&EPA
United a*K
Envirainwnlal Protection
Agency
Welfare Risk and Exposure Assessment for
Ozone
Second External Review Draft

-------
                                                   EPA-452/P-14-003 a
                                                        February 2014
Welfare Risk and Exposure Assessment for Ozone
                Second External Review Draft
                U.S. Environmental Protection Agency
                    Office of Air and Radiation
              Office of Air Quality Planning and Standards
              Health and Environmental Impacts Division
                     Risk and Benefits Group
             Research Triangle Park, North Carolina 27711

-------
                                        DISCLAIMER
       This draft document has been prepared by staff from the Risk and Benefits Group, Health and
Environmental Impacts Division, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. Any findings and conclusions are those of the authors and do not necessarily reflect
the views of the Agency. This draft document is being circulated to facilitate discussion with the Clean
Air Scientific Advisory Committee to inform the EPA's consideration of the ozone National Ambient
Air Quality Standards.

       This information is distributed for the purposes of pre-dissemination peer review under
applicable information quality guidelines.  It has not been formally disseminated by EPA.  It does not
represent and should not be construed to represent any Agency determination or policy.

       Questions related to this preliminary draft document should be addressed to Travis Smith, U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, C539-07, Research
Triangle Park, North Carolina 27711 (email: smith.jtravis@epa.gov).

-------
This page left intentionally blank.

-------
                           TABLE OF CONTENTS
1     INTRODUCTION	1-1
1.1    HISTORY	1-3
1.2    CURRENT RISK AND EXPOSURE ASSESSMENTS: GOALS AND PLANNED
      APPROACH	1-4
1.3    ORGANIZATION OF DOCUMENT	1-5
2     FRAMEWORK	2-1
2.1    O3 CHEMISTRY	2-2
2.2    SOURCES OF O3 AND O3 PRECURSORS	2-3
2.3    ECOLOGICAL EFFECTS	2-5
2.4    ECOSYSTEM SERVICES	2-7
3     SCOPE	3-1
3.1    OVERVIEW OF RISK AND EXPOSURE ASSESSMENTS FROM PREVIOUS
      REVIEW	3-2
      3.1.1  Exposure Characterization	3-3
      3.1.2  Assessment of Risks to Vegetation	3-4
3.2    OVERVIEW OF CURRENT ASSESSMENT PLAN	3-5
      3.2.1  Air Quality Considerations	3-6
      3.2.2  Relative TreeBiomass Loss and Crop Yield Loss	3-8
      3.2.3  Visible Foliar Injury	3-11
3.3    UNCERTAINTY AND VARIABILITY	3-13
4     AIR QUALITY CONSIDERATIONS	4-1
4.1    INTRODUCTION	4-1
4.2    OVERVIEW OF O3 MONITORING AND AIR QUALITY	4-1
4.3    OVERVIEW OF AIR QUALITY INPUTS TO RISK AND EXPOSURE
      ASSESSMENTS	4-4
      4.3.1  Air Quality Metrics	4-5
      4.3.2  Ambient Air Quality Measurements	4-7
      4.3.3  National-scale Air Quality Surfaces for Recent Air Quality	4-8
      4.3.4  Air Quality Adjustments to Meet Existing Primary and Potential Alternative
           Secondary O3 Standards	4-10
4.4    QUALITATIVE ASSESSMENT OF UNCERTAINTY	4-26
5     O3 RISK TO ECOSYSTEM SERVICES	5-1
5.1    INTRODUCTION	5-1
5.2    SUPPORTING SERVICES	5-3
      5.2.1  Net Primary Productivity	5-3

-------
      5.2.2  Community Composition and Habitat Provision	5-4
5.3    REGULATING SERVICES	5-5
      5.3.1  Hydrologic Cycle	5-5
      5.3.2  Pollination	5-6
      5.3.3  Fire Regulation	5-6
5.4    PROVISIONING SERVICES	5-9
5.5    CULTURAL SERVICES	5-13
      5.5.1  Non-Use Services	5-14
5.6    QUALITATIVE ASSESSMENT OF UNCERTAINTY	5-16
5.7    DISCUSSION	5-20
6     BIOMASS LOSS	6-1
6.1    INTRODUCTION	6-1
6.2    RELATIVE BIOMASS LOSS	6-2
      6.2.1  Species-Level Analyses	6-6
6.3    COMMERCIAL TIMBER EFFECTS	6-21
6.4    NON-TIMBER FOREST PRODUCTS	6-31
      6.4.1  Commercial Non-Timber Forest Products	6-34
      6.4.2  Informal Economy or Subsistence Use of Non-Timber Forest Products	6-36
6.5    AGRICULTURE	6-38
      6.5.1  Commercial Agriculture	6-38
6.6    CLIMATE REGULATION	6-46
      6.6.1  National Scale Forest Carbon Sequestration	6-46
      6.6.2  Urban Case Study Carbon Storage	6-50
6.7    URBAN CASE STUDY AIR POLLUTION REMOVAL	6-55
6.8    ECOSYSTEM-LEVEL EFFECTS	6-59
      6.8.1  Potential Biomass Loss in Federally Designated Areas	6-65
6.9    QUALITATIVE ASSESSMENT OF UNCERTAINTY	6-66
6.10   DISCUSSION	6-74
7     VISIBLE FOLIAR INJURY	7-1
7.1    INTRODUCTION	7-1
      7.1.1  Ecosystem Services	7-4
7.2    NATIONAL-SCALE ANALYSIS OF FOLIAR INJURY	7-10
      7.2.1  Forest Health Monitoring Network	7-10
      7.2.2  NOAA Palmer Z Drought Index	7-12
      7.2.3  Results	7-13
7.3    SCREENING-LEVEL ASSESSMENT OF VISIBLE FOLIAR INJURY IN

-------
      NATIONAL PARKS	7-20
      7.3.1   Screening Assessment Methods	7-21
      7.3.2   Screening Assessment Results and Discussion	7-28
      7.3.3   Limitations and Uncertainty Characterization for Screening-Level
             Assessment	7-35
7.4    NATIONAL PARK CASE STUDY AREAS	7-44
      7.4.1   Great Smoky Mountains National Park	7-47
      7.4.2   Rocky Mountain National Park	7-55
      7.4.3   Sequoia and Kings Canyon National Parks	7-62
7.5    QUALITATIVE ASSESSMENT OF UNCERTAINTY	7-67
7.6    DISCUSSION	7-72
8   SYNTHESIS OF RESULTS	8-1
8.1    INTRODUCTION	8-1
8.2    SUMMARY OF ANALYSES AND KEY RESULTS	8-2
      8.2.1 National-Scale Analyses	8-3
      8.2.2 Case Study-Scale Analyses	8-14
8.3    PATTERNS OF RISK	8-16
      8.3.1 Risk Patterns Across or Between Geographic Areas	8-17
      8.3.2 Risk Patterns Across Years	8-21
      8.3.3 Risk Patterns Across Alternative W126 Standard Levels	8-23
8.4    REPRESENTATIVENESS	8-27
      8.4.1 Species Representativeness	8-27
      8.4.2 Geographic Representativeness	8-28
      8.4.3 Temporal Representativeness	8-29
8.5    OVERALL  CONFIDENCE IN WELFARE EXPOSURE AND RISK RESULTS	8-30
      8.5.1 Uncertainties in Air Quality Analyses	8-30
      8.5.2 Uncertainties in Biomass Loss  Analyses	8-31
      8.5.3 Uncertainties in Visible Foliar Injury Analyses	8-32
8.6    CONCLUSIONS	8-33
9   REFERENCES	9-1
                                           in

-------
                                   LIST OF TABLES
Table 4-1     Highest 2006-2008 Average W126 Concentrations in Observed and Existing
             Standard Air Quality Adjustment Scenarios; Highest 2006-2008 8-hour 63 Design
             Values in Observed and Potential Alternative Standard Air Quality Adjustment
             Scenarios	4-14
Table 4-2     Summary of Qualitative Uncertainty Analysis of Key Air Quality Elements in the
             O3 NAAQS Risk Assessment	4-27
Table 5-1     Responses to NSRE Wildlife Value Questions	5-4
Table 5-2     Area of Moderate to High-Fire Threat, Mixed Conifer Forest for Existing and
             Alternative Standard Levels (in km2)	5-8
Table 5-3     Area (km2) 'At Risk' of High Pine Beetle Loss at Various W126
             Index Values	5-12
Table 5-4     Tree Basal Area Considered 'At Risk' of High Pine Beetle Loss ByW126 Index
             Values after Just Meeting the Existing and Alternative Standard Levels (in
             millions of square feet)	5-13
Table 5-5     NSRE Responses to Non-Use Value Questions For Forests	5-15
Table 5-6     Value Components for WTP for Extensive Protection Program for Southern
             Appalachian Spruce-Fir Forests	5-16
Table 5-7     Summary of Qualitative Uncertainty Analysis in Semi-Quantitative Ecosystem
             Services Assessments	5-18
Table 6-1     Relative Biomass Loss Functions for Tree Species	6-4
Table 6-2     Relative Biomass Loss Functions for Crop Species	6-4
Table 6-3     Comparison of Adult to Seedling Biomass Loss	6-7
Table 6-4     Comparison of Seedling Biomass Loss to Adult Circumference	6-8
Table 6-5     Summary of Uncertainty in Seedling to Adult Tree Biomass Loss
             Comparisons	6-9
Table 6-6     Individual Species Relative Biomass Loss Values - Median, 75th Percentile,
             Maximum Percentages	6-16
Table 6-7     Mapping O3 Impacts to FASOMGHG Forest Types	6-23
Table 6-8     Percent Relative Yield Loss for Forest Types by Region for
             Modeled Scenarios	6-24
Table 6-9     Percent Relative Yield Gain for Forest Types by Region with Respect to the
             Existing Standard	6-25
Table 6-10    Percentage Changes in National Timber Prices	6-29
                                      IV

-------
Table 6-11    Consumer and Producer Surplus in Forestry, Million $2010	6-31
Table 6-12    Os Sensitive Trees and Their Uses	6-32
Table 6-13    Quantity of NTFP Harvested on U.S. Forest Service and Bureau of Land
             Management Land	6-35
Table 6-14    Definition of FASOMGHG Production Regions and Market Regions	6-39
Table 6-15    Mapping of O3 Impacts on Crops to FASOMGHG Crops	6-40
Table 6-16    Consumer and Producer Surplus in Agriculture (Million 2010$)	6-45
Table 6-17    Annualized Changes in Consumer and Producer Surplus in Agriculture and
             Forestry, 2010-2040, Million 2010$ (4% Discount Rate)	6-46
Table 6-18    Increase in Carbon Storage, MMtCC^e, Cumulative over 40 years	6-48
Table 6-19    Tree Species with Available C-R Functions in Selected Urban Study Areas.... 6-52
Table 6-20     63 Effects on Carbon Storage for Five Urban Areas over 25 Years
             (in millions of metric tons)	6-54
Table 6-21    Comparison of Pollutant Removal Between an Unadjusted Scenario and
             Alternative Simulations and Gains Between the Existing Standard and
             Alternatives (metric tons)	6-58
Table 6-22    Percent of Total Basal Area Covered by 12 Assessed Tree  Species	6-60
Table 6-23    Grid Cells With No Data That Exceed W126 Index Values under Recent
             Conditions	6-61
Table 6-24    Percent of Area Exceeding 2% Weighted Biomass Loss - Recent Conditions and
             Existing Standard	6-63
Table 6-25    Percent of Area Exceeding 2% Weighted Biomass Loss - Alternative W126
             Standard Levels	6-64
Table 6-26    Weighted RBL and Percent Cover in Class I Areas	6-66
Table 6-27    Summary of Qualitative Uncertainty Analysis in Relative Biomass Loss
             Assessments	6-68
Table 7-1     Percent of Cover Category Exceeding W126 Index Values	7-4
Table 7-2     National Outdoor Activity Participation	7-8
Table 7-3     National Expenditures for Wildlife Watching, Trail, and Camp-Related
             Recreation (in billions of 2010$)	7-10
Table 7-4     Summary of Biosite Index Values for 2006 to 2010 O3 Biomonitoring Sites. .7-14
Table 7-5     Censored Regression Results	7-16
Table 7-6     W126 Benchmark Criteria for Os Exposure and Relative Soil Moisture in 6
             Scenarios used in Screening-Level Assessment of Parks	7-27
Table 7-7     Percent of 214 Parks that Exceed Benchmark Criteria in Each Year
             (2006-2010) in 6 Scenarios	7-29
                                       v

-------
Table 7-8     Screening-level Foliar Injury Results in 42 Parks with an Os Monitor using 3
             Methods for Assigning Os Exposure to Each Parkin Base Scenario	7-34
Table 7-9     Foliar Injury Sensitivity Analyses for 214 Parks	7-38
Table 7-10    Foliar Injury Sensitivity Analyses for Soil Moisture in 57 Os Monitors
             in Parks   	7-40
Table 7-11    Value of Most Frequent Visitor Activities at Great Smoky Mountains
             National Park	7-48
Table 7-12    Visitor Spending and Local Area Economic Impact of GRSM	7-48
Table 7-13    Median Travel Cost for GRSM Visitors	7-49
Table 7-14    Geographic Area of GRSM after Just Meeting Existing and
             Alternative Standard Levels (km2)	7-54
Table 7-15    Value of Most Frequent Visitor Activities at ROMO	7-55
Table 7-16    Visitor Spending and Local Area Economic Impact of ROMO	7-56
Table 7-17    Median Travel Cost for ROMO Visitors	7-56
Table 7-18    Geographic Area of ROMO after Just Meeting Existing and
             Alternative Standard Levels (km2)	7-58
Table 7-19    Value of Most Frequent Visitor Activities at Sequoia and Kings
             Canyon National Parks	7-62
Table 7-20    Visitor Spending and Local Area Economic Impact of SEKI	7-63
Table 7-21    Median Travel Cost for SEKI Visitors	7-63
Table 7-22    Geographic Area of SEKI after Just Meeting Existing and
             Alternative Standard Levels  (km2)	7-64
Table 7-23    Summary of Qualitative Uncertainty  Analysis in Visible Foliar Injury
             Assessments	7-68
Table 8-1     Summary of Os-Exposure Risk Across Alternative W126 Standards Relative to
             Just Meeting Existing Standard - National-Scale Analyses	8-24
Table 8-2     Summary of Os-Exposure Risk Across Alternative Standards Relative to Just
             Meeting Existing Standard - Case Study-Scale Analyses	8-26
                                       VI

-------
                                  LIST OF FIGURES

Figure 2-1    Causal Determinations for Os Welfare Effects	2-6
Figure 2-2    Linkages Between Ecosystem Services Categories and Components of
             Human Weil-Being	2-9

Figure 4-1    Map of U.S. Ambient O3 Monitoring Sites in Operation During 2006-2010	4-3
Figure 4-2    Map of Monitored 8-hour O3 Design Values for the 2006-2008 Period	4-4
Figure 4-3    Flowchart of Air Quality Data Processing for Different Parts of the Welfare Risk
             and Exposure Assessments	4-5
Figure 4-4    Monitored 2006-2008 Average W126 Concentrations in ppm-hrs	4-8
Figure 4-5    National Surface of Observed 2006-2008 Average W126 Concentrations,
             in ppm-hrs	4-9
Figure 4-6    Map of the 9 NOAA Climate Regions Used in National-scale Air Quality
             Adjustments	4-13
Figure 4-7    National Surface of 2006-2008 Average W126 Concentrations (in ppm-hrs)
             Adjusted to Just Meet the Existing O3 Standard of 75 ppb	4-16
Figure 4-8    Difference in ppm-hrs Between National Surface of Observed 2006-2008
             Average W126 Concentrations and National Surface of 2006-2008 Average
             W126 Concentrations Adjusted to Just Meet the Existing O3 Standard
             of 75 ppb	4-17
Figure 4-9    National Surface of 2006-2008 Average W126 concentrations (in ppm-hrs)
             Adjusted to Just Meet the Potential Alternative Standard of 15 ppm-hrs	4-18
Figure 4-10   Difference in ppm-hrs Between National Surface of 2006-2008 Average W126
             Concentrations Adjusted to Just Meet Existing O3 Standard of 75 ppb and
             National Surface of 2006-2008 Average W126 Concentrations Adjusted to Just
             Meet Potential Alternative Standard of 15 ppm-hrs	4-19
Figure 4-11   National Surface of 2006-2008 Average W126 Concentrations (in ppm-hrs)
             Adjusted to Just Meet Potential Alternative Standard of 11 ppm-hrs	4-20
Figure 4-12   Difference in ppm-hrs Between National Surface of 2006-2008 Average W126
             Concentrations Adjusted to Just Meet the Existing O3 Standard of 75 ppb and
             National Surface of 2006-2008 Average W126 Concentrations Adjusted to Just
             Meet the Potential Alternative Standard of 11 ppm-hrs	4-21
Figure 4-13   National Surface of 2006-2008 Average W126 concentrations (in ppm-hrs)
             Adjusted to Just Meet the Potential Alternative Standard of 7 ppm-hrs	4-22
                                      vn

-------
Figure 4-14   Difference in ppm-hrs Between National Surface of 2006-2008 Average W126
             Concentrations Adjusted to Just Meet the Existing O3 Standard of 75 ppb and
             National Surface of 2006-2008 Average W126 Concentrations Adjusted to Just
             Meet the Potential Alternative Standard of 7 ppm-hrs	4-23
Figure 4-15   Empirical Probability Density and Cumulative Distribution Functions for
             Monitored 2006-2008 8-hour O3 Design Values, and the 2006-2008 8-hour O3
             Design Values After Adjusting to Just Meet the Existing and Potential Alternative
             Standards	4-24
Figure 4-16   Empirical Probability Density and Cumulative Distribution Functions for
             Monitored 2006-2008 Average W126 Concentrations, and the 2006-2008
             Average W126 Concentrations After Adjusting to Just Meet the Existing and
             Potential Alternative Standards	4-25
Figure 5-1    Conceptual Diagram of the Major Pathway through which O3 Enters Plants
             and the Major Endpoints that O3 May Affect in Plants and Ecosystems	5-1
Figure 5-2    Relationship between Ecological Effects of O3 Exposure and
             Ecosystem Services	5-2
Figure 5-3    Overlap of W126 Index Values for the Existing Standard and Alternative W126
             Standards, Fire Threat > 2, and Mixed Conifer Forest	5-7
Figure 5-4    Location of Fires in 2010 in Mixed Conifer Forest Areas (under Recent O3
             Conditions)	5-9
Figure 5-5    Southern Pine Beetle Damage	5-10
Figure 5-6    Southern Pine Beetle Damage	5-10
Figure 5-7    W126  Index Values for Just Meeting the Existing and Alternative Standards in
             Areas Considered 'At Risk'  of High Basal Area Loss (>25%Loss)	5-12
Figure 5-8    Percent of Forest Land in the US by Ownership Category, 2007	5-14
Figure 6-1    Conceptual Diagram of Relationship of Relative Biomass Loss to Ecosystem
             Services  	6-1
Figure 6-2    Relative Biomass Loss Functions for 12 Tree Species	6-5
Figure 6-3    Relative Biomass Loss Functions for 10 Crop Species	6-6
Figure 6-4    W126  Index Values for Alternative Percent Biomass Loss for Tree Species ... 6-11
Figure 6-5    W126  Index Values for Alternative Percent Biomass Loss for Crop Species... 6-11
Figure 6-6    Relative Biomass Loss of Ponderosa Pine (Pinusponderosd) Seedlings under
             Recent Ambient W126 Index Values (2006 - 2008)	6-13
Figure 6-7    Relative Biomass Loss of Ponderosa Pine with O3 Exposure After Simulating
             Meeting the Existing (8-hr) Primary Standard (75 ppb)	6-13
                                      vin

-------
Figure 6-8    Relative Biomass Loss of Ponderosa Pine with O3 Exposure After Simulating
             Meeting an Alternative Secondary Standard of 15 ppm-hrs (after Meeting
             Existing O3 Standard)	6-14
Figure 6-9    Relative Biomass Loss of Ponderosa Pine with O3 Exposure After Simulating
             Meeting an Alternative Secondary Standard of 11 ppm-hrs (after Meeting
             Existing O3 Standard)	6-14
Figure 6-10   Relative Biomass Loss of Ponderosa Pine with O3 Exposure After Simulating
             Meeting an Alternative Secondary Standard of 7 ppm-hrs (after Meeting Existing
             O3 Standard)	6-15
Figure 6-11   Relative Biomass Loss of Ponderosa Pine at the Existing Primary and Alternative
             Secondary Standards	6-18
Figure 6-12   Proportion of Current Standard, Ponderosa Pine - Recent Conditions and
             Alternative Secondary Standards	6-19
Figure 6-13   Three-Year Compounded Relative Biomass Loss - Southeast and Southwest
             Regions  	6-20
Figure 6-14   Individual and 3-Year Average W126 Index Values - Southeast and Southwest
             Regions  	6-21
Figure 6-15   RYGfor Softwoods by Region	6-27
Figure 6-16   RYG for Hardwoods by Region	6-28
Figure 6-17   Percentage Changes in Corn RYG with Respect to 75 ppb	6-42
Figure 6-18   Percentage Changes in Corn RYG with Respect to 75 ppb  	6-43
Figure 7-1    Relationship between Visible Foliar Injury and Ecosystem Services	7-2
Figure 7-2    Tree Species Sensitive to Foliar Injury	7-3
Figure 7-3    Examples of Foliar Injury from O3 Exposure	7-6
Figure 7-4    Examples of Southern Bark Beetle Damage	7-7
Figure 7-5    Os Biomonitoring Sites	7-12
Figure 7-6    344 Climate Divisions with Palmer Z Soil Moisture Data	7-13
Figure 7-7    General Relationship of O3 (ppm-hrs) and Biosite Index	7-14
Figure 7-8    General Relationship of Average Palmer Z (April to August) and
             Biosite Index	7-15
Figure 7-9    Cumulative Proportion of Sites with Foliar Injury Present, by Year	7-17
Figure 7-10   Cumulative Proportion of Sites with Elevated Foliar Injury, by Year	7-18
Figure 7-11   Cumulative Proportion of Sites with Foliar Injury Present,
             by Moisture Category	7-19
Figure 7-12   Cumulative Proportion of Sites with Elevated Foliar Injury,
             by Moisture Category	7-19
                                       IX

-------
Figure 7-13   Cumulative Proportion of Sites with Foliar Injury Present,
             by Climate Region	7-20
Figure 7-14   Timeframe of W126 Estimates for 57 Monitors Located in Parks	7-22
Figure 7-15   214 National Parks included in the Screening-Level Assessment	7-24
Figure 7-16   Distribution of Os and Soil Moisture in 214 Parks by Year	7-25
Figure 7-17   Presence of Os-Sensitive Species in Parks	7-28
Figure 7-18   Screening-Level Results for Foliar Injury in 214 Parks in 6 Scenarios	7-30
Figure 7-19   Parks Exceeding Benchmark Criteria for at least 3 years by Scenario
             and Climate Region	7-31
Figure 7-20   Foliar Injury Results Maps for 6 Scenarios for 214 Parks	7-32
Figure 7-21   Percent of 214 Parks at Different W126 Levels with Adjustments for
             Existing and Alternative Standards (3-year average)	7-43
Figure 7-22   Os Exposure in Ten Most Visited National Parks after Just Meeting
             Existing and Alternative Standards (3-year average)	7-44
Figure 7-23   Identification of W126 Index Value where 10 Percent of Biosites Show
             Any Foliar Injury	7-46
Figure 7-24   Cover of Sensitive Species in GRSM	7-51
Figure 7-25   Trail Cover of Sensitive Species in GRSM	7-52
Figure 7-26   GRSM Trail Kilometers by Species Cover Category	7-53
Figure 7-27   Sensitive Vegetation Cover in GRSM Scenic Overlooks (3km)	7-54
Figure 7-28   Sensitive Species Cover in ROMO	7-59
Figure 7-29   ROMO Sensitive Species Trail Cover	7-60
Figure 7-30   ROMO Trail Cover by Sensitive Species Type	7-61
Figure 7-31   Sensitive Species Cover in SEKI	7-65
Figure 7-32   Sensitive Species Trail Cover in SEKI	7-66
Figure 7-33   SEKFs Sensitive Species Cover by Type	7-67
Figure 8-1    Map of 9 NOAA Climate Regions used in National-Scale Air
             Quality Adjustments  	8-5
Figure 8-2    Cumulative Proportion of Sites with Visible Foliar Injury Present, by Moisture
             Category	8-12
Figure 8-3    National Surface of 2006-2008 Average W126 Index values Adjusted to Just
             Meet the Alternative Standard of 15 ppm-hrs	8-19
Figure 8-4    National Surface of 2006-2008 Average W126 Index values Adjusted to Just
             Meet the Alternative Standard of 11 ppm-hrs	8-20
Figure 8-5    Cumulative Proportion of Sites with Foliar Injury Present, by Year	8-22

-------
                            APPENDICES


4A:   SPATIAL FIELDS FOR THE W126 METRIC

5A:   LARGER MAPS OF FIRE THREAT AND BASAL AREA LOSS

6A:   MAPS OF INDIVIDUAL TREE SPECIES

6B:   ASSESSMENT OF THE IMPACTS OF ALTERNATIVE OZONE
     CONCENTRATIONS ON THE U.S. FOREST AND AGRICULTURE SECTORS

6C:   SUPPLY CURVE SHIFTS

6D:   iTREE MODEL

6E:   CLASS I AREAS AND WEIGHTED RBL AT CURRENT STANDARD
     AND ALTERNATIVE W126 STANDARD LEVELS

7A:   ADDITIONAL INFORMATION FOR SCREENING-LEVEL ASSESSMENT OF
     VISIBLE FOLIAR INJURY IN NATIONAL PARKS

7B:   NATIONAL PARKS CASE STUDY LARGE SCALE MAPS
                             XI

-------
                     LIST OF ACRONYMS/ABBREVIATIONS
AGSIM
AQCD
AQS
BLM
CAA
CALFIRE
CASAC
CASTNET
C.F.R.
CH4
CMAQ
CO
C-R
CSTR
EGU
EPA
FACE
FASOMGHG
FHM
FHTET
FIA
FR
GIS
GRSM
HDDM
HNO3
HO2
IMPLAN®
IRP
ISA
i-Tree
MEA
MSA
NAAQS
NCDC
NCLAN
NCore
NEI
FHWAR
NHEERL-WED

NOAA
NOx
NPP
Agriculture Simulation Model
Air Quality Criteria Document
Air Quality System
Bureau of Land Management
Clean Air Act
California Department of Forestry and Fire Protection
Clean Air Scientific Advisory Committee
Clean Air Status and Trends Network
Code of Federal Regulations
Methane
Community Multi-Scale Air Quality
Carbon Monoxide
Concentration-Response
Continuous Stirred Tank Reactors
Electric Generating Unit
Environmental Protection Agency
Free- Air Carbon Dioxide/Ozone Enrichment
Forest and Agriculture Sector Optimization Model with Greenhouse Gases
Forest Health Monitoring
Forest Health Technology Enterprise Team
Forest Inventory and Analysis
Federal Register
Geographic Information System
Great Smoky Mountains National Park
Higher-Order Decoupled Direct Method
Nitric Acid
Hydro-Peroxy Radical
Impact Analysis for Planning Model
Integrated Review Plan
Integrated Science  Assessment
Urban Forestry Analysis and Benefits Assessment Tool
Millennium Ecosystem Assessment
Metropolitan Statistical Area
National Ambient Air Quality Standards
National Climatic Data Center
National Crop Loss Assessment Network
National Core
National Emissions Inventory
National Survey of Fishing, Hunting, and Wildlife Associated Recreation
National Health and Environmental Effects Laboratory - Western Ecology
Division
National Oceanographic and Atmospheric Administration
Oxides of Nitrogen
Net Primary Productivity
                                     xn

-------
NFS
NSRE
NTFP
03
OAQPS
OBP
OH
OIF
OTC
PA
PAMS
ppb
ppm-hrs
POMS
RBL
REA
ROMO
RYG
RYL
SEKI
SLAMS
SOx
SPMS
STE
TREGRO
UNESCO
U.S.
USDA
U.S. EPA
USFS
USGS
VegBank
VNA
voc
WHO
W126

WTP
ZELIG
National Park Service
National Survey on Recreation and the Environment
Non-Timber Forest Products
Ozone
Office of Air Quality Planning and Standards
Ozone Biomonitoring Program
Hydroxl Radical
Outdoor Industry Foundation
Open-Top Camber
Policy Assessment
Photochemical Assessment Monitoring Stations
Parts per Billion
Parts per Million Hours
Portable O3 Monitoring System
Relative Biomass Loss
Risk and Exposure Assessment
Rocky Mountains National Park
Relative Yield Gain
Relative Yield Loss
Sequoia/Kings Canyon National Parks
State and Local Monitoring Stations
Oxides of Sulfur
Special Purpose Monitoring Stations
Stratosphere-Troposphere Exchange
Tree Growth Model
United National Education,  Scientific, and Cultural Organization
United States
United States Department of Agriculture
United States Environmental Protection Agency
United States Forest Service
United States Geological Society
Vegetation Plot Database
Voronoi Neighbor Averaging
Volatile Organic Compound
World Health Organization
Cumulative Integrated Exposure Index with a Sigmoidal Weighting
Function
Willingness-to-Pay
A Forest Succession Simulation Model
                                     xin

-------
This page left intentionally blank.

-------
 1                                   1   INTRODUCTION

 2          The U.S. Environmental Protection Agency (EPA) is presently conducting a review of
 3    the national ambient air quality standard (NAAQS) for ozone (Os) and related photochemical
 4    oxidants.  The NAAQS review process includes four key phases: planning, science assessment,
 5    risk/exposure assessment, and policy assessment/rulemaking.1 This process and the overall plan
 6    for this review of the Os NAAQS are presented in the Integrated Review Plan for the Ozone
 1    National Ambient Air Quality Standards (TRP, US EPA, 201 la). The IRP additionally presents
 8    the schedule for the review; identifies key policy-relevant issues; and discusses the key scientific,
 9    technical, and policy documents.  These documents include an Integrated Science Assessment
10    (ISA), Risk and Exposure Assessments (REAs), and a Policy Assessment (PA). This draft
11    Welfare REA is one of the two quantitative REAs developed for the review by EPA's Office of
12    Air Quality Planning and Standards (OAQPS); the second is a Health REA. This draft Welfare
13    REA focuses on assessments to inform consideration of the review of the secondary (welfare-
14    based) NAAQS for O3.
15          The existing secondary standard for Os is set identical to the primary standard at a level
16    of 0.075 ppm, based on the annual fourth-highest daily maximum 8-hour average concentration,
17    averaged over three years (73 FR 16436). The EPA initiated the current review of the Os
18    NAAQS on September 29, 2008 with an announcement of the development of an Os ISA and a
19    public workshop to discuss policy-relevant science to inform EPA's integrated plan for the
20    review of the O3 NAAQS (73 FR 56581).  Discussions at the workshop, held on October 29-30,
21    2008, informed identification of key policy issues and questions to frame the review of the Os
22    NAAQS.  Drawing from the workshop discussions, EPA developed a draft and then final IRP
23    (U.S. EPA, 201 la).2 In early 2013, EPA completed the Integrated Science Assessment for Ozone
24    and Related Photochemical Oxidants (ISA, U.S. EPA, 2013). The ISA provides a concise
       1 For more information on the NAAQS review process, see http://www.epa.gov/ttn/naaqs/review.html.
       2 On March 30, 2009, EPA held a public consultation with the CASAC O3 Panel on the draft IRP. The final IRP
      took into consideration comments received from CASAC and the public on the draft plan, as well as input from
      senior Agency managers.
                                                1-1

-------
 1    review, synthesis, and evaluation of the most policy-relevant science to serve as a scientific
 2    foundation for the review of the NAAQS.  The scientific and technical information in the ISA,
 3    including that newly available since the previous review on the welfare effects of O3, includes
 4    information on exposure, physiological mechanisms by which Os might adversely impact
 5    vegetation, and an evaluation of the ecological evidence, including information on reported
 6    concentration-response (C-R) relationships for Os-related changes in plant biomass.
 7          The REA is a concise presentation  of the conceptual model, scope, methods,  key results,
 8    observations, and related uncertainties associated with the quantitative analyses performed. This
 9    REA builds upon the welfare effects evidence presented and assessed in the ISA, as well as
10    CAS AC advice (Samet, 2011) and public comments on a scope and methods planning document
11    for the REA (here after, "Scope and Methods Plan", U.S. EPA, 201 Ib). Preparation of this
12    second draft REA draws upon the final ISA and reflects consideration of CAS AC and public
13    comments on the first draft REA (Frey and Samet, 2012). This second draft welfare REA is
14    being released, concurrently with the second draft health REA and second draft PA,  for review
15    by the CASAC O3 Panel at a public meeting scheduled for March 25-27, 2014, and for public
16    comment.
17          The second draft PA presents a staff evaluation and preliminary staff conclusions of the
18    policy implications of the key scientific and technical information in the ISA and second draft
19    REAs. When final, the PA is intended to help "bridge the gap" between the Agency's scientific
20    assessments presented in the ISA and REAs and the judgments required of the EPA
21    Administrator in determining whether it is  appropriate to retain or revise the NAAQS.  The PA
22    integrates and interprets the information from the ISA and REAs to frame policy options for
23    consideration by the Administrator. In so doing, the PA recognizes that the selection of a
24    specific approach to reaching final decisions on primary and secondary NAAQS will reflect the
25    judgments of the Administrator. The development of the various scientific, technical and policy
26    documents and their roles in informing this NAAQS review are described in more detail in the
27    second draft PA.
                                               1-2

-------
 1     1.1   HISTORY
 2          As part of the previous O3 NAAQS review completed in 2008, EPA's OAQPS conducted
 3    quantitative risk and exposure assessments to estimate risks to human welfare based on
 4    ecological effects associated with exposure to ambient O3 (U.S. EPA 2007a, U.S. EPA 2007b).
 5    The assessment scope and methodology were developed with considerable input from CASAC
 6    and the public, with CASAC generally concluding that the exposure assessment reflected
 7    generally-accepted modeling approaches, and that the risk assessments were well done, balanced
 8    and reasonably communicated (Henderson, 2006a).  The final quantitative risk and exposure
 9    assessments took into consideration CASAC advice (Henderson, 2006a; Henderson, 2006b) and
10    public comments on two drafts of the risk and exposure assessments.
11          The assessments conducted as part of the previous review focused on national-level O3-
12    related impacts to sensitive vegetation and their associated ecosystems.  The vegetation exposure
13    assessment was performed using an interpolation approach that included information from
14    ambient monitoring networks and results from air quality modeling. The vegetation risk
15    assessment included both tree and crop analyses.  The tree risk analysis included three distinct
16    lines of evidence: (1) observations of visible foliar injury in the field linked to monitored O3 air
17    quality for the years 2001 - 2004; (2) estimates of seedling growth loss under then-current and
18    alternative O3 exposure conditions; and (3) simulated mature tree growth reductions using the
19    TREGRO model to simulate the effect of meeting alternative air quality standards on the
20    predicted annual growth of mature trees from three different species. The crop risk analysis
21    included estimates of crop yields under current and alternative O3 exposure conditions. The
22    assessments also analyzed the associated changes in economic value upon meeting the levels of
23    various alternative standards using an agricultural sector economic model.3
24          Based on the 2006 Air Quality Criteria for Ozone (U.S. EPA, 2006), the 2007 Staff Paper
25    (U.S. EPA, 2007) and related technical support documents (including the risk and exposure
26    assessments), EPA published a proposed decision in the Federal Register on July 11, 2007 (72
        3 We addressed key observations and insights from the O3 risk assessment, in addition to important caveats and
      limitations, in Section II.B of the Final Rule notice (73 FR 16440 to 16443, March 27, 2008).
                                                1-3

-------
 1   FR 37818). The EPA proposed to revise the level of the primary standard to a level within the
 2   range of 0.075 to 0.070 ppm. Two options were proposed for the secondary standard: (1)
 3   replacing the current standard with a cumulative, seasonal standard, expressed as an index of the
 4   annual sum of weighted hourly concentrations cumulated over 12 daylight hours during the
 5   consecutive 3-month period within the Os season with the maximum index value (W126), set at a
 6   level within the range of 7 to 21 ppm-hours, and (2) setting the secondary standard identical to
 7   the revised primary standard. EPA completed the review with publication of a final decision on
 8   March 27, 2008 (73 FR 16436), revising the level of the 8-hour primary O3 standard from 0.08
 9   ppm to 0.075 ppm, as the 3-year average of the fourth highest daily maximum 8-hour average
10   concentration, and revising the secondary standard to be identical to the revised primary
11   standard.
12          In May 2008, state, public health, environmental, and industry petitioners filed suit
13   against EPA regarding the 2008 decision. At EPA's request, the consolidated cases were held in
14   abeyance pending EPA's reconsideration of the 2008 decision. The Administrator issued a
15   notice of proposed rulemaking to reconsider the 2008 final decision on January 6, 2010.  EPA
16   held three public hearings. The Agency solicited CAS AC review of the proposed rule on January
17   25, 2010 and additional CAS AC advice on January 26, 2011. On September 2, 2011, the Office
18   of Management and Budget returned the draft final  rule on reconsideration to EPA for further
19   consideration. EPA decided to coordinate further proceedings on its voluntary rulemaking on
20   reconsideration with the ongoing periodic review, by deferring the completion of its voluntary
21   rulemaking on reconsideration until it completes its statutorily-required periodic review. In light
22   of that, the litigation on the 2008 final decision proceeded.  On July 23, 2013, the Court ruled on
23   the litigation of the 2008 decision, denying the petitioners suit except with respect to the
24   secondary standard, which was remanded to the Agency for reconsideration. The second draft
25   PA provides additional description of the court ruling with regard to the secondary standard.

26     1 2   CURRENT RISK AND EXPOSURE ASSESSMENTS: GOALS AND PLANNED
27          APPROACH
28          This second draft REA provides an assessment of exposure and risk associated with
29   recent ambient concentrations of Os and Os air quality simulated to just meet the existing

                                                1-4

-------
 1    secondary 63 standard and just meeting potential alternative 63 standards based on
 2    recommendations provided in the first draft of the PA. To inform the PA regarding the adequacy
 3    of existing standards and the potential for reductions in adverse effects associated with
 4    alternative standards that might be considered, the goals of the current quantitative welfare REA
 5    are to (1) provide estimates of the ecological effects of Os exposure across a range of
 6    environments; (2)  provide estimates of ecological effects within selected case study areas; (3)
 7    provide estimates of the effects of Os exposure on specific urban and non-urban ecosystem
 8    services based on the causal ecological effects; and (4) develop a better understanding of the
 9    response of ecological systems and ecosystem services to changing 63 exposure. This current
10    quantitative risk and exposure assessment builds on the approach used and lessons learned in the
11    previous O3 risk assessment and focuses on improving the characterization of the overall
12    confidence in the risk estimates, including related uncertainties, by improving the methods and
13    data used in the analyses; this current risk and exposure assessment also incorporates the range
14    of ecosystem effects and expands the characterization of adversity to include consideration of
15    impacts to ecosystem services.  This assessment considers a variety of welfare endpoints for
16    which,  in our judgment, there is  adequate information to develop quantitative risk estimates that
17    can meaningfully inform the review of the secondary Os NAAQS.

18     1.3    ORGANIZATION  OF DOCUMENT
19           The remainder of this document is organized into chapters.   Chapter 2 provides a
20    conceptual framework for the risk and exposure assessment, including discussions of O3
21    chemistry, sources of Os precursors, ecological exposure pathways and uptake into plants,
22    ecological effects, and ecosystem services endpoints associated with Os.  This conceptual
23    framework sets the stage for the scope of the risk and exposure assessments. Chapter 3 provides
24    an overview of the scope of the quantitative risk and exposure assessments, including a summary
25    of the previous risk and exposure assessments and an overview of the current risk and exposure
26    assessments.  Chapter 4 discusses air quality considerations relevant to the exposure and risk
27    assessments, including available 63 monitoring data and important air quality inputs to the risk
28    and exposure assessments. Chapter 5 describes the ecological effects of O3 exposure and the
29    associated ecosystem services, including the ecosystem services for which data and methods for
30    incremental analysis of direct Os are not yet available. Chapter 6 provides quantitative analysis
                                                 1-5

-------
1    of the biomass loss effects of 63 and the ecosystem services affected by this loss, such as
2    provision of food and fiber, carbon sequestration and storage, and pollution removal. Chapter 7
3    provides quantitative assessments of the effects of O3 on foliar injury and associated ecosystem
4    services, particularly cultural services related to recreation and the three selected National Park
5    case studies.  Chapter 8 provides an integrated discussion of the risk estimates generated in these
6    analyses, drawing on the results of the quantitative analyses and incorporating considerations
7    from the qualitative discussion of ecosystem services.
                                                 1-6

-------
 1                                     2   FRAMEWORK

 2          In this chapter, we summarize the conceptual framework for assessing exposures of
 3    ecosystems to Os and the associated risks to public welfare.  This conceptual framework includes
 4    elements related to characterizing: (1) Os chemistry (Section 2.1); (2) important sources of Os
 5    precursors, including oxides of nitrogen (NOX) and volatile organic compounds (VOC) (Section
 6    2.2); (3) Os-induced effects occurring on Os-sensitive species and in their associated ecosystems
 7    (Section 2.3); and (4) ecosystem services that are likely to be negatively impacted by changes in
 8    ecological functions resulting from 63 exposures (Section 2.4).  We conclude the chapter with
 9    key observations relevant for developing the scope of the quantitative risk and  exposure
10    assessments.
11          In the previous review of the secondary standards, we focused the ecological risk
12    assessment on estimating changes in biomass loss in forest tree species and yield loss in
13    agricultural crops, as well as qualitatively considering effects on ecosystem services. In this
14    review, EPA expanded the analysis to consider the broader array of impacts on ecosystem
15    services resulting from known effects of Os exposure on ecosystem  functions.  This expanded
16    scope is addressed in the risk assessment by quantifying the risks not just to ecosystems, but also
17    to the aspects of public welfare dependent on those ecosystems, i.e., services. EPA has started
18    using an ecosystem services framework to help inform determinations of the adversity to public
19    welfare associated with changes  in ecosystem functions (Rea et al, 2012). The Risk and
20    Exposure Assessment conducted as  part of the Review of the Secondary National Ambient Air
21    Quality Standards for Oxides of Nitrogen and Oxides of Sulfur (U.S. EPA, 2009) presented
22    detailed discussions  of how ecosystem  services and public welfare are related and how an
23    ecosystem services framework may  be  employed to evaluate effects on welfare. In this risk
24    assessment we will identify the ecosystem services associated with the  ecological effects caused
25    by Os exposure for both the national scale assessment and the more  refined case study areas.
26    These services may be characterized as: supporting services that are necessary  for all other
27    services (e.g., primary production); cultural services including existence and bequest values,
28    aesthetic values, and recreation values, among others; provisioning services (e.g., food and
29    timber); and regulating services such as climate regulation or hydrologic cycle  (Millenium
30    Ecosystem Assessment, 2005).

                                                     2-1

-------
 1     2.1   O3 CHEMISTRY
 2          O3 occurs naturally in the stratosphere where it provides protection against harmful solar
 3    ultraviolet radiation; Os is also formed closer to the Earth's surface in the troposphere by both
 4    natural and anthropogenic sources. Os is not emitted directly into the air, but is created when its
 5    two primary precursors, VOC and NOX, combine in the presence of sunlight. VOC and NOX are,
 6    for the most part, emitted directly into the atmosphere. Carbon monoxide (CO) and methane
 7    (CH4) are also important for O3 formation (U.S. EPA, 2013, section 3.2.2).
 8          Rather than varying directly with emissions of its precursors,  Os changes in a nonlinear
 9    fashion with the concentrations of its precursors.  Nitrogen oxide emissions lead to both the
10    formation and destruction of Os, depending on the local quantities of NOX, VOC, and radicals
11    such as the hydroxyl (OH) and hydro-peroxy (HO2) radicals. In areas dominated by fresh NOX
12    emissions, these radicals are removed via the production of nitric acid (HNOs), which lowers the
13    Os formation rate. The reduction in, or scavenging of, Os by this reaction is called "titration"
14    and is often found in downtown metropolitan areas, especially near busy streets and roads, and in
15    power plant plumes. Titration is usually short-lived and confined to areas close to strong NOX
16    sources; titration results in localized valleys in which Os  concentrations are low compared to
17    surrounding areas. Consequently, Os response to reductions in NOX emissions is complex and
18    may include Os decreases at some times and locations and Os increases to fill in the local valleys
19    of low Os. In contrast, in areas with low NOX concentrations, such as remote continental areas
20    and rural and  suburban areas downwind of urban centers, the net production of Os varies directly
21    with NOX concentrations and typically increases with increasing NOX emissions.
22          In general, the rate of Os  production is limited by the concentration of VOC or NOX, and
23    Os formation based on these two precursors depends on the relative sources of OH and NOX.
24    When OH radicals are abundant  and are not depleted by reaction with NOX and/or other species,
25    O3 production is "NOx-limited" (U.S. EPA, 2013, section 3.2.4). In this NOx-limited
26    circumstance, Os concentrations are most effectively reduced by lowering NOX emissions rather
27    than by lowering VOC emissions.  When OH and other radicals are not abundant, either through
28    low production or reactions with NOX and other species, Os production is referred to as "VOC-
29    limited", "radical-limited", or "NOx-saturated" (Jaegle et al., 2001), and O3 is most effectively
30    reduced by lowering VOC  emissions. However, even in NOx-saturated conditions, very large
31    decreases in NOX emissions can cause the Os formation regime to become NOx-limited.

                                                     2-2

-------
 1    Consequently, large reductions in NOX emissions can make further emissions reductions more
 2    effective at reducing 63.  Between the NOx-limited and NOx-saturated extremes there is a range
 3    where Oj is relatively insensitive to marginal changes in both NOX and VOC emissions.
 4          In rural areas and downwind of urban areas, 63 production is generally NOx-limited.
 5    This is particularly true in rural areas such as national parks, national forests, and state parks
 6    where VOC emissions from vegetation are high and anthropogenic NOX emissions are relatively
 7    low.  Due to lower chemical scavenging in non-urban areas, O3 tends to persist longer in rural
 8    than in urban areas and tends to lead to higher cumulative exposures in rural areas than in urban
 9    areas (U.S. EPA, 2013, Section 3.6.2.2).
10          We focused the analyses in the welfare risk and exposure assessments on the W126 Os
11    exposure metric. The W126 metric is a seasonal sum of hourly O3 concentrations, designed to
12    measure the cumulative effects of Os exposure on vulnerable plant and tree species.  The W126
13    metric uses a sigmoidal weighting function to place less emphasis on exposure to low
14    concentrations and more emphasis on exposure to high  concentrations.

15    2.2   SOURCES OF O3 AND O3 PRECURSORS
16          63 precursor  emissions can be divided into anthropogenic and natural source categories,
17    with natural sources further divided into biogenic emissions (from vegetation, microbes, and
18    animals) and abiotic  emissions (from biomass burning,  lightning, and geogenic sources). The
19    anthropogenic precursors of O3 originate from a wide variety of stationary and mobile sources.
20          In urban areas, both biogenic and anthropogenic VOC emissions are relevant to 63
21    formation. Hundreds of VOC are emitted by evaporation and combustion processes from a large
22    number of anthropogenic sources. Based on the 2005 national emissions inventory (NEI),
23    solvent use and highway vehicles are the two main sources of VOC emissions, with roughly
24    equal contributions to total emissions (U.S. EPA, 2013, Figure 3-2).  The emissions inventory
25    categories of "miscellaneous" (which includes agriculture and forestry, wildfires, prescribed
26    burns, and structural  fires) and off-highway mobile sources are the next two largest contributing
27    emissions categories, with a combined total of over 5.5  million metric tons a year (MT/year).
28          In rural areas and at the global scale, VOC emissions from vegetation are much larger
29    than those from anthropogenic sources. In the 2005 NEI, U.S. rural emissions from biogenic
                                                     2-3

-------
 1    sources were 29 MT/year, and emissions of VOC from anthropogenic sources were
 2    approximately 17 MT/year (wildfires constitute -1/6 of that total). Vegetation emits substantial
 3    quantities of VOC, such as isoprene and other terpenoid and sesqui-terpenoid compounds. Most
 4    biogenic emissions occur during the summer because of they depend on temperature and incident
 5    sunlight. Biogenic emissions are also higher in southern and eastern states than in northern and
 6    western states for these reasons and because of species variations.
 7          Anthropogenic NOX emissions are associated with combustion processes.  Based on the
 8    2005 NEI, the three largest sources of NOX emissions in the U.S. are on-road and off-road mobile
 9    sources (e.g., construction and agricultural equipment) and electric power generation plants
10    (electric generating units, or EGUs) (U.S. EPA, 2013, Figure 3-2). Emissions of NOX are highest
11    in areas with a high density of power plants and in urban regions with high traffic density.
12    However, it is not possible to make an overall statement about their relative impacts on Os in all
13    local areas because there are fewer EGUs than mobile sources, particularly in the west and south,
14    and because of the nonlinear chemistry discussed in Section 2.1.
15          Major natural sources of NOX in the U.S.  include lightning, soils, and wildfires. Biogenic
16    NOX emissions are generally highest during the summer and occur across the entire country,
17    including areas where anthropogenic emissions are low. It should be noted that uncertainties in
18    estimating natural NOX emissions are much larger than uncertainties in estimating anthropogenic
19    NOX emissions.
20          Os  concentrations in a region are affected both by local formation and by transport from
21    surrounding areas.  Os transport occurs on many  spatial scales, including local transport between
22    cities, regional transport over large regions of the U.S., and international/long-range transport.  In
23    addition, Os is also transferred from the stratosphere into the troposphere, which is rich in Os,
24    through stratosphere-troposphere exchange (STE). These inversions or "foldings" usually occur
25    behind cold fronts, bringing stratospheric air with them (U.S. EPA, 2013, section 3.4.1.1).
26    Contribution to Os concentrations in an area from STE are defined as being part of background
27    O3 (U.S. EPA, 2013, section 3.4).
28          Rural areas, such as national parks, national forests, and state parks, tend to be less
29    directly affected by anthropogenic pollution sources than urban sites.  However, they can be
30    regularly affected by transport of Os or Os precursors from upwind urban areas. In addition,
                                                      2-4

-------
 1    biogenic VOC emissions tend to be higher in rural areas, and major anthropogenic sources of 63
 2    precursor emissions such as highways, power plants, biomass combustion, and oil and gas
 3    operations are commonly found in rural areas, adding to the O3 produced in these areas.  Areas at
 4    higher elevations, such as many of the national parks in the western U.S., can also be affected
 5    more significantly by international transport of Os or stratospheric intrusions that transport Os
 6    into the area (U.S. EPA, 2013, section 3.7.3).

 7    2.3   ECOLOGICAL EFFECTS
 8          Recent studies reviewed in the ISA support and strengthen the findings reported in the
 9    2006 O3 Air Quality Criteria Document (AQCD) (U.S. EPA, 2006a).  The most significant new
10    body of evidence since the 2006 Oj AQCD comes from research on molecular mechanisms of
11    the biochemical and physiological changes observed in many plant species in response to Os
12    exposure.  These newer molecular studies not only provide very important information regarding
13    the many mechanisms of plant responses to Os, they also allow for the analysis of interactions
14    between various biochemical pathways that are induced in response to 03. However, many of
15    these studies have been conducted in artificial conditions with model plants, which are typically
16    exposed to very high, short doses of 63 and are not quantifiable as part of this risk assessment.
17          Chapter 9 of the O3 ISA (U.S. EPA, 2013) provides a detailed review of the effects of O3
18    on vegetation including the major pathways of exposure and known ecological and ecosystem
19    effects. In general, Os is taken up through the stomata into the leaves. Once inside the leaves, Os
20    affects a number of biological and physiological processes, including photosynthesis.  This leads,
21    in some cases, to visible foliar injury as well as reduced plant growth, which are the main
22    ecological effects assessed in this review. Visible foliar injury  and reduced growth can lead to a
23    reduction in ecosystem services, including crop and timber yield loss, decreased carbon
24    sequestration, alteration in community composition, and loss of recreational or cultural value.
25         Overall causal determinations are made based on the full  range of evidence including
26    controlled exposure studies and ecological studies. Figure 2-1 shows the Os welfare effects that
27    have been categorized by strength of evidence for causality in the O3 ISA (U.S. EPA, 2013,
28    Chapter 2). These determinations support causal or likely causal relationships between exposure
29    to Os and ecological and ecosystem-level effects.
30
                                                     2-5

-------
1
2
                 Not likely
                                Inadequate
                                  to infer
                                                             Reduced carbon
                                                             sequestration in
                                                             terrestrial ecosystems
                                                             Alteration of
                                                             terrestrial ecosystem
                                                             water cycling
                                                             Alteration of
                                                             terrestrial community
                                                             composition
Suggestive
Likely
      Figure 2-1      Causal Determinations for O3 Welfare Effects
                                                                             Visible foliar injury
                                                                             effects
                                                                             Reduced vegetation growth
                                                                             Reduced productivity in
                                                                             terrestrial ecosystems
                                                                             Reduced yield and
                                                                             quality of agricultural
                                                                             crops
                                                                             Alteration of below
                                                                             ground biogeochemical
                                                                             cycles
Causal
 4           The adequate characterization of the effects of 63 on plants for the purpose of setting air
 5    quality standards depends not only on the choice of the index used (i.e., W126) to summarize 63
 6    concentrations (Section 9.5 of the O3 ISA), but also on quantifying the response of the plant
 7    variables of interest at specific values of the selected index. The factors that determine the
 8    response of plants to Os exposure include species, genotype and other genetic characteristics,
 9    biochemical and physiological status, previous and current exposure to other stressors, and
10    characteristics of the exposure.
1 1           Quantitative characterization of exposure-response in the 2006 Os AQCD was based on
12    experimental data generated for projects conducted by the National Crop Loss Assessment
13    Network (NCLAN) and EPA's National Health and Environmental Effects Research Laboratory,
14    Western Ecology Division (NHEERL-WED) that used open-top chambers (OTCs) to expose
15    crops and trees seedling to 63. In recent years, additional yield and growth results for soybean
16    and aspen, respectively, (two of the species that provided extensive exposure-response
17    information in those projects) have become available from studies that used free-air carbon
18    dioxide/ozone enrichment (FACE) technology, which is  intended to provide conditions much
19    closer to natural environments (Pregitzer et al., 2008; Morgan et al., 2006; Morgan et al., 2004;
20    Dickson et al., 2000). The results of these FACE studies provided support for the earlier
21    findings reported in the OTC studies.
                                                       2-6

-------
 1          The quantitative exposure-response relationships described in the 2006 63 AQCD have
 2    not changed in the current ISA, with the exception of the addition of one new species.  The
 3    exposure-response models are summarized in the final ISA (U.S. EPA, 2013) and are computed
 4    using the W126 metric, cumulated over 90 days. These response functions provide an adequate
 5    basis for quantifying biomass loss damages.
 6          Visible foliar injury resulting from exposure to 63 has also been well characterized and
 7    documented over several decades of research on many tree, shrub, herbaceous, and crop species
 8    (U.S. EPA, 2006, 1996a, 1984, 1978). O3-induced visible foliar injury symptoms on certain
 9    bioindicator plant species are considered diagnostic as they have been verified experimentally in
10    exposure-response studies, using exposure methodologies such as continuous stirred tank
11    reactors (CSTRs), OTCs, and free-air fumigation. Experimental evidence has clearly established
12    a consistent association of visible injury with 63 exposure, with greater exposure often resulting
13    in greater and more prevalent injury.  This REA assesses the risk of visible foliar injury at
14    differing concentrations of 63 using U.S. Forest Service biomonitoring data along with soil
15    moisture information to establish certain risk benchmarks. However, without robust
16    concentration-response functions, a detailed quantitative assessment that can be applied across a
17    range of ecosystems for foliar injury is not currently possible.

18     2.4   ECOSYSTEM SERVICES
19          The Risk and Exposure Assessment conducted as part of the Review of the Secondary
20    National Ambient Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur evaluates
21    the benefits received from the resources and processes that are supplied by ecosystems.
22    Collectively, these benefits are known as ecosystem services and include products or provisions,
23    such as food  and fiber; processes that regulate ecosystems, such as carbon sequestration;  cultural
24    enrichment; and supportive processes for services, such as nutrient cycling. Ecosystem services
25    are distinct from other ecosystem products and functions because there is human demand for
26    these services. In the Millennium Ecosystem Assessment (MEA), ecosystem services are
27    classified into four main categories:
28          •   Provisioning — includes products obtained from ecosystems, such  as the production
29              of food and water.
                                                     2-7

-------
 1           •   Regulating — includes benefits obtained from the regulation of ecosystem processes,
 2              such as the control of climate and disease.

 3           •   Cultural — includes the nonmaterial benefits that people obtain from ecosystems
 4              through spiritual enrichment, cognitive development, reflection, recreation, and
 5              aesthetic experiences.

 6           •   Supporting — includes those services necessary for the production of all other
 7              ecosystem services, such as nutrient cycles and crop pollination (MEA, 2005).

