United States
ti'&r'Kpl Environmental Protection
JF lk Agency
EPA/600/R-21/044
April 2021
Comparative Assessment of the
Impacts of Prescribed Fire Versus
Wildfire (CAIF): A Case Study in
the Western U.S.
April 2021
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
DISCLAIMER
This document is a peer review draft, for review purposes only. This information is distributed
solely for the purpose of predissemination peer review under applicable information quality guidelines. It
has not been formally disseminated by U.S. EPA. It does not represent and should not be construed to
represent any Agency determination or policy. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
ii
DRAFT: Do Not Cite or Quote
-------
FUNDING
This work was supported in part by interagency agreements with the U.S. Department of
Agriculture (#92532401) and Department of the Interior (#92533501).
iii
DRAFT: Do Not Cite or Quote
-------
CONTENTS
LIST OF TABLES viii
LIST OF FIGURES ix
COMPARATIVE ASSESSMENT OF THE IMPACTS OF PRESCRIBED FIRE VERSUS WILDFIRE
(CAIF): A CASE STUDY IN THE WESTERN U.S. xiii
Executive Direction xiii
Scientific and Technical Direction xiii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xiv
Authors xiv
Contributors xviii
Reviewers xviii
Quality Assurance (QA) and Peer Review Support xix
EXECUTIVE SUMMARY ES-1
CHAPTER 1 INTRODUCTION 1-1
1.1 Background 1-1
1.2 Rationale 1-1
1.3 Novel Approach 1-3
1.4 Goals of This Report 1-8
1.5 References 1-9
CHAPTER 2 CONCEPTUAL FRAMEWORK FOR EVALUATING AND COMPARING
DIFFERENT FIRE MANAGEMENT STRATEGIES 2-1
2.1 Introduction 2-1
2.2 Expected Value Framework 2-2
2.3 Components of the Conceptual Framework 2-4
2.3.1 Baseline Wildland Fuels Vegetation and Resource Management Conditions 2-6
2.3.2 Types of Fires 2-6
2.3.3 Fire Management Strategies 2-8
2.3.4 Effects of Fire 2-10
2.3.5 Programs to Mitigate Exposures and Impacts 2-13
2.3.6 Implementing the Conceptual Framework 2-13
2.4 References 2-17
CHAPTER 3 FIRE REGIMES, FIRE EFFECTS, AND A HISTORY OF FUELS AND FIRE
MANAGEMENT IN DRY FORESTS OF THE PONDEROSA PINE REGION 3-1
3.1 Fire Regimes and Ecological Condition of Forests 3-1
3.1.1 Historic Fire Regimes in the Ponderosa Pine Region 3-4
3.1.2 Historic Forest Conditions 3-6
3.1.3 Fire Influences on Forest Structure and Composition 3-7
3.1.4 Ecosystem Resilience/Resistance to Fire 3-7
3.1.5 Changes to Historic Fire Regimes 3-7
3.2 Land Management Approaches to Reducing Fire Risks 3-9
3.2.1 Need for Fuel Treatments 3-9
3.2.2 Land Management Activities Affect Fire Behavior 3-10
iv
DRAFT: Do Not Cite or Quote
-------
CONTENTS (Continued)
3.3 Forest Characteristics for Timber Crater 6 (TC6) and the Rough Fires 3-15
3.3.1 Timber Crater 6 (TC6): Crater Lake National Park/Fremont-Winema National Forest 3-16
3.3.2 Rough Fire: Sierra and Sequoia National Forests and Kings Canyon National Park 3-17
3.4 Conclusions 3-20
3.5 References 3-22
CHAPTER 4 AIR QUALITY MONITORING OF WILDLAND FIRE SMOKE 4-1
4.1 Introduction 4-1
4.2 Objectives of Air Quality Monitoring 4-4
4.2.1 Public Reporting of Air Quality through the Air Quality Index (AQI) 4-4
4.2.2 Analyzing Air Quality Trends 4-6
4.2.3 Informing Fire Management 4-7
4.2.4 Quantifying the Impact of Wildland Fires on Air Quality 4-8
4.3 Ambient Air Quality Monitoring Capabilities 4-9
4.3.1 Overview 4-9
4.3.2 U.S. EPA Routine Regulatory Monitoring Networks 4-9
4.3.3 Temporary/Incident Response Measurements 4-11
4.3.4 Sensors 4-12
4.3.5 Remote Sensing/Satellite Data 4-14
4.4 Ambient Air Quality Monitoring Data Availability and Quality 4-22
4.4.1 Overview 4-22
4.4.2 U.S. EPA Routine Regulatory Data Availability 4-22
4.4.3 Temporary/Incident Response Data Availability 4-23
4.4.4 Sensor Data Availability 4-24
4.4.5 Remote Sensing Data Availability 4-25
4.4.6 Measurement Data Quality 4-25
4.5 Challenges in Ambient Smoke Monitoring 4-26
4.6 Recommendations 4-27
4.7 References 4-29
CHAPTER 5 AIR QUALITY MODELING OF WILDLAND FIRE 5-1
5.1 Background 5-1
5.1.1 Emissions of Wildland Fires 5-1
5.1.2 Using Air Quality Models to Estimate Wildland Fire PM2.5 and Ozone Impacts 5-3
5.1.3 Case Study: TimberCrater6 (TC6)Fire 5-4
5.1.4 Prescribed Fire near Crater Lake National Park 5-6
5.1.5 Case Study: RoughFire 5-8
5.2 Methodology 5-10
5.2.1 Fuels (Fuel Characteristic Classification System [FCCS]) 5-12
5.2.2 Characterizing Surface Fuel Loads for Use in the BlueSky Pipeline 5-12
5.2.3 Fuel Consumption and Fire Emissions (BlueSky Pipeline) 5-15
5.2.4 Pile/Slash Burn Emissions 5-16
5.2.5 Air Quality Modeling System 5-17
5.3 Results—Case Studies 5-18
5.3.1 Timber Crater 6 (TC6) Air Quality Impacts 5-20
5.3.2 Rough Fire Air Quality Impacts 5-28
5.4 Limitations, Implications, and Recommendations 5-38
5.5 References 5-41
CHAPTER 6 WILDLAND FIRE SMOKE EXPOSURE CHARACTERIZATION AND HEALTH
AND ECOLOGICAL IMPACTS 6-1
6.1 Introduction 6-1
v
DRAFT: Do Not Cite or Quote
-------
CONTENTS (Continued)
6.2 Wildfire Smoke Exposure and Health 6-2
6.2.1 Characterization of Wildfire Smoke Exposures 6-3
6.2.2 Health Effects Attributed to Wildfire Smoke Exposure 6-6
6.2.3 Summary 6-14
6.3 Mitigation of Prescribed Fire and Wildfire Smoke Exposure to Reduce Public Health Impacts 6-15
6.3.1 Framework for Estimating the Impact of Actions to Reduce Smoke Exposure 6-16
6.3.2 Individual and Community Actions to Reduce Smoke Exposure 6-18
6.3.3 Estimating the Overall Exposure Reduction to Wildfire Smoke for Individual-Level
Actions 6-25
6.3.4 Uncertainties and Limitations in Estimating Exposure Reduction to Wildland Fire Smoke 6-26
6.4 Ecological Effects Attributed to Wildfire Smoke and Deposition of Pollutants 6-28
6.4.1 Particulate Matter (PM) 6-28
6.4.2 Effects of Ozone (O3) from Fires 6-30
6.4.3 Atmospheric Deposition of Ash 6-31
6.4.4 Uncertainties and Limitations in the Ecological Effects Evidence 6-34
6.5 References 6-35
CHAPTER 7 ECOLOGICAL, WELFARE, AND OTHER DIRECT DAMAGES OF FIRE AND
SMOKE 7-1
7.1 Introduction 7-1
7.2 Wildland Firefighter Exposure to Smoke during Prescribed Fires and Wildfires 7-1
7.2.1 Health Hazards of Exposure to Smoke 7-2
7.2.2 Smoke Exposure at U.S. Prescribed Fires versus Wildfires 7-2
7.2.3 Management Implications 7-4
7.3 Economic Burden of Wildfire 7-4
7.3.1 Economics of Wildfire: Management Implications 7-7
7.3.2 Management Cost Categories 7-9
7.3.3 Wildfire Loss Categories 7-14
7.3.4 Magnitudes, Gaps, and Uncertainty 7-21
7.4 References 7-24
CHAPTER 8 ESTIMATED PUBLIC HEALTH IMPACTS 8-1
8.1 Introduction 8-1
8.2 Benefits Mapping and Analysis Program—Community Edition (BenMAP-CE) Analysis 8-1
8.2.1 Health Impact Function 8-2
8.2.2 Air Quality Modeling 8-3
8.2.3 Effect Coefficients 8-4
8.2.4 Baseline Incidence and Prevalence Data 8-5
8.2.5 Assigning PM2.5 Concentrations to the Population 8-5
8.2.6 Economic Analysis 8-6
8.3 Results from Case Study Fire Analyses 8-6
8.3.1 Main Results 8-7
8.3.2 Sensitivity Analyses 8-10
8.3.3 PM2.5 Exposure Reduction Sensitivity Analysis 8-13
8.4 Summary 8-15
8.5 References 8-17
CHAPTER 9 INTEGRATED SYNTHESIS 9-1
9.1 Introduction 9-1
9.2 Overview of Results 9-2
9.2.1 Timber Crater 6 (TC6) Case Study 9-3
9.2.2 Rough Fire Case Study 9-8
vi
DRAFT: Do Not Cite or Quote
-------
CONTENTS (Continued)
9.3 Limitations in Examining Differences between Prescribed Fire and Wildfire Impacts 9-11
9.3.1 Implementing the Conceptual Framework 9-13
9.3.2 Overarching Limitations 9-14
9.3.3 Identified Data Gaps and Uncertainties 9-17
9.4 Key Insights from Case Study Analyses 9-19
9.5 Future Directions 9-20
9.6 References 9-22
APPENDIX A-l
A.l. Supplemental Information for CHAPTER 1 A-l
A,2. Supplemental Information for CHAPTER 2 A-l
A, 3. Supplemental Information for CHAPTER 3 A-3
A.4. Supplemental Information for CHAPTER 4 A-4
A.4.1. Example State and Local Sponsored Smoke Blogs A-13
A.4.2. U.S. EPA PM2.5 Mass Monitoring A-14
A.4.3. U.S. EPA PM2.5 Speciation Monitoring A-15
A.4.4. U.S. EPA Criteria Gas Monitoring A-17
A, 5. Supplemental Information for CHAPTER 5 A-18
A. 5.1. Supplemental Tables for Chapter 5 A-18
A.5.2. Supplemental Materials for Section 5.2.2: Surface Fuel Loads A-30
A, 6, Supplemental Information for CHAPTER 6 A-59
A.6.1. Supplemental Information for Section 6.2 A-59
A.6.2. Supplemental Information for Section 6.3 A-69
A, 7. Supplemental Information for CHAPTER 7 A- 81
A, 8. Supplemental Information for CHAPTER 8 A- 81
A,9. Supplemental Information for CHAPTER 9 A-82
A, 10. References A-82
vii
DRAFT: Do Not Cite or Quote
-------
LIST OF TABLES
Table 2-1 Expected effects associated with wildland fire: quantified and unqualified for the case
study analyses. 2-14
Table 3-1 Fire regime groups and descriptions. 3-2
Table 4-1 Understanding the U.S. Environmental Protection Agency Air Quality Index (AQI): An
example for PM2 5. 4-5
Table 5-1 Emissions and fuel consumption for three different types of slash/pile burn fuel geometry
assumptions for the Timber Crater 6 (TC6) case study area. 5-16
Table 5-2 Wildfire and prescribed fires modeled as part of the Timber Crater 6 (TC6) and Rough
Fire case studies. 5-19
Table 6-1 Summary of data available for various exposure reduction actions. 6-25
Table 7 -1 The economic burden of wildland fires. 7-6
Table 7-2 Magnitude and uncertainty associated with the economic burden of wildfire at the
national level. 7-21
Table 8-1 Key data inputs for Benefits Mapping and Analysis Program—Community Edition
(BenMAP-CE) used to estimate health impacts for the case studies. 8-2
Table 8-2 Estimated counts of PM2 5 premature deaths and illnesses (95% confidence interval). 8-8
Table 8-3 Estimated counts of ozone (O3) premature deaths and illnesses (95% confidence interval). 8-9
Table 8-4 Estimated value of PM2 5 and ozone-related premature deaths and illnesses (95%
confidence interval; millions of 2015 dollars). 8-10
Table 8-5 Estimated value of wildfire-specific PM25 illnesses (95% Confidence interval; 2015
dollars) from sensitivity analyses. 8-13
Table 8-6 Overall reduction in total health impacts attributed to PM2 5 from wildfire smoke for the
Timber Crater 6 (TC6) Fire case study. 8-14
Table 8-7 Overall reduction in total health impacts attributed to PM2 5 from wildfire smoke for the
Rough Fire case study. 8-15
viii DRAFT: Do Not Cite or Quote
-------
LIST OF FIGURES
Figure 1-1
Figure 1-2
Figure 1-3
Figure 2-1
Figure 3-1
Figure 3-2
Figure 3-3
Figure 3-4
Figure 3-5
Figure 3-6
Figure 3-7
Figure 3-8
Figure 4-1
Figure 4-2
Figure 4-3
Overall approach to comparing fire management strategies in case study analyses. 1-5
Map of fire perimeters of hypothetical scenarios and actual fire for the Timber Crater 6
(TC6) Fire case study. 1-6
Map of fire perimeters for the Rough Fire case study. 1-7
Conceptual framework for evaluating and comparing fire management strategies. 2-5
Fire Regime Groups characterizing the presumed historical fire regimes within
landscapes based on interactions between vegetation dynamics, fire spread, fire effects,
and spatial context. 3-3
The ponderosa pine region as defined by the distribution of Pinus ponderosa in Oregon,
Washington, and northern California (94,000 km2, 11% of the land area shown) based on
2017 Gradient Nearest Neighbor maps. 3-5
Historic photo showing the open character of old growth ponderosa pine resulting from
high-frequency, low-intensity fire on the Klamath Indian Reservation in south-central
Oregon in the 1930s (left) and present-day ponderosa pine forest 10-15 years after
natural fire Ochoco National Forest, central Oregon (right). 3-6
Comparison of differences between a fire-suppressed and ecologically managed forest. 3-12
Prescribed fire in ponderosa pine, Deschutes National Forest. 3-13
Timber Crater 6 (TC6) Fire, Crater Lake National Park, and adjacent Fremont-Winema
National Forest. 3-17
Rough Fire: Sierra and Sequoia National Forests and Kings Canyon National Park. 3-19
Tree species maps for the area of the Rough Fire. 3-20
AirNow Fire and Smoke site displaying the October 7, 2020 layers of PM25 monitors
across central California for (a) regulatory Federal Equivalent Method (FEM) instruments
[circles], (b) with additional California Air Resources Board (CARB) and U.S. Forest
Service (USFS) temporary monitors [triangles], and (c) with the addition of Purple Air
sensors [squares]. 4-2
Tracking of Air Quality Index (AQI) in Oregon during the 2020 wildfire season (a) and
the cumulative annual Oregon population exposure to PM2 5 (b) showing the clear impact
of wildland fire events. 4-7
Image of surface Air Quality Index (AQI) for PM2 5 from U.S. EPA AirNow over plotted
with Air Quality Index (AQI) for PM2 5 derived from National Oceanic and Atmosphere
Administration (NOAA) aerosol optical depth from Visible Infrared Imaging Radiometer
Suite (VIIRS) instruments (Soumi-NPP and NOAA-20 satellites) for September 15,
2020. 4-19
ix DRAFT: Do Not Cite or Quote
-------
LIST OF FIGURE (Continued)
Figure 4-4 Image of western U.S. wildfire smoke transported to the Northeast U.S. as captured in the
Visual Infrared Imaging Radiometer Suite (VIIRS) True Color Image over plotted with
VIIRS aerosol optical depth for September 16, 2020. 4-21
Figure 5-1 Daily fire perimeters for the smaller Timber Crater 6 (TC6) hypothetical fire
(Scenario 1). 5-5
Figure 5-2 Daily fire perimeters for the larger Timber Crater 6 (TC6) hypothetical fires
(Scenarios 2a and 2b). 5-6
Figure 5-3 Fire perimeter of the actual Timber Crater 6 (TC6) Fire, multiple wildfires that yielded
positive resource benefits, and multiple prescribed fires. 5-7
Figure 5-4 Schematic showing the 2015 Rough Fire, 2010 Sheep Complex Fire, and Boulder Creek
Unit 1 prescribed burn unit in relation to large urban areas in central California. 5-9
Figure 5-5 Modeling framework used to characterize wildland fire emissions and air quality impacts
for case study analyses. 5-11
Figure 5-6 Fuel Characteristic Classification System (FCCS) fuel type data and Visualizing
Ecosystem Land Management Assessments (VELMA) fuel load data were merged to
produce fuelbed inputs for the BlueSky Pipeline. 5-14
Figure 5-7 Episode average PM2 5 and maximum daily 8-hour average (MDA8) ozone (O3) predicted
by the modeling system and measured by routine surface monitors for the 2018 modeling
period used for the Timber Crater 6 (TC6) scenarios. 5-20
Figure 5-8 Episode average PM2 5 impacts and aggregate population exposure from the actual
Timber Crater 6 (TC6) Fire and the difference between the actual fire and largest (2b)
and smaller (1) hypothetical scenarios. 5-21
Figure 5-9 Episode average maximum daily 8-hour average (MDA8) ozone (O3) impacts and
aggregate population exposure from the actual Timber Crater 6 (TC6) Fire and the
difference between the actual fire and largest (2b) and smaller (1) hypothetical scenarios. 5-22
Figure 5-10 Daily average PM2 5 ambient (top row) impacts and estimates of aggregate population
exposure (bottom row) from the Timber Crater 6 (TC6) case study scenarios (left) and
prescribed fire scenarios (right). 5-24
Figure 5-11 Maximum daily 8-hour average (MDA8) ozone (O3) ambient (top row) impacts and
estimates of aggregate population exposure (bottom row) from the Timber Crater 6 (TC6)
case study scenarios (left) and prescribed fire scenarios (right). 5-25
Figure 5-12 Daily average PM2 5 (left) and maximum daily 8-hour average (MDA8) ozone (O3)
(right) ambient (top row) impacts and estimates of aggregate population exposure
(bottom row) from hypothetical pile burns from the Timber Crater 6 (TC6) case study
area. 5-27
Figure 5-13 Episode average PM2 5 predicted by the modeling system (from all actual sources) and
measured by routine surface monitors (left) and fire specific modeled impacts (right) for
the actual Rough Fire (top), actual Sheep Complex Fire (middle), and hypothetical
Boulder Creek Unit 1 prescribed fire (bottom). 5-29
Figure 5-14 Episode average maximum daily 8-hour average (MDA8) ozone (O3) predicted by the
modeling system (from all sources) and measured by routine surface monitors (left) and
x
DRAFT: Do Not Cite or Quote
-------
LIST OF FIGURE (Continued)
fire-specific modeled impacts (right) for the 2015 modeling period used for the Rough
Fire scenarios (top), 2010 modeling period for the actual Sheep Complex Fire (middle),
and 2014 modeling period for the hypothetical Boulder Creek Unit 1 prescribed fire
(bottom). 5-30
Figure 5-15 Episode average PM2 5 impacts from the actual Rough Fire and the difference between
the actual scenario and smaller (Scenario 1) and larger (Scenario 2) hypothetical
scenarios. 5-32
Figure 5-16 Episode average maximum daily 8-hour average (MDA8) ozone (O3) impacts from the
actual Rough Fire and the difference between the actual scenario and smaller (Scenario 1)
and larger (Scenario 2) hypothetical scenarios. 5-33
Figure 5-17 Daily average PM25 ambient (left) and maximum daily 8-hour average (MDA8) ozone
(O3) (right) impacts and aggregate population exposure (bottom row) from the Rough
Fire scenarios. 5-34
Figure 5-18 Daily average PM2 5 observations and model predictions at monitors in the Central Valley
of California for August and September 2015. 5-36
Figure 5-19 Daily average PM25 ambient (left) and maximum daily 8-hour average (MDA8) ozone
(O3) (right) impacts and aggregate population exposure (bottom row) from the
hypothetical Boulder Creek Unit 1 prescribed fire. 5-37
Figure 5-20 Daily average ambient PM2 5 (left) and maximum daily 8-hour average (MDA8) ozone
(O3) (right) concentrations and estimates of aggregate population exposure (bottom row)
from the 2010 Sheep Complex Fire. 5-38
Figure 6-1 Odds ratios and relative risks from U.S.-based epidemiologic studies examining the
relationship between short-term wildfire smoke exposure and combinations of
respiratory-related diseases and asthma emergency department visits and hospital
admissions. 6-8
Figure 6-2 Odds ratios and relative risks from U.S.-based epidemiologic studies examining the
relationship between short-term wildfire smoke exposure and respiratory-related
emergency department visits and hospital admissions. 6-10
Figure 6-3 Odds ratios and relative risks from U.S.-based Epidemiologic studies examining the
relationship between short-term wildfire smoke exposure and cardiovascular-related
emergency department (ED) visits and hospital admissions. 6-12
Figure 6-4 Considerations for estimating potential reduction in wildfire smoke exposure due to
actions and interventions. 6-17
Figure 6-5 Summary of individual-level wildfire smoke exposure reduction actions and
effectiveness. 6-19
Figure 6-6 Percentage of the population taking a specific exposure reduction action as a function of
the characteristics of the surveyed population. 6-22
Figure 6-7 Comparison of estimated percent overall PM2 5 exposure reduction by action. 6-26
Figure 7-1 Billion-dollar wildfire event losses (1980-2020). 7-7
Figure 7-2 Illustrative example of the Cost plus Loss (C+L) Model of wildfire management. 7-9
xi DRAFT: Do Not Cite or Quote
-------
LIST OF FIGURE (Continued)
Figure 8-1 Estimated health impacts from sensitivity analyses using health impact functions based
on ambient PM2 5 exposures versus wildfire-specific PM2 5 exposures for the Timber
Crater 6 (TC6) Fire case study. 8-11
Figure 8-2 Estimated health impacts from sensitivity analyses using health impact functions based
on ambient PM2 5 exposures versus wildfire-specific PM2 5 exposures for the Rough Fire
case study. 8-12
Figure 9-1 Surface fuel loading in untreated forests in the Timber Crater 6 (TC6) Fire study area in
Crater Lake National Park. 9-5
Figure 9-2 Decadal-scale representation of the age structure of lodegpole pine (LP) and ponderosa
pine (PP) aggregated for a 200-ha study area within the Timber Crater 6 (TC6) Fire
perimeter. 9-7
Figure 9-3 Decline in fire frequency in mixed conifer forest (from nearby Sequoia and Kings
Canyon National Parks) starting around 1860. 9-10
Figure 9-4 Conceptual framework for evaluating and comparing fire management strategies. 9-12
Figure 9-5 Conceptual diagram presented by Hunter and Robles (2020) for assessing the impacts of
prescribed fire compared to wildfire. 9-15
Figure 9-6 Acres burned by wildfire (red) and prescribed fire (green) in the U.S. in 2017. 9-17
xii
DRAFT: Do Not Cite or Quote
-------
COMPARATIVE ASSESSMENT OF THE IMPACTS
OF PRESCRIBED FIRE VERSUS WILDFIRE (CAIF):
A CASE STUDY IN THE WESTERN U.S.
Executive Direction
Dr. Wayne E. Cascio (Executive Lead)—Director, Center for Public Health and
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Bryan Hubbell (Executive Lead)—National Program Director, Air and Energy National
Research Program, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Peter Lahm (Executive Lead)—Fire & Aviation Management, U.S. Forest Service, U.S.
Department of Agriculture, Washington, DC
Dr. Peter Teensma (Executive Lead)—Office ofWildland Fire, U.S. Department of Interior,
Washington, DC
Scientific and Technical Direction
Mr. Jason Sacks (Assessment Lead)—Center for Public Health and Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Christina Baghdikian (Deputy Assessment Lead)—Center for Public Health and
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Kirk Baker—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Long—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Lars Perlmutt—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Kelly Widener—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
xiii
DRAFT: Do Not Cite or Quote
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
Authors
Chapter 1
Ms. Christina Baghdikian—Center for Public Health and Environmental Assessment, Office
of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Bryan Hubbell—National Program Director, Air and Energy National Research
Program, Office of Research and Development, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Vivian Phan—Oak Ridge Associated Universities, Center for Public Health and
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Corvallis, OR
Mr. Jason Sacks—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Chapter 2
Dr. Bryan Hubbell—National Program Director, Air and Energy National Research
Program, Office of Research and Development, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Peter Lahm—National Headquarters Fire & Aviation Management, U.S. Forest Service,
U.S. Department of Agriculture, Washington DC
Mr. James Menakis— National Headquarters Fire & Aviation Management, U.S. Forest
Service, U.S. Department of Agriculture, Washington, DC (Detached)
Mr. David Mueller—Department of Interior, Bureau of Land Management, Boise, ID
Dr. Peter Teensma—Office of Wildland Fire, U.S. Department of Interior, Washington, DC
Mr. Jason Sacks—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Chapter 3
Dr. Peter Beedlow—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Jonathan Halama—Spatial Synergy Solutions, LLC, Independence, OR
Dr. James Markwiese—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Robert McKane—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Mr. David Mueller—Department of Interior, Bureau of Land Management, Boise, ID
xiv
DRAFT: Do Not Cite or Quote
-------
Ms. Vivian Phan—Oak Ridge Associated Universities, Center for Public Health and
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Corvallis, OR
Dr. Peter Teensma—Office of Wildland Fire, U.S. Department of Interior, Washington, DC
Chapter 4
Mr. Tim Hanley—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Ali Kamal—Office of Transportation and Air Quality, Office of Air and Radiation,
U.S. Environmental Protection Agency, Ann Arbor, MI
Mr. Peter Lahm—Fire & Aviation Management, U.S. Forest Service, U.S. Department of
Agriculture, Washington, DC
Dr. Matt Landis—Center for Environmental Measurement and Modeling, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Sim Larkin—U.S. Forest Service, U.S. Department of Agriculture, Seattle, WA
Dr. Russell Long—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. James Szykman—Center for Environmental Solutions and Emergency Response, Office
of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Shawn Urbanski—Rocky Mountain Research Station, U.S. Forest Service, U.S.
Department of Agriculture, Missoula, MT
Chapter 5
Dr. Kirk Baker—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Bradley Barnhart—Independent Contractor, Corvallis, OR
Mr. James Beidler—General Dynamics Information Technology, Durham, NC
Mr. Anthony Bova—U.S. Forest Service, U.S. Department of Agriculture, Seattle, WA
Dr. Allen Brookes—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Mr. Kevin Djang—General Dynamics Information Technology, Corvallis, OR
Dr. Calvin Farris—National Park Service, Department of Interior, Regions 8, 9, 10, and 12
Klamath Falls, OR
Dr. Jonathan Halama—Spatial Synergy Solutions, LLC, Independence, OR
Dr. Amara Holder—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Shannon Koplitz—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
xv
DRAFT: Do Not Cite or Quote
-------
Dr. Robert McKane—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Thien Khoi Nguyen—Air and Radiation Division, Region 9, U.S. Environmental
Protection Agency, San Francisco, CA
Mr. Richard Pasquale—Fuels & Fire, U.S. Forest Service, U.S. Department of Agriculture,
Lakeview, OR
Ms. Vivian Phan—Oak Ridge Associated Universities, Center for Public Health and
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Corvallis, OR
Dr. George Pouliot—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Susan Prichard—School of Environmental and Forest Sciences, University of
Washington, Seattle, WA
Ms. Dana Skelly—Regional Fuels Program Manager, Pacific Northwest Region, U.S. Forest
Service, U.S. Department of Agriculture, Portland, OR
Dr. Leland Tarnay—Region 5 Remote Sensing Laboratory, U.S. Forest Service, U.S.
Department of Agriculture, McClellan, CA
Mr. Jeffrey Vukovich—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Chapter 6
Dr. Christopher Andersen—Center for Public Health and Environmental Assessment, Office
of Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Peter Beedlow—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Janet Burke—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Jeffrey Herrick—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Amara Holder—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Stephen LeDuc—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. James Markwiese—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Ana Rappold—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
xvi
DRAFT: Do Not Cite or Quote
-------
Mr. Jason Sacks—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Chapter 7
Dr. David Butry—Applied Economics Office, Engineering Laboratory, National Institute of
Standards and Technology, Gaithersburg, MD
Dr. Roger Ottmar—Pacific Wildland Fire Sciences Laboratory, U.S. Forest Service, U.S.
Department of Agriculture, Seattle, WA
Dr. Jeffrey Prestemon—Southern Research Station, U.S. Forest Service, U.S. Department of
Agriculture, Asheville, NC
Mr. Timothy Reinhardt—Wood Environment & Infrastructure Solutions, Inc., Seattle, WA
Chapter 8
Dr. Janet Burke—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Mr. Neal Fann—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Amara Holder—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Ali Kamal—Office of Transportation and Air Quality, Office of Air and Radiation,
U.S. Environmental Protection Agency, Ann Arbor, MI
Mr. Brian Keaveny—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Lisa Thompson—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Chapter 9
Dr. Kirk Baker—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Calvin Farris—National Park Service, Department of Interior, Regions 8, 9, 10, and 12
Klamath Falls, OR
Dr. Bryan Hubbell—National Program Director, Air and Energy National Research
Program, Office of Research and Development, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jason Sacks—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Leland Tarnay—Region 5 Remote Sensing Laboratory, U.S. Forest Service, U.S.
Department of Agriculture, McClellan, CA
xvii
DRAFT: Do Not Cite or Quote
-------
Contributors
Mr. Aranya Ahmed—Oak Ridge Institute for Science and Education, Center for
Environmental Measurement and Modeling, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Randy Comeleo—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Becky Estes—Central Sierra Province, U.S. Forest Service, U.S. Department of
Agriculture, Placerville, CA
Ms. Julie Fieldsteel—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Mr. Floyd Gregor—U.S. Forest Service, U.S. Department of Agriculture, Klamath Falls, OR
Mr. James Klungness-Mshoi—U.S. Forest Service, U.S. Department of Agriculture,
Klamath Falls, OR
Mr. Eric Knerr—U.S. Forest Service, U.S. Department of Agriculture, Klamath Falls, OR
Dr. Marc Meyer—Pacific Northwest Region, U.S. Forest Service, U.S. Department of
Agriculture, Bishop, CA
Ms. Danielle Moore—Senior Environmental Employment Program, Center for Public Health
and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Kevin Park—Oak Ridge Associated Universities, Inc., Center for Public Health and
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Jenna Strawbridge—Oak Ridge Associated Universities, Inc., Center for Public Health
and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Reviewers
Mr. Lance Avey—U.S. EPA Region 7, U.S. Environmental Protection Agency, Kansas City,
MO
Ms. Katie Boaggio—Oak Ridge Institute for Science and Education, Center for Public
Health and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Alison Clune—Office of Radiation and Indoor Air, Office of Air and Radiation, U.S.
Environmental Protection Agency, Washington, DC
Dr. Stephanie Deflorio-Barker—Center for Public Health and Environmental Assessment,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Robert Elleman—U.S. EPA Region 10, U.S. Environmental Protection Agency, Seattle,
WA
Mr. Rick Gillam—U.S. EPA Region 4, U.S. Environmental Protection Agency, Atlanta, GA
xviii
DRAFT: Do Not Cite or Quote
-------
Mr. Michael McGown—Idaho Operations Office, U.S. EPA Region 10, U.S. Environmental
Protection Agency, Boise, ID
Dr. Anna Mebust—U.S. EPA Region 9, U.S. Environmental Protection Agency, San
Francisco, CA
Dr. Thomas Pierce—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Gail Robarge—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Susan Stone—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. John Vandenberg—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Alan Vette—Center for Environmental Measurement and Modeling, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Christopher Weaver—Center for Public Health and Environmental Assessment, Office
of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Quality Assurance (QA) and Peer Review Support
Ms. Christine Alvarez—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Andrea Bartolotti—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Anna Champlin—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Washington, DC
Ms. Rebecca Daniels—Center for Computational Toxicology, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Cheryl Itkin—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. David Lehmann—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Jenia McBrian—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Elizabeth Sams—Immediate Office of the Assistant Administrator, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
xix
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
EXECUTIVE SUMMARY
In January 2020, the Wildland Fire Leadership Council (WFLC), an intergovernmental
committee formed to support the implementation and coordination of Federal Fire Management Policy
and chaired by senior leadership in the U.S. Department of Agriculture (USDA) and Department of the
Interior (DOI), requested that the U.S. Environmental Protection Agency (U.S. EPA) lead an assessment
that would characterize and compare the impacts of wildland fires under different fire management
strategies, including prescribed fire. In this role, the U.S. EPA, in collaboration with the U.S. Forest
Service (USFS), DOI, and the National Institute of Standards and Technology (NIST) conducted an
assessment, focusing on the smoke impacts of prescribed fire and wildfire, while also recognizing the
direct fire impacts of each, as a means to help inform future land management and fire management
strategies.
Comparative Assessment of the Impacts of Prescribed Fire Versus Wildfire (CAIF): A Case Study
in the Western U.S. consists of a qualitative and, as feasible, quantitative assessment of the air quality and
health impacts of wildland fire (i.e., prescribed fire and wildfire), along with an integrated discussion of
topics that are important to consider in the context of comparing different fire management strategies
including:
• A conceptual framework and model for evaluating different fire management strategies
• Background information on different fire regimes, including land management practices, and the
associated impacts (both beneficial and detrimental) due to fire
• A discussion of air quality monitoring as it pertains to prescribed fire and wildfire including
current monitoring capabilities and resources available to obtain information on air quality
measurements and pollutant concentrations
• Characterization of epidemiologic evidence of health effects, specifically within the U.S.,
attributed to wildfire smoke exposures along with quantitative information on public health
measures that could be instituted to reduce individual and population-level exposures to wildfire
smoke
• Characterization of ecological impacts attributed to wildfire smoke
• A broad overview of the direct fire impacts of wildfire with a focus on firefighter health and
safety and societal impacts (i.e., economic and welfare impacts)
The qualitative discussions presented above set the stage for the main component of the
assessment, which is a novel modeling analysis focusing on case study fires in the western U.S. The first
case study analysis focuses on a small fire (-3,000 acres) that occurred in Oregon, the Timber Crater 6
(TC6) Fire, from July 21-26, 2018 while the second case study focuses on the Rough Fire which occurred
in California from July 31-October 1, 2015 and burned significantly more acres than the TC6 Fire
(-150,000 acres). Both case study fires were selected because they represented fires managed by USFS
and DOI, and occurred on federal land. The TC6 Fire was selected because there is extensive data on land
management, fuel treatment, prescribed fire, and wildfire activity; whereas, the Rough Fire was selected
ES-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
because it represented a larger fire to allow for a scaling up of the modeling approach developed for the
TC6 Fire. For both case studies, hypothetical scenarios assuming different fire management strategies that
could have resulted in smaller or larger actual fires were developed based on expert judgment. These
hypothetical scenarios allowed for a comparison of the air quality, specifically fine particulate matter
(PM2.5; particulate matter with a nominal mean aerodynamic diameter <2.5 (.un) and ozone, and
associated health impacts with the actual case study fires, as well as prescribed fires in each location,
using U.S. EPA's Environmental Benefits Mapping and Analysis Program—Community Edition
(BenMAP-CE).
The case study analyses presented within this assessment demonstrate the importance of having
refined information on prescribed fire activity to support air quality modeling of wildland fires. Within
the area of each case study, air quality modeling indicates that the overall air quality impacts of wildland
fires stem primarily from PM2 5. Wildfires, such as the TC6 Fire, that occur in more remote locations and
not near large population centers result in relatively small air quality and health impacts compared to
larger fires, such as the Rough Fire. The estimated societal economic value of damages of illnesses and
deaths attributed to smoke from each actual fire were:
• TC6 Fire: $18 million (M; 95% confidence intervals [CI]: $2 M to $47 M)
• Rough Fire: $3,000 M (95% CI: $260 M to $7,900 M)
The larger size of the Rough Fire and its closer proximity to population centers provided for a
more meaningful comparison of the air quality and health impacts of different fire management strategies.
Initial evidence indicates that a smaller wildfire adjacent to the Rough Fire that yielded positive resource
benefits did not substantially reduce the overall fire perimeter of the Rough Fire, and thus minimally
reduced the public health impacts. Addition of a prescribed fire targeted in a specific location to reduce
fire spread, in combination with a wildfire that yielded resource benefits, could have dramatically reduced
the overall size of the Rough Fire, resulting in an approximate 40% reduction in excess respiratory- and
cardiovascular-related emergency department visits and hospital admissions, and premature deaths. The
hypothetical scenarios for both case studies demonstrate that prescribed fires targeted for specific
locations can have an effect on reducing the overall size of a wildfire. Although prescribed fires are timed
for days with specific meteorological conditions to reduce population exposures to smoke, analyses show
that air quality and public health impacts, while small, are still observed. The estimated societal economic
value of damages of illnesses and deaths attributed to smoke from prescribed fires in each case study
were:
• TC6 Fire Case Study: $4 M (95% CI: $0 to $9 M)
• Rough Fire Case Study: $60 M (95% CI: $5 M to $160 M)
Lastly, although not extensively examined within this assessment, preliminary analyses
demonstrate that campaigns promoting actions and interventions to reduce or mitigate exposure to
ES-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
wildfire smoke can result in public health benefits, with potential reductions in population PM2.5 exposures ranging from 15
to 30%.
It is important to recognize that the results of this assessment are limited to the geographic
locations of the case study fires that have unique land management practices and resulting fire behavior
that is specific to the ecosystems of each. In addition, although the results of this assessment demonstrate
differences in the air quality and health impacts attributed to different fire management strategies, this
analysis was unable to take into consideration key relationships between prescribed fire and wildfire that
should be considered in future analyses. The analyses conducted within this assessment also treat
prescribed fire activity as occurring at one point in time and does not take into consideration the temporal
and spatial patterns of likely fire management strategies that include prescribed fire. Therefore, analyses
do not consider how prescribed fires intersect with wildfire activity, including the probability of a wildfire
occurring within the spatial domain of prescribed fires. As a result, the comparison of costs and benefits
from smoke impacts between prescribed fires and hypothetical scenarios presented within this assessment
is based on case studies where a wildfire occurred and does not take into consideration how the
relationship between costs and benefits could differ in instances where wildfires have not yet occurred.
Overall, this assessment demonstrates the positive impact that interagency collaborations can
have on complex issues at the intersection of land management and environmental public health, such as
wildland fire. This initial assessment lays the foundation for future collaborative research and analyses by
the partnering agencies to inform future land management and fire management strategies with the goal of
reducing the air quality and health impacts attributed to wildland fire smoke.
ES-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
CHAPTER 1
INTRODUCTION
1.1 Background
The Wildland Fire Leadership Council (WFLC) was established in 2002 by "the Secretaries of
Agriculture and the Interior to provide an intergovernmental committee [consisting of Federal, state,
tribal, county, and municipal government officials] to support the implementation and coordination of
Federal Fire Management Policy" (F&R. 2020b). The U.S. Department of Agriculture and the
Department of the Interior (DOI) are official members and the cochairs of WFLC. One of the aims of
WFLC is to improve communication and coordination with the public, specifically as it pertains to the
understanding of the benefits and tradeoffs of prescribed fire versus wildfire.
At the request of WFLC, in January 2020, the U.S. Environmental Protection Agency (U.S. EPA)
was asked to lead an assessment that would characterize and compare the impacts1 of different fire
management strategies, including prescribed fire. In this role, U.S. EPA would lead the development of
Comparative Assessment of the Impacts of Prescribed Fire Versus Wildfire (CAIF): A Case Study in the
Western U.S. in coordination with the U.S. Forest Service (USFS) and DOI, and with contributions from
the National Institute of Standards and Technology (NIST). This report would provide a better
understanding of the health and environmental impacts of wildland fire (i.e., prescribed fire and wildfire),
specifically pertaining to smoke. The interagency approach being used to conduct this assessment is
critical as USFS and DOI are experts in understanding various aspects of fire (e.g., fire management, fire
planning, fire effects and ecology, incident response), NIST is an expert in the direct and indirect
damages attributed to fire, and U.S. EPA provides expertise in understanding the public health and
environmental impacts of fire, especially smoke. This collaborative effort has allowed for the leveraging
of areas of expertise that are essential to characterizing complicated system-level impacts across the
varying fire management strategies, and established the interagency linkages needed for future research
activities.
1.2 Rationale
Fire has been used as a land management tool to return nutrients to the soil and remove detritus
and excess fuels to reduce wildfire risk and effects, and to manage wildlife habitats and watersheds. Prior
to modern land management, fire had been used for these same purposes and a myriad of other purposes
1 Within this assessment, the term "impacts" refers to the main quantitative results, which includes the estimated air
pollutant concentrations from the air quality modeling and the number of health events and associated economic
values calculated using U.S. EPA's Environmental Benefits Mapping and Analysis Program—Community Edition
(BenMAP-CE). The term "effects" is used to denote the other positive and negative consequences of wildland fire.
1-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
by Native Americans for millennia (Agee. 1993; Lewis. 1985. 1973). Over time our relationship with
wildland fire, and the smoke that comes from these fires has become more complicated. A confluence of
events have all contributed to increasing the likelihood of wildfire ignitions, including but not limited to,
a history of fire suppression that has left a backlog of fuel; a changing climate with warmer temperatures;
and humans moving at increasing rates into the line, area, or zone where structures and other human
development meet or intermingle with undeveloped wildland or vegetative fuels, referred to as the
wildland-urban interface [WUI; F&R (2020a) 1.
Over the past 30 years, on average approximately 5 million acres of wildlands in the U.S. have
burned annually, with over 9 million acres burned in 2020 (Hoover and Hanson. 2021; NIFC. 2018).
Although the number of fires has not changed significantly over this period, the size and intensity of the
fires have increased as a result of higher temperatures, drought, earlier snowmelt due to climate change,
and historically high fuel loading (e.g., undergrowth, tree density) Landis et al. (2017).
Although wildfire can be beneficial, it can also detrimentally impact ecosystems, damage animal
habitats, decrease water quality and quantity, and in some instances create conditions leading to increased
overland water flow and flooding. Additionally, with the rapid expansion of the WUI, wildfires are
increasingly encroaching on American communities, posing threats to lives, critical infrastructure, and
property (Lewis et al.. 2018). The direct effects of fire itself are compounded by the equally significant
effects of the smoke generated from fires, which can travel transcontinental distances and has been shown
to have significant adverse effects on public health (U.S. EPA. 2019b). As the risk that wildfire poses to
property and health has increased, especially when a wildfire is severe and catastrophic, the need to
address this growing risk has also increased. At the same time, there is a need to recognize and maintain
the ecological benefits of fire, which has always been a part of the natural landscape.
Various fire management strategies have been employed over time with the overall goal of
reducing the potential for negative effects of wildfire, such as the overall size of a wildfire and the direct
fire effects. These actions, which include prescribed fire and pile burns from thinning activities, have
associated risks, specifically to air quality and corresponding health and environmental effects. Prescribed
fire is perceived as lower risk compared to wildfire because the timing and area to be burned are managed
to limit smoke impacts (i.e., dispersed both spatially and temporally). Prescribed fires are conducted when
meteorological conditions are favorable, smoke production (fuel consumption) is less, atmospheric
conditions support adequate smoke dispersion, and wind patterns allow smoke to move away from
sensitive areas (e.g., populated areas, hospitals, schools, roadways). While prescribed fire is considered
low risk, it is important to note that there is a risk continuum for wildfire that can change daily based on
fire behavior resulting in a dynamic set of management actions. As a result, wildfire management can
shift between full suppression efforts and, if conditions allow (e.g., wet fuel, anticipated precipitation),
management that may achieve resource benefits. To date, limited information exists that allows for a
direct, systematic, and comprehensive comparison of the air quality and associated health impacts of
smoke from prescribed fire and wildfire. Together, prescribed fires and wildfire are how fire plays a role
1-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
in the natural ecosystem. To ensure the effective use of prescribed fires to reduce the risk of catastrophic
wildfire, decision makers need information on the air quality impacts associated with fire management
strategies that include prescribed fires compared to strategies that do not.
Numerous research activities have focused on examining the nexus between fire, smoke, and
ecological and health impacts. These activities have focused on this complex issue by examining how
various conditions (e.g., fuel type, temperature, moisture) influence the subsequent emissions from a fire,
how these emissions move over various geographic scales and topographies, the toxicologic and
ecotoxicologic effects from smoke exposure, and population-level health impacts of smoke exposure.
Recent studies have also evaluated actions and interventions that can be instituted to reduce the public
health impacts during smoke episodes by melding together social science, behavioral science, and health
risk communication. While all these activities have led to significant advancements in the science, the
overall air quality impacts of different fire management strategies, which consist of different land
management practices, including prescribed fire, are not well characterized. As a result, this complicates
the decision-making process in determining the appropriate fire management strategy and land
management action to implement at governmental levels ranging from local to federal.
1.3 Novel Approach
The CAIF Report represents a unique opportunity to bring together experts spanning multiple
disciplines related to fire science (e.g., air quality, monitoring, modeling, health effects, ecological
effects) to conduct an integrated interagency assessment. The focus of this report consists of a novel
modeling approach to estimate the air quality impacts, specifically of fine particulate matter (i.e., PM2 5
[particulate matter with a nominal mean aerodynamic diameter < 2.5 (.un |) and ozone, in response to
different fire management strategies, and the associated health and economic impacts. To conduct such an
analysis, this report will focus on two case study fires, both of which occurred in the Western U.S.:
(1) Timber Crater 6 (TC6) Fire that occurred from July 21-26, 2018 in Oregon; and (2) Rough Fire that
occurred from July 31-October 1, 2015, in California. These fires were selected, in part, because they
represented interagency fires managed by both USFS and DOI, and both had data available, to varying
degree, on previous land management practices. Due to the difference in the scale of these two fires, the
TC6 Fire burning approximately 3,000 acres and the Rough Fire burning approximately 150,000 acres,
and the different land management and fire management strategies employed in both locations there will
be slight differences in the resolution of the analyses and the analytical approaches between the fires.
The modeling component of the analysis, which is the main focus of this report, will estimate
PM2 5 and ozone concentrations for the actual fire and compare those air quality impacts to hypothetical
scenarios based on different fire management strategies resulting in smaller or larger fires for each of the
case studies, as depicted in Figure 1-1. In addition to the hypothetical smaller and larger fires, analyses
also examine prescribed fire activity, and in the case of the Rough Fire, the perimeter included the
1-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
footprint of a recent wildfire that burned at lower intensity and yielded positive resource benefits. For
both case studies, the prescribed fire analyses do not account for the episodic nature of prescribed fires
that are conducted over years to decades to keep fuel loads at a level needed for fire suppression
opportunities. For the TC6 Fire case study this resulted in the multiple prescribed fires that occurred over
many years in the vicinity of the TC6 Fire to be modeled as individual events within the same month
when prescribed fire activity was known to have occurred. This is in contrast to the Rough Fire case study
where the focus was on the modeling of a single prescribed fire event that was planned but did not occur.
Prescribed fire activity was treated this way within this assessment for numerous reasons including
current limitations in the ability to account for the timing of retrospective prescribed fire activity and
sparseness of available data. Further, the prescribed fires examined within these case studies are not
intended to account for the entirety of a spatial area needed to prevent the spread of a larger wildfire in
both areas.
For both the TC6 and Rough Fire analyses, to facilitate comparison of impacts across the
different fires being examined within the case study areas, the region-wide air quality impacts (i.e., PM2 5
and ozone) will be compared to a baseline of ambient air pollution with no case study area fire. This
approach allows for an estimation of the burden associated with each of the case study fires and a direct
comparison of the health impacts and associated economic values, across each fire and hypothetical
scenario using U.S. EPA's Environmental Benefits Mapping and Analysis Program—Community Edition
[BenMAP-CE; U.S. EPA (2019aYI.
For the TC6 Fire the hypothetical scenarios developed consist of: (1) a smaller hypothetical TC6
Fire in a heavily managed area (e.g., most prescribed fire activity), which would equate to a wildfire with
less fuel, a smaller fire perimeter, and less daily emissions; (2a) a larger hypothetical TC6 Fire, but not
the "worst-case" scenario, due to no land management which would equate to a wildfire with more fuel, a
larger fire perimeter, and more daily emissions; and (2b) a much larger, hypothetical "worst-case"
scenario TC6 Fire with no land management (i.e., no prescribed fire) which would equate to a wildfire
with the most fuel, largest fire perimeter, and largest daily emissions (Figure 1-2). In addition to each of
these scenarios, analyses will also include an examination of only the prescribed fires that occurred
around the TC6 Fire, for a comparison of air quality and health impacts between prescribed fires and the
actual wildfire. These prescribed fires were selected based on actual historical prescribed fire activity in
this area as a preliminary comparison point to the TC6 Fire and hypothetical scenarios.
1-4
DRAFT: Do Not Cite or Quote
-------
Baseline
Scenario
Hypothetical
Larger Fire
Ambient Air Pollution - No Case Study Area Fire Activity
Note; Black outline = actual fire perimeter, green outline = hypothetical smaller fire perimeter; dotted purple outline = hypothetical
larger fire perimeter.
Figure 1-1 Overall approach to comparing fire management strategies in
case study analyses.
1-5
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Elevation (meters)
2572
1950
1770
1400
~ Actual TC6
Fire Perimeter
Scenario 1
(Small fire)
Scenario 2a
(Large fire)
Scenario 2b
(Largest fire)
Diamond Lake
Antelope
Desert
Sources: Esn. Airbus DS. USGS. NGA, NASA CGIAR. N Robinson, NCEAS, NLS. OS, NMA, Geodatastyreisen.
Rijkswaterstaat, GSA> Geoland.. f-EMA, Intermap and the.GIS user community.. Sources; tsri. HERE, Garmin.
' fiFAO, NOAA, USGS, >6 OppnStree'Map contributors, anti the GIS User Community, and USPS. |
S _Cartogiaphy by Vivian Phan
Mollys
Ridge
Diamond Lake
Junction
Figure 1-2 Map of fire perimeters of hypothetical scenarios and actual fire for
the Timber Crater 6 (TC6) Fire case study.
The Rough Fire was selected for the second case study because its larger size and location
provides an opportunity to assess impacts on a larger downwind population and evaluate differences in
both air quality and health impacts for different hypothetical fire management strategies versus the TC6
Fire. For the Rough Fire, analyses will encompass the actual fire, which occurred over approximately
2 months, and the impact of multiple hypothetical scenarios, representing different land management
practices, on both the spread of the Rough Fire and corresponding air quality impacts. In comparing air
quality impacts between the actual Rough Fire and the hypothetical scenarios, the entire 2 months that
encompassed the Rough Fire will be modeled with the air quality impacts diverging at the point where the
Rough Fire would have reached the perimeters of two fires considered within this case study, the Boulder
Creek Prescribed Fire and the Sheep Complex Fire. Within the Rough Fire area there was no previous
prescribed fire activity, as a result, this case study models the proposed Boulder Creek Prescribed Fire,
which was a prescribed fire that USFS had planned, but did not occur; and the Sheep Complex Fire,
which is a wildfire that occurred in 2010 due to a lightning strike and as a result of wet fuel conditions
1-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
resulted in resource benefits. Hypothetical Scenario 1, also referred to as the smaller hypothetical Rough
Fire, revolves around examining the combined impact of a prescribed fire and wildfire that resulted in
resource benefits (i.e., reduced fuels) on reducing the spread and air quality impacts of the Rough Fire.
Hypothetical Scenario 2, also referred to as the larger hypothetical Rough Fire, will allow for the fire
perimeter of the Rough Fire to progress into the area of the Sheep Complex Fire as if both the Boulder
Creek Prescribed Fire and Sheep Complex Fire did not occur. In addition to comparing each hypothetical
scenario to the actual Rough Fire, air quality and health impacts will also be compared individually to the
Boulder Creek Prescribed Fire and the Sheep Complex Fire. Figure 1-3, depicts the fire perimeters that
are examined in the Rough Fire case study.
Elevation (meters)
—1 3550
I 1 2015 Rough Fire
| | Perimeter
•vstwsj 2013 Boulder Creek
to£X2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
individuals taking precautionary measures to reduce smoke exposure can vary between wildfire and
prescribed fire events depending on the presence and effectiveness of public health messaging as well as
the amount of lead time available for messaging to inform the public and the public's ability to act on that
messaging. As a result, it is important to also consider the potential public health implications of actions
or interventions that could be employed to reduce population exposure to smoke when evaluating
tradeoffs between wildfire and prescribed fire. Therefore, a crude estimation of the potential public health
benefits that could be realized in each case study analysis was conducted for different actions meant to
reduce or mitigate smoke exposure. For the actual TC6 and Rough Fires, the deployment of Air Resource
Advisors (ARAs) by the USFS, in combination with the respective state and local air quality agencies,
efforts are taken to predict smoke impacts, and warn the public of the hazards of smoke and the benefit of
minimizing exposure. The examination of smoke exposure reduction actions within this assessment does
not reflect a formal analysis of post-fire effectiveness of public health messaging for either the TC6 or
Rough Fires.
Although the comparison of air quality impacts and associated health and economic impacts
between the different fire management strategies represents the main output of the CAIF Report, in order
to put the results in the proper context, the report also captures qualitatively, and in some cases
quantitatively, other factors that can influence a full accounting of the benefits and damages associated
with each fire management strategy. This includes information pertaining to baseline forest conditions, air
quality monitoring of fires, direct fire effects on health, damages due to fire and smoke, and ecosystem
benefits and damages.
1.4 Goals of This Report
The goal of the CAIF Report is to provide an initial quantitative assessment of the air quality and
associated health, and economic impacts attributed to different fire management strategies, including
prescribed fire, through an extensive modeling exercise. This quantitative assessment will be
supplemented with qualitative discussions to highlight the current state of the science that informs this
assessment, and identify deficiencies that if addressed, can further inform analyses of fire management
strategies. The collective assessment within this report of the benefits and damages associated with both
fire and smoke can contribute to a fuller characterization of the benefits and tradeoffs of different fire
management strategies.
This report represents an initial step in the process of conducting assessments to characterize the
impacts of different fire management strategies to inform both public health actions to reduce population
exposures to wildfire smoke, and future land management decisions. By attempting to more fully account
for the impacts of different fire management strategies, tradeoffs can be assessed to ensure the appropriate
land management actions are taken to maintain forest health and minimize the public health impacts
attributed to wildland fire smoke.
1-8
DRAFT: Do Not Cite or Quote
-------
1.5 References
Agee. JK. (1993). Fire ecology of Pacific Northwest forests. Washington, DC: Island Press.
F&R (Forests and Rangelands). (2020a). Glossary of terms: W. Available online at
https://www.forestsandrangelands.gOv/resources/glossary/w.shtml (accessed January 19, 2021).
F&R (Forests and Rangelands). (2020b). Wildland Fire Leadership Council. Available online at
https://www.forestsandrangelands.gov/leadership/
Hoover. K: Hanson. LA. (2021). Wildfire statistics. (CRS In Focus, IF10244). Washington, DC: Library of
Congress, Congressional Research Service. https://fas.org/sgp/crs/misc/IF10244.pdf
Landis. MS: Edgerton. ES: White. EM: Wentworth. GR: Sullivan. AP: Dillner. AM. (2017). The impact of the
2016 Fort McMurray Horse River Wildfire on ambient air pollution levels in the Athabasca Oil Sands
Region, Alberta, Canada. Sci Total Environ 618: 1665-1676.
http://dx.doi.Org/10.1016/j.scitotenv.2017.10.008
Lewis. HT. (1973). Patterns of Indian burning in California: Ecology and ethnohistory. Ramona, CA: Ballena
Press.
Lewis. HT. (1985). Why Indians burned: Specific versus general reasons. In Proceedings, Symposium and
Workshop on Wilderness Fire (pp. 75-80). (General Technical Report INT-GTR-182). Ogden, UT: U.S.
Forest Service, Intermountain Forest and Range Experiment Station, https://www.frames.gov/catalog/2658
Lewis. KLM: Reidmiller. PR: Avery. CW. (2018). Appendix 2: Information in the Fourth National Climate
Assessment. In DR Reidmiller; CW Avery; DR Easterling; KE Kunkel; KLM Lewis; TK Maycock; BC
Stewart (Eds.), Impacts, risks, and adaptation in the United States: Fourth National Climate Assessment,
volume II (pp. 1410-1412). Washington, DC: U.S. Global Change Research Program.
http://dx.doi.org/10.7930/NCA4.2018.AP2
NIFC (National Interagency Fire Center). (2018). Fire information and statistics: Wildland fires and acres (1926-
2017). Available online at https://www.nifc.gov/fireInfo/fireInfo_stats_totalFires.html
U.S. EPA (U.S. Environmental Protection Agency). (2019a). Environmental Benefits Mapping and Analysis
Program - Community Edition (BenMAP-CE) (Version 1.5) [Computer Program], Washington, DC.
Retrieved from https://www.epa.gov/benmap/benmap-community-edition
U.S. EPA (U.S. Environmental Protection Agency). (2019b). Wildfire smoke: A guide for public health officials,
revised 2019 [EPA Report]. (EPA-452/R-19-901). Washington, DC: U.S. Environmental Protection Agency,
Office of Research and Development, https://www.airnow.gov/publications/wildfire-smoke-guide/wildfire-
smoke-a-guide-for-public-health-officials/
1-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
CHAPTER 2 CONCEPTUAL FRAMEWORK FOR
EVALUATING AND COMPARING
DIFFERENT FIRE MANAGEMENT
STRATEGIES
2.1 Introduction
Fire is an important element of the natural landscape and is highly influenced by both natural and
anthropogenic factors. Fire management decisions are made at multiple governance levels to influence the
types of fires that affect different vegetative systems. Goals include increasing overall forest and
rangeland health and resilience and reducing the potential for the occurrence of uncontrolled and often
catastrophic wildfire. Current federal fire policy recognizes the importance of wildland fire
(i.e., prescribed fire and wildfire) "as an essential ecological process and natural change agent that will be
incorporated into the (land management) planning process [which includes the development of] Fire
Management Plans (FMPs), programs, and activities supporting] Land and Resource Management Plans
and their implementation" (Interagency Federal Wildland Fire Policy Review Working Group. 2001).
Different fire management strategies before, during, and after a fire can result in different effects on the
landscape and adjacent communities, including the smoke that results from wildland fire. Understanding
the effects of different fire management strategies, defined as a planned set of activities to achieve
resource objectives, can help fire managers make informed decisions that reduce adverse effects, both
directly from the fire itself as well as from the smoke it produces, while yielding desired ecological and
risk management benefits. In this chapter, we describe a conceptual framework for evaluating and
comparing different fire management strategies, using a range of metrics to characterize and quantify
effects.2 Fire management strategies are developed to achieve multiple objectives, including promotion of
ecological benefits, protection of lives and property, safe and effective responses that minimize risks to
firefighting personnel, and reduction in likelihood of severe and catastrophic wildfire. While the focus of
this assessment is on the quantification of the air quality and associated health impacts attributed to smoke
exposure, it is also important to recognize the broader effects (including both positive and negative
effects) of wildland fire in the process of considering different fire management strategies. Therefore,
subsequent chapters provide more in-depth discussions of the elements of this framework and its
implementation in comparing the effects of different fire management strategies.
2 Within this assessment, the term "impacts" refers to the main quantitative results, which includes the estimated air
pollutant concentrations from the air quality modeling and the number of health events and associated economic
values calculated using U.S. Environmental Protection Agency's (EPA's) Environmental Benefits Mapping and
Analysis Program—Community Edition (BenMAP—CE). The term "effects" is used to denote the other positive
and negative consequences of wildland fire.
2-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
The overarching question that guides the evaluation conducted within this framework is What are
the expected effects (both positive and negative) of alternative fire management strategies over both short
(during the event) and long term (post-event) time horizons? with an emphasis within this assessment on
the smoke impacts. Critical to this question are the ideas of expected positive and negative effects, a
recognition that fire needs to be viewed over a management-relevant temporal and spatial frame, and that
fire is inevitable and necessary. While some effects can be quantified and monetized broadly
(i.e., nationally), and thus used in a more traditional cost-benefit comparison, it is important to recognize
that this can be challenging when examining the effects of individual fires. Many of the effects of
wildland fire are not easily quantified or assigned a dollar value. As a result, while this assessment
estimates the air quality and the dollar value of health impacts of smoke for quantitative comparisons, it
also provides additional qualitative discussions of other effects (i.e., positive and negative) of both direct
fire and smoke.
2.2 Expected Value Framework
An expected value (EV) framework is used within this assessment because of the inherent
stochastic nature of fire in the landscape. While in many cases, a wildfire is likely to occur given a
sufficient time horizon, both the timing and location of a wildfire event is uncertain as compared to
prescribed fires which are planned events that occur at specific times of the year and in specific locations.
Wildfires can also reburn the same area with very different outcomes because of the reduction or increase
in fuel loads. For example, many of the prescribed fires in the southeastern U.S. are maintenance burns
designed to keep fuel loads low and occur on a fairly frequent basis. A range of periodicity between
wildfires had been established for different ecosystems; however, under a changing climate, the previous
assumptions on potential risk of wildfires are often challenged. The management of wildland fire can
result in a desired outcome (positive effect) or an undesirable outcome (negative effect). Fire management
strategies such as prescribed fires can reduce the uncertainty in outcomes from fires. When comparing
strategies, both stochastic and nonstochastic elements need to be expressed in a way that allows for
equivalent comparison. In atypical cost-benefit framework, comparisons between alternatives requires a
complete accounting for all costs and benefits, both direct and indirect. The conceptual framework used in
this assessment aims to provide a full accounting of the overall effects of wildland fire; however, the
ability to quantify all elements is limited. As such, this chapter emphasizes the elements that will form the
basis of and be incorporated into the main component of this assessment, the quantitative comparison of
the smoke impacts of wildland fire. Key details of the inputs in this comparative analysis are the air
quality modeling and health impact analyses described in CHAPTER 5 and CHAPTER 8. respectively.
This focus on smoke impacts is to address a key gap in the overall knowledge base regarding wildland
fire management, however, this is not intended to suggest that the other positive and negative effects of
wildland fires and fire management strategies are less important. A full accounting of costs and benefits
of those strategies will require further development of models and methods to quantify effects across the
2-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
full range of domains, including ecological, health, safety, prevention, and risk to highly valued resources
and assets.
The expected value of a specific fire management strategy requires knowledge of (1) the impacts
effects associated with different fire types (e.g., prescribed fire vs. wildfire), (2) the effects associated
with different management techniques (e.g., targeted thinning, prescribed fires), and (3) probabilities of
these effects. Two other key concepts are fire ignition probabilities and the management of a wildfire
once it has ignited. Ignition probabilities, a key factor in determining risk from wildfires, indicate the
chance that a wildfire will occur over a specified time period within a defined spatial domain (Hunter and
Roblcs. 2020). In managing wildfire risk, land managers utilize an operational risk framework that gives
primary consideration to public and firefighter safety. This risk framework is intended to consider the
degree to which the extent, intensity (energy output), and severity (effects on ecosystems) of a wildfire
can be mitigated once started based on the land management plans, fire history fuel, and weather
conditions. Both ignition probability and management can be positively or negatively impacted by the fire
management strategy.
Within this report, costs of management strategies are defined as the specific economic
expenditures associated with implementing specific management actions. For example, the costs
associated with a management strategy that includes mechanical thinning would include but not be
limited to the costs of equipment and labor costs for equipment operators. Costs here do not refer to the
outcomes of management actions, but instead these outcomes are referred to as effects, which can be
either positive or negative (see Table 2-1). One consequence of a fire management strategy may be
reductions in future costs of fire management.
For this conceptual framework, the expected value (EVi) of effects (positive + negative) for a fire
management strategy M, is specified as:
BV/= PF/ + NF + PQNY ignition \MJ) x (F\Mi)
Equation 2-1
Where /'/¦', are prescribed fire-related effects conditional on Mi, NFt are nonfire effects from Mi,
/'(WF ignition|Mi) is the probability of wildfire ignition conditional on Mi, and b)M, are fire-related
effects conditional on Mi and land management objectives once a fire is ignited. Effects include all of the
positive and negative effects associated with a fire management action. In most applications, EV will be
expressed in dollars for comparison with the dollar costs of the management strategy, and because dollars
are a unit in which all damages can be theoretically expressed. Essentially, the expected value is the effect
of the fire management action itself plus the ignition-probability-weighted effects of wildfire conditional
2-3
DRAFT: Do Not Cite or Quote
-------
1 on the management strategy. For fire management strategies that do not include prescribed fire, the first
2 term will be zero.3
3 The net benefit of a fire management strategy is defined as A'F, - (where G is the cost of
4 management strategy Mi. Within this assessment, fire management costs are treated as a known quantity.
5 There is likely to be uncertainty in those fire management costs as well; however, addressing this
6 uncertainty is beyond the scope of the assessment.
7 2.3 Components of the Conceptual Framework
8 A graphical representation of the conceptual framework is presented in Figure 2-1. This figure is
9 meant to serve as an anchor for discussions of elements of the framework. The following discussions of
10 each element provide a short description and references to the chapters and sections of this report that
11 provide more detailed qualitative discussions, and where possible, quantification methods and modeling
12 results.
3 There may be some nonsmoke or fire-related benefits and damages associated with other fire management
approaches such as mechanical thinning. We are not quantifying those impacts for this assessment.
DRAFT: Do Not Cite or Quote
2-4
-------
Improved forest health
Resource Benefits
Costs of management
actions
e.g. equipment and labor costs, fire
suppression costs, etc.
Land Management Plan
GHG emissions
Air quality
Ecosystem exposure
Ecosystem impacts
Baseline forest/
Ecological condition
' Non-smoke fire impacts
Watershed integrity
Firefighter health and
safety
Direct and indirect
economic damages
Ecological impacts
Mechanical thinning
Ecological benefits
No action
Smoke emissions
1
Management Decision
Prescribed fire
Wildfire
Severity/Extent Conditional on
Management Decision
Ability to mitigate
impacts
Human exposure
Ability to mitigate
impact
Probability of wildfire
ignition
Ability to mitigate
exposure
Non-fire adverse
impacts
GHG = greenhouse gas.
In the figure, forest management inputs are colored dark blue, management decisions and their nonsmoke related effects are colored white, resource benefits are colored green,
mitigation actions are colored light blue, fires are colored yellow and orange, fire damages are colored red, and smoke exposure related elements are colored gray. The green arrows
indicate positive effects, and the orange arrows indicate negative effects. Dotted lines represent linkages that may occur but are less certain that solid lines.
Figure 2-1 Conceptual framework for evaluating and comparing fire management strategies.
2-5
DRAFT: Do Not Cite or Quote
-------
1
2
2.3.1 Baseline Wildland Fuels Vegetation and Resource Management
Conditions
3 Baseline vegetation conditions, which are discussed in detail in CHAPTER 3. influence the
4 probability of a wildfire occurring and the intensity and characteristics of a wildfire, including smoke
5 generation. These wildland fuels vegetation conditions include location, size, density, stand composition,
6 ladder fuels4, height to live crown, understory condition, and surface fuel loads. Other vegetation and
7 resource management attributes included in land management plans (see CHAPTER 3) or that influence
8 the management and outcomes of a fire include distance from the wildland to populated areas
9 (e.g., location in or relative to the wildland-urban interface [WUI]); proximity to Superfund sites, mining
10 sites, and other legacy contaminant sites; distance to watersheds that provide community drinking water,
11 plant and wildlife habitats, infrastructure, and consideration of positive impacts from fire (e.g., restoring
12 ecosystems, fuels reduction).
13 2.3.2 Types of Fires
14 There are two types of wildland fire, as designated in statute 40 CFR § 50.1—Definitions (U.S.
15 EPA. 2020a). and by policy, as stated in National Wildfire Coordinating Group (NWCG) Glossary of
16 Wildland Fire fNWCG. 2021). The following two definitions will be used throughout this assessment in
17 order to remain consistent with their use in air quality regulation and in Federal wildland fire management
18 policy.
19 • Prescribed fire: Also referred to as planned fires, controlled burns, or prescribed burns, 40 CFR §
20 50. l(m) defines a prescribed fire as "any fire intentionally ignited by management actions in
21 accordance with applicable laws, policies, and regulations to meet specific land or resource
22 management objectives" (U.S. EPA. 2020b).
23 • Wildfire (natural and human caused): 40 CFR § 50. l(n) defines a wildfire as "... any fire started
24 by an unplanned ignition caused by lightning; volcanoes; other acts of nature; unauthorized
25 activity; or accidental, human-caused actions, or a prescribed fire that has developed into a
26 wildfire. A wildfire that predominantly occurs on wildland is a natural event" (U.S. EPA. 2020c).
27 Effects are expected to vary based on characteristics such as types of biomass burned, burn
28 conditions (e.g., temperature, humidity, wind), season, duration, intensity, and location relative to
29 populated areas (which can vary from minute to minute, day to day, and site to site) within each area
30 burned. Fires also vary based on the history of previous fire occurrences, the periodicity and intensity of
31 previous occurrences, and the management and land use history of the area in question. For the purposes
32 of this conceptual framework, the focus is on two different types of fires (i.e., prescribed fire and
4 Fuel that allows fires in low-growing vegetation to jump to taller vegetation.
2-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
wildfire), recognizing that within each category, there will be a high degree of variability based on these
characteristics.
Although rare, prescribed fires can be declared a wildfire when they are no longer meeting
objectives (e.g., escaping boundaries, intensity, smoke management). A 2013 report from the Wildland
Fire Lessons Learned Center (LLC reported that in 2012, only 0.08% of prescribed fires escaped their
planned boundaries (LLC. 2013). This includes all escapes on federal, state, tribal, and private lands that
were reported into the Wildland Fire LCC Incident Review Database, along with additional agency
notifications and media reports that were available.
Wildfires vary widely in their effects depending on location, meteorological conditions during the
fire, and the types of forests where they occur. A wildfire may be also be deemed "catastrophic"
(Wooten). resulting in severe economic, social, and ecological effects (Carey and Schumann. 2003).
including a high percentage of dead trees (Wooten). While there is a great deal of year-to-year variability,
in recent decades, wildfires have affected an increasing number of acres, with an average of 6.9 million
acres burned from 2000-2019 compared with an average of 3.2 million acres burned from 1980-1999
(NICC. 2019).
On February 13, 2009, the Guidance for Implementation of Federal Wildland Fire Management
Policy was issued (FEC. 2009). This guidance provides for consistent implementation of the 1995 Federal
Fire Policy and the 2001 update. By policy, management response to a wildfire on federal land is based
on objectives established in an applicable Land/Resource Management Plan (L/RMP) and or FMP. Fire
management objectives are affected by changes in fuels, weather, topography, varying social and political
understanding and involvement of other governmental jurisdictions that may have different missions and
objectives. Managers use a decision support process to guide and document wildfire management
decisions. The process includes land management objectives, situational awareness, analysis of hazards
and risk, defining of implementation actions and the fire management decision documentation and
rationale.
A full range of fire management strategies can be used to achieve L/RMP and FMP objectives.
Wildfire may be managed solely to meet protection objectives, such as protecting values at risk of loss by
suppressing the fire in the safest, most effective, and efficient way. The initial response may be as simple
as evaluating the location of the fire without further on-the-ground active suppression action in areas
where the fire is distant from valued assets that require action to protect or where the risks from exposure
for firefighters is higher than the value of the assets that would be protected. Wildfire may be managed
concurrently for one or more objectives, and the objectives can change as the fire spreads across the
landscape. For example, a wildfire can be managed for suppression to protect points of valued resources
while at the same time taking no action when or where resource values are being enhanced.
2-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
No matter how a wildfire is being managed, firefighter and public safety is the first priority. All
fire management activities and decisions must reflect this commitment. A fuller description of how
wildfire can be used as a land management tool can be found in CHAPTER 3.
2.3.3 Fire Management Strategies
Severity of fires is determined by a number of factors, some of which can be affected by
management practices (e.g., forest structure, fuels, vegetation composition) and others which cannot be
controlled (e.g., weather, location). Most fire management strategies focus on fuel load reduction, which
is a management strategy that involves "manipulation, including combustion, or removal of fuels to
reduce the likelihood of ignition and/or to lessen potential damage and resistance to control" (USFS.
2003a'). Fuel reduction strategies aim to reduce the probability of ignition and reduce the intensity and
uncontrolled spread of wildfires (Agee and Skinner. 2005). Thus, fuel reduction strategies directly affect
two key parameters in the framework, /'(WF Ignition) and /"(control). Two common practices for fuel
load reduction include prescribed fires and mechanical thinning.
2.3.3.1 Prescribed Fires
Prescribed fires, as defined in Section 2.3.2. are a fire management tool that uses planned,
controlled fires to reduce fuel loads and achieve the ecological benefits of fires while reducing the
potential for catastrophic uncontrolled fires. There is growing evidence that prescribed fires can reduce
surface fuels and reduce fire severity while maintaining or improving forest health (Hunter and Robles.
2020; Kalies and Kent. 2016; USFS. 2003b).
Prescriptions for fire are based on clearly defined objectives, which might include ecological
aspects such as habitat diversity and endangered species recovery, as well as fuel reduction to reduce the
potential of high intensity, high severity fires. Prescriptions also take into account environmental and
meteorological conditions, fuels, burn area, and planned approaches for suppression once objectives are
met to reduce potential adverse impacts, including those associated with smoke emissions (USFS. 2021;
U.S. EPA. 2020d). The effectiveness of prescribed fires in reducing the potential for severe fires is
dependent on weather patterns and ecosystem characteristics such as types of fuels, as well as the
interactions between them [e.g., drought may affect fuel moisture content; Fernandes and Botelho
(2003)1.
On federal and most state lands, prescribed fire is only used after thorough preplanning and only
by highly trained and experienced professionals. It is only implemented when conditions meet preplanned
elements and adequate contingencies are in place or confirmed by managers and Agency Administrators.
Go/no-go checklists are used to determine compliance with policies and the prescribed fire plan
parameters (NWCG. 2017).
2-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
2.3.3.2 Mechanical Fuel Reduction
Mechanical treatments to thin trees and remove fuels can be used in conjunction with prescribed
fires or be employed in places and times when prescribed fires cannot be used (Mclvcr et al.. 2013). They
require equipment as well as plans for disposal or utilization of significant quantities of small trees (Agee
and Skinner. 2005; Rummer et al.. 2003). Thinning trees can reduce surface fuel loads, and also reduce
risks of crown fires (fires that spread across tree canopies) which can cause severe damage. There are
multiple types of thinning that affect different aspects of forest composition, including low thinning that
removes small trees, crown thinning which removes medium size trees, and selection thinning, which
removes larger, more marketable trees (Agcc and Skinner. 2005). How the residual wood from the
thinning operations is disposed of can have a substantial impact on surface fuel availability with chipping
or burning of the unusable tops of trees having the greatest impact on reducing fuel loads.
There is limited observational data on the degree to which mechanical thinning, alone or in
conjunction with prescribed fires changes the probability of ignition or intensity and severity of fires.
Simulations have shown that removing small trees and "ladder fuels" (i.e., fuels that allow fires to climb
up to forest canopies) can be effective in reducing fire severity, especially when in conjunction with
prescribed fires (Agee and Skinner. 2005).
2.3.3.3 Fuel Treatment Effectiveness
In 2006, the U.S. Department of Agriculture Forest Service and Department of the Interior (DOI)
Bureau of Land Management (BLM) initiated a program to evaluate the effectiveness of hazardous fuel
treatments (prescribed fire and mechanical) designed to reduce the potential of high intensity, high
severity wildfires. When a fuel treatment is tested by wildfire, an evaluation is performed to determine the
effectiveness of the treatment in changing the fire behavior (e.g., going from a crown fire to a surface fire)
and/or helping manage the wildfire. In 2011, the Forest Service and the DOI land management agencies
(Bureau of Indian Affairs, BLM, Fish and Wildlife Service, and National Park Service) made the
effectiveness assessment mandatory whenever a wildfire impacted a previously treated area.
Since 2006, almost 14,860 assessments have been completed (IFTDSS. 2021). About 89% of the
fuel treatments were effective in changing fire behavior or helping with management of the wildfire or
both (IFTDSS. 2021). In addition, prescribed fire treatments were observed to be the most effective in
changing fire behavior and reducing overstory mortality from wildfires. Unfortunately, until recently, due
to limitations in reporting systems the ability to detect all wildfire fuel treatment interactions has been
limited, resulting in a significant under sampling of fuel treatment effectiveness monitoring, mostly on the
smaller fires (less than 1,000 acres).
2-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
2.3.4
Effects of Fire
Prescribed fires and wildfires have the potential for both positive and negative effects, although
the magnitude of potential effects differs. The goal of prescribed fires is to reduce the fuel loads that will
result in decreasing the frequency, intensity, and severity of a wildfire while providing for safe and
effective response to wildfire and protecting highly valued resources and assets. In general, positive
effects that occur directly from fire result from improvements in landscape/watershed health which yield
ecological benefits or ecosystem services. Negative effects occur both directly, as a result of the fire itself,
or indirectly, through emissions of smoke and ash. The magnitude, scale, and duration of these effects
will depend highly on the type of fire, the fuel conditions, the terrain, and the fire weather conditions, as
well as the location relative to the WUI, and downwind populations. Air quality impacts result from
smoke emissions that impact ambient concentrations of numerous pollutants including ozone and
particulate matter, specifically fine particulate matter (PM2 5 [particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |im |: see CHAPTER 4 and CHAPTER 5). which have
been shown to contribute to a wide variety of adverse health and ecological impacts [see CHAPTER 6;
Holm et al. (2021); Jaffe et al. (2020); Cascio (2018)1. The severe wildfires occurring in the western U.S.
over the past few years causing loss of life and property and the reversal of trends in air quality
improvements in the western states attributed to increasing wildfire emissions (McClurc and Jaffe. 2018)
have drawn the attention of the National Academies of Science, Medicine and Engineering ("NASEM.
2020) and other medical professional organizations (Kaufman et al.. 2020; Raiagopalan et al.. 2020; Rice
et al.. In Press) which are strongly advocating for attention to finding solutions to prevent such severe
wildfires while simultaneously mitigating the adverse effect of exposure to smoke.
2.3.4.1 Direct Fire Effects
2.3.4.1.1 Benefits to Wildland Ecosystems
Many wildland ecosystems have adapted to periodic fires. In fact, a number of tree species such
as pines depend on fire for reproduction, as do many shrubs and most grasses. Other species, such as
Sequoias, rely on periodic fires to open up forest canopies to allow saplings to grow and flourish. Open
canopies also support the growth of shade-intolerant plants and reduce the probability of crown fire. Fires
also convert brush and dead trees and plants to nutrient rich ash which can be beneficial to established
trees and provide essential nutrients for new forest growth. These nutrients are also important to support
soil microbes which increases the overall health of wildland ecosystems. Fires and smoke can also
remove invasive species not adapted to fires, as well as reduce populations of destructive insects and
diseases (Nearv et al.. 2005; Brown and Smith. 2000; Smith. 2000). In some cases, for example
cheatgrass, fires can also help to control invasive species (Nearv et al.. 2005; Brown and Smith. 2000;
2-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Smith. 2000; Young et al.. 1987). Detailed information on benefits of wildland fire on wildland
ecosystems is provided in CHAPTER 3.
2.3.4.1.2
Benefits to Fire Management (Post-Event)
As discussed in Section 2.3.3.1 and Section 2.3.3.3. prescribed fires are designed to reduce the
potential for severe fire damages by changing the behavior of a subsequent wildfire and making it easier
to manage. This can result in fewer risks to firefighting personnel during subsequent wildfires, as well as
reducing economic damages, ecological damages, and health impacts to populations from fires and poor
air quality caused by smoke.
Direct fire damages, described in CHAPTER 7. include effects to firefighters (including impacts
from direct smoke inhalation), effects to populations in the vicinity of fires, economic damages, and
ecological damages. Health impacts to firefighters can be immediate, due to extreme heat, burns,
asphyxiation, overexertion, or accidents, or can be delayed, due to smoke-related diseases such as cancers
and chronic conditions such as heart disease that may be associated with prolonged and repeated
exposures to extreme heat, overexertion, and stress (Domitrovich et al.. 2017). Effects to populations in
the vicinity of fires include deaths, injuries, and psychological damages (Thomas et al.. 2017). Economic
damages include the value of lost property; loss of marketable timber; direct and indirect costs of
evacuations, including business interruption; damages to infrastructure, such as downed power lines or
damaged roadways; and the value of lost recreational resources, due to either safety-related closures or
fire damage (Thomas et al.. 2017). Ecological damages can occur due to changes in vegetation
composition; conversion from one vegetation type (e.g., forest) to another (e.g., shrubs); damage to soils,
which could lead to flooding and degraded water quality and quantity; loss of habitat and endangered or
threatened species; increased susceptibility to insects and diseases; and climate-related damages resulting
from releases of greenhouse gases (GHGs) and loss of carbon sequestration potential (Thomas et al..
All wildland fires produce smoke and ash. The amount and composition of smoke can vary
between the types of fires due to the types of burn conditions and type, loading, and consumption of fuels.
Release height and transport of smoke can also vary between types of fires (as well as within types of
fires) depending on meteorological conditions and burn conditions. For example, plume rise will depend
on the temperature of the fire, and long-range transport of smoke will depend on wind speed and
2.3.4.1.3
Fire Damages
2017).
2.3.4.2 Effects from Smoke and Ash
2-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
direction, as well as plume rise. The impacts associated with smoke emissions will depend on the
emissions density, how far and in which direction the smoke travels, and on the proximity of a fire to
downwind populated areas. CHAPTER 4 and CHAPTER 5 describe approaches used to monitor and
model air quality impacts from wildland fire smoke.
2.3.4.2.1 Smoke-Related Effects
Smoke has immediate impacts directly in the vicinity of a fire, as well as impacts downwind of a
fire due to worsened air quality. There are smoke transport mechanisms which function under flaming and
smoldering phases of a fire. These phases are important in terms of emissions, how far the emissions will
transport and implication in terms of safety impacts such as roadway visibility and air quality. CHAPTER
4 and CHAPTER 5 describe the current state of knowledge about smoke contributions to poor air quality
based on monitoring and modeling. CHAPTER 6 describes health and ecological effects associated with
smoke and worsened air quality, but also recognizes that smoke can also have some positive impacts,
such as stimulating flowering of some perennial grasses and herbs and contributing to climate cooling.
2.3.4.2.2 Ash-Related Damages
Ash from fires, discussed in CHAPTER 6. can deposit on soils, water, vegetation, and man-made
structures and vehicles. Ash deposition can lead to increased nutrient availability in soils, and depending
on what types of materials are burned, can also lead to increased levels of metals. Ash deposition can also
affect water quality, either directly through ash residues entering water bodies, or through increases in
nutrient loadings that result from movement of excess nutrients through soils.
2.3.4.2.3 Effects on Greenhouse Gas (GHG) Emissions
Fires result in the release of a number of GHGs, both from burning of trees and other woody
biomass, as well as from soils. Greenhouse gases released include CO2, N2O, NOx, and methane.
Emissions are a function of climate, soil properties, and vegetation composition and management
practices. Emissions of GHGs occur both during the fire, as well as longer-term, due to changes in soil
and surface fuel carbon and nitrogen pool sizes, conversions from one vegetation type to another, and
changes in soil moisture and temperature associated with canopy removal. There are differences in plume
rise and fuels consumed between most wildfires and prescribed fires which result in substantially different
areas of impact as well as potential entrainment into long-range transport and retention of GHGs in the
upper atmosphere [see U.S. EPA (2012)1. A clear benefit of fuels treatments including prescribed fire,
which affect wildfire risk, is the potential to improve long-term carbon sequestration [see CARB (2015)1.
2-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
2.3.5
Programs to Mitigate Exposures and Impacts
Prescribed fires occur after extensive planning in an attempt to reduce population exposures to
smoke and provide an opportunity to reduce smoke exposures of downwind communities through public
health messaging campaigns. As a result, the ability of behavioral actions such as staying indoors or using
N95 facemasks when outdoors to mitigate exposures can play an important role in reducing the health
impacts associated with smoke emissions during prescribed fires. While there is some limited opportunity
to use these types of behavioral actions during wildfires, prescribed fires provide the opportunity to
increase those behaviors in at-risk populations through communication and public awareness activities.
Likewise, communities can increase readiness for smoke during prescribed fires through public
information messaging about nearby burning activity or through messaging campaigns to ensure
populations, especially those at increased risk, are taking measures to protect themselves. Consideration
of programs that increase awareness of prescribed fire events, including the projected path of smoke
plumes, could have a large influence on reducing health impacts.
Wildfire smoke also has some opportunities for mitigation of exposures and effects. The
implementation of the Interagency Wildland Fire Air Quality Response Program as authorized by
PL 116-9 March 12, 2019. Page 617; Section 1114(f), as well as efforts by U.S. Environmental Protection
Agency (U.S. EPA), state, tribal, and local air quality regulatory agencies and public health agencies warn
the public and at-risk populations of wildfire smoke exposures and ways to mitigate impacts. Through
these efforts, the public is becoming more aware of the risks of wildfire smoke exposure and air quality
and health impacts. CHAPTER 6 (Section 6.3) provides a discussion of the various actions and
interventions that can be employed by individuals to mitigate or reduce wildland fire smoke exposures.
2.3.6 Implementing the Conceptual Framework
Wildland fire results in a range of beneficial and detrimental effects, some of which can be
quantified, while others are more difficult to quantify. Table 2-1 lists the categories of impacts associated
with wildland fire, both the direct fire effects and those specific to smoke exposure, and highlights those
effects that are the focus of the quantitative analyses that revolve around the case study fires (i.e., Timber
Crater 6 and Rough Fires) examined within this assessment. The nature and magnitude of these effects
will be dependent on the type of fire experienced, the vegetation affected, and the timescale, but the
potential for these effects exists for both prescribed fires and wildfires. Effects can occur directly within
the fire boundary, adjacent to the fire, or distant from the fire, for example impacts of smoke emissions on
air quality or degradation of water quality. Additionally, effects can be within a few days, or over months
or years. Effects can be positive or negative with positive effects providing some advantage, which could
include restoring ecosystems or mitigating the risk or loss from a wildfire, while negative effects describe
detrimental consequences from a fire, which could include damages to public health, property, or
infrastructure. The conceptual framework outlined within this chapter described the linkages between the
2-13
DRAFT: Do Not Cite or Quote
-------
1 direct fire and smoke effects of wildland fire to lay the foundation for discussions in subsequent chapters
2 that qualitatively and quantitatively evaluate the effects of prescribed fire and wildfire in an attempt to
3 provide an overall comparison of the benefits and costs associated with different fire management
4 strategies, with a focus on the smoke impacts.
Table 2-1 Expected effects associated with wildland fire: quantified and
unquantified for the case study analyses.
Categories of Expected Effects
Firefighting
•
Firefighter safety
•
Firefighter injuries/fatalities
•
Firefighter health, both mental and physical
Economic
•
Evacuations
•
Property (e.g., structures)
•
Property (e.g., loss of ecosystem services)
•
Timber and grazing
TO
•
Infrastructure (e.g., powerlines, recreation, others)
O
£
•
Municipal watersheds (e.g., reservoirs, industry, agriculture, drinking)
LU
-o
0)
•
Tourism (e.g., recreation, lodging, restaurants, etc.)
+-»
C
•
Aesthetics (e.g., property value, view shed, etc,)
GJ
3
O"
•
Natural and cultural resources
z>
•
Fuel reduction—cost effective method of treating acres
•
Fuel reduction—treatment opportunities not limited to local markets'5
Ecological
•
Ecological services including game and endangered species
•
Ecosystem health and resiliency
•
Restoration/maintenance of historic natural fire regime
•
Invasive species
•
Climate change (e.g., GHGs, carbon)
•
Redistribution of toxics and nutrients (e.g., mercury, metals, sulfur, nitrogen)
•
Soil and water quality and quantity
2-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 2 1 (Continued): Expected effects associated with wildland fire: quantified
and unquantified for the case study analyses.
Categories of Expected Effects
TO
¦ (/>
Public Health: Direct Fire
• Injuries
15 &
Z>
+-»
O
• PM2.5 concentrations
£
LU
• Ozone concentrations
T3
c
(5
D
o
• Respiratory- and cardiovascular-related emergency department visits and hospital
admissions
• Premature mortality
GHG = greenhouse gas; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm.
aOf these unquantified effects, some are not discussed in this assessment.
This fuel reduction effect reflects the issue that in some locations fuel reduction options are limited by the lack of local markets
for products such as merchantable timber of biomass, resulting in prescribed fire and chipping as the only fuel reduction options
available. The presence of local markets reduces costs and increases the fuel reduction options available.
°Examining these effects represents the primary focus of this assessment.
See Section A.2 (Table A.2-1) for a more detailed version of this table that accounts for whether the effects listed result in
positive or negative impacts due to prescribed fire and wildfire.
Fully implementing the conceptual framework detailed within this chapter requires a diverse set
of data and models. The ultimate results would be a complete set of quantified, and in some cases
monetized impacts, specifically health impacts and corresponding economic values, associated with each
selected fire management strategy. However, for the purpose of this assessment the quantification is
limited to the smoke impacts associated with different fire management strategies and reflects a
comparison of only one area of negative effects and not a comprehensive, full accounting of both the
negative effects along with the positive effects of wildland fire. Monetization is useful because it provides
a consistent way to aggregate disparate effects. Economic theory and practice typically recommend
discounting of benefits and costs that occur in the future to account for societal time preferences,
e.g., benefits occurring today are in most cases valued higher than benefits occurring in the future (U.S.
EPA. 2014). Because of uncertainty regarding when wildfires occur relative to when prescribed fires
occur, it is challenging to determine the timeframes for comparing the two types of fires. For this
assessment, we present undiscounted dollar values, which assumes that benefits and costs of fire
management strategies all occur in the same current year. Comparisons would differ if prescribed fire
effects are assumed to occur earlier in time than wildfire effects. A full accounting of comparisons
between strategies would require aggregating all of the monetized benefits and damages for each fire
management strategy, and then computing the expected value of damages using Equation 2-1, and
differencing the expected values between strategies (e.g., fire management strategy /' will have benefits
2-15
DRAFT: Do Not Cite or Quote
-------
1 compared with fire management strategy j if EVi - EVj > 0). However, given the limited availability of
2 data to model many nonhealth endpoints, this assessment only aggregates the values of health endpoints
3 associated with air quality changes due to smoke.
4 Net benefits can also be compared between fire management strategies. With a complete set of
5 potential wildland vegetation management strategies, the optimal strategy will be the one with the highest
6 net benefits. Even with an incomplete set, fire management strategy /' is preferred to management strategy
7 j if NBi > Ml.
2-16
DRAFT: Do Not Cite or Quote
-------
2.4 References
Agee. JK: Skinner. CN. (2005). Basic principles of forest fuel reduction treatments. ForEcol Manage 211: 83-
96. http://dx.doi.Org/10.1016/i.foreco.2005.01.034
Brown. JK: Smith. JK. (2000). Wildland fire in ecosystems: Effects of fire on flora. (RMRS-GTR-42-vol. 2).
Ogden, UT: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station.
http://dx.doi.org/10.2737/RMRS-GTR-42-V2
CARB (California Air Resources Board). (2015). Compliance Offset Protocol U.S. Forest Projects. Sacramento,
CA.
https://ww2.arb.ca.gov/sites/default/files/classic/cc/capandtrade/protocols/usforest/forestprotocol2015.pdf
Carey. H: Schumann. M. (2003). Modifying wildfire behavior — The effectiveness of fuel treatments. (Working
Paper #2). Santa Fe, NM: National Community Forestry Center, Southwest Region.
https://foreststewardsguild.org/wp-content/uploads/2019/06/ModifVing Fire Behavior new.pdf
Cascio. WE. (2018). Wildland fire smoke and human health [Review]. Sci Total Environ 624: 586-595.
http://dx.doi.Org/10.1016/i.scitotenv.2017.12.086
Domitrovich. JW: Brovles. GA: Ottmar. RD: Reinhardt. TE: Naeher. LP: Kleinman. MT: Navarro. KM:
Mackav. CE: Adetona. O. (2017). Final report: Wildland fire smoke health effects on wildland firefighters
and the public. (JFSP project ID 13-1-02-14). Boise, ID: Joint Fire Science Program.
https://www.firescience.gov/proiects/13-l-02-14/proiect/13-l-02-14 final report.pdf
FEC (Fire Executive Council). (2009). Guidance for Implementation of Federal Wildland Fire Management
Policy. Washington, DC: U.S. Department of Agriculture, U.S. Department of the Interior.
https://www.nifc.gov/policies/policies documents/GIFWFMP.pdf
Fernandes. PM: Botelho. HS. (2003). A review of prescribed burning effectiveness in fire hazard reduction
[Review]. International Journal of Wildland Fire 12: 117-128. http://dx.doi.org/10.1071/WF02042
Holm. SM: Miller. MP: Balmes. JR. (2021). Health effects of wildfire smoke in children and public health tools:
A narrative review [Review]. J Expo Sci Environ Epidemiol 31: 1 -20. http://dx.doi.org/10.1038/s41370-02Q-
00267-4
Hunter. ME: Robles. MP. (2020). Tamm review: The effects of prescribed fire on wildfire regimes and impacts:
A framework for comparison [Review]. For Ecol Manage 475: 118435.
http://dx.doi.Org/10.1016/i.foreco.2020.118435
IFTDSS (Interagency Fuel Treatment Decision Support System). (2021). Interagency Fuel Treatment
Effectiveness Monitoring (FTEM) reporting systems. Retrieved from
https://iftdss.firenet.gov/landing page/index.html
Interagency Federal Wildland Fire Policy Review Working Group. (2001). Review and Update of the 1995
Federal Wildland Fire Management Policy. Boise, ID: National Interagency Fire Center.
https://www.doi.gov/sites/doi.gov/files/uploads/2001-wfm-policy-review.pdf
Jaffe. DA: O'Neill. SM: Larkin. NK: Holder. AL: Peterson. PL: Halofskv. JE: Rappold. AG. (2020). Wildfire
and prescribed burning impacts on air quality in the United States [Editorial]. J Air Waste Manag Assoc 70:
583-615. http://dx.doi.org/10.1080/10962247.2020.1749731
Kalies. EL: Kent. LLY. (2016). Tamm Review: Are fuel treatments effective at achieving ecological and social
objectives? A systematic review. ForEcol Manage 375: 84-95.
http://dx.doi.Org/10.1016/i.foreco.2016.05.021
Kaufman. JP: Elkind. M: Bhatnagar. A: Koehler. K: Balmes. JR: Sidney. S: Burroughs Pefia. MS: Pockerv.
PW: Hou. L: Brook. RP: Laden. F: Raiagopalan. S: Bishop Kendrick. K: Turner. JR. (2020). Guidance to
reduce the cardiovascular burden of ambient air pollutants: A policy statement from the American Heart
Association. Circulation 142: e432-e447. http://dx.doi.org/10.1161/CIR.0000000000000930
2-17
PRAFT: Po Not Cite or Quote
-------
LLC (Wildland Fire Lessons Learned Center). (2013). 2012 escaped prescribed fire review summary: Lessons
from escaped prescribed fires. Tucson, AZ.
http://www.ncprescribedfirecouncil.org/pdfs/2012 Escaped Prescribed Fire Review Summarv.pdf
McClure. CD: Jaffe. DA. (2018). US particulate matter air quality improves except in wildfire-prone areas. Proc
Natl Acad Sci USA 115: 7901-7906. http://dx.doi.org/10.1073/pnas.18Q4353115
Mclver. JD: Stephens. SL: Agee. JK: Barbour. J: Boerner. REJ: Edminster. CB: Erickson. KL: Farris. KL:
Fettig. CJ: Fiedler. CE: Haase. S: Hart. SC: Keelev. JE: Knapp. EE: Lehmkuhl. JF: Moghaddas. JJ: Otrosina.
W: Outcalt. KW: Schwilk. DW: Skinner. CN: Waldrop. TA: Weatherspoon. CP: Yaussv. DA: Youngblood.
A: Zack. S. (2013). Ecological effects of alternative fuel-reduction treatments: Highlights of the National Fire
and Fire Surrogate study (FFS) [Review]. International Journal of Wildland Fire 22: 63-82.
http://dx.doi.org/10.1071/WF11130
NASEM (National Academies of Sciences, Engineering, and Medicine). (2020). Implications of the California
wildfires for health, communities, and preparedness: Proceedings of a workshop. Washington, DC: National
Academies Press, http://dx.doi.org/10.17226/25622
Nearv. DG: Ryan. KC: DeBano. LF. (2005). Wildland fire in ecosystems: Effects of fire on soil and water.
(RMRS-GTR-42-vol. 4). Ogden, UT: U.S. Department of Agriculture, Forest Service, Rocky Mountain
Research Station. http://dx.doi.org/10.2737/RMRS-GTR-42-V4
NICC (National Interagency Coordination Center). (2019). Total wildland fires and acres (1926-2019). Available
online at https://www.nifc.gov/fireInfo/fireInfo stats totalFires.html (accessed February 4, 2021).
NWCG (National Wildfire Coordinating Group). (2017). Interagency Prescribed Fire Planning and
Implementation Procedures Guide. (PMS 484). National Wildfire Coordinating Group, Fuels Management
Committee, https://www.nwcg.gov/publications/484
NWCG (National Wildfire Coordinating Group). (2021). NWCG glossary of wildland fire, PMS 205. Available
online at https://www.nwcg.gov/glossary/a-z (accessed February 4, 2021).
Raiagopalan. S: Brauer. M: Bhatnagar. A: Bhatt. PL: Brook. JR: Huang. W: Mtinzel. T: Newbv. D: Siegel. J:
Brook. RD. (2020). Personal-level protective actions against particulate matter air pollution exposure: A
scientific statement from the American Heart Association. Circulation 142: e41 l-e431.
http://dx.doi.org/10.1161/CIR.000000000000Q931
Rice. MB: Henderson. SB: Lambert. AA: Cromar. KR: Hall. JA: Cascio. WE: Smith. PG: Marsh. BJ: Coefield.
S: Balmes. JR: Kamal. A: Gilmour. MI: Carlsten. C. : Navarro. K. M.: Collman. GW: Rappold. A: Miller.
MP: Stone. SL: Costa. PL. (In Press) Respiratory impacts of wildland fire smoke: Future challenges and
policy opportunities. An official American Thoracic Society workshop report. Ann Am Thorac Soc.
Rummer. B: Prestemon. J: May. P: Miles. P: Vissage. J: McRoberts. R: Liknes. G: Shepperd. WP: Ferguson. P:
Elliot. W: Miller. S: Reutebuch. S: Barbour. J: Fried. J: Stokes. B: Bilek. E: Skog. K. (2003). A strategic
assessment of forest biomass and fuel reduction treatments in western states. Washington, PC: U.S.
Pepartment of Agriculture, Forest Service, https://www.fs.usda.gov/treesearch/pubs/23846
Smith. JK. (2000). Wildland fire in ecosystems: Effects of fire on fauna. (RMRS-GTR-42-vol. 1). Ogden, UT:
U.S. Pepartment of Agriculture, Forest Service, Rocky Mountain Research Station.
http://dx.doi.org/10.2737/RMRS-GTR-42-Vl
Thomas. P: Butrv. P: Gilbert. S: Webb. P: Fung. J. (2017). The costs and losses of wildfires: A literature
review. (NIST Special Publication 1215). Gaithersburg, MP: National Institute of Standards and
Technology. http://dx.doi.org/10.6028/NIST.SP.1215
U.S. EPA (U.S. Environmental Protection Agency). (2012). Report to Congress on black carbon [EPA Report].
(EPA-450/R-12-001). Washington, PC. https://nepis.epa. gov/Exe/ZvPURL.cgi?Pockev=P 100EIJZ.txt
U.S. EPA (U.S. Environmental Protection Agency). (2014). Guidelines for preparing economic analyses [EPA
Report]. Washington, PC: U.S. Environmental Protection Agency, National Center for Environmental
Economics, https ://www.epa. gov/environmental-economics/guidelines-preparing-economic-analyses
2-18
PRAFT: Po Not Cite or Quote
-------
U.S. EPA. Definitions. 40 CFR § 50.1 (2020a). httos://www.govinfo.gov/app/details/CFR-2020-title4Q-
vol2/CFR-2020-title40-vol2-sec50-l
U.S. EPA. Definitions: Prescribed fire. 40 CFR § 50. Kin) (2020b). https://www.govinfo.gov/app/details/CFR-
2020-title40-vol2/CFR-2020-title40-vol2-sec50-l
U.S. EPA. Definitions: Wildfire. 40 CFR § 50.1(n) (2020c). https://www.govinfo.gov/app/details/CFR-202Q-
title40-vol2/CFR-2020-title40-vol2-sec50-l
U.S. EPA (U.S. Environmental Protection Agency). (2020d). Wildland fire: What is a prescribed fire? Available
online at https://www.nps.gov/articles/what-is-a-prescribed-fire.htm (accessed February 4, 2021).
USFS (U.S. Forest Service). (2003a). Fire terminology. Available online (accessed January 20, 2021).
USFS (U.S. Forest Service). (2003b). Influence of forest structure on wildfire behavior and the severity of its
effects: An overview. Washington, DC: U.S. Department of Agriculture, Forest Service.
https://www.fs.fed.us/proiects/hfi/docs/forest structure wildfire.pdf
USFS (U.S. Forest Service). (2021). Prescribed fire. Available online at https://www.fs.usda.gov/managing-
land/prescribed-fire (accessed February 4, 2021).
Wooten, G. Fire and fuels management: Definitions, ambiguous terminology and references. Washington, DC:
U.S. Department of the Interior, National Park Service.
https://www.nps.gov/olvm/learn/management/upload/fire-wildfire-definitions-2.pdf
Young. JA: Evans. RA: Eckert. RE. Jr: Kav. BL. (1987). Cheatgrass. Rangelands 9: 266-270.
2-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
CHAPTER 3 FIRE REGIMES, FIRE EFFECTS,
AND A HISTORY OF FUELS AND
FIRE MANAGEMENT IN DRY
FORESTS OF THE PONDEROSA
PINE REGION
3.1 Fire Regimes and Ecological Condition of Forests
Fire regimes are patterns of fire size, intensity, severity, recurrence or frequency and the resulting
ecological effects that are typical of vegetation assemblages in spatial scales from sites to broad regions of
the county (Agee. 1993). They are typically based on historical patterns, based on human observation,
ecological records, and geological records, depending on the length of available data, and they are
temporally dynamic, depending on longer-term vegetation and climatic distributions, as well as on long
periods of human interaction and resource use. Fire regimes have changed with climate over long time
periods; they are likely changing now as well, although we are not able to define the changes while they
are occurring. They influence forest recovery, succession, structure, and ecosystem functioning (Agcc.
1993). Fire regimes are influenced largely by climate, vegetation types and by topographic and geologic
features that either facilitate or restrict fire spread and vary by season and geographic region resulting
from regional weather patterns (Taylor and Skinner. 1998; Agcc. 1993).
There are numerous classification systems for describing fire regimes, often depending on the
context and purpose of the classification system. The most used classifications consider the frequency,
severity (or scale of ecological impacts) and a measure of spatial scale of wildfire in a natural or
quasi-natural condition, although many other variables have been used in classification schemes rRvan
and Opperman (2013); Table 3-1, Figure 3-1], Fire frequency, or the mean fire return interval, is a
measure of how often fire returns, on average, to a specific area. There may be a wide range around this
mean, which has important ecological implications for stand development and forest structure (Baker and
Ehle. 2001). Landscape fire rotation, often used to characterize fire regimes, refers to the years required
for a defined area to experience fire (Farris et al.. 2010) and helps to smoothen out variations over space
and time to help characterize typical fire regimes.
