weighted monthly average during three consecutive months in the growing season when
daytime O3 concentrations are the highest and plant growth is most likely to be affected. CASTNET sites
are located in rural areas and often in sensitive ecosystems where vegetation related effects are
significant. Figure 5 shows the W126 values from CASTNET sites in 2014.

11 | P a g e

-------
Figure 5. Maximum W126 value for 2014

Site riot pictured:
DEN417, AK 1

C. Ozone Trends

For the purpose of reporting long-term regional trends, CASTNET sites are labeled as "western" or
"eastern" depending on whether they are west or east of 100 degrees west longitude (Figure 6). Eastern
long-term sites have been operating since at least 1990, while Western long-term sites have been
operating since at least 1996.

12 | P a g e

-------
Figure 6. CASTNET Western and Eastern Reference Sites

Western Reference Sites

Eastern Reference Sites

Figure 7 shows the hourly trends in ambient 03 concentrations from 1990-2014 (eastern sites) and 1996-
2014 (western sites). Hourly O3 data from the 34 Eastern reference sites show an overall reduction in
concentrations since 2002. The Eastern reference sites realized a 22% reduction between 2000-2002
and 2012-2014. In 2014, the median fourth highest daily maximum 8-hour average for the Eastern
reference sites was 62.5 ppb, the lowest level in the history of the network. The western reference sites
do not show the same dramatic reductions in O3 concentrations. There was a 5% reduction in O3
concentrations as measured by the Western reference sites between 2000-2002 and 2012-2014. In
2014, the median fourth highest daily maximum 8-hour average was 65 ppb at the 16 western reference
sites.

13 | P a g e

-------
Figure 7. Annual trends in hourly ozone concentrations from the eastern (right) and western (left)
CASTNET sites.

D. Shelter Temperature

Continuous gas analyzers are designed to operate within a specific temperature range. EPA guidance
recommends that shelter temperatures should be maintained at 20°C to 30°C with a standard deviation
of ± 2°C over a 24 hour period; however, a larger temperature range may be acceptable depending on
the model of the analyzer (US EPA, 2013a). Shelter temperature is measured continuously at all
CASTNET sites submitting O3 data to AQS and are polled hourly using the on-site data logger. Examples
of hourly shelter temperature measurements from July 2014 are shown in Figure 8. Daily review of the
O3 concentration data includes verifying that the shelter temperature is within the recommended range
and additional review of O3 data is required when it is not. During the ozone seasons of 2012-2014, the
network met the shelter temperature criteria approximately 88 percent of the time. Field notes from
the independent site auditor's Technical System Audit (TSA) reports indicated that most of the failures
observed were due to the site's shelter temperature sensors having slow response times as the shelter
heating and cooling systems cycled on and off (EE&MS, 2013).

14 | P a g e

-------
Figure 8. Hourly shelter temperature measured at Arendtsville, PA (ARE128) and Big Bend National Park,
TX (BBE401) in July 2014.

Site:ARE128 Site: BBE401

In an effort to provide more stable shelter temperatures and improve data capture efficiency, EPA plans
to upgrade the heating and cooling systems at approximately 40 sites in 2015. The upgrade will include
re-wiring and replacement of the site shelter thermostat with solid state relays controlled by the data
logger at the site, which will allow remote monitoring and control of the heating and cooling systems.

3. Quality assurance

A. Overview

The purpose of the CASTNET quality assurance (QA) program is to ensure that all reported data are of
known and documented quality in order to meet the CASTNET objectives and to be reproducible and
comparable with data from other monitoring networks. The CASTNET QA program is managed by an
independent QA Manager and Project QA Supervisor. The QA manager routinely performs internal
systems audits, reviews concentration and audit data, and prepares QA reports to management.

The CASTNET QAPP revision 8.2 (AMEC, 2014b) is comprehensive and covers all aspects of the
monitoring program. The QAPP is reviewed and updated by the contractor annually. Details on field,
data, and laboratory operations, training, SOPs, system audits, and reporting are examples of
information that can be found in the QAPP (http://iava.epa.gov/castnet/documents.do).