 8           The concept of ecosystem services can be used to help define adverse effects as they
 9    pertain to NAAQS reviews. The most recent secondary NAAQS reviews have characterized
10    known or anticipated adverse effects to public welfare by assessing changes in ecosystem
11    structure or processes using a weight-of-evidence approach that includes both quantitative and
12    qualitative data. For example, the previous Os  NAAQS review evaluated changes in foliar
13    injury, growth loss, and biomass reduction on trees beyond the seedling stage using the
14    TREGRO model.  The presence or absence of foliar damage in counties meeting the existing
15    standard has been used as a way to evaluate the adequacy of the secondary NAAQS.
16    Characterizing a known or  anticipated adverse  effect to public welfare is an important
17    component of developing any secondary NAAQS. According to the Clean Air Act (CAA),
18    welfare effects include the following:
19           "Effects on soils, water, crops, vegetation, manmade materials, animals, wildlife,
20    weather, visibility, and climate, damage to and deterioration of property, and hazards to
21    transportation,  as well as effect on economic values and on personal comfort and well-being,
22    whether caused by transformation,  conversion, or combination with other air pollutants."
23    (Section 302(h))
24           In other words, welfare effects are those effects that are important to individuals and/or
25    society in general. Ecosystem services can be generally defined as the benefits that individuals
26    and organizations obtain from ecosystems. EPA has defined ecological goods and services as
27    the "outputs of ecological functions or processes that directly or indirectly contribute to social
28    welfare or have the potential to do so in the future.  Some outputs may be bought and sold, but
29    most are not marketed"  (U.S. EPA, 2006). Conceptually, changes in ecosystem services may be
                                                     2-8

-------
 1   used to aid in characterizing a known or anticipated adverse effect to public welfare.  In the
 2   context of this review, ecosystem services may also aid in assessing the magnitude and
 3   significance of a resource and in assessing how O3 concentrations may impact that resource.
 4          Figure 2-2 provides the World Resources Institute's schematic demonstrating the
 5   connections between the categories of ecosystem services and human well-being (MEA, 2005).
 6   The interrelatedness of these categories means that any one ecosystem may provide multiple
 7   services.  Changes in these services can impact human well-being by affecting security, health,
 8   social relationships, and access to basic material goods (MEA, 2005). The strength of the
 9   linkages, as indicated by arrow width, and the potential for mediation, as indicated by arrow
10   color, differ in different ecosystems and regions.
11
                                                          CONSTITUENTS OF WELL-BEING
12
13 Fig
14
ECOSYSTEM SERVICES
Provisioning
i • i £
FRESH WATER •
WOOD AND FIBER
FUEL
Supporting Regulating
NUTRIENT CYCLING CUMATE REGULATION
SOIL FORMATION FLOOD REGULATION
PRI^DUC™ SS^SSESS! *
...
Cultural
AESTHETIC
SPIRITUAL
EDUCATIONAL
RECREATIONAL
LIFE ON EARTH - BIODIVERSITY
are 2-2 Linkages Between Ecosystem
Weil-Being
^^ Security
^^^ • PERSONAL SAFETY
^^^ SECURE RESOURCE ACCESS
Mf. SECURITY FROM DISASTERS
TfcL
^^ Basic material
for good life
ADEQUATE LIVELIHOODS
SUFFCIENT NUTRITIOUS FOOD
SHELTER
ACCESS TO GOODS
i_/ v
Health
STRENGTH
FESJNG WELL
ACCESS TO CLEAN AIR
AND WATER
Good social relations
SOCIAL COHESION
MUTUAL RESPECT
ABILITY TO HELP OTHERS
Freedom
of choice
and action
OPPORTUNITY TO BE
ABLE TO ACHIEVE
WHAT AN INDIVIDUAL
VALUES DOING
AND BEING
Source; Millennium Ecosystem Assessment
Services Categories and Components of Human
15
16          The ecosystems of interest in this welfare risk and exposure assessment are impacted by
17   the effects of anthropogenic air pollution, which may alter the services provided by the
18   ecosystems in question.  For example, changes in forest conditions as a result of Os exposure
                                                     2-9

-------
 1   may affect supporting services such as net primary productivity; provisioning services such as
 2   timber production; regulating services such as climate regulation; provisioning services such as
 3   food; and cultural services such as recreation and ecotourism.
 4          Where possible, we developed linkages to ecosystem services from indicators of each
 5   effect identified in the ISA (U.S. EPA, 2013). These linkages were based on existing literature
 6   and models, focus on the services identified in the peer-reviewed literature, and are essential to
 7   any attempt to evaluate O3-induced changes on the quantity and/or quality of ecosystem services
 8   provided. According to EPA's Science Advisory Board Committee on Valuing the Protection of
 9   Ecological Systems and Services, these linkages are critical elements for determining the
10   valuation of benefits of EP A-regulated air pollutants (SAB CVPESS, 2009).
11          We have identified the primary ecosystem  service(s) potentially impacted by O3 for
12   major ecosystem types and components (i.e., terrestrial ecosystems, productivity) under
13   consideration in this risk and exposure assessment. The impacts associated with various
14   ecosystem services for each targeted effect are assessed in Chapters 5, 6, and 7 of this document
15   at a national scale and in the more refined case studies.
                                                     2-10

-------
 1                                        3   SCOPE

 2          This chapter provides an overview of the scope and key design elements of the welfare
 3   risk and exposure assessment. The design of this assessment began with a review of the risk and
 4   exposure assessments completed during the previous review of the National Ambient Air Quality
 5   Standard for Ozone (Os NAAQS) (U.S. EPA, 2007), with an emphasis on considering key
 6   limitations and sources of uncertainty recognized in that analysis.
 7          In October 2008, as an initial step in the current Os NAAQS review, the Environmental
 8   Protection Agency (EPA) invited outside experts, representing a broad range of expertise, to
 9   participate in a workshop with EPA staff to help inform EPA's plan for the review.  The
10   participants discussed key policy-relevant issues that would frame the review, as well as the most
11   relevant new science that would be available to inform our understanding of these issues. One
12   workshop session focused on planning for quantitative risk and exposure assessments, taking
13   into consideration what new research and/or improved methodologies would be available to
14   inform the design of a quantitative welfare risk and exposure risk assessment. Based in part on
15   the workshop discussions, EPA developed a draft Integrated Review Plan for the Ozone National
16   Ambient Air Quality Standards (IRP) (U.S. EPA, 2009) outlining the schedule, process, and key
17   policy-relevant questions that would frame this review. On November 13, 2009, EPA held a
18   consultation with the Clean Air Scientific Advisory  Committee (CASAC) on the draft IRP (74
19   FR 54562, October 22, 2009), which included opportunity for public comment. The final IRP
20   incorporated comments from CASAC (Samet, 2009) and the public on the draft plan, as well as
21   input from senior Agency managers. The final IRP included initial plans for the quantitative risk
22   and exposure assessments for both human health and welfare (U.S. EPA, 201 la, chapters 5 and
23   6).
24          As a next step in the design of these quantitative assessments, the Office of Air Quality
25   Planning and Standards (OAQPS) staff developed more detailed planning documents,  including
26   the following: Os National Ambient Air  Quality Standards: Scope and Methods Plan for Health
27   Risk and Exposure Assessment (Health Scope and Methods Plan; U.S. EPA, 201 Ib) and Os
28   National Ambient Air Quality Standards: Scope and Methods Plan for Welfare Risk and
29   Exposure Assessment (Welfare Scope and Methods Plan, U.S. EPA, 201 Ic).  These plans were
30   the subject of a May 19-20, 2011, consultation with  CASAC  (76 FR 23809, April 28, 2011).
                                                  3-1

-------
 1    Based on consideration of CAS AC (Samet, 2011) and public comments on these plans and
 2    information in the second draft Integrated Science Assessment (ISA), we modified the scope and
 3    design of the risk and exposure assessment and drafted a memo with updates to the information
 4    presented in these plans (Wegman, 2012). We further modified the scope in response to
 5    comments from CASAC on the first draft assessment (Frey and Samet, 2012a). These plans,
 6    together with the update memo and comments from CASAC and the public, provide the basis for
 7    the discussion of the scope of the risk and exposure assessment provided in this chapter.
 8          Section 3.1 of this chapter provides a brief overview of the risk and exposure assessment
 9    completed for the previous Os NAAQS review, including key limitations and uncertainties
10    associated with that analysis.  Section 3.2 provides a summary of the design of the current
11    exposure assessment, including the ecosystem services framework, assessments for biomass loss
12    and visible foliar injury. Section 3.3 provides an overview of the uncertainty and variability
13    assessments.

14     3.1  OVERVIEW OF RISK AND EXPOSURE ASSESSMENTS FROM PREVIOUS
15          REVIEW
16          The assessments conducted as part of the previous review focused on national-level  Os-
17    related impacts to sensitive vegetation and their associated ecosystems. The vegetation exposure
18    assessment was performed using an interpolation approach that included information from
19    ambient monitoring networks and results from air quality modeling. The vegetation risk
20    assessment included both tree and crop analyses. The tree risk analysis included three distinct
21    lines of evidence: (1) observations of visible foliar injury in  the field linked to monitored Os air
22    quality for the years 2001 - 2004; (2) estimates of tree seedling growth loss under then current
23    and alternative 63 exposure conditions; and (3) simulated mature tree growth reductions of
24    meeting alternative air quality standards on the predicted annual growth of mature trees from
25    three different species.  The crop risk analysis included estimates of crop yields under current
26    and alternative Os exposure conditions.  EPA analyzed the associated changes in economic value
27    upon meeting the levels of various alternative standards using an agricultural sector economic
28    model. Key elements and observations from these risk and exposure assessments are outlined in
29    the following sections.
                                                   3-2

-------
 1                3.1.1  Exposure Characterization
 2          In many rural and remote areas where sensitive species of vegetation can occur,
 3    monitoring coverage is limited. Thus, the 2007 Staff Paper (U.S. EPA, 2007) concluded that it
 4    was necessary to use an interpolation method to better characterize 63 concentrations over broad
 5    geographic areas and at the national scale.  Based on the significant difference in monitoring
 6    network density between the eastern and western U.S., the 2007 Staff Paper further concluded
 7    that it was appropriate to use separate interpolation techniques in these two regions.  EPA used
 8    monitoring data for the eastern interpolation, and in the western U.S., where rural monitoring is
 9    sparser, EPA used the Community Multi-scale Air Quality (CMAQ) model
10    (http://www.epa.gov/asmdnerl/CMAQ, Byun and Ching, 1999; Byun and Schere, 2006) to
11    develop  scaling factors to augment the monitor interpolation.
12          To evaluate changing vegetation exposures under selected air quality scenarios, EPA
13    conducted a number of analyses.  One analysis adjusted 2001 base year Os concentration
14    distributions using a rollback method (Rizzo, 2005, 2006) to reflect meeting the current and
15    alternative secondary standard options.  For the "just meet" and alternative 8-hour average
16    standard scenarios, EPA generated the associated maps of estimated 12-hour, W126 exposures.1
17          A second analysis in the 2007 Staff Paper identified the overlap between different forms
18    of the secondary standard.  The analysis was designed to evaluate the extent to which county -
19    level Os concentrations measured in terms  of various concentrations of the then current 8-hour
20    average form overlapped with  concentrations measured in terms of various concentrations of the
21    12-hour W126 cumulative, seasonal form.  This analysis found that the number of counties
22    meeting either one or both of the standard forms depended greatly on the level of the forms
23    selected as well as the air quality pattern that exists in a particular year or set of years. Thus, the
24    2007 Staff Paper indicated that it remained uncertain as to the extent to which air quality
25    improvements designed to reduce 8-hour average Os concentrations would also reduce Os
26    exposures measured by a seasonal, cumulative W126 index. The 2007 Staff Paper stated this
27    was an important consideration because: (1) the biological database stresses the importance of
28    cumulative, seasonal exposures in determining plant response; (2) plants have not been
29    specifically tested for the importance of daily maximum 8-hour 63 concentrations in relation to
                 See Section 4.3.1 for more information regarding the W126 O3 exposure metric.
                                                    3-3

-------
 1    plant response; and (3) the effects of attainment of an 8-hour standard in upwind urban areas on
 2    rural air quality distributions cannot be characterized with confidence because of the lack of
 3    monitoring data in rural and remote areas.
 4                3.1.2  Assessment of Risks to Vegetation
 5           The risk assessments in the previous review reflected the availability of several lines of
 6    evidence that provided a picture of the scope of Os-related vegetation risks for seedling, sapling
 7    and mature tree species growing in field settings and, indirectly, for forested ecosystems.  To
 8    assess visible foliar injury, the 2007 Staff Paper presented an assessment that combined USFS
 9    Forest Inventory and Analysis (FIA) biomonitoring site data with the county-level air quality
10    data for those counties containing the FIA biomonitoring sites.
11           EPA conducted separate assessments for seedlings and mature trees. To estimate growth
12    reductions in seedlings, EPA used concentration-response (C-R) functions developed from open-
13    top chamber (OTC) studies for biomass loss for available seedling tree species and from
14    information on tree growing regions derived from the U.S. Department of Agriculture's (USD A)
15    Atlas of United States Trees.  The C-R functions were then combined with proj ections of air
16    quality based on 2001 interpolated exposures. To estimate growth reductions in  mature trees,
17    EPA used a tree growth model (TREGRO) to evaluate the effect of changing Os  concentration
18    scenarios from just meeting alternative Os standards on the growth of mature trees.  TREGRO is
19    a process-based, individual tree growth simulation model (Weinstein et al, 1991) that is linked
20    with concurrent climate data to account for Os and climate/meteorology interactions on tree
21    growth. The model was run for a single western species (ponderosa pine) and two eastern
22    species (red maple and tulip poplar).  These three species were chosen based on the  availability
23    of species-specific parameterization in the model, their relative abundance in their respective
24    regions, and the importance of their associated ecosystem services.
25           To estimate yield loss in agricultural commodity, fruit and vegetable crops, EPA applied
26    information from the National Crop Loss Assessment Network (NCLAN) program and a  1996
27    California fruit and vegetable analysis to develop C-R functions. The crop risk assessment, like
28    the tree seedling assessment, combined C-R information on nine commodity crops and six fruit
29    and vegetable species with crop growing regions, and interpolated exposures during each  crop
30    growing season.

                                                    3-4

-------
 1          The 2007 Staff Paper also presented estimates of economic valuation for crops associated
 2    with the then current and alternative standards.  The Agriculture Simulation Model (AGSEVI)
 3    (Taylor, 1993) was used to calculate annual average changes in total undiscounted economic
 4    surplus for commodity crops and fruits and vegetables when then current and alternative
 5    standard levels were met.  The 2007 Staff Paper recognized that the modeled economic impacts
 6    from AGSIM had many associated uncertainties, which limited the usefulness of these estimates.

 7     3.2  OVERVIEW OF CURRENT ASSESSMENT PLAN
 8          Since the 2008 Os NAAQS review, new scientific information on the direct and indirect
 9    effects of 63 on  vegetation and ecosystems, respectively, has become available. With respect to
10    mature trees and forests, the information regarding 63 impacts to forest ecosystems has
11    continued to expand, including limited new evidence that implicates O3 as an indirect contributor
12    to decreases in stream flow resulting from direct impacts on whole tree-level water use.
13    Recently published results from the long-term FACE studies provide additional evidence
14    regarding chronic Os exposures in forests, including decreased tree heights, stem volumes
15    (Kubiske et al., 2006), seed weight and seed germination (Darbah et al., 2008, 2007); and
16    changes in tree community structure (Kubiske et al., 2007). In addition, a comparison, presented
17    in the ISA (Section 9.6.3), using recent data from Aspen FACE found that 63 effects on biomass
18    accumulation in aspen during the first seven years of the experiment closely agreed with the
19    exposure-response function based on data from earlier OTC experiments. In addition, recent
20    available data from annual field surveys conducted by the USFS to assess visible foliar injury to
21    selected tree species is available. In light of this more recent information, we are updating the
22    analysis that combines the USFS data with recent air quality data to determine the incidence of
23    visible foliar injury occurring across the U.S. at recent air quality concentrations and have
24    included new assessments that combine foliar injury information with soil moisture data.
25          One of the objectives of the risk assessment for a secondary NAAQS is to quantify the
26    risks to public welfare, including ecosystem services.  For example, the Risk and Exposure
27    Assessment for Review of the Secondary National Ambient Air Quality Standards for Oxides of
28    Nitrogen and Oxides of Sulfur (U.S. EPA, 2009) includes detailed discussions of how ecosystem
29    services and public welfare are related and how an ecosystem services  framework may be
30    employed to evaluate effects on welfare. To the extent applicable, we provide qualitative and/or
                                                   3-5

-------
 1    quantitative assessments of ecosystem services impacted by 63 to inform the current review. In
 2    Chapter 5 of this assessment, we identify and describe the ecosystem services associated with the
 3    ecological effects for which data and methods for incremental analysis of direct O3 are not yet
 4    available. For example, we overlay data on fire incidence, risk, and expenditures related to fires
 5    in California (CAL-FIRE with Os data to better characterize areas where Os may result in
 6    increased risks of fires.  Similarly, we also overlay data on bark beetle infestation with Os data.
 7    In chapters 6 and 7, we identify and describe the ecosystem services associated with the
 8    ecological effects for biomass loss and foliar injury, respectively, including national scale
 9    assessments and more refined case study areas.
10                 3.2.1  Air Quality Considerations
11           Air quality information and analyses are used to inform and support welfare-related
12    assessments.  The air quality information and analyses for this review build upon those in the
13    ISA and include: (1) summaries of recent ambient air quality data; (2) application of a
14    methodology to  extrapolate measured Os concentrations to areas without monitors, including
15    natural areas important to a welfare effects assessment such as national parks; and (3) adjustment
16    of air quality to simulate the distributions of 63 when just meeting existing or potential
17    alternative W126 secondary standards.  In this assessment, we use W126 as a shorthand for the
18    maximum consecutive 3-month, 12-hour daylight W126 index value. Consistent with the 2007
19    Staff Paper (U.S. EPA, 2007) and CASAC recommendation (Henderson et al., 2007), the air
20    quality analyses in this assessment focus on the W126 metric. We provide more information
21    regarding the  air quality analyses in Chapter 4.
22                   3.2.1.1  Recent Ambient Data
23           In addition to updating air quality summaries from the previous review, these air quality
24    analyses include summaries of the recent ambient measurements for 2006 to 2010 for the
25    existing form  of the standard and potential alternative form of secondary standard. The ambient
26    measurements are from monitor data from EPA's Air Quality System (AQS) database (which
27    includes National Park Service monitors) and the EPA's Clean Air Status and Trends Network
28    (CASTNET) network.  We provide more information regarding the air quality analyses in
29    section 4.3.2.
                                                    3-6

-------
 1                    3.2.1.2  National O3 Exposure Surfaces
 2           Since the previous review, the extent of monitoring coverage in non-urban areas has not
 3    significantly changed.  The vegetation exposure assessments rely on recent Os concentrations
 4    adjusted to simulate just meeting the existing standard and of potential alternative W126
 5    secondary standards. National-scale 63 surfaces are used as input to the national foliar injury
 6    assessments described in subsequent sections. To estimate 63 exposure in areas without
 7    monitors, particularly those gaps left by a sparse rural monitoring network in the western United
 8    States, we used a spatial interpolation technique, called Voronoi Neighbor Averaging (VNA),
 9    (Gold, 1997; Chen et al., 2004) to create an air quality surface for the contiguous United States.
10    We created annual W126 surfaces for each year between 2006 and 2010 and for a three year
11    average for 2006-2008 at a 12km grid resolution.  We provide more information regarding these
12    data in section 4.3.1.
13                    3.2.1.3  Simulation of Existing and Alternative Standards
14          To generate a national-scale spatial surface that simulates just attaining the existing
15    standard, a spatial surface of O3 for 2006-2008 was created using VNA and monitor
16    concentrations adjusted to reflect just meeting the existing standard. For potential alternative
17    secondary standards, we simulated just meeting W126 standard levels of 15 ppm-hrs, 11 ppm-
18    hrs, and 7 ppm-hrs at Os monitor locations,  assuming the monitors already met the existing
19    standard. We selected these standard levels  for analysis in this REA because CAS AC
20    recommended and supported a range of alternative W126 standard levels from 15 to 7 ppm-hrs
21    during the previous review.  These adjusted monitor values were then used to create a spatial
22    surface that provided W126 index values to areas without monitors. The adjusted surfaces are
23    used in several vegetation assessments, including the geographic analysis for fire risk and bark
24    beetle, the national and case study biomass loss assessments, and the park case studies for foliar
25    injury. Each of these surfaces represents the 3-year average W126 index values. We provide
26    more information  regarding these data in section 4.3.2.
                                                    3-7

-------
 1                3.2.2   Relative Tree Biomass Loss and Crop Yield Loss
 2                   3.2.2.1  National-Scale Assessment: Concentration-Response Functions
 3                           for Tree Seedlings and Crops
 4          In the 2007 Staff Paper, the EPA derived information on tree species growing regions
 5    from the USDA Atlas of United States Trees (Little,  1971). In this assessment, we use more
 6    recent information (2006-2008) from the USFS Forest Health Technology Enterprise Team
 7    (FHTET) to update growing ranges for the 12 tree species studied by National Health and
 8    Environmental Effects Research Laboratory, Western Ecology Division (NHEERL-WED). We
 9    combine the national O3 surface with seedling C-R functions for each of the tree species and
10    information on each tree species growing region to produce estimates of O3-induced seedling
11    biomass loss for each of the 12 tree species.  From this information, we generate GIS maps
12    depicting seedling biomass loss for each species for each air quality scenario. For crops, we
13    estimate yield loss for each of the 10 crop species from NCLAN. This analysis enabled direct
14    evaluation of estimated seedling biomass loss for trees and yield loss for crops expected to occur
15    under air quality exposure scenarios expressed in terms of recent air quality and, after simulation,
16    of just meeting the existing standard and potential alternative secondary standards. In addition,
17    this assessment can be used to determine the W126 benchmark values associated with 1 to 2
18    percent seedling biomass  loss for trees and 5 percent yield loss for crops. For biomass loss,
19    CASAC recommended that EPA should consider options for W126 standard levels based on
20    factors including a predicted  1 to 2 percent biomass loss for trees and a predicted 5 percent loss
21    of crop yield. Small losses for trees on a yearly basis compound over time and can result in
22    substantial biomass losses over the decades-long lifespan of a tree (Frey and Samet, 2012b).
23                   3.2.2.2  National Scale Assessment: National weighted RBL and Class I
24                           Areas
25          To assess overall ecosystem-level effects from biomass loss, we used FHTET data for
26    modeled predictions of stand density and basal area.  The resolution of the FHTET data is 1,000
27    square meter grids, and we summed these data into the larger CMAQ grid cells (12 km x 12 km).
28    For the individual species analyses, these data were used only as a predictor of presence or
29    absence. In the ecosystem-level analysis, these data were used to scale the biomass loss by the
30    proportion of total basal area for each species. We combined the RBL values for 12 tree species

                                                  3-8

-------
 1    into a weighted RBL rate and considered the weighted value in relation to proportion of basal
 2    area covered (as measured by proportion of geographic area with available data on species).  A
 3    weighted RBL value is a relatively straightforward metric to attempt to understand the potential
 4    ecological effect on some ecosystem services. We provide more information regarding the
 5    individual species analysis in section 6.2.1.3 and the combined analysis in 6.2.1.4.
 6          We also calculated an average weighted biomass loss for 12 tree species occurring in
 7    federally designated Class I areas using USFS estimates of the proportion of total basal area from
 8    FHTET. Out of 156 Class I areas nation-wide,  119 Class I areas had tree data available for this
 9    analysis.  This analysis was conducted for air quality exposure scenarios expressed in terms of
10    recent air quality (2006-2008) and after simulation of just meeting the existing standard and
11    potential  alternative secondary standards. We provide more information regarding this analysis
12    in section 6.8.1.1.
13                    3.2.2.3  National-Scale Assessment: Ecosystem Services
14          The national-level ecosystem services quantified in this review associated with biomass
15    and yield loss include provisioning services (e.g., timber and crops) and regulating services (e.g.,
16    carbon sequestration).  Where information is available, we describe the impacts on other
17    ecosystem services such as impacts on biodiversity, biological community composition, health of
18    forest ecosystems, aesthetic values of trees and plants, and the nutritive quality of forage crops.
19    We also describe the cultural ecosystem services associated with non-timber forest products. In
20    addition,  there is new preliminary evidence that O?, adversely affects the ability of pollinators to
21    find their targets, which could have broad implications for agriculture, horticulture, and forestry.
22          We use the Forest and Agricultural Sector Optimization Model Greenhouse Gas version
23    (FASOMGHG) model (Adams et al., 2005) to estimate Os impacts on the agriculture and
24    forestry sectors and quantify how O?, exposure to vegetation affects the provision of timber and
25    crops and carbon sequestration. F ASOM has been used recently in many evaluations of effects of
26    climate change on the timber and agriculture market sectors, in part because it accounts for the
27    tradeoffs between land use for forestry and agriculture.  Specifically, FASOM is a dynamic, non-
28    linear programming model designed for use by the EPA to evaluate welfare benefits and market
29    effects of Os-induced biomass loss in trees and of carbon sequestration in trees, understory,
30    forest floor, wood products and landfills that would occur under different agricultural and

                                                    3-9

-------
 1    forestry scenarios. Using this model, we calculate the economic impacts of yield changes
 2    between recent ambient 63 conditions and after simulating just meeting the existing 75 ppb
 3    standard and alternative W126 standards.
 4                    3.2.2.4  Case Study Areas: Five Urban Areas
 5           In selecting urban case study areas for more in-depth analysis of the ecosystem services
 6    associated with urban tree biomass loss, EPA relied on several criteria:
 7           •   Areas expected to have elevated W126 index values where ecological effects might
 8              be expected to occur.

 9           •   Occurrence of O3 sensitive tree species and/or species for which O3 concentration-
10              response curves have been generated.

11           •   Availability of vegetation information in the case study area.
12           •   Geographic coverage representing a cross section of the nation, including urban and
13              natural settings.

14           We use the i-Tree model to assess effects on regulating ecosystem services provided by
15    urban forests, including pollution removal and carbon storage and sequestration for the case
16    study areas.  The i-Tree model is a publicly available, peer-reviewed software suite developed by
17    the USFS and its partners to assess the ecosystem service impacts of urban forestry (available
18    here: http://www.itreetools.org/).  We collaborated with the USFS to vary the tree growth metric
19    in the model, which allows us to assess the effects of O3  exposure on the  ability of the forests in
20    the selected case study area to provide the services enumerated by the model. Specifically, we
21    estimate impacts on vegetation in Atlanta, Baltimore, Syracuse, the Chicago region, and the
22    urban areas of Tennessee. We present results for model runs representing recent ambient Os
23    conditions, just meeting the existing 75 ppb standard, and just meeting alternative W126
24    standards.
                                                    5-10

-------
 1                3.2.3  Visible Foliar Injury
 2                    3.2.3.1  National Analysis of Visible Foliar Injury
 3           To assess visible foliar injury (hereafter referred to as foliar injury) at a national scale, we
 4    compared data from the USFS Forest Health Monitoring Network (USFS, 2011) with O3
 5    exposure estimates and soil moisture data for 2006-2010.  For estimates of short-term soil
 6    moisture in the contiguous U.S., we use NOAA's Palmer Z drought index (NCDC, 2012b).
 7    Foliar injury sampling data were not available for several western states (Montana, Idaho,
 8    Wyoming, Nevada, Utah, Colorado, Arizona, New Mexico, Oklahoma, and portions of Texas).
 9    This analysis provides estimates of the presence and absence of foliar injury for each of the 5
10    years by soil moisture category, which provides insight into whether drought provides protection
11    from foliar injury. In addition, we estimated foliar injury by soil moisture category for elevated
12    foliar injury. Using this analysis, we derived multiple W126 benchmark s for evaluating foliar
13    injury at national parks in a screening-level assessment and three case studies.
14                    3.2.3.2  National Scale Screening-level Assessment of Visible Foliar Injury
15                            in 214 National Parks
16           A study by Kohut (2007) assessed the risk of O3-induced visible foliar injury on O3-
17    sensitive vegetation in 244 parks managed by the National Park Service (NPS). We modified this
18    screening-level assessment to use more recent O3 exposure and soil moisture data and to
19    incorporate benchmarks derived from the national-scale foliar injury analysis (described above in
20    section 3.2.3.1). Specifically, we use O3 monitoring data to create spatial surfaces of O3 exposure
21    and short-term soil moisture data (Palmer Z) (NCDC, 2012b) for 2006 to 2010. These data
22    reflect the contiguous U.S. only, which is a key reason why this assessment includes fewer parks
23    than Kohut (2007). Overall, the screening-level assessment includes 42 parks with O3 monitors
24    and 214 parks with O3 exposure estimated from the interpolated O3 surface. We combine these
25    data with lists from the NPS of the parks containing O3-sensitive vegetation species (NPS, 2003,
26    2006). Consistent with Kohut (2007), we consider the results for these parks without identified
27    species as potential until sensitive species are identified in field surveys at these parks.
28           Using the results of the national-scale foliar injury analysis, we derived six W126
29    benchmark scenarios for evaluating foliar injury risk at parks in this screening-level assessment.
30    One scenario reflects O3 exposure only, four scenarios reflect O3 exposure and soil moisture
                                                   3-11

-------
 1    jointly for different percentages of biosites with injury, and one scenario reflects 63 exposure
 2    and soil moisture jointly for elevated injury. For each of these scenarios, we identify the number
 3    of parks that exceed the benchmark criteria in each year.
 4                    3.2.3.3  National Scale Assessment: Ecosystem Services
 5           We use GIS mapping developed for the ecological effects analysis to illustrate where
 6    foliar injury may be occurring, and we cross reference those areas to national statistics for
 7    recreational use available through the National  Survey of Fishing, Hunting, and Wildlife-
 8    Associated Recreation (U.S. DOT, 2011) and the National  Survey on Recreation and the
 9    Environment (USDA, 2002).  We also scale the resulting estimates of cultural service provision
10    to the current population and values assigned using existing meta-data on willingness-to-pay
11    from the Recreation Values Database.2 We understand that these estimates are limited to current
12    levels of service provision and provide a snapshot of the overall magnitude of services
13    potentially affected by O3 exposure. Currently, estimates of service loss from recent O3
14    exposure is beyond the available data and resources, as is the calculation of changes in
15    ecosystem services that might result from meeting existing and alternative O3 standards.
16    However, the current losses in service from O3 exposure are embedded in estimates of the
17    current level of services.
18                    3.2.3.4  Case Study Analysis:  Three National Parks
19           In selecting case study areas for more in-depth analysis of the ecosystem services
20    associated with visible foliar injury, EPA relied on several criteria:
21           •  Areas expected to have elevated W126 index values where ecological effects might
22              be expected to occur.

23           •  Availability of vegetation mapping,  including estimates of species cover.
24           •  Geographic coverage representing a cross section of the nation, including urban and
25              natural settings.

26           •  Occurrence of O3 sensitive species and/or species for which O3 concentration-
27              response curves have been generated.
              Available at:  http://recvaluation.forestry.oregonstate.edu/.
                                                    3-12

-------
 1            We selected Great Smoky Mountains National Park, Rocky Mountain National Park,
 2    and Sequoia/Kings National Park. All three of these park units are in areas with elevated
 3    ambient W126 index values, have vegetation maps, and have species that are considered O3
 4    sensitive. We considered including Acadia National Park, but we determined it did not fit our
 5    selection criteria for Os exposure. Using GIS, we compare the NFS vegetation maps to the
 6    national Os surface to illustrate where foliar injury may be occurring, particularly with respect to
 7    park amenities such as trails. Ecological metrics quantified for each park include:
 8           •   Percent of vegetation cover affected by foliar injury.
 9           •   Percent of trails affected by foliar injury.
10           In national parks, foliar injury affects primarily cultural values that include existence,
11    bequest and recreational values. In addition, we describe the other nonuse values associated with
12    national parks including existence and bequest values. We also provide park-specific statistics
13    for recreational use available and estimates of service provision values using existing meta-data
14    on willingness-to-pay from Kaval and Loomis (2003). We understand that these estimates  are
15    limited to current levels of service provision. Estimates of service loss due to Os  exposure  are
16    beyond the available data and/or resources for many if not all ecosystem services listed above.

17      3.3   UNCERTAINTY AND VARIABILITY
18           An important issue associated with any ecological risk assessment is the characterization
19    of uncertainty and variability.  Variability refers to the heterogeneity in a variable of interest that
20    is inherent and cannot be reduced through further research. For example, there may be
21    variability among C-R functions describing the relationship between 63 and vegetation injury
22    across selected study areas.  This variability may be due to differences in ecosystems (e.g.,
23    species diversity, habitat heterogeneity, and rainfall), concentrations and distributions of Os
24    and/or co-pollutants, and/or other factors that vary either within or across ecosystems.
25           Uncertainty refers to the lack of knowledge regarding both the actual values of model
26    input variables (parameter uncertainty) and the physical systems or relationships (model
27    uncertainty - e.g., the shapes of concentration-response functions). In any risk assessment,
28    uncertainty is, ideally, reduced to the maximum extent possible, through improved measurement
29    of key parameters and ongoing model refinement.  However, significant uncertainty often
                                                    3-13

-------
 1    remains, and emphasis is then placed on characterizing the nature of that uncertainty and its
 2    impact on risk estimates. The characterization of uncertainty can include both qualitative and
 3    quantitative analyses, the latter requiring more detailed information and, often, the application of
 4    sophisticated analytical techniques. Sources of variability that are not fully reflected in the risk
 5    assessment can consequently introduce uncertainty into the analysis.
 6           The goal in designing a quantitative risk assessment is to reduce uncertainty to the extent
 7    possible and to incorporate the sources of variability into the analysis approach to insure that the
 8    risk estimates are representative of the actual response of an ecosystem (including the
 9    distribution  of that adverse response across the ecosystem). An additional aspect of variability
10    that is pertinent to this risk assessment is the degree to which the set of selected case study areas
11    provide coverage for the range of Os-related ecological risk across the U.S.
12           Recent guidance from the World Health Organization (WHO, 2008) presents a four-
13    tiered approach for characterizing uncertainty. With this four-tiered approach, the WHO
14    framework provides a means for systematically linking the characterization of uncertainty to the
15    sophistication of the underlying risk assessment, where the decision to proceed to the next tier is
16    based on the outcome of the previous tier's assessment. Ultimately,  the decision as to which tier
17    of uncertainty characterization to include in a risk assessment will depend both on the overall
18    sophistication of the risk assessment and the availability of information for characterizing the
19    various sources of uncertainty. We used the WHO guidance as a framework for developing the
20    approach used for characterizing uncertainty in this assessment. The four tiers described in the
21    WHO guidance include:
22           •  Tier 0: recommended for routine screening assessments, uses default uncertainty
23              factors (rather than developing site-specific uncertainty characterizations);

24           •  Tier 1: the lowest level of site-specific uncertainty characterization, involves
25              qualitative characterization of sources of uncertainty (e.g., a qualitative assessment of
26              the general magnitude and direction of the effect on risk results);

27           •  Tier 2: site-specific deterministic quantitative analysis involving sensitivity analysis,
28              interval-based assessment, and possibly probability bounded (high-and low-end)
29              assessment; and

                                                     3-14

-------
 1           •   Tier 3: uses probabilistic methods to characterize the effects on risk estimates of
 2              sources of uncertainty, individually and combined.

 3           In this assessment, we applied a variety of quantitative (WHO Tier 2) and qualitative
 4    (WHO Tierl) analyses to address uncertainty and variability in this assessment of Os-related
 5    ecological risks. In general, we attempted to quantify uncertainty and variability where we had
 6    sufficient data to do so and addressed these aspects qualitatively where we did not have data.
 7    Two analyses include quantitative assessments of uncertainty and variability. For the analysis of
 8    the alternative percentages of biomass and yield loss, we plotted the C-R relationship for 54 crop
 9    studies and 52 tree seedling studies to estimate the differences in within-species variability. We
10    also qualitatively compared the uncertainty in the relationship between C-R functions for tree
11    seedlings and the effects on adult trees. For the screening-level assessment of foliar injury, we
12    conducted several quantitative sensitivity analyses, including six scenarios reflecting different
13    degrees of injury and consideration of soil moisture, three approaches for estimating Os exposure
14    at monitored parks, three durations for soil moisture data, and two time periods evaluating
15    different years of analysis. We provide detailed tables characterizing the uncertainty inherent in
16    the risk and exposure analyses at the end of Chapters 4, 5, 6, and 7.
                                                     5-15

-------
 1                         4   AIR QUALITY CONSIDERATIONS

 2    4.1   INTRODUCTION
 3           Air quality information is used to assess exposures and ecological risks for national-scale
 4    air quality surfaces generated to estimate 2006-2008l average concentrations based on the W126
 5    exposure metric, which is defined later in this chapter.  These national-scale air quality surfaces
 6    are generated for five air quality scenarios by the methodology summarized in Section 4.3.1 and
 7    4.3.4 below.  The five scenarios are for recent air quality,  air quality adjusted  to just meet the
 8    current standard, and air quality further adjusted to just meet three different W126 index values:
 9    15 ppm-hrs,  11  ppm-hrs, and  7 ppm-hrs.  Additional national-scale air  quality  surfaces are
10    generated using  observed W126 concentrations  for individual  years from 2006-2010.  This
11    chapter describes the air quality information used in these analyses, providing an overview of
12    monitoring data and air quality (section 4.2), and an overview of air quality  inputs to the welfare
13    risk and exposure assessments (section 4.3).

14    4.2   OVERVIEW OF O3 MONITORING AND AIR QUALITY
15           To monitor compliance with the NAAQS, state and local environmental agencies operate
16    Os monitoring sites at various locations, depending on the population of the area and typical peak
17    O3 concentrations (US EPA, 2013, sections 3.5.6.1, 3.7.4).  In 2010, there were over 1,300 state,
18    local, and tribal 63 monitors reporting concentrations to EPA (US EPA, 2012a, Figures 3-21 and
19    3-22).  The minimum number of Os monitors required in a Metropolitan Statistical Area (MSA)
20    ranges  from zero, for areas with a population under 350,000 and with no recent  history of an 63
21    design  value greater  than 85% of the NAAQS, to four, for areas with a population greater than
22    10 million and an Oj design value greater than  85% of the NAAQS.2  In areas for which Oj
23    monitors are required, at least  one site must be designed to record the maximum concentration
24    for that particular metropolitan  area. Since Os concentrations are usually significantly lower in
      1 The focus was placed on the years of 2006-2008 based on availability of data during that time period.
      2The existing monitoring network requirements (40 CFR Part 58, Appendix D) have an urban focus and do not
      address siting in non-urban (rural) areas. States may operate ozone monitors in non-urban (rural) areas to meet other
      objectives (e.g., support for research studies of atmospheric chemistry or ecosystem impacts).
                                                     4-1

-------
 1   the colder months  of the year,  63 is required to be monitored  only  during the required 63
 2   monitoring season, which varies by state (US EPA, 2012a, section 3.5.6 and Figure 3-20).3
 3          While the existing U.S. O?, monitoring network has a largely urban focus,  to address
 4   ecosystem impacts of 63 such as biomass loss and foliar injury, it is equally important to focus
 5   on Os monitoring in rural areas. Figure 4-1  shows the location of all U.S. Os monitors operating
 6   during the 2006-2010 period. The gray dots which make up  over 80%  of the Os monitoring
 7   network are "State  and Local Monitoring Stations"  (SLAMS) monitors which  are largely
 8   operated by state and local governments to meet regulatory requirements and provide air quality
 9   information to public health agencies, and thus are largely focused on urban areas.  The blue dots
10   highlight two important subsets of the  SLAMS network: "National Core" (NCore) multipollutant
11   monitoring sites, and the "Photochemical Assessment Monitoring Stations" (PAMS) network.
12          The green dots represent the Clean Air Status and Trends Network (CASTNET) monitors
13   which are focused on rural areas.  There were about 80  CASTNET sites operating in  2010, with
14   sites in the Eastern U.S. being operated by EPA and sites in the Western U.S. being operated by
15   the National Park Service (NPS).  Finally, the black dots represent "Special Purpose Monitoring
16   Stations" (SPMS), which include about 20 rural monitors as part of the "Portable O3 Monitoring
17   System" (POMS) network operated by the  NPS.  Between the CASTNET, NCore, and POMS
18   networks, there were about 120 rural Os monitoring sites in the U.S. in 2010.
19
     3 Some States and Territories are required to operate ozone monitors year-round, including Arizona, California,
     Hawaii, Louisiana, Nevada, New Mexico, Puerto Rico, Texas, American Samoa, Guam and the Virgin Islands.
                                                   4-2

-------
 i
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
                                                                          *  SLAMS
                                                                          •  CASTNET
                                                                          •  NCORE/PAMS
                                                                          •  SPMS/OTHER
Figure 4-1   Map of U.S. ambient O3 monitoring sites in operation during the 2006-2010

       To determine whether or not the NAAQS have been met at an ambient Os monitoring
site, a statistic commonly referred to as a "design value" must  be  calculated based on  3
consecutive years of data collected from that site.  The form of the existing 63 NAAQS design
value statistic is the  3-year average of the annual 4th  highest daily maximum 8-hour 63
concentration in parts per billion (ppb), with decimal digits truncated. The existing primary and
secondary O3 NAAQS are met at an ambient monitoring site when the design value is less than
or equal to 75 ppb.4 Figure 4-2 shows the design values for the existing 8-hour Oj NAAQS for
all regulatory monitoring sites in the U.S. for the 2006-2008 period.  Monitors shown as red dots
had design values above the existing 63 NAAQS of 75 ppb in 2006-2008.
     4For more details on the data handling procedures used to calculate design values for the existing O3 NAAQS, see
     40 CFR Part 50, Appendix P.
                                                   4-3

-------
1
2   Figure 4-2
3
                   Map of monitored 8-hour Os design values for the 2006-2008 period
 4   4.3   OVERVIEW OF AIR QUALITY INPUTS TO RISK AND EXPOSURE
 5         ASSESSMENTS
 6          In this section, we summarize the air quality inputs for the welfare risk and exposure
 7   assessments, and discuss the methodology used to adjust air quality to meet the existing standard
 8   and potential alternative standards.  These steps are summarized in the flowchart in Figure 4-3
 9   and discussed in more detail in this section.
10          Section 4.3.1 describes the W126 metric upon which the potential alternative standards
11   are based.  Section 4.3.2 describes the ambient air quality  monitoring data used in the welfare
12   risk  and exposure assessments.   Section 4.3.3  describes the procedure used to  generate the
                                                   4-4

-------
 1
 2
 3
 4
national-scale air quality surfaces upon which several of the welfare risk and exposure analyses
are based, with further details in Appendix 4a.  Finally, section 4.4.4  summarizes the method
used to adjust observed air quality concentrations to just meet the existing standard and potential
alternative standards, and discusses the resulting distributions of adjusted W126 concentrations.
                                                                                    Check if a
                                                                                  Monitors meet
                                                                                    targeted
                                                                                  W126 alternative
                                                                                    standard

Roll back to targeted
W126 alternative
standard


1
Aggregate hourly O3
to W126 at each
monitor



Hourly rolled backO3
Values at all monitors
In region

 7    Figure 4-3
              Flowchart of air quality data processing for different parts of the welfare
              risk and exposure assessments.
10    4.3.1  Air Quality Metrics
11          EPA focused the analyses in the welfare risk and exposure assessments on the W126 Os
12    exposure metric. The W126 metric is a seasonal aggregate of hourly Os concentrations, designed
13    to measure the cumulative effects of 63 exposure on vulnerable plant and tree species, with units
14    in parts per million-hours (ppm-hrs).  The metric uses a logistic weighting function to place less
15    emphasis on exposure to low hourly  63 concentrations and more emphasis on exposure to high
16    hourly Oj concentrations (Lefohn et al,  1988).
                                                    4-5

-------
 1          The first step in calculating W126 concentrations was to sum the weighted hourly 63
 2    concentrations within each month, resulting in monthly index values.  Since most plant and tree
 3    species are not photochemically active during nighttime hours, only O3 concentrations observed
 4    during  daytime hours  (defined as 8:00  AM to 8:00  PM local time) were included in the
 5    summations.  The monthly W126 index values were calculated from the hourly Os concentration
 6    data as follows:
                                          N   19
                      Monthly W126 = Y Y	—,	
                              y          /-l/-l\ + 4403 * exp (-126 * Cdh)
                                         d=l h=8              r V         atlJ
 1    where TV is the number of days in the month,
 8          d is the day of the month (d = 1, 2,  ..., N),
 9          his the hour of the day (h = 0, 1, ..., 23),
10          CM is the hourly 63 concentration observed on day d, hour h, in parts per million.
11          Next, the monthly W126 index values were adjusted for missing data.  lfNm is defined as
12    the number of daytime Os  concentrations observed during month m  (i.e. the number of terms in
13    the monthly index summation), then the monthly data completeness rate is Vm = Nm /12 * N.
14    The monthly index values were adjusted by dividing them by their respective Vm. Monthly index
15    values were not computed if the monthly data completeness rate was less than 75% (Vm < 0.75).
16          Finally, the annual W126 index values were computed as the maximum sum of their
17    respective  adjusted monthly index values  occurring in  three consecutive months (i.e., January-
18    March, February-April, etc.).  Three-month periods spanning across two years (i.e., November-
19    January, December-February) were not considered, because the seasonal nature of O3 makes it
20    unlikely for the maximum values to occur  at that time of year. The annual W126 concentrations
21    were considered valid if the data met the annual data completeness requirements for the existing
22    standard.  Three-year W126 index values  are calculated by taking the average of annual W126
23    index values in the same three-month period in three consecutive years.5
      5 W126 calculations are slightly modified in the case of the model adjustment scenarios described in Section 4.3.4.
      When calculating W126 for the model adjustment cases, we first found the three-year average of each three-month
      period, and then selected the three-month period with the highest three-year average using the same three-month
      period for each of the three years.
                                                   4-6

-------
 1     4.3.2  Ambient Air Quality Measurements
 2           Air quality monitoring data from  1,468 U.S. ambient Os monitoring sites were retrieved
 3    for use in the risk and exposure assessments. The initial dataset was the same as the one used for
 4    the Health REA, which consisted of hourly Os concentrations in ppb collected between 1/1/2006
 5    and 12/31/2010 from these monitors.  Data for nearly 1,400 of these monitors were extracted
 6    from EPA's Air Quality  System (AQS) database6, while the remaining data came from EPA's
 7    Clean Air Status  and Trends Network (CASTNET)  database which consists of primarily rural
 8    monitoring sites.  While the CASTNET monitors did not begin reporting regulatory data to AQS
 9    until 2011, it is generally agreed that data collected from these monitors prior to  2011 is  of
10    comparable quality to the data reported to AQS.
11           Observations flagged in AQS as  having  been  affected by exceptional events were
12    included the initial  dataset, but were not used in design value calculations in accordance with
13    EPA's exceptional events policy.  Missing data intervals of 1  or 2  hours  in the initial dataset
14    were filled in using linear interpolation. These  short gaps often occur at regular intervals in the
15    ambient data due to an EPA requirement for monitoring agencies  to perform routine quality
16    control checks on their Os monitors.  Quality control checks  are typically performed between
17    midnight and 6:00 AM when Os concentrations are  low.  Missing data intervals of 3 hours  or
18    more were not replaced, and interpolated data values were not used in design values calculations.
19         Annual W126  concentrations were calculated from the ambient data for each year in the
20    2006-2010 period, as well  as 3-year averages of the 2006-2008 annual W126 concentrations.
21    Figure 4-4 shows the 2006-2008 average W126  concentrations  in ppm-hrs at all monitoring sites
22    in the contiguous  U.S.  Monitors outside of the contiguous U.S. were not included in the welfare
23    analyses since they fell outside of the CMAQ 12 km modeling domain, and were already well
24    below the existing and potential alternative standards.
25
      6 EPA's Air Quality System (AQS) database is a national repository for many types of air quality and related
      monitoring data.  AQS contains monitoring data for the six criteria pollutants dating back to the 1970's, as well as
      more recent additions such as PM2 5 speciation, air toxics, and meteorology data. At present, AQS receives hourly
      O3 monitoring data collected from nearly  1,400 monitors operated by over 100 state, local, and tribal air quality
      monitoring agencies.
                                                     4-7

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
  0             5            10            15            20             25
Figure 4-4   Monitored 2006-2008 average W126 concentrations in ppm-hrs
30+
 4.3.3   National-scale Air Quality Surfaces for Recent Air Quality
       In addition to  ambient  monitoring data, the welfare risk  and exposure  assessments
analyzed national-scale air quality surfaces. For the biomass loss analyses presented in Chapter
6, a national-scale surface was  generated from  the  monitored  2006-2008  average  W126
concentrations using the Voronoi Neighbor Averaging (VNA) technique (Gold, 1997; Chen et al,
2004) (Figure 4-5). For the foliar injury analysis presented in Chapter 7, national-scale surfaces
were generated from the monitored annual W126 concentrations for individual years 2006-2010,
also using VNA.  Maps of the annual W126 air quality surfaces for 2006-2010 are included in
Appendix 4a.
                                                   4-8

-------
 2   Figure 4-5    National surface of observed 2006-2008 average W126 concentrations, in
 3                 ppm-hrs
 5          In the 1st draft of the REA, the national-scale  air quality surfaces  were created  by
 6   "fusing" monitored 2006-2008 average W126 concentrations with annual W126 concentrations
 7   from  a 2007 CMAQ  model  simulation,  using  the  enhanced Voronoi Neighbor Averaging
 8   (eVNA) technique (Timin et al., 2010). The resulting surfaces contained estimates of the 2006-
 9   2008  average annual W126 concentrations  at a 12km grid cell resolution in the contiguous U.S.
10   modeling domain.   In this  draft,  the  air quality surfaces of the  2006-2008 average  W126
11   concentrations are based  solely on  monitored W126 concentrations and  do not include CMAQ
12   model predictions. The reason for this change from the first draft REA is discussed below.
13          In addition to the  VNA methodology, two alternative methods for creating the national -
14   scale  air quality surfaces were also considered: eVNA and Downscaler (Berrocal et al, 2012;
15   used in the health REA).  Both the eVNA and Downscaler methods were tested using updated
                                                  4-9

-------
 1    2007 12km CMAQ modeling7 that is described in detail in Appendix 4b of the Health REA.
 2    While each of the three methods had its own advantages and disadvantages, the VNA method
 3    was ultimately selected because large differences between the modeled W126 surface and the
 4    monitored W126 concentrations8 made the two "data fusion" methods more uncertain in some
 5    instances, whereas VNA did not suffer from this problem since it is based solely on monitored
 6    values. Technical justification for the change from eVNA to VNA, including a cross-validation
 7    analysis, and comparisons between the resulting air quality surfaces for these three methods, can
 8    be found in Appendix 4a.
 9     4.3.4  Air Quality Adjustments to Meet Existing Primary and Potential Alternative
10           Secondary O3 Standards
11          In addition to observed W126 levels, the risk and exposure assessments also consider the
12    relative change in  risk and exposure after adjusting air quality  to just meet the  existing O3
13    standard of 75 ppb, and further adjusting air quality to just meet  possible alternative standards
14    with forms based on the W126 metric and levels of 15 ppm-hrs, 11  ppm-hrs, and 7 ppm-hrs.  The
15    sections below summarize the methodology used to adjust observed air quality concentrations to
16    just meet the existing  standard and potential alternative standards,  and discuss the  resulting
17    adjusted distributions of W126 concentrations.  More  details  on these  inputs are  provided in
18    Appendix 4A.
19              4.3.4.1  Adjustment Methods
20          The model-based HDDM  O3 adjustment approach  used for  this analysis is the same
21    general methodology developed for evaluating air  quality distributions that  could  occur if
22    meeting various alternate levels of the primary standard.  This methodology is described in detail
23    in Chapter 4 and Appendix 4d of the health REA. There are a few key differences between the
24    adjustments made in the health REA and those performed here.  First, the adjustments in health
25    REA focused on 15 urban case study areas while those used in the welfare  REA cover all
26    monitoring sites across the US. In the  health REA, a uniform reduction of U.S. anthropogenic
      7 The updated CMAQ modeling used wildfire emissions based on a multi-year average instead of 2007-specific
      wildfires.
      8 The 2007 CMAQ simulation over-predicted W126 values by an average of 4 ppm-hrs in monitored locations. A
      more in depth model evaluation of CMAQ W126 values is provided in Appendix 4b.
                                                   4-10

-------
 1    emissions was  applied to all sites  within  an urban  area.   By  applying  equal proportional
 2    decreases in emissions throughout the contiguous U.S., we were able to estimate how hourly 63
 3    concentrations would respond to changes in  ambient NOx  and VOC  concentrations without
 4    simulating  a specific control  strategy.   Note that the HDDM-adjustment approach  was not
 5    designed to produce an optimal control scenario but instead aimed to characterize a potential
 6    distribution of air quality across a region when  all monitors are meeting the existing standard and
 7    potential alternative standards. In this analysis,  we recognize the regional nature of W126 values,
 8    thus  we determined the requisite level of U.S. emissions  reduction independently  for nine
 9    distinct  regions  of the contiguous  U.S. (Figure 4-6) based on the  National Oceanic  and
10    Atmospheric Administration  (NOAA)  climate  regions  (Karl  and Koss,  1984).    NOAA
11    characterizes each  region as being "climatically  consistent" and  routinely uses these regions to
12    describe regional climate trends.  These regions were deemed an appropriate delineation for this
13    analysis since geographic  patterns  of both Os  and plant species  are driven by climatic features
14    such as  temperature and precipitation.  Analogous to the procedure used in the health REA for
15    the urban case study areas, a single NOx emissions perturbation was used to adjust ambient air
16    quality data at all  Os monitoring  sites  for each region and  standard.  The  magnitude of this
17    emissions perturbation was  determined independently  for each region  and  standard  by
18    determining the smallest perturbation necessary to bring all sites into attainment of the existing
19    standard or the potential alternative standards.  By evaluating the effect of U.S. anthropogenic
20    emissions reductions on all monitoring sites within a region, our analysis incorporates the effects
21    of emissions reductions on both local Os production and regional transport.  Since each region is
22    treated independently, the effects of the emissions reductions required to bring a particular region
23    down to the targeted  standard levels  do not affect  other regions which require less drastic
24    emissions reductions.  In portions of the country  with lower W126 values than nearby locations,
25    the emissions perturbation determined by the "controlling"  monitor in the region may be larger
26    than the emissions  reductions that would be required if the nine climate regions were replaced by
27    many smaller localized areas.  However, by considering larger regions, we are able to account
28    for the fact that nearby emissions reductions will  affect O3 monitors already meeting the targeted
29    standard level.9
      9 Another proponent for the use of large regions is that the air quality adjustments are computationally intensive, and
                                                    4-11

-------
 1          A second distinction between the welfare air quality adjustments and those in the health
 2    REA is  that only U.S. anthropogenic NOx emissions reductions were applied in the HDDM
 3    adjustment methodology for the welfare assessment (i.e. changes in U.S.  anthropogenic VOC
 4    emissions changes were not considered). NOx emissions reductions are believed to be the most
 5    effective method for reducing Os regionally, since most areas outside of urban population centers
 6    tend to be NOx limited in terms of Os formation.
 7          Finally, it should be noted that this analysis includes adjustment to  four standard  levels:
 8    1) the existing standard of 75 ppb based on the 3-year average of the 4th  highest 8-hour daily
 9    maximum Os concentration, 2) a W126-based standard with a level of 15 ppm-hrs, 3) a W126-
10    based standard with a level of 11 ppm-hrs, and 4) a W126-based standard with a level of 7 ppm-
11    hrs.  The  2006-2008 average W126 concentrations  and 4th highest 8-hour daily maximum Oj,
12    concentrations were calculated for every monitor in  each adjusted air quality scenario. For the
13    analysis of each of the W126 standards, we started with W126 air quality values resulting from
14    emission reductions required to just meet the existing standard at all monitors in the region, and
15    only applied the HDDM adjustments to those regions where all sites were not already below the
16    targeted W126 standard. In some cases, the emissions reductions necessary to meet the existing
17    standard resulted in W126 values below the level of one or more potential  alternative standards
18    at all monitors within  the region. In those cases, there is no change in air quality between the
19    scenario meeting the  existing  standard  and  the  scenario meeting the  potential  alternative
20    standard.
21          National-scale  spatial surfaces that represent  2006-2008 W126 concentrations when just
22    meeting the existing standard and the potential alternate standards (at the highest monitor in the
23    region) were then created using the monitor values from the appropriate adjustment scenario and
24    the Voronoi Neighbor Averaging (VNA) spatial interpolation technique.  Additional details on
25    the VNA technique can be found in Appendix 4A.  Note that since each  region was adjusted
26    independently, in some cases distinct boundaries may be visible in the adjusted surfaces.  These
27    boundaries may be obscured to some degree due to the VNA interpolation procedure.
28
      focusing on a small number of large regions, rather than many localized areas, greatly reduces the problem size.
                                                   4-12

-------
       Legend
           Central
           East North Central
           Northeast
           Northwest
           South
           Southeast
           Southwest
           West
           West North Central
 2
 3
 4
Figure 4-6    Map of the 9 NOAA climate regions (Karl and Koss, 1984) used in the
              national-scale air quality adjustments
 5              4.3.4.2   Results
 6           Table 4-1 shows the highest monitored 2006-2008 average W126 concentration in each
 7    region  for observed air quality and air quality adjusted to meet the existing 63 standard of 75
 8    ppb, and the highest monitored 2006-2008 8-hour Os design value in each region for observed air
 9    quality and air quality adjusted to meet alternative standards based on the W126 metric with
10    levels of 15  ppm-hrs, 11  ppm-hrs, and 7 ppm-hrs.  Recall that the adjusted air quality surfaces
11    used in the welfare risk  and exposure analyses adjusted  each region  down to the existing  Os
12    standard before applying  additional reductions to meet the alternative standards. So effectively,
13    Table 4-1 shows which standard was the "controlling"  standard in each region.  For example,
                                                    4-13