3-1
DRAFT: Do Not Cite or Quote
-------
Table 3-1 Fire regime groups and descriptions.
Group
Frequency
Severity
Severity Description
I
0-35 yr
Low/mixed
Generally low-severity fires replacing
less than 25% of the dominant
overstory vegetation; can include
mixed-severity fires that replace up to
75% of the overstory
II
0-35 yr
Replacement
High-severity fires replacing greater
than 75% of the dominant overstory
vegetation
III
35-200 yr
Mixed/low
Generally mixed-severity can also
include low-severity fires
IV
35-200 yr
Replacement
High-severity fires
V
200+ yr
Replacement/any severity
Generally replacement-severity; can
include any severity type in this
frequency range
Source: Hann et al. (2008)
3-2
DRAFT: Do Not Cite or Quote
-------
LANDFIRE: Fire Regime Groups
FRG = Fire Regime Group; LF = LANDFIRE.
Note: FRG definitions best approximate the definitions outlined in the Interagency Fire Regime Condition Class Guidebook and
refined to create discrete, mutually exclusive criteria appropriate for use with LF's fire frequency and severity data products.
Source: LF (2012).
Figure 3-1 Fire Regime Groups characterizing the presumed historical fire
regimes within landscapes based on interactions between
vegetation dynamics, fire spread, fire effects, and spatial context.
1
2 Fire severity is determined by either a visual estimate or measured assessment of fire effects on
3 soils and vegetation (Table 3-1). Fire intensity, a major factor in severity, is a measure of heat or energy
4 released (kW) per unit length (m) along the fireline and can be estimated by measuring flame length as
5 the flaming front passes a known point (Rothermel and Deeming. 1980). Fligh-intensitv fires (e.g., long
6 flame lengths), for example, result in more consumption and charring of surface fuel, increased exposure
7 of soil and alteration of soil properties, and more damage and mortality of trees and other vegetation.
8 While duration of burning at a given site has profound implications for fire severity and smoke
9 production, duration is much more difficult to observe and to characterize than fire intensity.
3-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.1.1
Historic Fire Regimes in the Ponderosa Pine Region
This chapter focuses on the characterization of ponderosa pine ecosystems because they are very
well understood and comprise a large portion of the ecosystem within the two case study areas that form
the basis of the quantitative analyses within this assessment (see CHAPTER 5 and CHAPTER 8). At a
finer resolution, the case studies do contain some different forest types as well as shrub, grass, and
understory vegetation components. However, these areas represent much smaller areas than ponderosa
pine and dry mixed conifer forest.
Historically, ponderosa pine (Pinus ponderosa P. & C. Lawson) forests and much of the adjacent
dry mixed conifer zone experienced frequent, mixed to low-intensity fire (Agee. 1993). Periodic fires
consumed accumulated fuels, thinned young seedlings and saplings, and consumed shrubs and herbaceous
plant material, leaving the large, fire-resistant trees intact. Some individual large trees or small groups of
large trees may have been directly killed or stressed by fire and later attacked and killed by bark beetles
(Mungcr. 1917). This fire regime aligns geographically with the current distribution of ponderosa pine,
which occupies 76,997 km2 (14.7% of the land area) in Oregon and Washington, and approximately
94,200 km2 (11% of the land area) in northern California (Figure 3-2). For the purposes of this
assessment, the area occupied by these forests is collectively referred to as the ponderosa pine region.
The continental climate of the ponderosa pine region is semiarid and is largely controlled by a
rain shadow effect from the Cascade, Coast and Sierra mountain ranges to the west. Annual summer
droughts are a common characteristic as less than 20% of precipitation falls during May-September,
based on precipitation data from the Parameter-Elevation Relationships on Independent Slopes Model
(PRISM) Climate Group at Oregon State University (Daly et al.. 2008). Historically, low-severity surface
fires were more frequent and burned over larger areas compared to nondrought years (Hagmann et al..
2019; Johnston. 2017; Heverdahl et al.. 2008; McKenzie et al.. 2004). However, drought is usually not the
sole or ultimate cause of most tree mortality, but it interacts with pests and diseases, collectively termed
biological disturbance agents (BDAs), to influence tree mortality (Kolb et al.. 2016). These factors,
drought and BDAs, account for much of the tree mortality throughout the region (Hessburg etal.. 1994).
3-4
DRAFT: Do Not Cite or Quote
-------
125"W
120°W
115"W
a;- '\
Seattle , - £y'3 ." -
/J;/k ' . i J*
• V ¦ ? . • . :/$& y "-r* .
W--A S.m H r N c T O X/
. . ,v. W
JEr^— I v j
m .:•* • .;• ••• ' ' -
.„>•• jit " . •
..JTv «r8E&; .VA
'it. , .
JVv.
•%v •
«E;*r
, _*>* * * * ™
O R F. G O \
IDAHO
TC 6
Fire
Crater,
Lake
NP
0
A Fire Locations
1990 WUI
2010 WUI
Highways
^ Ponderosa Pine
N E V A D 'A
• r. •'
¦:
.. - ' Mjjr
'¦ V"-"*V
„San Francisco
100 200
P Kings p.
. . panyon)^\ Sheep
'*' • NP < J Complex
V and
. , Rough Fires
WUI = wildland-urban interface.
Source: https://lemma.forestrv.oreaonstate.edu LEMMA (2020).
Note: Distribution and expansion of the WUI (Radeloff et al . 2018) in Washington, Oregon and California has increased from 41,318
to 50,856 km2 (23%) between 1990 to 2010 and is depicted in orange and red with approximately 8.3% of the WUI in the Ponderosa
Pine Region as of 2010. The growth of the WUI in the ponderosa pine region, from 3,072 km2 in 1990 to 4,211 km2 in 2010 (37%),
highlights how recent fire activity in dry fire prone forests impacts an expanding human population. Locations of the Timber Crater 6
and Rough fires are identified by red triangles.
Figure 3-2 The ponderosa pine region as defined by the distribution of Pinus
ponderosa in Oregon, Washington, and northern California
(94,000 km2, 11% of the land area shown) based on 2017 Gradient
Nearest Neighbor maps.
3-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
3.1.2
Historic Forest Conditions
Historically, forests in the ponderosa pine region consisted of multiaged stands with a structural
backbone of large old-growth trees that persisted because of resistance to frequent and extensive fires,
severe and prolonged droughts, and BDAs. Douglas fir (Pseudotsuga menziesii [Mirb. | Franco), grand fir
{Abies grcindis [Douglas ex D. Don] Lindl.), and white fir (Abies concolor [Gordon & Glend. | Lindl. ex
Hildebr) are common associates of ponderosa pine at higher elevations across the region (Safford and
Stevens. 2017; Franklin and Dvrness. 1988). while blending to pinyon and juniper woodlands at lower
elevations (Miller et al.. 2019). Presettlement forests throughout the region were characterized by open,
park-like stands of large-diameter trees with a few seedlings and saplings in the understory. Stands were
typically uneven-aged, with many stands containing a few large individual trees 400 to 600 years old
(Youngblood et al.. 2004; Arno et al.. 1997). Historic photos show the open character of old growth
ponderosa pine on the Klamath Indian Reservation in south-central Oregon in the early 20th century and
current old growth (Figure 3-3).
Source: left, BIA photo; right, photo; PA Beedlow,
Figure 3-3 Historic photo showing the open character of old growth
ponderosa pine resulting from high-frequency, low-intensity fire
on the Klamath Indian Reservation in south-central Oregon in the
1930s (left) and present-day ponderosa pine forest 10-15 years
after natural fire Ochoco National Forest, central Oregon (right).
3-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
3.1.3
Fire Influences on Forest Structure and Composition
Comparing forest conditions under a frequent low-severity fire regime with infrequent mixed- to
high-severity fire illustrates how an open and heterogeneous structure historically resulted in resistant
forest conditions over long time periods and across the ponderosa pine region. Patches of high-severity
fire historically were small and rare (Hcverdahl et al.. 2019; Merschel et al.. 2018; Agee. 1993) because
fire maintained low surface and canopy fuel loads (Johnston et al.. 2016). there was heterogeneity in
horizontal structure at fine (Churchill et al.. 2013) and coarse scales (Hcssburg et al.. 2005). and because
most trees were large and, consequently, fire-resistant (Hag m arm et al.. 2014. 2013).
3.1.4 Ecosystem Resilience/Resistance to Fire
Resilience is the capacity of an ecosystem to recover its essential characteristics following a
disturbance, whereas resistance is the property of an ecosystem to remain essentially unchanged when
disturbed. Resistance is often thought of as a component of resilience, but the two ecological processes
are distinct mechanisms that maintain the essential characteristics of an ecosystem including taxonomic
composition, structure, ecosystem function, and process rates (Holling. 1973). Within the ponderosa pine
region, open forest structure and fine scale heterogeneity predominated historically, and this conveyed
resistance to fire and other disturbances at fine scales (Koontz et al.. 2020). as well as broadly across
entire landscapes (Hcssburg et al.. 2005). However, after years of fire exclusion, in addition to logging
and livestock grazing, low intensity surface fires have been excluded in many areas, resulting in dense
stands that show both reduced resistance and resilience because of changes in species composition (Busse
et al.. 2009). Extreme severe fire is now much more likely to occur, reflecting decreased resistance.
3.1.5 Changes to Historic Fire Regimes
3.1.5.1 Land Management Practices
Forest ecosystems in the ponderosa pine region have undergone structural and functional changes
in the last 140 years since settlement (Hessburg and Agee. 2003). Heavy grazing in the late 1800s and
early 1900s, active fire suppression after the 1910 fires, and other land uses have disrupted the natural fire
regime in these ecosystems. Tree establishment and survival increased in the late 19th and early 20th
centuries resulting in denser forests characterized by increased homogeneity in horizontal structure,
increased canopy layering and connectivity, inter-tree competition, and canopy cover. This densification
combined with widespread logging of large and old fire-resistant trees (Naficv et al.. 2010; Hessburg and
Agee. 2003) contributed to mesophication—a shift from drought and fire-resistant shade intolerant
species to shade tolerant species adapted to competition but not as resistant to drought and fire (Nowacki
3-7
DRAFT: Do Not Cite or Quote
-------
1 and Abrams. 2008). Aggressive fire suppression since 1910 ensured that densification and mesophication
2 continued to the present conditions. The forests of today are the cumulative result of tree establishment
3 and growth versus mortality from drought, pests and diseases, fire, and land management [e.g., timber
4 harvesting, thinning, prescribed fire; Merschel et al. (2021)1.
5 3.1.5.2 Habitat Fragmentation from Human Population Growth
6 Wildfires pose the greatest risk to people in the wildland-urban interface (WUI)—the area where
7 houses are in or near wildland vegetation (Radcloff et al.. 2005). It is the fastest growing land use type in
8 the conterminous U.S. From 1990 to 2010 new houses in the WUI increased by 41%, from 30.8 to
9 43.4 million and land area increased 33%, from 581,000 to 770,000 km2 (Radcloff et al.. 2018). A more
10 current study estimates -49 million residential homes in the WUI, a number that has been increasing by
11 roughly 350,000 houses per year over the last two decades (Burke et al.. 2021). In the ponderosa pine
12 region of Oregon, Washington, and California (Figure 3-2) the land area of WUI increased by 37%
13 between 1990 and 2010 to 4,211 km2.
14 3.1.5.3 Invasive Species and Encroachment
15 Invasive species can establish permanency within ponderosa pine landscapes, but less frequently
16 than within other biomes. The conditions required for invasive species to dominate ponderosa pine
17 landscapes is complex. Many site features favor invasive plant suppression such as frequent small to
18 moderate fires, fire resistant trees, rugged terrain, and high elevation (Zouhar et al.. 2008). The
19 establishment of invasive species within ponderosa pine region has been minimal, likely due to "less
20 activity by humans, relatively intact shrub and tree canopies, [and] harsh climatcs"(Zouhar et al.. 2008).
21 Sites that do contain abundant levels of invasive plants have usually been disturbed first by human
22 activity (Keelev et al.. 2003; Moore and Gerlagh. 2001). Moderating fire intensity and targeting areas of
23 high severity for remediation may reduce post-fire invasive plant outbreaks (Svmstad et al.. 2014).
24 3.1.5.4 Weather and Changing Climatic Conditions
25 Topography, fire weather, and fuels have generally not limited chronic low-severity fire even in
26 relatively cool-moist environments where relatively fire susceptible Douglas fir and true fir (Abies spp.)
27 were common prior to fire exclusion (Hagmann et al.. 2019; Merschel et al.. 2018; Johnston et al.. 2016;
28 Heverdahl et al.. 2008). However, in the last 30 to 35 years, the West has seen a steady rise in the
29 intensity of wildfires as well as area burned, tied to human-caused climate change (Goss et al.. 2020).
30 Drought conditions occurred in 15 of 18 years during 2000-2017 as air temperature was increasing at
31 0.3°C/decade (Abatzoglou and Williams. 2016). The years from 2000-2018 contained the driest 19-year
3-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
period in western North America since the late 1500s (Williams et al.. 2020). Recent drought in western
North America was partially a product of natural variability, but its concurrence with anthropogenic
warming resulted in intensity and duration on par with the most extreme drought events since 800 CE
(Williams et al.. 2020). As climate continues to warm in the 21st century, drought and related impacts to
forests are projected to increase (Luce et al.. 2016). Further, increasing drought severity in combination
with climate-driven fungal pathogens and insect pests are exacerbating the fire hazard (Allen et al.. 2019)
Historic fire regimes were compiled locally, regionally, and nationally, with extensive records,
studies, and modelling. Changes observed in the past few decades are ongoing. Mapping and
characterizing the changes on a large scale would likely become obsolete before this could be completed.
Representation of projected future climate and expected fire regimes is important, but well beyond the
scope of this assessment.
3.2 Land Management Approaches to Reducing Fire Risks
Fire is an important tool to improve forest conditions, reduce fuels and decrease the threat of
large, high-severity wildland fires (Vaillant and Reinhardt. 2017). Fire managers have used natural
ignitions as a key component in the restoration of historical forest conditions and fuel loadings. The 2009
Policy Guidance (FEC. 2009) provided federal land management agencies and their state partners greater
flexibility to use natural ignitions to meet resource objectives through strategies other than full
suppression. Though some land managers have increasingly used wildfire to meet resource objectives
since the 1970s (Hunter and Robles. 2020; Collins et al.. 2007). managers more commonly resort to full
suppression strategies—a result of current land-management policies and local land use planning (Mever
et al.. 2015; Thompson et al.. 2013). However, land management and planning policies are beginning to
be revised to be more inclusive of using wildfire to meet resource objectives.
3.2.1 Need for Fuel Treatments
With increasing growth of the WUI, long-term suppression of wildfires and resulting forest
changes, and an era of increasing drought, wildfire has become a profound ecological and social issue in
forests formerly dominated by frequent low intensity wildfires (Moritz et al.. 2014). In response to
decades of fire suppression, resulting in a fire deficit, and increasing periods of drought, wildfires have
tended to become both larger and more severe. Compared to the area burned historically, there exists
today an enormous fire deficit in the region, especially for low-severity fire. The fire deficit extends to a
vast portion of dry forests of the conterminous U.S. (Kolden. 2019). Historically, open forests
characterized by larger trees was the most common structural condition in the ponderosa pine region
(Hagmann et al.. 2014. 2013). However, in some high elevation and alpine forests, humid temperate
forests, and shrublands, there may not be a deficit, and may indeed be a surplus of fire; this is beyond the
3-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
scope of the current assessment and the two case studies. Tree regeneration and growth in the absence of
frequent low-intensity fire in contemporary times has resulted in the loss of open resistant forest structure
and composition, sparse woodlands, and nonforest cover (Stcvcns-Rumann ct al.. 20IS). Wildfires in
these denser forests tend to be more severe and have a greater chance of converting forested areas to
different vegetation types, such as becoming shrublands in drier areas (Morcira et al.. 2020; Parks and
Abatzoglou. 2020).
3.2.2 Land Management Activities Affect Fire Behavior
Forest policy and management practices are slowly changing from predominantly fire
suppression to managing fire and associated risks to communities (Thompson et al.. 2018; Ingalsbcc and
Raja. 2015). Prior to Euro-American settlement, many dry forests of the western U.S. were maintained by
frequent low-to-moderate severity fires (i.e., cultural fires) often set by indigenous tribes (Hcssburg et al..
2005). Native American tribes used cultural fires to purposely burn forests and grasslands to promote
habitat diversity, environmental stability, predictability, and maintenance of ecotones, but perhaps the
most important effect may have been the lack of advanced fire suppression technology (Raish ct al..
2005). The absence of fire suppression allowed the natural progression of wildfires, both lightening and
human caused, across the landscape. Over the last 140 years, forests in the ponderosa pine region have
changed immensely and bear little resemblance to their presettlement condition. The original old-growth
ponderosa pine forests were once considered an endless resource to early pioneers and settlers, and the
vast "yellow pine" forests were utilized to fuel economic growth and the development of western North
America. Past and current land use activities along with active fire suppression eliminated natural surface
fires from these forests.
Fire exclusion over the last century in ponderosa pine forests has allowed for the buildup of
surface fuels on the forest floor and shrub cover and tree regeneration to increase. This buildup has
created "fuel ladders" where surface fuels are now connected to the overstory canopy by dense understory
and mid-story saplings and medium-sized trees. As a result, it is easier for surface fires to move up and
torch tree crowns and, under the right weather conditions and topographic setting, support active
crown-to-crown fire spread.
Removing accumulated surface fuels or targeting the removal of specific brush fuels (such as
bitterbrush [Purshia tridentata (Pursh) DC.] because of its high energy content), reduces flame lengths
making it more difficult to initiate torching of tree crowns. Also, the higher the base of tree crowns, the
more difficult it is for surface flames to torch tree crowns. Once a fire begins torching and moving up into
the canopy, the rate of spread and density of the crowns determines the likelihood of an actively moving
crown fire. Increasing the space between tree crowns reduces the ability for fire to spread from tree crown
to tree crown and allows a crown fire to transition back to a surface fire.
3-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Currently, the forest area being managed to reduce density, restore large ponderosa pine trees,
and reintroduce low-intensity, frequent fire is very small compared to the forest area experiencing
continued densification and succession. Fire is not being adopted into management practices at a scale
necessary to affect the fire deficit in the western U.S. and reduce the potential for more wildfire disasters;
the area burned by prescribed fire has actually decreased in the Pacific Northwest from 1998-2018
(Koldcn. 2019).
Land Managers have tools and methods to improve fire resilience and resistance in the ponderosa
pine region: these include reducing surface fuels, removing ladder fuels, leaving large, fire resistant trees,
and spacing tree crowns (see CHAPTER 7 for economic considerations). These conditions can be
achieved with a variety of methods including prescribed fire, use of naturally ignited wildfire to achieve
land management objectives, mechanical thinning, and biological control (Agcc and Skinner. 2005). The
use of multiple tools to reintroduce fire as a natural process in fire-prone forests has come to be known as
ecological forestry (Kelsev. 2019) and involves targeted removal of forest fuels plus implementation of
prescribed fire and managed wildfire where it is safe to do so (Figure 3-4).
3-11
DRAFT: Do Not Cite or Quote
-------
Fire-suppressed Forest
Ecologically managed Forest
... I. rL.I «. J i I..
Note: Fire-suppressed forest (left): Forests become dense with thickets of young trees and shrubs in the understory and are prone
to high-severity fires that can kill most of the trees. Ecologically managed forest (right): Strategic thinning the understory can reduce
overall fuel load so fire can safely be reintroduced to maintain healthy forests. (Kelsev. 2019).
©The Nature Conservancy, copyright permission pending.
Figure 3-4 Comparison of differences between a fire-suppressed and
ecologically managed forest.
3-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
3.2.2.1
Prescribed Fire
Prescribed fire is one of the most widely advocated management practices for meeting land
management goals and objectives and has a long and rich tradition rooted in indigenous and local
ecological knowledge. The scientific literature has repeatedly reported that prescribed lire is often the
most effective means to reduce fuels and wildfire hazard in order to restore sustainable ecological
functioning to fire-adapted ecosystems in the U.S. following a century of fire suppression (Kolden. 2019).
As defined in CHAPTER 2. a prescribed fire is "any fire intentionally ignited by management
actions in accordance with applicable laws, policies, and regulations to meet specific land or resource
management objectives" (U.S. EPA. 2020). Prescribed fire is used on the landscape to remove
accumulated surface fuels, consume slash generated from thinning activities, kill and thin out encroaching
trees in the understory, and rejuvenate herbaceous plants and shrubs (Sackctt and Haase. 1998; Walstad et
al.. 1990; Ffolliott and Thorud. 1977). Prescribed fire also scorches and kills lower branches of trees,
which, in the long run, results in lifting the canopy much like pruning, increasing the height from the
forest floor to the lower canopy and increasing fire resistance (Figure 3-5).
Source: Photo: PA Beedlow.
Figure 3-5
Prescribed fire in ponderosa pine, Deschutes National Forest.
3-13 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Periodic burning can prevent the development of fuel ladders and can be used to maintain
fire-resilient stands. However, in most forests of the ponderosa pine region prescribed fire is limited as an
initial treatment to reduce fuel loads because of heavy accumulations of surface and ladder fuels. In many
cases, other mechanical treatments are needed prior to prescribed fire to reduce fuels to a level that will
allow fire to be used without unnecessary damage to the forest.
The ability to control fire while minimizing human exposure to smoke and achieving the desired
ecological results are central goals of prescribed burning (Long et al.. 2018). On federal and most state
lands, prescribed fire is only used after thorough preplanning and only by highly trained and experienced
professionals (TSTWCG. 2017). Go/no-go checklists are used to determine compliance with policies and the
prescribed fire plan parameters. Based on these guidelines, prescribed fire is only implemented when
weather conditions are favorable, such as good smoke clearance conditions, moderate temperatures or
even dry fuel conditions that result in rapid consumption and ventilation, and an incoming cool/moist
weather pattern. Further, burning when smoke is not being produced by many wildfires over a large area
is favored to reduce the magnitude and duration of smoke exposure. In the much of the western U.S,
spring or after the start of fall rain provide good opportunities to manage wildfires due to environmental
conditions resulting in low-severity, shorter duration fires. In the ponderosa pine region, most prescribed
burns are conducted in the spring and late fall because personnel are available and weather conditions are
favorable.
3.2.2.2 Mechanical Treatments
Prescribed fire as a restoration tool, while often the cheapest to implement, is not practical in
many cases due to limited burning seasons, excessive accumulation of fuel due to fire exclusion, concerns
over potential undesirable fire effects, concerns about human exposure to smoke, visibility, and the
chance that a fire will escape and cause damage. Mechanical treatments can create a variety of
uneven-aged or even-aged stand structures depending on the desired treatment goals (e.g., fuel reduction
to meet fire behavior goals), wildlife habitat maintenance requirements (e.g., for endangered species), and
restoration of spatial and structural condition (Huggett et al.. 2008). They require equipment as well as
plans for disposal or utilization of significant quantities of small trees (Agee and Skinner. 2005).
How the residual wood from the thinning operations is disposed can have a large impact on
surface fuel availability with chipping or burning of the unusable tops of trees having the greatest impact
on reducing fuel loads. Mechanical treatments include activities, such as cutting and piling or stacking
trees, cutting and piling brush, pruning lower branches of trees, and creating fuel breaks based on
treatment objectives. Typically, mechanical treatments are emphasized in the WUI, while both
mechanical and fire treatments, alone or in combination, are emphasized in adjacent lands from which
wildland fire might spread into the WUI (Barros et al.. 2019).
3-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
3.2.2.3 Biological and Chemical Control
Biological control involves the intentional use of domestic animals, insects, nematodes, mites, or
pathogenic agents such as bacteria or fungus that can cause diseases in plants to reduce or eliminate
vegetation. Biological controls are used mostly to control invasive plants but can be used to control native
vegetation for fire management purposes. For instance, cattle may be used for target grazing in defined
areas for the creation of fuel breaks on rangelands and in some instances in forested lands.
In addition to natural agents, chemical agents, such as herbicides can also be used to kill or injure
plants to meet land management objectives. Herbicides can be categorized as selective or nonselective.
Selective herbicides kill only a specific type of plant, such as broad-leaved plants, while nonselective can
kill all plants. Only those herbicides approved for use can be used to manipulate vegetation to meet land
management goals and objectives.
3.2.2.4 Natural Ignitions
Remote forest areas as well as designated wilderness areas and national parks provide the best
opportunities for taking advantage of natural ignitions to reduce fuel loads. While fire managers may
choose to suppress fire inside or outside of wilderness areas, it is also federal policy to use fire "to protect,
maintain, and enhance resources and, as nearly as possible, be allowed to function in its natural ecological
role" (FEC. 2009). The very definition of wilderness in the Wilderness Act of 1964, as an area "managed
so as to preserve its natural conditions and which generally appears to have been affected primarily by the
forces of nature" aligns closely with federal fire policy and thus are often an excellent location to achieve
this goal. Moreover, wilderness area ignitions are often in steep, rugged terrain too dangerous for
firefighters to attack directly or limit in the technologies and equipment that can be deployed.
Agencies permit lightning-caused fires to play a natural ecological role to reduce the risks and
consequences of wildfire both within and outside wilderness areas. Fire managers seek to prevent fires
from causing damage to nearby communities. In pursuit of that goal, Minimum Impact Suppression
Techniques are implemented that cause the least alteration of the wilderness resource and the least
disturbance to the land surface, air quality, and visitor solitude (USFS. 2007). The initial response to
lightning caused wildfires is suppression if they occur in a landscape without a fire management plan, do
not meet certain conditions, or cannot achieve land management objectives.
3.3 Forest Characteristics for Timber Crater 6 (TC6) and the
Rough Fires
This assessment focuses on a quantitative analysis of the air quality and associated health impacts
of smoke by examining two case study fires (see CHAPTER 1). both of which occurred in the western
3-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
U.S.: (1) the Timber Crater 6 (TC6) Fire that occurred from July 21-26, 2018 in Oregon; and the
(2) Rough Fire that occurred from July 31 to October 1, 2015, in California. The Timber Crater 6 Fire
burned approximately 3,000 acres in Crater Lake National Park from July 21 to July 26, 2018. The Rough
fire burned in parts of the Sierra National Forest, Sequoia National Forest, and Kings Canyon National
Park between July 31 and October 1, 2015 (https://www.nps.gov/seki/learn/nature/rough-fire-interactive-
map.htm). burning approximately 150,000 acres. These fires were chosen as case studies because they
were on federal land previously subjected to fuel reduction management. Both areas are in dry forests
characteristic of the ponderosa pine region. The following sections describe the forest characteristics of
the case study areas. Additional details on the burn characteristics of each case study fire are provided in
CHAPTER 5.
3.3.1 Timber Crater 6 (TC6): Crater Lake National Park/Fremont-
Winema National Forest
Crater Lake National Park spans the divide of the Cascade Mountains in central Oregon. Forests
in the western part follow an elevational gradient from low elevation Douglas fir forests, to mixed conifer
forests dominated by red fir (Abies magnified A. Murr.), to mountain hemlock (Tsuga mertensiana
[Bong.] Carriere) dominated stands at high elevation (Forrcstcl et al.. 2017). The eastern portions of the
park are dominated by ponderosa pine grading into mix-conifer forests at higher elevations. Forests in
which ponderosa pine is a dominant tree principally occur up to 1,675 m elevation (A dam us et al.. 2013).
Ponderosa pine forests can contain a mixture of ponderosa pine, white fir (Abies concolor [Gord. &
Glend.] Lindl . ex. Hildebr.), and scattered sugar pine (Pinus lambertiana Douglas) and Douglas fir.
Where ponderosa pine shares dominance with these species, the forests can be called mixed conifer. On
the east side of the park, lodgepole pine (Pinus contorta var. murrayana [Grev. & Balfi] Engelm.) is a
common associate with ponderosa pine, and understory species may include the Great Basin shrubs,
including antelope bitterbrush (Purshia tridentata [Pursh] DC.), the montane chaparral shrub, greenleaf
manzanita (Arctostaphylos patula Greene), and a greater abundance of native grasses.
The Timber Crater 6 Fire started with a lightning strike in the northeast portion of the park on
July 15, 2018 and within 4 days spread into a nearby section of the Fremont-Winema National Forest
(Figure 3-6). The fire spread to property where the U.S. Forest Service had invested in fuel treatments
starting in the 1990s. Treatments included mowing and small tree thinning followed by pile burnings and
prescribed burning. The fire had the potential to burn about 81 km2, but because of the fuel treatments, it
was contained to 12.7 km2 (Dc lam arte r. 2019).
3-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
t
U22J
Elevation (meters)
ftif
~
Timber Crater 6
Burn Area
1//0
r |
Worst Case
• 1
Countcrlactual
1400
O»arnond l*kc
Njj.onjl P*ni
(•a)
Dumcnd Lj
Junction
P' M
Antelope
Desert
Figure 3-6 Timber Crater 6 (TC6) Fire, Crater Lake National Park, and
adjacent Fremont-Winema National Forest.
3.3.2 Rough Fire: Sierra and Sequoia National Forests and Kings
Canyon National Park
In the Sierra Nevada, especially on the west slope exposed to moisture off the ocean, much of the
area where ponderosa pine occurs is considered mixed conifer (Safford and Stevens. 2017). often referred
to in California as Yellow Pine Mixed Conifer. The Rough fire burned a substantial area of the Kings
Canyon, one of the deepest canyons in California, in a footprint that spanned an elevational gradient of
more than 2,100 m, from -300 m above sea level (ASL) to just under 2,500 m ASL (Figure 3-7). On the
western side of the southern Sierra Nevada Mountain range, which is exposed to storms and prevailing
winds coming off the Pacific Ocean, this area of the canyon encompasses a steep precipitation gradient,
with a distinct Mediterranean annual pattern allowing for high productivity, but also requiring robust
summer drought tolerance. This precipitation gradient and moisture availability pattern in turn drives a
diverse range of vegetation assemblages and growth strategies, from grassland and oak woodlands in the
lower elevations (-300 to 1,200 m ASL) to highly productive yellow pine (ponderosa pine) mixed conifer
3-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
(including giant sequoia [Sequoiadendron giganteum (Lindl.)] Buchholz) in the mid-elevations (1,200 to
2,100 m ASL) to red fir and lodgepole pine typical of boreal forest in the higher elevations (over 2,100 m
ASL) of the Rough Fire footprint. Pure ponderosa pine stands occur in the lower to mid-elevations but
comprise a relatively small fraction (-7%) of the total area burned by the Rough Fire. However,
ponderosa pine (and its higher-elevation cousin, the Jeffrey pine) is often an important component of the
mixed conifer zone, which comprises a majority (-33%) of the forested area burned by the Rough Fire
(Huang et al.. 2018; Safford and Stevens. 2017).
Throughout this complex and highly heterogeneous matrix of vegetation types (Figure 3-8), fuels
generally increase with elevation from under 2,000 mg/km2 in the lower elevation oak woodlands and
grasslands to over 18,000 mg/km2 in the mixed conifer and upper montane vegetation of the mid-upper
elevations. The overall amount of those fuels was also likely enhanced by an unprecedented mortality
event, wherein about 30% of the area burned by the 2015 Rough Fire had experienced at least 10%
canopy cover loss due to tree mortality from 2011-2014 (and likely much more in some of the ponderosa
pine dominated stands). Ponderosa pine and mixed conifer stands in the lower to mid- elevations in
particular appeared to have suffered the most uniformly severe mortality (Fettig et al.. 2019; Paz-Kagan et
al.. 2017). Though the proximate cause of this mortality was likely a bark beetle infestation that
opportunistically attacked trees weakened by several years of antecedent drought from 2012 through 2015
(Rcstaino et al.. 2019). these lower elevations also experienced chronically phytotoxic levels of ozone and
nitrogen deposition for decades [e.g., Yates et al. (2020); Panek et al. (2013); Peterson et al. (1991)1.
which likely contributed to their susceptibility to those beetles and the drought(Joncs et al.. 2004). By the
time the Rough Fire burned in 2015, this mortality event was in the "red needle" phase mortality event,
wherein the tree canopy consisted of dead or dry needles and twigs, which contributed to increased crown
fire potential and higher forest fire severity (Stephens et al.. 2018; USFS. 2016). Torching potential and
ember production were also thought to have occurred in areas affected by tree mortality (Reiner. 2017). In
the years post-fire, dead trees not consumed in the fire decay and the coarser "gray phase" fuels fall to the
ground increasing fuel loads potentially contributing to larger scale, "mass fire" events similar the more
recent 2020 Creek Fire (Stephens et al.. 2018).
3-18
DRAFT: Do Not Cite or Quote
-------
10 Kilometers
Elevation (meters)
—1 3550
I I 2015 Rough Fire
' ' Largest Perimeter
I 1 2013 Boulder Creek
' ' Prescribed Fire
~ 2010 Sheep
Complex Fire
Pine
Hat
Lake
Kings Canyon,
Scuroes: tsn Alitous DS.HJSGS.
. "i Sijkswawrstaat GS^ Geolsrd
I ' ;AO, NOAA? 1KGS OprnVrrrtM.ip rnnlnr»..to
f Y Ci;-lo:?Mniiv liy Viv in
Figure 3-7 Rough Fire: Sierra and Sequoia National Forests and Kings
Canyon National Park.
3-19
DRAFT: Do Not Cite or Quote
-------
&w~
7 Kilometers
Legend
Forest Type in
the Rough Fire
Footprint (2011)
Value
183-Western
Juniper
184-Juniper
Woodland
185-Pinyon
Juniper
201-Douglas
1 Fir
| 221-Ponder...
I Pine
| 222-lncense
I Cedar
l 261-White Fir
1 270-Mtn
' Hemlock
| 281-Lodgep..
I pine
301-Western
Hemlock
| 342-Giant
I Sequoia
| 361-Knobco..
I Pine
I 365Foxtail-b...
I Pine
| 367-Whiteb...
I Pine
I 371-California
Mixed Conifer
I 921-Gray Pine
922-California
Black Oak
(Woodland)
—j 923-0 reg on
White Oak
924-BlueOak
I 925-Decidu...
I Oak woodland
i 931-coast Live
I Oak
932-Canyon
Live Oak
941-Tan oak
942-Califomia
Laurel
i 951-Pacific
I Madrone
953-M ountain
Brush
Woodland
997-FVS
] Other
Hardwoods
999-Not
Stocked
Note: based on Forest Inventory and Analysis (FIA) and satellite data, see Huang et al. (2018), copyright permission pending.
Figure 3-8 Tree species maps for the area of the Rough Fire.
1
2 3.4 Conclusions
3 From an ecological perspective, restoration of frequent low-severity fire is an essential to
4 restoring sustainable ecosystems in the dry forests of the ponderosa pine region. However, extensive
5 densification and mesophication of these dry ecosystems due to land management practices in the 20th
6 century, followed by an increase in wildland fire frequency and severity, drought, invasive species, pests
7 and diseases, as well as the rapid expansion of the WUI pose serious ecological and socioeconomic
8 challenges to human wellbeing in the 21st century. Key to living with fire in the ponderosa pine region is
9 an all-lands and all ownerships approach to forest management planning that helps determine where
3-20
DRAFT: Do Not Cite or Quote
-------
1 prescribed fire and mechanical treatments are appropriate and should be prioritized, and where fires can
2 safely be allowed to burn (Dunn et al.. 2020).
3-21
DRAFT: Do Not Cite or Quote
-------
3.5 References
Abatzoglou. JT: Williams. AP. (2016). Impact of anthropogenic climate change on wildfire across western US
forests. Proc Natl Acad Sci USA 113: 11770-11775. http://dx.doi.org/10.1073/pnas. 1607171113
Adamus. PR: Odion. DC: Jones. GV: Groshong. LC: Reid. R. (2013). Crater Lake National Park: Natural
resource condition assessment. (Natural Resource Report NPS/NRSS/WRD/NRR—2013/724). Fort Collins,
CO: U.S. Department of Interior, National Park Service.
https://irma.nps.gov/DataStore/Reference/Profile/2204346
Agee. JK. (1993). Fire ecology of Pacific Northwest forests. Washington, DC: Island Press.
Agee. JK: Skinner. CN. (2005). Basic principles of forest fuel reduction treatments. For Ecol Manage 211: 83 -
96. http://dx.doi.Org/10.1016/i.foreco.2005.01.034
Allen. I: Chhin. S: Zhang. J. (2019). Fire and forest management in montane forests of the northwestern states
and California, USA [Review]. Fire 2: 17. http://dx.doi.org/10.3390/fire2020Q17
Arno. SF: Smith. HY: Krebs. MA. (1997). Old growth ponderosa pine and western larch stand structures:
Influences of pre-1900 fires and fire exclusion. (Research Paper INT-RP-495). Ogden, UT: U.S. Department
of Agriculture, Forest Service, Intermountain Research Station.
https://www.fs.fed.us/rm/pubs/rmrs gtr292/int rp495.pdf
Baker. WL: Ehle. D. (2001). Uncertainty in surface-fire history: The case of ponderosa pine forests in the
western United States. Can J For Res 31: 1205-1226. http://dx.doi.org/10.1139/x01-046
Barros. AMG: Ager. AA: Day. MA: Palaiologou. P. (2019). Improving long-term fuel treatment effectiveness in
the National Forest System through quantitative prioritization. For Ecol Manage 433: 514-527.
http://dx.doi.Org/10.1016/i.foreco.2018.10.041
Burke. M: Driscoll. A: Heft-Neal. S: Xue. JN: Burnev. J: Wara. M. (2021). The changing risk and burden of
wildfire in the United States. Proc Natl Acad Sci USA 118: e2011048118.
http://dx.doi.org/10.1073/pnas.2011048118
Busse. MP: Cochran. PH: Hopkins. WE: Johnsoa WH: Riegel. GM: Fiddler. GO: Ratcliff. AW: Shestak. CJ.
(2009). Developing resilient ponderosa pine forests with mechanical thinning and prescribed fire in central
Oregon's pumice region. Can J For Res 39: 1171-1185. http://dx.doi.org/10.1139/X09-044
Churchill. DJ: Larson. AJ: Dahlgreen. MC: Franklia JF: Hessburg. PF: Lutz. JA. (2013). Restoring forest
resilience: From reference spatial patterns to silvicultural prescriptions and monitoring. For Ecol Manage
291: 442-457. http://dx.doi.Org/10.1016/i.foreco.2012.ll.007
Collins. BM: Kelly. M: van Wagtendonk. JW: Stephens. SL. (2007). Spatial patterns of large natural fires in
Sierra Nevada wilderness areas. Landsc Ecol 22: 545-557. http://dx.doi.org/10.1007/sl0980-006-9Q47-5
Daly. C: Halbleib. M: Smith. JI: Gibson. WP: Doggett. MK: Taylor. GH: Curtis. J: Pasteris. PP. (2008).
Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous
United States. Int J Climatol 28: 2031-2064. http://dx.doi.org/10.1002/ioc. 1688
Delamarter. C. (2019). 2018 Timber Crater 6 fire proves how fuel treatments make a difference. Available online
at https://www.kdrv.com/content/news/Fuel-treatments-prevent-fire-from-spreading-to-more-than-6-times-
originallv--510791421.html (accessed January 21, 2021).
Dunn. CJ: O'Connor. CD: Abrams. J: Thompson. MP: Calkia DE: Johnston. JD: Stratton. R: Gilbertson-Dav. J.
(2020). Wildfire risk science facilitates adaptation of fire-prone social-ecological systems to the new fire
reality. Environ Res Lett 15: 025001. http://dx.doi.org/10.1088/1748-9326/ab6498
Farris. CA: Baisan. CH: Falk. DA: Yool. SR: Swetnam. TW. (2010). Spatial and temporal corroboration of a
fire-scar-based fire history in a frequently burned ponderosa pine forest. Ecol Appl 20: 1598-1614.
http://dx.doi.org/10.1890/09-1535.1
3-22
DRAFT: Do Not Cite or Quote
-------
FEC (Fire Executive Council). (2009). Guidance for Implementation of Federal Wildland Fire Management
Policy. Washington, DC: U.S. Department of Agriculture, U.S. Department of the Interior.
https://www.nifc.gov/policies/policies documents/GIFWFMP.pdf
Fettig. CJ: Mortenson. LA: Bulaon. BM: Foulk. PB. (2019). Tree mortality following drought in the central and
southern Sierra Nevada, California, US. For Ecol Manage 432: 164-178.
http://dx.doi.Org/10.1016/i.foreco.2018.09.006
Ffolliott. PF: Thorud. DB. (1977). Water yield improvement by vegetation management. Water Resour Bull 13:
563-571. http://dx.doi.org/10.1111/i. 1752-1688.1977.tb05568.x
Forrestel. AB: Andrus. RA: Fry. PL: Stephens. SL. (2017). Fire history and forest structure along an elevational
gradient in the southern Cascade Range, Oregon, USA. Fire Ecology 13: 1-15.
http://dx.doi.org/10.4996/fireecology.1301001
Franklia JF: Dvrness. CT. (1988). Natural vegetation of Oregon and Washington. Corvallis, OR: Oregon State
University Press, https://ir.librarv.oregonstate.edu/concern/technical reports/2v23wl3z
Goss. M: Swain. PL: Abatzoglou. JT: Sarhadi. A. li: Koldea CA: Williams. AP: Diffenbaugh. NS. (2020).
Climate change is increasing the likelihood of extreme autumn wildfire conditions across California. Environ
Res Lett 15: 094016. http://dx.doi.org/10.1088/1748-9326/ab83a7
Hagmann. RK: Franklin. JF: Johnson. KN. (2013). Historical structure and composition of ponderosa pine and
mixed-conifer forests in south-central Oregon. For Ecol Manage 304: 492-504.
http://dx.doi.Org/10.1016/i.foreco.2013.04.005
Hagmann. RK: Franklin. JF: Johnson. KN. (2014). Historical conditions in mixed-conifer forests on the eastern
slopes of the northern Oregon Cascade Range, USA. For Ecol Manage 330: 158-170.
http ://dx. doi.org/10.1016/i .foreco.2014.06.044
Hagmana RK: Merschel. AG: Reillv. MJ. (2019). Historical patterns of fire severity and forest structure and
composition in a landscape structured by frequent large fires: Pumice Plateau ecoregion, Oregon, USA.
Landsc Ecol 34: 551-568. http://dx.doi.org/10.1007/sl0980-019-0Q791-l
Hann. WJ: Shliskv. A: Havlina. P: Schon. K: Barrett. SW: PeMeo. TE: Pohl. K: Menakis. JP: Hamilton. P:
Jones. J: Levesaue. M: Frame. CK. (2008). Interagency Fire Regime Condition Class (FRCC) guidebook,
version 1.3.0. Hann, WJ; Shlisky, A; Havlina, P; Schon, K; Barrett, SW; PeMeo, TE; Pohl, K; Menakis, JP;
Hamilton, P; Jones, J; Levesque, M; Frame, CK. https://www.frames.gov/catalog/7793
Hessburg. PF: Agee. JK. (2003). An environmental narrative of Inland Northwest United States forests, 1800-
2000. For Ecol Manage 178: 23-59. http://dx.doi.org/10.1016/S0378-1127(03)00052-5
Hessburg. PF: Agee. JK: Franklia JF. (2005). Pry forests and wildland fires of the inland Northwest USA:
Contrasting the landscape ecology of the pre-settlement and modern eras. For Ecol Manage 211: 117-139.
http://dx.doi.Org/10.1016/i.foreco.2005.02.016
Hessburg. PF: Mitchell. RG: Filip. GM. (1994). Historical and current roles of insects and pathogens in eastern
Oregon and Washington forested landscapes. (General Technical Report PNW-GTR-327). Portland, OR:
Pacific Northwest Research Station.
Heverdahl. EK: Loehmaa RA: Falk. PA. (2019). A multi-century history of fire regimes along a transect of
mixed-conifer forests in central Oregon, USA. Can J For Res 49: 76-86. http://dx.doi.org/10.1139/cifr-2018-
0193
Heverdahl. EK: McKenzie. P: Paniels. LP: Hessl. A: Littell. JS: Mantua. NJ. (2008). Climate drivers of
regionally synchronous fires in the inland Northwest (1651-1900). International Journal of Wildland Fire 17:
40-49. http://dx.doi.org/10.1071/WF07024
Holling. CS. (1973). Resilience and stability of ecological systems. Annu Rev Ecol Evol Systemat 4: 1-23.
http://dx.doi.org/10.1146/annurev.es.04.110173.000245
Huang. SL: Ramirez. C: McElhanev. M: Evans. K. (2018). F3: Simulating spatiotemporal forest change from
field inventory, remote sensing, growth modeling, and management actions. For Ecol Manage 415-416: 26-
37. http://dx.doi.Org/10.1016/i.foreco.2018.02.026
3-23
PRAFT: Po Not Cite or Quote
-------
Huggett. RJ. Jr: Abt. KL: Shepperd. W. (2008). Efficacy of mechanical fuel treatments for reducing wildfire
hazard. Forest Pol Econ 10: 408-414. http://dx.doi.Org/10.1016/i.forpol.2008.03.003
Hunter. ME: Robles. MP. (2020). Tamm review: The effects of prescribed fire on wildfire regimes and impacts:
A framework for comparison [Review]. For Ecol Manage 475: 118435.
http://dx.doi.org/10.1016/i.foreco.2020.118435
Ingalsbee. T: Raia. U. (2015). The rising costs of wildfire suppression and the case for ecological fire use. In DA
DellaSalla; CT Hanson (Eds.), The ecological importance of mixed-severity fires: Nature's phoenix (pp. 348-
371). Amsterdam, Netherlands: Elsevier. http://dx.doi.org/10.1016/B978-0-12-802749-3.00Q12-8
Johnston. JD. (2017). Forest succession along a productivity gradient following fire exclusion. For Ecol Manage
392: 45-57. http://dx.doi.Org/10.1016/i.foreco.2017.02.050
Johnston. JD: Bailey. JD: Dunn. CJ. (2016). Influence of fire disturbance and biophysical heterogeneity on pre-
settlement ponderosa pine and mixed conifer forests. Ecosphere 7: e01581.
http://dx.doi.org/10.1002/ecs2.1581
Jones. ME: Paine. TP: Fena ME: Poth. MA. (2004). Influence of ozone and nitrogen deposition on bark beetle
activity under drought conditions. For Ecol Manage 200: 67-76.
http ://dx. doi.org/10.1016/i .foreco.2004.06.003
Keelev. JE: Lubin. D: Fotheringham. CJ. (2003). Fire and grazing impacts on plant diversity and alien plant
invasions in the southern Sierra Nevada [Review]. Ecol Appl 13: 1355-1374. http://dx.doi.org/10.1890/02-
5002
Kelsev. R. (2019). Wildfires and forest resilience: The case for ecological forestry in the Sierra Nevada.
Sacramento, CA: The Nature Conservancy.
https://www.scienceforconservation.org/assets/downloads/WildfireForestResilience 2019 Kelsev 2.pdf
Kolb. TE: Fettig. CJ: Avres. MP: Bentz. BJ: Hicke. JA: Mathiasen. R: Stewart. JE: Weed. AS. (2016). Observed
and anticipated impacts of drought on forest insects and diseases in the United States. For Ecol Manage 380:
321-334. http://dx.doi.Org/10.1016/i.foreco.2016.04.051
Kolden. CA. (2019). We're not doing enough prescribed fire in the Western United States to mitigate wildfire
risk. Fire 2: 30. http://dx.doi.org/10.3390/fire202003Q
Koontz. MJ: North. MP: Werner. CM: Fick. SE: Latimer. AM. (2020). Local forest structure variability
increases resilience to wildfire in dry western U.S. coniferous forests. Ecol Lett 23: 483-494.
http://dx.doi.org/10. Ill 1/ele. 13447
LEMMA (Landscaping Ecology, Modeling, Mapping & Analysis). (2020). Landscaping Ecology, Modeling,
Mapping & Analysis (LEMMA): GNN maps and data. Available online at
https://lemma.forestrv.oregonstate.edu/data (accessed January 22, 2021).
LF (Landfire). (2012). Fire regime groups (FRG). Available online at
https://landfire.gov/geoareasmaps/2012/CQNUS FRG cl2.pdf (accessed February 4, 2021).
Long. JW: Tarnav. LW: North. MP. (2018). Aligning smoke management with ecological and public health
goals. J Forest 116: 76-86. http://dx.doi.org/10.5849/iof.16-042
Luce. CH: Vose. JM: Pederson. N: Campbell. J: Millar. C: Kormos. P: Woods. R. (2016). Contributing factors
for drought in United States forest ecosystems under projected future climates and their uncertainty. For Ecol
Manage 380: 299-308. http://dx.doi.Org/10.1016/i.foreco.2016.05.020
McKenzie. D: Hessl. AE: Peterson. PL: Agee. JK: Lehmkuhl. JF: Kellogg. LKB: Kernan. J. (2004). Fire and
climatic variability in the inland Pacific Northwest: Integrating science and management. Corvallis, OR: U.S.
Pepartment of Agriculture, Forest Service, Pacific Northwest Research Station.
https://cig.uw.edu/publications/fire-and-climatic-variabilitv-in-the-inland-pacific-northwest-integrating-
science-and-management/
Merschel. A: Beedlow. PA: Shaw. PC: Woodruff. PR: Lee. H: Cline. S: Comeleo. R: Hagmann. KR: Reillv.
MJ. (2021). An ecological perspective on living with fire in ponderosa pine forests of Oregon and
Washington: Resistance, gone but not forgotten. Trees For People.
3-24
PRAFT: Po Not Cite or Quote
-------
Merschel. AG: Heverdahl. EK: Spies. TA: Loehman. RA. (2018). Influence of landscape structure, topography,
and forest type on spatial variation in historical fire regimes, Central Oregon, USA. Landsc Ecol 33: 1195-
1209. http://dx.doi.org/10.1007/sl0980-018-Q656-6
Mever. MP: Roberts. SL: Wills. R: Brooks. M: Winford. EM. (2015). Principles of effective USA federal fire
management plans. Fire Ecology 11: 59-83. http://dx.doi.org/10.4996/fireecology. 1102059
Miller. RF: Chambers. JC: Evers. L: Williams. CJ: Snyder. KA: Roundv. BA: Pierson. FB. (2019). The ecology,
history, ecohydrology, and management of pinyon and juniper woodlands in the Great Basin and Northern
Colorado Plateau of the western United States. (Gen. Tech. Rep. RMRS-GTR-403). Fort Collins, CO: U.S.
Department of Agriculture, Forest Service, Rocky Mountain Research Station.
https://www.fs.usda.gov/treesearch/pubs/59333
Moore. PE: Gerlagh. JD. (2001). Exotic species threat assessment in Sequoia, Kings Canyon, and Yosemite
national parks. In D Harmon (Ed.), Crossing boundaries in park management: Proceedings of the 11th
Conference on Research and Resource Management in Parks and on Public Lands (pp. 96-103). Hancock,
MI: The George Wright Society, http://www.georgewright.org/17moore.pdf
Moreira. F: Ascoli. D: Safford. H: Adams. MA: Moreno. JM: Pereira. JMC: Catrv. FX: Armesto. J: Bond. W:
Gonzalez. ME: Curt. T: Koutsias. N: McCaw. L: Price. O: Pausas. JG: Rigolot. E: Stephens. S: Tavsanoglu.
C: Valleio. VR: Van Wilgen. BW: Xanthopoulos. G: Fernandes. PM. (2020). Wildfire management in
Mediterranean-type regions: Paradigm change needed [Editorial]. Environ Res Lett 15: 011001.
http://dx.doi.org/10.1088/1748-9326/ab541e
Moritz. MA: Batllori. E: Bradstock. RA: Gill. AM: Handmer. J: Hessburg. PF: Leonard. J: McCaffrey. S: Odion.
DC: Schoennagel. T: Syphard. AD. (2014). Learning to coexist with wildfire [Review]. Nature 515: 58-66.
http://dx.doi.org/10.1038/naturel3946
Munger. TT. (1917). Western yellow pine in Oregon. Washington, DC: U.S. Department of Agriculture.
http://dx.doi.org/10.5962/bhl.title.108047
Naficv. C: Sala. A: Keeling. EG: Graham. J: DeLuca. TH. (2010). Interactive effects of historical logging and
fire exclusion on ponderosa pine forest structure in the northern Rockies. Ecol Appl 20: 1851-1864.
http://dx.doi.org/10.1890/09-0217.1
Nowacki. GJ: Abrams. MP. (2008). The demise of fire and "Mesophication" of forests in the eastern United
States. Bioscience 58: 123-138. http://dx.doi.org/10.1641/B580207
NWCG (National Wildfire Coordinating Group). (2017). Interagency Prescribed Fire Planning and
Implementation Procedures Guide. (PMS 484). National Wildfire Coordinating Group, Fuels Management
Committee, https://www.nwcg.gov/publications/484
Panek. J: Saah. D: Esperanza. A: Bvtnerowicz. A: Fraczek. W: Cisneros. R. (2013). Ozone distribution in remote
ecologically vulnerable terrain of the southern Sierra Nevada, CA. Environ Pollut 182: 343-356.
http://dx.doi.Org/10.1016/i.envpol.2013.07.028
Parks. SA: Abatzoglou. JT. (2020). Warmer and drier fire seasons contribute to increases in area burned at high
severity in western US forests from 1985 to 2017. Geophys Res Lett 47: e2020GL089858.
http://dx.doi.org/10.1029/2020GLQ89858
Paz-Kagan. T: Brodrick. PG: Vaughn. NR: Das. AJ: Stephenson. NL: Nvdick. KR: Asner. GP. (2017). What
mediates tree mortality during drought in the southern Sierra Nevada? Ecol Appl 27: 2443-2457.
http://dx.doi.org/10.1002/eap. 1620
Petersoa PL: Arbaugh. MJ: Robinson. LJ. (1991). Regional growth changes in ozone-stressed ponderosa pine
(Pinus ponderosa) in the Sierra Nevada, California, USA. Holocene 1: 50-61.
http://dx.doi.org/10.1177/0959683691001001Q7
Radeloff. VC: Hammer. RB: Stewart. SI: Fried. JS: Holcomb. SS: McKeefrv. JF. (2005). The wildland-urban
interface in the United States. Ecol Appl 15: 799-805. http://dx.doi.org/10.1890/04-1413
3-25
DRAFT: Do Not Cite or Quote
-------
Radeloff. VC: Helmers. DP: Kramer. HA: Mockria MH: Alexandre. PM: Bar-Massada. A: Butsic. V:
Hawbaker. TJ: Martinuzzi. S: Syphard. AD: Stewart. SI. (2018). Rapid growth of the US wildland-urban
interface raises wildfire risk. Proc Natl Acad Sci USA 115: 3314-3319.
http://dx.doi.org/10.1073/pnas. 1718850115
Raish. C: Gonzalez-Caban. A: Condie. CJ. (2005). The importance of traditional fire use and management
practices for contemporary land managers in the American Southwest. Glob Env Chang Part B: Env Hazards
6: 115-122. http://dx.doi.Org/10.1016/i.hazards.2005.10.004
Reiner. A. (2017). Fire behavior in tree mortality. Washington, DC: U.S. Department of Agriculture, Forest
Service.
https://www.fs.fed.us/adaptivemanagement/reports/fbat/FIRESCOPE treemortalitv Nov2017 share.pdf
Restaino. C: Young. D: Estes. B: Gross. S: Wuenschel. A: Mever. M: Safford. H. (2019). Forest structure and
climate mediate drought-induced tree mortality in forests of the Sierra Nevada, USA. Ecol Appl 29: e01902.
http://dx.doi.org/10.1002/eap.1902
Rothermel. RC: Deeming. JE. (1980). Measuring and interpreting fire behavior for correlation with fire effects.
(Gen. Tech. Rep. INT-93). Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Forest
and Range Experiment Station, https://www.frames.gov/catalog/6153
Ryan. KC: Qpperman. TS. (2013). LANDFIRE — A national vegetation/fuels data base for use in fuels
treatment, restoration, and suppression planning. For Ecol Manage 294: 208-216.
http://dx.doi.Org/10.1016/i.foreco.2012.ll.003
Sackett. SS: Haase. SM. (1998). Two case histories for using prescribed fire to restore ponderosa pine
ecosystems in northern Arizona. In TL Pruden; LA Brennan (Eds.), Fire in ecosystem management: Shifting
the paradigm from suppression to prescription (pp. 380-389). Tallahassee, FL: Tall Timbers Research
Station, https://www.fs.usda.gov/treesearch/pubs/23291
Safford. HP: Stevens. JT. (2017). Natural range of variation for yellow pine and mixed-conifer forests in the
Sierra Nevada, southern Cascades, and Modoc and Inyo National Forests, California, USA. (PSW-GTR-
256). Albany, CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest Research Station.
https://www.fs.fed.us/psw/publications/documents/psw gtr256/psw gtr256.pdf
Stephens. SL: Collins. BM: Fettig. CJ: Finney. MA: Hoffman. CM: Knapp. EE: North. MP: Safford. H:
Wavman. RB. (2018). Drought, tree mortality, and wildfire in forests adapted to frequent fire. Bioscience 68:
77-88. http://dx.doi.org/10.1093/biosci/bixl46
Stevens-Rumann. CS: Kemp. KB: Higuera. PE: Harvey. BJ: Rother. MT: Donato. DC: Morgan. P: Veblen. TT.
(2018). Evidence for declining forest resilience to wildfires under climate change. Ecol Lett 21: 243-252.
http://dx.doi.org/10. Ill 1/ele. 12889
Svmstad. AJ: Newton. WE: Swansoa DJ. (2014). Strategies for preventing invasive plant outbreaks after
prescribed fire in ponderosa pine forest. For Ecol Manage 324: 81-88.
http://dx.doi.Org/10.1016/i.foreco.2014.04.022
Taylor. AH: Skinner. CN. (1998). Fire history and landscape dynamics in a late-successional reserve, Klamath
Mountains, California, USA. For Ecol Manage 111: 285-301. http://dx.doi.org/10.1016/S0378-
1127(98)00342-9
Thompsoa MP: MacGregor. DG: Dunn. CJ: Calkin. DE: Phipps. J. (2018). Rethinking the wildland fire
management system. J Forest 116: 382-390. http://dx.doi.org/10.1093/iofore/Ivy020
Thompson. MP: Scott. J: Kaiden. JD: Gilbertson-Dav. JW. (2013). A polygon-based modeling approach to
assess exposure of resources and assets to wildfire. Natural Hazards 67: 627-644.
http://dx.doi.org/10.1007/sllQ69-013-0593-2
U.S. EPA. Definitions: Prescribed fire. 40 CFR § 50. l(m) (2020). https://www.govinfo.gov/app/details/CFR-
2020-title40-vol2/CFR-2020-title40-vol2-sec50-l
3-26
DRAFT: Do Not Cite or Quote
-------
USFS (U.S. Forest Service). (2007). Forest Service Manual (FSM) 2300: Recreation, wilderness, and related
resource management. Chapter 2320: Wilderness management. Section 2324.23: Fire management activities.
(FSM 2320). Washington, DC: U.S. Department of Agriculture, Forest Service.
https://www.fs.fed.us/im/directives/fsm/2300/2320.doc
USFS (U.S. Forest Service). (2016). 2015 Rough Fire Sierra and Sequoia National Forests and Kings Canyon
National Park: Fire Behavior Assessment Team summary report. Washington, DC: U.S. Department of
Agriculture, Forest Service.
https://www.fs.fed.us/adaptivemanagement/reports/fbat/2015RoughFire FBAT Summary Final 2Mar2016.
pdf
Vaillant. NM: Reinhardt. ED. (2017). An evaluation of the forest service hazardous fuels treatment program—
Are we treating enough to promote resiliency or reduce hazard? J Forest 115: 300-308.
http://dx.doi.org/10.5849/iof.16-067
Walstad. JD: Radosevich. SR: Sandberg. DV. (1990). Natural and prescribed fire in Pacific Northwest forests.
Corvallis, OR: Oregon State University Press.
Williams. AP: Cook. ER: Smerdon. JE: Cook. BI: Abatzoglou. JT: Bolles. K: Baek. SH: Badger. AM: Livneh.
B. (2020). Large contribution from anthropogenic warming to an emerging North American megadrought.
Science 368: 314-318. http://dx.doi.org/10.1126/science.aaz9600
Yates. EL: Iraci. LT: Tarnav. LW: Burlev. JD: Parworth. C: Rvoo. JM. (2020). The effect of an upwind non-
attainment area on ozone in California's Sierra Nevada Mountains. Atmos Environ 230: 117426.
http://dx.doi.Org/10.1016/i.atmosenv.2020.117426
Youngblood. A: Max. T: Coe. K. (2004). Stand structure in eastside old-growth ponderosa pine forests of
Oregon and northern California. ForEcol Manage 199: 191-217.
http://dx.doi.Org/10.1016/i.foreco.2004.05.056
Zouhar. K: Smith. JK: Sutherland. S: Brooks. ML. (2008). Wildland fire in ecosystems: Fire and nonnative
invasive plants. (RMRS-GTR-42-vol. 6). Ogden, UT: U.S. Department of Agriculture, Forest Service, Rocky
Mountain Research Station. http://dx.doi.org/10.2737/RMRS-GTR-42-V6
3-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
CHAPTER 4 AIR QUALITY MONITORING OF
WILDLAND FIRE SMOKE
4.1 Introduction
Wildland fires (prescribed fire and wildfire) can produce significant air pollution emissions which
may pose health risks to fire crews, first responders, and nearby residents, as well as downwind
populations (see CHAPTER 5. CHAPTER 6. CHAPTER 7. CHAPTER 8). Wildland fire smoke is a
complex mixture of thousands of different organic, inorganic, gaseous, and particulate phase compounds
(Reisen ct al.. 2015). The primary constituents of emitted wildland fire smoke that impact air quality are
fine particulate matter (PM) with a nominal mean aerodynamic diameter less than or equal to 2.5 |im
(PM2 5), carbon monoxide (CO), oxides of nitrogen, and volatile organic compounds Urbanski (2014).
The secondary photochemical formation of PM2 5 and ozone (O3) from wildland fire emission precursors
is also a concern (Liu et al.. 2017; Alvarado et al.. 2015; Jaffe and Wigder. 2012).
The U.S. Environmental Protection Agency (U.S. EPA) and its partners at state, local and tribal
monitoring agencies manage several routine regulatory monitoring networks. Each of these ambient air
monitoring networks have regulatory requirements and policy objectives that dictate decisions on the
location and pollutants measured at each site. The implementation of the network objectives results in
monitoring sites that are predominantly concentrated in larger population centers where anthropogenic air
pollution sources are concentrated (Figure 4-la). The relatively high cost of establishing and maintaining
regulatory monitoring sites limits their overall numbers and reach into smaller and more remote
communities. The Code of Federal Regulations [CFR; U.S. EPA (2016)1 require the use of U.S. EPA
designated Federal Reference Method (FRM) or Federal Equivalent Method (FEM) instruments for
regulatory National Ambient Air Quality Standards (NAAQS) compliance monitoring. However, some
flexibility is provided to monitoring agencies in using nonregulatory PM measurements when reporting
the U.S. EPA Air Quality Index (AQI) as detailed in 40 CFR Appendix G to Part 58. Although there are
efforts by individual state, local, and tribal monitoring agencies, U.S. EPA currently has no national air
quality monitoring programs specifically designed to evaluate wildland fire smoke impacts. There are no
national programs designed to require the establishment of new sites in smoke prone areas, no grant
opportunities to otherwise encourage optional supplemental smoke monitoring, and no program to
evaluate the performance of designated FRM/FEM monitoring instruments in smoke. As a result, even
though U.S. EPA and its state, local, and tribal partners developed and maintain a relatively advanced set
of regulatory air monitoring networks, remote wildland firefighter camps and smaller population centers
impacted by smoke in most instances lack adequate observational air quality data, and those instances
where regulatory monitors are present the accuracy of the reported smoke impacted air pollution data is
uncertain (Landis et al.. 2017; Long et al.. In Press).
4-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
° °o
, «$-•
'1-^ o 1
ogo * o«~-
ri§ 9&V" eT~"
p* t o.
p«*~ -o
• Rx
o /t>
.£
O »"T
Q
A • ^
lO
~
I A fe . £
• . ° % , D
li o04 . .io
"b ^ o .- f
JSU 0 ^ A
ogo^ o*~ „•
,.ff?.«3rdr- a a* •
AO o.
n
O-
o\
• o A 4
O-tf
f~
O;
A 0 ,0
* •
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
websites ("smoke blogs") dedicated to providing the public with information on wildfire smoke impacts
(Section A.4.1). The material delivered by these smoke blogs varies from state to state with the sites
leveraging smoke and fire observations and forecast products from a variety of sources. National
coverage is provided by the AirNow website, which uses ambient regulatory air quality monitoring data
and calculates AQI values based on current measurements of six NAAQS pollutant indicators (particulate
matter with a nominal aerodynamic diameter less than or equal to 10 |im [PMio], PM2.5, CO, nitrogen
dioxide (NO2), O3, sulfur dioxide (SO2) to inform the public of the current air quality conditions and what
associated health effects may be of concern. AirNow also provides modeled forecasts for future air
quality and links to numerous resources for understanding air quality during smoke episodes and
protecting public health [e.g., U.S. EPA (20196)1. The accuracy of the U.S. EPA AQI and the
appropriateness of the associated AirNow public health messaging are a direct function of the underlying
measured observational air quality data and spatial interpolation models.
The U.S. EPA and USFS have partnered to develop the AirNow Fire and Smoke Map
Ihttps://rirc.airno\v.gov/:AirNow (2021 b) I through a pilot project that incorporates temporary monitors
(Figure 4-lb) and beginning in 2020 air quality sensor data (Figure 4-lc), initially from PurpleAir PM2 5
measurements (PurpleAir. 2021). to provide spatially improved AQI and associated public health
messaging during wildfire season. The associated public health messaging on the site is augmented
through direct access to the IWFAQRP daily Smoke Outlooks for specific incident impacted areas.
PurpleAir and similar commercially available air quality sensors have just started to be evaluated under
high smoke concentration conditions in laboratory and field studies. These evaluations demonstrate the
sensors' variable accuracies under different smoke impact conditions but highlight their potential for
providing timely public health messaging during wildland fire smoke events after calibration of reported
raw results (Landis et al.. 2021; Delp and Singer. 2020; Holder et al.. 2020; Mehadi et al.. 2019). Remote
sensing observations also provide the opportunity to visualize the downwind transport of wildland fire
smoke and inform potential impacts on ambient air quality (Wu et al.. 2018; Krstic and Henderson. 2015;
Mei et al.. 2012; Liu et al.. 2009). The primary shortcoming of these satellite-based total air column
measurements is there is no definitive way to know whether the observed plume is impacting surface air
quality conditions or being transported aloft. Additionally, visible band measurements are only available
during daylight hours, and plumes are only detectable at relatively high concentration. The emergence of
a ground-based ceilometer network, the Unified Ceilometer Network (UCN); https://alg.umbc.edu/ucn/).
through a collaboration between U.S. EPA, University of Maryland, Baltimore County (UMBC), National
Aeronautics and Space Administration (NASA), and National Oceanic and Atmospheric Administration
(NOAA) will provide three-dimensional aerosol backscatter profile measurements at numerous sites
across the U.S. The ceilometer measurements will allow characterization of the smoke plume heights,
including multiple layers, when smoke is transported over the sites.
This chapter summarizes current national regulatory ambient air quality measurement
infrastructure, nonregulatory temporary incident response measurement capabilities, air quality sensor
capabilities, and remote sensing products and their utility in estimating the impact of wildland fire smoke
4-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
on air quality. Limiting exposure is the principal measure available to mitigate human health impacts of
smoke, and real-time measurements of air quality are critical to providing actionable guidance to
communities for minimizing population exposure. Air quality data from the current discrete federal, state,
local, and tribal monitoring programs, remote sensing products, and ad hoc air quality sensor
manufacturer's public web portal data are the basis for deterministic air quality model development and
validation (CHAPTER 5) and wildland fire smoke exposure and health assessment research (CHAPTER
6). This chapter will also describe the current availability of air quality monitoring data, the relative
accuracy of different types of monitoring instruments, public availability of measurement data, gaps in
smoke monitoring capabilities, the challenges of ambient smoke monitoring, and provide
recommendations to improve future ambient monitoring and data curation efforts for better
characterization of the air quality impacts from wildland fire smoke.
4.2 Objectives of Air Quality Monitoring
4.2.1 Public Reporting of Air Quality through the Air Quality Index
(AQI)
The Clean Air Act (CAA) requires U.S. EPA to protect public health and welfare by
promulgating NAAQS for common harmful pollutants and establishing a uniform AQI for reporting of air
quality for CO, NO2, O3, PM2.5, PM10, and SO2. AQI values (0-500) are calculated individually for each
of the five major air pollutants and are based on measured or forecast concentrations. The single AQI
value reported on the multiagency AirNow represents the current highest individually calculated pollutant
value (NowCast AQI) and is used to communicate how clean or polluted the air is, and guidance for
planning outdoor activities (Table 4-1). During wildland fire smoke events PM2 5 is typically the primary
pollutant responsible for the elevated AQI values and the specific suggested intervention strategies to
lower population PM2 5 exposures to smoke and resulting negative health outcomes.
4-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 4-1 Understanding the U.S. Environmental Protection Agency Air Quality
Index (AQI): An example for PM2.5.
Level of Concern
Air quality
conditions are:
AQI Color
As symbolized by
this color
Value of Index
When the AQI is in
this range:
PM2.5 (|jg/m3)
With a 24-h
concentration of:
PM10 (|jg/m3)
With a 24-h
concentration of:
Good
Green
0-50
0.0-12.0
0-54
Moderate
Yellow
51-100
12.1-35.4
55-154
Unhealthy for
sensitive groups
Orange
101-150
35.5-55.4
155-254
Unhealthy
Red
151-200
55.5-150.4
255-354
Very unhealthy
Purple
201-300
150.5-250.4
355-424
Hazardous
Maroon
301-400
401-500®
250.5-350.4
350.5-500.4
425-504
505-604
|jm/m3 = micrograms per cubic meter; AQI = Air Quality Index; h = hour; PM = particulate matter; PM25 = particulate matter with a
nominal mean aerodynamic diameter greater than 2.5 |jm and less than or equal to 10 |jm; PM10 = particulate matter with a
nominal aerodynamic diameter less than or equal to 10 |jm; SHL = significant harm level.
aAn index value of 500 represents the SHL. SHL's are those ambient concentrations of air pollutants that present an imminent
and substantial endangerment to public health or welfare, or to the environment, as established in 40 CFR 51.151 (U.S. EPA.
2001). For PM there is only a published SHL for PM10.
State, local, and tribal agencies regularly monitor and report their air quality data to U.S. EPA for
the calculation of AQI. However, most monitoring agency networks are designed around urbanized areas
known as Core Based Statistical Areas (CBSAs). These networks are typically designed to evaluate the
pollution exposure associated with anthropogenic pollution sources under meteorological conditions of
pollution maxima as required by the CFR (U.S. EPA. 2015c). Air pollutant monitoring networks such as
for PM2.5 also include upwind, downwind, and transport sites for each state. State, local, and tribal
agencies report all available data for calculation of the AQI as well as to understand transport into and out
of their monitored jurisdictions. However, a major limitation of many state, local, and tribal agency
networks is that the network design requirements associated with urbanized areas concentrate sites inside
of major population centers. The focus on urbanized areas as well as the large geographical footprint of
unmonitored rural areas often results in very limited or no monitors in areas adversely impacted by
wildland fire smoke. Specifically, a recent U.S. Government Accountability Office report found that
2,120 of the 3,142 counties (67.5%) in the U.S. had no regulatory monitor (GAP. 2020).
AQI values for PM2.5 and O3 presented on the AirNow website, App, or widget that are entitled
"current air quality" are calculated using the U.S. EPA NowCast algorithms. The full AQI is based on
averaging times used for the NAAQS: 24-hour local midnight to midnight average for PM2.5 and PM10;
max 8-hour avg for CO and O3; and max 1-hour avg for NO2 and SO2. The NowCast algorithm is
complex, but is designed to approximate the full AQI, but to also be more responsive to recent data trends
and to be calculable after each new hour's data (U.S. EPA. 2020d). The NowCast algorithms dynamically
4-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
scale the duration of past hourly monitoring data used to calculate the Current AQI based on the observed
temporal trend of ambient concentrations, using longer time averages during stable concentrations and
shorter time averages when air quality is changing rapidly. In practice they often approximate a 3-hour
running average. The hourly updated NowCast PM2 5 and O3 AQI values are useful during wildland fire
events when downwind ambient air quality can change abruptly but do not necessarily reflect
up-to-the-minute current conditions.
4.2.2 Analyzing Air Quality Trends
In addition to directly informing AQI, regulatory network air quality data collected from fixed
state, local, and tribal monitoring stations with at least several years of data allow for the characterization
of air quality trends and provides context for understanding wildland fire smoke conditions. U.S. EPA
maintains an annual air trends report in the form of an interactive web application [e.g.,.
https://gispub.epa.gov/air/trendsreport/2020/; U.S. EPA (2020gYI. The online report features a suite of
visualization tools that allow the user to:
• Learn about air pollution and how it can affect our health and environment.
• Compare key air emissions to gross domestic product, vehicle miles traveled, population, and
energy consumption back to 1970.
• Take a closer look at how the number of days with unhealthy air has dropped since 2000 in
35 major U.S. cities.
• Explore how air quality and emissions have changed over time for each of the common air
pollutants.
• Check out air trends where you live.
Information about long-term air quality trends can be useful in determining the extent to which
air quality management strategies are helping reduce concentrations of pollutants to the levels specified
by the NAAQS. Online resources are also available that present daily trends in air quality during wildland
fire events that can be used to estimate daily (Figure 4-2a) and year-to-date (Figure 4-2b) population
exposure rhttps://tools.airfire.org; regional air quality and historical tools; USFS (2021bYI. The AirFire
resource can be used to contextualize the current air quality conditions during large wildland fire events
and the dramatic impact on population PM2 5 exposure like that from a September 2020 wildfire event on
the state of Oregon presented in Figure 4-2b.
4-6
DRAFT: Do Not Cite or Quote
-------
Oregon Air Quality
state population living within each AQ1 level
4,000.000 „
01-01 02-01 03-01 04-01 05-01 06-01 07-01 08-01 09-01 10-01 11-01 12-01
2005 - 2018 — 2019 — 2020 — 2021
AQI = Air Quality Index; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm;
USFS = U.S. Forest Service; WRCC = Western Regional Climate Center.
Colors shown in Figure 4-2a are U.S. EPA AQI categories (Table 4-1) and grey indicates no data.
Source: https://covid.airfire.ora/tracking/. IWFAQRP (2021.; site uses U.S. EPA regulatory monitor data and an analysis of
LANDSCAN population data within 20 km of each monitoring site.
Figure 4-2 Tracking of Air Quality Index (AQI) in Oregon during the 2020
wildfire season (a) and the cumulative annual Oregon population
exposure to PM2.5 (b) showing the clear impact of wildland fire
events.
1 4.2.3 Informing Fire Management
2 Smoke from wildland fires can impact the health and safety of fire personnel and the public,
3 interfere with fire suppression operations and transportation, and disrupt local economies (USFS. 2020a).
4 Because of the scale of these impacts, such as those seen during the 2020 western U.S. wildfire season,
5 these impacts can become the focus of fire management, air quality regulators, and public health officials.
6 The USFS led the development of the IWFAQRP to address the air quality impacts of wildfires on the
7 American public. IWFAQRP uses emergency deployable air quality monitoring equipment, state-of-the-
4-7
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
art wildland fire smoke dispersion models, and ARAs for dispatch to ongoing wildfires to develop and
publicly disseminate smoke impact information (USFS. 2020a). ARAs are technical specialists that
deploy nationwide to large wildfires to assist with understanding and predicting smoke impacts on local
communities and fire personnel. They work on Incident Management Teams with their public
information, fire behavior, and fire weather specialists as well as coordinate with local emergency
response, air regulatory, and public health agencies to provide timely smoke outlooks that address the
public health risks and concerns from smoke (USFS. 2020a). In areas without an existing PM2 5 monitor,
ARAs deploy temporary PM monitors to provide real-time information on air quality to assist local
officials and communities make informed decisions to minimize their exposure to smoke. ARAs are also
a point of contact for the public and commonly present smoke information at public meetings and address
smoke-related concerns of local citizens.
Smoke impacts from prescribed fires also present significant challenges to land management
agencies. Prescribed fire is an important tool for achieving key management objectives such as ecosystem
restoration and maintenance and hazardous fuel reduction. Smoke management concerns are among the
top impediments to prescribed burning rMelvin (2018. 2015); see CHAPTER 31. While nuisance smoke
is the most common smoke issue, prescribed fires can subject local communities and sensitive
populations to unhealthy levels of PM2 5 (Melvin. 2018. 2015). Prescribed fire smoke can also endanger
public safety by reducing visibility on roadways leading to serious and/or fatal traffic accidents
(Bartolome et al.. 2019; Ashley et al.. 2015). Additionally, unlike wildfires, prescribed fires are
considered a controllable emission source and the resultant smoke can trigger a regulatory violation of
NAAQS. However, the 2016 Exceptional Events Rule states that prescribed fire on wildland can be a
human-caused event eligible for treatment as an exceptional event, and properly managed prescribed fires
are generally less likely than wildfires to cause or contribute to an exceedance or a violation of the
NAAQS (U.S. EPA. 2019d). In instances where smoke from a prescribed fire leads to an exceedance or
violation of a NAAQS, and all rule criteria are satisfied, air agencies or federal land managers can prepare
an exceptional events demonstration and request the event-influenced data to be excluded from the data
set used for certain regulatory determinations. To help mitigate these deleterious smoke impacts and
obtain observational data to improve smoke management techniques and tools, land management
agencies and atmospheric researchers have increasingly begun to deploy temporary smoke monitors as
part of operational prescribed burns (Pearson. 2021).
4.2.4 Quantifying the Impact of Wildland Fires on Air Quality
One of the key objectives of the U.S. EPA regulatory air monitoring program is quantifying
specific anthropogenic source impacts on NAAQS pollutant concentrations. However, there are no
existing national monitoring programs specifically designed to evaluate air pollutant impacts from
wildfires or prescribed fire programs even though the U.S. EPA National Emissions Inventory (NEI) has
reported wildland fires contribute a substantial amount to total national annual CO (30-43%) and PM2 5
4-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
(32-44%) emissions from 2011-2017 (U.S. EPA. 2021b). However, it remains unclear how emissions of
these pollutants from wildland fires translates to overall contributions to annual ambient concentrations.
To date, U.S. EPA has not undertaken a national measurement-based integrated assessment into the
impact of wildland fire emissions on (1) ambient air quality, (2) regulatory NAAQS compliance, or
(3) negative human health outcomes. There are numerous local/regional assessments in the scientific
literature that document the deleterious changes on ambient air quality, human exposures, and human
health outcomes related to specific wildfire events (Stowel 1 et al.. 2019; Landis et al.. 2017; Reid et al..
2016; Cisneros et al.. 2012; Rappold et al.. 2011). Incremental progress in the examination of the
local/regional impact of wildland fire smoke on ambient air quality, human exposure, and human health
effects is being made (Johnson et al.. 2020). However, in the absence of a national measurement-based
assessment the full impact of wildfire smoke remains largely unknown on a national scale, particularly at
population centers that are often distant from wildfire events.
4.3 Ambient Air Quality Monitoring Capabilities
4.3.1 Overview
The fundamental understanding of wildland fire source emission estimates, the impacts of smoke
on air quality, human exposures and health outcomes, and the ability to develop and validate predictive
deterministic air quality models, are predicated on accurate measurements of air pollutants in smoke.
While there are no U.S. EPA national air quality monitoring networks focused on wildland fire smoke,
there are several discrete federal, state, local, and tribal monitoring programs and ad hoc air quality sensor
networks that provide critical observational air quality data during wildland fire events. This section
summarizes current national regulatory ambient air quality measurement infrastructure, nonregulatory
temporary incident response measurement capabilities, air quality sensor capabilities, and remote-sensing
products and their utility in estimating the impact of wildland fire smoke on air quality.
4.3.2 U.S. EPA Routine Regulatory Monitoring Networks
U.S. EPA has established NAAQS for the criteria pollutants and maintains multiple national
regulatory air pollution monitoring networks as required by the CAA that are set forth in Title 40, Part 50
of the Code of Federal Regulations (U.S. EPA. 2016). Each monitoring network has associated regulatory
requirements and policy objectives that dictate decisions on the location of and pollutants measured at
each site. In addition to reporting the AQI in large population centers, key monitoring objectives include
NAAQS compliance, trend analysis, quantifying specific source impacts, and improving the performance
of air quality forecast models. National monitoring of air quality is accomplished through a partnership of
U.S. EPA delegated federal, regional, state, city, and tribal stakeholder organizations. U.S. EPA
4-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
regulatory monitoring is carried out as part of a national network of approximately 4,400 monitoring sites,
called the State and Local Air Monitoring Stations (SLAMS). The air quality data obtained from these
sites are reported to U.S. EPA's Air Quality System (AQS) database, along with other information, and
are used for determining compliance with the NAAQS, assessing effectiveness of State Implementation
Plans (SIPs) in addressing NAAQS nonattainment areas, characterizing local, state, and national air
quality status and trends, and associating health and environmental damage with air quality
levels/concentrations.
To assure the accuracy, integrity, and uniformity of the SLAMS air quality monitoring data
collected, the U.S. EPA has established one or more FRM for measuring each of the six criteria
pollutants. These FRMs are set forth in appendices to 40 CFR Part 50 and specify a measurement
technique to be implemented in commercially produced monitoring instruments (U.S. EPA. 2020f. h, i, j.,
k, 2011a. b, c). These monitoring instruments must be shown to meet specific performance requirements
in addition to other requirements detailed in the U.S. EPA regulations at 40 CFR Part 53 (U.S. EPA.
2019a). in which case the instrument may be designated by the U.S. EPA as an FRM analyzer. To
encourage innovation and development of new air quality monitoring methods, the U.S. EPA has also
provided for FEMs. An FEM is not constrained to the specific measurement technique of the
corresponding FRM. However, an FEM must meet the same or similar performance requirements as
specified for the corresponding FRM, and in addition, it must show a high degree of comparability to
collocated FRM measurements at one or more field testing sites under typical ambient conditions. These
FEM requirements are also detailed in 40 CFR Part 53 (U.S. EPA. 2019a). and a monitor that is shown to
meet all applicable requirements may be designated by the U.S. EPA as an FEM monitor. A current
listing of all designated FRMs and FEMs as of December 2020 can be found at
https://www.epa.gov/sites/production/files/2019-08/documcnts/dcsignatcd reference and-
equivalent methods.pdf (U.S. EPA. 20206).
The siting criteria and regulatory monitoring methodologies are briefly described here and a
detailed discussion of specific air pollution networks, criteria air pollutants measured, and measurement
methods are provided in Section A.4.2 (PM2 5 Mass Monitoring), Section A.4.3 (PM2 5 Speciation
Monitoring), and Section A.4.4 (Criteria Gas Monitoring). The U.S. EPA PM2 5 monitoring program is
the largest component of the national monitoring infrastructure and PM2 5 monitors are mostly sited in
urban areas at the neighborhood scale as defined in 40 CFR Appendix D to Part 58 (U.S. EPA. 2015a. b),
where typical PM2 5 concentrations are reasonably homogeneous throughout an entire subregion in the
absence of wildland fire smoke. There are four main components of the U.S. EPA PM2 5 monitoring
program: 24-hour integrated filter-based FRM samplers, continuous FEM mass instrument measurements
reported as 1-hour concentrations, continuous non-FEM mass instrument measurements reported as
1-hour concentrations, and 24-hour integrated filter-based Chemical Speciation Network (CSN) samplers.
Continuous PM2 5 FEM and criteria gas FRM/FEM (Table A.4-1) real-time data support NAAQS
compliance and is integrated with non-FEM continuous PM2 5 data to support public AQI communication
and air quality smoke forecasting on AirNow (AirNow. 2021a). The top three states using the non-FEM
4-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
continuous PM2 5 instruments for AQI reporting are Washington (n = 47), Oregon (n = 45), and California
(n = 43) primarily to communicate air quality impacts from wildland fire smoke.
4.3.3 Temporary/Incident Response Measurements
The IWFAQRP provides significant incident response smoke monitoring capabilities by
maintaining a cache of smoke monitoring equipment for nationwide deployment by their ARA personnel.
The IWFAQRP smoke monitor cache consists of -40 Met One Instruments (Grants Pass, OR)
E-SAMPLER and E-BAM nonregulatory PM25 samplers (USFS. 2020a). These PM2 5 samplers are
available upon request to land management agency administrators and incident management teams for
monitoring wildfires and to federal land managers conducting prescribed burns. The monitors are
typically used by ARAs supporting wildfire incident management teams. The ARA will often consult
with local land managers, air quality regulatory agencies, and public health officials for guidance on
positioning temporary smoke monitors. When siting monitors, ARAs attempt to meet the same siting
criteria as used for regulatory FRM/FEM monitors like avoiding interferences from other emission
sources or physical barriers that may obstruct air flow around the sampler (U.S. EPA. 2012). Deployment
of either the E-BAM or E-SAMPLER requires a land line power hookup, but since these monitors are
intended to inform communities, this infrastructure requirement is not typically an issue as they are often
deployed at fire stations, schools, or other municipal buildings. The E-BAM and E-SAMPLER both
upload their measurement data by satellite to a cloud-based data acquisition system where it is reported on
an hourly basis. The data must then pass a quality assurance (QA) check before being publicly distributed
through the www.airfire.org or the www.fire.airnow.gov websites (see Section 4.4.3V
Beyond the national smoke monitoring resources offered by IWFAQRP, jurisdictions within
federal agencies (e.g., USFS Regions), states, and tribal agencies also maintain and deploy PM samplers
for monitoring smoke impacts from wildfires and prescribed burns. The PM samplers used include
ThermoFisher Scientific (Franklin, MA) DataRAM pDR-1500 and Met One Instruments BAM-1020,
E-BAM, and E-SAMPLER systems (USFS. 2020b). Remote telemetry and satellite data transmission are
used to gather and present the raw data in "near-real-time" (USFS. 2020b). These data are also collected
and integrated into the www.fire.airnow.gov websites (see Section 4.4.3). Several states have programs to
monitor smoke from wildland fires. The most extensive program is in California, where the California Air
Resources Board (CARB) and local air districts have >100 E-BAM samplers available for deployment to
monitor smoke impacts (Pearson. 2021). The CARB program initially targeted wildfires but was
expanded in response to state legislation (California SB-1260. 2018) which sought to increase hazardous
fuels reduction and included funding for prescribed fire smoke monitoring. Other states with monitoring
efforts for wildfire and prescribed fire include Alaska, Arizona, Colorado, Idaho, Montana, Nevada, New
Mexico, Oregon, and Washington as well as some tribes (Section A.4. IV Even with the deployment of
temporary PM monitors supplied by federal, tribal, state, or local agencies, gaps in air quality
observations often persist, especially in lower population foothill and mountain communities. To address
4-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
monitoring gaps, ARAs began deploying PM sensors (Purple Air) in 2020 to estimate the air quality in
communities that previously would have gone without monitoring. Likewise, states and local agencies are
augmenting existing monitoring networks with air quality sensors. In 2018, CARB initiated a pilot
program that distributed several hundred PM sensors to augment existing air quality monitoring networks
and capabilities (Pearson. 2021).
Historically, smoke monitoring for wildfire response has relied on existing and temporarily
deployed stationary monitors to provide air quality information to incident command and state/local
public health officials. The relatively small number of local monitors, the dynamic nature of smoke
emissions and transport, and the dispersed nature of firefighting personnel and downwind communities
make predicting exposures challenging. Mobile monitoring capabilities are another strategic approach to
measure and communicate real-time smoke information. Personal monitoring of firefighters, vehicle
mounted instruments, and airborne drones are all viable mobile monitoring platforms. These approaches
have been used in focused research studies (Apte et al.. 2017; Navarro et al.. 2016; Villa et al.. 2016) but
not for routine operations. Research into both built-for-purpose and commercially available
small-form-factor air quality measurement systems have been reported by Landis et al. (2021) and Holder
et al. (2020). respectively; and others are working on continuous reading mobile air quality platforms (2B
Tech. 2021; Mui et al.. 2021; Apte et al.. 2017).
With exception of the BAM-1020, monitors used by federal, state, local, and tribal agencies for
temporary smoke monitoring are not expected to be U.S. EPA designated FEM monitors. Across all
agencies the most frequently deployed monitors are E-BAM and E-SAMPLERS. The performance of
both samplers in measuring PM2 5 in fresh smoke has been evaluated in limited laboratory testing which
indicates high correlation (r2 >0.9) and relatively low bias range for the E-BAM (1-21%) and
E-SAMPLERS (8-18%) compared to reference FRM/FEM monitors across concentration ranges of
20-1,700 (ig/m3 (Mehadi et al.. 2019; Trent. 2006. 2003). However, it is unclear how the samplers
perform across the natural range of smoke properties (chemical composition, size distribution) and how
performance may vary over extended periods of sampling in smoke impacted environments. Additionally,
federal interagency smoke monitor inventories also include DataRAM monitors, and laboratory
evaluation of these monitors indicates they overestimate PM2 5 in smoke by a factor of ~2 (Trent. 2003).
4.3.4 Sensors
Over the last decade there has been rapid development of miniaturized, user-friendly air quality
sensor systems (Karagulian et al.. 2019; Malings et al.. 2019; Baron and Saffell. 2017; Williams et al..
2015). Significant advancements in internal gas and PM sensor components, compact microprocessors,
power supply/management, wireless data telemetry, advanced statistical data fusion/analysis, real-time
sensor calibration, and graphical data interfaces hint at the future potential of accurate small form factor
integrated air quality sensor systems. This technology is being developed for a variety of potential
4-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
applications, including human exposure assessment (Moraw ska et al.. 2018). industrial emissions (Thoma
ct al.. 2016). local source impact estimation (Feinberg et al.. 2019). and to increase the spatial density of
outdoor monitoring networks (Bart et al.. 2014; Mead et al.. 2013). Some manufacturers of air quality
sensor systems have built cloud-data systems and public websites to host measurement data and allow
public access (2B Tech. 2021; Kunak. 2021; PurpleAir. 2021). The large number of installed sensors and
centralized data hosting capabilities of PurpleAir (Draper, UT) led the U.S. EPA and USFS to launch a
pilot project to provide data from air quality sensors calibrated with U.S. EPA's correction equation
(Barkiohn et al.. 2020) and the derived AQI and NowCast on the AirNow Fire and Smoke Map. The goal
of the pilot project was to provide additional AQI (PM2 5 only) information during wildfires in those areas
not adequately served by regulatory monitoring sites (AirNow. 2021a).
However, the reliability, accuracy, stability, and longevity of many types of air quality sensors
under smoke conditions is largely unknown. Routine performance testing of many air quality sensors, to
date, has been mostly limited to typical ambient conditions (Collier-Oxandale et al.. 2020; Zamoraetal..
2019; Feinberg et al.. 2018; Jiao et al.. 2016; Williams et al.. 2015). with more limited assessment of
certain technologies at higher ambient concentrations (Johnson et al.. 2018; Zheng et al.. 2018). These
previously published findings have indicated, in some cases, high correlation between collocated sensors
and FRM/FEM reference monitors; however, there are also many sensor test results that exhibit
measurement artifacts (Hossain et al.. 2016; Lin et al.. 2015; Spinelle et al.. 2015; Mead et al.. 2013).
inconsistency among identical sensors (Savahi et al.. 2019; Castell et al.. 2016; Williams et al.. 2015).
drift over time (Savahi et al.. 2019; Feinberg et al.. 2018; Artursson et al.. 2000). sensitivity to
environmental conditions [e.g., temperature, relative humidity; Wei et al. (2018); Cross et al. (2017)1. and
limitations to upper limit measurement capabilities (Zou et al.. 2020; Schweizer et al.. 2016). U.S. EPA
has endeavored to improve the reliability, consistency, and accuracy of air quality sensor data by
regularly engaging academia, industry, nonprofit groups, community-based organizations, and regulatory
agencies to develop recommendations, performance targets, and best practices (Duvall et al.. 2021a. b;
Williams et al.. 2019; Clements et al.. 2017). U.S. EPA has also created an online Air Sensor Toolbox
(U.S. EPA. 2021a) as a clearinghouse for information on the use of air quality sensors. The U.S. EPA Air
Sensor Toolbox aims to improve the operation, data collection, and quality assurance of air sensor data by
providing resources such as the Air Sensors Guidebook, Standard Operating Procedures (SOPs) for
sensors, Sensor Performance Targets and Test Protocols, Sensor Collocation Guide, Sensor Evaluation
Reports, Quality Assurance Handbook and Guidance Documents for Citizen Science Projects, and air
sensor loan opportunities to enable the public to learn about air quality in their communities (U.S. EPA.
2020b).
More recently U.S. EPA partnered with other U.S. federal agencies (Centers for Disease Control
and Prevention (CDC), NASA, National Park Service (NPS), NOAA, USFS) to sponsor the Wildland
Fire Sensor Challenge to advance wildland fire air measurement technology to be easier to deploy,
suitable to use for high concentration events, and durable to withstand difficult field conditions, with the
ability to report high time resolution data continuously and wirelessly (Landis et al.. 2021). The Wildland
4-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Fire Sensor Challenge encouraged innovation worldwide to develop sensor prototypes capable of
measuring PM2 5, CO, carbon dioxide (CO2), and O3 during wildfire episodes. The raw PM2 5 sensor
accuracies of the three winners ranged from -22-32%, while smoke specific U.S. EPA regression
calibrations improved the accuracies to -75-83% demonstrating the potential of these systems in
providing reasonable accuracies over conditions that are typical during wildland fire events (Landis et al..
2021). Select commercially available PM2 5 sensors have also been evaluated versus collocated FEM
measurements under smoke conditions that highlight their potential for providing useful public health
messaging during wildland fire smoke events (Dclp and Singer. 2020; Holder et al.. 2020; Mehadi et al..
2019). These research studies like the Wildland Fire Sensor Challenge have shown that raw PM2 5 sensor
data requires post-processing using smoke specific calibration functions to account for differences in
aerosol chemistry, particle size distribution, aerosol density, and optical properties. The range of smoke
specific calibration correction factors are summarized in Table A.4-2 and show a broad range of
responses depending on specific fire conditions and reference instruments used. As more information is
gathered on the performance and calibration of air quality sensors in wildland fire smoke, the utility of
their reported air quality measurements for informing public health messaging will improve.
4.3.5 Remote Sensing/Satellite Data
Remote sensing is the science of acquiring information about the Earth's surface or atmosphere
without actually being in contact with it and requires a source of reflected, emitted, or absorbed and
re-emitted energy which interacts with the geophysical parameter being measured, such as aerosols
(e.g., PM25) in the atmosphere. As a result, remote sensing allows for the estimation of wildfire smoke
impacts in areas of the country that lack other sources of ground-based observational data. The two types
of remote sensing are referred to as passive and active. Passive remote sensing uses the sun as the energy
source, where the solar radiation is reflected by the Earth's surface or scattered in the atmosphere for
visible wavelengths or absorbed and then re-emitted from the earth surfaces for thermal infrared
wavelengths. Because measurements made in the visible wavelengths require reflected solar radiation,
they can only be conducted during daylight hours. Active remote sensing techniques require their own
energy source, where the emitted radiation is directed at the target of interest and reflected back to the
instrument. In active remote sensing, lasers often provide this energy source, and the pulsing energy can
provide information on 3-dimensional structure of the geophysical variable being measured; this
technique is called Light Detection and Ranging (LiDAR).
4.3.5.1 Satellite Measurements
Satellite observations from both low earth and geostationary orbit have become an important set
of measurements for monitoring air pollutant abundances and transport across large spatial scales.
Instruments aboard low-earth-orbit satellites most often provide once-a-day observations over the region
4-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
of interest and with large swaths provide global coverage every day. Geostationary satellites are in
fixed-orbit position relative to the earth and used to observe phenomena which require high-temporal
resolution observations, such as severe weather and disasters, such as wildland fires. Most satellite
instruments used to provide information on air quality are passive remote sensing instruments and span
wavelengths in the ultraviolet-visible (UV/VIS) range or the thermal infrared. The visible wavelengths are
used to provide true color imagery which can be used to identify smoke from wildland fires in a
cloud-free scene. More quantitative measurements involve the measure of backscatter radiances in the
UV/VIS or thermal infrared emission through the atmosphere to provide a derived geophysical column
measurement dependent on physics-based retrieval algorithms (Martin. 2008). Over the past two decades,
satellite column measurements are increasingly being used to provide near-surface information on
pollutants such as PM25, O3, NO2, SO2, CO, and formaldehyde (CH2O). Polarimetric, multispectral,
multidirectional, and active remote sensing observations bring information on the aerosol amount, size,
type, and vertical distributions of column abundances of the geophysical parameter of interest.
One of the major challenges of passive remote measurements from satellites is resolving the
vertical distribution of the parameter of interest, and in some cases the sensitivity of the satellite
measurement to the lowermost atmosphere, which is the region with substantial variability and the most
relevant for gaining an understanding of ambient air quality and subsequent public health impacts
(Martin. 2008). Nevertheless, it is well documented that satellite-derived geophysical parameters of
column integrated abundances such as aerosol optical depth (AOD) can be used to constrain estimates of
near-surface pollution concentration, especially when used in combination with model (chemical transport
or statistical)-based predictions, and provide valuable information on the horizontal distribution of the
pollutant burdens because of the satellite instrument synoptic field-of-view. Smoke plume height
characteristics from the Moderate Resolution Imaging Spectroradiometer (MODIS) Multi-Angle
Implementation of Atmospheric Correction (MAIAC) algorithm (Lvapustin et al.. 2019). which is based
on 11 -|im absorption of fire-emitted gases in the plume, have shown potential for improving surface PM2 5
concentration estimates derived from AOD (Cheeseman et al.. 2020). LiDARs aboard satellites have a
unique capability of resolving the vertical distribution of aerosol in the atmosphere and can make
measurement during the day and night (Winker et al.. 2010) but have very limited spatial coverage to
capture wildland fire plumes (Raffuse et al.. 2012). TROPOspheric Monitoring Instrument (TROPOMI)
provides an aerosol height index which is a developing data product (Griffin et al.. 2020) and offers the
ability to diagnose the aerosol plume height to help assess if the majority of aerosol seen by satellite is
within the boundary layer or being transported aloft.
4.3.5.1.1 Correct Reflectance True Color Imagery—Smoke Plume
Identification and Tracking
Satellite-corrected reflectance from visible wavelengths, also referred to as true color imagery,
from both geostationary and low earth (polar) orbiting satellite instruments is one of the basic satellite
4-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
data products used to identify the spatial extent of smoke plumes and transport from wildland fires. In
addition to providing access to true color images from Geostationary Operational Environmental
Satellites (GOES), satellite analysts atNOAA develop a daily smoke analysis over the contiguous U.S.
and adjacent area of Canada through their Hazard Mapping System (HMS) Fire and Smoke Product
website (https://www.ospo.noaa.gov/Products/land/hms.html#maps'). The HMS products use multiple
data inputs to create a digitized data product displaying the extent of visible smoke and a qualitative
classification on the density of the smoke as low, medium, or high based on visible opacity determined by
the analysts (NOAA. 2020). In combination with surface observation of pollutants or visibility these
imagery-based data products can help identify areas with air quality impacts from wildland fires, but
alone these products provide no information on air quality at the surface.
4.3.5.1.2 Satellite (Geophysical) Composition Observations
AOD is an integrated measure of extinction through the atmosphere that is the derived
geophysical variable from satellite instruments most relevant to PM2 5 and PM10 mass concentrations.
Both operational and research algorithms are used to generate AOD from passive satellite sensors such as
MODIS, Multiangle Imaging Spectroradiometer (MISR), Visual Infrared Imaging Radiometer Suite
(VIIRS), and the GOES Advanced Baseline Imager (ABI). Deriving near-surface PM concentrations from
AOD values is difficult. The challenges in deriving PM concentrations from AOD values are related to
uncertainties in the microphysical (intrinsic) properties of the particles, their size distribution, aerosol
type, and hygroscopic state, as well as key extrinsic properties, such the vertical profile distribution (Hoff
and Christopher. 2009). Few measurements studies have examined the uncertainties associated with the
use of AOD measurements to estimate ground-based PM2 5. Early results from the NASA
DISCOVER-AQ mission over the urban Baltimore region (Crumevrolle et al.. 2014) found accurate
quantification of the aerosol mixed layer height is critical for predicting PM2 5 concentrations, with
aerosol type variability being of lesser importance. In addition, the results indicate the presence of aerosol
layers above the boundary layer introduced significant uncertainties in surface PM2 5 concentrations
estimates when using a column-integrated AOD measurements, and that active remote-sensing techniques
such as LiDARs can provide a characterization of aerosol layers to improve upon the PM2 5 estimates. The
transport of smoke plumes can often result is stratified aerosol layers, including aerosol layers above the
boundary layer, so proper characterization of such aerosol layer structure remains a critical variable in use
of satellite AOD to predict surface PM2 5 concentrations.
Geophysical retrievals of trace gas column abundance from satellite have seen great
improvements in spatial resolution over the past several decades with European Space Agency (ESA)
Sentinel 5 Precursor TROPOMI launched in 2017 now producing global daily NO2 observations at a
resolution (7 x 3.5 km) consistent with chemical transport modeling used for wildland fire air quality
forecasting. In addition to NO2, TROPOMIs standard trace gas data products include column abundances
of CH2O, SO2, CO, and methane at varying spatial resolutions (Lcvclt et al.. 2006). The NASA
4-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Tropospheric Emissions: Monitoring Pollution (TEMPO) mission, scheduled to launch in mid-2022 into a
geostationary orbit, will provide hourly observations of NO2 and O3, across the North American continent
during daytime (Zoogman et al.. 2017).
Because there are significant portions of the U.S. that have no continuous surface monitors, a
very active stream of research developed in the early 2000s focused on the use of column abundances of
the aerosol (AOD) and trace gas satellite data products to aid in the predictions of pollutant surface
concentrations. Over the past 20 years many research groups have developed a multitude of methods to
model surface PM2 5 concentrations using AOD from numerous satellites as discussed in a review by Chu
et al. (2016) with a primary focus on estimates of annual PM2 5 concentrations. Some research groups
have continued to improve upon the methods over time as inputs in the methods have improved (Hammer
et al.. 2020). including estimates of PM chemical composition (van Donkelaar et al.. 2019) and have
made their data sets available for public use (Atmospheric Composition Analysis Group. 2021). Similar
research efforts combine chemical transport model and NO2 column abundances to infer surface
concentrations (Cooper et al.. 2020). Most of these research efforts focus on predictions of annual means
for these pollutants versus daily predictions of surface concentrations. For example, Alvarado et al.
(2020) has recently demonstrated the transport and tracking of several trace gases associated with
wildland fires in western Canada.
Useful and actionable information for wildland fire pollutants requires daily predictions of
surface pollutants, such as PM2 5, or more temporally resolved information because of the diurnal nature
of wildland fire emissions and meteorological transport patterns (Baker et al.. 2019). Methods focused on
the use of satellite data to aid in daily pollutant surface predictions during wildland fires have been
demonstrated on a very limited basis through a case-study approach, and include simple regression
analysis (Raffuse et al.. 2013). machine learning techniques (Reid et al.. 2015). and generalized
geographically weighted regression models (Gupta etal.. 2018; Gupta and Christopher. 2009). all with
moderate success. Resolving the apportionment of AOD impacting the surface concentrations is
complicated because of long-range and high-altitude transport of aerosols which often occurs for wildland
fire events. While not associated with surface PM2 5 predictions for wildland fires, Jin et al. (2019) used a
geophysical approach to estimate daily surface PM2 5 concentrations and conducted a detailed assessment
of uncertainties using this approach, estimating uncertainties in the modeled PM2 5/AOD led to an error of
1 (ig/m3 in daily PM2 5 predictions, and while satellite AOD uncertainties produced errors of 8 (ig/m3.
