c FPA
d str^s EPA/600/R-22/213 | September 2022 | www.
Environmental
Protection Agency
The Enhanced Air Sensor
Guidebook
Office of Research and Development
Center for Environmental Measurement and Modeling
-------
EPA/600/R-22/213
September 2022
The Enhanced Air Sensor Guidebook
By
Andrea Clements and Rachelle Duvall
Center for Environmental Measurement and Modeling
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Danny Greene
Eastern Research Group, Inc.
Morrisville, NC 27560
Tim Dye
TD Environmental Services, LLC
Petaluma, CA 94952
-------
Disclaimer
This document presents work performed by the United States Environmental Protection
Agency (U.S. EPA) Office of Research and Development (ORD) and the Office of Air
Quality Planning and Standards (OAPQS) with technical support provided by Eastern
Research Group through two task orders: Task Order 68HERH19F0257 and Task Order
68HERH21F0093 both under EPA Contract No. 68HERD19A0001. Any mention of trade
names, manufacturers or products does not imply an endorsement by the U.S. Government
or the U.S. EPA. The U.S. EPA and its employees do not endorse any commercial
products, services, or enterprises. This document has been reviewed in accordance with
U.S. EPA policy and approved for publication.
-------
Table of Contents
Disclaimer,. ii
List of Tables vi
List of Figures vii
Acronyms arid Abbreviations x
Acknowledgements xii
Executive Summary xiii
Chapter 1 Introduction to Air Sensors and the Guidebook 1
1.1 Background on Air Sensors 2
1.2 Purpose of this Enhanced Guidebook 4
1.3 Differences Between the 2014 Guidebook and the Enhanced Version 4
1.4 Intended Audience 6
Chapter 2 Air Quality 101 ..7
2.1 Overview of Outdoor Air Quality and Air Pollution 8
2.1.1 The Pollutant Lifecycle 15
2.1.2 Differences in Pollutant Concentrations Over Time ...16
2.2 Pollutant Effects on Health and the Environment 19
2.3 Outdoor Air Pollution Monitoring 25
2.4 Air Quality Standards and Indices 30
2.5 The Air Quality Index (AQI) 35
Chapter 3 Monitoring Using Air Sensors 39
3.1 Planning and Conducting Air Monitoring 40
3.2 Question: Determining a Purpose for Monitoring. .......42
3.3 Plan: Developing a Plan 45
3.4 Plan: Selecting an Air Sensor 49
3.4.1 Target Pollutant and Sensor Performance 51
3.4.2 General Features of a Sensor 54
3.5 Setup: Locating Sites for Air Sensors 58
3.5.1 Installing Air Sensors..... 59
3.5.2 Specifics for Designing a Network of Air Sensors .......61
3.6 Setup: Collocation and Correction 66
3.6.1 Air Sensor Collocation 68
iii
-------
3.6.2 Correction of Sensor Data 73
3.7 Collect: Data Collection, Quality Assurance/Quality Control, and Data Management
78
3.7.1 Data Collection Activities 79
3.7.2 Checks to Ensure Quality Assurance and Quality Control ....80
3.7.3 Data Management System 84
3.8 Evaluate: Analyzing, Interpreting, Communicating, and Acting on Results. ....89
3.8.1 Analyze and Interpret Data..... ...90
3.8.2 Communicating Results 92
3.8.3 Take Action 93
Chapter 4 Sensor Performance Guidance 98
4.1 Overview of Sensor Performance 99
4.2 Sensor Performance Evaluations..... ...99
4.3 Approaches Used to Evaluate Sensor Performance 102
4.3.1 U.S. EPA Recommendations on Evaluating Sensor Performance 104
4.3.2 Guidance from other Organizations on Evaluating Sensor Performance ...105
4.4 How to Select Sensors Based on Evaluation Reports or Information 106
Appendix A: Resources A-1
A. 1 Introduction to Air Sensors A-1
A.2 Air Quality 101 A-1
A.2.1 Outdoor Air Quality and Air Pollution A-1
A.2.2 Health And Environmental Effects of Air Pollution A-3
A.2.3 Air Pollution Monitoring A-5
A.2.4 Air Quality Standards and indices A-6
A.2.5 The U.S. Air Quality Index (AQI) A-7
A.3 Monitoring Using Air Sensors A-9
A.3.1 Question: Determining a Purpose For Monitoring....... .A-9
A.3.2 Plan: Developing a Plan A-10
A.3.3 Plan: Selecting an Air Sensor....... ...A-11
A.3.4 Setup: Locating Sites for Air Sensors A-12
A.3.5 Setup: Collocation and Correction A-14
A.3.6 Collect: Data Collection, Quality Assurance/Quality Control, and Data
Management A-15
A.3.7 Evaluate: Analyzing, Interpreting, Communicating, and Acting on Results A-16
iv
-------
A. 4 Sensor Performance Guidance A-19
A.4.1 Sensor Performance Evaluations A-19
A.4.2 Approaches Used to Evaluate Sensor Performance A-20
A.4.3 Summarizing Sensor Performance Evaluation Results using U.S. EPA's
Targets Reports A-21
Appendix B: Questions to Consider When Planning for and Collecting Air Sensor Data, and
Sharing Your Results B-1
B.1 Planning ..B-1
B.2 Working with Governmental Officials B-1
B.3 Setting up Monitoring Locations.. B-2
B.4 Collecting Data B-2
B.5 Conducting Quality Control B-3
B.6 Evaluating Data B-3
B.7 Other B-3
Appendix C: Checklists C-1
C. 1 What to Look for in an Air Sensor? C-1
C.2 What to Look for in a User Manual? C-3
C.3 How to Maintain Your Air Sensor? C-5
Appendix D: Data Handling and Air Quality Index (AQI) Calculations D-1
D.1 Data Processing D-1
D. 1.1 Data Quality Assurance (QA) D-2
D. 1.2 Data Aggregation D-3
D.2 AQI Calculations D-5
D.2.1 Background D-5
D.2.2 Computing the AQI D-7
Appendix E: Interpreting Sensor Performance Evaluation Results E-1
E.1 Deployment Details ..E-5
E.2 Time Series Plots E-8
E.3 Scatter Plots E-11
E.4 Performance Evaluation Metrics and Target Values E-13
E.5 Meteorological Conditions During the Evaluation E-15
Appendix F: Glossary F-1
v
-------
List of Tables
Table 1-1. Overview of Non-Regulatory Supplemental and informational Monitoring
Applications (NSIM) for Air Sensors 3
Table 2-1. Common Air Pollutants, Their Sources, and Concentration Ranges to Expect in
Outdoor Air 12
Table 2-2. Health and Environmental Effects of Select Common Air Pollutants 19
Table 2-3. Comparison Between Reference Monitors and Air Sensors 26
Table 2-4. U.S. EPA National Ambient Air Quality Standards (NAAQS; current as of
9/30/2022) 32
Table 2-5. The Air Quality Index (AQI) Levels of Health Concern, Numerical Values, and
Meanings 35
Table 2-6. Pollutant-Specific Sensitive Groups for the AQI Greater than 100 (Additional
information available on the AirNow website) 36
Table 3-1. Common Topics and Information Included in an Air Monitoring Plan .........46
Table 3-2. Common Quality Control (QC) Checks ....81
Table 4-1. Common Approaches for Evaluating Air Sensor Performance ..102
Table D-1. Example Breakpoints for PM2.5.... D-9
vi
-------
List of Figures
Figure 1 -1. Typical Air Sensor Components 2
Figure 1-2. Examples of How Users Can Deploy Air Sensors 4
Figure 2-1. Poor Versus Good Air Quality 8
Figure 2-2. Typical Movement of Warm Air in the Atmosphere Versus Warm Air Trapped by
a Temperature Inversion 9
Figure 2-3. Atmospheric Conditions and Their Impacts on Pollutant Concentrations .........10
Figure 2-4. Sources of Primary and Secondary Pollutants (Adapted from:
https://www. mrgscience.com/ess-topic-63-photochem ical-smog. htm I) 11
Figure 2-5. Particulate Matter Size Ranges 12
Figure 2-6. The Pollutant Lifecycle from Source to Impact on People and the Environment
16
Figure 2-7. Typical Concentrations for O3 and PM2.5 During Different Time Periods 17
Figure 2-8. Common Types of Air Monitoring Instruments and Their Characteristics 26
Figure 2-9. Different Air Monitoring Locations for Outdoor Air (Near Source, Mobile,
Ambient, and Background) and Indoor Air (Occupational and Residential) ....27
Figure 2-10. U.S. EPAs AirData Air Quality Monitors website - Active PM2.5 Continuous
Monitoring Stations (as of 9/30/2022) ....28
Figure 3-1. Five Steps Recommended for Planning Air Monitoring Projects Using Air
Sensors 40
Figure 3-2. Example of Adding Details to Your Question or Objective 43
Figure 3-3. Questions to Consider Before Purchasing an Air Sensor 50
Figure 3-4. Illustration of Air Sensor Bias, Accuracy, and Precision 52
Figure 3-5. Example of Noisy Measurement Data 53
Figure 3-6. Example of an Air Sensor's Response Time 54
Figure 3-7. Logistical Considerations and Tips for Installing an Air Sensor 59
vii
-------
Figure 3-8. Example Maps for Placing Air Sensors for Networks of Different Scales
Depending on the Purpose: a) Regional/Urban Network, b) Neighborhood Network, and c)
Microscale, Small Area Network .........63
Figure 3-9. Different Types of Air Sensor Collocation Strategies .........69
Figure 3-10. Example of the Ordinary Least-Squares Regression 74
Figure 3-11, Example of a Sensor That Shows No Agreement with the Reference
Instrument 75
Figure 3-12. Scatter Plot Showing that an Air Sensor has a Linear Response at Lower
Concentrations and a Non-linear Response at Higher Concentrations 76
Figure 3-13. Examples of Air Sensor Data Corrections 77
Figure 3-14. Definitions of Quality Assurance and Quality Control 80
Figure 3-15. Major Components and Functions of a Data Management System (DMS) 84
Figure 3-16. Common Visualization Methods for Air Quality Data 91
Figure 3-17. Factors that Can Contribute to how Individuals or Communities use Air Sensor
Data for Personal Action (Source: Understanding social and behavioral drivers and impacts
of air quality sensor use)...... ....94
Figure 4-1. Air Sensors on Tripods (in foreground) with Reference Instruments (in the
background) to Evaluate Sensor Performance. Photo Credit: South Coast Air Quality
Monitoring District (AQMD) 99
Figure 4-2. Common Concerns Related to Sensor Performance 100
Figure 4-3. Flow Chart for Considering an Air Sensor Based on Performance 109
Figure D-1. Time Series of Ozone (O3) Concentrations Showing a "Spike" in Concentration
that is an Outlier in the Data. ..D-3
Figure D-2. Time Series Showing Raw PM2.5 Data with Block and Rolling Averages ....... D-4
Figure D-3. Comparison of the Traditional AQI and Color-Accessible AQI Color Scale
Presented in Color, Grey-Scale, and on a Map of the South Coast Air Basin (South Coast
AQMD Press Release - May 2022) D-6
Figure D-4. Flow Chart Showing How to Compute the AQI D-7
Figure E-1. Page 1 of U.S. EPA's Base Testing Reporting Template for PM2.5 Sensors -
Deployment Details and Visual Plots of Sensor Performance E-2
viii
-------
Figure E-2. Page 2 of U.S. EPA's Base Testing Reporting Template for PM2.5 Sensors -
Tables and Graphs Summarizing Sensor Performance E-3
Figure E-3. Page 3 of U.S. EPA's Base Testing Reporting Template for PM2.5 Sensors -
Table Documenting Supplemental Materials and Information E-4
Figure E-4. Testing Organization and Site Information Details of the Reporting Template E-5
Figure E-5. Sensor Information Details of the Reporting Template......... E-6
Figure E-6. FRM/FEM Information Details of the Reporting Template.... E-7
Figure E-7. Time Series Plots in the Reporting Template E-8
Figure E-8. Scatter Plot in the Reporting Template E-12
Figure E-9. Performance Metrics in the Reporting Template..... E-14
Figure E-10. Tabular Summary of Sensor Performance Metrics on Page 2 of the Reporting
Template E-15
Figure E-11. Meteorological Conditions in the Reporting Template E-16
ix
-------
Acronyms and Abbreviations
AAPCA
Association of Air Pollution Control Agencies
AMTIC
Ambient Monitoring Technology Information Center
API
Application Programming interface
AQI
Air Quality Index
AQMD
Air Quality Management District
AQ-SPEC
Air Quality Sensor Performance Evaluation Center
AQS
Air Quality System
ATSDR
Agency for Toxic Substances and Disease Registry
b
intercept
BC
black carbon
BTEX
benzene, toluene, ethylbenzene, and xylene
CAA
Clean Air Act
CARB
California Air Resources Board
CEMM
Center for Environmental Measurement and Modeling
CEN
European Committee for Standardization
CFR
Code of Federal Regulations
ch4
methane
cm3
cubic centimeter
CO
carbon monoxide
C02
carbon dioxide
CSN
Chemical Speciation Network
CV
coefficient of variation
°C
degrees Celsius
DIY
Do-it-Yourself
DMS
Data Management System
EC
electrochemical
EU
European Union
FAQs
Frequently Asked Questions
FEM
Federal Equivalent Method
FRM
Federal Reference Method
GPS
global positioning system
GMI
Global Methane Initiative
h2s
hydrogen sulfide
HAPs
hazardous air pollutants
HEI
Health Effects Institute
Hg
mercury
ID
identification
IMPROVE
Interagency Monitoring of Protected Visual Environments
IQ
intelligence quotient
ISA
Integrated Science Assessment
IRIS
Integrated Risk Information System
IT
information technology
JRC
Joint Research Centre
LoRa
Low-power Wide Area Network
m
slope
MAE
mean absolute error
MBE
mean bias error
MEE
China Ministry of Ecology and Environment
MeHg
methylmercury
MOS
metal oxide sensors
X
-------
N
number of data points
NAAQS
National Ambient Air Quality Standards
NACAA
National Association of Clean Air Agencies
NATTS
National Air Toxics Trends Stations
NCore
National Core Multipollutant Monitoring Network
nh3
ammonia
NIOSH
National Institute for Occupational Safety and Health
N02
nitrogen dioxide
NOx
oxides of nitrogen
NSIM
non-regulatory supplemental and informational monitoring
NSRDB
National Solar Radiation Database
NTAA
National Tribal Air Association
OAQPS
Office of Air Quality Planning and Standards
ORD
Office of Research and Development
03
ozone
PAHs
polycyclic aromatic hydrocarbons
Pb
lead
PID
photoionization detector
PM
particulate matter
PM1.0
particles with diameters generally less than 1.0 micrometers
PM2.5
particles with diameters generally less than 2.5 micrometers; also called fine
particulate matter or fine PM
PM10
particles with diameters generally less than 10 micrometers
PPb
parts per billion
ppm
parts per million
QA
quality assurance
QAPP
Quality Assurance Project Plan
QC
quality control
R2
coefficient of determination
RETIGO
REal Time GeOspatial Data Viewer
RH
relative humidity
RMSE
root mean square error
SD
standard deviation
SIP
State Implementation Plan
SLAMS
State and Local Air Monitoring Station
S02
sulfur dioxide
sox
sulfur oxides or oxides of sulfur
SOP
standard operating procedure
T
temperature
tvoc
total volatile organic compounds
UFP
ultrafine particles; particles with diameters generally less than 0.1 micrometers
UL
Underwriters Laboratories
UV
ultraviolet
|jg/m3
micrograms per cubic meter
(jm
micrometers
U.S.
United States
U.S. EPA
United States Environmental Protection Agency
USGS
United States Geological Survey
VOCs
volatile organic compounds
WHO
World Health Organization
xi
-------
Acknowledgements
The authors acknowledge the Eastern Research Group technical staff associated with Task
Order 68HERH19F0257 and Task Order 68HERH21F0093 (both under EPA Contract No.
68HERD19A0001) for their research efforts and graphics development (Mindy Mitchell
lead) included in this document. This effort was jointly led by the United States
Environmental Protection Agency (U.S. EPA) Office of Research and Development (ORD),
Center for Environmental Measurement and Modeling (CEMM) and the Office of Air Quality
Planning and Standards (OAQPS). OAQPS staff including Kristen Benedict, Ron Evans,
Amanda Kaufman, Colin Barrette, and Corey Mocka are acknowledged for contributions
supporting the development of this document. Libby Nessley (U.S. EPA/ORD/CEMM) and
Trisha Curran (U.S. EPA/OAQPS) are recognized for quality assurance support in
developing this document.
We acknowledge the following U.S. EPA internal reviewers: Amanda Kaufman, Rachael
Leta-Graham, Karoline Barkjohn, Samuel Frederick (former National Student Service
Contractor assigned to U.S. EPA), Amara Holder, Rich Baldauf, Ethan McMahon (formerly
with U.S. EPA), Robert Judge (retired), Ryan Brown, Marta Fuoco, Sheila Batka, Dena
Vallano, Idalia Perez, Ken Davidson, Dave Nash, Susan Stone, Rachel Mclntosh-
Kastrinsky, Karen Wesson, Deirdre Murphy, Brian Keaveny, and Laureen Burton. We also
acknowledge the following external reviewers: Dr. Edmund Seto and Ms. Orly Stampfer
(University of Washington, School of Public Health); and Dr. Vasileios Papapostolou and
team members (South Coast Air Quality Management District, Air Quality Sensor
Performance Evaluation Center).
Lastly, we would like to acknowledge the authors of the original 2014 Air Sensor Guidebook
including: Ron Williams (retired from U.S. EPA); Vasu Kilaru and Emily Snyder (U.S.
EPA/ORD); Amanda Kaufman (U.S. EPA/OAQPS); Timothy Dye (TD Environmental
Services); Andrew Rutter (formerly with Sonoma Technology; deceased) and Ashley
Russell (formerly with Sonoma Technology); and Hilary Hafner (Sonoma Technology).
xii
-------
Executive Summary
In 2014, the United States Environmental Protection Agency (U.S. EPA) published the
original Air Sensor Guidebook to help those interested in using sensors to collect air quality
measurements and interpret sensor data. The Air Sensor Guidebook has been one of the
most popular resources on the U.S. EPA's Air Sensor Toolbox website. The guidebook was
intended to provide basic foundational knowledge on topics including:
• Background information on common air pollutants and air quality
• Selecting appropriate sensors for different applications
• Data quality considerations, and
• Sensor performance for different applications
The initial target audience for the Air Sensor Guidebook was limited to participatory
scientists and sensor manufacturers/developers. Since 2014, the sensor user community
has grown to include individuals, communities, schools, researchers, environmental
agencies (e.g., air quality, environmental quality, natural resources, health), industry,
medical professionals, emergency responders, technology developers, data integrators,
and more.
Recognizing the ever-increasing availability of sensors, expanding scientific knowledge,
and availability of best practices to support sensor use, the U.S. EPA significantly updated
the 2014 Guidebook. The refreshed version, The Enhanced Air Sensor Guidebook,
includes updated content and new topics that incorporate best practices, current
knowledge, and recommendations on sensors.
The goal of the enhanced Guidebook is to support users in planning and collecting air
quality measurements using air sensors. This Guidebook can help sensor users:
• Learn the basics of air quality, air pollution monitoring, and air sensors
• Plan and conduct an air quality monitoring study
• Select, setup, and use air sensors
• Analyze, interpret, communicate, and act on results
• Understand the basics of air sensor performance
The enhanced Guidebook also contains expanded resources and content in the
Appendices including: 1) General resources (e.g., air quality information, air sensor
performance, data analysis tools), 2) a list of 'Questions to Consider when Planning,
Collecting, and Sharing Your Data' to prepare sensor users for questions to expect from
others, 3) Checklists for quickly determining how to select sensors, items to look for in a
user manual, and maintaining air sensors, 4) information on data handling and the Air
Quality Index, 5) education on Interpreting Sensor Performance Evaluation Results, and 6)
a glossary of terms used in the guidebook.
While this guidebook may not exhaustively answer every question air sensor users may
have, it is provided as a resource to help support the user community in effectively using
this class of technology to support their air quality monitoring needs.
xiii
-------
Chapter 1
Introduction to Air Sensors and the
Guidebook
This new and expanded Air Sensor Guidebook empowers communities, environmental
agency officials, researchers, students, educators, and others to plan for, select, and
operate air sensors to meet specific needs.
This chapter provides:
• Background information about air sensors,
• Our purpose in providing this Enhanced Air Sensor Guidebook,
• Key differences between the 2014 Guidebook and this enhanced version, and
• The intended audience of the guidebook.
1
-------
1.1 Background on Air Sensors
Air sensors are a class of non-regulatory technology that are lower in cost, portable, and
generally easier to operate than monitors used for regulatory monitoring purposes. Air
sensors and regulatory monitors differ in that regulatory monitors are the gold standard and
are designed to meet strict performance requirements for use in regulatory monitoring.
These differences are discussed in more detail in Chapter 2.
Air sensors typically provide relatively quick or instant air pollutant concentration
measurements and allow for measurement of air quality in more locations. The term air
sensor often describes an integrated set of hardware and software that uses one or more
sensing elements (sometimes called sensors or other terms) to detect or measure pollutant
concentrations. Figure 1-1 shows the typical air sensor components, which vary from one
manufacturer to another. Most air sensors include a power source, components to detect
air pollutants and weather parameters, electronics to transmit data (e.g., cellular), and a
microprocessor to control the devices. Some air sensors may include a battery, display
screen, and a global positioning system (GPS) component to determine location. Sensors
that transmit data may connect to cloud servers that store, process, and provide access to
data. Additionally, data may be displayed on maps or graphical plots.
Display screen
Communications
(cellular, WiFi, etc.)
Power transformer
(solar, land power)
Enclosure
Microprocessor
Battery/backup power
Weather sensors
Air quality
detection devices
Figure 1-1. Typical Air Sensor Components
2
-------
Advancements in microprocessors and miniaturization have led to a rapid expansion in the
availability of air sensors to measure a variety of air pollutants. As air sensors have become
more accessible worldwide, there has been a dramatic increase in their use for measuring
air quality conditions and there is greater access to publicly available sensor data sets.
The United States Environmental Protection Agency (U.S. EPA) has identified that a
primary use of air sensors is for non-regulatory supplemental and informational monitoring
(NSIM) applications. Table 1-1 summarizes examples of these applications. Other potential
applications for air sensors include mobile monitoring, personal exposure monitoring,
indoor air monitoring, among others. Figure 1-2 shows examples of how air sensors may
be used for NSIM and other applications.
Table 1-1. Overview of Non-Regulatory Supplemental and Informational Monitoring
Applications (NSIM) for Air Sensors
Category
Description
Common Examples
Spatiotemporal
Variability
Characterizing a pollutant concentration over
a geographic area and/or time.
Is pollution higher in the morning at a
location?
• Daily trends
• Gradient studies
• Air quality forecasting
• Participatory science*
• Education
Comparison
Analyzing differences and/or similarities in
air pollution characteristics against a
threshold value or between different
networks, locations, regions, time periods,
etc.
Does a location show high pollution levels,
but other locations do not?
• Hotspot detection
• Data fusion
• Emergency response
• Supplemental monitoring
Long-term
Trend
Characterizing changes in pollutant
concentrations over a long time.
How did pollution concentrations change at a
location over a 5-year period?
• Long-term changes
• Epidemiological studies
• Model verification
"Participatory science is also referred to as citizen science, community science, volunteer
monitoring, or public participation in scientific research (Source:
https://www.epa.gov/participatory-science).
3
-------
Fixed
Outdoors
i
a \
// \
Mobile
Indoors
Indoors/
Outdoors
JL
AM
a
Figure 1-2. Examples of How Users Can Deploy Air Sensors
1.2 Purpose of this Enhanced Guidebook
This Guidebook provides information on air quality monitoring using air sensors. When
sited, installed, configured, and operated properly, maintained carefully, and with thoughtful
project planning, air sensors can provide useful information for a range of air quality
applications. The purpose of this Guidebook is to provide information to those interested in
using air sensors, including the basics of air quality, planning and conducting air
monitoring, selecting air sensors, considerations for sensor performance, and more.
The Guidebook identifies the best practices for using air sensors and provides
recommendations for planning and implementing a study to save time, effort, and money
and ultimately help users collect useful data. This Guidebook also offers resources to help
you with the many tasks needed to produce, correct, use, and interpret air sensor data,
including sensor selection, operating procedures, illustrated figures, checklists, and links to
external resources.
1.3 Differences Between the 2014 Guidebook and the Enhanced
Version
This Enhanced Air Sensor Guidebook reflects information gathered from more recent
studies, best practices, and scientific literature for rapidly evolving air sensor technologies
that have become more available since U.S. EPA released the 2014 Guidebook. This
Guidebook includes input from the air quality community, air quality agencies and experts,
and U.S. EPA scientists and technical experts.
4
-------
The U.S. EPA updated Chapter 2 with the latest air quality and health science and
expanded air monitoring network and instrument descriptions.
We added new topics or significantly updated existing Guidebook information to incorporate
best practices, current knowledge, and recommendations as follows:
• Monitoring Using Air Sensors (Chapter 3)
o Planning and Conducting an Air Quality Monitoring Study identifies all
tasks involved in monitoring air quality successfully,
o Determining the Purpose for Monitoring helps develop an objective for
monitoring and matching the air sensor technology to that objective.
• Sensor Performance Guidance (Chapter 4) provides information about why
sensor performance evaluations are necessary and how they are conducted, where
results can be found, and how those results can be used to make informed
purchasing decisions.
Changes to Appendices include:
• Expanded Appendix A, which lists Resources including air quality information,
sources of real-time and historical data, air sensor performance, data analysis tools,
health effect resources, resources for educators, and economic impacts of air
pollution.
• Expanded Appendix B, which contains a list of Questions to Consider when
Planning, Collecting, and Sharing Your Data to anticipate questions to expect
from others. Answering these questions helps you plan, ensure credibility in your
data and results, and allows others to use your data.
• Added the new Appendix C containing checklists for quickly determining 1) What to
Look for in an Air Sensor, 2) What to Look for in a User Manual, and 3) How to
Maintain Air Sensors.
• Added the new Appendix D to provide information on Data Handling and Air
Quality Index (AQI) Calculations.
• Added the new Appendix E which provides education on Interpreting Sensor
Performance Evaluation Results.
• Added the new Appendix F which is a Glossary of definitions for commonly used
terms in this document.
Lastly, a List of Abbreviations is provided at the beginning of the document for
easy reference.
5
-------
1.4 Intended Audience
The target audience for the 2014 Air Sensor Guidebook was primarily participatory
scientists (i.e., citizen science, community science, volunteer monitoring, and public
participation in scientific research) and sensor manufacturers/developers. Since 2014, air
sensors are more available, and the number and types of applications continue to expand
rapidly. The intended audience for this enhanced Guidebook Includes participatory
scientists, environmental agency officials, researchers, health professionals, emergency
responders, technology developers, educators, and the public.
Resources for More Information
• U.S. EPA's Air Sensor Toolbox
o Information and resources for topics related to air sensors; Includes links to
other organizations and resources that sensor users may find helpful
o https://www.epa.gov/air-sensor-toolbox
6
-------
Chapter 2
Air Quality 101
Air quality is a complex subject as it involves emission of air pollutants, chemical and
physical transformation of pollutants, and atmospheric conditions which can move or trap
pollutants and affect the speed of chemical reactions. Air quality is measured in a variety of
ways. The impacts of air quality on health and the environment vary based on
concentration and pollutant type.
This chapter provides:
• Basic knowledge of outdoor air quality,
• Summary of the health and environmental impacts of select air pollutants,
• Overview of the different types of air quality monitoring approaches,
• A review of air quality regulations and indices, and
• Specific information about the Air Quality Index (AQI).
7
-------
2.1 Overview of Outdoor Air Quality and Air Pollution
Air quality is a term used to
describe how much pollution is
present in the air (Figure 2-1). We
care about air quality because air
pollutants can affect our health
and our environment. As indicated
by the World Health Organization
(WHO), air pollution is a leading
cause of death. Several scientific
studies (e.g., epidemiologic,
exposure) link air pollution to a
range of health problems including
decreased lung function,
aggravation of respiratory and
cardiovascular diseases,
increased asthma incidence and
severity and premature mortality, Figure 2-1. Poor Versus Good Air Quality
among many other effects. In
addition to causing adverse health effects, air pollutants can also cause adverse
environmental effects such as reduced visibility and damage to plant and animal life. Acidic
pollutants deposited on the ground, predominantly from rain, harm both land and water
ecology and even structures. Furthermore, some pollutants also affect the Earth's energy
balance, impacting global climate conditions. See Section 2.2 for further discussion of
health and environmental effects.
Air pollution consists of a complex mixture of different chemicals in the form of solid
particles (in a range of sizes), liquid droplets, and gases. Air pollution is produced as a
result of human-made (i.e., anthropogenic) and naturally occurring pollutant sources.
Examples of anthropogenic sources include electricity-generating power plants, cars and
trucks, and oil and gas production facilities. Natural pollutant sources include wildfires, dust
storms, volcanic activity, and biogenic sources (e.g., plants, soils).
Some of these pollutants are short-lived in the atmosphere (i.e., hours to days), while
others are long-lived (i.e., years). Like the weather, air quality changes from day to day, or
even hour to hour. The amount of time that a particular pollutant remains in the atmosphere
depends on its reactivity with other substances and its tendency to deposit on a surface
(also called deposition). These factors are governed by the type of pollutant and weather
conditions, including temperature, sunlight, precipitation, humidity, and wind speed. Strong
winds can decrease concentrations by diluting or dispersing pollutants over a larger
geographic area, whereas stagnant air (e.g., having no air flow) can lead to pollutant
concentrations that gradually increase.
Poor air quality
More pollution
Good air quality
Less pollution
8
-------
One of the most common causes of stagnant air is a temperature inversion, illustrated in
Figure 2-2, which occurs when a layer of cooler air is trapped close to the ground by a layer
of warmer air above. Inversions most commonly form overnight when clear skies allow air
at the surface to cool faster than the air above. An inversion can last all day, or even for
several days. When the air cannot rise, pollution at the surface is trapped and can
accumulate, leading to higher concentrations of pollution near the surface.
Cool
i <
^^WarrTI
er
18
k
f Cooler
~
b|A
Warm
Cool Pj
ft. idUB ¦
Warm Air Rising
Rising Warm Air Inhibited
by Inversion
Figure 2-2. Typical Movement of Warm Air in the Atmosphere Versus Warm
Air Trapped by a Temperature Inversion
Understanding how weather conditions can influence pollutant concentrations and
measurement of pollution is important for gathering accurate information and interpreting
trends in data. Figure 2-3 summarizes how conditions in the atmosphere can impact
pollutant concentrations.
9
-------
TP
Winds transport and disperse pollutants.
Pollutant concentrations may increase
during stagnant conditions.
Warmer air can speed up chemical
reactions in the atmosphere.
Moisture in the atmosphere and
/
clouds affect chemical reactions of
C )
some pollutants.
Sunshine causes some pollutants to
undergo chemical reactions, resulting
in the development of pollutants or
V J
mixtures like ozone or photochemical
' ~ x
smog.
Temperature inversions can trap pollution
near the ground and increase pollutant
concentrations. Without an inversion, air
can move upward and pollutants can mix
and disperse thereby decreasing
concentrations near the ground.
Figure 2-3. Atmospheric Conditions and Their Impacts on Pollutant Concentrations
Air pollutants are generally characterized as either primary or secondary pollutants, as
shown in Figure 2-4. Primary pollutants are emitted directly from a source. Secondary
pollutants are formed in the atmosphere by chemical reactions after release from an
emission source and, in the outdoor environment, are often found downwind from a source.
Pollutant concentrations can vary significantly over space and time because of variations in
local emissions, proximity to pollutant sources, chemical reactions, and weather conditions.
10
-------
Primary
Pollutants
Carbon Monoxide
CO Nitric Oxide
Sulfur Dioxide NO
S02 Nitrogen Dioxide
no2
Amrnia Particulate Matter I
NH- PM ,
Volatile Organic Compounds
5 VOCs
Volcanoes
Particulate Matter
PM
Dry and Wet
Deposition
Trains/Railway!
Communities
Wind
lanpQ^
Airplanes
Secondary
Pollutants
Nitric Acid
HNO.
Ozone
03
Ammonium
NH,+
Wildfires
Trees
Agriculture
mo Vehicle Exhaust
Figure 2-4. Sources of Primary and Secondary Pollutants (Adapted from:
https://www.mrqscience.com/ess-topic-63-photochemical-smoq.html)
Concentration is the metric for reporting the amount of a pollutant in the air and represents
the weight or number of molecules of a pollutant in a volume of air. Common units include
micrograms per cubic meter (pg/rn3), parts per million (ppm), and parts per billion (ppb)
Less common units include number of particles per cubic centimeter (#particles/cm3). For
example, a concentration of 43 pg/m3 is the weight of 43 micrograms (a microgram is one
millionth of a gram) per cubic meter of air, and parts per billion is the number of units of
mass of a pollutant per 1 billion units of the total mass of the air. Units of pg/m3 are often
used for particulate matter (PM) pollution and ppm and ppb are often used for gaseous
pollution.
11
-------
Common pollutants of concern in
outdoor air include PM (see Figure 2-5),
ground-level ozone (O3), sulfur dioxide
(SO2), nitrogen dioxide (NO2), carbon
monoxide (CO), lead (Pb), ammonia
(NH3), volatile organic compounds
(VOCs), mercury (Hg), airborne
particles, and more. Table 2-1 lists
these and other common air pollutants,
sources, and the range expected in the
outdoor air. While PM can range in size,
it is typically characterized into one of
two groups: PM10 and PM2.5. PM10
particles have diameters that are
generally less than 10 micrometers
(pm) while PM2.5 (also called "fine
particulate matter" or "fine PM") have
diameters generally less than 2.5 pm
(see Figure 2-5).
Table 2-1. Common Air Pollutants, Their Sources, and Concentration Ranges to
Expect in Outdoor Air
Pollutant (Abbreviation)
Examples of Outdoor Sources
Typical Hourly Outdoor
Concentration Range to
Expect within the U.S.
Ammonia (NH3)
Agriculture, animal husbandry, fertilizers,
and mobile sources
0 to 3 pg/m3
Benzene
Gasoline, evaporative losses from
above-ground storage tanks, and mobile
sources
0 to 7 pg/m3
(0.03 to 2.3 ppb)
Black Carbon (BC)
Biomass burning and mobile sources
0 to 15 pg/m3
Carbon Dioxide (CO2)
Fuel combustion from electric utilities
and mobile sources
350 to 600 ppm
Carbon Monoxide (CO)*
Incomplete fuel combustion from mobile
sources and industrial processes
0 to 0.3 ppm
Hydrogen Sulfide (H2S)
Natural sources (e.g., volcanoes, hot
springs, bacterial breakdown of organic
matter) and industrial sources (e.g.,
refineries, natural gas plants,
petrochemical plants, food processing,
tanneries)
0 to 20 ppm
HUMAN HAIR
50-70 nm
(microns) in diameter
c PM2.5
Combustion particles, organic
compounds, metals, etc.
2.5 (im (microns) in diameter
Wpm10
Dust, pollen, mold, etc.
<10 (im (microns) in diameter
90 |if71 (microns) in diameter
FINE BEACH SAND
Image courtesy of the U.S. EPA
Figure 2-5. Particulate Matter Size Ranges
12
-------
Pollutant (Abbreviation)
Examples of Outdoor Sources
Typical Hourly Outdoor
Concentration Range to
Expect within the U.S.
Lead (Pb)*
Smelting, aviation gasoline, waste
incinerators, electric utilities, and lead-
acid batteries
0 to 0.1 pg/m3
Mercury (Hg)
Combustion of coal, oil, and wood
0.001 to 0.17 pg/m3
Methane (CH4)
Industry (e.g., natural gas operations),
agriculture, and waste management
1,500 to 2,000 ppb
Nitrogen Dioxide (NO2)*
Fuel combustion from mobile sources
and electric utilities
0 to 50 ppb
Ozone (O3)*
Formed via ultraviolet (UV) radiation
in sunlight and the presence of other key
pollutants (e.g., nitrogen oxides, volatile
organic compounds)
0 to 125 ppb
Particulate Matter
(PM2.5)*
Fuel combustion (mobile sources,
electric utilities, industrial processes),
dust, agriculture, fires, and formation in
the atmosphere due to chemical
reactions
0 to 40 pg/m3
(100 to 1,000 pg/m3
near wildfires)
Particulate Matter
(PM10T
Dust (e.g., agriculture, roads,
construction), brake/tire and engine
wear from mobile sources, and fires
0 to 100 pg/m3
(500 to 1,000+ pg/m3
in dust storms)
Sulfur Dioxide (SO2)*
Fuel combustion from electric utilities,
refineries, and industrial processes
0 to 100 ppb
(100 to 5,000 ppb near
active volcanoes)
Ultrafine Particles (UFP)
Fuel combustion (mobile sources,
industries), gasoline evaporation, and
solvent usage
3,000
to 200,000 particles/cubic
centimeter (cm3)
Volatile Organic
Compounds (VOCs)
Fuel combustion (mobile sources,
industries), gasoline evaporation,
solvents, and consumer products
5 to 100 pg/m3
"Criteria pollutant regulated by the U.S. EPA (see Section 2.4)
13
-------
VOCs are not a single gas species but
are comprised of thousands of
chemicals. Total VOC (tVOC)
measurement is an estimated
concentration of several different VOC
species. VOCs are commonly found in
commercial products (e.g., paints,
refrigerants, fuels), consumer products
(e.g., cleaning supplies, deodorants,
hair products), and used in industrial
processes (e.g., chemical production,
petroleum refining, fuel combustion)
that evaporate into the air. VOCs occur indoors and outdoors. Some common VOCs
include:
• Benzene, Toluene, Ethylbenzene, and Xylene (BJEX) are four VOCs that are
normally grouped as they are often found together. The primary sources of BTEX
are on-road and non-road gasoline vehicles and engines, petroleum
transport/storage, and solvent usage.
• Biogenic VOCs are emissions created by some type of biological activity. Examples
include emissions resulting from trees, vegetation, and microbial activity in soils.
Emissions from biogenic sources can react in the atmosphere to form O3 and PM
pollutants.
• Formaldehyde is a colorless, flammable gas at room temperature that has a strong
odor. Formaldehyde is a byproduct of combustion (e.g., emission gas stoves,
kerosene space heaters, cigarette smoke). Some commercial products (e.g., glues,
paints, building materials) also release formaldehyde.
Sometimes, concerns about new and emerging pollutants arise as researchers and
scientists identify links with adverse health effects or environmental impacts. Some
examples of emerging pollutants include:
• Black carbon (BC) is a type of particle produced from incomplete combustion and is
emitted from sources such as diesel engines and wildfires. BC is almost entirely
made of carbon and is strongly light absorbing. BC absorbs solar radiation and may
lead to heating in the atmosphere (i.e., radiative forcing).
• Methane (CH4), the simplest hydrocarbon consisting of one carbon atom and four
hydrogen atoms, accounts for approximately 10 percent of all U.S. greenhouse gas
emissions from human activities, including leaks from natural gas systems and
raising of livestock.
What are mobile sources?
Mobile sources include vehicles on the road
(i.e., on-road) like motorcycles, passenger
cars and trucks, and commercial trucks and
buses. Mobile sources also include vehicles
not on roads (i.e., non-road) like excavators,
aircraft, locomotives, marine vessels, other
heavy equipment, recreation vehicles (e.g.,
snowmobiles, all-terrain vehicles), and small
engines and tools (e.g., lawnmowers).
14
-------
• Polycyclic aromatic
hydrocarbons (PAHs) are a
class of chemicals that occur
naturally in coal, crude oil, and
gasoline and diesel fuel. They are
also produced by combustion of
coal, oil, gas, wood, garbage, and
tobacco, and by high-temperature
cooking of meat and other foods.
PAHs are a concern because
they persist in the environment for
long periods of time.
2.1.1 The Pollutant Lifecycle
After being em itted from a natural or
anthropogenic source, air pollutants are
transported in the atmosphere and can
ultimately impact the environment and
human population, as shown in Figure 2-
6. Many processes throughout this
lifecycle can make it challenging to
understand pollution and its sources.
While in the atmosphere, pollutants can undergo chemical reactions with other gases and
particles. Atmospheric conditions can affect these chemical reactions (e.g., sunlight is
needed to create O3) and weather controls the movement and dispersion of pollutants from
upwind locations (where air moves through before it goes over an area of interest) to
downwind locations (where air goes after moving over an area of interest). Local geography
can channel and direct the air pollutants to locations that can impact people and the
environment. Eventually pollutants are removed from the atmosphere by deposition onto
earth's surface or as people breathe in pollution. One of the biggest challenges in air quality
monitoring is determining the origin of a measured pollutant, which requires an
understanding of all factors—weather, atmospheric chemistry, and geography—that
affected the pollutants on their journey from the original source to the measurement
location.
/\
What are the Typical Characteristics of
Traffic Emissions?
Pollutants directly em itted from cars, trucks
and other mobile sources are found in
higher concentrations near major roads.
Examples of directly emitted pollutants
include PM, CO, oxides of nitrogen (NOx),
and benzene, though hundreds of
chemicals are emitted by motor vehicles.
Motor sources also emit compounds that
lead to the formation of other pollutants in
the atmosphere, such as NO2, which is
found in elevated concentrations near major
roads, and O3, which forms further
downwind. Beyond vehicle tailpipe and
evaporative emissions, roadway traffic also
emits brake and tire debris and can throw
road dust into the air. Individually and in
combination, many of the pollutants found
near roadways have been associated with
adverse health effects.