-------
 1   when all monitors in the Central region were adjusted to meet the existing standard, the highest
 2   resulting W126 value was 14 ppm-hrs.  Thus, in the Central region, no further adjustments were
 3   necessary to meet the alternative standard of 15 ppm-hrs, but further adjustments were necessary
 4   to meet the alternative standards of 11 ppm-hrs and 7-ppm-hrs.
 5   Table 4-1     Highest 2006-2008 average W126 concentrations in the observed and existing
 6                 standard air quality adjustment scenarios; highest 2006-2008 8-hour Os
 7                 design values in the observed and potential alternative standard air quality
 8                 adjustment scenarios
Region
Central
East North Central
Northeast
Northwest
Southeast
South
Southwest
West
West North Central
Highest W126 value (ppm-
hrs)
Observed
18.3
13.8
17.9
6.6
22.2
18.1
24.3
48.6
12.2
75 ppb
adjustment
14.0
6.4
2.6
3.8
11.9
6.4
17.7
18.9
9.3
Highest 8-hour maximum-based design value (ppb)
Observed
88
86
92
76
95
91
86
119
80
15 ppm-hr
adjustment
83
86
94
76
81
89
71
71
80
11 ppm-hr
adjustment
72
83
89
76
74
91
65
66
79
7 ppm-hr
adjustment
66
76
76
76
67
79
62
61
72
10          From Table 4-1, it can be inferred that while each of the  9  regions had at least one
11   monitor with 2006-2008 air quality data not meeting the existing Os standard, there were 3
12   regions (East North Central, Northwest, West North Central) with all monitors meeting the
13   potential alternative standard with a W126 level of 15 ppm-hrs based  on 2006-2008 air quality
14   data.  Furthermore, all monitors in the Northwest region met the alternative standards of 11 ppm-
15   hrs and 7-ppm-hrs based on 2006-2008 ambient data.  When the air quality was adjusted to meet
16   the existing standard, only two regions (West  and Southwest)  had monitors with  W126
17   concentrations remaining  above 15 ppm-hrs.  In addition,  there were 4  regions (East North
18   Central, Northeast,  Northwest, and South) that already met 7 ppm-hrs when air quality was
19   adjusted to meet the existing standard.
                                                  4-14

-------
 1          Figure 4-7 shows the national-scale 2006-2008 average W126 surface adjusted to just
 2    meet the existing 63 standard of 75 ppb using the HDDM adjustment procedure described in
 3    Section  4.3.2.1,  and Figure 4-8 shows the  difference between  the recent air  quality surface
 4    (Figure  4-5) and Figure 4-7.  Figure 4-9, Figure 4-11, and Figure 4-13 show the 2006-2008
 5    average  W126 surfaces further adjusted to just meet 15 ppm-hrs, 11 ppm-hrs, and 7  ppm-hrs,
 6    respectively, while Figure 4-10, Figure 4-12, and Figure 4-14 show the differences between the
 7    surface adjusted to just meet the existing Os standard of 75 ppb, and the surfaces further adjusted
 8    to just meet the potential alternative standards based on the W126 metric with levels of 15 ppm-
 9    hrs, 11 ppm-hrs,  and 7 ppm-hrs. It is immediately apparent from these figures that the reductions
10    in W126 between recent air quality and air quality just meeting the existing standard (Figure 4-8)
11    are  much larger  than the additional reductions in W126 between  air quality just meeting the
12    existing standard and air quality  meeting the alternative standards (Figure 4-10, Figure 4-12,
13    Figure 4-14).
14          This  is further exemplified in Figure 4-15 and Figure 4-16, which show empirical
15    probability  density  and  cumulative distribution functions based on  the  monitored 8-hour O3
16    design values (Figure 4-15) and W126 concentrations (Figure 4-16) for each of the air quality
17    scenarios.  Both  sets of density functions show a large shift leftward going from observed air
18    quality to just meeting the existing standard, followed by much smaller leftward shifts from air
19    quality just meeting the existing  standard to air quality just meeting the potential alternative
20    standards. The shift between air  quality just meeting the existing standard and air quality just
21    meeting the potential alternative standard based on the W126 metric with a level of 15 ppm-hrs
22    is especially small, since only a few monitors in the Southwest and West regions did not meet a
23    W126 level of 15 ppm-hrs when air quality was adjusted to meet the existing standard.
                                                   4-15

-------
2
O


4
Figure 4-7    National surface of 2006-2008 average W126 concentrations (in ppm-hrs)
             adjusted to just meet the existing Os standard of 75 ppb
                                               4-16

-------
3   Figure 4-8   Difference in ppm-hrs between the national surface of observed 2006-2008
4                average W126 concentrations and the national surface of 2006-2008 average
5                W126 concentrations adjusted to just meet the existing Os standard of 75
6                ppb
                                               4-17

-------
1
2
3

4
Figure 4-9    National surface of 2006-2008 average W126 concentrations (in ppm-hrs)
             adjusted to just meet the potential alternative standard of 15 ppm-hrs
                                               4-18

-------
1
2
3
4
5
Figure 4-10   Difference in ppm-hrs between the national surface of 2006-2008 average
             W126 concentrations adjusted to just meet the existing Os standard of 75
             ppb and the national surface of 2006-2008 average W126 concentrations
             adjusted to just meet the potential alternative standard of 15 ppm-hrs
                                               4-19

-------
1
2
3

4
5
Figure 4-11   National surface of 2006-2008 average W126 concentrations (in ppm-hrs)
             adjusted to just meet the potential alternative standard of 11 ppm-hrs
                                               4-20

-------
1
2
3
4
5
Figure 4-12   Difference in ppm-hrs between the national surface of 2006-2008 average
             W126 concentrations adjusted to just meet the existing Os standard of 75
             ppb and the national surface of 2006-2008 average W126 concentrations
             adjusted to just meet the potential alternative standard of 11 ppm-hrs
                                               4-21

-------
1
2
3

4
5
Figure 4-13   National surface of 2006-2008 average W126 concentrations (in ppm-hrs)
             adjusted to just meet the potential alternative standard of 7 ppm-hrs
                                               4-22

-------
1
2
3
4
5
Figure 4-14   Difference in ppm-hrs between the national surface of 2006-2008 average
             W126 concentrations adjusted to just meet the existing Os standard of 75
             ppb and the national surface of 2006-2008 average W126 concentrations
             adjusted to just meet the potential alternative standard of 7 ppm-hrs
                                               4-23

-------
         g
        ff"*
        1= O
                                                                            Observed
                                                                            75 ppb
                                                                            15 ppm-hrs
                                                                            11 ppm-hrs
                                                                            7 ppm-hrs
                     i
                    40
                                  i                 I
                                 60               80
                                      03 Design Value (ppb)
                                                     100
               120
         o
         r-j
         o
         CD
        fD
        2<=,
        to co
       'CO
       01
              Observed
              75 ppb
              15 ppm-hrs
              11 ppm-hrs
              7 ppm-hrs
                            i
                           20
                                       i
                                      40
                                         I
                                        60
 i
30
100
1
2
3
4
5
                                               Percentile
Figure 4-15
Empirical probability density and cumulative distribution functions for the
monitored 2006-2008 8-hour O3 design values, and the 2006-2008 8-hour O3
design values after adjusting to just meet the existing and potential
alternative standards
                                                  4-24

-------
         CO
         o
        IS
        (D
                                                             Observed
                                                             75ppb
                                                             15 ppm-hrs
                                                             11 ppm-hrs
                                                             7 ppm-hrs
                                         7            15
                                            W1 26 (ppm-hrs)
                                                                      i
                                                                     50
         o
         CO
        Q. '
       •-_-•
       CO
Observed
75ppb
15 ppm-hrs
11 ppm-hrs
7 ppm-hrs
                            i
                           20
                         i
                        40
 I
60
 i
80
100
                                               Percentile
2   Figure 4-16   Empirical probability density and cumulative distribution functions for the
3                 monitored 2006-2008 average W126 concentrations, and the 2006-2008
4                 average W126 concentrations after adjusting to just meet the existing and
5                 potential alternative standards. Note W126 concentrations are displayed using
6                 a square root scale.
                                                  4-25

-------
 1    4.4   QUALITATIVE ASSESSMENT OF UNCERTAINTY
 2          As noted in Chapter 3, we have based the design of the uncertainty analysis for this
 3    assessment on the framework outlined in the WHO guidance (WHO, 2008).  For this qualitative
 4    uncertainty analysis, we have described each key source of uncertainty and qualitatively assessed
 5    its potential impact (including both the magnitude and direction of the impact) on risk results, as
 6    specified in the WHO guidance. In general, this assessment includes qualitative discussions of
 7    the potential impact of uncertainty on the results (WHO Tierl) and quantitative  sensitivity
 8    analyses where we have sufficient data (WHO Tier 2).
 9          Table 4-2 includes the key sources  of uncertainty identified for the Os REA. For each
10    source of uncertainty, we have (a) provided a description, (b) estimated the direction of influence
11    (over, under, both, or unknown) and magnitude  (low, medium, high) of the potential impact of
12    each source of uncertainty on the  risk estimates, (c) assessed the degree of uncertainty (low,
13    medium, or high) associated with the knowledge-base (i.e., assessed how well we understand
14    each source of uncertainty), and (d) provided comments further clarifying the qualitative
15    assessment presented. The categories used in describing the  potential magnitude of impact for
16    specific sources of uncertainty on risk estimates (i.e., low, medium,  or high) reflect our
17    consensus on the degree to which a particular source could produce  a sufficient  impact on risk
18    estimates to influence the interpretation of those estimates in the context of the secondary Os
19    NAAQS review. Where appropriate, we have included references to specific sources of
20    information considered in arriving at a ranking and classification for a particular source of
21    uncertainty.
                                                   4-26

-------
1     Table 4-2  Summary of Qualitative Uncertainty Analysis of Key Air Quality Elements in the Os NAAQS Risk Assessment
        Source
          Description
                                                               Potential influence of
                                                           uncertainty on risk estimates
Direction
Magnitude
Knowledge-
    Base
uncertainty*
                         Comments

   (KB: knowledge base, INF: influence of uncertainty on risk
                          estimates)
A. Ambient air quality
measurement data
O3 concentrations measured by
ambient monitoring instruments
have inherent uncertainties
associated with them.  Additional
uncertainties due to other factors
may include:

- monitoring network design

- required O3 monitoring seasons

- monitor malfunctions

- wildfire and smoke impacts

- interpolation of missing data
  Both
   Low
    Low
KB: O3 measurements are assumed to be accurate to within 1A of the
instrument's Method Detection Limit (MDL), which is 2.5 ppb for
most instruments. EPA requires that routine quality assurance checks
are performed on all regulatory instruments, and that all data reported
to AQS are certified by both the monitoring agency and the
corresponding EPA regional office.  See 40 CFR Part 58, Appendix A
for details. The CASTNET monitoring data were subject to their own
quality assurance requirements, and these data are generally believed
to be of comparable quality to the regulatory data stored in AQS.

KB: Monitor malfunctions sometimes occur causing periods of
missing data or poor data quality. Monitoring data affected by
malfunctions are usually flagged by the monitoring agency and
removed from AQS. In addition, the AQS database managers run
several routines to identify suspicious data for potential removal.

KB: There is a known tendency for smoke produced from wildfires to
cause interference in O3 instruments. Measurements collected by O3
analyzers were reported to be biased high by 5.1-6.6 ppb per 100
Hg/m of PM2.5 from wildfire smoke (Payton, 2007). However,
smoke concentrations high enough to cause significant interferences
are infrequent and the overall impact is believed to be minimal.

KB: Missing intervals of 1 or 2 hours in the measurement data were
interpolated, which may  cause some additional uncertainty.
However, due to the short length of the interpolation periods, and the
tendency for these periods to occur at night when O3 concentrations
are low, the overall impact is believed to be minimal.

INF: EPA's current O3 monitoring network requirements (40 CFR
Part 58, Appendix D) are primarily focused on urban areas. Rural
areas where O3 concentrations are lower tend to be under-represented
by the current monitoring network.  The network requirements also
state that at least one monitor within each urban area must be sited to
capture the highest O3 concentrations in that area, which may cause
some bias toward higher measured concentrations.

INF: Each state has  a required O3 monitoring season which varies in
length from May - September to year-round. Some states turn their
O3 monitors off during months outside of the required season, while
                                                                                     4-27

-------
        Source
           Description
                                                                Potential influence of
                                                             uncertainty on risk estimates
Direction
Magnitude
 Knowledge-
    Base
uncertainty*
                          Comments

   (KB: knowledge base, INF: influence of uncertainty on risk
                          estimates)
                                                                                                            others leave them on. This can cause differences in the amount of
                                                                                                            data available throughout the year across states, especially in months
                                                                                                            outside of the required O3 monitoring season.
B. Veronoi Neighbor
Averaging (VNA)
spatial fields
VNA is a spatial interpolation
technique used to estimate W126
concentrations in unmonitored
areas, which has inherent
uncertainty
  Both
   Low-
 Medium
Low-Medium
KB: VNA interpolates 2006-2008 average W126 values estimated
from hourly ambient air quality measurements at each CMAQ grid
cell in each of the 9 NOAA climate regions. The VNA estimates are
weighted based on distance from neighboring monitoring sites, thus
the uncertainty tends to increase with distance from the monitoring
sites becomes greater.  As a result, there is less uncertainty in the
VNA estimates near urban areas where the monitoring networks are
dense, and more uncertainty in sparsely populated areas where
monitors are further apart, particularly in the Western U.S.
C.CMAQ modeling
Model predictions from CMAQ,
like all deterministic photochemical
models, have both parametric and
structural uncertainty associated
with them
                                                                  Both
                  Low-
                Medium
              Low-Medium
                KB: Structural uncertainties are uncertainties in the representation of
                physical and chemical processes in the model. These include: choice
                of chemical mechanism used to characterize reactions in the
                atmosphere, choice of land surface model and choice of planetary
                boundary layer model.

                KB: Parametric uncertainties include uncertainties in model inputs
                (hourly meteorological fields, hourly 3-D gridded emissions,  initial
                conditions, and boundary conditions)

                KB: Uncertainties due to initial conditions are minimized by using a
                10 day ramp-up period from which model results are not used.

                KB: Evaluations of models against observed pollutant concentrations
                build confidence that the model performs with reasonable accuracy
                despite the uncertainties listed above. A comprehensive model
                evaluation provided in Appendix 4-B of the hREA shows generally
                acceptable model performance which is equivalent or better than
                typical state-of-the science regional modeling simulations as
                summarized in Simon et al (2012). However, both under-estimations
                and over-estimations do occur at some times and locations. Generally
                the largest mean biases occur on low ozone days during the summer
                season. In addition, the model did not fully capture rare wintertime
                high ozone events occurring in the Western U.S. Both of these types
                of biases are not likely to substantially affect W126 performance
                since low ozone days are not heavily weighted in the W126
                calculation and since the highest 3-month W126 values were  only
                                                                                       4-28

-------
        Source
           Description
                                                                Potential influence of
                                                            uncertainty on risk estimates
Direction
Magnitude
 Knowledge-
    Base
uncertainty*
                          Comments

   (KB: knowledge base, INF: influence of uncertainty on risk
                          estimates)
                                                                                                            calculated for April-October in this analysis.
D. Higher Order
Decoupled Direct
Method (HDDM)
HDDM allows for the
approximation of ozone
concentrations under alternate
emissions scenarios without re-
running the model simulation
multiple times using different
emissions inputs. This
approximation becomes less
accurate for larger emissions
perturbations especially under
nonlinear chemistry conditions.
  Both
   Low-
  Medium
Low-Medium
KB: To accommodate increasing uncertainty at larger emissions
perturbations, the HDDM modeling was performed at three distinct
emissions levels to allow for a better characterization of ozone
response over the entire range of emissions levels. The replication of
brute force10 hourly ozone concentration model results by the HDDM
approximation was quantified for 50% and 90% NOx cut conditions
for each urban case study areas (as shown in Appendix 4-D of the
hREA). At 50% NOx cut conditions, HDDM using information from
these multiple simulations predicted hourly ozone concentrations with
a mean bias and a mean error less than +/- 1 ppb in all urban case
study areas compared to brute force model simulations. At 90% NOx
cut conditions, HDDM using  information from these multiple
simulations predicted hourly ozone concentrations with a mean bias
less than +/- 3ppb and a mean error less than +/- 4 ppb in all urban
case study areas.
E. Application of
HDDM sensitivities to
ambient data
In order to apply modeled
sensitivities to ambient
measurements, regressions were
developed which relate ozone
response to emissions perturbations
with ambient ozone concentrations
for every season, hour-of-the-day
and monitor  location. Applying
ozone responses based on this
relationship adds uncertainty.
  Both
Medium
  Medium
KB: Preliminary work showed that the relationships developed with
these regressions were generally statistically significant for most
season, hour-of-the-day, and monitor location combinations for 2005
modeling in Detroit and Charlotte.  Statistical significance was not
evaluated for each regression in this analysis  since there were over
280,000 regressions created (1300 monitors x 2 sensitivity
coefficients x 3 emissions levels x 3 seasons  x 12 hours = 280,800
regressions). Statistics can quantify the goodness of fit for the
modeled relationships and can quantify the uncertainty in response at
any given ozone concentration based on variability in model results at
that portion of the distribution for each regression. However it is not
possible to quantify the applicability of this modeled relationship to
the actual atmosphere.

KB: The regression model provided both a central tendency and a
standard error value for ozone response at each measured hourly
ozone concentration.  The base analysis used  the central tendency
which will inherently dampen some of the variability in ozone	
     10 Brute force model concentrations refer to model results obtained by changing the emissions inputs and re-running the CMAQ model.  HDDM concentration
     estimates are an approximation of the model results that would be obtained by re-running the simulation with different inputs.
                                                                                       4-29

-------
        Source
           Description
                                                                Potential influence of
                                                             uncertainty on risk estimates
Direction
Magnitude
 Knowledge-
    Base
uncertainty*
                          Comments

   (KB: knowledge base, INF: influence of uncertainty on risk
                          estimates)
                                                                                                            response.  The standard error of each sensitivity coefficient was
                                                                                                            propagated through the calculation of predicted ozone concentrations
                                                                                                            at various standard levels.  These standard errors reflect the amount
                                                                                                            of variability that is lost due to the use of a central tendency. Since
                                                                                                            emissions reductions increased for lower standard levels the standard
                                                                                                            errors were larger for adjustments to lower standards. Mean (95th
                                                                                                            percentile) standard errors of hourly ozone for the 75 ppb adjustment
                                                                                                            case ranged from 0.13 (0.29) to 1.02 (2.11) ppb in the 9 climate
                                                                                                            regions. Mean (95th percentile) standard errors of hourly ozone for
                                                                                                            the 7 ppmh adjustment case ranged  from 0.23 (0.5) to 1.02 (2.14)
                                                                                                            ppb. The largest standard errors occurred in the northeast and west
                                                                                                            regions.

                                                                                                            INF: The NOx emissions reductions resulted in both increases and
                                                                                                            decreases in ozone depending on the time and location. In cases
                                                                                                            where the use of the central tendency of response reduced the total
                                                                                                            estimated emissions reductions required to achieve a given standard
                                                                                                            level, we expect that the benefits of reducing high ozone
                                                                                                            concentrations and the disbenefits of increasing low ozone would be
                                                                                                            generally underestimated.  Since the weighting function used to
                                                                                                            calculate W126 amplifies the importance of hourly concentrations
                                                                                                            above 50-60 ppb and dampens the importance of hourly
                                                                                                            concentrations below 50 ppb, this behavior would lead to an
                                                                                                            underestimation of the W126 metric. In contrast, in cases where the
                                                                                                            use of the central tendency of response increased the total estimated
                                                                                                            emissions reductions  required to achieve a given standard, we expect
                                                                                                            that the W126 metric would be overestimated.
F. Applying modeled
sensitivities to un-
modeled time periods
Relationships between ozone
response and hourly ozone
concentration were developed based
on 7 months of modeling: April-
October 2007. These relationships
were applied to ambient data from
2006-2008.
  Both
   Low-
 Medium
Low-Medium
KB: The seven months that were modeled capture a variety of
meteorological conditions. In cases where other years have more
frequent occurrences of certain types of conditions, the regressions
should be able to account for this. For instance, if a monitor only had
2-3 high ozone days associated with sunny, high pressure conditions
in the 2007 modeling but had 30-40 of those days in another year, the
regression may be more uncertain at those high ozone values but
should still be able to capture the central tendency which can be
applied to the more frequent occurances in other years.  If, on the
other hand, the meteorology/ozone conditions in another year were
completely outside the range of conditions captured in the model,
then the regression based on modeled conditions might not be able to
                                                                                       4-30

-------
        Source
          Description
                                                              Potential influence of
                                                           uncertainty on risk estimates
Direction
Magnitude
Knowledge-
    Base
uncertainty*
                         Comments

   (KB: knowledge base, INF: influence of uncertainty on risk
                         estimates)
                                                                                                        capture those conditions.

                                                                                                        KB: If emissions change drastically between the modeled period and
                                                                                                        the time of the ambient data measurements this could also change the
                                                                                                        relationship between ozone response and ozone concentrations. The
                                                                                                        regressions derived from the 2007 modeling period are only applied
                                                                                                        to measurements made within one year of the modeled time period.
                                                                                                        Although some emissions changes did occur over this time period, we
                                                                                                        believe it is still reasonable to apply 2007 modeling to this relatively
                                                                                                        small window of measurements which occurs before and after the
                                                                                                        modeling.
 G. Assumptions of
 across-the-board
 emissions reductions
Ozone response is modeled for
across-the-board reductions11 in
U.S. anthropogenic NOx. These
across-the-board cuts do not reflect
actual emissions control strategies.
  Both
 Medium
  Medium
KB: The form, locations, and timing of emissions reductions that
would be undertaken to meet various levels of the ozone standard are
unknown.  The across-the-board emissions reductions bring levels
down uniformly across time and space to show how ozone would
respond to changes in ambient levels of precursor species but do not
reflect spatial and temporal heterogeneity that may occur in local and
regional emissions reductions.
1     * Refers to the degree of uncertainty associated with our understanding of the phenomenon, in the context of assessing and characterizing its uncertainty. Sources
2     classified as having a "low" impact would not be expected to impact the interpretation of risk estimates in the context of the O3 NAAQS review; sources
3     classified as having a "medium" impact have the potential to change the interpretation; and sources classified as "high" are likely to influence the interpretation
4     of risk in the context of the O3 NAAQS review.
      11 "Across the board" emission reductions refer to equal percentage NOx emissions cuts in all source categories and all locations at all times.

                                                                                    4-31

-------
 1                        5   O3 RISK TO ECOSYSTEM SERVICES


 2      5.1   INTRODUCTION

 3           The EPA is using an ecosystem services framework as described in Chapter 2 to help

 4    define how the damage to ecosystems informs determinations of the adversity to public welfare

 5    associated with changes in ecosystem functions. Figure 9-1 of the ISA (U.S. EPA, 2013) is

 6    reproduced below (Figure 5-1) as a summary of exposure and effects that lead to potential loss of

 7    ecosystem services. Figure numbers in this figure refer to Chapter 9 of the ISA.
                            03 exposure
                        03 uptake & physiology (Fig 9-2)
                        •Antioxidant metabolism up-regulated
                        •Decreased photosynthesis
                        •Decreased stomatal conductance
                        or sluggish stomatal response
                      Effects on leaves
                      •Visible leaf injury
                      •Altered leaf production
                      •Altered leaf chemical composition
                        Plant growth (Fig 9.8)
                        •Decreased biomass accumulation
                        •Altered reproduction
                        •Altered carbon allocation
                        •Altered crop quality
                       Belowground processes (Fig 9.8)
                       •Altered litter production and decomposition
                       •Altered soil carbon and nutrient cycling
                       •Altered soil fauna and microbial communities
                                                                    Affected ecosystem services
                                                                    •Decreased productivity
                                                                    •Decreased C sequestration
                                                                    •Altered water cycling (Fig 9-7)
                                                                    •Altered community composition
                                                                    (i.e., plant, insects microbe)
10
11
12

13
Figure 5-1      Conceptual Diagram of the Major Pathway through which O3 Enters Plants and the Major
Endpoints that O3 May Affect in Plants and Ecosystems
                                                      5-1

-------
 1           This chapter focuses primarily on those ecosystem services potentially at risk from 63
 2    exposure that we were only able to assess qualitatively, due to a lack of sufficient data, methods,
 3    or resources to allow quantification of the incremental effects of O3. It also includes semi-
 4    qualitative GIS driven assessments of the potential impacts of Os on risks of fire and bark beetle
 5    damage and identifies additional adverse effects associated with Os exposure that we are not able
 6    to assess, even qualitatively.  In contrast, Chapters 6 and 7 provide quantitative assessments for
 7    risks related to tree biomass loss, timber and crop yield loss and visible foliar injury. Figure 5-2
 8    illustrates the relationships between the ecological effects of 63 and the anticipated ecosystem
 9    services  impacts that will be discussed in the following sections.
10
11
12
13
14
15
Figure 5-2
                       Ecological Effects
                       • Biomass Loss (Chapter 6)
                       • Foliar Injury (Chapter 7)
                        Supporting Services
                        • Primary Productivity
                                                            Regulating Services
                                                            • Carbon Sequestration
                                                            • Pollution Removal
                                                             Cultural Services
                                                             • Recreational Use
                                                       Provisioning Services
                                                       •Agricultural Harvest
                                                       • Timber Production
                                                                                 Additional Assessments
                                                                                 Sup port ing Services
                                                                                 • Soil Formation
                                                                                 • Community Structure
                                                                                 • Primary Productivity
                                                                                 Regulating Services
                                                                                 * Nutrient Cycling
                                                                                 • Water Regulation
                                                                                 • Pollination
                                                                                 • Fire Regulation
                                                                                 Cultural Services
                                                                                 • Aesthetic Services
                                                                                 * Non-Use Values
                                                                                 Provisioning Services
                                                                                 • Insect Damage
                                                                                 • Non-Timber Uses
Relationship between Ecological Effects of O3 Exposure and Ecosystem Services
       While most of the impacts of O3 on these services cannot be specifically quantified, it is
important to provide an understanding of the magnitude and significance of the services that  may
be negatively impacted by Os exposures. For many services, we can estimate the current total
                                                        5-2

-------
 1    magnitude and, for some, we can estimate the current value of the services in question. The
 2    estimates of current service provision will reflect the loss of services occurring from historical
 3    and current O3 exposure and provide context for the importance of any potential impacts of O3
 4    on those services, e.g., if the total value of a service is small, the likely impact of O3 exposure
 5    will also be small.  Likewise, if the total value is large, there is a higher potential for significant
 6    damage, even if the relative contribution of O3 as a stressor is small.  Also, in some cases we can
 7    provide information on locations where high O3 exposures occur in conjunction with significant
 8    ecosystem service  impairment. Specifically, we can provide information on areas where high
 9    W126 index values may have the greatest contribution to the service impairment caused by fires
10    in California and bark beetle damage in forests. This assessment will address O3 impacts on
11    ecosystem services following the framework of the Millennium Ecosystem Assessment (MEA,
12    2005). In line with the framework, the subsequent sections are divided into supporting,
13    regulating, provisioning, and cultural ecosystem services.

14     5.2  SUPPORTING SERVICES
15          Supporting services are the services needed by all of the other ecosystem services. Other
16    categories of services have relatively direct or short-term impacts on humans, while the impacts
17    on public welfare from supporting services are generally either indirect or occur over a long
18    time. The next sections describe potential impacts of O3 on some of these supporting services.
19                5.2.1   Net Primary Productivity
20          Primary productivity underlies the provision of many subsequent ecosystem services that
21    are highly valued by the public, including provision of food and timber.  The ISA determined
22    that biomass loss due to O3 exposure may reduce net primary productivity (NPP). According to
23    the ISA (U.S. EPA, 2013), when compared to 1860's era preindustrial conditions,  NPP in U.S.
24    Mid-Atlantic temperate forests decreased 7-8 percent per year from 1991-2000 due to O3
25    exposure, even with growth stimulation provided by elevated carbon dioxide and nitrogen
26    deposition.  Also, compared to a presumed pristine condition in 1860, NPP for the conterminous
27    U.S from 1950-1995 decreased as much as 13 percent per year in some areas in the agricultural
28    region of the Midwest during the mid-summer.  While there are models available to help
29    quantify changes in NPP and in the hydrologic cycle discussed in Section 5.3.1 we were not able

                                                   5-3

-------
 1    to attempt quantification of NPP or hydrology due to resource limitations.  Additionally these
 2    services are more difficult to interpret in ways that are meaningful to people.
 3                5.2.2   Community Composition and Habitat Provision
 4           Community composition or structure is also affected by 63 exposure. Since species vary
 5    in their response to 63, those species that are more resistant to the negative effects of 63 are able
 6    to out-compete more susceptible species. For example, according to studies cited in the ISA
 7    (U.S. EPA, 2013), the San Bernardino area community composition in high- O3 sites has shifted
 8    toward Os. tolerant species such as white fir, sugar pine, and incense cedar at the expense of
 9    ponderosa and Jeffrey pine. Changes in community composition underlie possible changes in
10    associated services such as herbivore grazing, production of preferred species of timber, and
11    preservation of unique or endangered communities or species, among others.
12            The National Survey on Recreation and the Environment (NSRE) is an ongoing survey
13    of a random sample of adults over the age of 16 on their interactions with the environment that
14    provides data on the values survey respondents place on the provision of habitat for wild plants
15    and animals. Table 5-1 summarizes the responses to survey questions regarding the value of
16    wildlife habitat and preservation of unique or endangered species.
17    Table 5-1 Responses to NSRE Wildlife Value Questions
Service
Wildlife Habitat
Preserving Unique Wild Plants and
Animals
Protecting Rare or Endangered Species
Percent of Respondents Considering the Service
Important
Extremely
Important
51
44
50
Very
Important
36
36
33
Moderately
Important
9
13
11
Total
96
93
94
18    *The remaining respondents felt these services were not important.
19          There exist meta-analyses on the monetary values Americans place on threatened and
20    endangered species. One such study (Richardson and Loomis, 2009) estimates the average
21    annual willingness to pay (WTP) for a number of species.  The authors report a wide range of
22    values dependent on the change in the size of the species population, type of species, and
                                                    5-4

-------
 1    whether visitors or households are valuing the species. The average annual WTP for surveyed
 2    species ranged from $9/year for striped shiner for Wisconsin households to $26I/year for
 3    Washington state households value for anadromous fish, such as salmon, in constant 2010$.

 4     5.3   REGULATING SERVICES
 5           Regulating services as defined by the MEA (2005) are those services that regulate
 6    ecosystem processes.  Services such as air quality, water, climate, erosion, and pollination
 7    regulation fit within this category. The next sections describe potential impacts of Os on some of
 8    these services.
 9                 5.3.1   Hydrologic Cycle
10           Regulation of the water cycle is another ecosystem service that can be adversely affected
11    by the effects of Os on plants. Studies of Os-impacted forests in eastern Tennessee in or near the
12    Great Smoky Mountains has shown that ambient Os exposures resulted in increased water use in
13    Os-sensitive species which led to decreased modeled late-season stream flow in those
14    watersheds. The increased water use resulted from a sluggish stomatal response that increases
15    water loss, which in turn increases water requirements (U.S. EPA, 2013). Ecosystem services
16    potentially affected by such a loss in stream flow could include habitat for species (e.g., trout)
17    that depend on an optimum stream flow or temperature. Additional downstream effects could
18    potentially include a reduction in the quantity and/or quality of water available for irrigation or
19    drinking and for recreational use.  Conversely, one model study reported in the ISA (U.S. EPA,
20    2013) associate  reduced stomatal aperture from Os exposure combined with nitrogen limitation
21    with decreased water loss, which in turn increased runoff; increased runoff could lead to more
22    soil erosion.  Regardless of the response, water cycling in forests is affected by 63 exposure and
23    has impacts on ecosystem services associated with both water quality and quantity.  As part of
24    the NSRE, the United States Forest Service (USFS) and the National Oceanographic and
25    Atmospheric Administration (NOAA) jointly surveyed Americans, age 16 and over, for their
26    report on Uses and Values of Wildlife and Wilderness in the United States. The NSRE
27    specifically asked respondents to rank the importance of water quality as a benefit of wilderness.
28    Ninety  one percent of respondents ranked water quality protection as either extremely or very
29    important; less than 1 percent of respondents ranked this service as not important at all.

                                                    5-5

-------
 1                 532  Pollination
 2           The ISA (U.S. EPA, 2013) identifies Os as a possible agent affecting the travel distance
 3    of and the loss of specificity of volatile organic compounds emitted by plants, some of which act
 4    as scent cues for pollinators.  While it is not possible to explicitly calculate the loss of pollination
 5    services resulting from this negative effect on scent cues, the loss is reflected in the current
 6    estimated value of $18.3 billion (2010$) for all pollination services, managed and wild, in North
 7    America (U.S., Canada, and Bermuda) (Gallai et al., 2009).
 8                 533  Fire Regulation
 9           Fire regime regulation is also negatively affected by 63 exposure.  Grulke et al.  (2009)
10    reported various lines of evidence indicating that 63 exposure may contribute to southern
11    California forest susceptibility to wildfires by increasing leaf turnover rates and litter. This, in
12    turn, creates increased fuel loads on the forest floor, O3-increased drought stress, and increased
13    susceptibility to bark beetle attacks. According to the National Interagency Fire Center
14    (http ://www. nifc. gov/firelnfo/firelnfo_stati sties .html), in 2010 in the United States over 3
15    million acres burned in wildland fires and an additional 2 million acres were burned in
16    prescribed fires.  Over the 5-year period from 2004 to 2008, Southern California alone
17    experienced, on average, over 4,000 fires per year burning, on average, over 400,000 acres per
18    fire (National Association of State Foresters [NASF], 2009).
19           The short-term benefits of reducing the 03-related fire risks include the value of avoided
20    residential property damages; avoided damages to timber, rangeland, and wildlife resources;
21    avoided losses from fire-related air quality impairments; avoided deaths and injury due  to fire;
22    improved outdoor recreation opportunities; and savings in costs associated with fighting the fires
23    and protecting lives and property.  For example, the California Department of Forestry  and Fire
24    Protection (CAL FIRE) estimated that average annual losses to homes due to wildfire from 1984
25    to 1994 were  $226 million (CAL FIRE,  1996) and were over $263 million in 2007 (CAL FIRE,
26    2008) in inflation adjusted 2010$. In fiscal year 2008, CAL FIRE's budgeted costs for  fire
27    suppression activities were nearly $304 million 2010 dollars (CAL FIRE,  2008).  CAL  FIRE also
28    estimates fire risk in the state on a -1 to 5 scale, with 2 being moderate risk.  Using GIS, we
29    developed maps that overlay the area of California with mixed conifer forest and the fire risk
30    area calculated by CAL FIRE. We then generated maps overlaying the current ambient Os
31    conditions and the modeled alternative scenarios with the areas of mixed conifer forest  that have
                                                    5-6

-------
 1
 2
 3
 4
 6
 7
a fire risk in the moderate and higher range. These maps allow us to calculate the area of mixed
conifer forests with moderate to high fire risk and high W126 index values under various
scenarios. Figure 5-3 shows W126 index values after just meeting the existing and alternative
standard levels in areas in California with fire risk greater than 2 on CAL FIRE's scale.
        Existing Standard
             11-15
                                      W12615ppm-hr
                                                                  W12611 and7ppm-hr
Figure 5-3     Overlap of W126 Index Values for the Existing Standard and Alternative W126 Standards,
Fire Threat > 2, and Mixed Conifer Forest
 9           The highest fire risk and highest W126 index values overlap with each other, as well as
10    with significant portions of mixed conifer forest. Under recent conditions, over 97 percent of
11    mixed conifer forests (21,800 square kilometers) have W126 index values over 7 ppm-hrs and a
12    moderate to severe fire risk, and 74 percent (16,500 square kilometers) have W126 index values
13    over 15 ppm-hrs with moderate to severe fire risk.  When we simulate just meeting the existing
14    standard almost all of the area of mixed conifer forest where there is a moderate to high fire
15    threat sees a reduction in 63 to below a W126 index value of 7 ppm-hrs.  At the adjusted
16    alternative W126 standard level of 15 ppm-hrs all but 40 km2 are under a W126 index value of 7
17    ppm-hrs and at 1 lor 7 ppm-hrs all of the moderate to high fire threat area is under 7 ppm-hrs.
18    Table 5-2 summarizes the reductions in areas of moderate to high-fire threat, mixed conifer
19    forests  at the existing and alternative standard levels.
20
                                                    5-7

-------
 1
 2
Table 5-2      Area of Moderate to High-Fire Threat, Mixed Conifer Forest for Existing and Alternative
Standard Levels (in km2)

Recent Conditions
Existing Standard
(75 ppb)
15 ppm-hrs
1 Ippm-hrs
7 ppm-hrs
<7ppm-hrs
482
22,180
22,257
22,297
22,297
7-llppm-hrs
2,542
117
40
0
0
11-15 ppm-hrs
5,271
0
0
0
0
>15 ppm-hrs
16,544
0
0
0
0
 4
 5
 6
 7
10
11
12
13
14
15
16
17
18
       In the long term, decreased frequency of fires could result in an increase in property
values in fire-prone areas. Mueller et al. (2007) conducted a hedonic pricing study to determine
whether increasing numbers of wildfires affect house prices in southern California. They
estimated that house prices would decrease 9.7 percent after one fire and 22.7 percent after a
second wildfire within 1.75 miles of a house in the study area. After the second fire, the housing
prices took between 5 and 7 years to recover.
       Figure 5-4 shows the locations of fires in the mixed conifer forest range in 2010. There
were 961 fires detected in these areas, including many in the national parks.  While we can't
conclude that Os reductions would have prevented these fires because there are many
contributing factors, we can conclude that under the air quality adjusted scenario just meeting the
existing standard will  in many areas,  decrease the role of Os as a contributing factor by reducing
the W126 index value to below 7 in most areas. Meeting alternative W126 standards results in
small to no additional reductions in the area of forests above a 7 ppm-hrs W126 standard level.
Additionally, long- term decreases in wildfire would be expected to yield outdoor recreation
benefits consistent with the discussion of scenic beauty in subsequent  sections.

-------
 1
 2
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
                       *   2010 Fires
                           <7
                           11-15
                           Mixed Conifer Range
Figure 5-4 Location of Fires in 2010 in Mixed Conifer Forest Areas (under Recent O3 Conditions)

  5.4   PROVISIONING SERVICES
       Provisioning services include market goods, such as forest and agricultural products. The
direct impact of Os-induced biomass and yield loss can be predicted for the commercial timber
and agriculture markets, respectively, using the Forest and Agriculture Optimization Model
(FASOM). This model provides a national-scale estimate of the effects of Os on these two
market sectors, including producer and consumer surplus estimates (see Section 6.3 for a
discussion of producer and consumer surplus).  Chapter 6 of this document provides detailed
analyses of the potential impact of biomass and yield loss on these services. Non-timber forest
products (NTFP), such as foliage and branches used for arts and crafts or edible fruits, nuts,  and
berries, can be affected by the impact of Os through biomass loss, foliar injury, insect attack, fire
regime changes, and effects on reproduction. Acknowledging that services lost in this sector can
be the result of interacting effects  of Os with other stressors, we also have included details for the
magnitude of the NTFP services in Chapter 6.
                                              5-9

-------
12
13
14
26
27
28
29
30
31
      Figure 5-5  Southern Pine Beetle Damage
      Courtesy: Ronald F. Billings, Texas Forest Service
      Bugwood.org
                                                         In addition to the direct effects of
                                                  O3 on tree growth, O3 causes increased
                                                  susceptibility to infestation by some
                                                  chewing insects (U.S. EPA, 2006). This
                                             &';';  potentially includes species that are not
                                                  considered sensitive to either biomass loss
                                                  or foliar injury such as Douglas fir.
                                                Chewing insects include the southern pine
                                                beetle and western bark beetle, species that
                                          11    are of particular interest to commercial
timber producers and consumers. These infestations can cause economically significant damage
to tree stands and the associated timber production.  Figure 5-5 and Figure 5-6 illustrate the
damage caused by southern pine beetles in parts of the south.
                                                      According to the USFS Report on the
                                                southern pine beetle (Coulson and Klepzig,
                                                2011), "Economic impacts to timber
                                                producers and wood-products firms  are
                                                essential to consider because the SPB causes
                                                extensive mortality in forests that have high
                                                commercial value in a region  with the most
                                                active timber market in the world."  The
                                                economic impacts of beetle outbreaks are
                                                multidimensional. In the short-term, the
                                                      surge in timber supply caused by owners
Figure 5-6 Southern Pine Beetle Damage
Courtesy: Ronald F. Billings, Texas Forest Service.
Bugwood.org
harvesting damaged timber depresses prices for timber and benefits consumers. In the long-
term, beetle outbreaks reduce the stock of timber available for harvest, raising timber prices to
the benefit of producers and the detriment of consumers.
       The USFS further reports that over the 28 years covered in their analysis (1977-2004),
because of beetle outbreaks, timber producers have incurred losses of about $1.4 billion, or about
$49 million per year, and wood-using firms have gained about $966 million, or about $35
                                                   5-10

-------
 1   million per year.  This results in a $15 million per year net negative economic impact.  (All
 2   dollar values are reported in constant 2010$.) These annual figures mask that most of the
 3   economic impacts result from a few catastrophic outbreaks, causing the impacts to pulse through
 4   the system in large chunks rather than being evenly distributed over the years. It is not possible
 5   to attribute a portion of these impacts resulting from the effect of Os on trees' susceptibility to
 6   insect attack; however, such losses are already reflected in the losses cited, and any welfare gains
 7   from decreased Os would positively impact the net economic impact.
 8          In the western United States, (Vsensitive ponderosa and Jeffrey pines are subject to
 9   attack by bark beetles. Ozone exposure increases susceptibility to these insect infestations in
10   sensitive species. Figure 5-7 shows areas considered 'at risk' of losing 25 percent or more basal
11   area in the contiguous United States to the top seven pine beetle species over the next 15 years
12   (pine beetle projections were calculated by the Forest Health Technology Enterprise Team).
13   Under recent conditions,  approximately 48,000 km2 have W126 index values above 15 ppm-hrs.
14   After just meeting the existing standard, all areas are under a W126 index value of 7 ppm-hrs
15   with the exception of about 4,000 km2 in Arizona and Colorado. After just meeting an
16   alternative standard level of 15 ppm-hrs, no area is above 7 ppm-hrs.  Table 5-3  and Table 5-4
17   provide summaries of areas at risk of higher pine beetle loss and millions of square feet of basal
18   tree area at high risk at various W126 index values.
19
                                                   5-11

-------

                                                                                        W126: 15, Il,and7ppm-hr
1

2    Figure 5-7      W126 Index Values for Just Meeting the Existing and Alternative Standards in Areas
3    Considered 'At Risk' of High Basal Area Loss (>25% Loss)
4
     Table 5-3
Area (km2) 'At Risk' of High Pine Beetle Loss at Various W126 Index Values

Recent Conditions
Existing Standard
(75 ppb)
15 ppm-hrs
1 Ippm-hrs
7 ppm-hrs
<7 ppm-hrs
3,456
80,640
84,528
84,528
84,528
7-1 Ippm-hrs
19,440
3,888
0
0
0
11-15 ppm-hrs
13,536
0
0
0
0
>15 ppm-hrs
48,096
0
0
0
0
                                                        5-12

-------
 1
 2
Table 5-4      Tree Basal Area Considered 'At Risk' of High Pine Beetle Loss ByW126 Index Values after
Just Meeting the Existing and Alternative Standard Levels (in millions of square feet)

Recent Conditions
Existing Standard
(75 ppb)
15 ppm-hrs
1 Ippm-hrs
7ppm-hrs
<7 ppm-hrs
90
982
1,091
1,091
1,091
7-llppm-hrs
368
110
0
0
0
11-15 ppm-hrs
145
0
0
0
0
>15 ppm-hrs
488
0
0
0
0
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
       In 2006, California was the largest producer of ponderosa and Jeffrey pine timber from
public lands. California accounted for 99 million board feet of saw logs - almost 40 percent of
the total U.S. production (U.S. Forest Service, 2009, available at:
http://srsfia2.fs.fed.us/php/tpo_2009/tpo_rpa_int2.php). California also experiences high W126
index values that may contribute to susceptibility to bark beetle attack.  It is not possible to
attribute a quantified impact of Os exposure to economic loss from bark beetle damage because
that impact is already reflected in the loss attributed to bark beetle infestation.  Reducing Os
impacts would likely reduce economic loss to California timber production.
       Figures 5-5 and 5-6 illustrate the impact insect outbreaks can have on aesthetic values
such as scenic beauty, as well as to the impacts on timber production. As shown in the NOx/SOx
Policy Assessment (U.S. EPA, 201 le), the value of the impact of 63  and insect attack
susceptibility on aesthetic values may be even greater than the market value of the timber.  We
will address timber production effects from reduced growth rates in Chapter 6  and effects of
foliar injury on related ecosystem services in Chapter 7.
18      5.5   CULTURAL SERVICES
19           Cultural services include non-use values (i.e., existence and bequest values) that can be
20    directly or indirectly impacted by Os exposure. According to responses to the NSRE, a large
21    majority of Americans wishes to preserve natural or pristine areas, even if they do not intend to
22    visit these areas. Outdoor recreation is another cultural service that may be affected by 63
23    exposure.  Foliar injury caused by 63 exposure and insect attack aided by 63 exposure may have
24    negative impacts on people's satisfaction with outdoor activities, especially those activities
25    associated with the quality of natural environments.
                                                    5-13

-------
 1           According to the National Report on Sustainable Forests (USDA, 2011) there are
 2    approximately 751 million acres of forest lands in the U.S., one-third of which is federally
 3    owned (Figure 5-8). All of these lands are assumed to be protected to some degree, but specific
 4    protections apply to wilderness areas, which comprise about 20 percent of public land. Of the
 5    remaining lands, 7 percent is protected as national parks; 13 percent is designated as wildlife
 6    refuges; and 60 percent is protected, managed forests, including national forests, Bureau of Land
 7    Management lands, and other state and local government lands. The protections afford
 8    preservation of cultural, social, and spiritual values.
                                 Other corporate,  Corporate forest   ^Local, 1.5%
                                               industry, 6.8% ^
                                                              State, 9.2%
                                        Family
                                       individual,
                                        35.1%

                 Other noncorporate,
                       2.9%
 9                                Forest land ownership (percent)
10    Figure 5-8      Percent of Forest Land in the US by Ownership Category, 2007
11    Source: USFS (Almost all forest lands are open for some form of recreation, although access may be restricted.)
12

13                 5.5.1   Non-Use Services
14                  The NSRE surveys also track American's attitudes toward various benefits
15    derived from the environment, including non-use values.  When people value a resource even
16    though they may never visit the resource or derive any tangible benefit from it, they perceive an
17    existence service. When the resource is valued as a legacy to future generations, a bequest
18    service exists.  Additionally, there exists an option value to knowing that you may visit a
19    resource at some point in the future. Data provided by the NSRE indicates that Americans have
20    very strong preferences for existence, bequest, and option services related to forests.
21    Significantly, according to the survey, only 5 percent of Americans rate wood products as the
22    most important value of public forests and wilderness areas, and for private forests, only 20
                                                     5-14

-------
1
2
3
4
5
6
7
      percent of respondents rated wood products as most important. Table 5-5 details the survey
      responses to these questions.
      Table 5-5      NSRE Responses to Non-Use Value Questions For Forests
Service
Existence
Option
Bequest
Percent of Respondents Considering the Service Important
Extremely
Important
36
36
81
Very Important
38
37
12
Moderately
Important
18
17
4
Total
92
90
97
     *Remaining respondents felt these services were not important.
             Studies (Haefele et al., 1991, Holmes and Kramer, 1995) indicate that the American
      public places a high value on protecting forests and wilderness areas from the damaging effects
 8    of air pollution. These studies assess willingness-to-pay (WTP) for spruce-fir forest protection in
 9    the southeast from air pollution and insect damage and confirm that the non-use values held by
10    the survey respondents were in fact greater than the use or recreation values.  The survey
11    presented respondents with a sheet of color photographs representing three stages of forest
12    decline and explained that, without forest protection programs, high-elevation spruce forests
13    would all decline to worst conditions. Two potential forest protection programs were proposed.
14    The first program (minimal program) would protect the forests along road and trail corridors
15    spanning approximately one-third of the ecosystem at risk. This level of protection may be most
16    appealing to recreational users. The second level of protection (more extensive program) was for
17    the entire ecosystem and may be most appealing to those who value the continued existence of
18    the entire ecosystem.  Median household WTP was estimated to be roughly $29 (in 2007 dollars)
19    for the minimal program and $44 for the more extensive program. Respondents were then asked
20    to decompose their value for the extensive program into use, bequest, and existence values. The
21    results were 13 percent for use value, 30 percent for bequest, and 57 percent for existence value
22    (Table 5-6).
23          While these studies are specific to damage due to excess nitrogen deposition and the
24    wooly balsam adelgid (a pest in Fraser fir), the results are relevant to Os exposure in forests
                                                  5-15

-------
 1
 2
 3
 4
 5
 6
because the effects are similar. In the southeast, loblolly pine is a prevalent species and 63 foliar
injury can cause visible damage.  Ozone exposure may also result in trees more susceptible to
insect attack, which in the southeast would include damage caused by the southern pine beetle.
Table 5-6      Value Components for WTP for Extensive Protection Program for Southern Appalachian
Spruce-Fir Forests
Type of Value
Use
Bequest
Existence
Total
Proportion of WTP
0.13
0.30
0.57
1.0
Component Value ($2007)
5.72
13.20
25.08
44.00
 8      5.6   QUALITATIVE ASSESSMENT OF UNCERTAINTY
 9           As noted in Chapter 3, we have based the design of the uncertainty analysis for this
10    assessment on the framework outlined in the WHO guidance (WHO, 2008). For this qualitative
11    uncertainty analysis, we have described each key source of uncertainty and qualitatively assessed
12    its potential impact (including both the magnitude and direction of the impact) on risk results, as
13    specified in the WHO guidance. In general, this assessment includes qualitative discussions of
14    the potential  impact of uncertainty on the results (WHO Tierl) and quantitative sensitivity
15    analyses where we have sufficient data (WHO Tier 2).
16           Table 5-7 includes the key sources of uncertainty identified for the O3 REA. For each
17    source of uncertainty, we have (a) provided a description, (b) estimated the direction of influence
18    (over, under, both, or unknown) and magnitude (low, medium, high) of the potential impact of
19    each source of uncertainty  on the risk estimates, (c) assessed the degree of uncertainty (low,
20    medium,  or high)  associated with the knowledge-base (i.e., assessed how well we understand
21    each source of uncertainty), and (d) provided comments further clarifying the qualitative
22    assessment presented. The  categories used in describing the potential magnitude of impact for
23    specific sources of uncertainty on risk estimates (i.e., low, medium, or high) reflect our
                                                    5-16

-------
1    consensus on the degree to which a particular source could produce a sufficient impact on risk
2    estimates to influence the interpretation of those estimates in the context of the secondary O3
3    NAAQS review. Where appropriate, we have included references to specific sources of
4    information considered in arriving at a ranking and classification for a particular source of
5    uncertainty.
                                                   5-17

-------
1    Table 5-7  Summary of Qualitative Uncertainty Analysis in Semi-Quantitative Ecosystem Services Assessments
           Source
              Description
  Potential influence of
   uncertainty on risk
        estimates
                                                       Direction
                                              Magnitude
             Knowledge-
                 Base
                       Comments (KB: knowledge base, INF: influence of
                             uncertainty on risk estimates)
A.  National W126
surfaces
The fire risk and bark beetle
analyses in this chapter use the
national W126 surfaces for
recent conditions and adjusted to
just meet the existing standard
and alternative W126 standards.
Both
Low-
Medium
Low-Medium
KB and INF: See Chapter 4 for more details.
B. Incremental impact
of O3 on ecosystem
services
Many ecosystem services
affected by O3 exposure are
discussed qualitatively or semi-
quantitatively, including
supporting services (e.g., net
primary productivity and
community composition),
regulating services (e.g.,
hydrologic cycle and
pollination), and cultural
services (e.g., recreation and
non-use).
Under
High
Low
 KB: The O3 ISA concludes that there is a causal relationship
 between O3 exposure and productivity in terrestrial ecosystems
 and biogeochemical cycles, and a likely to be causal
 relationship between O3 exposure and terrestrial water cycling
 and terrestrial community composition (U.S. EPA, 2011).
 However, we do not have sufficient data, methods, or resources
 to adequately quantify the incremental effects of changes in O3
 on many ecosystem services.
 INF: For many services, we can estimate the current total
 magnitude and, for some, we can estimate the current
 monetized value. The estimates of current service provision
 will reflect the loss of services occurring from historical and
 current O3 exposure and provide context for the importance of
 any potential impacts of O3 on those services, e.g., if the total
 value of a service is small, the total value of the likely impact of
 O3 exposure will also be small. Likewise, if the total value is
 large, there is a higher potential for significant damage, even if
 the relative contribution of O3 as a stressor is small.
                                                                              5-18

-------
           Source
             Description
  Potential influence of
   uncertainty on risk
        estimates
                                                       Direction
                                              Magnitude
             Knowledge-
                 Base
                       Comments (KB: knowledge base, INF: influence of
                             uncertainty on risk estimates)
C. Areas with fire risk
in California
Maps of areas with moderate and
higher fire risk have uncertainty,
and thus the potential overlap
with areas with higher W126
index values and mixed conifer
forests are also uncertain.
Unknown
Medium
High
KB: California's fire risk maps are systematically developed
including consideration of factors such as defensible space,
non-flammable roofs, and ignition resistant construction reduce
fire risk. (See
http://www.fire.ca.gov/firejrevention/firejrevention wildlan
d zones developmentphp).
INF: In 2010, over 3 million acres burned in wildland fires
(NIFC, 2010). The economic value of homes lost due to
wildfire and fire suppression activities can be hundreds of
millions of dollars per year in California (CAL Fire, 2006,
2007, 2008).
D. Areas at risk due to
bark beetle
In the western U.S., O3-sensitive
ponderosa and Jeffrey pines are
subject to attack by bark beetles.
Maps that identify areas
considered 'at risk' of losing 25
percent or more basal area to
pine beetle have uncertainty, and
thus the potential area of overlap
with areas with higher W126
index values are also uncertain.
Unknown
Medium
Medium
KB: O3 causes increased susceptibility to infestation by some
chewing insects (U.S. EPA, 2006, 2013), including the southern
pine beetle and the western bark beetle. It is not possible to
attribute a portion of these impacts resulting from the effect of
O3 on trees' susceptibility to insect attack; however, such losses
are already reflected in the  losses cited, and any welfare gains
from decreased O3 would positively impact these numbers.
INF: Insect infestations can cause economically significant
damage to tree stands and the associated timber production.
USFS estimates a$15 million per year net negative economic
impact due to bark beetle infestations (Coulson and Klepzig,
2011).
                                                                               5-19

-------
 1     5.7  DISCUSSION
 2          Ozone damage to vegetation and ecosystems from recent conditions causes widespread
 3    impacts on an array of ecosystem services. Biomass loss impacts numerous services, including
 4    supporting and regulating services such as net primary productivity, community composition,
 5    habitat, and climate regulation.  The provisioning services of timber production can be affected
 6    by the increased susceptibility to insect attack caused by Os exposure. Non-use values, including
 7    existence and bequest values, are also affected by the damage to scenic beauty caused by insect
 8    attack (an indirect effect of Os) and foliar injury (a direct effect).  Below we offer a few
 9    observations on the challenges of explicitly valuing ecosystem services, highlight the importance
10    of continuing to consider the services in our assessments, and indicate where additional analyses
11    and discussion on valuing the ecosystem services are located in this document.
12
13          •   Most of the impacts of Os exposure on ecosystem services cannot be specifically
14              quantified, but it is very important to provide an understanding of the magnitude and
15              significance of the services that may be harmed by Os exposure. For many ecosystem
16              services, we can estimate the current total magnitude and, for some, we can estimate
17              the current value of the services in question.

18          •   The impacts on public welfare from supporting services are generally either indirect
19              or occur over a long time. The ISA determined that biomass loss due to Os exposure
20              may have adverse effects on net primary productivity. But because of data and
21              methodology limitations, the loss of value to the public from incremental changes in
22              Os exposure on NPP on a national level is unquantifiable. Also, we were not able to
23              quantify the impacts of Os exposure on community composition.

24          •   Regulating ecosystem services include hydrologic cycle, pollination, and fire
25              regulation. Hydrologic, or water cycling in forests is affected by Os exposure and has
26              impacts on ecosystem services associated with both water quality and quantity.
27              While the NSRE results show that 91 percent of respondents rank water quality
28              protection as either extremely important or very  important, because of data and
29              methodology limitations, the loss of value to the public from incremental changes in
                                                   5-20

-------
 1              Os exposure on water cycling is not quantifiable. F'orpollination services, it is not
 2              possible to explicitly calculate the loss of pollination resulting from 63 exposure, but
 3              the loss is reflected in the current total estimated value of $18.3 billion (2010$) for
 4              pollination services in North America. Lastly, fire regulation is negatively affected
 5              by Os exposure through forest susceptibility to wildfires, drought stress, and insect
 6              attack. The value of this ecosystem service is reflected in avoided damage to
 7              residential property, timber, rangeland, and wildfire fighting resources, as well as
 8              improved outdoor recreation opportunities. As an example, the California
 9              Department of Forestry and Fire Protection (CAL FIRE) estimated that average
10              annual losses to homes due to wildfire from 1984 to 1994 were $163 million (CAL
11              FIRE, 1996) and were over $250 million in 2007 (CAL FIRE, 2008).  In fiscal year
12              2008, CAL FIRE's costs for fire suppression activities were nearly $300 million
13              (CAL FIRE, 2008).

14           •   Provisioning services include market goods, such as forest and agriculture products.
15              The direct impact of Os-induced biomass loss can be predicted for the commercial
16              timber and agriculture markets using the Forest and Agriculture Optimization Model.
17              Chapter 6 of this document provides detailed analyses of the potential impact of
18              biomass and yield loss on these services. In addition, non-timber forest products
19              (NTFP), such as foliage and branches used for arts and crafts or edible fruits, nuts,
20              and berries, can be affected by the impact of O3  through biomass loss, foliar injury,
21              insect attack, fire regime changes, and effects on reproduction.  We include details for
22              the magnitude  of the NTFP services in Chapter 6.

23           •   In addition, to  estimate the magnitude of insect attacks related to Os exposure  on
24              provisioning services, such as forest products, we reviewed the USFS Report on the
25              Southern Pine Beetle (Coulson and Klepzig, 2011). The USFS further reports that
26              over the 28 years covered in their analysis  (1977-2004), because of beetle outbreaks,
27              timber producers have incurred losses of about $1.4 billion, or about $49 million per
28              year, and wood-using firms have gained about $966 million, or about $35 million per
                                                   5-21

-------
 1              year.  This results in a $15 million per year net negative economic impact.1 While it
 2              is not possible to attribute a portion of these impacts resulting from the effect of 63 on
 3              trees' susceptibility to insect attack, these losses are reflected in the values cited.