However, none of these efforts provide an ongoing source of data.
The U.S. EPA AirNow Program application called the AirNow Satellite Data Processor (ASDP)
(Pasch et al.. 2013) integrates AOD from the MODIS instruments (Terra and Aqua) with a focus on
improving the accuracy of daily ground-level PM2 5 concentrations. The ASDP approach based to provide
PM2 5 predictions uses climatological scaling factors (van Donkelaar et al.. 2019) from GEOS-Chem. The
NOAA Aerosol Watch provides access to a variety of relevant GEOS (16 and 17) and VIIRS (S-NPP and
NOAA-20) data products, including true color imagery and AOD retrievals which can be overlaid with
4-17
DRAFT: Do Not Cite or Quote
-------
1 AirNow PM2 5 data to help assess if the satellite data and surface concentrations are spatially correlated in
2 time and space which is an indication that the smoke extent observed by the satellite is at or near the
3 surface impacting ground level air quality. Figure 4-3 is a result of recent efforts by NOAA Aerosol
4 Watch to product an operational daily satellite derived PM2 5 product for September 15, 2020 during the
5 Oregon wildland fires. This approach aggregates VIIRS AOD from two polar orbiting satellites, S-NPP
6 and NOAA-20, and applies a regression algorithm from available surface PM2 5 data to produce a daily
7 satellite derived PM2 5 field.
4-18
DRAFT: Do Not Cite or Quote
-------
Good Moderate USG Unhealthy V, Unhealthy Hazardous
0.0 12.0 35.5 55.5 150.5 250.5 500.0
Daily PM2.5 (pg/m3)
PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 pm.
Note; This figure captures spatial extent of poor air quality associated with several large western wildland fire complexes over the
western U.S. Closed circles in the plot represent surface monitors of PM2.5.
Image Source: Shobha Kondragunta (NOAA/NESIDS).
Figure 4-3 Image of surface Air Quality Index (AQI) for PM2.5 from U.S. EPA
AirNow over plotted with Air Quality Index (AQI) for PM2.5 derived
from National Oceanic and Atmosphere Administration (NOAA)
aerosol optical depth from Visible Infrared Imaging Radiometer
Suite (VIIRS) instruments (Soumi-NPP and NOAA-20 satellites) for
September 15, 2020.
1 4.3.5.2 Ground-Based Measurements
2 Ground-based remote sensing networks across the U.S. serve a wide range of functions, such as
3 the highly operational surface weather observation stations which contain several remote-sensing
4 instruments in combination with in situ instruments used to provide continuous observations to generate
5 routine weather reports to more research based networks, such as the NASA Micro-Pulse LiDAR
6 Network (MPLNET). a small federated network of compact LiDARs designed to measure aerosol and
7 cloud vertical structure, and boundary layer heights. The combination of these networks provides relevant
4-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
data on surface visibility, the vertical distribution of aerosols, boundary layer heights, AOD, and total
column NAAQS gaseous pollutants.
The combined Automated Surface Observing System (ASOS)/Automated Weather Observing
System (AWOS) networks consist of over 1,000 sites across the U.S., with ASOS containing over
900 sites. The primary remote sensing measurement at ASOS/AWOS sites is surface visibility. The
visibility measurement uses a forward scatter sensor and detector to measure the attenuation of light by
scattering and absorption at the wavelength of 550 nm. The sensor measures a 1-minute avg extinction
coefficient and reports a 10-minute avg. The 550 nm measurement is very sensitive to PM2 5 and therefore
can be used to understand reduced visibility caused by wildland fire smoke. The ASOS/AWOS sites also
used ceilometers for reporting cloud-based heights. Ceilometers are a type of LiDAR, capable of
providing vertical profile information on aerosols in the troposphere through the attenuation of
backscatter from aerosols. While NOAA operates a large network of ceilometers as part of ASOS, the
instruments are not currently configured to report the aerosol backscatter profiles, which can be used to
define aerosol layer heights and derive a mixing layer height/planetary boundary layer height. Ceilometer
technology is being implemented by U.S. EPA Photochemical Assessment Monitoring Station (PAMS)
program.
Recent updates to the U.S. EPA PAMS network require the stations to measure and report an
hourly mixing layer height. This measurement requirement will be fully implemented by June 2021 and is
primarily being satisfied through the installation of ceilometers across the network sites. While state and
local agencies are required to only report an hourly mixing layer height, an U.S. EPA collaboration with
the UMBC, NASA, and NOAA is focused on the development of a near-real-time data system to archive
and display ceilometer backscatter profiles, aerosol layer heights, and planetary boundary layer heights
(PBLH) from PAMS and non-PAMS ceilometers into the UCN rhttps://alg.umbc.edu/ucn/; UMBC
(2021)1. The UCN will use a common algorithm to determine PBLH (Caiccdo et al.. 2020) and display
near-real-time aerosol backscatter vertical profiles which can be used to track the vertical structure of
aerosol plumes, including wildland fire smoke, as the plumes are transported across the U.S. as shown in
Figure 4-4.
The NASA MPLNET is a global federated LiDAR network which supports research and the
NASA Earth Observing System (EOS) program ("Wielicki etal.. 1995). Value-added network data sets are
made available to the community via an online repository rhttp://mplnet.gsfc.nasa.gov; NASA (2021);
Campbell et al. (2008)1. The micropulse LiDAR operated at 532 nm in contrast to ceilometers which
operate in the 900 nm range or 1,064 nm, which provides the micropulse LiDAR system the benefit of
being more sensitive to PM2 5.
4-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
luHt CHMlSk
Note: Aerosol backscatter profiles from ceilometers located at air quality monitoring sites (a) Bristol, PA; (b) Philadelphia, PA;
(c) Edgewood, MD show the smoke layer being transported above the boundary layer, with little to no impacts to surface air quality.
Source: Unified Ceilometer Network—https://ala.umbc.edu/ucn/. UMBC (2021).
Figure 4-4 Image of western U.S. wildfire smoke transported to the Northeast
U.S. as captured in the Visual Infrared Imaging Radiometer Suite
(VIIRS) True Color Image over plotted with VIIRS aerosol optical
depth for September 16, 2020.
For over two decades the NASA AErosol RObotic NETwork (AERONET), a federated
association of ground-based sun and sky scanning radiometer, has provided high-temporal-resolution
measurements of the optical, microphysical, and radiative properties of aerosols. One of the primary data
products, columnar AOD, is used as a primary validation resource for satellite validation of AOD. The
Angstrom exponent from the measurements can be used to provide an estimate of the dominant aerosol
size within the AOD measurement (Giles et al.. 2019). Similar to AERONET, the Pandonia Global
Network (PGN) is an emerging federated global network of ground-based spectrometers lead by NASA
and the ESA and was developed for validation of trace gas column abundances from satellites such as
TROPOMI (Judd et al.. 2020; Zhao et al.. 2020). The instrument, called pandora, is a UV/VIS
spectrometer and currently provides near-real-time data products of total column O s and NOa,
tropospheric NOa, and a derived NO2 surface concentration, with tropospheric column CH2O moving
from a research data product to a standard data product in the coming year (Szvkman et al.. 2019). The
number of AERONET and PGN sites across the U.S. can vary on a year-to-year basis as both instruments
4-21
DRAFT; Do Not Cite or Quote
-------
1
2
are often used to support research field campaigns, at the end of 2020 AERONET reported approximately
100 active sites and PGN 14 active sites.
3 4.4 Ambient Air Quality Monitoring Data Availability and
4 Quality
5 4.4.1 Overview
6 Observational air quality data is used in many facets of wildland fire smoke management from
7 first-responder force protection and public health messaging where real-time data availability is critical, to
8 regulatory NAAQS review and public health research (e.g., epidemiologic studies) where delayed data
9 access is acceptable but rigorous data quality assurance/quality control (QA/QC) review is required. This
10 section discusses observational air quality data availability and relative data quality that is routinely used
11 by wildland fire smoke managers, public health officials, and researchers.
12 4.4.2 U.S. EPA Routine Regulatory Data Availability
13 As described above, near real-time measurements of PM2 5 and O3 are reported from state, local
14 and tribal air monitoring agencies to AirNow (Table A.4-3). The data are then made publicly available
15 through NowCast reporting of the AQI. The raw hourly data for PM2 5 and O3 as well as all other reported
16 real-time air pollution and meteorological parameters are stored and available to the AirNow technical
17 community through the website www.AirNowTech.org. AirNow-Tech is a password-protected website
18 for air quality data management analysis, and decision support. AirNow-Tech is primarily used by the
19 federal, state, tribal, and local air quality organizations that provide data and forecasts to the AirNow
20 system, as well as researchers and other air data users. Automated availability of large amounts of
21 AirNow data can be accomplished by registered users through accessing the AirNow application
22 programming interface (API). There are important distinctions between the AirNow data system and the
23 AQS database described below. First, to ensure real-time availability of data in AirNow, data are reported
24 as soon as practical after the end of each hour. Therefore, data are available to support forecasting and
25 reporting of the AQI but are not used for regulatory decisions until all QA/QC checks are performed and
26 validation of data is certified by the responsible state/local/tribal agency. Second, data reported to AirNow
27 include many monitoring stations for communities outside the U.S. For example, air monitoring programs
28 for Canadian Provinces and cities report their PM2 5 and O3 data to AirNow. However, data from outside
29 the U.S. are usually not reported to the AQS data system described below.
30 U.S. EPA's long-term repository of data is provided by the AQS. The AQS contains ambient air
31 pollution data collected by state, local, and tribal air pollution monitoring agencies. The data set includes
4-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
data from both automated methods reported to AirNow, but also from manual methods where data are not
available for several weeks to months due to post-sampling laboratory analysis. In addition to pollutant
concentrations and meteorological data, AQS contains descriptive information about each monitoring
station (including its geographic location and its operator), and data quality assurance/quality control
information. While data are reported to AirNow within minutes after the end of an hour, data are not
required to be reported to AQS until 90 days past the end of a calendar quarter. This lag and difference in
data reporting allow monitoring agencies the time needed to validate ambient air monitoring data for
NAAQS compliance. Data reported to both AQS and AirNow are matched on a routine basis with AQS
data overwriting any reported data to AirNow. This allows monitoring agencies the opportunity to
invalidate data in one location while ensuring validation decisions are carried through to both databases.
By May 1st of each year, monitoring agencies are required to "certify" (U.S. EPA. 2019c) their criteria
pollutant data used for NAAQS compliance determinations so that it is available to use in design value
calculations. A user friendly portal to access reports and data from AQS data is available at:
https://www .epa. gov/outdoor-air-qualitv-data .
4.4.3 Temporary/Incident Response Data Availability
The temporary PM2 5 monitors deployed by federal, state, tribal, and local agencies for incident
response typically report hourly data through satellite communications. The AirNow Fire and Smoke Map
project, a collaborative effort between IWFAQRP and U.S. EPA collects these data through the AirSis
and Western Regional Climate Center (WRCC) data feeds (AirNow. 2021b. c). Following quality
assurance checks including flow rate, internal humidity, battery levels, and measurement values within a
feasible range, the data are made available through the AirNow Fire and Smoke Map. PM data from
permanent monitors (Section 4.3.2) obtained through the U.S. EPA AirNow system (Section 4.4.2). as
well as sensors (Section 4.4.4). are also included in AirNow Fire and Smoke Map. The system also
provides the locations of large fire incidents from the U.S. National Interagency Fire Center's active
incident feed and satellite based active fire detections and smoke plume locations (Section 4.3.5.1.1') from
the NOAA Hazard Mapping System. Currently, the system functions as an operational data viewer—data
is not available for download and viewing is limited to data <10 days old. Data downloads (<10 days old)
for temporary and permanent monitors are available through the USFS AirFire V4.1 smoke monitoring
system rhttps://tools.airfire.org/monitoring; USFS (2021a)!. a predecessor to the AirNow Fire Map.
Limited historical PM data from some temporary monitors can be accessed through the WRCC
rhttps ://wrcc .dri .edu/cgi-bin/smoke ,pl; WRCC (2021)1. No comprehensive archive of temporary PM2 5
monitor data is currently available to researchers, land managers, or the public.
The www.airfire.org website provides visualization tools for ARAs to evaluate temporal and
spatial smoke trends, and how PM concentrations vary between observational surface measurements and
smoke prediction model estimates. The temporospatial trends and smoke model performance are
important for ARAs to contextualize with current fire conditions and observed smoke production during
4-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
large wildfire events. Diurnal smoke behavior is particularly important for predicting how the smoke will
impact some areas, especially when the smoke dispersion is dominated by terrain driven winds in foothill
and mountain communities. A limitation to the publicly available AirNow data is that it reflects the
NowCast concentration, not necessarily the current concentration at any given time, so it could be
anywhere from 1 to 3 hours behind providing the appropriate trend. This is important for public health
officials when tracking concentrations, especially when they are trying to provide schools and athletics
information on whether outdoor activities are safe or if air quality is remaining in the unhealthy range.
AirNow does provide another link to the information—this is the primary public-facing sites and
resources provided to better understand the trends in air quality. The forecasting reported on AirNow by
local air pollution control districts are quite often accurate for a 24-hour period; however, there is a
limitation in how the reporting area is determined. Quite often the reporting area is based on the largest
metropolitan area with an air quality forecast. This forecast may be an accurate estimate of smoke impact
if the area has uniform terrain. However, reporting areas for cities in the foothills or neighborhoods with
substantial elevation change the actual smoke concentration may be substantially different than predicted
due to terrain induced drainage flows. So even when air quality improves in the closest metropolitan area,
the smoke may linger and take longer to dissipate in certain areas which may change the 24-hour estimate
of the AQI. In foothill communities, when there are terrain-driven winds, these communities will often
see delayed AQI improvement compared to centrally located monitors because of how the smoke will
transport with upslope and downslope winds following typical diurnal patterns. The delay in AQI
improvement has been particularly evident during extended periods of high pressure over fires where
smoke continues to hang in the valleys over a period of days and sometimes weeks. Therefore, the smoke
reporting for certain areas, especially in the wildland-urban interface (WUI) and in foothill communities
provide AQI prediction challenges where actual air quality is not adequality represented by the closest
central monitoring site.
4.4.4 Sensor Data Availability
The current business use case for most commercially available air quality sensors involve either
local data storage for end user use only or vendor specific cloud-based data telemetry, storage, quality
assurance review, and graphical presentation of summary monitoring data. Most air quality sensor
manufactures that maintain cloud-based systems do so to provide secure storage and analysis tools for
each end user. 2B Tech (2021). Clarity (2021). and PurpleAir (2021) are examples of manufacturers that
do allow the end users to choose whether to keep their monitoring data private or allow for public
dissemination of their data through each manufacturers proprietary map-based web portals as part of the
sensor registration process. The Environmental Defense Fund (EDF), OpenAQ, and other
nongovernmental organizations have undertaken independent initiatives that advocate for the
development of a centralized repositories of data collected from ambient air quality sensors that includes
the development of data standards and definitions of terms with the vision of making integrated air quality
4-24
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
sensor from all manufacturers publicly available (EPF Air Sensor Workgroup. 2021; QpenAO. 2021).
The value of publicly available sensor data was demonstrated by U.S. EPA and USFS as part of their pilot
AirNow Fire and Smoke Map project in 2020 (AirNow. 2021a). The pilot used the data from a single
manufacturer (PurpleAir) due to the relatively large number of deployed sensors, documented PM2 5
sensor performance (Barkjohn et al.. 2020). and the public availability of their data. However, there is
currently no centralized publicly accessible air quality data repository from ambient sensors that are
available for wildland fire incident teams, air quality regulators, researchers, or public health officials to
access during wildland fire events.
4.4.5 Remote Sensing Data Availability
Data latency and reliable data availability are critical attributes for the use of satellite data,
particularly for air quality uses associated with smoke plume tracking and improved predictions of
pollutants distributions during active wildland fires. Operational satellite instruments such as VIIRS,
GOES-ABI and TROPOMI are designed for low data latency and reliable data availability because of the
reliance on such instruments to inform weather and air quality forecast. Such considerations are usually
not a high priority for research satellites, however the direct broadcast X-band downlink and
near-real-time science data production software International MODIS/AIRS processing package (Strabala
et al.. 2003) implemented for the MODIS sensors aboard the Terra and Aqua satellites facilitated use of
the data for tracking wildland fire plumes to improve PM2 5 forecast (Al-Saadi et al.. 2005). The
availability and latency for satellite and ground based remote sensing data is summarized in Table A.4-4
and Table A.4-5. respectively.
4.4.6 Measurement Data Quality
FRM and FEM methods include instrument design requirements, strict performance
specifications, and routine calibration and maintenance requirements. In addition, monitoring
requirements (U.S. EPA. 2019b) prescribe routine onsite auditing of instrument performance, rigorous
data quality assurance/quality control review of all regulatory measurements, and adherence to siting
criteria (e.g., distance from obstructions). Monitoring agencies carry out and perform ambient air
monitoring in accordance with the U.S. EPA's requirements and guidance as well as often meeting their
own state monitoring needs that may go beyond the minimum federal requirements. As previously stated,
air quality data obtained from state, local, and tribal monitoring sites are reported to U.S. EPA's AQS
database, along with other information, and are used for determining compliance with the NAAQS,
assessing effectiveness of mitigation strategies, characterizing local, state, and national air quality status
and trends, and associating public health outcomes with air pollution concentrations/population
exposures. Therefore, regulatory measurements are the highest quality air pollution measurements
available.
4-25
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Nonregulatory instruments used for temporary incident response measurements like the ARA
deployed E-BAM and E-SAMPLERS are maintained/calibrated to manufacturer specifications by
IWFAQRP at their Lakewood, CO facility prior to field deployment each fire season. The USFS
conducted tests on these two devices ranging from 30 to 1,700 (.ig/nr1 under smoke conditions against a
gravimetric, filter-based U.S. EPA FRM sampler (Trent. 2006. 2003). On average, both the E-SAMPLER
and the E-BAM samplers overestimated the mass concentration by approximately 13% over the FRM, yet
correlation coefficients were very high, over 0.96 and over 0.99, respectively (Trent. 2003). The
IWFAQRP instruments are received, installed, and maintained by trained ARA professionals following
established program SOPs. The data quality from other state, local, or tribal agency temporary incident
response instruments are expected to be of similar quality to IWFAQRP deployments when following
their established training and instrument SOPs.
Raw PM2 5 concentration data from air quality sensors is generally considered qualitative during
wildland fire smoke events due to the general lack of smoke specific performance testing, routine
maintenance and calibration procedures, and data QA/QC screening and validation. There is also a
recognition that certain sensor systems are better categorized by objective testing organizations such as
U.S. EPA (U.S. EPA. 2020c). the South Coast Air Quality Management District's Air Quality Sensor
Performance Evaluation Center (SCAOMD. 2021). and the European Commission Joint Research Center
(JRC. 2021). and that sensor networks deployed, characterized relative to FRM/FEM measurements, and
maintained by governmental/professional organizations may be of higher quality. Adoption of formal air
quality sensor performance targets, calibration, maintenance, data quality review guidelines, and
certification requirements that are currently being investigated by U.S. EPA (Duval 1 et al.. 2021a. b)
would provide a path forward for ensuring that future air quality sensor data would better serve the
observational air quality monitoring requirements of the wildland fire smoke management community.
4.5 Challenges in Ambient Smoke Monitoring
Wildland fire smoke events can produce extreme near-field air pollutant concentrations that
exceed monitoring instrument linear dynamic range and reporting limits, cause analytical interference(s),
and generally increase the uncertainty in reported air pollution concentrations. In many areas of the
country wildfire smoke is responsible for the highest air pollution concentration values experienced and
may dominate the local populations exposure to air pollution (e.g., PM2 5) on an annual basis. Some initial
evaluations of UV-photometric FEM O3 instruments (Landis et al.. 2017; Long et al.. In Press) and visible
spectrum FEM PM2 5 instruments (Landis et al.. 2021) have documented measurement accuracy
degradation under smoke conditions. In addition, wildland fire smoke events present many inherent
measurement, quality assurance, data latency, data integration, data availability, and communication
challenges for land management, wildland fire smoke management, air quality management, and public
health officials including:
4-26
DRAFT: Do Not Cite or Quote
-------
1 • Wildland fire events and downwind smoke impact zones occur disproportionally in areas of the
2 U.S. having diffuse population centers and lacking U.S. EPA regulatory air quality monitoring
3 infrastructure typically used to measure AQI and communicate appropriate public health
4 messages. Complex terrain and unpredictable smoke plume behavior can also complicate accurate
5 determination and spatial interpolation of AQI and the associated public health recommendations
6 for limiting smoke exposure.
7 • Wildland fire smoke can be highly spatially and temporally variable. Smoke can be confined to
8 topographic areas such as valleys or in specific vertical or meteorological layers (e.g., inversions),
9 meaning that air quality monitors only a few kilometers apart can report dramatically different
10 concentrations. Smoke concentrations can change substantially over short time periods as fire
11 activity and meteorological dispersion changes make it difficult to predict and manage hazardous
12 conditions (e.g., measured average hourly concentration values may not match the experience of
13 smoke even at that location due to subhourly temporal fluctuations).
14 • Wildland fire smoke can transport for long distances. Smoke plumes from specific wildfires have
15 been traced across continental or even oceanic/transcontinental scales. Air pollution
16 concentrations can be significantly elevated thousands of km away without an obvious connection
17 to distant fire events.
18 • The availability, validity, comparability, and integration of observational air quality
19 measurements during wildland fire events is improving (e.g., sensor data pilot, smoke modeling
20 tools); however, there is a long way to go to enable real-time (low latency), integrated, and
21 publicly available data and modeling tools that are required for effective management activities at
22 local, state, and regional scales.
23 The air quality monitoring challenges during wildland fire events are inherently linked to the
24 associated limitations in current U.S. EPA regulatory monitoring networks. The objectives of these
25 networks do not include smoke monitoring. The current network designs that prioritize densely populated
26 urban and suburban areas where most anthropogenic air pollution sources are concentrated result in a lack
27 of network site density and spatial/elevation distribution of monitors in more remote areas where wildland
28 fire events are more likely to occur. Issues with data telemetry, latency, and QA/QC review culminate to
29 create a situation where wildland fire smoke impacts are not well captured by existing regulatory
30 networks. Temporarily emplaced monitors, remote sensing, and air quality sensors offer future
31 opportunities to supplement regulatory monitoring infrastructure. However, as discussed above, these
32 observational monitoring technologies have their own issues with accuracy, reliability, and availability of
33 measured concentration values and the ability to quickly emplace and telemeter data to fill the most
34 important gaps in spatial coverage.
4.6 Recommendations
35 Currently, the fundamental understanding of wildland fire source emissions, the impact of smoke
36 on ambient air quality, the estimation of human exposures, the quantification of adverse health outcomes,
37 and the ability to develop and validate predictive deterministic air quality models are predicated on
38 accurate measurements of criteria air pollutants and their precursors in smoke. This chapter presented and
39 discussed the contemporary sources of ambient air quality monitoring data, the relative accuracy of data
4-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
sources, the latency and availability of data, and the tools for accessing and analyzing air pollution
monitoring data and smoke dispersion modeling in the U.S. While U.S. EPA's current regulatory
monitoring network objectives do not include smoke monitoring, it is evident that recent advances in
measurement technologies, cloud computing capabilities, and online data accessibility tools have
improved the national capacity to measure, predict, and disseminate public health information on smoke
impacts from wildland fire events. However, it is also clear that there are fundamental gaps in the ability
to (1) accurately measure air quality impacts from wildland fire smoke over relevant spatial and temporal
scales, (2) integrate and archive available observational data streams into common data format standards,
and (3) provide timely access to integrated data analysis and visualization tools necessary for smoke
management and public health officials to take effective control and abatement actions.
Based on these gaps, enumerated below are several actions that could help address the identified
challenges and advance national capabilities for wildland fire smoke monitoring:
• Establishment of a program to evaluate the performance of U.S. EPA designated FRM/FEM
regulatory monitors under wildland fire smoke conditions.
• Inclusion of wildland fire smoke monitoring as an air quality monitoring objective for areas of the
country with recurring wildland fire smoke impacts.
• Establishment of guidelines for the evaluation of commonly used commercially available
nonregulatory instruments and air quality sensors under wildland fire smoke conditions.
• Establishment of data and QA/QC review standards for commonly used commercially available
nonregulatory instruments and air quality sensors.
• Development of mobile air quality monitoring capabilities around wildland fire events as an
added capability for ARAs working on large incidents particularly in more remote areas with
limited existing monitoring infrastructure. Mobile wildland fire smoke measurements would
provide public health officials the means to inform the placement of their temporary stationary
monitors, evaluate the wildland fire smoke exposure risks across multiple communities, and to
provide timely and actionable public safety information.
• Collaborative effort across federal agencies (e.g., U.S. EPA, USFS, NOAA, NASA) to establish
common data sharing agreements for remote sensing data.
• Development of a publicly available cloud-based data integration and visualization platform for
all available regulatory, nonregulatory, air quality sensor, and remote-sensing data streams for
wildland fire smoke management and wildland fire smoke impact analysis. AirNow serves some
of this capacity now and could be enhanced with the suggested functionalities.
Through these actions it is possible to chart a collaborative interagency path forward in
addressing current wildland smoke monitoring challenges such as unknown accuracy of air pollution
monitors in wildland fire smoke, lack of network site density and spatial/elevation distribution of
monitors, data telemetry and latency issues, and the availability and comparability of wildland fire smoke
impacted monitoring data products. In addition, the collaborative nature of the proposed actions would
allow for the formation of a constructive community of wildland fire smoke practitioners and researchers
focused on improving the quality, integration, and availability of air quality monitoring data.
4-28
DRAFT: Do Not Cite or Quote
-------
4.7 References
2B Tech (2B Technologies). (2021). AQTreks. Available online at htto://aatreks.com/index.html (accessed
January 14, 2021).
AirNow. (2021a). AirNow. Available online at http://www.airnow.gov/ (accessed January 14, 2021).
AirNow. (2021b). AirNow Fire and Smoke Map. Available online at https://fire.airnow.gov/ (accessed February
10, 2021).
AirNow. (2021c). AirNow: Fires. Available online at https://www.airnow.gov/fires/ (accessed February 11,
2021).
Al-Saadi. J: Szvkman. J: Pierce. RB: Kittaka. C: Neil. D: Chu. DA: Remer. L: Gumlev. L: Prins. E: Weinstock.
L: Macdonald. C: Wavland. R: Dimmick. F: Fishman. J. (2005). Improving national air quality forecasts with
satellite aerosol observations. Bull Am Meteorol Soc 86: 1249-1262. http://dx.doi.org/10.1175/BAMS-86-9-
1249
Alvarado. LMA: Richter. A: Vrekoussis. M: Hilboll. A: Hedegaard. ABK: Schneising. O: Burrows. JP. (2020).
Unexpected long-range transport of glyoxal and formaldehyde observed from the Copernicus Sentinel-5
Precursor satellite during the 2018 Canadian wildfires. Atmos Chem Phys 20: 2057-2072.
http://dx.doi.org/10.5194/acp-20-2057-202Q
Alvarado. MJ: Lonsdale. CR: Yokelson. RJ: Akagi. SK: Coe. H: Craven. JS: Fischer. EV: McMeeking. GR:
Seinfeld. JH: Soni. T: Taylor. JW: Weise. PR: Wold. CE. (2015). Investigating the links between ozone and
organic aerosol chemistry in a biomass burning plume from a prescribed fire in California chaparral. Atmos
Chem Phys 15: 6667-6688. http://dx.doi.org/10.5194/acp-15-6667-2015
Apte. JS: Messier. KP: Gani. S: Brauer. M: Kirchstetter. TW: Lunden. MM: Marshall. JD: Portier. CJ:
Vermeulen. RCH: Hamburg. SP. (2017). High-resolution air pollution mapping with Google street view cars:
Exploiting big data [Abstract]. Environ Sci Technol 51: 6999-7008.
http://dx.doi.org/10.1021/acs.est.7b00891
Artursson. T: Eklov. T: Lundstrom. I: Martcnsson. P: Siostrom. M: Holmberg. M. (2000). Drift correction for
gas sensors using multivariate methods. Journal of Chemometrics 14: 711-723.
http://dx.doi.org/10.1002/1099-128X(200009/12)14:5/6<711:: AID-CEM607>3,0.CO;2-4
Ashley. WS: Strader. S: Dziubla. DC: Haberlie. A. (2015). Driving blind: Weather-related vision hazards and
fatal motor vehicle crashes. Bull Am Meteorol Soc 96: 755-778. http://dx.doi.org/10.1175/BAMS-D-14-
00026.1
Atmospheric Composition Analysis Group. (2021). Surface PM2.5 datasets. Halifax, NS: Dalhousie University.
Retrieved from http://fizz.phys.dal.ca/~atmos/martin/7page id=140
Baker. KR: Koplitz. SN: Foley. KM: Avev. L: Hawkins. A. (2019). Characterizing grassland fire activity in the
Flint Hills region and air quality using satellite and routine surface monitor data. Sci Total Environ 659:
1555-1566. http://dx.doi.Org/10.1016/i.scitotenv.2018.12.427
Barkiohn. KK: Gantt. B: Clements. AL. (2020). Development and Application of a United States wide correction
for PM2.5 data collected with the PurpleAir sensor. Atmos Meas Tech. http://dx.doi.org/10.5194/amt-202Q-
413
Baron. R: Saffell. J. (2017). Amperometric gas sensors as a low cost emerging technology platform for air
quality monitoring applications: A review [Review]. ACS Sens 2: 1553-1566.
http://dx.doi.org/10.1021/acssensors.7b0062Q
Bart. M: Williams. DE: Ainslie. B: McKendrv. I: Salmond. J: Grange. SK: Alavi-Shoshtari. M: Stevn. D:
Henshaw. GS. (2014). High density ozone monitoring using gas sensitive semi-conductor sensors in the
Lower Fraser Valley, British Columbia. Environ Sci Technol 48: 3970-3977.
http://dx.doi.org/10.1021/es404610t
4-29
DRAFT: Do Not Cite or Quote
-------
Bartolome. C: Princevac. M: Weise. PR: Mahalingam. S: Ghasemian. M: Venkatram. A: Vu. H: Aguilar. G.
(2019). Laboratory and numerical modeling of the formation of superfog from wildland fires. Fire Safety
Journal 106: 94-104. http://dx.doi.Org/10.1016/i.firesaf.2019.04.009
Caicedo. V: Delgado. R: Sakai. R: Knepp. T: Williams. D: Cavender. K: Lefer. B: Szvkman. J. (2020). An
automated common algorithm for planetary boundary layer retrievals using aerosol lidars in support of the
US EPA photochemical assessment monitoring stations program. J Atmos Ocean Tech 37: 1847-1864.
http://dx.doi.org/10.1175/JTECH-D-20-0050.1
California SB-1260. Fire prevention and protection: Prescribed burns. California SB-1260. (2018).
https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill id=201720180SB1260
Campbell. JR: Sassen. K: Welton. EJ. (2008). Elevated cloud and aerosol layer retrievals from Micropulse Lidar
Signal Profiles. J Atmos Ocean Tech 25: 685-700. http://dx.doi.org/10.1175/2007JTECHA1034.1
Castell. N: Dauge. FR: Schneider. P: Vogt. M: Lerner. U: Fishbain. B: Brodav. D: Bartonova. A. (2016). Can
commercial low-cost sensor platforms contribute to air quality monitoring and exposure estimates? Environ
Int 99: 293-302. http://dx.doi.org/10.1016/i.envint.2016.12.007
Cheeseman. M: Ford. B: Volckens. J: Lvapustin. A: Pierce. JR. (2020). The relationship between MAI AC smoke
plume heights and surface PM. Geophys Res Lett 47: e2020GL088949.
http://dx.doi.org/10.1029/2020GLQ88949
Chu. Y: Liu. Y: Li. X: Liu. Z: Lu. H: Lu. Y: Mao. Z: Chen. X: Li. N: Ren. M: Liu. F: Tian. L: Zhu. Z: Xiang. H
(2016). A review on predicting ground PM2.5 concentration using satellite aerosol optical depth [Review].
Atmosphere (Basel) 7: 129. http://dx.doi.org/10.3390/atmos7100129
Cisneros. R: Schweizer. D: Zhong. S: Hammond. K: Perez. MA: Guo. O: Traina. S: Bvtnerowicz. A: Bennett.
PH. (2012). Analysing the effects of the 2002 McNally fire on air quality in the San Joaquin Valley and
southern Sierra Nevada, California. International Journal of Wildland Fire 21: 1065-1075.
http://dx.doi.org/10.1071/WF11025
Clarity (Clarity Movement). (2021). Clarity. Available online at https://www.claritv.io/ (accessed February 10,
2021).
Clements. AL: Griswold. WG: RS. A: Johnston. JE: Herring. MM: Thorson. J: Collier-Oxandale. A: Hannigan.
M. (2017). Low-cost air quality monitoring tools: From research to practice (a workshop summary)
[Review]. Sensors 17: 2478. http://dx.doi.org/10.3390/sl7112478
Collier-Oxandale. A: Feenstra. B: Papapostolou. V: Zhang. H: Kuang. M: Boghossian. B: Polidori. A. (2020).
Field and laboratory performance evaluations of 28 gas-phase air quality sensors by the AQ-SPEC program.
Atmos Environ 220: 117092. http://dx.doi.Org/10.1016/i.atmosenv.2019.117092
Cooper. MJ: Martin. R: Henze. DK: Jones. DBA. (2020). Effects of a priori profile shape assumptions on
comparisons between satellite N02 columns and model simulations. Atmos Chem Phys 20: 7231-7241.
http://dx.doi.org/10.5194/acp-20-7231-2020
Cross. ES: Williams. LR: Lewis. DK: Magoon. GR: Onasch. TB: Kaminskv. ML: Worsnop. PR: Javne. JT.
(2017). Use of electrochemical sensors for measurement of air pollution: Correcting interference response
and validating measurements. Atmos Meas Tech 10: 3575-3588. http://dx.doi.org/10.5194/amt-10-3575-2017
Crumevrolle. S: Chen. G: Ziemba. L: Beversdorf. A: Thornhill. L: Winstead. E: Moore. RH: Shook. MA:
Hudgins. C: Anderson. BE. (2014). Factors that influence surface PM2.5 values inferred from satellite
observations: perspective gained for the US Baltimore-Washington metropolitan area during DISCOVER-
AQ. Atmos Chem Phys 14: 2139-2153. http://dx.doi.org/10.5194/acp-14-2139-2014
Delp. WW: Singer. BC. (2020). Wildfire smoke adjustment factors for low-cost and professional
PM(2.5)monitors with optical sensors. Sensors 20: 3683. http://dx.doi.org/10.3390/s2Q133683
4-30
DRAFT: Do Not Cite or Quote
-------
Duvall. R: Clements. A: Hagler. G: Kamal. A: Kilaru. V: Goodman. L: Frederick. S: Johnson Barkiohn. K:
VonWald. I: Greene. D: Dye. T. (2021a). Performance testing protocols, metrics, and target values for fine
particulate matter air sensors: Use in ambient, outdoor, fixed site, non-regulatory supplemental and
informational monitoring applications [EPA Report]. (EPA/600/R-20/280). Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development.
https://cfpub.epa.gov/si/si public record Report.cfm?dirEntrvId=350785&Lab=CEMM
Duvall. R: Clements. A: Hagler. G: Kamal. A: Kilaru. V: Goodman. L: Frederick. S: Johnson Barkiohn. K:
VonWald. I: Greene. D: Dye. T. (2021b). Performance testing protocols, metrics, and target values for ozone
air sensors: Use in ambient, outdoor, fixed site, non-regulatory and informational monitoring applications
[EPA Report]. (EPA/600/R-20/279). Washington, DC: U.S. Environmental Protection Agency, Office of
Research and Development.
https://cfpub.epa.gov/si/si public record Report.cfm?dirEntrvId=350784&Lab=CEMM
EPF Air Sensor Workgroup (Environmental Defense Fund, Air Sensor Workgroup). (2021). Air Quality Data
Commons (AQDC) [Database]. Retrieved from https://aadatacommons.org/
Feinberg. S: Williams. R: Hagler. GSW: Rickard. J: Brown. R: Garver. D: Harshfield. G: Stauffer. P: Mattsoa
E: Judge. R: Garvev. S. (2018). Long-term evaluation of air sensor technology under ambient conditions in
Denver, Colorado. Atmos Meas Tech 11: 4605-4615. http://dx.doi.org/10.5194/amt-ll-4605-2Q18
Feinberg. SN: Williams. R. on: Hagler. G: Low. J: Smith. L: Brown. R: Garver. D: Davis. M: Morton. M:
Schaefer. J. oe: Campbell. J. (2019). Examining spatiotemporal variability of urban particulate matter and
application of high-time resolution data from a network of low-cost air pollution sensors. Atmos Environ
213: 579-584. http://dx.doi.Org/10.1016/i.atmosenv.2019.06.026
GAP (U.S. General Accounting Office). (2020). Air pollution: Opportunities to better sustain and modernize the
national air quality monitoring system. (GAO-21-38). Washington, DC: U.S. Government Accountability
Office, https ://www. gao. gov/assets/gao-21 -3 8.pdf
Giles. DM: Sinvuk. A: Sorokin. MG: Schafer. JS: Smirnov. A: Slutsker. I: Eck. TF: Holben. BN: Lewis. JR:
Campbell. JR: Weltoa EJ: Korkin. SV: Lvapustia AI. (2019). Advancements in the Aerosol Robotic
Network (AERONET) Version 3 database - automated near-real-time quality control algorithm with
improved cloud screening for Sun photometer aerosol optical depth (AOD) measurements. Atmos Meas Tech
12: 169-209. http://dx.doi.org/10.5194/amt-12-169-2019
Griffin. D: Sioris. C: Chen. J: Dickson. N: Kovachik. A: de Graaf. M: Nanda. S: Veefkind. P: Dammers. E:
McLinden. CA: Makar. P: Akingunola. A. (2020). The 2018 fire season in North America as seen by
TROPOMI: Aerosol layer height intercomparisons and evaluation of model-derived plume heights. Atmos
Meas Tech 13: 1427-1445. http://dx.doi.org/10.5194/amt-13-1427-2020
Gupta. P: Christopher. SA. (2009). Particulate matter air quality assessment using integrated surface, satellite,
and meteorological products: Multiple regression approach. J Geophys Res Atmos 114: D14205.
http://dx.doi.org/10.1029/2008JD011496
Gupta. P: Doraiswamv. P: Lew. R: Pikelnava. O: Maibach. J: Feenstra. B: Polidori. A: Kiros. F: Mills. KC.
(2018). Impact of California fires on local and regional air quality: The role of a low-cost sensor network and
satellite observations. Geohealth2: 172-181. http://dx.doi.org/10.1029/2018GH00Q136
Hammer. MS: van Donkelaar. A: Li. C: Lvapustia A: Saver. AM: Hsu. NC: Lew. RC: Garav. MJ:
Kalashnikova. OV: Kahn. RA: Brauer. M: Apte. JS: Henze. DK: Zhang. L: Zhang. O: Ford. B: Pierce. JR:
Martia RV. (2020). Global estimates and long-term trends of fine particulate matter concentrations (1998-
2018). Environ Sci Technol 54: 7879-7890. http://dx.doi.org/10.1021/acs.est.0c01764
Hoff. RM: Christopher. SA. (2009). Remote sensing of particulate pollution from space: Have we reached the
Promised Land? [Review]. J Air Waste Manag Assoc 59: 645-675. http://dx.doi.org/10.3155/1047-
3289.59.6.645
Holder. AL: Mebust. AK: Maghran. LA: Mcgown. MR: Stewart. KE: Vallano. DM: Elleman. RA: Baker. KR.
(2020). Field evaluation of low-cost particulate matter sensors for measuring wildfire smoke. Sensors 20:
4796. http://dx.doi.org/10.3390/s2Q174796
4-31
DRAFT: Do Not Cite or Quote
-------
Hossain. M: Saffell. J: Baron. R. (2016). Differentiating N02 and 03 at low cost air quality amperometric gas
sensors. ACS Sens 1: 1291-1294. http://dx.doi.org/10.1021/acssensors.6b00603
IWFAORP (Interagency Wildland Fire Air Quality Response Program). (2021). Smoke COVID Dashboard.
Available online at https://covid.airfire.org/tracking/ (accessed February 10, 2021).
Jaffe. DA: Wigder. NL. (2012). Ozone production from wildfires: A critical review [Review]. Atmos Environ
51: 1-10. http://dx.doi.Org/10.1016/i.atmosenv.2011.ll.063
Jiao. W: Hagler. G: Williams. R: Sharpe. R: Brown. R: Garver. D: Judge. R: Caudill. M: Rickard. J: Davis. M:
Weinstock. L: Zimmer-Dauphinee. S: Buckley. K. (2016). Community Air Sensor Network (CAIRSENSE)
project: evaluation of low-cost sensor performance in a suburban environment in the southeastern United
States. Atmos Meas Tech 9: 5281-5292. http://dx.doi.org/10.5194/amt-9-5281-2016
Jin. X: Fiore. AM: Curci. G: Lvapustin. A: Civerolo. K: Ku. M: van Donkelaar. A: Martin. R. (2019). Assessing
uncertainties of a geophysical approach to estimate surface fine particulate matter distributions from satellite-
observed aerosol optical depth. Atmos Chem Phys 19: 295-313. http://dx.doi.org/10.5194/acp-19-295-2019
Johnson. AL: Abramson. MJ: Dennekamp. M: Williamson. GJ: Guo. Y. (2020). Particulate matter modelling
techniques for epidemiological studies of openbiomass fire smoke exposure: A review. Air Qual Atmos
Health 13: 35-75. http://dx.doi.org/10.1007/sll869-019-0Q771-z
Johnson. KK: Bergin. MH: Russell. AG: Hagler. GSW. (2018). Field test of several low-cost particulate matter
sensors in high and low concentration urban environments. Aerosol Air Qual Res 18: 565-578.
http://dx.doi.org/10.4209/aaar.2017.10.Q418
JRC (Joint Research Centre). (2021). Joint Research Centre (JRC) Publications Repository. Available online at
https://publications.irc.ec.europa.eu/repositorv/ (accessed February 11, 2021).
Judd. LM: Al-Saadi. JA: Szvkman. JJ: Valin. LC: Janz. SJ: Kowalewski. MG: Eskes. HJ: Veefkind. JP: Cede. A:
Mueller. M: Gebetsberger. M: Swap. R: Pierce. RB: Nowlan. CR: Abad. GG: Nehrir. A: Williams. D.
(2020). Evaluating Sentinel-5P TROPOMI tropospheric N02 column densities with airborne and Pandora
spectrometers near New York City and Long Island Sound. Atmos Meas Tech 13: 6113-6140.
http://dx.doi.org/10.5194/amt-13-6113-2020
Karagulian. F: Barbiere. M: Kotsev. A: Spinelle. L: Gerboles. M: Lagler. F: Redon. N: Crunaire. S: Borowiak.
A. (2019). Review of the performance of low-cost sensors for air quality monitoring. Atmosphere (Basel) 10:
506. http://dx.doi.org/10.3390/atmosl00905Q6
Krstic. N: Henderson. SB. (2015). Use of MODIS data to assess atmospheric aerosol before, during, and after
community evacuations related to wildfire smoke. Rem Sens Environ 166: 1-7.
http://dx.doi.Org/10.1016/i.rse.2015.05.017
Kunak (Kunak Technologies). (2021). Kunak Air Cloud: Air quality and noise monitoring software. Available
online at https://www.kunak.es/en/products/ambient-monitoring/air-aualitv-and-noise-software/
Landis. MS: Long. RW: Krug. J: Colon. M: Vanderpool. R: Habel. A: Urbanski. S. (2021). The U.S. EPA
Wildland Fire Sensor Challenge: Performance and evaluation of solver submitted multi-pollutant sensor
systems. Atmos Environ 247: 118165. http://dx.doi.Org/10.1016/i.atmosenv.2020.118165
Landis. MS: Edgerton. ES: White. EM: Wentworth. GR: Sullivan. AP: Dillner. AM. (2017). The impact of the
2016 Fort McMurray Horse River Wildfire on ambient air pollution levels in the Athabasca Oil Sands
Region, Alberta, Canada. Sci Total Environ 618: 1665-1676.
http://dx.doi.Org/10.1016/i.scitotenv.2017.10.008
Levelt. PF: Yen Den Oord. GHJ: Dobber. MR: Malkki. A: Visser. H: De Vries. J: Stammes. P: Lundell. JOY:
Saari. H. (2006). The ozone monitoring instrument. IEEE Trans Geosci Remote Sens 44: 1093-1101.
http://dx.doi.org/10.1109/TGRS.2006.872333
Lin. C: Gillespie. J: Schuder. MP: Duberstein. W: Beverland. I J: Heal. MR. (2015). Evaluation and calibration
of Aeroqual series 500 portable gas sensors for accurate measurement of ambient ozone and nitrogen
dioxide. Atmos Environ 100: 111-116. http://dx.doi.Org/10.1016/i.atmosenv.2014.ll.002
4-32
DRAFT: Do Not Cite or Quote
-------
Liu. X: Huev. LG: Yokelson. RJ: Selimovic. V: Simpson. 1J: Mtiller. M: Jimenez. JL: Campuzano-Jost. P:
Beversdorf. AJ: Blake. PR: Butterfield. Z: Choi. Y: Crounse. JD: Day. DA: Diskin. GS: Dubev. MK;
Fortner. E: Hanisco. TF: Hu. W: King. LE: Kleinman. L: Meinardi. S: Mikovinv. T: Onasch. TB: Palm. BB:
Peischl. J: Pollack. IB: Rverson. TB: Sachse. GW: Sedlacek. AJ: Shilling. JE: Springston. S: St Clair. JM:
Tanner. DJ: Teng. AP: Wennberg. PO: Wisthaler. A: Wolfe. GM. (2017). Airborne measurements of western
US wildfire emissions: Comparison with prescribed burning and air quality implications. J Geophys Res
Atmos 122: 6108-6129. http://dx.doi.org/10.1002/2016JDQ26315
Liu. Y: Kahn. RA: Chaloulakou. A: Koutrakis. P. (2009). Analysis of the impact of the forest fires in August
2007 on air quality of Athens using multi-sensor aerosol remote sensing data, meteorology and surface
observations. Atmos Environ 43: 3310-3318. http://dx.doi.Org/10.1016/i.atmosenv.2009.04.010
Long. RW: Whitehill. A: Habel. A: Urbanski. S: Hallidav. H: Colon. M: Kaushik. S: Landis. MS. (In Press)
Comparison of ozone measurement methods in biomass smoke: An evaluation under field and laboratory
conditions. Atmospheric Measurement Techniques Discussions, http://dx.doi.org/10.5194/amt-2020-383
Lvapustin. A: Wang. YJ: Korkin. S: Kahn. R: Winker. D. (2019). MAIAC thermal technique for smoke injection
height from MODIS. IEEE Geosci Remote Sens Lett 17: 730-734.
http://dx.doi.org/10.1109/LGRS.2019.2936332
Malings. C: Tanzer. R: Haurvliuk. A: Kumar. SPN: Zimmerman. N: Kara. LB: Presto. AA: Subramanian. R.
(2019). Development of a general calibration model and long-term performance evaluation of low-cost
sensors for air pollutant gas monitoring. Atmos Meas Tech 12: 903-920. http://dx.doi.org/10.5194/amt-12-
903-2019
Martia RV. (2008). Satellite remote sensing of surface air quality [Review]. Atmos Environ 42: 7823-7843.
http://dx.doi.Org/10.1016/i.atmosenv.2008.07.018
Mead. MI: Popoola. QAM: Stewart. GB: Landshoff. P: Calleia. M: Haves. M: Baldovi. JJ: Mcleod. MW:
Hodgson. TF: Dicks. J: Lewis. A: Cohen. J: Baron. R: Saffell. J. R.: Jones. RL. (2013). The use of
electrochemical sensors for monitoring urban air quality in low-cost, high-density networks. Atmos Environ
70: 186-203. http://dx.doi.Org/10.1016/i.atmosenv.2012.ll.060
Mehadi. A: Moosmtiller. H: Campbell. DE: Ham. W: Schweizer. D: Tarnav. L: Hunter. J. (2019). Laboratory
and field evaluation of real-time and near real-time PM2.5 smoke monitors. J Air Waste Manag Assoc 70:
158-179. http://dx.doi.org/10.1080/10962247.2019.1654Q36
Mei. L: Xue. Y: de Leeuw. G: Guang. J: Wang. Y: Li. Y: Xu. H: Yang. L: Hou. T: He. X: Wu. C: Dong. J:
Chen. Z. (2012). Corrigendum to "Integration of remote sensing data and surface observations to estimate the
impact of the Russian wildfires over Europe and Asia during August 2010" published in Biogeosciences, 8,
3771-3791, 2011 [Erratum], Biogeosciences 9: 41-45. http://dx.doi.org/10.5194/bg-9-41-2012
Melvia MA. (2015). National Prescribed Fire Use Survey Report. (Technical Report 02-15). Coalition of
Prescribed Fire Councils, http://www.prescribedfire.net/resources-links
Melvia MA. (2018). National Prescribed Fire Use Survey Report. (Technical Report 03-18). Coalition of
Prescribed Fire Councils, http://www.prescribedfire.net/resources-links
Morawska. L: Thai. PK: Liu. XT: Asumadu-Sakvi. A: Avoko. G: Bartonova. A: Bedini. A: Chai. FH:
Christensen. B: Dunbabin. M: Gao. J: Hagler. GSW: Javaratne. R: Kumar. P: Lau. AKH: Louie. PKK:
Mazaheri. M: Ning. Z: Motta. N: Mullins. B: Rahman. MM: Ristovski. Z: Shafiei. M: Tiondronegoro. D:
Westerdahl. D: Williams. R. (2018). Applications of low-cost sensing technologies for air quality monitoring
and exposure assessment: How far have they gone? [Review]. Environ Int 116: 286-299.
http://dx.doi.Org/10.1016/i.envint.2018.04.018
Mui. W: DerBoghossian. B: Collier-Oxandale. A: Boddeker. S: Low. J: Papapostolou. V: Polidori. A. (2021).
Development of a performance evaluation protocol for air sensors deployed on a Google street view car.
Environ Sci Technol 55: 1477-1486. http://dx.doi.org/10.1021/acs.est.0c05955
NASA (National Aeronautics and Space Administration). (2021). MPLNET: The NASA Micro-Pulse Lidar
Network. Available online at https://mplnet.gsfc.nasa.gov/ (accessed January 26, 2021).
4-33
DRAFT: Do Not Cite or Quote
-------
Navarro. KM: Cisneros. R: O'Neill. SM: Schweizer. D: Larkin. NK: Balmes. J. R. (2016). Air-quality impacts
and intake fraction of PM2.5 during the 2013 Rim Megafire. Environ Sci Technol 50: 11965-11973.
http://dx.doi.org/10.1021/acs.est.6b02252
NOAA (National Oceanic and Atmospheric Administration). (2020). Hazard Mapping System Fire and Smoke
Product. Available online at https://www.ospo.noaa.gov/Products/land/hms.html (accessed February 11,
2021).
QpenAO. (2021). OpenAQ. Available online at https://openaa.org/#/ (accessed February 11, 2021).
Pasch. AN: Zahn. PH: DeWinter. JL: Haderman. MP: Dye. TS: Szvkman. JJ: White. JE: Dickerson. P: van
Donkelaar. A: Martin. RV. (2013). AirNow satellite data processor: Improving EPA's AirNow air quality
index maps using NASA/NOAA satellite data and air quality model predictions. Poster presented at 12th
Annual CMAS Conference, October 28-30, 2013, Chapel Hill, NC.
Pearson. C. (2021). Description of CARB smoke monitoring [Personal communication], Pearson, C.
PurpleAir. (2021). PurpleAir: Real-time air quality monitoring. Available online at https://www2.purpleair.com/
(accessed February 10, 2021).
Raffuse. SM: Craig. KJ: Larkin. NK: Strand. TT: Sullivan. DC. oe: Wheeler. NJM: Solomon. R. (2012). An
evaluation of modeled plume injection height with satellite-derived observed plume height. Atmosphere
(Basel) 3: 103-123. http://dx.doi.org/10.3390/atmos3010103
Raffuse. SM: McCarthy. MC: Craig. KJ: DeWinter. JL: Jumbam. LK: Fruin. S: Gauderman. WJ: Lurmann. FW.
(2013). High-resolution MODIS aerosol retrieval during wildfire events in California for use in exposure
assessment. J GeophysRes Atmos 118: 11242-11255. http://dx.doi.org/10.1002/igrd.50862
Rappold. A: Stone. SL: Cascio. WE: Neas. LM: Kilaru. VJ: Carrawav. MS: Szvkman. JJ: Ising. A: Cleve. WE:
Meredith. JT: Vaughan-Batten. H: Devneka. L: Devlin. RB. (2011). Peatbog wildfire smoke exposure in
rural North Carolina is associated with cardiopulmonary emergency department visits assessed through
syndromic surveillance. Environ Health Perspect 119: 1415-1420. http://dx.doi.org/10.1289/ehp. 1003206
Reid. CE: Jerrett. M: Petersen. ML: Pfister. GG: Morefield. PE: Tager. IB: Raffuse. SM: Balmes. JR. (2015).
Spatiotemporal prediction of fine particulate matter during the 2008 northern California wildfires using
machine learning. Environ Sci Technol 49: 3887-3896. http://dx.doi.org/10.1021/es5Q5846r
Reid. CE: Jerrett. M: Tager. IB: Petersen. ML: Mann. JK: Balmes. JR. (2016). Differential respiratory health
effects from the 2008 northern California wildfires: A spatiotemporal approach. Environ Res 150: 227-235.
http://dx.doi.Org/10.1016/i.envres.2016.06.012
Reisen. F: Duran. SM: Flannigan. M: Elliott. C: Rideout. K. (2015). Wildfire smoke and public health risk
[Review]. International Journal of Wildland Fire 24: 1029-1044. http://dx.doi.org/10.1071/WF15Q34
Savahi. T: Butterfield. A: Kelly. KE. (2019). Long-term field evaluation of the Plantower PMS low-cost
particulate matter sensors. Environ Pollut 245: 932-940. http://dx.doi.Org/10.1016/i.envpol.2018.ll.065
SCAOMD (South Coast Air Quality Management District). (2021). Air Quality Sensor Performance Evaluation
Center (AQ-SPEC). Available online at http://www.aamd.gov/aa-spec/ (accessed February 11, 2021).
Schweizer. D: Cisneros. R: Shaw. G. (2016). A comparative analysis of temporary and permanent beta
attenuation monitors: The importance of understanding data and equipment limitations when creating PM2.5
air quality health advisories. Atmos Pollut Res 7: 865-875. http://dx.doi.Org/10.1016/i.apr.2016.02.003
Spinelle. L: Gerboles. M: Aleixandre. M. (2015). Performance evaluation of amperometric sensors for the
monitoring of 03 and N02 in ambient air at ppb level. In G Urban; J Wollenstein; J Kieninger (Eds.),
Eurosensors 2015 (pp. 480-483). Amsterdam, Netherlands: Elsevier Science.
http://dx.doi.Org/10.1016/i.proeng.2015.08.676
Stowell. JD: Geng. G: Saikawa. E: Chang. HH: Fu. J: Yang. CE: Zhu. O: Liu. Y: Strickland. MJ. (2019).
Associations of wildfire smoke PM2.5 exposure with cardiorespiratory events in Colorado 2011-2014.
Environ Int 133: 105151. http://dx.doi.org/10.1016/i.envint.2019.105151
4-34
DRAFT: Do Not Cite or Quote
-------
Strabala. KI: Gumlev. LE; Rink. T: Huang. HL: Dengel. R. (2003). MODIS direct broadcast products and
applications. Proc SPIE 4891: 402-412. http://dx.doi.org/10.1117/12.466347
Szvkman. J: Swap. R: Lefer. B: Valin. L: Lee. S: Fioletov. V: Zhao. X: Davies. J: Williams. D: Abuhassan.
N: Shalabv. L: Cede. A: Tiefengraber. M: Mueller. M: Kotsakis. A: Santos. F: Robinson. J. (2019).
Pandora connecting in-situ and satellite monitoring in support of the Canada-U.S. Air Quality Agreement
[Magazine]. EM Magazine, 2019.
Thoma. ED: Brantley. HL: Oliver. KD: Whitaker. DA: Mukeriee. S: Mitchell. B: Wu. T: Sauier. B: Escobar. E:
Cousett. TA: Gross-Davis. CA: Schmidt. H: Sosna. D: Weiss. H. (2016). South Philadelphia passive sampler
and sensor study. J Air Waste Manag Assoc 66: 959-970. http://dx.doi.org/10.1080/10962247.2016.1184724
Trent. A. (2003). Laboratory evaluation of real-time smoke particulate monitors. (0325-2834-MTDC). Missoula,
MT: U.S. Department of Agriculture, Forest Service, Technology & Development Program.
https://www.fs.fed.us/t-d/pubs/pdfpubs/pdf03252834/pdf03252834dpi72.pdf
Trent. A. (2006). Smoke particulate monitors: 2006 update. (0625-2842-MTDC). Missoula, MT: U.S.
Department of Agriculture, Forest Service, Technology & Development Program, https://www.fs.fed.us/t-
d/pubs/pdfpubs/pdf06252842/pdf06252842dpi300.pdf
U.S. EPA. Significant harm levels. 40 CFR § 51.151 (2001). https://www.govinfo.gov/app/details/CFR-2001 -
title40-vol2/CFR-2001 -title40-vol2-sec51-151
U.S. EPA. Measurement principle and calibration procedure for the measurement of nitrogen dioxide in the
atmosphere (gas phase chemiluminescence), 40 CFR pt. 50, app. F (201 la).
https://www.govinfo.gov/app/details/CFR-2011-title40-vol2/CFR-2011-title40-vol2-part50-appF
U.S. EPA. Measurement principle and calibration procedure for the measurement of ozone in the atmosphere. 40
CFR pt. 50. app. D (201 lb). https://www.govinfo.gov/app/details/CFR-2011 -title40-vol2/CFR-2011 -title40-
vol2-part50-appD
U.S. EPA. Reference measurement principle and calibration procedure for the measurement of sulfur dioxide in
the atmosphere (ultraviolet fluorescence method), 40 CFRpt. 50, app. A-l (2011c).
https://www.govinfo.gov/app/details/CFR-2011-title40-vol2/CFR-2011-title40-vol2-part50-appA
U.S. EPA. Probe and monitoring path siting criteria for ambient air quality monitoring. 40 CFR pt. 58. app. E
(2012). https://www.govinfo.gov/app/details/CFR-2012-title40-vol6/CFR-2012-title40-vol6-part58-appE
U.S. EPA. Fine particulate matter (PM2.5) design criteria: General Requirements: Neighborhood scale, 40 CFR
pt. 58, app. D, 4.7.1(c)(3) (2015a). https://www.govinfo.gov/app/details/CFR-2015-title40-vol6/CFR-2015-
title40-vol6-part58-appD
U.S. EPA. Fine particulate matter (PM2.5) design criteria: Specific design criteria for PM2.5: At least one
monitoring station is to be sited at neighborhood or larger scale in an area of expected maximum
concentration, 40 CFRpt. 58, app. D, 4.7.1(b)(1) (2015b). https://www.govinfo.gov/app/details/CFR-2015-
title40-vol6/CFR-2015-title40-vol6-part58-appD
U.S. EPA. Network design criteria for ambient air quality monitoring. 40 CFR pt. 58. app. D (2015c).
https://www.govinfo.gov/app/details/CFR-2015-title40-vol6/CFR-2015-title40-vol6-part58-appD
U.S. EPA. National primary and secondary ambient air quality standards. 40 CFR pt. 50 (2016).
https://www.govinfo.gov/app/details/CFR-2016-title40-vol2/CFR-2016-title40-vol2-part5Q
U.S. EPA. Ambient air monitoring reference and equivalent methods. 40 CFR pt. 53 (2019a).
https://www.govinfo.gov/app/details/CFR-2019-title40-vol6/CFR-2019-title40-vol6-part53
U.S. EPA. Ambient air quality surveillance. 40 CFR pt. 58 (2019b). https://www.govinfo.gov/app/details/CFR-
2019-title40-vol6/CFR-2019-title40-vol6-part58
U.S. EPA. Annual air monitoring data certification. 40 CFR § 58.15 (2019c).
https://www. govinfo. gov/app/details/CFR-2019-title40-vol6/CFR-2019-title40-vol6-sec58-15
4-35
DRAFT: Do Not Cite or Quote
-------
U.S. EPA (U.S. Environmental Protection Agency). (2019d). Exceptional events guidance: Prescribed fire on
wildland that may influence ozone and particulate matter concentrations [EPA Report]. (EPA-457/B-19-004).
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, https://www.epa.gov/air-aualitv-analvsis/exceptional-events-guidance-prescribed-fire-wildland-
mav-influence-ozone-and
U.S. EPA (U.S. Environmental Protection Agency). (2019e). Wildfire smoke: A guide for public health officials,
revised 2019 [EPA Report]. (EPA-452/R-19-901). Washington, DC: U.S. Environmental Protection Agency,
Office of Research and Development, https://www.airnow.gov/publications/wildfire-smoke-guide/wildfire-
smoke-a-guide-for-public-health-officials/
U.S. EPA (U.S. Environmental Protection Agency). (2020a). Air data: Air quality data collected at outdoor
monitors across the US. Available online at https://www.epa.gov/outdoor-air-qualitv-data (accessed February
11,2021).
U.S. EPA (U.S. Environmental Protection Agency). (2020b). Air sensor toolbox: Air sensor loan programs.
Available online at https://www.epa.gov/air-sensor-toolbox/air-sensor-loan-programs (accessed February 10,
2021).
U.S. EPA (U.S. Environmental Protection Agency). (2020c). Air Sensor Toolbox: Evaluation of emerging air
sensor performance. Available online at https://www.epa.gov/air-sensor-toolbox/evaluation-emerging-air-
sensor-performance (accessed February 11, 2021).
U.S. EPA (U.S. Environmental Protection Agency). (2020d). How is the NowCast algorithm used to report
current air quality? Available online at
https://usepa.servicenowservices.com/airnow?id=kb article&svsparm article=KB0011856
U.S. EPA (U.S. Environmental Protection Agency). (2020e). List of designated reference and equivalent
methods [EPA Report]. Research Triangle Park, NC. https://www.epa.gov/sites/production/files/2019-
08/documents/designated reference and-eauivalent methods.pdf
U.S. EPA. Measurement principle and calibration procedure for the measurement of carbon monoxide in the
atmosphere (non-dispersive infrared photometry), 40 CFR pt. 50, app. C (2020f).
https://www.govinfo.gov/app/details/CFR-2020-title40-vol2/CFR-2020-title40-vol2-part5Q-appC
U.S. EPA (U.S. Environmental Protection Agency). (2020g). Our Nation's Air 2020 [EPA Report]. Washington,
DC. https://gispub.epa. gov/air/trendsreport/2020/
U.S. EPA. Reference method for the determination of fine particulate matter as PM2.5 in the atmosphere. 40
CFR pt. 50. app. L (2020h). https://www.govinfo.gOv/app/details/CFR-2020-title40-vol2/CFR-2020-title40-
vol2-part50-appL
U.S. EPA. Reference method for the determination of lead in total suspended particulate matter. 40 CFR pt. 50.
app. G (2020i). https ://www. govinfo. gov/app/details/CFR-2020-title40-vol2/CFR-2020-title40-vol2-part50-
appG
U.S. EPA. Reference method for the determination of particulate matter as PM10 in the atmosphere. 40 CFR pt.
50. app. J (2020j). https://www.govinfo.gov/app/details/CFR-2020-title40-vol2/CFR-2020-title4Q-vol2-
part50-appJ
U.S. EPA. Reference method for the determination of sulfur dioxide in the atmosphere (pararosaniline method),
40 CFR pt. 50, app. A-2 (2020k). https://www.govinfo.gov/app/details/CFR-2020-title40-vol2/CFR-202Q-
title40-vol2-part50-appA-id28
U.S. EPA (U.S. Environmental Protection Agency). (2021a). Air Sensor Toolbox. Available online at
https://www.epa.gov/air-sensor-toolbox (accessed February 10, 2021).
U.S. EPA (U.S. Environmental Protection Agency). (2021b). National Emissions Inventory (NEI). Available
online at https://www.epa.gov/air-emissions-inventories/national-emissions-inventorv-nei (accessed February
10, 2021).
4-36
DRAFT: Do Not Cite or Quote
-------
UMBC. EPA.. NASA & NOAA. (University of Maryland Baltimore County, U.S. Environmental Protection
Agency, National Aeronautics and Space Administration, and National Atmospheric and Oceanic
Administration). (2021). Unified Ceilometer Network (UCN). Available online at https://alg.umbc.edu/ucn/
(accessed January 26, 2021).
Urbanski. S. (2014). Wildlandfire emissions, carbon, and climate: Emission factors. For Ecol Manage 317: 51-
60. http://dx.doi.Org/10.1016/i.foreco.2013.05.045
USFS (U.S. Forest Service). (2020a). Interagency Wildland Fire Air Quality Response Program (IWFAQRP).
Available online at https://www.wildlandfiresmoke.net
USFS (U.S. Forest Service). (2021a). Airfire V4.1 smoke monitoring system. Available online at
https://tools.airfire.org/monitoring (accessed January 26, 2021).
USFS (U.S. Forest Service). (2021b). AirFire: Wildland fire/air quality tools. Available online at
https://portal.airfire.org/ (accessed February 10, 2021).
USFS. FWS.. BLM & NPS. (U.S. Forest Service, U.S. Fish and Wildlife Service, Bureau of Land Management,
and National Park Service). (2020b). Interagency Real Time Smoke Monitoring (IRTSM). Available online
at https://app.airsis.com/USFS/ (accessed February 10, 2021).
van Donkelaar. A: Martia RV: Li. C: Burnett. RT. (2019). Regional estimates of chemical composition of fine
particulate matter using a combined geoscience-statistical method with information from satellites, models,
and monitors. Environ Sci Technol 53: 2595-2611. http://dx.doi.org/10.1021/acs.est.8b06392
Villa. TF: Gonzalez. F: Miliievic. B: Ristovski. ZD: Morawska. L. (2016). An overview of small unmanned
aerial vehicles for air quality measurements: Present applications and future prospectives. Sensors 16.
http://dx.doi.org/10.3390/sl6071072
Wei. P: Ning. Z: Ye. S: Sun. L: Yang. F: Wong. KC: Westerdahl. D: Louie. PKK. (2018). Impact analysis of
temperature and humidity conditions on electrochemical sensor response in ambient air quality monitoring.
Sensors 18: 59. http://dx.doi.org/10.3390/sl8020Q59
Wielicki. BA: Cess. RD: King. MP: Randall. DA: Harrison. EF. (1995). Mission to planet Earth: Role of clouds
and radiation in climate [Review]. Bull Am Meteorol Soc 76: 2125-2153. http://dx.doi.org/10.1175/1520-
0477(1995)076<2125:MTPERO>2.0.CQ:2
Williams. R: Duvall. R: Kilaru. V: Hagler. G: Hassinger. L: Benedict. K: Rice. J: Kaufman. A: Judge. R: Pierce.
G: Allen. G: Bergin. M: Cohen. RC: Fransioli. P: Gerboles. M: Habre. R: Hannigan. M: Jack. D: Louie. P:
Martia NA: Penza. M: Polidori. A: Subramanian. R: Rav. K: Schauer. J: Seto. E: Thurstoa G: Turner. J:
Wexler. AS: Ning. Z. (2019). Deliberating performance targets workshop: Potential paths for emerging
PM2.5 and 03 air sensor progress. Atmospheric Environment: X 2: 100031.
http://dx.doi.Org/10.1016/i.aeaoa.2019.100031
Williams. R: Kaufman. A: Hanlev. T: Rice. J: Garvev. S. (2015). Evaluation of field-deployed low cost PM
sensors [EPA Report]. (EPA/600/R-14/464). Research Triangle Park, NC: U.S. Enviromental Protection
Agency, Office of Research and Development, National Exposure Research Laboratory.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults/titleDetail/PB20151021Q4.xhtml
Winker. DM: Pelon. J: Coaklev. JA. Jr: Ackerman. SA: Charlson. RJ: Colarco. PR: Flamant. P: Fu. O: Hoff.
RM: Kittaka. C: Kubar. TL: Le Treut. H: McCormick. MP: Megie. G: Poole. L: Powell. K: Trepte. C:
Vaughan. MA: Wielicki. BA. (2010). The CALIPSO mission: A global 3D view of aerosols and clouds. Bull
Am Meteorol Soc 91: 1211-1229. http://dx.doi.org/10.1175/2010BAMS3009.1
WRCC (Western Regional Climate Center). (2021). Fire Cache Smoke Monitor Archive. Available online at
https://wrcc.dri.edu/cgi-bin/smoke.pl (accessed January 26, 2021).
Wu. Y: Arapi. A: Huang. J: Gross. B: Mosharv. F. (2018). Intra-continental wildfire smoke transport and impact
on local air quality observed by ground-based and satellite remote sensing in New York City. Atmos Environ
187: 266-281. http://dx.doi.Org/10.1016/i.atmosenv.2018.06.006
4-37
DRAFT: Do Not Cite or Quote
-------
Zamora. ML: Xiong. FLZ: Gentner. D: Kerkez. B: Kohrman-Glaser. J: Koehler. K. (2019). Field and laboratory
evaluations of the low-cost Plantower particulate matter sensor. Environ Sci Technol 53: 838-849.
http://dx.doi.org/10.1021/acs.est.8b05174
Zhao. X: Griffig D: Fioletov. V: Mclinden. C: Cede. A: Tiefengraber. M: Mueller. M: Bognar. K: Strong. K:
Boersma. F: Eskes. H: Davies. J: Qgyu. A: Lee. SC. (2020). Assessment of the quality of TROPOMI high-
spatial-resolution N02 data products in the Greater Toronto Area. Atmos Meas Tech 13: 2131-2159.
http://dx.doi.org/10.5194/amt-13-2131-2020
Zheng. T: Bergia MH: Johnson. KK: Tripathi. SN: Shirodkar. S: Landis. MS: Sutaria. R: Carlson. DE. (2018).
Field evaluation of low-cost particulate matter sensors in high-and low-concentration environments. Atmos
Meas Tech 11: 4823-4846. http://dx.doi.org/10.5194/amt-ll-4823-2018
Zoogman. P: Liu. X: Suleiman. RM: Pennington. WF: Flittner. DE: Al-Saadi. JA: Hilton. BB: Nicks. DK:
Newchurch. MJ: Carr. JL: Janz. SJ: Andraschko. MR: Arola. A: Baker. BP: Canova. BP: Miller. CC: Cohen.
RC: Davis. JE: Dussault. ME: Edwards. DP: Fishman. J: Ghulam. A: Gonzalez Abad. G: Grutter. M:
Herman. JR. JR: Houck. J: Jacob. DJ: Joiner. J: Kerridge. BJ: Kim. J: Krotkov. NA: Lamsal. L: Li. C:
Lindfors. A: Martin. RV: McElrov. CT: McLindea C: Natrai. V: Neil. DO: Nowlan. CR: O'Sullivan. EJ:
Palmer. PI: Pierce. RB: Pippin. MR: Saiz-Lopez. A: Spurr. RJD: Szvkman. JJ: Torres. O: Veefkind. JP:
Veihelmann. B: Wang. H: Wang. J: Chance. K. (2017). Tropospheric emissions: Monitoring of pollution
(TEMPO). J Quant Spectrosc Radiat Transf 186: 17-39. http://dx.doi.Org/10.1016/i.iasrt.2016.05.008
Zou. Y: Young. M: Chen. J: Liu. J: May. A: Clark. JD. (2020). Examining the functional range of commercially
available low-cost airborne particle sensors and consequences for monitoring of indoor air quality in
residences. Indoor Air 30: 213-234. http://dx.doi.org/10. Ill 1/ina. 12621
4-38
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
CHAPTER 5 AIR QUALITY MODELING OF
WILDLAND FIRE
5.1 Background
Wildland fires (i.e., prescribed fire and wildfire) directly emit fine particulate matter (PlVfv
particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |im ) and gaseous
pollutants emitted from fires can also form secondary PM2 5 and ozone (O3) in the atmosphere (Prichard et
al.. 2019; Urbanski. 2014; Hu et al.. 2008). Estimating emissions and concentrations of pollutants formed
from wildfires is challenging due to variability in fuel consumed, fuel types, fuel moisture, plume
dynamics, and complex nonlinear chemistry (Prichard et al.. 2019; Jiang et al.. 2012). A realistic
characterization of O3 and other secondary pollutant formation in a wildfire plume is also dependent on
capturing the plume's surrounding chemical and physical environment, factors that evolve as the plume
moves further downwind from the fire.
Prescribed fire is a relatively efficient, cost-effective tool implemented by land managers for a
range of uses, including ecosystem maintenance (Kobziar et al.. 2015) and wildfire mitigation (Prichard et
al.. 2010). While the use of prescribed fire as a land management tool is common in some parts of the
contiguous U.S., both the specific land management goals and the response of the landscape to prescribed
fire can vary significantly (Ryan et al.. 2013). For these reasons, it has been historically difficult to
synthesize both the environmental trade-offs between wildfire and prescribed fire, as well as the
behavioral influence of prescribed fire on wildfire activity (e.g., changes in intensity, risk of ignition, fire
size, etc.).
This chapter presents a novel analysis evaluating air quality trade-offs across multiple fire
management strategies for two wildfires: Timber Crater 6 (TC6) Fire in 2018 and Rough Fire in 2015.
CHAPTER 3 contains general details and maps describing these fires. In both cases, detailed alternative
burn scenarios were developed with fuel information from multiple sources. Actual and alternative burn
scenarios were then simulated with air quality modeling to estimate surface concentrations of PM2 5 and
O3. Comparing the air quality impacts across different burn scenarios for each fire case study offers
insights into relative air quality impacts from hypothetical land management approaches, although
downwind transport and resulting air quality impacts near population centers can be strongly influenced
by locally specific features like terrain and meteorology.
5.1.1 Emissions of Wildland Fires
The relative amounts and chemical composition of emissions depend upon the fuel
characteristics, combustion conditions, and meteorological conditions (Urbanski. 2014). Additionally,
5-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
these factors are interrelated; for example, the combustion intensity is dependent on the meteorological
conditions (temperature, relative humidity and wind conditions) and the fuel characteristics [structure,
moisture, and loading; Surawski et al. (2015)1. Meteorology can also modulate combustion conditions,
with strong winds increasing the rate and extent of spread and peak heat release rate (i.e., intensity).
The modified combustion efficiency, defined as MCE = excess CCh/fexcess CO + excess CO2], is
widely used as an indicator of combustion conditions. MCEs greater than 0.9 are generally considered
flaming dominated and lower MCEs are smoldering dominated. Grasses and other fine fuels (<1/4"
diameter and large surface to volume ratios) tend to burn in the flaming phase. Coarse wood, duff, and
organic soils tend to burn in the smoldering phase. Fuel loading, density, and geometry also impact the
combustion phase (e.g., densely packed fine fuels will smolder). Many wildland fires burn in landscapes
with a variety of fuel types, structures, and moisture content and will therefore exhibit both flaming and
smoldering conditions simultaneously.
The fuel moisture content is a critical factor that impacts combustion conditions. Energy is lost in
evaporating the water in the fuel rather than volatilizing fuel components needed to sustain flaming
combustion. Fuels with higher moisture content take longer to ignite, may smolder before transitioning to
flaming, have shorter flaming and longer smoldering durations, lower peak heat release rate, as well as
lower and more variable fuel consumption (Possell and Bell. 2013; Chen et al.. 2010). Additionally, the
moisture content impacts the composition of the emissions, carbon monoxide (CO), volatile organic
compound (VOC), ammonia (NH3), and particulate matter (PM) emission factors increase while carbon
dioxide (CO2), nitrogen oxides (NOx), and elemental carbon (EC) generally decrease with increasing
moisture (May et al.. 2019; Tihav-Felicelli et al.. 2017; Chen et al.. 2010). PM emissions are especially
sensitive to fuel moisture. PM emission factors can be larger than CO emission factors for some fine fuels
(e.g., litter, pine needles, etc.) at high fuel moistures [e.g., above 60% dry basis; Chen et al. (2010)1.
Most emission factor compilations group emission factors by ecoregions to aggregate the impacts
of fuel chemistry, structure, and to some extent moisture (Prichard et al.. 2020; Andreae. 2019; Akagi et
al.. 2011). The emissions model will then predict how much fuel is consumed (or emitted) during the
flaming or smoldering phases. The Smoke Emissions Reference Application (SERA) described in
Prichard et al. (2020) is the most extensive compilation of smoke emission factors for North American
fires to date. However, knowledge gaps persist for emissions factors for wildland fires as summarized by
rPrichard et al. (2020); derived from Figure 1, Table 3], resulting in limited information with respect to:
• Wildfire emission factors: Emission factors are predominantly from laboratory studies (72% of
the observations); field data are 85% from prescribed fires and 15% from wildfires.
• Smoldering emission factors: Smoldering emission factors account for 31% of the prescribed fire
observations and 50% of wildfire observations, but wildfires have no emission factors for
long-term (residual) smoldering conditions.
• Fuels that tend to smolder: Most emission factor observations are from western conifer forests,
eastern conifer forests, and shrublands, but there are few observations for duff, coarse woody
debris, and peat from these regions.
5-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
• PM and VOC speciation: PM composition data is largely limited to black carbon and limited
VOC data exists, particularly for field data; most emission factor observations are the major
pollutants of CO, CO2, methane (CH4), and PM2 5, but a range of compounds have over
100 observations (propene, acetylene, methanol, formaldehyde, NH3, NO, NO2, NOx, hydrogen
cyanide [HCN], sulfur dioxide [SO2]).
In comparing emissions from wildfires and prescribed fires the different meteorological
conditions, potentially different fuels, and combustion conditions mean that emission factors will be
different for each type of fire, even in the same region. For example, Urbanski (2014) compared MCEs
for wildfires and prescribed fires in northwest conifer forests and found an average MCE of 0.883 ± 0.010
for wildfires and 0.935 ± 0.017 for prescribed fires. However, the type of fuel that is consumed may be an
important factor because prescribed fires and wildfires in the northern Rocky Mountains both had lower
MCEs (-0.87) that were attributed to a larger fraction of coarse woody debris (Urbanski. 2013). While
meteorological, fuel, and combustion parameters are factored into emissions estimation, there is still a
need to understand whether emissions modeling systems accurately capture the differences between fire
types.
5.1.2 Using Air Quality Models to Estimate Wildland Fire PM2.5 and
Ozone Impacts
Quantifying the contribution of wildland fire to ambient O3 and PM2 5 is important for air quality
alerts, air quality mitigation programs, and multiple regulatory programs including National Ambient Air
Quality Standards (NAAQS) and Regional Haze. It is important to understand how wildland fires impact
air quality and regional haze so that anticipated changes in land management (i.e., more prescribed fires)
could potentially minimize air quality degradation while still meeting ecological goals as well as
potentially reducing the impact of wildfire.
Photochemical grid models can provide information about how air quality would change based on
changes in emissions due to different types of land management choices (Hu et al.. 2008). The
Community Multiscale Air Quality [CMAQ; www.epa.gov/cmaq; U.S. EPA (2020a)l model includes
emissions, chemical reactions, and physical processes such as deposition and transport. The CMAQ
model has been used to estimate the air quality impact of wildland fires as a collective source group
(Kelly et al.. 2019) and for specific fires (Baker et al.. 2018; Zhou et al.. 2018; Baker et al.. 2016).
Photochemical grid models provide continuous spatial and temporal estimates of smoke impacts
from wildfires, which is particularly useful in areas not covered by ambient measurements (O'Dell et al..
2019). However, fire behavior and associated smoke characteristics can vary substantially by region
(Brev et al.. 2018). Accurately representing wildfire smoke in models for different geographic areas is an
ongoing effort that will continue to be important as landscapes evolve due to climate change and human
development (Ford et al.. 2018; Liu et al.. 2016; Yue et al.. 2013). Detailed case studies, like the two in
this report, provide some constraints on the representation of wildfire smoke in models for specific areas,
5-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
but more work is required to improve these estimates at both regional and global scales (Liu et al.. 2020;
Garcia-Menendez et al.. 2014. 2013).
Previous applications of CMAQ for specific fire plumes show a reasonable representation of local
to continental scale transport (Kelly et al.. 2019; Baker etal.. 2016). The modeling system treatment of
plume rise and transport is best when supplied with accurate activity data including fire size and timing
(Baker et al.. 2018; Zhou et al.. 2018). Performance related to PM2 5 impacts from wildland fire are mixed
and do not seem systematically biased high or low (Baker et al.. 2018; Koplitz et al.. 2018; Wilkins et al..
2018; Baker et al.. 2016). This modeling system tends to overestimate O3 impacts from wildland fire at
the surface (Baker et al.. 2018; Baker et al.. 2016). Predicting wildland fire impacts on O3 is challenging
because formation can be highly variable in time and space. Fresh nitric oxide emissions at the fire tend to
inhibit O3 formation as chemical destruction reactions outpace production. As the plume moves further
downwind O3 may be formed at the edges of the plume where sunlight and precursors are abundant.
Atmospheric transport processes are also important as O3 may be formed in smoke plumes but not
necessarily mix to the surface.
5.1.3 Case Study: Timber Crater 6 (TC6) Fire
The TC6 Fire burned approximately 3,000 acres in Crater Lake National Park from July 21 to
July 26, 2018. The fire covered lands managed by multiple Federal agencies. This fire was chosen as a
case study for this report because land managers in the area determined that reduced fuel loading from
previously managed land slowed fire progression enough to allow for successful suppression
(e.g., burning out fire lines). As a result, the TC6 Fire had a smaller total area burned than might have
occurred without those suppression efforts.
Three hypothetical scenarios, as detailed in 0, were developed to examine the air quality impacts
of different fire management strategies compared to the actual TC6 Fire. Scenario 1 assumed a smaller
and shorter duration fire than the actual fire, attributed to less fuel from more intensive land management
(Figure 5-1). Scenarios 2a and 2b assume more fuel in the area due to a lack of past land management.
Both Scenarios 2a and 2b are larger than the actual fire and are longer in duration than the actual fire.
Hypothetical Scenario 2b is the largest fire extending outward to a contingency perimeter where fire
suppression would be aided by roadways and other existing fire breaks. All of these scenarios used fuel
data based on a consistent approach which is described in the following sections.
5-4
DRAFT: Do Not Cite or Quote
-------
Scenario 1
1 s
1
v
*
% H/Tr
«
j
\
V
\
*
\
- N \
~
Day 1 hypothetical perimeter
Day 2 hypothetical perimeter
Day 3 hypothetical perimeter
Actual TC6 perimeter
buppression coniainment perimeter
TC6 = Timber Crater 6.
Note: The fire perimeter of the actual Timber Crater 6 Fire is also shown as the dashed line. The solid gray outline shows the fire
suppression contingency perimeter which is considered the maximum extent of wildfires in this area. The total area assumed burned
with Scenario 1 is delineated by the Day 3 perimeter.
Figure 5-1 Daily fire perimeters for the smaller Timber Crater 6 (TC6)
hypothetical fire (Scenario 1).
Each of the hypothetical scenarios were based on expert judgement of land managers familiar
with Crater Lake National Park, the fuels in the area, meteorology during the TC6 Fire, existing fire
breaks (e.g., roadways), and additional suppression techniques that would have been employed if the fire
5-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
had spread faster than the actual fire. Actual fire perimeters from the TC6 Fire were used for the first
3 days of the larger hypothetical scenarios with Day 3 being the actual final perimeter of Timber Crater 6.
These hypothetical scenarios were not based on fire behavior or fire spread models. Two hypothetical
scenarios (2a and 2b) were developed to represent larger fires than the actual fire (Figure 5-2). Both larger
hypothetical scenarios (2a and 2b) cover a larger area and extend more days than the actual fire.
Scenario 2a - Daily Fire Perimeters
Day 1 perimeter
Day 2 perimeter
Day 3 perimeter (actual TC6 perimeter)
Day 4 perimeter
Day 5 perimeter
Day 6 perimeter
Suppression containment perimeter
Scenario 2b - Daily Fire Perimeters
- Day 1 perimeter
- Day 2 perimeter
Day 3 perimeter (actual TC6 perimeter)
Day 4 perimeter
- Day 5 perimeter
Day 6 perimeter
- Day 7 perimeter
- Day 8 perimeter
Suppression containment perimeter
TC6 = Timber Crater 6.
Note: The solid gray outline shows the fire suppression contingency perimeter which is considered the maximum extent of wildfires
in this area.
Figure 5-2 Daily fire perimeters for the larger Timber Crater 6 (TC6)
hypothetical fires (Scenarios 2a and 2b).
5.1.4
Prescribed Fire near Crater Lake National Park
Land management practices in and near Crater Lake National Park include prescribed fire and
mechanical thinning. Some of the leftover fuel from mechanical thinning is sold as timber and some is
burned in slash piles during the winter. Multiple prescribed burns have been conducted in the area (Figure
5-3), some of which intersect the TC6 Fire perimeter: Cornerstone in 2007 (no specific dates known),
Timber Crater 1 and 2 in 2001 (no specific dates known), and Timber Crater 1978 in 1978 (no specific
dates known). More recent prescribed fires (not named) were conducted in this area in September 2019
(13-15 and 26-28). Because the days of the September 2019 prescribed fires were presumed to match
5-6
DRAFT: Do Not Cite or Quote
-------
1 criteria for prescribed fire in the region, this time period was used for modeling both actual prescribed
2 fires during that period and provided a basis for modeling other prescribed burn units from previous
3 years. Each prescribed fire (e.g., actual 2019 prescribed fires, Cornerstone, Timber Crater 1 and 2, and
4 Timber Crater 1978) were modeled for these 2019 dates but in separate model simulations so they would
5 not interact with each other.
Rx = prescribed burn.
Figure 5-3 Fire perimeter of the actual Timber Crater 6 (TC6) Fire, multiple
wildfires that yielded positive resource benefits, and multiple
prescribed fires.
5-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
5.1.5
Case Study: Rough Fire
The Rough Fire burned in parts of the Sierra National Forest, Sequoia National Forest, and Kings
Canyon National Park between July 31 and October 1, 2015
rhttps://www.nps.gov/seki/learn/nature/rough-fire-interactive-map.htm'); NPS (2016)1. This wildfire
covered approximately 150,000 acres of land managed by multiple Federal agencies. The Rough Fire was
chosen as a complement to the TC6 Fire due to its much larger size and duration.
Land managers were able to suppress the Rough Fire at several points where land had been
previously managed. One such area was the Sheep Complex Fire in 2010 (-9,000 acres), which resulted
in less available fuel and provided a break to stop fire progression. The Sheep Complex Fire in 2010 was
a multimonth wildfire that burned at lower intensity, had slow progression related to moist fuels from
heavy rains in the area earlier that year, and yielded positive resource benefits.
One alternative hypothetical scenario for the Rough Fire (Scenario 1) consists of a smaller Rough
Fire under the assumption that a planned prescribed fire (Boulder Creek Unit 1 prescribed fire unit),
which did not occur, had occurred prior to the Rough Fire. This smaller fire hypothetical scenario
assumes that when the Rough Fire got to the area of the Boulder Creek Unit 1 prescribed fire unit,
progression downslope toward the Central Valley of California would have stopped. Fire perimeters are
shown for the Rough Fire, Sheep Complex Fire, and Boulder Creek Unit 1 Prescribed Fire area in Figure
5-4.
Another hypothetical scenario for the Rough Fire (Scenario 2) was a larger fire that progressed
through the area of the Sheep Complex Fire with an assumption that fuels were dry and fuel loading
would be similar to the surrounding area as if the Sheep Complex Fire had not happened. The
hypothetical larger Rough Fire includes the actual Rough Fire in addition to the area of the Sheep
Complex Fire. The hypothetical wildfire version of the Sheep Complex Fire was based on the original
spatial extent of the Sheep Complex Fire. The Sheep Complex Fire activity data was aggregated to the
total event/fuelbed/location, then a daily fraction of total acres from the Rough Fire (from September 1 to
10, 2015) to the Sheep Complex Fire aggregated activity data was applied to each of these combined
factors. This means that the Sheep Complex Fire kept the same total area and fuel beds but was
temporalized like the Rough Fire activity between September 1 and 10, 2015. This allowed the Sheep
Complex area to be burned as part of the Rough Fire at the point the actual Rough Fire progressed to this
area and beyond.
5-8
DRAFT: Do Not Cite or Quote
-------
Rough Fire 2015
Sheep Complex Fire 2010
Boulder Creek Unit 1
Merced
Fresno
Visalia
Bakersfield
Figure 5-4 Schematic showing the 2015 Rough Fire, 2010 Sheep Complex
Fire, and Boulder Creek Unit 1 prescribed burn unit in relation to
large urban areas in central California.
A prescribed fire (Boulder Creek Unit 1) was originally planned for an area adjacent to the
footprint of the Sheep Complex 2010 fire in 2013. This planned prescribed fire represents the minimum
amount of prescribed fire activity that was needed to create the suppression anchor that underpins the
smaller hypothetical scenario (Scenario 1) as the initial prescription plan for the area called for
5-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
approximately 5 more years of prescribed fire activity in the area (USFS. 2014). Boulder Creek
Unit lincluded a 3,200-acre area that was intended to restore fire and reduce fuels and fire behavior for
this steep and inaccessible area of the Boulder Creek drainage, through which the Rough Fire
subsequently burned. Since this prescribed fire unit was never burned, a series of days in the fall of 2014
were selected that matched meteorological conditions for a prescribed burn in the area of the Boulder
Creek prescribed fire burn unit. September 30 to October 3, 2014 were selected as days matching
meteorology appropriate for this prescribed fire burn unit.
5.2 Methodology
The air quality surfaces for PM2 5 and ozone for the TC6 Fire and Rough Fire, each hypothetical
scenario, and the prescribed fires, were produced using the modeling framework detailed in Figure 5-5,
which shows the connectivity and relationships between various tools and models used to develop case
study fire emissions. Fire location and timing was based on incident information where available and
supplemented with data generated by the Satellite Mapping Automated Reanalysis Tool for Fire Incident
Reconciliation Version 2 (SmartFire2; [SF2]) tool (Raffusc et al.. 2009). SF2 reconciles data from
satellite sensors and ground-based reports to use the strengths of both types of data while avoiding
double-counting of fires (Lark in et al.. 2020; Larkin et al.. 2009).
5-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Major Steps of the Process Model(s) for each step of the process
Fire location & size
Incident information, SmartFire2
Fuel type
FCCSv3
Fuel loading
VELMA/FCCSv3
Fuel consumed
CONSUME
Fire emission factors
Other emissions
Chemistry & transport
SERA database
SMOKE (emissions from mobile, EGUs , etc.)
CMAQ photochemical grid model
Human health impacts
BenMAP
BenMAP = Environmental Benefits Mapping and Analysis Program; CMAQ = Community Multiscale Air Quality; FCCS = Fuel
Characteristic Classification System; SERA = Smoke Emissions Reference Application; SMOKE = Sparse Matrix Operator Kernel
Emissions; VELMA = Visualizing Ecosystem Land Management Assessments.
Figure 5-5 Modeling framework used to characterize wildland fire emissions
and air quality impacts for case study analyses.
The BlueSky Pipeline (https: //github. com/pnwairfire/blue sky) is a version of the BlueSky
Framework rearchitected as a pipeable collection of stand-alone modules. The original BlueSky
Framework was Java based whereas the pipeline is based on Python. The BlueSky Pipeline estimates fuel
type, fuel loading, fuel consumption, and emissions for each fire. Fuel type is based on the Fuel
Characteristic Classification System (FCCS). Fuel loading is based on a combination of FCCS and
Visualizing Ecosystem Land Management Assessments (VELMA) model output. Fuel consumption is
based on the CONSUME module in the BlueSky Pipeline. BlueSky Pipeline provides daily total
emissions of CO, NOx, SO2, NH3, VOC, and primary PM2.5 for each wildfire and prescribed fire. Case
study fire emission factors are based on the SERA database (Prichard et al.. 2020).
Daily emissions were processed for input to the CMAQ photochemical model using the Sparse
Matrix Operator Kernel Emissions [SMOKE, https://www.cmascenter.org/smoke/; CMAS (2020)1
emissions model, which also provided emissions of other wildland fires, biogenic, and anthropogenic
emissions. The CMAQ model uses emissions generated by the SMOKE model and meteorological data
generated by the Weather Research and Forecasting (WRF) model to transport and deposit emissions
injected to the model and estimate chemical transformation. The output from the photochemical model
was processed for input to U.S. EPA's Environmental Benefits Mapping and Analysis
5-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
Program—Community Edition [BenMAP-CE; U.S. EPA (2019a)l to estimate the human health impacts
related to specific fire scenarios for each case study fire (see CHAPTER 8). More details about fuels,
emissions, and photochemical modeling follow in subsequent sections of this chapter.
4 5.2.1 Fuels (Fuel Characteristic Classification System [FCCS])
5 The FCCS contains a reference library of wildland fuelbeds that can be used for wildland fire
6 planning and smoke management decisions (Ottmar et al.. 2007). The FCCS calculator within the Fuel
7 and Fire Tools |https://\\\\\\ .fs.usda.gov/pn\\/tools/fucl-and-firc-tools-fft: FERA (2020)1 is used to
8 produce a fuel loadings input file for CONSUME v5.0, a fuel consumption module within the BlueSky
9 Pipeline (Prichard et al.. 2021).
10 Although the LANDFIRE system (LF. 2008). contains a FCCS fuelbed layer, it does not include
11 recent small wildfires and prescribed fires. To support emissions trade-offs analyses, we created four
12 separate 30-m FCCS fuelbed raster layers to represent each of the scenarios evaluated in the Timber
13 Crater 6 case study.
14 To represent prewildfire fuelbed layers for each of the four scenarios, we assigned base FCCS
15 fuelbeds Table A.5 FUELS-1 based on the 2014 LANDFIRE Existing Vegetation Type (EVT) layer (LF.
16 2014). We then used an existing Python script developed to update the base fuelbeds to represent canopy
17 and surface fuel changes associated with recent wildfires and prescribed burns within the study area,
18 including the 2010 Phoenix and 2014 Founders Day fires Table A.5 FUELS-2. For the TC6 Fire smaller
19 fire scenario (Scenario 1), fuelbeds were assigned to represent a recent prescribed fire over the entire
20 scenario area so that fuel loading would be more like an area post-prescribed fire rather than multiyear
21 fuel buildup. Fuel loading was not similarly modified for the Rough Fire scenarios. A Python script was
22 used to update fuelbeds to recent low-severity prescribed burns immediately post-disturbance (111),
23 recent high-severity wildfires within 0-5 years (132) and older high-severity wildfires within 5-10 years
24 (133).
25 5.2.2 Characterizing Surface Fuel Loads for Use in the BlueSky
26 Pipeline
27 Surface fuel load characterization is an important component of modeling air quality impacts
28 associated with wildfires and prescribed fires. The most commonly used tool for estimating surface fuel
29 loads in the U.S. is the FCCS (Ottmar et al.. 2007). which characterizes available fuel loading for various
30 vegetation classification categories across a landscape and includes both vegetation type (e.g., Ponderosa
31 Pine, Red Alder) and fuel load category (e.g., canopy, shrubs, nonwoody).
5-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
While FCCS captures the general diversity of available fuels found throughout the U.S., the fuel
loadings are assumed to be homogenous within each vegetation type. Studies suggest that FCCS and
other vegetation classification-based approaches do not fully characterize the spatial and temporal
variability of fuels, site-specific conditions, and the presence of disturbances such as harvests and
prescribed fires (Lutes et al.. 2009; Brown and See. 1986. 1981). In light of these considerations, an
ecohydrological modeling approach was implemented for this assessment to supplement existing FCCS
data, specifically to characterize spatial and temporal variations more fully in forest fuel loads arising
from site-specific biophysical and disturbance conditions.
The VELMA model is a spatially distributed (grid-based) ecohydrological model that simulates
integrated responses of vegetation, soil, and hydrologic components to various inputs of land use, soil,
and climate (McKanc et al.. 2014). It has been widely applied to many terrestrial ecosystem types,
including forests, grasslands, agricultural floodplains, and alpine and urban landscapes. Particularly in
western U.S. forests and grasslands, VELMA has simulated effects of fire and harvest and subsequent
spatial and temporal dynamics of ecosystem recovery (McKanc et al.. 2020; Yee et al.. 2017; Barnhart et
al.. 2015; Abdelnour et al.. 2013; Abdelnour et al.. 2011).
VELMA was used here to simulate aboveground biomass for the two case study fires (i.e., the
TC6 Fire in Oregon and the Rough Fire in California). In addition, for the Rough Fire case study VELMA
modeling was conducted for additional fires within the actual Rough Fire vicinity to support the
development of hypothetical scenarios. This additional modeling included the areas of the Sheep
Complex Fire and the area within the proposed Boulder Creek prescribed fire. More detailed information
on the actual and hypothetical fuel treatments and boundaries are described in CHAPTER 3 and in the
present chapter. For each case study area, VELMA was spatially initialized using high-resolution (30-m),
aboveground total (live and dead) biomass developed for western forest ecosystems (California, Oregon,
Washington) by the Landscape Ecology, Modeling, Mapping, and Analysis (LEMMA) project at Oregon
State University (LEMMA. 2020; Kennedy et al.. 2018; Davis et al.. 2015). LEMMA forest biomass map
data are developed and updated annually using state-of-the-science, satellite-based change-detection
technology (Landsat), calibrated using the U.S. Forest Service (USFS) Forest Inventory and Analysis
(FIA) regional network of forest biomass plot measurements (Bell et al.. 2018).
Extensive validation of LEMMA-mapped biomass predictions has previously been performed for
the western U.S., including the Deschutes National Forest near the TC6 case study area (Bell et al.. 2018).
In validation tests, LEMMA-initialized VELMA TC6 application closely simulated aboveground biomass
pools and rates of accumulation published for this dry coniferous forest ecoregion (Smithwick et al..
2002). This is important because VELMA was initialized for the 2018 TC6 Fire based on 2010 LEMMA
biomass data, primarily to allow for potential future prefire fuel reduction simulation treatments using
VELMA. The LEMMA aboveground live and dead forest biomass data for the Sheep Complex and
Rough fires corresponded to the actual 5 years, 2010 and 2015, respectively. See APPENDIX for details.
5-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Following initialization for each case study site, VELMA's LEMMA-based overstory fuel-load
estimates were merged with FCCS surface fuel load estimates, specifically to replace FCCS forest
overstory fuel load estimates assumed to be homogenous within each vegetation type, rather than on
location-specific data (Lutes et al.. 2009; Brown and Sec. 1986. 1981).
Figure 5-6 generally illustrates how VELMA and FCCS data products were merged and fed into
the BlueSky Pipeline. The combined VELMA-FCCS fuelbed database for each site was used as an input
to BlueSky Pipeline, specifically to the CONSUME model, to simulate air quality impacts associated with
wildfire and prescribed fire simulations. The resulting BlueSky Pipeline input data comprised a raster
map of fuelbed classifications and a comma-separated-value (CSV) look-up file of fuel loadings for
various fuelbed categories (e.g., canopy, shrubs, nonwoody vegetation, woody fuels, litter/lichen/moss,
ground fuels) that include merged FCCS and VELMA fuel type and fuel load data, respectively. The
combined use of FCCS and VELMA for this purpose plays to the strengths of both models, together
representing the best available science for estimating fine-scale horizontal and vertical distributions of
fuelbed types and loadings (Bell et al.. 2018; Ottmar et al.. 2007).
Fuel Type, Management Code
| mo. 133 Temperate Pacific subalpine-montane wet meadow
| boo u.i Idaho fescue-California oatgrass grassland
1319133 Pacific silver fir-Sitka alder forest
| jis.133 Showy sedge-black alpine sedge grassland
12A3133 Engelmann spruce-Douglas-fir-white fir-ponderosa pine forest
123*113 Pacific silver fir-mountain hemlock forest ..
1239.132 Pacific silver fir-mountain hemlock forest i3!UGSKy r ip6llfl6
1237.133 Huckleberry heather shrubland
> 18 Other Fuelbeds
g Carbon/m2
¦ 10,000
Management Code
111: Rx burn
132: 0-5 years since WF burn
133: 5-10 years since WF burn
Synthesis of fuel type details and loadings will better
characterize smoke emissions and dispersion
FCCS Simulated
Fuel Types
VELMA Simulated
Aboveground
FCCS = Fuel Characteristic Classification System; g Carbon/m2 = grams carbon per square meter; Rx = prescribed burn;
VELMA = Visualizing Ecosystem Land Management Assessments; WF = wildland fire.
Note: Example shown is for the Timber Crater 6 (TC6) Fire case study in Oregon.
Figure 5-6 Fuel Characteristic Classification System (FCCS) fuel type data
and Visualizing Ecosystem Land Management Assessments
(VELMA) fuel load data were merged to produce fuelbed inputs
for the BlueSky Pipeline.
5-14
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
In summary, whereas FCCS performs well at providing estimates of management-sensitive
surface and understory fuel types and loads, VELMA performs well at estimating overstory/canopy fuel
loads by virtue of its use of LEMMA initialization and mechanistically modeled live and dead biomass
dynamics. Additional details on the methods used to develop LEMMA-initialized VELMA for both case
studies and associated VELMA-FCCS fuelbed databases are located in the Appendix (see Section A.5).
5.2.3 Fuel Consumption and Fire Emissions (BlueSky Pipeline)
The BlueSky Pipeline Version 4.2.14 was used support this project. The BlueSky Pipeline is a set
of python-based stand-alone modules that can be linked or piped together in a series so that the output of
one module becomes the input of the next. For all the fire emission scenarios, the BlueSky Pipeline was
used to calculate consumption, to calculate emission factors and to calculate emissions using
geo-references area burned input. Generally, as fire data flow through the modules within the BlueSky
Pipeline, the modules add to the data without modifying what was already defined. The consumption
module used was CONSUME Version 5.0.2. Fuel loadings were based on either FCCS v3 with LandFire
vl.4 fuel beds or FCCS v4 with USFS fuel beds. Emissions were based on University of Washington
SERA emission factors for the case study fires and Fire Emission Production Simulator (FEPS) v2 for all
other fires.
5.2.3.1 Temporal Profile for Timber Crater 6 (TC6) Fire
Fire hotspot characterization data from the Geostationary Operational Environmental Satellite
(GOES)-16 Advanced Baseline Imager (ABI) were obtained in Network Common Data Form (NetCDF)
format from Amazon's AWS S3 file system at s3://noaa-goesl6/ABI-L2-FDCC/2018/ (GOES-R
Algorithm Working Group. 2018). The data set comprises latitude, longitude, fire radiative power (FRP),
estimated fire area, fire temperature, and a data quality factor (DQF) for each pixel. The fire data are
derived (i.e., not directly measured) products of the GOES-ABI. The algorithms for deriving fire data and
data quality are described elsewhere (Schmidt et al.. 2013). Data from 15-29 July 2018 were extracted
from within a bounding box defined by the points (43.03°N, 122.1°W) and (43.1°N, 121.9°W), centered
roughly on the centroid of the final Timber Crater 6 Fire perimeter. Although data are typically available
at 5-minute intervals, there are often large temporal discontinuities due to absence of detection because of
issues such as low fire power, glare or obscuration by smoke or clouds. After filtering for validity
(DQF = 0), 166 data points were available for the analysis. Analysis was performed using Python 3 code
and libraries.