CASTNET data quality indicators include precision, accuracy, bias, completeness, representativeness,
and comparability. CASTNET data are evaluated against the data quality indicators (DQI) and the QA
results are reported quarterly and annually to assess overall measurement uncertainty (AMEC, 2013;
2014a; 2015). The CASTNET contractor reports the O3 precision, as calculated by results from the 1-
point QC checks, by site in the quarterly QA report. The accuracy and bias of each O3 monitor is reported
by the independent auditor and summarized in the Audit Program Annual Report (EE&MS, 2015). A
summary of the DQIs presented in the quarterly reports is included below.

15 | P a g e

-------
B. Precision

Ozone precision is reported as the 90 percent confidence limit (CL) of the coefficient of variation (CV) as
measured by the 1-point QC checks (40 CFR Part 58, Appendix A 4.1.2). The 1-pt QC check is the
difference between a known O3 concentration and the response of the O3 analyzer. For a site to meet
the acceptance criterion, the 90% CL of the CV must be < 7%. The analyzer is challenged with 60 ppb of
O3 during the daily 1-point QC check, which is considered representative of the ambient concentrations
measured within the network. Prior to 2014, the analyzers were challenged at 90 ppb of O3.

The overall network precision is shown in Figure 9 for 2012-2014. Results from the 1-point QC checks
are loaded into AQS quarterly. Additional data review is required for sites that do not meet the 7%
criterion. In 2012 and 2013 all sites met the 7% acceptance criterion. In 2014, all sites except Cadiz
(CDZ171, KY, 8.9%) met the acceptance criteria.

Figure 9. Box Plot showing all CASTNET precision estimates for 2012-2014. The mean (diamond) and
median (line) are shown for each year.

Precision may also be estimated as the relative percent difference (RPD) between the expected
concentration and the analyzer response. Figure 10 shows the annual RPD for all CASTNET sites using
the 1-point QC checks. The median RPD for 2012, 2013 and 2014 was -0.38, -0.010, and 0.21 ppb,
respectively.

16 | P a g e

-------
Figure 10. Relative percent difference between expected response and analyzer response for the 1-point
QC checks at all CASTNET sites. The mean (diamond) and median (line) are shown for each year.

C. Bias

The bias estimate is also calculated using results from the 1-point QC checks. A site is required to meet a
95 percent CL of the absolute bias estimate (40 CFR Part 58 App A sec 4.1.3). A site meets the
acceptance criterion if the absolute bias is < 7%.

A positive or negative direction is assigned to the bias estimate when the signs of both the 25th and 75th
percentiles of the percent differences for each site are in the same direction. No direction is assigned if
the percentiles are of different signs. Signed bias results, by site, for 2014 are shown in Table 2. Sites are
shaded green if the bias estimate was positive, orange if the bias estimate was negative, and not shaded
if the bias estimate had no sign. Each site met the 7% acceptance criterion for 2014.

17 | P a g e

-------
Table 2. The bias estimate calculated from one-point QC checks for 2014.

Site

State

Bias (%)

Site

State

Bias (%)

Site

State

Bias (%)

DEN417

1.8

MCK131

1.2

CHE185

3.1

SND152

MCK231

1.4

ARE128

1.2

CAD150

1.7

BEL116

4.5

KEF112

1.8

CHA467

2.8

BWR139

1.9

LRL117

0.8

GRC474

3.1

ACA416

0.7

MKG113

1.3

PET427

0.9

ASH135

0.8

PSU106

3.5

JOT403

1.9

ANA115

1.7

WNC429

0.4

LAV410

1.4

HOX148

1.2

ESP127

1.4

PIN414

1.9

UVL124

1.5

GRS420

0.9

SEK430

1.2

VOY413

1.2

SPD111

1.9

YOS404

0.7

CVL151

2.6

ALC188

2.9

GTH161

1.8

GLR468

3.3

BBE401

1.5

MEV405

2.3

BFT142

1.8

PAL190

ROM206

2.3

CND125

2.3

CAN407

2.7

ROM406

2.1

COW 137

1.3

DIN431

0.9

ABT147

1.8

PNF126

2.8

PED108

2.6

IRL141

1.3

THR422

3.1

SHN418

2.6

SUM156

SAN189

1.2

VPI120

1.2

GAS153

0.9

WST109

1.8

PRK134

1.8

ALH157

1.9

WSP144

1.4

CDR119

1.3

BVL130

3.6

GRB411

2.4

PAR107

2.6

STK138

0.6

CTH110

1.7

BAS601

2.6

SAL133

HWF187

4.9

CNT169

1.5

VIN140

1.2

DCP114

1.9

NEC602

CDZ171

3.3

OXF122

1.6

PND165

2.7

CKT136

3.1

QAK172

2.2

YEL408

1.4

MAC426

1.8

D. Accuracy

1. Semi-Annual Site Visits

Approximately every six months, technicians managed by the Field Operations Manager perform semi-
annual performance checks to the on-site analyzer and reverify the on-site transfer standard, calibrate
the on-site analyzer to the traveling transfer standard (Level 2) as needed, and verify the data logger and
the shelter temperature probe using NIST traceable standards. These results are used to perform final
validation on the hourly O3 data.