15
-------
Chemical Reactions
in The Atmosphere ^
//T\S
2A+B=D#
9 0 C+D=E
Movement by
Weather
A+B=C
limn
mint
Emitted Pollutants _ Formation
©V- I
Dispersion
Emissions Deposition
Human made and
Natural Sources
Effect on Local
Geography
M
Impacts On People
and Environment
Figure 2-6. The Pollutant Lifecycle from Source to Impact on People and the
Environment
2.1.2 Differences in Pollutant Concentrations Over Time
Pollutant concentrations may vary significantly depending on the time of day, day of the
week, and season. These differences can be attributed to changes in emissions patterns,
atmospheric conditions (e.g., mixing height, temperature, sunlight), the source's activity
schedule (e.g., daily traffic rush hour patterns), and chemical reactions.
As shown in Figure 2-7. O3 typically varies slowly on an hour-to-hour basis but often
undergoes a diurnal change (i.e., daily cycle) from low concentrations at night to higher
concentrations during the day. The day-to-day difference in emissions of O3 precursors
(e.g., vehicle emissions) can produce lower O3 concentrations on weekends than
weekdays. As sunlight and heat help convert emissions into O3, concentrations are usually
higher during summer.
16
-------
PM2.5 can change more rapidly on an hour-to-
hour basis, even on a minute-to-minute basis,
due to local sources and atmospheric
conditions (see Figure 2-7 and Figure 2-3).
PM2 5 concentrations are generally higher at
night and in the morning due to calm winds
and pollution trapped below a nighttime
temperature inversion. From day to day,
PM2.5 concentrations can build up and then
change rapidly due to increasing winds or air
mass changes (e.g., cold front passage).
Lastly, PM2.5 concentrations can vary based
on season. For instance, wintertime
concentrations may be high due to residential
wood-burning activities in cold climates. Summertime concentrations may be high due to
secondary particle formation in the Southeast U.S., seasonal crop burning, agricultural
activities, or wildfire smoke.
v_
Tip: Consider variations in pollutant
concentrations when developing an
air monitoring plan
Knowing the daily, weekly, and
seasonal variations in pollutant
concentrations can help you develop a
plan for conducting air monitoring. This
information can help guide the time
and conditions under which
measurements should be taken. See
Chapter 3 for more information on
monitoring using air sensors.
j
Hour
Day
PM
2.5
r
W
1 hr 2 hr
1 r-
1 hr 2 hr
Midnight
Noon
Midnight
Midnight Noon Midnight
Week
Year
JFMAMJJASOND
JFMAMJJASOND
Figure 2-7. Typical Concentrations for O3 and PM2.5 During Different Time Periods
17
-------
Resources for More Information
• U.S. EPA Air Quality Planning and Standards Website
o Provides additional information regarding air quality and pollutants
o https://www3.epa.gov/airqualitv/
• U.S. EPA National Air Quality - Status and Trends of Key Air Pollutants
Website
o Provides air quality trends, reports, and summaries for criteria air pollutants
o https://www.epa.gov/air-trends
• U.S. EPA AirNow Website
o Provides a variety of resources on air quality including air quality information
at local, state, national, and world views, air quality and health, maps and
data, educational resources, and more
o https://www.airnow.gov/
• Wildfire Smoke: A Guide for Public Health Officials, U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, EPA-452/R-19-
901, August 2019
o Document provides guidance to tribal, state, and local public health officials,
and other interested groups (e.g., health professionals, air quality officials,
public) in preparing for wildfire smoke events and in communicating health
risks and taking measures to protect the public during smoke events
o https://www.airnow.gov/sites/default/files/2021-05/wildfire-smoke-guide-
revised-2019.pdf
• U.S. EPA Mobile Source Pollution and Related Health Effects Website
o Overviews mobile sources of air pollution, summarizes health effects
associated with exposure to mobile source emissions, provides data and
modeling resources, and information on programs to reduce mobile source
pollution
o https://www.epa.gov/mobile-source-pollution
• U.S. EPA Near-Roadway and Other Near-Source Pollution Website
o Overview of research on near-roadway pollution from cars, trucks, and other
mobile sources and frequently asked questions about near-roadway air
pollution and health effects
o https://www.epa.gov/air-research/research-near-roadwav-and-other-near-
source-air-pollution
18
-------
• Near-Roadway Air Pollution and Health: Frequently Asked Questions, U.S.
Environmental Protection Agency, Office of Transportation and Air Quality, EPA-
420-F-14-044, August 2014
o Document provides U.S. EPA's responses to frequently asked questions
received from the public regarding exposure to near-roadway air pollution
o https://nepis.epa.gov/Exe/ZvPDF.cgi/P100NFFD.PDF?Dockev=P100NFFD.P
DF
• Report to Congress on Black Carbon, U.S. Environmental Protection Agency,
EPA-450/R-12-001, March 2012
o Document summarizes available scientific information on the current and
future impacts of black carbon (BC) and evaluates the effectiveness of
available BC mitigation approaches and technologies
o https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100EIJZ.txt
• U.S. EPA Integrated Science Assessments (ISAs) for Criteria Air Pollutants
Website
o Reports that summarize scientific information that is the foundation for
reviewing the National Ambient Air Quality Standards (NAAQS) for criteria
pollutants; ISAs are an important resource for state and local health agencies,
other federal agencies, and international health organizations
o https://www.epa.gov/isa
2.2 Pollutant Effects on Health and the Environment
A broad range of health and environmental effects have been observed following exposure
to air pollutants or mixtures of air pollutants. Health effects vary by the type of pollutant or
mixture of pollutants, concentration, and exposure time, which can be short term (hours to
weeks) or long term (months to years). Potential health effects associated with air pollution
exposures include decreased lung function, aggravation of respiratory and cardiovascular
diseases, and increased asthma incidence and severity, among a variety of other effects.
Table 2-2, provided by EPA's Office of Air Quality Planning and Standards Health and
Environmental Impacts Division, summarizes the health and environmental effects of
common air pollutants.
Table 2-2. Health and Environmental Effects of Select Common Air Pollutants
Pollutant
Health Effects
Environmental Effects
Ammonia (Nhh)
• Can cause severe irritation of the
skin, eyes, nose, and throat
• Precursor to secondary particulate
formation
• Contributes to acid deposition of soils
and surface waters
Benzene,
Toluene,
Ethylbenzene,
• Known human carcinogen
• Short-term exposure can cause
drowsiness, dizziness, headaches,
• Can contribute to formation of ground-
level O3
19
-------
Pollutant
Health Effects
Environmental Effects
and Xylene
(BTEX)
and irritation to the eyes, skin, and
respiratory tract
• Long-term exposure can cause blood
disorders and adverse effects to the
reproductive system
Black Carbon
(BC)
• A component of PM (see health
effects of PM)
• A component of PM (see
environmental effects of PM)
• Contributes to climate change causing
changes in patterns of rain and clouds
• As BC deposits in the Arctic, the
particles cover the snow and ice,
decreasing the Earth's ability to reflect
warming rays of the sun, while
absorbing heat and hastening melting
Carbon Monoxide
(CO)
• Breathing air with a high
concentration reduces the amount of
oxygen that can be transported in the
blood stream to critical organs like
the heart and brain
• At very high levels, which are
possible indoors or in other enclosed
environments, CO can cause
dizziness, confusion,
unconsciousness, and death
• Short-term exposure to elevated CO
may result in reduced oxygen to the
heart accompanied by chest pain
(also known as angina)
• There is no secondary standard as the
concentrations that would have any
effect on flora or fauna are many times
higher than what might occur in
outdoor air
Ground-level
Ozone (O3)
• Elevated concentrations cause
respiratory effects and can:
0 Trigger responses such as throat
irritation or burning sensation in
the airways, coughing, difficulty
breathing, and airway
inflammation
0 Reduce lung function and harm
lung tissue
0 Aggravate bronchitis,
emphysema, and asthma,
increasing the risk of asthma
attacks and associated outcomes
in affected individuals
• Elevated concentrations can:
0 Affect sensitive vegetation and
ecosystems, including in parks and
wilderness areas, by reducing
photosynthesis and slowing growth
0 Increase sensitive plants' risk from
other stressors such as disease or
insects
0 Cause blemishes or stippling on
leaves of sensitive plants
• O3 is a greenhouse gas, with
associated potential for effects on
climate
Lead (Pb)
• Elevated exposures can adversely
affect the nervous system, kidney
function, immune system,
reproductive and developmental
systems, and the cardiovascular
system
• Elevated exposure can also affect the
oxygen carrying capacity of the blood
• The effects most likely to be
encountered in current populations
are neurological effects in children.
Infants and young children are
especially sensitive to lead
• Elevated Pb in the environment can
result in effects on:
0 Terrestrial and aquatic animal
behavior, reproduction,
development, and survival
0 Plant growth
20
-------
Pollutant
Health Effects
Environmental Effects
exposures, which may contribute to
behavioral problems, learning
deficits, and lowered intelligence
quotient (IQ)
Mercury (Hg)
• Exposure to methylmercury (MeHg)
occurs when people eat fish and fish
products with high levels of MeHg
• The primary effect is damage to the
central nervous system:
0 At very high exposures,
symptoms may include blurred
vision, malaise, speech
difficulties, and constriction of the
visual field
0 For infants born to pregnant
people with high levels of
methylmercury, effects such as
mental retardation, ataxia,
constriction of the visual field,
blindness, and cerebral palsy. At
lower MeHg concentrations,
effects such as developmental
delays and abnormal reflexes
• Deposits to soil and bodies of water
• MeHg can accumulate in fish which are
then consumed by birds, mammals,
and predators and, at very high
concentrations, cause effects in
animals that include:
0 Death
0 Reduced reproduction
0 Slower growth and development
0 Abnormal behavior
Methane (CH-t)
• Not generally considered a toxic gas
• Very high exposures can cause
headaches, dizziness, nausea,
vomiting, and, in severe cases,
respiratory problems and loss of
consciousness
• Greenhouse gas that contributes to
climate change, particularly warming of
the Earth
Nitrogen Dioxide
(NO2)
• Exposure to high concentrations of
NO2 can irritate airways in the
respiratory system
• Exposure to elevated concentrations
over short periods of time can
aggravate respiratory diseases,
particularly asthma, leading to
respiratory symptoms (e.g., coughing,
wheezing, difficulty breathing)
• Longer exposures to elevated
concentrations may contribute to the
development of asthma and
potentially increase susceptibility to
respiratory infections
• Contributes to the formation of
ground-level O3 and PM (see health
effects of PM)
• NO2 and other oxides of nitrogen (NOx)
interact with water, oxygen, and other
chemicals in the atmosphere to form
acid rain, which can harm sensitive
ecosystems such as lakes and forests
• Can contribute to nutrient pollution in
coastal waters
• Contributes to the formation of ground-
level O3 and PM (see environmental
effects of PM)
21
-------
Pollutant
Health Effects
Environmental Effects
Particulate Matter
(PM2.3 and PM10)
• Exposure can affect both the lungs
and the heart. Numerous scientific
studies have linked particle pollution
to a variety of problems, including:
0 Premature death in people with
heart or lung disease
0 Nonfatal heart attacks
0 Irregular heartbeat
0 Aggravated asthma
0 Decreased lung function
0 Increased respiratory symptoms,
such as irritation of the airways,
coughing, or difficulty breathing
• Can reduce visibility (create haze)
• Can be carried over long distances by
wind and then settle on ground or
water. Depending on its chemical
composition, the effects of this settling
may include:
0 Acid rain-like effects, such as
making lakes and streams acidic
0 Effects on the nutrient balance in
coastal waters and large river
basins or in soils, with related
effects on sensitive forests,
ecosystems, and farm crops
• Can have effects on climate, including
radiative forcing
• Can stain and damage stone and other
materials
Sulfur Dioxide
(S02)
• Short-term exposure to elevated
concentrations can harm the
respiratory system and make
breathing difficult
• Contributes to the formation of PM
(see health effects of PM)
• At high concentrations, gaseous sulfur
oxides (SOx) can harm trees and
plants by damaging foliage and
decreasing growth
• SO2 and other SOx interact with water,
oxygen, and other chemicals in the
atmosphere to form acid rain, which
can harm sensitive ecosystems such
as lakes and forests
• Contributes to the formation of PM
(see environmental effects of PM)
Ultrafine Particles
(UFP)
• Included in PM25 mass (see health
effects of PM)
• There is some evidence associated
with respiratory, cardiovascular, and
nervous system effects, but research
on UFP-associated health effects is
still emerging
• Included in PM2 5 mass (see
environmental effects of PM)
Volatile Organic
Compounds
(VOCs)
• Some are air toxic pollutants that
cause cancer and/or other serious
health effects
• Contributes to the formation of
ground-level O3 and PM (see health
effects of PM)
• Contributes to the formation of ground-
level O3 and PM (see environmental
effects of PM)
Resources for More Information
• U.S. EPA Criteria Air Pollutants Website
o Provides detailed information on the six criteria pollutants including basic
information, health and environmental effects, technical documents, setting
and reviewing the standards, implementing the standards, and current air
quality designations
o https://www.epa.gov/criteria-air-pollutants
22
-------
Health Effects of Ozone (O3) Pollution Website
o Provides detailed information on health effects of breathing air containing O3
o https://www.epa.gov/ground-level-ozone-pollution/health-effects-ozone-
pollution
Health and Environmental Effects of Particulate Matter (PM) Website
o Provides detailed information on health and environmental effects of PM
o https://www.epa.gov/pm-pollution/health-and-environmental-effects-
particulate-matter-pm
Basic Information about Nitrogen Dioxide (NO2) Website
o Provides basic information on NO2 including health and environmental effects
o https://www.epa.goV/no2-pollution/basic-information-about-no2#Effects
Basic Information about Sulfur Dioxide (SO2) Website
o Provides basic information on SO2 including health and environmental effects
o https://www.epa.gOv/so2-pollution/sulfur-dioxide-basics#effects
Basic Information about Carbon Monoxide (CO) Outdoor Air Pollution Website
o Provides basic information on CO including health and environmental effects
o https://www.epa.gov/co-pollution/basic-information-about-carbon-monoxide-
co-outdoor-air-pollution#Effects
Basic Information about Lead (Pb) Air Pollution Website
o Provides basic information on Pb including health and environmental effects
o https://www.epa.gov/co-pollution/basic-information-about-carbon-monoxide-
co-outdoor-air-pollution#Effects
Report on the Environment - Volatile Organic Compounds (VOCs) Emissions
Website
o Provides detailed information on sources, health and environmental effects,
and emissions estimates of VOCs
o https://cfpub.epa.gov/roe/indicator.cfm?i=23#1
Definition of VOC Website
o Provides a detailed information on the definition of VOCs as outlined in air
pollution regulations
o https://www.epa.gov/air-emissions-inventories/what-definition-voc
Health Effects of Exposures to Mercury Website
o Provides detailed information on health effects of exposure to mercury
o https://www.epa.gov/mercurv/health-effects-exposures-mercurv
23
-------
Integrated Risk Information System (IRIS) Methylmercury (MeHg) Summary
Website
o Provides health assessment information on MeHg based on review of toxicity
data
o https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=73
Agency for Toxic Substances and Disease Registry (ATSDR) Public Health
Statement for Benzene Website
o Provides information about benzene and effects of exposure to it
o https://wwwn.cdc.gov/TSP/PHS/PHS.aspx?phsid=37&toxid=14
Integrated Risk Information System (IRIS) Benzene Summary Website
o Provides health assessment information on benzene based on review of
toxicity data
o https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=276
Global Methane Initiative (GMI) Website
o Provides information on GMI, description of methane and mitigation
approaches, methane sites around the globe, and more
o https://www.epa.gov/gmi
Report to Congress on Black Carbon, U.S. Environmental Protection Agency,
EPA-450/R-12-001, March 2012
o Report provides summary on black carbon, health and environmental effects,
emissions, mitigation overview, and more
o https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100EIJZ.txt
Integrated Science Assessment (ISA) for Particulate Matter, U.S. Environmental
Protection Agency, EPA/600/R-19/188, December 2019
o ISA provides detailed information on particulate matter including sources,
ambient levels, health and environmental effects, and more
o https://www.epa.gov/isa/integrated-science-assessment-isa-particulate-matter
Traffic-Related Air Pollution: A Critical Review of the Literature on Emissions,
Exposure, and Health Effects, HEI Panel on the Health Effects of Traffic-Related
Air Pollution, HEI Special Report 17, Health Effects Institute (HEI), January 2010
o Report provides a summary and synthesis of information on air pollution from
traffic and its health effects
o https://www.healtheffects.org/publication/traffic-related-air-pollution-critical-
review-literature-emissions-exposure-and-health
24
-------
2.3 Outdoor Air Pollution Monitoring
Air pollution monitoring is the detection of pollutant levels by measuring the quantity and
types of certain pollutants in the outdoor air. Different methods and instruments are used to
measure pollutants and it is critical to match a specific application's measurement needs
with a device that provides sufficient accuracy, reliability, and traceability.
Some common measurement approaches for monitoring air quality from fixed locations and
mobile platforms (e.g., a vehicle) are shown in Figure 2-8. Several categories of
measurement methods are described in more detail below.
• Reference monitors are used to determine compliance with the National Ambient
Air Quality Standards (NAAQS; see Section 2.4) and are designated as either
Federal Reference Method (FRM) or Federal Equivalent Method (FEM) monitors.
These monitors must meet strict operating and performance requirements, as
outlined in the U.S. Code of Federal Regulations (40 CFR Parts 50, 53, and Part 58).
Reference monitors produce very high quality, accurate data. PM FRM samplers
collect particles on a filter over a period of time (typically 24 hours), whereas
continuous FEM monitors detect pollutant concentrations on a more frequent basis
(e.g., continuously every hour). See the U.S. EPA's Ambient Monitoring Technology
Information Center (AMTIC) for a list of designated FRM/FEM monitors.
Precise operation, siting, and quality assurance and quality control (QA/QC) are
necessary to ensure that reference monitors produce accurate data.
• Research instruments describe a wide range of technologies ranging from lower-
cost air sensor technologies to mid-range prototype instruments to high-end
laboratory-type instruments modified for use in the field. These instruments are often
built or designed for specific applications. They may be designed to explore a specific
research question or as prototypes designed to measure pollutants that do not have
an existing measurement method. These instruments are most often operated by
experts to achieve the performance needs for their application. After demonstrating
strong comparability with FRM/FEM instruments, they can be used as a reference
monitor if/when it is not feasible to deploy an FRM/FEM. For example, an instrument
like an E-BAM may be considered a near-reference instrument because it has
comparable performance to FRM/FEM instruments, but it is designed to be operated
outdoors and designed to be portable so that it can be quickly deployed to measure
wildfire smoke in more rugged terrain. It does not have an FEM designation, but has
many of the features of an FEM, and when operated by trained staff can provide
valuable data for this application.
• Remote sensing is a method for measuring pollutants at a distance without physical
contact (e.g., by measuring reflected or emitted light). Remote sensing can be useful
for detecting PM, gaseous criteria pollutants, and some VOCs. Remote sensing can
be deployed aboard aircraft, satellite-based platforms in orbit, or at ground-based
sites.
25
-------
• Air sensors are a class of technology that are lower in cost, more portable, and
generally easier to operate than reference monitors or research instruments. Air
sensors can measure some, but not all, air pollutants including PM and gases.
However, the accuracy, lifetime, and reliability of sensors varies due to the
underlying measurement technology, quality of different air sensor components,
environmental conditions, and methods of operation. Table 2-3 summarizes the
main differences between air sensors and reference monitors.
Reference
(Certified)
Sampler Continuous
Remote Sensing
Ground- Space-
based based
XI
ID III
Air Sensor
i
Accuracy
Complexity
Measurement
Frequency
Relative Cost
High
High
Daily
$$
High
High
Sub hourly to
hourly
$$S
High
High
Sub hourly
$$$$
Varies
High
Hourly to Daily
$$$
Varies
Low
Sub hourly
$
Figure 2-8. Common Types of Air Monitoring Instruments and Their Characteristics
Table 2-3. Comparison Between Reference Monitors and Air Sensors
Consideration
Reference Monitors
Air Sensors
Typical Purchase
Cost
$15,000 to $40,000 (U.S. Dollars)
$100 to $5,000 (U.S. Dollars)
Staff Training for
Operation
Highly trained technical staff
Little or no training
(user guide/manual may help)
Operating
Expense
Expensive - need for shelter,
technical staff, maintenance,
repair, quality assurance
Less expensive - need for
device replacement or repair,
data streaming, data
management
Siting Location
Fixed location
(building/trailer needed)
More portable
(with basic weather protection)
Data Quality
Known and consistent quality in a
variety of conditions
Unknown and may vary from
sensor to sensor in different
weather conditions and
pollution environments
Operating Lifetime
10 years or more
(calibrated and operated to
maintain accuracy)
Short (pollutant dependent; <1-
3 years) (may become less
accurate over time)
Used for
Regulatory
Monitoring
Yes
No
26
-------
The concentration of many pollutants varies over long or even short distances, therefore
deciding where to place instruments is very important. The location of where pollutants are
measured determines what is measured. As shown in Figure 2-9, air monitoring can occur
in different locations and represent different conditions.
Occupational Near Mobile Indoor Ambient
Source
Figure 2-9. Different Air Monitoring Locations for Outdoor Air (Near Source, Mobile,
Ambient, and Background) and Indoor Air (Occupational and Residential)
Concentrations for most pollutants will
almost always be highest near the source,
decreasing rapidly within the first few
hundred feet away from the source. If
multiple sources are widely distributed
within a given area, pollutant
concentrations may be more similar, but
will still be different from location to
location. Other factors, including geography
and local atmospheric conditions will also
influence concentrations.
An air pollution monitoring network is a collection of sites equipped with instrumentation
for measuring one or more pollutants. Monitoring networks are designed and operated for
specific purposes and objectives and use instruments that can meet the goals necessary
for the objective (e.g., use of FRM or FEM for monitoring the NAAQS as required by the
Clean Air Act. CAA) because of the stringent needs for accuracy and completeness. U.S.
EPA's AirData Air Quality Monitors website (Figure 2-10) provides an interactive display of
air monitoring locations and monitor-specific information for the different regulatory air
monitoring networks using the AirData Map.
—
Tip: Carefully consider where to locate
air sensors when conducting an air
monitoring study
Carefully locating an air sensor will play a
significant role in determining whether the
data collected represent the location and
are useful. Section 3.5 provides further
discussion regarding where and how to
properly place air sensor devices.
27
-------
Within the United States (U.S.), Puerto Rico,
and the U.S. Virgin Islands, ambient
(outdoor) concentrations of criteria air
pollutants are measured at more than 4,000
monitoring stations owned and operated by
tribal, state, local, or environmental
agencies. Each of these sites measures one
or more of the criteria air pollutants [O3, PM
(PM2.5 and PM10), CO, NO2, SO2, Pb] using
instruments that have been designated as FRMs or FEMs. In managing these instruments,
agencies follow stringent siting criteria, extensive operational plans, and quality assurance
procedures (e.g., calibration, maintenance, audits, data validation) to ensure that they
produce high quality data. EPA's QA Handbook further describes these procedures.
Agencies send quality assured hourly or daily measurements of pollutant concentrations to
EPA's database called the Air Quality System (AOS). AirData retrieves the air quality data
from AQS and provides the public with easy access to daily and annual data summaries.
Continuous Monitoring Stations (as of 9/30/2022)
Several other air monitoring networks serving other needs also exist within the U.S.
including
• National Air Toxics Trends Stations (HAT fS) which are set up across the U.S. to
monitor air toxics. Under the Clean Air Act (CAA). U.S. EPA regulates a list of 187
hazardous air pollutants (HAPs), also called air toxics. The principal objective of the
NATTS network is to provide long-term monitoring data across representative areas
of the country to establish overall trends for priority pollutants such as benzene,
formaldehyde, 1,3-butadiene, hexavalent chromium, PAHs, and others.
f S
Tip: You can access outdoor air
monitoring data in several ways
Current conditions and forecast data are
available through AirNow.gov. Historical
data are available as pre-generated data
files, through the AQS API, or through
AirData reports and visualizations.
28
-------
• The Interagency Monitoring of Protected Visual Environments (IMPROVE)
monitoring program which was initiated in response to the CAA and Regional Haze
Rule to establish current visibility and aerosol conditions, identify chemical species
and anthropogenic emission sources responsible for visibility impairment, and
document long-term trends for assessing progress toward visibility goals for national
parks and wilderness areas. The IMPROVE network uses identical reference
samplers to collect 24-hour PM filters every three days that are analyzed for
chemical components in PM.
• The Chemical Speciation Network (CSN) which collects PM2.5filter samples that
are analyzed for chemical components in PM (e.g., metals, ions, carbon). These
data are collected to assess trends, link health effects to PM2.5 constituents,
characterize annual and seasonal spatial variation of PM2.5, and other applications.
• The National Core Multipollutant Monitoring Network (NCore) which is a network
that integrates several advanced measurement systems for PM, gaseous pollutants,
and meteorology. NCore sites collect data to track long-term trends, support long-
term health assessments, support scientific studies, and other applications.
Resources for More Information
• U.S. EPA Ambient Monitoring Technology Information Center (AMTIC) Website
o Contains technical information regarding ambient air monitoring programs,
including the networks of state and local air monitoring stations (SLAMS),
monitoring methods, and QA/QC procedures
o https://www.epa.gov/amtic
• U.S. EPA Ambient Air Monitoring Website
o Overviews the reasons for why monitoring ambient air quality is important and
provides links to U.S. EPA's AMTIC, Air Quality System (AQS), Air Data,
AirNow, and AirNow International websites
o https://www.epa.gov/air-qualitv-management-process/managing-air-qualitv-
ambient-air-monitorinq
• Overview of the Clean Air Act (CAA) Website
o Provides an in-depth overview of the CAA including history and requirements,
role of science and technology, role of state, local, tribal and federal
government, and more
o https://www.epa.gov/clean-air-act-overview
29
-------
• Videos on Sources of Air Quality Information and Air Sensor Measurements,
Data Quality, and Interpretation
o Educational videos, in both English and Spanish, that can be used to learn
how U.S. EPA collects and uses air quality data, how air quality health risks
are communicated, and how to interpret data collected using air sensors
o https://www.epa.gov/air-sensor-toolbox/videos-air-sensor-measurements-
data-qualitv-and-interpretation
• Understanding Air Quality and Monitoring Video, South Coast Air Quality
Management District (South Coast AQMD), Air Quality Sensor Performance
Evaluation Center (AQ-SPEC), September 2021
o Educational video providing background on air quality, criteria pollutants,
pollutant sources and health effects, air quality monitoring technologies, and
the role of air sensors
o https://www.voutube.com/watch?v=2r0XxQm50IE
• California Air Resources Board (CARB) Outline of Measurement Technologies
o Online resource that discusses air monitoring applications, applicable
measurement technologies, and their relative availability and cost developed
by CARB to support community air monitoring conducted under California
Assembly Bill 617
o https://ww2.arb.ca.gov/capp-resource-center/communitv-air-
monitoring/outline-of-measurement-technologies
• Hazardous Air Pollutants (HAPs) Website
o Provides detailed information on HAPs including the list of HAPs, health an
environmental effects, sources and exposures, data, and more
o https://www.epa.gov/haps
• Regional Haze Program Website
o Provides information on the Regional Haze Rule and Program, list of the
national parks and wilderness areas covered by the program, and more
o https://www.epa.gov/visibilitv/regional-haze-program
2.4 Air Quality Standards and Indices
Determining the health implications of air quality measurements is complex. Fortunately,
there are ways to put air quality measurements into context. Government agencies conduct
extensive analyses of the research on health effects of air pollutants (e.g., U.S. EPA's
Integrated Science Assessments, the 2021 Air Quality in Europe report prepared by the
European Union). This research and high-quality data are used to establish the standards
and indices to protect public health. Many countries set their own air quality standards. In
addition, the WHO has established air quality standards.
30
-------
Major U.S. air quality standards arid indices include:
• National Ambient Air Quality Standards (NAAQS). The CAA, which was last
amended in 1990, requires the U.S. EPA to set NAAQS for air pollutants considered
harmful to public health and the environment. The CAA identifies two types of
NAAQS: primary and secondary standards. Primary standards provide public health
protection, including protecting the health of
"sensitive" populations such as asthmatics,
children, and the elderly. Secondary standards
provide public welfare protection, including
protection against decreased visibility and
damage to animals, crops, vegetation, and
buildings. The U.S. EPA has set NAAQS for six
principal pollutants, called criteria air
pollutants PM (PM2.5 and PM10), O3, CO, SO2,
NO2, and Pb. As required by the CAA, the U.S.
EPA reviews and revises the standards, if
appropriate, every 5 years. The NAAQS are
summarized in "able 2-4.
• Air Quality Index (AQI). The AQI was established by the U.S. EPA as a method to
translate pollution measurements into potential health effects. The AQI is a numeric
scale for reporting air quality that describes how clean or polluted the air is at a given
location, and any associated health effects that may result from exposure to the air.
Section 2.5 and Appendix D provide more information on the AQI and how it is
calculated.
• National Institute for
Occupational Safety and Health
(NIOSH) NIOSH has established
guidelines and recommendations
for preventing work-related injury
and illness, including exposure to
air pollutants. In general, these
guidelines often represent shorter
time periods because they relate to
occupational locations and
schedules (e.g., workdays).
Why Aren't there NAAQS
for all Air Pollutants?
Other common air pollutants
shown in Table 2-1 (e.g.,
CH4, VOCs, benzene. Hg,
NH3, BC, UFP) are not
criteria pollutants; therefore,
the U.S. EPA has not
established NAAQS for
these pollutants.
How are NAAQS and NIOSH
Guidelines Different?
NIOSH guidelines are established for
some of the pollutants shown in Table
2-1 and some additional pollutants as
well. NIOSH concentration levels are
different than the NAAQS because
they are established for occupational
exposures only.
31
-------
Table 2-4. U.S. EPA National Ambient Air Quality Standards (NAAQS; current as of
9/30/2022)
Air Pollutant
Primary or
Secondary
Standard*
Averaging
Time
Concentration
Level
Form**
Carbon Monoxide (CO)
Primary
8 hours
9 ppm
Not to be exceeded
more than once per year
1 hour
35 ppm
Lead (Pb)
Primary and
Secondary
Roiling 3
month
average
0.15 (.ig/m3
Not to be exceeded
Nitrogen Dioxide (NO2)
Primary
1 hour
100 ppb
98th percentile of 1-hour
daily maximum
concentrations,
averaged over 3 years
Primary and
Secondary
1 year
53 ppb
Annual mean
Ozone (O3)
Primary and
Secondary
8 hours
70 ppb
Annual fourth-highest
daily maximum 8-hour
concentration, averaged
over 3 years
Particulate Matter (PM2.5)
Primary
1 year
12.0 jj,g/m3
Annual mean, averaged
over 3 years
Secondary
1 year
15.0 ,ug/m3
Annual mean, averaged
over 3 years
Primary and
Secondary
24 hours
35 jag/m3
98th percentile, averaged
over 3 years
Particulate Matter (PM10)
Primary and
Secondary
24 hours
150 |ag/m3
Not to be exceeded
more than once per year
on average over 3 years
Sulfur Dioxide (SO2)
Primary
1 hour
75 ppb
99th percentile of 1-hour
daily maximum
concentrations,
averaged over 3 years
Secondary
3 hours
0.5 ppm
Not to be exceeded
more than once per year
The primary standard provides public healt
h protection, including protecting the health of sensitive
populations such as asthmatics, children, and the elderly. The secondary standard provides public
welfare protection, including protection against decreased visibility and damage to animals, crops,
vegetation, and buildings.
"See the AQS data dictionary for the definitions and calculations of these forms.
32
-------
In the U.S. only data from
properly operated, cited, and
maintained FRM and FEM
instruments are used to
determine compliance with
the NAAQS. When
comparing air quality
measurements to these
standards, it is important to
ensure that the
measurements align with the
standard. Every standard has
an associated concentration
level for a specified averaging
time period (e.g., 1 hour, 24
hours, 1 year) and a location
(e.g., ambient, occupational). For example, Table 2-4 shows the related averaging time
period, and concentration level for each criteria air pollutant for the NAAQS.
How Can I Compare Air Sensor Measurements to
the NAAQS or AQI for Informational Purposes?
When comparing measurements from an air sensor to
the NAAQS or AQI, it is important to remember that air
sensors may over- or under-estimate pollutant
concentrations (see Section 3.6). Therefore, sensor
data must be cleaned and corrected and then averaged
to match the time average specified for the pollutant
and air quality standard or index. For example, to
compare O3 air sensor measurements provided every
minute to the 8-hour NAAQS for O3 of 70 ppb, you
would need to clean and correct the O3 sensor data
and then calculate an 8-hour average from the 1-
minute sensor measurements before comparing.
What Happens if an Air Pollutant Measurement is Above the NAAQS
Concentration Level for the Specified Averaging Period?
Each NAAQS has a 'form' (see Table 2-4) which is a criterion for how many times the
standard may be exceeded in a certain timeframe. Even though a measured
concentration may exceed the NAAQS (called an exceedance), it does not constitute a
NAAQS violation. So, what is a NAAQS exceedance vs. a NAAQS violation?
A NAAQS exceedance occurs when a measured concentration exceeds the
concentration level for the averaging period specified by the NAAQS. For example, an
exceedance of the short-term (24-hour) PM2.5 NAAQS occurs when the PM2.5
concentration measured at a regulatory air monitoring location is greater than 35 ng/m3.
Air monitoring agencies must report NAAQS exceedances to the public.
A NAAQS violation occurs when a measured concentration level exceeds the
concentration level for the specified NAAQS averaging period for specific criteria over a
specified timeframe. For example:
A violation of the 24-hour PM2.5 NAAQS occurs when the 3-year average of the annual
98th percentile 24-hour concentration is greater than 35 pg/m3
A violation of the 1-year PM2.5 NAAQS occurs when the annual mean averaged over 3
years is greater than 12 pg/m3
An area that has a NAAQS violation for any given criteria air pollutant, can potentially
be designated as nonattainment (not meeting the NAAQS) for that pollutant and may
need to address in State Implementation Plans (SIPs) how they will reach attainment.
33
-------
Resources for More Information
• WHO National Air Quality Standards Tool
o An interactive tool providing an international map of current national air quality
standards for criteria pollutants for various averaging times
o https://www.who.int/tools/air-quality-standards
• Air Quality in Europe 2021
o An annual assessment of recent air quality trends at both European and
national levels
o https://www.eea.europa.eu//publications/air-quality-in-europe-2021
• Air Quality System Data Dictionary
o The AQS Data Dictionary describes the fields typically encountered by AQS
users and are listed in alphabetical order; field definitions and calculation
algorithms are provided as appropriate
o https://aqs.epa.gov/aqsweb/documents/AQS Data Dictionary.html
• U.S. EPA National Ambient Air Quality Standards (NAAQS) Table
o A webpage detailing the NAAQS for six criteria pollutants which includes
details from Table 2-4 but is a resource that will be updated if the standards
change
o https://www.epa.gov/criteria-air-pollutants/naaqs-table
• The National Institute for Occupational Safety and Health (NIOSH)
o The Occupational Safety and Health Act of 1970 established NIOSH as a
research agency focused on the study of worker safety and health, and
empowering employers and workers to create safe and healthy workplaces
o https://www.cdc.gov/niosh/index.htm
• Center for Disease Control and Prevention (CDC) National Environmental
Public Health Tracking - Air Quality
o CDC works closely with the U.S. Environmental Protection Agency, the
National Aeronautics and Space Administration (NASA), the National Oceanic
and Atmospheric Association (NOAA), and the National Weather Service to
provide air quality data on the Tracking Network and to better understand how
air pollution affects our health
o https://www.cdc.gov/nceh/trackinq/topics/AirQuality.htm
34
-------
• Greenbook: Nonattainment Areas for Criteria Pollutants
o The EPA Green Book provides detailed information about area National
Ambient Air Quality Standards (NAAQS) designations, classifications and
nonattainment status
o https://www.epa.gov/green-book
2.5 The Air Quality Index (AQI)
As previously described, air pollution can have a number of serious health impacts. For all
criteria pollutants except Pb, U.S. EPA has established the AQI. The AQI scale runs from 0
to 500, with higher AQI values indicating greater levels of air pollution and associated
health concerns (Table 2-5). For example, an AQI value of 50 or below represents "Good"
air quality, while an AQI value over 300 represents "Hazardous" air quality.
Table 2-5. The Air Quality Index (AQI) Levels of Health Concern, Numerical Values,
and Meanings
Daily AQI
Color
Levels of
Concern
Values of
Index
Description of Air Quality
Green
Good
Oto 50
Air quality is satisfactory, and air
pollution poses little or no risk.
Yellow
Moderate
51 to 100
Air quality is acceptable. However,
there may be a risk for some people,
particularly those who are unusually
sensitive to air pollution.
Orange
Unhealthy for
Sensitive Groups
101 to 150
Members of sensitive groups may
experience health effects. The general
public is less likely to be affected.
Red
Unhealthy
151 to 200
Some members of the general public
may experience health effects;
members of sensitive groups may
experience more serious health
effects.
Purple
Very Unhealthy
201 to 300
Health alert: The risk of health effects
is increased for everyone.
Maroon
Hazardous
301 and
higher
Health warning of emergency
conditions: everyone is more likely to
be affected.
The AQI is divided into six color-coded categories, as shown in Table 2-5, with each
category corresponding to a different level of health concern. The color coding allows the
public to quickly determine whether air quality is reaching unhealthy levels in their
communities. The specific concentration breakpoints for each of the six levels vary by
pollutant.
35
-------
For each pollutant, an AQI value of 100 generally corresponds to an ambient air
concentration equal to the level of the short-term NAAQS for protection of public health.
AQI values at or below 100 are generally considered to be satisfactory. AQI values above
100 are considered to be unhealthy for sensitive groups of people (see Table 2-6), then
unhealthy for everyone as AQI values reach higher levels.
Table 2-6. Pollutant-Specific Sensitive Groups for the AQI Greater than 100
(Additional information available on the AirNow website)
Pollutant
At-Risk Populations
Carbon Monoxide (CO)
People with heart disease
Nitrogen Dioxide (NO2)
People with asthma, children, and older adults
Ozone (O3)
People with lung disease, children and teenagers, older
adults, people who are active outdoors (including outdoor
workers), people with certain genetic variants, and people
with diets limited in certain nutrients
Particulate Matter (PM2.5)
People with heart or lung disease, older adults, children, and
people of low socioeconomic status
Particulate Matter (PM10)
Sulfur Dioxide (SO2)
People with asthma, children, and older adults
U.S. EPA calculates the AQI values based on air pollution data that is averaged over 1, 8,
or 24 hours, depending on the pollutant (see Table 2-4). The reason for the different
averaging times is that different pollutants affect the human body in different ways. For
example, O3 can cause coughing, sore or scratchy throat, inflamed airways, or difficulty
breathing within hours to a day after exposure. On the other hand, SO2 can cause
breathing difficulty, wheezing, and chest tightness within 5 minutes. The variation in
symptom offset is due to the specific way that the body reacts to exposure.
The AQI uses a formula to convert the averaged measurements (e.g., 24-hour average for
PM2.5, 8-hour average for O3) to the corresponding AQI value. For real-time data, the U.S.
EPA developed a method to estimate AQI from short-term averages called the NowCast
AQI. The NowCast AQI shows air quality for the most current hour available by using a
calculation that involves multiple hours of past data. The NowCast AQI uses longer
averages during periods of stable air quality and shorter averages when air quality is
changing rapidly, such as during a wildfire event. Easy-to-use online calculators are
available that allow users to either calculate AQI values from measured concentrations or
vice versa.