Fire radiative power is proportional to the rate of fuel consumption in wildland fires (Kremens et
al.. 2012). To derive a characteristic fuel consumption curve, valid FRP values from all days in the data
set were binned by hour. A mean value and standard deviation of FRP was calculated for each hour. Valid
5-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
detections were available for only the hours 11:00 a.m. to 9:00 p.m. PDT each day over the time span of
the fire. It is possible that fire radiative powers were too low and/or weather conditions were not favorable
for detections outside of that range of times. A Weibull-like curve function (Barnctt. 2002) was fitted to
the hourly mean FRP values using the "curve fit" method from the SciPy (Version 1.4.1) Optimize
library. To facilitate curve fitting, mean FRP values outside of the available time range were extrapolated
using a linear ramp between the end values (Numpy v. 1.18.5 "pad" function).
The resulting fitted curve gives a realistic profile of diurnal fuel consumption and indicates that,
on average, peak FRP, and therefore fuel consumption, occurred around 3:00 p.m. PDT. However, the
FRP curves for any given day of the fire varied considerably from the fitted curve, as indicated by the
large variations in FRP during the afternoon hours. As important, with only 166 data points there is a
relative paucity of GOES satellite data for this fire, suggesting that the resulting consumption curve
should be used with caution.
5.2.4 Pile/Slash Burn Emissions
Typical practices for collecting fuel leftover from mechanical thinning operations include
collecting the debris into three types of piles: machine landing piles (largest), machine grappling piles,
and hand piles (smallest). Each of these practices are common at Crater Lake National Park and
surrounding areas. Typical geometry for each was provided by land managers in the region (shown in
Table 5-1).
Table 5-1
Emissions and fuel consumption for three different types of
slash/pile burn fuel geometry assumptions for the Timber Crater 6
(TC6) case study area.
File Name
Geometry
NOx
(tons)
PM2.5
(tons)
VOC
(tons)
CO
(tons)
NHs
(tons)
SO2
(tons)
Fuel
Consumption
(tons)
Machine
landing pile
50' x 100' x 25'
0.6256
2.1475
0.7200
12.0854
0.5942
0.3118
270.4
Machine
grappled
pile
15' x 15' x 10'
0.0113
0.0387
0.0130
0.2175
0.0107
0.0056
4.9
Hand_pile
5' x 5' x 5'
0.0004
0.0014
0.0005
0.0082
0.0004
0.0002
0.2
CO = carbon monoxide; NH3 = ammonia; NOx = nitrogen oxides; PM2 5 = particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |jm; S02 = sulfur dioxide; VOC = volatile organic compound.
5-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Pile burn emissions were based on the University of Washington (UW) tool
rhttps: //depts .Washington. edu/nwfire/pile s/ University of Washington (2014)1. This tool provides daily
emissions of PM2 5, CO, and VOC (also provides CO2 and CH4 which were not used in this analysis). The
CMAQ model needs emissions for NOx, NH3, and SO2, and these were estimated using BlueSky Pipeline
FEPS assuming a 70/15/15 flaming/smoldering/residual smoldering split for pile burns. A cubic shape
was assumed for each pile. Pile burns used the same diurnal profile as prescribed fire (late morning start
and ending in the early evening).
5.2.5 Air Quality Modeling System
The CMAQ Version 5.3.2 model was applied with aqueous phase chemistry (Fahev et al.. 2017).
inorganic thermodynamics (Fountoukis and Nenes. 2007). and gas phase chemistry based on the Carbon
Bond 6 Revision 3 mechanism (Emery et al.. 2015). The default option was used where photolysis rates
were attenuated in the presence of model predicted particulate matter (Baker etal.. 2016). Secondary
organic aerosol (SOA) treatment is a yield-based approach based on precursors including isoprene,
monoterpenes, sesquiterpenes, benzene, toluene, and xylenes. Some of the SOA becomes nonvolatile
through oligomerization processes (Carlton et al.. 2010). Primarily emitted organic aerosol is treated as
nonvolatile. The ratio of organic mass to organic carbon is assumed to be 1.7 for primary PM2 5 wildland
fire emissions (Simon and Bhave. 2012).
The WRF model was used to provide the modeling system meteorological inputs (Skamarock et
al.. 2008). Both CMAQ and WRF were applied with 35 layers to represent the vertical atmosphere from
the surface up to 50 mb. The WRF configuration used here has been evaluated and shown reasonable
performance for winds, temperature, and surface mixing layer height for the Pacific Northwest (Zhou et
al.. 2018) and California (Baker et al.. 2013). WRF was initialized with the 12-km North American
mesoscale (NAM) analysis product rhttps://www.ncdc.noaa.gov/data-access/model-data/model-
datasets/north-american-mesoscale-forecast-svstem-nam; NCEP (2021)1. CMAQ initialization and
boundary inflow conditions were extracted from coarser hemispheric CMAQ simulations.
Anthropogenic emissions in the model domain were based on the 2016 National Emission
Inventory (U.S. EPA. 2019b) with year-specific data used for electrical generating units based on
continuous emissions monitor data. Biogenic emissions were estimated with the Biogenic Emission
Inventory System Version 3.6.1, which has been shown to perform well for biogenic VOC in California
(Bash et al.. 2016). Emissions of wildland fires other than the case studies were based on daily fire
location and burn area information using the SmartFire2 system, which is largely based on satellite
products and incident information. Location, burn area, and date information is provided to the BlueSky
Pipeline to estimate fuel type, fuel moisture, and fuel consumption that is used to estimate daily emissions
(Urbanski. 2014) of CO, NOx, VOC, SO2, NH3, and PM2 5 based on FEPSv2 emission factors for each
noncase study wildfire and prescribed fire in the model domain (Larkin et al.. 2020).
5-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
SMOKE is used to apply a fire type-specific diurnal profile and allocates total emissions of NOx,
VOC, and PM2 5 to specific model species needed for chemical mechanisms. Speciation profiles are based
on those available in the SPECIATE database I https://\\\v\\ .cpa.gov/air-cinissions-inodcling/spcciatc;
U.S. EPA (2020b)l. NOx emissions were allocated 10% to NO and 90% to NO2. Speciation profiles for
VOC and primarily emitted PM2 5 are provided in Table A.5 SPECIATION-1. Daily total emissions were
allocated to specific hours of the day based on default profiles for wildfire and prescribed fire (Baker et
al.. 2020; Baker et al.. 2016). Fuel moisture is a global parameter that only varies by fire type (wildfire or
prescribed).
5.3 Results—Case Studies
For both the Timber Crater 6 and Rough Fire case studies total acres burned, PM2 5 emissions,
fuel, and fuel consumption are shown for the wildfire, alternative hypothetical scenarios, and areas that
had been managed in the past in Table 5-2.
Photochemical model predictions of baseline maximum daily 8-hour average (MDA8) O3 and
major components of speciated PM2 5 (total carbon, sulfate ion, and nitrate ion) were paired in time and
space with measurements from routine surface network monitors. This type of comparison provides
information about how well the modeling system is predicting air quality from wildland fire and other
sources. A reasonable representation of the chemical environment surrounding fire plumes is important to
best capture secondarily formed pollutants like O3 since wildland fire emit precursors of O3 (NOx and
VOC) that can react with other sources of pollution to form O3.
The photochemical modeling system generally compares well with ambient data for the various
episodes included in this assessment. Model performance metrics for daily model-observation pairs at
routine surface network monitors aggregated over each episode are shown in Table A.5-1. Each
prediction-observation pair is also shown with scatterplots for each species (Figure A.5 MPE-1 to Figure
A.5 MPE-6). Additional model performance information is provided as part of subsequent figures in this
section that show episode average surface level modeled PM2 5 and MDA8 O3 compared with
measurements made at routine monitors. The modeling system does well at replicating spatial gradients in
PM2 5 and O3. It also generally captures synoptic and day-to-day variability in measurements near each of
the case study fires. The performance metrics for these episodes is consistent with the performance shown
for this type of modeling system for monitors impacted by large wildfires in the western U.S. (Baker et
al.. 2018; Koplitz et al.. 2018; Baker et al.. 2016). Very little data exist on episodic model performance
for these areas during large wildfire events for performance comparison. However, performance metrics
of other studies completed over longer time frames and larger model domains are generally consistent
with those estimated for the modeling periods included in this assessment (Kelly et al.. 2019; Simon et al..
2012).
5-18
DRAFT: Do Not Cite or Quote
-------
Table 5-2 Wildfire and prescribed fires modeled as part of the Timber Crater 6 (TC6) and Rough Fire case studies.
Fire/Burn Unit
Name
Type
Modeled Time Period
Acres Burned
(acres)
Total Fuel
Consumption
(tons)
Total Fuel
(tons)
PM2.5 Emissions
(tons)
Timber Crater 6
Actual wildfire
July 15 to 31, 2018
3,123
213,454
145,985
1,869
TC6 hypothetical
smaller fire (1)
Hypothetical wildfire
July 15 to 31, 2018
1,237
37,954
91,419
1,041
TC6 hypothetical
larger fire (2a)
Hypothetical wildfire
July 15 to 31, 2018
20,878
468,843
1,249,089
12,794
TC6 hypothetical
larger fire (2b)
Hypothetical wildfire
July 15 to 31, 2018
27,373
727,180
1,825,606
20,015
Timber Crater 1978
Hypothetical prescribed fire
September 1 to 30, 2019
2,049
26,992
112,362
565
Cornerstone
Hypothetical prescribed fire
September 1 to 30, 2019
772
10,671
69,787
232
Timber Crater 1/2
Hypothetical prescribed fire
September 1 to 30, 2019
633
7,751
37,649
157
2019 actual
prescribed fires
Actual prescribed fire
September 1 to 30, 2019
886
6,206
20,955
117
Rough Fire
Actual fire
August 1 to September 30,
2015
145,438
3,284,638
7,128,199
85,638
Rough hypothetical
smaller fire (1)
Hypothetical wildfire
August 1 to September 30,
2015
113,349
2,631,258
6,450,696
68,949
Rough hypothetical
larger fire (2)
Hypothetical wildfire
August 1 to September 30,
2015
154,354
3,448,094
7,562,392
89,349
Boulder Creek Unit 1
Hypothetical prescribed fire
September 26 to October 7,
2014
3,289
30,163
90,452
499
Sheep Complex Fire
Actual fire
July 30 to September 30,
2010
8,916
103,037
434,193
2,344
PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; TC6 = Timber Crater 6.
5-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
5.3.1
Timber Crater 6 (TC6) Air Quality Impacts
A domain with 4-km-sized grid cells covering Oregon and northern California were applied for
the time period coinciding with the case study fire (July 2018). Initial conditions and boundary inflow
were extracted from a CMAQ simulation for a 12-km domain covering the continental U.S. for the entire
year of 2018.
Model predicted episode average PM2.5 and MDA8 O3 for the 2018 episode compared well with
routine surface monitor data (Figure 5-7). Large wildfires in southwest Oregon and northern California
resulted in a strong gradient in PMj 5 concentrations across the domain. Enhancements of O3 from
wildfire were less evident because meteorologic conditions during this period was favorable to regional
formation. Agreement between model predictions and measurements provides confidence that the actual
and hypothetical case study fires are being modeled in a realistic chemical and physical environment.
Episode average PM2.5 all sources Episode average MDA8 03 all sources
Max = maximum: MDA8 = maximum daily 8-hour average; pg/m3 = micrograms per cubic meter; 03 = ozone; PM25 = particulate
matter with a nominal mean aerodynamic diameter less than or equal to 2.5 pm; ppb = parts per billion.
Figure 5-7 Episode average PM2.5 and maximum daily 8-hour average (MDA8)
ozone (O3) predicted by the modeling system and measured by
routine surface monitors for the 2018 modeling period used for
the Timber Crater 6 (TC6) scenarios.
Episode average model predicted PM2j from the actual TC6 Fire and hypothetical scenarios are
shown in Figure 5-8 (top row). To assess population exposure to PM§,s produced by the TC6 Fire, model
predictions were also multiplied by gridded population to provide an estimate of aggregate population
5-20
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
exposure (Figure 5-8, bottom row). Figure 5-8 also shows the difference in episode average PM2.5.
between the largest and smallest hypothetical scenarios and the actual fire scenario. The spatial pattern of
differences between the largest hypothetical scenario (2b) and the actual TC6 Fire is strongly influenced
by days toward the end of the largest hypothetical fire scenario where nighttime winds blew smoke
southward toward the Oregon-California border. The spatial extent of impacts from the hypothetical
scenario 2a fire (not shown) are similar to hypothetical scenario 2b, but with a smaller magnitude of
change.
Ambient PM2.5 Ambient PM2.5 Ambient PM2.5
PM2.5 = particulate matter w/ith a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; TC6 = Timber Crater 6;
pg/m3 = micrograms per cubic meter.
Note: Ambient PM2.5 impacts are shown in the top row and aggregate population exposure in the bottom row where PM2.5 is
multiplied by gridded population.
Figure 5-8 Episode average PM2.5 impacts and aggregate population
exposure from the actual Timber Crater 6 (TC6) Fire and the
difference between the actual fire and largest (2b) and smaller (1)
hypothetical scenarios.
The Episode average model predicted MDA8 O3 from the TC6 Fire and hypothetical Scenarios 1
and 2b are shown in Figure 5-9 (top row). Model predictions are also multiplied by gridded population to
provide an estimate of aggregated population impacts. The spatial pattern of differences between the
5-21
DRAFT: Do Not Cite or Quote
-------
1 largest (2b) and actual scenario is strongly influenced by daytime winds blowing smoke eastward toward
2 the Oregon-Idaho border. This differs from the spatial extent ofPMai impacts because the largest PM2 5
3 concentrations are overnight when winds moved air toward the south. Impacts of the daytime wind
4 patterns dominate the spatial extent of O3 formation because these daytime winds coincide with solar
5 radiation, which is needed for photochemical O3 production.
Ambient MDA8 03
Ambient MDA8 03
Ambient MDA8 03
Actual fire (TC6)
Largest hypothetical (2b) - Actual fire (TC6)
Smaller hypothetical (1) - Actual fire (TC6)
Population Exposure MDA8 03
Population Exposure MDA8 03
Population Exposure MDA8 03
Actual fire (TC6)
Largest hypothetical (2b) - Actual fire (TC6)
Smaller hypothetical (1) - Actual fire (TC6)
MDA8 = maximum daily 8-hour average; 03 = ozone; ppb = parts per billion; TC6 = Timber Crater 6.
Note: Ambient MDA8 03 impacts are shown in the top row and aggregate population exposure in the bottom row where MDA8 03 is
multiplied by gridded population.
Figure 5-9 Episode average maximum daily 8-hour average (MDA8) ozone
(63) impacts and aggregate population exposure from the actual
Timber Crater 6 (TC6) Fire and the difference between the actual
fire and largest (2b) and smaller (1) hypothetical scenarios.
6 Without considering air quality impacts, based on this case study and other similar studies, results
7 indicate that land management, such as prescribed fire and mechanical thinning, reduce fuel, which means
8 less fuel is consumed when wildfires happen later. Less fuel available for wildfire consumption in turn
9 means less emissions and lower levels of downwind pollutants. Reduced fuel loading also can lead to
10 smaller fire perimeters, which is represented in the smaller fire hypothetical (Scenario 1) presented here.
5-22
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
This smaller perimeter is based on expert judgement for this hypothetical scenario and is not based on fire
behavior or fire spread models. Illustrating the change in air quality related to past land management
activity is challenging because spatial and temporal scales of both are quite different. For instance, many
prescribed fires may need to be conducted over many years to effectively minimize the rate of spread of
wildfire or reduce fuels enough to impact air quality. Further, only a single period of conducive
meteorology (September 2019) was used for the prescribed fire impacts, which does not capture the
variability possible if other years or time of year were chosen.
Figure 5-10 shows daily domain average PM2 5 ambient and aggregate population exposure from
the actual TC6 Fire and hypothetical fire scenarios compared with multiple prescribed fires. All the
prescribed fires were modeled in separate simulations with the same days in September 2019 when
prescribed fires were happening near Crater Lake National Park. Similar information is shown for MDA8
O3 in Figure 5-11. The daily average impacts only include grid cell-days where modeled fire impacts
exceed a threshold (0.01 (.ig/nr1 for PM2 5 and 0.01 ppb for MDA8 O3) so that the average does not include
large areas of the model domain with no fire impacts due to wind transport patterns.
5-23
DRAFT: Do Not Cite or Quote
-------
7/16/2018 7/19/2018 7/22/2018 7/25/2018 7/28/2018 8/1/2018
Episode Day
7/16/2018 7/19/2018 7/22/2018 7/25/2018 7/28/2018 8/1/2018
Episode Day
Ambient PM2.5
2.049 acre prescribed fire
772 acres prescribed fire
633 acres prescribed fire
Actual Sept. 2019 prescribed fires (8
B6 acres)
H 1 H
9/24/2019 9/26/2019 9/28'2019
Episode Day
Episode Day
|jg/rn3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to
2.5 |jm.
Figure 5-10 Daily average PM2.5 ambient (top row) impacts and estimates of
aggregate population exposure (bottom row) from the Timber
Crater 6 (TC6) case study scenarios (left) and prescribed fire
scenarios (right).
5-24
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
Ambient MDA8 03
Hypothetical Scenario 2b
Hypothetical Scenario 2a
Actual lire
Hypothetical scenario 1
Ambient MDA8 03
Episode Day
Episode Day
Population Exposure MDA8 03
Hypothetical Scenario 2b
Hypothetical Scenario 2a
Actual fire
Hypothetical Scenario 1
7/22/2018 7/25/2018
Episode Day
Population Exposure MDA8 03
— 2.049 acre prescribed fire
772 acres prescribed fire
— 633 acres prescribed fire
— Actual Sept. 2019 prescribed fires (886 acres)
IV1DA8 = maximum daily 8-hour average; 03 = ozone; ppb = parts per billion; Rx = prescribed fire..
Figure 5-11 Maximum daily 8-hour average (MDA8) ozone (Os) ambient (top
row) impacts and estimates of aggregate population exposure
(bottom row) from the Timber Crater 6 (TC6) case study scenarios
(left) and prescribed fire scenarios (right).
Daily aggregate population exposures are notably different than ambient impacts for July 20
when ambient concentrations were high, but winds did not transport smoke to populated areas. The
prescribed fires had high ambient impacts but did not impact highly populated areas in this case study.
The large estimated population exposures of the biggest hypothetical fires toward the middle and end of
the episode are related to larger fire size (e.g., more fuel consumption and emissions) and winds blowing
smoke towards populated areas on the additional simulation days.
The daily impacts of prescribed fire on PM2 5, particularly the estimated population exposures,
were typically lower than wildfire. However, the daily impacts of MDA8 O3 from prescribed fire were
sometimes comparable or even larger than the wildfire scenarios. This is due to the large amount of fuel
burned as part of the hypothetical prescribed fires on a single day compared to the daily amount of fuel
consumed by these small (small compared to the Rough Fire for instance) hypothetical wildfire scenarios.
5-25
DRAFT; Do Not Cite or Quote
-------
1 Further, the prescribed fire emissions are temporally allocated to daytime hours which means more of the
2 mass is available for photochemical reactions leading to O3 production compared to wildfire emissions
3 which are spread out over the entire day and night.
4 Figure 5-12 shows daily average PM2 5 and MDA8 O3 ambient impacts and estimates of
5 aggregate population exposure results from hypothetical slash burn piles in the area of the TC6 Fire.
6 These slash burns were based on common slash burning activity (pile type, size, geometry, and fuel; see
7 Table 5-1). A total of seven hypothetical pile burns were included in each model simulation. These piles
8 were not intended to relate to the amount of fuel from mechanical thinning activity in the area but rather
9 illustrate the potential impacts of slash burning on PM2 5 and MDA8 O3 on winter days when meteorology
10 conditions would be conducive for slash burns (e.g., snow cover, no rain, cold temperatures).
5-26
DRAFT: Do Not Cite or Quote
-------
Ambient PM2.5
7 Machine Landing Piles
7 Machine Grappling Piles
^ A
"I 1 1 1 r
2/17/2019 2/21/2019 2/25/2019 3/1/2019 3/5/2019
Episode Day
Ambient MDA8 03
7 Machine Landing Piles
7 Machine Grappling Piles
— A
Episode Day
t 1 1 1 r
2117/2019 2/21/2019 2/25/2019 3/1/2019 3/5/2019
s -
Population Exposure PM2.5
7 Machine Landing Piles
7 Machine Grappling Piles
2117/2019 2/21/2019 2/25/2019
Episode Day
1 r
3/1/2019 3/5/2019
8.
Q.
CL
Population Exposure MDAS 03
7 Machine Landing Piles
7 Machine Grappling Piles
A.
2/17/2019 2/21/2019 2/25/2019
Episode Day
1 r
3/1/2019 3/5/2019
MD08 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; 03 = ozone; PM2.5 = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 pm; ppb = parts per billion.
Figure 5-12 Daily average PM2.5 (left) and maximum daily 8-hour average
(MDA8) ozone (O3) (right) ambient (top row) impacts and
estimates of aggregate population exposure (bottom row) from
hypothetical pile burns from the Timber Crater 6 (TC6) case study
area.
5-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
5.3.2 Rough Fire Air Quality Impacts
The modeling system was applied for the 2015 Rough Fire, a hypothetical smaller Rough Fire
(Scenario 1), a hypothetical larger Rough Fire (Scenario 2), the 2010 Sheep Complex Fire, and a
hypothetical prescribed fire (Boulder Creek Unit 1) for a period matching ideal meteorological conditions
for prescribed fire in the fall of 2014. The larger Rough Fire hypothetical (Scenario 2) includes the actual
Rough Fire in its entirety and also includes the area of the Sheep Complex Fire, which did not burn as
part of the actual Rough Fire. The smaller Rough Fire hypothetical scenario (Scenario 1) eliminates
sections of the actual Rough Fire that were downslope of an area planned for prescribed fire (Boulder
Creek Unit 1) but never happened. This smaller fire hypothetical is based on the idea that if that
prescribed fire had happened before the Rough Fire, it would have provided a boundary for fire
suppression and stopped progression after that point downslope toward the Central Valley of California.
CMAQ was applied for a 12-km domain covering the continental U.S. Initial conditions and
boundary inflow were extracted from a coarser hemispheric scale photochemical model simulation. This
coarser grid spacing scale was selected for the larger Rough Fire case study because a larger domain was
used in anticipation of impacts much further downwind than the TC6 Fire case study. Model simulations
were done for periods coincident with case study fires in 2010 (Sheep Complex), 2014 (hypothetical
Boulder Creek Unit 1), and 2015 (actual Rough Fire). Model predicted episode average PM2 5 (Figure
5-13) and MDA8 O3 (Figure 5-14) for each episode compared well with routine surface monitor data.
This agreement between model predictions and measurements provides confidence that the actual and
hypothetical case study fires are being modeled in a realistic chemical and physical environment.
5-28
DRAFT: Do Not Cite or Quote
-------
Episode average PM2.5 all sources
h 50
30
- 10
pg/mr3
I- 10
pg/m3
- 25
20
- 15
- 10
jjg/m3
- 4
- 3
- 1
lig.'m3
Max = maximum; pg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 [jm.
Figure 5-13 Episode average PM2.5 predicted by the modeling system (from all
actual sources) and measured by routine surface monitors (left)
and fire specific modeled impacts (right) for the actual Rough Fire
(top), actual Sheep Complex Fire (middle), and hypothetical
Boulder Creek Unit 1 prescribed fire (bottom),
PM2.5 from Boulder Creek Unit 1 fire
Max = 30 ug'm3
Episode average PM2.5 all sources
Max = 136 ug/m3
5-29
DRAFT; Do Not Cite or Quote
-------
Episode average MDA8 03 all sources
Episode average MDA8 03 from Rough fire
Episode average MDA8 03 all sources
Episode average MDA8 03 from Sheep Complex fire
Episode average MDA8 03 all sources
Episode average MDA8 03 from Boulder Crek Unit 1 fire
Max = maximum; IV1DA8 = maximum daily 8-hour average; 03 = ozone; ppb = parts per billion.
Figure 5-14 Episode average maximum daily 8-hour average (MDAB) ozone
(63) predicted by the modeling system (from all sources) and
measured by routine surface monitors (left) and fire-specific
modeled impacts (right) for the 2015 modeling period used for the
Rough Fire scenarios (top), 2010 modeling period for the actual
Sheep Complex Fire (middle), and 2014 modeling period for the
hypothetical Boulder Creek Unit 1 prescribed fire (bottom).
5-30
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
The actual Rough and Sheep Complex fires spanned multiple months. The Rough Fire had much
larger downwind impacts which is related to the larger size of that fire in terms of acres burned and fuel
consumed. The largest impacts from each of the fires is at the fire location itself with concentrations
decreasing as distance from the fire increases. The episode average impacts for the hypothetical Boulder
Creek Unit 1 prescribed fire are averaged over a much shorter time period (10 days) compared to the
Rough and Sheep Complex fires which should be kept in consideration when comparing these spatial
plots.
Each of the fires modeled as part of this case study have some impacts on populated areas in the
Central Valley of California and further downwind toward the east. Some of the near-fire impacts on
population areas may be overstated due to the 12-km-sized grid cells used for this case study, which may
not capture complex terrain influenced meteorology and transport. This is particularly important to
consider for the hypothetical Boulder Creek Unit 1 fire since the days for this fire were selected based on
meteorology that was considered conducive to keeping air in the mountains and minimizing downslope
flow to the Central Valley.
Each of the fires modeled in this case study produce fairly small levels of MDA8 O3 compared
with regional levels measured at surface monitor sites during the same time periods (Figure 5-14). The
spatial nature of elevated MDA8 O3 in California suggest sources other than wildland fire
(e.g., anthropogenic, biogenic, lateral boundary inflow) contributed the most to ambient surface level O3.
Episode average model predicted PM2 5 from the actual Rough Fire is shown in Figure 5-15.
Model predictions are also multiplied by gridded population to provide an estimate of aggregated
population exposure. Figure 5-15 also shows the difference in episode average PM2 5 between the
hypothetical scenarios and actual fire scenario. Similar information is presented for MDA8 O3 in Figure
5-16.
5-31
DRAFT: Do Not Cite or Quote
-------
|jg/rn3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to
2.5 |jm.
Note: Ambient PM2.5 impacts are shown in the top row and aggregate population exposure in the bottom row where estimated PM2.5
concentrations are multiplied by gridded population.
Figure 5-15 Episode average PM2.5 impacts from the actual Rough Fire and
the difference between the actual scenario and smaller
(Scenario 1) and larger (Scenario 2) hypothetical scenarios.
5-32
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
avg = average; MDA8 = maximum daily 8-hour average; O3 = ozone; ppb = parts per billion.
Note: MDA8 O3 impacts are shown in the top row and aggregate population exposure in the bottom row where estimated MDA8 O3
concentrations are multiplied by gridded population.
Figure 5-16 Episode average maximum daily 8-hour average (MDAB) ozone
(Oa) impacts from the actual Rough Fire and the difference
between the actual scenario and smaller (Scenario 1) and larger
(Scenario 2) hypothetical scenarios.
The ambient impacts of the actual fires and hypothetical wildfire scenarios are highest in
California and decrease downwind as air moves smoke into the intermountain west and central plains.
When the impacts are multiplied by population most urban areas in the model domain have nonzero
impacts. This shows that very small concentrations of smoke in large population areas can result in
similar aggregated exposure to sparsely populated areas near the fire.
Rough Fire impacts on regional MDA8 O3 are highest near the fire with smaller impacts in the
Central Valley of California and central Nevada. Population impacts are also notable in large downwind
urban areas like Salt Lake City. A very small opposite response in MDAB O3 is seen in the northern
Central Valley of California for both alternative scenarios. This feature is magnified when applying
population due to the very large number of people living in that part of the state. In this situation, changes
in available oxidants and precursors has a small impact on model predicted O3.
5-33
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
Figure 5-17 shows daily domain average PM2.5 ambient impacts and aggregate population
exposure from the actual and hypothetical Rough Fire scenarios. Similar information is shown for MDA8
O3 in Figure 5-17.
Ambient PM2.5
Hypothetical larger fire
— Actual 2015 Rough fire
Hypothetical smaller fire
A/w^
/w.
AJis
wv;
/\ A
/ v\.
l\
8/1/2015 8/&2015 8/15/2015 8/22/2015 8/29/2015 9/5/2015 9/12/2015 9/19/2015 9/26/2015
Episode Day
Ambient MDA8 03
Hypothetical larger fire
— Actual 2015 Rough fire
Hypothetical smaller fire
:A
8/1/2015 8W2015 8/15/2015 8/22/2015 8/29/2015 9/5/2015 9/12/2015 9/19/2015 9/26/2015
Episode Day
Population Exposure PM2.5
Hypothetical larger fire
Actual 2015 Rough fire
Hypothetical smaller fire
J 1-
1=
a
\
s\
A
l\
/ \
Population Exposure MDA8 03
Hypothetical larger fire
Actual 2015 Rough fire
Hypothetical smaller fire
J*Lm
1 A
8/1/2015 a'8/2015 8/15/2015 8/22/2015 8/29/2015 9/5/2015 9/12/2015 9/19/2015 9/26/2015
Episode Day
8/1/2015 8/8/2015 8/15/2015 8/22/2015 8/29/2015 &5/2015 9/12/2015 9/19/2015 9/26/2015
Episode Day
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; O3 = ozone; PM2.5 = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 |jm; ppb = parts per billion.
Figure 5-17 Daily average PM2.5 ambient (left) and maximum daily 8-hour
average (MDA8) ozone (O3) (right) impacts and aggregate
population exposure (bottom row) from the Rough Fire scenarios.
Daily average impacts are the same for each scenario during the first month of the fire because
the emissions are the same. The alternative scenarios diverge from the actual fire at the beginning of
September. Aggregate population exposure is greatest when the model predicts impacts in the Central
Valley of California for a period in early September and again to a lesser extent in mid-September.
Ambient impacts are reduced in the smaller fire hypothetical scenario once the actual fire progresses to
the Boulder Creek Unit 1 area and increases in the larger fire hypothetical scenario when the actual fire
also includes the area of the Sheep Complex Fire.
5-34
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Figure 5-18 shows daily PM2 5 measurements and model predictions at multiple monitors in the
Central Valley of California. These monitors were selected to provide an indication about how well the
model captures smoke impacts from the Rough Fire. The model tends to overpredict PM2 5 impacts at
these monitors when a large contribution from the Rough Fire is predicted. However, there were days at
Visalia when the model underpredicted PM2 5 impacts toward the beginning of early September. These
overpredictions may be related to PM2 5 emissions, physical treatment of the plume (evaporation and
condensation processes), transport, grid resolution, or some combination of these factors. The large
estimated population exposures of the Rough Fire are most likely overstated during the early September
period of high modeled fire impacts in the Central Valley of California. Some of the model overprediction
at monitors that were impacted by smoke may be related to the model treating primarily emitted organic
aerosol as nonvolatile. If some amount of the primarily emitted organic aerosol was allowed to evaporate
in the model, then downwind surface concentrations would be smaller. This treatment would result in
model predictions closer to measurements as fire impact monitors were often over-predicted (Figure
5-18).
5-35
DRAFT: Do Not Cite or Quote
-------
80
E
. so -
+ Observation
• Model baseline
Visalia
+
+
¦g 40
CD
S 20
O
§•
++++
•"*"+•++ +
++1
r
o
O
0
"O
_Q)
100
- 80
- 60
40
20
0
100
80
|- 60
40
20
0
100
- 80
- 60
E40
20
n
I I I I I I I I I I I I I I I I I I I I I I I
I I I I I I I I
I I I I I I I I I I I I I I I I I I I I I I
08/01 08/06 08/111 08/16 08/21 08/26 08/31 09/05 09/10 09/15 09/20 09/25 09/30
Date
|jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to
2.5 pm.
Note: Model predictions are shaded by the percent contribution from the actual Rough Fire
Figure 5-18 Daily average PM2.5 observations and model predictions at
monitors in the Central Valley of California for August and
September 2015.
1
2 Ambient impacts of the hypothetical Boulder Creek Unit 1 prescribed fire (Figure 5-19) are
3 notably smaller on the last 2 days than the first 3 days. Aggregate population exposures are high on 1 day
4 toward the end of the prescribed fire when winds blew smoke toward the Central Valley of California. It
5-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
is possible that the grid resolution used in this study may exaggerate estimates of population exposure as
terrain-influenced meteorology may be not well resolved with 12-km-sized grid cells for this particular
fire. The 12-km-sized grid cell resolution was chosen for the Rough Fire related scenarios to capture
potential continental scale impacts at the expense of capturing near-fire orographic effects. While daily air
quality impacts from the Boulder Creek Unit 1 prescribed fire are similar in magnitude with some days of
the Rough Fire, the estimates of population exposure are much smaller due to the meteorology on those
days not transporting smoke to large population areas in central California and isolated to a much smaller
number of days.
Ambient PM2.5
Hypothetical Boulder Creek Unit 1 lire
9/27/2014 9/30/2014 10/03/2014
Episode Day
Ambient MDA8 03
Hypothetical Boulder Creek Unit 1 fire
10/03/2014
Episode Day
Population Exposure MDA8 03
<5. 8 Hypothetical Boulder Creek Unit 1 fire |
&
10/03/2014
Episode Day
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; 03 = ozone; PM2.5 = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 |jm.
Figure 5-19 Daily average PM2.5 ambient (left) and maximum daily 8-hour
average (MDA8) ozone (O3) (right) impacts and aggregate
population exposure (bottom row) from the hypothetical Boulder
Creek Unit 1 prescribed fire.
Daily air quality impacts of the actual Sheep Complex Fire in 2010 (Figure 5-20) are fairly steady
with respect to ambient concentrations and aggregate population exposure. A short period of high PM and
O3 impacts in populated areas was evident at the end of the fire in late September when the model
predicted winds transporting smoke to more populated areas of the Central Valley in California. The daily
5-37
DRAFT: Do Not Cite or Quote
-------
ambient concentrations of the Sheep Complex Fire tend to be lower than the Rough Fire and aggregate
population exposures are much lower than the Rough Fire. This is attributed to the smaller amount of
biomass burned on a given day during the Sheep Complex Fire compared with the Rough Fire.
Ambient PM2.5
Actual 2010 Sheep Complex fire
A/V
T ! 1 1 1 1 I 1
7/30/2010 8/6/2010 8/13/2010 8/20/2010 8/27/2010 9/3/2010 9/10/2010 9/17/2010 9/24/2010
Ambient MDA8 03
Actual 2010 Sheep Complex fire
-1 r~
7/30/2010 8/6/2010 &13/2010 8/20/2010 8/27/2010 9/3/2010 9/10/2010 9/17/2010 9/24/2010
Population Exposure PM2.5
Actual 2010 Sheep Complex fire
-8.
Q. o
s H
Population Exposure MDA8 03
Actual 2010 Sheep Complex fire
7/30/2010 8/6/2010 8/13/2010 8/20/2010 8/27/2010 9/3/2010 9/10/2010 9/17/2010 9/24/2010
Episode Day
7/30/2010 8/6/2010 8/13/2010 8/20/2010 8/27/2010 9/3/2010 9/10/2010 9/17/2010 9/24/2010
Episode Day
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; 03 = ozone; PM2.5 = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 |jm; ppb = parts per billion.
Figure 5-20 Daily average ambient PM2.5 (left) and maximum daily 8-hour
average (MDA8) ozone (O3) (right) concentrations and estimates
of aggregate population exposure (bottom row) from the 2010
Sheep Complex Fire.
5.4
Limitations, Implications, and Recommendations
Since the air quality impacts of these wildfire and prescribed fire scenarios occur over different
time scales the aggregation of impacts is presented later in this report in the section covering human
health effects (see CHAPTER 8) with a synthesis of the results of the air quality modeling and health
5-38
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
impact analyses in CHAPTER 9. A summary of highlights from the air quality modeling of the case study
fires follows.
• Surface fuel load characterization is an important component of modeling air quality impacts
associated with wildfires and prescribed fires.
• Outputs from two established fuel load characterization models, FCCS and VELMA, were
merged and fed into the Blue Sky Pipeline to simulate air quality impacts associated with wildfire
and prescribed fire simulations for the TC6 and Rough Fire case studies.
• Whereas FCCS excels at providing estimates of management-sensitive surface and understory
fuel types and loads, VELMA excels at characterizing overstory/canopy fuel loads through its use
of linked forest inventory and satellite-based (LEMMA) data. The combined use of FCCS and
VELMA for this purpose plays to the strengths of both models to better characterize fine-scale
horizontal and vertical distributions of fuelbed types and loadings.
• Applied photochemical grid model to estimate PM2 5 and O3 impacts from an actual wildfire in
Oregon and California
• Photochemical model was also used to estimate how PM2 5 and O3 impacts change for
hypothetical smaller and larger realizations of the actual fires
In considering the assumptions and approach used in the air quality modeling for the case studies
presented in this report, it is important to consider the limitations of these analyses to ensure the results
are interpreted in the proper context. The prescribed fire impacts presented here represent a small subset
of meteorological conditions, fuel loadings, and timing choices and may not be reflective of potential
impacts on air quality in other areas or under different conditions.
For example, despite widespread prescribed fire activity in the southeastern U.S., there are
currently no areas in the Southeast that are noncompliant with the PM or O3 National Ambient Air
Quality Standard. This widespread regional compliance with existing NAAQS across the Southeast
suggests that carefully chosen timing of prescribed fire coupled with anthropogenic control programs can
provide an opportunity for meeting land management goals without compromising public health.
However, when prescribed burning activity is concentrated into a small window of time, which is typical
for example in the Flint Hills region of central Kansas, the enormous amount of fuel being burned on a
few days has led to downwind monitors with O3 and PM2 5 sometimes exceeding the level of the NAAQS
(Baker et al.. 2019).
One challenge related to scale is understanding how the case study information provided in this
report would translate to larger fires (size, duration) or larger regions where many fires would be on the
landscape. The case studies within this assessment are somewhat limited in considering trade-offs over
time because land management techniques would be conducted over multiple years to meet historical fire
return interval goals while these case studies are episodic. Further, information about how many
acres/total fuel needs to be burned in addition to the time interval between burns is needed to place the
information here into a broader context of land management and air quality impacts.
5-39
DRAFT: Do Not Cite or Quote
-------
1 Future studies should attempt to include emissions related to fire suppression activity and model
2 near-fire impacts using a horizontal grid resolution that would best capture complex terrain impacts on
3 wind patterns.
4 While the interactions between prescribed burns and wildfire characteristics is an active area of
5 research (Hunter and Robles. 2020). more information is needed to understand and apply these dynamics
6 quantitatively in air quality models, especially at the regional and national scales. The lack of a
7 generalizable, mechanistic understanding of the influence of prescribed burning and other land treatments
8 on wildfire activity (and consequently on air pollution due to wildfires) remains a major source of
9 uncertainty when projecting future changes in fire-related air quality impacts, especially in areas where
10 prescribed burning is a common practice.
5-40
DRAFT: Do Not Cite or Quote
-------
5.5 References
Abdelnour. A: McKane. RB: Stieglitz. M: Pan. F: Cheng. Y. (2013). Effects of harvest on carbon and nitrogen
dynamics in a Pacific Northwest forest catchment. Water Resour Res 49: 1292-1313.
http://dx.doi.org/10.1029/2012WR012994
Abdelnour. A: Stieglitz. M: Pan. F: McKane. R. (2011). Catchment hydrological responses to forest harvest
amount and spatial pattern. Water Resour Res 47: W09521. http://dx.doi.org/10.1029/2010WR010165
Akagi. SK: Yokelson. RJ: Wiedinmver. C: Alvarado. MJ: Reid. JS: Karl. T: Crounse. JD: Wennberg. PO.
(2011). Emission factors for open and domestic biomass burning for use in atmospheric models. Atmos
ChemPhys 11: 4039-4072. http://dx.doi.org/10.5194/acp-ll-4039-2011
Andreae. MO. (2019). Emission of trace gases and aerosols from biomass burning - An updated assessment.
Atmos ChemPhys 19: 8523-8546. http://dx.doi.org/10.5194/acp-19-8523-2019
Baker. K: Rao. V: Beidler. J: Vukovich. J: Koplitz. S: Avev. L. (2020). Illustrating wildland fire air quality
impacts using an EPA emission inventory [Magazine]. EM: Environmental Manager, 24, 26-31.
Baker. KR: Koplitz. SN: Foley. KM: Avev. L: Hawkins. A. (2019). Characterizing grassland fire activity in the
Flint Hills region and air quality using satellite and routine surface monitor data. Sci Total Environ 659:
1555-1566. http://dx.doi.Org/10.1016/i.scitotenv.2018.12.427
Baker. KR: Misenis. C: Obland. MP: Ferrare. RA: Scaring. A. mvJ: Kelly. JT. (2013). Evaluation of surface and
upper air fine scale WRF meteorological modeling of the May and June 2010 CalNex period in California.
Atmos Environ 80: 299-309. http://dx.doi.Org/10.1016/i.atmosenv.2013.08.006
Baker. KR: Woody. MC: Tonnesen. GS: Hutzell. W: Pve. HOT: Beaver. MR: Pouliot. G: Pierce. T. (2016).
Contribution of regional-scale fire events to ozone and PM2.5 air quality estimated by photochemical
modeling approaches. Atmos Environ 140: 539-554. http://dx.doi.Org/10.1016/i.atmosenv.2016.06.032
Baker. KR: Woody. MC: Valin. L: Szvkman. J: Yates. EL: Iraci. LT: Choi. HP: Soia. AJ: Koplitz. SN: Zhou. L:
Campuzano-Jost. P: Jimenez. JL: Hair. JW. (2018). Photochemical model evaluation of 2013 California wild
fire air quality impacts using surface, aircraft, and satellite data. Sci Total Environ 637-638: 1137-1149.
http://dx.doi.Org/10.1016/i.scitotenv.2018.05.048
Barnett. CR. (2002). BFD curve: A new empirical model for fire compartment temperatures. Fire Safety Journal
37: 437-463. http://dx.doi.org/10.1016/S0379-7112(02)00006-l
Barnhart. BL: McKane. R: Brookes. A: Schumaker. N: Papenfus. M: Pettus. P: Halama. J: Powers. B: Diang. K:
Groskinskv. B: Grier. G: Hawkins. A: Tapp. J: Watson. D: Gross. T: Goodia D: Mohler. R. (2015).
Integrated modeling to assess the ecological and air quality trade-offs of agricultural burning in the Flint
Hills of eastern Kansas. Abstract presented at American Geophysical Union Fall Meeting, December 14-18,
2015, San Francisco, CA.
Bash. JO: Baker. KR: Beaver. MR. (2016). Evaluation of improved land use and canopy representation in BEIS
v3.61 with biogenic VOC measurements in California. GMD 9: 2191-2207. http://dx.doi.org/10.5194/gmd-9-
2191-2016
Bell. DM: Gregory. MJ: Kane. V: Kane. J: Kennedy. RE: Roberts. HM: Yang. Z. (2018). Multiscale divergence
between Landsat- and lidar-based biomass mapping is related to regional variation in canopy cover and
composition. Carbon Balance and Management 13: 15. http://dx.doi.org/10.1186/sl3021-018-0104-6
Brev. SJ: Ruminski. M: Atwood. SA: Fischer. EV. (2018). Connecting smoke plumes to sources using Hazard
Mapping System (HMS) smoke and fire location data over North America. Atmos Chem Phys 18: 1745-
1761. http://dx.doi.org/10.5194/acp-18-1745-2018
Brown. JK: See. TE. (1981). Downed dead woody fuel and biomass in the northern Rocky Mountains. (INT-
GTR-117). Ogden, UT: U.S. Dept. of Agriculture, Forest Service, Intermountain Forest and Range
Experiment Station.
5-41
DRAFT: Do Not Cite or Quote
-------
Brown. JK: See. TE. (1986). Surface fuel loadings and predicted fire behavior for vegetation types in the
northern Rocky Mountains. (Research Note INT-358). Ogden, UT: U.S. Dept. of Agriculture, Forest Service,
Intermountain Forest and Range Experiment Station.
Carlton. AG: Bhave. PV: Napelenok. SL: Ednev. EQ: Sarwar. G: Pinder. RW: Pouliot. GA: Houvoux. M.
(2010). Model representation of secondary organic aerosol in CMAQv4.7. Environ Sci Technol 44: 8553 -
8560. http://dx.doi.org/10.1021/esl00636q
Chen. LWA: Verburg. P: Shackelford. A: Zhu. D: Susfalk. R: Chow. JC: Watson. JG. (2010). Moisture effects
on carbon and nitrogen emission from burning of wildland biomass. Atmos Chem Phys 10: 6617-6625.
http://dx.doi.org/10.5194/acp-10-6617-2010
CMAS (University of North Carolina, Community Modeling and Analysis System). (2020). SMOKE (Sparse
Matrix Operator Kerner Emissions) Modeling System (Version 4.8) [Computer Program]. Chapel Hill, NC:
University of North Carolina, Institute for the Environment. Retrieved from
https://www.cmascenter.org/smoke/
Davis. RJ: Ohmann. JL: Kennedy. RE: Cohen. WB: Gregory. MJ: Yang. Z: Roberts. HM: Gray. AN: Spies. TA.
(2015). Northwest Forest Plan-The first 20 years (1994-2013): Status and trends of late-successional and old-
growth forests. (PNW-GTR-911). Portland, OR: Department of Agriculture, Forest Service, Pacific
Northwest Research Station, http://dx.doi.org/10.2737/PNW-GTR-911
Emery. C: Jung. J: Koo. B: Yarwood. G. (2015). Improvements to CAMx snow cover treatments and carbon
bond chemical mechanism for winter ozone. (UDAQ PO 480 52000000001). Novato, CA: Ramboll Environ.
http://www.camx.com/files/udaa snowchem final 6augl5.pdf
Fahev. KM: Carlton. AG: Pve. HOT: Baeka. J: Hutzell. WT: Stanier. CO: Baker. KR: Appel. KW: Jaoui. M:
Offenberg. JH. (2017). A framework for expanding aqueous chemistry in the Community Multiscale Air
Quality (CMAQ) model version 5.1. GMD 10: 1587-1605. http://dx.doi.org/10.5194/gmd-10-1587-2017
FERA (Fire and Environmental Research Applications Team). (2020). Fuel and Fire Tools (FFT) (Version
2.0.2017) [Computer Program], Seattle, WA: University of Washington, College of Environmental and
Forest Resources. Retrieved from https://www.fs.usda.gov/pnw/tools/fuel-and-fire-tools-fft
Ford. B: Val Martin. M: Zelaskv. SE: Fischer. EV: Anenberg. SC: Heald. CL: Pierce. JR. (2018). Future fire
impacts on smoke concentrations, visibility, and health in the contiguous United States. Geohealth 2: 229-
247. http://dx.doi.org/10.1029/2018GH00Q144
Fountoukis. C: Nenes. A. (2007). ISORROPIAII: A computationally efficient thermodynamic equilibrium
model for K+-Ca2+-Mg2+-Nh(4)(+)-Na+-S042~N03~Cl~H20 aerosols. Atmos Chem Phys 7: 4639-4659.
http://dx.doi.org/10.5194/acp-7-4639-2007
Garcia-Menendez. F: Hu. Y: Odman. MT. (2013). Simulating smoke transport from wildland fires with a
regional-scale air quality model: Sensitivity to uncertain wind fields. J Geophys Res Atmos 118: 6493 -6504.
http://dx.doi.org/10.1002/igrd.50524
Garcia-Menendez. F: Hu. Y: Odman. MT. (2014). Simulating smoke transport from wildland fires with a
regional-scale air quality model: Sensitivity to spatiotemporal allocation of fire emissions. Sci Total Environ
493: 544-553. http://dx.doi.Org/10.1016/i.scitotenv.2014.05.108
GOES-R Algorithm Working Group. GOES-R Program Office.. (2018). NOAA GOES-R Series Advanced
Baseline Imager (ABI) Level 2 Fire/Hot Spot Characterization (FDC). Silver Spring, MD: National Oceanic
and Atmospheric Administration, National Centers for Environmental Information. Retrieved from
https://doi.org/10.7289/V5X065CR
Hu. Y: Odman. MT: Chang. ME: Jackson. W: Lee. S: Edgertoa ES: Baumann. K: Russell. AG. (2008).
Simulation of air quality impacts from prescribed fires on an urban area. Environ Sci Technol 42: 3676-3682.
http://dx.doi.org/10.1021/es071703k
Hunter. ME: Robles. MD. (2020). Tamm review: The effects of prescribed fire on wildfire regimes and impacts:
A framework for comparison [Review]. For Ecol Manage 475: 118435.
http://dx.doi.Org/10.1016/i.foreco.2020.118435
5-42
DRAFT: Do Not Cite or Quote
-------
Jiang. X: Wiedinmver. C: Carlton. AG. (2012). Aerosols from fires: An examination of the effects on ozone
photochemistry in the Western United States. Environ Sci Technol 46: 11878-11886.
http://dx.doi.org/10.1021/es3Q1541k
Kelly. JT: Koplitz. SN: Baker. KR: Holder. AL: Pve. HOT: Murphy. BN: Bash. JO: Henderson. BH: Possiel. N:
Simon. H: Evth. AM: Jang. C: Phillips. S: Timin. B. (2019). Assessing PM2.5 model performance for the
conterminous U.S. with comparison to model performance statistics from 2007-2015. Atmos Environ 214:
116872. http://dx.doi.Org/10.1016/i.atmosenv.2019.116872
Kennedy. RE: Ohmann. J: Gregory. M: Roberts. H: Yang. Z: Bell. DM: Kane. V: Hughes. MJ: Cohen. WB:
Powell. S: Neeti. N: Larrue. T: Hooper. S: Kane. J: Miller. PL: Perkins. J: Braaten. J: Seidl. R. (2018). An
empirical, integrated forest biomass monitoring system. Environ Res Lett 13: 025004.
http://dx.doi.org/10.1088/1748-9326/aa9d9e
Kobziar. LN: Godwia D: Taylor. L: Watts. AC. (2015). Perspectives on trends, effectiveness, and impediments
to prescribed burning in the southern U.S. Forests 6: 561-580. http://dx.doi.org/10.3390/f6030561
Koplitz. SN: Nolte. CG: Pouliot. GA: Vukovich. JM: Beidler. J. (2018). Influence of uncertainties in burned area
estimates on modeled wildland fire PM2.5 and ozone pollution in the contiguous U.S. Atmos Environ 191:
328-339. http://dx.doi.Org/10.1016/i.atmosenv.2018.08.020
Kremens. RL: Dickinson. MB: Bova. AS. (2012). Radiant flux density, energy density and fuel consumption in
mixed-oak forest surface fires. International Journal of Wildland Fire 21: 722-730.
http://dx.doi.org/10.1071/WF10143
Larkin. NK: O'neill. SM: Solomon. R: Raffuse. S: Strand. T: Sullivan. DC: Krull. C: Rorig. M: Peterson. J:
Ferguson. SA. (2009). The BlueSky smoke modeling framework. International Journal of Wildland Fire 18:
906. http://dx.doi.org/10.1071/WF07086
Larkin. NK: Raffuse. SM: Huang. S: Pavlovic. N: Lahm. P: Rao. V. (2020). The Comprehensive Fire
Information Reconciled Emissions (CFIRE) inventory: Wildland fire emissions developed for the 2011 and
2014 U.S. National Emissions Inventory. J Air Waste Manag Assoc 70: 1165-1185.
http://dx.doi.org/10.1080/10962247.2020.1802365
LEMMA (Landscaping Ecology, Modeling, Mapping & Analysis). (2020). Landscaping Ecology, Modeling,
Mapping & Analysis (LEMMA): GNN maps and data. Available online at
https://lemma.forestrv.oregonstate.edu/data (accessed January 22, 2021).
LF (Landfire). (2008). Fuel characteristic classification system (FCCS) fuelbeds (Version 1.1.0). Washington,
DC: U.S. Department of Agriculture and U.S. Department of the Interior. Retrieved from
https://landfire.gov/fccs.php
LF (Landfire). (2014). Existing vegetation type (EVT) (Version 1.4.0). Washington, DC: U.S. Department of
Agriculture and U.S. Department of the Interior. Retrieved from https://www.landfire.gov/evt.php
Liu. JC: Micklev. LJ: Sulprizio. MP: Dominici. F: Yue. X. u: Ebisu. K: Anderson. GB: Khan. RFA: Bravo. MA:
Bell. ML. (2016). Particulate air pollution from wildfires in the Western US under climate change. Clim
Change 138: 655-666. http://dx.doi.org/10.1007/slQ584-016-1762-6
Liu. T: Micklev. LJ: Marlier. ME: DeFries. RS: Khan. MF: Latif. MT: Karambelas. A. (2020). Diagnosing
spatial biases and uncertainties in global fire emissions inventories: Indonesia as regional case study. Rem
Sens Environ 237: 111557. http://dx.doi.Org/10.1016/i.rse.2019.111557
Lutes. DC: Keane. RE: Caratti. JF. (2009). A surface fuel classification for estimating fire effects. International
Journal of Wildland Fire 18: 802-814. http://dx.doi.org/10.1071/WF08Q62
May. N: Ellicott. E: Gollner. M. (2019). An examination of fuel moisture, energy release and emissions during
laboratory burning of live wildland fuels. International Journal of Wildland Fire 28: 187-197.
http://dx.doi.org/10.1071/WF18Q84
5-43
DRAFT: Do Not Cite or Quote
-------
McKane. R: Brookes. A: Diang. K: Stieglitz. M: Abdelnour. A: Pan. F: Halama. J: Pettus. P: Phillips. D. (2014).
Visualizing Ecosystem Land Management Assessments (VELMA) v. 2.0: User manual and technical
documentation. (Document control number L-PESD-30840-QP-1-2). Corvallis, OR: U.S. Environmental
Protection Agency, National Health and Environmental Effects Research Laboratory.
https://www.epa.gov/sites/production/files/2016-01/documents/velma 2.0 user manual.pdf
McKane. RB: Brookes. AF: Diang. KS: Halama. JJ: Pettus. PB: Barnhart. BL: Russell. M: Vache. KB: Bolte. JP.
(2020). An integrated multi-model decision support framework for evaluating ecosystem-based management
options for coupled human-natural systems. In TG O'Higgins; M Lago; TH DeWitt (Eds.), Ecosystem-based
management, ecosystem services and aquatic biodiversity: Theory, tools and applications (pp. 255-274). New
York, NY: Spring. http://dx.doi.org/10.1007/978-3-03Q-45843-0 13
NCEP (National Weather Service, National Centers for Environmental Prediction). (2021). North American
Mesoscale Forecast System (NAM). Silver Spring, MD: National Oceanic and Atmospheric Administration,
National Centers for Environmental Information. Retrieved from https ://www. ncdc .noaa. gov/data-
access/model-data/model-datasets/north-american-mesoscale-forecast-svstem-nam
NPS (U.S. National Park Service). (2016). Rough Fire interactive map. Available online at
https://www.nps.gov/seki/learn/nature/rough-fire-interactive-map.htm (accessed January 27, 2021).
O'Dell. K: Ford. B: Fischer. EV: Pierce. JR. (2019). Contribution of wildland-fire smoke to US PM2.5 and its
influence on recent trends. Environ Sci Technol 53: 1797-1804. http://dx.doi.org/10.1021/acs.est.8b05430
Ottmar. RD: Sandberg. DV: Riccardi. CL: Prichard. SJ. (2007). An overview of the fuel characteristic
classification system—quantifying, classifying, and creating fuelbeds for resource planning. Can J For Res
37: 2383-2393. http://dx.doi.org/10.1139/X07-077
Possell. M: Bell. TL. (2013). The influence of fuel moisture content on the combustion of Eucalyptus foliage.
International Journal of Wildland Fire 22: 343-352. http://dx.doi.org/10.1071/WF12Q77
Prichard. S: Larkin. NS: Ottmar. R: French. NHF: Baker. K: Brown. T: Clements. C: Dickinson. M: Hudak. A:
Kochanski. A: Linn. R: Liu. Y: Potter. B: Mell. W: Tanzer. D: Urbanski. S: Watts. A. (2019). The fire and
smoke model evaluation experiment-a plan for integrated, large fire-atmosphere field campaigns [Review].
Atmosphere (Basel) 10: 66. http://dx.doi.org/10.3390/atmoslQ020066
Prichard. SJ: Andreu. AG: Drve. B. (2021). Consume 5.0 technical documentation (in prep.). Corvallis, OR: U.S.
Department of Agriculture Forest Service Pacific Northwest Research Station.
Prichard. SJ: O'Neill. SM: Eagle. P: Andreu. AG: Drve. B: Dubowv. J: Urbanski. S: Strand. TM. (2020).
Wildland fire emission factors in North America: Synthesis of existing data, measurement needs and
management applications. International Journal of Wildland Fire 29: 132-147.
http://dx.doi.org/10.1071/WF19Q66
Prichard. SJ: Peterson. PL: Jacobson. K. (2010). Fuel treatments reduce the severity of wildfire effects in dry
mixed conifer forest, Washington, USA. Can J For Res 40: 1615-1626. http://dx.doi.org/10.1139/X10-109
Raffuse. S: Prvden. DA: Sullivan. DC: Larkin. NK: Strand. T: Solomon. R. (2009). SMARTFIRE algorithm
description. (STI-905517- 3719). Washington, DC: U.S. Environmental Protection Agency.
https://firesmoke.ca/smartfire/pdfs/SMARTFIRE Algorithm Description Final.pdf
Ryan. KC: Knapp. EE: Varner. JM. (2013). Prescribed fire in North American forests and woodlands: History,
current practice, and challenges. Front Ecol Environ 11: E15-E24. http://dx.doi.org/10.1890/12Q329
Schmidt. C: Hoffman. J: Prins. E: Lindstrom. S. (2013). GOES-R advanced baseline imager (ABI) algorithm
theoretical basis document for fire/hot spot characterization, version 2.6. College Park, MD: National
Oceanic and Atmospheric Administration, Center for Satellite Applications and Research.
https://www.star.nesdis.noaa.gov/goesr/documents/ATBDs/Baseline/ATBD GOES-
R FIRE v2.6 Qct2013.pdf
Simon. H: Baker. KR: Phillips. S. (2012). Compilation and interpretation of photochemical model performance
statistics published between 2006 and 2012. Atmos Environ 61: 124-139.
http://dx.doi.Org/10.1016/i.atmosenv.2012.07.012
5-44
DRAFT: Do Not Cite or Quote
-------
Simon. H: Bhave. PV. (2012). Simulating the degree of oxidation in atmospheric organic particles. Environ Sci
Techno146: 331-339. http://dx.doi.org/10.1021/es2Q2361w
Skamarock. WC: Klemp. JB: Dudhia. J: Gill. DO: Barker. DM: Duda. MG: Huang. X: Wang. W: Powers. JG.
(2008). A description of the advanced research WRF version 3. (NCAR/TN-475+STR). Boulder, Colorado:
U.S. National Center for Atmospheric Research. http://dx.doi.org/10.5065/D68S4MVH
Smithwick. EAH: Harmon. ME: Remillard. SM: Acker. SA: Franklin. JF. (2002). Potential upper bounds of
carbon stores in forests of the Pacific Northwest. Ecol Appl 12: 1303-1317. http://dx.doi.org/10.189Q/1051-
0761(2002)01211303 :PUBOCS12.Q.CO:2
Surawski. NC: Sullivan. AL: Mever. CP: Roxburgh. SH: Polglase. PJ. (2015). Greenhouse gas emissions from
laboratory-scale fires in wildland fuels depend on fire spread mode and phase of combustion. Atmos Chem
Phys 15: 5259-5273. http://dx.doi.org/10.5194/acp-15-5259-2015
Tihav-Felicelli. V: Santoni. PA: Gerandi. G: Barboni. T. (2017). Smoke emissions due to burning of green waste
in the Mediterranean area: Influence of fuel moisture content and fuel mass. Atmos Environ 159: 92-106.
http://dx.doi.Org/10.1016/i.atmosenv.2017.04.002
U.S. EPA (U.S. Environmental Protection Agency). (2019a). Environmental Benefits Mapping and Analysis
Program - Community Edition (BenMAP-CE) (Version 1.5) [Computer Program], Washington, DC.
Retrieved from https://www.epa.gov/benmap/benmap-communitv-edition
U.S. EPA (U.S. Environmental Protection Agency). (2019b). Technical support document (TSD): Preparation of
emissions inventories for the version 7.2 2016 - North American emissions modeling platform.
https://www.epa.gov/sites/production/files/2019-
09/documents/2016v7.2 regionalhaze emismod tsd 508.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2020a). CMAQ: The Community Multiscale Air Quality
Modeling System (Version 5.3.2) [Computer Program]: Zenodo. Retrieved from
https://www.epa.gov/cmaa/download-cmaa
U.S. EPA (U.S. Environmental Protection Agency). (2020b). SPECIATE (Version 5.1) [Database]. Washington,
DC. Retrieved from https://www.epa.gov/air-emissions-modeling/speciate-2
University of Washington. (2014). Piled fuels biomass and emissions calculator. Available online at
https://depts.washington.edu/nwfire/piles/ (accessed January 27, 2021).
Urbanski. S. (2014). Wildland fire emissions, carbon, and climate: Emission factors. For Ecol Manage 317: 51-
60. http://dx.doi.Org/10.1016/i.foreco.2013.05.045
Urbanski. SP. (2013). Combustion efficiency and emission factors for wildfire-season fires in mixed conifer
forests of the northern Rocky Mountains, US. Atmos Chem Phys 13: 7241-7262.
http://dx.doi.org/10.5194/acp-13-7241-2013
USFS (U.S. Forest Service). (2014). News release: Forest begins prescribed burning in Boulder Creek drainage.
Dunlap, CA: U.S. Forest Service, Sequoia National Forest and Giant Sequoia National Monument.
Wilkins. JL: Pouliot. G: Foley. K: Appel. W: Pierce. T. (2018). The impact of US wildland fires on ozone and
particulate matter: A comparison of measurements and CMAQ model predictions from 2008 to 2012.
International Journal of Wildland Fire 27: 684-698. http://dx.doi.org/10.1071/WF18053
Yee. S: Bousauin. J: Bruins. R: Canfield. TJ: DeWitt. TH: de Jesus-Crespo. R: Dyson. B: Fulford. R: Harwell.
M: Hoffman. J: Littles. CJ: Johnston. JM: McKane. RB: Green. L: Russel. M: Sharpe. L: Seeteram. N:
Tashie. A: Williams. K. (2017). Practical strategies for integrating final ecosystem goods and services into
community decision-making. (EPA/600/R-17/266). Washington, DC: U.S. Environmental Protection
Agency, https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P 100SGRC.txt
Yue. X: Micklev. LJ: Logan. JA: Kaplan. JO. (2013). Ensemble projections of wildfire activity and
carbonaceous aerosol concentrations over the western United States in the mid-21st century. Atmos Environ
77: 767-780. http://dx.doi.Org/10.1016/i.atmosenv.2013.06.003
5-45
DRAFT: Do Not Cite or Quote
-------
Zhou. L: Baker. KR: Napelenok. SL: Pouliot. G: Elleman. R: O'Neill. SM: Urbanski. SP: Wong. DC. (2018).
Modeling crop residue burning experiments to evaluate smoke emissions and plume transport. Sci Total
Environ 627: 523-533. http://dx.doi.Org/10.1016/i.scitotenv.2018.01.237
5-46
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
CHAPTER 6 WILDLAND FIRE SMOKE
EXPOSURE CHARACTERIZATION
AND HEALTH AND ECOLOGICAL
IMPACTS
6.1 Introduction
Wildland fire (i.e., prescribed fire and wildfire) smoke can have detrimental effects on both
human and ecological health, but can also provide ecological benefits (see Section 6.4.1.2') as well as
cultural benefits when used as part of indigenous cultural fires (Raish et al.. 2005). While the health
impacts of wildfire smoke exposure can be quantitatively estimated using the Environmental Benefits
Mapping and Analysis Program—Community Edition rBenMAP-CE;U.S. EPA (2019a)l. it is much more
challenging to quantify the potential ecological impacts. This chapter summarizes the health effects
attributed to wildfire smoke exposure, with a focus on U.S.-based epidemiologic studies; characterizes the
different actions and interventions that can be employed at a population and individual level to reduce
smoke exposure; and highlights the ecological impacts attributed to wildfire smoke. The evaluation of
epidemiologic studies is meant to inform the estimation of potential health impacts of smoke from
wildfires based on different fire management strategies, including prescribed fire, within the case study
areas using BenMAP-CE (see CHAPTER S).
In assessing the evidence base spanning both human and ecological health, the current
understanding of impacts from wildland fire smoke primarily stems from studies examining effects due to
exposures to ambient fine particulate matter (PM2 5; particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |im ), with a growing body of evidence focusing specifically on wildfire
smoke, and only a few studies focusing on prescribed fires. While smoke also contains precursors that can
lead to ozone formation downwind from a wildland fire (see CHAPTER 5). fewer studies have examined
wildfire-specific health impacts attributed to ozone. However, extensive evidence demonstrating health
effects from ambient ozone exposures indicates the potential for ozone formed from wildfires to result in
an additional significant public health burden (U.S. EPA. 2020a).
The extent of prescribed fire and wildfire smoke exposure depends on proximity to the fire and
the location (i.e., not everyone is exposed to smoke from fires), duration, and intensity of smoke plumes.
Therefore, it is plausible that individuals can take actions to reduce or mitigate exposure to smoke from
prescribed fires or wildfires. In addition to identifying the potential human health impacts of smoke
exposure, this chapter also evaluates and characterizes the effectiveness of various actions that can be
taken to reduce smoke exposure and subsequently protect public health.
6-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
6.2
Wildfire Smoke Exposure and Health
Scientific evidence examining the health effects attributed to wildfire smoke exposure has grown
significantly in recent years. The underpinnings of this evidence are rooted in the decades of research
examining the health effects of ambient air pollutants, many of which are components of wildfire smoke.
Of these components, particulate matter, specifically PM2 5, is a main component and has been shown to
have a significant public health impact, which is demonstrated by the range of health effects attributed to
PM2 5 exposure, including respiratory and cardiovascular effects, as well as mortality (Jaffe et al.. 2020;
U.S. EPA. 2019b). Recent epidemiologic and experimental studies examining the health effects of
wildfire smoke exposure report findings that are generally consistent with the broad body of evidence
from studies examining short-term PM2 5 exposure (U.S. EPA. 2019c; Black et al.. 2017; Reid et al..
2016). The consistency in results across studies of wildfire smoke exposure and PM2 5 are further
supported by studies that compared the health effects associations between various sources of PM2 5,
including wildfire smoke, and ambient PM2 5. These studies have not provided evidence indicating
differences in the risk of health effects between different sources of PM2 5 and total ambient PM2 5
(DeFlorio-Barker et al.. 2019; U.S. EPA. 2019b). However, it is important to note that experimental
studies have provided evidence of differential toxicity and mutagenicity due to both the flaming and
smoldering of different individual fuel sources, which may be important to consider when examining the
trade-offs between prescribed fire and wildfire (Kim et al.. 2018).
Most studies that examine the health effects of wildfire smoke exposure at the population level
focus broadly on wildfire smoke, without accounting for potential differences in smoke emissions
between prescribed fire and wildland fire. A recent epidemiologic study conducted by Prunicki et al.
(2019) reported initial evidence of differences in markers of immune function, DNA methylation, and
worsened respiratory outcomes in school-aged children in Fresno, CA exposed to wildfire smoke
compared to prescribed fire smoke. The difference in effects observed coincided with lower
concentrations of air pollutants from prescribed fires compared to wildfires. However, it is unclear what
aspects of the difference between prescribed fires and wildfires resulted in the differential health effects
(e.g., differences in duration, air pollutant concentrations, fuel types, burn conditions). Although Prunicki
et al. (2019) provides initial evidence of potential differences in subclinical effects due to prescribed fire
versus wildfire smoke exposure, it remains unclear if there are differences in more overt health effects
(e.g., hospital admissions, mortality) between the two fire types.
Overall, wildfire smoke exposure studies report results that are generally consistent with
epidemiologic studies of short-term PM2 5 exposure, and are also remarkably consistent with each other
considering the large degree of variability in both the exposure indicators (e.g., PM2 5, wildfire-specific
PM2 5, smoke day) and exposure assessment methodologies employed. This variability across smoke
indicator metrics directly influences the utility of results in quantitative assessments, such as a risk
assessment or a cost-benefit analysis. Furthermore, as noted previously, there is limited evidence
regarding the health effects attributed to ozone derived from wildland fire smoke even though there is
6-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
extensive evidence of numerous health effects from studies of ambient ozone exposure (U.S. EPA.
2020a). As a result, this section consists of an evaluation of epidemiologic studies conducted within the
U.S., published through December 2020, that could be used, either alone or in combination with studies
of ambient PM2 5 and ozone, in a quantitative assessment of the potential health impacts associated with
different fire management strategies in the case study areas (i.e., Timber Crater 6 [TC6] Fire and Rough
Fire), identified in earlier chapters, using U.S. EPA's Environmental BenMAP-CE, (see CHAPTER 7V
Based on the majority of the wildfire smoke epidemiologic studies focusing on PM2 5 and because of the
consistency in health effects between studies of short-term PM2 5 and wildfire smoke exposure the
epidemiologic studies evaluated within this section consist of those that examined health outcomes where
the most recent U.S. EPA Integrated Science Assessment for Particulate Matter concluded that the
evidence indicates either a "causal relationship" or "likely to be causal relationship" (i.e., respiratory and
cardiovascular effects, and mortality). This approach is in line with the criteria used by the U.S. EPA in
the process of conducting BenMAP-CE analyses.
This assessment of the health effects of wildfire smoke exposure is not intended to be an
exhaustive review of the evidence. Recent reviews and interagency efforts have extensively characterized
the current state of the science with respect to the health effects attributed to wildfire smoke exposure
(Jaffe et al.. 2020; U.S. EPA. 2019c; Reid et al.. 2016). In addition, it is important to recognize that the
evaluation with this assessment does not rely on the numerous animal toxicological studies conducted to
date that focused on examining health effects from exposures consisting of wildfire smoke from fuel
sources commonly found in the U.S. (e.g., individual tree species) or real-world wildfire smoke.
6.2.1 Characterization of Wildfire Smoke Exposures
Wildfires are often natural, spontaneous events, which has complicated the ability of
epidemiologic studies to characterize population exposures to wildfire smoke. As a result, studies have
used a variety of approaches to estimate wildfire smoke exposure in terms of both the exposure indicator
and exposure assessment methodology used (Table A.6-1). While the exposure assessment approaches
used across studies vary in complexity and in the specificity of the indicator in representing wildfire
smoke exposure, epidemiologic studies report generally consistent associations between short-term
wildfire smoke exposure and health effects (Section 6.2.2V
6.2.1.1 Exposure Indicator
Within epidemiologic studies, the exposure indicator is a quantity meant to represent exposure to
an environmental contaminant. For wildfire smoke, which consists of a complex mixture of pollutants,
various indicators have been used as a surrogate for wildfire smoke exposure. These indicators vary in
specificity and sensitivity with respect to how well they represent exposure to wildfire smoke. Because of
6-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
the public health implications of exposure to PM2 5 and PM2 5 being a main component of wildfire smoke,
studies often rely on the use of some form of PM2 5 as an exposure indicator. Some epidemiologic studies
used monitored or modeled PM2 5 concentrations as the exposure indicator (Alman et al.. 2016; Reid et
al.. 2016; Delfino et al.. 2009) while other studies used wildfire or smoke-specific PM2 5, which consisted
of removing PM2 5 derived from other PM2 5 sources from the concentrations estimated (Stow ell et al..
2019; Gan et al.. 2017; Rappold et al.. 2012). Additionally, some studies use PM2 5 concentrations or
estimate a range of PM2 5 concentrations from an atmospheric model to develop an exposure indicator
based on classifying days as either smoke or nonsmoke days or by assigning each day a level of smoke
density (i.e., light, medium, or dense). In these studies the defining of days by smoke status often
depended on using criteria to define specific ranges of PM2 5 concentrations that are considered indicative
of wildfire smoke exposure (Jones et al.. 2020; Wettstein et al.. 2018; Liu et al.. 2017b; Liu et al.. 2017a).
The use of a broad exposure indicator, such as smoke days, may be more representative of the
multipollutant nature of wildfire smoke. However, to date there has been no indication that any one
exposure indicator represents a better surrogate of wildfire smoke exposure than another. Overall, the
variability in the exposure indicator used across studies partly reflects the difficulty in examining the
health effects of wildfire smoke exposure, and the air quality data available, or lack thereof in some
instances (see CHAPTER 4).
6.2.1.2 Exposure Assessment Methodology
Estimating wildfire smoke exposure for epidemiologic studies is challenging because wildfire
smoke is spatially and temporally dynamic and areas impacted by wildfire smoke often have few ambient
monitoring sites because most air quality monitors reside in urban locations (see CHAPTER 4). As a
result, epidemiologic studies have resorted to using numerous methods that vary in complexity to assign
exposures (Table A.6-1). In contrast, due to the planned nature of prescribed fires, monitors could be
deployed to capture population exposure, but to date have not been widely used in this capacity and the
data is not always reported (see CHAPTER 4).
Consistent with many epidemiologic studies of ambient air pollution, a few studies examined
relationships between short-term wildfire smoke exposure and health effects using monitored PM2 5
concentrations and some approach to assign exposures to a defined spatial extent, whether that be a city or
ZIP code (Leibel et al.. 2020; Zu et al.. 2016). Most epidemiologic studies focusing on wildfire smoke
exposure use exposure models that rely on data from multiple sources and are often referred to as hybrid
exposure models. These models use both monitoring and modeling data, and in some instances satellite
measurements to take advantage of all the data available to estimate wildfire or smoke-specific PM2 5
concentrations. Incorporating all these data sources into the model allows for a broader spatial extent to
be included in epidemiologic studies instead of being limited to only those locations within reasonable
proximity to air quality monitors, which are primarily in urban centers. Relatively few of the studies that
used hybrid exposure models evaluated model performance, but those studies that did indicate the models
6-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
performed rather well fsee Table A.6-1; Reid et al. (2019); Stowell et al. (2019); Gan et al. (2017); Reid
etal. (2016).
A few epidemiologic studies relied on other approaches to estimating wildfire smoke exposure.
While often included as a component in the exposure model to estimate wildfire smoke, one study used
only satellite measurements (i.e., aerosol optical depth [AOD]) to identify areas that were impacted by a
smoke plume (Rappold et al.. 2011). The remaining studies used various models that were developed to
examine wildfire smoke exposures by estimating either wildfire-specific PM2 5 exposure or smoke
exposure more broadly. Studies that estimated wildfire-specific PM2 5 exposure used the Wildland Fire
Emissions Information System (WFEIS) or National Oceanic and Atmospheric Administration's
(NOAA's) Smoke Forecasting System (SFS) in combination with the transport and dispersion model
HYSPLIT (Hutchinson et al.. 2018; Tinling et al.. 2016; Rappold et al.. 2012) while studies that focused
on smoke days used NOAA's Hazard Mapping System (HMS) to characterize exposure based on the
density of smoke (Jones et al.. 2020; Wettstein et al.. 2018).
Regardless of the exposure assessment approach used, the results across epidemiologic studies
provide evidence supporting a relationship between wildfire smoke exposure and various health effects
(see Section 6.2.2). However, the variability in the exposure approaches used does not allow for the
results from some epidemiologic studies, such as those that used indicator variables to represent wildfire
smoke, to be used for the development of wildfire-specific health impact functions in BenMAP-CE
analyses.
6.2.1.3 Uncertainties and Limitations in Characterizing Wildfire Smoke
Exposure
A challenge in estimating wildland fire smoke exposure, as detailed within CHAPTER 4. is the
fact the current ambient monitoring network was not designed with the goal of measuring smoke from
wildfires or public health surveillance. As a result, as noted within this section, epidemiologic studies
have relied on a variety of approaches to estimate smoke exposure, whether through PM2 5 concentration
data collected from the ambient monitoring network, predicted concentrations from photochemical
transport models, satellite measurements or a combination each, as well as estimations of smoke plumes.
Although results across epidemiologic studies are consistent regardless of the approach used to assign
exposure (see Section 6.2.2). there are inherent uncertainties across each of the approaches employed,
with one of the larger uncertainties being how well exposures represented by smoke plumes reflect PM2 5
concentrations experienced on the ground. However, a recent study by Larsen et al. (2018) that examined
PM2 5 monitoring and smoke plume data and found initial evidence that monitored values on the ground
reflect the presence of smoke plumes in the vertical column measured by satellites. In the future, more
detailed evaluations of the different approaches that can be used and a characterization of their strengths
and weaknesses will aid in further supporting the interpretation of results from epidemiologic studies.
6-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
6.2.2
Health Effects Attributed to Wildfire Smoke Exposure
In the context of wildfires, most U.S.-based studies focus on short-term or daily exposures
(i.e., 24-hour average). Across these studies, the primary pollutant of interest is PM2 5, with only one
study focusing on ozone (Reid et al.. 2019). Studies examining exposure durations shorter than a 24-hour
average, often referred to as subdaily exposures, have been limited to epidemiologic and controlled
human exposure studies of ambient PM25 focusing on subclinical measures of heart or lung function and
not overt population-level effects, such as hospital admissions or mortality (U.S. EPA. 2019b). Therefore,
these studies of subdaily exposures do not directly inform the relationship between shorter duration
wildfire smoke exposures and health effects due to the difficulty in linking a change in a subclinical effect
to an overt health outcome. As a result, most of the evidence informing the current understanding of
health effects attributed to wildfire smoke exposure stems from epidemiologic studies primarily focusing
on examining exposures over single-day or multiday lags ranging from 0 to 5 days.
The focus on examining health effects attributed to short-term wildfire smoke exposures has
resulted in a relative lack of information on the health effects due to repeated wildfire smoke exposures
(i.e., over many days, weeks, or months); the long-term health effects of wildfire smoke exposure from a
single wildfire event; and the health effects due to long-term wildfire exposures over many months and
multiple fire seasons. To date, studies have not examined the impact of repeated wildfire smoke exposure
on health; whereas, an initial study provides preliminary evidence that a wildfire smoke event with high
PM2 5 concentrations may detrimentally impact health, specifically lung function, over multiple
subsequent years (Orr et al.. 2020). The examination of longer-term exposures to wildfire smoke has been
limited to a recent study indicating increased risk of mortality in hemodialysis patients as cumulative
exposures increase up to 30 days (Xi et al.. 2020). analyses of subclinical effects (e.g., changes in lung
function) in wildland fire fighters over multiple fire seasons (Adetona et al.. 2016). and a study examining
the potential implications of wildfire smoke exposure on the influenza season (Landguth et al.. 2020).
6.2.2.1 Respiratory Effects
Most studies to date specifically examining the health effects of wildfire smoke exposure focus
on respiratory-related outcomes (e.g., emergency department [ED] visits, hospital admissions, and
medication use). In addition to the wildfire-specific evidence, there is extensive evidence spanning both
experimental and epidemiologic studies focusing on short-term exposures to ambient PM2 5 demonstrating
a range of respiratory effects, with the strongest evidence supporting relationships with exacerbations of
asthma and chronic obstructive pulmonary disease (COPD), as well as respiratory mortality (U.S. EPA.
2019b).
Epidemiologic studies that examined relationships between short-term wildfire smoke exposure
and respiratory-related outcomes also provide evidence of positive associations, which are consistent with
6-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
the results of studies focusing on ambient PM2 5. The pattern of associations across studies of wildfire
smoke and ambient PM2 5 are generally observed within the first few days after exposure [i.e., at lags in
the range of 0-2 days; U.S. EPA (2019b); Figure 7-1 and Figure 7-2], However, there has been limited
examination of longer durations of exposure (i.e., exposures over multiple days) for both ambient PM2 5
and wildfire exposures and respiratory effects. Initial evidence indicates respiratory effects of larger
magnitude due to prolonged exposure (i.e., over a series of days with lags ranging from 0 to 5 days),
which is important to consider when examining wildfire smoke exposure that often lasts for many weeks
or months (U.S. EPA. 2019b; Rappold et al.. 2011).
Across the epidemiologic studies examining respiratory-related outcomes, the most extensive
evidence comes from studies examining combinations of respiratory-related diseases (i.e., all
International Classification of Diseases [ICD] codes for the entire range of respiratory diseases or a subset
of ICD codes for only a few respiratory diseases, noted as "all respiratory" in Figure 6-1) and asthma.
These studies provide consistent evidence of positive associations for both ED visits and hospital
admissions when using different exposure indicators, including smoke/wildfire PM2 5 or ambient PM2 5
(Figure 6-1). Some of the studies evaluated examined whether there was evidence of differential risk
across age groups (Table A.6-1). and while in some instances the magnitude of the association was
reported to be larger for a specific age range, the results presented in Figure 6-1, capture the main results
of each study.
In addition to the studies that relied on PM2 5 to develop the exposure indicator, studies that used
alternative exposure indicators or applied different techniques to identify wildfire smoke exposures
provide supporting evidence of a relationship between short-term wildfire smoke exposure and respiratory
effects. Studies that used the exposure indicator of smoke plume or smoke density reported evidence of
consistent positive associations when examining both combinations of respiratory-related diseases and
asthma (Wettstein et al.. 2018; Rappold et al.. 2011). Instead of examining associations with respiratory
outcomes, Leibel et al. (2020) in a study conducted in San Diego County, CA examined, and
subsequently reported, evidence of excess ED visits and urgent care visits for combinations of
respiratory-related diseases during wildfire periods compared to a control period. Lastly, Liu et al.
(2017a) reported a positive association between wildfire PM2 5 and respiratory disease hospital admissions
when 2 consecutive days of wildfire PM25 concentrations (i.e., a smoke wave) were greater than
37 (ig/m3.
6-7
DRAFT: Do Not Cite or Quote
-------
Gan etaL
Gan etal
Gan etal (2017)c
Stowefl etal. (2019)d,e
TinHng etal. (2016)
Tmling etal- (2016)
Hutchinson et aL (2018)
Delfino et aL (2009)
Reid et aL (2016)
Reid et al. (2019)
Deflorio-Barker et aL (201?)f
Deflorio-Barker et aL (2019,
Deflorio-Barker et al. (2019,
Ainan etal. (2016)
Reid et aL (2016)
Reid et aL (2019)
Gan etaL (2017}a
Gan etaL (2017)b
Gan et aL (2017)c
Stowefl et aL (2019)d.e
Gan etal. (2020)j
Gan et3. (2020)1
Gan et al. (2020)
Tmling etal. (2016)
Tinling etal.
018)
in etal. (2020*
an et3. (2020)i
in et al. (2020}
al. (2016}
Hutchinson etal, (2(
Gan etal. (2020)k
Delfino et al. (2009)
Reid et aL (2016)
Reid et aL (2019)
Deflorio-Barker et aL (2019)f
Deflorio-Barker et al. (201912
Deflorio-Barker et aL (20l9)n
Aknan etal. (2016)
" " '201®
Location
Washington
Washington
Washington
Colorado
North Carolina (28 counties)
North Carolina (28 counties)
San Diego, CA
S. California
N. California (781 ZCTA)
N. California (753 zip codes)
595 U.S. counties
134 U.S. counties
134 U.S. counties
Colorado
N.California (781 ZCTA)
N. California (753 zip codes)
Washington
Washington
Washington
Colorado
Oregon
Oregon
Oregon
North Carolina (2S counties)
North Carolina (.28 counties)
San Diego, CA
Oregon
S. California
N. California (781 ZCTA)
N. California (7:>3 zip codes)
341 U.S. counties
92 U.S. counties
92 U.S. counties
Colorado
N. California (781 ZCTA) x
California (753 zip codes)
N. C:
Age
All
All
All
AH
<18
18-
0-64
a8
.AH
65+
65+
65-
$
All
tf
All
1
.AH
<18
18-
0-64
All
AS
65+
65t
65+
.All
ft
Lag
0
0
0
0-2
0-2DL
0-2DL
0-2 (72-h MA)
0-l
1-2
1-2
0
0
0
0-2
1-2
1-2
0
0
0
0-2
0
0
0
0-2DL
0-2 (72-hLMA)
0
0-1
1-2
1-2
0
0
0
0-2
1-2
1-2
AD Respiratory
1.00 1.05 1.10
Odds Ratio/Relative Risk
DL = distributed lag; CMAQ = Community Multiscale Air Quality; ED = emergency department; GWR = geographically weighted
ridge regression; MA = moving average; |jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; WRF = Weather Research and Forecasting; ZCTA = ZIP-code tabulation area.
Black circles = studies that used smoke/wildfire PM25 as the exposure indicator; red circles = studies that used ambient PM25
measurements as the exposure indicator; solid circles = hospital admissions; open circles = ED visits; odds ratios and relative risks,
unless otherwise noted, are for a 10 |jg/m3 increase in smoke/wildfire or ambient PM2g concentrations.
Exposure estimated using WRF-Chem smoke.
bexposure estimated from kriging.
"exposure estimated using GWR smoke PM25.
dEstimate is for a 1 |jg/m3 increase in wildfire PM25.
Combination of Hospital Admissions and ED Visits.
fPM2.5 Tot-CMAQ with indicator variable for smoke day.
gPM2.5 Tot-CMAQ-Monitor with indicator variable for smoke day.
hPM25 from monitors with indicator variable for smoke day.
'Outpatient hospital admission.
'Inpatient hospital admission,
kOffice visit.
Figure 6-1 Odds ratios and relative risks from U.S.-based epidemiologic
studies examining the relationship between short-term wildfire
smoke exposure and combinations of respiratory-related
diseases and asthma emergency department visits and hospital
admissions.
2 Several epidemiologic studies also examined associations between short-term wildfire smoke
3 exposure and other respiratory diseases, including COPD, acute bronchitis, pneumonia, upper respiratory
4 infections (URIs), and respiratory symptoms. Consistent with the studies that examined all respiratory
5 diseases and asthma ED visits and hospital admissions, these studies indicate an increased risk following
6 exposure for a range of respiratory effects (Figure 6-2). Examples include Rappold et al. (2011). which
7 reported positive associations for COPD. pneumonia and acute bronchitis, and URI in North Carolina
6-8
DRAFT; Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
counties exposed to wildfire smoke estimated by using smoke plume data, as well as Liu et al. (2017b)
which reported positive associations for hospital admissions related to COPD and respiratory infection in
adults 65 years of age an older exposed to two or more consecutive days to wildfire PM2 5 concentrations
>37 (ig/m3.
While the most extensive examination of the health effects attributed to wildfire smoke exposure
is based on exposure indicators that rely on PM2 5, populations can experience exposure to additional
pollutants, such as ozone as a result of the mixture of pollutants emitted from wildfires undergoing
atmospheric reactions in the presence of sunlight (U.S. EPA. 2019c). There is extensive evidence
indicating a relationship between short-term ozone exposure and respiratory effects including changes in
lung function and asthma-related ED visits and hospital admissions (U.S. EPA. 2020a). A recent study by
Reid et al. (2019) examined the relationship between ozone produced from wildfire events and respiratory
health. The authors reported the strongest evidence of an association with asthma and combinations of
respiratory-related ED visits during the fire, but the results across all of the respiratory outcomes
examined were attenuated in copollutant models with PM2 5 even though the correlation between ozone
and PM2 5 was low (r = 0.195), indicating the complexity in examining health effects attributed to both
primary pollutants and secondary pollutants from wildfire smoke.
6-9
DRAFT: Do Not Cite or Quote
-------
Study
Location
Age
Lag
Gan et al (2017)a
Washington
All
0
Ganetal. (2017)b
Washington
AD
0
Gan etal. (2017)c
Washington
All
0
Stowell et al. (2019)d,e
Colorado
ah
0-2
Tilling etal. (2016)
North Carolina (28 counties)
18+
0-2
Reid et al. (2016)
N. California (781 ZCTA)
All
1-2
Reid et aL (2019)
N. California (753 zip codes)
All
1-2
Delfino et al. (2009)
S. California
20-99
0-1
Reid et al. (2016)
N. California (781 ZCTA)
All
1-2
Reid et al. (2019)
N. California (753 zip codes)
All
1-2
Stowell etal (2019)4e
Colorado
All
0-2
Delfino et al. (2009)
S. California
All
0-1
Reid et al. (2019)
N. California (753 zip codes)
All
1-2
Gan et al. (2017)a
Washington
All
0
Ganetal. (2017)b
Washington
All
0
Gan etal. (2017)c
Washington
All
0
Delfino et aL (2009)
S. California
All
0-1
Reid et aL (2019)
N. California (753 zip codes)
.Al
1-2
Reid et al. (2019)
N. California (753 zip codes)
Al
1-2
Stowell etal. (2019)d,e
Colorado
All
0-2
Tinling etaL (2016)
North Carolina (28 counties)
<18
0-2DL
Tilling etal. (2016)
North Carolina (28 counties)
18-
0-2DL
Alman et al. (2016)
Colorado
All
0
Tilling etal. (2016)
North Carolina (28 counties)
<18
0-2DL
Tilling etal. (2016)
North Carolina (28 counties)
18+
0-2DL
COPD
Acute Bronchitis
Pneumonia
ITU
Respiratory Symptoms
0.85 0.95 1.05 1.15 125
Odds Ratio/Relative Risk
135
COPD = chronic obstructive pulmonary disease; DL = distributed lag; ED = emergency department; pg/rri3 = micrograms per cubic
meter; GWR = geographically weighted ridge regression; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 |jm; URI = upper respiratory infection; WRF = Weather Research and Forecasting; ZCTA = ZIP-code tabulation
area.
Black circles = studies that used smoke/wildfire PM2.5 as the exposure indicator; red circles = studies that used ambient PM2.5
measurements as the exposure indicator; solid circles = hospital admissions; open circles = ED visits; odds ratios and relative risks,
unless otherwise noted, are for a 10 |jg/m3 increase in smoke/wildfire or ambient PM2.5 concentrations.
Exposure estimated using WRF-Chem smoke.
bexposure estimated from kriging.
°exposure estimated using GWR smoke PM2.5.
dEstimate is for a 1 pg/m3 increase in wildfire PM2.5.
Combination of hospital admissions and ED visits.
Figure 6-2 Odds ratios and relative risks from U.S.-based epidemiologic
studies examining the relationship between short-term wildfire
smoke exposure and respiratory-related emergency department
visits and hospital admissions.
6-10
DRAFT; Do Not Cite or Quote
-------
6.2.2.2
Cardiovascular Effects
1 There is extensive experimental and epidemiologic evidence indicating a relationship between
2 short-term PM2 5 exposure and cardiovascular effects, particularly for ischemic heart disease (IHD) and
3 heart failure (HF) as well as cardiovascular mortality (U.S. EPA. 2019b). While there is a more limited
4 evidence base related to the effects of wildfire smoke exposure on cardiovascular health, compared to
5 respiratory outcomes, these studies report generally positive associations albeit with wide confidence
6 intervals (Figure 6-3), with the magnitude of associations being relatively consistent to those reported in
7 studies of ambient PM2 5 (U.S. EPA. 2019b).
6-11
DRAFT: Do Not Cite or Quote
-------
Author
Stowell etal «2C19nb
Tailing etal i201o>
Deflone Barker etal 0019>e
Deflorte -Baikeretal i?019id
Deflone -Barker etal O01?te
Delfine et at ^2009)
Sro'Aell etal !2019fab
Delfino etal OO""*}
Almati etal «2«"ilct
Tmhig etal. 5,201 Dj
ReidctaL(20Itf)
Tmkig etal C'lO1
Stem ell etal OClQ;ab
Imlrng etal i2>"lo;
Stewell etal i2«ilQ»a>»
Location Age Lag
C oloi ado 65- 0-1
North C3i ehna ;ZS coimties) 18- 0-2DL
C''7 US coimties <55- 1
157 t" S counties 65- 1
157 US cornmes 65~ I
S California 45- 0-1
Colorado 65+ 0-1
S California 45^ 0-1
Colei ado Al 0-2
North Caiolma (25 counties) 18- 0-2DL
N. California (781 ZCTA) Al 1-2
North Caiobna <2$ counties) IS- 0-2DL
C clot ado Al 0-1
North C'aioana OS counties) 18- 0-2DL
Colorado Al 0-1
AH Cardiovascular
mn
Heart Failure
Hypeitensioa
Dysrhythmia
AMI
0.85
0.95 1.05 1.15 1.25
Odds Ratio/Relative Risk
AMI = acute myocardial infarction; DL = distributed lag; ED = emergency department; |jg/m3 = micrograms per cubic meter;
GWR = geographicaly weighted ridge regression; IHD = ischemic heart disease; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; WRF = Weather Research and Forecasting; ZCTA = ZIP-code tabulation area.
Black circles = studies that used smoke/wildfire PM25 as the exposure indicator; red circles = studies that used ambient PM25
measurements as the exposure indicator; solid circles = hospital admissions; open circles = ED visits; odds ratios and relative risks,
unless otherwise noted, are for a 10 |jg/m3 increase in smoke/wildfire or ambient PM2.5 concentrations.
aEstimate is for a 1 |jg/m3 increase in wildfire PM2.5.
bCombination of hospital admissions and ED visits.
°exposure estimated using WRF-Chem smoke.
dexposure estimated from kriging.
Exposure estimated using GWR smoke.
Figure 6-3 Odds ratios and relative risks from U.S.-based Epidemiologic
studies examining the relationship between short-term wildfire
smoke exposure and cardiovascular-related emergency
department (ED) visits and hospital admissions.
1 Several studies examining cardiovascular effects used indicators of smoke events to capture the
2 spatial and temporal extent of exposure CWettstein et al.. 2018; Liu et al.. 2017a; Rappold et al.. 2011). In
3 a study of 561 western U.S. counties, Liu et al. (2017a) did not report any evidence of an association
4 between total cardiovascular-related hospital admissions and smoke wave days (i.e., two consecutive days
5 with wildfire PM2 5 concentrations >20 |_ig/m3) in adults 65 years of age and older. However, in a study of
6 ED visits within 42 North Carolina counties, Rappold et al. (2011) reported an increased risk for
7 combined cardiovascular-related outcomes. When examining, cause-specific cardiovascular outcomes,
8 the authors reported the strongest evidence of an association for heart failure and myocardial infarction.
9 Similarly, in a study of eight California air basins Wettstein et al. (2018) reported an increased risk of ED
6-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
visits across combined cardiovascular outcomes at medium (PM2 5 concentrations between 10-19 (ig/m3)
and dense (PM2 5 concentrations >20 (ig/m3) smoke density. The authors observed positive associations in
all adults, but associations were larger in magnitude among individuals 65 years of age and older.
Additionally, Wettstein et al. (2018) reported a positive association with incidence of stroke among those
65 years and older following smoke exposure. The results of Rappold et al. (2011) and Wettstein et al.
(2018) indicate a need for additional exploration of: the effect of wildfire smoke exposure on
cardiovascular outcomes in older individuals; cause-specific cardiovascular outcomes; and the most
appropriate exposure indicator to represent wildfire smoke exposure when focusing on cardiovascular
outcomes.
In addition to the outcomes examined through studies of ED visits and hospital admissions, a
recent study by Jones et al. (2020) examined out-of-hospital cardiac arrests (OHCAs) attended by
emergency medical services (EMSs). The study was conducted across 14 California counties where daily
exposures were classified as light, medium, or high smoke density based on PM2 5 estimated from the
NOAA Hazard Mapping System. The authors reported positive associations with OHCA at multiple
single day lags on heavy smoke density days (i.e., estimated smoke PM2 5 concentrations >22 (ig/m3) with
the strongest evidence at lag 2 (odds ratio [OR]: 1.70 [95% confidence interval (CI): 1.18, 2.45]). There
was no evidence of associations when examining light or medium smoke density days.
6.2.2.3 Mortality
Across the epidemiologic studies conducted that examine the relationship between short-term
wildfire smoke exposure and health effects, to date, only a few U.S.-based studies examine mortality.
Although the evidence base for wildfire smoke exposure and mortality from studies conducted in the U.S.
is limited to a few studies, there is extensive evidence indicating a relationship between short-term PM2 5
exposure and mortality spanning both single and multicity studies conducted in diverse geographic
locations, populations with different demographic characteristics, and studies employing different
exposure assessment methodologies (U.S. EPA. 2019b).
Doubledav et al. (2020) conducted the most comprehensive assessment of mortality, in a study
conducted in Washington state that spanned multiple fire seasons and cause-specific mortality outcomes.
Using an exposure indicator that was based on defining smoke days versus nonsmoke days, in a
case-crossover analysis the authors reported evidence of a positive association for both respiratory disease
(OR: 1.09 [95% CI: 1.00, 1.18], lag 0) and COPD mortality (OR: 1.14 [95% CI: 1.02, 1.26]; lag 0), but no
evidence of an association for other mortality outcomes including total (nonaccidental), cardiovascular,
and IHD. Unlike Doubledav et al. (2020). which focused on wildfire events and mortality in the same
state, Zu et al. (2016) provided evidence to support the results of Doubledav et al. (2020) in a study that
examined the relationship between wildfire smoke exposure and mortality in a study of hemodialysis
patients with end-stage kidney disease (ESKD) that resided in U.S. counties near a major wildfire during
6-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
the years 2008 through 2012. The authors reported a positive association with all-cause mortality for a 10
(ig/m3 increase in wildfire specific PM2 5 that was similar in magnitude at both lag 0 (relative risk
[RR] = 1.04 [95% CI: 1.01, 1.07]) and for a distributed lag of 0-1 days(RR= 1.05 [95% CI: 1.01, 1.08]),
with limited evidence of an association for the other mortality outcomes examined.
6.2.2.4 Uncertainties and Limitations in the Health Effects Evidence
The current state of the science with respect to the health effects of wildland fire smoke exposure
stems from the large evidence base demonstrating a range of health effects, including respiratory and
cardiovascular effects and mortality, in response to short- and long-term PM2 5 exposure. Studies of
wildfire smoke exposure report results that are generally consistent with this larger evidence base, but
uncertainties remain in gaining a fuller understanding of the health effects of wildland fire smoke.
Although this section focuses exclusively on U.S.-based epidemiologic studies, it is important to
recognize it only represents a fraction of the studies conducted globally, but overall, the results of the
U.S.-based studies are consistent with this broader body of evidence.
As noted in Section 6.2.1. there is variability in both the exposure assessment approach and
exposure indicator employed across studies, which can complicate the interpretation of results across
studies. However, even with this variability results are generally consistent across studies and health
effects, specifically when examining short-term (i.e., daily) smoke exposure. While there is a general
understanding of the health effects attributed to short-term smoke exposure, to date, there has been
limited investigation and evidence for other exposure durations, including subdaily (i.e., <24-hour
average), repeated high exposures over many days, and exposures over multiple fire seasons or years.
Additional research focusing on other exposure durations can aid in informing land management
strategies, such as prescribed fire; the potential health implications of smoke exposure from single
wildfire events, as well as fires over multiple years; and further enhance public health messaging
campaigns. Lastly, although current evidence does not indicate a difference in the health effects between
ambient PM2 5 exposure and other source-based exposures, such as wildfire smoke (U.S. EPA. 2019b). as
wildfires continue to encroach upon the wildland-urban interface (WUI) the complex smoke mixture
could change as structures and cars are burned, potentially resulting in different health risks.
6.2.3 Summary
Decades of research on the health effects attributed to exposure to ambient air pollution,
specifically PM2 5 and ozone, provide a strong evidence base for the health effects that could be observed
due to short-term (i.e., daily) and long-term (i.e., months to years) exposure to wildland fire smoke.
U.S.-based epidemiologic studies, which represent a fraction of the studies conducted globally, examining
the health effects attributed to short-term wildfire smoke exposure have extensively examined and
6-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
consistently report evidence of associations with respiratory-related health effects, including respiratory
and asthma ED visits and hospital admissions, regardless of the exposure indicator used
(e.g., wildfire-specific PM25, smoke density, etc.). Recent studies examining short-term wildfire smoke
exposure also provide growing evidence of cardiovascular effects, with more limited evidence for
mortality. Overall, there are few studies that examined the health effects associated with exposure to
smoke from prescribed fires, therefore, it remains unclear whether there are differential health effects
from smoke from prescribed fires compared to wildfires.
Studies of wildfire smoke have not examined the health implications of long-term exposure, such
as from wildfires that last multiple months (e.g., the Rough Fire) or over multiple fire seasons, but
evidence from studies of long-term PM2 5 exposure indicate these types of wildfire events could also
result in mortality impacts. Additionally, although there is limited evidence of health effects attributed
specifically to ozone produced from wildfire smoke, there is extensive evidence demonstrating health
effects from exposure to ambient ozone exposure, indicating potential additional public health impacts
from wildfire smoke.
6.3 Mitigation of Prescribed Fire and Wildfire Smoke Exposure
to Reduce Public Health Impacts
Characterizing exposure to wildfire smoke is instrumental in examining health effects, and
epidemiologic studies have typically used levels of smoke or the concentration of PM2 5 in outdoor
ambient air as the exposure estimate (Section 6.2). In addition to these studies focusing on relationships
between wildfire smoke exposure and health outcomes, several studies have examined individual and
community actions that can be taken to reduce or mitigate exposure to smoke during wildfire events. For
example, people spend most of their time indoors at home, work or school (Klepeis et al.. 2001). where
smoke exposure can be reduced relative to outdoors depending on factors such as building ventilation and
use of air filtration (U.S. EPA. 2020c'). This section describes a framework for, and the type of data
needed to quantify the potential public health benefits of actions that reduce or mitigate smoke exposure.
These actions, also often referred to as interventions, consist of some form of individual behavioral
change, such as staying indoors with windows and doors closed or reducing activity levels; the use of
personal protective measures, such as a respirator; using a portable air cleaner indoors or the extended use
of a heating, ventilation, and air conditioning (HVAC) system equipped with a high efficiency particle
filter; or community-level interventions (e.g., providing clean air spaces). While each of these actions can
reduce wildfire smoke exposure for an individual by some percent, the overall fraction of the population
taking preventative measures depends on many factors, such as population demographics, access to
interventions, whether smoke is visible or can be smelled, and perceived risk of smoke exposure, all of
which may also be impacted by public health messaging campaigns. Of these factors, the timing, content,
and extent of public health messaging campaigns may represent a major difference in how prescribed fire
6-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
and wildfire events are managed. However, whether there are differences in the percent of the population
taking actions between the fire types has not been assessed.
The following sections provide an overview of a framework that captures the factors that need to
be accounted for in order to estimate the potential reduction in overall smoke exposure for a population
during both prescribed fire and wildfire events. Additionally, it evaluates and summarizes results from
studies that provide data on how often people take action during smoke events and the exposure reduction
that occurs from those actions (see Appendix A.6.2 for details on study inclusion criteria). The
information presented within this section will be used to provide a crude estimate of the potential
reduction in health impacts in the case study areas that could be achieved due to specific actions or
interventions to reduce smoke exposure (see Section 8.3.3). However, it is important to recognize that the
estimation of the potential reduction in public health impacts attributed to smoke exposure within this
assessment is not meant to reflect a formal analysis of post-fire effectiveness of public health messaging
by Air Resource Advisors (ARAs) deployed by the U.S. Forest Service, in combination with respective
state and local air quality agencies, for either the TC6 or Rough fires.