18 | P a g e

-------
2. Independent PE Results

The Audit Agency performs annual PEs in accordance with 40 CFR Part 58 Appendix A Section 3.2.2 and
EPA's Quality Assurance Handbook for Air Pollution Measurement Systems: Volume II and submits these
results to AQS on a quarterly basis (US EPA, 2013a). In 2010, EPA issued a memorandum expanding the
allowable audit levels required in the CFR Appendix A from 5 to 10 (US EPA, 2010). The auditor is
required to select audit levels that bracket 80 percent of the ambient data; however, the audit levels do
not need to be consecutive. In 2011, EPA issued a second memorandum revising the acceptance criteria
for the expanded audit levels (US EPA, 2011). For levels 1 and 2 (which includes the range of 4 to 19
ppb), the acceptance criteria is ±1.5 ppb difference or ± 15 percent difference, whichever is greater. The
acceptance criteria for levels 3-10 remains ±15 percent difference.

Figure 11. Estimated Bias in O3 concentrations from PE Audit Results for All CASTNET Sites

Estimated Bias from Annual PE for All CASTNET Sites

CO 0 -

-5 -

-10

201 2 201 3 2014

The bias is estimated from the PE values for the years 2012, 2013, and 2014 and displayed in Figure 11.
The median and mean for all sites have been within ±1% for all years. Ninety percent of all sites have
realized a bias within ±6% for the 3 years shown.

E. Completeness

Completeness is a measure of the amount of valid data obtained from a measurement system compared
to the amount that was expected to be obtained under correct, normal conditions. For comparison with
the NAAQS for a given 3-year period, a site must meet two completeness criteria: having at least 75
percent valid data for each of the three ozone seasons comprising the period, and having at least 90
percent valid data for all three ozone seasons combined. For the 2012-2014 time period, 64 out of 81

19 | P a g e

-------
(78 percent) CASTNET O3 sites met both of these completeness criteria. Three sites did not meet the
completeness criteria because they were decommissioned during the 3-year time period. These sites
were Howland, ME (HOW132), Konza Prairie, KS (KNZ184), and Mount Rainier, WA (MOR409).
Additionally, three sites did not meet completeness criteria because they began collecting ozone
measurements after the beginning of the 2012 ozone season, including Basin, WY (BAS601), New Castle,
WY (NEC602), and Dinosaur National Monument, UT (DIN431).

4. Precursor Measurements and Meteorology

A. NOv Monitoring

Reactive nitrogen compounds are precursors for both O3 and PM2.5 formation. Total reactive oxidized
nitrogen (NOy) is defined as NOx (NO + NO2) plus NOz(PAN, HNO3, HNO2, PPN, other organic nitrates, and
NO2"). EPA and NPS operate eight trace-level continuous NOy analyzers at CASTNET sites (Figure 1). Great
Smokies National Park, TN (GRS420) operated by NPS and Bondville, IL (BVL130) operated by EPA are
also NCore stations. GRS420, TN and BVL130, IL also measure trace SO2 and CO as part of the NCore
suite of measurements. The Beltsville, MD (BEL116) NOy analyzer has been converted to an "enhanced"
NOy analyzer which includes a heated stainless steel converter (TNX), Light Emitting Diode (LED)
converter (NOx) and molybdenum converter (NOx). The sample stream switches between each converter
(or no converter for NO) to measure or calculate speciated reactive nitrogen, including NOy, NO2, NOx,
TNX, NHX, NO, and NOz. BEL116 also measures continuous trace SO2 concentrations.

Total reactive oxidized nitrogen (NOy) is measured using a thermal molybdenum converter at the inlet to
convert reactive nitrogen species to NO followed by the detection of NO by chemiluminescence. The
EPA-sponsored CASTNET sites with trace NOy each have a Teledyne (API) T200U chemiluminescence
analyzer, 701H zero air system, and a T700U multi-gas calibrator in addition to the typical suite of
CASTNET equipment (e.g., data logger, ozone analyzer, etc.).