 4           •   Outdoor recreation is a cultural service that may be affected by Oj exposure. Foliar
 5              injury caused by O3 exposure and insect attack aided by O3 exposure may have
 6              negative impacts on people's satisfaction with outdoor activities, especially those
 7              activities associated with the quality of natural environments.  These impacts are
 8              discussed in Chapter 7 on foliar injury.  In addition, some cultural services, such as
 9              existence or bequest services, lend themselves to evaluating total importance and
10              measuring total value, but assessing the impact of 63 effects on these services is not
11              currently possible.
              All values are reported in constant 2010$.
                                                     5-22

-------
 1                                        6   BIOMASS LOSS

 2    6.1     INTRODUCTION

 3            The previous O3 AQCDs (U.S. EPA,  1996, 2006) and current O3 ISA (U.S. EPA, 2013)
 4    concluded that there is strong and consistent evidence that ambient O3 decrease photosynthesis
 5    and growth in numerous plant species, but the magnitude of the effects are variable both across

 6    species and across regions of the U.S.
 7            The ecosystem services most directly affected by biomass loss include: (1) habitat
 8    provision for wildlife, particularly habitat for threatened or endangered wildlife, (2) carbon
 9    storage, (3) provision  of food and fiber,  and (4) pollution removal (see Figure 6-1).  Although we

10    cannot quantify reduction in habitat provision due to O3 exposure on either a national or case

11    study scale, there is evidence that this service is important to the public. In the cases of carbon
                                                         Ecological Effect
                                                           Biomass Loss
          Ambient
           Ozone
          Exposure
         (Chapter 4)
                Provisioning Services
                • Non-Timber Uses
                * Commercial Non-Timber
                 Forest Products
                • Informal Economy of Non-
                 Timber Forest Products
                Ecosystem Level Effects
                • Weighted Biomass
                 Loss
                • Species Diversity
                * Community Structure
                • Ecosystem functioning
                                   Regulating Services
                                   •Carbon Sequestration
                                   •Changes in Carbon
                                   Sequestration in 5 Urban Areas

                                   •Pollution Removal
                                   •Changesin Pollution Removal
                                   in 5 Urban Areas
Provisioning Services
•Timber Production
•Changes in National Yield and Prices
* Impacts on Producers and
Consumers

•Agricultural Harvest
• Changes in National Yield and Prices
• Impacts on Producers and
Consumers

Regulating Services
• Carbon Sequestration
• National Changes in Carbon
Sequestration
12

13
14
15
Figure 6-1     Conceptual Diagram of Relationship of Relative Biomass Loss to Ecosystem
                Services [The dashed box indicates those services for which direct
                quantification was not possible.]
                                                           6-1

-------
 1    storage and food and fiber provision, the analyses presented here used the concentration-
 2    response  (C-R) functions developed for trees and crops to model, at the national scale, the
 3    approximate loss of services and the marginal benefits of alternative levels of a W126 standard.
 4          We included national parks at the case-study scale, as well as Class I areas. Class I areas
 5    are designated as areas in which visibility has been determined to be of important value (C.F.R.
 6    40, 81.400). The determination is primarily based on air quality limitations on visibility, but in
 7    this assessment we are using them in the context of protected areas of interest to address
 8    potential impacts. The national parks are meant to be preserved for the enjoyment of present and
 9    future generations, as well as for the unique or sensitive ecosystems and species in the parks.
10    The parks are not a source of food or fiber production and are not included in the analysis of
11    those services.  And although the parks do provide carbon sequestration and storage and
12    pollution  removal, neither of the models for these ecosystem services available for this review
13    was able to include national parks.  The model used for the urban case study areas allows
14    analysis of carbon sequestration and storage and pollution removal services; it does not include
15    habitat provision or food and fiber production.
16          The remainder of this Chapter includes Section 6.2 - Relative Biomass Loss; Section 6.3
17    - Commercial Timber Effects; Section 6.4 - Non-Timber Forest Products; Section 6.5 -
18    Agriculture; 6.6 - Climate Regulation; Section 6.7 - Urban Case Study Air Pollution Removal;
19    and Section 6.8 - Ecosystem Level Effects.

20    6.2    RELATIVE BIOMASS LOSS
21          The 1996 and 2006 Os AQCDs relied extensively on results from analyses conducted on
22    commercial crop species for the National Crop Loss Assessment Network (NCLAN) and on
23    analyses of tree seedling species conducted by the EPA's National Health and Environmental
24    Effects Laboratory Western Ecology Division (NHEERL/WED). Results from these studies
25    have been published  in numerous publications, including Lee et al. (1994; 1989,  1988b, 1987),
26    Hogsett et al. (1997), Lee and Hogsett (1999), Heck et al. (1984), Rawlings and Cure (1985),
27    Lesser et  al. (1990), and Gumpertz and Rawlings (1992). Those analyses concluded that a three-
28    parameter Weibull model is the most appropriate model for the response of absolute yield and
29    growth to 63 exposure because of the interpretability of its parameters, its flexibility (given the
                                                     6-2

-------
 1    small number of parameters), and its tractability for estimation.  See equation 6-1 for an example
 2    of a three-parameter Weibull model .
 3
 5                                                                              Equation 6-1
 6
 7          In addition, if the intercept term, a, is removed, the model estimates relative yield or
 8   biomass without any further reparameterization. Formulating the model in terms of relative yield
 9   or biomass loss (RBL) in relation to the 3-month W126 index is essential for comparing
10   exposure-response across species or genotypes or for experiments for which absolute values of
1 1   the response may vary greatly. See equation 6-2 for the reformulated model.
12
13                                    RBL = 1 - exp[-(W126/ti)p]
14                                                                              Equation 6-2
15          In the 1996 and 2006 Os AQCDs, the two-parameter model of RBL was used to derive
16   common models for multiple species, multiple genotypes within species, and multiple locations.
17   Relative biomass loss (RBL) functions for the 12 tree species used in this assessment are
18   presented in Table 6-1 (see the ISA (U.S. EPA, 2013) for a more extensive review of the
19   calculation of the C-R functions), and RBL functions for the 10 crop species used in this
20   assessment are presented in Table 6-2.
21
                                                    6-3

-------
1   Table 6-1     Relative Biomass Loss Functions for Tree Species
Species
Red Maple (Acer rubrum)
Sugar Maple (Acer saccharum)
Red Alder (Alnus rubra)
Tulip Poplar (Liriodendron tulipifera)
Ponderosa Pine (Pinus ponder osa)
Eastern White Pine (Pinus strobus)
Loblolly Pine (Pinus taeda)
Virginia Pine (Pinus virginiand)
Eastern Cottonwood (Populus deltoides)
Quaking Aspen (Populus tremuloides)
Black Cherry (Prunus serotina)
Douglas Fir (Pseudotsuga menzeiesii)
RBL Function





1~__,-r /\\T-t ->/:/~\Pl
— exp[-^wizo/TiJ j





TI (ppm)
318.12
36.35
179.06
51.38
159.63
63.23
3,966.3
1,714.64
10.10
109.81
38.92
106.83
P
1.3756
5.7785
1.2377
2.0889
1.1900
1.6582
1.000
1.0000
1.7793
1.2198
0.9921
5.9631
2
3
Table 6-2     Relative Biomass Loss Functions for Crop Species
Species
Barley
Field Corn
Cotton
Kidney Bean
Lettuce
Peanut
Potato
Grain Sorghum
Soybean
Winter Wheat
RBL Function




p
exp[ (W126/T,) ]




TI (ppm)
6,998.5
97.9
96.1
43.1
54.6
96.8
99.5
205.3
110.2
53.4
P
1.388
2.968
1.482
2.219
4.917
1.890
1.242
1.957
1.359
2.367
4
5
6
7
       Figure 6-2 shows a comparison of W126 median RBL response functions for the tree
species used in this assessment, and Figure 6-3 shows a comparison of W126 median RBL
response functions for the crop species used in this assessment. The figures illustrate how the
two parameters affect the shape of the resulting curves. Differences in the shapes of these curves
                                                  6-4

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
are important for understanding differences in the analyses presented later in this chapter. The
two parameters of the RBL equation (Equation 6-2) control the shape of the resulting curve. The
value of r| in the RBL function affects the inflection point of the curve, and P affects the
steepness of the curve. Species with smaller values of P (e.g., Virginia Pine) or species with r|
values that are above the normal range of ambient W126  measurements (e.g., Ponderosa Pine
and Red Alder) have response functions with more gradual and consistent slopes. This results in
a more constant rate of change in RBL over a range of Os exposure consistent with ambient
exposure concentrations.
       In contrast, the species with larger P values (e.g.,  Sugar Maple) have response functions
that behave more like thresholds, with large changes in RBL over a small range of W126 index
values and relatively small changes at other index values. In these cases the "threshold" is
determined by the t] parameter of the  model. In the example of Eastern Cottonwood, P is
relatively low, but because r| is also very low relative to the other species, the resulting C-R
curve has a very steep gradient relative to other species with similar P values.
          CO
          0

          CD
          O
      00
      01
          CNI
          d
          p
          d
             Q  Red Maple
             •  Sugar Maple
             •  Red Alder
             •  Tulip Poplar
             n  Ponderosa Pine
             •  White Pine
             •  Loblolly Pine
             d  Virginia Pine
             •  Cottonwood
             •  Aspen
             •  Black Cherry
             d  Douglas Fir
                 0
                         10
20             30
     W126
40
50
16   Figure 6-2 Relative Biomass Loss Functions for 12 Tree Species
17
                                                     6-5

-------
          CO
          0

          CD
          O
      00
      01
          CN)
          O
•  Barley
•  Field Corn
•  Cotton
•  Kidney Bean
•  Lettuce
•  Peanut
•  Potato
d  Grain Sorghum
n  Soybean
•  Winter wheat
                0
            10
20
30
40
50
 1                                                W126
 2   Figure 6-3 Relative Biomass Loss Functions for 10 Crop Species
 O
 4               6.2.1     Species-Level Analyses
 5                   6.2.1.1   Comparison of seedling to adult tree biomass loss
 6          The response functions for tree species used in this analysis are all based on seedlings
 7   grown in open top chambers (OTC). Since the 2006  O3 AQCD (U.S. EPA, 2006), several studies
 8   were published based on the Aspen Free-Air Carbon Dioxide Enrichment (FACE)1 experiment
 9   using "free air," Os  and CC>2 exposures in a planted forest in Wisconsin. Overall, the studies at
10   the Aspen FACE experimental site were consistent with many of the open-top chamber (OTC)
11   studies that were the foundation of previous Os NAAQS reviews. These results strengthen our
12   understanding of Os effects on forests and demonstrate the relevance of the knowledge gained
13   from Aspen tree seedlings grown in OTC studies.
14          In the 2006 AQCD (U.S. EPA, 2006), the TREGRO and ZELIG models were used to
15   simulate growth of adult trees. For this analysis we did not conduct new TREGRO or ZELIG
16   simulations. We used several existing publications, which modeled tree species used in this
17   analysis. For this analysis, we calculated the W126 index values from the hourly concentrations
     1 The Aspen FACE experiment is a multidisciplinary study to assess the effects of increasing tropospheric O3 and
       carbon dioxide levels on the structure and function of northern forest ecosystems.
                                                    6-6

-------
 1
 2
 3
 4
 5
at the monitors used in the studies. The seedling RBL was calculated from this W126 and
compared to the study results and adjusted to reflect an annualized RBL. The results are
summarized below in Table 6-3.

Table 6-3     Comparison of Adult to Seedling Biomass Loss
Study

Constable and
Taylor, 1997






Weinstein et
al., 2001









W126

0.18
8.98
46.37
89.40
149.22




0.32
15.38
59.17

0.32
15.38
59.17

0.32
15.38
59.17
Adult RBL
TREGRO
0%
0.3%
3.1%
6.4%
12.1%




Tulip Polar
4.7%
10.6%
16.8%
Red Maple
2.5%
4.9%
8.2%
Black Cherry
0.2%
0.3%
0.5%
Adult RBL
ZELIG
N/A






Tulip Polar
+3.2%
5.3%
11.2%
Red Maple
0%
15.6%
15.6%
Black Cherry
11.2%
4.2%
+9.1%
Seedling RBL

0.03%
3.2%
20.5%
39.5%
60.3%




Tulip Poplar
0%
7.7%
73.89%
Red Maple
0.01%
1.5%
9.4%
Black Cherry
0.9%
32.8%
78.0%
Comments

This study used TREGRO
and included the western
and eastern subspecies of
Ponderosa Pine. Os data
were not available for the
western subspecies, which
was found to be more
sensitive than the eastern
subspecies. The seedling
C-R function used does
not differentiate between
subspecies.
This study used TREGRO
and ZELIG to model
Tulip Poplar, Red Maple,
and Black Cherry.









 6
 7
 8
 9
10
11
12
       These studies indicate that overall, the seedling biomass loss values are much more
consistent with the adult loss, as estimated by TREGRO and ZELIG, at lower W126 index
values. The Constable and Taylor (1997) study implies that for the eastern subspecies of
Ponderosa Pine, the seedling RBL rate overestimates the adult RBL rate. 63 data for the western
subspecies were not available, but Constable and Taylor (1997) found the western subspecies to
be more sensitive. The Weinstein et al. (2001) study indicates that the seedling RBL estimates
                                                   6-7

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
are comparable to the adult estimates, except at higher W126 index values of 63 for Tulip
Poplar. The Black Cherry results are an exception, which tells us that this species is much less
sensitive as an adult than as a seedling.  As such, the seedling RBL rate would overestimate RBL
loss in adult trees. One other study (Samuelson and Edwards, 1993) on Red Oak, another
hardwood species, found the exact opposite pattern — adult trees are much more sensitive to Os-
related biomass loss than seedlings.
      Mclaughlin et al. (2007) completed a  study assessing the interactive effects of Os and
climate on tree growth and water use. We used the monitored Os concentrations in this study to
calculate the W126 index value and then used these values to compare the predicted seedling
RBL to observations in the study. The study did not use absolute biomass loss, instead relying on
measurements of circumference to address growth.  In addition, the results were presented as
comparisons in growth in 2002 and 2003 relative to 2001. Table 6-4 presents a summary of the
results.

Table 6-4     Comparison of Seedling Biomass Loss to Adult Circumference
Species
Tulip Poplar
Tulip Poplar
Tulip Poplar
Black Cherry
Red Maple
Sugar Maple
W126
2001
23.31
19.78
14.71
14.71
14.71
14.71
2002
39.82
32.14
17.50
17.50
17.50
17.50
2003
20.15
11.25
9.22
9.22
9.22
9.22
Study Results
(% change in
circumference)
2002
-26%
-49.6%
-62%
-75%
-59.6%
-43.5%
2003
-38%
7.5%
N/A
N/A
N/A
N/A
RBL (seedling)
2001
-17.5%
-12.7%
-7.1%
-31.7%
-1.5%
-0.5%
2002
-44.4%
-31.3%
-10.0%
-36.4%
-1.8%
-1.5%
2003
-13.9%
-4.1%
-2.7%
-21.3%
-0.8%
-0.04%
Comparison
2002
-60.7%
-59.4%
-72.8%
-41.5%
-58.4%
-97.5%
2003
32.4%
210%
N/A
N/A
N/A
N/A
16
17
18
19
20
21
       Relative to the observed changes in circumference, the seedling RBL estimates are mixed
for Tulip Polar. Loss was overestimated in 2002 but was underestimated in 2003. The results for
Sugar Maple were similar to Tulip Poplar, with loss overestimated in 2002. In contrast to the
TREGRO results presented above, the results in this study found much greater loss in Black
Cherry, and the seedling RBL underestimated the change for adult trees in 2002. The results for
                                                    6-8

-------
    Red Maple were very similar for 2002. Table 6-5 summarizes the uncertainty for all species used

    in this study.
4   Table 6-5     Summary of Uncertainty in Seedling to Adult Tree Biomass Loss
5                 Comparisons
Species
Red Maple (Acer rubrum)
Sugar Maple (Acer saccharum)
Red Alder (Alnus rubra)
Tulip Poplar (Liriodendron
tulipifera)
Ponderosa Pine (Pinus ponder osa)
Eastern White Pine (Pinus strobus)
Loblolly Pine (Pinus taeda)
Virginia Pine (Pinus virginiand)
Eastern Cottonwood (Populus
deltoides)
Quaking Aspen (Populus
tremuloides)
Black Cherry (Prunus serotina)
Douglas Fir (Pseudotsuga
menzeiesii)
Summary of Seedling-Adult Uncertainty
Seedling C-R functions underestimated RBL relative to estimates of adult
biomass loss from TREGRO and ZELIG. The seedling RBL was comparable
to field results of changes in circumference.
No TREGRO data were available. Seedling RBL overestimated loss compared
to field results of changes in circumference.
No data were available.
Seedling C-R functions underestimated RBL relative to results from TREGRO
and ZELIG and lower W126 index values of O3, and overestimated RBL at
the very high index values. Seedling RBL overestimated loss compared to
field results of changes in circumference in 2002, but underestimated loss in
2003.
Seedling C-R functions overestimated RBL relative to TREGRO results for
the eastern subspecies. Data were not available for the western subspecies, but
the western subspecies is known to be more sensitive.
No data were available.
No comparable data were available; however this species is very non-sensitive
as measured by the seedling C-R function, so the risk of overestimating loss is
low.
No comparable data were available; however this species is very non-sensitive
as measured by the seedling C-R function, so the risk of overestimating loss is
low.
No data were available. This species is very sensitive as measured by the
seedling C-R function, so the risk of overestimating loss is high.
OTC studies found very consistent biomass loss between seedlings and adult
trees.
Seedling C-R functions overestimated RBL relative to results from TREGRO
and ZELIG, except the ZELIG results at the lowest W126 index values.
Seedling RBL underestimated loss relative to field results of changes in
circumference.
No comparable data were available; however this species is very non-sensitive
as measured by the seedling C-R function, so the risk of overestimating loss is
low.
                                                  6-9

-------
 2                   6.2.1.2   W126for Different levels of Biomass Loss
 3           The C-R functions can be plotted as a function of the percent biomass loss against
 4    varying W126 index values. This allows us to compare the W126 index values associated with a
 5    range of biomass loss values. Figure 6-4 and Figure 6-5 reflect two separate graphical
 6    representations of these results for trees and crops respectively.
 7           In each graph, the red line represents the median W126 index value associated with the
 8    percent biomass value on the x-axis when all 54 crop studies or 52 tree seedling studies are
 9    included. The green line is the value when only the composite C-R function is used for each of
10    the species included (10 crop species and 12 tree species). The grey lines are included as
11    sensitivity analyses to assess the effect of within-species variability. For each grey line, a C-R
12    function for each species was randomly selected from the available studies, with the resulting
13    line representing the median value of the 12 tree species  and 10 crops. For some species only one
14    study was available (e.g., Red Maple), and for other species there were as many as 11 studies
15    available (Ponderosa Pine). The process was repeated 1,000 times, and the median value is
16    plotted as the red points for biomass loss values of 1% to 7%,  and 10%. The error bar associated
17    with the points represents the 25th and 75th percentiles. For tree and crop species, the median
18    W126 index values are similar, when using all of the studies or just the composite C-R function
19    for each species; however, the median value is higher when within-species variability is
20    included.
21
                                                      6-10

-------
1

2   Figure 6-4
3
                    .1  8
                    -a

             i   i   i   i   ;   ;   i   i   i      i   i    •   i   '
         0   1   2   3  4   5   6   7   8   9   10  11  12  13  14  15

                          Percent Biomass Loss

W126 Index Values for Alternative Percent Biomass Loss for Tree Species
                r   i   i   i
                2   3  4   5
                                                     i   i      i   i   i   i   i
                                                     8   9  10  11  12  13  14  15
                                            Percent Biomass Loss
5   Figure 6-5     W126 Index Values for Alternative Percent Biomass Loss for Crop Species
                                                    6-11

-------
 1
 2                   6.2.1.3   Individual Species Analyses
 3          Using GIS (ESRI®, ArcMAP™ 10), we used the C-R functions listed in Table 6-1 to
 4    generate RBL surfaces for the 12 trees species. We created the surfaces using recent ambient 63
 5    conditions based on monitored data from 2006 through 2008 and the four Oj rollback surfaces
 6    simulating just meeting the existing 8-hr secondary standard of 75 ppb (4th highest daily
 7    maximum) and three alternative W126 scenarios of 7, 11 and 15 ppm-hrs (see Chapter 4 for a
 8    more detailed description of the Os surfaces). We present the maps for one species, Ponderosa
 9    Pine, to illustrate the results (see Figure 6-6, Figure 6-7, Figure 6-8, Figure 6-9, and Figure 6-10).
10    RBL surfaces for 10 species are presented in Appendix 6A (Maps of Individual Tree Species). It
11    is important to  note that these maps represent the RBL value for one tree species within each
12    CMAQ grid cell represented,  so these maps should be interpreted as indicating potential risk to
13    individual trees of that species growing in that area.
14          We based the ranges for the species on data from the Forest Health Technology
15    Enterprise Team (FHTET) of the USFS (http://www.fs.fed.us/foresthealth/technology/). These
16    data provide modeled predictions of stand density  and basal  area. The modeled data were
17    estimated in 1,000 square meter grids for individual tree species, as well as total basal area. We
18    summed these values into the larger CMAQ grid cells (12 km x 12 km) used for the Os surfaces.
19    For the individual  species analyses, these data were used only as a predictor of presence or
20    absence. In the ecosystem level analysis presented in Section 6.8 these values were used to scale
21    the biomass loss by the proportion of total basal area for each species.
22          Overall, the western tree species have more fragmented habitats than the eastern species.
23    The areas in southern California have the highest W126 index values, which can be seen as the
24    very high areas of RBL in Figure 6-6. The eastern tree species had less fragmented ranges and
25    areas of elevated RBL that were more easily attributed to urban areas (e.g., Atlanta, GA and
26    Charlotte, NC) or to the Tennessee Valley Authority region.  In addition to the two western
27    species not illustrated here, we include maps for the eastern species in Appendix 6A.
                                                     6-12

-------
                                Ponderosa Pine (Pinus ponderosa) (Recent Conditions)
                   RBL
                       I 0.006660-0.020314
                       | 0.020315-0032036
                       0,032037 • 0.045889
                       0.045890 - 0 062250
                       0062251 -0.082266
                       | 0082267-0.113164
                       | 0.113165-0.164S12
                       I 0.164813-0.243435
2
3
Figure 6-6      Relative Biomass Loss of Ponderosa Pine (Pinus ponderosa) Seedlings under
                 Recent Ambient W126 Index Values (2006 - 2008)
4
5
6
                                 Ponderosa Pine (Pinus ponderosa) (Current Standard)
                   RBL
                       | 0.000767 - 0.003953
                       0003954 -0005914
                       0,005915 - 0008482
                       0008483-0.011989
                       0,011990-0.015976
                       0,015977-0021390
                       | 0.021391 -0028521
                       I 0028522-0040549
Figure 6-7
Relative Biomass Loss of Ponderosa Pine with Os Exposure After
Simulating Meeting the Existing (8-hr) Primary Standard (75 ppb)
                                                            6-13

-------
2
3
4
Figure 6-8
6
7
8
9
Figure 6-9
                                  Ponderosa Pine (Pinus ponderosa) (15 ppm-hrs)
                   RBL
                       | 0000729-0003815

                       | 0 003816 • 0 005624
                       0005625-0007219
                       0007220-0009488

                       0009489-0012511
                       0 012512 -0015953
                       | 0015954-0020754

                       I 0 020755 - 0,032525
Relative Biomass Loss of Ponderosa Pine with Os Exposure After
Simulating Meeting an Alternative Secondary Standard of 15 ppm-hrs
(after Meeting Existing Os Standard)
                                  Ponderosa Pine (Pinus ponderosa) (11 ppm-hrs)
                   RBL
                       | 0 000687 .0 002635

                       | 0002635-0004073
                       0 004074 - 0 005543
                       0 005544 -0007192

                       0.007193-0.009559

                       0009560-0.012992
                       | 0012993-0019336

                       I 0019337-0032525
Relative Biomass Loss of Ponderosa Pine with Os Exposure After
Simulating Meeting an Alternative Secondary Standard of 11 ppm-hrs
(after Meeting Existing O3 Standard)
                                                          6-14

-------
1
                                  Ponderosa Pine (Pinus ponderosa) (7 ppm-hrs)
                 RBL
                     I 0.000643-0.002341
                     | 0.002342 • 0.003536
                     0.003537.0.004690
                     0.004691 - 0.005921
                     0005922-0.007591
                     0.007592-0.010159
                     | 0010160-0.014979
                     I 0014980-0.024900
2    Figure 6-10
4
Relative Biomass Loss of Ponderosa Pine with Os Exposure After
Simulating Meeting an Alternative Secondary Standard of 7 ppm-hrs (after
Meeting Existing Os Standard)
                                                          6-15

-------
                                                                         th
1   Table 6-6     Individual Species Relative Biomass Loss Values - Median, 75  Percentile, Maximum Percentages

Species
Red Maple
Sugar Maple
Red Alder
Tulip Poplar
Ponderosa Pine
White Pine
Loblolly Pine
Virginia Pine
Cottonwood
Aspen
Black Cherry
Douglas Fir
Relative Biomass Loss
(Median/75th Percentile/Maximum Percentages)
Recent Conditions
0.95/1.25/3.49
0.06/0.22/3.96
0.83/1.15/10.10
5.20/6.88/24.68
3.71/5.93/24.34
3.33/5.58/14.70
0.30/0.36/0.71
0.77/0.88/1.63
58.32/74.03/99.79
3.71/6.54/27.51
23.97/28.54/51.51
<0.01/<0.01/0.46
75ppb
0.08/0.17/0.77
<0.01/<0.01/0.07
0.32/0.40/0.78
0.17/0.35/2.79
0.67/1.18/4.05
0.10/0.40/2.66
0.05/0.07/0.17
0.15/0.20/0.54
5.93/11.97/65.90
0.47/1.14/5.85
4.89/7.94/23.90
<0.01/<0.01/<0.01
15 ppm-hrs
0.08/0.17/0.77
<0.01/<0.01/0.07
0.32/0.40/0.78
0.17/0.35/2.79
0.65/0.94/3.25
0.10/0.40/2.66
0.05/0.07/0.17
0.15/0.20/0.54
5.87/11.68/65.90
0.46/1.03/4.22
4.89/7.94/23.90
<0.01/<0.01/<0.01
11 ppm-hrs
0.08/0.13/0.70
<0.01/<0.01/0.01
0.32/0.40/0.78
0.12/0.21/2.40
0.56/0.69/3.25
0.10/0.30/2.05
0.05/0.06/0.15
0.12/0.16/0.50
5.26/8.06/53.33
0.45/0.82/3.89
4.51/6.31/19.42
<0.01/<0.01/<0.01
7 ppm-hrs
0.05/0.08/0.39
<0.01/<0.01/<0.01
0.31/0.39/0.78
0.05/0.09/0.93
0.50/0.58/2.49
0.09/0.17/1.60
0.04/0.05/0.09
0.08/0.10/0.32
3.74/5.06/35.29
0.43/0.72/3.03
3.41/4.41/13.68
<0.01/<0.01/<0.01
                                                6-16

-------
             Table 6-6 above includes individual species relative biomass loss values at the median,
      the 75th percentile, and the maximum for the 12 tree species for which we have C-R functions.
      We include the relative biomass loss values for each species at recent conditions, when adjusted
      to just meet the existing standard of 75 ppb, and when adjusted to meet potential alternative
      standard levels of 15, 11, and 7 ppm-hrs.2  For Ponderosa Pine, at recent conditions, the median
 6    value is 3.71 percent RBL, the 75th percentile value is 5.93 percent RBL, and the maximum
 7    value is 24.24 percent RBL.  When adjusted to just meet the existing standard, the median value
      is 0.67 percent RBL, the 75th percentile value is 1.18 percent RBL, and the maximum value is
      4.05 percent RBL; when adjusted to meet a potential alternative standard level of 15 ppm-hrs,
      the median value is 0.65 percent RBL, the 75th percentile value is 0.94 percent RBL, and the
      maximum value is 3.25 percent RBL; and when adjusted to meet a potential alternative standard
      level of 7 ppm-hrs, the median value is 0.50 percent RL, the 75th percentile value is 0.58 percent
      RBL, and the maximum value is 2.49 percent RBL.  In addition, RBL values for each scenario
      can be viewed  across the entire distribution within each species (Figure 6-11) or as a proportion
      of the current standard (Figure 6-12). Figure 6-11 and Figure 6-12 use Ponderosa Pine as an
16    example -  plots for the other 11 species are included in Appendix 6A.
 9
10
11
12
13
14
15
       W126 calculations are slightly modified in the case of the model adjustment scenarios described in Chapter 4,
      Section 4.3.4. When calculating W126 for the model adjustment cases, we first found the three-year average of each
      three-month period, and then selected the three-month period with the highest three-year average using the same
      three-month period for each of the three years. In this way, the five scenarios are for recent air quality, air quality
      adjusted to just meet the current standard, and air quality further adjusted to just meet three different W126 index
      values: 15 ppm-hrs, 11 ppm-hrs, and 7 ppm-hrs.
                                                       6-17

-------
                                                Recent Conditions
                             g
                             o -1
                            Jl
                               0.00      0.05      0.10      0.15      0.20       0.25
                                                 75 |>|>l> Scenario
                             g =
                             CD -
                             rffrf
                                0.00      0.01      0.02      0.03      0.04       0.05
                                                15 ppm-hr Scenario
                          o
                                  ^m
                          SH    0.00      0.01      0.02      0.03      0.04       0.05
                                               11 |>|>m -hi Seen,11 io
                                0.00      0.01      0.02      0.03      0.04       0.05
                                                7 |>|>m-lii Scenario
1
2
3
4
5
                                 J
                           0.00      0.01       0.02      0.03      0.04       0.05
                                 Relative Biomass Loss, Ponderosa pine

Figure 6-11    Relative Biomass Loss of Ponderosa Pine at the Existing Primary and
                Alternative Secondary Standards [RBL in this figure is plotted as a
                proportion relative to no Os exposure.]
                                                         6-18

-------
                                        Recent Conditions
                         o
                         o
I
k
1 1 1 1 1 1 1
0 10 20 30 40 SO 60
                      (D
                         o
                         § H
0.0
                                       15 ppm-hr Scenario
si
:1
i i
>1 0.0 0.2
0
c

-------
 1   used the values associated with just meeting the existing standard of 75 ppb.  Within each region
 2   we calculated both the W126 value at each monitor in the region for each year and the three-year
 3   average W126 value using the method described in Chapter 4. The results, depicted in Figure
 4   6-13 below, show that the use of the three-year average W126 index value may underestimate
 5   RBL values slightly, but the approach does not account for moisture levels or other
 6   environmental factors that could affect biomass loss.  Figure 6-14 shows the air quality data that
 7   was used in this analysis.  In both regions and in all three years, the three-year average W126
 8   value is sometimes above and sometimes below the individual year W126 index value.

                               3 Year Compounded Relative Biomass Loss
 9
10
ll
12
13
14
            CD
            
-------
                          Comparison of W126 Values, 75 ppb Scenario
       CD
       
  JJV
                                                                   10
                                                                          15
                                     3 Year Average W126
 1   Figure 6-14   Individual and 3-Year Average W126 Index Values - Southeast and
 2                 Southwest Regions
 3   6.3    COMMERCIAL TIMBER EFFECTS
 4          We used the Forest and Agricultural Sectors Optimization Model with Greenhouse Gases
 5   (FASOMGHG) (Adams et al., 2005) to calculate the resulting market-based welfare effects of Os
 6   exposure in the forestry and agriculture sectors of the United States under the scenarios outlined
 7   below. This section provides a summary of the results of those analyses. The current crop/forest
 8   budgets, which include all inputs to production and the resulting products, included in
 9   FASOMGHG are considered the budgets under recent ambient conditions. To model the effects
10   of changing W126 index values on the forestry sector, two primary and three alternative
11   scenarios were constructed and run through the model:
12          •   a base scenario, consistent with recent ambient conditions;
13          -a scenario with crop and forest yields for Os exposures after simulating just meeting
14             the existing standard of 75 ppb (4th highest daily maximum) and

15          •   three scenarios that represent O3 exposure after just meeting alternative W126-based
16             standard levels -  15,  11, and 7 ppm-hrs.
                                                   6-21

-------
 1          We used the 63 C-R functions for tree seedlings to calculate relative yield loss (RYL),
 2   which is equivalent to relative biomass loss, for FASOMGHG trees over their entire life span. To
 3   derive the FASOMGHG region-level RYLs for trees under each scenario, we used FASOMGHG
 4   region Os values along with the mapping in Table 6-7. For additional details on FASOMGHG,
 5   including a map of the FASOMGHG regions, see Appendix 6B (FASOMGHG Full Results).
 6          We calculate the FASOMGHG region-level RYLs for each tree species listed in the first
 7   column of Table 6-7 by extracting county-level W126 concentrations from the CMAQ air quality
 8   surfaces, using only the portion of each county that is identified as forested in the GIS data
 9   utilized and used the simple average across county Os values (forested portions of each county)
10   for all counties falling in a given FASOM region to represent the region-level Os impacts on
11   forests.  Then the region-level W126 O3 values are applied to tree species present in that region to
12   calculate RYLs.  Then, we calculate a simple average of RYLs for each tree  species mapped to a
13   FASOMGHG forest type in a given region. The mapping of tree species to FASOMGHG forest
14   types is based on "Atlas of United States Trees" (Little, 1971, 1976, 1977, 1978).
15
16
                                                    6-22

-------
 1   Table 6-7      Mapping O3 Impacts to FASOMGHG Forest Types
Tree Species Used for
Estimating O3 Impacts
Black Cherry, Tulip Poplar
Douglas Fir
Eastern White Pine
Ponderosa Pine
Quaking Aspen
Quaking Aspen, Black Cherry, Red Maple,
Sugar Maple, Tulip Poplar
Red Alder
Red Maple
Virginia Pine
Virginia Pine, Eastern White Pine
Virginia Pine, Eastern White Pine
FASOMGHG Forest Type
Upland Hardwood
Douglas Fir
Softwood
Softwood
Hardwood
Hardwood
Hardwood
Bottomland Hardwood
Natural Pine, Oak-Pine, Planted
Pine
Natural Pine, Oak-Pine, Planted
Pine
Softwood
FASOMGHG Region(s)
SC, SE
PNWW
CB,LS
PNWE, PNWW, PSW, RM
RM
CB, LS, NE
PNWE, PNWW, PSW
SC, SE
SC
SE
NE
 4
 5
 6
 7
 8
 9
10
11
12
13
14
       Table 6-8 presents the region-specific RYLs for the forest types by region.  At the
existing standard the highest yield loss occurs in upland hardwood forests in the South Central
and Southeast regions at over three percent per year. The next highest yield losses at the existing
standard occur in Corn Belt hardwoods with just over two percent loss per year and in hard- and
softwoods of the Rocky Mountain region at an average loss across all sensitive forests of slightly
over 1 percent loss per year. With the exception of the Rocky Mountain region, which has yield
losses reduced to under 1 percent per year, yield losses do not appreciably change at the 15 ppm-
hrs alternative.  This is primarily because most areas have W126 index values lower than 15
ppm-hrs after just meeting the existing standard. The Corn Belt forests remain at about 1.5
percent loss at 11 ppm-hrs and the South Central and Southeastern forests continue to experience
yield losses between 1 and 2 percent even after just meeting an alternative standard level of 7
ppm-hrs.
                                                    6-23

-------
1   Table 6-8     Percent Relative Yield Loss for Forest Types by Region for Modeled
2                 Scenarios
Forest Type
Douglas Fir
Natural Pine

Oak/Pine

Other Softwoods
Planted Pine

Softwoods





Bottomland Hardwoods

Hardwoods






Upland Hardwoods

Region
PNWW
SC
SE
SC
SE
PNWW
SC
SE
CB
LS
NE
RM
PSW
PNWE
SC
SE
PNWW
CB
LS
NE
RM
PSW
PNWE
SC
SE
Existing Standard
(75 ppb)
0.00
0.15
0.28
0.15
0.28
0.49
0.15
0.28
0.78
0.13
0.05
1.13
0.40
0.52
0.13
0.12
0.34
2.10
0.69
0.41
1.59
0.27
0.36
3.25
3.07
W126
15 ppm-hrs
0.00
0.15
0.28
0.15
0.28
0.48
0.15
0.28
0.78
0.13
0.05
0.91
0.36
0.50
0.13
0.12
0.34
2.10
0.69
0.41
1.27
0.25
0.35
3.25
3.07
11 ppm-hrs
0.00
0.12
0.24
0.12
0.24
0.48
0.12
0.24
0.46
0.13
0.04
0.64
0.31
0.48
0.10
0.10
0.34
1.51
0.69
0.33
0.88
0.22
0.34
2.71
2.79
7 ppm-hrs
0.00
0.09
0.13
0.09
0.13
0.48
0.09
0.13
0.23
0.13
0.02
0.53
0.28
0.47
0.06
0.06
0.33
0.98
0.67
0.25
0.73
0.19
0.33
2.00
1.85
                                                  6-24

-------
1   Table 6-9     Percent Relative Yield Gain for Forest Types by Region with Respect to the
2                 Existing Standard
Forest Type
Douglas Fir
Natural Pine

Oak/Pine

Other Softwoods
Planted Pine

Softwoods





Bottom Hardwoods

Hardwoods






Upland Hardwoods

Region
PNWW
SC
SE
SC
SE
PNWW
SC
SE
CB
LS
NE
RM
PSW
PNWE
SC
SE
PNWW
CB
LS
NE
RM
PSW
PNWE
SC
SE
W126
15 ppm-hrs - ES
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.23
0.04
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.35
0.03
0.01
0.01
0.01
11 ppm-hrs - ES
0.00
0.02
0.04
0.02
0.04
0.01
0.02
0.04
0.35
0.00
0.01
0.52
0.09
0.04
0.03
0.01
0.01
0.65
0.00
0.09
0.77
0.06
0.03
0.65
0.34
7 ppm-hrs - ES
0.00
0.06
0.16
0.06
0.16
0.01
0.06
0.16
0.59
0.00
0.02
0.63
0.13
0.05
0.06
0.06
0.01
1.22
0.02
0.17
0.93
0.09
0.04
1.48
1.48
                                                 6-25

-------
 1          Yield gains associated with meeting alternative W126 standards compared to meeting the
 2    existing standard are relatively small on a percentage change basis, especially in the 15 ppm-hrs
 3    scenario where the highest change is 0.35 percent per year.  At 11 ppm-hrs the yield gains are
 4    larger with gains between 0.35 and 0.77 percent for the most affected regions. The 7 ppm-hrs
 5    scenario generates yield gains between 0.59 and 1.48 percent for the Corn Belt, Rocky Mountain,
 6    South Central, and Southeast regions. These results are presented in Table 6-9 and  graphically in
 7    Figure 6-15 and Figure 6-16. While the yield gains for the alternative scenarios are small
 8    relative to the baseline of the existing standard, when applied nationally to forest production they
 9    result in increased forest production at every alternative in all years until the last period modeled
10    in 2040 as shown in Table 6-10. The change in relative yield between the existing standard and
11    the alternative scenarios results in changes in timber harvests and prices, as shown in Table 6-10.
12    In general, harvests increase and prices decrease with resulting changes in consumer and
13    producer welfare.
14
15
16
17
                                                      6-26

-------
     Legend
          FASOM Region
                3
1
2   Figure 6-15    RYG for Softwoods by Region
                                                6-27

-------
     Legend
         FASOM Region
     •A.

1
2   Figure 6-16   RYG for Hardwoods by Region
                                               6-28

-------
 1   Table 6-10    Percentage Changes in National Timber Prices
Product
Hardwood saw logs

15 ppm-hrs
11 ppm-hrs
7ppm-hrs
Hardwood pulp logs

15 ppm-hrs
11 ppm-hrs
7 ppm-hrs
Softwood saw logs

15 ppm-hrs
11 ppm-hrs
7 ppm-hrs
Softwood pulp logs

15 ppm-hrs
11 ppm-hrs
7 ppm-hrs
Policy
75 ppb
2010
0.69
2020
0.65
2030
0.39
2040
0.19
Change with Respect to Existing Standard



75 ppb
-0.28
-0.79
-1.59
0.24
0.13
0.13
-2.60
0.44
-0.16
-2.52
-8.72
0.22
0.94
-1.51
-7.12
0.12
Change with Respect to Existing Standard



75 ppb
0.00
-0.87
-2.10
2.31
-0.15
-1.95
-3.52
1.91
-0.08
-2.06
-4.92
1.60
-0.08
-2.64
-6.23
1.31
Change with Respect to Existing Standard



75 ppb
-0.09
-0.26
-0.46
1.42
-0.33
-1.24
-1.54
1.12
-0.44
-1.32
-1.91
1.34
-0.69
-1.40
-2.28
0.94
Change with Respect to Existing Standard



-0.14
-0.43
-1.03
0.12
0.13
-0.42
0.15
-0.19
-0.82
0.18
-0.51
-2.17
 2
 3
 4
 5
 6
 7
 8
 9
10
11
       Table 6-11 shows the estimated welfare changes brought about by the simulation
scenarios. Consumer and producer welfare in the forest sector are more affected by the
alternative scenario environments than the agricultural sector (see Section 6.5). In general,
consumer welfare increases in both the forest and agricultural sectors as higher productivity
tends to increase total production and reduce market prices. Because demand for most forestry
and agricultural commodities is inelastic, producer welfare tends to decline with higher
productivity as the effect of falling prices  on profits more than outweighs the effects of higher
production levels. In other words consumers do not increase their demand for the product enough
in response to the falling prices created by increases production to offset the producer's loss of
                                                     6-29

-------
1
2
3
4
revenue. The increase in consumer welfare is much larger than the loss of producer welfare
resulting in net welfare gains in the forestry sector nationally.
          Welfare economics focuses on the optimal allocation of resources and goods and how
   those allocations affect total social welfare. Total welfare is also referred to as economic surplus,
   which is the overall benefit a society, composed of consumers and producers, receives when a
   good or service is bought or sold, given a quantity provided and  a market price. Economic
   surplus is divided into two parts:  consumer and producer surplus.
          Consumers like to feel like they are getting a good deal on the goods and services they
   buy, and consumer surplus is an economic measure of this satisfaction. For example, assume a
   consumer goes out shopping for a CD player and he or she is willing to spend $250.  When the
   shopper finds that the CD player is on sale for $150, economists would say that this shopper has a
   consumer surplus of $100, e.g., the difference between the $150 sale price and the $250 the
   consumer was willing to spend.
          Producer surplus refers to the benefit a producer receives from providing a good or
   service at a market price when they would have been willing to sell that good or service at a lower
   price. For example, if the amount the producer is willing to sell  the CD player for is $75, and the
   producer sells the CD player for $150, the producer surplus is $75, e.g., the $150 sale price less
   the $75 price at which the producer was willing to sell.
                    Pnce
                                                               Supply
                                                              Demand
                                                       Quantity
                                                    6-30

-------
 1
 2
Table 6-11    Consumer and Producer Surplus in Forestry, Million $2010
Product
Consumer
surplus

15 ppm-hrs
11 ppm-hrs
7ppm-hrs
Producer
surplus

15 ppm-hrs
11 ppm-hrs
7 ppm-hrs
Policy
75ppb




75 ppb




2010
721,339
2015
793,234
2020
809,271
2025
826,375
2030
875,620
2035
894,705
2040
934,882
Change with Respect to Existing Standard
7
44
86
93,322
31
48
187
121,476
118
360
694
153,997
105
202
224
146,275
2
688
734
145,913
6
56
91
146,115
597
712
779
133,132
Change with Respect to Existing Standard
-11
-41
-136
-7
20
-48
-141
-503
-892
-161
-178
-37
15
-880
-786
-46
55
156
-839
-858
-766
 4          Key uncertainties in this approach are discussed in Section 6.6.1.  It should be noted that
 5    since public lands are not affected within the model, the estimates presented would likely be
 6    higher if public lands were included.3 See Appendix 6B for a full discussion of the model and
 7    methodology.

 8    6.4    NON-TIMBER FOREST PRODUCTS
 9          Non-timber forest products (NTFP) such as foliage and branches used for arts and crafts,
10    or edible fruits, nuts, and berries can be affected by the impact of Os through biomass loss, foliar
11    injury, insect attack, fire regime changes, and effects on reproduction. The USDA has assessed
12    the harvest and market value of these products in commercial markets (Emery, 2003).  A
13    significant portion of NTFP is also valuable to subsistence gatherers.  Subsistence practices are
14    much more difficult to assess because these forest users are not required to obtain a permit for
15    use of federal public lands; as such the impacts are more difficult to enumerate. Because permits
     3 The FASOMGHG model includes 348.6 million acres of private, managed forests. The USFS estimates that there
       are approximately 751 million forest acres in the United States (USDA, 2011).
                                                     6-31

-------
1
2
3
4
5
6
or contracts are not required for gathering activities for personal use the analyses done by USDA
are not able to account for the subsistence use of NTFP.
        In Table 6-12 we list some of the uses of the tree species known to be sensitive to the
effects of Os on biomass.  These species have a wide variety of uses ranging from the value of
the timber produced to medicinal uses.
7    Table 6-12     O3 Sensitive Trees and Their Uses
     Tree Species
                   O3 Effect
Uses
     Black Cherry
     Prunus serotina
                  Biomass loss,
                  Visible foliar
                  injury
Cabinets, furniture, paneling, veneers, crafts, toys
Cough remedy, tonic, sedative
Flavor for rum and brandy
Wine making and jellies
Food for song birds, game birds, and mammals
     Douglas Fir
     Pseudotsuga
     menziesii
                  Biomass loss
Commercial timber
Medicinal uses, spiritual and cultural uses for several Native American tribes
Spotted owl habitat
Food for mammals including antelope and mountain sheep
     Eastern
     Cottonwood
     Populus deltoides
                  Biomass loss
Containers, pulp, and plywood
Erosion control and windbreaks
Quick shade for recreation areas
Beaver dams and food
     Eastern White
     Pine
     Pinus strobus
                  Biomass loss
Commercial timber, furniture, woodworking, and Christmas trees
Medicinal uses as expectorant and antiseptic
Food for song birds and mammals
Used to stabilize strip mine soils
     Hemlock
     Tsuga canadensis
                  Biomass loss
Commercial logging for pulp
Habitat for deer, ruffled grouse, and turkeys
Important ornamental species
     Hickory
                  Biomass loss
Used in furniture and cabinets, fuelwood, and charcoal
Edible nuts
Food for ducks, quail, wild turkeys and many mammals
     Ponderosa Pine
     Pinus ponderosa
                  Biomass loss,
                  Visible foliar
                  injury
Lumber for cabinets and construction
Ornamental and erosion control use
Recreation areas
Food for many bird species, including the red-winged blackbird, chickadee,
finches, and nuthatches
                                                          6-32

-------
     Tree Species
                   O3 Effect
Uses
     Quaking Aspen
     Populus
     tremuloides
                   Biomass loss,
                   Visible foliar
                   injury
Commercial logging for pulp, flake-board, pallets, boxes, and plywood
Products including matchsticks, tongue depressors, and ice cream sticks
Valued for its white bark and brilliant fall color
Important as a fire break
Habitat for variety of wildlife
Traditional native American use as a food source
     Red Alder
     Alnus rubra
                   Biomass loss,
                   Visible foliar
                   injury
Commercial use in products such as furniture, cabinets, and millwork
Preferred for smoked salmon
Dyes for baskets, hides, moccasins
Medicinal use for rheumatic pain, diarrhea, stomach cramps - the bark
contains salicin, a chemical similar to aspirin
Roots used for baskets
Food for mammals and birds - dam and lodge construction for beavers
Conservation and erosion control
     Red Maple
     Acer rubrum
                   Biomass loss
Revegetation and landscaping especially riparian buffer
     Red Oak
     Quercus rubrum
                   Biomass loss
Important for hardwood lumber for furniture, flooring, cabinets
Food, cover, and nesting sites for birds and mammals
Bark used by Native Americans for medicine for heart problems, bronchial
infections or as an astringent, disinfectant, and cleanser
     Short Leaf Pine
     Pinus echinata
                   Biomass loss
Second only to loblolly pine in standing timber volume
Used for lumber, plywood, pulpwood, boxes, crates, and ornamental
vegetation
Habitat and food for bobwhite quail, mourning dove, other song birds and
mammals
Older trees with red heart rot provide red-cockaded woodpecker cavity trees
     Sugar Maple
     Acer saccharum
                   Biomass loss
Commercial syrup production
Native Americans used sap as a candy, beverage - fresh or fermented into
beer, soured into vinegar and used to cook meat
Valued for its fall foliage and as an ornamental
Commercial logging for furniture, flooring, paneling, and veneer
Woodenware, musical instruments
Food and habitat for many birds and mammals
     Virginia Pine
     Pinus virginiana
                   Biomass loss,
                   Visible foliar
                   injury
Pulpwood, strip mine spoil banks and severely eroded soils
Nesting for woodpeckers, food for songbirds and small mammals
1
2
     Yellow (Tulip)
     Poplar
     Liriodendron
     tulipifera
                   Biomass loss,
                   Visible foliar
                   injury
Furniture stock, veneer, and pulpwood
Street, shade, or ornamental tree - unusual flowers
Food for wildlife
Rapid growth for reforestation projects
Sources: USDA-NRCS, 2013; Burns, 1990; Hall and Braham, 1998.
                                                             6-33

-------
 1                6.4.1    Commercial Non-Timber Forest Products
 2                 In addition to timber, forests provide many other products that are harvested for
 3    commercial or subsistence activities.  These products include:
 4           •   edible fruits, nuts, berries, and sap,
 5           •   foliage, needles, boughs, and bark,
 6           •   transplants,
 7           •   grass, hay, alfalfa, and forage,
 8           •   herbs and medicinals,
 9           •   fuelwood, posts and poles, and
10           •   Christmas trees.
11           For the 2010 National Report on Sustainable Forests (USDA, 2011) these products were
12    divided into several categories including nursery and landscaping uses; arts, crafts, and floral
13    uses; regeneration and silviculture uses. Table 6-13 details selected categories of NTFP
14    harvested by permit in 2007. These harvests are reported in measures relevant to the specific
15    articles, i.e., bushels of cones, tons of foliage or boughs, or individual transplants.
16
17
                                                      6-34

-------
1
2
Table 6-13    Quantity of NTFP Harvested on U.S.
              Management Land
Forest Service and Bureau of Land
Product Category
Arts, crafts, and florals


Christmas trees

Edible Fruits, nuts, berries, and sap


Fuelwood

Grass, hay, and alfalfa
Forage
Herbs and medicinals
Nursery and landscape


Regeneration and silviculture




Posts and poles


Unit
Bushels
Pounds
Tons
Each
Lineal foot
Bushels
Pounds
Syrup Taps
ccf
Cords
Pounds
Tons
Pounds
Each
Pounds
Tons
Bushels
ccf
Each
Pounds
Tons
ccf
Each
Lineal foot
Harvest All U.S.
70,222
3,442,125
620,773
151,274
94.758
250
1,614,565
10,686
35,800
417,692
4,265,952
480
101,365
766,645
25,689
316
7,627
8
21,265
247,543
110,873
5,281
1,684,618
326,312
4
5
6
7
Note: ccf = 100 cubic feet  Source: USD A 2011

       According to the ISA, Os exposure causes biomass loss in sensitive woody and
herbaceous species, which in turn could affect forest products used for arts, crafts, and florals.
For example, Douglas Fir and Red Alder, among others, are used on the Pacific Coast for arts
and crafts, particularly holiday crafts and decorations. The effects of 63 on plant reproduction
                                                   6-35

-------
 1    (see ISA, Table 9-1, 2013) could affect the supply of seeds, berries, and cones.  Foliar injury
 2    impacts on (Vsensitive plants would potentially affect the harvest of leaves, needles, and
 3    flowers from these plants for decorative uses.  The visible injury and early senescence caused by
 4    Os in some evergreens may also reduce the value of a whole tree such as Christmas trees.
 5    Likewise the same Oj effects would reduce the harvest of edible fruits, nuts, berries, and sap.4
 6    The use of native grasses as forage is a significant aspect of forest-land management in the
 7    western U.S. (Alexander et al. 2002).  Os effects on community composition, particularly
 8    changes in the ratio of grasses to forbs (broad-leaved herbs  other than a grass), and nutritive
 9    quality of grasses can have effects on rangeland quality for  some herbivores (Krupa et al., 2004,
10    Sanz et al., 2005) and therefore effects on grazing efficiency.  The negative impacts of 63 on
11    plants would similarly affect the harvest in the rest of the categories.
12          According to the U.S. Census Bureau's County Business Patterns data from 2006, this
13    activity is captured in the industry code 1132 — forest nurseries and gathering of forest products -
14    -  and employed 2,098 people, accounting for an annual payroll of $71,657,000 ($2006) with an
15    average annual income of $34,155 (U.S. Census Bureau, 2006).
16          The USD A estimates the proportion of the national supply of NTFP represented by USFS
17    and BLM lands is approximately 10 percent.  Retail values for NTFPs harvested on USFS and
18    BLM lands are approximately $1.4 billion (2010$). These estimates are very rough and are based
19    only on permit or contract sales.  These estimates could be low due to harvests taken without
20    permit or contract and sold through complex commodity chains that can combine wild-harvested
21    and agriculturally grown commodities. It is important to note that while we cannot estimate the
22    loss of production and value to this sector due to Os exposures, these losses are already reflected
23    in the harvest and values reported.
24                6.4.2    Informal Economy or Subsistence Use of Non-Timber Forest Products
25          Most people gathering NTFPs are  doing so for personal use (Baumflek et al., 2010;
26    USD A, 2011).  By one estimate (Baumflek et al., 2010) up to 80 percent of the people collecting
27    NTFPs in Oregon and Washington are collecting for personal reasons.  Such personal use may be
28    characterized either as part of the informal economy or as subsistence activity. Participants in
29    the informal economy may earn a wage or salary and participate in gathering NTFPs for reasons
      1 To name a few, this category includes blueberries, pine nuts, and sap for maple syrup.

                                                     6-36

-------
 1    other than recreation (Brown et al., 1998). The term subsistence has usually been applied to
 2    special groups such as Native Americans or the Hmong people and has generally been
 3    understood to imply extreme poverty such that these activities are essential to the necessities of
 4    life (Freeman, 1993).  However, Freeman points out researchers stress that economic goals are
 5    only a part of the impetus for these activities.
 6           Brown et al. (1998) proposed a composite definition for the terms that captures both the
 7    informal economy, as practiced by those who are not necessarily a part of a special population,
 8    and subsistence, as generally referenced to those special populations.
 9
10           "Subsistence refers to activities in addition to, not in place of, wage labor engaged in on a
11           more or less regular basis by group members known to each other in  order to maintain a
12           desired and/or normative level of social and economic existence."
13
14    This definition allows consideration of the cultural and social aspects of subsistence lifestyles.
15    These non-economic benefits range from maintenance of social ties and relationships through
16    shared activity to family cohesiveness to retreatism and a sense of self-reliance for the individual
17    practitioner (Brown et al., 1998).
18           While there is general acknowledgement of subsistence activities by Native Americans
19    and specific treaty rights for tribes guaranteeing access to lands for hunting, fishing, and
20    gathering, there has been a lack of research focused on other populations (Emery and Pierce,
21    2005).  However, there are some studies that clarify that subsistence activities provide valued
22    resources for a variety of people in the coterminous United States.  Baumflek et al. (2010) and
23    Alexander et al.  (2011) have documented the collection and use of culturally and economically
24    important NTFPs in Maine and the eastern United States, respectively.  Brown et al. (1998)
25    reports on subsistence activities among residents of the Mississippi Delta.  Emery (2003) and
26    Hufford (2000) examine activities in the Appalachians, and Pena (1999) reports activities by
27    Latinos in the Southwest.
28           As with the commercial harvest of NTFPs, subsistence gathering of these forest products
29    can potentially be affected by the adverse effects of Os on growth,  reproduction, and foliar injury
30    to the sensitive plants in use for nutrition, medicine, cultural, and decorative purposes. It is
31    important to note that some plants may have more than one use or significance.  For example, the
                                                     6-37

-------
 1    Mi'kmaq and Maliseet Indian tribes in Maine do not differentiate between blueberries'
 2    nutritional, medicinal, and spiritual uses.  Blueberries are a food and a medicine that is often
 3    incorporated into ceremonies (Baumflek et al., 2010).  And while we cannot quantify the size of
 4    the harvest of subsistence-gathered items or monetize the loss of benefit due to Os effects, a
 5    comparison to the commercial harvest detailed in section 6.4.1 may provide perspective on the
 6    significance of these activities to the people who engage in them.