6.3.1 Framework for Estimating the Impact of Actions to Reduce
Smoke Exposure
Estimating potential reductions in wildfire or prescribed fire smoke exposure that a population
could experience as a result of actions, or interventions, is based on a series of events and assumptions
(Figure 6-4). The overall exposure reduction for a population will be determined by the likelihood and/or
ability to take a particular action combined with the exposure reduction effectiveness of the action. There
are multiple factors that influence both of these elements, but one major driver for any action is the
awareness of the need to take action.
6-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Drivers for Actions
Public Health
Messages with
Action/Intervention
Information
Public Awareness of
Wildfire Smoke
Demographic Factors
(e.g., Age, Pre-existing
Heart or Lung Disease)
/
Va of Population
that Takes
Actions/
Interventions
\
Housing Characteristics
(e.g., Age of Housing Stock,
HVAC Prevalence)
Effectiveness of
Actions/
Interventions
Taken
/
Access or availability
, Purchase/own portable air cleaners,
higher MERV filters for HVAC)
HVAC = heating, ventilation, and air conditioning; MERV = minimum efficiency reporting value; PM25 = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 2.5 |jm.
Figure 6-4 Considerations for estimating potential reduction in wildfire
smoke exposure due to actions and interventions.
Information dissemination, specifically focusing on the potential risks of wildfire smoke exposure
and actions a population can take is the initial step that can ultimately dictate whether individuals take
actions to reduce exposure. However, public health messaging on its own is not enough if the proper
information is not conveyed. The limited assessment of public health messaging campaigns has shown
that only 14-46% of wildfire smoke related messages disseminated by government and media entities
indicate the individual and administrative actions that can be taken to reduce smoke exposure (Van
Deventer et al.. In Press). Additionally, people may take actions to reduce exposure regardless of public
health messaging campaigns as a result of general awareness of the presence of wildfire smoke (Kolbe
and Gilchrist. 2009; Kiinzli et al.. 2003). Both public health messaging and general awareness of smoke
factor into the percent of the population that takes an action or institutes an intervention to reduce
exposure.
Whether or not people take actions to reduce smoke exposure can depend on their knowledge of
the potential impact of environmental exposures on their health (Rappold et al.. 2019) as well as their
personal experiences with smoke, perceptions of risk and level of self-efficacy (Hano et al.. 2020). This is
often a reflection of the age or underlying health status of an individual or family member. While not
directly factored into an estimation of potential exposure reductions due to various actions, it is important
to acknowledge that the population-level response to taking actions could vary within the population
based on socio-demographic factors. Additionally, the ability to take actions depends on the accessibility
and availability of interventions, such as portable air cleaners and high efficiency HVAC filters. Access
6-17 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
and availability may depend on having the financial means to purchase interventions, but also whether
programs for providing interventions exist within an area. Even low-cost interventions can have barriers
to their use, such as staying indoors with doors and windows closed without air conditioning when smoke
and high temperatures co-occur.
There are numerous actions people can take to reduce exposure to smoke, with a large degree of
variability in the efficacy of each (Xu et al.. 2020; Laumbach. 2019). The primary focus for several
actions is reducing indoor PM2 5 concentrations while at home where people spend most of their time.
Housing characteristics, such as age of the home and presence and type of HVAC system, influence the
infiltration of particles indoors under normal conditions, and also influence the efficacy and availability of
these actions for reducing smoke exposure in homes (Joseph et al.. 2020; U.S. EPA. 2020c; Davison et
al.. In Press). Therefore, housing characteristics of the geographic area impacted by smoke is another
important factor, and if variability in the housing stock is not accounted for in some way then estimates of
exposure reduction could be under- or overestimated.
Embedded within the potential actions and interventions a population within a defined geographic
area may take to reduce smoke exposure are a series of operating conditions that can directly influence
the overall percent reduction in smoke, specifically PM2 5. These conditions, such as how often a
building's HVAC system runs, and whether a high-efficiency filter and/or portable air cleaner is used, are
important to consider when constructing potential exposure reduction scenarios. Similar to housing
characteristics, the distribution of these operating conditions can vary throughout the population being
examined, may depend upon public awareness of the presence of smoke, and can contribute to over- or
underestimation of the overall exposure reduction of a particular action.
At each step of the process of developing scenarios to estimate the influence of actions to reduce
smoke exposure there are decision points that rely on both data from published studies and assumptions
regarding the population being examined. Outlining these decision points will allow for a clear
articulation of the factors that influence each exposure reduction scenario and the ability to construct
scenarios meant to represent the range of exposure reductions that could be experienced.
6.3.2 Individual and Community Actions to Reduce Smoke Exposure
In identifying the overall percent reduction in smoke exposure that can be achieved in response to
public health information dissemination, the key factors to consider are the actions that can be taken at
both the individual and community level and the effectiveness of those actions in reducing exposure,
particularly to PM2 5. Recent publications by Xu et al. (2020) and Laumbach (2019) provide overviews of
the actions that individuals can take to reduce smoke exposure, which are delineated into four broad
categories according to the hierarchy of controls traditionally used for occupational hazards (NIQSH.
2015): elimination, engineering controls, administrative controls, and personal protective equipment. As
depicted in Figure 6-5, there are a range of smoke exposure reductions that can be achieved depending on
6-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
the approach instituted, but for each there are limitations and concerns that should be considered. This
section characterizes the broader body of studies that examined the effectiveness of information
dissemination and various exposure reduction actions, which collectively provides evidence that supports
the range of smoke exposure reductions that could be achieved if individuals are well informed and take
the necessary steps to reduce/mitigate exposure.
Most
effective
Elimination
Reduces exposure by 100%
Engineering controls
Reduce exposure by 20 to 90%,
depending on quality of filters
or air cleaners
Personal Actions
Relocation
Close doors and windows
Set air conditioners in recirculation mode
Use portable air cleaners with HEPA filter
or central air conditioner with filters
Administrative controls
Reduce exposure by approximately 50%
Personal protective equipment
Reduces exposure by >90% if well fitted
but nearly 0% if poorly fitted
Stay indoors
Avoid heavy or prolonged
physical activity
Limitations or Concerns
Relocation increases costs and stress and has
unpredictable duration.
Wildfire particulate matter and ozone may extend
thousands of kilometers.
Relocation may not be feasible.
Effectiveness varies greatly with ventilation and filtration rates.
Most filters reduce only particulate matter and not gaseous
pollutants (e.g., ozone).
Cost is prohibitive for some.
Strategies are less effective for "leaky" houses.
Exposure to indoor air pollution (e.g., cooking smoke and aldehydes
from paints and furnishings) is increased.
Insufficient physical activity may lead to adverse health effects.
Strategies are impractical for outdoor workers.
Least
effective
Only certain face masks (e.g., N95 or P100) can reduce exposure to particulate matter.
Effectiveness depends on fit, and fit testing is not generally available.
Masks cannot protect against gaseous pollutants.
Masks may provide a false sense of security and thus increase outdoor time
and actual exposure.
Masks may cause physical stress due to increased work of breathing, heat,
and discomfort.
Masks are not suitable for children, people with facial hair, and those with lung
or heart diseases.
Cost is prohibitive for some.
HEPA = high-efficiency particulate air.
Source: Xu et aL (2020). copyright permission pending.
Figure 6-5 Summary of individuai-levei wildfire smoke exposure reduction
actions and effectiveness.
6.3.2.1 Factors That Influence Taking Actions to Reduce Smoke Exposure
Several studies examined how awareness of smoke, whether by direct observation or through
public service announcements (PSAs), can translate into a population taking exposure reduction actions.
Most of the information available on the effectiveness of PSAs stems from studies conducted in
California or in Australia where wildfires impact large population centers and occur on a near yearly
basis. Of the available studies, all were conducted in the context of wildfire with no information currently
available on the likelihood of actions taken in response to prescribed fire smoke. Studies on prescribed
fire have focused on the factors governing the tolerance of smoke and optimal risk communication (01 sen
et aL 2017; Blades et aL 2014). rather than exposure reduction actions taken in response to smoke.
6-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Studies of exposure reduction actions are often conducted through retrospective surveys of
communities impacted by major wildfires to determine population awareness of smoke, PSAs or other
health risk communications, and the resulting actions as a function of the messaging medium (Kolbc and
Gilchrist. 2009; Mott et al.. 2002). content (Sugerman et al.. 2012). and the characteristics of the
community (Kolbc and Gilchrist. 2009). These studies have investigated the impact of population
demographics (e.g., age, gender, income level, etc.), pre-existing conditions, and experiencing symptoms
on the type and extent of exposure reduction action taken.
Across studies that examined PSAs, in most communities the awareness of PSAs was high
(74-88% of those surveyed recalled a PSA) with many people (43-98%, Table A.6-2) taking some
exposure reduction action in response to a PSA, but the most effective method of communication varied
by community. Television was the most effective communication medium in studies conducted in San
Diego [77%; Sugerman et al. (2012)1 and Australia [68%; Kolbe and Gilchrist (2009)1 while radio was
the most effective medium in a rural tribal community in northern California (Mott et al.. 2002).
However, in this community a wide variety of information sources (e.g., the medical establishment,
friends and family, and the workplace) were recalled in greater frequency than television, demonstrating
the impact of the community type on the most effective method of risk communication.
When considering the implications of the demographic composition of a population, older adults
were less likely to be aware of PSAs with only 58% of those over 75 years of age aware of the PSA
compared to 74% for the entire population (Kolbe and Gilchrist. 2009). Those with pre-existing
conditions (81%) were also found to be slightly less likely than those without a pre-existing condition
(85%) to be aware of PSAs (Mott et al.. 2002). While it is important to be aware of the message, message
comprehension is also extremely important when considering whether individuals take the necessary
actions to protect themselves. Sugerman et al. (2012) observed that message comprehension was reduced
in those that did not speak the primary language or when the message was too technical in nature
(e.g., stay inside vs. run HVAC system more often).
Overall, across studies it was found that most people aware of a PSA took some action to reduce
exposure [66-98%; Sugerman et al. (2012); Kolbe and Gilchrist (2009); Mott et al. (2002)1. The
awareness of smoke also prompted people to act. For example, in Australia of the 76% of the population
that took an exposure reduction action 43% did so due to the PSA and 28% due to the presence of smoke
(Kolbe and Gilchrist. 2009). Furthermore, the percentage of people reducing outdoor activities, closing
doors and windows, and evacuating was similar between those responding to the smoke and those
responding to a PSA (Kolbe and Gilchrist. 2009). The more technical actions were much more likely to
be used by those aware of the PSA than those that were not, like using a mask (8.1% aware of PSA vs.
1.3% not aware) or using ceiling fans [10.5% aware of PSA vs. 2.9% not aware; Kolbe and Gilchrist
(2009)1.
The most commonly used actions were those that are easiest to carry out, including reducing or
avoiding outside activity, and staying inside or closing windows and doors (Figure 6-6, Table A.6-2). On
6-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
average for the total surveyed population, the least likely actions were using an air cleaner (10-34%) or
respirator [7-14%;. Sugerman et al. (2012)1 found that more technical actions (e.g., use home air
conditioning, use HEPA air filtration, wear N95 mask during ash clean up) were least likely to be done in
part due to a poor recall of the PSA and a difficulty understanding the PSA. Accessibility to measures that
can reduce exposure, such as an HVAC system, air cleaner or respirators/masks, while not formally
characterized in any of the studies evaluated, may significantly impact the probability of an individual
taking an exposure reduction action.
As depicted in Figure 6-6, there is a wide distribution in the percentage of each population taking
action to reduce or mitigate smoke exposure, which is in large part due to the different populations
surveyed. People actively experiencing symptoms due to wildfire smoke were much more likely to take
an action than the general population. This is most striking for actions that require equipment, like air
cleaners or respirators. For example, Rappold et al. (2019) reported that 86% of people with four or more
symptoms used an air cleaner versus 24% of the average population (averaged across all studies).
Most studies provided some indication of the smoke concentration and duration in the community
[e.g., from Mott et al. (2002)1 which reported PMio (particulate matter with a nominal aerodynamic
diameter less than or equal to 10 ^m ): 2 days PM2 5 >425 |ig/m3 and 15 days >128 |ig/nr\ assuming PM2 5
is 85% of PM10 concentrations as detailed in Lutes (2014). However, inconsistent reporting prevents the
determination of a clear association between smoke exposure (duration or peak concentration) and the
probability of taking a particular action. The level and duration of smoke exposure are likely major
determinants in what actions a community will take and important factors to be considered in future
studies.
6-21
DRAFT: Do Not Cite or Quote
-------
100
£
_o
o
c
bO
a
H
^ y s .~
&
,oN
•V
& *
$ ^
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
of nonresidential building types were limited to a few studies focusing on office buildings, with no other
building types (e.g., schools) examined. Studies that examined the effectiveness of masks for reducing
exposure to particles in air have primarily been conducted for occupational exposure and other purposes
(Allen and Barn. 2020). and not specifically for examining their effect in reducing smoke exposure within
the general population.
Reviews by Xu et al. (2020) and Laumbach (2019) compare percent reductions for various
actions that could be taken to reduce or mitigate smoke exposure. Elimination of smoke exposure can be
achieved by relocation (exposure reduction = 100%), while engineering controls such as closing windows
and doors or indoor air filtration can also be effective (20-80% exposure reduction), as are administrative
controls such as staying indoors and avoiding outdoor activities (-50% exposure reduction). Additionally,
both Xu et al. (2020) and Laumbach (2019) noted that wearing N95 or PI00 masks can be 90% effective
or more, but only if properly fitted along with other limitations (e.g., not suitable for children). The results
reported in Xu et al. (2020) and Laumbach (2019) are generally consistent with the levels of effectiveness
for the different actions reported in recent studies.
U.S. EPA (2018) reviewed residential measurement studies that used portable air cleaners and
central HVAC system filters to reduce indoor PM2 5 exposures overall, not PM2 5 specific to wildfire
smoke. Portable air cleaners were found to substantially reduce indoor concentrations of PM of both
indoor and outdoor origin, often reducing indoor PM2 5 concentrations by around 50% on average.
Residential measurement studies that examined portable air cleaner effectiveness in homes during
wildfire smoke events (Barn et al.. 2008; Henderson et al.. 2005) also reported a similar percent reduction
in indoor PM2 5 concentrations with the elevated outdoor PM2 5 concentrations during these events. Barn
et al. (2016) also reviewed many of the same studies as U.S. EPA (2018) and concluded that portable air
cleaners can reduce indoor PM2 5 concentrations by 32-88% and recommended their use during fire
events.
U.S. EPA (2018) also noted a few residential measurement studies that showed higher efficiency
central HVAC system filters such as those rated minimum efficiency reporting value (MERV) 13 or
above can reduce indoor PM2 5 concentrations. Singer etal. (2017) reported a 90% reduction in PM2 5
using HVAC filtration with high efficiency MERV filters in a single test house in California during
typical ambient PM2 5 concentrations, which was comparable to running a portable air cleaner in the
home. However, results from a recent study by Alavv and Siegel (2020) showed actual in-home
effectiveness of HVAC filtration for PM2 5 was much lower (average -40%) and varied widely across
homes even for filters with the same MERV rating depending on the home. Filter performance was
strongly linked to home- and system-specific parameters including ventilation rate and system runtime.
Of the studies evaluated, Reisen et al. (2019) is the only available residential measurement study
that examined the effectiveness of closing windows and doors during a smoke event. However, the study
only included four homes in Australia that experienced smoke due to a prescribed fire. Simple infiltration
6-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
modeling of the measurements showed that remaining indoors with windows and doors closed reduced
exposure to peak PM2 5 concentrations by 29 to 76% across the homes and that a tighter house, in terms of
reduced ventilation, provided greater protection against particle infiltration.
A comprehensive residential modeling study by Fisk and Chan (2017b) compared central HVAC
system filtration and portable air cleaners for six different home type scenarios during a wildfire smoke
event in California. The combined effect of continuous HVAC fan use with a high efficiency (MERV 12)
filter and continuous portable air cleaner was most effective (62% reduction in PM2 5), while continuous
portable air cleaner use in homes without forced-air HVAC systems provided 45% reduction in PM2 5
concentrations.
While most of the studies conducted focus on examining the effectiveness of interventions in
residential locations, a few studies examined the effectiveness of HVAC systems and filters in office
buildings during wildfire smoke events. Stauffer et al. (2020) compared offices with and without portable
air cleaners during a wildfire season. They reported 73 and 92% reduction in PM2 5 concentrations indoors
with portable air cleaner use for daytime and nighttime, respectively. Pantelic et al. (2019) reported a 60%
reduction in PM2 5 for a mechanically ventilated office building and higher efficiency filters compared to
a naturally ventilated building during a wildfire.
Fisk and Chan (2017a) conducted a modeling study comparing improved filtration using filters in
residential forced-air systems and/or portable air cleaners for homes and higher efficiency filters in
commercial buildings in three U.S. cities (Los Angeles, Houston, Elizabeth, NJ) for ambient PM2 5
concentrations. Additional higher efficiency filtration in other buildings only slightly reduced overall
PM2 5 exposures due to the amount of time spent in these locations compared to at home.
In summary, although limited in number, studies that examined the effectiveness of actions or
interventions to reduce PM2 5 exposure provide relevant data for considering the potential implications of
public health messaging campaigns and the most effective actions to recommend to the public to reduce
exposure to wildfire smoke (see Table A.6-3). Portable air cleaners were shown to reduce residential
indoor PM2 5 concentrations from -40-90%, depending on the study and home characteristics. Increasing
filtration efficiency in residential forced-air systems and/or running the system more/continuously can
also reduce indoor PM2 5 concentrations by a similar percent, but data from these studies were more
variable between homes and efficiency of the filters. The data also suggest office buildings with high-
efficiency filters in HVAC systems or that use portable air cleaners can achieve a similar reduction in
indoor PM2 5 concentrations (-60-90%) as homes. Lastly, there is limited data to fully assess the
effectiveness of only closing windows and doors and staying inside as a means to reduce wildfire smoke
exposure.
6-24
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
6.3.3 Estimating the Overall Exposure Reduction to Wildfire Smoke for
Individual-Level Actions
Although the available data on individual and community actions that can be taken to reduce
smoke exposure is currently limited, the data detailed within this section provides information on many of
the factors to consider that are depicted in Figure 6-4 for estimating the potential impact of
actions/interventions on reducing PM2 5 exposure from wildfire smoke. An approximation of the overall
percent reduction in PM2 5 exposure for a population that could be achieved by individual-level actions
can be estimated by combining the data on the likelihood of taking actions in response to smoke with the
effectiveness of the various actions (Table 6-1). However, it is important to recognize that across the
studies evaluated there was a wide range of data on both the likelihood and effectiveness of exposure
reduction actions (see, Table A.6-2 and Table A.6-3). Therefore, the values reported in Table 6-1
represent the average with standard deviation across studies for likelihood and effectiveness of the
different actions and interventions.
Table 6-1 Summary of data available for various exposure reduction actions.
Exposure Reduction
Action
Likelihood of Taking
Action in Response
to Wildfire3
Mean ± SD
Effectiveness of Action13
Mean ± SD
Average Overall Exposure
Reduction0
Reduced activity
70.3% ± 15.5
No data
-
Stayed inside
64.0% ± 12.9
49.8% ± 22.8
31.8%
Ran home HVAC system
38.0% ± 31.1d
64.0% ± 32.8
24%
Evacuated
24.4% ± 18.7
100%
24%
Used air cleaner
23.8% ± 10.7
63.7% ±21.0
15%
Used respirator
10.3% ± 3.5
No datae
-
HVAC = heating, ventilation, and air conditioning; SD = standard deviation.
aFrom studies in Table A.6-2 for respondents regardless of health history or status.
bFrom studies in Table A.6-3.
°Average likelihood of taking the action multiplied by the average effectiveness of the action.
dMay include the use of other air conditioning systems in addition to HVAC systems.
eNo data available on the effectiveness of respirators for reducing wildfire smoke exposure.
For each exposure reduction action, the average overall percent exposure reduction, at the
population-level, was calculated by multiplying the average likelihood of taking the action by the average
effectiveness of the action. Although simplistic, this approach provides an initial comparison that shows
6-25
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
the more effective actions are generally less likely to be used, resulting in a lower overall exposure
reduction (Figure 6-7). For example, the data on portable air cleaner use showed an average -64%
reduction in PM2.5 but the likelihood of using them was -24% on average, resulting in an overall exposure
reduction of -15%. The exposure reduction action with the highest average overall percent reduction was
staying inside (-32%), due to the greater likelihood of people taking this action (-64% on average) and its
relative effectiveness (-50% on average). It should be noted that combining these two types of study data
assumes a reasonable match between the interventions reported in the survey studies of PSA effectiveness
and those evaluated in the effectiveness studies, which may not be appropriate in all cases.
100
80
c
0)
u
L_
0)
CL
60
40
20
Stayed inside
Ran home HVAC
system
I Likelihood of taking
action
I Effectiveness of
action
I Overall exposure
reduction
Evacuated Used air cleaner
HVAC = heating, ventilation, and air conditioning; PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or
equal to 2.5 |jm.
Figure 6-7 Comparison of estimated percent overall PM2.5 exposure
reduction by action.
6.3.4 Uncertainties and Limitations in Estimating Exposure Reduction
to Wildland Fire Smoke
While it is clear from Figure 6-7 that there are actions that can be taken at the individual level that
could substantially reduce overall population exposure to wildfire smoke, there are multiple assumptions
and limitations that should be considered in the process of using this information to estimate the potential
public health benefit of messaging campaigns. The studies conducted to date examining actions and
interventions to reduce wildfire smoke exposure, specifically PM2.5, have been conducted over a limited
6-26
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
geographic scale, and, as such, may not be transferable across locations. However, the limited
geographic scale of available studies could be accounted for by including location specific information in
an analysis, such as detailed information on the housing stock (e.g., age, type of HVAC, etc.), population
demographics, and community characteristics (e.g., urban vs. rural). Additionally, the average overall
exposure reductions presented do not account for the likelihood that taking actions may differ
significantly between wildfire and prescribed fire smoke events, due to potential differences in public
health messaging campaigns for each fire type (e.g., PSAs in preparation for prescribed fires are not
uniform across locations). These potential differences between wildfire and prescribed fire may also
include differences in the effectiveness of an action or intervention due to variability in PM2 5
concentrations. Specifically, the effectiveness may be reduced at the very high PM2 5 concentrations
associated with large wildfire events.
Perhaps the greatest difference in potential smoke exposure reductions can be attributed to the
different level of public awareness of smoke for the two different types of fires. Smoke from prescribed
fires may be present for a short duration, as little as several hours, and at lower concentrations that may
not be noticeable. Alternatively, wildfires may lead to prolonged high smoke concentrations with
noticeable odor and visibility impacts. Wildfires are often reported on by the local news media, which
may include public service announcements about actions to reduce smoke exposure. Additionally, most
major wildfire incidents have an ARA that develops and disseminates information on smoke forecasts, air
quality, and messaging to address public health concerns. The ARA generates daily smoke reports that are
posted online on InciWeb rhttps://inciweb.nwcg.gov; NIFC (2021)1. state smoke blogs, and on fire
information boards through impacted communities. Prescribed fires are not as widely publicized and
depending on the state or local regulations may be conducted without any notifications or alerts to the
surrounding community. Therefore, public awareness of prescribed fires may be very limited, greatly
reducing the potential for exposure mitigation actions to be taken.
Another difference is that wildfires and prescribed fires often occur at different times of the year
when residential ventilation rates may vary. In the study areas and many parts of the western U.S.,
wildfires largely occur during July through October (Jaffe et al.. 2020; Ryan et al.. 2013). Prescribed fires
are often done in the late fall or early spring during cooler weather (see Section 3.2.2.1). while pile burns
of mechanically thinned biomass are typically done in the winter months. These different seasons of the
year for the fire types may have ambient conditions that lead to different behaviors with respect to home
ventilation (Marr et al.. 2012; Yamamoto et al.. 2010). In areas where residential air conditioning systems
are not prevalent, wildfires may frequently coincide with time periods when ventilation rates may be
highest as windows and doors would be opened to cool the indoor environment. In contrast, in areas
where air conditioning systems are prevalent, prescribed fires may coincide with time periods when
ventilation rates may be greater due to window and door opening during the more temperate months.
In addition to the differences in smoke exposure between fire types noted above, there are also
data gaps that complicate the ability to quantitatively estimate the overall exposure reduction that could
6-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
be achieved. Within this assessment, a crude approach is taken to estimate the potential public health
impact of different actions and interventions to reduce smoke exposure, but it does not account for the
fact that in reality a combination of these actions or interventions will be employed across the population
(see Section 8.3.3). As depicted in Figure 6-4, a real-world estimation of the overall percent reduction in
smoke exposure requires multiple pieces of data including demographic data, housing characteristic data,
and data on access or availability to various actions or interventions. Therefore, each of these pieces of
data will vary depending on geographic location, demonstrating that a one-size fits all approach is not
ideal, but can provide an estimation of the potential public health implications of reducing smoke
exposure using different actions or interventions as presented within this assessment. In the future, as
more data is collected on how people respond to wildfire smoke, such as through the SmokeSense app
I https://\\\v\\ .cpa.gov/air-rcscarch/sinokc-scnsc-studv-citizcn-scicncc-proicct-using-inobilc-app: U.S.
EPA (2020b)l. it could be possible to more fully account for and quantify the actions taken by individuals
affected by smoke through data analysis or exposure modeling, and subsequently assess the potential
overall smoke exposure reduction for a population.
6.4 Ecological Effects Attributed to Wildfire Smoke and
Deposition of Pollutants
Wildfire smoke and the deposition of pollutants on plants and animals in terrestrial and aquatic
environments can have a range of effects. For example, pathogenic fungi have been shown to be
aerosolized on smoke particulates and transported downwind from wildfires. Forest pests can be
stimulated by smoke where it serves as an attractant to pyrophilous beetle species that are adapted to
reproduce in the downed lumber and freshly burned wood following a fire (Lesk et al.. 2017; Hart. 1998;
Evans. 1971). In addition to effects on lower trophic levels, smoke effects have also been documented to
occur in vertebrates. After the fires of 1988 in Yellowstone National Park, for example, hundreds of large
mammals, including elk, moose, mule deer, and bison were found dead: autopsy evidence suggested that
smoke inhalation killed nearly all these animals (Singer and Schullerv. 1989). Smoke inhalation has also
been associated with mortality in raptors as a result of promoting subsequent fungal infection in lungs
following smoke exposure (Kinne et al.. 2010). While numerous adverse effects from wildfire emissions
have been documented, smoke can also have a stimulatory effect on the environment (McLauchlan et al..
2020). The following sections more fully characterize the direct effects of wildfire smoke and deposition
of pollutants on plants and animals in terrestrial and aquatic environments.
6.4.1 Particulate Matter (PM)
Wildland fire is an increasing source of particulate matter emissions (see CHAPTER 5). a
substantial fraction of which is represented by particulate matter, specifically PM2 5, which have been
shown to have a variety of impacts on the environment (Bond and Keane. 2017). While this section will
6-28
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
focus on aquatic and ecological effects, it's important to recognize the potential climatological impact of
wildfire smoke. Particulate matter generated from wildfires has been shown to affect cloud cover and ice
nucleation and interact with solar radiation through absorption and scattering. Specifically, the deposition
of the PM2 5 component black carbon has been shown to increase soil temperature through absorption of
solar radiation (U.S. EPA. 2012). This is noteworthy in Pacific Northwest forests given the increasing rate
and quantity of black carbon deposition in today's unprecedented fire regime because increased soil
temperature is associated with concurrent decreases in tree growth which is a precursor to tree mortality.
In contrast, certain highly reflective PM components in the atmosphere can scatter incoming solar
radiation with much of that energy returning to space and can have an overall cooling effect on the
climate (U.S. EPA. 2019b). Two of the better studied aspects of the effects of wildfire smoke on the
environment include the transport of microorganisms on smoke particles following a fire and
smoke-stimulated flowering/seed germination seed release.
6.4.1.1 Transport of Bacteria and Fungi from Soil and Plants through
Smoke
A relatively recent advancement in fire ecology includes the nascent field of pyroaerobiology,
which considers the living component of smoke particles generated from wildfires; specifically, microbes
aerosolized and transported on particles by wildland fire (Hu et al.. 2020; Kobziar and Thompson. 2020;
Kobziaretal.. 2018). Recent studies have documented elevated concentrations of bacteria and fungi in
smoke from burning of woody materials (Mirskava and Agranovski. 2020) and in smoke from coniferous
forest wildfires (Kobziar et al.. 2018) through the collection of microorganisms on passive samplers
oriented to the wind direction. The authors showed that microbial counts following the fire were
significantly elevated above ambient conditions and decreased with distance from the fire's flaming front.
It has been hypothesized that these microorganisms could represent an infectious risk to the public
(Kobziar and Thompson. 2020) while also serving as an important inoculum for reseeding the soil flora
following a fire event.
6.4.1.2 Smoke-Stimulated Flowering/Seed Germination and Seed Release
In their review of the ecological effects of fire, Bond and Keane (2017) noted that flowering is
common among perennial grasses and herbs, some species of which only flower when cued by smoke
(Chou et al.. 2012). Wildfire smoke also stimulates germination in soil seedbanks for many species
adapted to fire-prone forests and shrublands such as those in California (Keelev et al.. 2005). In such
environments, seedling recruitment from seed banks is one of the primary means of regeneration
following a wildfire event. In fire adapted environments, germination is the result of either heat shock or
exposure to combustion products in smoke. However, germination can also be stimulated in some fire
adapted species through direct deposition on the seed or as a result of smoke binding to soil particles and
6-29
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
subsequent aqueous or atmospheric transfer to seeds (Kcclcv et al.. 2005). While dozens of individual
chemicals and particulate matter comprise wildfire smoke, Keelev and Fotheringham (1997) showed that
it is the nitrogen oxides (NOx) present as trace gases in smoke that are responsible for seed germination.
6.4.2 Effects of Ozone (O3) from Fires
Wildfire smoke consists of numerous components (see CHAPTER 4 and CHAPTER 5).
including volatile organic compounds (VOCs) and NOx which can increase ozone production downwind
following a wildfire event (Jaffc and Wigder. 2012). Because VOCs are ubiquitous in most geographic
locations, it is generally thought that NOx concentrations are the rate-limiting factor to ozone formation.
The amount of NOx produced during a wildfire is a function of the N content of the fuel, which varies by
species, age and type of ecosystem, and the intensity of the burn. Higher temperature fires tend to produce
more oxidized forms of N than lower intensity burns, and therefore are thought to produce more of the
NOx precursors available for ozone production (Jaffc and Wigder. 2012). Lower intensity burns tend to
produce more oxygenated VOCs. The difference in ozone concentrations that can occur depending on fire
type is reflected in the different hypothetical scenarios examined in the case studies presented within this
assessment (see CHAPTER 5).
There is overwhelming evidence linking tropospheric ozone with reductions in growth and
productivity in both agricultural and natural ecosystems (U.S. EPA. 2020a). Ecological effects of ozone
can be observed across multiple scales of biological organization, from cellular to individual organism to
the level of communities and ecosystems. Ozone can affect both aboveground and belowground processes
leading to changes in productivity, carbon sequestration, biogeochemical cycling and hydrology.
At the plant level, ozone enters the leaves through stomates, and quickly disassociates in the leaf
apoplast into hydrogen peroxide (H2O2), organic radicals, and other reactive compounds that damage
cellular membranes (Wohlgemuth et al.. 2002; Hippeli and Elstner. 1996). Through both direct effects on
stomatal regulation (Grulke. 1999) as well as chloroplast degradation, ozone can decrease photosynthesis
and metabolism (Matvssek and Innes. 1999). Reductions in photosynthesis and overall carbon
assimilation leads to decreased growth, but also can result in a shift in allocation of carbon resources
within the plant, particularly to roots. These shifts in carbon allocation can lead to a change in the
physiological functioning of the plant, including changes in gene regulation (U.S. EPA. 2020a; Andersen.
2003).
The direct effects of ozone on carbon assimilation and plant growth can subsequently alter the
competitiveness of individuals in ecosystems. A reduction in carbon allocation to roots can alter
rhizosphere interactions and symbiotic associations, both potentially leading to changes in nutrient uptake
(U.S. EPA. 2020a). Changes in nutrient uptake therefore can lead to further reductions in growth, and to a
change in the competitive stature of the plant. Because not all species are equally susceptible to ozone,
there is often a shift in the competitive structure of ecosystems exposed to ozone, with sensitive species
6-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
dropping out and ozone tolerant species becoming more competitive. Through these direct and indirect
effects, both ecosystem structure and function can be altered by ozone stress downwind of a wildfire.
Ozone also influences the susceptibility of natural ecosystems to future wildfire. Ozone stress
often results in early senescence of leaves, which can increase fuel load in conifer forests such as
ponderosa pine that shed older leaf whorls in response to ozone stress (Miller et al.. 1982). Ozone
sensitive species are also more susceptible to other stresses, such as insects and pathogens, increasing tree
mortality and potentially increasing the fuel load in stressed ecosystems. Since ozone tends to reduce
carbon allocation to roots, ozone stressed plants also can become drought stressed, further increasing their
susceptibility to other stresses and to wildfire. This can be a positive feedback loop in that as fire occurs,
ozone is produced in smoke, potentially leading to susceptibility in future fire events. In addition to
ozone, many studies have documented the release of hazardous organic and inorganic chemicals from
combustion of biomass through wildfire and there is a vast literature on the toxicity of organic chemicals
and heavy metals on plants and animals. There are, however, no studies available reporting demonstrable
ecological effects from hazardous pollutants released or generated from wildland fire.
6.4.3 Atmospheric Deposition of Ash
The most immediate effect of wildland fire on the land surface is the removal of vegetation and
the subsequent deposition of a layer of charcoal or ash (De Sales et al.. 2019). Ash is the particulate
residue that consists of mineral and charred organic materials formed when carbon fuels are burned (Bodi
et al.. 2014). Characteristics of ash are affected by the type of fuel burned and intensity of combustion,
with low-intensity fires yielding ash of greater organic content and hotter fires resulting in more
mineralized material. In forested environments, the mass of ash deposition following a fire can range
from 2-9% of woody biomass (Raison. 1979).
Ash that is deposited on the ground is incorporated into soil where vegetation has burned. Given
its high mobility, however, ash is also readily transported downwind and downstream where it can
influence habitats far removed from areas burned by wildfire. Ash deposition is becoming an increasingly
common input into ecosystems and it can have a dramatic effect on the biogeochemical cycling of
nutrients and minerals in forested soils. This section considers the ecological effects of ash on soil
chemistry and structure, nutrient flux, microbial activity, and plant growth.
6.4.3.1 Soil Chemistry and Structure
Ash deposition following wildfire can profoundly change soil characteristics. In a study of ion
release from burning plant material, Grier and Cole (1971) demonstrated greatly increased concentrations
of ions entering the soil, which were adsorbed in the uppermost soil horizons and caused major chemical
changes such as the influx of basic ions increasing soil pH. In a study of wildfire sites in California, Ulerv
6-31
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
et al. (1993) showed that ash deposition raised soil pH by as much as 3 pH units (to pH 10.5) compared
with unburned soil. More basic pH increases the solubility of soil organic carbon (Andersson et al.. 1994)
and increases the number of binding sites in soil that can hold cationic micronutrients (Raison. 1979).
The physical deposition of ash can act to increase soil water repellency, preventing the infiltration
of meteoric precipitation (Doerretal.. 2000) and decreasing the potential for nutrient leaching. Another
consequence of ash deposition on soil structure is an increase in bulk density, which is the soil mass
divided by the bulk volume of the sample (g/cm3). The bulk density of soil increases with ash deposition
because soil aggregates collapse and the ash clogs pore-space voids, both of which serve to decrease soil
porosity and permeability (Vcrina and Javakumar. 2012). Factors like increased soil hydrophobicity and
increased density that limit the infiltration of meteoric water would help to retain otherwise leached soil
nutrients.
6.4.3.2 Stimulation of Microbiological Activity and Plant Growth
Although it is widely accepted that fire stimulates microbial activity (Bodi et al.. 2014). most
research on wildfire's effect considers soil heating only where, in extremely hot fires, sterilization of the
upper soil layers can occur (Mataix-Solera et al.. 2009). Far fewer studies address the effects of ash
deposition on soil microbiota and nutrient processing. Compared to the growth of fungal organisms that
would occur at lower soil pH, Jokinen et al. (2006) suggested that the increased soil pH and nutrient and
carbon availability from ash deposition stimulated bacterial respiration. Bacteria proliferate more quickly
than fungi and their ability to capitalize on a new carbon pool, such as ash-mobilized organic carbon,
would favor bacterial growth suggesting an inhibitory effect of ash deposition on fungal microflora.
Mycorrhizal fungi, however, have a symbiotic relationship with plants that depends on the latter's ability
to produce carbohydrates through photosynthesis and share sugars with the fungus. In this relationship,
plants receive water and nutrients from the soil by the extensive network of fungal mycelial hyphae. The
plant's provision of carbohydrates to mycorrhizae makes this group of fungi competitive with bacteria in
an otherwise challenging post-fire environment for fungal organisms.
New tree growth in burned forests is highly dependent upon mycorrhizal symbiosis and the
fungal colonization of burned areas is relatively well documented. In a study of burned pine forests in
northern California, Grogan et al. (2000) found that wildfire disturbance resulted in marked changes in
mycorrhizal community composition and a significant increase in the relative biomass of
mycorrhizal-ascomycetous fungi. Additionally, in an experiment to examine the effects of
ectomycorrhizal colonization and fire on the growth of Bishop pine seedlings (Pinus muricata) in
northern California, Peav et al. (2010) showed that the percent nitrogen in needles was greatest in
treatments with an ectomycorrhizal inoculum regardless of whether ash was added to soil. These results
underly the critical relationship of pine forests and their dependence on mycorrhizal associations.
6-32
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Immediately downwind of fires, larger particles of ash are deposited onto vegetation with a
concomitant observable soiling of leaves, which can adversely affect photosynthesis and plant growth.
Nutrients may be released from combustible fuels after fire and transported as ash by atmospheric
deposition to stimulate vegetation growth (Bodi et al.. 2014). The amounts of calcium, nitrogen,
phosphorous, potassium, magnesium and sulfur released by burning forest vegetation are elevated in
relation to both the total and available quantities of these elements in soils (Raison and McGaritv. 1980:
Raison. 1979). The addition of nutrients in ash tend to stimulate plant growth although germination may
be inhibited by deposited ash, due perhaps to ash's hydrophobicity and osmotic pressure excluding water
from the seed, the presence of toxic elements in ash, and/or elevated pH (Bodi et al.. 2014).
Nitrogen is among the most important nutrients that can stimulate plant growth. Forested systems
rely on cycling the nitrogen locked in dead plant matter into more bioavailable forms. Compared to the
biological decay of plant remains, burning rapidly releases nutrients into a plant-available form. Nitrogen
from wildfires can represent over 30% of nitrogen deposition in forested systems of the Pacific Northwest
(Koplitz et al.. In Press) and growth of the predominant forest tree species (Douglas fir, Pseudotsuga
menziesii) in the Pacific Northwest is stimulated by nitrogen deposition. However, too much nitrogen can
be problematic and lead to nitrogen inputs exceeding the critical nitrogen load in Northwest forests and
ultimately decreased tree survival.
6.4.3.3 Ash Deposition and Water Quality
The aerial transport and deposition of materials in smoke and ash may also affect downwind
water quality. Increased runoff of ash, sediments, and chemical constituents following fire appear to be
the dominant mechanism by which water quality is affected (see Section 7.3.3.2.5 for further discussion
of fire effects on water quality, including potential effects on drinking water). Nevertheless, it is logical to
assume that some material could be deposited in wet or dry form onto the surfaces of downwind water
bodies, such as streams, lakes or reservoirs, or deposited on unburned terrestrial surfaces and
subsequently moved via overland or subsurface flow to water bodies. Post-fire increases in nutrient
deposition (Ranalli. 2004; Koplitz et al.. In Press), and wind dispersion of ash, nutrients, and sediments
(Roehner et al.. 2020; Bodi et al.. 2014) are suggestive of such a mechanism.
Though studies measuring this phenomenon are limited, several have reported water quality
changes potentially linked to aerial transport of materials from fires (Earl and Blinn. 2003; Lathrop. 1994;
Spencer and Hauer. 1991). For instance, Earl and Blinn (2003) found higher nutrient concentrations in an
unburned watershed in southwestern New Mexico, associated in time with a nearby fire. The authors
suggested that aerial transport of nutrients from the fire was likely responsible. Initiating sampling within
hours of a fire, Spencer and Hauer (1991) observed spikes in nitrogen and phosphorus in stream water,
before returning to background levels within several days to weeks. The authors concluded nitrogen from
the smoke diffused into the surface water, while phosphorus leached from ash deposited directly into the
6-33
DRAFT: Do Not Cite or Quote
-------
1 waterbodies. Nitrogen volatilizes at lower temperatures than phosphorus, likely explaining the differences
2 in method of transport of these two nutrients.
6.4.4 Uncertainties and Limitations in the Ecological Effects Evidence
3 There are considerable uncertainties and limitations in understanding the ecological effects of
4 smoke and ash on plants and animals. Ultimately ecosystems have adapted to fire regimes, but an
5 understanding of fire's immediate ecological effects are limited by a dearth of studies on the direct
6 ecological effects of smoke and ash. The influx of fire-liberated nutrients and hazardous pollutants on
7 terrestrial and aquatic receptors is just beginning to be investigated and the time frame over which fires
8 influence air and water chemistry is an area that warrants further investigation.
6-34
DRAFT: Do Not Cite or Quote
-------
6.5 References
Adetona. Q: Reinhardt. TE: Domitrovich. J: Brovles. G: Adetona. AM: Kleinman. MT: Ottmar. RD: Naeher. LP.
(2016). Review of the health effects of wildland fire smoke on wildland firefighters and the public [Review].
Inhal Toxicol 28: 95-139. http://dx.doi.org/10.3109/08958378.2016.1145771
Alaw. M: Siegel. JA. (2020). In-situ effectiveness of residential HVAC filters. Indoor Air 30: 156-166.
http://dx.doi.org/10. Ill 1/ina. 12617
Allen. RW: Barn. P. (2020). Individual- and household-level interventions to reduce air pollution exposures and
health risks: A review of the recent literature [Review]. Curr Environ Health Rep 7: 424-440.
http://dx.doi.org/10.1007/s40572-020-0Q296-z
Alman. BL: Pfister. G: Hao. H: Stowell. J: Hu. X: Liu. Y: Strickland. MJ. (2016). The association of wildfire
smoke with respiratory and cardiovascular emergency department visits in Colorado in 2012: A case
crossover study. Environ Health 15: 64. http://dx.doi.org/10.1186/sl2940-016-0146-8
Andersen. CP. (2003). Source-sink balance and carbon allocation below ground in plants exposed to ozone
[Review], New Phytol 157: 213-228. http://dx.doi.Org/10.1046/i.1469-8137.2003.00674.x
Andersson. S: Valeur. I: Nilsson. I. (1994). Influence of lime on soil respiration, leaching of DOC, and C/S
relationships in the mor humus of a haplic podsol. Environ Int 20: 81-88. http://dx.doi.org/10.1016/0160-
4120(94)90070-1
Barn. P: Larson. T: Noullett. M: Kennedy. S: Copes. R: Brauer. M. (2008). Infiltration of forest fire and
residential wood smoke: An evaluation of air cleaner effectiveness. J Expo Sci Environ Epidemiol 18: 503 -
511. http://dx.doi.org/10.1038/si.ies.750064Q
Barn. PK: Elliott. CT: Allen. RW: Kosatskv. T: Rideout. K: Henderson. SB. (2016). Portable air cleaners should
be at the forefront of the public health response to landscape fire smoke [Letter], Environ Health 15: 116.
http://dx.doi.org/10.1186/sl2940-016-0198-9
Black. C: Tesfaigzi. Y: Bassein. JA: Miller. LA. (2017). Wildfire smoke exposure and human health: Significant
gaps in research for a growing public health issue [Review]. Environ Toxicol Pharmacol 55: 186-195.
http://dx.doi.Org/10.1016/i.etap.2017.08.022
Blades. JJ: Shook. SR: Hall. TE. (2014). Smoke management of wildland and prescribed fire: Understanding
public preferences and trade-offs. Can J For Res 44: 1344-1355. http://dx.doi.org/10.1139/cifr-2014-0110
Bodi. MB: Martin. DA: Balfour. VN: Santin. C: Doerr. SH: Pereira. P: Cerda. A: Mataix-Solera. J. (2014). Wild
land fire ash: Production, composition and eco-hydro-geomorphic effects. Earth Sci Rev 130: 103-127.
http://dx.doi.Org/10.1016/i.earscirev.2013.12.007
Bond. WJ: Keane. RE. (2017). Fire, ecological effects of. In BD Roitberg (Ed.), Reference module in life
sciences (pp. 1-11). Amsterdam, Netherlands: Elsevier. http://dx.doi.org/10.1016/B978-0-12-8Q9633-
8.02098-7
Chou. YF: Cox. RD: Wester. DB. (2012). Smoke water and heat shock influence germination of shortgrass
prairie species. Rangeland Ecol Manag 65: 260-267. http://dx.doi.Org/10.2111/REM-D-ll-00093.l
Davison. G: Barkiohn. K: Hagler. GS: Holder. A: Coefield. S: Noonan. C: Hassett-Sipple. B. (In Press) Creating
clean air spaces during wildland fire smoke episodes: Web summit summary. Front Public Health.
http://dx.doi.org/10.3389/fpubh.2021.508971
De Sales. F: Okin. GS: Xue. Y: Dintwe. K. (2019). On the effects of wildfires on precipitation in Southern
Africa. ClimDynam 52: 951-967. http://dx.doi.org/10.1007/sQ0382-018-4174-7
DeFlorio-Barker. S: Crooks. J: Reves. J: Rappold. AG. (2019). Cardiopulmonary effects of fine particulate
matter exposure among older adults, during wildfire and non-wildfire periods, in the United States 2008-
2010. Environ Health Perspect 127: 37006. http://dx.doi.org/10.1289/EHP3860
6-35
DRAFT: Do Not Cite or Quote
-------
Delfino. RJ: Brummel. S: Wu. J: Stern. H: Ostro. B: Lipsett. M: Winer. A: Street. DH: Zhang. L: Tioa. T: Gillen.
PL. (2009). The relationship of respiratory and cardiovascular hospital admissions to the southern California
wildfires of 2003. Occup Environ Med 66: 189-197. http://dx.doi.org/10.1136/oem.2008.041376
Doerr. SH: Shakesbv. RA: Walsh. RPD. (2000). Soil water repellency: Its causes, characteristics and hydro-
geomorphological significance. Earth Sci Rev 51: 33-65. http://dx.doi.org/10.1016/S0012-8252(00)00011-8
Doubledav. A: Schulte. J: Sheppard. L: Kadlec. M: Dhammapala. R: Fox. J: Isaksen. TB. (2020). Mortality
associated with wildfire smoke exposure in Washington state, 2006-2017: A case-crossover study. Environ
Health 19: 4. http://dx.doi.org/10.1186/sl2940-020-0559-2
Earl. SR: Blinn. DW. (2003). Effects of wildfire ash on water chemistry and biota in South-Western U.S.A.
streams. Freshw Biol 48: 1015-1030. http://dx.doi.Org/10.1046/i.1365-2427.2003.01066.x
Evans. WG. (1971). The attraction of insects to forest fires. In Proceedings of the Tall Timbers Conference on
Ecological Animal Control by Habitat Management. Tallahassee, FL: Tall Timbers Research Station.
Fisk. WJ: Chan. WR. (2017a). Effectiveness and cost of reducing particle-related mortality with particle
filtration. Indoor Air 27: 909-920. http://dx.doi.org/10. Ill 1/ina. 12371
Fisk. WJ: Chan. WR. (2017b). Health benefits and costs of filtration interventions that reduce indoor exposure to
PM2.5 during wildfires. Indoor Air 27: 191-204. http://dx.doi.org/10. Ill 1/ina. 12285
Gan. RW: Ford. B: Lassman. W: Pfister. G: Vaidvanathan. A: Fischer. E: Volckens. J: Pierce. JR: Magzamen. S.
(2017). Comparison of wildfire smoke estimation methods and associations with cardiopulmonary-related
hospital admissions. Geohealth 1: 122-136. http://dx.doi.org/10.1002/2017GH000Q73
Grier. CC: Cole. DW. (1971). Influence of slash burning on ion transport in a forest soil. NW Sci 45: 100-106.
http://dx.doi.Org/BCI:BCI197253019441
Grogan. P: Baar. J: Bruns. TP. (2000). Below-ground ectomycorrhizal community structure in a recently burned
bishop pine forest. J Ecol 88: 1051-1062. http://dx.doi.Org/10.1046/i.1365-2745.2000.00511.x
Grulke. NE. (1999). Physiological responses of ponderosa pine to gradients of environmental stressors. In PR
Miller; JR McBride (Eds.), Oxidant air pollution impacts in the montane forests of southern California: A
case study of the San Bernardino Mountains (pp. 126-163). New York, NY: Springer.
http://dx.doi.org/10.10Q7/978-l-4612-1436-6 7
Hano. MC: Prince. SE: Wei. L: Hubbell. BJ: Rappold. AG. (2020). Knowing your audience: A typology of
smoke sense participants to inform wildfire smoke health risk communication. Front Public Health 8: 143.
http://dx.doi.org/10.3389/fpubh.2020.00143
Hart. S. (1998). Beetle mania: An attraction to fire. Bioscience 48: 3-5. http://dx.doi.org/10.2307/1313221
Henderson. DE: Milford. JB: Miller. SL. (2005). Prescribed burns and wildfires in Colorado: Impacts of
mitigation measures on indoor air particulate matter. J Air Waste Manag Assoc 55: 1516-1526.
http://dx.doi.org/10.1080/10473289.20Q5.10464746
Hippeli. S: Elstner. EF. (1996). Mechanisms of oxygen activation during plant stress: Biochemical effects of air
pollutants. J Plant Physiol 148: 249-257. http://dx.doi.org/10.1016/S0176-1617(96)80250-l
Hu. W: Wang. ZH: Huang. S: Ren. LJ: Yue. SY: Li. P: Xie. OR: Zhao. WY: Wei. LF: Ren. H: Wu. LB: Deng.
JJ: Fu. PO. (2020). Biological aerosol particles in polluted regions. Curr Pollut Rep 6: 65-89.
http://dx.doi.org/10.1007/s40726-020-0Q138-4
Hutchinson. JA: Vargo. J: Milet. M: French. NHF: Billmire. M: Johnson. J: Hoshiko. S. (2018). The San Diego
2007 wildfires and Medi-Cal emergency department presentations, inpatient hospitalizations, and outpatient
visits: An observational study of smoke exposure periods and a bidirectional case-crossover analysis. PLoS
Med 15: el002601. http://dx.doi.org/10.1371/iournal.pmed.1002601
Jaffe. DA: O'Neill. SM: Larkin. NK: Holder. AL: Peterson. PL: Halofskv. JE: Rappold. AG. (2020). Wildfire
and prescribed burning impacts on air quality in the United States [Editorial]. J Air Waste Manag Assoc 70:
583-615. http://dx.doi.org/10.1080/10962247.2020.1749731
6-36
DRAFT: Do Not Cite or Quote
-------
Jaffe. DA: Wigder. NL. (2012). Ozone production from wildfires: A critical review [Review]. Atmos Environ
51: 1-10. http://dx.doi.Org/10.1016/i.atmosenv.2011.ll.063
Jokinen. HK: Kiikkila. O: Fritze. H. (2006). Exploring the mechanisms behind elevated microbial activity after
wood ash application. Soil Biol Biochem 38: 2285-2291. http://dx.doi.Org/10.1016/i.soilbio.2006.02.007
Jones. CG: Rappold. AG: Vargo. J: Cascio. WE: Kharrazi. M: McNallv. B: Hoshiko. S. (2020). Out-of-hospital
cardiac arrests and wildfire-related particulate matter during 2015-2017 California wildfires. J Am Heart
Assoc 9: e014125. http://dx.doi.org/10.1161/JAHA.119.014125
Joseph. G: Schramm. PJ: Vaidvanathan. A: Brevsse. P: Goodwin. B. (2020). Evidence on the use of indoor air
filtration as an intervention for wildfire smoke pollutant exposure: A summary for health departments.
Atlanta, GA: Centers for Disease Control and Prevention, National Center for Environmental Health.
https://www.cdc.gov/air/wildfire-smoke/socialmedia/Wildfire-Air-Filtration-508.pdf
Keelev. JE: Fotheringham. CJ. (1997). Trace gas emissions and smoke-induced seed germination. Science 276:
1248-1250. http://dx.doi.org/10.1126/science.276.5316.1248
Keelev. JE: McGinnis. TW: Bollens. KA. (2005). Seed germination of Sierra Nevada postfire chaparral species.
Madrono 52: 175-181. http://dx.doi.org/10.3120/0024-9637(2005)52ri75:SGOSNP12.0.CQ:2
Kim. YH: Warren. SH: Krantz. OT: King. C: Jaskot. R: Preston. WT: George. BJ: Havs. MP: Landis. MS:
Higuchi. M: DeMarini. DM: Gilmour. MI. (2018). Mutagenicity and lung toxicity of smoldering vs. flaming
emissions from various biomass fuels: Implications for health effects from wildland fires. Environ Health
Perspect 126: 017011. http://dx.doi.org/10.1289/EHP2200
Kinne. J: Bailey. TA: Kilgallon. C: Louagie. E: Wernerv. U. (2010). Aspergillosis in raptors after smoke-
inhalation injury. Falco (35): 22-24.
Klepeis. NE: Nelsoa WC: Ott. WR: Robinson. JP: Tsang. AM: Switzer. P: Behar. JV: Hern. SC: Engelmana
WH. (2001). The National Human Activity Pattern Survey (NHAPS): A resource for assessing exposure to
environmental pollutants. J Expo Anal Environ Epidemiol 11: 231-252.
http://dx.doi.org/10.1038/si.iea.7500165
Kobziar. LN: Pingree. MRA: Larson. H: Dreaden. TJ: Green. S: Smith. JA. (2018). Pyroaerobiology: The
aerosolization and transport of viable microbial life by wildland fire. Ecosphere 9: e02507.
http://dx.doi.org/10.1002/ecs2.2507
Kobziar. LN: Thompson. GR. III. (2020). Wildfire smoke, a potential infectious agent [Editorial]. Science 370:
1408-1410. http://dx.doi.org/10.1126/science.abe8116
Kolbe. A: Gilchrist. KL. (2009). An extreme bushfire smoke pollution event: Health impacts and public health
challenges. NSW Public Health Bulletin 20: 19-23. http://dx.doi.org/10.1071/nb08061
Koplitz. S: Nolte. CG: Sabo. RD: Clark. CM: Horn. KJ: Thomas. RO: Newcomer-Johnson. TA. . (In Press) The
contribution of wildland fire emissions to deposition in the U.S.: Implications for tree growth and survival in
the Northwest.
Ktinzli. N: Avol. E: Wu. J: Gaudermaa WJ: Rappaport. E: Millstein. J: Bennion. J: McConnell. R: Gilliland.
FD: Berhane. K: Lurmann. F: Winer. A: Peters. JM. (2003). Health effects of the 2003 southern California
wildfire on children. Am J Respir Crit Care Med 174: 1221-1228. http://dx.doi.org/10.1164/rccm.200604-
5190C
Landguth. EL: Holden. ZA: Graham. J: Stark. B: Mokhtari. EB: Kaleczvc. E: Anderson. S: Urbanski. S: Jolly.
M: Semmens. EO: Warren. DA: Swanson. A: Stone. E: Noonan. C. (2020). The delayed effect of wildfire
season particulate matter on subsequent influenza season in a mountain west region of the USA. Environ Int
139: 105668. http://dx.doi.Org/10.1016/i.envint.2020.105668
Larsen. AE: Reich. BJ: Ruminski. M: Rappold. AG. (2018). Impacts of fire smoke plumes on regional air
quality, 2006-2013. J Expo Sci Environ Epidemiol 28: 319-327. http://dx.doi.org/10.1038/s41370-017-0Q13-
x
Lathrop. RG. (1994). Impacts of the 1988 wildfires on the water-quality of Yellowstone and Lewis Lakes,
Wyoming. International Journal of Wildland Fire 4: 169-175. http://dx.doi.org/10.1071/WF994Q169
6-37
DRAFT: Do Not Cite or Quote
-------
Laumbach. RJ. (2019). Clearing the air on personal interventions to reduce exposure to wildfire smoke
[Editorial]. Ann Am Thorac Soc 16: 815-818. http://dx.doi.org/10.1513/AnnalsATS.201812-894PS
Leibel. S: Nguyen. M: Brick. W: Parker. J: Ilango. S: Aguilera. R: Gershunov. A: Benmarhnia. T. (2020).
Increase in pediatric respiratory visits associated with Santa Ana wind-driven wildfire smoke and PM2.5
levels in San Diego County. Ann Am Thorac Soc 17: 313 -320.
http://dx.doi.org/10.1513/Anna1sATS.201902-150QC
Lesk. C: Coffel. E: D'Amato. AW: Dodds. K: Hortoa R. (2017). Threats to North American forests from
southern pine beetle with warming winters. Nat Clim Chang 7: 713-717.
http://dx.doi.org/10.1038/NCLIMATE3375
Liu. JC: Wilson. A: Micklev. LJ: Dominici. F: Ebisu. K: Wang. Y: Sulprizio. MP: Peng. RD: Yue. X: Son. JY:
Anderson. GB: Bell. ML. (2017a). Wildfire-specific fine particulate matter and risk of hospital admissions in
urban and rural counties. Epidemiology 28: 77-85. http://dx.doi.org/10.1097/EDE.000000000000Q556
Liu. JC: Wilson. A: Micklev. LJ: Ebisu. K: Sulprizio. MP: Wang. Y: Peng. RD: Yue. X: Dominici. F: Bell. ML.
(2017b). Who among the elderly is most vulnerable to exposure and health risks of PM2.5 from wildfire
smoke? Am J Epidemiol 186: 730-735. http://dx.doi.org/10.1093/aie/kwxl41
Lutes. DC. (2014). First Order Fire Effects Model (FOFEM) mapping tool, version 6.1: User guide. Missoula,
MT: U.S. Forest Service, Rocky Mountain Research Station, Fire Modeling Institute.
https://www.frames.gov/catalog/17743
Marr. D: Mason. M: Moslev. R. on: Liu. X. (2012). The influence of opening windows and doors on the natural
ventilation rate of a residential building. HVAC R Res 18: 195-203.
http://dx.doi.org/10.108Q/10789669.2011.585423
Mataix-Solera. J: Guerrero. C: Garcia-Orenes. F: Barcenas. GM: Torres. MP. (2009). Forest fire effects on soil
microbiology. In A Cerda; PRRobichaud (Eds.), Fire effects on soils and restoration strategies (pp. 133-
175). Enfield, NH: Science Publishers, https://www.tavlorfrancis.com/books/fire-effects-soils-restoration-
strategies-cerda/e/10.1201/9781439843338?refId=5936698c-e4aa-4976-a9d3-814c5352e9a5
Matvssek. R: Innes. JL. (1999). Ozone - A risk factor for trees and forests in Europe? Water Air Soil Pollut 116:
199-226. http://dx.doi.Org/10.1023/A:1005267214560
McLauchlan. KK: Higuera. PE: Miesel. J: Rogers. BM: Schweitzer. J: Shuman. JK: Teplev. AJ: Varner. JM:
Veblen. TT: Adalsteinsson. SA: Balch. JK: Baker. P: Batllori. E: Bigio. E: Brando. P: Cattau. M: Chipman.
ML: Coen. J: Crandall. R: Daniels. L: Enright. N: Gross. WS: Harvey. BJ: Hatten. JA: Hermana S: Hewitt.
RE: Kobziar. L: Landesmann. JB: Lorantv. MM: Maezumi. SY: Mearns. L: Moritz. M: Myers. JA: Pausas.
JG: Pellegrini. AFA: Piatt. WJ: Roozeboom. J: Safford. H: Santos. F: Scheller. RM: Sherriff. R: Smith. KG:
Smith. MP: Watts. AC. (2020). Fire as a fundamental ecological process: Research advances and frontiers
[Review], JEcol 108: 2047-2069. http://dx.doi.org/10.llll/1365-2745.13403
Miller. PR: Taylor. OC: Wilhour. RG. (1982). Oxidant air pollution effects on a western coniferous forest
ecosystem [EPA Report]. (EPA Report No. EPA-600/D-82-276). Corvallis, OR: U.S. Environmental
Protection Agency, Environmental Research Laboratory.
http://cfpub.epa.gov/si/ntislink.cfm?dirEntrvID=47047
Mirskava. E: Agranovski. IE. (2020). Generation of viable bacterial and fungal aerosols during biomass
combustion. Atmosphere (Basel) 11: 313. http://dx.doi.org/10.3390/atmosl 1030313
Mott. JA: Meyer. P: Manning. D: Redd. SC: Smith. EM: Gotway-Crawford. C: Chase. E. (2002). Wildland
forest fire smoke: Health effects and intervention evaluation, Hoopa, California, 1999. West J Med 176: 157-
162. http://dx.doi.Org/10.1136/ewim.176.3.157
NIFC (National Interagency Fire Center). (2021). InciWeb - Incident information system. Available online at
https://inciweb.nwcg.gov/ (accessed February 9, 2021).
NIOSH (National Institute for Occupational Safety and Health). (2015). Hierarchy of controls. Available online
at https://www.cdc.gov/niosh/topics/hierarchy/default.html (accessed March 10, 2021).
6-38
DRAFT: Do Not Cite or Quote
-------
Olsen. CS: Toman. E: Frederick. S. (2017). A multi-region analysis of factors that influence public acceptance of
smoke from different fire sources. International Journal of Wildland Fire 26: 364-374.
http://dx.doi.org/10.1071AVF16034
Orr. A: Migliaccio. CAL: Buford. M: Ballou. S: Migliaccio. CT. (2020). Sustained effects on lung function in
community members following exposure to hazardous PM2.5 levels from wildfire smoke. Toxics 8: 53.
http://dx.doi.org/10.3390/toxics8030Q53
Pantelic. J: Dawe. M: Licina. D. (2019). Use of IoT sensing and occupant surveys for determining the resilience
of buildings to forest fire generated PM2.5. PLoS ONE 14: e0223136.
http://dx.doi.org/10.1371/iournal.pone.0223136
Peav. KG: Bruns. TP: Garbelotto. M. (2010). Testing the ecological stability of ectomycorrhizal symbiosis:
Effects of heat, ash and mycorrhizal colonization on Pinus muricata seedling performance. Plant Soil 330:
291-302. http://dx.doi.org/10.1007/slll04-009-020Q-l
Prunicki. M: Kelsev. R: Lee. J: Zhou. X: Smith. E: Haddad. F: Wu. J: Nadeau. K. (2019). The impact of
prescribed fire versus wildfire on the immune and cardiovascular systems of children [Letter], Allergy 74:
1989-1991. http://dx.doi.org/10.1111/all. 13825
Raish. C: Gonzalez-Caban. A: Condie. CJ. (2005). The importance of traditional fire use and management
practices for contemporary land managers in the American Southwest. Glob Env Chang Part B: Env Hazards
6: 115-122. http://dx.doi.Org/10.1016/i.hazards.2005.10.004
Raison. RJ. (1979). Modification of the soil environment by vegetation fires, with particular reference to
nitrogen transformations: A review [Review]. Plant Soil 51: 73-108. http://dx.doi.org/10.1007/BF022Q5929
Raison. RJ: McGaritv. JW. (1980). Some effects of plant ash on the chemical properties of soils and aqueous
suspensions. Plant Soil 55: 339-352. http://dx.doi.org/10.1007/BF02182695
Ranalli. AJ. (2004). A summary of the scientific literature on the effects of fire on the concentration of nutrients
in surface waters. (Open-File Report 2004-1296). Reston, VA: U.S. Geological Survey.
https://pubs.usgs.gov/of/2004/1296/
Rappold. A: Stone. SL: Cascio. WE: Neas. LM: Kilaru. VJ: Carrawav. MS: Szvkman. JJ: Ising. A: Cleve. WE:
Meredith. JT: Vaughan-Batten. H: Devneka. L: Devlia RB. (2011). Peatbog wildfire smoke exposure in
rural North Carolina is associated with cardiopulmonary emergency department visits assessed through
syndromic surveillance. Environ Health Perspect 119: 1415-1420. http://dx.doi.org/10.1289/ehp. 1003206
Rappold. AG: Cascio. WE: Kilaru. VJ: Stone. SL: Neas. LM: Devlin. RB: Diaz-Sanchez. D. (2012). Cardio-
respiratory outcomes associated with exposure to wildfire smoke are modified by measures of community
health. Environ Health 11:71. http://dx.doi.org/10.1186/1476-069X-11-71
Rappold. AG: Hano. MC: Prince. S: Wei. L: Huang. SM: Baghdikian. C: Stearns. B: Gao. X: Hoshiko. S:
Cascio. WE: Diaz-Sanchez. D: Hubbell. B. (2019). Smoke sense initiative leverages citizen science to
address the growing wildfire-related public health problem. Geohealth 3: 443 -457.
http://dx.doi.org/10.1029/2019GH00Q199
Reid. CE: Brauer. M: Johnston. FH: Jerrett. M: Balmes. JR: Elliott. CT. (2016). Critical review of health impacts
of wildfire smoke exposure [Review]. Environ Health Perspect 124: 1334-1343.
http://dx.doi.org/10.1289/ehp. 1409277
Reid. CE: Considine. EM: Watson. GL: Telesca. D: Pfister. GG: Jerrett. M. (2019). Associations between
respiratory health and ozone and fine particulate matter during a wildfire event. Environ Int 129: 291-298.
http://dx.doi.Org/10.1016/i.envint.2019.04.033
Reisen. F: Powell. JC: Dennekamp. M: Johnston. FH: Wheeler. AJ. (2019). Is remaining indoors an effective
way of reducing exposure to fine particulate matter during biomass burning events? J Air Waste Manag
Assoc 69: 611-622. http://dx.doi.org/10.1080/10962247.2Q19.1567623
Richardson. LA: Champ. PA: Loomis. JB. (2012). The hidden cost of wildfires: Economic valuation of health
effects of wildfire smoke exposure in Southern California. J Forest Econ 18: 14-35.
http://dx.doi.Org/10.1016/i.ife.2011.05.002
6-39
DRAFT: Do Not Cite or Quote
-------
Roehner. C: Pierce. JL: Yager. EM. (2020). Spatial and temporal changes in aeolian redistribution of sediments
and nutrients following fire. Earth Surface Processes and Landforms 45: 2556-2571.
http://dx.doi.org/10.1002/esp.4913
Ryan. KC: Knapp. EE: Varner. JM. (2013). Prescribed fire in North American forests and woodlands: History,
current practice, and challenges. Front Ecol Environ 11: E15-E24. http://dx.doi.org/10.1890/12Q329
Singer. BC: Delp. WW: Black. PR: Walker. IS. (2017). Measured performance of filtration and ventilation
systems for fine and ultrafine particles and ozone in an unoccupied modern California house. Indoor Air 27:
780-790. http://dx.doi.org/10.1111/ina. 12359
Singer. FJ: Schullerv. P. (1989). Yellowstone wildlife: Populations in process. West Wildlands 15: 18-22.
Spencer. CN: Hauer. FR. (1991). Phosphorus and nitrogen dynamics in streams during a wildfire. J North Am
Benthol Soc 10: 24-30. http://dx.doi.org/10.2307/1467761
Stauffer. DA: Autenrieth. DA: Hart. JF: Capoccia. S. (2020). Control of wildfire-sourced PM2.5 in an office
setting using a commercially available portable air cleaner. J Occup Environ Hyg 17: 109-120.
http://dx.doi.org/10.1080/15459624.202Q.1722314
Stowell. JD: Geng. G: Saikawa. E: Chang. HH: Fu. J: Yang. CE: Zhu. O: Liu. Y: Strickland. MJ. (2019).
Associations of wildfire smoke PM2.5 exposure with cardiorespiratory events in Colorado 2011-2014.
Environ Int 133: 105151. http://dx.doi.org/10.1016/i.envint.2019.105151
Sugerman. DE: Keir. JM: Dee. PL: Lipman. H: Waterman. SH: Ginsberg. M: Fishbein. DB. (2012). Emergency
health risk communication during the 2007 San Diego wildfires: Comprehension, compliance, and recall. J
Health Commnn 17: 698-712. http://dx.doi.org/10.1080/1081073Q.2011.635777
Tinling. MA: West. JJ: Cascio. WE: Kilaru. V: Rappold. AG. (2016). Repeating cardiopulmonary health effects
in rural North Carolina population during a second large peat wildfire. Environ Health 15: 12.
http://dx.doi.org/10.1186/sl2940-016-0Q93-4
U.S. EPA (U.S. Environmental Protection Agency). (2012). Report to Congress on black carbon [EPA Report].
(EPA-450/R-12-001). Washington, DC. https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100EIJZ.txt
U.S. EPA (U.S. Environmental Protection Agency). (2018). Residential air cleaners: A technical summary [EPA
Report] (3rd ed.). (EPA 402-F-09-002). Washington, DC.
https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100UX5R.txt
U.S. EPA (U.S. Environmental Protection Agency). (2019a). Environmental Benefits Mapping and Analysis
Program - Community Edition (BenMAP-CE) (Version 1.5) [Computer Program], Washington, DC.
Retrieved from https://www.epa.gov/benmap/benmap-communitv-edition
U.S. EPA (U.S. Environmental Protection Agency). (2019b). Integrated Science Assessment (ISA) for
particulate matter (final report, Dec 2019) [EPA Report]. (EPA/600/R-19/188). Washington, DC.
https ://cfpub .epa. gov/ncea/isa/recordisplav ,cfm?deid=3 47534
U.S. EPA (U.S. Environmental Protection Agency). (2019c). Wildfire smoke: A guide for public health officials,
revised 2019 [EPA Report]. (EPA-452/R-19-901). Washington, DC: U.S. Environmental Protection Agency,
Office of Research and Development, https://www.airnow.gov/publications/wildfire-smoke-guide/wildfire-
smoke-a-guide-for-public-health-officials/
U.S. EPA (U.S. Environmental Protection Agency). (2020a). Integrated Science Assessment (ISA) for ozone and
related photochemical oxidants (final report, Apr 2020) [EPA Report]. (EPA/600/R-20/012). Washington,
DC. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=348522
U.S. EPA (U.S. Environmental Protection Agency). (2020b). Smoke Sense study: A citizen science project using
a mobile app. Available online at https://www.epa.gov/air-research/smoke-sense-studv-citizen-science-
proiect-using-mobile-app (accessed February 9, 2021).
U.S. EPA (U.S. Environmental Protection Agency). (2020c). Wildfires and indoor air quality (IAQ). Available
online at https://www.epa.gov/indoor-air-aualitv-iaa/wildfires-and-indoor-air-qualitv-iaa (accessed February
9, 2021).
6-40
DRAFT: Do Not Cite or Quote
-------
Ulerv. AL: Graham. RC: Amrhein. C. (1993). Wood-ash composition and soil-pH following intense burning.
Soil Sci 156: 358-364. http://dx.doi.org/10.1097/00010694-199311000-000Q8
Van Deventer. D: Marecaux. J: Doubledav. A: Errett. N: Isaksen. TMB. (In Press) Wildfire smoke risk
communication efficacy: A content analysis of Washington state's 2018 statewide smoke event public health
messaging. J Public Health Manag Pract. http://dx.doi.org/10.1097/PHH.0000000000001151
Verma. S: Javakumar. S. (2012). Impact of forest fire on physical, chemical and biological properties of soil: A
review [Review]. Proc Intern Acad Ecol Env Sci 2: 168-176.
Wettstein. ZS: Hoshiko. S: Fahimi. J: Harrison. RJ: Cascio. WE: Rappold. AG. (2018). Cardiovascular and
cerebrovascular emergency department visits associated with wildfire smoke exposure in California in 2015.
J Am Heart Assoc 7: e007492. http://dx.doi.org/10.1161/JAHA. 117.007492
Wohlgemuth. H: Mittelstrass. K: Kschieschan. S: Bender. J: Weigel. HJ: Overmver. K: Kangasiarvi. J:
Sandermann. H: Langebartels. C. (2002). Activation of an oxidative burst is a general feature of sensitive
plants exposed to the air pollutant ozone. Plant Cell Environ 25: 717-726. http://dx.doi.org/10.1046/i. 1365-
3040.2002.00859.x
Xi. Y: Kshirsagar. AY: Wade. TJ: Richardson. DB: Brookhart. MA: Wvatt. L: Rappold. AG. (2020). Mortality
in US hemodialysis patients following exposure to wildfire smoke. J Am Soc Nephrol 31: 1824-1835.
http://dx.doi.org/10.1681/ASN.2019101066
Xu. R: Yu. P: Abramson. MJ: Johnston. FH: Samet. JM: Bell. ML: Haines. A: Ebi. KL: Li. S: Guo. Y. (2020).
Wildfires, global climate change, and human health. N Engl J Med 383: 2173-2181.
http://dx.doi.org/10.1056/NEJMsr2Q28985
Yamamoto. N: Shendell. D: Winer. A: Zhang. J. (2010). Residential air exchange rates in three major US
metropolitan areas: Results from the Relationship Among Indoor, Outdoor, and Personal Air Study 1999-
2001. Indoor Air 20: 85-90. http://dx.doi.org/10.1111/i. 1600-0668.2009.00622.X
Zu. K: Tao. G: Long. C: Goodman. J: Valberg. P. (2016). Long-range fine particulate matter from the 2002
Quebec forest fires and daily mortality in Greater Boston and New York City. Air Qual Atmos Health 9: 213-
221. http://dx.doi.org/10.1007/sll869-015-Q332-9
6-41
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
CHAPTER 7 ECOLOGICAL, WELFARE, AND
OTHER DIRECT DAMAGES OF FIRE
AND SMOKE
7.1 Introduction
The primary focus of this assessment is a quantitative analysis of the smoke impacts, both air
quality and health, from wildland fires. As detailed in the conceptual framework outlined in CHAPTER 2.
in the process of examining the trade-offs between prescribed fire and wildfire it is also important to
consider the potential effects, both positive and negative, of the fire itself. While in this assessment it is
not possible to quantify these effects due to a dearth of location specific data, the qualitative
characterization of these additional effects helps add context to the overall examination of the trade-offs
of smoke impacts due to different fire management strategies.
This chapter discusses the direct fire damages (value of economic loss) that are often experienced
as a result of wildland fire. As detailed in CHAPTER 6 and quantitatively examined in CHAPTER 8. the
health effects and overall population impacts of smoke exposure are well characterized. However, in
addition to the potential health hazards of smoke to the general population, fire fighters are also subjected
to smoke exposure and other hazards in the process of trying to control and suppress a wildland fire
(Section 7.2).
While there are ecological benefits to fire (see CHAPTER 3). severe wildfires can adversely
impact ecosystems and lead to significant effects on public welfare and incur societal costs as listed in
Table 7-1, and discussed below. In considering the costs incurred due to wildfires, preparedness,
mitigation, and suppression efforts are included, along with numerous losses that have significant impacts
on society. Section 7.3. provides a broad discussion of these additional effects often experienced due to
wildfires.
7.2 Wildland Firefighter Exposure to Smoke during Prescribed
Fires and Wildfires
This section is a brief summary of the inhalation health hazards and management implications to
wildland firefighters exposed to smoke pollutants at wildfires and prescribed fires. The discussion focuses
on exposures to smoke from the combustion of natural fuels (with mention of soil dust) but does not
consider smoke exposures from the burning of man-made products encountered by structural and
wildland firefighters at wildland-urban interface (WUI) fires, or airborne hazards resulting from fires
burning across polluted soils.
7-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Similar to the population as a whole (see CHAPTER 6). smoke is both a short-term acute
irritation and a long-term chronic health hazard. In the past, firefighters believed smoke was only an
inconvenience, irritating the eyes and nose, causing coughing, and occasionally causing nausea and
headaches. Many of the exposure limits are established to prevent acute health effects. However, there is
evidence there may be serious chronic health effects, and potentially even a reduced life span from
long-term exposure to wildland fire smoke (Navarro et al.. 2019; Booze et al.. 2004).
Because most wildland firefighters are deployed to both wildfires and prescribed fires during
their career, an interesting question arises: are wildland firefighters exposed to more smoke, or more
hazardous smoke at wildfires or prescribed fires? This section will address this issue and discuss
management implications.
7.2.1 Health Hazards of Exposure to Smoke
Wildland fuels are composed of living and dead vegetation, and the burning of this fuel produces
smoke. In a complete combustion environment, fuels are consumed by fire and converted mostly to
carbon dioxide (CO2) and water vapor (H2O) with the release of heat. However, the combustion process
in wildland fires is never complete, and incomplete combustion produces dozens of chemicals and
hundreds of trace chemicals [Naeher et al. (2007); Reinhardt et al. (2000); Reinhardt and Ottmar (2000);
Sandberg and Dost (1990); see CHAPTER 4 and CHAPTER 51. Some of the combustion products may
present acute health hazards, others may present chronic health hazards, and some can be both. The main
inhalation hazards for wildland firefighters and other personnel at fire camp are carbon monoxide (CO)
and respiratory irritants such as particulate matter (PM) and several key gases: acrolein, formaldehyde,
and to a lesser extent nitrogen dioxide (NO2) and sulfur dioxide (SO2). Smoke also includes low
concentrations of many other potentially toxic, carcinogenic components such as polycyclic aromatic
hydrocarbons (PAHs), and although there is extensive scientific evidence indicating a relationship
between long-term exposure to ambient fine particulate matter (PM2 5; particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 |im)and lung cancer, the cancer risk of PM2 5 derived
from wildland fires remains unclear (U.S. EPA. 2019). Evidence to date indicates a PM occupational
exposure limit is likely to be lower than the Occupational Safety and Health Administration (OSHA)
standard for respirable nuisance dust (Kim et al.. 2018). In addition to PM, wildland firefighters must also
be protected against exposure to airborne soil dust, which can result in hazardous exposures to respirable
crystalline silica.
7.2.2 Smoke Exposure at U.S. Prescribed Fires versus Wildfires
A relatively small number of studies have examined acute health effects to firefighters from
smoke exposure during prescribed fire and wildfires across individual work shifts and entire fire seasons.
7-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Although these studies indicate declines in individual lung function, a general conclusion from these
studies is that smoke exposure does not exceed occupational exposure standards most of the time for both
fire types (Adctona et al.. 2016; Reinhardt et al.. 2000; Reinhardt and Ottmar. 2000). However, when
there is an exceedance, it is often related to the job assignment and duration of that assignment rather than
the type of fire. For example, it has been shown that direct attack, line holding, and extensive mop-up can
lead to high smoke exposures (Domitrovich et al.. 2017; Reinhardt and Ottmar. 2004; Reinhardt et al..
2000; Reinhardt and Ottmar. 2000). The job assigned to a wildland firefighter and the length of time the
individual is carrying out that task will often be the overriding determinant of exceeding occupational
exposure limits rather than the fire type.
Most of these studies however, only collected data from individuals during their work shift and
did not consider smoke exposures outside their work period, allowing for a potential misinterpretation of
the results. Ongoing research by National Institute for Occupational Safety and Health (NIOSH) is
looking into wildland firefighter smoke exposure effects beyond their work shift, considering exposure
during their off hours during a work assignment and extending the assessment of health effects to a
season and career worth of smoke exposure attributable to wildland fire incidents.
7.2.2.1 Daily Exposure
Wildland firefighter work shifts average approximately 12 hours with 7 hours on the fire line if
assigned to a prescribed fire (Reinhardt et al.. 2000). During the work shift, firefighters have the potential
to be exposed to smoke concentrations that are similar to wildfires and the exposures will depend on job
assignment and duration. However, the firefighter often will return to a clean air environment at their duty
station until the next work shift begins, reducing the 24-hour average exposure level. In contrast,
firefighters assigned to long-duration project wildfires average 14-hour work shifts with 10 hours on the
fire line (Reinhardt and Ottmar. 2000). The increase in total work shift hours and longer assignments on
the fire line increase the duration of exposure to smoke as compared to prescribed fires. An additional
concern is the potential for continued exposure after a firefighter returns to a dusty, smoke-filled fire
camp following their work shift, if poorly-sited fire camps are affected by smoke and inversion
conditions, increasing the 24-hour exposure. For example, if the Air Quality Index (AQI) during off-duty
exceeds 100 (i.e., orange: unhealthy for sensitive groups) due to PM in the fire camp, this can result in
firefighters experiencing continuous exposure to high PM concentrations. As a result, the constant
exposure to higher PM concentrations could result in greater long-term health consequences when
compared to the same individual deployed to a prescribed fire where the duration and concentration of
exposure is less. This could have greater long-term health consequences when compared to the same
individual deployed to a prescribed fire.
7-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
7.2.2.2 Career Exposure
Exposure to smoke over the career of an individual will depend on the number and duration of
assignments to both wildfire and prescribed fire incidents. Type 1 crews (generally with the most
experience, leadership, and availability) will generally be exposed to the most smoke because their
primary job is firefighting, and the majority of their work shifts will occur on both wildfires and
prescribed fires. Type 2 and type 3 fire crews (generally with less experience, leadership, and availability
than type 1 crews) are believed to have fewer overall wildfire and prescribed fire assignments resulting in
less overall exposure throughout their career (Navarro ct al.. 2019). Exposure limits to prevent chronic
health effects from career-long exposure patterns have yet to be established for PM exposure from
wildfire smoke.
7.2.3 Management Implications
Evidence confirms that wildland firefighters are exposed to a variety of pollutants and respirable
crystalline silica at levels that can exceed recommended exposure limits during deployment at both
wildfires and prescribed fires. It is common for short-term exposure (usually 15 minutes, where irritation,
chronic or irreversible tissue damage does not occur) or maximum exposure limits to be exceeded during
brief but intense exposures to smoke at both fire types (Henn et al.. 2019). This is often related to job
assignment and other associated factors such as the site fuel model, wind orientation (downwind being
higher), crew type, relative humidity, type of attack, and wind speed. The resulting acute or short-term
effects such as eye or respiratory irritation require management intervention to reduce the exposure.
Recent National Wildfire Coordinating Group guidance on smoke exposure during wildland firefighting
recommended a reduction in the acceptable exposure limit on a shift-average basis, and this may be
adjusted further as ongoing research is completed.
Smoke exposure, whether due to a prescribed fire or wildfire is both a health and safety issue for
firefighters, prescribed fire training classes, and annual refresher courses. A range of literature is available
to better understand the potential acute and chronic effects that may result from exceeding smoke
exposure limits, and how best to manage, limit exposure, and inform crew personnel (Sharkey. 1997).
7.3 Economic Burden of Wildfire
NIST Special Publication 1215 (Thomas et al.. 2017) quantified the burden on the U.S. economy
from wildfires. The economic burden includes wildfire induced damages and losses, and also the
management costs to suppress and mitigate ignition and fire spread (see Table 7-1). The annualized
burden was estimated to be between $71.1 billion to $347.8 billion in 2016 dollars ($77.4 billion to
$378.7 billion in 2020 dollars). The estimates were based on literature or data available in early- to
7-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
mid-2017. Not included, for example, were recent catastrophic wildfire incidents. [Note, however, the
estimates in Thomas et al. (2017) were significantly larger than the previous estimates found in NIST
Special Publication 1130 Hamins et al. (2012)1.
Based on NOAA billion-dollar weather and climate disaster data (Smith. 2020). which include
direct losses from insured and uninsured sources, since 1980 the largest losses from billion-dollar wildfire
disasters have all come since 2017 (Figure 7-1, note: there were no billion-dollar events prior to 1991).
Since 1980, no year experienced more than a single billion-dollar wildfire disaster (direct losses from a
single-event), meaning each year represents a single event in Figure 7-1. Accounting for more than just
direct losses, Wang et al. (2020) measured the impacts from the 17 largest wildfires in California during
2018 and estimated their direct, indirect, and health costs. They estimated the wildfires to have caused
$148.5 billion ($126.1 billion to $192.9 billion 95% confidence interval) in losses associated with direct
capital losses ($27.7 billion), health effects ($32.2 billion), and indirect economic effects [$88.6 billion;
Wang etal. (2020)1.
7-5
DRAFT: Do Not Cite or Quote
-------
Table 7-1 The economic burden of wildland fires.
Costs
Losses
Prevention
Direct
• Education and training
•
Deaths and injuries
• Detection
•
Psychological impacts
• Enforcement
•
Structure and infrastructure loss
• Equipment
•
Environmental impact
Mitigation
•
Habitat and wildlife loss
• Fuels management
•
Timber loss
• Insurance
•
Agricultural loss
• Disaster resilience
Indirect
Suppression
•
General economic impacts
• Federal
•
Evacuation costs
• State
•
Accelerated economic decline
• Municipal (paid)
•
Utility and pipeline interruption
• Rural (volunteer)
•
Transportation interruption
Cross-cutting
•
Government service interruption
• Legal
•
Psychological impacts (loss of amenities)
• R&D
•
Housing market impact
• Building codes and standards
•
Loss of ecosystem service
• Regulations
•
Increase risk of other hazards
•
Loss of tax base
•
Health impacts from fire retardant use
7-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
20
25
20
10
Source: Smith (2020). copyright permission pending.
Figure 7-1 Billion-dollar wildfire event losses (1980-2020).
It appears the economic burden from wildfire has been increasing over time. While the wildfires
of the last few years have been particularly devastating, the increasing ability in measurement science to
better account for wildfire impacts can also partly explain the increase in reported costs and losses. In
particular is the recognition of human-health impacts from wildfire smoke as economic loss that has been
underappreciated until recently.
The next section discusses economic issues related to wildfire management, followed by a section
on management costs, and then a section covering economic issues related to valuing wildfire net value
change (NYC).
7.3.1 Economics of Wildfire: Management Implications
Economics is a discipline concerned with the allocation of scare resources and the understanding
of trade-offs. Central to the economics of wildfire management is the search for the understanding of
7-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
trade-offs between management inputs (e.g., prevention and suppression) and the consequences of
unwanted wildfire ignitions (e.g., life-safety, acres burned, structure loss). The economics of wildfire
management is not a new concept. In 1916, Headlev (1916) discussed ideas of suppression effectiveness,
efficiency, and waste of effort. Sparhawk (1925) introduced the idea of the "Cost plus Loss" (C+L) model
as the management trade-off between prevention and prefire suppression expenditures, suppression
expenditures, and wildfire losses. A central finding of the C+L model is that prevention and prefire
suppression expenditures can be selected to minimize the sum of all costs (i.e., prevention, prefire
suppression, and suppression spending) plus the resulting wildfire losses to identify the optimal level of
management effort. The optimal level corresponds with the C+L minimum and it can be shown that at the
minimum, any other allocation of management resources will either result in (1) an increase in spending
that exceeds the expected avoided loss or (2) a reduction in spending that surpasses an increase in
expected loss. This concept of the C+L model is depicted in C+L = cost plus loss.
Figure 7-2, where the inputs of presuppression and suppression are independent inputs, and
presuppression expenditures are held constant (Donovan and Rideout. 2003; Sparhawk. 1925).
The C+L model has been revised several times [e.g., Gorte (2013); Gorte and Gorte (1979)1. with
modern depictions acknowledging the potential for positive impacts of wildfires, necessitating a change in
the term "loss" to NVC; (Rideout and Omi. 1990; Simard. 1976). While the graphical depiction of the
C+NVC is useful for illustration, it is less useful for identifying the minimum C+NVC when
presuppression expenditures are allowed to be unconstrained. Further, because management activities and
recent wildfire activity can have lasting impacts on the fuels, affecting future wildfire risk (Prestemon et
al.. 2002). intertemporal optimization is required. Intertemporal optimization introduces additional
considerations such as discounting and risk perception, which affect the optimal timing of forest
management activities (Mercer et al.. 2007; Amacher et al.. 2005a. b).
Two immediate challenges exist that make the identification of the optimal levels of intervention
difficult to determine. First, an understanding of the functional relationship between wildfire management
activities and the resulting NVC is needed. Second, and perhaps more fundamental, is that many of the
impacts from wildfire are not well known or measured, particularly indirect or cascading impacts.
However, additional challenges include (1) the costs and losses are not incurred by the same subsets of
the population, creating equity concerns and barriers to aligning economic interests and (2) the spatial,
temporal, and economic boundaries of the C+L loss model are hard to define.
7-8
DRAFT: Do Not Cite or Quote
-------
C+L = cost plus loss.
Figure 7-2 Illustrative example of the Cost plus Loss (C+L) Model of wildfire
management.
7.3.2 Management Cost Categories
1 Management cost categories include those expenditures spent on preparing for, mitigating,
2 suppressing, and recovering from wildfires. Presuppression activities include prevention and
3 preparedness. Suppression accounts for firefighter labor, equipment, firefighter training and wellness
4 programs, as well as the monetary equivalence of volunteer time from local, nonpaid fire departments.
5 Post-fire rehabilitation and recovery includes efforts to return lands to prefire functionality. The
6 "cross-cutting" cost category includes activities that impact multiple management activities: for
7 example, research and development efforts result in more effective suppression technologies, improved
8 building codes, and fire-resistant building products.
7-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
7.3.2.1 Preparedness and Prevention
At the federal level, prevention and mitigation activities, including wildfire detection and
education, are aggregated together in budget line items as "preparedness." Preparedness is considered to
be "comprise [d] [of] a range of tasks to ensure readiness for wildfire response, including workforce
preparation, equipment and resource management, and wildfire outlook conditions for forecasting"
(Hoover. 2020). For FY2020, preparedness spending was $1,672 billion dollars in total for the U.S. Forest
Service (80%) and the Department of Interior [20%; Hoover (2020)1.
Wildfire prevention activities include awareness efforts to promote fire safety to reduce
unintentional wildfire ignitions. Awareness programs, such as public service announcements and media
spots, community townhall-style presentations by wildfire prevention specialists, distribution of
brochures and flyers containing educational messaging, and community wildfire hazard assessment
performed by risk specialists have all been shown to reduce the numbers of human-caused unintentional
wildfire starts and generate positive economic return on investment (Prestation et al.. 2010). For example,
Prestemon et al. (2010) estimated that the benefit-cost ratio of prevention to be 35 to 1 on the margin. Abt
et al. (2015). who also accounted for law enforcement efforts and intentionally-set wildfires, found
benefits were 5 to 38 times larger than prevention costs. Prevention efforts have been shown to have
differential effects that vary by ignition cause type [e.g., escaped campfire, debris fire; Butrv and
Prestemon (2019); Abt et al. (2015)1 and the timing of activities can be exploited to yield larger economic
benefits (Butrv et al.. 2010b; Butrv et al.. 2010a) or coupled with other risk reduction activities, such as
fuels management (Butrv et al.. 2010b).
Early wildfire warning and detection systems, including aerial and satellite technologies, can lead
to improved firefighting response time, limiting fire growth after ignition or assist in monitoring wildfire
progression, and increase suppression effectiveness (Cardil et al.. 2019). Satellite-based wildfire detection
information has been shown to improve fire commanders' decision making during suppression activities,
yielding better firefighting safety and economic outcomes (Herr et al.. 2020). Steele and Stier (1998)
found that wildfire surveillance from fixed lookouts yielded benefit-cost ratios of 6 to 1 in terms of
reduced suppression costs and property losses.
Wildfire risk assessments and related tools can be used to identify occurrences of elevated
temporal or spatial (landscape-level) risks, by factors such as prior wildfire history, weather, climate, fuel
conditions, and other socioeconomic factors. Such information can be used to inform decisions on the
prepositioning of mitigation and suppression resources (Bavham et al.. 2020; Thomas etal.. 2011;
Prestemon and Butrv. 2005). Improved suppression response time can yield economic benefits by
reducing burned areas (Cardil et al.. 2019).
7-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
7.3.2.2 Mitigation
Mitigation activities are designed to reduce the consequences from wildfire (e.g., area burned,
value of economic loss). For wildfires, the primary mitigation approaches are fuels management,
insurance, and disaster assistance.
7.3.2.2.1 Fuels Management
Fuels management activities result in the reduction of hazardous fuels in forests. The reduction of
fuels can be accomplished by a number of methods, including prescribed burning and mechanical and
chemical thinning of materials (as discussed in CHAPTER 3). In FY2020, the federal government spent
$194.0 million on the line item "hazardous fuels/fuels management" on federal lands and the line item
"other Forest Service (FS) wildfire appropriations," which also includes fuels management that amounted
to $545.3 million (Hoover. 2020). Fuels management spending is not readily available at the state, local,
and private levels, nationally.
There is statistical evidence that fuel treatments can impact wildfire behavior (Mercer et al.. 2007;
Prestemon et al.. 2002). resulting in suppression cost savings in excess of treatment costs (Thompson et
al.. 2017; Taylor et al.. 2013; Butrv. 2009). Research into optimization has shown that with careful
planning, fuel treatments can be leveraged to yield larger economics returns, when considering timing
(Butrv et al.. 2010a) or when allowing for the sale of harvested materials after forest thinning (Prestemon
et al.. 2012). Beyond avoided suppression costs, Huang et al. (2013) identified additional benefits,
including fatalities avoided, timber loss avoided, regional economic impacts, rehabilitation costs avoided,
and carbon storage implications. In addition, Houtman et al. (2013) considered the impact of "free" fuel
treatments (i.e., wildfire that are allowed to continue to burn to achieve multiple objectives which can
include resource benefits) on future suppression costs avoided and found instances of large economic
returns. However, policies allowing for more wildfires to burn (wildland fire use) may be more
economically favorable with a low or zero discount rate. Furthermore, wildland fire use is controversial
and carries inherent risk. Current federal fire management policy, for example, allows for limited
wildland fire use (i.e., as long as the managers determine that it would not endanger the public). To
increase the amount of wildland fire use, the risk thresholds would need to be relaxed, potentially
resulting in more unintended losses of people, structures, and resources [see Houtman (2011)1.
Fuel modification also occurs on private land, often as part of a program to create an area around
a structure designed to reduce wildfire ignition and spread (i.e., "defensible space"). The major barriers to
use of defensible-space programs are related to cost, aesthetics, and privacy (Absher et al.. 2013; Kyle et
al.. 2010; Absher et al.. 2009). For some, climate change and risk perceptions have mediated some of the
resistance (Wolters et al.. 2017). while for others it is a familiarity with the programs and expectations of
its effectiveness that have led to acceptance. Stockmann et al. (2010) evaluated the cost-effectiveness of
various homeowner risk reduction strategies including fuels management and structure hardening. They
7-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
found that fuel reduction within 61m (200 ft) of the house was the most cost-effective. Nevertheless,
homeowner actions to reduce wildfire risk are potentially limited by the homeowners' own inaccurate
assessment of risk factors [e.g., Champ et al. (2009)1.
7.3.2.2.2
Insurance
In measuring the U.S. fire problem, the cost of insurance has typically been calculated as the
difference between premiums paid in and claims paid out (Hall. 2014). which constitutes overhead costs.
These costs would include employees' wages, underwriting expenses, administrative expenses, taxes,
real-estate expenses, legal expenses, and cost of capital. There are a number of insurance markets that are
exposed to wildland fire, including homeowner's insurance, commercial insurance, automobile insurance
(Hall. 2014). health and life insurance. Frequently, wildfire losses are reported as direct, insured losses.
Although insurance could be part of the solution to increased efforts to reduce overall risk to
wildland fire on private lands, very few firms offer insurance focused in particular on forests (Chen et al..
2014). A leading limiting factor to widespread adoption of such insurance is a lack of actuarial
information on wildfire risk at fine spatiotemporal scales. There is additionally a need to develop a better
understanding of the approaches for reducing moral hazard and adverse selection in the issuance of
policies. As a result, policies tend to be expensive and out of reach of small forestland owners, meaning
that an insurance-based incentive structure for reducing overall wildfire risks on private lands remains
elusive.
Disaster assistance is financial assistance provided by the federal government following a disaster
declaration. Because assistance can be used for things such as temporary housing, lodging expenses,
repair, replacement, housing construction, child-care, medical expenses, household items, clean-up, fuel,
vehicles, moving expenses, and other necessary expenses determined by the Federal Emergency
Management Agency (FEMA), care needs to be taken in tracking the economic burden of wildfires
because counting these costs or reimbursements directly and also as disaster assistance may result in
double counting.
In FY2020, at the federal level, suppression spending exceeded $1.4 billion dollars, split between
the U.S. Forest Service (73%) and the Department of Interior [27%; Hoover (2020)1. State suppression
expenditures are estimated at $1 billion to $2 billion a year (Gortc. 2013).
7.3.2.2.3
Disaster Assistance
7.3.2.3 Suppression
7-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
An estimate for local (municipal) fire departments is more difficult to determine. An
approximation can be calculated assuming the cost of wildfire prevention and suppression is proportional
to the incident volume of fire involving wildland fuels. In 2014, based on Zhuang et al. (2017). it is
estimated that career fire department expenditures amounted to $41.9 billion ($46.21 billion in 2020
dollars), and the value of volunteer (rural) fire departments is estimated at $46.9 billion [see "method 5"
used in Zhuang et al. (2017); $51.72 billion]. Based on reported call volume (27.8 million calls) reported
to the National Fire Incident Reporting System (NFIRS) data from 2018, fires involving natural
vegetation represented 0.8% of all calls (20% of all fire incidents). In combination with fire department
expenditures, this information could be used to estimate the amount spent to suppress wildland fires in
local jurisdictions.
Gebert et al. (2007) found suppression spending to be impacted by burned area, suppression
strategy, and region of the country. Statistical models developed to forecast U.S. Forest Service
suppression costs by region of the country show that forecasted suppression spending is influenced by
factors such as prior suppression expenditures, sea surface temperatures, and weather [e.g., temperature
and precipitation; Gebert and Black (2012); Abt et al. (2009)1 found that suppression strategy influences
total suppression costs for large wildfires, with direct suppression being the most expensive on a per acre
burned and per day basis but leads to smaller wildfire sizes and duration. However, studies have found
that overall suppression strategy can be complicated by other factors, which also impact total suppression
expenditures. For example, Liang et al. (2008) found that the percentage of private land within the burned
area influenced suppression expenditures on large wildfires, while Rossi and Kuusela (2020) indicated
that management risk attitudes (risk aversion) impact expenditures.
7.3.2.4 Post-Fire Rehabilitation and Recovery
Post-fire rehabilitation is funded at the federal level as part of "other activities," and in FY2020
the other activities amounted to $41.9 million. Also included in this line item are activities related to
research and development, construction and maintenance of fire facilities, and forest health management
(Hoover. 2020).
7.3.2.5 Cross-Cutting Cost Categories
There are several costs that cut across various organizations and categories. These include legal
costs, research, and regulations. Legal costs include the prosecution, defense, and incarceration of
fire-setters. In 2019, there were 785,500 prisoners in local prisons (Zeng and Minton. 2021) In 2019,
there were 1,430,805 prisoners in federal and state facilities, with 0.9% sentenced for "other" property
crimes, which include arson [all types; Carson (2020)1. The Bureau of Prisons (2018) estimated that the
7-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
average cost of incarceration for a federal inmate in fiscal year 2016 was $36,299.25 ($39,566.18 in 2020
dollars).
Many public and nonprofit organizations are involved in research and development to reduce the
costs and losses associated with wildland fires. For federal research and science agencies, some of these
costs are included in the $41.9 million "other activities" listed above (Hoover. 2020).
Each state has its own building codes and fire regulations, based on the international model
codes. In addition, some consumer products are built for fire safety. Zhuang et al. (2017) estimated in
2014 that fire-safety related costs for building construction were $57.4 billion ($63.30 billion in 2020
dollars) and for consumer products were $54.0 billion ($59.55 billion in 2020 dollars). This includes fire
safety from all ignition and risk sources. In a study comparing the construction costs of atypical house
with a "wildfire-resistant" house, Quarles and Pohl (2018) found that the costs components to total
slightly less expensive for the wildfire-resistant house ($79,230 vs. $81,140). The cost components
included the roof, exterior walls, deck, and landscaping. The largest savings were found for the exterior
walls, which more than offset increases to the other components.
7.3.3 Wildfire Loss Categories
Wildfire-induced losses are grouped into two categories: direct and indirect. Direct losses are
those that occur as a primary result of wildfire (e.g., structure loss), while indirect losses are those that
occur as a secondary, or cascading, result of wildfire (e.g., economic downturn due to business structure
loss). Indirect losses are often more difficult to quantify due to latency and many may only be realized
years after the wildfire.
7.3.3.1 Direct Losses
7.3.3.1.1 Fatalities and Injuries
The National Fire Protection Association (NFPA) reported 80 civilian (nonfire-service) fatalities
and 700 injuries in 2019 from fire incidents reported as "outside and other fires" (Ahrens and Evarts.
2020). The "outside and other fire" incident type includes wildland, grass, crop, timber, and rubbish fires.
The estimates are based on a survey to U.S. fire departments, meaning the fatalities and injuries would
tend to include those observed or reported immediately following the fire incident. Long term health
consequences made worse due to fire exposure, but not known until well after the incident, would not be
captured. In 2017, there were 10 firefighter deaths associated with wildland suppression activities (USFA.
2018). The Incident Management Situation Report system, which tracks data on wildfires in federal
7-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
jurisdictions, includes firefighter injuries. From 2003 to 2007, an average of 260 injuries per year were
reported (Britton. 2010).
7.3.3.1.2 Psychological Impacts
Studies from wildfires have found depression, post-traumatic stress disorder (PTSD), and other
anxiety disorders to have resulted from exposure to wildfire events. Estimates for civilian rates of PTSD
and other anxiety disorders after a disaster range from 30% (Cole. 2011) to 60% (Kuligow ski. 2017). with
effects sometimes taking years to manifest (Kuligowski. 2017). For first responders, rates of PTSD have
been estimated to occur in up to 20% of firefighters and paramedics (Rahman. 2016).
7.3.3.1.3 Structure and Infrastructure Loss
The National Interagency Coordination Center (NICC) reported 963 structures lost by wildfire in
2019, under the annual average of 2,593 (NICC. 2019). NICC reported 25,790 structures lost in 2018
(NICC. 2018) and 12,306 structures lost in 2017 (NICC. 2017). NICC does not provide dollar lost
estimates.
7.3.3.1.4 Environmental Impacts
Environmental impacts can take many forms, including impacts on vegetation, soil and erosion,
watershed, and carbon sequestration. Vegetation loss can create the need to reseed and regrow forest and
grasslands. Soil degradation can result in poor soil nutrients and vegetation growth. Both vegetation and
soil loss can result in erosion and increase the risk of mudslides (Ren et al.. 2011; Benda et al.. 2003).
Trees sequester carbon, but it can be released to the atmosphere if trees are burned. Wildfires can
decrease water quality through the introduction of carbon, metals, other containments, and changes to
nutrients, which can affect aquatic ecosystems and drinking water (Rhoades et al.. 2019b). In addition to
increased treatment costs for potable water, poor water quality can impact agricultural and industrial
operations (Bladon et al.. 2014). Treatment costs include the increased need for elimination of solids and
dissolved organic carbon in water impacted by discharge from burned forests and wildlands (Emelko et
al.. 2011). However, traditional water quality protection strategies would fail to recognize impacts,
requiring treatment, from wildfire (Emelko et al.. 2011)).
7.3.3.1.5 Timber and Agricultural Loss
Wildfires on lands managed for timber and agricultural purpose result in business losses. The
1998 Florida wildfires resulted in pine timber damage of between $300 to $ 500 million in 1998 dollars
7-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
($479 million to $798 million), which represented over half of the quantified costs and losses of the
wildfire event (Butrv et al.. 2001). The timber losses were from two effects: (1) value from the physical
loss of timber and (2) a price increase, due to scarcity, after all salvageable timber was sold. Prestemon et
al. (2006) evaluated salvage harvest scenarios following the 2000 Bitterroot wildfire and found similar
(direction of) impacts to consumers, owners of damaged stands, and owners of undamaged stands. They
demonstrate that the value of timber lost due to wildfire could be more than offset (in general welfare
effects) through salvage.