Trace NOy is audited twice per year by the CASTNET contractor and audited once every other year by an
independent 3rd party. The ambient data are submitted to AQS monthly and the QC results are
submitted quarterly. Trace-level precision is verified against the acceptance criteria in 40 CFR Part 58
Appendix A. The acceptance criterion is an upper 90 CL for the CV of 10%. The NOy analyzer is
challenged with 15 ppb NO every other day. The precision estimates for 2013 and 2014 for the trace gas
NO/NOy analyzers are shown in Table 3. Efforts to better understand the trace gas methods and quality
control procedures are on-going between CAMD, EPA's Office of Air Quality Planning and Standards
(OAQPS), the EPA Regions, the manufacturer, and contractors.

Table 3. Precision estimates (CV %)from EPA NOv sites calculated from Equation 3 in 40 CFR Part 58
Appendix A Section 4.1.3.

2013 Precision (%)

2014 Precision (%)

BEL116, MD

25.01

5.62

BVL130, IL

3.87

6.23

HWF187, NY

2.49

13.64

PND165, WY

4.27

3.17

PNF126, NC

N/A

4.26

20 | P a g e

-------
Daily trends in total reactive nitrogen vary between urban sites or sites influenced by mobile sources
and high elevation or remote sites. As shown in Figure 12, trace NOy data from BEL116 (a suburban site
NE of Washington DC) peaks in the morning and is associated with NOx emissions (e.g. vehicular traffic),
with a slow decline in total reactive nitrogen in the afternoon as O3 is produced through photochemical
reactions. There is a loss of O3 in the evening, which is more pronounced at BEL116 than ROM206. At
CASTNET high elevation sites, such as ROM206 (elevation 2,742 m), the peak NOy concentrations are
generally lower than BEL116, MD including a less pronounced morning peak. The diurnal O3
concentrations are usually less variable at ROM206, which is typical at high elevation sites. For example,
at high elevation sites, the absence of a shallow boundary layer can facilitate stratospheric ozone
intrusions at night (prevalent during the winter) along with reduced scavenging of O3 by NO2 (Ambrose
et. al., 2011, Brodin et. al., 2010). CASTNET NOy data are useful for model evaluation and validation, as
high elevation sites provide a unique data set to the modeling community.

Figure 12. Diurnal pattern ofC>3, NO and NOyfrom Beltsville, MD (left) and Rocky Mountain National
Park, CO (right) for 2014. Note, scales are different.

Beltsville, MD (2014)

12 -ii

0 2 4 6

10 12 14 16 18 20 22

¦Ozone

•NOY

•NO

Rocky Mountain National Park, CO (2014)

Ozone NOY NO

B. CASTNET meteorology

All NPS-sponsored and all BLM-WSO-sponsored CASTNET sites include meteorological measurements.
Six EPA-sponsored CASTNET sites: Beltsville, MD (BEL116); Bondville, IL(BVL130); Cherokee Nation
Stilwell, OK (CHE185); Indian River Lagoon, FL (IRL141); Palo Duro, TX (PAL190); and Pinedale, WY
(PND165) also collect meteorological data. The locations of the 32 CASTNET sites reporting
meteorological measurements with regulatory O3 are displayed in Figure 13. Historically, CASTNET on-
site meteorology has been used for calculating deposition velocities using the MLM (Meyers et al.,
1998). At sites without meteorological measurements, missing deposition velocity (Vd) values resulting
from missing meteorological data are replaced based on the results in Bowker et al. (2011), which
substitutes hour-specific historical averages of Vd for missing Vd values at specific sites. The substitution
procedure was shown to result in long-term, unbiased estimates of the annual mean Vd. In addition to
calculating deposition velocity values, on-site meteorology is particularly important for identifying
environmental conditions with high potential for ozone formation, pollutant transport (back
trajectories), and model validation (Zhang et al., 2011; Zoogman et al., 2014).

21 | P a g e

-------
On-site meteorology at a CASTNET site includes measurements of temperature (9m at EPA-sponsored
sites, 2m at most NPS-sponsored sites, and at select locations 9m and 2m), relative humidity, solar
radiation, precipitation, wind speed, wind direction, sigma theta (standard deviation of the wind
direction), and wetness reported as hourly averages.