 7    6.5    AGRICULTURE
 8                6.5.1    Commercial Agriculture
 9          Because the forestry and agriculture sectors are related, and trade-offs occur between the
10    sectors, we used the same FASOMGHG model runs outlined in the forestry/timber section
11    (Section 6.3) to  calculate the resulting market-based welfare effects of 63 exposure in the
12    agricultural sector of the United States. This section provides a summary of the results of the
13    agricultural sector analyses. We have included results at the national scale for both sectors and
14    at the regional and subregional scale for agriculture. Table 6-14 defines the production and
15    market regions available in FASOMGHG. The regional-scale analysis provides an estimate of
16    the changes due to alternative levels of the standard for  63 subregions and indicates the disparate
17    results between  regions. The full model results, including a county-level analysis, are reported in
18    Appendix 6B.
19
20
                                                     6-38

-------
 1   Table 6-14    Definition of FASOMGHG Production Regions and Market Regions
Key
NE
LS
CB
GP
SE
SC
SW
RM
PSW
PNWE
PNWW
Market Region
Northeast
Lake States
Corn Belt
Great Plains (agriculture
only)
Southeast
South Central
Southwest (agriculture
only)
Rocky Mountains
Pacific Southwest
Pacific Northwest — East
side
Pacific Northwest — West
side (forestry only)
Production Region (States/Sub regions)
Connecticut, Delaware, Maine, Maryland, Massachusetts, New
Hampshire, New Jersey, New York, Pennsylvania, Rhode Island,
Vermont, West Virginia
Michigan, Minnesota, Wisconsin
All regions in Illinois, Indiana, Iowa, Missouri, Ohio (IllinoisN, IllinoisS,
IndianaN, IndianaS, lowaW, lowaCent, lowaNE, lowaS, OhioNW,
OhioS, OhioNE)
Kansas, Nebraska, North Dakota, South Dakota
Virginia, North Carolina, South Carolina, Georgia, Florida
Alabama, Arkansas, Kentucky, Louisiana, Mississippi, Tennessee,
Eastern Texas
Oklahoma, All of Texas but the Eastern Part (Texas High Plains, Texas
Rolling Plains, Texas Central Blacklands, Texas Edwards Plateau, Texas
Coastal Bend, Texas South, Texas Trans Pecos)
Arizona, Colorado, Idaho, Montana, Nevada, New Mexico, Utah,
Wyoming
All regions in California (CaliforniaN, CaliforniaS)
Oregon and Washington, east of the Cascade mountain range
Oregon and Washington, west of the Cascade mountain range
 4
 5
 6
 7
 8
 9
10
11
12
13
14
       Using the modeled W126 index values in each subregion under the scenarios, for crops,
we first calculated the RYL in the 63 subregions that have C-R functions. For those crops that do
not have C-R functions, we assign them RYLs for each scenario based on the crop proxy
mapping shown in Table 6-15. In addition, for oranges, rice, and tomatoes, which have 63 C-R
functions that are not W126-based (they are defined based on alternative measures of O3
concentrations), we directly used the median RYG values under the "13 ppm-hrs" Os
concentration reported in Table G-7 of Lehrer et al. (2007).  In  addition, we updated RYLs for
crops with county-level production data and specific C-R functions by using production-
weighting. Production weighting applies a county's share of the region's total production to the
average so that counties with less production have a smaller impact on the average.
       The RYLs for proxy crops were calculated for each FASOMGHG subregion so they
could be used in calculating the yield losses for other crops that occur in those regions.  The
                                                    6-39

-------
 1
 2
 3
 4
weighted RYLs that were used for corn, cotton, winter wheat (hard winter wheat and soft red
winter wheat), sorghum, and soybeans in the model scenarios were calculated only for their
production regions. The values calculated in all 63 regions were weighted by production for
these crops, which eliminated regions with no production.
Table 6-15    Mapping of O3 Impacts on Crops to FASOMGHG Crops
CROPS
FASOMGHG Crops
WU6 Crops
Corn
Cotton
Potatoes
Winter Wheat
Sorghum
Soybeans
Aspen (tree)
Corn
Cotton
Potatoes
Soft White Wheat, Hard Red Winter Wheat, Soft Red Winter Wheat, Durum Wheat, Hard Red
Spring Wheat, Oats, Barley, Rye, Wheat Grazing, and Improved Pasture
Sorghum, Silage, Hay, Sugarcane, Sugar Beet, Switchgrass, Energy Sorghum, and Sweet Sorj
jhum
Soybeans, Canola
Hybrid Poplar, Willow (FASOMGHG places short-rotation woody biomass production in the
sector rather than in the forest sector)
crop
Non-W126 Crops
Oranges
Rice
Tomatoes
Orange Fresh/Processing, Grapefruit Fresh/Processing
Rice
Tomato Fresh/Processing
 6
 7
 8
 9
10
11
12
13
14
15
16
17
       The following figures (Figure 6-17 and Figure 6-18) present the yield loss relative to the
existing standard and yield gains for corn and soybeans under the various adjusted air quality
scenarios. We are using corn and soybeans to illustrate some of the interactions that occur
between crop responses to 63 reductions, production, prices, producer cropping decisions, and
welfare effects for both producers and consumers.  For full model results for all crops included in
the analysis see Appendix 6B. In general, the RYL and RYG are unchanged between the
existing 75 ppb standard and the 15 ppm-hrs W126 scenarios.  Also, note that in many cases,
subregions that show no change in yield for a given crop have no production of that crop in that
subregion in FASOMGHG. For example, soybeans are relatively sensitive to Os and there are
large reductions in 63 in California, but there are no impacts on soybean yields in that region
because no soybeans are produced in California in FASOMGHG.
                                                    6-40

-------
 1          Corn is relatively insensitive to (Vinduced yield losses at the existing standard or 15
 2    ppm-hrs. The highest loss occurs in California at 0.88 percent, while in the Corn Belt, Lake
 3    States, and Great Plains the highest loss occurs in southern Ohio with 0.34 percent. Because the
 4    yield losses are small due to corn's insensitivity to Os under the alternative W126 standard
 5    scenarios, the yield losses are virtually eliminated at all three alternative W126 standards. Yield
 6    gains associated with the alternative scenarios are almost nonexistent; the highest gain occurs in
 7    Arizona at 0.02 percent at the 7ppm-hrs level.
 8          Soybeans, on the other hand, are relatively sensitive to (Vinduced yield losses.  The
 9    highest losses at the existing  standard or 15 ppm-hrs occur in Colorado, southern Indiana,
10    Kentucky, and northwest Ohio at over 1 percent.  Yield losses remain under all scenarios for
11    W126, although for the 7 ppm-hrs scenario all losses are less than 0.6 percent.  Yield gains
12    across the alternative W126 standard levels generally range between 0.54 percent and 0.84
13    percent with northeast Ohio,  Tennesse, Kentucky, Illinois, and Indiana on the high end.
14    Colorado has the highest gain at 1.01 percent at the 7 ppm-hrs level and most soybean producing
15    states have at least small gains at every W126 scenario.
                                                      6-41

-------
     Legend
          FASOM Region

          FASOM Subregion


          °%    ?

         ^    3




         \    I
                2.
                Q.
     o»  q.
                O

                O
                c
                EC
                OI
                3
                a.


                I
1
2   Figure 6-17    Percentage Changes in Corn RYG with Respect to 75 ppb
                                                    6-42

-------
     Legend
     |    | FASOM Region
     |    | FASOM Subregion
1
2   Figure 6-18    Percentage Changes in Soybean RYG with Respect to 75 ppb
                                                6-43

-------
 1           In general, increased yield leads to increased supply and lower prices. Because corn does
 2    not lose or gain very much under any scenario one could expect that prices would remain
 3    relatively stable.  Soybeans, however, would experience yield gains in any scenario, and prices
 4    would likely fall. In the modeled scenarios soybean prices fall, and since consumer demand does
 5    not increase enough to offset the loss of revenue due to price decreases there is a net decrease in
 6    producer welfare, but consumers always benefit from falling prices. In response to falling
 7    soybean prices, the model predicts that producers would switch to less Os-sensitive crops with
 8    stable prices, such as corn, thereby increasing corn production.  See Appendix 6C for an
 9    explanation of the supply curve shift.
10           Overall, across the full agriculture sector, these changes in production are small, seldom
11    above 0.5 percent and usually 0.01 percent or less. The production increases lead to generally
12    lower prices, with price decreases greater than the change in production.  The drop in market
13    prices, while a loss for producers, represents a gain for consumers. In terms of producer and
14    consumer welfare across the agriculture sector, in nearly all cases producer welfare is negatively
15    affected.  Table 6-16 presents the overall  welfare gains and losses. For producers, the W126
16    alternatives occasion welfare gains in the middle years, 2020-2030, and welfare losses in all
17    other years. For consumers, however, the changes in production and prices occasion welfare
18    gains in all scenarios in all years.
19           Since the forestry and agriculture  sectors are interlinked  and factors affecting one sector
20    can lead to changes in the other, it is important to consider the overall effect of O3 changes in  the
21    context of producer and consumer welfare across both sectors.  The impacts on consumer surplus
22    are positive for both sectors, with benefits increasing with lower W126 alternatives. For producer
23    surplus, however, impacts are negative for the 15 ppm-hrs  and 11 ppm-hrs scenarios and  positive
24    for the 7 ppm-hrs case.  Table 6-17 presents the annualized surplus for both sectors.
25
26
                                                      6-44

-------
1
2
Table 6-16    Consumer and Producer Surplus in Agriculture (Million 2010$)
Product
Consumer
surplus




Producer
surplus




Policy
75ppb

15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
75ppb

15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
2010
1,918,082
2015
1,940,673
2020
1,968,142
2025
1,995,346
2030
2,023,022
2035
2,050,791
2040
2,076,018
Change with Respect to Existing Standard
15
19
-31
725,364
-2
24
46
831,565
1
13
36
815,072
6
51
104
863,165
-7
42
90
878,986
10
20
26
836,692
3
13
46
863,308
Change with Respect to Existing Standard
612
1,474
269
-1,255
-2,197
-1,873
980
1,013
1,780
-961
230
423
90
232
264
41
-3,413
-1,052
697
2,189
2,991
                                               6-45

-------
 1   Table 6-17    Annualized Changes in Consumer and Producer Surplus in Agriculture and
 2                 Forestry, 2010-2040, Million 2010$ (4% Discount Rate)
Product
Consumer surplus
Policy
75 ppb
Agriculture
NA
Forestry
NA
Total
NA
Change with Respect to Existing Standard



Producer surplus
15 ppm-hrs
1 1 ppm-hrs
7ppm-hrs
75 ppb
4.5
25.4
36.7
NA
88.1
236.9
344.0
NA
92.5
262.3
380.7
NA
Change with Respect to Existing Standard



Total surplus
15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
75 ppb
-4.7
-4.6
194.4
NA
-112.2
-264.4
-318.4
NA
-116.9
-269.0
-124.0
NA
Change with Respect to Existing Standard



15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
-0.2
20.8
231.1
-24.2
-27.5
25.6
-24.4
-6.7
256.7
 4   6.6    CLIMATE REGULATION
 5          Biomass loss due to Os exposure affects climate regulation by ecosystems by reducing
 6   carbon sequestration and storage.  More carbon stays in the atmosphere because carbon uptake
 7   by forests is reduced. The studies cited in the ISA demonstrate a consistent pattern of reduced
 8   carbon uptake because of 63 damage, with some of the largest reductions proj ected over North
 9   America. In one simulation (Sitch et al., 2007) the indirect radiative forcing due to 63 effects on
10   carbon uptake by plants are shown as even greater than the direct effect of 63 on climate change.
11               6.6.1    National Scale Forest Carbon Sequestration
12          FASOMGHG can calculate the difference in carbon sequestration by forests and
13   agriculture  due to biomass loss caused by 63 exposure. By comparing equilibriums under the
14   different scenarios outlined in Section 6.3, we can calculate changes in carbon sequestration
                                                    6-46

-------
 1    potential over time.  Details of FASOMGHG and the methodology for the analyses done for this
 2    risk and exposure assessment are available in Appendix 6B.
 3           The impacts of the simulations of meeting the existing and alternative secondary O3
 4    standards on carbon sequestration potential in U.S. forest and agricultural sectors are presented
 5    in Table 6-18, where numbers indicate increased sequestration. As  shown in the table, much
 6    greater sequestration changes are projected in the forest sector than in the agricultural sector. The
 7    15 ppm-hrs scenario does not appreciably increase carbon storage relative to just meeting the
 8    existing standard. The vast majority of the enhanced carbon sequestration potential under the
 9    alternative secondary standard scenarios lies in the forest biomass increases over time at the 11
10    and 7 ppm-hrs standard levels.  The forest carbon sequestration potential would increase between
11    593 and 1,602 million tons of CC>2 equivalents over 30 years after meeting the 11 or 7 ppm-hrs
12    standard level, respectively, compared to just meeting the existing Os standard.  On an annual
13    basis when just meeting the 11 ppm-hrs W126 standard level, total  forestry and agriculture
14    carbon storage is increased by about 20 million tons per year relative to just meeting the existing
15    Os standard; equivalent to taking about 4 million cars off the road as calculated by the EPA
16    Greenhouse Gas Equivalencies Calculator5 (U.S. EPA, 2013b).  When meeting the 7 ppm-hrs
17    W126 standard level, the increased annual carbon storage is about 53 million tons relative to just
18    meeting the existing 63 standard, or approximately 11 million fewer cars on the road.
19           The baseline stock of carbon storage decreases over time for agriculture because the
20    agriculture sector GHG emissions sources are released every year and soil carbon sequestration
21    stabilizes over the 30-year period. There are only small increases in net carbon storage compared
22    to the existing standard for each of the alternative scenarios modeled.
      1 Available at http://www.epa.gov/cleanenergv/energv-resources/calculator.html.

                                                      6-47

-------
 1   Table 6-18    Increase in Carbon Storage, MMtCO2e, Cumulative over 30 years
Product
Forestry




Agriculture




Policy
75 ppb
2010
74,679
2020
79,171
2030
84,863
2040
89,184
Change with Respect to Existing Standard
15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
75 ppb
1
19
50
18,748
0
103
305
15,363
16
312
832
12,002
13
593
1,602
8,469
Change with Respect to Existing Standard
15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
0
2
3
1
5
4
1
6
6
4
10
9
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

14
15
16
17
18
19
Key uncertainties in using the FASOMGHG in these analyses include:
•   Although the modeling system applied builds on existing models that have previously
    been used for assessments of Os impacts and reflects what we consider reasonable
    and appropriate assumptions, it is very important to recognize the considerable
    uncertainties and limitations surrounding the results of this study or any study
    assessing the potential impacts of changes in 63 concentrations on forest and
    agricultural production. First, the changes in W126 index values being used to
    calculate the agricultural and forest productivity responses are assumed to equal point
    estimates taken as exogenous to the economic modeling. However, these changes in
    the W126 index values were calculated using air quality simulation models where (as
    with any model) model parameter values are not known with certainty.

    Second, the  63 C-R functions applied to crops and trees used the median parameters
    from Lehrer et al. (2007)—the RYLs and RYGs calculated are thus representative of
    these median values, whereas there is actually a range of responsiveness to 63. Using
    alternative "low" and "high" O3 CRs would present data inputs that exhibit lower or
    higher Os impacts on crop and tree species biomass productivity, and thus the
    magnitude of exogenous environmental shocks under different policy scenarios. The
                                                    6-48

-------
 1              changing of the magnitude of exogenous shocks on U.S. agriculture and forestry
 2              systems would potentially lead to different economic equilibrium outcomes,
 3              especially if the shocks go beyond the buffering or adjustability of the system.

 4          •   Third, the use of crop proxy mapping and forest type mapping due to incomplete data,
 5              as specified in Section 4, adds to the uncertainty of these model results to the extent
 6              that actual crop-specific impacts differ from those of the proxy crop used. The current
 7              mappings of crops/tree species that have Os C-R functions, which are a subset of
 8              crops/tree species that are present in U.S. agriculture and forestry systems, present
 9              probable  "omission" biases. In particular, forest types that vary by region may have
10              been underrepresented by just a handful of tree species that have 63 C-R functions
11              specified. Moreover, due to data limitations, we are using a simple average of tree
12              RYLs for all forest types within a region, which is an imperfect estimate. For
13              instance,  in southern regions, poplar is far more common than black cherry for
14              hardwood forests.

15          •   Fourth, the potential changes in tree species mixes within a forest type that would be
16              made by landowners due to differential impacts associated with ground-level 63
17              exposure changes were not considered. Tree species that are less susceptible to
18              ground 63 damage may gain relative advantage over tree species that are more
19              sensitive  to ground 63. Thus, as time moves forward, the 63 impacts on forests may
20              get ameliorated because forests adapt to O3 environments - whether via forestry
21              industry management or through natural  processes.

22          •   Fifth, the international trade component in FASOMGHG assumes USDA-based
23              future projections under recent Os conditions.  This may present another uncertainty
24              for the model  results, especially when soybeans and wheat are among the major crop
25              commodities for U.S. exports and have relatively large responses to changed 63
26              environments. As a result, the exogenous RYGs obtained under 63 policy scenarios
27              for these  crops would present a potentially enhanced supply advantage, relative to
28              soybeans and  wheat produced in the rest of world. The general trade projections thus
29              may need to reflect these potential changes.
                                                     6-49

-------
 1           •   Finally, there is a very large number of parameters contained within the
 2              FASOMGHG modeling system, introducing further uncertainty regarding the best
 3              values to use for each parameter as well as potential interactions between parameters.

 4           To summarize, the uncertainty in crop and tree species' Os C-R functions, whether the
 5    crops/forest types are well-represented, the possibly of changes in such representations due to
 6    adaptation, and the potential changes in international trade for Os-sensitive crops present the
 7    known uncertainties for the model results. Careful consideration and sound judgment related to
 8    the potential implications of uncertainties is important for appropriate interpretation of model
 9    results.
10           In addition, it  should be noted that since public lands are not affected within
11    FASOMGHG the estimates presented would likely be higher if public lands were included.
12               6.6.2    Urban Case Study Carbon Storage
13           Urban forests  are subject to the adverse effects of Os exposure in the same ways as
14    forests in rural areas.  These urban forests provide a range of ecosystem services such as carbon
15    sequestration, pollution removal, building energy savings, and  reduced stormwater runoff.  The
16    analyses in this section focus on carbon sequestration. Pollution removal  services  are discussed
17    in section 6.7.  The i-Tree  model6 used in this analysis is a peer-reviewed suite of software tools
18    provided by USFS. We used data from five urban areas to estimate the effects of 63 (based on
19    CMAQ modeled W126 index surfaces) on carbon storage. We used the i-Tree Forecast model to
20    estimate tree growth and ecosystem services provided by trees  over a 25-year period, using for
21    the base case the measured inventory of trees in the area and standard growth rates over the 25-
22    year period. We adjusted the tree growth downward from the base case using the reduced
23    growth factors for the species present in each area for which we have C-R functions (only
24    species with W126 C-R functions were reduced). Unlike the FASOMGHG model, C-R
25    functions were not assigned to species in the study areas that do not have  specific C-R functions
26    available from the literature because the model does not account for dynamic interactions in the
27    community composition based on increased or decreased competitiveness of the species present.
28    We contrasted the differences between the scenarios for the 25-year period.  We ran  six scenarios
      6 Available at http://www.itreetools.org/.
                                                     6-50

-------
 1    simulating a scenario without (Vinduced changes in biomass, recent ambient conditions, a
 2    simulation of "just meeting" the existing standard, and just meeting three alternative W126
 3    standards of 15, 11, and 7 ppm-hrs. The model assumed an annual influx of between one and six
 4    trees/hectare/year and a three to four percent annual mortality rate. See Appendix 6D for details
 5    of the model  and the methodology employed for these case studies.
 6          We chose the five urban areas based on data availability and presence of species with a
 7    W126 C-R function. No urban area with available vegetation data had more than three qualified
 8    species present. The selected study areas include Baltimore, Syracuse, the Chicago region,
 9    Atlanta, and the urban areas of Tennessee. Table 6-19 shows details of the tree species present,
10    the percent of sensitive trees in the top ten species present, and the percent of sensitive trees in
11    the total species in each study area.
12
                                                     6-51

-------
1   Table 6-19    Tree Species with Available C-R Functions in Selected Urban Study Areas
Study Area
Top Ten
Occurring
Species
1
2
3
4
5
6
7
8
9
10
Species w/ C-R
Function %of
Top 10
Species w/ C-R
Function % of
Total Trees
Baltimore
American beech
Black locust
American elm
Tree of heaven
White ash
Black cherry
White mulberry
Northern red
oak
Red maple
White oak
8.5
11.2
Syracuse
European
buckthorn
Sugar maple
Black cherry
Boxelder
Norway maple
Northern white
cedar
Norway spruce
Staghorn sumac
Eastern
cottonwood
Eastern
hophornbeam
18.5
20.2
Chicago Region
European
buckthorn
Green ash
Boxelder
Black cherry
Hardwood
American elm
Sugar maple
White ash
Amur
honeysuckle
Silver maple
7.7
10.5
Atlanta
Sweetgum
Loblolly pine
Flowering
dogwood
Tulip tree
Water oak
Boxelder
Black cherry
White oak
Red maple
Southern red oak
6.6
8.9
Tennessee
Chinese privet
Virginia pine
Eastern red
cedar
Hackberry
Flowering
dogwood
Amur
honeysuckle
Winged elm
Red maple
Black tupelo
American beech
9.3
17.4
    Bold - species with C-R function, Italics - species known to be sensitive, no C-R function
                                                      6-52

-------
 1
 2          The largest differences in the modeled air quality are between the recent ambient
 3    conditions and meeting the existing standard.  The distribution of O3 air quality is not changed in
 4    most areas in the eastern U.S. when simulating meeting an alternative W126 standard of 15 ppm-
 5    hrs relative to the scenario of just meeting the existing Os standard. There are small incremental
 6    differences between just meeting the existing Os standard and just meeting alternative W126
 7    standards of 11 and 7 ppm-hrs.
 8          The model results for changes in carbon storage show substantial reductions in the
 9    capacity of these urban forests to sequester carbon for the simulation of "just meeting" the
10    existing  standard. Estimates for the five modeled areas at the existing standard or an alternative
11    standard of 15 ppm-hrs are about 3.5 million tons of carbon storage lost over 25 years (about
12    140,000 tons /year). At an alternative standard of 11 ppm-hrs, loss of carbon sequestration is
13    128,000 metric tons per year, and at an alternative standard of 7 ppm-hrs, the estimated loss is
14    112,000 metric tons per year of carbon storage services.
15          Three of the urban areas show gains in carbon storage at alternative W126 standards
16    below 15 ppm-hrs.  Syracuse and Baltimore do not realize gains because they are currently very
17    close to meeting the alternative standards.  Of the five areas modeled, the combined urban areas
18    of Tennessee have the largest estimated gains in carbon storage at almost 20,000 tons per year
19    when meeting the alternative standard of 7 ppm-hrs. The Chicago region gains about 6,400 tons
20    per year of carbon sequestration when meeting the alternative standard of 7 ppm-hrs. See Table
21    6-20 for details.
22          Compared to other activities, the yearly carbon  storage gains at 11 ppm-hrs for Atlanta
23    are only equivalent  to taking 50 cars per year off the road or  recycling about 90 more tons of
24    waste every year. At the 7 ppm-hrs standard level, Atlanta would need to remove 250 cars per
25    year to be equivalent to the gains from reduced 63.  The Chicago region would need to take 417
26    cars per year off the road. At 7 ppm-hrs, Chicago would need to remove more than 1,300 cars.
27    The urban areas of Tennessee would need about 1,800 fewer cars per year at the 11 ppm-hrs
28    standard level. To reach the carbon sequestration provided by the urban forests in Tennessee at
29    the 7 ppm-hrs standard level, Tennessee would need 4,000 fewer cars every year.
30          Baltimore and Syracuse would realize no gains  at the alternative standard levels chosen
31    for this analysis. Chicago and Atlanta are in the middle of the range of results.  In Tennessee, at
                                                     6-53

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
recent ambient conditions, the urban areas are all above a W126 standard of 15 ppm-hrs and
comprise a much larger area than the other four case study areas with a far larger tree population.
Thus the relative gains in carbon storage in Tennessee are far larger than the other case study
areas. Keeping in mind that of the 11 tree species for which we have C-R functions, only two to
three species were present in a given area comprising at most 18.5 percent of the total trees
present. It seems reasonable to conclude that the actual effect on carbon storage because of Os
exposure would be higher than the estimates modeled here.
       These results should not be combined with the results from the FASOMGHG model
discussed in Section 6.7.1. The methodology employed for the FASOMGHG runs assigned
values for 63  exposure C-R functions for species that do not have a function calculated in the
ISA. We did  this to ensure the dynamic trade-offs in the model functioned properly.  The i-Tree
model does not provide trade-offs between species,  so the species that do not have a C-R
function were not assigned values. This could lead  to an underestimation of the carbon storage
losses in i-Tree if the other species in the  study area are sensitive to Os exposure effects.
Alternatively  assigning C-R functions to species as  we did for the FASOMGHG runs would
likely produce an overestimation since many species, even within the same genus, may not be
sensitive to 63 effects.
Table 6-20     Os Effects on Carbon Storage for Five Urban Areas over 25 Years (in
              millions of metric tons)*
Region
Atlanta
Baltimore
Chicago
Region
Syracuse
Tennessee
Total
No03
Adjustment
(NOA)
1.426
0.578
19.560
0.169
20.568
42.302
Existing
Standard/15
ppm-hrs
(ES/15)
1.315
0.571
17.053
0.169
19.668
38.607
ES/15 v
NOA
-0.112
-0.007
-2.508
-0.0015
-0.900
-3.528
11 ppm-
hrs V
NOA
-0.106
-0.007
-2.457
-0.0015
-0.676
-3.247
7 ppm-
hrs v BC
-0.081
-0.007
-2.346
-0.0015
-0.410
-2.845
ESv
llppm-
hrs
0.006
0.00
0.05
0.00
0.22
0.276
ESv 7
ppm-hrs
0.03
0.00
0.16
0.00
0.49
0.68
21
       ES = Existing standard
                                                     6-54

-------
 1          In addition to its direct impacts on vegetation, O3 is a well-known GHG that contributes
 2    to climate warming (U.S. EPA, 2013). A change in the abundance of tropospheric O3 perturbs
 3    the radiative balance of the atmosphere, an effect quantified by the radiative forcing metric. The
 4    IPCC (2007) reported a radiative forcing of 0.35 W/m2 for the change in tropospheric O3 since
 5    the preindustrial era, ranking it third in importance after the greenhouse gases CC>2 (1.66 W/m2)
 6    and methane (CH4) (0.48 W/m2). The earth-atmosphere-ocean system responds to the radiative
 7    forcing with a climate response, typically expressed as a change in surface temperature. Finally,
 8    the climate response causes downstream climate-related ecosystem effects, such as redistribution
 9    of ecosystem characteristics because of temperature changes. While the global radiative forcing
10    impact of O3 is generally well understood, the downstream effects of the O3-induced climate
11    response on ecosystems remain highly uncertain.
12          Since O3 is not emitted directly but is photochemically formed in the atmosphere, it is
13    necessary to consider the climate effects of different O3 precursor emissions.  Controlling
14    methane, CO, and non-methane VOCs may be a promising means of simultaneously mitigating
15    climate change and reducing global O3 concentrations (West et al. 2007). Reducing these
16    precursors reduces global concentrations of the hydroxyl radical (OH), their main sink in the
17    atmosphere, feeding back on their lifetime and further reducing O3 production. NOx  reductions
18    decrease OH, leading to increased methane lifetime and increased O3 production globally in the
19    long-term. The resulting positive radiative forcing from increased methane may cancel or even
20    slightly exceed the negative forcing from decreased O3 globally (West et al. 2007). Of the O3
21    precursors, methane abatement reduces climate forcing most per unit of emissions reduction, as
22    methane produces O3 on decadal and global scales and is itself a strong climate forcer. Since
23    they may have different effects on concentrations of different species in the atmosphere, all O3
24    precursors must be considered in evaluating the net climate impact of emission sources or
25    mitigation strategies.

26    6.7    URBAN CASE STUDY AIR POLLUTION REMOVAL
27          In addition to sequestering and storing carbon, urban forests also remove pollutants from
28    the local atmosphere. The reduction in growth rates resulting from O3 exposure would reduce the
29    current and future amount of pollutants removed by these forests. We used the i-Tree model
                                                     6-55

-------
 1    described in Section 6.5.2 to estimate the removal of air pollutants by the forests in the urban
 2    areas discussed.
 3          The preliminary results for changes in air pollution removal estimates for carbon
 4    monoxide, nitrogen dioxide, Os, and sulfur dioxide show reduced capacity for these urban forest
 5    canopies to remove pollution (1) at recent ambient Os conditions and (2) after adjusting air
 6    quality to just meeting the existing standards and alternative standards.  These analyses show that
 7    even at the lowest scenario urban forest capacity to remove pollution is still reduced compared to
 8    a no ozone scenario. Because of the limitations in the availability of C-R functions for all of the
 9    common tree species in urban areas, and because of the limited number  of urban areas for which
10    the i-Tree model has been applied, these reductions only reflect a portion of the impacts on
11    pollution removal by urban forests in the U.S.  Though the model does include estimates for
12    particulate matter (PM), we do not include those estimates because the model does not yet
13    distinguish between PMio and PM2.5, and this distinction is important for evaluating the potential
14    health and welfare effects associated with PM. Estimates suggest that after meeting the existing
15    standard about 1,535 tons of air pollution removal capacity is lost annually (or about 38,384 tons
16    over 25 years) in  the five areas modeled. As in the simulations for carbon storage, Syracuse and
17    Baltimore see the least change in capacity with the urban areas of Tennessee reporting the largest
18    changes.  Syracuse and Baltimore have no change in pollution removal when meeting the
19    existing and the modeled alternatives. Atlanta and Chicago gain about 470 and 6,500 metric tons
20    of additional pollution removal after meeting the alternative W126 standard of 7 ppm-hrs
21    compared to meeting the existing standard, while Tennessee gains almost 12,000 tons of
22    potential pollution removal annually for the same comparison. For the 7 ppm-hrs scenario, about
23    51 percent of the  pollution removal capacity lost under the existing standard is regained. See
24    Table 6-21 for details.
25          We performed a simple analysis of the 63 removal potential to show how this process
26    might affect ambient air quality values. The analysis makes some general assumptions to
27    estimate order of magnitude effects of 63 removal by trees on 63 concentrations in the five urban
28    areas.  To make this calculation, the metric tons of O3 removed listed in Table 6-20 are spread
29    evenly over every hour in the 25-year tree lifetime to achieve an hourly  Os removal. Using the
30    ideal gas-law, this mass can be converted to an equivalent volume of gas assuming standard
31    temperature and pressure. Each urban area is treated as a well-mixed volume with the height
                                                     6-56

-------
 1   determined as the average maximum daytime boundary layer height7 extracted from an April-
 2   October 2007 Weather Research Forecasting (WRF) model simulation for each area of interest.
 3   The ratio of the O3 volume to the urban area air volume multiplied by 109 gives an equivalent
 4   concentration in ppbv. Table 6-21 shows that the effects on Os concentration are generally
 5   small; deposition to tree surfaces results in ambient Os concentration reductions ranging from
 6   0.08 ppbv in Tennessee to 0.52 ppbv in Chicago. Differences between the scenarios are minute.
 7   The base case numbers are consistent with previously published values from Song et al. (2008)
 8   who used a photochemical model to show that changes in land use from development in Austin,
 9   TX, might lead to a 0-0.3 ppbv change in 63 concentration due solely to deposition differences.
10   Some additional benefit may be achieved from cumulative effects, which are not accounted for
11   here (i.e., O3 removed at 9am will not only decrease concentrations instantaneously, but will  also
12   decrease the  starting concentration to some degree at 10am,  1 lam, etc. throughout the day). In
13   addition, changing the boundary layer height based on variability in this value could increase or
14   decrease the  ppbv estimates by a factor of two. But in any case, the values would still be small.
15
     7 The maximum daytime boundary layer height is the depth in the atmosphere over which air is well-mixed in the
       afternoon. The WRF modeling simulation showed that this depth was approximately 1700m in Atlanta, 1500m in
       Baltimore, 1150m in Chicago, 1350m in Syracuse, and 1750m in Tennessee.

                                                     6-57

-------
1   Table 6-21  Comparison of Pollutant Removal Between an Unadjusted Scenario and
2                Alternative Simulations and Gains Between the Existing Standard and
3                Alternatives (metric tons)

No O3 Adjustment
(NOA)
Existing
Standard/15
(ES/15)
NOA
V
ES/15
NOA v
11 ppm-
hrs
NOAv
7ppm-
hrs
ES/15 v
11 ppm-
hrs
ES/15 v
7 ppm-
hrs
CO
Atlanta
Baltimore
Chicago
Syracuse
Tennessee
1,482
186
8,620
55
12,854
1,429
186
8,001
55
12,626
-54
0
-619
0
-227
-50
0
-569
0
-97
-34
0
-476
0
62
3
0
142
0
131
9
0
235
0
290
NO2
Atlanta
Baltimore
Chicago
Syracuse
Tennessee
6,852
1,968
104,247
50
54,381
6,605
1,963
96,766
50
53,419
-248
-5
-7,481
0
-962
-231
-5
-6,883
0
-408
-159
-5
-5,758
0
263
16
5
598
0
554
88
5
1,723
0
1,226
03
Atlanta
Baltimore
Chicago
Syracuse
Tennessee
25,495
6,262
243,701
1,544
393,205
24,574
6,247
226,214
1,541
386,247
-921
-15
-17,488
-4
-6,957
-861
-15
-16,090
-4
-2,953
-591
-15
-13,460
-4
1902
60
0
1,398
0
4,004
331
0
4,028
0
8,860
S02
Atlanta
Baltimore
Chicago
Syracuse
Tennessee
3,380
852
29,675
71
59,371
3,257
850
27,546
71
58,320
-122
-2
-2,129
0
-1,050
-114
-2
-1,959
0
-446
-78
-2
-1,639
0
287
8
0
170
0
605
44
0
490
0
1,338
Total
Atlanta
Baltimore
Chicago
Syracuse
Tennessee
37,209
9,268
386,242
1,721
519,810
35,865
9,246
358,527
1,717
510,613
-1,344
-22
-27,817
-4
-9,197
-1,825
-22
-25,501
-4
-3,904
-862
-22
-21,333
-4
2,514
87
5
2,308
0
5,294
472
0
6,476
0
11,714
                                                6-58

-------
 1
 2           Key uncertainties in this approach include:
 3           •   C-R functions are available for only 11 species.  The urban areas chosen had a
 4              maximum of three of the 11 species present.8 This limitation neglects the effects of
 5              O3 on species where no C-R function is available. In the areas modeled that means
 6              that the majority of trees in the cities were not accounted for in the O3 damages.
 7              There are 2 species present that we know are sensitive but for which no C-R function
 8              is available.  This excludes 80 - 90 percent of the total trees present in the study areas.

 9           •   Uncertainties inherent within the models, both i-Tree and the CMAQ-generated air
10              quality surfaces.

11           -In addition this analysis does not account for the fact that many tree species are
12              biogenic sources of volatile organic compounds (VOC) that contribute to formation
13              of air pollution. Vegetation may account for as much as two-thirds of the VOC
14              production (Guenther et al., 2006). Carlton et al. (2010) found, however, that were
15              man-made pollutants not present biogenic pollution would drop by as a much as 50
16              percent.

17           If we were able to account for 63 damages to the species without a C-R function the
18    estimates would likely be higher.

19    6.8     ECOSYSTEM-LEVEL EFFECTS
20           To assess the risk to ecosystems from biomass loss, as opposed to the potential risk to
21    individual  tree species, we attempted to combine the RBL values into one metric. One factor in
22    assessing the risk to ecosystems is a measure of the overall abundance of each species.  As  a
23    measure of overall abundance, we used the basal area estimates described in Section 6.2.1 to
24    calculate the proportion of basal area for each of the 12 species assessed. Table 6-22 below
      8 Because of the timing of this analysis, we did not include Loblolly Pine (it would have been included in Atlanta).
      We did include Loblolly Pine for the other analyses in this draft assessment.
                                                      6-59

-------
 1    reflects, by region, the total basal area covered by the 12 tree species assessed. We separated the
 2    total basal area covered into different categories of percent cover of the species assessed. For
 3    example,  in the Southwest region, 13 percent of the total basal area assessed had less than 10
 4    percent cover of the 12 tree species; 7.1 percent of the total basal area assessed had between 10
 5    and 25 percent cover of the 12 tree species; 8.8 percent of the total basal area assessed had
 6    between 25 and 50 percent cover of the 12 tree species;  and 64.9 percent of total basal area
 7    assessed had no data on percent cover of the 12 tree species.  The Southwest and West regions
 8    had the largest percentages of total basal area assessed with no data on percent cover of tree
 9    species, and the Central and Northeast regions had the smallest percent of total basal area
10    assessed with no data on percent cover of tree species.
11
12    Table 6-22    Percent of Total Basal Area Covered by 12 Assessed Tree Species

Region
Central
East North Central
Northeast
Northwest
South
Southeast
Southwest
West
West North Central
All Regions
Percent of Total Basal Area Covered by 12 Assessed Tree Species
^10%
38.4%
33.4%
7.0%
4.5%
28.6%
16.0%
13.0%
10.0%
20.2%
20.3%
10% to 25%
32.0%
25.7%
22.1%
7.7%
4.0%
14.2%
7.1%
3.7%
8.0%
12.0%
25% to 50%
26.6%
27.5%
47.9%
20.0%
7.7%
48.1%
8.8%
7.0%
9.7%
19.1%
50% to 75%
2.2%
8.9%
22.2%
24.5%
7.7%
17.7%
5.1%
5.5%
8.2%
10.0%
> 75%
<0.1%
0.1%
0.5%
15.0%
0.9%
0.3%
1.2%
0.2%
6.5%
2.7%
No Data
0.7%
4.3%
0.3%
28.3%
51.2%
3.8%
64.9%
73.5%
47.4%
35.9%
13
14
15
16
17
18
19
20
21
       Data on basal area were available in over 64 percent of the cover area assessed, as
measured by the number of grid cells. To understand the potential W126 index values in the
percent of cover area not assessed, Table 6-23 includes information on the (i) number of grid
cells with no data on basal area above a certain amount and (ii) total number of grid cells with no
data on basal area. For those grid cells with no data on basal area, the table also shows, under
recent conditions, the number of grid cells with W126 index values that would exceed potential
alternative standards of 15, 11, and 7 ppm-hrs. In the Southwest, under recent conditions, 52
percent of the grid cells with no  data have W126 index values above 15 ppm-hrs, 95 percent
                                                     6-60

-------
 1
 2
 3
 4
 5
 6
 7
have W126 index values above 11 ppm-hrs, and 100 percent have W126 index values above 7
ppm-hrs. In contrast, in the East North Central, under recent conditions, no grid cells with no
data have W126 index values above 15 ppm-hrs, 1 percent have W126 index values above 11
ppm-hrs, and 3.5 percent have W126 index values above 7 ppm-hrs.
Table 6-23    Grid Cells With No Data That Exceed W126 Index Values under Recent
              Conditions
Region
Central
East North Central
Northeast
Northwest
South
Southeast
Southwest
West
West North Central
All Regions
Number of Grid
Cells w/No Data
Exceeding 10
sqft/acre Basal Area
(Total with No Data)
35 (35)
193 (198)
11(11)
709 (1,256)
4,329 (5,239)
198 (200)
2,854 (4,904)
2,315(3,550)
3,307(4,013)
13,951 (19,406)
Number of Grid Cells w/No Data that Exceed W126
Index Values Under Recent Conditions
> 7 ppm-hrs
34
7
11
779
4,638
59
4,904
3,452
1,870
15,754
> 11 ppm-hrs
11
2
11
451
1,945
15
4,662
3,274
1,158
11,529
> 15 ppm-hrs
3
0
6
189
27
3
2,572
2,680
283
5,763
 9
10
11
12
13
14
15
16
17
18
19
       We used the proportion of total basal area for each species to weight the RBL value for
that species in each grid cell. The weighted values for all species present in each grid cell were
added to generate a weighted RBL value for each grid cell. Table 6-24 provides a summary of
the percent of total basal area that exceeds a 2 percent weighted biomass loss under recent
conditions and when adjusted to just meet the current standard.  Table 6-25 provides a summary
of the percent of total basal area that exceeds a 2 percent biomass loss at potential alternative
standard levels of 15, 11, and 7 ppm-hrs. Note that for biomass loss,  CAS AC recommended that
EPA should consider options for W126 standard levels based on factors including a predicted
one to two percent biomass loss for trees and a predicted five percent loss of crop yield.  Small
losses for trees on a yearly basis compound over time and can result in substantial biomass losses
over the decades-long lifespan of a tree (Frey and Samet, 2012b). We chose to use the 2 percent
                                                    6-61

-------
 1   biomass loss recommendation in this analysis; however, the weighted RBL value is not the same
 2   as the individual species analysis (Section 6.2.1.3). These data are interpreted in a more relative
 3   manner where higher values represent a larger potential impact on the overall ecosystem.
 4          The data in Table 6-24 and Table 6-25 shows that the total area exceeding two percent
 5   biomass loss decreases, as expected, across air quality scenarios.  For example, for the Central
 6   region under recent conditions, a total of 23.7 percent of total basal area assessed would exceed a
 7   2 percent biomass loss and when adjusted to just meet the current standard, a total of 2.7 percent
 8   of total basal area assessed would exceed a 2 percent biomass loss. When adjusted to meet
 9   potential alternative standard levels of 15, 11, and 7 ppm-hrs, 2.7  percent, 1 percent and 0.1
10   percent, respectively, of total basal area assessed would exceed a 2 percent biomass loss.
11          While it is not possible to predict overall effects, the results from these analyses show the
12   weighted average RBL to be a potential predictor of risk in areas with a high proportion of
13   species included. As such, the percent of area exceeding one and two percent weighted RBL is
14   most at risk where the  species included account for more than 75 percent of the total basal area.
15
                                                     6-62

-------
1   Table 6-24    Percent of Area Exceeding 2% Weighted Biomass Loss - Recent Conditions
2                and Existing Standard


Region
Central
East North Central
Northeast
Northwest
South
Southeast
Southwest
West
West North Central
All Regions


Region
Central
East North Central
Northeast
Northwest
South
Southeast
Southwest
West
West North Central
All Regions
Percent of Area Exceeding 2% Weighted Biomass Loss
(Recent Conditions)
Cover Categories of 12 Assessed Tree Species
^10%
2.4%
1.1%
0.1%
0.0%
1.2%
<0.1%
0.1%
<0.1%
3.0%
1.1%
10% to 25%
11.0%
8.0%
0.6%
0.0%
0.5%
0.9%
0.3%
0.6%
3.1%
2.6%
25% to 50%
9.1%
3.5%
6.1%
0.3%
0.2%
6.2%
4.4%
2.0%
2.2%
3.4%
50 to 75%
1.2%
<0.1%
10.3%
1.0%
0.1%
1.4%
4.3%
1.8%
3.0%
2.2%
> 75%
<0.1%
0.0%
0.2%
1.5%
0.0%
0.0%
1.2%
0.2%
3.7%
0.9%
Total
23.7%
12.6%
17.3%
2.9%
2.1%
8.6%
10.3%
4.7%
15.0%
10.1%
Percent of Area Exceeding 2% Weighted Biomass Loss
(75 ppb Scenario)
Cover Categories of 12 Assessed Tree Species
<_10%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.1%
<0.1%
10% to 25%
1.3%
0.1%
0.0%
0.0%
<0.1%
0.0%
0.1%
0.0%
0.5%
0.2%
25% to 50%
1.3%
0.6%
0.1%
0.0%
0.1%
0.0%
0.1%
0.0%
1.0%
0.4%
50 to 75%
<0.1%
0.0%
0.1%
0.0%
0.0%
0.0%
0.1%
0.0%
0.3%
0.1%
> 75%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.3%
0.0%
0.3%
0.1%
Total
2.7%
0.6%
0.2%
0.0%
0.2%
0.0%
0.5%
0.0%
2.2%
0.8%
                                                6-63

-------
1   Table 6-25    Percent of Area Exceeding 2% Weighted Biomass Loss - Alternative W126
2                 Standard Levels


Region
Central
East North Central
Northeast
Northwest
South
Southeast
Southwest
West
West North Central
All Regions


Region
Central
East North Central
Northeast
Northwest
South
Southeast
Southwest
West
West North Central
All Regions


Region
Central
East North Central
Northeast
Percent of Area Exceeding 2% Weighted Biomass Loss
(15 ppm-hrs Scenario)
Cover Categories of 12 Assessed Tree Species
^10%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
<0.1%
<0.1%
10% to 25%
1.3%
0.1%
0.0%
0.0%
<0.1%
0.0%
0.1%
0.0%
0.4%
<0.1%
25% to 50%
1.3%
<0.1%
0.1%
0.0%
0.1%
0.0%
0.1%
0.0%
1.0%
0.2%
50% to 75%
<0.1%
0.0%
0.1%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.3%
0.1%
> 75%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.2%
<0.1%
Total
2.7%
0.6%
0.2%
0.0%
0.2%
0.0%
0.2%
0.0%
2.0%
0.7%
Percent of Area Exceeding 2% Weighted Biomass Loss
(11 ppm-hrs Scenario)
Cover Categories of 12 Assessed Tree Species
^10%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
<0.1%
<0.1%
10% to 25%
0.4%
<0.1%
0.0%
0.0%
<0.1%
0.0%
0.0%
0.0%
0.4%
0.1%
25% to 50%
0.4%
0.4%
0.0%
0.0%
0.1%
0.0%
<0.1%
0.0%
0.9%
0.3%
50% to 75%
<0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.3%
0.1%
> 75%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.2%
<0.1%
Total
1.0%
0.4%
0.0%
0.0%
0.1%
0.0%
0.1%
0.0%
1.8%
0.5%
Percent of Area Exceeding 2% Weighted Biomass Loss
(7 ppm-hrs Scenario)
Cover Categories of 12 Assessed Tree Species
<_10%
<0.1%
0.0%
0.0%
10% to
25% to 50%
25%
0.1% 0.1%
<0.1% 0.3%
0.0% 0.0%
50% to 75%
0.0%
0.0%
0.0%
> 75%
0.0%
0.0%
0.0%
Total
0.1%
0.3%
0.0%
                                                6-64

-------
Northwest
South
Southeast
Southwest
West
West North Central
All Regions
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
<0.1%
0.0%
0.1%
0.0%
0.0%
0.0%
0.5%
0.1%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.3%
0.1%
0.0%
0.0%
0.0%
<0.1%
0.0%
0.1%
<0.1%
0.0%
0.1%
0.0%
<0.1%
0.0%
1.0%
0.2%
 2          Two important things to note with respect to the weighted RBL analysis. First, the
 3    proportional basal area values do not account for total cover, only for the relative cover of the
 4    tree species present. This is most noticeable with Cottonwood and Ponderosa pine, which are
 5    near 100 percent cover in some areas; however, the absolute cover is very different. Ponderosa
 6    pine occurs in relatively high density in some grids, exceeding 100 square feet per acre, while
 7    Cottonwood is often less than 10 square feet per acre. This affects the direct interpretation of the
 8    values presented because the overall ecosystem effect may be very different, although equally
 9    important. It is important to remember with this data set that these numbers are only useful as a
10    very general estimate of potential effects. Second, this analysis only accounts for the 12 tree
11    species with C-R functions; other species may also be sensitive to Os exposure. It is also
12    possible other species that are not sensitive may be indirectly affected through changes in
13    community composition and competitive interactions.
14               6.8.1    Potential Biomass Loss in Federally Designated Areas
15                   6.8.1.1   Class I Areas
16          We analyzed federally designated Class I areas in relation to the W126 air quality surface
17    and the weighted RBL values. We completed the analyses of Class I areas in the same manner as
18    the analyses across the entire range of data;  however, we present the results as a count of the
19    Class I areas and not as a percentage of area. We treated each Class I area as an individual
20    geographic endpoint and calculated  an average weighted RBL for all Class I areas  with at least
21    one grid cell that had a non-zero weighted RBL. Data were available in 145 of the 156 Class I
22    areas.  A complete list of Class I areas and the weighted RBL values at the current standard and
23    alternative W126 standard levels is included in Appendix 6E.
                                                     6-65

-------
 1
 2
 3
 4
 5
 6
       Table 6-26 summarizes the number of Class I areas exceeding 1 percent and 2 percent
weighted RBL across varying percent cover of species and under recent conditions and when
adjusted to just meet the existing standard and potential alternative standard levels of 15, 11, an
7 ppm-hrs. The number of areas exceeding 1 percent and 2 percent decreases across air quality
and
scenarios.
Table 6-26    Weighted RBL and Percent Cover in Class I Areas

Percent
of Total
Basal
Area
No Data
<10
10 to 25
25 to 50
50 to 75
>75
Total

Areas

Class I
Areas
Covered

11
54
35
48
6
2

156

Number of Class I Areas Exceeding 1%
Weighted RBL


Recent
Conditions
-
6
8
20
1
1

36


75
ppb
-
2
0
1
0
1

4


15
ppm-
hrs
-
2
0
1
0
1

4


11
ppm-
hrs
-
2
0
0
0
1

3


7
ppm-
hrs
-
1
0
0
0
1

2

Number of Class I Areas Exceeding 2%
Weighted RBL


Recent
Conditions
-
2
2
7
1
1

13


75
ppb
-
1
0
0
0
1

2


15
ppm-
hrs
-
1
0
0
0
1

2


11
ppm-
hrs
-
1
0
0
0
1

2


7
ppm-
hrs
-
0
0
0
0
1

1

 7
 8   6.9    QUALITATIVE ASSESSMENT OF UNCERTAINTY
 9          As noted in Chapter 3, we have based the design of the uncertainty analysis for this
10   assessment on the framework outlined in the WHO guidance (WHO, 2008).  For this qualitative
11   uncertainty analysis, we have described each key source of uncertainty and qualitatively assessed
12   its potential impact (including both the magnitude and direction of the impact) on risk results, as
13   specified in the WHO guidance. In general, this assessment includes qualitative discussions of
14   the potential impact of uncertainty on the results (WHO Tierl) and quantitative  sensitivity
15   analyses where we have sufficient data (WHO Tier 2).
16           Table 6-27 includes the key sources of uncertainty identified for the Os REA. For each
17   source of uncertainty, we have (a) provided a description, (b) estimated the direction of influence
18   (over, under,  both, or unknown) and magnitude (low, medium, high) of the potential impact of
19   each source of uncertainty on the risk estimates, (c) assessed the degree of uncertainty (low,
                                                     6-66

-------
 1   medium, or high) associated with the knowledge-base (i.e., assessed how well we understand
 2   each source of uncertainty), and (d) provided comments further clarifying the qualitative
 3   assessment presented. The categories used in describing the potential magnitude of impact for
 4   specific sources of uncertainty on risk estimates (i.e., low, medium, or high) reflect our
 5   consensus on the degree to which a particular source could produce a sufficient impact on risk
 6   estimates to influence the interpretation of those estimates in the context of the secondary Os
 7   NAAQS review. Where appropriate, we have included references to specific sources of
 8   information considered in arriving at a ranking and classification for a particular source of
 9   uncertainty.
10
                                                      6-67

-------
1    Table 6-27   Summary of Qualitative Uncertainty Analysis in Relative Biomass Loss Assessments
       Source
         Description
   Potential influence of
    uncertainty on risk
         estimates
                                                      Direction
                                               Magnitude
               Knowledge-
                  Base
                   Comments (KB: knowledge base, INF: influence of
                             uncertainty on risk estimates)
A.  National W126
surfaces
The biomass loss analyses in
this chapter use the national
W126 surfaces for recent
conditions and adjusted to just
meet the existing standard and
alternative W126 standards.
Both
Low-
Medium
Low-medium
KB and INF: See Chapter 4 for more details.
B. Shape of the C-R
function for biomass
loss for different
species
Biomass loss and yield loss
estimates are highly sensitive to
the parameters in the C-R
function.
Unknown
High
Medium
KB: We conducted sensitivity analyses for 10 crops (in 54
studies) and 12 tree species (in 52 studies), which showed
high intraspecific and interspecific variability. Some species
only had one study, while other species had many studies.
INF: The resulting C-R functions for the included species
were mostly of intermediate sensitivity, with only a few
species considered very sensitive and several that showed
little or no sensitivity to O3. This range of sensitivities was
consistent with the additional studies included in the ISA, but
further studies are needed to determine how accurately this
reflects the larger suite of tree species in the U.S.
C. Absence of C-R
functions for many
O3-sensitive species
C-R functions are available for
only 12 tree species, thus the
majority of trees in the modeled
urban areas and Class I areas
were not incorporated.
Under
Medium-
High
Medium-Low
KB: We are certain that there are additional sensitive species
based on studies cited in the ISA that reported effects.
However, the studies of additional sensitive species did not
provide sufficient information to generate C-R functions.
Therefore, we are certain that we are underestimating tree
biomass loss in urban areas and Class I areas.
INF: Eighty to 90 percent of the total trees in the urban case
study areas are excluded from the analysis. There are 2 tree
species in the case study areas that we know are sensitive but
for which no C-R function is available. The magnitude of the
influence is dependent on the community composition in each
area.
                                                             6-68

-------
       Source
         Description
   Potential influence of
    uncertainty on risk
         estimates
                                                      Direction
                                               Magnitude
               Knowledge-
                  Base
                   Comments (KB: knowledge base, INF: influence of
                             uncertainty on risk estimates)
D. Using C-R
functions for tree
seedlings rather than
adult trees
C-R functions for trees are
based on analyses of tree
seedlings, but most biomass
impacts are from effects on
adult trees.
Both
Low-
Medium
Medium
KB and INF: In general, estimates of relative biomass loss
(RBL) in tree seedlings are comparable to the estimates for
adult trees, with a few exceptions such as black cherry.  Some
species overestimate RBL in adult trees and some species
underestimate RBL.
E. Urban tree
inventory in iTree
The base inventory of urban
trees, including species and
distribution, in iTree has
uncertainty.
Unknown
Low
High
KB: The urban tree inventories included in the iTree analyses
are based on field counts and measurements of trees in the
specific urban areas analyzed (personal communication,
Nowak, 6/2011). Tree census data (e.g., Baltimore, Syracuse,
Chicago, and Atlanta) are generally considered less uncertain
than modeled tree inventories (e.g., urban areas of
Tennessee).
INF: The iTree model estimates carbon sequestration and
pollution removal services provided by urban forests. These
services are based on tree growth and pollution removal
functions that are specific to the forest structure in each urban
area, including the species composition, number of trees, and
diameter distribution of trees. Uncertainties in the tree
inventory are propagated into the estimates of carbon
sequestration and pollution removal based on those
inventories.
F. Pollution removal
functions in iTree
The functions applied in iTree
to estimate pollution removal
are uncertain and vary by
species.
Unknown
Medium
Medium
KB: Pollution removal is calculated based on field, pollution
concentration, and meteorological data. The pollution
removal functions in iTree are from Nowak et al. (2006).
INF: iTree estimates that 1,535 tons/year of pollution are
removed from the urban case study areas at the existing
standard. Nowak et al. (2006) provides an indication of the
ranges of pollution removal in the literature.
                                                             6-69

-------
       Source
         Description
   Potential influence of
    uncertainty on risk
         estimates
                                                     Direction
                                              Magnitude
              Knowledge-
                  Base
                  Comments (KB: knowledge base, INF: influence of
                            uncertainty on risk estimates)
G. VOC emissions
from trees
Many tree species are biogenic
sources of volatile organic
compounds (VOC) that
contribute to formation of
ozone. Additional VOC
emissions associated with
biomass gains are not
addressed.
Over
(generally)
Medium
High
KB:  According to the O3 ISA (U.S. EPA, 2013, section
3.2.1), vegetation emits substantial quantities of VOCs, and
the 2005 NEI approximately 29 MT/year of VOC emissions
were from biogenic sources.
INF: Vegetation may account for as much as two-thirds of the
VOC production (Guenther et al., 2006).  Carlton et al. (2010)
found, however, that if man-made pollutants were not present,
O3 attributable to biogenic emissions would drop by as a
much as 50 percent.
H. Carbon
sequestration
functions in iTree and
FASOM
The functions applied in the
models to estimate carbon
sequestration are uncertain and
vary by species.
Unknown
Medium
Medium
KB: The studies in the ISA show a consistent pattern of
reduced carbon uptake due to O3 damage, with large
reductions projected over North America. The forest carbon
accounting component of FASOMGHG is largely derived
from the U.S. Forest Service's Forestry Carbon (FORCARB)
modeling system, which is an empirical model of forest
carbon budgets simulated across regions, forest types, land
classes, forest age classes, ownership groups, and carbon
pools. Multiple equations for individual species were
combined to produce one predictive equation for a wide range
of diameters for individual species. Formulas were combined
to prevent disjointed sequestration estimates that can occur
when calculations switch between individual biomass
equations. If no allometric equation could be found for an
individual species, the average of results from equations of
the same genus is used. If no genus equations are found, the
average  of results from all broadleaf or conifer equations is
used.
INF: We estimate  that carbon storage would increase by 13
million metric tons and 1.6 billion metric tons over 40 years
after just meeting the existing and the alternative standard
level of 7 ppm-hrs, respectively. The process of combining
the individual formulas produced results that were typically
within 2% of the original estimates.
                                                            6-70

-------
       Source
         Description
   Potential influence of
    uncertainty on risk
         estimates
                                                     Direction
                                              Magnitude
              Knowledge-
                  Base
                  Comments (KB: knowledge base, INF: influence of
                            uncertainty on risk estimates)
I. Use of median C-R
functions for crops in
FASOM
FASOMGHG incorporates
median parameters from Lehrer
etal. (2007) in the C-R
functions for oranges, rice, and
tomatoes. Using alternative C-R
functions would result in lower
or higher O3 impacts on crop
and tree species biomass
productivity, which would
potentially lead to different
economic equilibrium
outcomes.
Both
Low
Low
KB: These 3 crops have C-R functions based on O3 metrics
other than W126, as reported in Lehrer (2007).
INF: Use of the median function could affect the estimates for
those crops specifically. No other crop estimates are based on
these functions.
J. Crop proxy and
forest type
assumptions
The crops/tree species modeled
are only a subset of species
present in U.S. agriculture and
forestry systems. Actual
impacts may differ from those
of the crop proxy or the forest
type. Further, FASOMGHG
modeling used a simple average
of tree RYLs for all forest types
within a region.
Both
Medium-
High
Low
KB: Aggregation of crop and tree species was conducted
based on recommendations from CASAC (Frey and Samet,
2012a). As stated by CASAC, it is not feasible to obtain C-R
functions for all species, and there is no reliable mechanism
to infer C-R relationships in a novel species even from
knowledge of a closely-related species.
INF: Total economic surplus is estimated to decrease by $24
million or increase by as much as $257 million between 2010
and 2040. It is unclear how using actual species information
rather than proxy species would affect these estimates.
However, consistent with CASAC recommendation, we did
not assign the most sensitive C-R relationships to the proxy
species.
                                                            6-71

-------
       Source
         Description
   Potential influence of
    uncertainty on risk
         estimates
                                                     Direction
                                              Magnitude
              Knowledge-
                  Base
                  Comments (KB: knowledge base, INF: influence of
                            uncertainty on risk estimates)
K. FASOMGHG
does not model
agriculture/ forestry
on public lands
Because public lands are not
affected within the model, the
estimates of changes in
consumer and producer surplus
would likely be higher if public
lands were included.
Under
Medium
Medium-Low
KB: The model assumes that O3 biomass effects would have
little influence on harvest decisions because timber harvests
on public lands are set by the relevant government regulating
body (Forest Service, Bureau of Land Management, etc).
INF: The FASOMGHG model includes 349 million acres of
private, managed forests. The USFS estimates that there are
approximately 751 million forest acres in the U.S., but only a
small portion of this public land is logged for timber.
L. Forest adaptation
toO3
FASOMGHG modeling does
not reflect changes in tree
species mixes within a forest
type made by natural adaptation
and adaptive management by
landowners due to O3. Less
sensitive tree species may gain
relative advantage over more
sensitive species.
Unknown
Low
Low
KB: The ISA finds that the evidence is sufficient to conclude
that O3 causes changes in community composition favoring
O3 tolerant species over sensitive species. The KBs for
natural adaptation and adaptive management are different,
and the relative dominance of one over the other would differ
depending on the degree of active management.
INF: Over time, the O3 impacts on forests may be reduced as
forests adapt to O3 environments through forest management
or natural processes.
M. International trade
projections in
FASOMGHG
FASOMGHG reflects future
international trade projections
by USDA based on recent O3
conditions. Soybeans and wheat
are major crop exports and have
relatively large responses to O3,
which are not reflected in the
trade projections.
Both
Medium
Medium
KB: Although FASOMGHG includes international trade for
major commodities, the international trade projections do not
reflect the potential for increased exports associated with
increased yield from reduced O3 exposure. The world trade
quantities  data in the model have been updated to reflect more
recent trade data for specific commodities in the literature
since the original data from the USDA SWOPSDVI model
(Roningen, 1986).
INF: Increased exports could increase producer surplus but
the impacts on consumer surplus are unclear.
                                                            6-72

-------
       Source
         Description
   Potential influence of
    uncertainty on risk
         estimates
                                                    Direction
                                             Magnitude
              Knowledge-
                  Base
                  Comments (KB: knowledge base, INF: influence of
                            uncertainty on risk estimates)
N. Estimates of tree
basal area used to
assess larger scale
ecosystem effects
Estimates of basal area were
modeled by the FHTET at a
scale of 240 m2. These values
were aggregated to the 144
(12x12) km2 CMAQ grid.
Unknown
Low-
Medium
Low
KB: USDA's FHTET has been actively working to refine
their models to estimate basal area for individual tree species
and total basal area nationwide.
INF: The effect on risk estimates would vary between
ecosystems, depending on community composition, total
basal area and the ecosystem services being affected. Due to
the overall large number of CMAQ cells included for each
species, the overall estimates presented here would likely be
small.
                                                           6-73

-------
 1    6.10   DISCUSSION
 2          Relative Biomass Loss:
 3          •   We compared seedling RBL to results from several studies with mixed results. The
 4              studies indicate that overall the seedling biomass loss values are much more
 5              consistent with the adult loss at lower W126 index values.