7.3.3.2 Indirect Losses
7.3.3.2.1 General Economic Impacts
Wildfires, and disasters in general, can have long lasting impacts on an economy. They can
include business interruption (temporary and permanent closures) and supply chain impacts. Supply chain
disruption can affect businesses and customers far removed from the wildfire threatened areas.
Butrv et al. (2001) found the 1998 Florida wildfires impacted the tourism and service sectors. In
an analysis of the 2002 Haymen Fire [Colorado; Kent et al. (2003)1 found the wildfire induced overall
employment growth of 0.5%, by creating shifts in the economy resulting in a decline in average wages by
3%. Focusing on employment and wage dynamics, Davis et al. (2014) examined the impact of the 2008
large wildfires in Trinity County, California. They found that employment in the natural resource sector
increased by 30%, while average wages fell by 19%; whereas wage growth was experienced in the other
sectors, again demonstrating disparate effects. Borgschulte et al. (In Press) found that wildfire smoke
impacts annual labor income and employment in the U.S. and estimates the economic loss to be four
times that from mortality ($83 billion in 2020 dollars).
Nielsen-Pincus et al. (2014) explored the economic impacts of large wildfires (fires where
suppression exceed $1.0 million) in the western U.S. states by economic sector. For counties with
populations under 250,000, they found sectors with employment increases included natural resources and
mining; trade, transportation, and utilities; information services; financial services; and federal
employment. Sectors that lost employment included construction, manufacturing, professional and
business services, education and health services, and leisure and hospitality services. For larger counties,
total employment was reduced after a large wildfire by 0.04%.
Loomis et al. (2001) found in a study of visitors to forests in Colorado that hikers and mountain
bikers responded with fewer visits in areas with crown fires, but the time since the fire also played a role.
Englin et al. (2008) and Englin et al. (2001) found the linkage to recreation demand is time dependent,
with recent wildfires correlated with increased visitation and older wildfires linked to fewer, with Englin
et al. (2001) also noting a rebound effect with the oldest wildfires. Hesseln et al. (2003) found crown and
7-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
prescribed fires reduced visitation, visitation to the recovery area, but consumer surplus differed between
hikers (increased) and mountain bikers (decreased) in New Mexico. In Montana, Hesseln et al. (2004)
found hikers decreased visitations due to crown fire, but increased visitations due to prescribed fire. They
found mountain bikers displayed the opposite pattern.
7.3.3.2.2 Evacuations
Evacuation costs include temporary lodging and travel to and from the impacted area. Kent et al.
(2003) found the Hayman Fire in Colorado resulted in other expenditures, which included evacuation, that
were estimated to be up to $14 million ($19.5 million in 2020 dollars). In addition to expenditures,
McCaffrey et al. (2015) mentioned the nonmonetary expenditures, including the "logistical" and
"emotional" toll of fire evacuation.
7.3.3.2.3 Lost Natural Amenities
National forests provide a stream of values including historic, use and recreational, and existence
(value someone places on knowing something exists whether or not they may ever visit or use). Some of
these values can be monetized in the form of entrance and use fees. The National Parks were estimated to
be worth $92 billion dollars [$100 billion in 2020 dollars; Haefele et al. (2016)1.
7.3.3.2.4 Housing Market
Hedonic analyses that relate home sales prices to nonmarket amenities and other property
attributes can detect the values of environmental goods and services not directly traded in markets.
Several studies have evaluated the effect of wildfire risk on home sales prices, with the expectation that
higher risk lowers sales prices, all else being equal. Loomis (2004) compared housing sale prices before
and after the 1996 Buffalo Creek Fire (Colorado) and found a price decline between 13 to 15% of
undamaged homes near the wildfire. Kim and Wells (2005). in a study of the greater Flagstaff area
(Arizona), found moderate crown canopy closure (40 to 69%) was shown as preferred by home buyers;
whereas high crown canopy closure (70% and higher), which posed a higher wildfire risk, was shown to
decrease sale prices.
Meldrum et al. (2015) explored whether wildfire risk perceptions of residents of homes in Ouray
County, in southwestern Colorado, aligned with professionals' data-based assessments of wildfire risk
based on features of the home and property, including whether the property had vegetation nearby.
Residents underestimated the risks of wildfire nearby. In many other aspects of the property's features,
residents' perceptions were generally not highly correlated with the assessments of the professionals. The
implication is that economic motivations to undertake risk reduction efforts would be lower than if risk
7-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
were more accurately quantified by residents. Donovan et al. (2007) compared housing sales prices before
and after homes were rated based on wildfire risk in Colorado Springs, CO. They found that the
availability of risk information was correlated with a decrease of a representative home sales value by
13.7%. Champ et al. (2009) explored whether home prices in Colorado Springs, CO were aligned with
risks of wildfire. They found that homebuyers prefer risky locations due to their favorable amenities
besides fire (e.g., topography) but that homebuyers were less cognizant of wildfire risks than objective
assessments would identify. Although these homebuyers preferred less fire-prone building materials, they
tended to undervalue features of their properties from the perspective of wildfire risk reduction.
Hierpe et al. (2016). in a study of house prices in four western cities, found that the sales of
homes with medium forest density (34 to 66%) within 100 m of a house was associated with lower sales
prices; yet, homes with high forest density (67% and greater) within 500 m of a house was associated
with higher sales prices. Stetleretal. (2010) estimated home sales prices in Montana and found that
distance to the wildfire, time since, size of fire, and whether the home was within sight distance of the
wildfire affected home sales, for an average price loss of-13.7% for a home within 5 km of the fire.
Kalhoretal. (2018) evaluated the impact of visible fire scars from the 2000 Cerro Grande fire
(New Mexico) on assessed house values in 2013. They found impact of the previous damage equated to a
1.7 to 4.4% decline in assessed house value, while measures of future wildfire risk were found to be
correlated to an increase in assessed house value by 0.3 to 0.4%. The latter impact was attributed to the
crown area likely accounting for the aesthetic value of vegetation.
7.3.3.2.5 Loss of Ecosystem Services
Ecosystem services are generally defined as "any positive benefit that wildlife or ecosystems
provides to people" (NWF. 2017). Few studies exist on a national scale. Most tend to be regional in scope
and not specific to wildfire. For example, Loomis et al. (2000) evaluated the value of better watershed
services for a 45-mile section of the Platte River, Desvousges et al. (1983) valued lake preservation,
Moore and McCarl (1987) valued the preservation of Mono Lake ecosystem, and Hanemann et al. (1991)
valued increased salmon stock in the San Joaquin River. Such examples provide methods that could be
used to value avoided losses to ecosystem services from wildfire mitigation.
Wildfire Fire and Prescribed Fire Impacts on Forest Health and Wildlife
Studies in the ponderosa pine ecoregion of California, Oregon, and Washington have shown that
fire management based on low-intensity prescribed fire coupled with mechanical thinning can, over time,
approximate historical landscape conditions that are much less susceptible to catastrophic fires (Prichard
et al.. 2017a; Prichard et al.. 2017b; Allen et al.. 2002). Where it is feasible to use such practices,
low-severity fires can promote important wildlife habitat and forest health benefits (Pausas and Kcclcv.
2019). These ecological benefits include improvements in habitat quality for threatened and endangered
7-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
species (Pausas and Keelev. 2019); reductions in ground layer and understory "ladder" fuels; reduced
losses of forest floor nutrient capital and water holding capacity (Murphy et al.. 2006); and increased
forest resistance to drought, pests and diseases, all of which are being exacerbated by climate change
(Spies et al.. 2019; Yose et al.. 2019).
To date, prescribed low-intensity fire and thinning treatments have not been adopted into local,
state, and federal forest management practices at a scale necessary to affect the overall fire deficit, and
associated fuel load excess, in western forests. The potential impacts of ignoring the fire deficit is
underscored by the growing body of evidence for the role of climate change in amplifying recent
increases in the frequency and intensity of wildland fires (Kolden. 2019; Abatzoglou and Williams. 2016)
and consequent impacts on ecological benefits associated with low-intensity fire regimes.
Water Resources
Wildfire can both directly and indirectly affect water resources as well. Direct effects can occur
via downwind smoke and ash deposition on the surface of waterbodies (see Section 6.4). and damage to
drinking water infrastructure. Indirectly, fire affects water resources primarily through increased runoff of
water and other materials into nearby waterbodies. Together, these direct and indirect effects can alter the
physical, chemical, and biological characteristics of water resources, and by doing so, impact their end
use, such as for recreation, aquatic life, and drinking water.
The direct effects of fire on drinking water infrastructure is an area of rising concern. Fires can
damage water treatment facilities or water supply lines, for example. In two locations in California (Santa
Rosa and Paradise), benzene and other volatile organic compounds (VOCs) were detected in tapwater
post-fire, with concentrations of benzene exceeding federal and state drinking water standards (Proctor et
al.. 2020). This was likely caused by the partial melting of plastic water-supply lines to homes and
infiltration of hot gas and other materials when the supply system became depressurized (Proctor et al..
2020). As fires become more frequent, they are increasingly likely to burn into urbanized areas, and direct
effects on drinking water infrastructure could be become more common.
The indirect effects of fire are more widespread currently, including the indirect effects on
waterbodies used as drinking water sources. Fire-prone ecosystems are major sources of the national
water supply. Fire impacts on forested watersheds are particularly concerning as these watersheds provide
50% of the water consumed in the lower 48 states. Most of these watersheds are at high risk from wildfire
now or in the near future (Hallema et al.. 2018).
Fire can impact the physical supply and timing of water delivery by altering runoff and
streamflow. The loss of ground layer vegetation and canopy leaf biomass reduces evapotranspiration,
potentially resulting in pronounced increases in runoff and flood severity (Stevens. 2013). Moreover, on
some soil types, intense wildfires can dramatically increase runoff by increasing water repellency of
near-surface soil layers, a condition that can persist for years (Certini. 2005). Depending on fire severity,
7-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
rainfall patterns, and watershed soil and land cover characteristics, post-fire streamflow can increase in
the days, months and years following fire (Niemever et al.. 2020). Fire can also change the amount and
timing of snowmelt. For instance, mountain snowpack beneath charred forests absorbed more solar
energy, causing earlier melt and snow disappearance in >11% of forests in the western seasonal snow
zone over the past two decades (Gleason et al.. 2019). Fire and climate change impacts on snowpack can
also have significant impacts on late summer runoff when it is most needed by fish and wildlife (Pausas
and Keelev. 2019).
By increasing runoff and flow, fires can also increase erosion and delivery of sediments, ash, and
other constituents to downslope ecosystems. The increased sediment loads and land destabilization that
can occur post-fire (Ren et al.. 2011; Benda etal.. 2003) may be characterized by a large influx of
suspended solids to headwater streams, termed "slurry flows", up to 700,000 mg/L in magnitude (Rinne.
1996). A wide variety of chemical constituents are often mobilized along with the sediments and ash. This
includes nutrients and cations, heavy metals, organic compounds, like polycyclic aromatic hydrocarbons
(PAHs), and dissolved organic carbon (Smith et al.. 2011). Besides direct additions to water resources,
fire can indirectly increase disinfection byproducts (DBPs), compounds that form during drinking water
treatment when disinfectants (e.g., chlorine, chloramine) react with organic carbon and nitrogen
compounds present in higher concentrations post-fire (Bladon et al.. 2014). Some DBPs pose health risks,
with the potential to cause certain cancers, reproductive issues, and anemia.
Encroachment of wildfire into the wildland/urban interface can also release largely unknown
types and quantities of anthropogenic contaminants into streams. Combustion of houses, buildings,
vehicles, waste sites and other infrastructure present risks from hazardous chemicals, such as benzene and
VOCs, as well as heavy metals (Proctor et al.. 2020; Uzun et al.. 2020). Finally, the use of fire retardants
may also increase nutrient and chemical loading to post-fire landscapes.
Beyond physical and chemical changes, fires can also indirectly alter biological assemblages in
downstream waters. Fire can increase coarse woody debris in streams (Young. 1994). positively
impacting long-term habitat for fish, yet over the shorter term, fish and macroinvertebrate populations
typically decline post-fire [e.g., Rinne (1996)1. Concomitantly, burning in riparian areas can increase light
levels to streams, and studies have often recorded increases in stream temperatures post-fire [e.g.,
Dunham et al. (2007)1. This could negatively affect cold-water fish species, like salmonids (Beakes et al..
2014). Combined with the increased light and temperature, an influx of nutrients and sediment can also
promote harmful algal blooms and the production of cyanotoxins (Bladon et al.. 2014; Smith et al.. 2011).
These cyanotoxins both contaminate drinking water and negatively affect aquatic life.
While wildfire has been a part of the natural ecology of many ecosystems for millennia, the
increase in fire frequency, intensity and area burned can have deleterious effects on water resources,
altering their physical, chemical, and biological characteristics. In general, the more severe the fire, the
more likely downstream waters will be affected. Water quality impacts generally are most pronounced in
the first few years post-fire but may persist for more than a decade in some cases (Rhoades et al.. 2019a;
7-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Smith etal.. 2011). Increased concentrations of nutrients, heavy metals, organic compounds like benzene,
and DBPs pose particular risks, along with increased algal blooms and cyanotoxins. Communities will
need to be aware—and plan for—the potential for post-fire contamination of water resources. The
provisioning of safe drinking water from burned watersheds may require additional treatment
infrastructure and increased operations and maintenance costs to remediate effects.
7.3.3.2.6 Other
Other impacts of wildfires include accelerated economic decline, loss of utilities and
transportation systems, disruption to government services, interference with military operations
(e.g., smoke visibility issues), cascading natural hazard risks (e.g., increase risk of mudslide or growth of
invasive species), loss of tax base due to housing and building stock, and health and environmental
impacts from fire retardants. Many of these impacts are not well-defined or monetized.
7.3.4 Magnitudes, Gaps, and Uncertainty
Table 7-2 shows estimated magnitudes of value of the costs and losses and levels of uncertainty
in their measurement or ability to measure at a national scale [reproduced from Thomas et al. (2017)1.
The estimated magnitudes and uncertainties were based on the values found in the report, and where not
available, were estimated using expert judgment of the report authors. The largest cost and loss categories
were fuel treatments and defensible space, suppression, economic value of deaths and injuries, evacuation
costs, and housing market impacts. The largest sources of uncertainty tended to be indirect economic
effects, insurance, and some of the cross-cutting categories (e.g., building codes and standards,
regulations).
While there exists a significant literature detailing components of the costs and losses associated
with wildland fire, producing an annual national estimate, which could be tracked over time to evaluate
management success, is difficult at this time without introducing large sources of uncertainty in the
estimates. However, it does appear that the economic burden from wildland fire is increasing over time.
Table 7-2
Magnitude and uncertainty associated with the economic burden of
wildfire at the national level.
Order of Magnitude Uncertainty
Costs
Preparedness
$$$$ ?
7-21
DRAFT: Do Not Cite or Quote
-------
Table 7-2 (Continued): Magnitude and uncertainty associated with the economic
burden of wildfire at the national level.
Order of Magnitude
Uncertainty
Mitigation
Fuels management
Fuel treatments (Rx fire, thinning)
$$$
?
Defensible space/firewise
$$$$
???
Insurance
$$
????
Disaster assistance
$
??
Suppression
Fire departments (labor, equipment,
training)
Federal
$$$$
?
State
$$$$
?
Municipal (professional)
$$$$
???
Rural (volunteer)
$$$$
???
Cross-cutting
Legal
Prosecution
$$
??
Incarceration
$$$
??
Civil/liability
$$
????
Science/research and development
$$
???
Building codes and standards
$$
????
Regulations (e.g., zoning)
$$
????
Losses
Direct
Deaths and injuries (civilian and
firefighter)
$$$$
??
Psychological impacts (PTSD)
$$
???
Structure and infrastructure loss
$$$
???
7-22
DRAFT: Do Not Cite or Quote
-------
Table 7-2 (Continued): Magnitude and uncertainty associated with the economic
burden of wildfire at the national level.
Order of Magnitude
Uncertainty
Environmental impact
$$$
????
Habitat and wildlife loss
$$
????
Timber loss
$$$$
???
Agriculture loss
$$$
????
Remediation/cleanup
$$
???
Indirect
General Economic impacts (business
interruption, tourism, supply chain)
$$$
????
Evacuation costs
$$$$
???
Accelerated economic decline of
community
$$$
????
Utility and pipeline interruption
(electricity, gas, water, oil)
$$$
????
Transportation interruption (e.g., roads
and rail)
$$
????
Government service interruption
(including education)
$$
????
Psychological impacts (loss of natural
amenities)
$$
????
Housing market impact (loss due to fire
risk)
$$$$
???
Loss of ecosystem services (e.g.,
watershed/water service)
$$$
????
Increased risk of other hazards (e.g.,
mudslide, invasive species)
$$$
????
Decrease in tax base (structure loss or
decline in value of structure)
$$$
???
Decrease in government services
$$$
????
Health/environmental impacts from use
of fire retardants/suppressants
$$$
????
PTSD = post-traumatic stress disorder; Rx = prescribed.
Note: Classification of "order of magnitude": $ =
-------
7.4 References
Abatzoglou. JT: Williams. AP. (2016). Impact of anthropogenic climate change on wildfire across western US
forests. Proc Natl Acad Sci USA 113: 11770-11775. http://dx.doi.org/10.1073/pnas. 1607171113
Absher. JD: Vaske. JJ: Lyon. KM. (2013). Overcoming barriers to firewise actions by residents. Final report to
joint fire science program. (10-3-01-15). Joint Fire Science Program, http://www.firescience.gov/proiects/10-
3-01-15/proiect/10-3-01-15 final report.pdf
Absher. JD: Vaske. JJ: Shelby. LB. (2009). Residents' responses to wildland fire programs: A review of
cognitive and behavioral studies. In General Technical Report. (PSW-GTR-223). Albany, CA: Pacific
Southwest Research Station. http://dx.doi.org/10.2737/PSW-GTR-223
Abt. KL: Butrv. DT: Prestemon. JP: Scranton. S. (2015). Effect of fire prevention programs on accidental and
incendiary wildfires on tribal lands in the United States. International Journal of Wildland Fire 24: 749-762.
http://dx.doi.org/10.1071/WF14168
Abt. KL: Prestemon. JP: Gebert. KM. (2009). Wildfire suppression cost forecasts for the US forest service. J
Forest 107: 173-178.
Adetona. O: Reinhardt. TE: Domitrovich. J: Brovles. G: Adetona. AM: Kleinman. MT: Ottmar. RD: Naeher. LP.
(2016). Review of the health effects of wildland fire smoke on wildland firefighters and the public [Review].
Inhal Toxicol 28: 95-139. http://dx.doi.org/10.3109/08958378.2016.1145771
Ahrens. M: Evarts. B. (2020). Ahrens, M., and B. Evarts. 2020. "Fire loss in the United States During 2019."
National Fire Protection Association. https://www.nfpa.org//-/media/Files/News-and-Research/Fire-statistics-
and-reports/US-Fire-Problem/osFireLoss.pdf. In NFPA Research. National Fire Protection Association.
https://www.nfpa.org//-/media/Files/News-and-Research/Fire-statistics-and-reports/US-Fire-
Problem/osFireLoss.pdf
Allen. CD: Savage. M: Falk. DA: Suckling. KF: Swetnam. TW: Schulke. T: Stacev. PB: Morgan. P: Hoffman.
M: Klingel. JT. (2002). Ecological restoration of Southwestern ponderosa pine ecosystems: A broad
perspective [Review]. Ecol Appl 12: 1418-1433. http://dx.doi.org/10.189Q/1051-
0761(2002)012[1418:EROSPP12.0.CQ:2
Amacher. GS: Malik. AS: Haight. RG. (2005a). Forest landowner decisions and the value of information under
fire risk. Can J For Res 35: 2603-2615. http://dx.doi.org/10.1139/x05-143
Amacher. GS: Malik. AS: Haight. RG. (2005b). Not getting burned: The importance of fire prevention in forest
management. Land Econ 81: 284-302. http://dx.doi.Org/10.3368/le.81.2.284
Bavham. J: Belval. EJ: Thompson. MP: Dunn. C: Stonesifer. CS: Calkin. DE. (2020). Weather, risk, and
resource orders on large wildland fires in the western US. Forests 11: 169.
http://dx.doi.org/10.3390/fll020169
Beakes. MP: Moore. JW: Haves. SA: Sogard. SM. (2014). Wildfire and the effects of shifting stream
temperature on salmonids. Ecosphere 5: 63. http://dx.doi.Org/10.1890/ES13-00325.l
Benda. L: Miller. D: Bigelow. P: Andras. K. (2003). Effects of post-wildfire erosion on channel environments,
Boise River, Idaho. For Ecol Manage 178: 105-119. http://dx.doi.org/10.1016/S0378-l 127(03)00056-2
Bladoa KD: Emelko. MB: Silins. U: Stone. M. (2014). Wildfire and the future of water supply. Environ Sci
Techno148: 8936-8943. http://dx.doi.org/10.1021/es50Q130g
Booze. TF: Reinhardt. TE: Quiring. SJ: Ottmar. RD. (2004). A screening-level assessment of the health risks of
chronic smoke exposure for wildland firefighters. J Occup Environ Hyg 1: 296-305.
http://dx.doi.org/10.1080/1545962049044250Q
Borgschulte. M: Molitor. D: Zou. EY. (In Press) Air pollution and the labor market: Evidence from wildfire
smoke. Rev Econ Stat.
7-24
DRAFT: Do Not Cite or Quote
-------
Britton. CL. (2010) Risk factors for injury among federal wildland firefighters in the United States. (Doctoral
Dissertation). University of Iowa, Iowa City, IA. Retrieved from https://doi.org/10.17077/etd.iog4mln5
Bureau of Prisons (U.S. Department of Justice, Federal Bureau of Prisons). (2018). Annual determination of
average cost of incarceration. Fed Reg 83: 18863.
Butrv. DT. (2009). Fighting fire with fire: Estimating the efficacy of wildfire mitigation programs using
propensity scores. Environ Ecol Stat 16: 291-319. http://dx.doi.org/10.1007/sl0651-007-0Q83-3
Butrv. DT: Mercer. DE: Prestemon. JR: Pve. JM: Holmes. TP. (2001). What is the price of catastrophic wildfire?
J Forest 99: 9-17. http://dx.doi.org/10.1093/iof/99.11.9
Butrv. DT: Prestemon. J. (2019). Economics of WUI/Wildfire prevention and education. In SL Manzello (Ed.),
Encyclopedia of Wildfires and Wildland-Urban Interface (WUI) Fires. Cham, Switzerland: Springer.
http://dx.doi.org/10.1007/978-3-319-51727-8 105-1
Butrv. DT: Prestemon. JP: Abt. KL. (2010a). Optimal timing of wildfire prevention education. In G Perona; CA
Brebbia (Eds.), WIT Transactions on Ecology and the Environment (pp. 197-206). Ashurst, England: WIT
Press. http://dx.doi.org/10.2495/FIVA100181
Butrv. DT: Prestemon. JP: Abt. KL: Sutphen. R. (2010b). Economic optimisation of wildfire intervention
activities. International Journal of Wildland Fire 19: 659-672. http://dx.doi.org/10.1071/WF09090
Cardil. A: Lorente. M: Boucher. D: Boucher. J: Gauthier. S. (2019). Factors influencing fire suppression success
in the province of Quebec (Canada). Can J For Res 49: 531-542. http://dx.doi.org/10.1139/cifr-2018-0272
Carson. EA. (2020). Prisoners in 2019. (NCJ 255115). Washington, DC: U.S. Department of Justice, Office of
Justice Programs, Bureau of Justice Statistics, https://www.bis.gov/content/pub/pdf/pl9.pdf
Certini. G. (2005). Effects of fire on properties of forest soils: A review. Oecologia 143: 1-10.
http://dx.doi.org/10.1007/sQ0442-004-1788-8
Champ. PA: Donovan. GH: Barth. CM. (2009). Homebuyers and wildfire risk: A Colorado Springs case study.
Society and Natural Resources 23: 58-70. http://dx.doi.org/10.1080/089419208Q2179766
Chen. X: Goodwin. BK: Prestemon. JP. (2014). Is timber insurable? A study of wildfire risks in the U.S. forest
sector using spatio-temporal models. Am J Agric Econ 96: 213-231. http://dx.doi.org/10.1093/aiae/aat087
Cole. V. (2011). PTSD and natural disasters, Defense Centers of Excellence for Psychological Health and
Traumatic Brain Injury. Presentation presented at Defense Centers of Excellence for Psychological Health
and Traumatic Brain Injury (DCoE) Webinar, August 25, 2011, Virtual.
Davis. EJ: Moselev. C: Nielsen-Pincus. M: Jakes. PJ. (2014). The community economic impacts of large
wildfires: A case study from Trinity County, California. Society and Natural Resources 27: 983-993.
http://dx.doi.org/10.1080/08941920.2014.9Q5812
Desvousges. W: Smith. VK: McGivnev. M. (1983). A comparison of alternative approaches to estimating
recreation and related benefits of water quality improvements. (EPA 230 83 001). Washington, DC: US
Environmental Protection Agency.
https://nepis.epa.gOv/Exe/ZvNET.exe/910 lEUFI.txt?ZvActionD=ZvDocument&Client=EPA&Index=1981%
20Thru%201985&Docs=&Ouerv=&Time=&EndTime=&SearchMethod=l&TocRestrict=n&Toc=&TocEntr
v=&OField=&OFieldYear=&OFieldMonth=&OFieldDav=&UseOField=&IntOFieldOp=0&ExtOFieldOp=Q
&XmlOuerv=&File=D%3 A%5 CZYFILES%5 CINDEX%20D AT A%5 C81THRU85 %5 CTXT%5 C00000024
%5C9101EUFI.txt&User=ANONYMOUS&Password=anonvmous&SortMethod=h%7C-
&MaximumDocuments=l&FuzzvDegree=0&ImageOualitv=r75g8/r75g8/xl50vl50gl6/i425&Displav=hpfr
&DefSeekPage=x&SearchBack=ZvActionL&Back=ZvActionS&BackDesc=Results%20page&MaximumPa
ges=l&ZvEntrv=l#
Domitrovich. JW: Brovles. GA: Ottmar. RD: Reinhardt. TE: Naeher. LP: Kleinman. MT: Navarro. KM:
Mackav. CE: Adetona. O. (2017). Final report: Wildland fire smoke health effects on wildland firefighters
and the public. (JFSP project ID 13-1-02-14). Boise, ID: Joint Fire Science Program.
https://www.firescience.gov/proiects/13-l-02-14/proiect/13-l-02-14 final report.pdf
7-25
DRAFT: Do Not Cite or Quote
-------
Donovan. GH: Champ. PA: Butrv. DT. (2007). Wildfire risk and housing prices: A case study from Colorado
Springs. LandEcon83: 217-233.
Donovan. GH: Rideout. DB. (2003). A reformulation of the Cost plus Net Value Change (C plus NVC) model of
wildfire economics. Forest Sci 49: 318-323. http://dx.doi.Org/10.1093/forestscience/49.2.318
Dunham. JB: Rosenberger. AE: Luce. CH: Rieman. BE. (2007). Influences of wildfire and channel
reorganization on spatial and temporal variation in stream temperature and the distribution of fish and
amphibians. Ecosystems 10: 335-346. http://dx.doi.org/10.1007/sl0021-007-9Q29-8
Emelko. MB: Silins. U: Bladon. KD: Stone. M. (2011). Implications of land disturbance on drinking water
treatability in a changing climate: Demonstrating the need for "source water supply and protection"
strategies. Water Res 45: 461-472. http://dx.doi.Org/10.1016/i.watres.2010.08.051
Englia J: Holmes. TP: Lutz. J. (2008). Wildfire and the economic value of wilderness recreation. In TP Holmes;
JP Prestemon; KL Abt (Eds.), Economics of forest disturbances: Wildfires, storms and invasive species (pp.
191-208). Dordrecht, Netherlands: Springer, http://dx.doi.org/10.1007/978-1-4020-4370-3 10
Englia J: Loomis. J: Gonzalez-Caban. A. (2001). The dynamic path of recreational values following a forest
fire: A comparative analysis of states in the Intermountain West. Can J For Res 31: 1837-1844.
http://dx.doi.org/10.1139/cifr-31-10-1837
Gebert. KM: Black. AE. (2012). Effect of suppression strategies on federal wildland fire expenditures. J Forest
110: 65-73. http://dx.doi.org/10.5849/iof.10-068
Gebert. KM: Calkin. DE: Yoder. J. (2007). Estimating suppression expenditures for individual large wildland
fires. Western Journal of Applied Forestry 22: 188-196. http://dx.doi.org/10.1093/wiaf/22.3.188
Gleason. KE: McConnell. JR: Arienzo. MM: Chellman. N: Calvin. WM. (2019). Four-fold increase in solar
forcing on snow in western U.S. burned forests since 1999. Nat Commun 10: 2026.
http://dx.doi.org/10.1038/s41467-019-09935-v
Gorte. JK: Gorte. RW. (1979). Application of economic techniques to fire management—A status review and
evaluation. In General Technical Report. (INT-GTR-53). Ogden, UT: Department of Agriculture Forest
Service, https://www.fs.fed.us/rm/pubs int/int gtr053.pdf
Gorte. R. (2013). The rising cost of wildfire protection. Bozeman, MT: Headwaters Economics.
https://headwaterseconomics.org/wp-content/uploads/fire-costs-background-report.pdf
Haefele. M: Loomis. JB: Bilmes. L. (2016). Total economic valuation of the National Park Service lands and
programs: Results of a survey of the American public. In Harvard Kennedy School Working Papers. (16-
024). Elsevier, http://dx.doi.org/10.2139/ssrn.2821124
Hall. JR. (2014). The total cost of fire in the United States. (NFPA No. USS13). Quincy, MA: National Fire
Protection Association, https://www.maine.gov/dps/fmo/sites/maine.gov.dps.fmo/files/inline-
files/research/documents/nfpa totalcostfire 2014.pdf
Hallema. DW: Sun. G: Caldwell. PV: Norman. SP: Cohen. EC: Liu. Y: Bladon. KD: McNultv. SG. (2018).
Burned forests impact water supplies. Nat Commun 9: 1307. http://dx.doi.org/10.1038/s41467-018-03735-6
Hamins. A: Averill. J: Brvner. N: Gann. R: Butrv. D: Davis. R: Amon. F: Gilman. J: Maranghides. A: Mell. W:
Madrvkowski. D: Manzello. S: Yang. J: Bundv. M. (2012). Reducing the risk of fire in buildings and
communities: A strategic roadmap to guide and prioritize research (pp. 171). National Institute of Standards
and Technology.
Hanemann. M: Loomis. J: Kanninen. B. (1991). Statistical efficiency of double-bounded dichotomous choice
contingent valuation. Am J Agric Econ 73: 1255-1263. http://dx.doi.org/10.2307/1242453
Headlev. R. (1916). Fire suppression district 5. In Scientific Journal. US Department of Agriculture Forest
Service.
Henn. SA: Butler. C: Li. J: Sussell. A: Hale. C: Brovles. G: Reinhardt. T. (2019). Carbon monoxide exposures
among U.S. wildland firefighters by work, fire, and environmental characteristics and conditions. J Occup
EnvironHyg 16: 793-803. http://dx.doi.org/10.1080/1549624.2019.167Q833
7-26
DRAFT: Do Not Cite or Quote
-------
Herr. V: Kochanski. AK: Miller. VV: Mccrea. R: O'Brien. D: Mandel. J. (2020). A method for estimating the
socioeconomic impact of Earth observations in wildland fire suppression decisions. International Journal of
Wildland Fire 29: 282-293. http://dx.doi.org/10.1071AVF18237
Hesseln. H: Loomis. JB: Gonzalez-Caban. A. (2004). The effects of fire on recreation demand in Montana.
Western Journal of Applied Forestry 19: 47-53. http://dx.doi.org/10.1093/wiaf/19.1.47
Hesseln. H: Loomis. JB: Gonzalez-Caban. A: Alexander. S. (2003). Wildfire effects on hiking and biking
demand in New Mexico: A travel cost study. J Environ Manage 69: 359-368.
http://dx.doi.Org/10.1016/i.ienvman.2003.09.012
Hierpe. E: Kim. YS. u: Dunn. L. (2016). Forest density preferences of homebuyers in the wildland-urban
interface. Forest Pol Econ 70: 56-66. http://dx.doi.Org/10.1016/i.forpol.2016.05.012
Hoover. K. (2020). Federal wildfire management: Ten-year funding trends and issues. (R46583). Congressional
Research Service. http://dx.doi.Org/https://crsreports.congress.gov/product/pdf/R/R46583
Houtman. RM. (2011) Letting wildfires burn: Modeling the change in future suppression costs as the result of a
suppress versus a let-burn management choice. (Master's Thesis). Oregon State University, Corvallis, OR.
Retrieved from https://ir.librarv.oregonstate.edu/concern/graduate thesis or dissertations/z890rw891
Houtman. RM: Montgomery. CA: Gagnon. AR: Calkin. DE: Dietterich. TG: McGregor. S: Crowley. M. (2013).
Allowing a wildfire to burn: Estimating the effect on future fire suppression costs. International Journal of
Wildland Fire 22: 871-882. http://dx.doi.org/10.1071/WF12157
Huang. CH: Finkral. A: Sorensen. C: Kolb. T. (2013). Toward full economic valuation of forest fuels-reduction
treatments. J Environ Manage 130: 221-231. http://dx.doi.Org/10.1016/i.ienvman.2013.08.052
Kalhor. E: Horn. BP: Valentia V: Berrens. RP. (2018). Investigating the effects of both historical wildfire
damage and future wildfire risk on housing values. International Journal of Ecological Economics &
Statistics 39.
Kent. B: Gebert. K: McCaffrey. S: Martin. W: Calkin. D: Schuster. E: Martin. I: Wise Bender. H: Alward. G:
Kumagai. Y: Cohn. PJ: Carroll. M: Williams. D: Ekarius. C. (2003). Social and economic issues of the
Hayman Fire. In General Technical Report (pp. 315-396). (RMRS-GTR-114). US Department of Agriculture
Forest Service, https://www.fs.fed.us/rm/value/docs/hayman fire social%20and%20economic%20issues.pdf
Kim. YH: Warren. SH: Krantz. OT: King. C: Jaskot. R: Preston. WT: George. BJ: Hays. MP: Landis. MS:
Higuchi. M: DeMarini. DM: Gilmour. MI. (2018). Mutagenicity and lung toxicity of smoldering vs. flaming
emissions from various biomass fuels: Implications for health effects from wildland fires. Environ Health
Perspect 126: 017011. http://dx.doi.org/10.1289/EHP2200
Kim. YS: Wells. A. (2005). The impact of forest density on property values. J Forest 103: 146-151.
http://dx.doi.Org/10.1093/iof/103.3.146
Kolden. CA. (2019). We're not doing enough prescribed fire in the Western United States to mitigate wildfire
risk. Fire 2: 30. http://dx.doi.org/10.3390/fire202003Q
Kuligowski. E. (2017). Burning down the silos: Integrating new perspectives from the social sciences into
human behavior in fire research. Fire and Materials 41: 389-411. http://dx.doi.org/10.10Q2/fam.2392
Kyle. GT: Theodori. GL: Absher. JD: Jun. J. (2010). The influence of home and community attachment on
firewise behavior. Society and Natural Resources 23: 1075-1092.
http://dx.doi.org/10.1080/089419209Q2724974
Liang. JJ: Calkin. DE: Gebert. KM: Venn. TJ: Silverstein. RP. (2008). Factors influencing large wildland fire
suppression expenditures. International Journal of Wildland Fire 17: 650-659.
http://dx.doi.org/10.1071/WF0701Q
Loomis. J. (2004). Do nearby forest fires cause a reduction in residential property values? 10: 149-157.
http://dx.doi.Org/10.1016/i.ife.2004.08.001
7-27
DRAFT: Do Not Cite or Quote
-------
Loomis. J: Kent. P: Strange. L: Fausch. K: Covich. A. (2000). Measuring the total economic value of restoring
ecosystem services in an impaired river basin: results from a contingent valuation survey. Ecol Econ 33: 103 -
117. http://dx.doi.org/10.1016/80921-8009(99100131-7
Loomis. JB: Gonzalez Caban. A: Englin. JE. (2001). Testing for differential effects of forest fires on hiking and
mountain biking demand and benefits. Western Journal of Agricultural Economics 26: 508-522.
http://dx.doi.org/10.22004/ag.econ.31049
McCaffrey. S: Rhodes. A: Stidham. M. (2015). Wildfire evacuation and its alternatives: Perspectives from four
United States' communities. International Journal of Wildland Fire 24: 170-178.
http://dx.doi.org/10.1071/WF13050
Meldrum. JR: Champ. PA: Brenkert-Smith. H: Warziniack. T: Barth. CM: Falk. LC. (2015). Understanding gaps
between the risk perceptions of wildland-urban interface (WUI) residents and wildfire professionals. Risk
Anal 35: 1746-1761. http://dx.doi.org/10.1111/risa. 12370
Mercer. DE: Prestemon. JP: Butrv. DT: Pve. JM. (2007). Evaluating alternative prescribed burning policies to
reduce net economic damages from wildfire. Am J Agric Econ 89: 63-77. http://dx.doi.org/10. Ill 1/i. 1467-
8276.2007.00963.x
Moore. WB: McCarl. BA. (1987). Off-site costs of soil erosion: A case study of the Willamette valley. Western
Journal of Agricultural Economics 12: 42-49.
Murphy. JD: Johnson. DW: Miller. WW: Walker. RF: Carroll. EF: Blank. RR. (2006). Wildfire effects on soil
nutrients and leaching in a Tahoe Basin watershed. J Environ Qual 35: 479-489.
http://dx.doi.org/10.2134/ieq2005.0144
Naeher. LP: Brauer. M: Lipsett. M: Zelikoff. JT: Simpson. CD: Koenig. JO: Smith. KR. (2007). Woodsmoke
health effects: A review [Review]. Inhal Toxicol 19: 67-106. http://dx.doi.org/10.1080/0895837060Q985875
Navarro. KM: Kleinman. MT: Mackav. CE: Reinhardt. TE: Balmes. JR: Brovles. GA: Ottmar. RD: Naher. LP:
Domitrovich. JW. (2019). Wildland firefighter smoke exposure and risk of lung cancer and cardiovascular
disease mortality. Environ Res 173: 462-468. http://dx.doi.Org/10.1016/i.envres.2019.03.060
NICC (National Interagency Coordination Center). (2017). National Interagency Coordination Center wildland
fire summary and statistics annual report 2017.
https://www.predictiveservices.nifc.gov/intelligence/2017 statssumm/intro summarvl7.pdf
NICC (National Interagency Coordination Center). (2018). National Interagency Coordination Center wildland
fire summary and statistics annual report 2018.
https://www.predictiveservices.nifc.gov/intelligence/2018 statssumm/intro summary 18.pdf
NICC (National Interagency Coordination Center). (2019). National Interagency Coordination Center wildland
fire summary and statistics annual report 2019.
https://www.predictiveservices.nifc.gov/intelligence/2019 statssumm/intro summary 19.pdf
Nielsen-Pincus. M. ax: Moselev. C: Gebert. K. (2014). Job growth and loss across sectors and time in the
western US: The impact of large wildfires. Forest Pol Econ 38: 199-206.
http://dx.doi.Org/10.1016/i.forpol.2013.08.010
Niemever. RJ: Bladoa KD: Woodsmith. RD. (2020). Long-term hydrologic recovery after wildfire and post-fire
forest management in the interior Pacific Northwest. Hydrolog Process 34: 1182-1197.
http://dx.doi.org/10.10Q2/hyp.13665
NWF (National Wildlife Federation). (2017). Ecosystem services. Available online at
https://www.nwf.org/Wildlife/Wildlife-Conservation/Ecosystem-Services.aspx (accessed July 9, 1905).
Pausas. JG: Keelev. JE. (2019). Wildfires as an ecosystem service. Front Ecol Environ 17: 289-295.
http://dx.doi.org/10.1002/fee.2044
Prestemon. JP: Abt. KL: Barbour. RJ. (2012). Quantifying the net economic benefits of mechanical wildfire
hazard treatments on timberlands of the western United States. Forest Pol Econ 21: 44-53.
http://dx.doi.Org/10.1016/i.forpol.2012.02.006
7-28
DRAFT: Do Not Cite or Quote
-------
Prestemon. JP: Entry. DT. (2005). Time to burn: Modeling wildland arson as an autoregressive crime function.
Am J Agric Econ 87: 756-770. http://dx.doi.org/10.1111/i. 1467-8276.2005.00760.X
Prestemon. JP: Butrv. DT: Abt. KL: Sutohen. R. (2010). Net benefits of wildfire prevention education efforts.
Forest Sci 56: 181-192. http://dx.doi.Org/10.1093/forestscience/56.2.181
Prestemon. JP: Pve. JM: Butrv. DT: Holmes. TP: Mercer. DE. (2002). Understanding broadscale wildfire risks in
a human-dominated landscape. Forest Sci 48: 685-693.
Prestemon. JP: Wear. DN: Stewart. FJ: Holmes. TP. (2006). Wildfire, timber salvage, and the economics of
expediency. Forest Pol Econ 8: 312-322. http://dx.doi.Org/10.1016/i.forpol.2004.07.003
Prichard. SJ: Kennedy. MC: Wright. CS: Cronan. JB: Ottmar. RD. (2017a). Predicting forest floor and woody
fuel consumption from prescribed burns in southern and western pine ecosystems of the United States. For
Ecol Manage 405: 328-338. http://dx.doi.Org/10.1016/i.foreco.2017.09.025
Prichard. SJ: Stevens-Rumann. CS: Hessburg. PF. (2017b). Tamm Review: Shifting global fire regimes: Lessons
from reburns and research needs. For Ecol Manage 396: 217-233.
http://dx.doi.Org/10.1016/i.foreco.2017.03.035
Proctor. CR: Lee. J: Yu. D: Shah. AD: Whelton. AJ. (2020). Wildfire caused widespread drinking water
distribution network contamination. AWWA Water Sci 2: el 183. http://dx.doi.org/10.1002/aws2.1183
Ouarles. SL: Pohl. K. (2018). Building a wildfire-resistant home: Codes and costs. Bozeman, MT: Headwaters
Economics, https://headwaterseconomics.org/wildfire/homes-risk/building-costs-codes/
Rahman. F. (2016). New study estimates 20 percent of firefighters, paramedics have PTSD. Rahman, F.
https://www.firerescuel.com/health/articles/new-studv-estimates-20-percent-of-firefighters-paramedics-
have-ptsd-
6zvmMnUZ7sWwZib6/#:~:text=New%20studv%20estimates%2020%20percent%20of%20firefighters%2C
%20paramedics%20have%20PTSD.-
The%20IAFF%20studv&text=LAS%20VEGAS%20%E2%80%94%20Firefighters%20experience%20post.I
nternational%20Association%20of%20Fire%20Fighters.
Reinhardt. TE: Ottmar. RD. (2000). Smoke exposure at western wildfires. (PNW-RP-525). Portland, OR: U.S.
Department of Agriculture, Forest Service, Pacific Northwest Research Station.
http://www.fs.fed.us/pnw/pubs/pnw rp525.pdf
Reinhardt. TE: Ottmar. RD. (2004). Baseline measurements of smoke exposure among wildland firefighters. J
Occup Environ Hyg 1: 593-606. http://dx.doi.org/10.1080/1545962049049Q101
Reinhardt. TE: Ottmar. RD: Hanneman. AJS. (2000). Smoke exposure among firefighters at prescribed burns in
the Pacific Northwest. (PNW-RP-526). Portland, OR: U.S. Department of Agriculture, Forest Service,
Pacific Northwest Research Station, http://depts.washington.edu/wildfire/resources/smokefire.pdf
Ren. D: Fu. R: Leslie. LM: Dickinson. RE. (2011). Modeling the mudslide aftermath of the 2007 Southern
California Wildfires. Natural Hazards 57: 327-343. http://dx.doi.org/10.1007/sllQ69-010-9615-5
Rhoades. CC: Chow. AT: Covino. TP: Fegel. TS: Pierson. DN: Rhea. AE. (2019a). The legacy of a severe
wildfire on stream nitrogen and carbon in headwater catchments. Ecosystems 22: 643-657.
http://dx.doi.org/10.1007/slQ021-018-0293-6
Rhoades. CC: Nunes. JP: Silins. U: Doerr. SH. (2019b). The influence of wildfire on water quality and
watershed processes: New insights and remaining challenges. International Journal of Wildland Fire 28: 721-
725. http://dx.doi.org/10.1071/WFv28nlQ FO
Rideout. DB: Omi. PN. (1990). Alternative expressions for the economic theory of forest fire management. 36:
614-624. http://dx.doi.Org/10.1093/forestscience/36.3.614
Rinne. JN. (1996). Short-term effects of wildfire on fishes and aquatic macroinvertebrates in the southwestern
United States. North American Journal of Fisheries Management 16: 653-658.
http://dx.doi.org/10.1577/1548-8675(1996)016<0653:MBSTEO>2.3.CO:2
7-29
DRAFT: Do Not Cite or Quote
-------
Rossi. D: Kuusela. OP. (2020). The influence of risk attitudes on suppression spending and on wildland fire
program budgeting. Forest Pol Econ 113. http://dx.doi.Org/10.1016/i.forpol.2019.102087
Sandberg. DV: Dost. FN. (1990). Effects of prescribed fire on air quality and human health. In JD Walstad;
SR Radosevich; DV Sandberg (Eds.), Natural and prescribed fire in the Pacific Northwest forests (pp. 191-
298). Corvallis, OR: Oregon State University Press.
Sharkey. B. (1997). Health hazards of smoke: Recommendations of the consensus conference, April 1997.
(9751-2836-MTDC). Missoula, MT: U.S. Department of Agriculture, Forest Service, Technology and
Development Program.
Simard. AJ. (1976). Wildland fire management: the economics of policy alternatives. Ottawa, Ontario: Canadian
Forestry Service, https ://d 1 ied5 g 1 xfgpx8.cloudfront.net/pdfs/24010.pdf
Smith. AB. (2020). U.S. billion-dollar weather and climate disasters, 1980 - present (NCEI Accession 0209268):
National Oceanic and Atmospheric Administration, National Centers for Environmental Information.
Retrieved from https://doi.org/10.25921/stkw-7w73
Smith. HG: Sheridan. GJ: Lane. PNJ: Nvman. P: Havdon. S. (2011). Wildfire effects on water quality in forest
catchments: A review with implications for water supply [Review]. J Hydrol 396: 170-192.
http://dx.doi.Org/10.1016/i.ihvdrol.2010.10.043
Sparhawk. WN. (1925). The use of liability ratings in planning forest fire protection. Emmitsburg, MD: National
Emergency Training Center.
https://books.google.com/books/about/The Use of Liability Ratings in Planning.html?id=7EqvJOAACA
AJ
Spies. TA: Long. JW: Charnlev. S: Hessburg. PF: Marcot. BG: Reeves. GH: Lesmeister. DB: Reillv. MJ:
Cervenv. LK: Stine. PA: Raphael. MG. (2019). Twenty-five years of the Northwest Forest Plan: What have
we learned? [Review]. Front Ecol Environ 17: 511-520. http://dx.doi.org/10.1002/fee.2101
Steele. TW: Stier. JC. (1998). An economic evaluation of public and organized wildfire detection in Wisconsin.
International Journal of Wildland Fire 8: 205-215. http://dx.doi.org/10.1071/WF9980205
Stetler. KM: Venn. TJ: Calkin. DE. (2010). The effects of wildfire and environmental amenities on property
values in northwest Montana, USA. Ecol Econ 69: 2233-2243.
http://dx.doi.Org/10.1016/i.ecolecon.2010.06.009
Stevens. MR. (2013). Analysis of postfire hydrology, water quality, and sediment transport for selected streams
in areas of the 2002 Hayman and Hinman fires, Colorado. (Scientific Investigations Report 2012-5267).
Reston, VA: U.S. Geological Survey, http://dx.doi.org/10.3133/sir20125267
Stockmann. K: Burchfield. J: Calkin. D: Venn. T. (2010). Guiding preventative wildland fire mitigation policy
and decisions with an economic modeling system. Forest Pol Econ 12: 147-154.
http://dx.doi.Org/10.1016/i.forpol.2009.09.009
Taylor. MH: Rollins. K: Kobavashi. M: Tausch. RJ. (2013). The economics of fuel management: Wildfire,
invasive plants, and the dynamics of sagebrush rangelands in the western United States. J Environ Manage
126: 157-173. http://dx.doi.Org/10.1016/i.ienvman.2013.03.044
Thomas. D: Butrv. D: Gilbert. S: Webb. D: Fung. J. (2017). The costs and losses of wildfires: A literature
review. (NIST Special Publication 1215). Gaithersburg, MD: National Institute of Standards and
Technology. http://dx.doi.org/10.6028/NIST.SP.1215
Thomas. PS: Butrv. DT: Prestemon. JP. (2011). Enticing arsonists with broken windows and social disorder.
Fire Technology 47: 255-273. http://dx.doi.org/10.1007/slQ694-010-0145-l
Thompsoa MP: Rilev. KL: Loeffler. D: Haas. JR. (2017). Modeling fuel treatment leverage: Encounter rates,
risk reduction, and suppression cost impacts. Forests 8: 469. http://dx.doi.org/10.3390/f8120469
U.S. EPA (U.S. Environmental Protection Agency). (2019). Integrated Science Assessment (ISA) for particulate
matter (final report, Dec 2019) [EPA Report]. (EPA/600/R-19/188). Washington, DC.
https ://cfpub .epa. gov/ncea/isa/recordisplav ,cfm?deid=3 47534
7-30
DRAFT: Do Not Cite or Quote
-------
USFA (US Fire Administration). (2018). Firefighter fatalities in the United States in 2017. Emmitsburg, MD:
Federal Emergency Management Agency.
https://www.usfa.fema.gov/downloads/pdf/publications/ff fatl7.pdf
Uzun. H: Dahlgren. RA: Olivares. C: Erdem. CU: Karanfil. T: Chow. AT. (2020). Two years of post-wildfire
impacts on dissolved organic matter, nitrogen, and precursors of disinfection by-products in California
stream waters. Water Res 181: 115891. http://dx.doi.Org/10.1016/i.watres.2020.115891
Vose. JM: Peterson. PL: Luce. CH: Patel-Wevnand. T. (2019). Effects of drought on forests and rangelands in
the United States: Translating science into management responses. (Gen. Tech. Rep. WO-98). Washington,
DC: U.S. Department of Agriculture, Forest Service. http://dx.doi.org/10.2737/WO-GTR-98
Wang. D: Guan. D: Zhu. S: Kinnon. MM. ac: Geng. G: Zhang. O: Zheng. H: Lei. T: Shao. S: Gong. P: Davis. SJ.
(2020). Economic footprint of California wildfires in 2018. Nature Sustainability.
http://dx.doi.org/10.1038/s41893-020-0Q646-7
Wolters. EA: Steel. BS: Weston. D: Brunsond. M. (2017). Determinants of residential firewise behaviors in
central Oregon. The Social Science Journal 54: 168-178. http://dx.doi.Org/10.1016/i.soscii.2016.12.004
Young. MK. (1994). Movement and characteristics of stream-borne coarse woody debris in adjacent burned and
undisturbed watersheds in Wyoming. Can J For Res 24: 1933-1938. http://dx.doi.org/10.1139/x94-248
Zeng. Z: Minton. TP. (2021). Jail inmates in 2019. (NCJ 255608). Washington, DC: U.S. Department of Justice,
Office of Justice Programs, Bureau of Justice Statistics, https://www.bis. gov/content/pub/pdf/ii 19.pdf
Zhuang. J: Pavvappalli. VM: Behrendt. A: Lukasiewicz. K. (2017). The total cost of fire in the United States.
(FPRF-2017-21). Quincy, MA: National Fire Protection Association, https://www.nfpa.org//-
/media/Files/News-and-Research/Fire-statistics-and-reports/US-Fire-Problem/RFTotalCost.pdf
7-31
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
CHAPTER 8 ESTIMATED PUBLIC HEALTH
IMPACTS
8.1 Introduction
A main goal of this assessment is to provide a quantitative comparison of the estimated health
impacts and associated economic values attributed to smoke from wildland fire (i.e., wildfire and
prescribed fire) under different fire management strategies by focusing on two case study fires: Timber
Crater 6 (TC6) and Rough fires. Previous chapters of this assessment describe in detail the air quality
impacts of each case study fire and defined hypothetical scenarios meant to reflect different fire
management strategies (CHAPTER 5) and the health effects of wildfire smoke (CHAPTER 6). which
collectively represent key inputs to the process of quantitatively estimating health impacts. Within this
chapter, the information presented in previous chapters is used to conduct analyses using U.S.
Environmental Protection Agency's (U.S. EPA's) Environmental Benefits Mapping and Analysis
Program—Community Edition (BenMAP-CE) to provide additional insight on the overall public health
impacts of wildland fire smoke and how those impacts can vary depending on the fire management
strategy employed.
8.2 Benefits Mapping and Analysis Program—Community
Edition (BenMAP-CE) Analysis
BenMAP-CE quantifies the number and economic value of air pollution-related premature deaths
and illnesses (Sacks et al.. 2018). The program draws upon a library of preinstalled and user-imported
input parameters (Table 8-1) to systematize the procedure for calculating the estimated health impact and
then valuing the resulting counts of adverse effects. The sections below describe the steps to configuring
and running BenMAP-CE to estimate the number, and corresponding economic impact, of wildland
fire-related particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |im
(PM2 5) and ozone-attributable effects.
8-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
Table 8-1 Key data inputs for Benefits Mapping and Analysis
Program—Community Edition (BenMAP-CE) used to estimate health
impacts for the case studies.
Data Input
Source
Air quality data
Modeled PM2.5 and ozone concentrations from each case study3
Population counts
U.S. census data allocated to air quality model grid cells, stratified by
race, sex, age, and ethnicity and projected to the Year 2021
Risk coefficients
Concentration-response relationships from U.S-based air pollution
epidemiologic studies examining PM2.5, ozone, and wildfire-specific
PM2.5b
Baseline rates of death and disease
Centers for Disease Control and Prevention provided death rates, and
Healthcare Cost and Utilization Program provided hospital visit rates
for all other areas
BenMAP-CE = Benefits Mapping and Analysis Program—Community Edition; PM2 5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm.
aFor more information see CHAPTER 5
bFor more information on epidemiologic studies examining wildfire-specific PM25 see CHAPTER 6
8.2.1 Health Impact Function
This analysis estimates the number of wildfire and prescribed fire-attributable premature deaths
and illnesses associated with the TC6 and Rough fire case studies using a health impact function. The
following example Equation_8-l) details the approach for calculating PM2 5-attributable premature deaths;
the approach for quantifying PM-attributable morbidity impacts and ozone-related mortality and
morbidity impacts is identical except for the ages for which the function is calculated, as detailed below.
Counts of PM2 5-attributable total deaths (y/;) are calculated for period /' (/' = 2021) among individuals of
all ages (0-99) (a) in each county j (J = 1,...,J where J is the total number of counties) as:
yij = llayiJa
Vija = m°ija X (e/? "i_1) X Pm,
Equation 8-1
where mo,]L, is the daily baseline all-cause mortality rate for individuals aged a = 0-99 in county j in
Year /' stratified in 10-year age groups, (3 is the risk coefficient for all-cause mortality for adults associated
with PM2 5 exposure, Q is annual mean PM2 5 concentration in county j in Year and is the number
of residents aged a = 0-99 in county j in Year /' stratified into 5-year age groups. When calculating
impacts, the program assigns the 10-year stratified death rate to the corresponding 5-year stratified
8-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
population bin. The health impact function used to calculate all other impacts is identical to Equation 8-1,
except for the effect coefficient. The program performs a Monte Carlo analysis by randomly sampling
5,000 times from a distribution constructed from the standard error reported for each study; the resulting
distribution is then used to report 95% confidence intervals.
The function above is calculated using BenMAP-CE (vl.5.5.1), atool that contains the baseline
incidence rates, population counts, and health impact functions needed to quantify counts of PM2 5 and
ozone attributable deaths and respiratory hospital admissions (U.S. EPA. 2019; Sacks et al.. 2018). This
approach to quantifying air pollution health impacts, and the adverse effects of wildland fires in
particular, has been used within the peer-reviewed literature (Farm et al.. 2019; Fann et al.. 2018; Berman
et al.. 2012). The following sections describe the specification of each input parameter within
BenMAP-CE for the purposes of the analyses conducted within this assessment.
8.2.2 Air Quality Modeling
The emissions inputs and photochemical modeling simulations performed to predict the PM2 5
and ozone concentrations attributable to each case study fire, prescribed fire activity in each location, and
defined hypothetical scenarios are detailed in CHAPTER 5. As noted in 0, for each hypothetical scenario,
wildfire specific air quality impacts (the delta used to estimate the change in health impacts) is calculated
using a baseline of no case study fire to estimate the burden attributed to the actual fire, prescribed fires,
and hypothetical scenarios for each case study. BenMAP-CE used the model-predicted daily mean PM2 5
and model-predicted daily 8-hour max ozone concentration to quantify health impacts for the following
actual fire and hypothetical scenarios for each case study:
TC6 Case Study
• Actual TC6 Fire
• Hypothetical Scenario 1 (small): a smaller hypothetical TC6 Fire in a heavily managed area (most
prescribed fire/fire managed for resource benefits), which would equate to a wildfire with less
fuel, a smaller fire perimeter, and less daily emissions
• Hypothetical Scenario 2a (large): a larger hypothetical TC6 Fire, but not the "worst-case"
scenario, due to no land management which would equate to a wildfire with more fuel, a larger
fire perimeter, and more daily emissions
• Hypothetical Scenario 2b (largest): a much larger, hypothetical "worst-case" scenario TC6 Fire
with no land management (i.e., no prescribed fire/managed fire) which would equate to a wildfire
with the most fuel, largest fire perimeter, and largest daily emissions
• Prescribed fires: three prescribed fires that occurred in the past and one prescribed fire that
occurred in 2019, all modeled to occur on the same days in September 2019 that fit prescription
conditions
8-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Rough Fire Case Study
• Actual Rough Fire
• Hypothetical Scenario 1 (small): a small hypothetical Rough Fire that represents the combined
impact of the proposed Boulder Creek Prescribed Fire and the Sheep Complex Fire, a wildfire
managed for resource benefits, on reducing the overall size of the Rough Fire
• Hypothetical Scenario 2 (large): a large hypothetical Rough Fire that allows for the fire perimeter
of the Rough Fire to progress into the area of the Sheep Complex Fire as if both the Boulder
Creek Prescribed Fire and Sheep Complex Fire did not occur
• Boulder Creek Prescribed Fire: a proposed prescribed fire that was planned, but did not occur in
the fall of 2014
• Sheep Complex Fire: a wildfire that occurred in 2010 due to a lightning strike and because of wet
fuel conditions was effectively managed to achieve the same objectives as a prescribed fire
8.2.3 Effect Coefficients
This analysis quantifies an array of adverse health effects attributable to PM2 5 and ozone
exposures, including premature death and morbidity outcomes. For the main analysis, the chosen studies
examine the health effects associated with ambient exposures to PM2 5 and ozone and have been used in
recent U.S. EPA benefits analyses. U.S. EPA recently published a Technical Support Document that
provides a detailed description of the Agency's systematic evaluation of the epidemiologic literature and
the concentration-response (C-R) relationships used to develop health impact functions (U.S. EPA. 2021).
In summary for PM2 5, analyses focus on the following outcomes: short-term PM2 5 exposure and
mortality, all ages (Zanobetti and Schwartz. 2009); long-term PM2 5 exposure and mortality, ages
30-99 (Turner et al.. 2016); respiratory-related emergency department (ED) visits, all ages (Krall et al..
2016); cardiovascular-related ED visits, all ages (Qstro et al.. 2016); respiratory-related hospital
admissions, ages 0-18 years (Qstro et al.. 2009); and cardiovascular-related hospital admissions, ages
65 and over (Bell et al.. 2015). For ozone, analyses focus on short-term ozone exposure and respiratory
mortality, all ages (Katsouvanni et al.. 2009); long-term ozone exposure and respiratory mortality, ages
30-99 (Turner et al.. 2016); respiratory-related ED visits, all ages (Barry et al.. 2019); and
respiratory-related hospital admissions, ages 65 and over (Katsouvanni et al.. 2009).
The analysis quantifies the same morbidity impacts for each case study scenario. However,
because the length of the actual TC6 and Rough fires varied, the analysis quantifies mortality impacts
differently for each case study. Because the TC6 Fire only lasted a few days, mortality impacts are
quantified using a short-term PM2 5 exposure function. However, because the Rough Fire lasted multiple
months, mortality impacts are quantified using a long-term PM2 5 exposure function. Mortality impacts
due to short-term PM2 5 exposure are not quantified in the Rough Fire case study analyses to prevent the
double counting of mortality impacts.
8-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Whereas the main analyses rely on health impact functions derived from epidemiologic studies of
ambient PM2 5 exposures, the sensitivity analysis examined whether estimated health impacts differed
when using health impact functions derived from epidemiologic studies that specifically examined
wildfire smoke exposure (i.e., wildfire-specific PM2 5). In the sensitivity analysis, only respiratory and
cardiovascular outcomes are quantified because among the epidemiologic studies evaluated in CHAPTER
6 (see Section 6.2). only these studies used an exposure indicator of wildfire PM2 5 and were suitable for
use within BenMAP-CE (i.e., were conducted in locations similar to the case studies and represented
health outcomes with available incidence data). Of the available respiratory-related ED visits studies that
used wildfire PM2 5 as the exposure indicator, none examined all respiratory-related ED visits; as a result,
the sensitivity analysis quantified asthma ED visits using a risk coefficient from a study conducted by
Reid et al. (2019) in northern California. With respect to hospital admissions, respiratory-related hospital
admissions were quantified using a risk coefficient from a study conducted by Gan et al. (2017) in
Washington state, and cardiovascular-related hospital admissions were quantified using a risk coefficient
from a study focusing on a wildfire event in southern California conducted by Delfino et al. (2009).
8.2.4 Baseline Incidence and Prevalence Data
The epidemiologic studies noted above report estimates of risk (i.e., effect coefficients or (3
coefficients) that are expressed as being relative to a baseline rate. In this analysis, these effect
coefficients were used to quantify cases of ED visits, hospital admissions and premature deaths, and thus
baseline rates of all-cause mortality, ED visits, and hospital admissions were used in the estimation of
these health impacts. County-level age-stratified all-cause death rates were obtained from the Centers for
Disease Control Wide-ranging ONline Data for Epidemiologic Research (WONDER) database (CDC.
2016) for the Year 2010, while ED visit and hospital visit rates were obtained from the Healthcare Cost
and Utilization Program (HCUP), which consists of a mixture of county, state and regional rates.
8.2.5 Assigning PM2.5 Concentrations to the Population
Changes in population-level exposure are quantified by assigning the predicted PM2 5
concentrations to the U.S. census-reported population in each 4-km by 4-km model grid cell for the TC6
Fire case study and 12 km by 12 km in the Rough Fire case study (see CHAPTER 5 for a detailed
description of the air quality modeling simulations). As a first step, the PopGrid population preprocessing
tool was used to assign U.S. census-reported population counts at the census block level to each air
quality model grid cell. These population counts were stratified by age, sex, race, and ethnicity. The
census-reported population counts for the Year 2010 were used and then counts were projected to the
Year 2020 using forecast population from Woods & Poole (2016).
8-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
To calculate wildland fire PM2 5 concentrations, concentrations were weighted to the size of the
population exposed to wildland fire PM2 5 concentrations for all counties combined (()) in Year i as
_ Yj) Cjj X Pjj
Pi
Equation 8-2
where Q is the wildfire-attributable annual mean PM2 5 concentration in county j in Year is the
population in county j in Year i, and is the total population over all counties combined in Year i.
8.2.6 Economic Analysis
The value of avoided premature deaths was estimated using a Value of Statistical Life (VSL)
recommended by the U.S. EPA's Guidelines for Preparing Economic Analyses (U.S. EPA. 2014).
Following U.S. EPA guidelines, this value was indexed to the inflation and income year of the analysis.
Using a 2015 inflation year and assuming 2020 income levels, a VSL of $9.5 millions (M) was used. To
value changes in respiratory hospital admissions, a cost of illness estimate was used, which is consistent
with the approached used by the U.S. EPA in its Regulatory Impact Analysis for the PM2 5 National
Ambient Air Quality Standards (U.S. EPA. 2013). This value of $36,000 reflects the direct medical costs
associated with the hospital visit as well as lost earnings. Following this same approach, we estimate the
value of cardiovascular hospital admissions to fall between $41,000 and $42,000 depending on the age of
onset. Finally, we quantify the value of emergency department visits using a simple average of two Cost
of Illness values reported by Smith et al. (1997) and Stanford et al. (1999). which produces a value of
$430.
8.3 Results from Case Study Fire Analyses
The sections present the estimated health impacts and corresponding economic values from the
BenMAP-CE analyses for each of the actual fires, hypothetical scenarios, prescribed fires, and wildfire
that yielded positive resource benefits for each case study. The main results presented in Section 8.3.1 are
based on risk coefficients from epidemiologic studies used by U.S. EPA in previous benefits analyses as
noted above; while Section 8.3.2 presents results from the sensitivity analyses using risk coefficients from
studies examining wildfire-specific PM2 5 and alternative epidemiologic studies examining ambient ozone
exposure. Lastly, building off the discussion presented in CHAPTER 6 (see Section 6.3). Section 8.3.3
estimates the potential reduction in health impacts presented that could be achieved through the
implementation of various actions or interventions to reduce or mitigate wildland fire smoke exposure.
8-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
8.3.1
Main Results
The estimated number and value of wildfire-related health impacts varies across the scenarios and
the pollutant assessed. PM2 5-attributable effects are consistently larger than those quantified for ozone.
The estimated number of premature deaths, ED visits, and hospital admissions are larger for the Rough
Fire scenarios than they are for the TC6 Fire scenarios; this can be attributed to differences in the
magnitude of the fires, the duration of each fire, and the population density around each fire. For the TC6
Fire scenarios, fractional counts of air pollution-attributable effects are presented to illustrate the small,
but meaningful, differences in impacts among the scenarios.
The dollar value of fires for the TC6 Fire case study is as large as $100 M while the value of the
Rough Fire case study is as large as $3 billions (B). These values represent the sum of the medical costs
and productivity losses associated with the ED visits and hospital admissions and the of value air
pollution-attributable deaths. This latter value is quantified using a Value of Statistical Life, which is a
measure of an individual's willingness to pay to reduce the risk of dying prematurely by a small amount;
it is not the value of any individual life.
8-7
DRAFT: Do Not Cite or Quote
-------
Table 8-2 Estimated counts of PM2.5 premature deaths and illnesses (95%
confidence interval).
ED Visits Hospital Admissions Mortality
Case
Study
Scenario
Respiratory
Cardiovascular
Respiratory
Cardiovascular
Short Term
Long Term
Actual fire
0.2
(0.0 to 0.4)
0.1
(-0.0 to 0.2)
0.0
(0.0 to 0.0)
0.0
(0.0 to 0.1)
0.04
(0.01 to 0.08)
...
<0
0
Scenario 1
(small)
0.1
(0.0 to 0.2)
0.1
(-0.0 to 0.1)
0.0
(0.0 to 0.0)
0.0
(0.0 to 0.0)
0.03
(0.01 to 0.5)
...
0
+¦»
re
0
Scenario 2a
(large)
0.8
(0.2 to 1.6)
0.4
(-0.1 to 0.9)
0.1
(0.0 to 0.1)
0.2
(0.1 to 0.2)
0.16
(0.01 to 0.32)
...
0
&
E
i-
Scenario 2b
(largest)
1.2
(0.2 to 2.5)
0.6
(-0.2 to 1.3)
0.1
(0.1 to 0.2)
0.3
(0.2 to 0.3)
0.25
(0.01 to 0.49)
...
Prescribed
fires
0.04
(0.01 to 0.08)
0.02
(-0.01 to 0.05)
0.00
(0.00 to 0.01)
0.01
(0.01 to 0.01)
0.01
(0.001 to 0.02)
...
Actual fire
47.3
(9.3 to 98.5)
19.7
(-7.6 to 46.0)
6.9
(3.0 to 10.7)
8.6
(6.2 to 10.9)
...
80.0
(53.6 to 105.4)
Scenario 1
(small)
28.2
(5.5 to 58.7)
11.8
(-4.6 to 27.6)
4.2
(1.8 to 6.5)
5.0
(3.6 to 6.3)
...
48.1
(32.2 to 63.4)
S
Li.
.c
Scenario 2
(large)
49.8
(9.8 to 103.7)
20.7
(-8.0 to 48.4)
7.3
(3.2 to 11.2)
9.1
(6.6 to 11.5)
...
84.3
(56.5 to 111.1)
U)
D
O
a.
Sheep
Complex
Fire
6.6
(1.3 to 13.7)
2.7
(-1.0 to 6.2)
0.9
(0.4 to 1.4)
0.9
(0.7 to 1.2)
—
10.1
(6.7 to 13.3)
Boulder
Creek
Prescribed
Fire
1.1
(0.2 to 2.4)
0.5
(-0.2 to 1.1)
0.2
(0.1 to 0.3)
0.2
(0.2 to 0.3)
1.9
(1.3 to 2.5)
ED = emergency department; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to
2.5 |jm; TC6 = Timber Crater 6.
8-8 DRAFT: Do Not Cite or Quote
-------
Table 8-3 Estimated counts of ozone (O3) premature deaths and illnesses (95%
confidence interval).
Mortality
Respiratory Respiratory Hospital
Case Study Scenario ED Visits Admissions Short Term Long Term
Actual fire
0.06
(0.02 to 0.1)
0.0
(-0.0 to 0.0)
0.0
(-0.0 to 0.0)
-------
Table 8-4 Estimated value of PM2.5 and ozone-related premature deaths and
illnesses (95% confidence interval; millions of 2015 dollars).
Sum of Value of Morbidity Impacts and Value of:
Case
Study
Scenario
Short-Term Exposure
Mortality ($)
Long-Term Exposure
Mortality ($)
Actual Fire
18
(2 to 47)
-
<0
0
1-
Scenario 1 (small)
10
(1 to 26)
-
0
+¦»
re
0
Scenario 2a (large)
66
(6 to 170)
-
0
&
E
i-
Scenario 2b (largest)
100
(9 to 270)
-
Prescribed Fires
4
(0 to 9)
-
Actual Fire
...
3,000
(260 to 7,900)
82
Scenario 1 (small)
...
1,800
(160 to 4,700)
Li.
.c
U)
D
O
Scenario 2 (large)
...
3,100
(270 to 8,300)
a:
Sheep Complex Fire
...
350
(20 to 960)
Boulder Creek Prescribed
Fire
...
60
(5 to 160)
PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; TC6 = Timber Crater 6.
8.3.2 Sensitivity Analyses
1 As noted above, the results presented within this section include estimates derived from health
2 impact functions based on risk coefficients from epidemiologic studies that examined exposures to
3 wildfire-specific PM2 5 as a comparison to results from health impact functions based on ambient PM2 5
4 exposures. Compared to the main analysis results, using the wildfire-specific PM2 5 functions resulted in
5 an increase in the estimated impacts for each case study (TC6: Figure 8-1; Rough Fire: Figure 8-2). This
6 difference in estimated health impacts between studies examining ambient and wildfire-specific PM2 5
8-10
DRAFT: Do Not Cite or Quote
-------
1 exposures could be attributed to a steeper C-R relationship at the higher short-term PM2 5 concentrations
2 experienced during wildfire events or the behavior of individuals exposed to PMis during a wildfire
3 event. However, additional research focused on examining the C-R relationship for wildfire smoke
4 exposure is required to fully grasp the differences between the main analysis and sensitivity analysis
5 results. The corresponding economic values from the sensitivity analyses are presented in Table 8-5, but
6 these values are not directly comparable to the main analysis because the sensitivity analyses did not
7 estimate premature deaths as noted in Section 8.2.3.
Timber Crater 6 (TC6) Fire Case Study Sensitivity Analyses
Respiratory ED Visits Asthma ED Visits Respiratory Hospital Admissions Cardiovascular Hospital Admissions
2.9
OA T
h
-0.1
vy/ -i? J9
•4-
I
//>° ^
• •
1
.|4"
vyyy
>v
"V r$> <5
I 1
1
1
c- \ r/b
5 5
1
^ ¦<&
.x <&> jC cv
cJ-> <
h °>, &
ED = emergency department; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 (jm;
TC6 = Timber Crater 6.
Figure 8-1 Estimated health impacts from sensitivity analyses using health
impact functions based on ambient PM2.5 exposures versus
wildfire-specific PM2.5 exposures for the Timber Crater 6 (TC6)
Fire case study.
8-11
DRAFT: Do Not Cite or Quote
-------
Rough Fire Case Study Sensitivity Analyses
Respiratory ED Visits
i
Asthma ED Visits
i
Respiratory Hospital Admissions
Cardiovascular Hospital Admissions
I
60
J.
I 5 i . . [11}
i . 5
• •
<> ?
• ^ . I® . 4® <4^ . I® , y .«•» -.« X -rt ..O rC>x x >P ^ *
«#* A^V
-%>°
r o>V
a*. -*
•#VV <0°"
*" Ou
.0°.^
<6°"
9
<^> C?
^ ^ ,x
r oV
• t®1 .<€»
o" A'
/s#/ «c
®r
lS° c/V"
£. -s
><•&>
s^VV
'r
ED = emergency department; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 pm.
Figure 8-2 Estimated health impacts from sensitivity analyses using health
impact functions based on ambient PM2.5 exposures versus
wildfire-specific PM2.5 exposures for the Rough Fire case study.
8-12
DRAFT: Do Not Cite or Quote
-------
Table 8-5 Estimated value of wildfire-specific PM2.5 illnesses (95% Confidence
interval; 2015 dollars) from sensitivity analyses.
Case Study
Scenario
Sum of Value of Morbidity Impacts ($)
Actual fire
8,600
(-76 to 17,000)
<0
0
<0
Scenario 1 (small)
5,100
(-59 to 10,000)
O
+-»
rc
1_
O
Scenario 2a (large)
35,000
(-220 to 69,000)
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Using the average overall exposure reduction that could be achieved due to various exposure
reduction actions, presented in Table 6-1, the potential reduction in public health impacts that could be
achieved are estimated for both case study fires, the corresponding hypothetical scenarios, and the
prescribed fires (either actual or hypothetical) conducted in each location. The estimated overall reduction
in total health impacts in Table 8-6 and Table 8-7 assume a linear relationship between population
exposure concentrations and estimated health impacts such that the percent reduction in PM2 5 exposure
corresponds to an equivalent percent reduction in health impacts. Also as noted in Section 6.3.3. the
reduction in health impacts presented in Table 8-6 and Table 8-7 correspond to an average overall
exposure reduction based on data from available studies and accounts for both the magnitude of the
intervention and the likelihood that this intervention is employed. The exposure reductions presented do
not account for differences in communication efforts between wildfires and prescribed fires or that
different concentrations may impact the likelihood of taking action as well as factors specific to the case
study areas (e.g., population demographics and housing stock) that can influence the corresponding
exposure reduction for these actions. Additionally, the estimation of the reduction in potential public
health impacts attributed to smoke exposure for each actual fire is not meant to reflect a formal analysis of
post-fire effectiveness of public health messaging by Air Resource Advisors (ARAs) deployed by the
U.S. Forest Service, in combination with respective state and local air quality agencies, for either the TC6
or Rough fires, but instead an estimation of the potential implications of exposure reduction actions on
reducing the overall public health impact of smoke.
Table 8-6 Overall reduction in total health impacts attributed to PM2.5 from
wildfire smoke for the Timber Crater 6 (TC6) Fire case study.
Exposure Reduction Action
(Overall Exposure
Reduction; %)
Hypothetical Scenarios
Actual
Fire
1
(small)
2a
(large)
2b
(largest)
Prescribed
Fires
Total health impacts3
0.34
0.23
1.66
2.45
0.08
Stayed inside (31.8%)
-0.11
-0.07
-0.53
-0.78
-0.03
Ran home HVAC system (24%)
-0.08
-0.06
-0.40
-0.59
-0.02
Evacuated (24%)
-0.08
-0.06
-0.40
-0.59
-0.02
Used air cleaner (15%)
-0.05
-0.03
-0.25
-0.37
-0.01
ED = emergency department; HVAC = heating, ventilation, and air conditioning; PM25 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; TC6 = Timber Crater 6.
aTotal number of health impacts represents the sum of ED visits, hospital admissions, and mortality detailed in Table 8-2; negative
values in the table represent the estimated overall reduction in total impacts.
Corresponding 95% confidence intervals are not presented because these results represent and illustrative example.
8-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 8-7 Overall reduction in total health impacts attributed to PM2.5 from
wildfire smoke for the Rough Fire case study.
Exposure Reduction Action
(Overall Exposure
Reduction; %)
Hypothetical Scenarios
Rough
Fire
1
(small)
2
(large)
~ Boulder Creek
Sheep Fire—Prescribed
Complex Fire Fire3
Total health impacts'5
162.5
97.3
171.5
21.2
Stayed inside (31.8%)
-51.7
-30.9
-54.5
-6.7
Ran home HVAC system (24%)
-39.0
-23.4
-41.2
-5.1
Evacuated (24%)
-39.0
-23.4
-41.2
-5.1
Used air cleaner (15%)
-24.4
-14.6
-25.7
-3.2
ED = emergency department; HVAC = heating, ventilation, and air conditioning; PM25 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; TC6 = Timber Crater 6.
aThe health impacts for the Boulder Creek analysis were negative, hence they are not reported in this table.
Corresponding 95% confidence intervals are not presented because these results represent and illustrative
example.
Total number of health impacts represents the sum of ED visits, hospital admissions, and mortality detailed in
Table 8-2; negative values in the table represent the estimated overall reduction in total impacts.
8.4 Summary
The analyses presented within this chapter estimate the potential public health impacts and
associated economic values attributed to smoke exposure, focusing specifically on PM2 5 and ozone, from
wildland fire within the case study areas of the TC6 and Rough fires. Analyses for both case studies,
which build off the assessment of the air quality impacts of each actual fire, hypothetical scenarios, and
prescribed fires presented in CHAPTER 5. demonstrate that health impacts are dominated by exposure to
PM2 5 from wildland fire smoke.
The results of the case study analyses indicate that proximity to population centers and
atmospheric conditions (e.g., wind patterns) influence the magnitude of health impacts attributed to
smoke. Building off the air quality modeling analyses presented in CHAPTER 5 that depict differences in
both PM2 5 concentrations and population exposures, the corresponding BenMAP-CE analyses indicate
that fire management strategies targeted to reduce the spread and overall size of wildfires, as depicted in
the smaller hypothetical fires, can result in substantial differences in the health impacts and corresponding
economic values when compared to the actual fires. Even though prescribed fires in both case study areas,
and wildfires managed for resource benefits (i.e., Sheep Complex Fire), are shown to contribute to an
8-15
DRAFT: Do Not Cite or Quote
-------
1 estimated reduction in health impacts from wildfires, it is important to recognize that these fires are not
2 without risk and do also contribute to health impacts, albeit smaller in number.
3 Sensitivity analyses that explore potential differences in estimated health impacts between health
4 impact functions derived from epidemiologic studies of ambient PM2 5 and wildfire-specific PM2 5,
5 provide evidence of potentially larger estimated impacts when using wildfire-specific PM2 5 health impact
6 functions. Additional analyses that provide a crude estimation of the potential implications of actions or
7 interventions to reduce and mitigate wildland fire smoke exposure demonstrate the potential public health
8 benefits of messaging campaigns to the public. However, for both sensitivity analyses, additional research
9 is warranted to more fully assess the implications of using ambient and wildfire-specific PM2 5 health
10 impact functions, and to provide a more representative estimation of the potential public health benefits of
11 actions or interventions to reduce wildfire smoke exposure.
8-16
DRAFT: Do Not Cite or Quote
-------
8.5 References
Barry. V: Klein. M: Winguist. A: Chang. HH: Mulholland. JA: Talbott. EQ: Rager. JR: Tolbert. PE: Sarnat. SE.
(2019). Characterization of the concentration-response curve for ambient ozone and acute respiratory
morbidity in 5 US cities. J Expo Sci Environ Epidemiol 29: 267-277. http://dx.doi.org/10.1038/s41370-Q18-
0048-7
Bell. ML: Son. JY: Peng. RD: Wang. Y: Dominici. F. (2015). Ambient PM2.5 and risk of hospital admissions:
Do risks differ for men and women? Epidemiology 26: 575-579.
http://dx.doi.org/10.1097/EDE.0000000000000310
Berman. JD: Fann. N: Hollingsworth. JW: Pinkerton. KE: Rom. WN: Szema. AM: Brevsse. PN: White. RH:
Curriero. FC. (2012). Health benefits from large-scale ozone reduction in the United States. Environ Health
Perspect 120: 1404-1410. http://dx.doi.org/10.1289/ehp. 1104851
CDC (Centers for Disease Control and Prevention). (2016). CDC WONDER (data from years 1980-2010)
[Database]. Atlanta, GA. Retrieved from http://wonder.cdc.gov/
Delfino. RJ: Brummel. S: Wu. J: Stern. H: Ostro. B: Lipsett. M: Winer. A: Street. DH: Zhang. L: Tioa. T: Gillen.
PL. (2009). The relationship of respiratory and cardiovascular hospital admissions to the southern California
wildfires of 2003. Occup Environ Med 66: 189-197. http://dx.doi.org/10.1136/oem.2008.041376
Fann. N: Alman. B: Broome. RA: Morgan. GG: Johnston. FH: Pouliot. G: Rappold. AG. (2018). The health
impacts and economic value of wildland fire episodes in the U.S.: 2008-2012. Sci Total Environ 610-611:
802-809. http://dx.doi.Org/10.1016/i.scitotenv.2017.08.024
Fann. N: Coffman. E: Haiat. A: Kim. SY. (2019). Change in fine particle-related premature deaths among US
population subgroups between 1980 and 2010. Air Qual Atmos Health 12: 673-682.
http://dx.doi.org/10.1007/sll869-019-0Q686-9
Gan. RW: Ford. B: Lassman. W: Pfister. G: Vaidvanathan. A: Fischer. E: Volckens. J: Pierce. JR: Magzamen. S.
(2017). Comparison of wildfire smoke estimation methods and associations with cardiopulmonary-related
hospital admissions. Geohealth 1: 122-136. http://dx.doi.org/10.1002/2017GH000Q73
Katsouvanni. K: Samet. JM: Anderson. HR: Atkinson. R: Le Tertre. A: Medina. S: Samoli. E: Touloumi. G:
Burnett. RT: Krewski. D: Ramsay. T: Dominici. F: Peng. RD: Schwartz. J: Zanobetti. A. (2009). Air
pollution and health: A European and North American approach (APHENA). (HEI Research Report 142).
Boston, MA: Health Effects Institute, https://www.healtheffects.org/publication/air-pollution-and-health-
european-and-north-american-approach
Krall. JR: Mulholland. JA: Russell. AG: Balachandran. S: Winguist. A: Tolbert. PE: Waller. LA: Sarnat. SE.
(2016). Associations between source-specific fine particulate matter and emergency department visits for
respiratory disease in four U.S. cities. Environ Health Perspect 125: 97-103.
http://dx.doi.org/10.1289/EHP271
Ostro. B: Malig. B: Hasheminassab. S: Berger. K: Chang. E: Sioutas. C. (2016). Associations of source-specific
fine particulate matter with emergency department visits in California. Am J Epidemiol 184: 450-459.
http://dx.doi.org/10.1093/aie/kwv343
Ostro. B: Roth. L: Malig. B: Marty. M. (2009). The effects of fine particle components on respiratory hospital
admissions in children. Environ Health Perspect 117: 475-480. http://dx.doi.org/10.1289/ehp. 11848
Reid. CE: Considine. EM: Watson. GL: Telesca. D: Pfister. GG: Jerrett. M. (2019). Associations between
respiratory health and ozone and fine particulate matter during a wildfire event. Environ Int 129: 291-298.
http://dx.doi.Org/10.1016/i.envint.2019.04.033
Sacks. JD: Lloyd. JM: Zhu. Y: Anderton. J: Jang. CJ: Hubbell. B: Fann. N. (2018). The Environmental Benefits
Mapping and Analysis Program - Community Edition (BenMAP-CE): A tool to estimate the health and
economic benefits of reducing air pollution. Environ Modell Softw 104: 118-129.
http://dx.doi.Org/10.1016/i.envsoft.2018.02.009
8-17
DRAFT: Do Not Cite or Quote
-------
Smith. DH: Malone. DC: Lawson. KA: Okamoto. LJ: Battista. C: Saunders. WB. (1997). A national estimate of
the economic costs of asthma. Am J Respir Crit Care Med 156: 787-793.
http://dx.doi.org/10.1164/airccm. 156.3.9611072
Stanford. R: McLaughlin. T: Okamoto. LJ. (1999). The cost of asthma in the emergency department and
hospital. Am J Respir Crit Care Med 160: 211-215. http://dx.doi.Org/10.l 164/airccm. 160.1.9811040
Turner. MC: Jerrett. M: Pope. A. Ill: Krewski. D: Gapstur. SM: Diver. WR: Beckerman. BS: Marshall. JD: Su.
J: Crouse. PL: Burnett. RT. (2016). Long-term ozone exposure and mortality in a large prospective study.
Am J Respir Crit Care Med 193: 1134-1142. http://dx.doi.org/10.1164/rccm.201508-1633QC
U.S. EPA (U.S. Environmental Protection Agency). (2013). Regulatory impact analysis for the final revisions to
the National Ambient Air Quality Standards for particulate matter [EPA Report]. (EPA-452/R-12-005).
Research Triangle Park, NC: U.S Environmental Protection Agency, Office of Air Quality Planning and
Standards, Health and Environmental Impacts Division.
https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100G5UQ.txt
U.S. EPA (U.S. Environmental Protection Agency). (2014). Guidelines for preparing economic analyses [EPA
Report]. Washington, DC: U.S. Environmental Protection Agency, National Center for Environmental
Economics, https://www.epa.gov/environmental-economics/guidelines-preparing-economic-analvses
U.S. EPA (U.S. Environmental Protection Agency). (2019). Environmental Benefits Mapping and Analysis
Program - Community Edition (BenMAP-CE) (Version 1.5) [Computer Program], Washington, DC.
Retrieved from https://www.epa.gov/benmap/benmap-communitv-edition
U.S. EPA (U.S. Environmental Protection Agency). (2021). Technical Support Document (TSD) for the final
revised Cross-State Air Pollution Rule update for the 2008 ozone season NAAQS: Estimating PM2.5- and
ozone-attributable health benefits. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Air and Radiation, https://www.epa.gov/sites/production/files/2021-
03/documents/estimating pm2.5- and ozone-attributable health benefits tsd.pdf
Woods & Poole (Woods & Poole Economics). (2016). Complete demographic database [Database]. Washington,
DC. Retrieved from http://www.woodsandpoole.com/index.php
Zanobetti. A: Schwartz. J. (2009). The effect of fine and coarse particulate air pollution on mortality: A national
analysis. Environ Health Perspect 117: 1-40. http://dx.doi.org/10.1289/ehp.08001Q8
8-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
CHAPTER 9
INTEGRATED SYNTHESIS
9.1 Introduction
The focus of this chapter is to summarize and synthesize the information presented in the
previous chapters that either directly informed the quantitative analyses examining the air quality impacts
and corresponding health impacts of smoke from wildland fire (i.e., wildfire and prescribed fire) under
different fire management strategies, or provided ancillary information that allowed for the overall results
of the analyses to be put into the proper context.5 Overall, this assessment demonstrates the successful
application of a novel modeling approach in the examination of two case study fires to provide a
quantitatively estimate the differences in air quality and health impacts based on different fire
management strategies.
In theory, an assessment of the air quality impacts and the corresponding human health impacts of
prescribed fire compared to wildfire may seem relatively straightforward. However, the question is
layered with complexities in both the development of analyses and the interpretation of results due to
numerous factors including spatial and temporal differences between prescribed fire and wildfire along
with the overall management objectives of each (i.e., either suppression objectives or resource
objectives), which is dynamic and can change daily depending on various factors (e.g., fire behavior, as
detailed in CHAPTER 2 and CHAPTER 3). As a result, it is important to recognize that while the
analyses conducted within this report represent an incremental advancement in the overall understanding
of the health implications of smoke from wildland fire on surrounding populations, the results are based
on a novel modeling approach that required assumptions and decisions based on expert judgment,
particularly with respect to fire spread in the design of hypothetical scenarios for each case study.
The preceding chapters of this report were organized around characterizing the components that
are important to consider in the process of examining the air quality impacts and corresponding health
impacts of smoke from wildland fire under different fire management strategies. In estimating differences
between the air quality impacts of prescribed fire and wildfire, this assessment took a holistic approach of
identifying all of the factors and impacts (both positive and negative) that should be accounted for in the
process of examining different fire management strategies through the development of a conceptual
framework (CHAPTER 2; Figure 2-1, Figure 9-4 in this chapter). CHAPTER 3 through CHAPTER 8
then described the current state of the science with respect to implementing this framework with the goal
of employing the best available science and data to estimate as many of those impacts as feasible. A fuller
5 Within this assessment, the term "impacts" refers to the main quantitative results, which includes the estimated air
pollutant concentrations from the air quality modeling and the number of health events and associated economic
values calculated using U.S. Environmental Protection Agency's (U.S. EPA's) Environmental Benefits Mapping
and Analysis Program—Community Edition (BenMAP-CE). The term "effects" is used to denote the other positive
and negative consequences of wildland fire.
9-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
accounting of benefits and costs of fire management strategies, which is not the focus of this assessment,
would quantitatively address the remaining components of the conceptual framework, including
management costs, direct fire effects, and ecological effects.
While the results of this assessment are extremely informative in addressing the larger question of
whether there are differences in the public health impacts of wildland fire smoke for different fire
management strategies, it is also important to ensure the results are interpreted appropriately and to
recognize that the effects characterized represent only a portion of the broader societal, human health, and
ecological effects of wildland fire events. Therefore, subsequent sections of this chapter provide an
overview of the results of the analyses; broadly assesses the limitations and uncertainties surrounding the
examination of the air quality and corresponding public health impacts of prescribed fire and wildfire;
identifies the limitations and gaps in knowledge and data that shaped the implementation of the
conceptual framework (Figure 9-4); highlights key insights from the case study analyses; and outlines
additional areas of research that could further enhance the characterization of the impacts of smoke from
wildland fire.
9.2 Overview of Results
The overall goal of the case study analyses conducted within this assessment is twofold:
(1) develop a modeling framework to examine the air quality and health impacts of smoke from wildland
fire under different fire management strategies and (2) demonstrate the application of the modeling
framework for wildfires that encompass different spatial and temporal scales. Because the analyses
conducted within this assessment focus on wildfires of different spatial extent that occurred in two
different geographic locations (i.e., Oregon and California), it is important to clearly note that the results
are specific to the locations of the two case study fires and the land management practices that were used
prior to both fires occurring. Therefore, the results of these analyses cannot be extrapolated to other
geographic locations without consideration of differences in land management practices (including
history) and environmental variables (e.g., geography, vegetation, fire regime, climate, and weather).
Across each of the case studies, the air quality modeling, and subsequently health impact analyses
using U.S. Environmental Protection Agency's (U.S. EPA's) Environmental Benefits Mapping and
Analysis Program—Community Edition (BenMAP-CE), clearly depicts that air quality impacts attributed
to wildland fire smoke are dominated by changes in fine particulate matter (PM2 5; particulate matter with
a nominal mean aerodynamic diameter less than or equal to 2.5 |im) concentrations (see Section 5.3).
Although ozone is formed downwind of a smoke plume as a result of many of its precursors being
emitted in wildland fire smoke (see CHAPTER 4 and CHAPTER 5). the incremental contributions to
concentrations often do not result in substantial public health impacts. This is because the magnitude of
population-level health impacts depends on the intersection of smoke plumes that have elevated PM2 5 and
ozone concentrations over time with population density. Fires that result in smoke plumes, or elevated
9-2
DRAFT: Do Not Cite or Quote
-------
1 ozone concentrations downwind of a smoke plume, that do not intersect with high population areas or last
2 only a few days are less likely to have as substantial health impacts as fires affecting larger populations
3 for longer periods. This concept of duration of fire multiplied by population density represents the main
4 driver behind the difference in results between the Timber Crater 6 (TC6) and Rough fire case studies
5 discussed in more detail below.
6 As a reminder, both case study fires were selected because they occurred on federal land and
7 were managed by multiple federal agencies. Additionally, the TC6 Fire was selected because there is
8 extensive data that had been collected on the land management practices employed, including prescribed
9 fire activity, within the area, and in combination with the small size of the fire, allowed for a finer
10 resolution analysis (i.e., at the 4-km scale). In comparison, the Rough Fire was selected to provide an
11 examination of a larger fire, in terms of duration and size, but there was no actual prescribed fire activity
12 in the area. However, with the Sheep Complex Fire yielding positive resource benefits and detailed
13 information available on the proposed Boulder Creek Prescribed Fire it was possible to develop
14 hypothetical scenarios for the Rough Fire case study that were consistent with the hypothetical scenarios
15 developed for the TC6 Fire case study (i.e., a smaller and larger fire based on different land management
16 strategies).
9.2.1 Timber Crater 6 (TC6) Case Study
17 The analysis of the TC6 Fire case study focused on estimating the air quality and health impacts
18 attributed to the actual TC6 Fire, as well as hypothetical TC6 Fire scenarios based on assumptions
19 surrounding fire spread and fuel availability that was rooted in the detailed land management data for the
20 area (see Section 5.1.3). resulting in the following hypothetical scenarios:
21 • Hypothetical Scenario 1 (small): a smaller hypothetical TC6 Fire in a heavily managed area
22 (i.e., most prescribed fire), which would equate to a wildfire with less fuel, a smaller fire
23 perimeter, and less daily emissions
24 • Hypothetical Scenario 2a (large): a larger hypothetical TC6 Fire, but not the "worst-case"
25 scenario, due to no land management which would equate to a wildfire with more fuel, a larger
26 fire perimeter, and more daily emissions
27 • Hypothetical Scenario 2b (largest): a much larger, hypothetical "worst-case" scenario TC6 Fire
28 with no land management (i.e., no prescribed fire) which would equate to a wildfire with the most
29 fuel, largest fire perimeter, and largest daily emissions
30 Even with the detailed land management data available, in devising the hypothetical scenarios for this
31 case study, expert judgment was used to determine the daily fire perimeters and the overall burn perimeter
32 for each scenario, which was influenced by the prescribed fire history within the area.
33 One of the main differences between the TC6 Fire and Rough Fire case studies was the
34 availability of data on prescribed fire activity. Although there was information on the prescribed fire
9-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
activity within the vicinity of the TC6 Fire that could have impacted the spread of the fire, these fires
occurred over many years, with one dating back to 1978 (see Section 5.1.4V As a result, to compare the
prescribed fire smoke impacts with the actual TC6 Fire and hypothetical scenarios, all prescribed fire
activity was modeled for the same month and year (i.e., September 2019). This approach was used
because there was detailed data both on the days in September 2019 that fit prescription requirements and
for which a prescribed fire occurred. However, employing this strategy does not take into consideration
the rate of prescribed fire activity and ignores the episodic nature of prescribed fires compared to
wildfires, which is one of the overarching challenges of an analysis devised to compare the air quality and
health impacts of prescribed fire to wildfire (see Section 9.3.1).
The air quality modeling demonstrates that there are clear differences in the air quality impacts
between the actual TC6 Fire and each of the hypothetical scenarios, with the larger fire scenarios
(Hypothetical Scenarios 2a and 2b) resulting in higher concentrations for a longer duration, specifically
for PM2 5. This observation is also consistent when comparing the actual TC6 Fire and hypothetical
scenarios with the air quality impacts from the prescribed fires. The difference in the modeled air quality
impacts between the prescribed fires and the actual TC6 Fire and hypothetical scenarios can be attributed
to the short duration of each prescribed fire combined with the fact these fires were scheduled on days
that met specific criteria aimed at minimizing population exposure (e.g., meteorology conducive for
ventilation and dilution of pollutants).
While there are differences in air quality impacts across each of the scenarios examined, within
the vicinity of the TC6 Fire, population density is relatively small, and the examination of aggregate
population exposures which combines the influence of daily weather patterns, fire duration, and
population proximity to fires, shows that the overall potential public health impacts attributed to smoke
exposure would be small relative to larger fires (see Section 5.3.1; see Figure 5-10 and Figure 5-11). This
observation from the air quality modeling is reflected in the BenMAP-CE analysis for the actual TC6 Fire
and the hypothetical scenarios where the overall health impacts and corresponding economic values are
small (see Table 8-2i Table 8-3, and Table 8-4). From a health impact perspective while the overall values
are <1 for most health outcomes for PM2 5, and for all health outcomes for ozone across each fire type,
when examining the economic impact there is a more notable difference between the actual TC6 Fire,
prescribed fires, and each hypothetical scenario. This difference reflects the high value placed on
reductions in the risk of premature death. Even small changes in risk can have economic value because
one statistical premature death is valued at $9.5 million.
Although the small smoke-related health impacts from the TC6 Fire can be attributed to the small
population density within the case study area, the land management activities employed over time were
instrumental in reducing the fuel available and the overall fire perimeter, which equated to smaller air
quality impacts. Untreated forests within the TC6 Fire case study area are characterized by high fuel loads
(live and dead) that pose significant challenge to fire managers. Combined with hot, dry summers and few
natural barriers to fire spread, these spatially contiguous fuel loads create conditions ripe for large fire
9-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
growth. Baseline surface fuel loads (dead and down biomass) in untreated stands vary along a
productivity gradient, ranging from an average of 38 tons per acre in pure ponderosa pine to 46 tons per
acre in mixed ponderosa pine/lodgepole pine forests (the most widespread type), and up to 56 tons per
acre in the more productive upper elevation mixed conifer forest types (see Figure 9-1). Standing tree
densities (live and snags) averaged 881 to 2,899 trees per hectare across the same gradient. These
conditions were typical of what was encountered during the rapid initial growth of the TC6 Fire where no
fuels treatments had occurred. The extensive fuel treatment network employed in other parts of the area
prevented these conditions from occurring across the entire TC6 Fire footprint.
PP/LP LP MCL
Forest Type
TC6 = Timber Crater 6.
Note: Forest types from left to right are: PP = ponderosa pine (n = 4 plots), PP/LP = mixed ponderosa/lodgepole (n = 5),
LP = lodgepole pine (n = 9), MCL = lower mixed conifer (n = 13), MCU = upper mixed conifer (n = 8).
Source: National Park Service Long-Term Monitoring Plots (Farris. 2017).
Figure 9-1 Surface fuel loading in untreated forests in the Timber Crater 6
(TC6) Fire study area in Crater Lake National Park.
This high contemporary fuel loading within the TC6 Fire case study area is an artifact of more
than a century of ubiquitous fire exclusion in the region beginning in the late 1800s. Prior to about 1890,
fires were frequent across this landscape and resulted in limited broad-scale tree density and surface fuel
accumulation. Hagmann et al. (2019) recently conducted a detailed fire history study just 30 km east of
9-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
the TC6 Fire area consisting of nearly identical terrain and forest composition. They found that years in
which fires burned >20,000 hectares occurred every 9.5 years on average for the period 1700 to 1918, and
seven fire years burned >40,000 hectares. These large, predominantly low-intensity fires were associated
with drought years; more frequent but smaller fires occurred in interim periods. At a finer scale, fire
frequency and intensity varied along gradients of productivity and surface fuel continuity ranging from 7
to 25 years. Most high severity burning was restricted to pockets of dense lodgepole pine. This pattern
resulted in relatively small patch mosaics of denser stands within a matrix of open, low density stands.
This same gradient occurs in the TC6 Fire study area and is consistent with a description of fire
occurrence and fire severity mosaics by Agee (1981).
The cessation of frequent low intensity fires across the greater Pumice Plateau Ecoregion in
central Oregon began in the late 1800s but varied locally (Omcrnik and Griffith. 2014). Initially,
extensive fire exclusion occurred indirectly as a result of changing land uses (such as heavy grazing and
logging and development) and the displacement of Native American populations. Direct fire suppression
activities continued with the onset of formal fire suppression activities in the early 1900s. Heverdahl et al.
(2014) and Merschel et al. (2018) documented sharp declines in fire occurrence in similar forests
approximately 100 km north in the 1880s. Hagmann et al. (2019) also documented declines beginning in
the late 1800s adjacent to the TC6 Fire area. There was a single large fire in 1918, but even this fire did
not impact the TC6 Fire footprint.
The results of Heverdahl et al. (2014). Merschel et al. (2018). and Hagmann et al. (2019) are
consistent with local tree demography and recruitment data from Crater Lake National Park. In a
200-hectare study area burned by the TC6 Fire (along the U.S. Forest Service [USFS]/National Park
Service [NPS] boundary) Kipfmueller (2014) documented a major pulse in tree recruitment in the 1880s,
with high levels of recruitment continuing through the 1950s in the absence of fire (Figure 9-2). This
recruitment pulse likely reflects a widespread fire exclusion signature and is consistent with the onset of
major recruitment pulses elsewhere in Crater Lake National Park (Forrestel et al.. 2017). High
contemporary fuel loads in the study area are a direct legacy of this broad fire-exclusion recruitment
cohort. In the absence of 20th century fires, this cohort created high tree densities across the formerly
more heterogenous mosaic of productivity gradients. Many of these trees were converted to surface fuels
following density-dependent thinning and periodic insect outbreaks in recent years. Moreover, continuous
vertical fuel continuity from extensive ladder fuels create high crown fire initiation risk.
9-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
io
Lodgepole Pine
(n=996)
1450
15
1500
1550 1600 1650
1700
1750 1800 1850 1900 1950
2000
IO
Ponderosa Pine
(n=ll2)
1450
1500
1550 1600 1650
1700
1750 1800 1850 1900 1950
2000
Decade
ha = hectare.
Source: Kipfmueller (2014).
Figure 9-2 Decadal-scale representation of the age structure of lodegpole
pine (LP) and ponderosa pine (PP) aggregated for a 200-ha study
area within the Timber Crater 6 (TC6) Fire perimeter.
Despite the high potential for major fires in the area, fire suppression was largely successful in
the TC6 landscape as most ignitions were kept small (some of which were aided by the fire management
strategies employed within the area). For example, in 2011 alone, there were seven lighting fires
suppressed within 4 km of the TC6 Fire ignition. Thus, the only substantial burned acreage and reduction
in fuel loads on the NPS side resulted from management-ignited prescribed burning and a lightning fire
that yielded positive resource benefits (these efforts have been focused along the park boundary to
facilitate future management of more lightning fires). On the USFS side, a combination of prescribed
burns and mechanical fuels treatments have reduced fuel loads. Prescribed burning has reduced surface
fuel loads by an average of 20% in low-productivity ponderosa pine to an average of 69% in lodgepole
pine forests. Corresponding tree densities have been reduced by an average of 25% in ponderosa pine to
78% for lodgepole pine. The duration of treatments varies across a productivity gradient, but typically
reach 75% of prefire levels within 15 years on productive sites. The fire management challenges and
impacts of fire suppression fuel loading and potential fire behavior in the TC6 Fire area are similar to
other coniferous forest types in the western U.S.