36
-------
Resources for More Information
• AirNow Air Quality Index (AQI) Website
o Provides information on AQI basics, air pollutants, action days, and other
resources
o https://www.airnow.gov/aqi/
• Technical Assistance Document for Reporting of Daily Air Quality - the Air
Quality Index (AQI), U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, EPA 454/B-18-007, September 2018
o Document provides guidance to aid local agencies in calculating and
reporting the AQI as required in the Code of Federal Regulations (CFR)
o https://www.airnow.gov/publications/air-gualitv-index/technical-assistance-
document-for-reportinq-the-dailv-aqi/
• AirNow AQI Calculator
o Online tool that converts user-specified AQI values into an equivalent
concentration or converts concentration into AQI values. The tool also
provides the corresponding AQI Category (e.g., good, moderate), health
effects, and cautionary statements
o https://www.airnow.gov/aqi/aqi-calculator/
• AirNow - Using the Air Quality Index Website
o Provides an overview of the AQI, AQI forecasts, and the NowCast AQI and
how to use these tools to assess local air quality and plan for outdoor
activities; links on the page provide technical information about NowCast
algorithms and leads to a github code library for calculating the NowCast for
03
o https://www.airnow.gov/agi/agi-basics/using-air-gualitv-index/
• Air Quality Index: A Guide to Air Quality and Your Health, U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, EPA-456/F-14-002,
February 2014
o Booklet that discusses the importance of air quality and provides an overview
of the AQI; health effects of exposure to ozone (O3), particulate matter (PM),
carbon monoxide (CO), and sulfur dioxide (SO2); and suggested actions to
reduce exposure to unhealthy air for each AQI Category
o https://www.airnow.gov/sites/default/files/2018-04/agi brochure 02 14 O.pdf
37
-------
• Air Quality Guide for Nitrogen Dioxide (NO2), U.S. Environmental Protection
Agency, Office of Air and Radiation, EPA-456/F-11-003, February 2011
o Booklet that overviews actions to reduce exposure NO2 near roadways for
each AQI category, provides an overview of NO2 sources and health effects,
and provides tips for reducing NO2 emissions
o https://www.airnow.gov/sites/default/files/2018-06/no2.pdf
• Air Quality Guide for Ozone (O3), U.S. Environmental Protection Agency, Office of
Air and Radiation, EPA-456/F-15-006, August 2015
o Booklet that overviews actions to reduce exposure to O3 for each AQI
category, provides an overview of O3 sources and health effects, and
provides tips for reducing pollution from O3
o https://www.airnow.gov/sites/defauit/fiies/2021-03/air-qualitv-
quide ozone 2015.pdf
• Air Quality Guide for Particle Pollution, U.S. Environmental Protection Agency,
Office of Air and Radiation, EPA-456/F-15-005, August 2015
o Booklet overviews the actions to reduce exposure to particle pollution for
each AQI category, provides an overview of pollution sources, and overviews
health effects and tips for reducing particle pollution
o https://www.airnow.gov/sites/default/files/2021-03/air-quality-
quide pm 2015.pdf
38
-------
Chapter 3
Monitoring Using Air Sensors
Air quality monitoring, whether deploying a single air sensor or a network of sensors,
requires planning. There are many steps involved in planning and conducted an air
monitoring study and this chapter walks through each of those steps and will include
information, considerations, and advice for each. This chapter focuses on monitoring
projects which use a sensor(s) in a stationary, outdoor environment. Additional
considerations may be important if you are using sensors In other applications (e.g., mobile
monitoring, personal exposure, indoor air monitoring).
This chapter provides an overview of the steps involved in planning a monitoring study
including:
• Question: determining a purpose for monitoring
• Plan: developing a monitoring plan including guidance for selecting a sensor to
support the plan which measures the target pollutant with the general features
required
• Setup: locating a monitoring site(s), install a sensor(s), designing a sensor
network(s), and planning and conducting a collocation to determine how to interpret
the sensor data
• Collect: reviewing data collection activities, common quality control and assurance
checks, and how a data management system can support these tasks
• Evaluate: analyzing, interpreting, communicating, and acting on your monitoring
results
39
-------
3.1 Planning and Conducting Air Monitoring
Air monitoring using sensors can be complicated and requires advance planning to be
successful. This planning is a critical component of quality assurance (QA) and is
necessary to produce useful and high-quaiity data. The planning steps and activities
outlined in this section enable users to collect quality data, build trust in the data, and allow
others to use the data, as applicable.
As shown in Figure 3-1, there are five key steps in planning for air monitoring. This
advance planning will save time and money and ensure that useful measurement data are
collected or that the purpose of the data collection is met. The amount of time spent on
each step in Figure 3-1 depends on your purpose. If deploying a single air sensor to seek
knowledge for educational purposes, you can quickly consider some of the details outlined
in this section to select, deploy, and operate your sensor. A more complex project requires
spending time on each step and addressing the topics and recommendations presented in
each section.
Quest/'o/i
\\
Analyze,
interpret,
communicate
results, or take
action
Determine your
purpose for
monitoring or
question you
seeking to
answer
o
our
- the
are
j Develop
an approach^p
to obtain data.
If needed, plan
measurements,
identify quality
control tasks, and
select sensors
ui
o
Collect
measurements,
review data, and
conduct
maintenance
Select sites, setup
and test sensors,
and check and
compare sensor
measurements
Figure 3-1. Five Steps Recommended for Planning Air Monitoring Projects Using Air
Sensors
40
-------
The recommended steps for
monitoring include the following:
1. Question. It is essential
to take the time to
establish and document
the question or
questions to be
answered by the air
monitoring study before
developing a plan to
collect measurements.
A simple question such
as, "Are ozone (O3)
concentrations in my
neighborhood higher during the afternoon than in the morning?", can help you
get started. See Section 3.2 for more guidance and tips on determining a
monitoring purpose.
2. Plan. With the question posed, develop a detailed plan for how to obtain the
measurements. The plan is the foundation for collecting quality data and provides
a roadmap with information about staffing, sensor selection and deployment,
data processing, data validation, quality control (QC) tasks, and more. See
Section 3.3 for details on the components to include in a plan. Section 3.4
describes how to select an air sensor.
3. Setup. To ensure useful results, you should select a monitoring location(s) to
measure the atmosphere or source of interest with minimal interference from the
surroundings (e.g., buildings, trees). A well-placed site is key for obtaining
representative data of the area being monitored. See Section 3.5 for tips on
placing air sensors. Setup will often include a site where sensors can be
operated side-by-side with a reference instrument, also known as collocation, to
collect data useful for checking sensor operation and for developing data
correction equations. See Section 3.6 for guidance on checking sensors via
collocation and data correction activities.
4. Collect. With a measurement plan clearly defined and sensors properly located,
it is time to collect data. Measuring air quality is not as easy as just turning on the
sensor and collecting measurements. Periodically checking sensors throughout
the course of a project is important to make sure they are functioning properly
and collecting quality data. See Section 3.7 for details on collecting, managing,
and quality-controlling data.
5. Evaluate. The approach for analyzing and presenting data is critical to
successfully communicating the results and, ultimately, achieving your
/ \
Do all Air Sensor Projects Need to Follow the
Five Steps in the Planning Wheel (Figure 3-1)?
No. Although most sensor projects would benefit
from following the five steps in the planning wheel,
there are some situations where sensor data is
used for educational or informational purposes
only. For example, you may be interested in using
a sensor to understand when the air quality is
better or worse in your neighborhood. An accurate
pollutant concentration may not be needed. This
type of use case does not require as much
planning.
41
-------
objective(s) for collecting air quality data. Section 3.8 provides details about
analyzing and interpreting data and communicating results.
You may find that during or after evaluating your data, some unanswered questions remain.
As shown in Figure 3-1. you may need to Revise/Update Your Question and adjust your
plan and monitoring activities accordingly. Quality Assurance (QA) includes all the steps
you perform to plan and manage the project and to collect, assess, and review data.
Completing these steps increases the likelihood of collecting credible and useful data.
3.2 Question: Determining a Purpose for Monitoring
It is essential for air sensor users to ask questions and to provide a clear monitoring goal
before beginning data collection. Asking what data may already exist and why new air
quality data are needed is important before purchasing an air sensor. Defining questions to
be answered will help identify the pollutant of interest, the field conditions likely to be
encountered, the data collection period, the type of measurements needed (e.g., short-term
vs. long-term, stationary measurements), and the desired quality of these measurements.
All of these data collection characteristics will determine the air sensor(s) best suited for
your purpose.
As you consider your monitoring purpose, you should also consider what you will do with
the information collected. Are there specific people, groups, organizations, or companies
with whom you will share your findings? Are there actions that you hope to inform and
inspire in yourself or others? What are some of those potential actions? These intentions
may shape your question, help you recruit team members, and inform your data quality
needs. Do not wait until data has been collected to determine how you will use it!
There are many purposes for monitoring with air sensors, including, but not limited to,
general interest in air quality, education, and participatory science engagement; identifying
an air pollutant of concern (e.g., hotspot identification); supplementing reference
instruments; and conducting research. Consider the following to help identify a question
that defines why monitoring air pollution might be needed:
• What is my air quality concern or suspicion?
• What do I already know about the air quality concern (being as specific as
possible about when, where, and what)?
• What is not known about the situation that I want to understand?
• What do I think may have caused the situation? What are other possible
causes?
• What are my desired outcomes for monitoring?
• Where are the nearest reference instruments and what pollutants do those
instruments measure?
• Do the pollutants measured by nearest reference instruments reflect the sources
that are of concern (see Table 2-1)?
42
-------
Developing a good question is a process of a well-defined path to achieving the desired
monitoring objective. Some users may start with curiosity or concern, a hunch, or suspicion
about air quality, as well as ideas about the outcome(s). A good question has the following
characteristics:
• It seeks to understand.
• It addresses a concern or suspicion.
• It can be answered using available resources (e.g., time, budget, skills).
You may need to modify the initial question to make it more specific and detailed as
possible, as shown in Figure 3-2. Several attempts may be necessary to develop a
question that has these characteristics. This upfront work helps ensure that the project
achieves the desired outcome.
"What is the air pollution
in my neighborhood?"
Better
"What time of day is
PM25 higher in my
neighborhood?
Does that vary by day
of the week?"
"When does PM25 in
my neighborhood
reach unhealthy levels
and from what
direction is the wind
blowing. What are the
potential causes of the
higher PM25?"
Figure 3-2. Example of Adding Details to Your Question or Objective
43
-------
Resources for More Information
• Handbook for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Handbook that covers common expectations for quality assurance and
documentation and best management practices for organizations that train
and use volunteers in the collection of environmental data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csqapphandbook 3 5 19 mmedits.pdf
• Guidebook for Developing a Community Air Monitoring Network: Steps,
Lessons, and Recommendations from the Imperial County Community Air
Monitoring Project, Tracking California, October 2018
o Outlines the process and considerations for creating an air monitoring
network using air sensors
o https://trackinacalifornia.org/cms/file/imperial-air-proiect/auidebook
• Community in Action: A Comprehensive Guidebook on Air Quality Sensors,
South Coast Air Quality Management District (South Coast AQMD), Air Quality
Sensor Performance Evaluation Center (AQ-SPEC), September 2021
o Guidebook for community organizations that covers planning for monitoring
using sensors; sensor deployment, use, and maintenance; and data handling,
interpretation, and communication
o http://www.aamd.gov/aa-spec/special-proiects/star-arant
• Air Sensor Stories, University of Rochester, University of North Carolina at Chapel
Hill, University of Texas Medical Branch, Columbia University, and WE ACT for
Environmental Justice, 2018
o Workshop guide and supporting materials to assist diverse audiences
understand the potential of air sensors in addressing community concerns
about particulate matter pollution; includes an air monitoring action plan
worksheet to help groups think through key questions
o https://www.urmc.rochester.edu/environmental-health-sciences/communitv-
engagement-core/proiects-partnerships/air-sensor-stories-workshop.aspx
• Appendix B: Questions to Consider When Planning for and Collecting Air
Sensor Data, and Sharing Your Results {this document)
o Provides a list of questions for consideration to help sensor users better plan,
collect, and share data
44
-------
3.3 Plan: Developing a Plan
A plan should include details on the pollutant(s) and environmental parameters [e.g.,
temperature (T), relative humidity (RH), wind speed, wind direction] to be monitored, where
the data will be collected, and how the data collection will be conducted. Whether mounting
a single sensor on the side of a house to monitor air pollutants in wildfire smoke or
deploying an air sensor network to assess air quality across a community (e.g.,
neighborhood, town, city, region) a detailed plan will help ensure that all tasks are
completed and all sensors and supporting instruments (e.g., reference monitors, weather
instruments) are collecting useful data for the desired application.
Creating a plan is a valuable process that can help minimize complications later.
Developing a specific monitoring plan will also allow air sensor users to share the project
design with others before investing time and money. Plans can vary in complexity as
needed for a project and the intended use of the data. If possible, air sensor users should
share their plan with an expert(s) (e.g.,
local air monitoring agency, university
professor, environmental consultant)
who is willing to provide ideas and
constructive feedback. Air sensor users
should also consider sharing their plan
with the audience for whom the results
are intended. A plan can help air
sensor users identify potential
problems at an early stage. Such an
effort is worthwhile because it is likely
that end users will need some or all of
this information to answer questions about measurements when presenting results.
As shown in Table 3-1. a plan can include many topics that encompass all steps for air
monitoring. If you deploy a single sensor, the plan could simply guide you to consider the
topics in the table. However, a detailed plan is necessary for more extensive air sensor
networks, typically involving more people, organizations, and resources. In this case, the
plan becomes a critical tool for success.
What is a Quality Assurance Project Plan
(QAPP)?
A QAPP is a written document that explains
how organizations ensure, using quality
assurance (QA) and quality control (QC)
activities, that the data collected can be used
for its intended purpose, A QAPP gives more
confidence that the data collected will meet
the project objectives and help others
understand the data quality.
45
-------
Table 3-1. Common Topics and Information Included in an Air Monitoring Plan
Topic
Information to Include
Purpose and Organizational Topics
Purpose for
monitoring
State the specific environmental topic/problem that is to be investigated,
the decision to be made, or the outcome to be achieved using the
sensor data. (See Section 3.2)
Project/task
organization
Determine the roles and responsibilities of all key players in the project.
Engagement with
local partners
Solicit insights from tribal/state/local/ air quality or health agencies,
universities, research organizations, or others. Engage them early and
discuss the proiect and desired outcomes. (See Appendix B)
Project/task
description
Summarize the work, objectives, schedule (timeline), and expected
outcomes.
Data quality
objectives and criteria
Define: 1) Why data are needed? 2) Does this data already exist? 3)
What measurements are needed and what do they need to represent?
4) Is there a certain level of accuracy needed? (See Section 3.2)
Contingency planning
Determine backup plans if something changes during the project (e.g.,
What happens if staff depart? How to deal with sensors that fail? What
happens if my site or equipment are vandalized?). Prepare for various
potential outcomes and plan troubleshooting plans.
Training and
experience
Identify any training and/or certification requirements (e.g., sensor
operation, programming courses through Coursera).
Documentation and
records
Determine how air monitoring activities will be documented. This could
include standard operating procedures (SOPs), quality
assurance/quality control (QA/QC) forms, site logbooks, etc.
Setup Topics
Measurement
methods
Describe equipment and measurement methods used in the monitoring
network. (See Section 2.3)
Siting criteria
Discuss the criteria for placing air sensors also considering site
security/safety. (See Section 3.5.1)
Monitoring location(s)
Discuss the monitoring location or locations (for a network) selected and
rationale. (See Section 3.5)
Instrument/equipment
testing and
inspection
Identify and describe how you will select the air sensor(s) and test and
inspect them to determine that they are workinq properly. (See Section
3J)
Instrument/equipment
calibration and/or
correction
Determine collocation locations and establish the calibration and/or
collocation and data correction methods. (See Section 3.6)
46
-------
Topic
Information to Include
Other data needed
Identify types of data that originate from other sources that may be used
in the analysis. These data could include nearby reference monitor data,
weather data, and/or traffic counts.
Collection Topics
Maintenance and
operations
List the methods or procedures that will be used to maintain and operate
air sensors, including site visits, routine maintenance, emergency
maintenance, daily data reviews, periodic collocation, etc. (See
Appendix C)
Quality control (QC)
Describe the tvpes of QC checks performed. (See Section 3.7.2)
Data processing and
access
Understand how the data are processed, stored, and adjusted. Decide
how vou will access the data and who will own the data. (See Section
3.7.3 and Appendix C)
Verification and
validation methods
Describe the methods or procedures that will be used to verify and
validate data during the collection period. (See Section 3.7.2)
Data management
Determine how the air monitoring data will be managed, tracing the path
of data generation in the field to the final data use and end storage. (See
Section 3.7.3 and Appendix C)
Evaluation Topics
Data analysis
methods and
visualization
Describe the methods or steps used to answer question(s) (e.g., data
processing needs, visualization software needs). (See Section 3.8)
Compare results with
original objective(s)
Describe how the results obtained from this project will be reconciled
with the proiect's data guality obiective(s). (See Section 3.2)
Evaluation,
communication, and
action
Describe how the results of the air monitoring project will be used. (See
Section 3.8)
47
-------
Resources for More Information
• Guidance for Quality Assurance Project Plans (QA/G-5), U.S. Environmental
Protection Agency, EPA/240/R-02/009, December 2002
o Provides guidance on developing a Quality Assurance Project Plan (QAPP),
which is an important part of the planning process for air quality monitoring
projects
o https://www.epa.gov/sites/default/files/2015-06/documents/g5-final.pdf
• Examples for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Collection of examples that provide tools and procedures to help community
science organizations properly document the quality of data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgappexamples3 5 19 mmedits.pdf
• Templates for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Templates that provide tools and procedures to help properly document the
quality of data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgapptemplates3 5 19 mmedits.pdf
o Editable templates: https://www.epa.gov/citizen-science/gualitv-assurance-
handbook-and-guidance-documents-citizen-science-proiects
• Community Science Air Monitoring
o Guidance, provided by the New Jersey Department of Environmental
Protection Division of Air Quality - Air Monitoring, on using air sensors for
community projects; includes approaches to using sensors, types of sensors
available, interpreting sensor data, four types of sensor projects and data
quality assurance plan templates for each, and other helpful links
o https://www.ni.gov/dep/airmon/communitv-science.html
48
-------
• Air Quality Agencies
o Websites that provide a list of state, local, and/or tribal agencies that manage
air quality
o U.S. Environmental Protection Agency: https://www.epa.gov/aboutepa/health-
and-environmental-agencies-us-states-and-territories
o National Tribal Air Association (NTAA). https://www.ntaatribalair.org/
o National Association of Clean Air Agencies (NA CAA):
https://www.4cleanair.org/agencies/
o Association of Air Pollution Control Agencies (AAPCA):
https://cleanairact.org/about/
3.4 Plan: Selecting an Air Sensor
Before purchasing an air sensor, evaluating specific criteria will help you match air sensors
to your application of interest and purpose for collecting data. Figure 3-3 provides an
overview of six important questions to ask before purchasing an air sensor. The key sensor
selection criteria are discussed in additional detail below.
49
-------
Six Questions to Ask Before You Buy
a Lower-Cost Air Sensor
GL0
O
What is the purpose?
• Education and information
• Hotspot identification
• Personal exposure
• Participatory science
What pollutant or pollutants do
you want to measure?
Particulate matter
• A gas (ozone, nitrogen
dioxide)
• Total volatile organic
compounds (VOCs)
What are some of the
features you should consider?
• Size, weight, and portability
• Demonstrated accuracy in the
real-world
• Weatherproof
• Power source
• Storage capacity and
wireless transmission
• Maintenance requirements
How can you check the performance
of your lower-cost air sensor?
• Compare results to a nearby
regulatory monitor
• Conduct periodic quality
control checks
Dr? j
V
• Check weather and other conditions
that may impact performance
• Periodically review and evaluate data
for errors or problems
How much do lower-cost air
sensors typically cost?
• $150-$ 1,500 (1-2 pollutants)
• $500-$2,500 (I-3 pollutants)
• $2,500-$ 10,000 (4 or more
pollutants or I pollutant)
What should you look for in a
user manual?
Type of pollutants measured
General operating instructions
How to store and recover data
Conditions of operation
Expected performance
Customer service support
$
lata
Learn more about how to select
and use an air sensor:
Air Sensor Toolbox —
https://www.epa.gov/air-sensor-toolbox
Air Sensor Guidebook —
https.7/www.epa.gov/air-sensor-toolbox/
how-use-air-sensors-air-sensor-guidebook
AIR
SENSOR
TOOLBOX
v>EPA
United States
Environmental Protection
Agency
Figure 3-3. Questions to Consider Before Purchasing an Air Sensor
50
-------
3.4.1 Target Pollutant and Sensor Performance
Selecting a target pollutant. The target pollutant(s) to be measured by the sensor will
depend on the question asked and the purpose for monitoring. Consult Tables 2-1 and ^2,
which identify common sources and health effects of various pollutants. It is important to
keep in mind that a sensor's cost may depend on the types and number of pollutants
selected. For each target pollutant, consider the other factors below (e.g., detection limit,
measurement range, accuracy) to determine if a sensor will meet your monitoring needs.
/ \
How Do Air Sensors Work?
Sensor design and performance is an active area of research and innovation. New
methods may be introduced over time.
PM (PM-io, PM2.5) sensors: Currently, PM is typically measured using an optical
approach where light scattered by a particle(s) is used to estimate the particle mass
concentration. The amount of light scattered can vary due to the size, shape, and
chemical composition of the particles. This method can only detect a narrow range of
particle sizes. Over time, particles may build up within the sensor causing changes in
performance. Lifetimes vary but are often between 1 to 4 years. PM sensors are typically
useful for outdoor, indoor, and smoke monitoring applications. Currently, many sensors
will not detect PM10 or dust.
Gas (O3, NO2, CO, SO2) sensors: Currently, most gas pollutants are measured using
either electrochemical (EC) or metal oxide sensors (MOSs). They may respond to the
target pollutant, changes in T or RH, or other interferent gases. Sensors may lose
sensitivity over time whether in use or not and should be collocated at least seasonally.
Lifetimes may be between 6 months and 2 years. Measurements at low concentrations
are often difficult and the most successful application for some sensors may be near
sources.
Total VOC (tVOC) sensors: Currently, most tVOC sensors are MOS or photoionization
detector (PID) sensors. They detect a wide variety of VOCs but do not measure all VOCs
and are more responsive to some compounds than others. Therefore, the reported
concentration cannot be attributed to a specific compound or even the sum of several
compounds. Sensors may also respond to changes in T, RH, other interferent gases.
Currently, the most successful applications match upwind and downwind measurements
near a source to look for spikes which may indicate a VOC emission event. Lifetimes
may be months to a year for PID sensors and longer for MOS.
Measurement range and detection limit. Air pollutants can often be present at very low
or very high concentrations in the ambient air. The measurement range refers to the lowest
and highest pollutant concentrations that a device can measure. A sensor will be most
useful when it measures a target pollutant over the full range of concentrations commonly
found in the atmosphere (consult Table 2-1 under "Range to Expect" for each pollutant).
Depending on proximity to a pollution source, the sensor's ability to measure either very
low or very high concentrations is essential. The detection limit is the lowest concentration
51
-------
of a pollutant that a device can routinely detect. It is important to consult the manufacturer's
specifications for the detection limit to ensure that the air sensor can measure lower
concentrations that are typically found within the U.S.
Sensor accuracy. Accuracy describes the agreement between the sensor's pollutant
concentration measurement and the concentration measured by the reference instrument.
The accuracy of a sensor* is determined by two components: precision and bias. Precision
refers to how well a set of sensors reproduces the measurement of a pollutant under
identical conditions (e.g., same concentration and temperature). Bias refers to
measurement error. For example, a sensor may always measure a little higher or lower
than the true concentration. Figure 3-4 shows example plots illustrating accuracy, precision,
and bias. Before purchasing a sensor, buyers should evaluate the air sensor's precision
and bias using the manufacturer's specifications, evaluation reports, and published
literature (further discussion in Chapter 4). Also, users should conduct checks of the
precision and bias to qualify the air sensor's accuracy. See Sections 3.6.1, 3.6.2, and
Chapter 4, which describe how to perform these checks through collocation.
ID
DC
o
CO
2
m
(O
CO
Precision
LESS
MORE
time ~
More bias and less precision:
Sensor 1 overestimates the concentration (positive
bias), Sensor 2 underestimates the concentration
(negative bias). Sensors 1 and 2 do not follow the
same pattern (poor precision).
time ~
More precision but more bias:
Sensors 1 and 2 agree well (good precision) but
overestimate the concentration (positive bias).
time ~
Less bias but less precision:
Sensors 1 and 2 do not follow the same pattern, nor
agree with the truth (poor precision).
time ~
Accurate! Less bias and more precision:
Sensors 1 and 2 agree well (good precision) and
are very close to the truth (low bias).
Reference
Instrument
Legend
Sensor 1
Sensor 2
Figure 3-4. Illustration of Air Sensor Bias, Accuracy, and Precision
52
-------
Users should be aware that a sensor's accuracy, precision, and bias can change over time.
For example, exposure to warm temperatures or humid air may lead to a gradual increase
in sensor bias. Ozone (O3) sensors, in particular, can become less sensitive over time
either because of age or fluctuations in temperature and humidity. Particulate matter (PM)
sensor bias may change if pollutant source, particle type, or particle size changes. The
sensor may also experience interference from other pollutants in the atmosphere, leading
to inaccurate concentration measurements (see Section 3.7.2). Some sensor
manufacturers provide an expiration date, after which the sensor measurements are no
longer likely to be accurate.
Air sensor data are called "noisy" when the data fluctuates from one value to another.
Noise in the data can be caused by multiple factors including electrical interferences,
sensor precision, rapid weather changes, and averaging period. To reduce the noise, air
sensor data are often averaged over longer time periods to make it easier to see trends in
the data. For example, 1 -minute PM2.5 data can oscillate rapidly from one minute to the
next. Averaging 60 of these 1- minute data points to create a 1-hour average will reduce the
noise in the data. Figure 3-5 shows an example of noisy versus less noisy data.
Figure 3-5. Example of Noisy Measurement Data
Calibration or collocation and data correction. Calibration is a procedure that checks
and adjusts an instrument's settings so that the measurements produced are comparable
to a certified standard. Collocation is the process of checking the performance of an air
sensor by installing and operating a sensor in close proximity to a reference instrument(s).
Data correction involves adjusting the air sensor data to increase its accuracy relative to a
known reference value. Section 3.6 provides guidance on calibration, collocation, and
correction. Before purchasing a sensor, users should determine whether the manufacturer
53
-------
has calibrated or corrected the sensor. In addition, users should fully understand when and
how collocation should be performed and how to correct the air sensor's measurements.
Talk to the sensor manufacturer about the method, frequency, and any additional costs for
the calibration or collocation and correction services.
Response time. A sensor may be quick or slow to detect changing pollutant
concentrations in the air. A sensor that responds quickly (i.e., high time resolution) may be
useful for mobile monitoring and observing very rapid (e.g., seconds to minutes) changes in
pollutant concentrations at fixed sites (such as near roadways with heavy traffic), as shown
in Figure 3-6. A sensor that responds slowly (i.e., low time resolution) may be more suited
to stationary monitoring where pollutant concentrations often change more gradually (e.g.,
minutes to hours). Specific data collection goals and purposes will determine which type of
sensor is best.
Figure 3-6. Example of an Air Sensor's Response Time
3.4.2 General Features of a Sensor
Durability. Sensors vary in size, shape, durability, and quality of construction. Durability
refers to an air sensor's ability to be shipped, moved, and to endure wear and tear and
continue to perform. For example, sensors are often tossed around during shipping and
some components or parts could dislodge inside causing anything from communication or
data logging issues to complete destruction. Sensors that are worn by the user or are
deployed for mobile monitoring on vehicles might be shaken, hit against other objects, or
dropped and must be designed to handle these impacts. All sensors measuring outdoor air
quality are likely to be exposed to variable weather conditions such as wind, heat, cold,
moisture, and dust and should be built to handle this exposure. A user manual or
manufacturer's specification sheet should provide details on the general durability of the
sensor.
Fast-response sensor
Slow-response sensor
\
\
\
Time
54
-------
Enclosure. An enclosure is a case or structure that contains the sensor and its
components and protects the components from water, light, temperature variations (e.g., by
adding heaters or cooling fans), and electromagnetic noise. The sensor enclosure must
allow air to reach the sensing components while shielding the components from weather
effects. The materials, design of the enclosure, and sensor orientation (e.g., air inlet
location, air flow path) may affect measured pollutant concentration levels and response
time. For example, certain types of plastics and coatings might react with the pollutant of
interest or release the pollutant, interfering with the air sensor's measurements. In addition,
the enclosure may impact the internal T or RH, potentially also impacting the air sensor's
measurements. Sensors that are exposed to ambient conditions for an extended period of
time may experience a build-up of dust, dirt, ice/snow, and other debris near the sensor
inlet. This may alter the accuracy and bias of the sensor, and users should ensure that a
sensor's inlet remains clear of obstruction.
Ease-of-use. A wide variety of people with different levels of experience may use an air
sensor and it is important to understand how easy or difficult it is to operate a sensor.
Everyone, especially less experienced users, appreciate sensors that are easier to use.
Determine whether any special expertise (e.g., technician, programmer) or tools (e.g.,
ladders, computers with specific software, special screwdrivers) are needed to operate or
maintain the sensor both in the short term and long term.
Power. Power requirements vary for sensors and include plug-in, battery, or solar power.
The choice of power options will depend on a user's application. Some sensors may alter
their sampling frequency depending on the type of power supply used. This may result in
some sensors logging data at intervals spaced longer apart when configured for battery or
solar-powered operation. Plug-in devices are best suited for stationary monitoring
applications with access to a wall power outlet; however, users need to ensure that power
is available and easily accessible at the installation site. Battery-powered devices are often
suitable for mobile applications or short-term data collection activities, although users
should be aware of how long the battery lasts after charging and at what point the charge is
too low to fully operate the sensor. For solar-powered devices, users should consult the
manufacturer to ensure proper sizing of the solar components for the device and the
available sunlight at the monitoring location (e.g., latitude, longitude, season) and
information about proper placement, orientation, and maintenance. If you need to design
your own solar option, some commercial solar kits may be available which contain solar
panels and batteries. Consult the supplier to learn more about proper sizing, including a
margin of error for hours without adequate sunlight (e.g., cloudy or smoky days, overnight).
Keep in mind that a solar power solution can be costly.
Display. Some sensors do not include a data display and require users to visit a website or
use an app to view data instead. Others feature a screen or display allowing users to view
sensor information, real-time data, and/or view historical data. Some sensors include lights
which indicate power or may change color depending on pollutant concentration and these
lights may be paired with a data screen or be the only form of display. Consider whether a
display is necessary for your project.
55
-------
Data transmission. There are several options available for data transmission. Options
vary from sensor to sensor and include, but are not limited to, cellular, WiFi, Bluetooth,
satellite, and low-power wide-area network (LoRa). Some sensors store data on the unit
itself (e.g., local on-board storage, memory card) and data must be transferred manually.
When selecting sensors, users should consider their application and whether the sensor's
data transmission methods will suit their needs, cost constraints (e.g., subscription costs
associated with cellular services), and will work in their desired monitoring location.
Data access. There are a variety of data storage options available that may influence data
access options. Sensors with on-board data storage require physical data download. Other
sensors communicate data to central servers and data can be accessed by remote
download or call from an Application Programming Interface (API). Users should consider
how the data can be accessed, who has permission to access, who has data ownership
rights, and how long the data will be available. Once data can be accessed, users will need
to fully understand the data format, data analysis, and visualization options. For devices
that share information with the public, carefully consider what information is shown as there
may be privacy concerns (e.g., sharing a specific address).
Data handling. Conversion of information from the raw sensor signal to the final reported
pollutant concentration happens in a variety of ways but often involves some kind of
mathematical equation or model. These methods may depend only on data collected on-
board the sensor or may include other data (e.g., nearby weather station). Users should
ask manufacturers to describe how data is processed and any of these other data
dependencies to understand whether that data will be available in their study area. Sensor
manufacturers may choose to make their data handling methods public or keep them
proprietary.
Cost. A sensor's cost may vary greatly depending on the pollutant measured and degree of
accuracy and sensitivity needed. Even for sensors measuring the same pollutant, the costs
can vary depending on the device's features. Some sensor manufacturers offer different
purchasing options, including buy, lease, or rent. Users should be aware that there are
upfront costs (e.g., purchasing the sensor and sensor components) and long-term costs
that can include, but are not limited to, repair or replacement of the sensor and their
components, calibration services, data transmission charges (e.g., cellular service), or data
hosting and storage fees (e.g., cloud storage on a manufacturer server or other server).
Additionally, other potential costs (or time) could include data analysis, interpretation, and
communication of air sensor data. Of course, costs increase if more sensors are needed
(e.g., sensor networks).
56
-------
Resources for More Information
• Chapter 4: Sensor Performance Guidance (this document)
o Provides an overview of laboratory and field sensor performance evaluations;
performance characteristics needed for spatiotemporal variability,
comparison, and long-term trend NSIM applications; and U.S. EPA's
recommendations for sensor testing protocols, performance metrics, and
targets
• Appendix C: Choosing Air Sensors (this document)
o Provides checklists for: (1) what to look for in a sensor before buying, (2) what
to look for in a sensor user manual, and (3) sensor maintenance to ensure
proper functionality and reliable performance
• Performance Testing Protocols, Metrics, and Target Values for Ozone Air
Sensors - Use in Ambient, Outdoor, Fixed Site, Non-Regulatory Supplemental
and Informational Monitoring Applications, U.S. Environmental Protection
Agency, EPA/600/R-20/279. February 2021
o Provides recommended testing protocols (field and laboratory), performance
metrics (parameters used to describe sensor data quality), and target levels
to evaluate ozone air sensors
o https://cfpub.epa.gov/si/si public record Report.cfm?dirEntrvld=350784&Lab
=CEMM
• U.S. EPA's Performance Targets and Testing Protocols Website
o Summary of the U.S. EPA's research on recommended testing protocols,
metrics, and target values for evaluating the performance of air sensors
o https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-
testing-protocols
• Air Quality Sensor Performance Evaluation Center (AQ-SPEC) of the South
Coast Air Quality Management District (South Coast AQMD) Website
o Website for the AQ-SPEC program which conducts laboratory and field
evaluations of air sensors and provides information to the public regarding
actual sensor performance and the advantages and potential limitations of
using air sensors. AQ-SPEC is operated by South Coast AQMD
o http://www.agmd.gov/ag-spec
57
-------
• The National Solar Radiation Data Base (NSRDB), Sengupta, M., Y. Xie, A.
Lopez, A. Habte, G. Maclaurin, and J. Shelby. Renewable and Sustainable Energy
Reviews 89 (2018): 51-60
o Paper reviews the complete package of surface observations, models, and
satellite data used for the NSRDB - an open dataset of solar radiation and
meteorological data over the United States and regions of the surrounding
countries
o https://www.sciencedirect.com/science/article/pii/S136403211830087X
3.5 Setup: Locating Sites for Air Sensors
Finding locations to set up air sensors, whether a single
air sensor or a network of sensors (and other
instruments), is a critical task. Finding suitable sites
enables air sensors to collect useful data representing
the surrounding conditions, ensures the sensor has
power (and internet access, if needed), provides security
for the sensor, allows easy access for maintenance, and
adds credibility to the data.
Section 3.5.1 provides recommendations for placing an
air sensor at a specific site. Section 3.5.2 provides
information on designing a network of air sensors (i.e.,
multiple air sensors placed at different locations). This
entire process, including evaluating logistics like site
access and security, can be time consuming so users
should start early to find suitable locations.
/ \
What is an Example of How you Would Select Monitoring Locations Based on Your
Purpose for Monitoring?
Let's say your study question is "What are the typical particulate matter concentrations in
my city? Where are concentrations highest?"
Here are some potential monitoring locations based on the study question:
• Where people live/work
• Near locations where many people gather
• Near susceptible and sensitive populations (e.g., schools, hospitals)
• Between gaps in an existing regulatory monitoring network
• A background site (e.g., an area not impacted by the suspected pollution source)
• Near the actual (or expected) maximum concentration
• At or near emissions source(s) of concern
• Upwind/downwind of emissions source(s)
• Next to a site with reference monitors for collocation/correction activities
V V
58
-------
3.5.1 Installing Air Sensors
Users should carefully place a sensor or instrument in a location where it can reliably and
safely measure the ambient air or source of interest with minimal interference from the
location's surroundings. A well-placed sensor yields data that are representative of the air
quality in the area being monitored.
Air pollution concentrations can be affected considerably by local sources (e.g., fire pit,
grill), buildings, and structures, among other factors. These factors may vary based on the
target pollutant or monitoring goal and users should consider the potential effects of these
factors when choosing a monitoring location. The data will be most useful if the sensor can
measure the pollutant of interest with little impact from other sources at the site. Figure 3-7
provides eight key considerations when placing an air sensor at a specific location.
9
GENERAL LOCATIONS
Considering the purpose of the measurements,
where should the site be located?
6
SECURITY
Ensure that the sensor is secure and protected
from vandalism or theft and that site
operators can remain safe.
Determine the requirements (permissions, keys, etc.)
to visit the site to install/service the sensor(s).
PLACEMENT
In general, place the sensor at least 6 feet above
ground level, rooftop, or other objects and away from
obstructions, vegetation, or emission sources that
would interfere with the measurement.
©
Determine requirements and establish power early.
Take photos of the sensor installation
and area around the site.
COMMUNICATIONS
DOCUMENTATION
Ensure reliable communications
(cellular, Wi-Fi, etc.) before installation.
| Record information about the site, including latitude,
longitude, elevation, nearby obstacles,
date of installation, etc.
Figure 3-7. Logistical Considerations and Tips for Installing an Air Sensor
Location. Before setting up a sensor, it is useful to consider your monitoring goals since
they can impact your ideal location selection. For example, a sensor that will be used to
monitor for emissions from idling buses may be setup in a different location than one used
to estimate the local ambient air quality index (AQI).
Access. Although easy to use, air sensors are generally not something you can "set up
and forget." You will want to access your site to install and periodically check on the sensor.
If you do not control the site, you will want to determine permissions, access requirements,
and any limitations on access frequency or timing during the planning stage. Some users
have found formal access agreements helpful in explicitly defining these conditions.
59
-------
Power. Air sensors may need to be plugged in, may have solar panels, or may offer both
options. Some sensors that offer power options may operate differently depending on
which option is used (e.g., the data reporting frequency may change). Be sure to consult
the sensor manufacturer to understand the implications. It can be expensive and time
consuming to deliver power to a location that does not have the existing infrastructure.
Available outlets should be tested rather than assuming they work. Consider using a surge
protected power strip so that others can also use the outlet without unplugging your sensor.
Extension cords may be needed for optimal sensor placement safety (e.g., trip hazard, fire
risk). Water and electricity don't mix so be sure to consider electrical safety and water
proofing for all connections. Solar panels may not be adequate if your location does not get
enough sun and they will need periodic maintenance to remove dust. Areas that experience
public safety power shutoffs may benefit from solar power to prevent monitoring
interruptions.
Communications. Sensors may communicate data to a cloud-based interface using a
variety of technologies (e.g., cellular, WiFi, LoRa). Some sensors may offer just one option,
while other sensors may provide multiple options. Be sure to consult the manufacturer to
understand specific requirements such as network limitations (e.g., 2G, 5G), carrier
limitations (e.g., Verizon, AT&T), area coverage (U.S. and international), and signal
strength needs. If supplying your own mobile hotspot,
you may also want to know the typical data use to
properly estimate costs and if the sensor settings can
be adjusted to reduce data use.
Security. Sensors and their peripheral equipment (such
as solar panels) are subject to tampering and theft. A
small sign describing your project and the device may
help. Users will want to consider placing sensors in
secure locations. Ideas include mounting a sensor
overhead out of arms reach, in an inconspicuous
location, or behind a locked gate or fence. When
considering secure locations, keep in mind that sensors
need a free flow of air, and consider your physical
safety when visiting the area or even while climbing a
ladder or stepstool for installation or maintenance.
Placement. It is ideal to place sensors near the typical
breathing zone height (3-6 feet). Sensors should be
placed away from pollutant sources (e.g., fire pit, grill)
or pollution sinks (e.g., trees, shrub barrier) to get a more representative measure of air
quality within the local area. Sensors should also be located to allow for free air flow to the
sensor. Avoid placing sensors near high voltage power lines, which may create electronic
interferences. Consider what hardware might be needed to mount the sensor (e.g., tripods,
poles). Note that some locations (e.g., on top of buildings) may have specific engineering
requirements to withstand wind, etc.
/ \
Tip: Also consider time-based
factors when locating
monitoring sites
Sensors are commonly used to
monitor air quality before and
after implementation of
emission reduction programs
(e.g., anti-idling campaigns),
changes in land use, and
changes in industry. Don't
forget that collecting
measurements before AND
after will be useful to
understand air quality impacts.
Also consider collecting
measurements during different
seasons (e.g., summer vs.
winter).
V /
60
-------
Photos. Photos of the sensor deployment may assist you with data interpretation later. Be
sure to photograph nearby features that may impact the sensor readings. Outdoors, this
may be nearby buildings, roadways, or landscapes. Indoors, this may be building features
like windows, doors, and exhaust vents. The photos should also capture the typical use of
an area or room where the sensors are placed. Make sure to have written or verbal consent
when taking photos of anyone in the community, especially children.
Additional Documentation. A deployment log can assist you in recording notes about
sensor placement (e.g., location, height, date of installation) and maintenance (e.g.,
cleaning, component replacement). It's easiest to track or tag this information by assigning
each sensor an ID (e.g., serial number, user given name). You may also want to capture
more information about how the area is used. Also consider that temporary activities (e.g.,
road work, construction activities, cleaning, cooking) may impact the area and confuse data
interpretation, so keep notes while the sensor is in use.
3.5.2 Specifics for Designing a Network of Air Sensors
An air sensor network is made up of two or more sensors placed at several different
locations in an area to gain more information about variations in pollutant concentrations.
Examples of a network include deploying air sensors throughout a neighborhood to gather
general knowledge of air pollution levels or, designing a monitoring network to locate the
potential source of pollution impacting a location.
Selecting general locations for an air sensor network may seem daunting at first. This
section outlines considerations to help users design a successful and purposeful sensor
network.
Start by answering some questions about why and where to collect measurements, such
as:
• What types of changes in air quality do I expect in the area?
• How do I expect one site to be different than another?
• Where should I put sensors to measure and show these differences?
• What is the typical or prevailing wind flow in the area and how might winds transport
pollutants?
• Do I need meteorological data to interpret the air quality data?
• Could I use data from existing air monitoring networks?
• What is my budget for air sensors and approximately how many can I afford? (Some
air sensors may fail, so plan for extras or replacements.)
-------
Identify the general locations on a map to place air sensors. Consider the following tips:
• Spread out the deployment locations to get good spatial coverage.
• Avoid hyperlocal sources (e.g., smoking stations, grills) and locations where winds
can channel and trap pollutants unless that is your specific research question.