Figure 13. Sites with on-site meteorology and regulatory 05 measured by CASTNET (May 2015)

CASTNET Ozone Sites with Meteorology

YEL408

BAS601

^ NEC602

PND165 WNC4jfi

DIN431

BEL116

Shdfcl8

CAN407

MEV405

Agency sponsor

# BLM

# EPA
¦ NPS

5. Summary

The CASTNET ozone monitoring program provides critical information to stakeholders and has met its
primary monitoring objectives through consistent, long-term measurements since 1989. The rural
ozone monitors detect regional air quality signals, provide a unique data set for evaluating the effects of
O3 on vegetation and ecosystems, and are used to evaluate the primary and secondary O3 NAAQS.
Federal land managers use CASTNET data to assess environmental conditions and risk of air quality
impacts on nationally-recognized sensitive areas and other federal lands. Other stakeholders and

22 | P a g e

-------
participants include Tribes, States, other federal agencies, and universities who use CASTNET data to
evaluate air quality models and determine human health and environmental risks in their areas.

With over 25 years of data from many of its sites, CASTNET has measured a significant reduction in
regional O3 concentrations in the Eastern US in response to emission control programs, allowing policy
makers to assess the effectiveness of these programs for improving air quality and reducing negative
impacts of air pollution. Moving forward, CASTNET data in the coming years will allow stakeholders to
evaluate the effectiveness of current policies and ongoing emission reduction programs such as the
Cross-State Air Pollution Rule.

The enhanced quality assurance program implemented in 2011 has improved the quality of CASTNET O3
monitoring data. Data quality indicators indicate that most CASTNET sites are meeting the network
quality assurance criteria for accuracy, bias, and precision for 2012-2014, the most recent 3-year period
available. While 78 percent of CASTNET sites met the completeness criteria for inclusion in NAAQS
nonattainment decisions for 2012-2014, many of the sites that did not meet the criteria were due to the
difficulty in operating technical equipment in remote and unattended locations. Efforts to improve the
data capture efficiency throughout the network are ongoing.

CASTNET remains committed to improving our understanding of reactive nitrogen and other O3 and
PM2.5 precursors in the ambient environment. Eight monitoring sites already provide continuous NOy
data, and several of these sites also measure continuous SO2 and CO. In addition, a speciated NOy
monitor is being developed and tested for broader deployment within the network. Expanded use of
these and other continuous monitors will enhance the utility of CASTNET data in model evaluation and
development.

CASTNET has been a stable platform for regional air monitoring for over 25 years and the program
continues to evolve within the constraints of budgets, regulatory demands, and agency priorities.
Developing long-term solutions to improve the cost-effectiveness of routine measurements and
leveraging existing and new partnerships has been crucial for the continuity of CASTNET, and these
attributes will continue to be important over the next five years as CASTNET strives to improve data
capture, enhance the types of measurements collected, and expand into areas with limited air quality
monitoring.

23 | P a g e

-------
References

Ambrose, J.L.; Reidmiller, D.R.; Jaffe, D.A. (2011). "Causes of high O3 in the lower free troposphere over
the Pacific Northwest as observed at the Mt. Bachelor Observatory." Atmospheric Environment. 45:
5302-5315.

AMEC Foster Wheeler, Inc. (2013). "Summary of Quarterly Operations (October through December) with
2012 Annual Summary."

http://epa.gov/castnet/iavaweb/docs/QA Quarterly 2012 Q4-Annual.pdf
Accessed May 2015.

AMEC Foster Wheeler, Inc. (2014a). "Summary of Quarterly Operations (October through December)
with 2013 Annual Summary."

http://epa.gov/castnet/iavaweb/docs/QA Quarterly 2013 Q4.pdf
Accessed May 2015.

AMEC Foster Wheeler Inc. (2014b) "CASTNET Quality Assurance Project Plan Revision 8.2"
http://epa.gov/castnet/iavaweb/docs/qapp v8-2 Main bodv.pdf.

Accessed May 2015.

AMEC Foster Wheeler, Inc. (2015). "Summary of Quarterly Operations (October through December) with
2014 Annual Summary."

http://epa.gov/castnet/iavaweb/docs/QA Quarterly 2014 Q4.pdf
Accessed May 2015.