 6          •   The Constable and Taylor (1997) study implies that for the eastern subspecies of
 7              Ponderosa Pine, the seedling RBL rate could possibly overestimate the adult RBL
 8              rate.

 9          •   The Weinstein et al. (2001) study indicates that the seedling RBL estimates are
10              comparable to the adult estimates, except at higher W126 index values for Tulip
11              Poplar. The Black Cherry results are an exception, which tells us that this species is
12              much less sensitive as an adult than as a seedling.

13          •   Another study (Samuelson and Edwards, 1993) on Red Oak found the exact opposite
14              pattern — adult trees are much more sensitive to Os-related biomass loss than
15              seedlings.

16          •   Overall, the western tree species have more fragmented habitats than the eastern
17              species. The areas in southern California have the highest W126 index values. The
18              eastern tree species had less fragmented ranges and areas of elevated RBL that were
19              more easily attributed to urban areas (e.g. Atlanta, GA and Charlotte,  NC) or to the
20              Tennessee Valley Authority Region.

21          Commercial Timber Effects:
22          •   At the existing standard of 75 ppb the highest yield loss occurs in upland hardwood
23              forests in the South Central and Southeast regions at over 3 percent per year. The next
24              highest yield losses occur in Corn Belt hardwoods with just over 2 percent loss per
25              year and in hard- and softwoods of the Rocky Mountain region at an average loss
26              across all sensitive forests of slightly  over 1 percent loss per year.  With the exception
27              of the Rocky Mountain region, yield losses do not appreciably change when meeting
                                                     6-74

-------
 1              the 15 ppm-hrs alternative incremental to meeting the existing standard. Yield gains
 2              associated with meeting alternative W126 standards are relatively small on a
 3              percentage change basis, especially in the 15 ppm-hrs scenario where the highest
 4              change is 0.35 percent per year.

 5           •   Consumer and producer welfare in the forest sector are more affected by meeting
 6              alternative W126 standards incremental to meeting the existing standard than the
 7              agricultural sector. In general, consumer welfare increases in both the forest and
 8              agricultural sectors as higher productivity tends to increase total production and
 9              reduce market prices. Because demand for most forestry and agricultural
10              commodities is inelastic, producer welfare tends to decline with higher productivity
11              as the effect of falling prices on profits more than outweighs the effects of higher
12              product! on 1 evel s.

13           Climate Regulation:
14           •   For national-scale carbon sequestration, much greater changes in carbon sequestration
15              are projected in the forest sector than in the agricultural sector. The 15 ppm-hrs
16              scenario  does not appreciably increase carbon storage relative  to just meeting the
17              existing standard.  The vast majority of the enhanced carbon sequestration potential
18              under the scenarios is from increased forest biomass due to the yield increases
19              accruing  to forests over time at the 11 and 7 ppm-hrs alternative W126 standards.
20              The forest carbon sequestration potential would increase between 593 and 1,602
21              million tons of CC>2 equivalents over 30 years after meeting the 11  or 7 ppm-hrs
22              W126 standard level, respectively.

23           •   For the urban case study areas, estimates suggest that in the five modeled areas
24              relative to recent conditions, at the existing standard or at an alternative W126
25              standard  level  of 15 ppm-hrs about 3.5 million tons of carbon storage will be lost over
26              25 years  (about 140,000 tons/year). At an alternative W126 standard level of 11
27              ppm-hrs, loss of carbon sequestration is approximately 128,000 metric tons per year,
28              and meeting an alternative W126 standard of 7 ppm-hrs results in the loss of 112,000
29              metric tons per year of carbon storage services.
                                                      6-75

-------
 1          •   Of the five areas modeled, the combined urban areas of Tennessee have the largest
 2              estimated gains in carbon storage at almost 20,000 tons per year when meeting an
 3              alternative W126 standard of 7 ppm-hrs relative to the existing standard.

 4          Urban Case Study Air Pollution Removal:
 5          •   Estimates from i-Tree indicate that at the existing standard about 1,535 tons of air
 6              pollution removal capacity is lost annually in the five areas modeled.  Syracuse and
 7              Baltimore have no change in pollution removal when meeting the existing standard
 8              and the modeled alternatives. Atlanta and Chicago gain about 470 and 6,500 metric
 9              tons of additional pollution removal when meeting the 7 ppm-hrs W126 alternative
10              standard compared to the existing standard, while Tennessee gains almost 12,000 tons
11              of potential pollution removal annually for this scenario.  Under the 7 ppm-hrs
12              scenario, about 51 percent of the pollution removal capacity lost under the existing
13              standard is regained.

14          Agriculture:
15          •   Among the major crops, winter wheat and soybeans are more sensitive to ambient 63
16              levels than corn and sorghum. California, the Northeast, and the Rocky Mountain
17              regions generally have the highest yield losses.

18          •   For winter wheat, the highest loss occurs in California at 15 percent. In the
19              Northeast, the losses range from  7.65 percent in Maryland to 3.69 percent in
20              Pennsylvania, with 6.43 percent in Delaware and 6.55 percent in New Jersey. In the
21              Rocky Mountain region, the losses in Utah are 7.26 percent.  When the W126
22              scenarios are modeled, the yield losses are almost eliminated at all values of W126.

23          •   For soybeans, the highest loss occurs in Maryland at 8.3 percent. In the Northeast,
24              the losses range from 8.3 percent in Maryland to 5.38 percent in Pennsylvania, with
25              7.65 percent in Delaware and 7.76 percent in New Jersey.  In the Corn Belt the
26              highest loss occurs in southern Indiana at 5.1 percent. In the Rocky Mountain region,
27              the losses in Colorado are 6.73 percent.  Yield losses remain under all scenarios for
28              W126, although for the 7 ppm-hrs scenario all losses are less than 0.6 percent.
                                                     6-76

-------
 1           •   For corn, the highest loss occurs in California at 0.88 percent.  In the Northeast, the
 2              losses range from 0.68 percent in Maryland to 0.26 percent in Pennsylvania, with
 3              0.56 percent in Delaware and 0.48 percent in New Jersey. In the Corn Belt, Lake
 4              States, and Great Plains the highest loss occurs in southern Ohio at 0.34 percent. And
 5              in the Rocky Mountain region, the losses range from 0.67 percent in Utah to 0.42
 6              percent in Nevada, with 0.45 percent in Colorado.  When the W126 scenarios are
 7              modeled, the yield losses are virtually eliminated at all values of W126 and
 8              subsequent yield gains are almost nonexistent.

 9           •   In general, increased yield leads to increased supply and lower prices.  Because corn
10              does not lose or gain very much under any scenario prices are likely to remain
11              relatively stable. Soybeans, however, would experience yield gains in any scenario
12              and prices would likely fall. In response to falling soybean prices, the model predicts
13              that producers would switch to less Os-sensitive crops with stable prices, such as
14              corn, thereby increasing corn production.

15           •   For producers, the W126  alternatives results in welfare gains in the middle years,
16              2020-2030, and welfare losses  in all other years. For consumers, however, the
17              changes in production and prices results in welfare gains in all scenarios in all years.
                                                      6-77

-------
 1                              7   VISIBLE FOLIAR INJURY

 2     7.1   INTRODUCTION
 3          Visible foliar injury resulting from exposure to ozone (63) has been well characterized
 4    and documented over several decades on many tree, shrub, herbaceous, and crop species (U.S.
 5    EPA, 2013, 2006,  1996, 1984, 1978). Visible foliar injury symptoms are considered diagnostic
 6    as they have been verified experimentally in exposure-response studies using exposure
 7    methodologies such as continuous stirred-tank reactors (CSTRs), open-top chambers (OTCs),
 8    and free-air fumigation (see Section 9.2 of the ISA for more detail on exposure methodologies).
 9    Although the majority of (Vinduced visible foliar injury occurrence has been observed on
10    seedlings and small plants, many studies have reported visible injury of mature coniferous trees,
11    primarily in the western U.S. (Arbaugh et al., 1998), and of mature deciduous trees in eastern
12    North America (Schaub et al., 2005; Vollenweider et al., 2003; Chappelka et al., 1999a;
13    Chappelka et al., 1999b; Somers et al., 1998; Hildebrand et al., 1996).
14          The ecosystem services most likely to be affected by Os-induced foliar injury are
15    aesthetic value and outdoor recreation. Aesthetic value and recreation services depend on the
16    perceived scenic beauty of the environment. Studies of Americans' perception of scenic beauty
17    are quite consistent (Ribe, 1994) in their findings - people tend to have a reliable set of
18    preferences for forest and vegetation with fewer damaged or dead trees and plants. Aesthetic
19    value not related to recreation includes the scenic value of vistas observed as people go about
20    their daily lives and the scenic value of the views of open space near and around homes. Many
21    outdoor recreation activities directly depend on the scenic value of the area, in particular scenic
22    viewing, wildlife watching, hiking, and camping. These activities are enjoyed by millions of
23    Americans every year and generate millions of dollars in economic value (OIF, 2012; NFS,
24    2002a, 2002b, 2002c). Figure 7-1 illustrates the relationship between foliar injury and ecosystem
25    services as discussed in this chapter.
                                                     7-1

-------
              Ambient
               Ozone
              Exposure
         Ecological Effect
         Visible Foliar Injury
      Ecosystem Level Effects
 • National-scale Analysis of Foliar Injury
         Cultural Services
         * Recreational Use
         * National Values of Trip and
         Equipment-Related Expenditures
         for Wildlife-Watching:, Trail, and
         Camping-Activities
 Screening-Level Assessment
of Foliar Injury in National
Parks
Cultural Services
* Recreational Use
* For 3 National Parks Case Studies,
Data on Activities, Travel and Local
Expenditures, and Local Economic
Impact
 1                                                                     	
 2    Figure 7-1     Relationship between Visible Foliar Injury and Ecosystem Services
 3
 4           The significance of 63 injury at the leaf and whole-plant levels depends on how much of
 5    the total leaf area of the plant has been affected, as well as the plant's age, size, developmental
 6    stage, and degree of functional redundancy among the existing leaf area. Previous O3 Air
 7    Quality Criteria Documents (AQCDs) and the Os Integrated Science Assessment (ISA) for have
 8    noted the difficulty in relating visible foliar injury symptoms to other vegetation effects such as
 9    individual plant growth, stand growth, or ecosystem characteristics (U.S. EPA, 2013, 2006,
10    1996). As a result, it is not currently possible to determine, with consistency across species and
11    environments, what degree of injury at the leaf level has significance to the vigor of the whole
12    plant. However, in some cases, visible foliar symptoms have been correlated with decreased
13    vegetative growth (Somers et al., 1998; Karnosky et al.,  1996; Peterson et al., 1987; Benoit et al.,
14    1982) and with impaired reproductive function (Chappelka, 2002; Black et al., 2000).
15    Conversely, the lack of visible injury does not always indicate a lack of phytotoxic effects from
16    Os or a lack of non-visible Os  effects (Gregg et al., 2006).
17           The National Park Service (NFS) published a list of trees and plants considered sensitive
18    because they exhibit foliar injury at or near ambient concentrations in fumigation chambers or
19    they have been observed to exhibit symptoms in the field by more than one observer. This list
20    includes many species not included in Table 6-10, such as various milkweed species, asters,
                                                       7-2

-------
 1    coneflowers, huckleberry, evening primrose, Tree-of-heaven, redbud, blackberry, willow, and
 2    many others.  Many of these species are important for non-timber forest products, recreation, and
 3    aesthetic value among other ecosystem services.
 4          Based on the NFS sensitive species list (NFS, 2003), data from the Forest Health
 5    Technology Enterprise Team of the U.S. Forest Service (described in Chapter 6, Section 6.2.1.3)
 6    were available for 15 tree species. Figure 7-2 illustrates the percent of total basal area that is
 7    accounted for by these 15 species, which include Ponderosa Pine, Loblolly Pine, Virginia Pine,
 8    Red Alder, Tulip Poplar, Aspen, Black Cherry, Jack Pine, Table Mountain Pine, Pitch Pine,
 9    White Ash, Green Ash, Sweetgum, California Black Oak, and Sassafrass.
                                      15 Tree Species Sensitive to Foliar Injury
             Percent of Basal Area
                 D%
               | 0.1% - 10%
                 10.1%-25%
                 25.1%-50%
             ^B 50.1%-75%
             ^H 75.1%- 100%
10
11
12
13
Figure 7-2     Tree Species Sensitive to Foliar Injury

       Table 7-1 summarizes the overall cover of the 15 tree species and the percent of area in
each cover category that exceeds varying W126 index values. It is important to note that there
                                                      7-3

-------
 1
 2
 3
are additional tree species that are known to be sensitive for which cover data were not available,
and there are many non-tree species listed in the NFS report that are not addressed in this
analysis.
Table 7-1      Percent of Cover Category Exceeding W126 Index Values
Cover Category
(percent of total basal
area accounted for by
the 15 species
included)
None Present
Less than 10%
10% to 25%
25% to 50%
50% to 75%
Greater than 75%
National
Distribution
34.5%
26.0%
17.0%
12.8%
7.9%
1.8%
Percent of Cover Category Area Exceeding W126 Index
Values
> 7 ppm-hrs
85.0%
65.4%
73.4%
79.3%
83.6%
82.4%
>10 ppm-hrs
71.1%
39.7%
52.7%
60.9%
57.2%
41.6%
> 15 ppm-hrs
31.4%
9.4%
13.9%
20.7%
15.7%
9.5%
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
       In addition to direct impacts on foliar injury, 63 exposure contributes to trees'
susceptibility to insect infestation.  These infestations can affect scenic beauty and the services
associated with the perceived beauty of the environment. Foliar injury and insect attack can
occur separately or in conjunction with one another and are briefly discussed together in the next
section of this chapter, Section 7.1.1, on ecosystem services impacts. The remainder of this
chapter provides details on the analyses we conducted and includes Section 7.2 - National-Scale
Analysis of Foliar Injury; Section 7.3 -Screening-level Assessment of Visible Foliar Injury in
National Parks; and Section 7.4 -National Park Case Study Areas, including Great Smoky
Mountains National Park, Rocky Mountain National  Park, and Sequoia and Kings Canyon
National Parks. The national park case studies include discussions of the potential value of the
ecosystem services affected by foliar injury resulting from Os exposure.
           7.1.1     Ecosystem Services
                 7.1.1.1   Aesthetic Value
       Aesthetic value services not related to recreation include the view of the landscape from
houses, as individuals commute, and as individuals go about their daily routine in a nearby
community. Studies find that scenic landscapes are capitalized into the price of housing. Studies
also document the existence of housing price premiums associated with proximity to forest and
                                                      7-4

-------
 1    open space (Acharya and Bennett, 2001; Geoghegan, Wainger, and Bockstael, 1997; Irwin,
 2    2002; Mansfield et al., 2005; Smith et al., 2002; Tyrvainen and Miettinen, 2000). In addition,
 3    according to Butler (2008), approximately 65 percent of private forest owners rate providing
 4    scenic beauty as either a very important or important reason for their ownership of forest land.
 5           These aesthetic value services are at risk of impairment because of Os-induced damage:
 6    directly due to foliar injury, and indirectly due to increased susceptibility to insect attack.  Data
 7    are not available to explicitly quantify these negative effects; however, the damage is included in
 8    the price premium mentioned. In other words, without such damage, the associated price
 9    premium for scenic beauty that is incorporated into housing prices is likely higher.
10                    7.1.1.2  Recreation
11           With few exceptions, publicly owned forests are open for some form of recreation.
12    Based on the analysis done for the USDA National Report on Sustainable Forests (USDA, 2011),
13    almost all of the 751 million acres of forest lands in the U.S. are at least partially managed for
14    recreation. Of these 751 million acres, 44 percent are publicly owned (federal, state, or local).
15    Scenic quality has been found to be strongly correlated to recreation potential and the likelihood
16    of visiting recreation settings, and the correlations apply to both active and passive recreational
17    pursuits (Ribe, 1994). According to Ribe (1994), differences in scenic beauty account for 90
18    percent of the variation in participant satisfaction across all recreation types.
19           Americans enjoy a wide variety of outdoor pursuits many of which are subject to
20    negative impacts resulting from O?, exposure, especially the effects on foliage, insect
21    susceptibility, habitat, and community composition. The effects related to scenic beauty  (foliar
22    injury and insect damage) affect not only the scenery viewing, but also satisfaction with other
23    scenery-dependent activities. Ninety-seven percent of National Survey on Recreation and the
24    Environment (NSRE) survey respondents rated scenic beauty as an important or extremely
25    important aspect of their wilderness experience.
26           Perceptions of scenic beauty depend on a number of forest attributes,  including the
27    appearance of forest health, the effects of air pollution and insect damage, visual variety, species
28    variety, and lush ground cover (Ribe, 1989). The ISA concludes that there is a causal
29    relationship between 63 exposure and visible foliar injury. Figure 7-3 shows the effects of foliar
30    injury on ponderosa pine, milkweed, and tulip poplar.
                                                      7-5

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
       The presence of downed wood, whether caused by 63 mortality, insect attack, or other
causes, has a negative impact on scenic beauty assessments (Ribe, 1989; Buyhoff, et al., 1982).
Figure 7-4 shows the effects of southern bark beetle damage.  Species composition of forests
may also influence preferences.  According to Ribe (1994) these preferences may be affected by
cultural, regional, or contextual expectations, which would include the expectation of the
presence of certain species in specific areas (e.g., the presence of ponderosa pine in California).
In addition, there is a positive effect on preferences for ground cover rather than bare or
disturbed soil (Brown and Daniel, 1984, 1986).  Thus, the reduced value of scenic beauty from
(Vinduced effects on sensitive plants, by way of foliar injury, extends beyond large trees to the
grasses, forbs, ferns, and shrubs that comprise the understory of a forest setting.
       In Peterson et al. (1987), where O3-exposure had resulted in foliar injury to ponderosa
pines in the San Bernardino Forest, survey participants were asked to: (1) rank preferences for
scenic views, (2) rate their recreation experiences, (3) state how decreases in tree quality would
affect their visitation,  and (4) specify whether they would be willing to pay for programs to
mitigate the damage.  This survey showed that visible foliar injury had a negative impact on
perceptions of scenic beauty and a nonzero value for willingness to pay for programs to improve
forest aesthetics damaged by 63.
Figure 7-3 Examples of Foliar Injury from Os Exposure
Courtesy: NFS, Air Resources Division
                                                      7-6

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
Figure 7-4  Examples of Southern Bark Beetle Damage
Courtesy: Ronald F. Billings, Texas Forest Service. Bugwood.org

       The NSRE provides estimates of participation for many recreation activities.  According
to the survey some of the most popular outdoor activities are walking, including day hiking and
backpacking; camping; bird watching; wildlife watching; and nature viewing.  Participant
satisfaction with these activities depends wholly or partially on the quality of the natural scenery.
Table 7-2 summarizes the survey results, for these and other popular activities, including the
percent participation and the number of participants nationally, the number of days participants
engage in recreation activities annually, and their willingness-to-pay (WTP) for their
participation.
                                                      7-7

-------
 1    Table 7-2     National Outdoor Activity Participation
Activity
Day Hiking
Backpacking
Picnicking
Camping (Developed
and Primitive Sites)
Visit a Wilderness
Area
Birdwatching/
Photography
Wildlife Watching/
Photography
Natural Vegetation
Viewing/
Photography
Natural Scenery
Viewing/
Photography
Sightseeing
Gathering
(Mushrooms, Berries,
Firewood)
Percent
Participation
32.4
10.4
54.9
42.3
32.0
31.8
44.2
43.9
59.6
50.8
28.6
Number of
Participants
(in millions)
69.1
22.2
116.9
90.1
68.2
67.7
94.2
93.6
126.9
108.2
60.9
Number of
Activity
Days (in
millions)
2,508
224.0
935.2
757.5
975.4
5,828.1
3,616.5
5,720.8
7,119.7
2,055.0
852.7
Mean
WTP/Day
(in 2010$)
$60.63
$13.33
$20.70
$19.98
N/A
$49.74
$48.72
N/A
N/A
$45.94
N/A
Mean Total
Participation
Value
(in millions of
2010$)
$152,060
$2,986
$19,359
$15,135
N/A
$289,773
$176,196
N/A
N/A
$94,407
N/A
 2
 O

 4
 5
 6
 7
 8

 9

10

11

12
Source: NSRE 2010 and 2003 National Report on Sustainable Forest Management.  2003 National
Report: Documentation for Indicators 35, 36, 37, 42, and 43 available at:
http://warnell.forestry.uga.edu/nrrt/NSRE/MontrealIndDoc.PDF  and Recreation Values Database
available at: http://recvaluation.forestry.oregonstate.edu/
N/A = not available

       The relationship between scenic beauty and recreation satisfaction for camping was

quantified by Daniel et al. (1989) in a contingent valuation study.  The authors surveyed campers

regarding their perceptions of scenic beauty, as indicated by a photo array of scenes along a

spectrum of scenic beauty, and their WTP to camp in certain areas. All else being equal, scenic

beauty and WTP demonstrated a nearly perfect linear relationship (correlation coefficient of
                                                      7-8

-------
 1    0.96). This suggests that campers would likely have a greater WTP for recreation experiences in
 2    areas where scenic beauty is less damaged by 63.  Since as mentioned previously, Ribe (1994)
 3    found that scenic beauty plays a strong role in recreation satisfaction and explains 90 percent of
 4    the difference in recreation satisfaction among all types of outdoor recreation, there is reason to
 5    believe that this linear relationship between scenic beauty and WTP would hold across all
 6    recreation types.  We believe that it would follow that decreases in Os damage would generate
 7    benefits to all recreators. We cannot estimate the incremental impact of reducing Os damage to
 8    scenic beauty and subsequent recreation demand; however, given the large number of outdoor
 9    recreation participants and their substantial WTP for recreation, even very small increments of
10    change in WTP or activity days should generate significant benefit to these recreators.
11           Another resource for estimating the economic value of consumers' recreation experiences
12    is the data available on actual expenditures for recreation and the total economic impact of
13    recreation activities. Economic impacts across the national economy can be estimated using the
14    IMPLAN® model (MIG Inc, 1999).l  For this document we refer to analyses done for the 2011
15    National  Survey of Fishing, Hunting, and Wildlife-Associated Recreation (FHWAR) (U.S.
16    Department of the Interior, U.S. Fish and Wildlife Service, and U.S.  Department of Commerce,
17    2011) and an analysis performed by Southwick and Associates  for the Outdoor Industry
18    Foundation (OIF), The Economic Contribution of Active Outdoor Recreation - Technical Report
19    on Methods and Findings (OIF, 2012).
20           The FHWAR and the OIF report provide estimates of trip and equipment-related annual
21    expenditures for wildlife watching activities in the U.S. The OIF report also provides estimates
22    of recreators'  annual expenditures on trail-related activities, camping, bicycling, snow-related
23    sports, and paddle sports.  For this review, we include the data on trail-related activities and
24    camping as the most relevant for analysis of Os-related damages.
25           As shown in Table 7-3, the total expenditures across wildlife watching activities, trail-
26    based activities, and camp-based activities are approximately $230 billion dollars annually.
27    While we cannot estimate the magnitude of the impacts of Os damage to the scenic beauty, the
28    losses are reflected in the values reported.
      1 IMPLAN® is a commercially available input-output model that has been used by the Department of Interior, the
       National Park Service, and other government agencies in their analyses of economic impacts.

                                                      7-9

-------
 1
 2
Table 7-3     National Expenditures for Wildlife Watching, Trail, and Camp-Related
              Recreation (in billions of 2010$)
Expenditure Type
Trip-Related
Equipment & Services
Other Expenditures
Total for All Expenditures
Wildlife-Watching11
$16.7
$26.3
$10.2

Trail0
$53.7
$6.3
N/R

Campc
$109.3
$8.3
N/R

Total0
$179.7
$40.9
$10.2
$230.8
 4
 5
 6
 9
10
11
12
13
a Data from 2011 FHWAR
b Data from 2012 OIF report
N/R = not reported

       The impact of these expenditures has a multiplier effect through the economy, which was
estimated by OIF using the IMPLAN® model.2 The model estimates the flow of goods and
money through the economy at scales from local to national. According to the OIF report
(2012), trail activities generated over $190 billion in total economic activity, including $97
billion in salaries, and wages. The same report estimates the total economic activity generated
by camping-related recreation at $346 billion, including $175 billion in salaries, and wages. The
total economic activity estimates also include state and federal tax revenues.
14     7.2   NATIONAL-SCALE ANALYSIS OF FOLIAR INJURY
15          To assess foliar injury at a national scale, we compared data from the Forest Health
16    Monitoring Network (USFS, 2011) with O3 exposure estimates for individual years, described in
17    Section 4.3.1.2, and soil moisture data, which was estimated using NOAA's Palmer Z drought
18    index (NCDC, 2012b).
19              7.2.1     Forest Health Monitoring Network
20            The only national-scale data set pertaining to foliar injury is from the USDA Forest
21     Service's (USFS) Ozone Biomonitoring Program (OBP). This effort was completed as part of
22     the Forest Inventory and Analysis (FIA) and Forest Health Monitoring (FHM) programs (see
23     Figure 7-5 for O3 biomonitoring sites). The OBP used a number of bioindicator species (O3-
24     sensitive plants) to monitor the potential impacts of O3 on our nation's forests. The field
      2 Assumptions and Caveats to the IMPLAN®Results: Statistics on the precision of the final economic impacts were
      not produced by OIF because of feasibility issues. Harris Interactive survey results combine several parameters
      from the data, and outside data from the U.S. Bureau of the Census' population estimates and IMPLAN multipliers
      were used.
                                                     7-10

-------
 1     methods, sampling procedures, and analytical techniques were consistent across sites and
 2     between years (USFS, 2011).
 3            We obtained data on foliar injury from the USFS for the five years from 2006 to 2010.
 4     Because of privacy laws that require the exact location information of sampling sites
 5     ("biosites") to not be made public, the data were assigned to the CMAQ grid used for the Os
 6     exposure surface by the USFS (USFS, 2013).  Data were not available for California, Oregon,
 7     and Washington, so we used the publically available data. In those states we assigned the data
 8     to the CMAQ grid based on  the publically available geographic coordinates, which are masked
 9     for privacy concerns as mentioned above; the data in those states have additional uncertainty
10     relating the 63 and Palmer Z drought index data to the foliar injury data.  Also, because
11     sampling was discontinued in some states prior to this analysis, we did not include data for
12     most of the western states (Montana,  Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, New
13     Mexico, Oklahoma, and portions of Texas).
14           The biosite index is calculated from a combination of the proportion of leaves affected
15     on individual bioindicator plants. In order to calculate the biosite index, at least 30 individual
16     plants of two bioindicator species must be present at each biosite.  The mean severity of
17     symptoms ranges from a score of zero to a score of 100 (USFS, 2011).
                                                     7-11

-------
 2   Figure 7-5   Os Biomonitoring Sites
 3   Note: Sites are shown as the CMAQ grid cell in which they occur. Some biosites were sampled in more than one
 4   year, but are indicated on this figure only as the most recent year sampled.
 5              7.2.2     NOAA Palmer Z Drought Index
 6          The Palmer Z drought index represents the difference between monthly soil moisture and
 7   long-term average soil moisture (Palmer, 1965). These data typically range from -4 to +4, with
 8   positive values representing more wetness than normal and negative values representing more
 9   dryness than normal. Values between -1.25  and +1.0 could be interpreted as normal soil
10   moisture, whereas values beyond the range from -2.75 to +3.5 could be interpreted as extreme
11   drought and extremely moist, respectively (NCDC, 2012c).
12          The soil moisture index is calculated for each of the 344 climate regions divisions within
13   the contiguous U.S.  defined by the National Climatic Data Center (NCDC) (NOAA, 2012a).
14   Because we did not  have soil moisture data  outside of the continental  contiguous U.S., we did
15   not evaluate parks in Alaska, Hawaii, Puerto Rico, or Guam. We identify the NCDC climate
16   divisions with Palmer Z data in Figure 7-6.
                                                    7-12

-------
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
Figure 7-6    344 Climate Divisions with Palmer Z Soil Moisture Data
 Source: NCDC, 2012a

           7.2.3     Results
      Data were available for a total of 5,284 biosites across the five years from 2006 - 2010
(Table 7-4, Figure 7-5).  Table 7-4 summarizes the biosite index values for each year.  The
categories used in Table 7-4 follow the USFS risk categories with the exception of including a
separate category for a biosite index of zero, or no damage.  We included the data to highlight
that across all of the sites, over 81 percent of the observations recorded no foliar injury. This
percentage was similar across all of the years, with a low value of 78 percent and a high value of
85 percent. The data showed no clear relationship between Os and biosite index (Figure 7-7), as
well as no clear relationship between Os and the Palmer Z drought index (measured as an
average value of the months from April to August (Figure 7-8)).
                                                     7-13

-------
1   Table 7-4      Summary of Biosite Index Values for 2006 to 2010 O3 Biomonitoring Sites.
2   Categories modified from USFS (Smith et al., 2008)
Biosite Index
0
<5
5 to 15
15 to 25
>25
Damage
None
Very Light
Light
Moderate
Heavy
Total
2006
744
139
41
15
12
951
2007
769
131
29
6
4
939
2008
796
98
29
8
4
935
2009
902
135
61
6
8
1,112
2010
1,075
183
65
12
12
1,347
Total
4,286
686
225
47
40
5,284
3
4
5
                         o
                         LO -
                         o
                         o
                     X
                     
-------
                          o
                          LO -
                          o
                          o
                      X
                      
-------
 1    Table 7-5    Censored Regression Results
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
Coefficient
Intercept
W126
Palmer Z (Apr-Aug)
W 126: Palmer Z

W126
Palmer Z (Apr-Aug)
W 126: Palmer Z
Intercept Estimate
-22.5967
0.7307
1.8357
0.1357
Marginal Effect
0.1178
0.2960
0.0219
Std. Error
0.8934
0.0613
0.4850
0.0437

0.0099
0.0777
0.0070
t-value
-25.293
11.919
3.785
3.104

11.918
3.812
3.093
P
< 0.0001
0.0001
0.0002
0.0019

0.0001
0.0001
0.0020
       To further assess the relationship between 63 and foliar injury, we conducted a
cumulative analysis (Figures Figure 7-9 through Figure 7-12).  In these analyses, we ordered the
data by W126 index value, then for each W126 index value we calculated the proportion of sites
exceeding the selected biosite index value for all observations at or below that W126 index
value. We repeated this using an index value greater than zero, indicating presence of any foliar
injury, and an index value > 5, corresponding to a USFS cutoff for elevated injury (USFS,  2011).
In this analysis, we split the data into individual years, as well as into moisture categories;  the
moisture categories followed NOAA's Palmer Z drought index, with values less than -1.24
considered dry, values greater than or equal to 1 considered wet, and values between those
considered normal.
       When looking only at presence/absence of foliar injury ("any injury") (Figure 7-9), with
the  exception of 2008, the proportion of sites across all W126 index values exceeds 15 percent;
in 2006, it exceeds 20 percent, while in 2008 the proportion of sites with foliar injury across all
W126 index values was just below 15 percent.
       Although the overall percentages are much lower, when we use a biosite index of 5 or
greater as the benchmark (Figure 7-10), we see a similar pattern in which there is a rapid increase
in the proportion of sites exceeding a biosite index of 5 at W126 index values below 10 ppm-hrs.
In both cases, presence/absence and biosite index >  5, the data for 2007 show a more gradual
                                                     7-16

-------
 1   increase in proportion. The more gradual increase and relatively low overall proportions in 2007
 2   can at least be partly explained by 2007 being the driest year in the analysis. In contrast, 2008
 3   was a normal moisture and average O3 year among years in this analysis, which does not explain
 4   the consistently low proportions in 2008.
 5          There are two important observations that can be made in both of these analyses: (1) The
 6   proportion of sites exhibiting foliar injury rises rapidly at increasing W126 index values below
 7   10 ppm-hrs, and (2) there is relatively little change in the proportions above W126 index values
 8   of 20 ppm-hrs.
                                             Biosite Index > 0
10   Figure 7-9
11
                       I  £
                       'w  o
                       "6
                       c
                       _0
                       o
                       Q-  O
                       ol  o
                          8 -I
                          ci
                          8 -
                                         10
                                                   20
                                               W126 (ppm-hrs)
                                                             30
                                                                        40
Cumulative Proportion of Sites with Foliar Injury Present, by Year
                                                      7-17

-------
                                             Biosite Index > 5
                       o
                       2
                       CL
                          8 -I
                          0
                          p -
                          0
                                         10
                                                   20
                                               W126(ppm-hrs)
                                                             30
                                                                        40
 1
 2   Figure 7-10   Cumulative Proportion of Sites with Elevated Foliar Injury, by Year
 3
 4          When categorized by moisture categories, as defined by the average Palmer Z drought
 5   index, the data show a more distinct pattern.  Similar to the analysis by individual years, the most
 6   rapid increase in the proportion occurs at W126 index values below 10 ppm-hrs, but the moisture
 7   category has a much greater effect on the overall proportion (Figure 7-11 and Figure 7-12). In
 8   both analyses, there is again relatively little change in the proportion beyond a W126 of 20 ppm-
 9   hrs in normal and dry years.
10          The data for normal moisture sites are very similar to the dataset as a whole, with an
11   overall proportion of close to 18 percent for presence/absence, and close to 6 percent for sites
12   exceeding a biosite index of 5. Sites classified as wet (average Palmer Z > 1) have much higher
13   overall proportions at both any injury and elevated injury and a much more rapid increase in
14   proportion of sites with foliar injury present, exceeding 20 percent at W126 index values under 5
15   ppm-hrs. At sites considered dry (average Palmer Z < -1.24), the overall proportions are much
16   lower, around 10 percent and 4 percent for presence/absence and exceeding an index of 5. This
17   indicates that drought does provide protection from foliar injury as discussed in the ISA (U.S.
18   EPA, 2013), but not entirely.
                                                     7-18

-------
                                                 Biosite Index > 0
                           o
                           1   o
                           I   5
                              8 _
                                            10
                                                       20

                                                   W126(ppm-hrs)
                                                                  30
                                                                            40
2    Figure 7-11    Cumulative Proportion of Sites with Foliar Injury Present, by Moisture
3                   Category
                                              Biosite Index > 5
                       2
                       CL
                          s
                          O
                          O -
                          O
                                         10
                                                   20

                                               W126(ppm-hrs)
                                                              30
                                                                         40
5    Figure 7-12    Cumulative Proportion of Sites with Elevated Foliar Injury, by Moisture
6                   Category
                                                      7-19

-------
 1
 2
 3
 4
 5
 6
 7
 9
10
11

12
13
14
15
16
17
       In Figure 7-13, we provide the data separated by NOAA climate regions (Karl and Koss,
1984). Although we had data for most regions of the contiguous U.S., we did not have data for
the Southwest and limited data for the West and West North Central regions. For example, from
2006 to 2010, there were over 1,000 biosite index values each for the Northeast and Central
regions and no biosite index values  for the Southwest. In general, the regions show a similar
pattern: the proportion of biosites showing foliar injury increases steeply with W126 index
values up to approximately 10 ppm-hrs and is relatively constant  at W126 index levels abovelO
ppm-hrs.
                                        Biosite Index > 0
                           in
                           b
                           o
                           o
                           in
                           p _
                           o
                           o
                           d
                                                       • All
                                                       D Central
                                                       • East North Central
                                                       • Northeast
                                                       • Northwest
                                                       • South
                                                       • Southeast
                                                       a West
                                                       O West North Centra
                                         10
                                                    20
                                                W126 (ppm-hrs)
                                                              30
                                                                         40
Figure 7-13   Cumulative Proportion of Sites with Foliar Injury Present, by Climate
              Region

 7.3   SCREENING-LEVEL ASSESSMENT OF VISIBLE FOLIAR INJURY IN
       NATIONAL PARKS
       A study by Kohut (2007) assessed the risk of (Vinduced visible foliar injury on 63
bioindicators (i.e., (Vsensitive vegetation) in 244 parks managed by the NFS. Specifically,
Kohut (2007) estimated 63 exposure using hourly 63 monitoring data collected at 35 parks from
1995 to 1999, estimated O3 exposure at 209 additional parks using kriging, a spatial interpolation
                                                     7-20

-------
 1    technique, and qualitatively assessed risk. Kohut applied a subjective evaluation based on three
 2    criteria: (1) the frequency of exceedance of foliar injury "thresholds"4 using several 63 exposure
 3    metrics (i.e., SUM06, W126 and N100), (2) the extent that low soil moisture constrains O3
 4    uptake during periods of high exposure, and (3) the presence of Os sensitive species within each
 5    park. Based on these criteria, Kohut (2007) concluded that the risk of visible foliar injury was
 6    high in 65 parks (27 percent), moderate in 46 parks (19 percent), and low in 131 parks (54
 7    percent).
 8           In this assessment, we applied a modified screening-level approach using more recent 63
 9    exposure and soil moisture data for 214 parks in the contiguous U.S.5 Consistent with advice
10    from CAS AC (Frey and Samet, 2012a), we modified the approach used by Kohut (2007) to
11    apply the W126 metric alone, and, in doing so, we chose foliar injury benchmarks derived from
12    the analysis in section 7.2 that assesses soil moisture quantitatively.6
13               7.3.1      Screening Assessment Methods
14                     7.3.1.1 O3 Exposure
15           As described in Section 4.3.1.3, we used recent 63 monitoring data (2006-2010) to create
16    spatial surfaces of O3 exposure using the VNA interpolation method, which covers the
17    contiguous U.S. with a spatial resolution of 12 km by 12 km for each of the five years. This
18    method allowed us to assess parks in the contiguous U.S., including parks without Os monitors
19    located within their park boundaries. We provide the W126 estimates at each park by year in
20    Appendix 7A.
21                     7.3.1.2 Soil Moisture
22           As described in section 9.4.2 of the ISA (U.S.  EPA, 2013), soil moisture is a major
23    modifying factor for Os-induced visible foliar injury. Low soil moisture can limit the amount of
24    Os entering the leaf, which can decrease the incidence and severity of foliar injury during periods
      4 Kohut (2007) uses the term "foliar injury thresholds". In this assessment, we use the term "benchmarks" in order to
       avoid implying that foliar injury could not occur below these levels.
      5 The parks assessed here include lands managed by the NFS in the continental U.S., which includes National Parks,
       Monuments, Seashores, Scenic Rivers, Historic Parks, Battlefields, Reservations, Recreation Areas, Memorials,
       Parkways, Military Parks, Preserves, and Scenic Trails.
      6 We applied different foliar injury benchmarks in this assessment after further investigation into the benchmarks
       applied in Kohut (2007), which were derived from biomass loss rather than visible foliar injury. Kohut cited a
       threshold of 5.9 ppm-hrs for highly sensitive species from Lefohn (1997), which was based on the lowest W126
       estimate corresponding to a 10% growth loss for black cherry. For soil moisture, Kohut (2007) qualitatively
       assessed whether there appeared to be an inverse relationship between soil moisture and high O3 exposure.

                                                        7-21

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
of drought. To incorporate short-term soil moisture into the screening-level assessment, we
applied Palmer Z data for 2006 to 2010 (NCDC, 2012b). These data are for the contiguous U.S.
only.
       Short-term estimates of soil moisture are highly variable over time, even from month to
month within a single year. For this reason, we used an average estimate of soil moisture to
reflect the cumulative nature of foliar injury in each park in each year. To determine the
appropriate timeframe for the soil moisture average, we identified the months corresponding to
the highest W126 estimate at each park with an O3 monitor. As shown in Figure 7-14, the
highest 3-month W126 estimate for 98 percent of monitored parks occurred between March and
September, which roughly corresponds to the growing season. Based on this information, we
applied the 7-month average from March to September for each year in the screening-level
assessment for all parks. For parks with Os monitors, we also conducted sensitivity analyses
applying the 5-month soil moisture average from April to August and the 3-month soil moisture
average corresponding to the specific 3-months in the highest W126 estimate at that monitor. We
provide the average soil moisture estimates for each park by year and the timeframe of W126 for
monitored parks in Appendix 7A. We also provide figures illustrating the difference in soil
moisture across the 7-month, 5-month, and 3-month timeframes by year in Appendix 7A.
    40%
        •
        JFM   FMA  MAM  AMJ   MJJ   JJA    JAS   ASO   SON  OND
                      3 months used in W126 estimate
Figure 7-14    Timeframe of W126 Estimates for 57 Monitors Located in Parks
                                                    7-22

-------
 1                     7.3.1.3   GIS Analysis
 2           Using GIS (ESRI® ArcMAP™ 10), we spatially overlaid the interpolated O3 exposure
 3    surface and soil moisture  data (NCDC, 2012b) with NFS boundaries (USGS, 2003) to link these
 4    data to each park. First, we dissolved all of the internal boundaries for each park such that each
 5    park only had one park boundary. Next,  we spatially joined the soil moisture data and the
 6    gridded O3 exposure data  with the park boundaries, creating an average soil moisture estimate
 7    and O3 exposure estimate  at each park. To identify the parks with O3 monitors, we spatially
 8    overlaid the O3 monitor data with the NFS park boundaries and included only those monitors
 9    located within the park boundaries.7 We excluded all parks outside of the contiguous U.S.
10    because of the absence of soil moisture data, resulting in 42 parks with O3 monitors and 214
11    parks with  O3 exposure estimated from the interpolated surface.8 Figure 7-15 identifies the 214
12    parks included  in this assessment, including the 42 parks with O3 monitors. In Figure 7-16, we
13    provide the distribution of O3 exposure and average soil moisture estimates for the  214 parks for
14    each year in this assessment, noting the range of "near normal" soil moisture conditions as
15    defined by  NCDC (NOAA, 2012c).
      7 There are 57 O3 monitors located within NFS parks, and an additional 7 monitors are located within 1km of the
       park boundaries. Some monitors (e.g., at Rocky Mountain National Park) have addresses that imply locations
       within park boundaries but are actually located just outside the NFS boundary. We did not include the monitors
       located just outside of the parks in the monitored park assessment. In addition, nine parks contained more than one
       O3 monitor. We provide the O3 exposure and soil moisture data for the 57 monitors located within NFS parks in
       Appendix 7A.
      8 Along coastlines, the shapefile for soil moisture is more generalized than the shapefile for O3 exposure. Therefore,
       we manually linked the soil moisture data to (a) 8 seashore parks in order to include them in the 214 park
       assessment and (b) 4 park monitors for the 42 park assessment.

                                                        7-23

-------
1
2

3
                                                                                                 HOFR

                                                                                               DEWA

                                                                                              VAI
                                                                                         ALPOGETT
                                                                                        FONE.CHOH
                       GRBA
                                   COLM
                            CARE^BLCA  FLFO
                        LZIONBRCA         GRSA BEOl
                             GLCA MEVE
                          GRCA  _.J.        CAVO
               LAME  vm,-n
             MOJA     WUPA  "'•"CHCU

                    TUZ, PEFO
                                 ELMAPETRPE°

                                     SAPU

                                GICL
                                    WHSA
                              CHIR
                                      IGlSWtf
              Not monitored (172 parks, 80%)

              Monitored (42 parks, 20%)
(*Parks identified by park code, which are provided in Appendix 7A. Not all park labels shown due to overlap.

National Parks are prioritized in mapping.)
4    Figure 7-15     214 National Parks included in the Screening-Level Assessment
                                                            7-24

-------
   4.00 i

  Wet
   3.00 -
   -4.00
  Dry   0
                          2006
                 10         20         30

                             W126
  4.00 i

 Wet
  3.00 -


~ 2.00


  1.00 -


  0.00 -


 -1.00 -


 -2.00 -


 -3.00 -
                          2007
  -4.00
Dry
                                                                          •   t«
               10        20        30

                           W126
   4.00 -\
  Wet
   3.00 -


 "s 2'°°

 i i.oo
 <
 So
 Sr o.oo
 |

 17-1.00


 I -2.00
 a.

   -3.00 -
   -4.00
  Dry   0
                           2008
                 10         20         30         40

                             W126
  Wet
   3.00 -
NJ
b
o
r Z (average M

H1 O H1
b b b
o o o
   -4.00
  Dry   0
                            2010
                      2006-2010
  4.00 -|

Wet

  3.00 -


| 2.00


i i.oo

So
5 o.oo


N -1.00
OJ
E
"3 -2.00
  -4.00
 Dry
                                                                                    20         30

                                                                                      W126
• 2006

• 2007

 2008

• 2009

• 2010
(Shaded area represents "near normal" soil moisture (-1.25 > Palmer Z > 1)


Figure 7-16     Distribution of Os and Soil Moisture in 214 Parks by Year
                                                             7-25

-------
 1                    7.3.1.4  Benchmark Criteria
 2           For each park, we evaluated whether Os exposure exceeded certain foliar injury
 3    benchmark criteria in each year (2006-2010). Specifically, we derived 6 scenarios from the foliar
 4    injury assessment in section 7.2 for evaluation in this screening-level assessment. The base
 5    scenario as representing the W126  above which there was a consistent percentage (17.7 percent)
 6    of all biosites showing foliar injury, regardless of soil moisture. The other 5 scenarios explicitly
 7    consider soil moisture and represent W126 benchmarks corresponding to different percentages of
 8    biosites showing injury and the degree of injury (e.g., any injury or elevated injury). Four of
 9    these scenarios reflect the special status of parks as areas designated for protection, and thus
10    apply benchmarks corresponding to any visible foliar injury at certain percentages of biosites
11    (i.e., 5%, 10%, 15%, and 20%). One scenario represents elevated injury (i.e., visible foliar injury
12    exceeded a biosite index of 5), which is consistent with a USFS's biosite index cut-off for foliar
13    injury.9 These scenarios represent the full range of percentages of biosites showing visible foliar
14    injury in the assessment in section 7.2. In total, we evaluated 13 different W126 benchmarks
15    associated with the 6 foliar injury risk scenarios.
16           Table 7-6 provides the benchmark criteria for O3 exposure (as W126) and short-term,
17    relative soil moisture (Palmer Z) for each of these 6 scenarios. We provide the figures
18    corresponding to the benchmark criteria for each scenario in Appendix 7A.
19
20
      ' For further discussion of the biosite index, refer to section 7.2.
                                                      7-26

-------
 1
 2
Table 7-6      W126 Benchmark Criteria for Os Exposure and Relative Soil Moisture in 6
                Scenarios used in Screening-Level Assessment of Parks
      Scenario
          Description
    Normal
   (Palmer Z
 between-1.25
     andl)
    Wet
(Palmer Z> 1)
     Dry
(Palmer Z<-1)
      Base
           17.7% of all biosites in foliar injury analysis
           showed any injury (the W126 index value
           above which a consistent percentage of all
           biosites in the foliar injury analysis showed
           any injury)
                W126>10.46
         (soil moisture not considered)
      5% of
      biosites
           5% of biosites in foliar injury analysis showed
           any injury, reflects soil moisture
           categorization
W126>3.05
W126>3.76
W126>6.16
      10% of
      biosites
           10% of biosites in foliar injury analysis
           showed any injury, reflects soil moisture
           categorization
W126>5.94
W126>4.42
W126>24.61
      15% of
      biosites
           15% of biosites in foliar injury analysis
           showed any injury, reflects soil moisture
           categorization
W126>8.18
W126>4.69
N/A
      20% of
      biosites
          20% of biosites in foliar injury analysis
          showed any injury, reflects soil moisture
          categorization
N/A
W126>5.65
N/A
      5% of
      biosites,
      Injury > 5
           5% of biosites in foliar injury analysis showed
           injury equal or greater than 5 on the biosite
           injury index (e.g., 5% of leaf shows injury in
           10% of the leaves), reflects soil moisture
           categorization
W126>12.23
W126>7.02
W126>46.87
 3

 4

 5

 6

 7

 8

 9

10

11
                  7.3.1.5  Sensitive Vegetation Species

        Consistent with Kohut (2007), we identify the parks containing Os-sensitive vegetation

species (NFS, 2003, 2006b) and consider the results for parks without species as potential until

species are identified in field  surveys at these parks. In addition, we conducted a sensitivity

analysis where parks without sensitive species are assumed to not exceed the benchmark criteria.

Based on the NFS lists, 95 percent of the parks in this assessment contain at least one sensitive

species. We identify the parks with and without currently identified sensitive species in Figure

7-17. NFS (2003) defines a sensitive species as "species that typically exhibit foliar injury at or
                                                          7-27

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
near ambient ozone concentrations in fumigation chambers and/or are species for which ozone
foliar injury symptoms in the field have been documented by more than one expert observer."
According to NFS (2003), the list of sensitive species is limited in number of species because
few species from natural ecosystems have been fumigated in chambers or examined in the field
for Os symptoms.
                         	
                         3SPFOBU
                      /  T1CA  7*T
                                  ROMO
                      BA
                            ARCHCOUV
                        CARECANY| BLCA
       Species present (203 parks, 95%)
       Species not present (11 parks, 5%)
(Parks identified by park code. Not all park labels shown due to overlap. National Parks are prioritized in mapping.
Data source: NFS, 2003, 2006b)
Figure 7-17    Presence of Os-Sensitive Species in Parks

           7.3.2      Screening Assessment Results and Discussion
       Similar to Kohut (2007), we evaluated how often 63 exposure exceeded certain
benchmark criteria, the soil moisture conditions during high exposure periods, and the presence
of sensitive vegetation species. However, we updated the data for Os exposure, soil moisture, and
benchmark criteria.
       As shown in Figure 7-18, in the assessment of 214 parks, 11 percent exceeded the
benchmark criteria in the base scenario for all 5 years, 39 percent for at least 4 years, 58 percent
                                                     7-28

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
for at least 3 years, 70 percent for at least 2 years, and 83 percent for at least 1 year. Results for
each of the other scenarios vary. As the required percentage of biosites showing foliar injury
increases, the percentage of parks exceeding the benchmark criteria decreases. Similarly, as the
degree of injury increases (i.e., from any injury to elevated injury > 5), a smaller percentage of
parks exceed the benchmark criteria. As shown in Table 7-7, the percentage of parks exceeding
the benchmark criteria in any given year varies by scenario because Os exposure and soil
moisture vary by year.
       To compare geographic differences across the scenarios, we provide a graph of the
geographic breakdown for parks exceeding the benchmark criteria for at least 3 years in Figure
7-19 and maps of the full results for each scenario in Figure 7-20 to compare geographic
differences across the scenarios. Detailed results for each park and scenario, including additional
figures, are provided in Appendix 7A.
Table 7-7     Percent of 214 Parks that Exceed Benchmark Criteria in Each Year (2006-
              2010) in 6 Scenarios
Scenario
Base
5% of biosites, injury>5
5% of biosites, any injury
10% of biosites, any injury
15% of biosites, any injury
20% of biosites, any injury
2006
80%
65%
97%
81%
77%
8%
2007
69%
31%
96%
56%
48%
7%
2008
58%
43%
95%
83%
72%
12%
2009
12%
7%
96%
53%
33%
13%
2010
41%
29%
96%
80%
65%
4%
16
                                                     7-29

-------
                100%
1
2
3
         5%ofbiosites,
           any injury
         15%ofbiosites,
           any injury
         20%ofbiosites
           any injury
                                # Years Exceeding Criteria

Figure 7-18    Screening-Level Results for Foliar Injury in 214 Parks in 6 Scenarios
                                                      7-30

-------
               ro
               o.

               tt
                   200
                   150
                   100
                    50
0 -
• East North Central
• Northwest
• West North Central
West
• Central
• South
• Northeast
• Southeast
• Southwest
MM
All 214
Parks
8
11
20
19
17
21
27
39
52
Base
0
2
6
12
8
1
19
27
49
5%
biosites,
any injury
8
9
20
16
17
21
27
39
52
10%
biosites,
any injury
3
3
14
12
16
19
26
29
50
15%
biosites,
any injury
1
3
9
8
15
12
26
22
49
20%
biosites,
any injury
0
0
0
0
0
1
3
0
0
5%
biosites,
lnjury=5
0
1
0
4
3
3
15
10
36
                                             Scenario
2   Figure 7-19    Parks Exceeding Benchmark Criteria for at least 3 years by Scenario and

3                 Climate Region
                                                  7-31

-------
                      Key: All 5 years 4 years 3 years 2 years 1 year No years

                                                                  a
                                                                  JW-UDL
                                                           .'MM.    „£   »-4
                                                            1KA '  - -
                                                  a ra,   ™.  ..n..-""


                                                   UWS^5^ ;  ""!r^;w-
                                                                            „  .
15%ofbiosites,
any injury
                                                    20%ofbiosites,
                                                    any injury
5    Figure 7-20    Foliar Injury Results Maps for 6 Scenarios for 214 Parks

6    (Parks identified by park code. Not all park labels shown due to overlap. National Parks are prioritized in mapping.

7    Larger maps available in Appendix 7A.)
                                                      7-32

-------
 1          In the assessment of 42 parks with 63 monitors based on the interpolated surface, 24
 2    percent exceeded the benchmark criteria in the base scenario for all 5 years, 36 percent for at
 3    least 4 years, 57 percent for at least 3 years, 69 percent for at least 2 years, and 81 percent for at
 4    least 1 year. These results are generally similar to the results for the 214 park assessment for the
 5    base scenario, except that the monitored park analysis showed a higher fraction of parks that
 6    exceeded the benchmark criteria for all 5 years rather than at least 4 years. We provide the results
 7    of the monitored park assessment in Table 7-8. We also evaluated three different methods for
 8    assigning 63 exposure to parks with monitors: interpolated surface, highest monitor, and average
 9    monitor. The results using these methods are discussed in more detail in section 7.3.3.1.
10
11
                                                      7-33

-------
1   Table 7-8     Screening-level Foliar Injury Results in 42 Parks with an Os Monitor using 3
2                 Methods for Assigning Os Exposure to Each Park in Base Scenario.*
Park Name
Acadia National Park*
Agate Fossil Beds National Monument
Badlands National Park*
Big Bend National Park
Blue Ridge Parkway
Canyonlands National Park
Cape Cod National Seashore
Carlsbad Caverns National Park
City of Rocks National Reserve
Colorado National Monument
Congaree National Park
Cowpens National Battlefield
Craters of the Moon National Monument
Cumberland Gap National Historical Park
Death Valley National Park
Devil's Tower National Monument
Dinosaur National Monument
Glacier National Park*
Grand Canyon National Park
Great Basin National Park
Great Smoky Mountains National Park*
Indiana Dunes National Lakeshore
Joshua Tree National Park*
Lassen Volcanic National Park
Mesa Verde National Park
Mojave National Preserve
Mount Rainier Wilderness
State
ME
NE
SD
TX
NC
UT
MA
NC
ID
CO
sc
sc
ID
KY
CA
WY
CO
MT
AZ
NV
TN
IN
CA
CA
CO
CA
WA
Years with
Monitoring
Data
5
o
J
5
5
5
5
5
4
1
4
5
5
4
4
5
3
4
5
5
5
5
5
5
5
5
4
5
# Years Exceeding Benchmark
Criteria for Base Scenario
Interpolation
0
2
1
1
o
J
5
3
1
4
3
o
3
2
3
o
J
5
2
4
0
5
5
3
1
5
4
5
5
0
Highest
Monitor
1
1
1
3
1
5
3
2
0
2
2
2
1
1
5
0
2
0
4
4
4
1
5
3
5
4
0
Average
Monitor
0
1
0
3
1
5
3
2
0
2
2
2
1
1
5
0
2
0
4
4
3
1
5
3
5
4
0
                                                 7-34

-------
Park Name
Olympic National Park*
Padre Island National Seashore
Petrified Forest National Park
Pinnacles National Monument
Saguaro National Park
Saratoga National Historical Park
Scotts Bluff National Monument
Sequoia-Kings Canyon National Park*
Shenandoah National Park
Theodore Roosevelt National Park*
Tonto National Monument
Voyageurs National Park
Wind Cave National Park
Yellowstone National Park
Yosemite National Park*
Summary Results by O3 Exposure Method
State
WA
TX
AZ
CA
AZ
NY
NE
CA
VA
ND
AZ
MN
SD
WY
CA
All 5 years
At least 4 years
At least 3 years
At least 2 years
At least 1 year
No years
Years with
Monitoring
Data
1
2
5
5
5
5
1
5
5
5
5
5
5
5
5
71%
86%
93%
95%
100%
0%
# Years Exceeding Benchmark
Criteria for Base Scenario
Interpolation
0
0
4
3
4
0
3
5
2
0
5
0
2
1
5
24%
36%
57%
69%
81%
19%
Highest
Monitor
0
0
4
4
5
0
0
5
4
0
5
0
2
2
5
19%
36%
43%
60%
76%
24%
Average
Monitor
0
0
4
4
5
0
0
5
4
0
5
0
2
2
5
19%
33%
43%
60%
71%
29%
1
2
3
4
5
6
7
* More than one O3 monitor within park boundaries (shown as bold).