• If there is an area of concern (e.g., pollutant source, area of suspected higher
concentrations), locate sensors near/inside the area of concern AND upwind and
downwind of the area (see Section 2.1 and Figure 3-8) so that meaningful
comparisons can be made.
• Account for factors that affect safety when installing and maintaining the sensors
that include access to facility, security, signage, weather conditions (e.g., lightning),
etc.
• Locate a reference instrument for future collocation activities (see Section 3.6).
Seek input from your local air quality agency, a university professor, environmental
consultant, or other experts who are useful resources to help design effective sensor
networks.
• Contact your tribal, state, or local air quality agency.
• Contact professors in academic institutions with expertise in air quality such as
environmental studies, engineering, atmospheric science, or other sciences.
Air sensor networks may have different scales depending on the purpose for collecting
measurements. Figure 3-8 shows examples of air sensor networks of different scales for
three examples.
62
-------
b. Neighborhood
©0©
Collocation
with Reference
Downwind
Prevailing Upwind
Winds
Downwind
Upwind
1 mile
c. Microscale
Downwind
Prevailing
Winds
Background
Upwind
I Downwind
©0
Collocation
with Reference
Downwind
' i
100 feet
©AirSensor 0
Figure 3-8. Example Maps for Placing Air Sensors for Networks of Different Scales
Depending on the Purpose: a) Regional/Urban Network, b) Neighborhood Network,
and c) Microscale, Small Area Network
in each panel, Figure 3-8 identifies an optimal sensor collocation site. Generally, collocation
is recommended for any monitoring activity to make sure the air sensor is reporting
accurate data (see Section 3.6.1 for more information). The examples for the different
scales of sensor networks and potential collocation approaches are described below.
Figure 3-8a shows an example of using an air sensor to fill in an air monitoring gap. This
regional and urban network consisting of reference instruments located in a larger city, an
industrial area, and a background site. Air sensors can fill in the gap in this network by
providing air quality information for a small town that does not have existing reference
monitors. In this case, the air quality at the background site is most similar to that of the
small town experiencing similar meteorological conditions (e.g., T, RH), pollutants, and
pollutant concentrations. The pollution sources would be different, and the pollutant
concentrations would likely be higher, at the other two reference sites. Thus, in this
example the collocation is conducted at the background site collocating a sensor with the
reference instrument at that location.
63
-------
Figure 3-8b shows an example sensor network for a neighborhood or residential area
concerned about pollutant emissions from a nearby facility. The network measures air
quality around the facility (both upwind and downwind) and in the residential area located
downwind. The network also includes two air sensors collocated next to a reference
instrument. The collocation provides data to assess the accuracy of the sensors and
develop corrections for the other sensors in the network (see Section 3.6 for more details
about collocation and correction). The two collocated sensors provide replicate
measurements which allows for some additional quality assurance because the
measurements can be compared.
In Figure 3-8b, users can potentially determine the impact on air quality due to the facility of
concern by comparing the upwind and downwind pollutant concentrations. For this type of
example network, knowledge of the prevailing wind direction is critical and users will likely
need to place air sensors at several locations to make upwind and downwind
measurements under a variety of wind conditions. Locations upwind of the facility represent
the background air quality conditions before the air flows over the area of concern. The
locations downwind of the facility represent the air quality after winds pick up pollutant
contributions from the facility. In this case, wind speed and wind direction measurements
would be helpful for data interpretation.
Figure 3-8c shows an example of a microscale (small area) network using air sensor
measurements to assess the contributions to air pollution due to vehicle traffic at an
intersection. A background sensor measures the general air quality in the area outside of
the intersection of concern. An upwind air sensor measures the air quality before the air
flows through the intersection, and downwind air sensors measure the air quality after the
intersection. Comparing the upwind and downwind concentrations allows users to evaluate
the traffic's contribution to air quality. As with the neighborhood network in Figure 3-8b,
users will need to consider the direction of prevailing winds to determine appropriate
background, upwind, and downwind sensor locations. In this case, sensors can be
collocated at a nearby reference site which experiences similar pollutant concentrations
and meteorological conditions and is also impacted by nearby traffic sources.
For networks where the objective is to compare sensor measurements collected at different
sites, it is critical to first evaluate the precision, or agreement of the measurements made
by all the sensors. This can be accomplished by collocating all air sensors to evaluate the
precision and bias of each air sensor's data and then using that relationship (e.g., equation)
to apply correction(s) to all the sensors used in the network so that the data from all of the
sensor are more comparable (see Section 3.6.2 for more information including additional
considerations and tips). By taking this one step further and collocating the air sensors with
a reference instrument, a correction equation can be developed to make the sensor data
more comparable with the reference data thereby allowing comparisons to be made among
both the sensor and reference instruments.
64
-------
In summary, a wide range of networks can be designed to provide general air quality
information in an area, collect data for specific research questions, and more. Collocation is
highly recommended to make sure the air sensor(s) is reporting accurate data.
Resources for More Information
• U.S. Code of Federal Regulations (CFR), Title 40 (Protection of Environment),
Chapter 1 (Environmental Protection Agency), Subchapter C (Air Programs),
Part 58 (Ambient Air Quality Surveillance)
o Specifies the regulatory requirements for the U.S. ambient air quality
monitoring network including quality assurance procedures for operating air
quality monitors and handling data; methodology and operating schedules for
monitoring instruments; criteria for siting monitoring instruments; and air
quality data reporting requirements
o https://www.ecfr.gov/current/title-40/chapter-l/subchapter-C/part-
58#ap40.6.58.0000 Onbspnbspnbsp.e
• Quality Assurance Handbook for Air Pollution Measurement Systems, Volume
II, Ambient Air Quality Monitoring Program, U.S. Environmental Protection Agency,
EPA-454/B-17-001, January 2017
o Handbook provides additional information and guidance (including pollutant-
specific spatial scale characteristics) to assist tribal, state, and local
monitoring organizations in developing and implementing a quality
management system for the Ambient Air Quality Surveillance Program
described in 40 CFR Part 58
o https://www.epa.gov/sites/default/files/2020-
10/documents/final handbook document 1 17.pdf
• Air Quality Agencies
o Websites provide a list of state, local, and/or tribal agencies that manage air
quality
o U.S. Environmental Protection Agency: https://www.epa.gov/aboutepa/health-
and-environmental-agencies-us-states-and-territories
o National Tribal Air Association (NTAA): https://www.ntaatribalair.org/
o National Association of Clean Air Agencies (NACAA):
https://www.4cleanair.org/agencies/
o Association of Air Pollution Control Agencies (AAPCA):
https://cleanairact.org/about/
• Blueprint for the Development and Implementation of Distributed Sensor
Networks, U.S. National Institute of Standards and Technology Global Cities Team
Challenge Transportation SuperCluster
65
-------
o Blueprint that summarizes lessons learned, best practices, and research
questions for developing and implementing sensor networks
o https://static1 .sauarespace.com/static/5967c18bff7c50a0244ff42c/t/5ad7c41 c
758d464041c7e58a/1524089886422/Distributed Sensor Networks Recomm
endations.pdf
• U.S. EPA Guide to Siting and Installing Air Sensors
o Information and considerations for locating an air sensor in both outdoor and
indoor locations
o https://www.epa.aov/air-sensor-toolbox/auide-sitina-and-installing-air-sensors
• South Coast Air Quality Management District - Sensor Siting and Installation
Guide
o Guidance on how to locate and install air sensors: http://www.agmd.gov/ag-
spec/resources/related-documents
o English: http://www.aqmd.gov/docs/default-source/aq-spec/resources-
page/aq-spec-sensor-siting-and-installation-guide_v1-0-(english).pdf
o Spanish: http://www.agmd.gov/docs/default-source/ag-spec/resources-
page/sensor-siting-and-installation-guide v1 -Q-(spanish).pdf
• U.S. EPA Air Sensor Toolbox - Air Sensor Research Grants and Challenges
Website
o Website provides information on grants and challenges related to air research
and air sensors
o https://www.epa.gov/air-sensor-toolbox/air-sensor-research-grants-and-
challenges
3.6 Setup: Collocation and Correction
Air monitoring instruments need periodic checks to ensure they are functioning correctly
and generating high-quality data. Environmental agencies, which are responsible for
operating reference monitors, routinely
calibrate the instruments by testing them with
certified and known concentration standards.
They then use the calibration results to
adjust the instrument settings to match the
certified or known concentration. This process
is regularly repeated and monitored to ensure
highly accurate data.
What are Some Commonly Used
Terms for Reference Instruments?
• Reference Monitor
• Federal Reference Method (FRM)
• Federal Equivalent Method (FEM)
• FRM/FEM Monitors
66
-------
Air sensors also need periodic checks but often cannot be calibrated in the same way as
reference monitors. Instead, many air sensors are collocated or operated side-by-side with
a reference monitor to see if they produce comparable data. Instead of adjusting instrument
settings, which is often not possible for sensors, the raw data produced by a sensor may
need to be adjusted (such as applying a multiplier and additive factor to the sensor raw
data) to improve accuracy. This data adjustment, also called a correction, allows the
sensor data to better match the reference monitor data. The Collocation-Correction
process of collecting data and adjusting sensor measurements is described in the following
sections.
I |
What are Key Definitions Related to the Collocation-Correction Process?
Calibration - procedures for checking and adjusting a reference instrument's
settings so that the measurements produced are comparable to a certified standard
value.
Collocation - checking the performance of an air sensor by installing and operating
the air sensor close to a reference instrument.
Correction - adjusting air sensor data to increase its accuracy relative to a known
reference value.
""Important Note: Sometimes these terms are used interchangeably but they have
different and distinct meanings.
1
67
-------
3.6.1 Air Sensor Collocation
Collocation involves checking the performance of an air sensor(s) by placing it near or
beside a reference instrument and operating them at the same time and place under real-
world conditions. Collocations may take place at existing air quality monitoring sites around
the U.S. but require developing relationships with tribal, state, or local air monitoring
agencies who may have varying constraints for allowing collocations (e.g., space or power
limitations, access requirements, liability issues). Alternatively, it may be possible to work
with academic partners or contractors to set up and operate a reference instrument
specifically for your project. Be sure to follow established quality control and assurance
procedures when operating the reference instruments. Ideally, sensors would be setup
within about 20 meters of horizontal distance and 1 meter or vertical distance from the
reference instrument. However, airflow to the reference instrument and sensors must be
unobstructed.
As shown in Figure 3-9. a range of potential
collocation options exist to meet various
logistical and budgetary constraints. A
combination of approaches can be used
depending on the length of the project, the
desired data quality, and other project
constraints or needs.
s \
What if a Reference Instrument is
Unavailable?
Dependent on resources, it is possible
to work with an experienced partner to
setup and operate a reference
instrument.
Alternatively, all sensors can be
collocated together even without a
reference instrument to better
understand how well their
measurements compare. With this
information, you can conduct an air
monitoring study making comparisons
within your network of sensors. But,
you'll have less confidence in the
concentrations reported and be less
able to compare results with other
studies or data sources.
v >
68
-------
Kev
Collocation Strategy
sensor
0 reference
instrument
~-_©
Q 00
0
®0"a°
~
sensor
transfer
Periodic All
Sensors
Continuous
Subset
Reference
Transfer
Sensor Transfer
An air sensor
All air sensors
Some air sensors
A reference
collocated with a
V yes
™ somewhat
X no
$ cost
maintenance
operate next to a
reference
instrument for
short periods
before and after
the study and/or
periodically.
are continuously
operated next to a
reference
instrument while
others are
deployed to other
locations.
instrument visits
each air sensor for
a short period(s).
reference
instrument, with
known
performance
characteristics,
visits each sensor
location for a
short period(s).
Continually check
sensor performance
X
fSJ
X
X
Capture a wide
range of weather &
pollution conditions
/V
V
fSJ
/V
All sensors tested
at the same time
/
f\j
X
X
All sensors tested
against reference
instrument
V
y
V
X
All sensors tested
at their sites
X
X
V
y
Additional
equipment costs
$
$
$$$
$$
Frequent operator
maintenance
w
\\
w\
\w
Figure 3-9. Different Types of Air Sensor Collocation Strategies
Periodic All Sensor Strategy. All sensors are collocated with a fixed reference instrument
at the beginning and end of the monitoring study. Depending on the length of the study, the
collocation may happen periodically (e.g., seasonally, every 6 months, annually) throughout
the study.
Strengths:
• All sensors are tested at the same time letting you know how they compare.
• All sensors are compared to a reference instrument for a limited time.
• There are no additional equipment costs if you can use an existing reference
instrument.
• Sensors from smaller networks can be moved without major effort.
Weaknesses:
• Weather and air pollution conditions during the collocation may not be representative
of the actual conditions encountered by the sensors when deployed at their sites.
• It will be more difficult to detect subtle changes in sensor performance over time or
changes in performance due to a unique or short-term pollution events (e.g., short-
lived seasonal dust storm) because a sensor is not permanently located next to the
reference instrument.
69
-------
• Moving sensors back and forth between the collocation site and their permanent
sites can be labor intensive and increase the likelihood of damage.
• For large networks, there may not be enough space or power available at the
collocation site for all of the sensors to be collocated at once.
Continuous Subset Strategy. All sensors are first collocated at a reference site. Then,
some sensors are continuously operated next to a reference instrument while others are
deployed to a different location(s).
Strengths:
• Because some sensors remain collocated with a reference instrument, sensors are
tested under a wide range of weather and pollution conditions and you can detect
performance changes over time. However, this approach assumes that all sensors
perform similarly to one(s) that are continuously collocated.
• All sensors are tested at the same time letting you know how they compare.
• All sensors are compared to a reference instrument; some only for a limited time.
• There are no additional equipment costs if you can use an existing reference
instrument.
• Sensors from smaller networks can be moved without major effort.
Weaknesses:
• Moving sensors between the collocation site and their permanent sites can be labor
intensive and increase the likelihood of damage.
• For large sensor networks, there may not be enough space or power available at the
site for all of the sensors to be collocated at once.
• If the collocated sensor fails and needs to be replaced, you no longer know how the
new sensor's performance compared to the other sensors in the network. You might
consider leaving several sensors collocated with the reference instrument.
Potential Modification:
• Space and power constraints may dictate that not all sensors can be collocated at
the reference site at once. Alternatively, sensors can be collocated in batches. Or
the sensors can be collocated elsewhere to understand how the sensors compare to
one another and only the small subset then goes on to be collocated at the
reference site.
• A sensor network may grow with time or portions of the network may need to be
replaced due to sensor age. In these cases, batches of sensors could be collocated
just before the network is expanded/replaced.
Reference Transfer Strategy. A reference instrument visits each sensor for a short period
of time. This strategy can be useful for characterizing the performance of a network of
sensors over the course of the long-term study.
70
-------
Strengths:
• All sensors are compared to a reference instrument for a limited time; both the
sensor and reference instrument experience the same pollution sources and
concentrations and weather conditions
during collocation.
• Sensors do not need to be moved to
another location after their initial
deployment, thereby minimizing the
chances of damage.
Weaknesses:
• Weather and air pollution conditions or
sensor performance may change
between collocation periods.
• Does not test all sensors at the same time, under the same conditions.
• Can be costly to obtain, operate, move, and maintain a reference instrument(s).
• Some sensor sites may not be able to accommodate a collocated reference
instrument (e.g., the sensor is mounted on a pole or in an unsecured area).
Sensor Transfer Strategy. A sensor, or research grade instrument, with known
performance characteristics, is brought to each location where a sensor is deployed, in
order to best know the sensor performance characteristics, sensors used in this strategy
are usually left collocated with a reference instrument when not being moved around the
network.
Strengths:
• All sensors are compared to a sensor or research grade instrument with known
performance for a limited time; both experience the same pollution source and
concentrations and weather conditions during collocation.
• It is less costly and labor intensive to transport a sensor or research grade
instrument around the network.
Weaknesses:
• Assumes that the performance of the traveling sensor or research grade instrument
does not change when moved from site to site which may not be true if pollution
sources or concentrations change.
• Difficult to detect subtle changes in performance over time.
• The deployed sensors are not tested against a reference instrument, which makes it
more difficult to quantify the accuracy of each sensor.
• Sensors are not tested at the same time, so you cannot determine how one sensor
compares to another.
f S
Tip: Collocate sensors in a setting
similar to where they are deployed
Make sure that the collocation site has
similar characteristics to where the
sensors are deployed. For example,
do not collocate sensors near a road
and then apply those results to a
sensor network in a rural area.
8 s
71
-------
Collocation involves some critical questions to consider such as
1. What reference instrument should I use?
2. How frequently should I collocate the sensors? Continuously (e.g., every day) or
periodically (e.g., different seasons)?
3. How long should I collocate the sensors to measure enough variation in the full
range of weather conditions (e.g., T, RH), and the full range of pollutant
concentrations?
4. Should I collocate all sensors simultaneously at one central location or move a
reference instrument to each air sensor site (i.e., test each sensor in its
surroundings)?
5. Where should I collocate sensors so that the location resembles my sensor
deployment area?
6. How much effort and what type of equipment (or access to it) is involved?
Some general tips for collocation include:
• Collocate for an adequate number of days to characterize a sensor's response
over a range of weather and pollutant concentration conditions. Some suggested
number of collocation days for data collected at different time intervals include;
o 24-hour data: About 30 days of collocation
o 1 -hour data: About 14 days of collocation
o 5-minute data: About 7 days of collocation
• Locate the sensor as close to the reference instrument as possible so that the
devices are measuring the same air quality. If you are using an existing air quality
monitoring site, you will need to work closely with your tribal, state, or local
environmental agency to access their reference instruments and data repository.
Keep in mind that some agencies may not allow the general public to collocate
sensors at their sites.
• Measure meteorological conditions (T, RH,
wind speed, wind direction) because the
sensor performance may be affected by
weather conditions. The tribal, state, or local,
environmental agency may measure these at
their air monitoring site. Measurements may
also be available from a nearby airport or
national weather service or similar data
service.
• Consider using analytical tools to help you
evaluate your collocation data (e.g., U.S.
EPA's Macro Analysis Tool).
See U.S. EPA's Air Sensor Collocation Guide for additional information.
72
-------
3.6.2 Correction of Sensor Data
The collocation results can be used to correct the sensor data to more closely match the
data from the reference instrument. This correction process helps account for known bias
and unknown interferences from weather and other pollutants and is typically done by
developing an algorithm. An algorithm can be a simple equation or more sophisticated
process (e.g., set of rules, machine learning) that is applied to the sensor data. This section
further discusses the process of correcting sensor data.
The first step in correcting sensor data is to compare the collocation data obtained from the
sensors and reference monitor, making sure to align the sampling times and averaging
periods for the two data sources. Be sure to note the time zones and whether daylight
savings time is used. For example, to compare 1-minute air sensor data to 1-hour
reference data, users will need to calculate a 1 -hour average of the sensor data (i.e., the
average of 60, 1-minute values). Users should ensure that the averaging methods match
for both sensor and reference data. Averages may be time-beginning (e.g., a 1-hour
average computed from measurements collected between 12:00 and 12:59 is assigned to
the 12:00 hour) or time-ending (e.g., a 1-hour average computed from measurements
collected between 12:00 and 12:59 is assigned to the 13:00 hour).
Sensor and reference monitor data can be compared by creating a scatter plot to visualize
the relationship. Figure 3-10 shows an example scatter plots. In this Figure, data from the
reference instrument, in this case the T640x, is shown along the horizontal x-axis and
sensor data is shown on the vertical y-axis. Each dot shown in the Figure represents a
measurement taken at some point in time. By drawing lines vertically and horizontally from
each dot, you can find the reference concentration and sensor concentration reported at
each time interval. If the sensor and reference measurements perfectly agreed with one
another, all of the dots would lie along the 1:1 line shown in grey. Often, there is
disagreement and you find that the sensor may always measure a little higher or lower
than the reference instrument. This disagreement is called bias (see Section 3.4.1) and
sometimes an algorithm can be created to correct the sensor data and eliminate the bias.
You may see that the dots cluster along a line or curve. If so, it may be possible to create
an algorithm or equation that describes the relationship between the air sensor and
reference data. The most common, and often suitable, correction equation is the ordinary
least-squares linear regression equation. A function for determining a linear regression
equation is well established in many software packages (e.g., Excel, R) and available using
the U.S. EPA Excel-based Macro Analysis Tool as well.
73
-------
What is the Ordinary Least-Squares Regression Equation?
A common correction algorithm is the "best-fit" line equation derived using the
ordinary least-squares regression:
y = mx + b
Where:
y = the sensor measurement
m = the slope of the regression line
x = the reference monitor measurement
b = the y-intercept of the regression line
Slope (m), which can be positive or negative, shows how similar the sensor
measurements are to the reference measurements. The closer "m" is to 1, the more
the sensor responds like the reference instrument. Intercept (b) shows what the
sensor measurement will be, on average, when the reference instrument measures
zero concentration. Together the intercept and the slope describe the sensor bias.
Linear regression equations also include the coefficient of determination (R2),
ranging from 0 to 1. R2 measures the amount of data scatter and how closely the data
points are to the "best fit" line. R2 values nearer to 1 indicate stronger correlations or
less data scatter.
In Figure 3-10 panel A, the slope is > 1 (m = 2.02) and sensor measurements are higher
than reference monitor measurements when concentrations are above 8.5 ng/m3. This is
an example of a sensor that overestimates pollutant concentrations. In panel B, the slope is
< 1 (m = 0.22) and all sensor measurements are less than the reference monitor
measurements. This is an example of a sensor that underestimates pollutant
concentrations. A slope equal to 1 would indicate perfect agreement between sensor and
reference data.
A measure of how close
the data points are to the
slope-intercept line,
represented by R2, is the
called the coefficient of
determination. This value
describes the amount of
scatter in the data.
y = 2.02x-8.63,
R2 = 0.89 ;
15
10
(A)
1:1 line
Here, the sensor
dp
measures 30 ng/m3
whereas the
regulatory monitor
(T640x) measures
• 7
20 ng/m3
10 20 30
T640x PM2.5 ing/m3)
40
25
20
>15
^ 10
Q_
O
a c
c 5
(B)
y = 0.22x-0.65
R2 = 0.73
1:1 line X
Here, the sensor
measures 2.5 |j.g/m3
whereas the
regulatory monitor
measures 15 (ig/m3
0 5 10 15 20 25
T640x PM2.5 (Vg/rn3)
Figure 3-10. Example of the Ordinary Least-Squares Regression
74
-------
R2 ranges from 0 arid 1, where values closer to 1 indicate stronger correlation, or
agreement, between the sensor and reference data. A value closer to 0 indicates a lack of
agreement.
Some sensors, like the example shown in
Figure 3-11, may not show strong agreement
with the reference instrument (R2 closer to 0).
Sometimes scatter plots look like a scatter of
points rather than a grouping of points that
resemble a line. This often means that an
ordinary least-squares linear regression
equation is not appropriate for correcting the
sensor data. This can mean that the data is
not correctable or useable but, sometimes it
can mean that a more complex algorithm or
equation may be needed. For instance, a
user may choose to apply a multi-linear
regression because the sensor response is
related not only to pollutant concentration but
also another variable like RH or the
concentration of another pollutant.
Developing correction algorithms is an active
research area with new approaches and
methods likely to evolve over time. Some
research has applied multi-linear regressions,
polynomial fits, more complex models, and even
machine learning. Understanding the correction
algorithm is essential for ensuring traceability
from raw to corrected data and for understanding
the factors that influence the measurement.
Complex or "black box" algorithms may sound
promising, but often simple, understandable
algorithms are preferred.
Note that an air sensor may respond differently
over the full concentration range. For example,
Figure 3-12 shows a linear response relative to
the reference data at lower concentrations and a
non-linear response above this concentration.
Thus, a single type of correction equation may
not apply over the full range of concentrations.
25
20
y = 0.08x*1.98
R2 = 0.02
00
=l
—' 15
10
o
to
c
QJ
IS)
5 10
T640x PM
Relative Humidity (%)
15 20 25
2.5 (|ig/m3)
20 40 60 80 100
Figure 3-11. Example of a Sensor
That Shows No Agreement with
the Reference Instrument
Are There Other Metrics That
Describe Sensor Performance?
Error is a measure of the
disagreement between the
pollutant concentrations reported
by the sensor and the reference
instrument (see Appendix F).
Several metrics are used to
describe error including Root
Mean Square Error (RMSE),
Mean Bias Error (MBE), and
Mean Absolute Error (MAE),
among others. Each metric has a
slightly different definition and
calculation.
75
-------
o>
3
to
c
o
'¦p
ro
+J
C
CD
O
c
o
o
2
Q.
o
(/>
c
V
CO
>
/
/
/
•• •«
• •
• • /•«'••••«
•• • • •
• •• •
•
*/• - /
*/ •• x
7® /
'/•
//
\ •
\
Reference Monitor PM25 Concentrations (pg/m3)
Figure 3-12. Scatter Plot Showing that an Air Sensor has a Linear Response at Lower
Concentrations and a Non-linear Response at Higher Concentrations
Figure 3-13 presents an example correction using a linear regression equation derived from
sensor and reference data. In this example, the corrected sensor measurements are
obtained by rearranging the "best fit" line equation (i.e., y = mx + b) to solve for "x", where
"x" equals the corrected sensor measurement. Users can correct sensor data using the
algorithm so that it more closely matches the reference data. Note that some sensor
manufacturers apply data corrections on-board the sensor or in the cloud. Also, some data
management systems (see Section 3.7.3) can apply a correction. In either of these cases,
it is good practice to ask the manufacturer about any corrections performed on the data
and to fully understand how the air sensor data are corrected.
Measured Sensor Data — b
Corrected Sensor Data =
m
76
-------
1:1 line
Measured Data
Corrected Data
0
100
200
300
Reference PM2 5 (p.g rrr3)
Figure 3-13. Examples of Air Sensor Data Corrections
Sometimes scatter plots may resemble the examples shown in this section but have just a
few data points that appear far away from the linear regression line. These points are
referred to as outliers. Sometimes these data points can significantly change the results
and instead of the linear regression line passing through the middle of most of the data
points, it can seem to move off to one side. This indicates that those suspected outliers
should be investigated a little further before interpreting the results. Creating a time series
plot, or a plot of the sensor and reference instrument concentrations as a function of time,
may show data spikes (data points with higher concentrations) from one or more of the
instruments. Users may want to examine these spikes to consider if they are real or not.
Some spikes may represent elevated pollutant concentrations outdoors. For instance,
particulate matter (PM) concentrations may go up for a short period of time because of
nearby mowing activities. Other spikes may not reflect real changes in pollutant
concentrations and may instead indicate that something is wrong with the device like a data
logging or device error. If the spikes are believed to be false, users may consider if there is
a routine way of detecting and removing the false data that will not accidentally remove the
good data. If a routine method can be developed, it should be documented or written out so
that the method can be communicated, and all data will be treated similarly. Using a routine
method to identify and remove false spikes is a processed commonly called data cleaning.
After cleaning the data, the linear regression can be recalculated and applied to the
remaining data to complete the data correction.
Resources for More Information
• U.S. EPA Air Sensor Collocation Instruction Guide, U.S. Environmental
Protection Agency, Office of Research and Development
77
-------
o Resource provides background information, links to web-based supporting
materials, and instructions for evaluating the performance of air sensors by
comparing the measurements made by collocated sensors and reference
instruments
o https://www.epa.gov/air-sensor-toolbox/air-sensor-collocation-instruction-
quide
• U.S. EPA Air Sensor Collocation Macro Analysis Tool
o Excel-based tool that helps users compare data from air sensors to data from
reference instruments
o https://www.epa.gov/air-sensor-toolbox/air-sensor-collocation-macro-analvsis-
tool
• Community in Action: A Comprehensive Guidebook on Air Quality Sensors,
South Coast Air Quality Management District (South Coast AQMD), Air Quality
Sensor Performance Evaluation Center (AQ-SPEC), September 2021
o Guidebook for community organizations that covers planning for monitoring
using sensors; sensor deployment, use, and maintenance; and data handling,
interpretation, and communication
o http://www.aqmd.gov/ag-spec/special-proiects/star-grant
• South Coast AQMD Low-Cost Sensor Data Analysis Guide
o Guide that provides some brief instructions to help community scientists
interact with the data they are collecting as well as some questions to help
guide their analysis
o http://www.agmd.gov/docs/default-source/ag-spec/star-grant/air-gualitv-
sensor-data-analvsis-guide.pdf?sfvrsn=6
3.7 Collect: Data Collection, Quality Assurance/Quality Control, and
Data Management
With a question well-posed, a plan created, and sensors properly set up after collocation, it
is time to collect data. There are many activities involved in data collection beyond simply
turning on the sensor and collecting measurements. Users will need additional preparation
before and during data collection activities to ensure that useful data are collected.
This section discusses various data collection activities, quality assurance/quality control
(QA/QC) checks, and typical components of data management systems (DMSs).
78
-------
3.7.1 Data Collection Activities
Collecting good quality, complete, and ultimately useable data will require attention to
several oversight tasks after the air sensors begin operating. These tasks include:
Frequent data review. Reviewing data frequently (e.g., daily, weekly) lets you detect
problems early, notice trends in the data, ensure that maintenance activities are completed,
and become familiar with recurring patterns. For instance, plotting the data, whether in a
time series (i.e., a plot with the
pollutant concentrations on the
y-axis and the date and time on
the x-axis) or another form can
be a good place to start (see
Section 3.8 for plotting
options). You might see typical
patterns, such as low
concentrations during the
morning hours or identify when
high pollution episodes occur.
These data reviews help you
develop a general sense of air quality in an area under different conditions. When typical
conditions are known, it becomes easier to identify times when sensor readings are atypical
and why these atypical readings are occurring (e.g., Is an air sensor malfunctioning? Is
wildfire smoke present? Is a weather pattern responsible for higher levels?).
Maintenance. Like most other forms of technology, air sensors require preventive
maintenance to ensure proper functionality and reliable data collection. Maintenance
activities are necessary for both short- and long-term operations. Air sensor maintenance
can include regularly scheduled cleaning of surfaces or inlets to prevent the buildup of bugs
or dust, replacing filters, or replacing sensor detector components as they age.
Maintenance can also include examining site conditions for any changes (e.g., vandalism,
overgrown trees).
By properly maintaining an air sensor device, you can reduce errors in data collection,
extend the device's operating life, and save money that would otherwise be spent on
replacement parts and repair services. Typical air sensor maintenance activities are listed
in Appendix C and may also be provided by the manufacturer.
Troubleshooting. Problems with air sensors (e.g., failing to report data) will likely occur
and may require troubleshooting to resolve the problem and to continue collecting data.
Troubleshooting might include visiting the sensor, contacting the manufacturer, seeking
guidance from other air sensor users, or other activities. User manuals may also provide
tips on troubleshooting.
What are the Benefits of Frequent Data Review?
• Identify and resolve problems quickly
• Minimize data loss
• Learn what normal patterns look like
o Detect real, high-pollution events early
o Understand how air quality changes:
o During the day
o Weekend vs. weekday
^4
79
-------
Quality control (QC) checks. It is important to frequently review the data for problems
such as outliers (e.g., data that are significantly different from other data values), drift, etc.
Some sensor manufacturers may offer a software package or online user interface that
offers some automated checks of the data to assist in this process. Note that automated
checks may not catch subtle problems (e.g., a gas sensor slowly degrading and losing its
response) or may flag a real-life event or very high concentrations (e.g., high PM2.5
concentrations from wildfire smoke) as bad data. Do not solely rely on automatic QC
checks to identify issues with the data—always review the data frequently. Section 3.7.2
discusses additional QA and QC checks.
Periodic collocation. Collocation can help quantify the accuracy of a sensor while periodic
checks can help ensure that accuracy is not changing over time or in different conditions.
Users should develop a collocation approach or use the manufacturer's recommendation to
conduct a periodic collocation to check the quality of the air sensor's measurements.
Section 3.6 provides information on the process of collocation and how to correct data to
make it more accurate.
3.7.2 Checks to Ensure Quality Assurance and Quality Control
Quality assurance (QA) and quality control (QC) are essential components of a project
that will ensure that credible and useful data are collected (Figure 3-14). QA consists of
planned steps to manage the project and collect, assess, and review the data. An example
of QA is developing a plan for air monitoring (see Section 3.3) to ensure identification of all
tasks or steps to review air sensor data and confirm the sensor is operating properly. QC
includes steps taken to reduce error from the instruments or measurements during a
project. QC procedures are activities that include collocation, correction of data,
maintenance, automatic data checks, and data review. Essentially, QA is the planning and
QC is the action taken to produce high-quality data. QA/QC are important components of a
project that will help ensure that credible and useful data are collected.
Regardless of whether the user presents
the results as a written report, oral
presentation, or in conversation, users
should clearly describe the approach, the
measurements obtained, the QA/QC
checks in place, and the interpretation of
the data. If any of these components are
missing or not well executed, your data's
credibility will diminish.
Table 3-2 shows the recommended QC
checks that can be performed on an air
sensor and its data. The checks are
designed to catch problems early, correct
them, and produce a useful, high-quality data set.
Quality
Assurance
Quality
Control
4
Figure 3-14. Definitions of Quality Assurance
and Quality Control
80
-------
Table 3-2. Common Quality Control (QC) Checks
QC Check
Description
Units
Check that the sensor reports data in the correct units of measure.
Time
Check that the sensor reports data at the correct time and in the right time
zone. Check times after any seasonal time changes (e.g., daylight savings
time).
Timestamp
Determine the timestamp, which is the time when data are stamped (i.e.,
tagged) by an instrument. Measurements and data averages will have times
that either represent the beginning of the time period (time beginning) or the
end of the period (time ending).
Matching
Timestamps
Check the time zones and timestamps for each dataset to make sure they
are similar when comparing measurements made by different instruments.
Data Review
Check data frequently (e.g., daily, weekly) to detect problems early, identify
trends in the data, ensure that maintenance activities were completed, and
become familiar with recurring patterns (see Section 3.7.1).
Data
Completeness
Completeness measures the amount of data a sensor collects compared to
the amount of data that was possible to collect if the sensor operated
continuously, without data outages, during a period (e.g., 1-hour, 1-day). A
75% completeness level is a useful criterion to meet as the averaged data is
generally representative of that time period. For example, at least 45, 1-
minute measurements are needed to make a valid 1-hour average at 75%
completeness.
Automatic
Data Checks
Software can check data for problems and outliers. Check your data
management system for these and other data checks. Note that some data
checks may not catch subtle problems (e.g., a gas sensor degrading and
slowly losing its response) or may flag an infrequent event or very high
concentrations (e.g., high PM2.5concentrations from wildfire smoke) as bad
data. Do not solely rely on automatic QC to check data quality; always do
frequent manual data reviews.
Common automatic checks include:
• Range. Check the minimum and maximum concentrations expected
and recognize some air sensors may report slightly negative values.
• Rate of Change. Check the difference in data values from an air
sensor between two consecutive time periods (e.g., hours). Flag the
data if the difference, or rate of change, exceeds the value set by the
user. For example, it is unusual for PM2.5 concentrations to jump by
more than 100 |jg/m3 from one hour to the next unless a significant
source such as wildfire smoke or fireworks is present. Thus, if the
81
-------
QC Check
Description
value for the rate of change check is set to 100 |jg/m3, an increase
from 70 |jg/m3 to 200 |jg/m3 between consecutive hours would
exceed the rate-of-change check.
• Sticking. Check if data values are "stuck" at the same value for a
specified number of hours. Establish criteria for the number of
consecutive hours for which data can be reported at the same value.
For example, it is uncommon for PM10 concentrations to remain at the
same concentration for several consecutive hours. If the number of
concentration hours is set to three, and the PM10 concentration is the
same value for more than three consecutive hours, that could indicate
a stuck value.
• Duplicate sensor comparison. Some sensors incorporate two
identical sensing components inside which provide two separate
pollutant concentration measurements. Check the agreement
between the readings and flag data if the difference exceeds an
acceptable threshold.
• Buddy System. Check the difference between data values obtained
from a single location and the average data values obtained from
other nearby locations.
• Parameter-to-Parameter. Check two or more pollutants for known or
expected physical or chemical relationships. For example, PM2.5
should be less than PM10 measured at the same site and time, and
NO2 and O3 concentrations are often inversely correlated (i.e., when
O3 is high, NO2 is typically lower).
Manual Data
Validation
Evaluate the data quality during the collection phase of the project to identify
and correct potential problems that may arise. To accomplish this, analyze
data to identify seasonal, day/night, and weekday/weekend patterns and
weather changes. An absence of expected patterns may indicate a problem
with the sensor or with the measurement approach.
Many of the QC checks in Table 3-2 will help evaluate the quality of the data obtained
during the collection phase to identify and fix common problems found in air sensor data.
Common problems in the data are described below.
• Drift refers to a gradual positive or negative change in a sensor's response over
time due to various reasons (e.g., aging of the sensor component). Drift may lead
users to incorrectly conclude that concentrations have increased or decreased over
time. Some ways to reduce drift include frequently performing a collocation-
correction process (see Section 3.6) or conducting frequent maintenance on the
sensor.
• Interferents and Influences include factors that hinder, obstruct, or impede the
sensor's ability to provide high quality measurements. Other pollutants that interfere
with the measurement of the target pollutant are sometimes referred to as cross-
sensitivities. For example, oxidants (e.g., O3) in the air can interfere with
electrochemical sensors used to measure NO2 and high moisture content (e.g.,
above 85 percent RH) can cause PM air sensors using optical technologies to
overestimate PM concentrations. Debris, dirt, and insects can also impact sensor
performance. Interferents and influences may alter sensor accuracy, and a sensor
82
-------
can be impacted by several different factors simultaneously. Manufacturers
sometimes disclose which pollutants and weather conditions may impact sensor
performance, but manufacturers may not describe how much the sensor will be
affected. Before using a sensor to measure air quality, consider whether the possible
sensor interferents will be present in the air to be sampled and check with the
manufacturer about potential interferences and how to minimize their effects, if
possible.
• All measured parameters may not be reported in a dataset. Many sensors report
several pieces of data including pollutant concentration(s), temperature (T), relative
humidity (RH), and more. Because data may come from several components within
the device (e.g., pollutant concentration come from one component, T/RH from
another), it is possible pieces of the data stream may be missing while others are
present. Be sure to review each parameter separately to ensure all components are
reporting properly.
• Unexpected downtime due to for example, power loss and power surges, can
affect air sensor performance by causing sensors to shut down, restart, or interrupt
data transmission. For example, an air sensor that has a longer warm-up period may
show data gaps or inconsistent data when power loss occurs. Data gaps and
significant data loss associated with unexpected downtime may prevent users from
collecting a complete, valid data set.
• Unexpected problems can occur in many monitoring instruments, air sensors, and
other electronic equipment. In some cases, a concentration spike (e.g., data spike,
outlier) may be caused by
electronic or other device-
related issues; however, a
spike could be a valid
measurement and it can
sometimes be difficult to tell
the difference. Local
observations may help to
interpret the data. For
example, observing someone
smoking near a PM sensor
and a seeing a corresponding
short-term increase in PM
concentrations would indicate
that this is a valid
measurement (i.e., the
sensor responded to the
smoke), even though the
measurement does not
represent the general air
quality conditions at that time.
How do I Verify That my Sensor is Working
Properly if There are No Reference Monitors
Nearby?
Although we recommend that users carefully review
and compare the sensor data collected against
nearby reference instruments to ensure the data
appear reasonable, there are situations and
locations where this comparison is difficult because
there are no reference monitors nearby.
Based on resources (e.g., money, equipment,
expertise), users could
• Setup a reference instrument nearby
• Setup a sensor which was recently collocated
with a reference monitor and shown to
provide comparable results
• Compare trends and concentrations with
several nearby sensors and consider whether
similarities/differences are expected or
surprising.
83
-------
In summary, having QA/QC checks in place for a project is important as they help ensure
data quality and allow you to address common problems if they arise. Having QA/QC
information available is also important if it is requested by anyone who wants to use the
data.
3.7.3 Data Management System
Air sensors produce a large amount of data that must be routinely tracked and managed to
access, review, and use the data effectively. A data management system (DMS) is a
collection of procedures and software needed to acquire, process, and distribute data. A
DMS helps streamline data processing, provides QC and review tools, maintains digital
records and backups of the data, and displays, reviews, and facilitates sharing the data.
These features make it easier to use air sensor data and to identify instrument errors or
other problems early. A DMS also makes it easier to operate and manage a network of
multiple sensors simultaneously.
Figure 3-15 shows the key components and functions of a DMS. Note that a DMS may be
bundled with a sensor (e.g., manufacturer offered cloud data portal), purchased as a third-
party system, or be available as open-source software. Each of the functions shown in
Figure 3-15 play an essential role in operating either a single sensor or network of air
sensors and collecting useful data.
Figure 3-15. Major Components and Functions of a Data Management System (DMS)
The basic DMS functions consist of the following:
• Ingest data. A DMS acquires and reformats data. The DMS can either pull data
from an air sensor system or sensors can push data to the DMS. Think carefully
about data format as these choices can impact data storage volume and how easy it
is to use existing code to process and visualize the data.
84
-------
• Store data and metadata. After acquiring and reformatting data, the DMS needs to
store data and metadata (e.g., site locations, configurations, unit identifiers). The
storage system could be a spreadsheet, a file-based system, or a relational database
(i.e., database structured to recognize relations among data). These options provide
different levels of flexibility and scalability and may or may not require a license
agreement. Storing data can also include tracking changes to the data and metadata
(i.e., chain of custody) to help ensure data integrity.