Bowker, G.E., Schwede, D.B., Lear, G.G., Warren-Hicks, W.J., and Finkelstein, P.L. (2011). "Quality
Assurance Decisions with Air Models: A Case Study of Imputation of Missing Input Data Using EPA's
Multi-Layer Model." Water. Air. & Soil Pollution DOI 10.1007/sll270-011-0808-7

Brodin, M.; Helmig, D.; Oltmans, S. (2010). "Seasonal ozone behavior along an elevation gradient in the
Colorado Front Range Mountains." Atmospheric Environment. 44(39): 5305-5315.

Cooper, O. R., R.-S. Gao, D. Tarasick, T. Leblanc and C. Sweeney (2012). "Long-term ozone trends at rural
ozone monitoring sites across the United States, 1990-2010." Journal of Geophysical Research:
Atmospheres (1984-2012) 117(D22).

EE&MS, Inc. (2015). "CASTNET 2013 Annual Report."
http://epa.gov/castnet/iavaweb/docs/Audit Program Annual 2013.pdf
Accessed May 2015.

Fiore, A. M., D. J. Jacob, I. Bey, R. M. Yantosca, B. D. Field, A. C. Fusco and J. G. Wilkinson (2002).
"Background ozone over the United States in summer: Origin, trend, and contribution to pollution
episodes." Journal of Geophysical Research: Atmospheres 107(D15): ACH 11-1-ACH 11-25.

Lin, M., A. M. Fiore, L. W. Horowitz, O. R. Cooper, V. Naik, J. Holloway, B. J. Johnson, A. M. Middlebrook,
S. J. Oltmans and I. B. Pollack (2012). "Transport of Asian ozone pollution into surface air over the
western United States in spring." Journal of Geophysical Research: Atmospheres (1984-2012) 117(D4).

24 | P a g e

-------
Meyers, T.P., P. Finklestein, J. Clarke, T.G. Ellestad and P.F. Sims (1998). "A multilayer model for inferring
dry deposition using standard meteorological measurements." Journal of Geophysical Research.
103(D17): 22,645-22,661.

Rieder, H. E., A. M. Fiore, L. M. Polvani, J.-F. Lamarque and Y. Fang (2013). "Changes in the frequency
and return level of high ozone pollution events over the eastern United States following emission
controls." Environmental Research Letters 8(1): 014012.

U.S. Environmental Protection Agency (2010). "Use of Expanded List of Audit Levels for Annual
Performance Evaluation for SO2, NO2, O3, and CO as Described in 40 CFR Part 58 Appendix A Section
3.2.2."

http://www.epa.gov/ttn/amtic/files/ambient/pm25/datamang/TechMemoforPEAuditLevels.pdf
Accessed June 2015.

U.S. Environmental Protection Agency (2011). "Guidance on Statistics for Use at Audit Levels 1 and 2 of
the Expanded List of Audit Levels for Annual Performance Evaluation for SO2, NO2, O3, and CO as
Described in 40 CFR Part 58 Appendix A Section 3.2.2."

http://www.epa.gov/ttnamtil/files/ambient/pm25/datamang/20110217lowlevelstatmemo.pdf
Accessed May 2015.

U.S. Environmental Protection Agency (2013a). "Quality Assurance Handbook for Air Pollution
Measurement Systems, Volume II, Ambient Air Quality Monitoring Program." EPA-454/B-13-003.
http://www.epa.gov/ttn/amtic/files/ambient/pm25/qa/QA-Handbook-Vol-ll.pdf
Accessed May 2015.

U.S. Environmental Protection Agency (2013b). "Integrated Science Assessment of Ozone and Related
Photochemical Oxidants (Final Report)." EPA 600/R-10/076F.
http://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=247492
Accessed May 2015.

Zhang, L., D. J. Jacob, N. V. Downey, D. A. Wood, D. Blewitt, C. C. Carouge, A. van Donkelaar, D. Jones, L.
T. Murray and Y. Wang (2011). "Improved estimate of the policy-relevant background ozone in the
United States using the GEOS-Chem global model with 1/2 x 2/3 horizontal resolution over North
America." Atmospheric Environment 45(37): 6769-6776.

Zoogman, P., D. J. Jacob, K. Chance, X. Liu, M. Lin, A. Fiore and K. Travis (2014). "Monitoring high-ozone
events in the US Intermountain West using TEMPO geostationary satellite observations." Atmos. Chem.
Phvs. 14(12): 6261-6271.

25 | P a g e

-------