9-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
9.2.2
Rough Fire Case Study
The Rough Fire was selected because it represented a much larger fire than the TC6 Fire in terms
of both area burned (i.e., -150,000 acres) and duration (i.e., lasting ~2 months), which directly influenced
the amount of smoke produced and the potential for a larger aggregate population exposure. However,
compared to the TC6 Fire, there was more limited data available regarding previous land management
practices within the vicinity of the Rough Fire to inform the development of hypothetical scenarios. As a
result, the hypothetical scenarios devised for the Rough Fire are not based on the same type of land
management strategies employed in the TC6 Fire case study. Specifically, the reliance on a wildfire that
burned at lower intensity and yielded positive resource benefits (i.e., Sheep Complex Fire) and a proposed
prescribed fire that did not occur as planned (i.e., Boulder Creek Prescribed Fire). However, the use of the
proposed Boulder Creek Prescribed Fire and the Sheep Complex Fire achieved the same function of being
able to devise Rough Fire hypothetical scenarios indicative of a smaller and larger Rough Fire,
respectively, due to different land management strategies. The hypothetical scenarios for the Rough Fire
consisted of the following:
• Hypothetical Scenario 1 (small): a small hypothetical Rough Fire that represents the combined
impact of the proposed Boulder Creek Prescribed Fire and the Sheep Complex Fire, a wildfire
that yielded positive resource benefits, on reducing the overall size of the Rough Fire
• Hypothetical Scenario 2 (large): a large hypothetical Rough Fire that allows for the fire perimeter
of the Rough Fire to progress into the area of the Sheep Complex Fire as if both the Boulder
Creek Prescribed Fire and Sheep Complex Fire did not occur
Similar to the TC6 case study, when examining air quality impacts for the actual Rough Fire and
each hypothetical scenario, overall aggregate population exposures are greatest for PM2 5 (Figure 5-15)
even though ozone concentrations in this case study impact a larger geographic area. This difference can
be attributed to ozone only being produced through secondary atmospheric reactions downwind from
smoke events, whereas, PM2 5 is not only directly emitted by fires, which represents the predominate
downwind exposure, but it can also be produced through secondary atmospheric reactions. For both PM2 5
and ozone a similar temporal pattern of concentrations is observed between the actual Rough Fire and
hypothetical scenarios until later weeks in the duration of each fire, where there is a substantial reduction
in concentrations for Hypothetical Scenario 1 (small fire, Figure 5-17). Although there was not an actual
prescribed fire in the vicinity of the Rough Fire, air quality analyses of the proposed Boulder Creek
Prescribed Fire (Figure 5-19) and the Sheep Complex Fire (Figure 5-20), exhibit a shorter duration and
smaller exposure to PM2 5, respectively, compared to the actual fire and hypothetical scenarios.
The differences in the public health impacts between the actual fire, hypothetical scenarios, a
prescribed fire (i.e., Boulder Creek Prescribed Fire), and a wildfire that yielded positive resource benefits
(i.e., Sheep Complex Fire) are depicted in Table 8-2. The health impacts of the actual Rough Fire, which
reflects the occurrence of the Sheep Complex Fire, are relatively similar to hypothetical Scenario 2 (large
fire), which assumes the Sheep Complex Fire did not occur. This similarity can be attributed to the Sheep
Complex Fire not substantially affecting the overall spread and fire perimeter of the actual Rough Fire.
9-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
However, the results of Hypothetical Scenario 1 (smaller fire) demonstrates the potential benefit that
could occur, specifically the reduction in fire spread and perimeter, by strategically planning the location
of a prescribed fire. The modeling of the Boulder Creek Prescribed Fire shows that had that prescribed
fire occurred on the outskirts of the Sheep Complex Fire perimeter, it could have prevented the spread of
the Rough Fire, reducing air quality impacts and resulting in an approximate 40% reduction in health
impacts. However, it is important to recognize that both the Sheep Complex Fire and the Boulder Creek
Prescribed Fire scenarios did have detrimental effects on both air quality and health, although those
effects were smaller than those estimated for the actual Rough Fire and each hypothetical scenario.
In addition to the air quality and health impacts observed between the different hypothetical
scenarios of the Rough Fire case study, it is also important to take into consideration the impact of
different land management strategies on the forest ecology around the case study area. Beyond the
analogous examples in other parts of the U.S., the particular fire ecology and history of the dry forests of
the Sierra Nevada Mountain offer more context for the analysis of the Rough Fire area, and illuminates
how the results of the Rough Fire case study might be used to further understand how to minimize air
quality impacts from wildfire smoke, both in this area and in other dry forest regions. There is substantial
fuel available in these large, highly productive, west-facing Sierra Nevada drainages that can be released
into the air all at once, as witnessed during the Rough, Rim, and any number of megafires (i.e., fires with
>100,000 acres burned).
The spatial configuration, not just the amount of fuel is important as well. Fire adapted forest
stands are characterized not only by lower fuel loads, but fuels that are "packaged" into fire
adapted-clumps (also known as resilient forest structure) with gaps in between those clumps, resulting in
fire that burns more slowly across the land, rather than all at once. At larger landscape scales, a mosaic of
frequent, smaller, slower growing fires can contribute to reducing the number of megafires. Historically
and prehistorically there is overwhelming evidence that the forests of the Sierra Nevada, including the
area where the Rough Fire burned, experienced frequent burning. Local fire-scar chronologies indicate
that most fire years prior to the 20th century were characterized by relatively small, spatially clustered fire
events that were even smaller than the Sheep Complex Fire (Figure 9-3). Widespread fire events also
occurred periodically during severe droughts, as indicated by synchronous scarring across multiple
sample sites (Swetnam et al.. 2009). While there is uncertainty about the size or extent of these large fire
years, a major difference versus contemporary large fires is that they consisted predominantly of
low-severity burning (Mallek et al.. 2013).
9-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Fire History - Ponderosa Pine-Mixed Conifer Forest
Settlement Suppression
1 1 1 1 II 1
1
1
1
1 1 1 1 II 1
1
1
1
1 1 1 1 1 1 ? ^
r
1
1 |
II 1 1 1 1 Z Q)
1 "I
1 1
1
1
I I I I
1 1
I
l
1
II
1 1 ""
1 1
1
1 1 1 1
l i
II
1
1
II
1
I 4 m
I r
1
1 1
II
I
i
1 |
I 6 *
i i
i
1
1 1
I
1 1 M il M il II
1 1 1 1
1 1 1 1 | 1 1 1 1 | 1 1 1 1 | 1 1 1 1 | 1 1 1 1 | 1 1 1 1 | 1 1 1 1 | 1 1 1 1 |
1700 1800 year 1900 2000
Note: Reconstruction of past fire occurrence (tic marks) from fire scarred trees at six sites in the mixed conifer zone from
1700-2000.
Source: Sequoia & Kings Canyon National Parks (2005). copyright permission pending.
Figure 9-3 Decline in fire frequency in mixed conifer forest (from nearby
Sequoia and Kings Canyon National Parks) starting around 1860.
Impacts to air quality from these fires would have likely been similar in intensity, duration, and
spatial extent to impacts from the modeled Boulder Creek Prescribed and/or Sheep Complex fires, rather
than the Rough Fire, both because such fires spread more slowly and because the fuels over the area in
which they burned were substantially less than those currently observed in areas where fuels have
accumulated after 100 years of fire suppression (Stephens et al.. 2018). Additionally, these smaller,
frequent fires created a landscape-scale mosaic of fire footprints, wherein fires were limited in their size
by the footprints (and the removal of fuel within those footprints) of previous recent fires (Collins et al..
2009).
The hiatus from regular fire for the past 100 years has left substantial accumulated fuel on Sierra
Nevada forested landscapes, and as a result, frequent, small, regular fires are not feasible, leading to a
high potential for megafires (Stephens et al.. 2018; Liu et al.. 2016). The Sheep Complex Fire, compared
to these megafires like the Rough Fire, was quite small, but the cool, wet conditions under which it was
managed limited its ability to spread despite that fuel loading. Resulting air quality and public health
impacts were limited directly during the burning of Sheep Complex in 2010, but also contributed to some
reductions in impacts subsequently as the Rough Fire ran into its footprint in 2015. This illustrates the
principle that even limited and opportunistic reintroduction of fire to a landscape can reduce the overall
footprint of future fires, resulting in quantifiable air quality and public health benefits.
So far, at least in this case study, this work appears to qualitatively corroborate previous case
study analyses [e.g., Long et al. (2018); Schweizer and Cisneros (2014); Cisneros et al. (2012)1 showing
that daily emissions were much lower compared to those during the Rim fire, which, like the Rough Fire,
was ultimately contained by leveraging reduced fuels and fire behavior in previous fire footprints (Long
et al.. 2018). A limitation of the Rough Fire analyses is the regional-scale resolution (12-km-sized grid
9-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
cells) of the air quality modeling. This spatial resolution may not fully capture pollutant dispersion in
areas with complex terrain, such as the area of the Rough Fire, Sheep Complex Fire, and Boulder Creek
Prescribed Fire. When the model does not capture complex meteorology it is possible emissions from a
fire could be unrealistically dispersed over a larger area than would happen in reality and result in an
overestimation of impacts downwind of the fire and underestimate impacts at the fire itself. Implications
for this analysis would depend on the degree of over or underestimation of impacts in highly populated
areas of the central valley of California. Future work using higher resolution modeling (e.g., 2-km
resolution), and including a robust comparison of model predictions of PM2 5 to observed could provide a
more refined assessment of the magnitude of trade-offs between the Rough Fire scenarios presented
within this assessment.
In summary, in dry forest ecosystems, such as in the area of the Rough Fire, these landscapes will
experience some combination of prescribed fire and wildfire. The methodology for assessing public
health trade-offs of different fire management strategies developed in this assessment, if deployed on a
broader scale, landscape level analysis, could inform development of management strategies that
incorporate protection of regional air quality and public health.
9.3 Limitations in Examining Differences between Prescribed
Fire and Wildfire Impacts
Throughout this assessment, each chapter characterized the various components of the conceptual
framework presented in CHAPTER 2 (Figure 2-1, and also presented below Figure 9-4) to varying
degrees with some presenting a qualitative characterization of the state of the science, and others
providing a quantitative analysis specific to the case study areas. In identifying limitations in the analyses,
it is first necessary to review the information presented within each chapter and note which components
of the conceptual framework could be addressed broadly and more specifically within each of the case
study analyses (Section 9.3.1.). This approach then allows for a discussion of the overarching limitations
of the analysis (Section 9.3.2.) followed by a discussion of current gaps in the scientific literature that
were identified within this assessment (Section 9.3.3.). As the frequency of wildfires continues to grow,
along with the frequency of prescribed fire as a land management strategy, it is important to consider
these limitations and data gaps in the process of further refining the types of analyses conducted within
this report and in advancing the overall understanding of the impacts of wildland fires.
9-11
DRAFT: Do Not Cite or Quote
-------
Improved forest health
Ability to mitigate
impact
Resource Benefits
Ability to mitigate
exposure
Human exposure
Non-smoke fire impacts
Watershed integrity
Firefighter health and
safety
Ability to mitigate
impacts
Wildfire
Severity/Extent Conditional on
Management Decision
Direct and indirect
economic damages
Prescribed fire
Ecosystem exposure
Mechanical thinning
Ecological impacts
Probability of wildfire
ignition
No action
Ecosystem impacts
Non-fire adverse
impacts
Smoke emissions
Costs of management
actions
e.g. equipment and labor costs, fire
suppression costs, etc.
GHG emissions
Note: This is the same figure presented in CHAPTER 2. Figure 2-1. In the figure, forest management inputs are colored dark blue, management decisions and their nonsmoke related
effects are colored white, resource benefits are colored green, mitigation actions are colored light blue, fires are colored yellow and orange, fire damages are colored red, and smoke
exposure related elements are colored gray. The green arrows indicate positive effects, and the orange arrows indicate negative effects. Dotted lines represent linkages that may occur
but are less certain that solid lines.
Figure 9-4 Conceptual framework for evaluating and comparing fire management strategies.
9-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
9.3.1
Implementing the Conceptual Framework
The ability to implement the conceptual framework, originally outlined in CHAPTER 2. and the
degree to which quantitative information specific to the case study areas is available represents a key
aspect of the quantitative estimation of air quality impacts associated with different fire management
strategies. Each chapter presents information that is highly relevant to an assessment of the air quality
impacts between different fire management strategies; however, in many instances this information is not
specific to the case study areas and requires some degree of extrapolation.
Within this assessment, qualitative discussions are presented for multiple components of the
conceptual framework due to a lack of quantitative information specific to the case study areas. Moving
from left to right across the conceptual framework (Figure 2-1, and also Figure 9-4), CHAPTER 3
captures many of these initial components. This includes the baseline forest/ecological conditions of
ecosystems similar to the case study areas, provides background information on different fire
management decisions, and a history of fire activity, including the implementation of prescribed fire. In
addition, the qualitative discussion in CHAPTER 3 highlights the instances where a wildfire can yield
resource benefits, which are quantitatively evaluated in the Rough Fire case study through the modeling
of the Sheep Complex Fire (CHAPTER 5 and CHAPTER 8). and discusses how fire on the landscape can
contribute to improved forest health and result in ecological benefits.
The direct fire impacts of wildfire (CHAPTER 7). including firefighter health and safety, and
societal impacts including economic and ecological and welfare effects, while important to consider
broadly when making comparisons amongst different fire management strategies cannot be quantified at
the case study level. Although there are opportunities to mitigate these direct fire effects, they are not
accounted for within this assessment. The nonfire effects, which include greenhouse gas (GHG)
emissions (CHAPTER 3) and ash deposition (CHAPTER 6). are characterized qualitatively to varying
degrees, including the ecological effects of ash deposition.
The smoke emissions and corresponding modeling of air quality impacts (CHAPTER 5) represent
the key inputs to the quantitative analyses that form the backbone of this assessment. The results of the air
quality modeling directly inform both human and ecosystem exposure with only the resulting human
health impacts being quantitatively examined. However, this assessment also provides a qualitative
discussion of both health and ecosystem impacts attributed to smoke exposure (CHAPTER 6V The
current understanding of the health effects of wildland fire smoke exposure, as well as ambient PM2 5 and
ozone exposure, are subsequently used within BenMAP-CE to quantify the number of deaths and
illnesses attributed to smoke from the different scenarios examined within both case studies.
Additionally, scientific evidence supports the availability and efficacy of various actions and
interventions that can be employed at the individual and community level to mitigate the public health
impact of smoke exposure (CHAPTER 6). The overall population PM2 5 exposure reductions estimated
A-9-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
from these actions and interventions allows for a limited quantitative assessment of the potential public
health implications of promoting such measures (CHAPTER 8 V Although these actions and interventions
can be instituted for both wildfires and prescribed fires, the planned nature of prescribed fires enhances
opportunities for public engagement surrounding prescribed fires, and increases the opportunities to
inform populations at risk of wildfire smoke-related health effects of actions they can take to protect
themselves. In addition to the quantitative and qualitative discussions that directly support components of
the conceptual framework, this assessment also presents an overview of the current state of air quality
monitoring for wildland fire smoke. Although the discussion of air quality monitoring does not represent
a defined component of the conceptual framework, it is important to consider in the process of
interpreting both the air quality modeling output and epidemiologic studies examining the health effects
of smoke, which are the key inputs to the estimation of health impacts.
9.3.2 Overarching Limitations
As thoroughly detailed in CHAPTER 2. and noted in the previous section, the overall conceptual
framework for conducting this assessment identifies numerous factors to consider in examining trade-offs
between different fire management strategies, including prescribed fire, and the resulting effects, both
positive and negative. While many of these factors are characterized within this assessment, there are
spatial and temporal dimensions of fire management strategies that are not addressed. In addition, this
assessment does not assess the effect of fire management strategies on the probability of wildfire
occurrence (i.e., ignition probability), which is potentially a key factor in assessing differences in the
cumulative effects of those strategies as depicted in Equation 2-1 in CHAPTER 2. As recently discussed
in Hunter and Robles (2020). the comparison of positive and negative effects of prescribed fire and
wildfire is not a static comparison, but one that should be conducted by considering the spatial and
temporal aspect of prescribed fires and their interaction with the likelihood, severity, and magnitude of
wildfire over a specific time horizon.
In comparison to wildfires, which occur at one uncertain point in time, but can vary in length
from a few days to months, prescribed fires occur at planned times episodically over many years.
Prescribed fires are conducted to achieve a resource benefit (see CHAPTER 3). with one of the
overarching assumptions being that the prescribed fire will contribute to reducing the effect (e.g., size and
severity) of a future wildfire. However, to achieve this desired outcome requires a series of prescribed
fires over time that provide a patchwork of areas with less fuel, not an individual fire on its own, to
minimize the risk of a severe, catastrophic wildfire occurring within the vicinity of the prescribed fires
(see Figure 9-5).
A-9-14
DRAFT: Do Not Cite or Quote
-------
9
MAINTAINED PRESCRIBED FIRE
< 1 YEAR OLD
PRESCRIBED FIRE
< 1 YEAR OLD
PRESCRIBED FIRE
< 5 YEARS OLD
WILDFIRE
< 1 YEAR OLD
YEAR 5
YEAR 10
Source: Hunter and Robles (2020), copyright permission pending.
Figure 9-5 Conceptual diagram presented by Hunter and Robles (2020 for
assessing the impacts of prescribed fire compared to wildfire.
Fully accounting for the trade-offs of smoke impacts between prescribed fire and a wildfire
requires an understanding of the intersection of prescribed fire activity (both the total number of
prescribed fires and the frequency of prescribed fires) with a wildfire. While over a long enough time
period, the probability that a specific location will experience a wildfire can be substantial, there is still
uncertainty as to when that fire would occur and how severe it would be. Therefore, although prescribed
fires may reduce both the ignition probabilities and the severity of fires, they produce smoke that may, or
may not, have occurred due to a potential future wildfire. Focusing the analyses conducted within this
assessment around two previous wildfires and the land management strategies associated with each, did
not allow for the consideration of ignition probabilities along with the total number and frequency of
prescribed fires required to minimize the effects of a wildfire. Instead, these case studies address
hypothetical scenarios by asking how the effects of fires that did occur might have differed under
different types of fire management strategies. The information provided by these case studies is
informative in assessing the benefits of different fire management strategies given the occurrence of fire,
but does not address the uncertainty in the time horizon for fire in the landscape, nor the cumulative
effects on health of a series of prescribed fire activities.
A-9-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Although for the TC6 Fire case study, there was some information on prescribed fire activity over
time, the time window over which these fires occurred complicated the ability to conduct a direct
comparison of smoke impacts between prescribed fire and wildfire. As a result, for the TC6 Fire case
study, it is assumed that all prescribed fire activity, and subsequent smoke exposures, occurred at one
point in time (i.e., September 2019). For the Rough Fire case study the examination of prescribed fire
activity is purely hypothetical, as there was no actual prescribed fire activity in the vicinity of the fire.
However, by modeling the proposed Boulder Creek Prescribed Fire as if it actually occurred does provide
some indication of the potential impact of a prescribed fire on reducing the size of the actual Rough Fire.
Therefore, for both case studies, exposure to prescribed fire smoke is being treated as a static event and
not the episodic event it is in actuality.
The treatment of prescribed fires as events occuring at one point in time within this assessment,
out of both analytical convenience and sparseness of available data, also has ramifications from a health
perspective. The removal of the spatial and temporal pattern of prescribed fire activity does not allow for
the analyses conducted to consider that the location of prescribed fires varies on a year-to-year basis. By
excluding this variability in prescribed fire activity, it is not possible to account for the corresponding
spatial and temporal variability in population exposures to smoke that would occur, which could
potentially result in a different pattern of health impacts.
In addition to recognizing the spatial and temporal aspects of prescribed fires and wildfires, it is
also imperative to highlight the vastly different landscapes, in terms of both ecosystem composition
(e.g., forests vs. prairie) and the percent contribution of prescribed fire to total wildland fire activity
across the U.S. (Figure 9-6). The regional variability in the number of acres burned by prescribed fire and
wildfire nationally, specifically in areas with a higher percentage of prescribed fires such as the Southeast,
is an additional important consideration when examining air quality impacts associated with different fire
management strategies. The variability in the composition of fire activity nationally, clearly depicts why
the results of the case study analyses are not easily transferable to other parts of the country, especially to
areas where the number of acres burned is dominated by prescribed fires. Lastly, as noted earlier in this
section, the relationship between prescribed fires and wildfire ignition probabilities are unknown in the
case study areas and it is unclear how this relationship varies nationally, particularly in locations
dominated by prescribed fires.
A-9-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Source: Baker et al. (2020). copyright permission pending.
Figure 9-6 Acres burned by wildfire (red) and prescribed fire (green) in the
U.S. in 2017.
9.3.3 Identified Data Gaps and Uncertainties
In the process of developing the preceding chapters of this assessment, as well as the
development of the main modeling framework for the air quality and health impact analyses, gaps were
identified in the current scientific understanding of wildland fire smoke. Future efforts to collect data and
conduct studies to fill in these gaps could aid in future assessments and allow for a more extensive
quantitative estimation of impacts and trade-offs between prescribed fire and wildfire.
A main overarching data gap that filters into multiple aspects of this assessment, but does not
represent a key component of the conceptual framework, is the availability of ground-level air quality
monitoring data for wildfire smoke. The challenges associated with monitoring wildfire smoke (see
Section 4.5). and the resulting paucity of monitoring data, represents an important data gap because air
quality monitoring data is instrumental in the assessment of health effects through epidemiologic studies
as well as in air quality modeling to validate model predictions.
Even without a dense monitoring network to more fully capture the temporal and spatial patterns
of population-level exposures to wildfire smoke, epidemiologic studies have still been able to use
available air quality data (e.g., satellite, modeling, etc.) to assess the health effects of wildfire smoke.
While these studies have been extremely informative and valuable to build upon the broad understanding
of the health effects of ambient exposures to PM2.5 and ozone, uncertainties remain with respect to both
A-9-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
exposure assessment as well as a broader understanding of the health implications of exposures to
different durations of wildland fire smoke (e.g., repeated peak exposures over many days, exposures over
multiple fire seasons). Additionally, as reflected in the sensitivity analysis conducted in CHAPTER 8
(Section 8.3.2). additional epidemiologic studies that more fully capture wildfire smoke exposure can help
inform the concentration-response (C-R) relationship to better understand if there are differences
compared to the C-R relationship for ambient PM2 5 exposures that should be considered when examining
the public health impacts of smoke based on different fire management strategies. In addition, better
understanding of the differences in the composition of smoke resulting from different burn conditions
(e.g., fuel characteristics, moisture levels, and the health effects associated with different smoke
composition) can help improve the ability to differentiate between fire management strategies with and
without prescribed fire, and also strategies for designing prescribed fire programs to minimize negative
health impacts.
In considering the approach used within this assessment for the air quality modeling, the
assumptions that factored into the methods employed recognize the same overarching limitations
discussed in Section 9.3.2 (see Section 5.4). As noted earlier within this chapter, expert judgment was
relied upon heavily in the defining of the hypothetical scenarios for each of the case studies. In addition,
in the modeling of prescribed fires for both case studies, all prescribed fire activity over many years was
modeled for 1 month in the instance of the TC6 Fire case study or there was no prescribed fire activity in
the case of the Rough Fire case study, resulting on the reliance of a proposed prescribed fire that never
occurred. Results of analyses, such as those conducted within this assessment, could more fully capture
the differences between different land management and fire management strategies through data that can
capture the temporal and spatial scale of prescribed fire activity. Although a fuller accounting of
prescribed fire activity overtime and space is a key data gap, it also remains unclear how prescribed fire
activity could impact the size and duration of a wildfire. The relationship between prescribed fire activity
and its influence on wildfire size and duration, especially for larger fires (e.g., Rough Fire) represents a
key area that requires additional exploration and prevents extrapolation of results from these case studies
to other parts of the U.S.
In addition to the data gaps identified within this section, there are numerous ancillary issues
associated with wildfires that are not addressed within this assessment, but this does not diminish their
importance. For example, it is recognized that wildfires can lead to the resuspension of legacy pollutants,
such as asbestos, lead, and mercury. These pollutants have been shown to lead to a range of health effects,
but it remains unclear how much wildfires contribute to population-level exposures to these pollutants.
Additionally, over time the wildland-urban interface (WUI) has expanded rapidly in many parts of the
U.S. (Radeloff et al.. 2018). This expansion of the WUI has resulted in substantial portions of the
population now residing in locations that are considered high-fire-risk areas. The growth of the WUI not
only increases the risk of fire ignitions, but also direct fire effects. Although CHAPTER 7 broadly
captures direct fire effects, including those associated with the burning of structures that could be
experienced within the WUI, currently available information is not conducive to providing
A-9-1.8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
location-specific estimates of the costs of wildfire. Lastly, as wildfires infringe upon the WUI it can lead
to a change in the composition of smoke as homes and structures are burned and the likelihood of
populations being exposed to wildfire smoke.
9.4 Key Insights from Case Study Analyses
This assessment and the accompanying quantitative analyses, represent an incremental
advancement in the understanding of the air quality and health impacts of wildland fires under different
fire management strategies. As a reminder, the results of the analyses conducted within this assessment
are specific to the case study areas and are not intended to represent the air quality and health impacts that
would be observed in other locations around the U.S. The case studies were chosen to illustrate the type
and nature of air quality and health impacts associated with different fire management strategies.
Additionally, within this assessment it is important to reiterate that in examining the air quality and health
impacts attributed to prescribed fires, the analyses are retrospective and represent locations that
experienced a wildfire, and therefore, do not (1) account for the temporal and spatial variability of
prescribed fires occurring over many years that happens in reality or would happen in an ideal situation to
minimize the risk of catastrophic wildfire and (2) incorporate an estimate of uncertainty to account for the
probability that a wildfire may not occur in a location where there was prescribed fire activity. The case
study analyses conducted within this assessment support the following observations:
• To provide a reasonable estimation of air quality and health impacts from wildland fire,
location-specific information on fuels is needed to support air quality modeling.
• Smoke impacts on health are dependent upon population proximity to wildland fire events.
• Predicted concentrations of PM2 5 from prescribed fires are smaller in magnitude and shorter in
duration, and the estimated aggregate population exposure is smaller than for each hypothetical
scenario and the actual fires in both case studies.
o The smaller estimated aggregate population PM2.5 exposures for prescribed fires in both
case studies can be attributed to the small spatial extent of each prescribed fire and the
meteorological characteristics of the days in which the prescribed fires occurred.
o Although prescribed fires are timed for days with specific meterological conditions to
reduce population exposures to smoke, analyses show that air quality and public health
impacts are still observable.
• Within the case study areas, ozone produced from wildland fires is shown to have less impacts on
air quality and public health, providing additional support to the current public health focus being
on reducing exposures to PM2 5.
• Wildfires that are short in duration and size and not near large population centers, such as the
TC6 Fire, can still result in public health impacts, albeit substantially smaller than larger
wildfires, such as Rough Fire.
A-9-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
• Wildfires that yield positive resource benefits on their own (i.e., Sheep Complex Fire) could be
more effective in reducing future air quality and public health impacts when used in combination
with prescribed fires (i.e., Boulder Creek Prescribed Fire).
o Well designed prescribed fires targeted for specific locations, such as the proposed
Boulder Creek Prescribed Fire, along with the prescribed fires around the TC6 Fire, can
potentially reduce the size and resulting air quality and public health impacts of future
wildfires.
• Communicating the benefits of actions and interventions that reduce or mitigate PM2 5 exposures
can contribute to reducing the public health impacts attributed to wildland fire smoke if these
actions are more widely used by the population.
9.5 Future Directions
The analyses conducted within this assessment lay the foundation for future research efforts to
examine the air quality and corresponding public health impacts of smoke from wildland fire under
different fire management strategies. While the results of the quantitative analyses provide initial
evidence of differences in smoke impacts between prescribed fire and wildfire, additional research efforts
that attempt to address the following issues will further enhance the applicability of future analyses
examining the trade-offs between different fire management strategies:
• Identification and development of methods to account for the temporal (i.e., frequency) and
spatial component of prescribed fires and their relationship with wildfires. Related to this
advancement would include gaining a better understanding of how to capture the health effects of
repeated exposure to smoke from prescribed fires over many years and how that compares to the
health effects experienced during singular wildfire events.
• Enhanced characterization of the relationship between prescribed fire and wildfire on the
landscape. This would include analyses that examine specific spatial domains with prescribed
fires and the number of those locations that also experienced a wildfire, along with identifying
whether prescribed fires were able to reduce characteristics of the wildfire (e.g., size, intensity,
duration, etc.). This advancement would then allow for a greater understanding of the costs and
benefits of different fire management strategies with and without wildfire.
• Analyses that characterize the role of topography and meteorology, in combination with the
frequency of prescribed fires within a spatial domain, on the potential for population centers to
experience smoke impacts from wildland fires.
• Characterization of how air quality impacts differ between prescribed fire and wildfire in
different parts of the U.S., specifically in locations where prescribed fire is the dominant wildland
fire activity, to gain a better understanding of the ability to extrapolate results across geographic
locations.
• Centralized repository to capture prescribed fire data to enhance future assessments using more
recent data. Such a repository would include, but not be limited to, information on location,
timing (dates and approximate start and end time), actual acres burned, fuel type and loading
information, and any air quality monitoring data collected.
A-9-20
DRAFT: Do Not Cite or Quote
-------
1 In addition to these broad areas that require additional research to support future analyses, there
2 are overarching uncertainties and limitations identified in previous chapters that could further enhance our
3 understanding of the overall impacts of wildland fire smoke. These areas of additional research include
4 enhanced air quality monitoring capabilities for wildfire smoke, better characterization of wildland fire
5 smoke exposures for health studies, additional understanding of the health effects of wildfire smoke over
6 many seasons, and a fuller accounting for the role of public health actions and interventions in reducing or
7 mitigating wildland fire smoke exposure. Collectively, these broader research initiatives in combination
8 with those areas this assessment was unable to account for, noted within this section, would allow for a
9 fuller characterization of the air quality and health impacts due to different fire management strategies.
A-9-21
DRAFT: Do Not Cite or Quote
-------
9.6 References
Agee. JK. (1981). Initial effects of prescribed fire in a climax Pinus contorta forest: Crater Lake National Park.
(CPSU/UW 81-4). Seattle, WA: University of Washington, College of Forest Resources, Cooperative Park
Studies Unit, http://www.craterlakeinstitute.com/research-at-crater-lake/wildfire-at-crater-lake/effects-
prescribed-fire-lodgepole-forests/
Baker. K: Rao. V: Beidler. J: Vukovich. J: Koplitz. S: Avev. L. (2020). Illustrating wildland fire air quality
impacts using an EPA emission inventory [Magazine]. EM: Environmental Manager, 24, 26-31.
Cisneros. R: Schweizer. D: Zhong. S: Hammond. K: Perez. MA: Guo. O: Traina. S: Bvtnerowicz. A: Bennett.
PH. (2012). Analysing the effects of the 2002 McNally fire on air quality in the San Joaquin Valley and
southern Sierra Nevada, California. International Journal of Wildland Fire 21: 1065-1075.
http://dx.doi.org/10.1071/WF11025
Collins. BM: Miller. JD: Thode. AE: Kelly. M: van Wagtendonk. JW: Stephens. SL. (2009). Interactions among
wildland fires in a long-established Sierra Nevada natural fire area. Ecosystems 12: 114-128.
http://dx.doi.org/10.1007/sl0021-008-9211-7
Farris. C. (2017). Crater Lake National Park FFI fire effects monitoring data: U.S. Department of the Interior,
National Park Service. Retrieved from https://irma.nps.gov/DataStore/Reference/Profile/2239780
Forrestel. AB: Andrus. RA: Fry. PL: Stephens. SL. (2017). Fire history and forest structure along an elevational
gradient in the southern Cascade Range, Oregon, USA. Fire Ecology 13: 1-15.
http://dx.doi.org/10.4996/fireecology.1301001
Hagmann. RK: Merschel. AG: Reillv. MJ. (2019). Historical patterns of fire severity and forest structure and
composition in a landscape structured by frequent large fires: Pumice Plateau ecoregion, Oregon, USA.
Landsc Ecol 34: 551-568. http://dx.doi.org/10.1007/sl0980-019-00791-l
Heverdahl. EK: Loehmaa RA: Falk. DA. (2014). Mixed-severity fire in lodgepole-dominated forests: Are
historical regimes sustainable on Oregon's Pumice Plateau, USA? Can J For Res 44: 593-603.
http ://dx. doi.org/10.113 9/cjfr-2013-0413
Hunter. ME: Robles. MP. (2020). Tamm review: The effects of prescribed fire on wildfire regimes and impacts:
A framework for comparison [Review]. For Ecol Manage 475: 118435.
http://dx.doi.Org/10.1016/j.foreco.2020.118435
Kipfmueller. KF. (2014). Demography of a lodgepole pine-ponderosa pine forest ecotone in Crater Lake
National Park. Final report. (Great Lakes-Northern Forest Cooperative Ecosystem Studies Unit
J9322080905). Klamath County, OR: On file at Crater Lake National Park.
Liu. JC: Micklev. LJ: Sulprizio. MP: Dominici. F: Yue. X. u: Ebisu. K: Anderson. GB: Khan. RFA: Bravo. MA:
Bell. ML. (2016). Particulate air pollution from wildfires in the Western US under climate change. Clim
Change 138: 655-666. http://dx.doi.org/10.1007/sl0584-016-1762-6
Long. JW: Tarnav. LW: North. MP. (2018). Aligning smoke management with ecological and public health
goals. J Forest 116: 76-86. http://dx.doi.org/10.5849/jof.16-042
Mallek. C: Safford. H: Viers. J: Miller. J. av. (2013). Modern departures in fire severity and area vary by forest
type, Sierra Nevada and southern Cascades, California, USA. Ecosphere 4: 1-28.
http://dx.doi.org/10.1890/ES13-00217
Merschel. AG: Heverdahl. EK: Spies. TA: Loehman. RA. (2018). Influence of landscape structure, topography,
and forest type on spatial variation in historical fire regimes, Central Oregon, USA. Landsc Ecol 33: 1195 -
1209. http://dx.doi.org/10.1007/sl0980-018-0656-6
Omernik. JM: Griffith. GE. (2014). Ecoregions of the conterminous United States: Evolution of a hierarchical
spatial framework. Environ Manage 54: 1249-1266. http://dx.doi.org/10.1007/s00267-014-0364-l
A-9-22
DRAFT: Do Not Cite or Quote
-------
Radeloff. VC: Helmers. DP: Kramer. HA: Mockria MH: Alexandre. PM: Bar-Massada. A: Butsic. V:
Hawbaker. TJ: Martinuzzi. S: Syphard. AD: Stewart. SI. (2018). Rapid growth of the US wildland-urban
interface raises wildfire risk. Proc Natl Acad Sci USA 115: 3314-3319.
http://dx.doi.org/10.1073/pnas. 1718850115
Schweizer. D: Cisneros. R. (2014). Wildland fire management and air quality in the southern Sierra Nevada:
Using the Lion Fire as a case study with a multi-year perspective on PM2.5 impacts and fire policy. J
Environ Manage 144: 265-278. http://dx.doi.org/10.1016/jjenvman.2014.06.007
Sequoia & Kings Canyon National Parks. (2005). Fire and fuels management plan, 2004. U.S. Department of the
Interior, National Park Service, https://www.nps.gov/seki/learn/nature/upload/seki_ffmp_fmp_v0505.pdf
Stephens. SL: Collins. BM: Fettig. CJ: Finney. MA: Hoffman. CM: Knapp. EE: North. MP: Safford. H:
Wavman. RB. (2018). Drought, tree mortality, and wildfire in forests adapted to frequent fire. Bioscience 68:
77-88. http://dx.doi.org/10.1093/biosci/bixl46
Swetnam. TW: Baisan. CH: Caprio. AC: Brown. PM: Touchan. R: Anderson. RS: Hallett. DJ. (2009). Multi-
millennial fire history of the giant forest, Sequoia National Park, California, USA. Fire Ecology 5: 120-150.
http://dx.doi.org/10.4996/fireecology.0503120
A-9-23
DRAFT: Do Not Cite or Quote
-------
APPENDIX
A.1. SUPPLEMENTAL INFORMATION FOR CHAPTER 1
No supplemental information.
A.2. SUPPLEMENTAL INFORMATION FOR CHAPTER 2
Table A.2-1 represents a more detailed version of Table 2-1 that attempts to characterize whether
the impacts associated with wildland fire are negative or positive.
Table A.2-1 Positive and negative impacts associated with Wildland Fire3.
Prescribed Fire
Wildfire
Categories
During the
Event
Post-
Event13
During the
Event Post-Event
Firefighting
Firefighter safety
-
+
+ and/or -
Firefighter injuries/fatalities
-
+
+ and/or -
Firefighter health, both mental and physical (mental and
physical)
-
+
+ and/or -
Economic
Evacuations
NV
+
+ and/or -
Property (e.g., structures)
NV
+
+ and/or -
Property (e.g., loss of ecosystem services)
+ and/or -
+
+ and/or -
Timber and grazing
+ and/or -
+
+ and/or -
Infrastructure (e.g., powerlines, recreation, others)
NV
+
+ and/or -
Municipal watersheds (e.g., reservoirs, industry,
agriculture, drinking)
+
+
+ and/or -
Tourism (e.g., recreation, lodging, restaurants, etc.)
+ and/or -
+
+ and/or -
Aesthetics (e.g., property value, view shed, etc,)
+ and/or -
+
+ and/or -
A-l
DRAFT: Do Not Cite or Quote
-------
Table A.2-1 (Continued): Positive and Negative Impacts Associated with Wildland
Fire.3
Prescribed Fire Wildfire
Categories
During the
Event
Post-
Event13
During the
Event
Post-Event
Natural and cultural resources
+ and/or -
+
-
+ and/or -
Fuel reduction—cost effective method of treating acres
NV
+
+ and/or -
+ and/or -
Fuel reduction—treatment opportunities not limited to
markets
NV
+
+ and/or -
+ and/or -
Ecological
Ecological services including game and endangered
species
NV
+
+ and/or -
+ and/or -
Ecosystem health and resiliency
NV
+
+ and/or -
+ and/or -
Restoration/maintenance of historic natural fire regime
NV
+
+ and/or -
+ and/or -
Invasive species
+ and/or-
or NV
+
+ and/or -
+ and/or -
Climate change (e.g., GHG, carbon)
+ and/or -
+ and/or -
+ and/or -
+ and/or -
Redistribution of toxics and nutrients (e.g., mercury,
metals, sulfur, nitrogen)
"(?)
"(?)
"(?)
"(?)
Soil and water quality and quantity
+ and/or -
+ and/or -
+ and/or -
+ and/or -
Public Health: Direct Fire
Injuries
NV
+
-
+ and/or -
Hospitalizations
NV
+
-
+ and/or -
Premature mortality
NV
+
-
+ and/or -
Public Health: Air Quality
Hospitalizations and emergency department visits
-
+
-
+ and/or -
Premature mortality
-
+
-
+ and/or -
Nonfatal heart attacks/cerebrovascular events
-
+
-
+ and/or -
Asthma effects
-
+
-
+ and/or -
Other respiratory and illness effects
-
+
-
+ and/or -
A-2
DRAFT: Do Not Cite or Quote
-------
Table A.2-1 (Continued): Positive and Negative Impacts Associated with Wildland
Fire.3
Prescribed Fire
Wildfire
Categories
During the
Event
Post-
Event13
During the
Event Post-Event
Loss of work and school days
+
+ and/or -
GHG = greenhouse gas; NV = not available.
Note: Positive (+): providing some advantage (e.g., restoring ecosystems, mitigating the risk or loss from a wildfire, etc.). Negative
(-): negative consequences from a fire (e.g., property or infrastructure damage or loss).
aSigns on the impact categories are based on literature discussed throughout this report as well as expert judgements from the
report authors.
bPost-event includes impacts expected to occur as a result of reductions in the risk of more severe and damaging wildfires. For
example, reduced risk of severe wildfires reduces risks to firefighters, and reduces risks of poor air quality and related health
effects. Thus, a positive sign on the post-fire effects of prescribed fires on health categories does not indicate the fire itself
improves health, but rather than the reduction in risk of severe wildfires improves future public health.
°For many of the categories with an NA for prescribed fires, the impact will not be applicable as long as the prescribed fire remains
consistent with the management objectives. In the rare cases where prescribed fires are no longer meeting their objectives, they
can be reclassified as wildfires and will in those cases have the potential for additional negative impacts.
A.3. SUPPLEMENTAL INFORMATION FOR CHAPTER 3
No supplemental information.
A-3
DRAFT: Do Not Cite or Quote
-------
A.4.
SUPPLEMENTAL INFORMATION FOR CHAPTER 4
Table A.4-1
Criteria gas pollutant Federal Reference Methods (FRMs) and most widely employed Federal
Equivalent Methods (FEMs) used in U.S. EPA regulatory monitoring.
Pollutant Method Operating Principle
FRM Regulatory Citation
Notes
CO
Automated FRM
NDIR
40 CFR Part 50 Appendix C (U.S.
EPA. 2020a)
...
Automated FEM
Mercury replacement
UV photometry
...
Only existing CO FEM.
03
Automated FRM
Chemiluminescence
40 CFR Part 50 Appendix D (U.S.
EPA. 2011b)
Employs chemiluminescence reaction between ozone and
ethylene of NO. Ethylene chemiluminescence FRM
instruments no longer commercially available. NO
chemiluminescence method promulgated as a new FRM in
2015. NO chemiluminescence FRM instruments available
commercially.
Automated FEM
UV Photometry
...
Severe smoke interference resulting in overestimation of
ozone concentrations (Lonq et al.. In Press).
Automated FEM
Open-path DOAS
...
Employs open monitoring path length between 20-1,000 m
NO2
Automated FRM
Chemiluminescence
40 CFR Part 50 Appendix F (U.S.
EPA. 2011a)
Employs the catalytic conversion of NO2 to NO with
subsequent chemiluminescence detection of the reaction
between NO and O3. Known interference by higher oxides
of nitrogen (e.g., HNO3, HNO2, particulate nitrate).
A-4
DRAFT: Do Not Cite or Quote
-------
Table A.4-1 (Continued): Criteria gas pollutant Federal Reference Methods (FRMs) and most widely employed
Federal Equivalent Methods (FEMs) used in U.S. EPA regulatory monitoring.
Pollutant Method
Operating Principle
FRM Regulatory Citation
Notes
Automated FEM
Chemiluminescence
Employs the photolytic conversion of NO2 to NO with
subsequent chemiluminescence detection of the reaction
between NO and O3. Considered more specific for NO2
than the FRM and is a candidate for future FRM
consideration.
Automated FEM
Spectroscopic
...
Employs methods such as CAPS spectrometry.
Automated FEM
Open-path DOAS
...
Employs open monitoring path length between 50-1,000 m
SO2
Automated FRM
UV fluorescence
40 CFR Part 50 Appendix A-1 (U.S.
EPA. 2011c)
Previously an FEM, promulgated as a new FRM in 2010.
Manual FRM
Pararosaniline method
40 CFR Part 50 Appendix A-2 (U.S.
EPA. 2020b)
Manual wet chemical method not used at present time.
Automated FEM
UV fluorescence
...
Promulgated as a new FRM in 2010.
Automated FEM
Open-path DOAS
...
Employs open monitoring path length between 20-1,000 m.
CAPS = cavity attenuated phase shift; CFR = Code of Federal Regulations; CO = carbon monoxide; DOAS = differential optical absorption spectroscopy; FEM = Federal Equivalent
Method; FRM = Federal Reference Method; HN02 = nitrous acid; HN03 = nitric acid; NDIR = nondispersive infrared photometry; NO = nitric oxide; N02 = nitrogen dioxide;
03 = ozone; S02 = sulfur dioxide; UV = ultraviolet.
A-5
DRAFT: Do Not Cite or Quote
-------
Table A.4-2
Summary of low-cost sensors evaluated in biomass smoke.
Vendor
Model
Study Type
Max PM2.5
Reference Reference Regression Slope
Citation
Aeroqual
AQY 1
Field
-300 (|ig/m3)
FEM/nonFEM
0.54-2.18
Holder et al. (2020)
eLichens
IAQPS
Field
-150 (|ig/m3)
Multiple FEMs
-0.45-0.80
Delp and Sinqer (2020)
Purple Air
PA-II-SD
Field
-300 (|ig/m3)
FEM/nonFEM
0.93-1.61
Holder et al. (2020)
Purple Air
PA-II
Field
33 (|ig/m3)*
FEM
0.43
Mehadi et al. (2019)
Purple Air
PA-II
Field
-150 (|ig/m3)
Multiple FEMs
0.39-0.54
Deb and Sinaer (2020)
Sensit
RAMP
Field
-300 (|ig/m3)
FEM/nonFEM
0.77-1.48
Holder et al. (2020)
Sensit
RAMP
Chamber
-1,800 (|ig/m3)
FRM
1.35-2.43
Landis et al. (2021)
Thingy
Thingy AQ
Chamber
-1,800 (|ig/m3)
FRM
2.14-4.95
Landis et al. (2021)
Wicked Device
Air Quality Egg
Field
-150 (|ig/m3)
Multiple FEMs
-0.32-0.65
Deb and Sinaer (2020)
FEM = Federal Equivalent Method; FRM = Federal Reference Method,
t Daily average concentration.
1
A-6
DRAFT: Do Not Cite or Quote
-------
Table A.4-3 Summary of routine PM2.5 measurement methods and data availability.
Methods
PM2.5 FRMs
CSN and IMPROVE
Other PM2.5
Continuous
PM2.5 Continuous FEMs Methods
Sensor Networks
Method Specifications:
Manual or automated
Manual
Manual
Automated continuous Automated
continuous
Automated continuous
Measurement principle(s)
Gravimetric in
laboratory
Ion chromotography, x-ray
fluorescence, Thermal Optical
Reflectance all in laboratory
Key ones include: (3 Key ones include:
attenuation, TEOM, and LED (3 Attenuation,
broadband spectroscopy Nephelometers,
and TEOMs
Optical PM sensors
Method or manufacturer-
reported concentration
range
0-200 |jg/m3
however, in AQS
there are a few
values in the
Hazardous AQI
category
0-200 |jg/m3
BAM-Range: 0-1,000 |jg/m3 standard; up to
10,000 |jg/m3; T640-Range: 0.1-10,000 |jg/m3
Purple Air with U.S. EPA-
ORD correction equation
(Barkiohn et al.. 2020)
0-250 |jg/m3 range
(>250 |jg/m3 may
underestimate true PM2.5)
Manufacturer-reported
data resolution
0.1 Hg/3
0.1 |jg/m3
M1 BAM: 1 |jg/m3
TEOM and T640: 0.1 pg/m"3
0.1 |jg/m3
Data Attributes of Each Method:
Data availability (typical)
-1-3 mo after
sample collection
-3-6 mo after sample
collection
Hourly data are usually posted to AIRNow within
several minutes past the end of the hour
Near real time on Purple Air
web site
Hourly update on AIRNow fire
and smoke map
A-7
DRAFT: Do Not Cite or Quote
-------
Table A.4-3 (Continued): Summary of routine PM2.5 measurement methods and data availability.
Other PM2.5
Continuous
Methods
PM2.5 FRMs
CSN and IMPROVE
PM2.5 Continuous FEMs Methods
Sensor Networks
Data interval available
24-h midnight to
24-h midnight to midnight
Hourly data is collected and reported by
Sub hourly; data layer on
midnight local
local standard time. Most
AIRNow; some methods have subhourly data
AIRNow fire page is hourly
standard time.
sites operate every 3rd day;
available
Some sites operate
some CSN sites every 6th
(T640 has 1-min data available—smoothed in
daily, others every
day
rolling 10-min averages)
3rd or 6th day,
some QA samplers
every 12th day
Where are data available?
AQS—
AQS and UC Davis web
AQS, AIRNow, AIRNowTech, and many State
Purple air web site, AIRNow
https://www.epa.qo
site—
and local web sites—https://www.airnow.qov/
fire and smoke page—
v/aqs/obtaininq-
https://airqualitv.ucdavis.edu/
http: //a i rn owtech .oral
https://fire.airnow.qov/
aas-data
csn
(credentials required)
https://www.purpleair. com
httpsV/airqualitv.ucdavis.edu/i
m prove
Highest concentrations
There are seven
There are no cases in
Six cases reported in the In the Hazardous
NA
reported with this method
cases in the
Hazardous AQI category.
Hazardous AQI category. All AQI category
to AQS (2010-2019).
"Hazardous AQI
There are 13 cases in the
with a BAM in CA, MT, or there are 21 cases
category" all in AK,
"very unhealthy" AQI
WA. High = 557.1 |jg/m3. with a Correlated
CA, or OR. The
category. 8 by the IMPROVE
Nephelometer all
highest reported
method; high = 210.2 |jg/m3
in OR or WA, high
concentration was
all in CA and MT; 1 by a
reported = 570.3
411.7 |jg/m3.
SASS (CSN) at 206.7 |jg/m3
|jg/m3; 34 cases
in IL; and 4 cases listed as a
with a BAM all
generic filter-based method,
reported in AK,
high = 230 |jg/m3 all in CA
CA, ID, or MT,
and NV.
high = 642.0
|jg/m3; 1 case with
a TEOM at
252.0 |jg/m3 in ID.
Network Attributes:
U.S. Stations Reporting to
678
CSN = 143
678 305
NA
AQS (2020)
IMPROVE = 156
A-8
DRAFT: Do Not Cite or Quote
-------
Table A.4-3 (Continued): Summary of routine PM2.5 measurement methods and data availability.
Methods
PM2.5 FRMs
CSN and IMPROVE
PM2.5 Continuous FEMs
Other PM2.5
Continuous
Methods
Sensor Networks
Key network Design
features
Most sites are
population-
orientated locations
in CBSA's. Each
state should have a
background and
transport site
CSN includes STN, Ncore,
and supplemental sites (most
in CBSAs.)
IMPROVE supports regional
haze program with most sites
in Class 1 areas and national
parks. Some IMPROVE
protocol sites are operated in
lieu of CSN.
Same as FRM
Same as FRM. In
WA and OR
nephelometers are
often used to
supplement AQI
reporting in
communicates
where NAAQS
comparable data
are not required;
however, smoke
impacts may be of
concern.
Sites may exist anywhere
users report via internet to
Purple Air site. Users self-
describe if ambient air or
inside. Note: only sites
described as ambient air are
used in fire and smoke map
layer.
AQI = Air Quality Index; AQS = Air Quality System; CBSA = core-based statistical area; CSN = Chemical Speciation Network; IMPROVE = Interagency Monitoring of Protected Visual
Environments; LED = light-emitting diode; h = hour; min = minute; mo= month; NCore = National Core Network; ORD = Office of Research and Development; PM25= particulate matter
with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; QA = quality assurance; STN = Speciation Trends Network; TEOM = Tapered Element Oscillating
Microbalance.
A-9
DRAFT: Do Not Cite or Quote
-------
Table A.4-4 Overview of wildland fire relevant imagery/composition satellite data products.
System Content
Satellite Product
Instrument
NOAA Aerosol Watch
NOAA
JSTAR
Mapper
NOAA
Hazard
Mapping
System
NASA
LANCE/World
View
U.S. EPA
AirNow Tech
U.S. EPA
Remote
Sensing
Information
Gateway
Corrected Reflectance
True Color
GOES ABI
VIIRS
I, D
I, D
I, D
I, D
MODIS
I, D
I, D
Digitized Smoke
Analysis
ABI + VIIRS
I, D
Aerosol Optical Depth ABI
VIIRS
I, D
I, D
MODIS
I, D
I, D
Aerosol Detection
(smoke/dust)
ABI
VIIRS
I, D
A-10
DRAFT: Do Not Cite or Quote
-------
Table A.4-4 (Continued): Overview of wildland fire relevant imagery/composition satellite data products.
System Content
Satellite Product
Instrument NOAA Aerosol Watch
NOAA
JSTAR
Mapper
NOAA
Hazard
Mapping
System
NASA
LANCE/World
View
U.S. EPA
AirNow Tech
U.S. EPA
Remote
Sensing
Information
Gateway
Fire ABI
Characterization/Hot
Spots/Active Fires VIIRS
I, D
I, D
I, D
I, D
Al
TROPOMI
I, D
CO
NO2
I, D
Satellite predicted
PM2.5
I, D
I, D (ASDP)
Surface concentration
measurements from
AirNow or AQS (PM2.5,
O3, NO2, SO2)
AirNow
I (h PM2.5 only)
AQS
I, D
I, D
I, D
I, D
ABI = Advanced Baseline Imager; AI = aerosol index; AQS = Air Quality System; CO = carbon monoxide; D = data available; GOES = Geostationary
Operational Environmental Satellite; I = image available; MODIS = Moderate Resolution Imaging Spectroradiometer; NASA = National Aeronautics and
Space Administration; N02 = nitrogen dioxide; NOAA = National Oceanic and Atmospheric Administration; 03 = ozone; PM25 = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 2.5 |im. S02 = sulfur dioxide; VIIRS = Visual Infrared Imaging Radiometer Suite.
A-l 1
DRAFT: Do Not Cite or Quote
-------
Table A.4-5 Ground-based remote sensing networks vertical (profile and total column data).
Network and/or
Instrument
Lead
Organization
Total
Number
of Sites
in U.S.
Date
Latency
Initiated
Measurement
Relevant Constituent/
Properties
URL for Information on measurements/data
Automated Surface
Observing System
(ASOS)/
NOAA
900
Surface visibility
https://www.aviationweather.qov/metar?qis=off
Photochemical
Assessment
Monitoring Stations
U.S. EPA
-40
2021
Backscatter aerosol profiles (15 km),
PBLH, aerosol layer identification
httDs://ala. umbc.edu/ucn
MPLNET—Micro-
pulse LiDAR
Network
NASA
(federated)
35
2000
Aerosols and cloud layer heights
httD://mDlnet.asfc.nasa.aov/
AErosol RObotic
NETwork
(AERONET)
NASA
(federated)
-100
1998
Aerosol spectral optical depths, aerosol
size distributions, and precipitable water
httD://aeronet.asfc.nasa.aov/index.html
Pandonia Global
Network
NASA-ESA
14
Total Column O3, NO2, Tropospheric
Column NO2, HCHO, and Surface NO2
httDs://www.Dandonia-alobal-network.ora/
ESA = European Space Agency; HCHO = formaldehyde; LiDAR = Light Detection and Ranging; MPLENT = micro-pulse LiDAR; NASA = National Aeronautics and Space
Administration; NOAA = National Oceanic and Atmospheric Administration; N02 = nitrogen dioxide; 03 = ozone; PBLH = planetary boundary layer heights.
A-12
DRAFT: Do Not Cite or Quote
-------
A.4.1.
Example State and Local Sponsored Smoke Blogs
1 Information on general ambient air quality, the impact of wildland fire smoke on current ambient
2 air quality conditions, and air quality forecasts are available to the public through the multiagency
3 AirNow website as well as state and local websites. Several western states maintain websites ("smoke
4 blogs") dedicated to providing the public with information on wildfire smoke impacts (Examples listed
5 below). The material delivered by these smoke blogs varies from state to state with the sites leveraging
6 smoke and fire observations and forecast products from a variety of sources (e.g., AirNow, dedicated
7 state/local monitors). Below are some example state and local websites and smoke blogs that provide air
8 quality information to the public and are a resource during wildfire events with the landing page title in
9 parentheses.
10 • Alaska
11 (Wildfire Smoke—Particulate Matter Information)
12 https: //dec. alaska. go v/ai r/ai r-m on i to ri n g/w i 1 dfi re -smoke-info/
13 • Arizona
14 (Wildfire Support)
15 http ://www. azdeq. gov/node/2913
16 • California
17 Butte County Air Quality Management District (AQMD, Wildfires and Air Quality)
18 https://bcaqmd.org/resources-education/wildfires/
19 • North Coast Unified Air Quality Management District
20 http://www.ncuaqmd.org/index.php?page=wildfire
21 • Santa Barbara Pollution Control District, California (Today's Air Quality and Forecasts)
22 https://www.ourair.org/todavs-air-qualitv/
23 • South Coast Air Quality Management District, California (South Coast AQMD)
24 http://www.aqmd.gov/
25 • Ventura County Air Pollution Control District (VCAPD)
26 http ://www. vcapcd. org/
27 • Idaho
28 (Air Quality Index [AQI])
29 https://www.deq.idaho.gov/air-qualitv/air-qualitv-index/
30 • (Idaho Smoke Information)
31 http://idsmoke.blogspot.com/
32 • Montana
33 (Wildfire Smoke Update)
34 http ://svc .mt.gov/deq/todavsair/smokereport/mostrecentupdate.aspx
35 • (Montana Wildfire Smoke)
36 https: //www .montanawildfiresmoke. org/
A-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
• Nevada
(Northern Sierra Air Quality Management District)
https://mvairdistrict.com/
• New Mexico
(Wildfire and Prescribed Fire Smoke Resources)
https://www.env.nm.gov/air-qualitv/fire-smoke-links/
• North Carolina
(Air Quality)
https: //deq ,nc. gov/about/divisions/air-qualitv
• Oregon
(Oregon Smoke Information)
http://oregonsmoke.blogspot.com/
• South Carolina
(Wildfires—Protect Yourself)
https://scdhec.gov/disaster-preparedness/wildfires-protect-vourself
• Washington
(Washington Smoke Information)
https://wasmoke.blogspot.com/
A.4.2. U.S. EPA PM2.5 Mass Monitoring
The particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |im
(PM2 5) monitoring program is one of the major ambient air monitoring programs operated across the
country. For most urban locations PM2 5 monitors are sited at the neighborhood scale as defined in
40 Code of Federal Regulations (CFR) Appendix D to Part 58 (U.S. EPA. 2015). where PM2 5
concentrations are reasonably homogeneous throughout an entire urban subregion. In each CBS A with a
monitoring requirement, at least one PM2 5 monitoring station representing area-wide air quality is to be
sited in an area of expected maximum concentration.
There are three main components of the PM2 5 monitoring program: 24-hour integrated
filter-based Federal Reference Method (FRM) samplers, continuous Federal Equivalent Method (FEM)
mass instrument measurements reported as 1-hour concentrations, and 24-hour integrated filter-based
Chemical Speciation Network (CSN) samplers. The FRM data are primarily used for determining
NAAQS compliance, but also serve other important purposes such as developing trends and evaluating
the field performance FEM continuous mass instruments. Continuous FEM instrument data are also used
for determining NAAQS compliance and their real-time data support public AQI communication and air
quality forecasting on AirNow. FRMs have been available since the PM2 5 monitoring network
commenced operation in January of 1999, and PM2 5 continuous FEMs became commercially available in
2008. Many State and local agencies are transitioning their regulatory PM2 5 monitoring networks to
continuous FEMs. However, even if a monitoring agency chooses to run PM2 5 continuous FEMs at all
their stations, some FRMs are still required. For example, FRMs are required under quality assurance
A-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
requirements and at U.S. Environmental Protection Agency (U.S. EPA) National Core Network (NCore)
Stations (Appendices A and D to 40 CFR Part 58) (U.S. EPA. 2019. 2015).
The CSN and related Interagency Monitoring of Protected Visual Environments (IMPROVE)
network is used to provide chemical composition of the aerosol which serve a several objectives. The
CSN program is managed by U.S. EPA with field operations conducted by state and local agencies and
national contract laboratories responsible for shipping, handling, and analysis of samples. The IMPROVE
is operated by the Department of the Interior (DOI) under the direction of a multiagency federal/state
steering committee. The IMPROVE monitoring program supports the national goal of reducing haze to
near natural levels in National Parks and wilderness areas.
In 2020 there were 678 FRM filter-based samplers included in the U.S. EPA PM2 5 network that
provide 24-hour PM2 5 mass concentration data. Of these operating FRMs, 72 are providing daily PM2 5
data, 320 every 3rd day, 229 every 6th day, and 57 every 12th day. As of 2020, there are 983 continuous
PM2 5 mass monitors that provide hourly data on a near real-time basis reporting across the country. A
total of 678 of the PM2 5 continuous monitors are FEMs and therefore used both for comparison with the
NAAQS and to report the AQI. Another 305 monitors not approved as FEMs are operated primarily to
report the AQI. These legacy PM2 5 continuous monitors were largely purchased prior to the availability
of designated PM2 5 continuous FEMs instruments. The most widely used PM2 5 continuous monitor not
designated as an FEM is the Radiance Research (Seattle, WA) Model M903 nephelometer (locally
correlated to an FRM).
The first designated automated PM2 5 FEM instrument was the Met One Instruments (Grants Pass,
OR) Model BAM 1020 (14C (3 attenuation radiometric method) in 2008. The BAM 1020 and more
recently approved BAM 1022 account for approximately 50% and the Teledyne API (San Diego, CA)
Model T640/T640x account for approximately 30% of the nationally operating automated PM2 5 FEMs.
The U.S. EPA has approved a total of 11 PM2 5 automated methods as FEMs including beta attenuation
from multiple instrument manufacturers; optical methods such as the GRIMM Aerosol Technik (Ainring,
Germany) Model 180 and the Teledyne API Model T640/T640x; and methods employing the Thermo
Environmental (Franklin, MA) Model 1405 Tapered Element Oscillating Microbalance (TEOM) with a
Filter Dynamic Measurement System (FDMS).
A.4.3. U.S. EPA PM2.5 Speciation Monitoring
Particulate matter (PM) is the generic term for a broad class of chemically and physically diverse
substances that exist as liquid and/or solid particles over a wide range of sizes. Particles originate from a
variety of anthropogenic stationary and mobile sources, as well as from natural sources like wildfires.
Particles may be emitted directly or formed in the atmosphere by photochemical transformations of
gaseous precursors such as sulfur dioxide (SO2), NOx, ammonia (NH3), and volatile organic compounds
(VOCs). The chemical and physical properties of PM2 5 vary greatly with time, region, meteorology, and
A-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
source category. U.S. EPA implemented the CSN to investigate the chemical components of PM2 5 at
selected locations across the country. The CSN sample filters are analyzed for 33 trace elements using
energy dispersive x-ray fluorescence [EDXRF; Watson et al. (1999); Jaklevic et al. (1981)1. water soluble
major ions (e.g., ammonium, potassium, nitrate, sulfate) using ion chromatography [IC; U.S. EPA
(1999)1. and elemental carbon (EC)/organic carbon (OC) using thermal optical reflectance [TOR; Chow
et al. (1993); Huntzicker et al. (1982)1. Chemical composition can provide valuable information about the
sources and relative toxicity of PM2 5.
In 2020 the CSN continued routine long-term PM2 5 measurements at 143 predominately urban
locations. The major network components of the CSN include the Speciation Trends Network (STN),
NCore stations, and supplemental speciation sites. STN sites are intended to be long-term locations where
chemical section measurements are taken. NCore is a multipollutant network measuring PM2 5 mass,
criteria gases, and basic meteorology that has been in formal operation since January 1, 2011. Particle
measurements made at NCore include PM2 5 filter-based mass, which is largely the FRM, except in some
rural locations that utilize the IMPROVE program PM2 5 mass filter-based measurement; PM2 5 speciation
using either the CSN program or IMPROVE program; and coarse particulate matter (PM10-2.5; particulate
matter with a nominal aerodynamic diameter less than or equal to 10 ^m and greater than a nominal
2.5 |im) mass utilizing an FRM, FEM or IMPROVE samplers for some of the rural locations. As of 2020,
the NCore network includes a total of 78 stations of which 63 are in urban or suburban stations designed
to provide representative population exposure and another 15 rural stations designed to provide regional
background and transport information. The NCore network is deployed in all 50 States, District of
Columbia, and Puerto Rico with at least one station in each state and two or more stations in larger
population states (California, Florida, Illinois, Michigan, New York, North Carolina, Ohio, Pennsylvania,
and Texas). Both the STN and NCore networks which together comprise 76 locations with CSN
measurements are intended to remain in operation indefinitely. The CSN measurements at STN and
NCore stations operate on a l-in-3-day sampling schedule. Another approximately 67 CSN stations,
known as supplemental sites, are intended to be temporary locations used to support State Implementation
Plan (SIP) development and other local or regional monitoring objectives. Supplemental CSN stations
typically operate on a l-in-6-day sampling schedule.
Specific chemical components of PM2 5 are also measured through the IMPROVE monitoring
program which supports regional haze characterization and tracks changes in visibility in Class I areas
(e.g., large national parks) as well as many other rural and some urban areas. As of 2020, the IMPROVE
network includes 110 base network monitoring locations and additional 46 locations operated as
IMPROVE protocol sites where a state, local, or tribal monitoring agency has requested participation in
the program. These IMPROVE protocol sites operate the same way as the IMPROVE program, but they
may serve several monitoring objectives (e.g., SIP development) and are not explicitly tied to the
Regional Haze Program. Samplers at IMPROVE stations operate on a l-in-3-day sampling schedule.
Together, the CSN and IMPROVE data provide chemical species information for PM2 5 that are critical
for use in health and epidemiologic studies to help inform reviews of the primary PM NAAQS and can be
A-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
used to better understand visibility through calculation of light extinction using the IMPROVE algorithm
to support reviews of the secondary PM NAAQS.
A.4.4. U.S. EPA Criteria Gas Monitoring
Routine monitoring for criteria gases is performed at State and Local Air Monitoring Stations
(SLAMS) using designated FRMs and FEMs. Table A.4-1 provides information on the FRMs and most
widely deployed FEMs for the carbon monoxide (CO), ozone (O3), nitrogen dioxide (NO2), and SO2
criteria gases. The current FRM for measuring concentrations of CO in ambient air is based on
nondispersive infrared photometry (NDIR) and is detailed in 40 CFR Part 50 Appendix C (U.S. EPA.
2020a). To date only one FEM for CO has been designated and it is based upon mercury
replacement-ultraviolet (UV) photometry. For O3, the current FRM is based upon the chemiluminescent
reaction between O3 and ethylene or nitric oxide (NO) and is detailed in 40 CFR Part 50 Appendix D
(U.S. EPA. 2011^. Currently FRM instruments based upon ethylene chemiluminescence are not
available commercially, for this reason, an updated FRM that includes NO chemiluminescence was
promulgated in 2015. The most widely used O3 FEM is based upon UV photometry. This method,
however, has been shown to have severe interferences in smoke and may result in significant
overestimation of O3 concentrations in smoke impacted areas (Long et al.. In Press). The measurement
principle for the NO2 FRM detailed in 40 CFR Part 50 Appendix F (U.S. EPA. 2011a). consists of the
catalytic conversion of NO2 to NO followed by subsequent detection of the chemiluminescence reaction
of NO with O3. In addition to converting NO2 to NO prior to detection, this method also converts high
oxides of nitrogen (e.g., nitric acid [HNO3], nitrous acid [HNO2], particulate nitrate) to NO resulting in a
potential overestimation ofN02 concentrations. FEMs for NO2 involve direct spectroscopic measurement
of NO2 and the replacement of the catalytic converter with a more specific photolytic converter prior to
detection in the chemiluminescence method. Currently there are two FRMs for measuring concentrations
of SO2 in ambient air. The newer automated FRM is based on UV fluorescence and detailed in 40 CFR
Part 50 Appendix A-l (U.S. EPA. 2011c') and was promulgated in 2010. Prior to promulgation as an
FRM, the UV fluorescence method was the most widely used FEM. The second SO2 FRM is based upon
the manual wet-chemical pararosaniline method and detailed in 40 CFR Part 50 Appendix A-2 (U.S.
EPA. 2020b). Currently, this method is not employed in the routine monitoring of SO2. For O3, NO2, and
SO2 automated open-path FEMs also exist based upon differential optical absorption spectroscopy
(DOAS). These methods employ long measurement path lengths extending up to 1,000 m.
A-17
DRAFT: Do Not Cite or Quote
-------
A.5.
SUPPLEMENTAL INFORMATION FOR CHAPTER 5
A.5.1. Supplemental Tables for CHAPTER 5
Table A.5 FUELS-1. Crosswalk between LANDFIRE existing vegetation types (LANDFIRE, 2014 Existing
Vegetation Type) within the four scenario areas and an assigned Fuel Characteristic
Classification System (FCCS) fuelbed. Fuelbed descriptions for each of the base fuelbeds
can be found within the Fuel and Fire Tools (https://www.fs.usda.qov/pnw/tools/fuel-and-
fire-tools-fft).
EVTJD
EVT Name
FCCS ID
Fuelbed Name
11
Ba Open Water
0
Barren
31
Bab Barren
0
Barren
2001
Sps Inter-Mountain Basins Sparsely Vegetated Systems
0
Barren
2002
Sps Mediterranean California Sparsely Vegetated Systems
0
Barren
2003
Sps North Pacific Sparsely Vegetated Systems
0
Barren
2006
Sps Rocky Mountain Alpine/Montane Sparsely Vegetated Systems
0
Barren
2011
Tr Rocky Mountain Aspen Forest and Woodland
42
Quaking aspen/Engelmann spruce forest
2027
Tr Mediterranean California Dry-Mesic Mixed Conifer Forest and Woodland
37
Ponderosa pine-Jeffrey pine forest
2028
Tr Mediterranean California Mesic Mixed Conifer Forest and Woodland
214
Giant sequoia-white fir-sugar pine forest
2030
Tr Mediterranean California Lower Montane Conifer Forest and Woodland
16
Jeffrey pine-ponderosa pine-Douglas fir-CA black oak forest
A-18
DRAFT: Do Not Cite or Quote
-------
Table A.5 FUELS-1 (Continued): Crosswalk between LANDFIRE existing vegetation types (LANDFIRE, 2014
Existing Vegetation Type) within the four scenario areas and an assigned Fuel
Characteristic Classification System (FCCS) fuelbed. Fuelbed descriptions for
each of the base fuelbeds can be found within the Fuel and Fire Tools
(https://www.fs.usda.qov/pnw/tools/fuel-and-fire-tools-fft).
EVTJD
EVT Name
FCCS ID
Fuelbed Name
2032
Tr Mediterranean California Red Fir Forest
17
Red fir forest
2033
Tr Mediterranean California Subalpine Woodland
12
Red fir-mountain hemlock-lodgepole pine-western white pine forest
2037
Tr North Pacific Maritime Dry-Mesic Douglas Fir-Western Hemlock Forest
8
Western hemlock-Douglas fir-western redcedar/vine maple forest
2041
Tr North Pacific Mountain Hemlock Forest
238
Pacific silver fir-mountain hemlock forest
2042
Tr North Pacific Mesic Western Hemlock-Silver Fir Forest
238
Pacific silver fir-mountain hemlock forest
2043
Tr Mediterranean California Mixed Evergreen Forest
37
Ponderosa pine-Jeffrey pine forest
2044
Tr Northern California Mesic Subalpine Woodland
12
Red fir-mountain hemlock-lodgepole pine-western white pine forest
2045
Tr Northern Rocky Mountain Dry-Mesic Montane Mixed Conifer Forest
52
Douglas fir-Pacific ponderosa pine/oceanspray forest
2053
Tr Northern Rocky Mountain Ponderosa Pine Woodland and Savanna
53
Pacific ponderosa pine forest
2056
Tr Rocky Mountain Subalpine Mesic-Wet Spruce-Fir Forest and Woodland
59
Subalpine fir-Engelmann spruce-Douglas fir-lodgepole pine forest
2058
Tr Sierra Nevada Subalpine Lodgepole Pine Forest and Woodland
12
Red fir-mountain hemlock-lodgepole pine-western white pine forest
2068
Sh North Pacific Dry and Mesic Alpine Dwarf-Shrubland or Fell-Field or Meadow 319
Pacific silver fir-Sitka alder forest
2080
Sh Inter-Mountain Basins Big Sagebrush Shrubland
233
Sagebrush shrubland
2083
Sh North Pacific Avalanche Chute Shrubland
319
Pacific silver fir-Sitka alder forest
2084
Sh North Pacific Montane Shrubland
237
Huckleberry heather shrubland
2098
Sh California Montane Woodland and Chaparral
44
Scrub oak chaparral shrubland
2106
Sh Northern Rocky Mountain Montane-Foothill Deciduous Shrubland
331
Sitka alder-salmonberry shrubland
A-19
DRAFT: Do Not Cite or Quote
-------
Table A.5 FUELS-1 (Continued): Crosswalk between LANDFIRE existing vegetation types (LANDFIRE, 2014
Existing Vegetation Type) within the four scenario areas and an assigned Fuel
Characteristic Classification System (FCCS) fuelbed. Fuelbed descriptions for
each of the base fuelbeds can be found within the Fuel and Fire Tools
(https://www.fs.usda.qov/pnw/tools/fuel-and-fire-tools-fft).
EVTJD
EVT Name
FCCS ID
Fuelbed Name
2125
Sh Inter-Mountain Basins Big Sagebrush Steppe
233
Sagebrush shrubland
2138
He North Pacific Montane Grassland
315
Showy sedge-black alpine sedge grassland
2139
He Northern Rocky Mountain Lower Montane-Foothill-Valley Grassland
506
Idaho fescue-California oatgrass grassland
2145
He Rocky Mountain Subalpine-Montane Mesic Meadow
530
Temperate Pacific subalpine-montane wet meadow
2152
Tr California Montane Riparian Systems
319
Pacific silver fir-Sitka alder forest
2154
Tr Inter-Mountain Basins Montane Riparian Systems
319
Pacific silver fir-Sitka alder forest
2167
Tr Rocky Mountain Poor-Site Lodgepole Pine Forest
22
Mature lodgepole pine forest
2171
He North Pacific Alpine and Subalpine Dry Grassland
315
Showy sedge-black alpine sedge grassland
2172
Tr Sierran-lntermontane Desert Western White Pine-White Fir Woodland
273
Engelmann spruce-Douglas fir-white fir-ponderosa pine forest
2173
Tr North Pacific Wooded Volcanic Flowage
28
Ponderosa pine savanna
2174
Tr North Pacific Dry-Mesic Silver Fir-Western Hemlock-Douglas Fir Forest
8
Western hemlock-Douglas fir-western redcedar/vine maple forest
2181
He Introduced Upland Vegetation-Annual Grassland
57
Wheatgrass-cheatgrass grassland
2182
He Introduced Upland Vegetation-Perennial Grassland and Forbland
57
Wheatgrass-cheatgrass grassland
2902
Bau Developed-Low Intensity
0
Barren
2905
Bau Developed-Roads
0
Barren
2914
Dtc Urban Evergreen Forest
22
Mature lodgepole pine forest
2916
Dgr Urban Herbaceous
66
Bluebunch wheatgrass-bluegrass grassland
A-20
DRAFT: Do Not Cite or Quote
-------
Table A.5 FUELS-1 (Continued): Crosswalk between LANDFIRE existing vegetation types (LANDFIRE, 2014
Existing Vegetation Type) within the four scenario areas and an assigned Fuel
Characteristic Classification System (FCCS) fuelbed. Fuelbed descriptions for
each of the base fuelbeds can be found within the Fuel and Fire Tools
(https://www.fs.usda.qov/pnw/tools/fuel-and-fire-tools-fft).
EVTJD EVT Name
FCCS ID
Fuelbed Name
2917 Dsh Urban Shrubland
401
Holly-privet shrubland
2926 Dsh Developed Ruderal Shrubland
401
Holly-privet shrubland
EVT = existing vegetation type; FCCS = Fuel Characteristic Classification System.
1
A-21 DRAFT: Do Not Cite or Quote
-------
Table A.5 FUELS-2 Disturbance update rules for past prescribed burns and
wildfires.
Recent Low- Past
Severity Wildfire
FCCS ID Fuelbed Name Prescribed Burn 0-5 yr Past Wildfire 5-10 yr
8
Western hemlock-Douglas fir-western
redcedar/vine maple forest
8_111
8_132
8_133
12
Red fir-mountain hemlock-lodgepole pine-
western white pine forest
12_111
12_132
12_133
16
Jeffrey pine-ponderosa pine-Douglas fir—CA
black oak forest
16_111
16_132
16_133
17
Red fir forest
17_111
17_132
17_133
22
Mature lodgepole pine forest
22_i -i -i
22_132
22_133
28
Ponderosa pine savanna
28_111
28_132
28_133
37
Ponderosa pine-Jeffrey pine forest
37 111
37 132
37 133
42
Quaking aspen/Engelmann spruce forest
42_111
42_132
42_133
44
Scrub oak chaparral shrubland
44_111
44_132
44_133
52
Douglas fir-Pacific ponderosa pine/oceanspray
forest
52_111
52_132
52_133
53
Pacific ponderosa pine forest
53 111
53 132
53 133
57
Wheatgrass-cheatgrass grassland
57 111
57 132
57 133
59
Subalpine fir-Engelmann spruce-Douglas fir-
lodgepole pine forest
59 111
59 132
59 133
66
Bluebunch wheatgrass-bluegrass grassland
66_111
66_132
66_133
214
Giant sequoia-white fir-sugar pine forest
214 111
214_132
214_133
233
Sagebrush shrubland
233_111
233_132
233_133
237
Huckleberry heather shrubland
237_111
237_132
237_133
238
Pacific silver fir-mountain hemlock forest
238_111
238_132
238_133
273
Engelmann spruce-Douglas fir-white fir-
ponderosa pine forest
273_111
273_132
273_133
315
Showy sedge-black alpine sedge grassland
315 111
315 132
315 133
A-22
DRAFT: Do Not Cite or Quote
-------
Table A.5 FUELS-2 (Continued): Disturbance update rules for past prescribed
burns and wildfires.
Recent Low- Past
Severity Wildfire
FCCS ID Fuelbed Name Prescribed Burn 0-5 yr Past Wildfire 5-10 yr
319
Pacific silver fir-Sitka alder forest
319_
.111
319.
.132
319.
.133
331
Sitka alder-salmonberry shrubland
331_
.111
331.
.132
331.
.133
401
Holly-privet shrubland
401_
.111
401.
.132
401.
.133
506
Idaho fescue-California oatgrass grassland
1
CD
O
LO
.111
1
CD
O
LO
.132
1
CD
O
LO
.133
530
Temperate Pacific subalpine-montane wet
meadow
1
O
CO
LO
.111
1
O
CO
LO
.132
1
O
CO
LO
.133
FCCS = Fuel Characteristic Classification System.
Table A.5 SPECIATION-1 Speciation profiles used for converting volatile organic
compound (VOC) and PM2.5 to model species.
Mass
Profile ID
Pollutant
CB6 group
Fraction
95423
TOG
ALD2_PRIMARY
0.0223
95423
TOG
FORM_PRIMARY
0.0445
95423
TOG
SOAALK
0.009503
95423
TOG
ACET
0.0115
95423
TOG
ALD2
0.0223
95423
TOG
ALDX
0.036
95423
TOG
BENZ
0.005976
95423
TOG
CH4
0.0968
95423
TOG
ETH
0.0275
95423
TOG
ETHA
0.0132
95423
TOG
ETHY
0.006216
95423
TOG
ETOH
0.004761
95423
TOG
FORM
0.0445
95423
TOG
IOLE
0.0107
95423
TOG
ISOP
0.001913
95423
TOG
KET
0.005659
95423
TOG
MEOH
0.0501
95423
TOG
NAPH
0.006475
95423
TOG
NVOL
0.004562
95423
TOG
OLE
0.0553
95423
TOG
PAR
0.3296
95423
TOG
PRPA
0.004821
95423
TOG
TERP
0.0129
95423
TOG
TOL
0.0476
95423
TOG
UNR
0.1636
95423
TOG
XYLMN
0.038
Prescribed Fires: VOC->TOG factor = 1.14341685
Wild Fires
Mass
Profile ID Pollutant CB6 group Fraction
95424
TOG
ACET
0.0115
95424
TOG
ALD2
0.0224
95424
TOG
ALDX
0.0353
95424
TOG
BENZ
0.006012
95424
TOG
CH4
0.1095
95424
TOG
ETH
0.0273
95424
TOG
ETHA
0.0161
95424
TOG
ETHY
0.005622
95424
TOG
ETOH
0.004785
95424
TOG
FORM
0.0336
95424
TOG
IOLE
0.0108
95424
TOG
ISOP
0.001929
95424
TOG
KET
0.005694
95424
TOG
MEOH
0.0308
95424
TOG
NAPH
0.006505
95424
TOG
NVOL
0.004606
95424
TOG
OLE
0.0507
95424
TOG
PAR
0.343
95424
TOG
PRPA
0.007611
95424
TOG
TERP
0.013
95424
TOG
TOL
0.0492
95424
TOG
UNR
0.1647
95424
TOG
XYLMN
0.0393
Wildfires: VOC->TOG factor = 1.16417442
Wild and Prescribed Fires
CMAQ
Mass
Profile ID
Pollutant
specie
Fraction
3766AE6
PM2_
_5
PN03
2.810E-04
3766AE6
PM2_
_5
POC
4.688E-01
3766AE6
PM2_
_5
PSI
6.200E-04
3766AE6
PM2_
_5
PNA
1.220E-04
3766AE6
PM2_
_5
PS04
1.332E-03
3766AE6
PM2_
_5
PTI
1.500E-05
3766AE6
PM2_
_5
PNH4
1.105E-03
3766AE6
PM2_
_5
PEC
3.227E-02
3766AE6
PM2_
_5
PK
1.203E-03
3766AE6
PM2_
_5
PNCOM
3.281E-01
3766AE6
PM2_
_5
PAL
1.540E-04
3766AE6
PM2_
_5
PCA
3.693E-03
3766AE6
PM2_
_5
PCL
2.070E-03
3766AE6
PM2_
_5
PFE
1.800E-04
3766AE6
PM2_
_5
PMG
1.790E-04
3766AE6
PM2_
_5
PMN
5.000E-06
3766AE6
PM2
5
PMOTHR
1.599E-01
CMAQ = Community Multiscale Air Quality; PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal
to 2.5 |jm; VOC = volatile organic compound.
A-23
DRAFT: Do Not Cite or Quote
-------
Table A.5-1 Model performance metrics estimated for ozone and major speciated
components of PM2.5. Performance metrics include mean bias, mean
error, normalized mean bias, normalized mean error, and
correlations coefficient.
Normalized
Normalized
Mean
Mean
Mean Bias
Mean Error
Modeling Period
Specie
Data subset
N
Bias
Error
(%)
(%)
2
r
July 2018
MDA8 ozone
None (all data)
273
2.24
7.39
4.32
14.26
0.52
MDA8 ozone
Modeled MDA8 03 > 60 ppb
79
7.97
11.64
12.75
18.62
0.07
MDA8 ozone
Observed MDA8 03 > 60 ppb
89
-3.62
7.27
-5.42
10.90
0.16
PM2.5 nitrate ion
None (all data)
46
-0.07
0.22
-34.82
110.21
0.01
PM2.5 sulfate ion
None (all data)
46
-0.03
0.19
-5.96
41.28
0.04
PM2.5 total carbon
None (all data)
46
-1.20
4.29
-18.34
65.85
0.43
Sep 2019
MDA8 ozone
None (all data)
533
3.79
5.58
10.89
16.03
0.57
MDA8 ozone
Modeled MDA8 03 > 60 ppb
MDA8 ozone
Observed MDA8 03 > 60 ppb
PM2.5 nitrate ion
None (all data)
82
-0.03
0.10
-25.63
84.24
0.30
PM2.5 sulfate ion
None (all data)
82
0.23
0.26
73.66
84.20
0.15
PM2.5 total carbon
None (all data)
80
0.87
1.11
89.62
114.36
0.12
Feb/Mar 2019
MDA8 ozone
None (all data)
576
6.03
7.11
15.65
18.44
0.16
MDA8 ozone
Modeled MDA8 03 > 60 ppb
MDA8 ozone
Observed MDA8 03 > 60 ppb
PM2.5 nitrate ion
None (all data)
163
-0.14
0.15
-80.44
85.33
0.14
PM2.5 sulfate ion
None (all data)
167
0.30
0.31
164.85
166.27
0.63
PM2.5 total carbon
None (all data)
169
0.15
0.36
32.18
78.73
0.30
Aug/Sep 2015
MDA8 ozone
None (all data)
11,510
0.64
6.53
1.29
13.08
0.66
MDA8 ozone
Modeled MDA8 03 > 60 ppb
2,266
1.26
8.41
1.88
12.54
0.22
MDA8 ozone
Observed MDA8 03 > 60 ppb
2,660
-6.47
8.76
-9.24
12.50
0.31
PM2.5 nitrate ion
None (all data)
720
-0.39
0.47
-70.56
83.73
0.20
PM2.5 sulfate ion
None (all data)
722
-0.01
0.33
-0.92
44.37
0.25
PM2.5 total carbon
None (all data)
536
-0.57
1.82
-18.92
59.91
0.37
Aug/Sep 2010
MDA8 ozone
None (all data)
11,764
7.33
10.09
14.31
19.69
0.59
MDA8 ozone
Modeled MDA8 03 > 60 ppb
5,373
9.89
12.25
15.78
19.53
0.22
MDA8 ozone
Observed MDA8 03 > 60 ppb
3,582
2.76
8.80
3.91
12.47
0.27
PM2.5 nitrate ion
None (all data)
540
-0.16
0.24
-65.37
96.29
0.02
PM2.5 sulfate ion
None (all data)
541
-0.01
0.24
-2.31
42.02
0.16
PM2.5 total carbon
None (all data)
549
1.34
1.83
88.98
121.65
0.05
Oct 2014
MDA8 ozone
None (all data)
1,308
-2.46
7.68
-4.40
13.73
0.52
MDA8 ozone
Modeled MDA8 03 > 60 ppb
268
-2.99
7.89
-4.31
11.38
0.32
MDA8 ozone
Observed MDA8 03 > 60 ppb
503
-10.15
10.87
-14.54
15.57
0.30
PM2.5 nitrate ion
None (all data)
77
-0.21
0.32
-50.54
75.59
0.43
PM2.5 sulfate ion
None (all data)
77
-0.12
0.21
-23.80
42.91
0.26
PM2.5 total carbon
None (all data)
71
0.77
1.03
66.70
89.49
0.75
Metrics are aggregated over all monitors in the model domain for each modeling period.
A-24
DRAFT: Do Not Cite or Quote
-------
PP.12.5 total carbon
«-
£
CTi
A
"O
Measured Ijig/m \
PM2.5 sulfate Ion
«-
£
Q"
A
CO
d
o
Q
O
1—r
0.4 o.e 1.2
Measured |jig/m3}
S -
*r
31
o
MDA8 Ozone
*
base Iine2018_r3_TC6
1 1 1 1 J—
0 20 40 GO 80 100
Measured Ijig/m )
PM2.5 nitrate Ion
«-
CT
3,
ci
P
o
0.4 0.8 1.2
Msasured Ijig/m3}
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; TC6 = Timber Crater 6.
Model prediction-observation pairs represent monitor locations in the study area region during the 2018 modeling period used to
support the Timber Crater 6 scenarios.
Figure A.5 MPE-1 Daily average maximum daily 8-hour average (MDA8) ozone
and speciated components of PM2.5 including total carbon,
sulfate ion, and nitrate ion model predictions paired with
routine surface monitor data in space and time.
A-25
DRAFT: Do Not Cite or Quote
-------
PP.12.5 total carbon
MDA8 Ozone
«-
£
cn
3
"D
n"
£
cn
A
«—
£
CT
A
"D
0 5
-1 1—
10 15
t—1—1—1—1—1—1—r
0 10 30 50 70
Measured fjig/rrf)
Measured Ijig/m3)
PM2.5 sulfate Ion
PM2.5 nitrate Ion
0 _
20lfeffi.b3£3line2019 fl
0 _
201909.basalins2019 it
/
w
b
*<
• •
ri-
£
CO _
b
CO
rr:
5,
CD
*
•
0
• ¦
"D
ci
*
¦*r
• •
JB
* •
o
•»
-8
O ~~
% •
§
IN
» *
0
p _
Q
A
ci
0. _
CD
A-'
QJQ 0.2 0.4 0.6 0.8 1.0
Measured liig/m3)
QJQ 0.2 0.4 0.6 0.B 1.0
Measured Ijig/m3}
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm.
Model prediction-observation pairs represent monitor locations in the study area region during the 2019 fall modeling period used to
support the Timber Crater 6 prescribed fire scenarios.
Figure A.5 MPE-2 Daily average maximum daily 8-hour average (MDA8) ozone
and speciated components of PM2.5 including total carbon,
sulfate ion, and nitrate ion model predictions paired with
routine surface monitor data in space and time.
A-26
DRAFT: Do Not Cite or Quote
-------
PP.12.5 total carbon
MDA8 Ozone
«-
£
CTi
A
"O
Q
CO
Q
Q
o
Measured Ijig/m \
PM2.5 sulfate Ion
«-
£
cn
A
P
o
QjQ 0.5 1.0 1.5
Msasured liig/m3)
«—
5
JL'
8 -
3 -
8 -
R
p _
«-
£
A
O
rj
2019G2 .basel ine2019_r1
1—i—i—i—r
0 10 20 30 40 50 60
Measured Ijig/m3)
PM2.5 nitrate Ion
p
p
1.0 2JQ
Msasured Ijig/m^;
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm.
Model prediction-observation pairs represent monitor locations in the study area region during the 2019 winter modeling period used
to support the hypothetical slash/pile burn scenarios.
Figure A.5 MPE-3 Daily average maximum daily 8-hour average (MDA8) ozone
and speciated components of PM2.5 including total carbon,
sulfate ion, and nitrate ion model predictions paired with
routine surface monitor data in space and time.
A-27
DRAFT: Do Not Cite or Quote
-------
ri~
£
cn
3
P
o
PM2.5 total carbon
s -
bassi»te2015_rt
•
rf-
F
±1
Ui
A
Ui
A
TJ Q _
JE ™
-8
"~
JE
-8
£ S -
O -
•
§
MDA8 Ozone
t i r
10 20 30 40
Measured Ijig/m3)
PM2.5 sulfate ton
O
CO
O
CO
9
R
«-
CT
3,
1—i—i—i—r
0 30 to 60 B0 100
Measured Ijig/m3)
PM2.5 niIrate ton
Measured |jjg/m
0 1 2 3 4 5 6
Measured Ijig/m3}
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm.
Model prediction-observation pairs represent monitor locations in the study area region during the 2015 modeling period used to
support the Rough Fire scenarios.
Figure A.5 MPE-4 Daily average maximum daily 8-hour average (MDA8) ozone
and speciated components of PM2.5 including total carbon,
sulfate ion, and nitrate ion model predictions paired with
routine surface monitor data in space and time.
A-28
DRAFT: Do Not Cite or Quote
-------
PP.12.5 total carbon
MDA8 Ozone
«-
£
CTi
A
"O
CM
O
p
CM
€
en
a
in
c>
Q
o
bassline201Q r2
Measured Ijig/m }
PM2.5 sulfate Ion
basal ine2010 r2
«—
5
31
8 -
«-
CT
3,
50 100
Msasured Ijig/m3)
PM2.5 nitrate Ion
150
Measured Ijig/m
1 2 3
Msasured Ijig/m^;
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm.
Model prediction-observation pairs represent monitor locations in the study area region during the 2010 modeling period used to
support the Sheep Complex Fire scenario.
Figure A.5 MPE-5 Daily average maximum daily 8-hour average (MDA8) ozone
and speciated components of PM2.5 including total carbon,
sulfate ion, and nitrate ion model predictions paired with
routine surface monitor data in space and time.
A-29
DRAFT: Do Not Cite or Quote
-------
PP.12.5 total carbon
MDA8 Ozone
«-
£
CTi
A
"O
Measured Ijig/m \
PM2.5 sulfate Ion
«-
£
cn
A
—i—i—i 1—i—r
QjQ 0.4 0.8 1.2
Measured liig/m3)
"O
JL'
S -I
R -
«-
CT
3,
baseline2014_r1
1 1 1 T
0 20 40 60 80 100
Measured Ijig/m3)
PM2.5 nitrate Ion
1 2 3
Msasured
MDA8 = maximum daily 8-hour average; |jg/m3 = micrograms per cubic meter; PM2.5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm.
Model prediction-observation pairs represent monitor locations in the study area region during the 2014 modeling period used to
support the hypothetical Boulder Creek Unit 1 prescribed fire scenario.
Figure A.5 MPE-6
Daily average maximum daily 8-hour average (MDA8) ozone
and speciated components of PM2.5 including total carbon,
sulfate ion, and nitrate ion model predictions paired with
routine surface monitor data in space and time.
A.5.2. Supplemental Materials for Section 5.2.2: Surface Fuel Loads
A.5.2.1. Introduction
1 Supplementary materials included here for Section 5.2.2 provide additional details on methods
2 used to develop LEMMA-initialized Visualizing Ecosystem Land Management Assessments (VELMA)
3 applications and associated VELMA-Fuel Characteristic Classification System (FCCS) fuelbed databases
4 for the Timber Crater 6 (TC6), Rough, and Sheep Complex case study applications.
A-30
DRAFT: Do Not Cite or Quote
-------
1
2
Extensive technical and quality assurance documentation is referenced in U.S. EPA"s ScienceHub
data repository (https://catalog.data.gov/dataset/epa-sciencehub).
A.5.2.2. Quality Assurance Project Plan
3 U.S. EPA has established quality assurance requirements that must be followed within U.S. EPA
4 and by extramural contractors for all work performed that involves environmental data collection, use or
5 reporting, including modeling-related activities. Consistent with these requirements, all work performed
6 and reported herein using U.S. EPA s VELMA model follow the VELMA Modeling Quality Assurance
1 Project Plan (OAPP) (McKanc. 2020).
8 The VELMA Modeling QAPP describes quality assurance practices relevant to all VELMA
9 applications, such as those described in this report. These practices concern issues of data quality,
10 calibration, validation, propagation of error and other considerations outlined in the Table of Contents
11 (Figure A.5-1).
Table of Contents
Distribution List: iv
A. RESEARCH MANAGEMENT 1
A1 Organization 1
A2 Problem Definition/Background 2
A3 Research Description 10
A4 Data/Model Quality Objectives and Acceptance Criteria 10
A5 Special Training Requirements/Certification 12
A6 Documentation and Records 12
B. MEASUREMENT / DATA ACQUISITION 13
B1 Data Requirements for Model Input 14
B2 Secondary Data Requirements and Management Tasks 14
C. ASSESSMENT I OVERSIGHT 16
C1 Model Evaluation Methods and Activities 17
C2 Model Evaluation 17
C2.3 — Sensitivity Analysis 19
C3 Audits 19
C4 Reports to Management 20
D. EVALUATION of USABILITY 20
E. REQUIREMENT DIFFERENCES for CATEGORY B 20
VELMA Framework Page iv
L-PESD-30840-QP-1-2 May 2020
VELMA = Visualizing Ecosystem Land Management Assessments.
Figure A.5-1 Quality assurance topics addressed in the Visualizing Ecosystem
Land Management Assessments (VELMA) Modeling Quality
Assurance Project Plan (McKane. 2020).
A-31
DRAFT: Do Not Cite or Quote
-------
1
2
3
The QAPP also provides the U.S. EPA computer server secure location containing the VELMA
applications developed for this project. This information includes VELMA model input and output files
used for model calibration and validation, references and other documentation supporting these activities.
A.5.2.3. Methods
A.5.2.3.1. Characterizing Surface Fuel Load Estimates Using the
Fuel Characteristics Classification System
4 The FCCS is a consistent, scientifically based framework that provides a catalogue of fuelbeds
5 across the U.S. that coincide with various cover types, including grasslands, shrublands, woodlands, and
6 forests (Ottmar et al.. 2007). In FCCS, a fuelbed is defined as a relatively homogeneous landscape unit
7 that represents a unique combustion environment. Each fuelbed is separated into categories and
8 subcategories that depict the loading available for fuel and vary depending on the landscape unit being
9 represented (Figure A.5-2).
A-32
DRAFT: Do Not Cite or Quote
-------
Stratum
Category
CANOPY
SHRUBS
NONWOODY VEGETATION
I
Trees, snags, ladder fuels
Primary and secondary layers
Primary and secondary layers
WOODY FUELS
LITTER-LICHEN-MOSS
GROUND FUELS
All wood, sound wood, rotten
wood, stumps, and woody fuel
accumulations
Litter, lichen, and moss layers
Duff, basal accumulations, and
squirrel middens
FCCS = Fuel Characteristic Classification System.
Figure A.5-2 Fuelbed strata and categories included in the Fuel Characteristic
Classification System [FCCS; Ottmar et al. (2007)1.
The fuel load values (U.S. tons C/acre) for each fuelbed category are derived from scientific
literature, fuel databases, and expert knowledge. Further information can be found in Ottmar et al. (2007).
For the purposes of this investigation, the FCCS fuelbeds were matched to case stud}' regional boundaries
using existing vegetation type layers obtained from LANDFIRE (http://landfire.cr.usgs.gov/viewer/).
The resulting FCCS data then comprised a raster file that described unique identification codes that
represented various fuelbed types as well as a look-up table that provided fuelbed loading values (in
tons/acre) for each of the fuelbed categories and subcategories.
While FCCS captures the general diversity of available fuels found throughout the U.S., the fuel
loadings are summarized across all plots within a particular vegetation classification category. Studies
A-33
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
suggest that accuracy of FCCS and similar vegetation-based approaches are limited due to the high spatial
and temporal variability of fuels, site-specific conditions, and the presence of disturbances including
harvests, prescribed fires, and other disturbances (Lutes et al.. 2009; Brown and See. 1986. 1981).
A.5.2.4. Improving Fuel Characteristic Classification System (FCCS)
Surface Fuel Load Estimates Using Visualizing Ecosystem
Land Management Assessments (VELMA), a Spatially Explicit,
Process-Based Ecohydrological Model
Due to the limitations listed above, a model-based approach was explored to supplement existing
FCCS data to more accurately characterize surface fuel loads that could then be used to simulate air
quality impacts and the effects of prescribed fire for the various real and hypothetical case studies
described in CHAPTER 5 of this report.
A.5.2.4.1. Overview of Visualizing Environmental Land
Management Assessments (VELMA)
The VELMA model is a spatially distributed (grid-based) ecohydrological model that simulates
integrated daily responses of vegetation, soil, and hydrologic components to changes in climate, land use
and land cover. VELMA does this through its linkage of a land surface hydrology model with a terrestrial
biogeochemistry model. The hydrology model simulates water infiltration and redistribution,
evapotranspiration (ET) and surface and subsurface runoff. The biogeochemistry model simulates plant
growth and mortality, formation and turnover of detritus and soil organic matter, and associated cycling
of carbon and nutrients. The interaction of hydrological and biogeochemical processes in the model
constrain changes in ecosystem structure and function in response to various environmental changes
including management. VELMA simulates land management activities in a spatially and temporally
explicit manner, such as harvest, prescribed fire, and wildfire, among other potential treatments (McKane
et al.. 2014). VELMA has been applied in many terrestrial ecosystem types, including forests, grasslands,
agricultural lands, floodplains, alpine and urban landscapes (Barnhart et al.. 2021; Hoghooghi et al.. 2018;
McKane et al.. 2016; Barnhart et al.. 2015; Abdelnour et al.. 2013; Abdelnour et al.. 2011). Particularly in
forests and rangelands, it has been used to simulate effects of fire and harvest on ecosystem structure and
function and subsequent recovery, including impacts on ecosystem services vital to human health and
well-being (McKane et al.. 2018; Yee et al.. 2017).
As noted above, a main advantage of using VELMA to supplement FCCS surface fuel load
estimates are that FCCS data varies by fuelbed but each fuelbed does not vary spatially or temporally.
This means that two cells with the same fuelbed classification will give the exact same surface fuel load
estimations, regardless of their location in the watershed. Conversely, VELMA can be initialized using
spatially distributed aboveground biomass or forest age data that are location and condition-specific to a
A-34
DRAFT: Do Not Cite or Quote
-------
1 defining year. During a simulation, live and dead biomass pools within any watershed pixel can change
2 daily based as a function of water availability, temperature, soil type, and landscape position, as well as
3 any management actions (e.g., clearcutting, thinning, fire) that the user has specified. Therefore, VELMA
4 can capture spatial variations in live and dead biomass pools attributable to spatially and temporally
5 varying conditions within the landscape. For example, VELMA's forest harvest and forest burn tools
6 make it possible to simulate reductions in live and dead fuel loads and subsequent rates of recovery.
7 As discussed in Section 5.2.2. our goal in combining FCCS and VELMA fuelbed information is
8 to improve the accuracy of spatial and temporal surface fuel load estimates and, therefore, the accuracy of
9 the BlueSky and Community Multiscale Air Quality (CMAQ) air quality models and, ultimately, the
10 accuracy of Benefits Mapping and Analysis Program (BenMAP) and associated tools used to assess air
11 quality impacts on human health at local and regional scales (Figure A.5-3).
VELMA-FCCS -> BlueSky ->
Fuel loads, Fire,
Mgmt impacts Smoke
CMAQ -> BenMap
Atmospheric Human Health,
chemistry Economics
BenMAP = Benefits Mapping and Analysis Program; CMAQ = Community Multiscale Air Quality; FCCS = Fuel Characteristic
Classification System; VELMA = Visualizing Ecosystem Land Management Assessments.
Figure A.5-3 Generalized model-to-model workflow for this study.
A.5.2.4.1.1. Visualizing Ecosystem Land Management Assessments
(VELMA) Inputs and Initialization
12 Model inputs and simulation methods varied depending on the case study being
13 implemented—Timber Crater 6, Rough, or Sheep Complex. In this section we summarize the full range
14 of methods and discuss in subsequent sections how specific steps were implemented for each case study.
15 These steps include:
16 1. Acquire satellite-based LEMMA data to develop a spatial (30-m) description of total
17 aboveground forest biomass and stand age for a specified landscape and year (Figure A.5-4).
18 2. Use Step 1 LEMMA data to generate spatial carbon and nitrogen pools for VELMA's 13 plant
19 and soil state variables, per U.S. EPA VELMA documentation, How To Create VELMA Spatial
20 Chemistry Pools.docx (McKane et al.. 2014). This procedure resulted in carbon and nitrogen pool
21 look-up tables for stand ages ranging from 0 to 400 years-old. See Figure A.5-5 for an example
22 illustrating age-related (successional) changes in aboveground stem biomass.
23 3. Initialize VELMA using Step 2 spatial plant and soil carbon and nitrogen pool data. Initialization
24 also requires the additional environmental spatial data described in Table A.5-2.
A-35
DRAFT: Do Not Cite or Quote
-------
4. Use the fully initialized VELMA model (Step 3) to conduct specified actual and hypothetical fire
treatments for case study locations. Note: depending on a case study's end goals of combining
FCCS and VELMA fuelbed information, Steps 3 and 4 may not be necessary.
^ Forest Inventory Analysis (FIA)
LandTrendr
r ^efO_
xJ x =
X = FIA Variable of Interest
* For VELMA: Biomass or Age
References:
1. Landsat Science: https://landsat.esfc.nasa.gov/
2. FIA: https://www.fia.fs.fed.us/
3. LandTrendr:
http://geotrendr.ceoas.oregonstate.edu/landtrendr/
4. Kennedy, R.E., Yang, Z., Cohen, W.B., Pfaff, E.,
Braaten, J. and Nelson, P., (2012). "Spatial and
temporal patterns of forest disturbance and regrowth
within the area of the Northwest Forest Plan."
Remote Sensing of Environment, 122, pp.117-133.
5. LEMMA Project:
https://lemma.forestry.oregonstate.edu/data
6. VELMA Version 2.0 User Manual and Technical
Documentation, Appendix 3
<3
r!
GNN Model
(Gradient Nearest Neighbor)
Gridded Data of Variable X
Age as year, or
Biomass as grams of carbon/meter2
U
V^Ref:
Python Based Processing Tool
"SpatialPoolsCommandLine.py"
Results are VELMA Spatial Pools
Table 2. Additional nitroge
spatial
pools derived from C:N rat
os of above
ground carbon pools.