• Process data. After collected data are stored, the DMS needs to process and QC
data. This consists, for example, of applying data correction algorithms, creating data
averages (e.g., 1-hour, 8-hour, 24-hour), calculating summary
statistics, converting data to an Air Quality Index (AQI) for health messaging, or other
data processing and visualization needs. QC checks (see Table 3-2) are a necessary
part of data management and should be included to identify outliers and problems
with the data.
• Monitor network health. Features that help you quickly and easily identify problems
within a network of sensors are highly desirable especially when your network
consists of many sensors or when sensors are spread out over a large geographical
area. At a minimum, this feature should help identify sensors that are not reporting
data so that they can be repaired or replaced. More advanced features may identify
sensors that often fail the QC checks (see Table 3-2) or even help diagnose
problems for faster troubleshooting or repair (e.g., pinpoint the external component
that needs to be replaced).
• Control of the system. It is highly useful for users to have the ability to control the
DMS. System control features might include the ability to change system settings
(e.g., add/remove sites or users, change time zones), set or change QC criteria, edit
and correct data, generate statistical summaries and reports, and visualize data.
These types of control features enable users to oversee and adjust the DMS to meet
project needs.
• Distribute data: Retrieving and distributing the data is another important function of
a DMS. These activities can range from very simple ways to export data to more
sophisticated software called Application Program Interfaces (APIs). An API
allows interactions between the DMS and other remote software systems. The API
defines the kinds of calls or data requests, how to make them, the appropriate data
formats and conventions to follow, and so on. The API also allows other websites
and data systems on the internet to access data within the DMS.
Other recommended features of a DMS include:
• Data security. The ability to keep data safe from hacking, altering, or other
unwanted activities is important for data integrity.
• Data redundancy (backups). Given the potential for situations that can cause data
loss, data backups are critical to ensure that users can recover data.
• Cloud-based software that can be accessed from any location. It is ideal to have
software that can allow users to view data and/or instrument status remotely no
matter where a user is located. This will allow users to check monitor instruments for
85
-------
any problems, change instrument settings if needed, frequently review data, or carry
out other activities to manage a project.
• Version control features for tracking changes to the data. Manufacturers may
change the firmware or other settings on a device that can impact the data. Having a
way to document these changes will allows users to better understand their data.
• Data ownership terms and usage rights. Detailed information about who owns the
data, who has access, and if there are terms and conditions on its use will help
users determine if the DMS is acceptable for their proposed project.
• Tools for data QC, data review, and visualization. These tools can allow users to
frequently check the data to identify any issues and quickly look at trends in the
data.
• AQI calculations and formulas. AQI may be calculated differently using different
formulas. If users want to directly compare the AQI to the U.S. AQI, for example,
knowledge on how the AQI is calculated is important.
• Public website that includes data displays (e.g., maps, time-series plots, and
tables), responses to frequently asked questions, and health information. A
public DMS can help communities quickly view data for trends, health information,
and answer their questions about the data.
• Email or text alerts for missing data, high values, and other events. Instrument
problems, air quality events, or other issues can always occur in a monitoring
project. The ability to know when these events occur can help users troubleshoot
and resolve issues during a project.
Selecting a DMS that meets your current
and future needs involves several
considerations. First, consider the size of
the network. The larger the network of
sensors and volume of data generated, the
greater the management challenge, and a
DMS with automated procedures can help
minimize effort. You will also need to
consider security elements like who can
access the DMS and other special
information technology (IT) requirements
such as the physical location of the cloud
software. Some systems are easy to use, yet may not allow customization, while others are
fully customizable and require software programming experience. You will also need to
consider your budget for software, licenses, and recurring (e.g., monthly) fees. Make sure
to ask about fees associated with adding more network sites or parameters, and other cost
drivers such as who will be responsible for customizing and maintaining the software.
/ \
Tip: A DMS can add costs to a project,
so budget wisely
A DMS can be highly valuable for sensor
project, but it is important to recognize
that these systems can add additional
costs to a project. Depending on the
provider, pricing can vary and be a one-
time, monthly, or annual fee. Having a
good understanding of these costs and
pricing plans is important.
I J
86
-------
A range of options exist for a DMS based on your specific needs:
• Spreadsheets can be used as a DMS to handle a small amount of data collected
over a limited time period. Although a spreadsheet is easy to create, it can be
challenging to automate processes and scale the system for larger sensor networks.
• Cloud-based systems are widely available to handle data. Many large technology
companies provide generic solutions to manage data. Although not always explicitly
designed for air quality applications, these cloud-based systems offer many of the
basic DMS functions discussed above; however, they may need to be customized to
suit your needs.
• Sensor manufacturer DMS solutions are offered by some air sensor companies
and perform the functions shown in Figure 3-15. In addition, systems offered by
some companies can ingest data from other sensor and reference monitors. You will
need to assess the capabilities of their systems, understand how much
customization is possible (if any), and consider any additional costs.
• Commercial, air-quality focused systems are used by many air quality agencies
and other professionals that collect and manage air quality data. These systems are
highly customized to meet the needs of air quality monitoring, and the vendors
typically understand air quality concepts. Vendors usually charge a software license
and/or recurring fee to use their DMS and additional features.
• Open-source, air quality-focused systems can provide another solution for data
management. They may not have licensing costs, but there could be costs
associated with hosting the software. This approach allows for scaling up as a
sensor network becomes larger but may require experience with open-source
software to install, operate, and customize the system.
Whichever DMS approach is used, consider the following best practices for data
management:
• Plan how you will manage the data before deploying any air sensor - it takes
considerable time and effort to manage the data.
• Contact other organizations that are operating air sensors or sensor networks and
ask about what type of DMS they use and any recommendations they may have.
• Check with the air sensor manufacturer/vendor for DMS solutions they offer or if
they can recommend other DMS solutions that are compatible with their products.
• Check with your internal information technology (IT) department before
purchasing, contracting, or designing a DMS to ensure that it meets data security
needs and that the correct settings can be enabled on networks and computers.
• Look for an automated DMS to streamline routine tasks (e.g., data ingestion, QC,
data reporting).
87
-------
Resources for More Information
• AirSensor and DataViewer Tools (R package)
o AirSensor is an open-source R package that allows users to access historical
data, add spatial metadata, and visualize community monitoring data through
maps and plots
o DataViewer is an interactive web application that incorporates the
functionality and data plotting functions of the AirSensor for interpreting and
communicating community data collected by sensor networks
o https://github.eom/MazamaScience/AirSensor/tree/version-0.5
o https://github.com/MazamaScience/AirSensorShiny
o These papers summarize the development and enhancements of the
AirSensor and DataViewer tools:
¦ Feenstra et al, 2020 https://doi.Org/10.1016/i.envsoft.2020.104832
¦ Collier-Oxandale et al, 2022 https://d0i.0rg/l 0.1016/j.envsoft.2021.105256
• Data Policies for Public Participation in Scientific Research: A Primer,
DataONE Public Participation in Scientific Research Working Group, August 2013
o Guide that introduces data policies in the context of public participation in
scientific research or community science, provides examples, and best
practices for implementing data polices in community science projects
o https://safmc.net/wp-
content/uploads/2016/06/Bowseretal2013 DataPolicyPrimer.pdf
• Handbook for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Handbook that covers common expectations for guality assurance and
documentation and best management practices to level the playing field for
organizations that train and use volunteers in the collection of environmental
data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgapphandbook 3 5 19 mmedits.pdf
• Data Management Guide for Public Participation in Scientific Research,
DataONE Public Participation in Scientific Research Working Group, February 2013
o Guide that provides best practices and other considerations for data
management along the life cycle of community science projects
o https://www.dataone.org/sites/all/documents/DataONE-PPSR-
DataManagementGuide.pdf
88
-------
• USGS Guide to Data Management
o United States Geological Survey (USGS) website that provides guidance,
best practices, and tools for data management including in-depth training
modules and numerous data management example scenarios
o https://www2.usgs.gov/datamanagement
• Survey Report: Data Management in Citizen Science Projects, Chade S and
Tsinaraki C., Publications Office of the European Union, JRC101077, 2016
o Report summarizes the findings from a Joint Research Centre (JRC) survey
of community science projects completed primarily in European Union (EU)
countries
o https://publications.irc.ec.europa.eu/repositorv/handle/JRC101077
• Citizenscience.gov Website
o Government website that promotes crowdsourcing and citizen science across
the U.S, government; website catalogs government supported community
science projects, provides a toolkit to assist with project design and
maintenance, and serves as a gateway for community science practitioners
and coordinators across the government
o https://www.citizenscience.g0v/#
• U.S. EPA Guidance on Environmental Data Verification and Data Validation,
U.S. Environmental Protection Agency, EPA/240/R-02/004, November 2002
o Guidance document that specifies the agency-wide program for
environmental data QA and includes practical advice to individuals
implementing data verification and data validation
o https://www.epa.gov/sites/production/files/2015-06/documents/g8-final.pdf
3.8 Evaluate: Analyzing, Interpreting, Communicating, and Acting on
Results
Understanding air sensor data is as important as selecting and operating an air sensor.
You should plan early for how to process, analyze, and interpret the data and how you will
share and communicate the results. Do not wait until data have been collected to determine
how you will use the data. Evaluating the results may reveal unanswered questions that
revise or update your questions or other steps in your plan, as shown in Figure 3-1.
There are many methods to analyze, evaluate, and share results, but the choice of which
approach to use depends on the questions you are seeking to answer. Some analysis and
interpretation can be relatively simple, while others that involve complex evaluations and in-
depth interpretation can be a challenge to communicate. For example, a PIVh.sair sensor
outside a home can measure local concentrations and help users determine the times of
89
-------
day when PM2.5 levels are lowest. However, deploying an air sensor network consisting of
many sensors to detect areas of higher or lower concentrations will require much more
detailed data analyses and interpretation. Again, users should plan how they will analyze,
evaluate, and communicate their results in advance. The remainder of this section provides
guidance on methods and techniques for accomplishing these tasks and resources for
getting started.
3.8.1 Analyze and Interpret Data
Data analysis is generally comprised of processing, then visualizing the data. Processing
the data typically includes the following steps:
1. Data cleaning to prepare the data for analysis. Cleaning includes: a) QC checks
and validation of the data to remove problems (e.g., large negative values, high
values caused by sensor failure) and outliers, and b) checking timestamps and units.
2. Documenting any adjustments or changes to the data.
3. Acquiring data from other sources needed for the analysis. These data could
include corresponding meteorological data, traffic data, emissions information,
and/or other sources.
4. Averaging data to evaluate the "big picture" signals in the data.
5. Grouping data to summarize the data, or group or filter data to explore more
details. Some examples include grouping data by time of day, day of week, location,
and/or meteorological conditions.
6. Correlating data to begin evaluating the relationships between the air sensor data
and other data values. For example, correlating PM2.5 concentrations and wind
speed can show how different weather conditions are related to PM2.5
concentrations.
7. Comparing data to evaluate the air sensor data against different air quality
standards and indices like the AQI.
Air sensors produce a large amount of data, so visualizing these data can help you
understand what they mean. Many different types of visualizations can be used to explore
air sensor data. Figure 3-16 presents some of the most common visualization tools.
90
-------
Time Series Plots show changes in one or more parameters
with time. Useful in comparing trends (pollutants,
temperature, multiple sites, etc.).
Scatter Plots show the relationship between two
parameters. Color coding the dots can indicate a different
variable (humidity, temperature, etc.).
Calendar Plots give a big picture look at quality over a
month or longer period. Dates can be colored to indicate
higher or lower concentrations.
Maps show the spatial patterns of data across a region.
Plotting other data such as traffic count or locations of
emissions sources can help explain changes in the data.
Wind and Pollution Roses show the frequency of wind
direction and can be colored to show pollutant concentrations
or wind speed. Useful in showing where higher pollutant
concentrations come from.
Figure 3-16. Common Visualization Methods for Air Quality Data
Whether interpreting data in a table or graph, users should ask questions about what they
can see in these visual plots. Some questions to ask include:
• Where and when do high (or low) concentrations occur?
• Are there certain characteristics (e.g., weather conditions, emissions patterns, days
of the week, season) that lead to high concentrations?
• How do pollutant concentrations change during the day? Are they lower or higher
during nighttime, and why?
• What are the prevailing wind directions during a period of interest (e.g., the
occurrence of high PM2.5 concentrations), and where did the air originate?
• How do measured concentrations compare to concentrations from nearby reference
stations?
• How do measured concentrations compare to health indices like the AQI?
91
-------
3.8.2 Communicating Results
Sensor users may wish to communicate their findings to a variety of different audiences
including members of their community, government officials, regulators, industry, or others.
The approaches taken (e.g., brief or detailed summary, technical details) and methods
used (e.g., social media, presentations, written materials) may need to be tailored to each
audience to most effectively communicate information. Communicating the data and results
in an open and transparent way can help sensor users build trust in newer and rapidly
changing sensor technology. Consider the following elements when seeking to
communicate data:
1. State your purpose or objective. This information can help you communicate why
you conducted the study and why you made some of your decisions. It is crucial to
describe how air sensors provided sufficient data quality to meet the project
objective(s). Results should be tailored to answer your monitoring question.
2. Describe the monitoring setup and data. Providing a clear description of where
sensors were located and the data collected allows others to gain confidence in the
data and results. Make sure your study files/reports include information about sensor
locations, site photographs, QC checks, time stamps, units, formulas for calculated
values, etc. Some of this information can be documented in a data dictionary which
is a description of the parameters collected. It is useful to retain these records for
some time beyond the study time frame because it may be necessary to
retroactively adjust data as more information regarding sensor performance or data
correction becomes available.
3. Describe the data processing and analysis. When communicating data and
results, it is recommended that information on the following topics be addressed and
clearly documented:
• Data cleaning and corrections/data adjustments
• QC checks
• Data analysis and interpretation, including software/methods used
• Maintenance and operations
• Limitations of the data and air sensors
4. Visualize the data and share the results. There are many ways to visualize data
(e.g., graphs, tables, animations). How you chose to display your data may be based
on your project objective and/or the audience you are sharing your results with.
Some suggestions include:
92
-------
• Graphs: Time series plots that show sensor data compared to regulatory
monitor data; time series plots
that show how concentrations
change over time; maps that
show how concentrations differ
in space (see Figure 3-16).
• Tables: Show numerical sensor
data at different locations or at
different time to show areas
where/when concentrations
were elevated.
• Animations: Video-like images
that show pollutant
concentrations as they change
in time and/or space.
3.8.3 Take Action
One advantage of air sensors is that allow for measurement of air quality often in real-time
in more locations. This local information can empower a variety of decisions, specifically
behavioral changes to reduce emission of pollutants and to reduce your exposure to
pollutants. Both these actions can improve the air quality and improve your health. Here are
some examples of ways to take action based on air sensor data:
• Adjust your outdoor activities (e.g., walking, exercising, running errands)
biking, gardening,) if sensors measure higher pollution levels. Carry out your
activities when pollution levels are lower.
o Schools choose to have an activity indoors rather than outdoors or move
an activity indoors when sensors show that air pollution is high on a given
day.
o Avoid a busy road and use a different route if sensors are measuring
higher pollution on a given day.
• Adjust your indoor activities if sensors measure higher pollution levels.
Consider ways to adjust your habits or to clean the space to reduce pollution
levels.
o Grill outside or open windows to increase ventilation to reduce indoor
pollution from cooking,
o Choose cleaning products with fewer VOCs and fragrances,
o Reduce smoking and candle, incense or wood (fireplace) burning indoors
to eliminate a pollution source,
o More frequently clean pet hair or dust to reduce indoor PM2.5
concentrations.
o Run a mechanical (e.g., device with a filter) or low-ozone producing
electronic air cleaner if an indoor sensor measures elevated pollutant
93
-------
concentrations. The California Air Resources Board (CARB) provides a list
of certified air cleaning devices,
o Build a D1Y air cleaner if a nearby air sensor indicates elevated levels of
PIVte.sdue to wildfire smoke.
• Look for improvements in air quality after implementing a pollution
reduction strategy.
o Anti-idling programs at schools may reduce measured concentrations of
NO2 and PM2.5 in drop-off and pick-up areas,
o A vegetative barrier installed along a roadway may reduce the amount of
pollution impacting a populated area nearby,
o Fireplace change out programs or burn bans may reduce PM
concentrations.
Ways to use air sensor data to change behavior is an active area of research and is likely
to evolve as air sensors become more accurate and deployments more widespread. Air
sensor technologies may significantly shape how individuals and communities perceive and
respond to information about their air quality. Given the complexity of the science of air
pollution and rapidly evolving air sensor technology (e.g., improvements in ease-of-use,
data quality, interpretation), how individuals and communities use air sensor data for
personal action can depend on many factors, as shown in Figure 3-17.
Attributes of technology (cost,
quality, usability, availability)
Impacts of personalized
data on thoughts,
attitudes, decisions
Sensor design
Power dynamics
between communities,
polluters, regulators
Response
Community: Engagement, perceptions,
trust, demographics, history, disparities,
power, local knowledge
Communications
of findings
Marketing / sales of
technology
Citizen scientist,
technical expert roles
Social analysis of data
Data analysis
Attention /
information fatigue
Potential confirmation
and motivation biases
Figure 3-17. Factors that Can Contribute to how Individuals or Communities use Air
Sensor Data for Personal Action (Source: Understanding social and behavioral
drivers and impacts of air quality sensor use)
94
-------
Resources for More Information
• Personal Strategies to Minimize Effects of Air Pollution on Respiratory Health:
Advice for Providers, Patients and the Public, Carlsten C., S. Salvi, G.W.K.
Wong, K.F. Chung. European Respiratory Journal 55(6), 2020
o Paper provides guidance based on findings from published literature to assist
health care providers, patients, public health officials, and the public to reduce
exposure to indoor and outdoor air pollution
o https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7270362/
• The AirSensor Open-source R-package and DataViewer Web Application for
Interpreting Community Data Collected by Low-cost Sensor Networks,
Feenstra B., A. Collier-Oxandale, V. Papapostolou, D. Cocker, and A. Polidori.
Environmental Modelling & Software 134, 2020
o Paper summarizes the development of two software systems to assist in
visualizing and understanding air sensor data collected by community
networks. AirSensor is an open-source R package that allows users to access
historical data, add geospatial metadata, and visualize community monitoring
data using maps and plots. DataViewer is an interactive web application that
incorporates the functionality and data plotting functions of AirSensor for
interpreting and communicating community data collected by low-cost sensor
networks
o https://www.sciencedirect.com/science/article/pii/S1364815220308896
• AirSensor v1.0: Enhancements to the Open-Source R Package to Enable Deep
Understanding of the Long-Term Performance and Reliability of PurpleAir
Sensors, Collier-Oxandale A., B. Feenstra, , V. Papapostolou, and A. Polidori.
Environmental Modelling & Software 148, 2022
o Paper describes the enhancements made to the open-source R package
AirSensor (version 1.0) and the web application DataViewer (version 1.0.1).
to support data access, processing, analysis, and visualization for the
PurpleAir PA-II sensor. The paper also demonstrates how the enhancements
help track and assess the health of air sensors in real-time and historically
o https://www.sciencedirect.com/science/article/pii/S136481522100298X
95
-------
Understanding Social and Behavioral Drivers and Impacts of Air Quality
Sensor Use, Hubbell B.J., A. Kaufman, L. Rivers L, K. Schulte, G. Hagler, J.
Clougherty, W. Cascio, and D. Costa. Science of the Total Environment 621 (2018):
886-894
o Paper discusses the social science research conducted on air sensor use and
identifies: (1) research opportunities between the social and environmental
sciences and the entities involved in developing, testing, and deploying air
sensor technologies; (2) the challenges associated with sensor data
generation, interpretation, and analysis; and (3) collaboration opportunities for
communities and organizations to better understand the reasons and
approaches for using sensors and how technological innovations may
improve the ability to reduce exposures to air pollution
o https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6705391/
A Visual Analytics Approach for Station-Based Air Quality Data, Du, Y., C. Ma,
C. Wu, X. Xu, Y. Guo, Y, Zhou, and J. Li. Sensors 17(1), 2016
o Paper proposes a comprehensive visual analysis system (AirVis) for air
quality analysis that integrates several visual methods, such as map-based
views, calendar views, and trends views, to analyze multi-dimensional
spatiotemporal air quality data
o https://www.mdpi.eom/1424-8220/17/1/30/htm
AtmoVis: Visualization of Air Quality Data, Powley, B., Master of Science Thesis,
Victoria University of Wellington, New Zealand, 2019
o Document discusses the results from a search of literature regarding systems
and methods for visualizing and evaluating air pollution and presents AtmoVis
- a web-based system that includes visualizations for site view, line plot, heat
calendar, monthly rose, monthly averages, and data comparisons
o https://homepaaes.ecs.vuw.ac.nz/~djp/files/MSc BenPowlev 2019.pdf
Openair-An R Package for Air Quality Data Analysis, Carslaw, D C. and K.
Ropkins. Environmental Modelling & Software 27-28, 2012
o Openair is an R package used extensively in academia and in the public and
private sectors that analyzes air quality data and atmospheric composition
data
o https://davidcarslaw.github.io/openair
o https://bookdown.org/david carslaw/openair/
U.S. EPA Real Time Geospatial Data Viewer (RETIGO)
o REal Time GeOspatial Data Viewer (RETIGO) is a free, web-based tool that
can be used to explore stationary or mobile environmental data that you have
collected; nearby public air quality and meteorological data can be added to
the display
o https://www.epa.gov/hesc/real-time-geospatial-data-viewer-retigo
96
-------
U.S. EPA AirData: Air Quality Data Collected at Outdoor Monitoring Stations
Across the U.S.
o A website providing tools and access to recent and historical air quality
information for the U.S., Puerto Rico, and the U.S. Virgin Islands; view data
on an interactive mapping application; obtain information about each monitor;
and download daily and annual concentration data, AQI data, and speciated
particle pollution data (primarily from U.S. EPA's Air Quality System
database)
o https://www.epa.qov/outdoor-air-qualitv-data
Community in Action: A Comprehensive Guidebook on Air Quality Sensors,
South Coast Air Quality Management District (South Coast AQMD), Air Quality
Sensor Performance Evaluation Center (AQ-SPEC), September 2021
o Guidebook for community organizations that covers planning for monitoring
using sensors; sensor deployment, use, and maintenance: and data handling,
interpretation, and communication
o http://www.aqmd.gov/aq-spec/special-proiects/star-qrant
California Air Resources Board (CARB) List of CARB-Certified Air Cleaning
Devices
o A website providing a table of CARB-certified air cleaning devices, a
description of the difference between mechanical and electronic air cleaners,
and resources to help you select a safe and effective air cleaner
o https://ww2.arb.ca.oov/list-carb-certified-air-cleaninq-devices
U.S. EPA Research on Do-lt-Yourself (DIY) Air Cleaners to Reduce Wildfire
Smoke Indoors
o A website providing on overview of EPA research conducted to evaluate the
safety and effectiveness of DIY air cleaners; includes a summary and links to
the Underwriters Laboratories (UL) safety report findings, answers to
frequently asked questions, and links to helpful resources
o https://www.epa.gov/air-research/research-div-air-cleaners-reduce-wildfire-
srnoke-indoors
97
-------
Chapter 4
Sensor Performance Guidance
Choosing an air sensor that is most appropriate for an intended application can be
challenging. Because air sensor data quality is highly variable, sensor data is compared
with data from a reference instrument to describe sensor performance. Reviewing available
performance information can help users select a sensor that is appropriate for their
intended use.
This chapter provides:
• An overview of sensor performance guidance,
• Information about sensor performance evaluations,
• Approaches used to evaluate sensor performance, and
• Information about how to select sensors based on evaluation reports and other
information
98
-------
4.1 Overview of Sensor Performance
There are many air sensors available to the public and new options continue to become
available to meet the growing demand for affordable air monitoring technologies. Choosing
an air sensor that best fits your application of interest can be challenging. The quality of
sensor data can vary, with some sensors producing reliable and interpretable data and
others generating numbers that are not related to pollution concentrations. Because of
these differences, it is important to consider how a sensor performs before purchasing a
device. Sensor performance is a term used to describe how well a sensor works relative
to a reference instrument to determine how much confidence we should have in the data
produced by a sensor. Having an understanding of the approaches used to determine
sensor performance, where to find that information, and how to interpret the information
can help users select a sensor that best suits their application.
4.2 Sensor Performance Evaluations
A sensor performance evaluation is a
test that compares sensor data to
reference instrument data (see example
setup in Figure 4-1). Reference
instruments are used as they provide
highly accurate measurements and are
the "gold standard". Sensor
performance evaluations are needed
because air sensor data quality is highly
variable. Sensors and the way their data
is processed is constantly changing and
improving, which can also impact
performance. Sensor performance
evaluations can address common
concerns of sensors as summarized in
Figure 4-2. This information can also
help sensor users in the planning
phase of a project to select a sensor
(see Section 3.4 and Figure 3-3).
Figure 4-1. Air Sensors on Tripods (in
foreground) with Reference Instruments (in
the background) to Evaluate Sensor
Performance. Photo Credit: South Coast Air
Quality Monitoring District (AQMD)
99
-------
Ability to measure pollutant of
interest
• Does the sensor measure the pollutant of interest accurately
and reliably within the expected concentration range of the
application?
Performance under different
environmental conditions
• How do factors such as relative humidity, temperature, and
different pollutant concentrations and types impact sensor
measurements?
Ability to measure target
pollutant in a pollutant mixture
• Will the sensor measure the target pollutant in a mixture of
other pollutants?
Performance over time
• How does the sensor response change over time?
• When do the sensor readings become inaccurate or unreliable?
Performance out-of-the-box
• How does the sensor perform out-of-the-box?
• Are corrections or adjustments needed to provide more
accurate data?
Useful Lifetime
L.
• Does how long a sensor runs change how the sensor responds?
• Is the lifetime of the sensor impacted by concentration range
or whether the sensor is in use or not?
Figure 4-2. Common Concerns Related to Sensor Performance
A number of sensor performance evaluations have been conducted and results are publicly
available from many testing organizations, both within the U.S. and internationally,
including:
• U.S. EPA Office of Research and Development (ORD)
• South Coast Air Quality Management District (South Coast AQMD) Air Quality
Sensor Performance Evaluation Center (AQ-SPEC)
• European Commission Joint Research Centre (JRC)
• Airparif AIRLABS Microsensors Challenge
• Academic researchers
• Sensor manufacturers
We expect that additional evaluation centers, laboratories, or organizations may launch in
the coming years.
Sensor users should keep in mind that not all sensors, including do-it-yourself (DIY)
sensors, have performance evaluations. For evaluations that are available, perspective
users should check whether the test conditions were similar to their intended study area or
whether changes have been made to the product hardware or firmware since the
evaluation was conducted. Additionally, while a sensor may have a performance
evaluation, it is important to understand if the evaluation is unbiased and objective. In other
words, did the organization that conducted the evaluation do so in a fair and impartial
manner.
100
-------
Although testing organizations, like those previously mentioned, evaluate sensors in
different locations and can give users an idea of how a sensor is expected to perform, there
might be a need to do an evaluation on your own. If this is not feasible due to resources
(e.g., funding, equipment, expertise) or
other reasons, you may consider doing
the following:
• Ask a testing organization if they
have already conducted a sensor
evaluation for a device you are
interested in. Sometimes
evaluations have been conducted
but the results have not been
shared yet publicly.
• Ask a testing organization if they
are planning to evaluate a sensor
you are interested in using or if
they might be willing to conduct an
evaluation.
Sensor users should remember to check
and evaluate a sensor's performance
within the study area ideally before,
during, and after a study if possible. This
check is referred to as collocation (see
Section 3.6) and is a necessary part of
any sensor project.
What is the Difference Between a Sensor
Performance Evaluation and Collocation?
Sensor performance evaluations provide
information about how well a sensor
performs relative to a reference instrument..
A performance evaluation, at the core, is a
collocation. However, it is a collocation
conducted by an objective testing
organization for the specific purpose of
informing others about how well the sensor
agrees with reference instruments under the
test conditions. Manufacturers may use the
results to improve their sensor products.
Potential sensor users may use the results
to decide what sensor(s) to buy for a project.
It is a best practice to collocate air sensors
with reference instruments after purchase,
as discussed in Section 3.6, to check their
performance under outdoor conditions in the
desired monitoring location.
Resources for More Information
• U.S. EPA Air Sensor Performance Evaluations
o Collection of results for field and laboratory evaluations of air sensors
conducted by U.S. EPA Office of Research and Development
o https://www.epa.qov/air-sensor-toolbox/evaluation-emerging-air-sensor-
performance
• South Coast Air Quality Management District (South Coast AQMD) Air Quality
Sensor Performance Evaluation Center (AQ-SPEC)
o Field and laboratory evaluations of commercially available air sensors
conducted by AQ-SPEC
o http://www.agmd.gov/aq-spec
101
-------
• European Commission Joint Research Centre (JRC)
o Results for field and laboratory evaluations of air sensors primary in the form
of scientific journal articles and reports
o https://publications.irc.ec.europa.eu/repositorv/
• AirParif AIRLABS Microsensors Challenge
o Results from an international challenge that promotes innovation and helps
inform users on the performance of air sensors in different applications
o http://www.airlab.solutions/en/proiects/microsensor-challenqe
4.3 Approaches Used to Evaluate Sensor Performance
There are two main approaches for evaluating air sensor performance: field evaluations
and laboratory evaluations. Table 4-1 briefly summarizes these approaches and their
purpose in more detail.
Table 4-1. Common Approaches for Evaluating Air Sensor Performance
Evaluation
Approach
Description
Purpose
Field
Sensors evaluated in
the field at an
ambient (outdoor)
fixed site
• Gives information on how a sensor
performs in real-world, outdoor conditions
• Gives users information on how they might
expect a sensor to perform in similar
outdoor conditions
Laboratory
Sensors evaluated in
a controlled
laboratory setting
• Allows us to study a range of conditions
that may be more difficult to come across
outdoors
• Allows us to better understand certain
performance parameters that are difficult to
test outdoors
Field evaluations typically involve collocating one or more sensors side-by-side with a
reference instrument(s) for an extended period of time (e.g., days, months, years) outdoors.
In field evaluations, sensors experience typical changes in pollutant concentrations and
daily swings in temperature (T) and relative humidity (RH). Sensors may also experience
typical weather conditions (e.g., rain, fog, snow, high winds) that may impact how a sensor
performs. When operating outdoors, sensors will also experience different pollutant
mixtures which can test how well a sensor can detect the target pollutant. One downside of
a field evaluation is that outdoor conditions cannot be controlled. This means that it will be
impossible to understand how a sensor will perform under conditions that are not present.
For example, an evaluation during the summer will not tell you how a sensor may perform
in colder temperatures. Additionally, it can be difficult to figure out if changes in the
102
-------
pollutant concentration level, pollutant mixture, or environmental conditions are affecting
sensor performance, especially when these conditions change at the same time.
Laboratory evaluations are typically short-term evaluations (e.g., hours) conducted in an
environmentally controlled chamber. One or more sensors and the inlet(s) of a reference
instrument(s) are placed in the chamber and exposed to different test conditions (e.g., high
or low temperatures, high or low humidity, target and interferent pollutant concentrations
and types) and the measurements are compared. Laboratory evaluations are useful as they
can provide information on how sensors perform in specific environmental conditions that
may commonly or rarely happen outdoors. However, an environmental chamber can only
mimic specific real-world, outdoor environmental conditions. For example, a laboratory
cannot simulate the size and chemical composition of particles in the outdoor air nor
conditions like fog. While laboratory evaluations are extremely useful, they are more
expensive because they require complex equipment and skilled staff to run the tests. In
addition, the results represent sensor performance for a specific set of conditions which
may or may not occur outdoors.
As discussed, there are pros and cons associated with each sensor performance
evaluation approach. This is often why both approaches are recommended since they
complement each other, and the information collected overall is useful to better understand
sensor performance. Based on the purpose for monitoring, sensor users will need to decide
what performance evaluation approaches can best inform what sensors are selected for a
project. This decision may depend on the availability of resources (e.g., funding,
knowledge, expertise) and access to field sites and/or laboratory testing facilities. Ideally, at
a minimum, field testing in the location where sensors will be used is recommended.
103
-------
4.3.1 U.S. EPA Recommendations on Evaluating Sensor Performance
Recognizing the need for a
consistent approach for evaluating
air sensor performance, the U.S.
EPA published reports (herein
called Targets Reports') that
provide recommendations on how
to evaluate air sensors that
measure criteria pollutants. The
U.S. EPA's recommendations
provide a standardized, objective,
and streamlined approach for
evaluating air sensor performance.
The U.S. EPA based its
recommendations on the current
state-of-the-science, literature
reviews, findings from other sensor
evaluation organizations, sensor
standards/certification programs
(both existing and in development)
by other organizations, and the
U.S. EPA expertise in sensor
evaluation. The Targets Reports
include the following:
• Testing protocols - step-by-step instructions for setting up instruments, testing
instruments, and collecting data
• Performance metrics - parameters used to describe data quality and details on
how to calculate them
• Target values - recommended values that provide a benchmark to understand
sensor performance
The evaluations include both field testing (called base testing) and laboratory testing
(called enhanced testing). At minimum, base testing is recommended. The testing
protocols are specifically designed for sensors used in ambient, outdoor, fixed site
environments for non-regulatory supplemental and informational monitoring (NSIM)
applications (see Section 1.1, Table 1-1). As a brief summary, NSIM categories and
specific examples of applications include:
• Spatiotemporal Variability - daily trends, gradient studies, air quality forecasting,
participatory science, education
• Comparison - hotspot detection, data fusion, emergency response, supplemental
monitoring
• Long-term Trend - Long-term changes, epidemiological studies, model verification
/ \
What is the Difference Between EPA's
Recommendations for Evaluating Sensors and
Sensor Standards/Certification Programs?
U.S. EPA's recommended protocols are entirely
voluntary and testing results that meet some or all
of the targets does not constitute certification.
Further, U.S. EPA does not endorse or
recommend any specific product.
Setting sensor standards is a voluntary process
where technology testing methods are agreed
upon by authorities, manufacturers, customers,
and others invested in the performance of the
technology.
Certification is a process where an organization
carries out an agreed upon test methods set by
standards to make sure tests are conducted in the
same way every time. The certification process
often results in a certificate or specific label.
v J
104
-------
As part of the Targets Reports, the base and enhanced testing protocols have a reporting
template to encourage testers to present evaluation results using a similar format.
Information on how to interpret these reports is provided in Appendix E. Additionally, an
EPA-developed Python code library called sensortoolkit is available to help testing
organizations calculate the performance metrics and generate the evaluation report using
the reporting template. Links to the reports, reporting templates, and Python code library
are all available from EPA's Air Sensor Toolbox webpage.
The intended audience for U.S. EPA's Targets Reports includes testing organizations (e.g.,
routine evaluation organizations, sensor manufacturers); although, sensor users may also
choose to perform the testing protocols. Conducting the protocols is entirely voluntary.
Additionally, the results from the evaluations do not constitute certification or endorsement
by the U.S. EPA. The testing results are meant to inform sensor users.
4.3.2 Guidance from other Organizations on Evaluating Sensor Performance
A number of organizations have developed or are in the process of developing guidance on
conducting sensor performance evaluations, sensor targets/standards, or sensor
certification programs. Examples of these organizations include:
• ASTM International - developing standards of practice and test methods for field
and laboratory evaluations of ambient and indoor air sensors measuring common air
pollutants.
• European Union/European Committee for Standardization (EU/CEN) -
developing field and laboratory certification procedures for multiple tiers of sensor
applications.
• China Ministry of Ecology and Environment (MEE) - developed field and
laboratory test procedures and performance standards for air sensors.
Resources for More Information
• U.S. EPA Performance Testing Protocols, Metrics, and Target Values for Air
Sensors - Use in Ambient, Outdoor, Fixed Site, Non-Regulatory Supplemental
and Informational Monitoring Applications
o Reports that provide recommended testing protocols (field and laboratory),
performance metrics (parameters used to describe sensor data quality), and
target levels to evaluate air sensors that measure criteria air pollutants
o https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-
testinq-protocols
105
-------
• ASTM WK64899 "New Practice for Performance Evaluation of Ambient Air
Quality Sensors and Other Sensor-Based Instruments"
o Provides information on a practice for evaluating the performance of air
quality sensors in ambient air
o https://www.astm.org/workitem-wk64899
• EU/CEN/TC 264/WG 42 "Air quality - Performance evaluation of air quality
sensor systems - Part 1: Gaseous pollutants in ambient air"
o Outlines testing procedures and requirement for classifying performance of air
quality sensors for the monitoring of gaseous pollutants
o https://www.en-standard.eu/pd-cen-ts-17660-1 -2021 -air-gualitv-performance-
evaluation-of-air-qualitv-sensor-svstems-qaseous-pollutants-in-ambient-air/
• Air Quality Sensor Performance Evaluation Center (AQ-SPEC) of the South
Coast Air Quality Management District (South Coast AQMD) Website
o Website for the AQ-SPEC program which describes how field and laboratory
tests are conducted
o Field Evaluation Protocol and Reports: http://www.agmd.gov/ag-
spec/evaluations/field
o Laboratory Evaluation Protocol and Reports: http://www.agmd.gov/ag-
spec/evaluations/laboratorv
4.4 How to Select Sensors Based on Evaluation Reports or Information
Reviewing sensor performance evaluation reports and related information can help you
select an appropriate sensor for your project. Some questions to ask when reviewing an
evaluation report and determining whether a sensor would be suitable for your monitoring
project application include:
• Trends: How well do the changes in sensor measurements mimic the change in
pollutant concentrations that are measured by the reference instrument?
• Precision: How consistent are the concentration measurements obtained by
sensors of the same make, model, and firmware version that are operated under the
same field conditions?
106
-------
• Bias: How closely do the sensor measurements agree with measurements made by
a collocated reference instrument?
• Concentration Range: Does the
operating range of the sensor
cover the range of pollutant
concentrations expected at the
desired monitoring location?
• Meteorology: Is the sensor
response affected by
meteorological conditions (e.g.,
RH, T)?
• Specificity: Does the sensor
measure the target pollutant? Are
the sensor measurements affected
by an interferent(s)?
• Drift: Is the sensor response to
pollutant concentrations stable
over time?
Going over the objectives of your project (see Section 3.2) and your plan for obtaining
measurements (see Section 3 3), will help you decide on which information is most
important to consider when reviewing performance evaluation results. Let's consider two
examples:
• Example #1: You would like to
setup a network of sensors for a
long-term deployment in a
community to supplement an
existing regulatory monitoring
network For this case, you are
likely most interested in a
quantitative measurement of
pollutant concentrations. Therefore,
you would be most concerned with a
sensor's precision, bias, and
response to meteorology (i.e., T and
RH). The sensors should provide an
acceptable degree of agreement
between the network of sensors (precision) and between the sensors and reference
instrument measurements (bias) to allow you to confidently compare sensor data to
the regulatory monitoring network and to air quality and health standards. Because
the sensors need to provide reasonable measurements under all meteorological
107
What are the Definitions of Precision,
Bias, and Drift?
Bias: The systematic or persistent
disagreement between concentration
reported by the sensor and reference
instrument.
Precision: The variation around the mean
(average) of a set of measurements
obtained at the same time from two or more
sensors of the same type collocated under
the same environmental conditions.
Drift: A change in the response or
concentration reported by a sensor when
challenged by the same pollutant
concentration over an operating timeframe.
s \
What is the Difference Between
Quantitative and Qualitative
Measurements?
Quantitative measurements can be
expressed using numbers. For example, a
pollutant concentration expressed in parts
per billion (ppb).
Qualitative measurements are descriptive,
based on concepts, and often expressed in
words. For example, pollutant
concentrations described as "higher" or
"lower".
-------
conditions expected over the deployment period, you may be interested in knowing if
the sensor was tested under similar meteorological conditions and whether bias
changes as RH changes.
• Example #2: You would like to use one sensor to measure an air pollutant in
your backyard to determine when air quality is best for outdoor activities. In
this case, you are likely interested in a qualitative measurement. In other words, a
sensor that reliably provides information about the relative difference between high
and low pollutant concentrations (trends) might be sufficient.
Figure 4-3 provides a flow chart with considerations on how to select sensors based on
their performance. As mentioned in Section 4.1, many organizations and manufacturers
themselves conduct sensor performance evaluations. These can normally be found on the
organization or manufacturers' webpages or in scientific publications (e.g., presentations,
journal articles).
108
-------
Can you evaluate the sensor?
1
Suggest that the manufacturer
conduct the evaluation, enlist
some support to conduct the
evaluation, or choose a
different sensor
Conduct the
evaluation
and review
the results
Have modifications been made since this
evaluation (e.g., new sensors incorporated, new
firmware, new data correction algorithms)?
Do you understand what data, outside of the
sensor itself, was used to correct the data
during this evaluation? Is that information
available where you intend to use the sensor?
Existing data may not represent expected
performance where you intend to use the
sensor, local evaluation is essential
yes
Do the results suggest
sufficient performance for
your intended application?
no
Another sensor or measurement device
may be better for this application
yes
Was this a field or
laboratory evaluation?
Was the environment similar to where you
intend to use the sensor (e.g., concentration
range, pollutant sources, meteorology)?
Did the test environment cover a similar
range of conditions to where you intend to
use the sensor (e.g., concentration range,
temperature, relative humidity)?
Sensor is likely a good choice. Check other
sensor specs to be sure they meet your
needs. Local collocation before use is ideal,
Conduct additional evaluation
to be sure it meets your needs
Sensor is likely a good choice but the pollutant
mixture in the real world may be different. Check
other sensor specs to be sure they meet your
needs. Local evaluation is highly recommended.