           7.3.3     Limitations and Uncertainty Characterization for Screening-Level
                     Assessment
       This is a screening-level assessment that primarily relies on national-level data with
coarse spatial resolution. As such, these results should be interpreted within the context of the
analytical limitations and with the appropriate uncertainty characterization, as noted below.
                                                     7-35

-------
 1                    7.3.3.1  O3 Exposure
 2          As noted by Kohut (2007), monitoring provides the most accurate assessment of Os
 3    exposure in specific locations, but a single monitor may not reflect the differences in exposure
 4    throughout the park. For this reason, we compared the results of the assessment for parks with 63
 5    monitors located within the park boundaries using the interpolated surface with the results based
 6    on 63 monitor data. As noted above, we conducted sensitivity analyses to evaluate the impact of
 7    this analytical choice for parks with more than one monitor, using both the highest monitor in the
 8    park, and the average of the monitors in the park for the base scenario. As shown in Table  7-8,
 9    the results using the highest monitor and average monitor were generally similar but tended to
10    exceed the benchmark criteria for fewer years. This result is primarily due to the absence of
11    monitoring data for a few years in a few parks, which lowered the maximum number of years
12    that a park could exceed the benchmark criteria using the highest monitor and average monitor
13    methods. For the 9 parks with multiple monitors, only 2 parks exceeded the benchmark criteria
14    for a  different number of years using the highest monitor compared to using the average of the
15    monitors. For both Acadia National Park and Badlands National Park, the benchmark criteria
16    were exceeded for 1 year using the highest monitor but for no years based on the average of the
17    monitors. For the 30 parks with all  5 years of monitoring data, 17 parks had the same results
18    using all 3 methods, 5 parks had more years using the interpolation, 5 parks had more years
19    using either monitor method,  and 3 parks had more years using the highest monitor.
20          In the 214-park assessment, we used interpolated surfaces to estimate Os exposure at each
21    of the parks. As such, we are not able to identify which 3 months are included in the W126
22    estimate at all of the parks. This limitation has two important implications. First, we have to
23    make assumptions regarding the months included in the 63 exposure estimate to match with the
24    months in the soil moisture estimate. Second, a few areas in the West, such as Utah, can
25    experience high Os episodes during the  winter, when many plants are dormant with limited
26    opportunities for Oj uptake.10 However, because only a few areas of the country have high Oj
27    episodes in winter when many plants are dormant and some sensitive vegetation species do not
28    shed  leaves  in the winter (e.g., Pinusponderosa),  we believe that this limitation contributes only
29    a small amount of uncertainty to the overall results of the 214-park assessment.  In addition,
      10 There are 11 parks in Utah, which all exceeded the benchmark criteria for at least 4 years in the base scenario.
       Only one park has a monitor (Canyonlands), and the 3-month timeframes corresponding to the highest W126
       estimates for 2006 to 2010 occurred between March and July.

                                                     7-36

-------
 1   based on the assessment of 42 parks with 63 monitors, less than 2 percent of the highest W126
 2   estimates occurred outside of the March to September timeframe.
 3          Because W126 estimates can be highly variable from year to year, the selection of
 4   different analysis years for this analysis could lead to different results. In Table 7-9, we provide
 5   the sensitivity of the results for the base scenario by splitting the data into two timeframes. In
 6   general, more parks show higher Os exposure during the first 3 years of the assessed timeframes
 7   (i.e., 2006-2008) than the last 3 years (i.e., 2008-2010). However, assessing the exceedances in
 8   specific years across scenarios in Table 7-9 shows that the scenario affects which years have the
 9   highest percentage of parks that exceed the benchmark criteria.
10          For more information regarding uncertainty in the 63 exposure estimates, see Chapter 4.
                                                     7-37

-------
1   Table 7-9    Foliar Injury Sensitivity Analyses for 214 Parks
Screening Criteria
03
Exposure
and Soil
Moisture
Scenarios
03
Exposure
only
Timeframe
Sensitive
Species
Base (W126>10.46)
5% of biosites, injury>5
5% of biosites
10% of biosites
15% of biosites
20% of biosites
W126>15 (alternative standard)
W126>13 (alternative standard)
W126>11 (alternative standard)
W126>9 (alternative standard)
W126>7 (alternative standard)
W126>46.87 (5% of biosites, injury>5,
dry)
W126>24.61 (10% of biosites, dry)
W126>12.23 (5% of biosites, injury>5,
normal)
W126>8.18 (15% of biosites, normal)
W126>7.02 (5% of biosites, injury>5,
wet)
W126>6.16 (5% of biosites, dry)
W126>5.94 (10% of biosites, normal)
W126>5.65 (20% of biosites, wet)
W126>4.69 15% of biosites, wet)
W126>4.42 (10% of biosites, wet)
W126>3.76 (5% of biosites, wet)
W126>3.05 (5% of biosites, normal)
2006-2008
2008-2010
Base (0 if no species)
Percent of Parks Exceeding Benchmark Criteria by #
Years (2006-2010)
All 5
years
11%
0%
91%
27%
11%
0%
3%
3%
8%
21%
41%
0%
1%
5%
27%
40%
51%
57%
61%
80%
85%
91%
96%
0%
0%
10%
At
least 4
years
29%
13%
4%
32%
35%
0%
9%
21%
33%
55%
73%
0%
1%
23%
63%
72%
77%
78%
80%
87%
90%
94%
97%
0%
0%
37%
At
least 3
years
19%
21%
2%
21%
22%
2%
23%
36%
52%
70%
80%
0%
3%
41%
73%
80%
82%
84%
87%
92%
93%
97%
98%
55%
11%
54%
At
least 2
years
12%
18%
0%
10%
14%
7%
39%
51%
66%
77%
86%
0%
3%
57%
81%
86%
89%
91%
92%
95%
95%
98%
98%
70%
40%
66%
At
least 1
year
13%
26%
0%
5%
8%
24%
58%
72%
81%
87%
92%
0%
4%
75%
89%
92%
95%
95%
95%
97%
97%
98%
98%
83%
61%
79%
No
years
17%
22%
2%
5%
10%
66%
42%
28%
19%
13%
8%
100%
96%
25%
11%
8%
5%
5%
5%
3%
3%
2%
2%
17%
39%
21%
                                                  7-38

-------
 1                    7.3.3.2  Soil Moisture
 2           Evaluating soil moisture is more subjective than evaluating Os exposure because of its
 3    high spatial and temporal variability within the Os season. Due to the size of the NCDC climate
 4    divisions (e.g., potentially hundreds of miles wide), soil moisture will vary within each region
 5    and potentially even within a park. For example, some vegetation along riverbanks may still
 6    experience sufficient soil moisture during periods of drought to exhibit foliar injury. Due to the
 7    spatial resolution of the soil moisture regions, the inability to capture within-region variability in
 8    soil moisture adds some uncertainty to this assessment, but we are currently unable to quantify
 9    the magnitude of this uncertainty. Regarding temporal variability,  averaging the monthly values
10    from May to October for each year also adds some uncertainty to this assessment. For example,
11    the average is sensitive to skew by a single very wet or very dry month within that timeframe or
12    even a single precipitation episode within a month. To evaluate the sensitivity of the results to
13    different averaging times, we conducted an analysis using the 7-month,  5-month, and 3-month
14    soil moisture average for parks with Os monitors. As shown in Table 7-10, the results for the 57
15    Os monitors in parks are not very sensitive to the different timeframes for soil-moisture data for
16    the 6 scenarios, although the specific 3-month  data tend to show slightly fewer parks that exceed
17    the benchmark criteria for more years. On balance, we believe that the spatial and temporal
18    resolution for the soil moisture data is likely to underestimate the potential of foliar injury that
19    could occur along some areas such as stream banks.
20
                                                      7-39

-------
 1
 2
Table 7-10     Foliar Injury Sensitivity Analyses for Soil Moisture in 57
               Parks*
                                                                            Monitors in
Scenario and Soil Moisture Method



7-month
Palmer Z
(Mar-
Sept)



5 -month
Palmer Z
(Apr-Aug)



Specific 3-
Month
Palmer Z
(based on
monitor)


Base
5% of biosites, injury>5
5%ofbiosites
10% of biosites

15% of biosites
20% of biosites
5% of biosites, injury>5
5% of biosites

10% of biosites

15% of biosites
20% of biosites
5% of biosites, injury>5
5% of biosites

10% of biosites

15% of biosites
20% of biosites
Percent of Parks Exceeding Benchmark Criteria by #
(2006-2010)
All 5 At least 4 At least 3 At least 2 At least 1
years
18%
4%
54%
14%

4%
0%
4%
53%

12%

4%
0%
4%
56%

9%

2%
0%
years
33%
12%
74%
42%

26%
2%
11%
72%

33%

18%
0%
11%
72%

25%

9%
0%
years
40%
28%
79%
60%

46%
2%
25%
79%

58%

44%
2%
19%
79%

54%

37%
0%
years
56%
37%
86%
75%

68%
12%
39%
88%

75%

67%
7%
37%
88%

74%

65%
7%
year
74%
63%
96%
86%

81%
32%
63%
96%

84%

79%
40%
58%
96%

86%

81%
32%
Years
No years

26%
37%
4%
14%

19%
68%
37%
4%

16%

21%
60%
42%
4%

14%

19%
68%
 3
 4
 5
 6
 9
10
11
12
* Includes multiple monitors in 9 parks. The results for the base scenario are constant across soil moisture methods
because this scenario does not incorporate soil moisture.
       In addition, we are unaware of a clear threshold for drought below which visible foliar
injury would not occur. In general, low soil moisture reduces the potential for foliar injury, but
injury could still occur because plants must open their stomata even during periods of drought. In
addition, the degree of drought necessary to reduce potential injury is not clear. As shown in
Figure 7-11, foliar injury occurs even at lower soil moisture. In addition, we applied NOAA's
categorization for "near normal" for Palmer Z data in defining 5 of the 6 scenarios, as described
in section 7.2. However, NOAA's categorization for Palmer Z data has been described as "rather
                                                       7-40

-------
 1    arbitrary" (Karl,  1986).n Because we use these scenarios to assess the impact of soil moisture on
 2    the potential for foliar injury, any uncertainties in interpreting the underlying soil moisture data
 3    are embedded within those scenarios. Because using a different categorization would lead to
 4    different benchmark criteria for Os exposure, it is not clear whether this uncertainty could
 5    underestimate or overestimate the potential foliar injury.
 6                     7.3.3.3  Foliar Injury Benchmarks in 6 Scenarios
 7            This assessment relies upon the foliar injury benchmarks derived from the analysis in
 8    section 7.2. The precision in these benchmarks reflects the precision in the underlying soil
 9    moisture and Os  exposure data. Each of the uncertainties identified in that analysis extend to this
10    assessment. In particular, due to the absence of biosite injury data in the southwest region and
11    limited biosite data in the west and west north central regions (see Figure 7-5), the benchmarks
12    applied may not be applicable to these regions. This absence of data results in additional
13    uncertainty in extrapolating the national-level benchmarks to parks, particularly in the southwest
14    region. As shown in Figure 7-19,  many of the parks that exceed the benchmark criteria for at
15    least 3 years are located in the southwest region across many scenarios. In Table 7-9, we provide
16    the sensitivity of the results for 214 parks to 6 different scenarios reflecting different
17    considerations of Os exposure, soil moisture, and degree of foliar injury. This analysis indicates
18    that the results are highly sensitive to the selected foliar injury benchmarks in the 6 scenarios.
19                     7.3.3.4  Sensitive Species
20           As noted by NFS (2003), relatively few vegetation species have been evaluated for O3-
21    sensitive foliar injury in the field, and continuing fieldwork will likely identify additional
22    sensitive species. Because we  did not exclude parks without identified sensitive species, some
23    parks that exceed the benchmark criteria may not actually have high potential for foliar injury.
24    However, due to the small number of parks without sensitive species (i.e., only 11 parks,  or 5
25    percent) and on-going fieldwork,  the magnitude of this uncertainty is likely to be small. As
      11 From Karl (1986), (p.83): "It should be emphasized that these qualitative descriptions are rather arbitrary. It is
       important to realize that the Z-index is standardized across all 12 months. This means that it is quite possible and
       common for some months which typically have low precipitation, and/or low moisture reserves, and/or high
       potential evapotranspiration, and/or low run-off (i.e. northern locations in winter, arid areas during the dry season)
       to never have Z-indices less than minus two. Contrarily, areas and times of the year which typically have favorable
       moisture conditions, high reserves, ample precipitation, low potential evapotranspiration, and high runoff, can have
       very low Z indices (
-------
 1    shown in Table 7-8, assuming that parks without sensitive species do not exceed the benchmark
 2    criteria does not significantly change the results for the base scenario.
 3                     7.3.3.5  Evaluation of Alternative Standards
 4           This screening-level assessment does not evaluate the model-adjusted W126 spatial
 5    surfaces for the scenarios of just meeting the existing 75 ppb (4th highest daily maximum) or
 6    alternative W126 standards.12  Because this screening-level assessment relies on year-by-year
 7    estimates of O3 exposure and soil moisture, it would not be possible to evaluate these year-by-
 8    year impacts using the attainment scenario surfaces, which were derived from 3 years of model-
 9    adjusted W126 data. Nevertheless, we can make a few observations regarding the potential
10    implications of just meeting existing and alternative standards. For example, as shown in Table
11    7-8, 42 percent of parks did not exceed 15 ppm-hrs during 2006-2010 using annual W126 data.
12    In addition, none of the 214 parks would exceed the annual benchmark criteria for the base
13    scenario (W126>10.46 ppm-hrs) after adjusting air quality to meet the existing standard (3-year
14    average W126 data).13 Similarly, Figure  7-21 shows that only 8 parks exceed 7 ppm-hrs after
15    adjusting air quality to just meet the existing standard (3-year average). Figure 7-22 shows 63
16    exposure in the 10 most-visited parks after adjusting air quality to meet the existing and
17    alternative standards. We provide the results for just meeting the existing standard and
18    alternative W126 standards at  each of the 214 parks in Appendix 7 A.
      12 W126 calculations are slightly modified in the case of the model adjustment scenarios described in Chapter 4,
      Section 4.3.4.  When calculating W126 for the model adjustment cases, we first found the three-year average of each
      three-month period, and then selected the three-month period with the highest three-year average using the same
      three-month period for each of the three years. In this way, the five scenarios are for recent air quality, air quality
      adjusted to just meet the current standard, and air quality further adjusted to just meet three different W126 index
      values: 15 ppm-hrs, 11 ppm-hrs, and 7 ppm-hrs.
      13 See Sections 4.3.1, 4.3.2, and 4.3.4 for more information regarding the air quality adjustments.

                                                        7-42

-------
                100%
              (J
              as
              o>
              •a
              o>
              ;g


              1
              Q.
              V)
              ra
              Q.
             80%
             60%
                 40%
              °  20%
              01
              Q.
                  0%
                                                               11 ppm-hrs     7 ppm-hrs
                                                                     i>15 ppm-hrs


                                                                      13-15 ppm-hrs


                                                                      11-13 ppm-hrs


                                                                      9-11 ppm-hrs


                                                                     17-9 ppm-hrs


                                                                     I <7 ppm-hrs
2

3

4
Figure 7-21
                    2006-2008   Just meeting 75  15 ppm-hrs

                                   ppb

                                        Adjustment Scenarios
Percent of 214 Parks at Different W126 Levels with Adjustments for

Existing and Alternative Standards (3-year average)
                                                     7-43

-------
                  2006-2008     Just meeting 75     15 ppm-hrs
                                    ppb
                 •Great Smoky Mountains National Park      ^—
                 •Yosemite National Park                  ^—
                 •Rocky Mountain National Park
                  Olympic National Park
                  Acadia National Park
     11 ppm-hrs
7 ppm-hrs
•Grand Canyon National Park
•Yellowstone National Park
 Zion National Park
•Grand Teton National Park
 Cuyahoga Valley National Park
 1
 2    Figure 7-22   Os Exposure in Ten Most Visited National Parks after Just Meeting Existing
 3                  and Alternative Standards (3-year average)

 4     7.4    NATIONAL PARK CASE STUDY AREAS
 5           The national parks represent a set of resources the public has agreed are special areas in
 6    need of protection for this and future generations to experience and enjoy.14 Because of this
 7    status risks to park resources are of special concern, particularly for bequest and option services
 8    because these services are specifically referenced in the creation of the parks.  The NFS is
 9    responsible for the protection of all resources within the national park system.  These resources
10    include those that are related to and/or dependent upon good air quality, such as whole
11    ecosystems and ecosystem components.
12           Several laws and policies protect the natural resources in national parks. The NFS, in its
13    Organic Act (16 U.S.C. 1), is directed to conserve the scenery, natural and historic objects and
      14
       C.F.R. 40, 81.400 provides for visibility protection for federal Class I areas.
                                                       7-44

-------
 1    wildlife and to provide for the enjoyment of these resources unimpaired for current and future
 2    generations.  The Wilderness Act of 1964 (Public Law 88-577, 16 U.S. C.  1131-1136) asserts
 3    wilderness areas will be administered in such a manner as to leave them unimpaired and preserve
 4    them for the enjoyment of future generations. NFS Management Policies (2006) guide all NFS
 5    actions including natural resources management. In general, the NFS Management Policies
 6    reiterate the NFS Organic Act's mandate to manage the resources "unimpaired." Although we
 7    have not quantified the monetary value of the bequest or option services given the data and
 8    methodology limitations inherent in such an effort, the status afforded these special areas through
 9    these laws and policies is indicative of their value to the public.
10           The ecosystem service we can  quantify, with some qualifications, is the recent monetary
11    value of the total recreation opportunity provided by the parks. We cannot quantify the loss in
12    monetary value for these services associated with Os; however, the magnitude of the overall
13    value is informative in understanding the potential significance of any Os damage (see Chapter 5
14    for more discussion).  The NFS has collected data on visitation, recreational activities, and
15    expenditures for trips to parks and modeled the economic impacts to local communities around
16    parks.  The NSRE provides WTP estimates for the value of recreation activities specific to the
17    regions where parks are located. Together these data allow us to estimate the magnitude of the
18    recreation services provided by parks.  The loss of service provision or visitor satisfaction due to
19    63 injury to sensitive  species in the case study parks is reflected in these estimates.
20           The three parks we are highlighting for case study analysis, Great Smoky Mountains NP,
21    Rocky Mountain NP,  and Sequoia/Kings Canyon NP, represent different regions of the country,
22    different ecosystems,  and Os conditions. Each park contains species sensitive to Os injury. The
23    text boxes accompanying each section highlight some of the reasons these parks were chosen for
24    special protection.
25           For the case study areas, we used the (Vsensitive species list from the preceding section
26    and cover data from VegBank plots (see Section 7.2).  The  resulting maps give cover estimates
27    for (Vsensitive species at the finer scale of the NFS vegetation map. It is important to note that
28    the cover estimates are separated into vegetation stratum (e.g., herb, shrub, tree) and it is possible
29    to have more than one vegetation strata present in a location.  As such, it is possible to have
30    sensitive species cover at a higher cumulative proportion than is shown here. We also used the
31    benchmarks presented in section 7.2 to assess the effect of just meeting the existing and
                                                     7-45

-------
1    alternative standards on W126 index values in the case study parks. We used a benchmark of 10

2    percent of biosites exhibiting injury in a normal year as the basis for the analysis, which is

3    depicted in Figure 7-23.



                                        Biosite Index >  0
            —
            O
            c
            o
            t
            o
            Q.
            o
                LO
                c\|
                d
                o
                c\i
                d
                LO

                d
           o

           d
                LO
                o
                o
                p
                o
                                  10
                                            I

                                           20


                                      W126(ppm-hrs)
 I

30
 I

40
5
6

7
Figure 7-23   Identification of W126 Index Value where 10 Percent of Biosites show Any
              Foliar Injury
                                                    7-46

-------
 1               7.4.1     Great Smoky Mountains
 2                        National Park
 3          In 2010, the Great Smoky Mountains National
 4    Park (GRSM) welcomed approximately 9.5 million
 5    visitors (NFS, 2010) making it the most visited national
 6    park in America.
 7          The "whole park" services affected by potential
 8    Os impacts include the existence, option, and bequest
 9    values and habitat provision discussed in Chapter 5.
10    Recreation value specific to the park is discussed later in
11    this section.
12          The extent of sensitive species coverage in
13    GRSM is substantial. Showing the percent cover of
14    species sensitive to foliar injury and focusing the analysis
15    on areas where recreation services are provided can
16    provide some perspective on the potential level of harm
17    to scenic beauty and recreation satisfaction within the
18    Park.
19          The NP S 2002 Comprehensive Survey of the
20    American Public, Southeast Region Technical Report
21    includes responses from recent visitors to southeast parks
22    about the activities they pursued during their visits (NFS,
23    2002a).  Using the 2010 annual visitation rate from the
24    NFS survey (NFS, 2010) and the regional results from
25    the Kaval and Loomis (2003) report on recreational use
26    values compiled for the NFS, we estimated visitors'
27    WTP for various activities; we present the estimates in
28    Table 7-11. In addition to the activities listed in the
29    table, 19 percent, or 1.8 million park visitors, benefited
30    from educational services offered at the park by
  Mount Le Conte, Summer
  Great Smoky Mountains National Park
  Courtesy: NPS
  http://www.nps.gov/grsm/photosmultimedia/index.htm

Great Smoky Mountains National Park is
the most visited national park in America
and a UNESCO World Heritage Site. The
Park is valued for the diversity of its
vegetation and wildlife; the scenic beauty
of its mountains, including the famous fogs
that give the Smoky Mountains their name;
and the preservation of the remnants of
Southern Appalachian culture. It is also
subject to high ambient 63 levels. The
park has recent 63 levels ranging between
W126 levels of 10 -  18 ppm-hrs with a
mean level of 14.7 ppm-hrs.
                                                     7-47

-------
 1
 2
 3
 4
participating in a ranger-led nature tour, which suggests that visitors wish to understand the
ecosystems preserved in the park.
Table 7-11    Value of Most Frequent Visitor Activities at Great Smoky Mountains
              National Park
Activity
Sightseeing
Day Hiking
Camping
Picnicking
Total
Percent
Participation
82
40
19
50

Number of
Participants
(thousands)
7,790
3,800
1,805
4,750

Mean WTP
(in 2010$)
53.34
69.93
29.87
42.42

Total Value of
Participation
(millions of 2010$)
416
266
54
201
937
 5
 6          The report Economic Benefits to Local Communities from National Park Visitation and
 1   Payroll (NFS, 2011) provides estimates of visitor spending and economic impacts for each park
 8   in the system. Visitor spending and its economic impact to the surrounding area are provided in
 9   Table 7-12 for the GRSM.  In addition, Table 7-13 includes data on the median value that
10   visitors spend on food, gas, lodging, and other items.
11   Table 7-12     Visitor Spending and Local Area Economic Impact of GRSM
Public Use Data
2010 Recreation
Visits
9,463,538
2010 Overnight
Stays
393,812
Visitor Spending 2010a
All
Visitors
$818,195
Non-Local
Visitors
$792,547
Impacts on Non-Local Visitor
Spending
Jobs
11,367
Labor
Income"
$303,510
Economic
Impact"
$504,948
12
13
14
 ($OOOs)
Source: NFS (2011)
                                                   7-48

-------
 1    Table 7-13     Median Travel Cost for GRSM Visitors
Expense
Gas and Transportation
Lodging
Food and Drinks
Clothes, Gifts, and Souvenirs
Total Per Visitor Party
Median Expenditures (2010$)
$73
$182
$73
$61
$389
 2    Source: NFS (2002a)
 3
 4          Each of the activities discussed above is among those shown in the national-scale
 5    analysis to be strongly affected by visitor perceptions of scenic beauty.  As discussed in Section
 6    7.1.1.2 for visible Os damage (Peterson, 1987) and for visible nitrogen and adelgid damage (a
 7    pest in Fraser fir) (Haefele et al., 1991 and Holmes and Kramer, 1996) visitors have a non-zero
 8    WTP for reductions in the described scenic impairments. As in the national analysis, it is not
 9    possible to assess the extent of loss  of services from impairment of scenic beauty by 63;
10    however, for the park these losses are captured in the estimated values for spending, economic
11    impact, and WTP.
12          GRSM is prized, in part, for its rich species diversity. The large mix of species includes
13    37 Os-sensitive species across vegetative strata, and many areas contain several sensitive species.
14    For instance, there may be a sensitive tall shrub occurring under the canopy of a sensitive tree
15    and various sensitive short shrubs or herbaceous plants occurring in the area of the tall shrub. In
16    areas where sensitive species overlap, it is possible to have sensitive species coverage
17    substantially higher than coverage for any one category of vegetation.  Figure 7-24 shows the
18    park coverage of various sensitive species.  Nearly 40 percent of the Park's 2,185 km2 total area
19    has sensitive tree cover (canopy and subcanopy) greater than 20 percent. Of that, 232 km2 has
20    sensitive tree species cover between 20 percent and 40 percent.  Shrubs account for 491 km2 of
21    sensitive vegetation, with over 100 km2 having over 80 percent of the species present as
22    sensitive. While sensitive herbaceous species occur throughout the park, the percent cover rarely
23    exceeds 20 percent.
24          We can quantify the extent of the hiking trails in  areas where sensitive species are at risk
25    for foliar injury.  Of the approximately 1,287 km of trails in GRSM, including approximately
                                                      7-49

-------
 1    114 km of the Appalachian Trail, over 1,040 km, or about 81 percent of trail area, are in areas
 2    where species sensitive to foliar injury occur. Figure 7-25 shows a summary of the overlap of
 3    the hiking trails in the GRSM, including a portion of the Appalachian Trail, with the species
 4    cover index. The accompanying pie charts in Figure 7-26 show the number of trail kilometers in
 5    each cover category.  The categories likely most visible to hikers are subcanopy trees, shrubs,
 6    and herbaceous vegetation.  There are 311 km, or about 24 percent, of trail area where sensitive
 7    subcanopy tree cover accounts for over 20 percent of the tree species present.  Sensitive shrubs
 8    cover over 20 percent of 549 km of trail area, or about 43 percent of total area.
 9          Although we cannot quantify the incremental loss of hiker satisfaction with their
10    recreation experience, this analysis illustrates that very substantial numbers of trail kilometers
11    are potentially at risk. With 3.8 million hikers using the trails every year and those hikers willing
12    to pay over $266 million for that activity, even a small benefit of reducing Os  damage in the park
13    could result in a significant value.
                                                      7-50

-------
                 Canopy
Sub Canopy
               Tall Shrub
                                                                 Short Shrubs

1
2   Figure 7-24   Cover of Sensitive Species in GRSM
                                                                                       No data
                                                                                       0%
                                                                                       <20%
                                                                                       20%-40%
                                                                                       40%-60%
                                                                                       60%-80%
                                                                                       >80%
                                           7-51

-------
               Canopy
Sub Canopy
                                                                                               Wo data
                                                                                               0%
                                                                                               <20%
                                                                                               20%-40%
                                                                                               40%-60%
                                                                                               B0%-80%
                                                                                               >80%
                                                                                               Appalachian Trail
1
2   Figure 7-25   Trail Cover of Sensitive Species in GRSM
                                             7-52

-------
Tree Canopy
177.4 135.2 "No Data
315'6 ' 20% to 40%
t73g 6 • 40% to 60%
60% to 80%

• >80%



Tall Shrub
640135.2 • No Data
155.0 ^^ -0%
• <20%
20% to 40%
1032'5 • 40% to 60%
• 60 % to 80%
• >80%
Tree Subcanopy
176.2 135.2 • N° Data
r'°\
395 4 <20/°
20% to 40%
948.9 • 40% to 60%
• 60% to 80%

• >80%


Short Shrub
\ Il35.2 -No Data
155.0 ^M ^12.1 B0./0
<20%
20% to 40%
1 1"7 • 40% to 60%
• 60 % to 80%
• >80%



Herbaceous


55.6 135.2 • N° Data
QBO%
<20%
20% to 40%
40% to 60%
60% to 80%
• >80%


 1
 2   Figure 7-26    GRSM Trail Kilometers by Species Cover Category
 3
 4          One of the amenities provided by GRSM is the scenic views from the roads and trails —
 5   the views from the scenic overlooks are one of the major park attractions.  On a day with natural
 6   viewing conditions visitors can see about 150 km across the mountain ridges of North Carolina
 7   and Tennessee, far outside the borders of the park itself. On average viewing days visitors can
 8   still see about 40 km, again outside the park itself. Figure 7-27 shows the sensitive tree canopy
 9   cover within a 3 km buffer of the overlooks. Within these small buffers 78 km2 have sensitive
10   species cover over 20 percent. While there are no data on the number of visitors stopping at the
11   overlooks, almost 8 million visitors identify sightseeing as one of their activities in the Park.
12   With their collective WTP for this activity over $400 million, it seems reasonable to conclude
13   that park visitors substantially value the scenic quality of the overlooks. Os concentrations in
14   GRSM have been among the highest in the eastern U.S., sometimes twice as high as neighboring
15   cities such as Atlanta and Knoxville. Under recent conditions 44 percent, or 959 km2, of the park
16   has W126 index values above 15 ppm-hrs.  After just meeting the existing standard at 75 ppb,
                                                     7-53

-------
 1
 2
 3
 4
 5
 6
W126 index values are reduced such that no area is over 7 ppm-hrs.  Just meeting the alternative
of 15 ppm-hrs produces the same result as meeting the existing standard. The lower alternative
standards of 11 and 7 ppm-hrs result in the park having W126 index values under 3 ppm-hrs for
the entire park, with most of the park under 2 ppm-hrs after just meeting the 7 ppm-hrs standard
level.  See Table 7-14 for additional details.
 7
11
 8   Figure 7-27    Sensitive Vegetation Cover in GRSM Scenic Overlooks (3km)
10   Table 7-14     Geographic Area of GRSM after Just Meeting Existing and Alternative
              Standard Levels (km )

Recent conditions
(2006-2008)
Just meeting 75 ppb
15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
Under 5.94
ppm-hrs
0
2,185
2,185
2,185
2,185
Between 5.95
and 7 ppm-hrs
0
0
0
0
0
Between 7-
llppm-hrs
48
0
0
0
0
Between 11-15
ppm-hrs
1,178
0
0
0
0
Over 15
ppm-hrs
959
0
0
0
0
12
                                                   7-54

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
           7.4.2     Rocky Mountain National Park
       In 2010 Rocky Mountain National Park (ROMO) welcomed
3 million visitors (NFS, 2010) to its 1,075 km2 of mountain
ecosystems. ROMO allows visitors to enjoy vegetation and
wildlife unique to these ecosystems along over 483 km of hiking
trails.
       The NFS 2002 Comprehensive Survey of the American
Public, Intermountain Region Technical Report includes responses
from recent visitors to intermountain parks about the activities they
pursued during their visit (NFS, 2002b). As in the GRSM case
study, using the 2010 visitation rate from the NFS survey (NFS,
2010) and the regional results from the Kaval and Loomis (2003)
report on recreational use values compiled for the NFS, we present
estimates for visitors' WTP for various activities in Table 7-15.
Table 7-15  Value of Most Frequent Visitor Activities at
              ROMO
Activity
Sightseeing
Day Hiking
Camping
Picnicking
Total
Percent
Participation
85
51
27
38

Number of
Participants
(thousands)
2,550
1,520
810
1,140

Mean
WTP
(in
2010$)
$28.17
$46.03
$41.47
$33.77

Total Value of
Participation
(millions of
2010$)
$72
$70
$34
$38
$214
       In addition to the activities listed in Table 7-15, 11 percent
of, or 330,000, park visitors took advantage of educational services
offered at the park by participating in a ranger-led nature tour.
       Each of the activities discussed above are among those
shown in the national-scale analysis to be strongly affected by
visitor perceptions of scenic beauty.  As in the national analysis it is
                                                     7-55
  Sheep Lakes
  Courtesy: NFS
  http://www.nps.gov/romo/photosmulti
  media/index.htm

Rocky Mountain National Park
features riparian ecosystems
with 150 lakes and 450 stream
miles that support lush
vegetation. The montane
ecosystem includes pine
forests and grasslands, while
subalpine elevations present
spruce and fir trees weathered
by the elements. The alpine
ecosystems are too harsh for
trees, but support low  growing
plants. The park has recent Os
levels ranging between W126
levels of 2 - 54 ppm-hrs with a
mean level of 14.2 ppm-hrs.
          •-.5

-------
 1    not possible to assess the extent of loss of services due to impairment of scenic beauty due to 63
      damage; however those losses are captured in the estimated values for spending, economic
      impact, and WTP for the park.  If O3 impacts were lower these estimated values would likely be
      higher.
            The report Economic Benefits to Local Communities from National Park Visitation and
      Payroll (NFS, 201 1) provides estimates of visitor spending and economic impacts for each park
      in the system. Visitor spending and its economic impact to the surrounding area are given in
 8    Table 7-16 for the ROMO.  Table 7-17 includes data on the median value that visitors spend on
 9    food, gas, lodging, and other items.
10
      Table 7-16     Visitor Spending and Local Area Economic Impact of ROMO
Public Use Data
2010
Recreation
Visits
2,955,821
2010
Overnight
Stays
174,202
Visitor Spending 2010
All Visitors
229,032
Non-Local
Visitors
221,896
Impacts on Non-Local Visitor Spending
Jobs
3,316
Labor
Income"
$89,975
Economic
Impact"
$ 155,157
11
12
13
14
     a($OOOs)
     Source: NFS (2011)
     Table 7-17     Median Travel Cost for ROMO Visitors
Expense
Gas and Transportation
Lodging
Food and Drinks
Clothes, Gifts, and Souvenirs
Total per Visitor Party
Median Expenditures (in 2010$)
$63
$100
$63
$45
$271
15
16
17
18
19
20
21
     Source: NFS (2002b)

            Unlike GRSM, only 7 sensitive species provide cover in ROMO as depicted in Figure
     7-28. The most notable of these is Quaking Aspen, or Populus tremuloides. This is significant
     in that many of the visitors to ROMO visit specifically to see this tree in its fall foliage. In some
     areas of the park, cover of this species can reach 80 percent. The species is found, along with the
     other sensitive tree species silver wormwood and Scouler's willow, in all vegetative layers in the
                                                    7-56

-------
 1    park. Sensitive species cover in just the tree canopy, subcanopy, and tall shrub layers is over 40
 2    percent in 328 km2, or 30 percent, of the park.
 3          We were able to quantify the extent of the hiking trails present in areas where sensitive
 4    species are at risk for foliar injury.  Of the approximately 562 km of trails in ROMO, including
 5    approximately 87 km of the Continental Divide National Scenic Trail, over 242 km, or about 43
 6    percent of trail  area, are in areas where species sensitive to foliar injury in the canopy, subcanopy
 7    or tall shrub category occur in greater than 20 percent coverage.  Figure 7-29 maps the  hiking
 8    trails in ROMO, including the relevant portion of the Continental Divide National Scenic Trail
 9    overlaid with the species cover index.  The accompanying pie charts in Figure 7-30 show the
10    number of trail km in each cover category.
11          Again, although we are not able to quantify the impact of this scenic damage on hiker
12    satisfaction, given 1.5 million hikers in ROMO and their $70 million WTP for the hiking
13    experience, even a small improvement in the scenic value could be significant.  While we did not
14    map the scenic overlooks in ROMO, given the 2.5 million visitors who come to the park to
15    sightsee and the $72 million they are willing to pay for this activity, it is reasonable to conclude
16    that any improvement in the scenic quality of the vistas at the overlooks would be of significant
17    value.
18          Under recent conditions, all 1,067 km2 of the park have W126 index values over 15 ppm-
19    hrs. Meeting the existing standard would bring about 59 percent of the Park into the 7-15 ppm-
20    hrs range, with the remaining 440 km2 under 7 ppm-hrs. Assessing an alternative standard of 15
21    ppm-hrs would bring the entire park under 7 ppm-hrs. See Table 7-18 for a  summary of full
22    results.
23
                                                     7-57

-------
1   Table 7-18
2
Geographic Area of ROMO after Just Meeting Existing and Alternative
Standard Levels (km2)

Recent conditions
(2006-2008)
Just meeting 75 ppb
15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
Under 5.94
ppm-hrs
0
37
986
1,067
1,067
Between
5.95-7
ppm-hrs
0
403
81
0
0
Between 7-11
ppm-hrs
0
627
0
0
0
Between 11-15
ppm-hrs
0
0
0
0
0
Over 15
ppm-hrs
1,067
0
0
0
0
                                                 7-58

-------
1
     EmcrgentTree^
      \	'
             -•


                    Sub Canopy
                                     Short Shrub
                                                          -1
                                                         V
Herbaceous  \          /
               W^_,-__.,
Canopy
                                                                                —' •• ------
                                                    Tall Shrub
                                                                   No Data
                                                                   0%
                                                                   >20%
                                                                   20%-40%
                                                                   40%-60%
                                                                   60%-80%
                                                                   >80%
2   Figure 7-28    Sensitive Species Cover in ROMO
                                          7-59

-------
1
     Emergent Tree
      Dwarf Shrub
               Herbaceous
Sub Canopy
Canopy
2  Figure 7-29   ROMO Sensitive Species Trail Cover
                                       7-60

-------
                Trees Canopy                  Tree Subeanopy                Emergent Trees
         11.7  19.1-
         Ll./^-^'-^-v
            -C  Y°
           24.3->.^           • No Data              4-°


                      -1.5
                               20% to 40%                 262.7
            2glo               • 40% to 60%
                              • 60% to 80%        Igg4
                              • >80%
                               I <20%

20% to 40%
            	     40% to 60%
40% to 60%
                                                                                               • >80%
                                                               l>80%
             _394-i Tall Shrubs                     Short Shrubs                   Dwarf Shrubs

                                                                                  ^^•°2    .N
                             • o%                               BO%                     ^^.     man
                              <20%                             •<20%
                              20% to 40%                           20% to 40%              '            20% to 40%
                              40% to 60%        32?'9               • 40% to 60%
                              60 % to 80%                          • 60 % to 80%
                             • >80%                             • >80%                            • >80%
                                 Herbaceous
                                               <20%
                                               40%tof
3    Figure 7-30    ROMO Trail Cover by Sensitive Species Type
4
5
6
                                                            7-61

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
           7.4.3     Sequoia and Kings Canyon
                    National Parks
       Sequoia and Kings Canyon National Parks (SEKI)
are located in the southern Sierra Nevada Mountains east of
the San Joaquin Valley in California.  The two parks
welcomed 1.6 million visitors in 2010 (NFS, 2010) to
experience the beauty and diversity of some of California's
iconic ecosystems.
       The NFS 2002 Comprehensive Survey of the
American Public, Pacific West Region Technical Report
includes responses from recent visitors to western parks
about the activities they pursued during their visit (NFS,
2002c). By using the 2010 annual visitation rate from the
NFS survey and the regional results from the Kaval and
Loomis (2003) report on recreational use values compiled
for the NFS, we estimated visitors' WTP for various
activities; the results  are presented in Table 7-19.
Table 7-19   Value of Most Frequent Visitor Activities
              at Sequoia and Kings Canyon National
              Parks
Activity
Sightseeing
Day
Hiking
Camping
Picnicking
Total
Percent
Participation
81
58
33
45

Number of
Participants
(thousands)
1,300
928
528
720

Mean
WTP
(in
2010$)
$24.21
$27.77
$124.65
$76.72

Total Value of
Participation
(millions of
2010$)
$31
$26
$66
$55
$178
       In addition to the activities listed in Table 7-19, 14
percent of, or 224,000 park visitors availed themselves of

                                               7-62
  Kings Canyon
  Courtesy: NPS,
  http://www.nps.gov/seki/photosmultim
  edia/index.htm

The Sequoia and Kings
Canyon National Parks share a
boundary and natural
resources. The natural
resource features include the
giant sequoia trees (and other
species, including ponderosa
and Jeffrey pine). The varied
ecosystems from the top of
Mount Whitney to the marble
caverns provide habitat for a
rich diversity of species. The
park has recent O3 levels
ranging between W126 levels
of 34 - 53 ppm-hrs with a
mean level of 43ppm-hrs.

-------
 1    educational services offered at the park by participating in a ranger-led nature tour, which
 2    suggests that visitors wish to understand the ecosystems preserved in the park.
 3          Each of the activities discussed above is among the activities shown in the national-scale
 4    analysis to be strongly affected by visitor perceptions of scenic beauty. As in the national
 5    analysis, it is not possible to assess the extent of loss of services resulting from impairment of
 6    scenic beauty due to Os damage; however, these losses are captured in the estimated values for
 7    spending, economic impact, and WTP for the parks. If Os impacts were lower these estimated
 8    values would likely be higher.
 9          The report Economic Benefits to Local Communities from National Park Visitation and
10    Payroll (NFS, 2011) provides estimates of visitor spending and economic impacts for each park
11    in the system. Visitor spending and its economic impact to the surrounding area are provided in
12    Table 7-20 for SEKI.  In addition, Table 7-21 includes data on the median value that visitors
13    spend on good, gas, lodging, and other items.
14    Table 7-20    Visitor Spending and Local Area Economic Impact of SEKI
Public Use Data
2010
Recreation
Visits
1,320,156
2010
Overnight
Stays
438,677
Visitor Spending 2010a
All Visitors
$97,012
Non-Local
Visitors
$89,408
Impacts on Non-Local Visitor Spending
Jobs
1,283
Labor
Income"
$37,299
Economic
Impact"
$60,504
15
16
17
a($OOOs)
Source: NFS (2011)
18    Table 7-21
               Median Travel Cost for SEKI Visitors
Expense
Gas and Transportation
Lodging
Food and Drinks
Clothes, Gifts, and Souvenirs
Total per Visitor Party
Median Expenditures (in 2010$)
$75
$150
$98
$63
$386
19
20
21
Source: NFS (2002c)
                                                     7-63

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
       There are 12 identified sensitive species in SEKI.  The percent coverage of these species
is depicted in Figure 7-31.  Areas of the parks with sensitive species cover of over 20 percent in
the canopy comprise 646 km2, or about 20 percent of the total area of SEKI.  This area
encompasses about 285 km of the 1,287 km (22 percent) of hiking trails available to
approximately 928,000 hikers in the parks. Figure 7-32 depicts the sensitive species cover across
the trail system, including the portion of the John Muir Trail that crosses the Parks' 19 km,
which has sensitive species coverage over 20 percent.  Figure 7-33 shows the sensitive species
by type.
       Again, although we are not able to quantify the impact of this scenic damage on hiker
satisfaction for hikers in SEKI and their $26 million WTP for the experience, even a small
improvement in the scenic value could be significant.
       As in the previous case studies, moving from recent conditions to meeting the existing Os
standard results in a large change in the area of the parks with exposures above 15 ppm-hrs.  For
SEKI, this means the parks move from all areas experiencing exposures above  15 ppm-hrs to all
areas in the SEKI having exposures below 7 ppm-hrs.  At lower alternative standards, SEKI
moves to exposures below 3 ppm-hrs.  See Table 7-22 for additional details.
Table 7-22     Geographic Area of SEKI after Just Meeting Existing and Alternative
              Standard Levels  (km2)

Recent conditions
(2006-2008)
Just meeting 75 ppb
15 ppm-hrs
1 1 ppm-hrs
7 ppm-hrs
Under 5.94
ppm-hrs
0
3,466
3,466
3,466
3,466
Between 5.95-7
ppm-hrs
0
0
0
0
0
Between 7-11
ppm-hrs
0
0
0
0
0
Between 11-15
ppm-hrs
0
0
0
0
0
Over 15
ppm-hrs
3,466
0
0
0
0
19
                                                    7-64

-------
        Canopy
TallShrub
                                                                             Herbaceous

        •
i
2   Figure 7-31   Sensitive Species Cover in SEKI
                                                                No Data
                                                                0%
                                                                <20%
                                                                20%-40%
                                                                40%-60%
                                                                69%-80%
                                                                >80%
                                         7-65

-------
1
         Canopy
Tall Shrub
                                                                                 Herbaceous
                Sensitive Species Cover
                Tall Shrub
                ^^— No Data

                     <2Q%
                     20%-40%
                     40%-6D%
                     69°/^8D%
                — >80%
                     John Muir Trail	
2   Figure 7-32   Sensitive Species Trail Cover in SEKI
                                           7-66

-------
              Tree Canopy
                      Tall Shrub
                                 Herbaceous
              V"
         107.6^^1

         177.7
                                             1.9

          289.3
I No Data
10%
l<20%
 20% to 40%
 40% to 60%
 60 % to 80%
l>80%
                      1.9
      138.0
547.2
    w
I No Data
10%
I <20%
 20% to 40%
 40% to 60%
 60 % to 80%
I >80%
5.0
 fc
I No Data
10%
l<20%
 20% to 40%
 40% to 60%
 60 % to 80%
l>80%
 3    Figure 7-33    SEKI's Sensitive Species Cover by Type

 4     7.5    QUALITATIVE ASSESSMENT OF UNCERTAINTY
 5           As noted in Chapter 3, we have based the design of the uncertainty analysis for this
 6    assessment on the framework outlined in the WHO guidance (WHO, 2008). For this qualitative
 7    uncertainty analysis, we have described each key source of uncertainty and qualitatively assessed
 8    its potential impact (including both the magnitude and direction of the impact) on risk results, as
 9    specified in the WHO guidance. In general, this assessment includes qualitative discussions of
10    the potential impact of uncertainty on the results (WHO Tierl) and quantitative sensitivity
11    analyses where we have sufficient data (WHO Tier 2).
12           Table 7-23 includes the key sources of uncertainty identified for the O?, REA. For each
13    source of uncertainty, we have (a) provided a description, (b) estimated the direction of influence
14    (over, under, both, or unknown) and magnitude (low, medium, high) of the potential impact of
15    each source of uncertainty on the risk estimates, (c) assessed the degree of uncertainty (low,
16    medium,  or high) associated with the knowledge-base (i.e., assessed how well we understand
17    each source of uncertainty), and (d) provided comments further clarifying the qualitative
18    assessment presented. The categories used in describing the  potential magnitude of impact for
19    specific sources of uncertainty on risk  estimates (i.e., low, medium,  or high) reflect our
20    consensus on the degree to which a particular source could produce a sufficient impact on risk
21    estimates to influence the interpretation of those estimates in the context of the secondary Os
22    NAAQS review. Where appropriate, we have included references to specific sources of
23    information considered in arriving at a ranking and classification for a particular source of
24    uncertainty.
                                                      7-67

-------
1    Table 7-23     Summary of Qualitative Uncertainty Analysis in Visible Foliar Injury Assessments.
       Source
         Description
   Potential influence of
    uncertainty on risk
         estimates
                                                     Direction
                                              Magnitude
              Knowledge-
                  Base
                   Comments (KB: knowledge base, INF: influence of
                            uncertainty on risk estimates)
A.  National W126
surfaces
The foliar injury analyses in
this chapter use the interpolated
W126 surfaces for individual
years (2006-2010),  as well as
the surfaces for recent
conditions and adjusted to just
meet the existing standard and
alternative W126 standards.
Both
Low-
Medium
Low-medium
KB and INF: See Chapter 4 for more details.
B. Surveys of
recreational activities
Survey estimates of
participation rates, visitor
spending/economic impacts,
and willingness-to-pay are
inherently uncertain. These
surveys potential double-count
impacts based on the allocation
of expenditures across activities
but also potentially exclude
other activities with economic
value.
Both
Medium
Medium
KB: Each survey (NSRE, FHWAR, OIF, NFS, etc) uses
different survey methods, so it is not appropriate to generalize
across the surveys. In general, the national level surveys apply
standard approaches, which minimize potential bias.
INF: Since the surveys are in agreement that there are millions
of outdoor recreationists and billions of recreation days across
various recreation types even small changes induced by
changes in recreation satisfaction due to O3 injury to
recreation sites could potentially result in large changes in the
value of outdoor recreation.
C. Ozone sensitive
species
Only species identified as O3-
sensitive by NFS are included
in the analyses.
Under
Medium
Medium
KB: Relatively few vegetation species have been evaluated for
Os-sensitive foliar injury in the field and continuing fieldwork
will likely identify additional sensitive species (NFS, 2003).
INF: The identification of additional sensitive species would
likely increase the extent of foliar injury in additional
locations and the percentage of injured vegetation at a
location.
                                                            7-68

-------
       Source
         Description
    Potential influence of
    uncertainty on risk
         estimates
                                                     Direction
                                               Magnitude
               Knowledge-
                  Base
                   Comments (KB: knowledge base, INF: influence of
                             uncertainty on risk estimates)
D. Spatial assignment
of foliar injury
biosite data to 12x12
km grids
Because of privacy laws that
require the exact location
information of sampling sites to
not be made public, the data
were assigned to the CMAQ
grid by the USFS. Data in
California, Oregon, and
Washington were assigned to
the CMAQ grid based on
publically available geographic
coordinates; thus, these data
have a higher level of
uncertainty.
Both
Low
Medium-Low
KB: The FHM biosites are small relative to the 12x12 km
CMAQ grids. The publically available data have the latitude
and longitude fuzzed by up to 7km in any direction, so in
California, Washington and Oregon so it is possible these sites
were assigned to the wrong CMAQ grid. In the remaining
states, the CMAQ grid was assigned from the actual locality
data.
INF: Having precise geographic locations would reduce
uncertainty, but the direction is unclear. The sites would most
likely be assigned to an adjacent CMAQ grid cell. Due to the
interpolation of the surfaces, differences between adjacent
cells are relatively small, so the magnitude of this effect is
likely small.
E. Availability of
biosite sampling data
Because sampling was
discontinued in some states
prior to this analysis, we did not
include data for many western
states (Montana, Idaho,
Wyoming, Nevada, Utah,
Colorado, Arizona, New
Mexico, Oklahoma, and
portions of Texas).
Unknown
Medium
Low
KB: Due to the discontinued sampling, data are not available
in these areas. It appears unlikely that sampling will resume in
those regions at this time.
INF: It is unclear how the addition of biosites from these
states would affect the risk estimates. The absence of biosite
sampling data in the southwest region and limited data in the
west and west north central region results in national
benchmarks that may not be applicable to these region. The
southwest in particular has generally higher W126 index
values than other regions,  so data from that region would be
important. In addition, the southwest has many national parks.
F. Soil moisture
threshold for foliar
injury
Low soil moisture reduces the
potential for foliar injury, but
injury could still occur because
plants must open their stomata
even during periods of drought.
Over
High
Medium
KB: We are unaware of a clear threshold for drought below
which visible foliar injury would not occur. The national-scale
foliar injury analysis did not provide any evidence of a soil
moisture threshold for injury.
INF: If there is a threshold for drought, we may overestimate
foliar injury at lower levels of soil moisture.
                                                             7-69

-------
       Source
         Description
    Potential influence of
     uncertainty on risk
         estimates
                                                      Direction
                                               Magnitude
               Knowledge-
                  Base
                   Comments (KB: knowledge base, INF: influence of
                             uncertainty on risk estimates)
G. Spatial resolution
of soil moisture data
Some vegetation such as along
riverbanks may experience
sufficient soil moisture during
periods of drought to exhibit
foliar injury. In addition, we did
not have  soil moisture data for
Alaska, Hawaii, Puerto Rico, or
Guam.
Under
Medium
Medium
KB: Soil moisture has substantial spatial variation. The data
source for soil moisture are NOAA's 344 climate divisions,
which can be hundreds of miles wide. The inability to capture
within-division variability in soil moisture adds some
uncertainty to this assessment, particularly along riverbanks.
However, we are currently unable to quantify the magnitude
of this uncertainty.
INF: Soil moisture can vary,  even within small geographic
areas. It is most likely that soil moisture is underestimated in
areas considered to be in drought conditions, so if plants in
these areas exhibited foliar injury, the soil moisture would be
underestimated, which underestimate the importance of soil
moisture's effect on foliar injury.
H Time period for
soil moisture data
Short-term estimates of soil
moisture are highly variable
over time, even from month to
month within a single year.
Using averages contributes to a
potential temporal mismatch
between soil moisture and
injury.
Unknown
Low-
Medium
Low
KB: The average of monthly values is sensitive to skew by a
single very wet or very dry month within that timeframe or
even a single precipitation episode within a month. As shown
in a sensitivity analysis, parks are not very sensitive to the
different timeframes for soil-moisture data.
INF: Without much more precise sampling, it is difficult to
assess the effect of the soil moisture sampling period, but the
overall effect of averaging appears to normalize both very
high and very low moisture conditions, which would affect
these results in opposite directions.
I. Drought categories
The soil moisture categories
used to derive the foliar injury
benchmarks (i.e., wet, normal,
and dry) are uncertain.
Unknown
Unknown
Low
KB: NOAA's categorization for Palmer Z soil moisture data
has been described as "rather arbitrary" (Karl, 1986).
INF: Using a different categorization would lead to different
benchmark criteria for O3 exposure associated with foliar
injury, but it is not clear whether this uncertainty could
underestimate or overestimate the potential foliar injury.
                                                              7-70

-------
       Source
         Description
    Potential influence of
    uncertainty on risk
         estimates
                                                     Direction
                                               Magnitude
               Knowledge-
                  Base
                   Comments (KB: knowledge base, INF: influence of
                             uncertainty on risk estimates)
J.  Spatial resolution
for combining soil
moisture, biosite, and
ozone exposure data
For the national-scale foliar
injury assessment, we combined
data from different spatial
resolutions.
Unknown
Medium
Low
KB: In general, the biosite data is at a finer spatial resolution
(usually ~ .02 km2 than the ozone data (144 km2) and the soil
moisture data (hundreds of miles across).
INF: We used data at the finest spatial resolution available to
minimize this uncertainty.
K. Maps of
vegetation and
recreational areas
within parks
Maps of vegetation and
recreational areas that overlap
with areas with higher W126
index values are uncertain.
Unknown
Low
High
KB and INF: VegBank is the vegetation plot database of the
Ecological Society of America's Panel on Vegetation
Classification, and it consists of (1) actual plot records, (2)
vegetation types, and (3) all plant taxa. (See
http://vegbank.org/vegbank/general/info.html) Even though
the data quality of the vegetation maps are high, extrapolating
across the park using plant communities is uncertain due to
unqualified variation in the defined community. The spatial
resolution of the vegetation maps is higher than the gridded
ozone exposure maps (12km2).
                                                             7-71

-------
 1     7.6    DISCUSSION
 2           National-Scale Analysis of Foliar Injury:
 3           •   Using the data on biosites and the Palmer Z drought index, across all of the biosites
 4              (5,284 over five years from 2006-2010) over 81 percent of observations showed no
 5              foliar injury.  Using the full dataset including all observations with or without injury,
 6              the analysis showed no clear relationship between Os and the biosite index and no
 7              clear relationship between Os and the Palmer Z drought index. This largely reflects
 8              the fact that 63 is not a good predictor of the  presence or absence of foliar injury, but
 9              not necessarily that there is  no  relationship between the degree of injury and 63 in
10              plants that do show injury.

11           -To better understand the relationship between O3 and those biosites that did show
12              foliar injury, we conducted  a cumulative analysis. When analyzed by individual year
13              and looking at the presence/absence of foliar injury, the proportion of sites exhibiting
14              foliar injury rises rapidly (over 20 percent in  2010) at increasing W126 index values
15              up to 10 ppm-hrs.  Similarly, when looking at an elevated biosite index of > 5, the
16              proportion of sites exhibiting foliar injury rises rapidly (over 6 percent in  2010) at
17              increasing W126 index values below approximately 10 ppm-hrs.