Spatial Data Pool Name
Unit Type
N g leaf.asc
Ng/mJ
N e AsStem .as c
Ng/m2
N g BeStem.asc
N g/m3
N s root Lasc
Ng/m2
N b root 2.asc
Ng/m1
N g root 3asc
Ng/m2
N b root 4.asc
Ng/m2
N b Pet leaf.asc
Ng/m2
N g Pet AgStem.asc
Ng/m2
N b Pet BgStem lasc
Ng/m2
N s Pet BeStem Zasc
Ng/m3
N s Pet BeStem 3.asc
Ng/m2
N ° Pet_BgStem_4.asc
Ng/m2
N e Pet root l.asc
Ng/m1
N_g_Pet_root_2.asc
Ng/m3
N b Pet root 3.asc
Ng/m2
N ° Pet root 4.asc
Ng/m2
N s Humus lasc
Ng/m'
N B Hu mus_2_ asc
Ng/m1
N a Hu mu5_3. asc
Ng/m2
N g Hu mus_4. asc
Ng/m2
LEMMA = Landscape Ecology, Modeling, Mapping, and Analysis; NASA = National Aeronautics and Space Administration;
FIA = Forest Inventory Analysis; GNN = gradient nearest neighbor; VELMA = Visualizing Ecosystem Land Management
Assessments.
Figure A.5-4 Procedures for acquiring Landscape Ecology, Modeling, Mapping,
and Analysis (LEMMA; LandTrendr/gradient nearest neighbor
[GNN]) high-resolution (30-m) satellite data used to initialize
Visualizing Ecosystem Land Management Assessments (VELMA)
for the case studies described in this report.
A-36
DRAFT; Do Not Cite or Quote
-------
Table A.5-2 Spatial data type, source, and years used to initialize Visualizing
Ecosystem Land Management Assessments (VELMA) for case study
simulations.
VELMA Data Type
Source
Year
Timber Crater 6 Setup
Weather drivers
PRISM: precipitation and mean air temperature
2010 through 2019
httDs://Drism. oreaonstate.edu/exDlorer/
Elevation
USDA Data Gateway DEM):
httDs://dataaatewav.nrcs.usda.aov/GDGOrder.asDX
2019
Age
LEMMA:
httDs://lemma. forestrv.oreaonstate.edu/data
2010
Biomass
LEMMA:
https://lemma. forestry, oreaonstate.edu/data
*Note: LEMMA above ground biomass undergoes
a unit conversion, then processed through
VELMA's preprocessing tool
"Spatial_Pools_Py3_CommandLine.py" script.
2010
Coverage
Uniform
*Note: FCCS coverage for nonforested cells was
included during the combining of FCCS and
VELMA fuelbed information step.
NA
Soils
Uniform (TC6 Der Remillard (1999))
NA
Rough and Sheep Complex Setups
Age
LEMMA:
https://lemma. forestry, oreaonstate.edu/data
2012
Biomass
LEMMA:
https://lemma. forestry, oreaonstate.edu/data
*Note: LEMMA above ground biomass undergoes
a unit conversion, then processed through
VELMA's preprocessing tool
"Spatial_Pools_Py3_CommandLine.py" script.
2012
Coverage
Uniform
*Note: FCCS coverage for nonforested cells was
included during the combining of FCCS and
VELMA fuelbed information step.
NA
DEM = digital elevation model; FCCS = Fuel Characteristic Classification System; LEMMA = Landscape Ecology, Modeling,
Mapping, and Analysis; TC6 = Timber Crater 6; USDA = U.S. Department of Agriculture; VELMA = Visualizing Ecosystem Land
Management Assessments.
A-37
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Timber Crater 6 case study. This study site was set up using Steps 1 through 4 so that potential
alternate scenarios could be completed prior to the actual forest fire event date. Model initialization
occurs for the Year 2010 to leave open the possibility of simulating prefire land management actions prior
to the actual August 2018 actual fire. Simulations carried out to date were restricted to actual landscape
conditions.
Rough and Sheep Complex case studies. These case study sites were developed only up to Step 1,
above, then jumped directly to the step of combining FCCS and VELMA fuelbed information, described
in Section A.5.2.4.1.5. In this case, due to fortuitous data timing, the VELMA biomass data was acquired
from the time-zero LEMMA biomass and age data initialization and represented the forest state for the
actual scenario. If future work requires alternate scenarios of land management actions within these sites,
LEMMA biomass and age data initialization should occur for years preceding the actual fires to allow
VELMA to be initialized and set up to simulate prefire fuelbed treatments.
LEMMA data capture the effects of fine-scale annual changes in aboveground forest biomass
associated with fire, harvest, road construction and other disturbances that have occurred since 1990
across California, Oregon, and Washington. LEMMA data quality is keyed to U.S. Forest Service (USFS)
Forest Inventory and Analysis (FIA) survey data, along with extensive local- and regional-scale validation
against independent LiDAR-based forest survey methods (Bell et al.. 2018).
In practice, age-related biomass trajectories Figure A.5-5) take the form of look-up tables,
developed using the LEMMA-based procedure described for steps 1 and 2 in this section for initializing
spatial (30-m grid) carbon and nitrogen pools for VELMA's 13 plant and soil state variables across a
landscape. See Figure A.5-9 for a 3-D visualization of spatial variability in aboveground live forest
biomass for a LEMMA-initialized landscape for the TC6 case study. Figure A.5-10 is a histogram
showing the number of 30-m pixels represented in Figure A.5-9 across the full range of aboveground
biomass values for this case study domain (Figure A.5-8).
Note that age-related biomass trajectories, such as the example in Figure A.5-5. are used for the
sole purpose of spatially initializing time zero plant and soil carbon and nitrogen pools for landscapes
simulated using VELMA. Simulated trajectories from Day 1 forward are a function of environmental
forcing variables, such as climate, nutrient availability, and disturbances. For example, simulation of a
heavily irrigated and fertilized Ponderosa pine forest could potentially follow a steeper trajectory than that
shown for ponderosa pine (orange line) in Figure A.5-5.
A-38
DRAFT: Do Not Cite or Quote
-------
40000
35000
30000
? 25000
E
C 20000
O
.Q
i—
U 15000
oa
10000
5000
0
Year
VELMA = Visualizing Ecosystem Land Management Assessments; g carbon/m2 = grams of carbon per square meter.
Figure A.5-5 Age-related changes (successional trajectories) in aboveground
stem biomass for Douglas fir and ponderosa pine growing in
western and eastern Oregon, respectively.
A.5.2.4.1.2. Model Calibration and Performance
1 Prior to this study, Pacific Northwest VELMA applications focused on productive, high biomass
2 Douglas fir/western hemlock forest ecosystems growing on the moist west side of the Cascade Range in
3 Oregon and Washington (annual precipitation range -2,000-3,500 mm). For those applications a single
4 set of VELMA model parameters, calibrated for the HJ Andrews Experimental Forest (McKane et al..
5 2014; Abdelnour et al.. 2013; Abdelnour et al.. 2011). has accurately simulated hydrological and
6 biogeochemical responses across dozens of watersheds in western Oregon and Washington, after
7 accounting for location-specific climate and soil nutrient status (Figure A.5-6).
Age-related aboveground stem biomass trajectories used to spatially initialize
VELMA for contrasting forest ecosystems in Oregon
West-side Douglas fir
Janisch & Harmon 2002
East-side Ponderosa pine
Smithwick et al. 2002
0 50 100 150 200 250 300 350 400
A-39
DRAFT: Do Not Cite or Quote
-------
Snohomish River
•k Mashel River <¦
Washington
Trask River
-------
1 Regarding (2), deep volcanic Mazama ash soils in the vicinity of TC6/Crater Lake contain about 1/4 as
2 much soil nitrogen as HJ Andrews sandy loam soils (Rem il lard. 1999).
3 We ran the LEMMA-initialized TC6 VELMA from 2010 to 2100 to examine initial amounts and
4 long-term successional trajectories of live and dead forest biomass pools relevant to fuel load assessments
5 developed for this study (Figure A.5-7). Although no U.S. Forest Service Forest Inventory and Analysis
6 plots are located within the TC6 study area, published data describing observed biomass for mature
7 ponderosa pine forests at the U.S. Forest Service Pringle Falls Experimental Forest are available to assess
8 model performance.
A-41
DRAFT: Do Not Cite or Quote
-------
CM
E
O
o
CD
to
R -
CD
OJ
T
>
<
Q
< s
3
C
c
<
Foliage Biomass
HJ Andrews mature Douglas Fir
i Measured: Sollins et al. 1981
! Modeled: McKane et al. 2014
Eastern Oregon mature Ponderosa
Modeled TC6: this study
Measured: Smithwick et al. 2002
2020
2040
2060
2080
2100
Aboveground Live Woody Biomass
HJ Andrews mature Douglas Fir
Modeled: McKane et al. 2014
J Measured: Sollins et al. 1981
Eastern Oregon mature Ponderosa
Modeled TC6: this study
Measured: Smithwick et al. 2002
2020 2040 2060 2080 2100
Aboveground Dead Woody Biomass
HJ Andrews mature Douglas Fir
Measured: Sollins et al. 1981
(J) Modeled: McKane et al. 2014
Eastern Oregon mature Ponderosa
1
Modeled TC6: this study
i Measured: Smithwick et al.
2002
2020
2040
2060
Year
2080
2100
g C/m2 = grams of carbon per square meter; TC6 = Timber Crater 6; VELMA = Visualizing Ecosystem Land Management
Assessments.
Also shown are modeled and observed biomass data for the HJ Andrews mature Douglas fir/western hemlock forest in western
Oregon for which VELMA hydrological and biogeochemical parameters were calibrated (McKane et al„ 2014: Abdelnour et aL
2013: Abdelnour et aL. 2011) and only recently applied without changes to the TC6 site. See text for details.
Figure A.5-7
Visualizing Ecosystem Land Management Assessments (VELMA)
simulated biomass trajectories {2010 to 2100) for the Timber
Crater 6 (TC6) case study site versus observed biomass for a
mature eastern Oregon ponderosa pine forest [Pringle Falls
Experimental Forest reference stand PF29; Smithwick et al.
(2002)1.
A-42
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Modeled TC6 biomass trajectories from 2010 to 2100 for stand-level foliage, live aboveground
woody biomass, and dead aboveground woody biomass are in good agreement with long-term observed
targets for mature ponderosa pine near Pringle Falls, OR. The TC6 and Pringle Falls forest sites are
located on the same nutrient poor Mazama ash soil type, formed about 7,700 years ago when Mt. Mazama
erupted, leading to the formation of Crater Lake.
Also shown in Figure A.5-7 are modeled and observed biomass data for the HJ Andrews mature
Douglas fir/western hemlock forest site (Watershed 10) for which VELMA hydrological and
biogeochemical parameters were calibrated and applied to TC6. Taken together with the TC6 ponderosa
pine results, Figure A.5-7 indicates that the limited availabilities of water and nutrients in eastern Oregon
strongly constrain biomass growth and accumulation compared to conditions at the HJ Andrews site in
western Oregon.
These results are encouraging for future VELMA applications, suggesting that it will be possible
to use a single, broadly applicable set of VELMA parameters to closely approximate biomass and fuel
load dynamics across large landscapes that include steep, complex gradients of climate, soil, vegetation,
and disturbance histories. The availability of publicly-accessible spatial and temporal databases for all of
these variables—with LEMMA annual 30-m forest biomass estimates going back to 1990—make such
VELMA applications possible for essentially any forested site in California, Oregon, and Washington.
VELMA case study applications for the TC6, Rough, and Sheep Complex fires are discussed in
the following sections.
A.5.2.4.1.3. Case Study 1: Timber Crater 6 (TC6)
VELMA simulations were conducted for the landscape surrounding the Timber Crater 6 Fire that
occurred in south-central Oregon, near Crater Lake National Park, from July 21-26, 2018 (Figure A.5-8).
The TC6 actual fire burned -3,100 acres of forest cover dominated by mixed-age ponderosa pine and red
fir. VELMA was used to simulate biomass/fuel loads for two main boundaries, including the actual TC6
burn area (red area in Figure A.5-8) and the worst-case hypothetical scenario (dotted line in Figure A.5-
8).
A-43
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
1
J
EH.
-3J 3-Miles
0 13 5 Kilometers
Elevation (meters)
2572
~
Timber Crater 6
Bum Area
< 1 Worst Case
1 1 Counterfactual
Diamond Lake
r
Diamond La
K NGA MAS* f IjlM, N
FCCS = Fuel Characteristic Classification System; TC6 = Timber Crater 6; VELMA = Visualizing Environmental Land Management
Assessments.
Figure A.5-8 Study location of the actual Timber Crater 6 (TC6) Fire (red
shaded area) and the maximum extent of hypothetical fire
treatments, for which surface fuel load estimates were made by
harmonizing products from both the Fuel Characteristic
Classification System (FCCS) and the Visualizing Environmental
Land Management Assessments (VELMA) model.
As described in Section A.5.2.4.1.2. initial (time zero) aboveground total (live and dead) biomass
estimates for the TC6 region were obtained from gradient nearest neighbor (GNN) forest biomass and
species maps for 2010 from the Landscape Ecology, Modeling. Mapping, and Analysis (LEMMA)
project at Oregon State University (Kennedy et al.. 2018; Davis et al.. 2015).
The total simulation area was divided into four separate areas due to the large spatial extent and
since VELMA is a watershed model that depends on hydrologically created boundaries. Each of the four
areas were simulated separately by VELMA and the results were subsequently stitched together to
encapsulate the full fire boundary area. Each simulation began in 2009 to stabilize all pools prior to
initialization. Spatially distributed biomass quantities from LEMMA were then incorporated on 2010
Julian Day 1. Each simulation was then conducted until 2020, but the relevant surface fuel loads for 2018
Julian Day 201 (July 20, 2018), which represent the day prior to the start of the TC6 fire, were used for
A-44
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
subsequent analysis. Gridded inputs of elevation, land use/land cover, and soils were collected and
rescaled to match the 30-m resolution of the FCCS/LANDFIRE vegetation cover data it was intended to
supplement.
The study site digital elevation model (DEM) was clipped from the national elevation data set
(NED) acquired from the U.S. Geological Survey (USGS) and rescaled from a 1/3 arc-second resolution
to 30 m. The 30-m DEM was flat-processed using the JPDEM-Dredge processing tool (McKane et al..
2014; Pan et al.. 2012). JPDEM was also used to derive the stream network based on existing elevation
changes.
A single forest ecosystem calibration of VELMA was applied to TC6 that has been found to be
broadly applicable to Pacific Northwest coniferous forest types, including the ponderosa pine ecoregion
of eastern Oregon. Model initialization and validation details are described in Section A.5.2.4.1.2.
Daily precipitation and temperature drivers were obtained Oregon State University's PRISM
Climate Group for 2010-2020 and consist of climatologically aided interpolation (CAI) values that use
both long-term (30-year) averaging and radar measurements as inputs. For more information, see Daly et
al. (2008) and https://prism.oregonstate.edu/explorer/. No stream flow data were available for
hydrologic validation for this particular region. Nonetheless, VELMA's ability to model hydrologic
processes with minimal calibration has been shown to be regionally robust (Figure A.5-6).
VELMA's simulation outputs include a suite of environmental parameters that can be used to
model and better understand spatial and temporal variability in ecosystem properties that result from
differences in climate, wildfire, management, and other disturbances. Responses modeled include changes
in live and dead aboveground and belowground biomass components, stream flow, stream temperature,
stream nutrients and contaminant, and others.
For this case study, VELMA was used to simulate aboveground live and dead biomass pools
corresponding to fuel loadings for forest overstory trees (excluding near-surface fuels such as downed
coarse woody debris, shrubs, etc. that are not easily detected using Landsat-based satellite technology
such as LEMMA). These fuel categories were simulated at 30-m resolution and a daily time step. These
spatial and temporal resolutions can be aggregated to lower resolutions using spatial and temporal
averaging techniques.
Specifically, for the Timber Crater 6 application, VELMA's simulated aboveground live biomass
pools, including stem and leaf components, were exported as 30-m raster data sets. These represent the
aboveground live stem and leaf material across the TC6 region that are available on the day prior to the
actual TC6 fire, that is, July 20, 2018. Model performance tests shown in Figure A.5-7 demonstrate
VELMA's capabilities for accurately simulating aboveground biomass pools relevant to fuel load
estimation purposes. Figure A.5-9 shows VELMA's aboveground biomass simulations for the worst-case
hypothetical boundary associated with the TC6 fire.
A-45
DRAFT: Do Not Cite or Quote
-------
1 Figure A.5-10 is a histogram of aboveground stem values, which accounted for the majority of
2 the total aboveground live biomass.
g C/m2 = grams of carbon per square meter; FIA = Forest Inventory and Analysis; TC6 = Timber Crater 6; USFS = U.S. Forest
Service; VELMA = Visualizing Ecosystem Land Management Assessments.
The red line is the simulation boundary for hypothetical TC6 worst-case BlueSky Pipeline modeling scenarios. Spatial variations in
VELMA modeled aboveground biomass (g carbon/m2) range from near zero (white shading) to a maximum of ~10,000 g carbon/m2
(dark green), which corresponds to regional total biomass maxima for ponderosa/lodgepole pine-dominated forests measured on
permanent plots maintained by the FIA network (USFS reference) and by the Pringle Falls Research Natural Area (Smithwick et al..
Figure A.5-9 30-m resolution Visualizing Ecosystem Land Management
Assessments (VELMA) aboveground live forest biomass results
for the Timber Crater 6 (TC6) case study area for the day before
the beginning of the actual TC6 fire on July 20, 2018 (see also
Figure A.5-8).
A-46
DRAFT; Do Not Cite or Quote
-------
286
g Carbon/m2
9850
g Carbon/m2 = grams of carbon per square meter; TC6 = Timber Crater 6; VELMA = Visualizing Ecosystem Land Management
Assessments.
Vertical bars describe the number of 30-m grid cells for the range of biomass values shown on the y-axis. See Figure A.5-9 for
worst-case scenario boundary.
Figure A.5-10 Histogram of aboveground stem biomass simulated by
Visualizing Ecosystem Land Management Assessments (VELMA)
in the worst-case hypothetical scenario associated with the
Timber Crater 6 (TC6) Fire (simulation day: July 20, 2018).
1 Note that a regional maximum observed aboveground biomass of approximately 10,000 g C/m2
2 has been reported by Smithwick et al. (2002) at the nearby Pringle Falls Research Natural Area. Data for
3 this old-growth ponderosa pine forest was used to validate VELMA-simulated biomass in this study, as
4 described in Section A.5.2.4.1.2. Figure A.5-9 shows that this maximum biomass estimate corresponds
5 well with the western portion of the TC6 boundary, which are older and less disturbed. In fuel load terms,
6 this is equivalent to 44.6 U.S. tons C/acre or 89.2 U.S. tons dry wt./acre.
7 These VELMA simulations were used to supplement the FCCS surface fuel load estimations for
8 the TC6 region. The process by which the FCCS and VELMA data products were combined and exported
9 to the BlueSky Pipeline suite of air quality models are described in Section A.5.2.4.1.5.
A-47
DRAFT: Do Not Cite or Quote
-------
A.5.2.4.1.4. Case Study 2: Rough, Sheep Complex, and Boulder
Creek Fires
1 The second case study focused on the 2015 Rough Fire in the Sierra National Forest in California
2 and consisted of a total of 151,000 burned acres (Figure A.5-11). In late August of that year, the fire
3 expanded eastward, encountering areas partially burned in two earlier, less intense fires—the 2010 Sheep
4 Complex wildfire and the 2013 Boulder Creek Prescribed Fire. These earlier fires mostly reduced surface
5 fuels, likely preventing the speed and severity of the rapidly advancing Rough Fire in 2015, at least in
6 those particular areas and points to the east (Figure A.5-11). A National Park Service interactive story
7 map of the Rough Fire clearly illustrates these Rough Fire dynamics
8 (https://www.nps.gov/seki/learn/nature/rough-fire-interactive-map.htm).
Elevation (meters)
-i 3550
I I 2015 Ro»»gh Fire
' ' largest Pen meter
~ 2013 Boulder Creek
Prescribed Fire
~ 2010 Sheep
Complex Fire
Figure A.5-11 Study location of the 2015 Rough Fire, the 2010 Sheep Complex
Fire, and the 2013 Boulder Creek Prescribed Fire.
A-48
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
The fuelbed characterization objectives of this case study were to (1) use LEMMA and
VELMA-based methods to augment and improve accuracies of existing FCCS surface fuel load estimates
within the Rough, Sheep Complex, and Boulder Creek fire boundaries and (2) provide the combined
VELMA-FCCS fuelbed data to the BlueSky Pipeline CONSUME fire simulator.
To accomplish this, similarly to the TC6 case study, aboveground forest biomass estimates for
this case study were obtained from 30-m, satellite-derived forest biomass and species maps from the
Landscape Ecology, Modeling, Mapping, and Analysis (LEMMA) project at Oregon State University
(Kennedy et al.. 2018; Davis et al.. 2015; Kennedy et al.. 2012).
LEMMA data for 2012 were obtained for the extent of the Rough Fire boundary, whereas
LEMMA data for 2010 and 2013 were obtained for the Sheep Complex and Boulder Creek boundaries,
respectively.
As described for the TC6 case study (Section A.5.2.4.1.3). these LEMMA data were processed
through the VELMA Spatial_Pools_Py3_CommandLine.py Python tool that converted aboveground
biomass from a single layer into VELMA's 13 plant and soil carbon and nitrogen pools, which include
forest leaf biomass and aboveground stem wood (boles, branches, twigs) fuelbed categories.
The fuelbed data for aboveground stem and leaf biomass derived from this VELMA/LEMMA
method were directly merged with FCCS fuelbed categories, skipping the multiyear VELMA biomass
spin-up method applied to TC6. For TC6 there was a significant multiyear gap between the TC6 fire year
(2018) and the closest year of available LEMMA data (2010), which necessitated an 8-year VELMA
"spin-up" to account for growth and decay of live and dead biomass/fuelbeds during that time. Because
there was a closer overall match between fire years and corresponding LEMMA data years for the Rough,
Sheep Complex and Boulder Creek fires, it was not necessary to implement the VELMA spin-up step.
A.5.2.4.1.5. Process for Combining Fuel Characteristic
Classification System (FCCS) and Visualizing
Ecosystem Land Management Assessments (VELMA)
Surface Fuel Load Estimations for All Case Studies
A depiction of the process used to conjoin the FCCS and VELMA data for all case studies is
shown in Figure 5-6 from CHAPTER 5 of the Report. The process alters the original landscape units from
FCCS to include new fuelbed categories that incorporate different VELMA-simulated aboveground
biomass values. The CONSUME model within the BlueSky Pipeline is currently set up to accept inputs
using a standard FCCS data format; therefore, VELMA's spatial raster data were processed and
incorporated into the current FCCS data format to form a harmonized data product featuring spatially
variable surface fuel loads.
A-49
DRAFT: Do Not Cite or Quote
-------
1 VELMA's heterogeneous spatial maps of aboveground live stem and leaf biomass simulations
2 were processed into categories, then spatially merged with the FCCS classes. These tasks were carried out
3 in ArcGIS Pro and described below within the ESRI tool framework, though this data processing routine
4 could be performed in most GIS software.
5 First, VELMA biomass data were reclassified into discrete bins based on their value using the
6 "Reclassify" tool. The live aboveground stem and leaf biomass outputs were reclassified into 11 classes,
7 as shown in Table A.5-3.
Table A.5-3 Discrete bin classifications used for Visualizing Ecosystem Land
Management Assessments (VELMA) and Landscape Ecology,
Modeling, Mapping, and Analysis (LEMMA) aboveground biomass
values for each of the case studies.
Timber Crater 6
Rough and Sheep Complex
Bin Numbers
Stem
Leaf
Stem
Leaf
1
0-1,000
0-80
0-3,000
0-60
2
1,000-2,000
80-100
3,000-6,000
60-120
3
2,000-3,000
100-120
6,000-9,000
120-180
4
3,000-4,000
120-140
9,000-12,000
180-240
5
4,000-5,000
140-160
12,000-15,000
240-300
6
5,000-6,000
160-180
15,000-18,000
300-360
7
6,000-7,000
180-200
18,000-21,000
360-420
8
7,000-8,000
200-220
21,000-24,000
420-480
9
8,000-9,000
220-240
24,000-27,000
480-540
10
9,000-10,000
240-260
27,000-30,000
540-600
11
10,000-11,000
260-280
30,000-33,000
600-660
LEMMA = Landscape Ecology, Modeling, Mapping, and Analysis; VELMA = Visualizing Ecosystem Land Management
Assessments.
Note: The average value in each bin range was used as the actual value in the raster.(grams carbon per meters squared).
8
9 Once the VELMA data were reclassified into discrete bins based on their values, the FCCS
10 fuelbed identification raster was joined with the fuelbed loading look-up table that provided loadings for
A-50
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
each of the fuelbed categories using "Add Join". Then, both the VELMA outputs and the FCCS data were
joined together using "Intersect (Analysis)" after first converting to polygons using "Raster to Polygon
(Conversion)." The resulting output provides a combined polygon file with the attribute table containing
both sets of data in a spatially merged representation. The combined VELMA + FCCS polygon layer was
then converted back to a raster using the "Polygon to Raster (Conversion)" and exported as a final raster
layer, plus the tabular data was saved as an Excel file (.xlsx) using "Table to Excel."
While the raster file was now ready to be sent to CONSUME and the BlueSky Pipeline, a number
of processing steps were needed to adjust the exported attribute table so that VELMA information
replaced FCCS data for particular fuelbed categories and that the table followed the appropriate format.
At this step, care was taken to ensure that the units supplied by VELMA were correctly converted to those
used in FCCS. In particular, VELMA simulates aboveground biomass values as g C/m2, whereas FCCS
uses U.S. tons/acre and assumes dry weight biomass. Therefore, we conducted the conversion using the
relationship 1 g C/m2 = 0.0044609 U.S. tons/acre. That value was derived from 1 g = 1.10231 x 10 " U.S.
ton and 1 m2 = 0.000247105 acre. Alternatively, one can specify that 1 U.S. ton = 907,185 g and
1 acre = 4,046.86 m2. These tons of carbon then were converted to tons of dry weight biomass by
assuming that 0.5 g carbon are present in 1 g of dry weight biomass.
In addition, an R software (R Core Team. 2019) processing script was used to convert VELMA's
total aboveground biomass estimates (live stem and leaf) to the appropriate quantity to replace fuel load
defaults in FCCS.
Note that only forested fuelbeds were replaced using VELMA's simulated data, whereas all
grassland and savanna fuelbeds continued to use the standard FCCS inputs. Parameters and equations
from Jenkins et al. (2003) were used to derive component ratios for tree crowns for both hardwood and
softwood species, and these ratios were multiplied by the VELMA's total aboveground biomass for each
of the forested fuelbed classifications to replace the default FCCS "overstory_loading" category. The
"midstory_loading" and "understory_loading" categories were set to zero during replacement to avoid
double counting. All remaining fuelbed categories (e.g., snags, shrubs, litter, duff) continued to use FCCS
default values.
The final outputs of combining FCCS and VELMA data to provide surface fuel loads to the
CONSUME model in the BlueSky Pipeline consisted of two data products. The first was a new
FCCS + VELMA raster file that included new unique fuelbed identification numbers. These fuelbeds
incorporate both FCCS cover type specifications and VELMA's discrete biomass bins. The second data
product was a revised fuel loading look-up table that is used to specify loadings for each of the fuelbed
categories given in the raster. Both outputs can be found here:
file://aa/ord/ORD/DATA/PRIV/CPHEA WFLC Report Mate rial s/VELMA%20QutPut/.
A-51
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
A.5.2.5. Results
As mentioned in the previous methods section, the final outputs that combine FCCS and VELMA
data were used as inputs to the CONSUME model in the BlueSky Pipeline. Note that while FCCS
provides a number of fuel load categories for surface fuel loads (see Figure A.5-2). VELMA is used only
to modify the crown loading estimates for forested cover types. An in-depth comparison of the resulting
fuel loading changes is shown for each of the case studies in the following sections.
A.5.2.5.1. Case Study 1: Timber Crater 6 (TC6)
A comparison of the VELMA and FCCS crown loading estimates for all FCCS forested fuelbeds
that represent greater than 1% of the total TC6 actual fire boundary are shown in Table A.5-4. Note that
the VELMA values shown in the table are averages across all cells that have the FCCS fuelbed name.
VELMA's simulations tend to be generally higher than those from FCCS, where, for example, VELMA
predicts 15.28 U.S. tons/acre of overstory crown loading and FCCS predicts 9.51 U.S. tons/acre. An
exception is the red fir forest, for which FCCS estimates a loading of 24.97 U.S. tons/acre and VELMA
estimates an average value of 14.86 U.S. tons/acre.
Also, it is apparent that VELMA's simulated values are much greater than FCCS values for
fuelbeds characterized by prior disturbance—that is, fuelbeds denoted with "WF 5-10 YR." FCCS data
were obtained from 2012 and so these disturbance categories represent disturbances that occurred
between 2002 and 2007, which therefore may underestimate the actual biomass present during the TC6
fire in 2018.
A-52
DRAFT: Do Not Cite or Quote
-------
Table A.5-4 Comparison of Timber Crater 6 (TC6) study domain crown loading
estimates between Fuel Characteristic Classification System (FCCS)
and Visualizing Ecosystem Land Management Assessments
(VELMA) for all FCCS forested fuelbeds that represent greater than
1% of the total boundary area percentage.
FCCS Fuelbed Name
FCCS*
VELMA*
Area (%)
Pacific Ponderosa Pine
Forest
9.51
15.28
26
Red Fir Forest
24.97
14.86
15
Red Fir-Mountain Hemlock-
Lodgepole Pine-Western
White Pine Forest
16.21
16.11
9
Giant Sequoia—White
Fir—Sugar Pine Forest
9.36
14.43
5
WF 5-10 yr: Red Fir Forest
4.37
14.98
4
Pacific Silver Fir-Mountain
Hemlock Forest
9.38
16.00
3
WF 5-10 yr: Giant
Sequoia—White Fir—Sugar
Pine Forest
0.00
14.59
3
Mature Lodgepole Pine
Forest
4.13
19.07
3
WF 5-1 Oyr: Pacific
Ponderosa Pine Forest
1.48
16.54
3
Pacific Silver Fir-Sitka Alder
Forest
2.33
17.94
2
Ponderosa Pine—Jeffrey
Pine Forest
8.62
14.71
1
FCCS = Fuel Characteristic Classification System; TC6 = Timber Crater 6; VELMA = Visualizing Ecosystem Land Management
Assessments.
Note: The VELMA values represent the crown fuel loads estimated from VELMA's aboveground biomass simulations and Jenkins
Et Al. (2003) tree component ratios, while the FCCS values are the sum of the "overstoryjoading," "midstoryjoading," and
"understoryjoading" fuel load categories, all units are provided in U.S. tons/acre dry weight biomass.
A-53
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
A.5.2.5.2. Case Study 2: Sheep Complex and Rough Fires
A comparison of the FCCS and LEMMA crown loading estimates for all FCCS forested fuelbeds
that represent greater than 1% of the total Sheep Complex actual fire boundary are shown in Table A.5-5.
Note that LEMMA data were only produced for a subset of the total number of fuelbeds due to
lack of data of the component ratios available from Jenkins et al. (2003) that coincide with the FCCS
fuelbed names depicted in the table. When available, these ratios were multiplied by LEMMA's total
aboveground biomass for each of the forested fuelbed classifications to replace the default FCCS
category, as described previously. As with the TC6 case study, the VELMA/LEMMA crown loading
values are lower than the default FCCS values for Red Fir Forest fuelbed type (17.69 vs. 24.97 U.S.
tons/acre, respectively), whereas they match well for the ponderosa (8.61 vs. and 8.62 U.S. tons/acre)
Jeffrey Pine (8.13 vs. 8.33 U.S. tons/acre) mixes and are greater than the FCCS defaults for the Mature
Lodgepole Pine Forest type (18.25 vs. 4.13 U.S. tons/acre).
For the Rough Fire boundary, a comparison of the VELMA and FCCS crown loading estimates
for all FCCS forested fuelbeds that represent greater than 1% of the fire boundary are shown in Table
A.5-6.
A-54
DRAFT: Do Not Cite or Quote
-------
Table A.5-5 Comparison of Sheep Complex study domain crown loading
estimates between Fuel Characteristic Classification System (FCCS)
and Visualizing Ecosystem Land Management Assessments
(VELMA)/Landscape Ecology, Modeling, Mapping, and Analysis
(LEMMA) for all FCCS forested fuelbeds that represent greater than
1% of the total Sheep Complex boundary area percentage.
FCCS Fuelbed Name
FCCS
VELMA/LEMMA
Area (%)
Red Fir Forest
24.97
17.69
18
California Black Oak
Woodland
19.63
14
Douglas Fir-Sugar Pine-
Tanoak Forest
19.30
12
Ponderosa Pine-Jeffrey
Pine Forest
8.62
8.61
3
Douglas Fir-White Fir
Forest
20.94
3
Jeffrey Pine-Red Fir-White
Fir/Greenleaf-Snowbrush
Forest
14.38
3
Jeffrey Pine-Ponderosa
Pine-Douglas Fir-California
Black Oak Forest
8.33
8.13
2
Mature Lodgepole Pine
Forest
4.13
18.25
2
Douglas Fir/Ceanothus
Forest
3.75
2
Subalpine Fir-Lodgepole
Pine-Whitebark Pine-
Engelmann Spruce Forest
9.55
1
FCCS = Fuel Characteristic Classification System; LEMMA = Landscape Ecology, Modeling, Mapping, and Analysis;
VELMA = Visualizing Ecosystem Land Management Assessments.
Note: That tree component ratios used by Jenkins et al. (2003) were unavailable for some FCCS fuelbed cover types, and
therefore crown loading values could not be computed and are shown as blanks.
"The LEMMA values represent the crown fuel loads estimated from LEMMA's aboveground biomass estimates and Jenkins et al
(2003) tree component ratios, while the FCCS values are the sum of the "overstoryjoading," "midstoryjoading," and
"understoryjoading" fuel load categories. All units are provided in U.S. tons/acre dry weight biomass.
A-55
DRAFT: Do Not Cite or Quote
-------
Table A.5-6 Comparison of Rough study domain crown loading estimates
between Fuel Characteristic Classification System (FCCS) and
Visualizing Ecosystem Land Management Assessments
(VELMA)/Landscape Ecology, Modeling, Mapping, and Analysis
(LEMMA) for all FCCS forested fuelbeds that represent greater than
1% of the total Rough Fire area percentage.
Fuelbed Name
FCCS
VELMA/LEMMA
Area (%)
California Black Oak
Woodland
19.63
18
California Live Oak-Blue
Oak Woodland
1.21
17
Douglas Fir-Sugar Pine-
Tanoak Forest
19.30
17
Red Fir Forest
24.97
18.68
15
Jeffrey Pine-Ponderosa
Pine-Douglas Fir-California
Black Oak Forest
8.33
13.18
4
Jeffrey Pine-Red Fir-White
Fir/Greenleaf-Snowbrush
Forest
14.38
3
Douglas Fir-White Fir
Forest
20.94
2
Ponderosa Pine-Jeffrey
Pine Forest
8.62
13.59
2
Subalpine Fir-Lodgepole
Pine-Whitebark Pine-
Engelmann Spruce Forest
9.55
2
Mature Lodgepole Pine
Forest
4.13
19.43
2
Douglas Fir/Ceanothus
Forest
3.75
1
Black Cottonwood-Douglas
Fir-Quaking Aspen Forest
28.68
1
FCCS = Fuel Characteristic Classification System; LEMMA = Landscape Ecology, Modeling, Mapping, and Analysis;
VELMA = Visualizing Ecosystem Land Management Assessments.
LEMMA data were only updated for some cover types due to data availability for converting total aboveground biomass estimates
to crown loadings using equations from Jenkins et al. (2003).
The LEMMA values represent the crown fuel loads estimated from LEMMA's aboveground biomass estimates and Jenkins et al.
(2003) tree component ratios, while the FCCS values are the sum of the "overstoryjoading," "midstoryjoading," and
"understoryjoading" fuel load categories. All units are provided in U.S. tons/acre dry weight biomass.
A-56
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
As expected based on the previous case studies, VELMA estimates lower crown loading values
compared with the default FCCS values for Red Fir Forest. However, the remainder of comparisons show
that VELMA/LEMMA estimate higher crown loading values compared with the FCCS defaults. Further
validation is needed to confirm the canopy estimations from VELMA/LEMMA and their comparison
with the original estimates performed by FCCS. Also, note that the values in Table A.5-4. Table A.5-5.
and Table A.5-6 represent spatial averages of VELMA data for given FCCS cover types to simplify direct
comparison. The combined FCCS/VELMA data products sent to the BlueSky Pipeline, however, include
spatially distributed crown loading estimates that are not fully reflected in the previous tables.
A.5.2.6. Conclusions
The use of vegetation-based fuel load classification systems can be extremely helpful for air
quality modelers to simulate the air quality impacts of historical or projected wildfires. However, these
classification systems are inherently crafted to represent a wide variety of fuel loads across the entire U.S.
and therefore do not always capture the fine spatial and temporal heterogeneity associated with
landscape-level fuel load changes or disturbance patterns. In this study, we used a spatially distributed
ecohydrological landscape model (VELMA) to simulate aboveground live biomass and supplement
existing fuel load characterization data for the Timber Crater 6, Rough, Sheep Complex and Boulder
Creek fire boundaries. VELMA was initialized using LEMMA data that provided spatially distributed
estimates of live aboveground biomass corresponding to the regions of each of the case studies.
As shown in Figure A.5-7 and Figure A.5-9. VELMA fuel load estimates compare well with
measured data describing upper limits of aboveground biomass for ponderosa pine stands in eastern
Oregon (Smith wick et al.. 2002). In addition, VELMA crown loading estimates for forested fuelbeds were
compared with FCCS default values.
While differences exist between VELMA and default FCCS estimates for forest crowns and other
fuelbeds, further assessments of these estimates based on observed data would be beneficial to examine
the validity of surface fuel loads within the regional domain of this study.
Discussions with project partners and others familiar with the case study sites have so far turned
up no available georeferenced forest biomass data for assessing the accuracy of model-based estimates for
the case study sites. For example, there exists high-quality biomass data for Forest Inventory and Analysis
plots within the Rough Fire boundary, but precise coordinates for these plots are inaccessible for security
reasons.
Those challenges notwithstanding, it is important to emphasize that the ability of VELMA to
accurately simulate ecosystem responses across western and eastern Oregon using a single set of model
equations and parameter values provides the strongest possible test of a process-based modeling
framework (Section A.5.2.4.1.2). In essence, VELMA behaves similarly, though imperfectly, to real
A-57
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
ecosystems with regard to changes in structure and function in response to environmental changes,
whether in situ or across landscape gradients.
These results are encouraging for future VELMA applications, suggesting that it will be possible
to closely approximate biomass and fuel load dynamics across large landscapes that include steep,
complex gradients of climate, soil, vegetation, and disturbance histories. The availability of
publicly-accessible spatial and temporal databases for all of these variables—including LEMMA annual
30-m forest biomass estimates going back to 1990—make such VELMA applications possible for
essentially any forested site within some western states most hard hit by recent wildfires—California,
Oregon, and Washington.
Finally, VELMA is already capable of simulating real and hypothetical land management
practices and other disturbances (harvests, wild and prescribed fires, extreme climate events, etc.) at
multiple spatial scales. Therefore, future research could incorporate simulations of alternative prescribed
burning and mechanical thinning practices to explore local and regional impacts on fuel loads and
consequent air quality impacts. Additionally, since VELMA is designed to simulate ecohydrological
processes, it can also be used to assess effects of wild and prescribed fires on water quality and quantity,
thereby providing an opportunity for integrated air and water quality impact assessments on human
health.
A-58
DRAFT: Do Not Cite or Quote
-------
A.6. SUPPLEMENTAL INFORMATION FOR CHAPTER 6
1
A.6.1. Supplemental Information for Section 6.2
Table A.6-1 Study-specific details from U.S.-based epidemiologic studies examining associations between
wildfire smoke exposure and respiratory and cardiovascular-related health effects and mortality.
Study; Location; Fire Exposure Indicator Types of Air Quality
Yr Health Outcomes (Ages) Avg Time Data Used Exposure Assessment Methodology
ED Visits and Hospital Admissions; Medication Use
Alman et al. (2016);
Colorado;
2012 Wildfires
(6/5/12-7/6/12)
ED Visits: asthma and
wheeze, URI, pneumonia,
bronchitis, COPD,
RESPIRATORY disease,
AMI, IHD, dysrhythmia, CHF,
ischemic stroke, PVD, CVD
(All; 0-18; 19-64; 65+)
PM2.5
(24-h avg;
Modeled
1-h max)
WRF-Chem used to estimate PM2.5 concentrations at
12 x 12 km grid cells. Addresses for each patient
geocoded and assigned PM2.5 concentration from
respective grid cell.
Model evaluation: Model absolute bias (i.e., average
difference between model and monitored PM2.5
concentrations), 13 |jg/m3 for six monitoring stations
around Denver Metro Area, 13 |jg/m3 for two stations
north-east of Denver, and 19 |jg/m3 for the station
east of Denver.
A-59
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Health Outcomes (Ages)
Exposure Indicator
Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
DeFlorio-Barker et al.
(2019);
692 U.S. counties
within 200 km of
123 large fires;
>10,000 acres burned
(2008-2010)
HA: respiratory; asthma,
bronchitis and wheezing; all
CVD
(65+)
PM2.5 TotCMAQ;
PM2.5 Tot; PM2.5
TotCMAQ-M
(24-h avg)
Monitored
Modeled
(1) Ambient PM2.5 from monitoring stations (>4,000),
resulting in county-wide averages available for 178 of
692 counties; (2) PM2.5 estimated using CMAQ.
CMAQ estimated PM2.5 at 12 * 12 km grid cells—
estimated PM2.5 with all emissions (PM2.5 TotCMAQ)
and without wildfire (CMAQ NFCMAQ). CMAQ data
used to calculate area-weighted PM2.5 estimates for
each county. Difference between CMAQ estimates
represented fire-specific PM2.5 concentrations (PM2.5
FCMAQ). SmokeDay = PM2.5 FCMAQ > 5 |jg/m3.
Delfino et al. (2009);
Southern California;
2003 Wildfires
(Total:
10/1/03-11/15/03;
Prefire: 10/1-10/2; Fire:
10/21-10/30; Post-fire:
10/31-11/15)
HA: All respiratory, asthma,
acute bronchitis, COPD,
pneumonia, all CVD, IHD,
CHF, dysrhythmia,
cerebrovascular and stroke
PM2.5
(24-h avg)
Monitored
(All; 0-4; 5-
65-99)
19; 20-64;
Combination of monitoring data, continuous hourly
PM data at colocated or closely located sites, and
light extinction from visibility data. Meteorological
conditions and smoke data from MODIS at 250-m
resolution. For smoke periods created polygons from
smoke-covered areas and measured or estimated
PM2.5 concentrations from predictive models to assign
exposures at ZIP code centroid.
A-60
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Health Outcomes (Ages)
Exposure Indicator
Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
Gan etal. (2017):
Washington;
2012 Wildfires
(7/1/12-10/31/12)
HA: All respiratory, asthma,
COPD, pneumonia, acute
bronchitis, CVD, arrhythmia,
cerebrovascular disease, HF,
IHD, Ml
(All; <15; 15-65; 65+)
Smoke PM2.5
(24-h avg)
Modeled
Satellite
(1) WRF-Chem: Estimated daily PM2.5 at 15 * 15 km
grid cell, ran additional simulations with biomass
burning emissions turned off to estimate nonwildfire
smoke PM2.5.
Model evaluation: Slope = 0.67, R2 = 0.25
(2) Kriging in situ surface monitors: Interpolated
monitoring data (212 monitors) to 15 * 15 km grid
cells.
Model evaluation: Slope = 0.70, R2 = 0.69
(3) GWR: estimated PM2.5 concentrations at
15 x 15 km grid cells by combining kriged, AOD, and
WRF-CHEM estimates.
Model evaluation: Slope = 0.78; R2 = 0.66.
To distinguish wildfire PM2.5 for WRF-Chem
subtracted out nonsmoke PM2.5 produced by
WRF-Chem. For kriging and GWR methods
estimated background PM2.5 using NOAAs HMS to
identify days where wildfire smoke not near a monitor.
Smoke plumes in HMS accompanied by estimated
PM2.5 concentrations from atmospheric models.
Calculated median PM2.5 concentration for each
monitor on nonfire days, these concentrations were
interpolated by kriging for each grid cell. These
nonfire PM2.5 concentrations were subtracted from
PM2.5 concentrations for each method to estimate
PM2.5 attributed to smoke.
Gan et al. (2020);
Oregon;
Douglas Complex Fire
Big Windy Complex
Fire
(5/1/13-9/30/13)
HA: Asthma
Medication use: Short-acting
(32 agonists (SABA)
pharmacy refills
(All; <15; 15-65; 65+)
Smoke PM2.5
(24-h avg)
Modeled
Satellite
Similar method as Gan et al. (2017), focusing only on
the GWR method, estimated PM2.5 concentrations at
15 x 15 km grid cells by combining kriged, AOD, and
WRF-CHEM estimates. Monitors used in the analysis
consisted of both FRM and FEM monitors.
A-61
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Health Outcomes (Ages)
Exposure Indicator
Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
Hutchinson et al.
(2018):
San Diego, CA;
2007 Wildfires
ED Visits: Respiratory index,
Asthma
(0-64)
Wildfire PM2.5
(24-h avg)
Modeled
Wildfire Emissions from WFEIS were used in
HYSPLIT to estimate wildfire PM2.5 concentrations at
0.01° grid on an hourly basis. 24-h avg
concentrations calculated at the ZIP code level.
(9/1/07-11/29/07)
Leibel etal. (2020):
San Diego County, CA;
Lilac Fire
(2011-2017; Fire:
12/6/17-12/17/17)
ED Visits and Urgent care
visits: All respiratory
(0-19)
PM2.5
(24-h avg)
Monitored
24-h avg PM2.5 concentrations from 10 fixed site
monitors. PM2.5 concentrations interpolated using
inverse distance interpolation model using stations
within 12 miles from each population weighted ZIP
code centroid, concentrations than averaged and
assigned to each ZIP code. Monitors closest to each
centroid were given greater weight (weighted using
squared inverse distance).
Liu etal. (2017a):
561 Western U.S.
counties;
Wildfire season
(May-October,
2004-2009)
HA: All respiratory, All CVD
(65+)
Wildfire PM2.5; smoke
wave day vs.
nonsmoke wave day
Monitored
Modeled
GEOS-Chem predictions of "all-source PM2.5" and
"no-fire PM2.5" to -50 * 75 km grid cell. Ground based
or aircraft measurements used to validate model
results. Area weighted averaging used to convert
gridded predictions to county-level averages.
GEOS-Chem predictions biased low during extreme
events so model calibrated using county-average
monitoring data. Smoke wave defined as 2+
consecutive days of wildfire PM2.5 > 20 |jg/m3
(98th percentile, sensitivity analyses focusing on
23 |jg/m3 [98.5 percentile], 28 |jg/m3 [99th percentile],
and 37 |jg/m3 [99.5 percentile]).
A-62
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Health Outcomes (Ages)
Exposure Indicator
Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
(Liu etal.. 2017b):
561 western U.S.
counties;
Wildfire season
(May-October,
2004-2009)
HA: Respiratory (COPD and
respiratory tract infections)
(65-75; 75-84; 85+)
Wildfire PM2.5:
wave day vs.
nonsmoke wave day
smoke Monitored
Modeled
GEOS-Chem predictions of "all-source PM2.5" and
"no-fire PM2.5" to -50 * 75 km grid cell. Ground based
or aircraft measurements used to validate model
results. Area weighted averaging used to convert
gridded predictions to county-level averages.
GEOS-Chem predictions biased low during extreme
events so model calibrated using county-average
monitoring data. Smoke wave defined as 2+
consecutive days of wildfire PM2.5 > 37 |jg/m3
(99.5%).
Rappold etal. (2011);
42 North Carolina
counties
Peet Fire in Pocosin
Lakes National Wildlife
Refuge;
(6/1/08-7/14/08)
ED Visits: All respiratory, Smoke plume
COPD, pneumonia and acute
bronchitis, URIs, all CVD, Ml,
HF, dysrhythmia,
respiratory/other chest pain
symptoms
(All; <65; 65+)
Satellite
Half hour, AOD at 4 * 4 km averaged over daytime
hours to assign county-level exposure. AOD > 1.25
classified as high-density plume. Counties where at
least 25% of geographic area of county exceeded
AOD threshold were categorized as high-exposure
window. Counties with smoke exposure on at least
2 days classified as exposed (18 counties);
23 referent counties (15 exposed 1 day; 8 <1 day).
Rappold etal. (2012):
40 North Carolina
counties
Peet Fire in Pocosin
Lakes National Wildlife
Refuge;
(6/1/08-7/14/08)
ED visits: asthma, CHF
(>18; >44)
Wildfire PM2.5
(24-h avg)
Modeled
Satellite
PM2.5 concentrations obtained from NOAA SFS.
PM2.5 concentrations based on smoke dispersion
simulations from HYSPLIT, which relies on satellite
information of wildfire location. Hourly PM2.5
concentrations at 0.15 * 0.15° (-13.5 km) estimated
a lowest 100 m surface area averaged to generate
24-h avg concentrations. Daily averages for each
county calculated over county boundaries using
Monte Carlo approximation. HYSPLIT data not
available for 6/4, underestimating concentrations on
that day.
A-63
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire Exposure Indicator Types of Air Quality
Yr Health Outcomes (Ages) Avg Time Data Used Exposure Assessment Methodology
Reid etal. (2016):
Northern California,
781 ZCTA
(Air Basins:
Sacramento Valley,
San Francisco Bay
Area, Mountain
Counties, Lake County,
North Central Coast,
northern part of San
Joaquin Valley)
Thousands of wildfires
from lightning strikes
June 20-21, located in
Trinity Alps, Sierra NV
and Big Sur;
(Pre fire:
5/6/08-6/19/08; Fire:
6/20/08-7/31/08;
Post-fire:
8/1/08-9/15/08)
ED visits and HAs: All
respiratory, asthma, COPD,
pneumonia, all CVD, IHD,
CHF, dysrhythmias,
hypertension,
cerebrovascular disease
(All; <20; 65+)
PM2.5
(24-h avg)
Monitored
Modeled
Satellite
Data-adaptive machine learning employing 10-fold
CV. Used data from 112 monitoring stations as
dependent variable and predictor variables included
AOD from GEOS, WRF-Chem model output, various
meterological variables, Julian date, weekend, land
use types within 1 km, Xand Y coordinates,
elevation, and traffic counts. Used GBM with six most
predictive variables for the main model. Estimated
exposures at population-weighted centroid of
781 ZCTA.
Model evaluation: CV-R2 = 0.78,
CV-RMSE = 1.46 |jg/m3
A-64
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Health Outcomes (Ages)
Exposure Indicator
Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
Reid etal. (2019):
Northern California,
753 ZIP codes (Air
Basins: Sacramento
Valley, San Francisco
Bay Area, Mountain
Counties, Lake County,
North Central Coast,
northern part of San
Joaquin Valley);
Thousands of wildfires
from lightning strikes
June 20-21, located in
Trinity Alps, Sierra, NV
and Big Sur
(5/6/08-9/26/08)
ED visits: All respiratory,
asthma, COPD, pneumonia,
acute bronchitis, acute
respiratory infections
(AN)
PM2.5
(24-h avg)
O3
(8-h max)
Monitored
Modeled
Satellite
Used exposure model detailed in Reid et al. (2016).
Data-adaptive machine learning employing 10-fold
CV. Used data from 112 monitoring stations as
dependent variable and predictor variables included
AOD from GEOS, WRF-Chem model output, various
meterological variables, Julian date, weekend, land
use types within 1 km, Xand Y coordinates, elevation
and traffic counts. Used GBM with six most predictive
variables for the main model. Estimated exposures at
each ZIP code centroid.
Model evaluation: For PM2.5, CV-R2 = 0.78,
CV-RMSE = 1.46 |jg/m3. ForOs, CV-R2 = 0.83
Stowell et al. (2019):
Colorado;
Wildfire season
(April-September,
2011-2014)
ED visits and HAs: all
respiratory, asthma COPD,
URIs. bronchitis, IHD, AMI,
CHF, dysrhythmia,
peripherial/cerebrovasular
disease, all CVD
(All; 0-18; 19-64; 65+)
Smoke PM2.5
Monitored
Modeled
Satellite
Two model approach where data combined from
AOD from MAIAC, model simulations from CMAQ,
and ground based PM2.5 measurements. Model 1,
used random forest modeling to incorporate AOD
data, smoke mask, meterological fields, and land-use
variables. Second model used statistical downscaling
to calibrate CMAQ PM2.5 predictions. Exposure data
at 1 x 1 km grid cell. To estimate wildfire smoke
PM2.5, CMAQ scenarios with and without smoke and
dust particles. Difference between scenarios divided
by total PM2.5 to obtain smoke fraction which was
multiplied by total satellite-based PM2.sto obtain
smoke PM2.5 concentrations.
Model evaluation: For CMAQ predictions,
C\/-R2= 0.81; RMSE = 1.85 |jg/m. Random forest
model improved R2 from 0.65 to 0.92.
A-65
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Health Outcomes (Ages)
Exposure Indicator
Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
Tinlina et al. (2016):
28 North Carolina
counties with at least
one 24-h avg smoke
PM2.5 concentration >
20 |jg/m3;
Pains Bay Fire
(5/5/11-6/19/11)
ED visits: Respiratory/other
chest symptoms, all
respiratory, asthma, COPD,
URI, All CVD, dysrhythmia,
HF, hypertension
(All; <18; 18-64; 65+)
Wildfire PM2.e
Modeled
County-level daily wildfire PM2.5 estimated from
modeled predictions from NOAA SFS.
Wettstein et al. (2018):
Eight California Air
Basins
(Great Basin Valleys,
Lake County, Lake
Tahoe, Mountain
Counties, North Coast,
Northeast Plateau,
Sacramento Valley,
San Joaquin Valley);
2015 Wildfire season
(May-September,
2015)
ED visits: All CV,
hypertension, IHD, Ml,
dysrhythmia, HF, PE, All
cerebrovascular, ischemic
stroke, TIA, all respiratory
(19+; 45-64; 65+)
Smoke density
Modeled
Smoke plume data from NOAA HMS, assigning daily
maximum density to each ZIP code based off
estimated PM2.5 concentration data where
concentrations within the range of 0-10 |jg/m3
defined as light, 10.5-21.5 |jg/m3 defined as medium
, and 22+ |jg/m3 defined as dense.
Out-of-Hospital Evidents
Jones et al. (2020); OHCA
14 California counties; (19+)
Wildfires >50,000 acres
burned or >50 days
long
(May-October,
2015-2017)
Smoke day Modeled NOAA HMS used to detect plumes using visual range
of satellite images and assigned estimated smoke
PM2.5 density: light (0-10 |jg/m3); medium
(10.5-21.5 |jg/m3); and heavy (>22 |jg/m3). Used
geospatial intersect function to assign smoke data at
the census block group and then aggregated to
census tract, maximum smoke density used to define
exposure.
A-66 DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Exposure Indicator
Health Outcomes (Ages) Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
Mortality
Doubledav et al. (2020); Total (nonaccidental), Smoke day vs.
cardiovascular, IHD, nonsmoke day
Washington;
Wildfire season
(June-September,
2006-2017)
respiratory, asthma, COPD,
pneumonia, cerebrovascular
(AN)
Monitored
Modeled
4 x 4 km grid cells from AIRACT-4, each grid cell
assigned to 1 of 3 AQ monitors closest to each grid
cell out of 75 monitors in Washington. Grid cells
matched to nearby monitors based on agreement
between interpolated and monitored PM2.5. Each grid
cell then assigned the daily PM2.5 monitor
concentration. Smoke day defined as days with PM2.5
monitor concentrations > 20.4 |jg/m3, with additional
criteria if PM2.5 concentrations between 9 and
20.4 |jg/m3: (1: 2 of 3 days > 9 |jg/m3;
2: 1 day > 15 |jg/m3; 3: for urban areas at least
50% monitors > 9 |jg/m3).
Xi et al. (2020);
253 U.S. Counties;
(2008-2012)
All-cause, cardiac, vascular,
infection, other
(50+)
Wildfire PM2.5
Modeled
Ambient PM2.5 concentrations were predicted at
12 x 12 km grid cells using CMAQ with and without
wildland fire emissions. The difference between the
with and without wildland fire emissions represented
wildfire-specific PM2.5. Hourly concentrations were
averaged to calculate a daily county-level 24-h avg
PM2.5 concentration.
A-67
DRAFT: Do Not Cite or Quote
-------
Table A.6-1 (Continued): Study-specific details from U.S.-based epidemiologic studies examining associations
between wildfire smoke exposure and respiratory and cardiovascular-related health
effects and mortality.
Study; Location; Fire
Yr
Health Outcomes (Ages)
Exposure Indicator
Avg Time
Types of Air Quality
Data Used
Exposure Assessment Methodology
Zu etal. (2016):
New York, NY; Boston,
MA;
Total (nonaccidental)
(AH)
PM2.5
Monitored
Daily average PM2.5 concentrations across all
monitors in Boston and each borough in New York.
July 2002 Quebec
Wildfires
(July, 2001-2003)
AIRACT-4 = Air Indicator Report for Public Awareness and Community Tracking; AMI = acute myocardial infarction; AOD = aerosol optical depth; avg = average; CHF = congestive
heart failure; CMAQ = Community Multiscale Air Quality; COPD = chronic obstructive pulmonary disease; CV = cross-validation; CVD = cardiovascular disease; ED = emergency
department; FCMAQ = fused-CMAQ; FEM = Federal Equivalent Method; FRM = Federal Reference Method; GBM = Generalized Boosting Model; GEOS-Chem = Goddard Earth
Observing System with a global chemical transport model; GWR = geographically weighted regression; HA = hospital admissions; HF = heart failure; HMS = Hazard Mapping
System; HYSPLIT = Hybrid Single-Particle Lagrangian Integrated Trajectories; IHD = ischemic heart disease; MAIAC = Multiangle Implementation of Atmospheric Correction
algorithm; max = maximum; Ml = myocardial infarction; MODIS = Moderate Resolution Imaging Spectroradiometer; NOAA = National Oceanic and Atmospheric Administration;
OHCA = out-of-hospital cardiac arrest; PE = pulmonary embolism; PVD = peripheral vascular disease; PM25 = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 |jm; PM25 Tot = monitored PM2.5 data; PM25 TotCMAQ = PM25 estimated using CMAQ; PM2.5 TotCMAQ-M = PM2.5 estimated using CMAQ in locations and times
with monitoring data; SABA = short-acting (32 agonists; SFS = Smoke Forecasting System; TIA = transient ischemic attack; URIs = upper respiratory tract infections;
WFEIS = Wildland Fire Emissions Information System; WRF-Chem = Weather Research and Forecasting Model with Chemistry; yr = year; ZCTA = ZIP-code tabulation areas.
A-68
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
A.6.2.
Supplemental Information for Section 6.3
A literature review was conducted to identify published studies that provide data on individual
and community actions to reduce wildfire smoke exposure. The literature review was limited to studies
published from 2005 to May 2020 with keywords that included wildfire/prescribed fire and smoke, PM2 5,
and exposure, along with terms for actions/interventions (e.g., air filtration). Although several hundred
published studies were identified with the search terms, after reviewing the titles and abstracts only
243 publications were determined to be relevant to wildfire or prescribed fire smoke exposure. Of those,
26 specifically addressed some aspect of smoke exposure mitigation and were included in the discussion
within Section 6.3 of CHAPTER 6.
In order to be most informative in assessing the potential implications of public health messaging
campaigns that attempt to reduce/mitigate population exposure to wildfire smoke around the case study
areas, studies were limited to those conducted in the U.S. and Canada, with a few exceptions. Only three
publications were identified that surveyed the likelihood of taking action to reduce wildfire smoke
exposure in North America, so the literature review was expanded to include studies that were published
before 2005 and from other parts of the world. Two additional studies were included, one published in
2002 conducted in North America and one conducted in Australia. In addition, the only published study
with data on the effectiveness of staying indoors with windows and doors closed was conducted in
Australia and also was included.
A-69
DRAFT: Do Not Cite or Quote
-------
Table A.6-2 Likelihood of taking actions to reduce wildfire smoke exposure reported in recent studies.
Study Exposure Reduction Action
Percent
Population
Taking the
Action
Population Characteristics
Outdoor PM
Concentration.
Hg/m3
Behavioral Changes—Avoid Outdoor Activity
RaDDold et al. (2019) Avoided outdoor activity
61
SmokeSense App users with no reported health history
and no symptoms
NR 1
6
SmokeSense App users reported health history
90
SmokeSense App users experiencing four or more
symptoms
Richardson et al. (2012) Avoided outdoor recreation
78
Residents of five cities within the vicinity of Station Fire
in southern California
Maximum PM2.5
(Daily) = 82.9
(Hourly) = 223
Suaerman et al. (2012) Did not Dlav SDorts outside
88
Residents of San Diego County during the 2007 San
Diego Fires
PM2.5 2 (daily) >128
for 10 days
Maximum PM2.5
(daily) = 803.1
Mean PM2.5
(daily) = 89
(Hutchinson et al..
2018)
Kolbe and Gilchrist (2009) Reduced outdoor activities
54
Residents of Albury, New South Wales, Australia during
2003 bush fires
PM2.5 2 (daily) >128
for 9 days
Maximum PM2.5
2 (daily) = 597
A-70
DRAFT: Do Not Cite or Quote
-------
Table A.6-2 (Continued): Likelihood of taking actions to reduce wildfire smoke exposure reported in recent
studies.
Study
Exposure Reduction Action
Percent
Population
Taking the
Action
Population Characteristics
Outdoor PM
Concentration.
Hg/m3
Behavioral Changes—Stayed Inside/Closed Doors and Windows
RaDDold etal. (2019)
Stayed indoors
68
SmokeSense App users with no reported health history
and no symptoms
NR 1
70
SmokeSense App users reported health history
90
SmokeSense App users experiencing four or more
symptoms
Richardson et al. (2012)
Stayed inside
73
Residents of five cities within the vicinity of Station Fire
in southern California
Maximum PM2.5
(daily) = 82.9
(hourly) = 223
Suqerman et al. (2012)
Stayed inside
59
Residents of San Diego County during the 2007 San
PM2.5 2 (daily) >128
for 10 days
Maximum PM2.5
(daily) = 803.1
Mean PM2.5
(daily) = 89
(Hutchinson et al..
2018)
Kept windows closed
76
Kolbe and Gilchrist (2009)
Closed windows and doors
44
Residents of Albury, New South Wales, Australia during
2003 bush fires
PM2.5 2 (daily) >128
for 9 days
Maximum PM2.5
2 (daily) = 597
Mott et al. (2002)
Stayed inside
79
Residents of Hoopa, CA during 1999 wildfire that were
aware of public service announcements on smoke
impacts
PM2.5 2 (daily) >128
for 15 days
PM2.5 2 (daily) >425
for 2 days
A-71
DRAFT: Do Not Cite or Quote
-------
Table A.6-2 (Continued): Likelihood of taking actions to reduce wildfire smoke exposure reported in recent
studies.
Study
Exposure Reduction Action
Percent
Population
Taking the
Action
Population Characteristics
Outdoor PM
Concentration.
Hg/m3
Behavioral Changes—Evacuated
RaDDold etal. (2019)
Left area
30
SmokeSense App users with no reported health history
and no symptoms
NR 1
40
SmokeSense App users reported health history
65
SmokeSense App users experiencing four or more
symptoms
Richardson et al. (2012)
Evacuated
5.6
Residents of five cities within the vicinity of Station Fire
in southern California
Maximum PM2.5
(daily) = 82.9
(hourly) = 223
Kolbe and Gilchrist (2009)
Travelled out of area
14
Residents of Albury, New South Wales, Australia during
2003 bush fires
PM2.5 2 (daily) > 128
for 9 days
12
Residents of Albury, New South Wales Australia during
2003 bush fires who saw, heard, or read smoke
advisory
" Maximum PM2.5
2 (daily) = 597
Mott et al. (2002)
Evacuated area during smoke
48
Residents of Hoopa, CA during 1999 wildfire
PM2.5 2 (daily) >128
¦ for 15 days
PM2.5 2 (daily) >425
for 2 days
35
Residents of Hoopa, CA during 1999 wildfire that were
aware of public service announcements on smoke
impacts
44
Residents of Hoopa, CA during 1999 wildfire without a
pre-existing condition
58
Residents of Hoopa, CA during 1999 wildfire with a
pre-existing condition
A-72
DRAFT: Do Not Cite or Quote
-------
Table A.6-2 (Continued): Likelihood of taking actions to reduce wildfire smoke exposure reported in recent
studies.
Study Exposure Reduction Action
Percent
Population
Taking the
Action
Population Characteristics
Outdoor PM
Concentration.
Hg/m3
Exposure Reduction—Ran HVAC system
Richardson et al. (2012) Ran air conditioner more than
usual
60
Residents of five cities within the vicinity of Station Fire
in southern California
Maximum PM2.5
(daily) = 82.9
(hourly) = 223
Suaerman et al. (2012) Used home air conditioner
16
Residents of San Diego County during the 2007 San
Diego Fires
PM2.5 2 (daily) >128
for 10 days
Maximum PM2.5
(daily) = 803.1
Mean PM2.5
(daily) = 89
(Hutchinson et al..
2018)
Exposure Reduction—Used Air Cleaner
RaDDold et al. (2019) Ran an air cleaner
30
SmokeSense App users with no reported health history
and no symptoms
NR 1
52
SmokeSense App users reported health history
86
SmokeSense App users experiencing four or more
symptoms
Richardson et al. (2012) Used an air cleaner
21
Residents of five cities within the vicinity of Station Fire
in southern California
Maximum PM2.5
(daily) = 82.9
(hourly) = 223
A-73
DRAFT: Do Not Cite or Quote
-------
Table A.6-2 (Continued): Likelihood of taking actions to reduce wildfire smoke exposure reported in recent
studies.
Study
Exposure Reduction Action
Percent
Population
Taking the
Action
Population Characteristics
Outdoor PM
Concentration.
Hg/m3
Suqerman et al. (2012)
Used HEPA cleaner
10 Residents of San Diego County during the 2007 San
Diego Fires
PM2.5 2 (daily) >128
for 10 days
Maximum PM2.5
(daily) = 803.1
Mean PM2.5
(daily) = 89
(Hutchinson et al..
2018)
Mott et al. (2002)
Used HEPA cleaner
34%
26%
Residents of Hoopa, CA during 1999 wildfire
PM2.5 2 (daily) >128
for 15 days
Residents of Hoopa, CA during 1999 wildfire without a PM2.5 2 (daily) >425
pre-existing condition for 2 days
52% Residents of Hoopa, CA during 1999 with a pre-existing
condition
Exposure Reduction - Used Respirator/Mask
Rappold et al. (2019)
Wore a respirator
14 SmokeSense App users with no reported health history NR 1
and no symptoms
24
SmokeSense App users reported health history
80 SmokeSense App users experiencing four or more
symptoms
Richardson et al. (2012) Wore a mask
Residents of five cities within the vicinity of Station Fire
in southern California
Maximum PM2.5
(daily) = 82.9
(hourly) = 223
Mott et al. (2002)
Wore an N95 mask
10
Residents of Hoopa, CA during 1999 wildfire
PM2.5 2 (daily) >128
for 15 days
PM2.5 2 (daily) >425
for 2 days
A-74
DRAFT: Do Not Cite or Quote
-------
Table A.6-2 (Continued): Likelihood of taking actions to reduce wildfire smoke exposure reported in recent
studies.
Study
Exposure Reduction Action
Percent
Population
Taking the
Action
Population Characteristics
Outdoor PM
Concentration.
Hg/m3
Symptom Mitigation—Took Medicine
Kolbe and Gilchrist (2009)
Increased regular medication
1.6
Residents of Albury, New South Wales, Australia during
2003 bush fires
PM2.5 2 (daily) >128
for 9 days
Maximum PM2.5
2 (daily) = 597
2.3
Residents of Albury, New South Wales, Australia during
2003 bush fires who saw, heard, or read smoke
advisory
Richardson et al. (2012)
Took medicine
13
Residents of five cities within the vicinity of Station Fire
in southern California
Maximum PM2.5
(daily) = 82.9
(hourly) = 223
Messaging Effectiveness
Mott et al. (2002)
Took exposure reduction action
due to PSA
66
Residents of Hoopa, CA during 1999 wildfire
PM2.5 2 (daily) >128
for 15 days
PM2.5 2 (daily) >425
for 2 days
Kolbe and Gilchrist (2009)
Changed behavior due to
messaging
43
Residents of Albury, New South Wales, Australia during
2003 bush fires
PM2.5 2 (daily) >128
for 9 days
Maximum PM2.5
2 (daily) = 597
A-75
DRAFT: Do Not Cite or Quote
-------
Table A.6-2 (Continued): Likelihood of taking actions to reduce wildfire smoke exposure reported in recent
studies.
Percent
Population Outdoor PM
Taking the Concentration.
Study Exposure Reduction Action Action Population Characteristics M9^m3
Suqerman et al. (2012) Took at least one action from 98 Residents of San Diego County during the 2007 San PM2.5 2 (daily) > 128
messaging Diego Fires for 10 days
Maximum PM2.5
Took all actions from messaging 27 (daily) = 803.1
Mean PM2.5
(daily) = 89
(Hutchinson et al..
2018)
HEPA = high-efficiency particulate air; HVAC = heating, ventilation, and air conditioning; NR = PM2 5 concentrations not reported; PM = particulate matter; PM25 = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PSA = public service announcement.
Note: PM2.5 calculated assuming 85% of PM10 concentration (Lutes. 2014).
A-76
DRAFT: Do Not Cite or Quote
-------
Table A.6-3 Percent reduction in PM2.5 concentrations associated with actions/interventions reported in recent
studies.
Study
Intervention
Percent PM2.5
Reduction
Description of Comparison
Outdoor PM2.5
Concentration.
Hg/m3
Residential Measurement Studies
U.S. EPA (2018) Table 4
\From Park et al. (2017)1
Portable air cleaner with HEPA
filter
43a
Eight homes in California with HEPA filters with activated
carbon; eight homes without; 12-week intervention
U.S. EPA (2018) Table 5
\From Allen et al. (2011)1
Portable air cleaner with HEPA
filter
60
25 homes in British Columbia with HEPA filters during half
of study period and without HEPA filters during rest;
1-week intervention
11.2 (mean)
U.S. EPA (2018) Table 5
\From Weichenthal et al.
(2013)1
Portable air cleaner with
electrostatic precipitator
61
20 homes in Manitoba, Canada with Filtrete electrostatic
filters during half of study period and without filters during
rest; 1-week intervention
42.5
U.S. EPA (2018) Table 5
\From Kaibafzadeh et al.
(2015)1
Portable air cleaner with HEPA
filter
40
20 woodsmoke impacted homes in Vancouver with HEPA
filters during half of study period and without HEPA filters
during rest; 1-week intervention
5.0 HEPA off
3.9 HEPA on
Barn et al. (2008)
Table 2
Portable air cleaner with HEPA
filter
57.7
26 homes in British Columbia during forest fire (summer)
or wood smoke (winter); 1 day each with and without filter
3-91 (S)
<4-189(W)
Henderson et al. (2005)
Figure 7
ESP air cleaners
63-88
Eight homes in Colorado during wildfire or prescribed fire;
paired homes with and without air cleaners
6-38 (outside
during fire)
Sinqer et al. (2017)
HVAC with MERV13 at return (E)
88-93
Single test house in California; Reference system = HVAC
- MERV4 at return had 65-75% reduction
6-16(S)
8-31 (F/W)
Table 2
HVAC continuous with MERV16 at
supply (C)
96-97
A-77
DRAFT: Do Not Cite or Quote
-------
Table A.6-3 (Continued): Percent reduction in PM2.5 concentrations associated with actions/interventions
reported in recent studies.
Study
Intervention
Percent PM2.5
Reduction
Description of Comparison
Outdoor PM2.5
Concentration.
Hg/m3
Portable air cleaner with HEPA
90-94
Alavv and Sieqel (2020)
Figure 3
HVAC with MERV8, MERV11,
MERV14
16 MERV8
36 MERV8E
45 MERV11E
41 MERV14E
21 residences in Toronto; in situ effectiveness compared
to system off or no filter
Reisen et al. (2019)
Table 2
Window/door open
12b
Home: -98 yr old, 8 windows, 4 doors; air conditioner
(H10)
335.8 (h max.)
Windows open
56.7b
Home: 8 yr old, 16 windows, 4 doors; air conditioner (H11)
386.5 (h max.)
Windows/door open
38.5b
Home: 28 yr old, 4 windows, 2 doors; air conditioner (H12)
56.1 (h max.)
Closed
48.5b
Home: -30 yr old, 8 windows, 3 doors; air conditioner
(H16)
56.0 (h max.)
Windows open 20-60% of time
during wood smoke event
67.5-75.7b
Home: -23 yr old, 14 windows, 4 doors; air conditioner
(H21)
A-78
DRAFT: Do Not Cite or Quote
-------
Table A.6-3 (Continued): Percent reduction in PM2.5 concentrations associated with actions/interventions
reported in recent studies.
Study
Intervention
Percent PM2.5
Reduction
Description of Comparison
Outdoor PM2.5
Concentration.
Hg/m3
Residential Modeling Studies
Fiskand Chan (2017b)
Table 5
HVAC fan (continuous), low
efficiency filter (i1)
24
Comparator: Home with intermittent operatinq HVAC
system with typical low-efficiency particle filter (home B1
- mean = 29.2 |jg/m3)
56.9
HVAC fan (continuous), high
efficiency filter (i2)
47
HVAC fan (intermittent), high
efficiency filter (i3)
11
HVAC fan (continuous), low
efficiency filter, continuous portable
air cleaner (i4)
51
HVAC fan (continuous), high
efficiency filter, continuous portable
air cleaner (i5)
62
No forced air system, continuous
portable air cleaner (i6)
45
Comparator: Home with no HVAC. mav have moderate
window AC (home B2 mean = 31.9 |jg/m3)
Office Building Measurement Studies
Stauffer et al. (2020)
Table 4
Portable air cleaner
73 (day)
92 (night)
Offices with and without portable air cleaners during day
and night during wildfire season
17.5
Pantelic et al. (2019)
Figure 5 and text page 10
HVAC system with filters
60c
Office building with HVAC system with filters (MERV8,
Gas Phase filter, and MERV13) compared with an office
building with natural ventilation system
70 (4th St)
53 (Wurster)
A-79
DRAFT: Do Not Cite or Quote
-------
Table A.6-3 (Continued): Percent reduction in PM2.5 concentrations associated with actions/interventions
reported in recent studies.
Study Intervention
Percent PM2.5
Reduction
Description of Comparison
Outdoor PM2.5
Concentration.
Hg/m3
Modeling Studies Residential and Other Buildings
Fisk and Chan (2017a) Table Home HVAC MERV6 runninq 30%
S8 and S9. of time (i1 a, i1 b)
2-4
Comparison:
Home: HVAC MERV 6 operating 15-20% of the time, no
- HEPA portable air cleaner
Other buildings: MERV8
11.4 (LA)
10.0 (NJ)
Home HVAC MERV6 running
30-40% of time, HEPA portable air
cleaner (i5a, i5b)
27"31
10.4 (TX)
Home: HEPA portable air cleaner
(i4)
26"30
Homes: HVAC MERV6 running
15-20% of time, HEPA portable air
cleaner
Other buildings: MERV13 (i8)
7-9
Comparison:
Home: HVAC MERV 6 operating 15-20% of the time, no
HEPA portable air cleaner
Other buildings: MERV8
AC = air conditioning; ESP = electrostatic precipitator; HEPA = high-efficiency particulate air; HVAC = heating, ventilation, and air conditioning; PM2 5 = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 2.5 |jm.
aBased on average PM2 5 concentration difference between groups.
bBased on maximum hourly PM2.5.
Calculated as percent difference in median I/O ratios.
A-80
DRAFT: Do Not Cite or Quote
-------
A.7. SUPPLEMENTAL INFORMATION FOR CHAPTER 7
1 No supplemental information.
A.8. SUPPLEMENTAL INFORMATION FOR CHAPTER 8
Table A.8-1 Corresponding table of estimated wildfire-PM2.5 illnesses (95%
confidence interval) from sensitivity analyses presented in Figure 8-1
and Figure 8-2.
Hospital Admissions
Case Study
Scenario
Asthma
ED Visits
Respiratory
Cardiovascular
Actual Fire
0.4
(0.31 to 0.5)
0.17
(0.01 to 0.31)
0.07
(-0.01 to 0.14)
-------
A.9. SUPPLEMENTAL INFORMATION FOR CHAPTER 9
1 No supplemental information.
A.10. REFERENCES
Abdelnour. A: McKane. RB: Stieglitz. M: Pan. F: Cheng. Y. (2013). Effects of harvest on carbon and nitrogen
dynamics in a Pacific Northwest forest catchment. Water Resour Res 49: 1292-1313.
http://dx.doi.org/10.1029/2012WR012994
Abdelnour. A: Stieglitz. M: Pan. F: McKane. R. (2011). Catchment hydrological responses to forest harvest
amount and spatial pattern. Water Resour Res 47: W09521. http://dx.doi.org/10.1029/2010WR010165
Alaw. M: Siegel. JA. (2020). In-situ effectiveness of residential HVAC filters. Indoor Air 30: 156-166.
http://dx.doi.org/10. Ill 1/ina. 12617
Allen. RW: Carlsten. C: Karlen. B: Leckie. S: van Eeden. S: Vedal. S: Wong. I: Brauer. M. (2011). An air filter
intervention study of endothelial function among healthy adults in a woodsmoke-impacted community. Am J
Respir Crit Care Med 183: 1222-1230. http://dx.doi.org/10.1164/rccm.201010-1572QC
Alman. BL: Pfister. G: Hao. H: Stowell. J: Hu. X: Liu. Y: Strickland. MJ. (2016). The association of wildfire
smoke with respiratory and cardiovascular emergency department visits in Colorado in 2012: A case
crossover study. Environ Health 15: 64. http://dx.doi.org/10.1186/sl2940-016-0146-8
Barkiohn. KK: Gantt. B: Clements. AL. (2020). Development and Application of a United States wide correction
for PM2.5 data collected with the PurpleAir sensor. Atmos Meas Tech. http://dx.doi.org/10.5194/amt-202Q-
413
Barn. P: Larson. T: Noullett. M: Kennedy. S: Copes. R: Brauer. M. (2008). Infiltration of forest fire and
residential wood smoke: An evaluation of air cleaner effectiveness. J Expo Sci Environ Epidemiol 18: 503 -
511. http://dx.doi.org/10.1038/si.ies.750064Q
Barnhart. B: Pettus. P: Halama. J: McKane. R: Maver. P: Diang. K: Brookes. A: Moskal. LM. (2021). Modeling
the hydrologic effects of watershed-scale green roof implementation in the Pacific Northwest, United States.
J Environ Manage 277: 111418. http://dx.doi.org/10.1016/i.ienvman.2020.111418
Barnhart. BL: McKane. R: Brookes. A: Schumaker. N: Papenfus. M: Pettus. P: Halama. J: Powers. B: Diang. K:
Groskinskv. B: Grier. G: Hawkins. A: Tapp. J: Watson. D: Gross. T: Goodia D: Mohler. R. (2015).
Integrated modeling to assess the ecological and air quality trade-offs of agricultural burning in the Flint
Hills of eastern Kansas. Abstract presented at American Geophysical Union Fall Meeting, December 14-18,
2015, San Francisco, CA.
Bell. DM: Gregory. MJ: Kane. V: Kane. J: Kennedy. RE: Roberts. HM: Yang. Z. (2018). Multiscale divergence
between Landsat- and lidar-based biomass mapping is related to regional variation in canopy cover and
composition. Carbon Balance and Management 13: 15. http://dx.doi.org/10.1186/sl3021-018-0104-6
Brown. JK: See. TE. (1981). Downed dead woody fuel and biomass in the northern Rocky Mountains. (INT-
GTR-117). Ogden, UT: U.S. Dept. of Agriculture, Forest Service, Intermountain Forest and Range
Experiment Station.
Brown. JK: See. TE. (1986). Surface fuel loadings and predicted fire behavior for vegetation types in the
northern Rocky Mountains. (Research Note INT-358). Ogden, UT: U.S. Dept. of Agriculture, Forest Service,
Intermountain Forest and Range Experiment Station.
Chow. JC: Watson. JG: Pritchett. LC: Pierson. WR: Frazier. CA: Purcell. RG. (1993). The DRI thermal/optical
reflectance carbon analysis system: Description, evaluation and applications in U.S. air quality studies. In JP
Lodge, Jr. (Ed.), Fourth International Conference on Carbonaceous Particles in the Atmosphere (pp. 1185-
1201). Oxford, UK: Pergamon Press. http://dx.doi.org/10.1016/0960-1686(93)90245-T
A-82
DRAFT: Do Not Cite or Quote
-------
Daly. C: Halbleib. M: Smith. J1: Gibson. WP: Doggett. MK: Taylor. GH: Curtis. J: Pasteris. PP. (2008).
Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous
United States. Int J Climatol 28: 2031-2064. http://dx.doi.org/10.1002/ioc. 1688
Davis. RJ: Ohmann. JL: Kennedy. RE: Cohen. WB: Gregory. MJ: Yang. Z: Roberts. HM: Gray. AN: Spies. TA.
(2015). Northwest Forest Plan-The first 20 years (1994-2013): Status and trends of late-successional and old-
growth forests. (PNW-GTR-911). Portland, OR: Department of Agriculture, Forest Service, Pacific
Northwest Research Station, http://dx.doi.org/10.2737/PNW-GTR-911
DeFlorio-Barker. S: Crooks. J: Reyes. J: Rappold. AG. (2019). Cardiopulmonary effects of fine particulate
matter exposure among older adults, during wildfire and non-wildfire periods, in the United States 2008-
2010. Environ Health Perspect 127: 37006. http://dx.doi.org/10.1289/EHP3860
Delfino. RJ: Brummel. S: Wu. J: Stern. H: Ostro. B: Lipsett. M: Winer. A: Street. DH: Zhang. L: Tioa. T: Gillen.
PL. (2009). The relationship of respiratory and cardiovascular hospital admissions to the southern California
wildfires of 2003. Occup Environ Med 66: 189-197. http://dx.doi.org/10.1136/oem.2008.041376
Delp. WW: Singer. BC. (2020). Wildfire smoke adjustment factors for low-cost and professional
PM(2.5)monitors with optical sensors. Sensors 20: 3683. http://dx.doi.org/10.3390/s20133683
Doubledav. A: Schulte. J: Sheppard. L: Kadlec. M: Dhammapala. R: Fox. J: Isaksen. TB. (2020). Mortality
associated with wildfire smoke exposure in Washington state, 2006-2017: A case-crossover study. Environ
Health 19: 4. http://dx.doi.org/10.1186/sl2940-020-0559-2
Fisk. WJ: Chan. WR. (2017a). Effectiveness and cost of reducing particle-related mortality with particle
filtration. Indoor Air 27: 909-920. http://dx.doi.org/10. Ill 1/ina. 12371
Fisk. WJ: Chan. WR. (2017b). Health benefits and costs of filtration interventions that reduce indoor exposure to
PM2.5 during wildfires. Indoor Air 27: 191-204. http://dx.doi.org/10. Ill 1/ina. 12285
Gan. RW: Ford. B: Lassman. W: Pfister. G: Vaidvanathan. A: Fischer. E: Volckens. J: Pierce. JR: Magzamen. S.
(2017). Comparison of wildfire smoke estimation methods and associations with cardiopulmonary-related
hospital admissions. Geohealth 1: 122-136. http://dx.doi.org/10.1002/2017GH000Q73
Gan. RW: Liu. J: Ford. B: O'Dell. K: Vaidvanathan. A: Wilson. A: Volckens. J: Pfister. G: Fischer. EV: Pierce.
JR: Magzamen. S. (2020). The association between wildfire smoke exposure and asthma-specific medical
care utilization in Oregon during the 2013 wildfire season. J Expo Sci Environ Epidemiol 30: 618-628.
http://dx.doi.org/10.1038/s41370-020-021Q-x
Henderson. DE: Milford. JB: Miller. SL. (2005). Prescribed burns and wildfires in Colorado: Impacts of
mitigation measures on indoor air particulate matter. J Air Waste Manag Assoc 55: 1516-1526.
http://dx.doi.org/10.1080/10473289.20Q5.10464746
Hoghooghi. N: Golden. HE: Bledsoe. BP: Barnhart. BL: Brookes. AF: Diang. KS: Halama. JJ: McKane. RB:
Nietch. CT: Pettus. PP. (2018). Cumulative effects of low impact development on watershed hydrology in a
mixed land-cover system. Water 10: 991. http://dx.doi.org/10.3390/wlQ080991
Holder. AL: Mebust. AK: Maghran. LA: Mcgown. MR: Stewart. KE: Vallano. DM: Elleman. RA: Baker. KR.
(2020). Field evaluation of low-cost particulate matter sensors for measuring wildfire smoke. Sensors 20:
4796. http://dx.doi.org/10.3390/s2Q174796
Huntzicker. JJ: Johnson. RL: Shah. JJ: Carv. RA. (1982). Analysis of organic and elemental carbon in ambient
aerosols by a thermal-optical method. In GT Wolff; RL Klimisch (Eds.), Particulate carbon (pp. 79-88).
Boston, MA: Springer. http://dx.doi.org/10.10Q7/978-l-4684-4154-3 6
Hutchinson. JA: Vargo. J: Milet. M: French. NHF: Billmire. M: Johnson. J: Hoshiko. S. (2018). The San Diego
2007 wildfires and Medi-Cal emergency department presentations, inpatient hospitalizations, and outpatient
visits: An observational study of smoke exposure periods and a bidirectional case-crossover analysis. PLoS
Med 15: el002601. http://dx.doi.org/10.1371/iournal.pmed.1002601
Jaklevic. JM: Gatti. RC: Goulding. FS: Loo. BW: Thompson. AC. (1981). Aerosol analysis for the regional air
pollution study: Final report [EPA Report]. (EPA/600-S4-81-006). Washington, DC: US Environmental
Protection Agency. https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=2000TT0F.txt
A-83
DRAFT: Do Not Cite or Quote
-------
Jenkins. JC: Choinackv. DC: Heath. LS: Birdsev. RA. (2003). National-scale biomass estimators for United
States tree species [Review]. Forest Sci 49: 12-35. http://dx.doi.org/10.1093/forestscience/49.1.12
Jones. CG: Rappold. AG: Vargo. J: Cascio. WE: Kharrazi. M: McNallv. B: Hoshiko. S. (2020). Out-of-hospital
cardiac arrests and wildfire-related particulate matter during 2015-2017 California wildfires. J Am Heart
Assoc 9: e014125. http://dx.doi.org/10.1161/JAHA.119.014125
Kaibafzadeh. M: Brauer. M: Karlen. B: Carlsten. C: van Eeden. S: Allen. RW. (2015). The impacts of traffic-
related and woodsmoke particulate matter on measures of cardiovascular health: A HEPA filter intervention
study. Occup Environ Med 72: 394-400. http://dx.doi.org/10.1136/oemed-2014-102696
Kennedy. RE: Ohmann. J: Gregory. M: Roberts. H: Yang. Z: Bell. DM: Kane. V: Hughes. MJ: Cohen. WB:
Powell. S: Neeti. N: Larrue. T: Hooper. S: Kane. J: Miller. PL: Perkins. J: Braaten. J: Seidl. R. (2018). An
empirical, integrated forest biomass monitoring system. Environ Res Lett 13: 025004.
http://dx.doi.org/10.1088/1748-9326/aa9d9e
Kennedy. RE: Yang. Z: Cohea WB: Pfaff. E: Braaten. J: Nelson. P. (2012). Spatial and temporal patterns of
forest disturbance and regrowth within the area of the Northwest Forest Plan. Rem Sens Environ 122: 117-
133. http://dx.doi.Org/10.1016/i.rse.2011.09.024
Kolbe. A: Gilchrist. KL. (2009). An extreme bushfire smoke pollution event: Health impacts and public health
challenges. NSW Public Health Bulletin 20: 19-23. http://dx.doi.org/10.1071/nb08061
Landis. MS: Long. RW: Krug. J: Colon. M: Vanderpool. R: Habel. A: Urbanski. S. (2021). The U.S. EPA
Wildland Fire Sensor Challenge: Performance and evaluation of solver submitted multi-pollutant sensor
systems. Atmos Environ 247: 118165. http://dx.doi.Org/10.1016/i.atmosenv.2020.118165
Leibel. S: Nguyen. M: Brick. W: Parker. J: Ilango. S: Aguilera. R: Gershunov. A: Benmarhnia. T. (2020).
Increase in pediatric respiratory visits associated with Santa Ana wind-driven wildfire smoke and PM2.5
levels in San Diego County. Ann Am Thorac Soc 17: 313-320.
http://dx.doi.org/10.1513/AnnalsATS.201902-150QC
Liu. JC: Wilson. A: Micklev. LJ: Dominici. F: Ebisu. K: Wang. Y: Sulprizio. MP: Peng. RD: Yue. X: Son. JY:
Anderson. GB: Bell. ML. (2017a). Wildfire-specific fine particulate matter and risk of hospital admissions in
urban and rural counties. Epidemiology 28: 77-85. http://dx.doi.org/10.1097/EDE.000000000000Q556
Liu. JC: Wilson. A: Micklev. LJ: Ebisu. K: Sulprizio. MP: Wang. Y: Peng. RD: Yue. X: Dominici. F: Bell. ML.
(2017b). Who among the elderly is most vulnerable to exposure and health risks of PM2.5 from wildfire
smoke? Am J Epidemiol 186: 730-735. http://dx.doi.org/10.1093/aie/kwxl41
Long. RW: Whitehill. A: Habel. A: Urbanski. S: Hallidav. H: Coloa M: Kaushik. S: Landis. MS. (In Press)
Comparison of ozone measurement methods in biomass smoke: An evaluation under field and laboratory
conditions. Atmospheric Measurement Techniques Discussions, http://dx.doi.org/10.5194/amt-2020-383
Lutes. DC. (2014). First Order Fire Effects Model (FOFEM) mapping tool, version 6.1: User guide. Missoula,
MT: U.S. Forest Service, Rocky Mountain Research Station, Fire Modeling Institute.
https://www.frames.gov/catalog/17743
Lutes. DC: Keane. RE: Caratti. JF. (2009). A surface fuel classification for estimating fire effects. International
Journal of Wildland Fire 18: 802-814. http://dx.doi.org/10.1071/WF08062
McKane. B: Barnhart. B: Halama. J: Pettus. P: Brookes. A: Ebersole. J: Diang. K: Blair. G: Hall. J: Kane. J:
Swedeea P: Benson. L. (2016). Nisqually Community Forest VELMA modeling [Abstract]. Presentation
presented at 2016 South Sound Science Symposium, September 20, 2016, Olympia, WA.
McKane. B: Halama. J: Pettus. P: Barnhart. B: Brookes. A: Diang. K: Blair. G: Hall. J: Kane. J: Swedeen. P:
Benson. L. (2018). How Visualizing Ecosystem Land Management Assessments (VELMA) modeling
quantifies co-benefits and tradeoffs in Community Forest management. Presentation presented at Northwest
Community Forest Forum, May 10 - 11, 2018, Astoria, OR.
A-84
DRAFT: Do Not Cite or Quote
-------
McKane. R: Brookes. A: Diang. K: Stieglitz. M: Abdelnour. A: Pan. F: Halama. J: Pettus. P: Phillips. D. (2014).
Visualizing Ecosystem Land Management Assessments (VELMA) v. 2.0: User manual and technical
documentation. (Document control number L-PESD-30840-QP-1-2). Corvallis, OR: U.S. Environmental
Protection Agency, National Health and Environmental Effects Research Laboratory.
https://www.epa.gov/sites/production/files/2016-01/documents/velma 2.0 user manual.pdf
McKane. RB. (2020). Quality Assurance Project Plan (QAPP), VELMA Framework, version 1.2 [EPA Report].
(ORD QAPP ID No: L-PESD-30840-QP-1-2). Corvallis, OR: U.S. Environmental Protection Agency.
Mehadi. A: Moosmtiller. H: Campbell. DE: Ham. W: Schweizer. D: Tarnav. L: Hunter. J. (2019). Laboratory
and field evaluation of real-time and near real-time PM2.5 smoke monitors. J Air Waste Manag Assoc 70:
158-179. http://dx.doi.org/10.1080/10962247.2019.1654Q36
Mott. JA: Mever. P: Manning. D: Redd. SC: Smith. EM: Gotwav-Crawford. C: Chase. E. (2002). Wildland
forest fire smoke: Health effects and intervention evaluation, Hoopa, California, 1999. West J Med 176: 157-
162. http://dx.doi.org/10.1136/ewim. 176.3.157
Ottmar. RD: Sandberg. DV: Riccardi. CL: Prichard. SJ. (2007). An overview of the fuel characteristic
classification system—quantifying, classifying, and creating fuelbeds for resource planning. Can J For Res
37: 2383-2393. http://dx.doi.org/10.1139/X07-077
Pan. FF: Stieglitz. M: McKane. RB. (2012). An algorithm for treating flat areas and depressions in digital
elevation models using linear interpolation. Water Resour Res 48: W00L10.
http://dx.doi.org/10.1029/2011WR01Q735
Pantelic. J: Dawe. M: Licina. D. (2019). Use of IoT sensing and occupant surveys for determining the resilience
of buildings to forest fire generated PM2.5. PLoS ONE 14: e0223136.
http://dx.doi.org/10.1371/iournal.pone.0223136
Park. HK: Cheng. KC: Tetteh. AO: Hildemann. LM: Nadeau. KC. (2017). Effectiveness of air purifier on health
outcomes and indoor particles in homes of children with allergic diseases in Fresno, California: A pilot study.
J Asthma 54: 1-6. http://dx.doi.org/10.1080/02770903.2Q16.1218011
R Core Team (R Development Core Team). (2019). R: A language and environment for statistical computing.
Vienna, Austria: R Foundation for Statistical Computing. https://www.R-proiect.org/
Rappold. A: Stone. SL: Cascio. WE: Neas. LM: Kilaru. VJ: Carrawav. MS: Szvkman. JJ: Ising. A: Cleve. WE:
Meredith. JT: Vaughan-Batten. H: Devneka. L: Devlin. RB. (2011). Peatbog wildfire smoke exposure in
rural North Carolina is associated with cardiopulmonary emergency department visits assessed through
syndromic surveillance. Environ Health Perspect 119: 1415-1420. http://dx.doi.org/10.1289/ehp. 1003206
Rappold. AG: Cascio. WE: Kilaru. VJ: Stone. SL: Neas. LM: Devlin. RB: Diaz-Sanchez. D. (2012). Cardio-
respiratory outcomes associated with exposure to wildfire smoke are modified by measures of community
health. Environ Health 11:71. http://dx.doi.org/10.1186/1476-069X-11-71
Rappold. AG: Hano. MC: Prince. S: Wei. L: Huang. SM: Baghdikian. C: Stearns. B: Gao. X: Hoshiko. S:
Cascio. WE: Diaz-Sanchez. D: Hubbell. B. (2019). Smoke sense initiative leverages citizen science to
address the growing wildfire-related public health problem. Geohealth 3: 443 -457.
http://dx.doi.org/10.1029/2019GH00Q199
Reid. CE: Brauer. M: Johnston. FH: Jerrett. M: Balmes. JR: Elliott. CT. (2016). Critical review of health impacts
of wildfire smoke exposure [Review]. Environ Health Perspect 124: 1334-1343.
http://dx.doi.org/10.1289/ehp. 1409277
Reid. CE: Considine. EM: Watson. GL: Telesca. D: Pfister. GG: Jerrett. M. (2019). Associations between
respiratory health and ozone and fine particulate matter during a wildfire event. Environ Int 129: 291-298.
http://dx.doi.Org/10.1016/i.envint.2019.04.033
Reisen. F: Powell. JC: Dennekamp. M: Johnston. FH: Wheeler. AJ. (2019). Is remaining indoors an effective
way of reducing exposure to fine particulate matter during biomass burning events? J Air Waste Manag
Assoc 69: 611-622. http://dx.doi.org/10.1080/10962247.2Q19.1567623
A-85
DRAFT: Do Not Cite or Quote
-------
Remillard. SM. (1999) Soil carbon and nitrogen in old-growth forests in western Oregon and Washington.
(Master's Thesis). Oregon State University, Corvallis, OR. Retrieved from
https://ir.librarv.oregonstate.edu/concern/graduate thesis or dissertations/vh53wx881
Richardson. LA: Champ. PA: Loomis. JB. (2012). The hidden cost of wildfires: Economic valuation of health
effects of wildfire smoke exposure in Southern California. J Forest Econ 18: 14-35.
http://dx.doi.Org/10.1016/i.ife.2011.05.002
Singer. BC: Delp. WW: Black. PR: Walker. IS. (2017). Measured performance of filtration and ventilation
systems for fine and ultrafine particles and ozone in an unoccupied modern California house. Indoor Air 27:
780-790. http://dx.doi.org/10.1111/ina. 12359
Smithwick. EAH: Harmon. ME: Remillard. SM: Acker. SA: Franklin. JF. (2002). Potential upper bounds of
carbon stores in forests of the Pacific Northwest. Ecol Appl 12: 1303-1317. http://dx.doi.org/10.189Q/1051-
0761(2002)01211303 :PUBOCS12.Q.CO:2
Stauffer. DA: Autenrieth. DA: Hart. JF: Capoccia. S. (2020). Control of wildfire-sourced PM2.5 in an office
setting using a commercially available portable air cleaner. J Occup Environ Hyg 17: 109-120.
http://dx.doi.org/10.1080/15459624.202Q.1722314
Stowell. JD: Geng. G: Saikawa. E: Chang. HH: Fu. J: Yang. CE: Zhu. O: Liu. Y: Strickland. MJ. (2019).
Associations of wildfire smoke PM2.5 exposure with cardiorespiratory events in Colorado 2011-2014.
Environ Int 133: 105151. http://dx.doi.Org/10.1016/i.envint.2019.105151
Sugerman. DE: Keir. JM: Dee. PL: Lipman. H: Waterman. SH: Ginsberg. M: Fishbein. DB. (2012). Emergency
health risk communication during the 2007 San Diego wildfires: Comprehension, compliance, and recall. J
Health Commun 17: 698-712. http://dx.doi.org/10.1080/1081073Q.2011.635777
Tinling. MA: West. JJ: Cascio. WE: Kilaru. V: Rappold. AG. (2016). Repeating cardiopulmonary health effects
in rural North Carolina population during a second large peat wildfire. Environ Health 15: 12.
http://dx.doi.org/10.1186/sl2940-016-0Q93-4
U.S. EPA (U.S. Environmental Protection Agency). (1999). Particulate matter (PM2.5) speciation guidance:
Final draft (Edition 1) [EPA Report]. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards.
http://www.epa.gov/ttn/amtic/files/ambient/pm25/spec/specpln2.pdf
U.S. EPA. Measurement principle and calibration procedure for the measurement of nitrogen dioxide in the
atmosphere (gas phase chemiluminescence), 40 CFR pt. 50, app. F (201 la).
https://www.govinfo.gov/app/details/CFR-2011-title40-vol2/CFR-2011-title40-vol2-part50-appF
U.S. EPA. Measurement principle and calibration procedure for the measurement of ozone in the atmosphere. 40
CFR pt. 50. app. D (201 lb). https://www.govinfo.gov/app/details/CFR-2011 -title40-vol2/CFR-2011 -title40-
vol2-part50-appD
U.S. EPA. Reference measurement principle and calibration procedure for the measurement of sulfur dioxide in
the atmosphere (ultraviolet fluorescence method), 40 CFRpt. 50, app. A-l (2011c).
https://www.govinfo.gov/app/details/CFR-2011-title40-vol2/CFR-2011-title40-vol2-part50-appA
U.S. EPA. Network design criteria for ambient air quality monitoring. 40 CFR pt. 58. app. D (2015).
https://www.govinfo.gov/app/details/CFR-2015-title40-vol6/CFR-2015-title40-vol6-part58-appD
U.S. EPA (U.S. Environmental Protection Agency). (2018). Residential air cleaners: A technical summary [EPA
Report] (3rd ed.). (EPA 402-F-09-002). Washington, DC.
https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100UX5R.txt
U.S. EPA. Quality assurance requirements for monitors used in evaluations of National Ambient Air Quality
Standards. 40 CFR pt. 58. app. A (2019). https://www.govinfo.gov/app/details/CFR-2019-title40-vol6/CFR-
2019-title40-vol6-part58-appA
U.S. EPA. Measurement principle and calibration procedure for the measurement of carbon monoxide in the
atmosphere (non-dispersive infrared photometry), 40 CFR pt. 50, app. C (2020a).
https://www.govinfo.gov/app/details/CFR-2020-title40-vol2/CFR-2020-title40-vol2-part5Q-appC
A-86
DRAFT: Do Not Cite or Quote
-------
U.S. EPA. Reference method for the determination of sulfur dioxide in the atmosphere (pararosaniline method),
40 CFR pt. 50, app. A-2 (2020b). https://www.govinfo.gov/app/details/CFR-2020-title40-vol2/CFR-202Q-
title40-vol2-part50-appA-id28
Watson. JG: Chow. JC: Frazier. CA. (1999). X-ray fluorescence analysis of ambient air samples. In S
Landsberger; M Creatchman (Eds.), Elemental analysis of airborne particles (pp. 67-96). Amsterdam,
Netherlands: Gordon and Breach Science Publishers.
Weichenthal. S: Mallach. G: Kulka. R: Black. A: Wheeler. A: You. H: St-Jean. M: Kwiatkowski. R: Sharp. D.
(2013). A randomized double-blind crossover study of indoor air filtration and acute changes in
cardiorespiratory health in a First Nations community. Indoor Air 23: 175-184.
http://dx.doi.org/10. Ill 1/ina. 12019
Wettstein. ZS: Hoshiko. S: Fahimi. J: Harrison. RJ: Cascio. WE: Rappold. AG. (2018). Cardiovascular and
cerebrovascular emergency department visits associated with wildfire smoke exposure in California in 2015.
J Am Heart Assoc 7: e007492. http://dx.doi.org/10.1161/JAHA. 117.007492
Xi. Y: Kshirsagar. AY: Wade. TJ: Richardson. DB: Brookhart. MA: Wvatt. L: Rappold. AG. (2020). Mortality
in US hemodialysis patients following exposure to wildfire smoke. J Am Soc Nephrol 31: 1824-1835.
http://dx.doi.org/10.1681/ASN.2019101066
Yee. S: Bousauin. J: Bruins. R: Canfield. TJ: DeWitt. TH: de Jesus-Crespo. R: Dyson. B: Fulford. R: Harwell.
M: Hoffman. J: Littles. CJ: Johnston. JM: McKane. RB: Green. L: Russel. M: Sharpe. L: Seeteram. N:
Tashie. A: Williams. K. (2017). Practical strategies for integrating final ecosystem goods and services into
community decision-making. (EPA/600/R-17/266). Washington, DC: U.S. Environmental Protection
Agency. https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P100SGRC.txt
Zu. K: Tao. G: Long. C: Goodman. J: Valberg. P. (2016). Long-range fine particulate matter from the 2002
Quebec forest fires and daily mortality in Greater Boston and New York City. Air Qual Atmos Health 9: 213-
221. http://dx.doi.org/10.1007/sll869-015-Q332-9
A-87
DRAFT: Do Not Cite or Quote
------- |