Figure 4-3. Flow Chart for Considering an Air Sensor Based on Performance
109
-------
As you review sensor performance evaluations across different testing organizations, look
for as much information as possible to understand the performance of a device. Remember
that the level of detail and transparency can be different from one evaluation to the next.
Below are some important notes to keep in mind and questions to ask about the
evaluations to help you better understand the testing and results:
1. Evaluation protocols/procedures may be similar but not identical. Those
conducting the tests may have different protocols, laboratory test pollutant
concentrations, or report different performance metrics. Graphs or plots describing
performance may look similar but may be generated or plotted differently. The
terminology used and how the performance metric(s) is calcuated may vary. For
example, precision might be calculated as either the standard deviation, relative
standard deviation, or relative percent difference. Evaluations may use different
types of reference instruments. Use of a FRM/FEM instrument(s) will provide greater
confidence in actual pollutant concentrations than the use of another mid to lower-
cost sensor. There can be some differences in the results depending on which
FRM/FEM is used.
• Were both field and laboratory evaluations conducted?
• Was an FRM/FEM instrument used? Which one?
• Do you understand the performance metrics and how they were calculated?
2. Locations are not widespread. Generally, evaluations are conducted at a limited
number of locations which may or may not be similar to the environment where you
want to use the sensor. The evaluation results may not represent how a sensor
performs in the environment you will use them in if that environment is very different
(e.g., different pollutant concentration, composition, or sources).
• Where was the test conducted?
3. Environmental conditions are limited.
Consider whether the sensor was tested in
an environment similar to where you will be
using the sensor. Generally, sensor
evaluations are conducted outdoors where T,
RH, and other weather conditions vary.
These conditions may not be similar to the
environment where you want to place the
sensor(s). Similarly, evaluation results may
not represent sensor performance in the
environment you will use them in if that
environment is very different (e.g., T and RH
often outside of the tested range).
/ \
Tip: Find performance
evaluations that match the
conditions you expect in your
monitoring study
To the extent possible, find
performance evaluations for the
type of sensor and deployment
conditions (e.g., pollutant
concentration, RH, T) anticipated
in your monitoring study.
I J
110
-------
• Was the test conducted in an environment similar to where you will be using
the sensor?
4. Sensor performance is variable. It is a standard practice to test 3 or more identical
sensors at the same time because it provides information on the variation in
performance that can occur among identical sensors. Performance can vary based
on the make, model, and firmware version of the device and even among devices of
the same type. Past evaluation efforts have shown that two sensor packages that
use the same internal component can perform differently. These differences can be
a result of how the components are arranged, how they sample the air, or due to a
build up of heat inside of the sensor housing. Additionally, manufacturers may use a
mathematical equation or model to convert the output from the sensor into a
pollutant concentration. Differences in data processing can change sensor
performance. Firmware changes can also impact sensor performance.
• Did all sensors tested perform similarly?
• Did the sensors tested have the same make, model, configuration, and
firmware as the sensors you intend to use?
5. Evaluations are conducted for a finite amount of time. Typically, sensor
performance evaluations are conducted for 30 days or more as it allows for more
variation in pollutant concentrations and environmental conditions. Shorter
evaluations may not be able to capture these variations. Longer evaluations may be
needed to understand how sensor performance changes over time and sensor
lifetime. Unexpected events such as power outages, equipment failure or damage,
inclement or severe weather (e.g., thunderstorms, hurricanes), or pollution episodes
(e.g., fireworks, dust storms, volcanic eruption) can influence evaluation results or
cause missing data.
• How long was the evaluation (e.g., 7 days, 30 days, 1 year)?
6. Were there any difficulties or anomalies encountered during the test that might
influence the results? The evaluation may not represent your intended
application. A sensor may have been tested near a source (e.g., roadway, industry)
where pollutant concentrations are much higher than where you intend to use the
sensor.
• Was the sensor performance evaluation conducted in an enviroment similar
to your application of interest?
111
-------
Resources for More Information
• U.S. EPA Performance Testing Protocols, Metrics, and Target Values for Air
Sensors - Use in Ambient, Outdoor, Fixed Site, Non-Regulatory Supplemental
and Informational Monitoring Applications
o Reports that provide recommended testing protocols (field and laboratory),
performance metrics (parameters used to describe sensor data quality), and
target levels to evaluate air sensors that measure criteria air pollutants
o Appendices of the Targets Reports provide standardized reporting templates
o https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-
testing-protocols
• U.S.-EPA developed sensortoolkit python code library
o Code library for evaluating air sensor data collocated with reference
instruments; code can be used to calculate performance metrics
o GitHub: https://github.com/USEPA/sensortoolkit
o PyPI: https://pypi.org/proiect/sensortoolkit/
• U.S.-EPA PM2.5 Continuous FEM Monitor Comparability Assessments
o Tool providing a one-page technical report that assesses the comparability of
a PM2.5 continuous FEM monitors when collocated with an FRM sampler
o https://www.epa.gov/outdoar-air-gualitv-data/pm25-continuQus-monitor-
comparabilitv-assessments
• Air Quality Sensor Performance Evaluation Center (AQ-SPEC) of the South
Coast Air Quality Management District (South Coast AQMD) Website
o Website for the AQ-SPEC program which describes how field and laboratory
tests are conducted
o Field Evaluation Protocol and Reports: http://www.agmd.gov/ag-
spec/evaluations/field
o Laboratory Evaluation Protocol and Reports: http://www.agmd.gov/ag-
spec/evaluations/laboratorv
112
-------
Appendix A: Resources
A.1 Introduction to Air Sensors
• U.S. EPA's Air Sensor Toolbox
o Information and resources for topics related to
other organizations and resources that sensor
o https://www.epa.gov/air-sensor-toolbox
A.2 Air Quality 101
A.2.1 Outdoor Air Quality and Air Pollution
• U.S. EPA Air Quality Planning and Standards Website
o Provides additional information regarding air quality and pollutants
o https://www3.epa.gov/airaualitv/
• U.S. EPA National Air Quality - Status and Trends of Key Air Pollutants
Website
o Provides air quality trends, reports, and summaries for criteria air pollutants
o https://www.epa.gov/air-trends
• U.S. EPA AirNow Website
o Provides a variety of resources on air quality including air quality information
at local, state, national, and world views, air quality and health, maps and
data, educational resources, and more
o https://www.airnow.gov/
• Wildfire Smoke: A Guide for Public Health Officials, U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, EPA-452/R-19-
901, August 2019
o Document provides guidance to state, tribal, and local public health officials
and other interested groups (e.g., health professionals, air quality officials,
public) in preparing for wildfire smoke events and in communicating health
risks and taking measures to protect the public during smoke events
o https://www.aimow.gov/sites/default/files/2021-05/wildfire-smoke-guide-
revised-2019.pdf
air sensors; includes links to
users may find helpful
A-1
-------
U.S. EPA Mobile Source Pollution and Related Health Effects Website
o Overviews mobile sources of air pollution, summarizes health effects
associated with exposure to mobile source emissions, provides data and
modeling resources, and information on programs to reduce mobile source
pollution
o https://www.epa.gov/mobile-source-pollution
U.S. EPA Near-Roadway and Other Near-Source Pollution Website
o Overview of research on near-roadway pollution from cars, trucks, and other
mobile sources and frequently asked questions about near-roadway air
pollution and health effects
o https://www.epa.gov/air-research/research-near-roadwav-and-other-near-
source-air-poliution
Near-Roadway Air Pollution and Health: Frequently Asked Questions, U.S.
Environmental Protection Agency, Office of Transportation and Air Quality, EPA-
420-F-14-044, August 2014
o Document provides U.S. EPA's responses to frequently asked questions
received from the public regarding exposure to near-roadway air pollution
o https://nepis.epa.gov/Exe/ZvPDF.cgi/P100NFFD.PDF?Dockev=P100NFFD.P
DF
Report to Congress on Black Carbon, U.S. Environmental Protection Agency,
EPA-450/R-12-001, March 2012
o Document summarizes available scientific information on the current and
future impacts of black carbon (BC) and evaluates the effectiveness of
available BC mitigation approaches and technologies
o https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100EIJZ.txt
U.S. EPA Integrated Science Assessments (ISAs) for Criteria Air Pollutants
Website
o Reports that summarize scientific information that is the foundation for
reviewing the NAAQS for criteria pollutants; ISAs are an important resource
for state and local health agencies, other federal agencies, and international
health organizations
o https://www.epa.gov/isa
-------
A.2.2 Health And Environmental Effects of Air Pollution
• U.S. EPA Criteria Air Pollutants Website
o Provides detailed information on the six criteria pollutants including basic
information, health and environmental effects, technical documents, setting
and reviewing the standards, implementing the standards, and current air
quality designations
o https://www.epa.gov/criteria-air-pollutants
• Health Effects of Ozone (O3) Pollution Website
o Provides detailed information on health effects of breathing air containing O3
o https://www.epa.gov/ground-level-ozone-pollution/health-effects-ozone-
pollution
• Health and Environmental Effects of Particulate Matter (PM) Website
o Provides detailed information on health and environmental effects of PM
o https://www.epa.gov/pm-pollution/health-and-environmental-effects-
particulate-matter-pm
• Basic Information about Nitrogen Dioxide (NO2) Website
o Provides basic information on NO2 including health and environmental effects
o https://www.epa.goV/no2-pollution/basic-information-about-no2#Effects
• Basic Information about Sulfur Dioxide (SO2) Website
o Provides basic information on SO2 including health and environmental effects
o https://www.epa.gOv/so2-pollution/sulfur-dioxide-basics#effects
• Basic Information about Carbon Monoxide (CO) Outdoor Air Pollution Website
o Provides basic information on CO including health and environmental effects
o https://www.epa.gov/co-pollution/basic-information-about-carbon-monoxide-
co-outdoor-air-pollution#Effects
• Basic Information about Lead (Pb) Air Pollution Website
o Provides basic information on Pb including health and environmental effects
o https://www.epa.gov/co-pollution/basic-information-about-carbon-monoxide-
co-outdoor-air-pollution#Effects
• Report on the Environment - Volatile Organic Compounds (VOCs) Emissions
Website
o Provides detailed information on sources, health and environmental effects,
and emissions estimates of VOCs
o https://cfpub.epa.gov/roe/indicator.cfm?i=23#1
A-3
-------
Definition of VOC Website
o Provides a detailed information on the definition of VOCs as outlined in air
pollution regulations
o https://www.epa.gov/air-emissions-inventories/what-definition-voc
Health Effects of Exposures to Mercury Website
o Provides detailed information on health effects of exposure to mercury
o https://www.epa.gov/mercurv/health-effects-exposures-mercurv
Integrated Risk Information System (IRIS) Methylmercury (MeHg) Summary
Website
o Provides health assessment information on MeHg based on review of toxicity
data
o https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=73
Agency for Toxic Substances and Disease Registry (ATSDR) Public Health
Statement for Benzene Website
o Provides information about benzene and effects of exposure to it
o https://wwwn.cdc.gov/TSP/PHS/PHS.aspx?phsid=37&toxid=14
Integrated Risk Information System (IRIS) Benzene Summary Website
o Provides health assessment information on benzene based on review of
toxicity data
o https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=276
Global Methane Initiative (GMI) Website
o Provides information on GMI, description of methane and mitigation
approaches, methane sites around the globe, and more
o https://www.epa.gov/gmi
Report to Congress on Black Carbon, U.S. Environmental Protection Agency,
EPA-450/R-12-001, March 2012
o Report provides summary on black carbon, health and environmental effects,
emissions, mitigation overview, and more
o https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P 100EIJZ.txt
Integrated Science Assessment (ISA) for Particulate Matter, U.S. Environmental
Protection Agency, EPA/600/R-19/188, December 2019
o ISA provides detailed information on particulate matter including sources,
ambient levels, health and environmental effects, and more
o https://www.epa.gov/isa/integrated-science-assessment-isa-particulate-matter
-------
• Traffic-Related Air Pollution: A Critical Review of the Literature on Emissions,
Exposure, and Health Effects, HEI Panel on the Health Effects of Traffic-Related
Air Pollution, HEI Special Report 17, Health Effects Institute (HEI), January 2010
o Report provides a summary and synthesis of information on air pollution from
traffic and its health effects
o https://www.healtheffects.org/publication/traffic-related-air-pollution-critical-
review-literature-emissions-exposure-and-health
A.2.3 Air Pollution Monitoring
• U.S. EPA Ambient Monitoring Technology Information Center (AMTIC) Website
o Contains technical information regarding ambient air monitoring programs,
including the networks of state and local air monitoring stations (SLAMS),
monitoring methods, and QA/QC procedures
o https://www.epa.gov/amtic
• U.S. EPA Ambient Air Monitoring Website
o Overviews the reasons for why monitoring ambient air quality is important and
provides links to U.S. EPA's AMTIC, Air Quality System (AQS), Air Data,
AirNow, and AirNow International websites
o https://www.epa.gov/air-gualitv-management-process/managing-air-gualitv-
ambient-air-monitoring
• Overview of the Clean Air Act (CAA) Website
o Provides an in-depth overview of the CAA including history and requirements,
role of science and technology, role of state, local, tribal and federal
government, and more
o https://www.epa.gov/clean-air-act-overview
• Videos on Sources of Air Quality Information and Air Sensor Measurements,
Data Quality, and Interpretation
o Educational videos, in both English and Spanish, that can be used to learn
how U.S. EPA collects and uses air quality data, how air quality health risks
are communicated, and how to interpret data collected using air sensors
o https://www.epa.gov/air-sensor-toolbox/videos-air-sensor-measurements-
data-gualitv-and-interpretation
• Understanding Air Quality and Monitoring Video, South Coast Air Quality
Management District (South Coast AQMD), Air Quality Sensor Performance
Evaluation Center (AQ-SPEC), September 2021
A-5
-------
o Educational video providing background on air quality, criteria pollutants,
pollutant sources and health effects, air quality monitoring technologies, and
the role of air sensors
o https://www.voutube.com/watch?v=2r0XxQm50IE
• California Air Resources Board (CARB) Outline of Measurement Technologies
o Online resource that discusses air monitoring applications, applicable
measurement technologies, and their relative availability and cost developed
by CARB to support community air monitoring conducted under California
Assembly Bill 617
o https://ww2.arb.ca.gov/capp-resource-center/communitv-air-
monitorinq/outline-of-measurement-technoloqies
• Hazardous Air Pollutants (HAPs) Website
o Provides detailed information on HAPs including the list of HAPs, health an
environmental effects, sources and exposures, data, and more
o https://www.epa.gov/haps
• Regional Haze Program Website
o Provides information on the Regional Haze Rule and Program, list of the
national parks and wilderness areas covered by the program, and more
o https://www.epa.gov/visibilitv/regional-haze-program
A.2.4 Air Quality Standards and Indices
• WHO National Air Quality Standards Tool
o An interactive tool providing an international map of current national air quality
standards for criteria pollutants for various averaging times
o https://www.who.int/tools/air-gualitv-standards
• Air Quality in Europe 2021
o An annual assessment of recent air quality trends at both European and
national levels
o https://www.eea.europa.eu//publications/air-gualitv-in-europe-2021
• Air Quality System Data Dictionary
o The AQS Data Dictionary describes the fields typically encountered by AQS
users and are listed in alphabetical order; field definitions and calculation
algorithms are provided as appropriate
o https://ags.epa.gov/agsweb/documents/AQS Data Dictionarv.html
A-6
-------
• U.S. EPA National Ambient Air Quality Standards (NAAQS) Table
o A webpage detailing the NAAQS for six criteria pollutants which includes
details from Table 2-4 but is a resource that will be updated if the standards
change
o https://www.epa.gov/criteria-air-pollutants/naaqs-table
• The National Institute for Occupational Safety and Health (NIOSH)
o The Occupational Safety and Health Act of 1970 established NIOSH as a
research agency focused on the study of worker safety and health, and
empowering employers and workers to create safe and healthy workplaces
o https://www.cdc.gov/niosh/index.htm
• Center for Disease Control and Prevention (CDC) National Environmental
Public Health Tracking - Air Quality
o CDC works closely with the U.S. Environmental Protection Agency, the
National Aeronautics and Space Administration (NASA), the National Oceanic
and Atmospheric Association (NOAA), and the National Weather Service to
provide air quality data on the Tracking Network and to better understand how
air pollution affects our health
o https://www.cdc.gov/nceh/tracking/topics/AirQualitv.htm
• Greenbook: Nonattainment Areas for Criteria Pollutants
o The EPA Green Book provides detailed information about area National
Ambient Air Quality Standards (NAAQS) designations, classifications and
nonattainment status
o https://www.epa.gov/green-book
A.2.5 The U.S. Air Quality Index (AQI)
• AirNow Air Quality Index (AQI) Website
o Provides information on AQI basics, air pollutants, action days, and other
resources
o https://www.airnow.gov/agi/
• Technical Assistance Document for Reporting of Daily Air Quality - the Air
Quality Index (AQI), U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, EPA 454/B-18-007, September 2018
o Document provides guidance to aid local agencies in calculating and
reporting the AQI as required in the Code of Federal Regulations (CFR)
o https://www.airnow.gov/publications/air-gualitv-index/technical-assistance-
document-for-reporting-the-dailv-agi/
A-7
-------
AirNow AQI Calculator
o Online tool that converts user-specified AQI values into an equivalent
concentration or converts concentration into AQI values. The tool also
provides the corresponding AQI Category (e.g., good, moderate), health
effects, and cautionary statements
o https://www.airnow.gov/aqi/aqi-calculator/
AirNow - Using the Air Quality Index Website
o Provides an overview of the AQI, AQI forecasts, and the NowCast AQI and
how to use these tools to assess local air quality and plan for outdoor
activities; links on the page provide technical information about NowCast
algorithms and leads to a github code library for calculating the NowCast for
Os
o https://www.airnow.gov/aqi/aqi-basics/usinq-air-qualitv-index/
Air Quality Index: A Guide to Air Quality and Your Health, U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, EPA-456/F-14-002,
February 2014
o Booklet that discusses the importance of air quality and provides an overview
of the AQI; health effects of exposure to ozone (O3), particulate matter (PM),
carbon monoxide (CO), and sulfur dioxide (SO2); and suggested actions to
reduce exposure to unhealthy air for each AQI Category
o https://www.airnow.gov/sites/default/files/2018-04/agi brochure 02 14 O.pdf
Modified Air Quality Index, Improving Accessibility for People with Color
Vision Deficiencies, South Coast Air Quality Management District (South Coast
AQMD), May 2022
o Press release sharing South Coast AQMD's work to develop a modified
version of the AQI that accommodates individuals with color vision
deficiencies while still being similar to the traditional AQI color scale; the scale
was tested against eight common color impairments and remains
distinguishable in grayscale
o http://www.agmd.gov/docs/default-source/news-archive/2022/south-coast-
agmd-modified-AQI-05022022
Air Quality Guide for Nitrogen Dioxide (NO2), U.S. Environmental Protection
Agency, Office of Air and Radiation, EPA-456/F-11-003, February 2011
o Booklet that overviews actions to reduce exposure NO2 near roadways for
each AQI category, provides an overview of NO2 sources and health effects,
and provides tips for reducing NO2 emissions
o https://www.airnow.gov/sites/default/files/2018-06/no2.pdf
A-8
-------
• Air Quality Guide for Ozone (O3), U.S. Environmental Protection Agency, Office of
Air and Radiation, EPA-456/F-15-006, August 2015
o Booklet that overviews actions to reduce exposure to O3 for each AQI
category, provides an overview of O3 sources and health effects, and
provides tips for reducing pollution from O3
o https://www.airnow.gov/sites/default/files/2021-03/air-gualitv-
quide ozone 2015.pdf
• Air Quality Guide for Particle Pollution, U.S. Environmental Protection Agency,
Office of Air and Radiation, EPA-456/F-15-005, August 2015
o Booklet overviews the actions to reduce exposure to particle pollution for
each AQI category, provides an overview of pollution sources, and overviews
health effects and tips for reducing particle pollution
o https://www.airnow.gov/sites/default/files/2021-03/air-gualitv-
quide pm 2015.pdf
A.3 Monitoring Using Air Sensors
A.3.1 Question: Determining a Purpose For Monitoring
• Handbook for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Handbook that covers common expectations for quality assurance and
documentation and best management practices for organizations that train
and use volunteers in the collection of environmental data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgapphandbook 3 5 19 mmedits.pdf
• Guidebook for Developing a Community Air Monitoring Network: Steps,
Lessons, and Recommendations from the Imperial County Community Air
Monitoring Project, Tracking California, October 2018
o Outlines the process and considerations for creating an air monitoring
network using air sensors
o https://trackingcalifornia.org/cms/file/imperial-air-proiect/guidebook
• Community in Action: A Comprehensive Guidebook on Air Quality Sensors,
South Coast Air Quality Management District (South Coast AQMD), Air Quality
Sensor Performance Evaluation Center (AQ-SPEC), September 2021
o Guidebook for community organizations that covers planning for monitoring
using sensors; sensor deployment, use, and maintenance; and data handling,
interpretation, and communication
o http://www.agmd.gov/ag-spec/special-proiects/star-grant
A-9
-------
• Air Sensor Stories, University of Rochester, University of North Carolina at Chapel
Hill, University of Texas Medical Branch, Columbia University, and WE ACT for
Environmental Justice, 2018
o Workshop guide and supporting materials to assist diverse audiences
understand the potential of air sensors in addressing community concerns
about particulate matter pollution; includes an air monitoring action plan
worksheet to help groups think through key questions
o https://www.urmc.rochester.edu/environmental-health-sciences/communitv-
enqaqement-core/proiects-partnerships/air-sensor-stories-workshop.aspx
• Appendix B: Questions to Consider When Planning for and Collecting Air
Sensor Data, and Sharing Your Results {this document)
o Provides a list of questions for consideration to help sensor uses better plan,
collect, and share data
A.3.2 Plan: Developing a Plan
• Guidance for Quality Assurance Project Plans (QA/G-5), U.S. Environmental
Protection Agency, EPA/240/R-02/009, December 2002
o Provides guidance on developing a Quality Assurance Project Plan (QAPP),
which is an important part of the planning process for air quality monitoring
projects
o https://www.epa.gov/sites/default/files/2015-06/documents/g5-final.pdf
• Examples for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Collection of examples that provide tools and procedures to help community
science organizations properly document the quality of data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgappexamples3 5 19 mmedits.pdf
• Templates for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Templates that provide tools and procedures to help properly document the
quality of data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgapptemplates3 5 19 mmedits.pdf
o Editable templates: https://www.epa.gov/citizen-science/gualitv-assurance-
handbook-and-guidance-documents-citizen-science-proiects
A-10
-------
• Community Science Air Monitoring
o Guidance, provided by the New Jersey Department of Environmental
Protection Division of Air Quality - Air Monitoring, on using air sensors for
community projects; includes approaches to using sensors, types of sensors
available, interpreting sensor data, four types of sensor projects and data
quality assurance plan templates for each, and other helpful links
o https://www.ni.gov/dep/airmon/communitv-science.html
• Air Quality Agencies
o Websites that provide a list of state, local, and/or tribal agencies that manage
air quality
o U.S. Environmental Protection Agency, https://www.epa.gov/aboutepa/health-
and-environmental-aqencies-us-states-and-territories
o National Tribal Air Association (NTAA): https://www.ntaatribalair.org/
o National Association of Clean Air Agencies (NA CAA):
https://www.4cleanair.org/agencies/
o Association of Air Pollution Control Agencies (AAPCA):
https://cleanairact.org/about/
A.3.3 Plan: Selecting an Air Sensor
• Chapter 4: Sensor Performance Guidance (this document)
o Provides an overview of laboratory and field sensor performance evaluations;
performance characteristics needed for spatiotemporal variability,
comparison, and long-term trend NSIM applications; and U.S. EPA's
recommendations for sensor testing protocols, performance metrics, and
targets
• Appendix C: Choosing Air Sensors (this document)
o Provides checklists for: (1) what to look for in a sensor before buying, (2) what
to look for in a sensor user manual, and (3) sensor maintenance to ensure
proper functionality and reliable performance
• Performance Testing Protocols, Metrics, and Target Values for Ozone Air
Sensors - Use in Ambient, Outdoor, Fixed Site, Non-Regulatory Supplemental
and Informational Monitoring Applications, U.S. Environmental Protection
Agency, EPA/600/R-20/279. February 2021
o Provides recommended testing protocols (field and laboratory), performance
metrics (parameters used to describe sensor data quality), and target levels
to evaluate ozone air sensors
o https://cfpub.epa.gov/si/si public record Report.cfm?dirEntrvld=350784&Lab
=CEMM
A-11
-------
• U.S. EPA's Performance Targets and Testing Protocols Website
o Summary of the U.S. EPA's research on recommended testing protocols,
metrics, and target values for evaluating the performance of air sensors
o https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-
testing-protocols
• Air Quality Sensor Performance Evaluation Center (AQ-SPEC) of the South
Coast Air Quality Management District (South Coast AQMD) Website
o Website for the AQ-SPEC program which conducts laboratory and field
evaluations of air sensors and provides information to the public regarding
actual sensor performance and the advantages and potential limitations of
using air sensors. AQ-SPEC is operated by South Coast AQMD
o http://www.agmd.gov/ag-spec
• The National Solar Radiation Data Base (NSRDB), Sengupta, M., Y. Xie, A.
Lopez, A. Habte, G. Maclaurin, and J. Shelby. Renewable and Sustainable Energy
Reviews 89 (2018): 51-60
o Paper reviews the complete package of surface observations, models, and
satellite data used for the NSRDB - an open dataset of solar radiation and
meteorological data over the United States and regions of the surrounding
countries
o https://www.sciencedirect.com/science/article/pii/S136403211830087X
A.3.4 Setup: Locating Sites for Air Sensors
• U.S. Code of Federal Regulations (CFR), Title 40 (Protection of Environment),
Chapter 1 (Environmental Protection Agency), Subchapter C (Air Programs),
Part 58 (Ambient Air Quality Surveillance)
o Specifies the regulatory requirements for the U.S. ambient air quality
monitoring network including quality assurance procedures for operating air
quality monitors and handling data; methodology and operating schedules for
monitoring instruments; criteria for siting monitoring instruments; and air
quality data reporting requirements
o https://www.ecfr.gov/current/title-40/chapter-l/subchapter-C/part-
58#ap40.6.58.0000 Onbspnbspnbsp.e
• Quality Assurance Handbook for Air Pollution Measurement Systems, Volume
II, Ambient Air Quality Monitoring Program, U.S. Environmental Protection Agency,
EPA-454/B-17-001, January 2017
o Handbook provides additional information and guidance (including pollutant-
specific spatial scale characteristics) to assist tribal, state, and local
A-12
-------
monitoring organizations in developing and implementing a quality
management system for the Ambient Air Quality Surveillance Program
described in 40 CFR Part 58
o https://www.epa.gov/sites/default/files/2020-
10/documents/final handbook document 1 17.pdf
• Air Quality Agencies
o Websites provide a list of state, local, and/or tribal agencies that manage air
quality
o U.S. Environmental Protection Agency: https://www.epa.gov/aboutepa/health-
and-environmental-agencies-us-states-and-territories
o National Tribal Air Association (NTAA): https://www.ntaatribalair.org/
o National Association of Clean Air Agencies (NACAA):
https://www.4cleanair.org/agencies/
o Association of Air Pollution Control Agencies (AAPCA):
https://cleanairact.org/about/
• Blueprint for the Development and Implementation of Distributed Sensor
Networks, U.S. National Institute of Standards and Technology Global Cities Team
Challenge Transportation SuperCluster
o Blueprint that summarizes lessons learned, best practices, and research
questions for developing and implementing sensor networks
o https://static1 .sguarespace.com/static/5967c18bff7c50a0244ff42c/t/5ad7c41 c
758d464041c7e58a/1524089886422/Distributed Sensor Networks Recomm
endations.pdf
• U.S. EPA Guide to Siting and Installing Air Sensors
o Information and considerations for locating an air sensor in both outdoor and
indoor locations
o https://www.epa.gov/air-sensor-toolbox/guide-siting-and-installing-air-sensors
• South Coast Air Quality Management District - Sensor Siting and Installation
Guide
o Guidance on how to locate and install air sensors: http://www.agmd.gov/ag-
spec/resources/related-documents
o English: http://www.agmd.gov/docs/default-source/ag-spec/resources-
page/ag-spec-sensor-siting-and-installation-guide v1-0-(english).pdf
o Spanish: http://www.agmd.gov/docs/default-source/ag-spec/resources-
page/sensor-siting-and-installation-guide v1 -Q-(spanish).pdf
A-13
-------
• U.S. EPA Air Sensor Toolbox - Air Sensor Research Grants and Challenges
Website
o Website provides information on grants and challenges related to air research
and air sensors
o https://www.epa.gov/air-sensor-toolbox/air-sensor-research-grants-and-
challenges
A.3.5 Setup: Collocation and Correction
• U.S. EPA Air Sensor Collocation Instruction Guide, U.S. Environmental
Protection Agency, Office of Research and Development
o Resource provides background information, links to web-based supporting
materials, and instructions for evaluating the performance of air sensors by
comparing the measurements made by collocated sensors and reference
instruments
o https://www.epa.gov/air-sensor-toolbox/air-sensor-collocation-instruction-
guide
• U.S. EPA Air Sensor Collocation Macro Analysis Tool
o Excel-based tool that helps users compare data from air sensors to data from
reference instruments
o https://www.epa.gov/air-sensor-toolbox/air-sensor-collocation-macro-analvsis-
tool
• Community in Action: A Comprehensive Guidebook on Air Quality Sensors,
South Coast Air Quality Management District (South Coast AQMD), Air Quality
Sensor Performance Evaluation Center (AQ-SPEC), September 2021
o Guidebook for community organizations that covers planning for monitoring
using sensors; sensor deployment, use, and maintenance; and data handling,
interpretation, and communication
o http://www.agmd.gov/ag-spec/special-proiects/star-grant
• South Coast AQMD Low-Cost Sensor Data Analysis Guide
o Guide that provides some brief instructions to help community scientists
interact with the data they are collecting as well as some questions to help
guide their analysis
o http://www.agmd.gov/docs/default-source/ag-spec/star-grant/air-gualitv-
sensor-data-analvsis-guide.pdf?sfvrsn=6
A-14
-------
A.3.6 Collect: Data Collection, Quality Assurance/Quality Control, and Data
Management
• AirSensor and DataViewer Tools (R package)
o AirSensor is an open-source R package that allows users to access historical
data, add spatial metadata, and visualize community monitoring data through
maps and plots
o DataViewer is an interactive web application that incorporates the
functionality and data plotting functions of the AirSensor for interpreting and
communicating community data collected by sensor networks
o https://github.eom/MazamaScience/AirSensor/tree/version-0.5
o https://github.com/MazamaScience/AirSensorShinv
o These papers summarize the development and enhancements of the
AirSensor and DataViewer tools:
¦ Feenstra et al, 2020 https://doi.Org/10.1016/i.envsoft.2020.104832
¦ Collier-Oxandale et al, 2022 https://doi.org/10.1016/i.envsoft.2021.105256
• Data Policies for Public Participation in Scientific Research: A Primer,
DataONE Public Participation in Scientific Research Working Group, August 2013
o Guide that introduces data policies in the context of public participation in
scientific research or community science, provides examples, and best
practices for implementing data polices in community science projects
o https://safmc.net/wp-
content/uploads/2016/06/Bowseretal2013 DataPolicvPrimer.pdf
• Handbook for Citizen Science Quality Assurance and Documentation, U.S.
Environmental Protection Agency, EPA 206-B-18-001, March 2019
o Handbook that covers common expectations for quality assurance and
documentation and best management practices to level the playing field for
organizations that train and use volunteers in the collection of environmental
data
o https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgapphandbook 3 5 19 mmedits.pdf
• Data Management Guide for Public Participation in Scientific Research,
DataONE Public Participation in Scientific Research Working Group, February 2013
o Guide that provides best practices and other considerations for data
management along the life cycle of community science projects
o https://www.dataone.org/sites/all/documents/DataONE-PPSR-
DataManagementGuide.pdf
A-15
-------
• USGS Guide to Data Management
o United States Geological Survey (USGS) website that provides guidance,
best practices, and tools for data management including in-depth training
modules and numerous data management example scenarios
o https://www2.usgs.gov/datamanagement
• Survey Report: Data Management in Citizen Science Projects, Chade S and
Tsinaraki C., Publications Office of the European Union, JRC101077, 2016
o Report summarizes the findings from a Joint Research Centre (JRC) survey
of community science projects completed primarily in European Union (EU)
countries
o https://publications.irc.ec.europa.eu/repositorv/handle/JRC101077
• Citizenscience.gov Website
o Government website that promotes crowdsourcing and citizen science across
the U.S, government; website catalogs government supported community
science projects, provides a toolkit to assist with project design and
maintenance, and serves as a gateway for community science practitioners
and coordinators across the government
o https://www.citizenscience.g0v/#
• U.S. EPA Guidance on Environmental Data Verification and Data Validation,
U.S. Environmental Protection Agency, EPA/240/R-02/004, November 2002
o Guidance document that specifies the agency-wide program for
environmental data QA and includes practical advice to individuals
implementing data verification and data validation
o https://www.epa.gov/sites/production/files/2015-06/documents/g8-final.pdf
A.3.7 Evaluate: Analyzing, Interpreting, Communicating, and Acting on Results
• Personal Strategies to Minimise Effects of Air Pollution on Respiratory Health:
Advice for Providers, Patients and the Public, Carlsten C., S. Salvi, G.W.K.
Wong, K.F. Chung. European Respiratory Journal 55(6), 2020
o Paper provides guidance based on findings from published literature to assist
health care providers, patients, public health officials, and the public to reduce
exposure to indoor and outdoor air pollution
o https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7270362/
A-16
-------
• The AirSensor Open-source R-package and DataViewer Web Application for
Interpreting Community Data Collected by Low-cost Sensor Networks,
Feenstra B., A. Collier-Oxandale, V. Papapostolou, D. Cocker, and A. Polidori.
Environmental Modelling & Software 134, 2020
o Paper summarizes the development of two software systems to assist in
visualizing and understanding air sensor data collected by community
networks. AirSensor is an open-source R package that allows users to access
historical data, add geospatial metadata, and visualize community monitoring
data using maps and plots. DataViewer is an interactive web application that
incorporates the functionality and data plotting functions of AirSensor for
interpreting and communicating community data collected by low-cost sensor
networks
o https://www.sciencedirect.com/science/article/pii/S1364815220308896
• AirSensor v1.0: Enhancements to the Open-Source R Package to Enable Deep
Understanding of the Long-Term Performance and Reliability of PurpleAir
Sensors, Collier-Oxandale A., B. Feenstra, , V. Papapostolou, and A. Polidori.
Environmental Modelling & Software 148, 2022
o Paper describes the enhancements made to the open-source R package
AirSensor (version 1.0) and the web application DataViewer (version 1.0.1).
to support data access, processing, analysis, and visualization for the
PurpleAir PA-II sensor. The paper also demonstrates how the enhancements
help track and assess the health of air sensors in real-time and historically
o https://www.sciencedirect.com/science/article/pii/S136481522100298X
• Understanding Social and Behavioral Drivers and Impacts of Air Quality
Sensor Use, Hubbell B.J., A. Kaufman, L. Rivers L, K. Schulte, G. Hagler, J.
Clougherty, W. Cascio, and D. Costa. Science of the Total Environment 621 (2018):
886-894
o Paper discusses the social science research conducted on air sensor use and
identifies: (1) research opportunities between the social and environmental
sciences and the entities involved in developing, testing, and deploying air
sensor technologies; (2) the challenges associated with sensor data
generation, interpretation, and analysis; and (3) collaboration opportunities for
communities and organizations to better understand the reasons and
approaches for using sensors and how technological innovations may
improve the ability to reduce exposures to air pollution
o https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6705391/
A-17
-------
A Visual Analytics Approach for Station-Based Air Quality Data, Du, Y., C. Ma,
C. Wu, X. Xu, Y. Guo, Y, Zhou, and J. Li. Sensors 17(1), 2016
o Paper proposes a comprehensive visual analysis system (AirVis) for air
quality analysis that integrates several visual methods, such as map-based
views, calendar views, and trends views, to analyze multi-dimensional
spatiotemporal air quality data
o https://www.mdpi.eom/1424-8220/17/1/30/htm
AtmoVis: Visualization of Air Quality Data, Powley, B., Master of Science Thesis,
Victoria University of Wellington, New Zealand, 2019
o Document discusses the results from a search of literature regarding systems
and methods for visualizing and evaluating air pollution and presents AtmoVis
- a web-based system that includes visualizations for site view, line plot, heat
calendar, monthly rose, monthly averages, and data comparisons
o https://homepaaes.ecs.vuw.ac.nz/~djp/files/MSc BenPowlev 2019.pdf
Openair-An R Package for Air Quality Data Analysis, Carslaw, D C. and K.
Ropkins. Environmental Modelling & Software 27-28, 2012
o Openair is an R package used extensively in academia and in the public and
private sectors that analyzes air quality data and atmospheric composition
data
o https://davidcarslaw.github.io/openair
o https://bookdown.org/david carslaw/openair/
U.S. EPA Real Time Geospatial Data Viewer (RETIGO)
o REal Time GeOspatial Data Viewer (RETIGO) is a free, web-based tool that
can be used to explore stationary or mobile environmental data that you have
collected; nearby public air quality and meteorological data can be added to
the display
o https://www.epa.gov/hesc/real-time-geospatial-data-viewer-retigo
U.S. EPA AirData: Air Quality Data Collected at Outdoor Monitoring Stations
Across the U.S.
o A website providing tools and access to recent and historical air quality
information for the U.S., Puerto Rick, and the U.S. Virgin Islands; view data
on an interactive mapping application; obtain information about each monitor;
and download daily and annual concentration data, AQI data, and speciated
particle pollution data (primarily from U.S. EPA's Air Quality System
database)
o https://www.epa.gov/outdoor-air-gualitv-data
A-18
-------
• Community in Action: A Comprehensive Guidebook on Air Quality Sensors,
South Coast Air Quality Management District (South Coast AQMD), Air Quality
Sensor Performance Evaluation Center (AQ-SPEC), September 2021
o Guidebook for community organizations that covers planning for monitoring
using sensors; sensor deployment, use, and maintenance; and data handling,
interpretation, and communication
o http://www.aamd.gov/aa-spec/special-proiects/star-arant
• California Air Resources Board (CARB) List of CARB-Certified Air Cleaning
Devices
o A website providing a table of CARB-certified air cleaning devices, a
description of the difference between mechanical and electronic air cleaners,
and resources to help you select a safe and effective air cleaner
o https://ww2.arb.ca.gov/list-carb-certified-air-cleaning-devices
• U.S. EPA Research on Do-lt-Yourself (DIY) Air Cleaners to Reduce Wildfire
Smoke Indoors
o A website providing on overview of EPA research conducted to evaluate the
safety and effectiveness of DIY air cleaners; includes a summary and links to
the Underwriters Laboratories (UL) safety report findings, answers to
frequently asked questions, and links to helpful resources
o https://www.epa.gov/air-research/research-div-air-cleaners-reduce-wildfire-
smoke-indoors
A.4 Sensor Performance Guidance
A.4.1 Sensor Performance Evaluations
• U.S. EPA Air Sensor Performance Evaluations
o Collection of results for field and laboratory evaluations of air sensors
conducted by U.S. EPA Office of Research and Development
o https://www.epa.gov/air-sensor-toolbox/evaluation-emerging-air-sensor-
performance
• South Coast Air Quality Management District (South Coast AQMD) Air Quality
Sensor Performance Evaluation Center (AQ-SPEC)
o Field and laboratory evaluations of commercially available air sensors
conducted by AQ-SPEC
o http://www.agmd.gov/ag-spec
A-19
-------
• European Commission Joint Research Centre (JRC)
o Results for field and laboratory evaluations of air sensors primary in the form
of scientific journal articles and reports
o https://publications.irc.ec.europa.eu/repositorv/
• AirParif AIRLABS Microsensors Challenge
o Results from an international challenge that promotes innovation and helps
inform users on the performance of air sensors in different applications
o http://www.airlab.solutions/en/proiects/microsensor-challenqe
A.4.2 Approaches Used to Evaluate Sensor Performance
• U.S. EPA Performance Testing Protocols, Metrics, and Target Values for Air
Sensors - Use in Ambient, Outdoor, Fixed Site, Non-Regulatory Supplemental
and Informational Monitoring Applications
o Reports that provide recommended testing protocols (field and laboratory),
performance metrics (parameters used to describe sensor data quality), and
target levels to evaluate air sensors that measure criteria air pollutants
o https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-
testinq-protocols
• ASTM WK64899 "New Practice for Performance Evaluation of Ambient Air
Quality Sensors and Other Sensor-Based Instruments"
o Provides information on a practice for evaluating the performance of air
quality sensors in ambient air
o https://www.astm.org/workitem-wk64899
• EU/CEN/TC 264/WG 42 "Air quality - Performance evaluation of air quality
sensor systems - Part 1: Gaseous pollutants in ambient air"
o Outlines testing procedures and requirement for classifying performance of air
quality sensors for the monitoring of gaseous pollutants
o https://www.en-standard.eu/pd-cen-ts-17660-1 -2021 -air-qualitv-performance-
evaluation-of-air-qualitv-sensor-svstems-qaseous-pollutants-in-ambient-air/
A-20
-------
• Air Quality Sensor Performance Evaluation Center (AQ-SPEC) of the South
Coast Air Quality Management District (South Coast AQMD) Website
o Website for the AQ-SPEC program which describes how field and laboratory
tests are conducted
o Field Evaluation Protocol and Reports: http://www.agmd.gov/ag-
spec/evaluations/field
o Laboratory Evaluation Protocol and Reports: http://www.agmd.gov/ag-
spec/evaluations/laboratorv
A.4.3 Summarizing Sensor Performance Evaluation Results using U.S. EPA's
Targets Reports
• U.S. EPA Performance Testing Protocols, Metrics, and Target Values for Air
Sensors - Use in Ambient, Outdoor, Fixed Site, Non-Regulatory Supplemental
and Informational Monitoring Applications
o Reports that provide recommended testing protocols (field and laboratory),
performance metrics (parameters used to describe sensor data quality), and
target levels to evaluate air sensors that measure criteria air pollutants
o Appendices of the Targets Reports provide standardized reporting templates
o https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-
testing-protocols
• U.S.-EPA developed sensortoolkit python code library
o Code library for evaluating air sensor data collocated with reference
instruments; code can be used to calculate performance metrics
o GitHub: https://github.com/USEPA/sensortoolkit
o PyPI: https://pypi.org/proiect/sensortoolkit/
• U.S.-EPA PM2.5 Continuous FEM Monitor Comparability Assessments
o Tool providing a one-page technical report that assesses the comparability of
a PM2.5 continuous FEM monitors when collocated with an FRM sampler
o https://www.epa.gov/outdoor-air-gualitv-data/pm25-continuous-monitor-
comparabilitv-assessments
A-21
-------
Appendix B: Questions to Consider When Planning for and
Collecting Air Sensor Data, and Sharing Your Results
Getting input from others before you start collecting measurements will help you better plan
and collect data to meet your purpose. Below we provide a list of the types of questions to
consider. While this list is by no means exhaustive, answering these questions helps you
plan and ensures credibility in your data and results. These questions can also help you
respond to inquiries from others if you decide to share your plans, data, and results.
B.1 Planning (see Section 3.3)
• What is the purpose of the project and the question you want to answer?
• What existing research and data are available to help answer your question?
• What actions might you take depending on the research, data, or air monitoring
results?
• What pollutants will you measure? If you are interested in a particular source of air
pollution, have you checked that your selected pollutant is relevant to that source
(see Table 2-1)7
• Have you developed a plan for your monitoring activities (see Section 3.3)7
• What are the expected levels for the pollutant in the location of interest, including
background and peak concentrations, seasonal and day/night trends, and spatial
variability?
• Do you have procedures and instructions so that measurements are taken in a
consistent way (e.g., develop standard operating procedures)?
• Have you established clear roles and responsibilities for those involved in your
monitoring activities?
B.2 Working with Governmental Officials (see Section 3.3 and Section
3.8.2)
• Will you contact the state/local/tribal air monitoring agency during the planning
phase to obtain their input and recommendations?
• Will you consider alternative ways to answer the questions using other data sources
besides air sensors (e.g., traffic counts, health data, existing monitoring data)?
• Have you clearly defined your purpose for monitoring and expected outcomes? Is
collecting air sensor data the best way to achieve these outcomes?
• Are you expecting any agencies (local/state/federal/tribal) to use your data or
results? Have you spoken with the agencies to understand if that is possible?
• Have you filed a formal complaint with your state/local/tribal air quality agency or
other responsible organization (e.g., the department of health, fire department)?
B-1
-------
B.3 Setting up Monitoring Locations (see Section 3.5)
• Where will you collect the measurements?
• What will be the location of the sensor(s) (e.g., latitude, longitude, elevation, and
height of the sensor from ground level)?
• Will any obstructions nearby affect the airflow around the sensor?
• How will you select the site? What criteria or guidelines will you use?
• What nearby emission sources (e.g., roadways, industrial facilities) might affect the
sensor measurements?
• Are there other potential local sources near your site (e.g., dust from unpaved roads,
parking areas, street-sweeping activity) that might affect your measurements?
• Might anyone nearby be smoking (e.g., cigarettes, cigars) when and where you are
collecting the measurements?
• Will there be any periodic events (e.g., construction, fireworks, fires) that could affect
the data?
• What type of conditions will the measurements represent (e.g., outdoor, indoor,
occupational - see Section 2.3)7
B.4 Collecting Data (see Section 3.7)
• What instrument/sensor will you use (i.e., manufacturer, model, etc.)? Will these be
new devices, older, or refurbished?
• How long will you make your measurements (e.g., two weeks, two months)?
• Will you take the measurements at a fixed site or mobile platform (e.g., on a car, on
a person)? What type of environmental conditions will they represent?
• Will you receive adequate training on operating the device, maintaining it, and
troubleshooting any issues?
• Will you use a lab notebook, log sheets, or check lists to record instrument set up,
maintenance, and other additional data? Will these be in paper or electronic
formats? Where will this information be stored and who will have access?
• Will you collocate (i.e., place nearby) your sensor near reference monitors or other
trusted measurement systems to evaluate their performance? If so, where will you
collocate and how will you process and show your results?
• Will you make any adjustments, corrections, or calibrations to the data after
collection? Will you document the methods and techniques used?
• How will you estimate the precision and bias of your air sensor data?
• Will you read the data from the screen or store it in an electronic format?
• If applicable, how will you name the data files so that you can keep track of where
and when they were collected? How will you keep and protect raw, unedited data
from the sensors? How will you document how data was processed or corrected?
How will you track the data if you transfer it to others to analyze?
• What, if any, additional local data or observations will you collect (e.g., wind
measurements, site photos, global positioning system (GPS) coordinates, activity
logs, event logs, health information)?
B-2
-------
• What, if any, additional data sources will you draw from (e.g., meteorological data
from the National Weather Service, regulatory air monitoring data from a
state/local/tribal air monitoring agency)? Where will those data come from and how
will they be integrated with the measurements you are making?
• What type of file or database will you use to store the data?
• How will you ensure that each parameter has the correct units?
• How will you document the time standard [e.g., local standard time (LST),
Coordinated Universal Time (UTC)]?
B.5 Conducting Quality Control (see Section 3.7.2)
• What procedures will you use to ensure that the sensors measure high-quality data?
• Will you develop quality control (QC) criteria that the sensor data must meet?
• Will you have Standard Operating Procedures (SOPs) (i.e., detailed written
instructions and directions on how to perform a technical activity so that
measurements are obtained in a consistent way)?
• Will you average the data, and if so, how? How will you account for missing data,
negative values, and extreme outliers?
• What software will you use to process and QC your data (e.g., Microsoft Excel)?
• How will you correct or adjust the air sensor data?
• How will you document any changes or adjustments to the data?
• Will you use consistent data qualifiers to "flag" data that do not meet QC criteria?
• Will you document if there are any persistent problems with the data or significant
downtime?
• How will you record and resolve any data problems?
B.6 Evaluating Data (see Section 3.8)
• Do you need software to analyze the data (e.g., Microsoft Excel, R, Matlab, Python)?
Does someone on your team have the skills needed to use the required software?
• How will you analyze the data (e.g., create a scatterplot, create a time series plot,
compare with meteorological measurements)?
• How will you differentiate the source you are trying to measure from the background
pollutant concentrations?
• Will you publish your results or create any public communication materials?
B.7 Other
• Do you have any interaction (past, present, or future) with the entity potentially
responsible for creating the emissions?
• Will there be any limitations or restrictions for using your data?
• Who will be the primary contact if others have questions about the data?
• Is there additional support that would be helpful (e.g., data analysis or interpretation
support, collocation assistance)?
B-3
-------
Appendix C: Checklists
C.1 What to Look for in an Air Sensor?
Before buying an air sensor, use this checklist to help make sure you are purchasing a
sensor that meets your needs and produces data suitable for your application.
~ Sensor accuracy
Look for a sensor with demonstrated and documented performance under similar
environments and operating conditions that you anticipate encountering in your
monitoring application. Accuracy consists of precision and bias (see Section 3.6
and Chapter 4).
~ Pollutants of interest
Purchase sensors that can accurately detect the pollutants of interest with limited
interferences from weather conditions or other pollutants (see Section 3.4, Table 2-1,
and Table 2-2).
~ Detects high and low concentrations
Because pollutant concentrations can vary greatly, you will need to determine if a
sensor can detect a pollutant at the low-end of the concentration range and the
range's high end. See Table 2-1 for typical pollutant concentration ranges.
~ Reliability
The sensor(s) you select should be able to operate for extended periods with no or
minimal maintenance. Consider the warranty and replacement policy of the
manufacturer to ensure uninterrupted operation.
~ Instructions or user guide
Operation and maintenance instructions from the manufacturer enable you to set up,
conduct routine maintenance, fix and repair, and replace the sensors as needed,
instructions may also be provided for zeroing or calibrating the sensor or for quality
assuring, correcting, or processing of the data. Instructions should also include how to
access both real-time and historical data.
~ Power
Give careful consideration to the power requirements of the sensor. Power
requirements vary from plug-in, battery, or solar power and will depend on a user's
application. Plug-in devices are best suited for stationary monitoring applications;
however, make sure power is available and easily accessible. Battery-powered
devices are best for mobile applications or short-term data collection activities. Solar-
powered devices require proper size (of the solar panel and battery) and adequate
C-1
-------
orientation to sunlight. The choices and logistics are involved - plan enough time to
ensure that you can identify a solution that meets your needs.
~ Data Transmission
Many options exist for air sensor data communication from the sensor to the sensor
data repository and include cellular, WiFi, Bluetooth, satellite, low-power wide-area
network (LoRa), or other methods. Remember that you may want to test the
communications at the actual site and inquire about ongoing costs associated with
communicating data via cellular and other protocols. Ensure that you have the ability
to enable the necessary settings on your computer and/or WiFi network to allow you to
use the selected data transmission option (if applicable).
~ Transparent data processing
Sensor data often undergo processing, correction or adjustments, and averaging to
improve quality. Look for sensor manufacturers that provide descriptions of the
methods and approaches they use to process sensor data, especially related to
sensor data correction (sometimes called calibration) techniques. Section 3.6
discusses the calibration and correction methods and procedures.
~ Customer service
Seek to understand how responsive and knowledgeable the company will be to your
questions and needs. Consider the company's location (e.g., time zone), contact
methods (e.g., phone, email), and specific services they offer.
~ Ease-of-use
Look for intuitive and easy-to-use sensors and software to manage the data the air
sensor produces. Match the sensor's features with your needs. These features could
include weatherproofing, durability for handheld and mobile sensors, on-sensor
display (e.g., lights or digital readout), data management capabilities, etc.
~ Data Handling and Access
Determine how data processes and where it is stored and in what format. Identify the
data management system used to ingest, process, visualize, and distribute data.
Determine who can access the data, who has ownership rights of the data, privacy
terms, how long will the data be available.
~ Cost
The cost of sensor technology may vary greatly depending on the pollutant measured
and the degree of accuracy and sensitivity the user needs. Consider the initial
purchase price and the long-term operational costs, such as data transmission and
storage, maintenance, calibration services, data ownership, and repairing or replacing
sensor components.
C-2
-------
C.2 What to Look for in a User Manual?
A user manual should be comprehensive and clear and effectively describe the installation,
operation, and maintenance activities needed so that you can set up and run the sensor
optimally. Without a good user manual, you may have to spend more time figuring out how
to operate, troubleshoot, and/or repair your sensor. Request a user manual before
purchasing a sensor to ensure the device meets your needs. The following are
recommended items to look for in a user manual:
~ Performance specifications
The manual should include details about the sensor's accuracy, range of
measurement, minimum and maximum detection limits, sampling frequency, ambient
temperature range, power requirements, response time (i.e., how quickly the sensor
responds to changing conditions), sensor lifespan or expiration date, etc.
Demonstrations of sensor performance in real-world applications (ideally, independent
reports or scientific articles discussing sensor evaluation test results) are also very
important. Chapter 4 provides more details about air sensor performance.
~ Installation instructions
The manual should provide the procedures that describe where and how to install the
sensor and check that it is operating correctly.
~ Operating instructions
Details about how to ensure the sensor is running correctly and how to access its data
are critically important for a user manual.
~ Maintenance requirements
The manual should include a list of the activities, requirements, and specific
frequencies or schedule at which they must be performed. See Appendix C.3 for
more information regarding sensor maintenance.
~ Correction (sometimes called Calibration)
The manual should provide details about the process and method used to adjust or
correct data. Providing more information about the data correction helps build trust
and confidence in the data and the air sensor.
~ Data access
Look for details about how to view real-time and historical data, access backup raw
and processed data, and share data with other organizations. A user manual should
identify any data ownership terms and conditions and data privacy terms.
C-3
-------
~ Interferences
The manual should clearly describe the known (and potential) interfaces from weather
conditions (e.g., high humidity) and other pollutants in the air.
~ Customer service
Look for support information with contact details (e.g., methods, hours of operation)
and what is covered (or not covered) with customer support.
~ Limitations
The manual should describe restrictions or limitations of the sensors and its operation
and explicitly identity what is covered by the warranty.
~ Hazard Warnings
Check to see if the user manual provides any warnings concerning any hazards or
potential hazards that might be present during installation, calibration, operation,
maintenance, or troubleshooting of the sensor.
C-4
-------
C.3 How to Maintain Your Air Sensor?
Like most other forms of technology, air sensors require maintenance to ensure proper
functionality and reliable performance. These preventative actions associated with
maintenance are necessary for both short- and long-term operations. By properly caring for
an air sensor, you can reduce errors in data collection, extend the operating life of the
device, and save money that would otherwise be spent on replacement parts and repair
services.
Check with the air sensor manufacturer for protocols to maintain your device so it operates
properly and produces good data. Typical routine maintenance processes include:
~ Cleaning of internal and external surfaces and components to prevent the buildup of
bugs, dust, pollen, etc.
~ Replacement of filters and other consumables.
~ Replacement of the sensor when it fails or reaches the end of its service lifespan.
~ Replacement of rechargeable batteries.
~ Cleaning off dust and dirt from solar panels.
~ Collocation and correction of sensor data to improve data quality (see Section 3.8).
~ Visual inspection of data to identify odd patterns, a decrease in overall response,
drift in the baseline, and other unusual features. Instrument problems tend to
produce data that often look too regular and repeatable, or that change too abruptly
to be due to natural atmospheric phenomena.
~ Inspection of the sensor placement location to ensure that no significant changes
have occurred (e.g., tree growth, building changes, new local sources of pollutants).
~ Development and maintenance of a logbook to ensure maintenance occurs at
regular intervals.
C-5
-------
Appendix D: Data Handling and Air Quality Index (AQI)
Calculations
D.1 Data Processing
Working with data consists of reviewing and calculating values from the data to produce
meaningful information. With air sensors, data processing can be a very time-consuming
process because sensors can produce lots of data. For example, an air sensor network of
10 sites, with each site measuring two pollutants every minute, can produce over 5.3 million
data values over a one-year period. In contrast, a similar size network producing hourly
data only generates about 220,000 data values. Thus, it takes time to process the data to
glean insights, meaning, and information from the results. Software and planning can help
you efficiently and effectively process data. This section provides some tips for processing
data by harmonizing and aggregating the data.
When harmonizing data, you ensure that the data is of good quality and complete and are
comparable to other datasets. By
aggregating (e.g., averaging,
compiling) data, you can notice the
big-picture trends and see patterns in
the data. For example, if you want to
review 1 week of air sensor data, you
might want to aggregate the 1-minute
values by making hourly averages and
viewing those data to see the trends.
This aggregation results in reviewing
168 hourly averaged data points
instead of over 10,000 1-minute data
points. Once you see the "big picture,"
you can start looking at the details in
the 1-minute data.
When deciding how to view the data, consider your monitoring question. Some averages
will be more relevant to your investigation than others. For example, a project interested in
changes in air pollution before, during, and after school drop-off may find 5 or 10-minute
averages useful. A project focused on seasonal changes in air quality (e.g., residential
wood burning) may find weekly or monthly averages helpful.
When you compare data with standards, you will need to aggregate the data similar to the
aggregation required by the standard. For example, for fine particulate matter (PM2.5) data,
you will need to calculate a 24-hour average from the hourly data before comparing it to the
24-hour National Ambient Air Quality Standards (NAAQS).
/ \
What are Some Important Definitions
Related to Data Processing?
Data management is a collection of
procedures needed to acquire, process, and
distribute data.
Data harmonization is the process of
reviewing data for quality and completeness,
and then combining the data for querying and
viewing.
Data aggregation is the process of
compiling information to prepare combined
datasets for data processing,
s 4
D-1
-------
D.1.1 Data Quality Assurance (QA)
Data harmonization consists of the steps to prepare data for aggregation and subsequent
data analysis. Air sensor data may have issues that that need data cleaning. By reviewing
the data and addressing these problems, your data will be more trustworthy and resulting
analyses and conclusions will be more robust. Here are some suggestions to help
harmonize your data:
1. Quality control check and validate data by reviewing it using time series plots,
spatial plots, and statistical summaries to identify questionable data. Section 3.7.1
and 3.7.2 of this Guidebook discusses data review and quality control checks you
can apply.
2. Identify outliers or data points that are significantly different from other data values.
Outliers are typically easy to see in a time series plot of data (Figure D-1). It is
typically easier to find outliers in higher-time resolution data (e.g., 1-minute) rather
than averaged data (e.g., 1-hour and 24-hour averages). Identify these outliers and
decide whether or not to include them from data aggregation and analysis steps.
3. Address negative values which can occur because air sensors and instruments
have uncertainty, and at near-zero concentrations, you might see some slightly
negative values (e.g., -0.5 pg/m3). These slightly negative values may be valid and
you may wish to retain these values (consult the sensor manufacturer for guidance).
However, significant or persistent negative values likely indicate a problem with the
air sensor or instrument, and you should flag these values and exclude them from
subsequent data processing and analysis.
4. Address missing data which may occur when air sensors do not collect data due to
power outages, sensor malfunctions, loss of communications, etc. Some missing
data is common when using air sensors, but too much missing data means the data
and averages of the data will not represent actual air quality conditions. It is
recommended that you do not replace or fill in missing data with estimated values
(e.g., interpolated or extrapolated concentrations). Instead, seek to understand and
fix the problem causing the missing data.
5. Evaluate data completeness to ensure that enough data exist to represent air
quality conditions during that monitoring period. Completeness is a measure of the
amount of valid data obtained from a sensor compared to the amount expected to be
obtained under correct, normal conditions. Generally, at least 75 percent of the data
are needed to make a valid average (e.g., at least 45, 1-minute measurements are
needed to make a valid 1-hour average). When the 75 percent completeness
criterion is met, the resulting aggregation is usually considered representative of that
monitoring period.
D-2
-------
Figure D-1. Time Series of Ozone (O3) Concentrations Showing a "Spike" in
Concentration that is an Outlier in the Data
D.1.2 Data Aggregation
Data aggregation helps with the analysis and assessment of data by averaging, counting,
or filtering data. This process allows you see the more prominent trends in your data and
enables you to compare the aggregated values to standards and indexes. The most
common aggregation methods include:
• Averages of air quality data are typically calculated over a time period that ranges
from minutes to a year. There are several types of averages, as shown in Figure D-2.
a. Block average is a technique to reduce data points to a particular period by
computing the mean. For example, 24 hourly measurements of PM25 can be
averaged into a single 24-hour value that represents the mean PM2.5
concentration during that period. This type of averaging helps reduce the
number of data values you need to examine and can enable comparison to
standards like the NAAQS.
b. Rolling (moving) average is a technique to obtain an overall idea of the trends
in a data set and "smooth" some of the rapid changes. You can calculate it for
any period that corresponds to a standard or index (e.g., Air Quality index,
AQI). Typically, rolling average periods are 24 hours for PM2.5 and PM10, 8
hours for ozone (O3), and 3 months for lead (Pb).
A 24-hour rolling average on 1-hour data would result in 24 rolling averages over a
24-hour period, while biock averaging would result in only 1 block average.
D-3
-------
Day 1 | Day 2
Day 3
Day 4 | Day 5
Legend
— Hourly Data
Rolling Average
• 24-Hour Average
Figure D-2. Time Series Showing Raw PM2.5 Data with Block and Roiling Averages
• Counting is a method to characterize air quality conditions by summarizing the
number of "events" that occur above a threshold. For example, you might want to
count the number of days where the 8-hour average O3 concentrations are above 70
ppb, which is "Unhealthy for Sensitive Groups" on the AQI scale. You could also
count the number of days when PM2.5 averages fall within a specific category of the
AQI (e.g., number of "Good" days). The exact counting method will depend on your
application, but could include counting the number of hours, days, or sites above a
specific concentration.
D-4
-------
• Filtering or stratifying data is a
method to subset data by time or
location. This type of filtering
allows you to see differences in
aggregations that might result
from factors such as weather or
emission events (e.g., high traffic
times, wildfires). You might also
consider filtering by time (e.g.,
hour, day, weekday/weekend,
season). For example,
comparing averages of nitrogen
dioxide (NO2) concentrations by
the hour of the day and by
weekday and weekend may
show the effect of traffic activity
on local pollution conditions.
D.2 AQI Calculations
Tip: Consider ways that can make data
processing less time-consuming and
easier
Plan ahead as processing data can be very
time-consuming. This planning will save you
effort and reduce the chances that data
need to be re-measured or reprocessed.
Match the aggregation methods (e.g.,
averaging period) when comparing air
sensor data to other datasets, standards, or
indices.
Use software to facilitate data
processing Consider the costs of software.
Some software is free, whereas other
software is not. A data management system
(see Section 3.7.3 of this Guidebook) and
other analysis software (see Section 3.8 of
this Guidebook) can help process data.
D.2.1 Background
The Air Quality Index (AQI) is U.S. EPA's index for communicating daily air quality. It
provides statements for each category that tell you about air quality in your area, which
groups of people may be affected, and steps you can take to reduce your exposure to air
pollution. The AQI is calculated for five of the six criteria pollutants: O3, PM (including PM2.5
and PM10), CO,
NO2, and SO2.
The AQI scale runs
from 0 to 500, with
higher AQI values
indicating more
hazardous levels of
air pollution and
associated health
concerns. For each
pollutant, an AQI
value of 100
Are There Other Air Quality Indices in Different Countries?
Yes! Other countries have established different air quality
indices. Some examples include:
• Air Quality Health Index (Canada)
• Air Quality and Health Risk Index (Mexico)
• Air Pollution Index (Malaysia)
• Pollutant Standards Index (Singapore)
• European Air Quality Index (European Union)
generally corresponds to an ambient air concentration equal to the level of the short-term
NAAQS for protection of public health. AQI values at or below 100 are generally considered
to be satisfactory. For example, an AQI value of 50 or below represents "Good" air quality,
while an AQI value over 300 represents "Hazardous" air quality.
D-5
-------
The AQI is divided into six color-coded categories, as shown in Table 2-5, with each
category corresponding to a different level of health concern. The color-coding allows the
public to quickly determine whether air quality is reaching unhealthy levels in their
communities. The specific concentration breakpoints for each of the six levels vary by
pollutant.
Although the AQI color scale is required by law [40 Code of Federal Regulations (CFR)
Part 58.50 and 40 CFR Appendix G to Part 581, it is recognized that the colors may not be
accessible to all people. The South Coast Air Quality Management District (AQMD)
developed a modified version of the AQI coior scale that accommodates individuals with
color vision deficiencies. The modified scale was tested against eight common color
impairments and remains distinguishable in grayscale. Currently, the modified colors are
being piloted under the "View More Accessible AQI Colors" option on South Coast AQMD's
real-time air quality map and the ColorVision Assist option on the AirNow Fire and Smoke
Map.
Normal
Grey-scale
Traditional AQI
Color-accessible
AQI
Color-accessible
AQI (continuous)
Traditional AQI
Color-accessible
AQI
Color-accessible
AQI (continuous)
V4lt«y
Figure D-3. Comparison of the Traditional AQI and Color-Accessible AQI
Color Scale Presented in Color, Grey-Scale, and on a Map of the South Coast
Air Basin (South Coast AQMD Press Release - May 2022)
D-6
-------
D.2.2 Computing the AQI
This section explains how to calculate the AQI. Figure D-4 shows a flowchart of the process
for computing the AQI.
Hourly
concentration
value
Historical
Calculate the appropriate
average for the pollutant from
the hourly concentration data
Estimating AQI for
real-time (now) or
historical?
Real-time
Use NowCast to
estimate AQI
Use 0Q NowCast
Use PM NowCast
o
o
O)
c
O
<
o
I £
c o
d> o
o w
c
o
o
pm25
™,o
°3
NO,
CO
S02
24-hr
24-hr
1 and 8-hr
1-hr
8-hr
1 and 24-hr
average
average
averages
average
average
averages
Average hourly 03
concentrations
based on the
NowCast method
Average hourly
PM (PM2 5 or PM10)
concentrations
based on
NowCast method
M
Use AQI equation and breakpoints
to convert average concentration
to an AQI value
Figure D-4. Flow Chart Showing How to Compute the AQI
There are several important considerations when calculating the AQI using air sensor data:
1. Sensor data must first be corrected to be more comparable to FRM/FEM
reference instruments. The Clean Air Act requires air quality to be monitored
nationally within the U.S., using Federal Reference Method (FRM) or Federal
Equivalent Method (FEM) instruments. Sensor data must first be harmonized
(discussed above) and corrected (see Section 3.6.2) so that the data will be more
comparable to the ambient air quality network.
D-7
-------
2. Data is averaged to correspond to the NAAQS. For
historical data (i.e., not real-time), the U.S. EPA
calculates the AQI values from air data averages over
1, 8, or 24 hours using the AQI equation (see #3
below). The reason for the different averaging times is
that different pollutants affect the human body in
different ways as presented in Table 2-2.
3. NowCast AQI is used for PM and O3. For current air quality conditions, the U.S.
EPA calculates the AQI values using a method called NowCast to estimate short-
term averages that are then converted using the AQI formula (see below). Note that
the NowCast AQI is only calculated for O3 and PM because these pollutants
commonly drive the AQI. The NowCast calculations for PM and O3 are different as
these pollutants behave differently in the atmosphere and have different NAAQS
averaging times. For example, the NowCast AQI for PM shows air quality for the
most current hour available by using a calculation that involves multiple hours of
past data. The NowCast uses longer averages during periods of stable air quality
and shorter averages when air quality changes rapidly, such as during a wildfire
event.
Tip: Be aware that
reference instrument
data is the only data
that determines
NAAQS compliance
4, Online tools are available to calculate the AQI. Once you have calculated an
average concentration, use the following equation to compute the AQI for a given
pollutant. An online AQI calculator is available to help with this conversion.
AQI =
(AQIHi) -
CaqiL0)
(ConcHi) -
(ConcLo)
x (Conq) — (ConcLo)
+ Who)
Where:
Cone = Truncated concentration for a given pollutant
Condo = Concentration breakpoint that is less than or equal to Cone
ConcHi = Concentration breakpoint that is greater than or equal to Conci
AQI Lo — AQi value/breakpoint corresponding to ConcLo
AQIhi = AQI value/breakpoint corresponding to ConcHi
You can find the latest AQI-Concentration breakpoints here and the rules for
truncating concentrations are:
O3 (ppm) - truncate to 3 decimal places
PM2 s(|jg/m3) - truncate to 1 decimal place
PM10 (pg/m3) - truncate to integer
CO (ppm) - truncate to 1 decimal place
SO2 (ppb) - truncate to integer
NO2 (ppb) - truncate to integer
D-8
-------
For example, using an average PM2.5 value of 35.9 jjg/m3 and the PM2.5 breakpoints in
Table D-1 results in an AQI of 102:
(150) - (101)
(55.4) - (35.5)
x (35.9) - (35.5)
+ (101) = 1.02
Table D-1. Example Breakpoints for PM2.5
AQI Category
AQI Color
PM2.5
Coiiclo
(pg/m3)
PM2.5
ConcHi
(|jg/m3)
AQLo
AQIHi
Good
Green
0.0
12.0
0
50
Moderate
Yellow
12.1
35.4
51
100
Unhealthy For
Sensitive Groups
Orange
35.5
55.4
101
150
Unhealthy
Red
55.5
150.4
151
200
Very Unhealthy
Purple
150.5
250.4
201
300
Hazardous
Maroon
250.5
500.4
301
500
D-9
-------
Appendix E: Interpreting Sensor Performance Evaluation
Results
As mentioned in Section 4.3.1 the U.S. EPA's Targets Reports provide templates to
encourage a similar format for reporting sensor evaluation results. This Appendix will walk
you through a performance evaluation report for base testing (field evaluation) as this
testing is recommended, at a minimum. Using the reporting template as an example, this
Appendix will review the types of information that testing organizations are being asked to
report, why this information might be important to sensor users, and how sensor users
should use and interpret the information provided.
The base testing reporting template for fine particulate matter (PM2.5) sensors is shown in
Figures E-1, E-2, and E-3. Page 1 (Figure E-1) includes deployment details and several
graphs where testers can visually summarize sensor performance. Page 2 (Figure E-2)
contains tables for all of the calculated performance metrics and performance statistics.
There is also space for additional scatterplots so that a separate graph can be displayed for
each evaluated sensor. Page 3 (Figure E-3) includes a table where testers can share what
additional documentation is included with the evaluation report that would be helpful for
sensor users in interpreting the results. Testers can choose to attach additional information
or write a short description of that information in the table.
An example of a filled in reporting template is provided here using results from one of the
U.S. EPA's field evaluations of the AirBeam2 sensor compared to the T640x reference
instrument. The AirBeam2 is no longer available for purchase and has been chosen for
illustration purposes only.
E-1
-------
Testing Report - PM2 5 Base Testing
Manufacturer & Air Sensor Name
Deployment Number
Testing Organization
Contact Email / Phone Number
Date
Image of device
during
deployment
Deployment Details
Testing Organization and Site Information
Testing organization
(Name, Organization
Type, Contact website /
phone number / email)
Testing location
(City, State; Latitude &
Longitude)
AQSsite ID
Sampling timeframe
Sensor Information
FRM/FEM Monitor Information
Manufacturer,
model
Manufacturer, model
Device firmware
Sampling time
version
interval
Sampling time
Date of calibration
interval
Sensor serial
#1 #2 #3
Date of flowrate
numbers
Brief summary of issues
~
verification check
Issues
encountered
during
deployment?
Description, date(s)
of maintenance
activities
Time Series Plots
1-hour averaged PM2s
Scatter Plots: Comparison to FRM/FEM
1-hour averaged PM? 5
24-hour averaged PM2 5
We«kS Week 6 Woe* 7
Duration
24-hour averaged PM2 5
Range of FRM/FEM monitor concentrations over
duration of test (|jg/mJ)
Number of 24-hour periods in FRM/FEM monitor
measurements with a goal concentration of >25 ng/m3
Performance Metrics
Sensor - FRM/FEM Accuracy
Sensor - Sensor Precision
Meteorological Conditions During Deployment
Meteorological Influence
2 0.6
J 0.4
&08
e 0.6
or Relative Humidity (%)
Monitor Temperature (°C)
Number of 24-hour periods outside manufacturer-
listed temperature target criteria
Number of 24-hour periods outside manufacturer-
listed relative humidity target criteria
Number of paired, normalized concentration and
temperature values
Number of paired, normalized concentration and
relative humidity values
Sensor Serial ID 1
Sensor Serial ID 2
Sensor Serial ID 3
'For evaluations with greater than three sensors, grouping individual
whole evaluationgroup, such as RMSE, NRMSE, CV, and standard deviation.
into boxplots is recommended for displaying results. Note that this recommendation does not apply to m
cs computed as a single value for all se
Figure E-1. Page 1 of U.S. EPA's Base Testing Reporting Template for PM2.5 Sensors
- Deployment Details and Visual Plots of Sensor Performance
E-2
-------
Testing Report - PM2 5 Base Testing
Manufacturer & Air Sensor Name
Deployment Number
Testing Organization
Contact Email / Phone Number
Date
Image of device
during
deployment
Tabular Statistics
• Sensor- FRM/FEM Correlation
Bias and Linearity
Data Quality
R2
Slope
Intercept (b)
(pg/m3)
Uptime
(%)
Number of paired
sensor and FRM/FEM
concentration values
1-Hour
000
24-Hour
000
1-Hour 24-Hour
000 000
1-Hour 24-Hour
000 000
1-Hour 24-Hour
000 000
1-Hour 24-Hour
Metric Target Range
£ 0.70
£0.70
1.0 + 0.35 1.0 ±0.35
-5 < b < 5 -5 < b < 5
90%* 90%*
Sensor Serial #1
SensorSerial #2
Sensor Serial #3
Mean
Error
RMSE (pg/m3)
NRMSE (%)
1-Hour
0
24-Hour
0
1-Hour
0
24-Hour
0
Metric Target Range
s7
<,7
s 30
<30
Deployment Value
Device-specific metrics (computed for each sensor in evaluation)
ooo Metric value for none of devices tested falls within the target range
•oo Metric value for one of devices tested falls within the target range
• *0 Metric value for two of devices tested falls within the target range
• • • Metric value for three of devices tested falls within the target range
Single-valued metrics (computed via entire evaluation dataset)
O Indicates that the metric value is not within the target range
• Indicates that the metric value is within the target range
Sensor - Sensor Precision
Precision (between collocated sensors)
Data Quality
CV
(%)
SD
(pg/m3)
Uptime
(%)
Number of concurrently reported
sensor concentration values
1-Hour
0
24-Hour
0
1-Hour
0
24-Hour
0
1-Hour 24-Hour
0 0
1-Hour 24-Hour
Metric Target Range
<30
<30
<5
<5
90%* 90%*
Deployment Value
Individual Sensor - FRM/FEM Scatter Plots
Sensor Serial 1
Sensor Serial 2
Sensor Serial 3
FRM/FEM PM2 5 (pg/m3)
10
15
FRM/FEM PM2 5 (#jg/m3)
Relative Humidity (%)
20
40
0 ,/
0
FRM/FEM PM2 5 (vg/m3)
This value is only a recommendation for ensuring data quality and is not included in the list of target values discussed in Section 4 of the Performance Testing Protocols, Metrics, and Target Values for Fine Particulate Matter Air Sensors document.
Figure E-2. Page 2 of U.S. EPA's Base Testing Reporting Template for PM2.5 Sensors
- Tables and Graphs Summarizing Sensor Performance
E-3
-------
Testing Report - PM2 5 Base Testing
Manufacturer & Air Sensor Name
Deployment Number
Testing Organization
Contact Email / Phone Number
Date
Image of device
during
deployment
Supplemental Information
Additional documentation maybe attached or linked to digital versions alongside this report. Such documentation may includefield reports
and observations during the testing period, maintenance logs for sensors and FRM/FEM monitors, standard operating procedures, and other
documentation relevantto this testing report (see below for examples).
Supplemental
Documentation
Attached
Description & URL or file path to documentation
Field observations
~
Maintenance logs
~
Standard operating
procedure(s)
~
Photos of equipment setup
and testing
~
Product specification sheet(s)
~
Product manual(s)
~
Deployment issues
~
Data storage and transmission
method
~
Data correction approach
~
Data analysis/correction
scripts and version
~
Air MonitoringStation QAPP
~
Summary of FRM/FEM
monitor QC checks
~
Other documents
~
Figure E-3. Page 3 of U.S. EPA's Base Testing Reporting Template for PM2.5 Sensors
- Table Documenting Supplemental Materials and information
E-4
-------
E.1 Deployment Details
in the reporting template, you will find a summary of the testing organization name and
type, test site, air sensors, and reference instrument used (see Figures E-1, E-4, E-5,
and E-6).
Testing Organization and Site Information.
The first box includes the testing organization
and site information (Figure E-4) which is
designed to give users easy access to
information on who conducted the test, and
where and when the test was conducted.
Test organization gives users a point of
contact if they have questions about the
evaluation. Users can view the testing
organization name and type (e.g.,
manufacturer, routine testing facility,
government agency, academic institution).
This information can give users an idea of the
background (e.g.. level of experience,
objectivity of the tester) and the credentials of
the tester.
Testing location provides the site name, city and state, and latitude/longitude of the testing
site. This information may allow users to identify whether the climate at the testing
location(s) is similar to where they intend to use sensors. If the test is conducted at an
existing air monitoring site within the United States (U.S.), the report may list the U.S.
EPA's Air Quality System (AQS) Site ID This ID is a nine-digit number which uniquely
identifies the testing location and helps connect the reference instrument data within the
AQS database. For tests conducted in other countries, a similar ID or other identifier may
be provided that links the reference data to a database. Keep in mind that an AQS ID may
not be available if a testing organization sets up their own test site.
Testing Organization and Site Information
Testing organization
(Name, Organization type,
Contact website)
U.S. Environmental Protection
Agency - Office of Research and
Development
Federal Government
Air Sensor Toolbox I U.S. EPA Website
Testing location
(City, State, Latitude and
Longitude)
Ambient Monitoring Innovative
Research Station (AIRS)
RTP, NC
35.88951, -78.874572
AQS site ID
37 - 063 - 0099
Sampling timeframe
(MM-DD-YY)
06-09-21 to 07-02-21
Sensor data source
weekly downloads from the
AirCasting website
Reference data source
OAQPS file transfer
Figure E-4. Testing Organization
and Site Information Details of
the Reporting Template
E-5
-------
Sampling timeframe lists the start and end date/time of the sensor evaluation. This
information may allow users to identify whether the climate conditions during the test is
similar to the climate where they intend to use sensors. This information may also help
users identify whether climate (e.g., rainy seasons) or seasonal sources (e.g., seasonal
dust source, seasonal woodsmoke from home heating) may have impacted the test
results.
Sensor data source describes how the sensor data was obtained. For example, data
may have been collected from an on-board microSD card or through the manufacturer's
cloud server. Some sensors provide multiple ways of obtaining the data and there may
be differences within the data files depending on how they were obtained (e.g., different
time resolution, data processing method). Documenting the data source makes it easier
to access the data source especially in cases where issues are identified later.
Reference data source describes how the reference instrument data was obtained. For
example, data may have been transferred directly from an air monitoring agency that
collected it, downloaded from AirNow or AirNowTech, or downloaded from AQS. These
data sources have unique characteristics. For example, data downloaded from AirNow
is collected in real-time and several automated quality control (QC) checks are
performed before the data is posted. On the other hand, data downloaded from AQS
has undergone more quality assurance (QA) review and is certified quarterly. Therefore,
it is possible for the AirNow dataset to contain data that is later flagged as invalid and
removed from the AQS database. Documenting the data source makes it easier to
access the data and can help explain differences between similar analyses using
different data sources.
Sensor Information. The sensor
information column (Figure E-5) is
designed to give users a quick overview of
the sensor equipment tested and if there
were any issues with the equipment during
the evaluation.
Manufacturer, model applies to
commercially available sensors. Testers
are asked to list the manufacturer and
model of the sensor device. This is
important because improvements or
changes to the components of the sensor
or how those components are configured
within a device can impact sensor
performance.
Device firmware version is also requested. As mentioned previously, it is extremely
important to have a sense of how the data are processed from the raw sensor output
Sensor Information
Manufacturer, model
HabitatMap AirBeam2
Device firmware
version
Unknown (purchased Nov 2019)
Sampling time
interval
1-minute
Sensor serial
numbers
56C 63B 546
Issues encountered
during deployment?
~ Issues with deployment
Figure E-5. Sensor Information
Details of the Reporting Template
E-6
-------
into pollutant concentrations. This may happen on the device itself or within a
manufacturer's cloud-based data platform. Since manufacturers may change the
processing of data on-board, it is important that a performance evaluation includes the
configuration and data processing details. Testers can share supplemental information
on the firmware and data processing at the end of the reporting template.
Sampling time interval varies among sensors. Many sensors on the market provide data
very quickly (e.g.. a new data point every minute). Some devices produce data at a
steady time interval (i.e.. consistently every minute exactly on the minute) while others
do not. The sampling time interval box shows how often the sensor produces data.
Whatever the sampling time interval is, the data can be averaged to produce longer
time averages.
Sensor serial numbers capture the unique IDs for each device tested. EPA's Targets
Reports recommend testing at least 3 sensor devices simultaneously. This approach
allows users to better understand how sensor performance varies among identical
sensors. Recording the serial numbers helps testers connect the data files with the
sensor device being tested. For sensors that have remote data viewing platforms,
manufacturers often track the data and sensor configuration by the serial number. This
is also useful information for users when viewing graphs of the data or identifying which
sensors had issues during the testing.
Issues encountered during deployment allows testers to record any problems during the
testing. It is possible for something to go wrong during testing (e.g., sensor loses power,
a wasp makes a nest inside the sensor affecting result). If there were issues, testers
can indicate these by checking the box on the left. Additional information on these
issues can be described in more detail in the supplemental information section in the
reporting template.
FRM/FEM Information. The FRM/FEM
information column (Figure E-6) is designed
to give users a quick overview of the
reference instruments used during testing.
Manufacturer and model of the reference
instrument can be used to confirm that the
instrument has been designated as a
FRM/FEM by the U.S. EPA. The
manufacturer's website provides details on
how the pollutant concentration
measurement is made. This may help users
in interpreting similarities or differences in
testing results if the same sensor device was
tested with a different FRM/FEM monitor.
FRM/FEM Information
Manufacturer, model,
designation
Teledyne Advanced Pollution
Instrumentation T640x FEM
Sampling time
interval
1-hour averaging
Date of calibration
As required by 40 CFR Part 58 and the
Burdens Creek QAPP maintained by
OAQPS
Date of flowrate
verification check
Monthly as required by 40 CFR
Part 58 Appendix A
Description, date(s) of
maintenance activities
No maintenance activities
recorded during testing
Figure E-6. FRM/FEM Information
Details of the Reporting Template
E-7
-------
Sampling time interval shows how often the FRM/FEM monitor produces data.
Reference instruments for some pollutants may return data quickly (e.g., every minute)
while monitors for other pollutants may return data less frequently (e.g., every hour).
Date of calibration, date of flowrate verification check, and description, dates of
maintenance activities captures information on routine quality control (QC) procedures.
To maintain proper operation and high data quality, FRM and FEM monitors undergo
routine QC procedures such as
calibration checks, flowrate checks, and
regular maintenance. If the regular
schedule of maintenance and checks is
ignored, data quality can suffer. These
activities are conducted routinely at
existing air monitoring sites within the
U.S. but may not be as routine in other
locations or at temporary sites set up for
a limited number of evaluations. While
this section is designed to capture a
snapshot of information, testers can
provide additional information in the
supplemental section of the reporting
template or by citing an approved site
monitoring plan and checklist.
How Do I Find Which Instruments
are Designated as FRM/FEM
Monitors?
Instruments designated as FRM/FEM
monitors can be found on U.S. EPA's
Ambient Monitoring Technology
Information Center (AMTIC) on the
following webpage:
https://www.epa.gov/amtic/air-
monitorinq-methods-criteria-pollutants
The list of FRM/FEM monitors is
typically updated twice a year.
1-hour Averaged HabitatMap AirBeam2 PM2.5
tr 40
E
24-hour Averaged HabitatMap AirBeam2 PM2.5
06-19-21 06-24-21
— 56C — 63B — 546 — FEM
Figure E-7. Time Series Plots in the Reporting Template
E.2 Time Series Plots
The Targets Reports ask testers to create a time series plot (see example in Figure E-7)
which shows the data from each of the three identical sensors alongside data from the
reference monitor as a function of time. The purple, blue, and red lines represent the
E-8
-------
sensors arid the black line represents the reference monitor Some reports, like the
example in Figure E-7, may include more than one plot, showing data at different
averaging intervals (e.g., 1-hour, 8-hour, 24-hour), All averages should represent the
time period they describe. A data completeness level of at least 75% is recommended
when calculating these averages (see Section 3.7.2 and Table 3-2). Testers are asked
to share how they calculate data averages if they do not follow the recommendation in
the Targets Reports.
f \
Do other Testing Organizations Provide Details on the Instruments Evaluated
and Testing Conditions in their Evaluation Reports?
Yes! Most testing organizations document these details, but they may not be
presented all in one place. Some details may be held in private files, within longer
reports, or within various locations in a testing report. Here is one example report from
AQ-SPEC where similar information is presented within a draft field evaluation report.
Background
From 07/20/2018 to 09/19/2018, three HabitatMap AirBeam2 (hereinafter AirBeam2) sensors
were deployed at a SCAQMD stationary ambient monitoring site in Rubidoux and were run
side-by-side with three reference instruments measuring the same pollutants
AirBeam2 (3 units tested):
> Particle sensor (optical; non-FEM)
> PM sensor: Plantower PMS7003
> Each unit measures: PM, 0, PM25 and PM10 (pg/m3) Temperature
(°F), Relative Humidity (%) (measures T and RH inside of sensor)
^ Unit cost: ~$250
>Time resolution: 1-min
> Units IDs: F4F1,6FE0, 63CC
> Differences from 1st Generation:
¦ Different hardware (temp/RH sensor, PM sensor) and design
¦Firmware: 3.19.18AirBeam2
¦ Wi-Fi and cellular capabilities
¦ Different microcontroller
¦ Measures PM, 0, PM25and PM10mass conc. only
MetOne BAM Ireference instrument):
> Beta-attenuation monitor
(FEM PM2 5 & PM10)
> Measures PM25& PM10 (pg/m3)
> Unit cost: -$20,000
> Time resolution: 1-hr
GRIMM (reference instrument):
> Optical particle counter (FEM PM2 5)
> Measures PM, 0, PM2 5, and PMt0
(pg/m3)
> Cost: -$25,000 and up
> Time resolution: 1-min
¦ Teledvne API T640 (reference instrument):
> Optical particle counter (FEM PM2 5)
> Measures PM25& PM10 (pg/m3)
> Unit cost: -$21,000
> Time resolution: 1-min
2
E-9
-------
The primary purpose of the time series plot(s) is to show how measurements change
over time. These plots help users determine if the sensors capture the same variation in
measurements (trends) as the reference monitor over the testing period. If some time
periods show poor agreement, users should review the supplemental information in the
reporting template to understand if there were any issues during those time periods that
may have influenced sensor performance.
The time series plots may also
give users insight into whether
sensors over- or under-
estimate pollutant
concentrations. When sensor
measurements are lower than
the reference instrument
measurements, that is referred
to as underestimating
concentrations. When sensor
measurements are higher than
the reference monitor
measurements, that is referred
to as overestimating
concentrations. Some sensors
may show a combination of
underestimating and
overestimating concentrations.
Lastly, the time series plots
may show data spikes, or data
points with higher
concentrations, from one or
more of the instruments. Users
may want to pay close attention
to these spikes. Some spikes may represent elevated pollutant concentrations outdoors.
For example, particulate matter (PM) concentrations may go up for a short period of
time because of nearby mowing activities. Other spikes may not reflect real changes in
pollutant concentrations and may instead indicate that something is wrong with the
device. Past sensor evaluation efforts have found several sensors that report false
information occasionally as a result of incorrectly logged data or a device error. Users
should check the supplemental information in the testing report to identify if any of the
data spikes were removed from the dataset. If this information is not available, users
should contact the testing organization.
Are All Time Series Plots Similar?
Yes and no. Time series plots always show time
along the horizontal x-axis and pollutant
concentration along the vertical y-axis. There may
be differences in the colors used, time averaging
interval, and number of sensors plotted. The image
example below show an example from an AQ-
SPEC report. Occasionally you might also see a
secondary vertical y-axis if pollutant concentrations
are reported in different units or if multiple
pollutants are plotted on one graph.
-FEM GRIMM
'i miiiMiiwinwi
1/26/16 1/27/16 1/28/16 1/29/16 1/30/16 1/31/16 2/1/16 2/2/16 2/3/16 2/4/16
E-10
-------
E.3 Scatter Plots
The next part of the reporting template asks testers to create a scatter plot (see
example in Figure E-8) which graphs data from the sensors against data from the
reference monitor. In Figure E-8, 1-hour or 24-hour averaged data are shown. The
Targets Reports recommend making this plot with the sensor data on the y-axis and the
reference data on the x-axis. Other testing organizations may plot the data differently so
be careful to note which axis the data is plotted on when interpreting these plots. The
scatter plot includes the slope-intercept line calculated using the ordinary least-sguares
regression equation (see Section 3.6.2). As a reminder, the slope and intercept
describe the bias in the sensor data. If the slope is greater than 1, this means that the
sensor measurements are higher than reference instrument measurements (i.e., the
sensor overestimates concentrations). If the slope is less than 1, this means that the
sensor measurements are lower than the reference measurements (i.e., the sensor
underestimates concentrations)
In the reporting template, testers are asked to include the scatter plot for one of the
three sensors against the reference instrument data. Testers are encouraged to include
plots for each individual sensor in the supplemental information in the reporting
template. Users should check this information to understand how the other identical
sensors compared to the reference instrument.
The data points in some scatter plots may have different colors that represent another
parameter [e.g., temperature (T), relative humidity (RH)]. The purpose for this is to give
users more insight into data. The legend should describe what the colors represent. For
example, RH is known to affect the performance of some fine particulate matter (PM2.5)
sensors. In Figure E-8, the data points in the scatter plot have different colors based on
the RH levels (red points represent higher RH levels and blue points represent lower
RH levels). If RH affects the performance of a sensor, users may see red dots in one
area and blue dots in another.
E-11
-------
50
E
oi
HabitatMap AirBeam2 vs. FEM
1-hour PM2.5
: 1.05X-2.91
40
0.70
RMSE = 3.38
N = 512
30
a. 20
O
CO
10 10
10 20 30 40
FEM PM2.5 (MQ/m3)
50
20
15
10
O
CO
m
HabitatMap AirBeam2 vs. FEM
24-hour PM2.5
y= 1.13X-3.51
R2 = 0.73
w = 20
~ /
~ /
~ /
~ /
/ /
•V
/ ^
•
/'Z'
/
0 5 10 15
FEM PM2.5 (MQ/m3)
20
Relative Humidity (%)
1 1
40 60
20
80
100
Figure E-8. Scatter Plot in the Reporting Template
Some scatter plots, like those shown in Figure E-8, may include more details or
statistics about the data that is plotted. For example, the plot includes the number of
data points (N) plotted, and the root mean square error (RMSE), which is a measure of
error in the sensor measurements.
Although this scatter plot uses a linear regression to describe the relationship between
the sensor and reference data, some sensors or pollutant types may need more
complicated function such as a multilinear regression to describe the relationship. A
multilinear regression includes two or more variables to predict the outcome of another
variable.
E-12
-------
/ \
Do Some Testing Organizations Choose to Make Scatterplots Differently?
Yes. The recommendations provided by the U.S. EPA are voluntary and testing
organizations may choose to plot data differently. Some have chosen to plot
reference data on the y-axis and sensor data on the x-axis instead. For example, the
plot below is from a draft field evaluation conducted by AQ-SPEC from 2020. This
makes the bias interpretation different. More explicitly, because the axes are
switched, a slope less than 1 means that the sensor is overestimating concentrations.
For simplicity, users can pick a point on the graph and use the x and y coordinates to
determine if the sensor reads higher or lower than the reference measurement.
PM2.5 (1-hr mean, ng/m3)
50
v = 0.7899x + 3.7008
0 10 20 30 40 50
Unit F4F1
E.4 Performance Evaluation Metrics and Target Values
The Targets Report ask testers to plot the calculated performance metrics (see Figure
E-9). A series of plots are included which have dark gray shaded regions that indicate
the target value ranged. The plots can help users quickly identify whether a sensor
meets the target value for each metric, how close the sensor is to meeting the target
value (if the target is not met), and the total number of target values that the sensor
meets.
E-13
-------
1.0
R2
2.00
1.75
Slope
10.0
7.5
Intercept (pg/m3)
10
RMSE (pg/m3)
NRMSE (%)
50
CV {%)
10
SD (pg/m3)
60
0.8
1.50
5.0
8
50
40
8
» •
0.6
O
1.25
2.5
6
40
©
9
1.00
0 9
O A
0.0
6
30
0.4
0.75
-2.5
8 8
4
° •
20
4
0.50
0.25
-5.0
-7.5
20
0.2
2
10
10
© ©
2
0.0
0.00
-10.0
0
0
0
0
O 0
1-hour 24-hour
Averaging Interval
1-hour 24-hour
Averaging Interval
1-hour 24-hour
Averaging Interval
1-hour 24-hour
Averaging Interval
1-hour 24-hour
Averaging Interval
1-hour 24-hour
Averaging Interval
1-hour 24-hour
Averaging Interval
Figure E-9. Performance Metrics in the Reporting Template
When viewing this figure, dots
in the light gray space indicate
that the sensor does not meet
the target whereas dots in the
dark gray space indicate that
the sensor meets the target.
The numerical values for each
performance metric are also
shown in a table on the second
page of the report (see Figure
E-10). Target values are
included again within the table
for easy reference.
/ \
Do All Evaluations Use the Same Performance
Metrics and Target Values?
No. U.S. EPA's Targets Reports list specific
performance metrics, give detailed calculations for
each metric, and include target values for those
metrics. Testers using other protocols developed
by other organizations may report different
performance metrics. As an example, the image
below shows a report from AIRLAB which uses
different metrics and visualizations. If needed,
users should ask the testing organization to
explain the performance metric calculations and
interpretation of results.
Evaluation
Form factor
E-14
-------
Tabular Statistics
Sensor-FRM/FEM Correlation
Bias and Linearity
Data Quality
R2
Slope
Intercept
Uptime
(%)
Number of paired
sensor and FRM/FEM
concentration values
1-Hour
••o
24-Hour
•••
1-Hour
•••
24-Hour
• ••
1-Hour
•••
24-Hour
•••
1-Hour
•••
24-Hour
• ••
1-Hour
24-Hour
Metric Target Range
S0.70
£ 0.70
1.0 ±0.35
1.0 ± 0.35
-5 £ b £ 5
-5 £ b £ 5
75%*
75%*
Sensor 56C
0.70
0.73
1.05
1.13
-2.91
-3.51
95
87
512
20
Sensor 63B
0.68
0.71
1.15
1.17
-3.63
-3.80
90
83
485
19
Sensor 546
0.71
0.72
0.96
1.03
-3.01
-3.69
96
91
515
21
Mean
0.70
0.72
1.05
1.11
-3.18
-3.66
93
87
504
20
Error
RMSE
NRMSE
(ne/m1)
(%i
1-Hour
24-Hour
1-Hour
24-Hour
*
~
i*
-------
20
15
fO
jD
O
ra
_ra
CD
q:
0,0
25 50 75 100
Reference Relative Humidity {%)
Figure E-11. Meteorological Conditions in the Reporting Template
These graphs can help users understand how similar or different the RH and T
evaluation conditions were compared to where they would like to use the sensor. Users
should keep in mind that if the meteorological conditions are very different, sensors may
perform differently.
Evaluation reports may have additional graphs that show how meteorology impacts
sensor performance. The graphs may vary based on the sensor and/or pollutant type.
E-16
-------
Appendix F: Glossary
A-B-C-D-E-F-G-H-l-J-K-L-M-N-O-P-Q-R-S-I-U-V-W-X-Y-
Z
-A-
accuracy:
A measure of the agreement between the pollutant concentrations reported by the sensor and the reference
instrument. This includes a combination of random error (precision) and systematic error (bias) components
which are due to sampling and analytical operations. One way to measure this agreement is by calculating the
root mean square error. See Section 3.4.1.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
aerosol:
Solid or liquid droplets suspended in air originating from biogenic sources (e.g., sea salt spray, volcanoes) and
from anthropogenic sources (e.g., fossil fuel combustion, biomass burning). Aerosols can also form in the
atmosphere from reactions of chemical precursors (e.g., reaction of ammonia, sulfur dioxide, and water vapor
to form ammonium sulfate). See Section 2.3.
Source: https://earthobservatory.nasa.gov/features/Aerosols
air quality:
A relative measure of the amount of pollution present in the air. Good air quality means less air pollution, while
poor air quality means more air pollution. See Section 2.1.
Source: https://www3.epa.gov/airguality/cleanair.html
Air Quality Index (AQI):
U.S. EPA's index for reporting daily air quality that characterizes air pollution levels and associated health
effects that might be of concern. EPA calculates the AQI for five criteria pollutants. AQI values range from 0 to
500; the higher the AQI value, the greater the level of air pollution and the greater the health concern. AQI
values at or below 100 are generally thought of as satisfactory. When AQI values are above 100, air quality is
unhealthy: at first for certain sensitive groups of people, then for everyone as AQI values get higher. See
Sections 2.4 and Z5.
Source: https://www.airnow.gov/agi/agi-basics/
Air Quality System (AQS):
An electronic repository of ambient air pollution data collected by the U.S. EPA, state, local, and tribal air
pollution control agencies from over thousands of monitors The AQS also contains meteorological data,
descriptive information about each monitoring station (including its geographic location and its operator), and
data guality assurance/guality control information. See Section 2.3.
Source: https://www.epa.gov/ags
F-1
-------
air sensor:
A class of non-regulatory technology that is lower in cost, portable, and generally easier to operate than
regulatory monitors. Air sensors often provide relatively quick or instant air pollutant concentrations (both gas-
based and particulate matter) and allow air quality to be measured in more locations. The term 'air sensor'
often describes an integrated set of hardware and software that uses one or more sensing elements (also
sometimes called sensors) to detect or measure pollutant concentrations. See Section 1.1.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
air toxics:
See hazardous air pollutants.
anthropogenic emissions:
Emissions that originate from human activities or result from natural processes that have been affected by
human activities (e.g., fuel combustion, solvent use, biomass burning). See Section 2.1.
Source: https://www.epa.gov/sites/production/files/2021-02/documents/us-qhq-inventorv-2021-main-text.pdf
-B-
benzene, toluene, ethylbenzene, and xylene (BTEX):
Mixture of four volatile organic compounds (benzene, toluene, ethylbenzene, and xylene) that are normally
grouped as they are often found together. Primary sources of BTEX include on-road and non-road gasoline
vehicles and engines, petroleum transport/storage, and solvent usage. See Section 2.1.
Source:
https://mde.marvland.qov/proqrams/LAND/OilControl/Documents/BTEX%20Fact%20Sheet%202.12.07%202%
20pqs.pdf
bias:
The systematic (non-random) or persistent disagreement between the concentrations reported by the sensor
and reference instrument. It is often determined using the linear regression slope and intercept of a simple
linear regression, fitting sensor measurements (y-axis) to reference measurements (x-axis). See Section 3.4.1.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
biogenic emissions:
Air emissions that originate from a natural source such as vegetation, soils, volcanic emissions, lightning, sea
salt, etc. Also called natural sources. See Section 2.1.
Source: https://www.epa.gov/air-emissions-modelinq/bioqenic-emission-sources
black carbon (BC):
Most strongly light-absorbing component of particulate matter (PM) that is formed by the incomplete
combustion of fossil fuels, biofuels, and biomass. BC is emitted directly into the atmosphere in the form of fine
particulate matter (PM2.5). BC is the most effective form of PM, by mass, at absorbing solar energy - per unit of
mass in the atmosphere, BC can absorb a million times more energy than carbon dioxide (CO2). BC is a major
component of "soot", a complex light-absorbing mixture that also contains some organic carbon (OC). See
Section 2.1.
Source: https://www3.epa.gov/airqualitv/blackcarbon/basic.html
F-2
-------
-c-
calibration:
A procedure for checking and adjusting an instrument's settings so that the measurements produced are
comparable to a certified standard value. See Section 3.6.
Source: https://www3.epa.gov/ttnamti1/files/ambient/pm25/qa/vol2sec12.pdf
certification:
A process where an organization carries out an agreed upon test method(s) set by standards to make sure
tests are conducted in the same way every time. The certification process often results in a certificate or
specific label [e.g., ENERGY STAR label, Underwriter Laboratories (UL) listing)]. See Section 4.3.1.
citizen science:
See participatory science.
Code of Federal Regulations (CFR):
The codification of the general and permanent rules published in the Federal Register (a daily publication of
the U.S. Federal Government that issues proposed and final administrative regulations of federal agencies) by
the executive departments and agencies of the U.S. Federal Government. CFR is divided into 50 titles that
represent broad areas subject to Federal regulation, and each title is divided into chapters typically bearing the
name of the issuing agency. Each chapter is further subdivided into parts (and subparts, where needed) that
cover specific regulatory areas. See Section 2.3.
Source: https://www.archives.gov/federal-register/cfr/about.html
collocation:
The process by which a sensor and a reference instrument are operated at the same time and place under real
world conditions. The siting criteria (e.g., proximity and height of the sensor and the reference monitor) should
follow procedures outlined in 40 CFR Part 58 as closely as possible. For example, sensors should be placed
within 20 meters horizontal of the reference instrument, positioned such that the sample air inlets for the
sensors are within a height of ± 1 meter vertically of the sample air inlets of the reference instrument, and
placed as far as possible from any obstructions (e.g., trees, walls) to minimize spatial and wind turbulence
effects on sample collection. See Section 3.6.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
community science:
See participatory science.
comparability:
The level of overall agreement between two separate data sets. This term is often used to describe how well
sensor data compares with reference instrument data. Comparability is a combination of accuracy, precision,
linearity, and other performance metrics. See Section 3.4.1.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
F-3
-------
completeness:
In determining averages, completeness describes the amount of valid data obtained relative to the averaging
period. The purpose of the completeness threshold is to make sure that the average is representative of the
concentrations observed within the averaging period. For example, if a sensor collects measurements every 5
minutes, it can return 12 measurements every hour. To obtain 75 percent data completeness for a calculated
hourly average, at least 9 valid measurements are needed (i.e., 9/12 * 100 percent = 75 percent). See Section
3.7.2.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
concentration:
The metric for reporting the amount of a pollutant in the air. Concentration represents the weight or number of
molecules in a volume of air. Common units include microgram per cubic meter (|jg/m3), parts per million
(ppm), and parts per billion (ppb). For example, a concentration of 43 |jg/m3 is the weight of 43 micrograms (a
microgram is one millionth of a gram) per cubic meter of air. Parts per billion is the number of units of mass of
a pollutant per 1 billion units of the total mass of the air. See Section 2.1.
Source: https://en.wikipedia.org/wiki/Air pollutant concentrations
correction:
The adjustments to sensor measurement data to more closely match the measurement data collected by a
reference monitor. See Section 3.6.
criteria pollutant:
A group of six widespread and common air pollutants for which EPA established National Ambient Air Quality
Standards (NAAQS) under the Clean Air Act. The criteria pollutants include carbon monoxide (CO), lead (Pb),
ground-level ozone (O3), nitrogen dioxide (NO2), particulate matter (PIVksand PM10), and sulfur dioxide (SO2).
See Section 2.3.
Source: https://www.epa.gov/criteria-air-pollutants
cross-sensitivity:
Other pollutants that interfere with the measurement of the target pollutant. See Section 3.1.2.
-D-
data aggregation:
The process of compiling information to prepare combined datasets for data processing. See Appendix D.
data dictionary:
Detailed description of names and definitions of parameters collected in a database. See Section 3.8.2.
data handling:
The process of collecting, storing, processing, and presenting or reporting data. See Section 3.4.2.
data management system (DMS):
A collection of procedures and software needed to acquire, process, and distribute data. See Section 3.7.3.
F-4
-------
data quality objectives (DQO):
Quantitative acceptance criteria for the quality and quantity of data to be collected, relative to the ultimate use
of the data. See Section 3.3.
Source: https://www.epa.gov/sites/default/files/2015-06/documents/q5-final.pdf
data spikes:
Data points with higher concentrations from one or more instruments. See Section 3.6.2.
data transmission:
Sending or receiving data using a wireless or cable-based relay system. See Section 3.4.1.
data validation:
A pollutant- and sample-specific process that extends the evaluation of data beyond method, procedural, or
contractual compliance (i.e., data verification) to determine the analytical quality of a specific data set. See
Section 3.1.2.
Source: https://www.epa.gov/sites/default/files/2015-06/documents/q5-final.pdf
deployment:
The placement or arrangement of air sensors or instruments for a specific monitoring purpose or objective. See
Section 3.5.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
detection limit:
The lowest concentration that can be determined as being above zero by a single measurement at a stated
level of certainty. There are many types of detection limits, such as the Method Detection Limit (MDL) which is
typically defined as 99% confidence that the measurement is not instrument noise. See Section 3.4.1.
Source: https://www.epa.gov/sites/default/files/2015-06/documents/q5-final.pdf
detection range:
The lowest and highest pollutant concentration that an air sensor can reliably measure. Also called
measurement range. See Section 3.4.1.
downtime:
The period of time for which an air sensor or instrument is unavailable for use (e.g., during power interruption
or maintenance). See Section 3.1.2.
downwind:
Where air goes after moving over an area of interest. See Section 2.1.
Source: https://www.epa.gov/interstate-air-pollution-transport/what-interstate-air-pollution-transport
drift:
A change in the response or concentration reported by a sensor when challenged by the same pollutant
concentration over a period during which the sensor is operated continuously and without adjustment. See
Section 3.1.2.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
F-5
-------
-E-
electrochemical sensor:
A type of air sensor where the target gas interacts with an electrode thereby producing an electrical current
that is proportional to the concentration of the target gas. See Section 3.7.2.
Source: https://nepis.epa.gov/Exe/ZyPDF.egi/P100F2G5.PDF?Dockey=P100F2G5.PDF
error:
A measure of the disagreement between the pollutant concentrations reported by the sensor and the reference
instrument. One way to measure error is by calculating the root mean square error (RMSE). Additional ways
include calculating the mean bias error (MBE), mean absolute error (MAE), among others. See Section 3.4.1.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
exposure:
Contact of a chemical, physical, or biological agent (e.g., ozone) with the outer boundary of an organism.
Exposure is quantified as the concentration of the agent in the medium in contact integrated over the time
duration of that contact. See Section 2.5.
Source: https://www.epa.gov/sites/production/files/2014-11/documents/guidelines exp assessment.pdf
exceedance:
Occurs when a measured concentration of a criteria pollutant exceeds the concentration level for the averaging
period specified by the National Ambient Air Quality Standards (NAAQS). A NAAQS exceedance does not
constitute a NAAQS violation. See Section 2.4.
-F-
Federal Equivalent Method (FEM):
A method for measuring the concentration of an air pollutant in the ambient air that has been designated as an
equivalent method in accordance with 40 CFR Part 53. A FEM does not include a method for which an
equivalent method designation has been canceled in accordance with 40 CFR Parts 53.11 or 53.16. See
Section 2.3.
Source: https://www.ecfr.gov/current/title-40/chapter-l/subchapter-C/part-53
Federal Reference Method (FRM):
A method of sampling and analyzing the ambient air for an air pollutant that is specified as a reference method
in 40 CFR Part 50, or a method that has been designated as a reference method in accordance with 40 CFR
Part 53. A FRM does not include a method for which the U.S. EPA has cancelled a reference method
designation in accordance with 40 CFR Parts 53.11 or 53.16. See Section 2.3.
Source: https://www.ecfr.gov/current/title-40/chapter-l/subchapter-C/part-53
fine particulate matter (PM2.5):
See PM? 5.
F-6
-------
firmware:
A type of computer software or set of instructions programmed on a hardware device (e.g., an air sensor). See
Section 3.7.3.
Source:
https://www.techopedia.com/definition/2137/firmware#:~:text=Firmware%20is%20a%20type%20of,tasks%20a
nd%20functions%20as%20intended
-H-
hotspot:
An area of localized, increased pollutant concentrations (e.g., a congested roadway intersection). See Section
32.
Source: https://www.qreenfacts.org/qlossarv/abc/air-pollution-hot-spot.htm
hazardous air pollutants (HAPs):
Also called air toxics. Air pollutants that are known or expected to cause cancer or other serious health effects,
such as reproductive effects or birth defects, or adverse environmental and ecological effects. Examples of
toxic air pollutants include benzene (found in gasoline), perchloroethylene (emitted from some dry-cleaning
facilities), and methylene chloride (used as a solvent by a number of industries). Example of other listed air
toxics include dioxin, asbestos, toluene, and metals such as cadmium, mercury, chromium, and lead
compounds. See Section 2.3.
Source: https://www.epa.gov/haps
hyperlocal emissions source:
An emissions source that releases brief, but high pollutant concentrations that can influence the air quality of
nearby locations but are not representative of the larger area. See Section 3.5.2.
Source: https://www.epa.gov/air-sensor-toolbox/guide-siting-and-installing-air-sensors
-I-
interferent:
Any non-target pollutant(s) that might skew or influence a sensor's response to the target pollutant. Interferents
may have a positive or negative effect on a sensor signal. See Section 3.1.2.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
-L-
linear regression:
Also called simple linear regression or least-squares regression. A common statistical approach that models
the relationship between one variable as a function of another variable. Linear regression estimates the
equation: y = mx + b (where "b" is the y-intercept and "m" is the slope) by finding values for the parameters for
"b" and "m" that minimize the sum of the squared deviations between the observed responses and the linear
equation. See Section 3.6.2.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
F-7
-------
linearity:
A measure of the extent to which the measurements reported by a sensor can explain the concentrations
reported by the reference instrument. It is often quantified by the coefficient of determination (R2) obtained from
the simple linear regression fitting sensor measurements (y-axis) to reference instrument measurements (x-
axis) with values closer to 1 generally indicating better linearity. In some cases, sensor measurements can be
linear with a near perfect R2 but may differ significantly from the reference instrument measurements. For
example, a linear regression can result in an R2 of 0.99 and slope of 5. This indicates that the reported sensor
measurement is always 5 times higher than the reference instrument measurements. See Section 3.6.2.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
long-term:
A time period covering months to years. See Section 2.2.
-M-
maintenance:
Preventative actions taken to maintain sensor performance and deployment site conditions over the
measurement period. Maintenance can include regularly cleaning of internal surfaces to prevent the buildup of
bugs or dust, replacing filters, and examining site conditions for any changes (e.g., vandalism or overgrown
trees). Manufacturers may also provide air sensor maintenance activities. See Section 3.7.1.
measurement frequency:
The number of measurements collected per unit of time. See Section 2.3.
Source: https://www.epa.gov/air-emissions-monitoring-knowledge-base/basic-information-about-air-emissions-
monitoring
micrometer:
One millionth of a meter. See Section 2.1.
multilinear regression
A statistical approach that describes the relationship among two (2) or more variables to predict the outcome of
another variable. See Appendix E.
-N-
National Ambient Air Quality Standards (NAAQS):
Standards established by the U.S. EPA that apply to outdoor air throughout the United States. The Clean Air
Act (CAA) establishes two types of NAAQS. Primary standards set limits to protect public health, including the
health of "sensitive" populations such as asthmatics, children, and the elderly. Secondary standards set limits
to protect public welfare, including protection against decreased visibility and damage to animals, crops,
vegetation, and buildings. EPA has set NAAQS for the six criteria pollutants. See Section 2.4.
Source: https://epa.gov/naags
near-reference instrument:
An instrument that does not have a Federal Equivalent Method (FEM) designation but has many of the
features of an FEM, and when operated by trained staff, can provide air pollution data with sufficient accuracy
and quality. See Section 2.3.
F-8
-------
non-regulatory supplemental and informational monitoring (NSIM):
A term used by the U.S. EPA that describes monitoring applications conducted for purposes other than
demonstrating compliance with local, state, or federal air quality regulations. There are three NSIM categories
including spatiotemporal variability (e.g., daily trends, gradient studies, air quality forecasting, participatory
science, education), comparison (e.g., hotspot detection, data fusion, emergency response, supplemental
monitoring), and long-term trend (e.g., long-term changes, epidemiological studies, model verification). See
Section 1.1.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
-o-
optical technology:
A type of technology used in some particulate matter air sensors. A light receptor detects light scattered by
particles and the amount of light scattering (or absorption) is converted into particle count and mass
concentration values. See Section 3.7.2.
Source: http://www.agmd.gov/ag-spec/resources/operational-principles
overestimate:
Sensor measurements are higher than the reference instrument measurements. Also called over-report. See
Section 3.6.2.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
-P-
parameter:
Any of a set of physical properties whose values determine the characteristics or operation of an air sensor or
instrument. See Section 3.4.2.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
participatory science:
Activity that engages the public in advancing scientific knowledge by formulating research questions, collecting
data, and interpreting results. Other terms include citizen science, community science, volunteer monitoring, or
public participation in scientific research. See Section 1.1.
Source: https://www.epa.gov/participatory-science
performance evaluation:
A test that compares sensor data to reference instrument data. Reference instruments are used as they
provide highly accurate measurements and are the "gold standard". See Section 4.2.
Source: http://www.agmd.gov/ag-spec/home
performance metric:
A parameter used to describe the data quality of a measurement device. See Section 3.4.2.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
F-9
-------
performance targets:
Numeric benchmarks for assessing the measurement performance of an air sensor. See Section 3.4.2.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
personal exposure monitoring:
Measurements of an individual's exposure or contact with a health hazard. See Section 1.1.
Source: https://www.epa.gov/expobox/exposure-assessment-tools-approaches-direct-measurement-point-
contact-measurement
PM1.0:
Particles with diameters generally less than 1 micrometer (jjm). See Section 2.1.
PM2.5:
Also called fine particulate matter. Fine inhalable particles, with diameters generally less than 2.5 micrometers
(|jm). See Section 2.1.
PM10:
Inhalable particles, with diameters generally less than 10 micrometers (jjm). See Section 2.1.
polycyclic aromatic hydrocarbons (PAHs):
A group of over 100 different chemicals that are formed during the incomplete burning of coal, oil and gas,
garbage, or other organic substances like tobacco or charbroiled meat. PAHs are usually found as a mixture
containing two or more of these compounds, such as soot. Some PAHs are manufactured. These pure PAHs
usually exist as colorless, white, or pale yellow-green solids. PAHs are found in coal tar, crude oil, creosote,
and roofing tar, but a few are used in medicines or to make dyes, plastics, and pesticides. See Section 2.1.
Source: https://www.epa.gov/sites/production/files/2015-04/documents/walter atsdr pahs.pdf
precision:
Variation around the mean of a set of measurements obtained concurrently by two (2) or more sensors of the
same type collocated under the same sampling conditions. The consistency in measurements from identical
sensors is often quantified by standard deviation (SD) or the coefficient of variation (CV), with lower values
indicating a more precise measurement. See Section 3.4.1.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
primary air pollutant:
A pollutant that is emitted into the atmosphere directly from a source such as construction sites, unpaved
roads, smokestacks, or fires. See Section 2.1.
Source: https://www.epa.g0v/pmcourse/what-particle-pollution#where
primary standard:
A type of national ambient air quality standards (NAAQS) that provides public health protection, including
protecting the health of "sensitive" populations such as asthmatics, children, and the elderly. See Section 2.4.
Source: https://www.epa.gov/criteria-air-pollutants/naags-table
F-10
-------
-Q-
qualitative measurement:
A measurement that is descriptive, conceptual, and often expressed in words. For example, pollutant
concentrations described as "higher" or "lower". See Section 4.4.
quality assurance (QA):
Planned steps performed to manage a project and collect, assess, and review data to ensure that
measurements meet the data quality needed for the monitoring objective. An example QA activity is developing
a plan for air monitoring. See Section 3.7.2.
Source: https://www.epa.gov/sites/default/files/2015-06/documents/q5-final.pdf
quality assurance project plan (QAPP):
A plan that describes the activities of a monitoring project involved with the acquisition of environmental
information whether generated from direct measurements activities, collected from other sources, or compiled
from computerized databases and information systems. The QAPP documents the results of a project's
technical planning process, providing in one place a clear, concise, and complete plan for the environmental
data operation and its quality objectives and identifying key project personnel. The QAPP communicates the
specifications for implementation of the project design to all parties and ensures that the quality objectives are
achieved for the project. See Section 3.3.
Source: https://www.epa.gov/sites/default/files/2015-06/documents/q5-final.pdf
quality control (QC):
Steps performed to limit error from instruments or in measurements during a project. Examples of QC activities
include collocation, correction of data, maintenance, automatic data checks, and data review. See Section
3.7.1.
Source: https://www.epa.gov/sites/default/files/2015-06/documents/q5-final.pdf
quantitative measurement:
A measurement which can be expressed using numbers. For example, a pollutant concentration expressed in
parts per billion (ppb) or micrograms per cubic meter (|jg/m3). See Section 4.4.
-R-
radiative forcing:
A heating effect caused by greenhouse gases in the atmosphere. Radiative forcing is calculated in watts per
square meter, which represents the size of the energy imbalance in the atmosphere. See Section 2.2.
Source: https://www.epa.gov/climate-indicators/climate-chanqe-indicators-climate-forcinq
regulatory monitoring:
Monitoring conducted for the purposes of demonstrating compliance with the local, state, or federal air quality
regulations. See _.
Source: https://www.epa.gov/amtic/requlations-quidance-and-monitorinq-plans
F-11
-------
remote sensing:
The process of detecting and monitoring the physical characteristics of an area by measuring its reflected and
emitted energy or radiation from a distance (e.g., from satellite-based or aircraft-based instruments). See
Section 2.3.
Source: https://www.usgs.gov/faqs/what-remote-sensing-and-what-it-used7qt-news science products=Q#gt-
news science products
representativeness:
A description of how closely a sample reflects the characteristics of the whole. Although challenging to verify,
effort should be made to ensure that a sample is representative using techniques such as thorough mixing to
obtain homogeneity, duplicate analyses, etc. For example, the data completeness threshold suggested in this
report is meant to ensure that measurements averaged to longer time intervals are as representative as
possible by covering at least 75% of the time period. See Section 3.7.2.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
response time:
The amount of time required for a sensor to respond to a change in concentration. See Section 3.4.1.
Source: https://www.epa.qov/air-sensor-toolbox/air-sensor-performance-tarqets-and-testinq-protocols
-s-
scatter plot:
A plot that shows the relationship between two variables, one on the x-axis and one on the y-axis. For sensor
data, scatter plots can help explore the relationship between two parameters of interest, understand how a
sensor compares to a regulatory monitor, or understand if a sensor overestimates or underestimates pollutant
concentrations. See Section 3.6.2.
Source: https://cfpub.epa.gov/si/si public record report.cfm?dirEntryld=354208
secondary air pollutant:
A pollutant that is formed when other primary air pollutants react in the atmosphere. An example of a
secondary pollutant is ground-level ozone (O3), which forms from chemical reactions involving airborne
nitrogen oxides (N0X), airborne volatile organic compounds (VOCs), and sunlight. See Section 2.1.
Source: https://www.mrgscience.com/ess-topic-63-photochemical-smog.html
secondary standard:
A type of national ambient air quality standards (NAAQS) that provides public welfare protection, including
protection against decreased visibility and damage to animals, crops, vegetation, and buildings. See Section
2A.
Source: https://www.epa.gov/criteria-air-pollutants/naags-table
sensor lifespan:
The time period during which the air sensor is designed to function normally. See Appendix C.
sensor network:
Two (2) or more air sensors that collect pollutant concentration or other data (e.g., relative humidity,
temperature) from different locations and transmit the measurements to a central repository. See Section 3.3.
F-12
-------
sensor node:
An individual sensor within a sensor network. See Section 1.1.
short-term:
A time period covering seconds to weeks. See Section 2.2.
specification sheet:
A document that presents the detailed technical aspects and characteristics of an item or product. See Section
3.4.1.
specificity:
The ability of a sensor to measure the pollutant of interest (target pollutant). See Section 4.4.
standard operating procedure (SOP):
A set of written instructions that detail the one-time and repetitive activities to be conducted or followed within
an organization. An SOP provides individuals with the information to perform a job properly, which facilitates
consistent conformance to technical and quality system requirements and supports data quality. See Appendix
B.
Source: https://www.epa.gov/sites/production/files/2015-06/documents/g6-final.pdf
standards:
As related to sensors, a voluntary process where technology testing methods are agreed upon by authorities,
manufacturers, customers, and others invested in the performance of the technology. See Section 4.3.1.
State Implementation Plan (SIP):
A collection of regulations and documents used by a state, territory, or local air district to implement, maintain,
and enforce the National Ambient Air Quality Standards (NAAQS), and to fulfill other requirements of the Clean
Air Act. See Section 2.4.
Source: https://www.epa.gov/air-gualitv-implementation-plans/basic-information-about-air-guality-sips
supplemental monitoring:
An application where sensors are placed in locations that do not have an existing regulatory monitor(s).
Typically, the goal is to fill in gaps in areas where there are no or limited regulatory monitors or to identify
sources of interest or locations reguiring further study and/or monitoring. See Section 1.1.
Source: https://cfpub.epa.gov/si/si public record report.cfm?dirEntryld=354208
-T-
target pollutant:
A pollutant of interest for a measurement. See Section 3.4.1.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
time averaging interval:
The time period over which raw measurements are averaged. See Appendix E.
F-13
-------
-u-
ultrafine particles (UFP):
Particles in the atmosphere that have diameters generally less than 0.1 micrometer (jjm). UFP are created by
combustion processes and chemical reactions in the atmosphere. See Section 2.1.
Source: https://www.epa.gov/pmcourse/particle-pollution-exposure
underestimate:
Also referred to as under-report. Sensor measurements are lower than the reference monitor measurements.
See Section 3.6.2.
Source: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
upwind:
Where air moves through before it goes over an area of interest. See Section 2.1.
Source: https://www.epa.gov/interstate-air-pollution-transport/what-interstate-air-pollution-transport
-V-
violation:
Occurs when a measured concentration of a criteria pollutant exceeds the concentration level for the averaging
period specified by the National Ambient Air Quality Standards (NAAQS) for specific criteria over a specified
timeframe. See Section 2.4.
visibility:
The degree of perceived clarity (e.g., contrast, coloration, and texture elements) when viewing objects at a
distance. Visibility can be impaired by haze caused by air pollutant emissions from numerous sources
distributed over a wide geographic area. See Section 2.1.
Source: https://www.epa.gov/sites/production/files/2015-05/documents/1999hazefacts.pdf
volatile organic compounds (VOCs):
VOCs include a variety of chemicals that have a high vapor pressure and low water solubility. Many VOCs are
human-made chemicals that are used and produced in the manufacture of paints, pharmaceuticals, and
refrigerants. VOCs typically are industrial solvents, such as trichloroethylene; fuel oxygenates, such as methyl
tert-butyl ether (MTBE); or by-products produced by chlorination in water treatment, such as chloroform. VOCs
are often components of petroleum fuels, hydraulic fluids, paint thinners, and dry-cleaning agents. See Section
id-
Source: https://www.epa.gov/indoor-air-guality-iag/what-are-volatile-organic-compounds-vocs
verification:
The process of evaluating the completeness, correctness, and conformance/compliance of a specific data set
against the method, procedural, or contractual requirements. See Section 3.3.
Source: https://www.epa.gov/sites/default/files/2015-06/documents/g5-final.pdf
F-14
-------
&EPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
Recycled/Recyclable Printed on paper that contains a minimum of
50% postconsumer fiber content processed chlorine free
------- |