18           •   When categorized by moisture category, the results show a more distinct pattern.
19
20
21
22
23
24
25
wnen categorized oy moisture category, tne results snow a more distinct pattern.
Looking at both the presence/absence of foliar injury and an elevated biosite index of
> 5, there is a rapid increase in the proportion of sites exhibiting foliar injury at O3
below a W126 index value of 10 ppm-hrs.  Sites classified as wet have much higher
overall proportions at both any injury and elevated injury and a much more rapid
increase in proportion of sites with foliar injury present. At sites considered dry, the
overall proportions are much lower for presence/absence and an elevated biosite
index of > 5, potentially indicating that drought may provide protection from foliar
26              injury as discussed in the ISA.
27           •   This analysis suggests that reductions in W126 index values at or above the W126
28              benchmark of 10.46 ppm-hrs are unlikely to substantially reduce the prevalence of
                                                      7-72

-------
 1              foliar injury. Similarly, this analysis suggests that reductions in W126 index values
 2              below the base scenario benchmark are likely to relatively sharply reduce the
 3              prevalence of foliar injury.
8
 4          Screening-level Assessment of Visible Foliar Injury in National Parks:
 5          •   Based on NFS lists, 95 percent of the parks contain at least one Os-sensitive species.
 6          •   During 2006 to 2010, 58 percent of parks exceeded the benchmark W126
 7              corresponding to the base scenario (W126>10.46 ppm-hrs, 17.7 percent of biosites,
 8              without consideration of soil moisture, any injury) for at least three years. This
 9              analysis suggest that in order to substantially reduce the risk of foliar injury in these
10              parks, the W126 index values would need to be reduced to be below 10.46 ppm-hrs.

11          •   During 2006 to 2010, 98%, 80%, 68% and 2% of parks would exceed the benchmark
12              criteria corresponding to the 5%, 10%, 15%, and 20% prevalence scenarios for at
13              least 3 years.

14          •   For the elevated injury scenario, 34 percent of parks would exceed the benchmark
15              criteria (five percent of biosites, multiple moisture categories, elevated foliar injury)
16              for at least three years.

17          •   During 2006-2010, 42 percent of parks did not exceed 15 ppm-hrs.
18          •   None of the 214 parks would exceed the benchmark criteria for the base scenario
19              (W126>10.46 ppm-hrs) after adjustments to meet the existing standard at 75 ppb.
20              Only 8 parks exceed 7 ppm-hrs after adjustments to meet the existing standard at 75
21              ppb.

22          National Park Case Study Areas:
23          •   GRSM is prized, in part, for its  rich species diversity. The large mix of species
24              includes 37 (Vsensitive species and many areas contain several sensitive species.
25              With 3.8 million hikers using the trails every year and those hikers willing to pay over
26              $266 million for that activity, even a small benefit of reducing O3 damage in the park
27              could result in a significant value.
                                                     7-73

-------
 1          •  W126 index values in GRSM have been among the highest in the eastern U.S. - at
 2             times twice as high as neighboring cities such as Atlanta.  Under recent conditions, 44
 3             percent of the Park has W126 index values over 15 ppm-hrs.  After just meeting the
 4             existing standard, W126 index values are reduced such that no area is over 7
 5             ppm-hrs.

 6          •  Unlike GRSM, sensitive species cover in ROMO is driven by a few Os-sensitive
 7             species (7 species) and most notably by Quaking Aspen. This is significant in that
 8             many of the visitors to ROMO visit specifically to see this tree in its fall foliage.
 9             Given 1.5 million hikers in ROMO and their $70 million WTP for the hiking
10             experience, even a small improvement in the scenic value could be significant.

11          •  Under recent O3 conditions, all 1,067 km2 of ROMO have W126 index values over 15
12             ppm-hrs. Meeting the existing standard would bring about 59 percent of the Park into
13             the 7-15 ppm-hrs range, with the remaining 41 percent under 7 ppm-hrs. Assessing
14             an alternative standard of 15 ppm-hrs would bring the entire park under 7 ppm-
15             hrs.

16          •  SEKI is home to 12 identified sensitive species. Again, although we are not able to
17             quantify the impact of this scenic damage on hiker satisfaction for hikers in SEKI and
18             their $26 million WTP for the experience, even a small improvement in the scenic
19             value could be significant.

20          •  As in the previous national park case studies, moving from recent conditions to
21             meeting the existing O3 standard of 75 ppb results in a large change in the area of
22             SEKI with exposures above 15 ppm-hrs. For SEKI this means the parks move
23             from all areas experiencing exposures above 15 ppm-hrs to the SEKI having
24             exposures below 7 ppm-hrs.
                                                    7-74

-------
 1                                8   SYNTHESIS OF RESULTS
 2    8.1    Introduction
 3           The goals for this welfare risk and exposure assessment include characterizing ambient
 4    ozone (63) exposure and its relationship to ecological effects and estimating the resulting
 5    impacts to several ecosystem services. In particular,  we characterize ambient 63 exposures on
 6    two important ecological effects - biomass loss and foliar injury - and estimate impacts to the
 7    following ecosystem services: supporting, regulating, provisioning, and cultural services.  In the
 8    assessment, we conduct national- and regional-scale analyses to (1) characterize ambient Os
 9    exposure (Chapter 4); (2) quantify the effects of insect damage related to foliar injury (cultural
10    services) (Chapter 5); (3) consider the overall risk to  a subset of ecosystem services by
11    combining the relative biomass loss (RBL) rates for multiple tree species into one metric and
12    evaluating weighted RBL rates (Chapter 6); (4) estimate the market effects of biomass loss on
13    timber production and agricultural harvesting (provisioning services) and quantify the associated
14    economic effects (Chapter 6); (5) estimate the effect of biomass loss on carbon sequestration
15    (regulating service) (Chapter 6);  (6) estimate the effect of foliar injury and its impact on national
16    recreation (cultural  services) (Chapter 7); (7) derive potential W1261 benchmarks associated with
17    different combinations of the prevalence of biosites showing injury, the degree of foliar injury,
18    and different soil moisture considerations; and (8) apply these benchmark criteria to a  screening-
19    level assessment of foliar injury in 214 national parks (cultural services) (Chapter 7). In
20    addition, we conduct case study-scale analyses to  (1) characterize the effect of foliar injury on
21    forest susceptibility and fire regulation in California (regulating services) (Chapter 5); (2)
22    quantify the effects  of biomass loss on carbon sequestration and pollution removal  (regulating
23    services) in five urban areas (Chapter 6);  (3) quantify the effects  of relative biomass loss in Class
24    I areas (Chapter 7);  and  (4) assess the impacts of foliar injury on recreation in three national
25    parks (Chapter 7). In addition, in Chapters 5, 6, and 7 we also qualitatively assess additional
26    ecosystem services, including regulating services  such as hydrologic cycle and pollination;
      1 The W126 metric is a seasonal sum of hourly O3 concentrations, designed to measure the cumulative effects of O3
      exposure on vulnerable plant and tree species.  The W126 metric uses a sigmoidal weighting function to place less
      emphasis on exposure to low concentrations and more emphasis on exposure to high concentrations.
                                                  3-1

-------
 1    provisioning services such as commercial non-timber forest products; and cultural services with
 2    aesthetic and non-use values.
 3           To evaluate risk for the existing 8-hour daily maximum standard2 and alternative W126
 4    standards in this welfare risk and exposure assessment, we (1) quantified ecological effects based
 5    on relationships between ecological effect and the W126 metric, (2) quantified the impact of
 6    these ecological effects on ecosystem services, and (3) qualitatively assessed potential impacts to
 7    several additional ecosystem services.  The results from these assessments will help inform
 8    consideration of the adequacy of the existing Os standards and potential risk reductions
 9    associated with several alternative levels of the standard, using the W126 form. In addition, the
10    assessment (1) includes information (e.g., foliar injury analyses) that could be relevant to a three-
11    year average of a W126 standard, (2) addresses how just meeting alternative W126 standard
12    levels would affect exposures and welfare risks and associated ecosystem services, and (3)
13    addresses uncertainties and limitations in the available data.
14           To facilitate interpretation of these results, this chapter provides a synthesis of the various
15    results, focusing on comparing and contrasting results to identify common patterns or important
16    differences. These comparisons focus on patterns across different geographic areas of the U.S.,
17    across years of analysis, and across  alternative W126 standard levels. We evaluate the degree to
18    which the integrated results are representative of overall patterns of exposure and risk across
19    different types of ecosystems.  We also summarize overall confidence in the results, as well as
20    relative confidence between the different analyses.  The chapter  concludes with an overall
21    integrated characterization of risk in the context of key policy relevant questions. The remainder
22    of this chapter summarizes the results (Section 8.2)  and includes discussions on patterns of risk
23    (Section 8.3), representativeness (Section 8.4), confidence in the results (Section 8.5), and
24    integrated risk characterization (Section 8.6).
25    8.2    Summary  of Analyses and  Key Results
26           We conducted a variety of analyses to assess 63 welfare risk and exposure and to
27    estimate the relative change in risk and exposure resulting from air quality adjustments to just
28    meeting existing and alternative standards.  These analyses included national- and case study-
      2 The existing secondary standard for O3 is identical to the existing primary health-based standard, which is set at 75
      ppb for the 4th highest 8-hour daily maximum averaged over three years.
                                                  8-2

-------
 1    scale analyses addressing air quality, biomass loss, foliar injury, insect damage, fire risk, and
 2    recreation. The remainder of this section briefly summarizes the national- and case study-scale
 3    analyses and key results.
 4            8.2.1 National-Scale Analyses
 5               8.2.1.1    Air Quality Analyses
 6           The analyses used ambient air quality data from 2006 through 2008, as well as data
 7    adjusted to meet the current and potential alternative secondary standard levels.3  An HDDM
 8    adjustment methodology, similar to the one used in the Health Risk and Exposure Assessment
 9    (see Chapter 4, Section 4.3.4.1 for a discussion of the methodology), independently adjusted air
10    quality for nine climate regions as defined by the National Oceanic and Atmospheric
11    Administration (NOAA) and shown in Figure 8-1 below (reproduced from Chapter 4).4 We
12    considered these regions an appropriate delineation for our analyses because geographic patterns
13    of both 63 and plant species are often largely driven by climatic features such as temperature and
14    precipitation patterns.  The NOAA climate regions were used for all of the adjustments between
15    observed air  quality concentrations and air quality adjusted to just meet the existing and
16    alternative W126 standards.
17           In the air quality analyses in Chapter 4, we consider the changes across the distribution of
18    W126 index  values after adjusting air quality to just meet the existing standard and just meet
19    alternative W126 standard levels,  all 3-year averages. As indicated above, each climate region
20    was adjusted independently such that the entire region was adjusted based on the magnitude of
21    across-the-board reductions in U.S. anthropogenic NOx emissions required to bring the highest
22    monitor down to the targeted level. For the biomass loss analyses, we generated a national-scale
23    air quality surface that just meets the existing standard  using the Voronoi Neighbor Averaging
24    (VNA) interpolation technique to  fill in values between monitor locations. VNA national
      3 W126 calculations are slightly modified in the case of the model adjustment scenarios described in Chapter 4,
      Section 4.3.4. When calculating W126 for the model adjustment cases, we first found the three-year average of each
      three-month period, and then selected the three-month period with the highest three-year average using the same
      three-month period for each of the three years. In this way, the five scenarios are for recent air quality, air quality
      adjusted to just meet the current standard, and air quality further adjusted to just meet three different W126 index
      values: 15 ppm-hrs, 11 ppm-hrs, and 7 ppm-hrs.
      4 Many of the models and analytical tools used in the analyses include different definitions of geographic areas.  To
      the extent possible, we will refer to geographic areas by the nine climate regions based on National Oceanic and
      Atmospheric Administration (NOAA) National Climate Data Center (NCDC) regions in this chapter and note where
      definitions differ.
                                                   8-3

-------
 1    surfaces were also created for monitors adjusted to meet the current standard and for monitors
 2    adjusted to meet alternative W126 standard levels of 15, 11, and 7 ppm-hrs. During the last 63
 3    National Ambient Air Quality Standards review, the Clean Air Scientific Advisory Committee
 4    (CASAC) recommended and supported a range of alternative W126 standard levels from 15 to 7
 5    ppm-hrs.  The adjusted surfaces, based on monitored, three-year average W126 index values
 6    from 2006 through 2008, are used as inputs to several assessments (described below), including
 7    the geographic analysis to assess the effects of insect damage related to foliar injury, the
 8    national- and case study-scale biomass loss assessments, and the national park  case studies for
 9    foliar injury. For the national-scale and screening-level foliar injury analyses, we generated five
10    national-scale air quality surfaces from the monitored annual W126 index values (unadjusted) for
11    the individual years from 2006 through 2010, also using VNA. See Chapter 4, Section 4.3  for
12    more detailed discussions of the air quality analyses.
13          The  largest reduction in W126 index values occurs when moving from  recent ambient
14    conditions to meeting the existing secondary standard of 75 ppb (8-hour daily maximum).  After
15    adjusting to just meet the current standard, only two of the nine U.S. regions have W126 index
16    values remaining above 15 ppm-hrs (West — 18.9 ppm-hrs and Southwest - 17.7 ppm-hrs). The
17    Central region would meet an alternative W126 standard level of 15 ppm-hrs, but further air
18    quality adjustment would be needed for the Central region to meet alternative standards of 11
19    and 7 ppm-hrs. In addition, when adjusting to just meeting the existing standard, four regions
20    (East North Central, Northeast, Northwest, and South) would meet 7 ppm-hrs,  and two regions
21    (Southeast and West North Central) have index values between 9 and 12 ppm-hrs (Southeast -
22    11.9 ppm-hrs and West North Central - 9.3 ppm-hrs).
                                                8-4

-------
        Legend
           Central
           East North Cental
         •  Northeast
           Northwest
         •  South
         •  Southeast
         •  Southwest
           West
         •  West North Central
 2   Figure 8-1    Map of the 9 NOAA climate regions (Karl and Koss, 1984) used in the
 3   national-scale air quality adjustments (Chapter 4, Figure 4-6)
 4
 5              8.2.1.2    Forest Susceptibility to Insect Infestation
 6           In Chapter 5, we review information on 63 exposure and the increased susceptibility of
 7   forests to insect infestations. O3 exposure results in increased susceptibility to infestation by
 8   some chewing insects, including the southern pine beetle and western bark beetle.  These
 9   infestations can cause economically significant damage to tree stands and the associated timber
10   production. In the short term, the immediate increase in timber  supply that results from the
11   additional harvesting of damaged timber depresses prices for timber and benefits consumers. In
12   the longer term, the decrease in timber available  for harvest raises timber prices, harming
13   consumers and potentially benefitting some producers.  The United States Forest Service (USFS)
14   reports timber producers have incurred losses of about $1.4 billion (2010$), and wood-using
                                                 §-5

-------
 1    firms have gained about $966 million, due to beetle outbreaks between 1977 to 2004 (Coulson
 2    and Klepzig, 2011).  It is not possible to attribute a portion of these impacts resulting from the
 3    effect of O3 on trees' susceptibility to insect attack; however, the losses are embedded in the
 4    estimates cited.
 5          In addition, in Chapter 5 we provide summaries of area at risk of high pine beetle loss
 6    (i.e., high loss due to pine beetle damage), as well as millions of square feet of basal tree area at
 7    risk of high pine beetle loss after just meeting the existing and alternative standards. For area at
 8    risk of high pine beetle loss, under recent  ambient conditions approximately  57 percent of the at-
 9    risk area is at or above a W126 index value of 15 ppm-hrs; approximately 16 percent of the at-
10    risk area is at a W126 index value between 15  and 11 ppm-hrs; approximately 23 percent of the
11    at-risk area is at a W126 index value between 11 and 7 ppm-hrs; and approximately four percent
12    of the at-risk area is at a W126 index value below 7 ppm-hrs.  After just meeting the
13    existing  standard, approximately five percent of the at-risk area has W126 index value between
14    11 and 7 ppm-hrs, and no at-risk area is above a  W126 index value of 11 ppm-hrs.  When
15    adjusting to an alternative standard level of 15 ppm-hrs, no at-risk area is above a W126 index
16    value of 7 ppm-hrs. In terms of millions of square feet of tree basal area at risk of high pine
17    beetle loss, under recent ambient conditions, approximately 45 percent of the "at-risk square
18    feet" is at or above a W126 index value of 15 ppm-hrs; approximately 13 percent of "at-
19    risk square feet" is between 15 and 11 ppm-hrs; approximately 34 percent is  between 11 and 7
20    ppm-hrs;  and approximately eight percent is at a W126 index value below 7  ppm-hrs. After just
21    meeting the existing standard, approximately ten percent of the "at-risk square feet" is at a W126
22    index value between 11 and 7 ppm-hrs, and no square feet are above 11 ppm-hrs.
23              8.2.1.3    Biomass Loss
24          We reviewed several studies that modeled vegetation growth for several tree and crop
25    species.  For trees, we calculated seedling RBL associated with W126 index  values and
26    compared the seedling RBL values to the  study results for adult trees. Overall, seedling biomass
27    loss values are much more consistent with adult biomass loss at lower W126 index values. For
28    example,  for Tulip Poplar, at a W126 index value of 15 ppm-hrs, the adult biomass loss rate is
29    estimated to be 10.5 percent, and the seedling biomass loss rate is estimated to be 7.7 percent; at
30    a W126 index value of 59 ppm-hrs, the adult biomass loss rate is estimated to be 16.8 percent,

-------
 1    and the seedling biomass loss rate is estimated to be 74 percent. See Chapter 6, Section 6.2.1.1
 2    for additional information.
 3          For biomass loss, CAS AC recommended that EPA should consider options for W126
 4    standard levels based on factors including a predicted one to two percent biomass loss for trees
 5    and a predicted five percent loss of crop yield. Small losses for trees on a yearly basis compound
 6    over time and can result in substantial biomass losses over the decades-long lifespan of a tree
 7    (Frey and Samet, 2012b). To assess overall ecosystem-level effects from biomass loss, we
 8    weighted the RBL values for multiple tree species using basal area5 and combined them into a
 9    weighted RBL value and considered the weighted value in relation to the proportion of basal area
10    accounted for by the tree species. A weighted RBL value is a relatively straight-forward metric
11    to attempt to understand the potential ecological effect on some ecosystem services.  We
12    separated results into categories of different percentages of total basal area (e.g., <10 percent, 10
13    to 25 percent)and compared weighted RBL values against the one and two percent biomass loss
14    for trees recommended by CAS AC. In each category, the results indicate that of the area being
15    assessed the portion exceeding  benchmarks of one to two percent biomass loss decreases as
16    W126 index values decrease. For example, after just meeting the existing standard, 20.8 percent
17    and 12.4 percent of the total area being assessed exceeds benchmarks of one percent and two
18    percent biomass loss in trees, respectively.  After just meeting an alternative standard level of 7
19    ppm-hrs, 11.5 percent and 7.7 percent of the total area being assessed still exceeds benchmarks
20    of one and two percent biomass loss in trees, respectively.  It is important to note that the
21    proportional basal area values do not account for total cover, but rather the relative cover of the
22    tree species present. See Chapter 6, Section 6.8 for additional information.  We also analyzed
23    federally designated Class I areas by calculating an average weighted RBL value for 119 of the
24    156 Class I areas. The number of Class I areas that exceed one and two percent relative biomass
25    loss decreases as the alternative W126 standard levels become more stringent. See Chapter 6,
26    Section 6.8.1 for additional information.
27          Using the concentration-response (C-R) functions for tree seedlings and crops, we
28    determined the range of biomass  loss associated with just meeting the existing 8-hour daily
      5 Basal area is the term used in forest management that defines the area of a given section of land that is occupied by
      a cross-section of tree trunks and stems at their base. This typically includes a measurement taken at the diameter at
      breast height of a tree above the ground and includes the complete diameter of every tree, including the bark.

                                                 8-7

-------
 1    maximum standard and alternative W126 standard levels. To compare different levels of
 2    biomass loss to different W126 index values, we plotted the C-R functions as a function of the
 3    percent biomass loss against varying W126 index values.  For a one percent biomass loss for
 4    trees, the estimated W126 index values were between 4 and 10 ppm-hrs; for a two percent
 5    biomass loss for trees the estimated W126 index values were between 7 and 14 ppm-hrs; and for
 6    a five percent biomass loss for crops the estimated W126 index values were between 12 and 17
 7    ppm-hrs. See Chapter 6, Section 6.2.1.2 for additional information.
 8           Using the Forest and Agricultural Optimization Model with Greenhouse Gases
 9    (FASOMGHG), we conducted national-scale analyses to quantify the effects  of biomass loss on
10    timber production and agricultural harvesting, as well as on carbon sequestration.6 We used the
11    Os C-R functions for tree seedlings and crops to calculate relative yield loss (RYL), which is
12    equivalent to relative biomass loss. Because the forestry and agriculture sectors are related,  and
13    trade-offs occur between the sectors, we simultaneously calculated the resulting market-based
14    welfare effects of 63 exposure in the forestry and agriculture sectors.
15           In the analyses for commercial timber
16    production, because most areas have W126
17    index values lower than 15 ppm-hrs when
18    simulating meeting the existing standard,
19    relative yield losses (RYL) are below one
20    percent, with the exception of the Southwest,
21    Southeast, Central, and South regions (see text
22    box below for clarification on region names).
23    Relative yield losses remain above one percent
24    for the parts of the Southeast, Central, and
25    South regions at alternative W126 standard
26    levels of 15  and 11 ppm-hrs, and for the
27    Southeast and South regions at an alternative W126 standard level of 7 ppm-hrs.
The states included in the NOAA NCDC regions
and the states included in the FASOMGHG model
regions differ slightly. Below we align the
different region names. To be consistent across
summary discussions, we use the NCDC region
names.
NCDC
West
Southwest
Central
South
Southeast
Northeast
FASOMGHG
primarily Pacific Southwest
primarily Rocky Mountain
primarily Cornbelt
primarily South West and South
Central
primarily South Central and
Southeast
primarily Northeast
      6 FASOMGHG is a national-scale model that provides a complete representation of the U.S. forest and agricultural
      sectors' impacts of meeting alternative standards. FASOMGHG simulates the allocation of land over time to
      competing activities, e.g., production of different crops or livestock, in both the forest and agricultural sectors.

-------
 1           In the analyses for agricultural harvest, the largest yield changes occur when comparing
 2    recent ambient conditions to just meeting the existing standard.  Under recent ambient
 3    conditions, the West, Southwest, and Northeast regions generally have the highest yield losses.
 4    At alternative W126 standard levels of 15, 11, and 7 ppm-hrs, for winter wheat7 relative yield
 5    losses are less than the 5 percent loss recommended by CASAS, as well as less than one percent.
 6    For soybeans, when the W126 scenarios are modeled, yield losses above both 5 and 1 percent
 7    remain at 15 ppm-hrs for the Southwest and Central regions. Yield losses are reduced to below
 8    one percent at alternative W126 standard levels of 11 and 7 ppm-hrs.
 9           In addition to estimating changes in forestry and agricultural yields, FASOMGHG
10    estimates the changes in consumer and producer/farmer surplus associated with the change in
11    yields.8 Changes in yield affect individual tree species and crops, but the overall effect on forest
12    ecosystem productivity depends on the composition of forest stands and the relative sensitivity of
13    trees within those stands. Overall effect on agricultural yields and producer and consumer
14    surplus depends on the (1) ability of producers/farmers to substitute other crops that are less 63
15    sensitive and (2) responsiveness, or elasticity, of demand and supply. Relative to just meeting
16    the existing standard, W126 index values decrease in the Southwest, West, Central, Southeast,
17    South,  East North Central, and West North Central regions at alternative standard levels of 15,
18    11, and 7 ppm-hrs.  These decreases in W126 index values are estimated to result in changes in
19    patterns for agricultural production and resulting consumer and producer surplus.  For example,
20    with reductions W126 index values, wheat crops would likely increase in one of its major
21    production regions, the Southwest region. This expansion of wheat production may result in a
22    decrease in wheat production in the East North Central region. The East North Central region
23    would likely see production changes for other crops because the contraction in wheat production
24    makes  room for alternatives. Soybean production in the East North Central region would likely
25    expand, and this expansion would induce regional shifts of soybean production at the national
26    level, including decreases in soybean production in the West North Central and Central regions.
27    Generally the crop producers' surplus in the Central and Southwest regions would increase and
28    in the South region would decrease. Crop producers' surplus in the West North Central and East
29    North Central regions would fluctuate over time.
      7 Among the major crops, because winter wheat and soybeans are more sensitive to ambient O3 levels than other
      crops we include these crops for this discussion.
      8 See Chapter 6, Section 6.3 for a brief discussion of economic welfare and consumer and producer surplus.
                                                 8-9

-------
 1           Economic welfare impacts resulting from just meeting the existing and alternative
 2    standards were largely similar between the forestry and agricultural sectors — consumer surplus,
 3    or consumer gains, generally increased in both sectors because higher productivity under lower
 4    W126 index values increased total yields and reduced market prices. Because demand for most
 5    forestry and agricultural commodities is not highly responsive to changes in price, there were
 6    more cases where producer surplus, or producer gains,  decline. In some cases, lower prices
 7    reduce producer gains more than can be offset by higher yields.  For example, in 2040, the year
 8    with maximum changes in consumer and producer surplus, in the forestry sector at just meeting
 9    the existing standard, total producer surplus is estimated to be $133 billion and total consumer
10    surplus is estimated to be  $935 billion, or 7 times greater than producer surplus.  For the forestry
11    sector, when adjusting to meeting alternative W126 standard levels of 15, 11, and 7 ppm-hrs,
12    consumer surplus increases $597 million, $712 million, and $779 million (i.e., 0.06, 0.08, and
13    0.08 percent),  respectively, while producer surplus decreases $839 million,  $858 million, and
14    $766 million, (i.e., about 0.6 percent), respectively. All estimates are in 2010$.9
15           In the analysis for changes in carbon sequestration related to biomass loss, relative to just
16    meeting the existing standard, the 15 ppm-hrs  W126 alternative standard does not appreciably
17    increase carbon sequestration. The majority of the enhanced carbon sequestration potential is in
18    the forest biomass increases over time under alternative secondary W126 standard levels at 11
19    and 7 ppm-hrs. In the forestry sector, relative  to just meeting the existing standard (with
20    sequestration of 89 billion metric tons of CC>2  equivalents), at alternative W126 standard levels
21    of 11 and 7 ppm-hrs carbon sequestration potential is projected to increase 593 million and 1.6
22    billion metric tons of CC>2 equivalents over 30 years (i.e., 0.66 and 1.79 percent) respectively.
23    For the agricultural  sector, relative to just meeting the existing standard (with sequestration of 8
24    billion metric tons of CO2 equivalents), at alternative W126 standard levels of 11 and 7 ppm-hrs
25    carbon sequestration potential is projected to increase 9 and 10 million metric tons of CC>2
26    equivalents respectively over 30 years, or about 0.1 percent.
      9 FASOMGHG is an international model and the increase in productivity caused by a reduction in O3 results in a net
      increase in the present value of total global economic surplus (consumer + producer surplus). The reported producer
      surplus here is for U.S. producers only and benefits and costs accruing overseas are not included. Also, for any
      given year, there may be a decline in global consumer and producer surplus due to the effects on the dynamics of
      planting and harvesting decisions in the forestry sector.
                                                 8-10

-------
 1               8214    Visible Foliar Injury
 2           To assess the effects of visible foliar injury on recreation, we reviewed the National
 3    Survey on Recreation and the Environment (NSRE), as well as the 2006 National Survey of
 4    Fishing, Hunting, and Wildlife-Associated Recreation (FHWAR) and a 2006 analysis done for
 5    the Outdoor Industry Foundation (OIF).  According to the NSRE, some of the most popular
 6    outdoor activities are walking, including day hiking and backpacking; camping; bird watching;
 7    wildlife watching; and nature viewing. Participant satisfaction with these activities can depend
 8    on the quality of the natural scenery, which can be adversely affected by Os-related visible foliar
 9    injury. According to the FHWAR and the OIF reports, the total  expenditures across wildlife
10    watching activities, trail-based activities, and camp-based activities are approximately $200
11    billion dollars annually.  While we cannot quantify the magnitude of the impacts of Os damage
12    to the scenic beauty and outdoor recreation, the existing losses associated with current Os-related
13    foliar injury are reflected in reduced outdoor recreation expenditures.
14           To assess foliar injury at a national scale and identify potential W126 benchmarks, we
15    conducted several  analyses using a national data set on foliar injury from the USFS's Forest
16    Health Monitoring Network. We conducted the analyses using presence/absence of foliar injury,
17    as well as using a cutoff for elevated foliar injury.10 We also conducted analyses across years
18    and different soil moisture categories in NOAA climate divisions.11 Across years, when
19    assessing the presence or absence of foliar injury, at an alternative W126 standard level of 15
20    ppm-hrs between 12 and over 18 percent of sites indicated the presence of foliar injury; at an
21    alternative W126 standard level of 11 ppm-hrs between 12 and over 20 percent of sites indicated
22    the presence of foliar injury; and at an alternative W126 standard level of 7 ppm-hrs between 4
23    and over 20 percent of sites indicated the presence of foliar injury.12 Across years, when
24    assessing elevated foliar  injury, at an alternative W126 standard  level of 15 ppm-hrs between 3
25    and over 6 percent of sites show elevated foliar injury; at an alternative W126 standard level of
26    11 ppm-hrs between 2 and over 6 percent of sites show elevated foliar injury; and at an
27    alternative W126 standard level of 7 ppm-hrs between approximately 2 and over 6 percent of
      10 The elevated foliar injury corresponds to abiosite index of 5, which is consistent with a USFS cut-off for foliar
      injury.
      11 See Chapter 7, Section 7.2 for a more detailed discussion of the data onbiosites and foliar injury from the USFS
      and the Palmer Z drought index data from NOAA.
      12 See Chapter 7, Section 7.2.3 for additional discussion and Figure 7-8 for additional information. The proportion
      of sites with foliar injury present varies by year, creating these ranges for percent of sites with foliar injury present.
                                                  8-11

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
sites show elevated foliar injury. Generally, the results of all of these foliar injury analyses
demonstrate a similar pattern - the proportion of biosites showing foliar injury increases steeply
with W126 index values up to approximately 10 ppm-hrs and is relatively constant at W126 index
levels abovelO ppm-hrs.  This analysis suggests that reductions in W126 index values at or
above this benchmark (W126 > 10.46 ppm-hrs) are unlikely to substantially reduce the
prevalence of foliar injury. Similarly, this analysis suggests that reductions in W126 index
values below the base scenario benchmark are likely to relatively sharply reduce the prevalence
of foliar injury. Figure 8-2, which originally appears as Figure 7-10 in Chapter 7, shows the
pattern seen in the foliar injury analyses stratified by soil moisture category.  In addition, we see
similar patterns when the foliar injury is stratified by year and geographic region. See Section
7.2.3 for a more detailed discussion of the analyses.
                          5
                          H
                          1
                                                                Oiy
                                         10
Figure 8-2    Cumulative Proportion of Sites with Visible Foliar Injury Present, by
              Moisture Category
       We used the results of the national analysis to derive benchmarks for visible foliar injury
that we apply in a screening-level assessment and case studies of national parks.
We define six scenarios for evaluating potential W126 benchmarks, representing the full range of
the percentages of biosites showing visible foliar injury (i.e., any injury and elevated injury),
                                                 8-12

-------
 1    including five scenarios considering soil moisture. We defined the W126 benchmark for the
 2    "base scenario" as representing the point above which there was a consistent percentage (17.7
 3    percent) of biosites showing foliar injury, regardless of soil moisture. This analysis suggests that
 4    reductions in W126 index values at or above this benchmark (W126 > 10.46 ppm-hrs) are
 5    unlikely to substantially reduce the prevalence of foliar injury. Similarly, this analysis suggests
 6    that reductions in W126 index values below the base scenario benchmark are likely to relatively
 7    sharply reduce the prevalence of foliar injury.  We also looked at alternative scenarios based on 3
 8    different categories of soil moisture and the W126 index values associated with four different
 9    prevalences (e.g., 5%, 10%, 15% and 20% of biosites) of any foliar injury, and a final one based
10    on a 5% prevalence of foliar injury index greater than or equal to 5.  In total, the WREA
11    evaluated 13 different W126 benchmarks associated with the 6 foliar injury risk scenarios. The
12    W126 benchmarks across the six scenarios range from 3.05 ppm-hrs (five percent of biosites,
13    normal  moisture, any injury) up to 46.87 ppm-hrs (five percent of biosites, dry,  elevated injury).
14    See Table 7-5 for the specific benchmark criteria corresponding to each of the six scenarios.
15           The general approach in the screening-level assessment of national parks is derived from
16    Kohut (2007), but we apply more recent Os exposure and soil moisture data for  214 national
17    parks in the contiguous U.S. combined with the benchmarks derived from the national analysis.
18    Generally, benchmark scenarios corresponding to higher percentages of biosites showing foliar
19    injury show fewer parks that exceed the benchmark criteria for those scenarios.  During 2006 to
20    2010, 58 percent of parks exceeded the benchmark W126 corresponding to the base scenario
21    (W126>10.46 ppm-hrs,  17.7 percent of biosites, without consideration of soil moisture, any
22    injury)  for at least three years.  This analyses suggest that in order to substantially reduce the risk
23    of foliar injury in these parks, the W126 index values would need to be reduced to be below
24    10.46 ppm-hrs. In addition, 98%, 80%, 68% and 2% of parks would exceed the benchmark
25    criteria  corresponding to the 5%, 10%,  15%, and 20% prevalence scenarios for at least 3 years
26    within the 2006-2010 period. For the elevated injury scenario, 34 percent of parks would exceed
27    the benchmark criteria (five percent of biosites, multiple moisture categories, elevated foliar
28    injury)  for at least three years.  Because the screening-level assessment relies on annual estimates
29    of W126 index values and soil moisture, we cannot fully evaluate just meeting the existing and
30    alternative standards because they are based on the 3-year average air quality surfaces.
31    However, we can observe that after adjusting the W126 surfaces to just meet the existing
                                               8-13

-------
 1    standard, all of the 214 parks are below 10.46 ppm-hrs, which corresponds to the benchmark
 2    criteria for the base scenario.
 3           8.2.2  Case Study-Scale Analyses
 4              8221    Fire Regulation
 5           As indicated in Chapter 5, fire regime regulation is also negatively affected by 63
 6    exposure. For example, Grulke et al. (2009) reported various lines of evidence indicating that 63
 7    exposure may contribute to southern California forest susceptibility to wildfires by increasing
 8    leaf turnover rates and litter, increasing fuel loads on the forest floor.  According to the National
 9    Interagency Fire Center, in the U.S. in 2010 over 3 million acres burned in wildland fires and an
10    additional 2 million acres were burned in prescribed fires.  From 2004 to 2008, Southern
11    California alone experienced, on average, over 4,000 fires per year burning, on average,  over
12    400,000 acres per year. The  California Department of Forestry and Fire Protection (CAL FIRE)
13    estimated that losses to homes due to wildfire were over $250 million in 2007 (CAL FIRE,
14    2008).  In 2008, CAL FIRE's costs for fire suppression activities were nearly $300 million (CAL
15    FIRE, 2008).
16            We developed maps that overlay the mixed conifer forest area of California with areas
17    of moderate or high fire risk defined by CAL FIRE and with surfaces of recent conditions and
18    surfaces adjusted to just meet existing and alternative standards. The highest fire risk and
19    highest W126 index values overlap with each other, as well as with significant portions of mixed
20    conifer forest. Under recent conditions, over 97 percent of mixed  conifer forest area has O3
21    W126 index values over 7 ppm-hrs with a moderate to severe fire  risk, and 74 percent has Os
22    W126 index values over 15 ppm-hrs with a moderate to severe fire risk. When adjusted to just
23    meeting the existing standard,  almost all of the mixed conifer forest area with a moderate to high
24    fire risk shows a reduction in Os to below a W126 index value of 7 ppm-hrs. At the alternative
25    W126 standard level of 15 ppm-hrs, all but 0.18 percent of the area is less than 7 ppm-hrs, and at
26    alternative standard levels of 11 and 7 ppm-hrs all of the moderate to high fire threat area is less
27    than 7 ppm-hrs.
28              8222    Biomass Loss
29           Using the iTree model to estimate tree growth and ecosystem services provided by trees
30    over a 25-year period, we conducted case-study scale analyses to quantify the effects of biomass
                                               8-14

-------
 1    loss on carbon sequestration and pollution removal in five urban areas.13  See Appendix 6D for
 2    details on the iTree model and the methodology used for the case study analyses.
 3          We estimated the effects of O3-related biomass loss on carbon sequestration and ran six
 4    scenarios, including just meeting the existing standard and just meeting alternative W126
 5    standards of 15, 11, and 7 ppm-hrs.  While both urban and non-urban forests have the potential to
 6    remove pollutants from the atmosphere, using iTree we also estimated the effects of O3-related
 7    biomass loss on the potential to remove carbon monoxide, nitrogen dioxide, O3, and sulfur
 8    dioxide pollution in the five urban areas (1) at recent ambient O3 conditions and (2) after
 9    adjusting air quality to just meeting the existing standards and alternative W126 standard levels
10    of 15, 11, and 7 ppm-hrs.  As a supplement to the iTree analysis, we  also performed a simple
11    analysis of the O3 pollution removal potential to show how this process might affect ambient air
12    quality values. This analysis made some general assumptions to estimate order of magnitude
13    effects of O3 removal by trees in the five urban areas. The results indicate that the effects on O3
14    concentrations are small; when meeting the current standard, deposition to tree surfaces results in
15    ambient O3 concentration reductions ranging from 0.08 parts per billion by volume (ppbv) in
16    Tennessee to 0.52 ppbv in Chicago compared to O3 concentrations that would occur without any
17    deposition to trees in these cities.14  Relative changes in ambient O3 concentrations due to
18    changes in deposition to tree surfaces were much smaller.
19          Relative to just meeting the existing standard, three of the urban areas (Atlanta, Chicago,
20    and the urban areas of Tennessee) show gains in carbon sequestration at alternative W126
21    standard levels of 11 and 7 ppm-hrs. For example, relative to just meeting the existing standard,
22    Chicago gains about 6,400 tons of carbon sequestration per year at 7 ppm-hrs, and the urban
23    areas of Tennessee gain about 8,800 tons  of carbon sequestration per year at 11 ppm-hrs and
24    20,000 tons of carbon sequestration per year at 7 ppm-hrs.  Syracuse and Baltimore do not
25    realize gains in carbon sequestration because recent air quality almost meets the alternative
26    standards levels in those areas.  Similar to changes in carbon sequestration, Syracuse and
27    Baltimore have no change in pollution removal when just meeting the existing standard and the
28    W126 alternative standards. Atlanta, Chicago, and the urban areas of Tennessee show gains in
      13 The iTree model is a peer-reviewed suite of software tools provided by USFS.
      14 The ratio of O3 volume to urban area air volume multiplied by 10A9 gives the concentration in ppbv.
                                                8-15

-------
 1    potential pollution removal at alternative W126 standard levels of 11 and 7 ppm-hrs compared to
 2    meeting the existing standard.  For example, relative to just meeting the existing standard,
 3    Chicago gains about 2,300 metric tons of pollution removal annually at 11 ppm-hrs and 6,500
 4    metric tons of pollution removal annually at 7 ppm-hrs, and the urban areas of Tennessee gain
 5    about 5,300 metric tons of pollution removal annually at 11 ppm-hrs and 11,700 metric tons of
 6    pollution removal annually at 7 ppm-hrs.
 7              8223     Foliar Injury - Three National Parks
 8          In addition to the national-scale analysis, we also assess foliar injury at a case-study scale
 9    because national parks are designated as  special areas in need of protection. Specifically, we
10    assess (Vexposure risk at three national  parks - Great Smoky Mountains National Park
11    (GRSM), Rocky Mountain National Park (ROMO), and Sequoia/Kings Canyon National Parks
12    (SEKI). For each park, we assess the potential impact of Os-related foliar injury on recreation
13    (cultural services) by considering information on visitation patterns, recreational activities and
14    visitor expenditures. We include percent cover of species sensitive to foliar injury and focus on
15    the overlap between recreation areas within the park and elevated W126 index values.
16          In GRSM, there are 37 sensitive species across vegetative strata, and 2011 visitor
17    spending exceeded  $800 million. W126  index values in GRSM have been among the highest in
18    the eastern U.S. — under recent ambient conditions, 44 percent of GRSM has W126 index values
19    over 15 ppm-hrs. After adjustments to just meet the existing standard of 75 ppb, no area in
20    GRSM exceeds an alternative W126 standard level of 7 ppm-hrs. ROMO has seven sensitive
21    species, including Quaking Aspen.  In 2011 visitor spending at ROMO was over $170 million.
22    Under recent ambient conditions, all of ROMO has W126 index values over 15 ppm-hrs. When
23    adjusted to just meet the existing standard, 41 percent of the park would meet an alternative
24    W126 standard level of 7 ppm-hrs and 59 percent of the park would meet an alternative W126
25    standard level between 7 and 11 ppm-hrs. In SEKI there are  12 sensitive species across
26    vegetative strata, and 2011 visitor spending was over $97 million.  When adjusted to just meet
27    the existing standard, no area in SEKI has W126 index values above 7 ppm-hrs.
28       8.3 Patterns of Risk
29          Considering the national- and case study-scale analyses and appropriate benchmarks for
30    biomass loss and foliar injury, we reviewed whether there were patterns or trends in the risk and
                                               8-16

-------
 1    risk reductions - between geographic areas and across years and alternative standards. For
 2    biomass loss, CASAC recommended that EPA should consider options for W126 standard levels
 3    based on factors including a predicted one to two percent biomass loss for trees and a predicted
 4    five percent loss of crop yield.  Small losses for trees on a yearly basis compound over time and
 5    can result in  substantial biomass losses over the decades-long lifespan of a tree (Frey and Samet,
 6    2012b).  For trees, annual W126 index values for a one percent biomass loss range from
 7    approximately  4 to 10 ppm-hrs and for a two percent biomass loss range from approximately 7 to
 8    14 ppm-hrs.  For crops, annual W126 index values for a five percent biomass loss range from
 9    approximately  12 to 17 ppm-hrs. Based on this assessment, the pattern is that crops exceed
10    CASAC's benchmarks at higher W126 index values than trees, and suggests that meeting
11    alternative standards that are protective of trees will  also protect crops. Unlike biomass, CASAC
12    did not recommend a benchmark for foliar injury.  As a result, we developed a set of W126
13    benchmark criteria ("scenarios") associated with different combinations of the prevalence of
14    biosites showing injury, the degree of foliar injury, and different soil moisture considerations.
15           8.3.1  Risk Patterns Across or Between Geographic Areas
16          The geographic or spatial patterns of changes in W126 index values and changes in
17    ecosystem services and related economic welfare are slightly different.  Figure 8-3 and Figure
18    8-4, which originally appear as Figures 4-9 and 4-11 in Chapter 4, show the W126 index values
19    after being adjusted to just meeting alternative standards of 15 and 11 ppm-hrs. After adjusting
20    to just meeting an alternative standard of 15
21    ppm-hrs, the West, Southwest, and Central
22    regions show the highest W126 index values
23    between 11 and 15 ppm-hrs; after adjusting to
24    just meeting  an alternative standard level of 11
25    ppm-hrs, all  areas show W126 index values
26    below 11 ppm-hrs. The analyses of biomass  loss
27    and affected  timber and agricultural yields show
28    that most of the remaining risk after adjusting to
29    just meeting  an alternative standard level of 15
30    ppm-hrs is in the Southwest, South, Southeast,
General references to the eastern and western
U.S. and the states included in the NOAA
NCDC regions differ. For ease of discussion,
below we align the general U.S. region and
NCDC region references.
General U.S.
Western U.S.
Eastern U.S.
NCDC
Northwest
West
Southwest
West North Central
East North Central
Central
South
Southeast
Northeast
                                                8-17

-------
 1   and Central regions; after adjusting to just meeting an alternative standard level of 11 ppm-hrs,
 2   most of the remaining risk is in the South, Southeast, and Central regions.
 3          There is substantial heterogeneity in plant responses to Os, both within species, between
 4   species, and across regions of the U.S.  The (Vsensitive tree species are different in the eastern
 5   and western U.S. — the eastern U.S. has far more total species (see text box for clarification on
 6   region names). 63 exposure and risk are somewhat easier to assess in the eastern U.S. because of
 7   the availability of more data and the greater number of species to analyze.  In addition, there are
 8   more Os monitors in the eastern U.S. but fewer national parks. In the national-scale analyses for
 9   commercial timber production, because most areas have W126 index values below 15 ppm-hrs
10   after simulating just meeting the existing  standard, relative yield losses (RYL) are below one
11   percent, with the exception of the Southwest, Southeast, Central, and South regions. In part
12   because the South and Southeast regions have more forest land, RYL remain above one percent
13   for parts of those regions even after just meeting an alternative W126 standard level of 7 ppm-
14   hrs.
                                                8-18

-------
                                    15 ppm-hr Scenario
2   Figure 8-3   National Surface of 2006-2008 Average W126 Index values Adjusted to
3               Just Meet the Alternative Standard of 15 ppm-hrs
                                          8-19

-------
                                       11 ppm-hr Scenario
 2   Figure 8-4    National Surface of 2006-2008 Average W126 Index values Adjusted to Just
 3                 Meet the Alternative Standard of 11 ppm-hrs
 4
 5          The largest improvements in agricultural harvesting resulting from reduced Os exposure
 6   are likely to occur in the West, Southwest, South, Southeast, and Central regions because those
 7   regions (1) have the most sensitive crop species present, (2) have significant agricultural
 8   production, and (3) will experience the most significant air quality improvement between recent
 9   conditions and just meeting the existing secondary standard. For soybeans, when the W126
10   scenarios are modeled, yield losses above both five and one percent remain at 15 ppm-hrs for the
11   Southwest and Central regions. For all regions, yield losses are reduced to below five and one
12   percent at alternative W126 standard levels of 11 and 7 ppm-hrs.
                                               8-20

-------
 1          In the analyses using presence/absence of foliar injury and cutoff for elevated foliar
 2    injury, we analyzed the data sets across NOAA climate divisions. We did not have foliar injury
 3    data for the Southwest, and we had limited data for the West and West North Central.
 4    Similar to the analyses across years and moisture categories, across NOAA climate divisions the
 5    proportion of biosites showing foliar injury increases steeply with W126 index values up to
 6    approximately  10 ppm-hrs and is relatively constant at W126 index levels above 10 ppm-hrs.
 7           8.3.2         Risk Patterns Across Years
 8          Using the FASOMGHG model to calculate forestry and agricultural yield changes, we
 9    estimated changes in consumer and producer surplus from 2010 through 2040 for alternative
10    standard levels of 15, 11, and 7 ppm-hrs.  Over the period in the forestry sector, changes in
11    consumer surplus are always positive and range from <0.01 percent in 2010 for alternative
12    standard levels of 15 and 11 ppm-hrs up to 0.08 percent in 2040 for alternative standard levels of
13    11 and 7 ppm-hrs (relative to consumer surplus at just meeting the existing standard of $721
14    billion in 2010 and $934 billion in 2040 (2010$)).  Consumer surplus does not consistently
15    increase between 5-year periods  from 2010 to 2040.15  For example, while always a positive
16    value, consumer surplus decreases between 2025 and 2030, increases slightly between 2030 and
17    2035, and increases  significantly between 2035 and 2040. Changes in producer surplus are
18    generally negative and range from <-0.1 percent in 2010 for an alternative standard level of 7
19    ppm-hrs to -0.6 percent in 2040 for alternative standard levels of 15 and 11 ppm-hrs (relative to
20    producer surplus at just meeting  the existing standard of between $93 billion in 2010 and $133
21    billion in 2040).
22          In the agricultural sector  over the period, changes in consumer surplus are generally
23    positive and <0.01 percent (relative to consumer surplus at just meeting the existing standard of
24    between $1.9 trillion in 2010 and $2.1 trillion in 2040 (2010$)). Changes in producer surplus
25    vary and range from -0.2 percent in 2015 for alternative standard levels of 11 and 7 ppm-hrs to
26    0.25 and 0.35 percent in 2040 for alternative standard levels of 11 and 7 ppm-hrs (relative to
27    producer surplus at just meeting  the existing standard of between $725 billion in 2010 and $863
28    billion in 2040).  At just meeting the existing standard, total consumer and producer surplus
      15 FASOMGHG results include multi-period, multi-commodity results over 60 to 100 years in 5-year time intervals
      when running the combined forest-agriculture version of the model.
                                                8-21

-------
 1   values are much higher in the agricultural sector than in the forestry sector.  As a result, absolute
 2   changes in consumer and producer surplus values at alternative standard levels are much larger
 3   in the agricultural sector.  In the agricultural sector, over time and by alternative standard,
 4   changes in consumer surplus are largely positive, with approximately 15 percent of the estimates
 5   being minor negative changes.  Over time and by alternative standard, changes in producer
 6   surplus are mixed, with approximately 30 percent of the estimates being significant negative
 7   changes.  See Section 6.5 and Appendix 6B for additional discussion of these analyses.
 8          In the national-scale assessment to identify foliar injury benchmarks, we conducted
 9   analyses using a national data set on foliar injury.  Across years in the data set, we analyzed
10   presence/absence of foliar injury, as well as a cutoff for risk of elevated foliar injury.  Generally,
11   2010 showed a more dramatic rise in the proportion of sites showing the presence of foliar injury
12   or elevated foliar injury at W126 index values below 10 ppm-hrs, and 2006 through 2009
13   showed a more subtle pattern. Figure 8-5 below, which originally appears as Figure  7-8 in
14   Chapter 7, shows the pattern  for presence/absence of foliar injury across years.

                                             Biosite Index > 0
                        o
                        o
                        I  °
                        CL  O
                                         10
15
16   Figure 8-5
17
                              20
                          W126 (ppm-hrs)
                                                              30
                                                                        40
Cumulative Proportion of Sites with Foliar Injury Present, by Year
                                                8-22

-------
 1          In addition to the above foliar injury analyses, the screening-level assessment for 214
 2   national parks assessed foliar injury in individual years. This assessment, which was based on
 3   W126 index values and  soil moisture that varied temporally, concluded that O3-related foliar
 4   injury risk in parks was generally lower in the 2008-2010 time period than in the 2006-2008 time
 5   period. For the base scenario, 2009 represented the year with the lowest percentage of parks
 6   exceeding the benchmark criteria (i.e., only 12 percent of parks) and 2006 represented the year
 7   with the highest percentage of parks exceeding the benchmark criteria (i.e., 80 percent of parks).
 8   Further, this assessment determined that the 3-month timeframe corresponding to the highest
 9   W126 estimates in monitored parks occurred between March and September, which roughly
10   corresponds to the vegetation growing season.

11           8.3.3 Risk Patterns Across Alternative W126 Standard Levels
12          For the ecological effect of biomass loss, Os-related exposure and risk decrease at lower
13   alternative W126 standard levels. For the ecological effect of foliar injury, changes in Os-related
14   exposure and risk at lower alternative W126 standard levels are more challenging to directly
15   assess because we do not have concentration-response (C-R) functions to assess changes in foliar
16   injury across different W126 index values. However, we observe that after just meeting the
17   existing standard, all of the 214 parks are below 10.46 ppm-hrs, which corresponds to the W126
18   benchmark for the base scenario.  See Table 8-1 and Table 8-2 for a summary of risk across
19   alternative W126 standard levels for these two ecological effects.

-------
1    Table 8-1   Summary of O3-Exposure Risk Across Alternative W126 Standards Relative to Just Meeting Existing Standard -
2    National-Scale Analyses
                                                                         15 ppm-hrs
                                                                           11 ppm-hrs
                                                                                   7 ppm-hrs
       Ecological Effect
                    Biomass Loss
Average Weighted RBL Loss
for Tree Seedlings
(Section 6.8)
Percent of Covered Area exceeding 1
and 2 percent weighted RBL
declines by about 0.3 percent
Percent of Covered Area exceeding 1
and 2 percent weighted RBL declines
by between 0.5 and 1.3 percent
Percent of Covered Area exceeding
1 and 2 percent weighted RBL
declines by between 0.6 and 2
percent	
              Ecosystem Services
                     Provisioning
Timber Production
(Section 6.3)
For hardwoods and upland
hardwoods, RYL between 1 and
3.25 percent for Southeast, Central,
and South regions. All other regions
RYL below 1 percent.	
For hardwoods and upland hardwoods,
RYL between 1 and 3 percent for
Southeast, Central, and South regions.
All other regions RYL below 1 percent.
For upland hardwoods, RYL
around 2 percent for Southeast
region. All other regions RYL
below 1 percent.
                                  Consumer and Producer
                                  Surplus (2010$) - Forestry
                                  (Section 6.3)
                             Consumer surplus - in 2010 is $7
                             million, or 0.01% and in 2040 is
                             $597 million, or 0.06%

                             Producer surplus - in 2010 is -$11
                             million, or -0.01% and in 2040 is
                             -$839 million, or -0.6%	
                                   Consumer surplus - in 2010 is $44
                                   million, or 0.01% and in 2040 is $712
                                   million, or 0.08%

                                   Producer surplus - in 2010 is -$41
                                   million, or -0.04% and in 2040 is -$858
                                   million, or -0.6%	
                                      Consumer surplus - in 2010 is $86
                                      million, or 0.01% and in 2040 is
                                      $779 million, or 0.08%

                                      Producer surplus - in 2010 is
                                      -$136 million, or -0.15% and in
                                      2040 is -$766 million, or -0.6%
                                  Agricultural Harvest
                                  (Section 6.5)
                             For some sensitive crops (soybeans),
                             RYL remain > 1 percent in the
                             Southwest and Central regions. All
                             other regions RYL below 1 percent.
                                   For most sensitive crops, RYL < 1
                                   percent.
                                      For most sensitive crops, RYL < 1
                                      percent.
                                  Consumer and Producer
                                  Surplus (2010$) - Agriculture
                                  (Section 6.5)
                             Consumer surplus - in 2010 is $15
                             million, or <0.01% and in 2040 is $3
                             million, or <0.01%

                             Producer surplus - in 2010 is $612
                             million, or 0.08%; in 2015 is -$1,255
                             million, or -0.15%; and in 2040 is
                             $697 million, or 0.08%	
                                   Consumer surplus - in 2010 is $19
                                   million, or <0.01% and in 2040 is $13
                                   million, or <0.01%

                                   Producer surplus - in 2010 is $1,474
                                   million, or 0.2%; in 2015 is -$2,197
                                   million, or -0.26%; and in 2040 is
                                   $2,189 million, or 0.25%	
                                      Consumer surplus - in 2010 is
                                      -$31 million, or <0.01% and in
                                      2040 is $46 million, or <0.01%

                                      Producer surplus - in 2010 is $269
                                      million, or 0.04%; in 2015 is
                                      -$1,873 million, or -0.23%; and in
                                      2040 is $2,991 million, or 0.3%
                       Regulating
Carbon Sequestration
(Section 6.6.1)
Little change compared to just
meeting existing standard
In forestry sector, storage potential is
projected to increase 593 million metric
tons of CO2 equivalents (CO2e), or 0.66
percent, over 30 years.

In agricultural sector, storage potential is
projected to increase 9 million metric
tons of CO2e, or about 0.1 percent, over
30 years.
In forestry sector, storage potential
is projected to increase 1.6 billion
metric tons of CO2e, or 1.79
percent, over 30 years.

In agricultural sector, storage
potential is projected to increase 10
million metric tons of CO2e, or 0.1
percent, over 30 years.	
                                                                                     8-24

-------
2
3
4
Table 8-1 Summary of O3-Exposure Risk Across Alternative W126 Standards Relative to Just Meeting Existing Standard -
National-Scale Analyses, continued
                                                                        15 ppm-hrs
                                                                                                       11 ppm-hrs
                                                                                  7 ppm-hrs
       Ecological Effect
                     Foliar Injury
                            National Foliar Injury
                            Screening16
                            (Section 7.2)
Depending on year, between 12 and
18 percent of sites show
presence/absence of foliar injury and
between 3 and 6 percent of sites
show elevated foliar injury

Depending on moisture category,
between 7 and >20 percent of sites
show presence/absence of foliar
injury and between 3 and >6
percent of sites show elevated foliar
injury	
Depending on year, between 12 and
>20 percent of sites show
presence/absence of foliar injury and
between 2 and 6 percent of sites show
elevated foliar injury

Depending on moisture category,
between 7 and >20 percent of sites
show presence/absence of foliar injury
and between 3 and 5 percent of sites
show elevated foliar injury
Depending on year, between 4 and
> 20 percent of sites show
presence/absence of foliar injury
and between 2 and 6 percent of
sites show elevated foliar injury

Depending on moisture category,
between 7 and >20 percent of sites
show presence/absence of foliar
injury and between 3 and 6
percent of sites show elevated
foliar injury	
     16 This analysis is not relative to just meeting the existing standard, but is a national-scale analysis that summarizes foliar injury at different levels.
                                                                                    8-25

-------
1     Table 8-2   Summary of O3-Exposure Risk Across Alternative Standards Relative to Just Meeting Existing Standard -
2     Case Study-Scale Analyses
                                                                    15 ppm-hrs
                                                                        11 ppm-hrs
                                                                                        7 ppm-hrs
             Ecosystem Services
            Regulating (Biomass
                          Loss)
Carbon Sequestration
(Section 6.6.2)
W126 levels and carbon storage
potential do not change relative to
just meeting existing standard
Atlanta, Chicago, and the urban areas of
Tennessee show gains in carbon
sequestration. For example, urban areas of
Tennessee gain about 8,800 tons of
sequestration annually.

Syracuse and Baltimore do not realize
gains because recent W126 index values
almost meet the alternative standards
levels.
Atlanta, Chicago, and the urban areas
of Tennessee show gains in carbon
sequestration. For example, urban areas
of Tennessee gain about 20,000 tons of
sequestration annually.

Syracuse and Baltimore do not realize
gains because recent W126  index
values almost meet the alternative
standards levels.
                                 Pollution Removal
                                 (Section 6.7)
                        W126 index values and pollution
                        potential do not change relative to
                        just meeting existing standard
                                   Atlanta, Chicago, and the urban areas of
                                   Tennessee show gains in pollution
                                   removal. For example, urban areas of
                                   Tennessee gain about 5,300 tons of
                                   pollution removal annually.

                                   Syracuse and Baltimore do not realize
                                   gains because recent W126 index values
                                   almost meet the alternative standards
                                   levels.
                                         Atlanta, Chicago, and the urban areas
                                         of Tennessee show gains in pollution
                                         removal. For example, urban areas of
                                         Tennessee gain about 11,700 tons of
                                         pollution removal annually.

                                         Syracuse and Baltimore do not realize
                                         gains because recent W126 index
                                         values almost meet the alternative
                                         standards levels.
             Ecosystem Services
          Cultural (Foliar Injury)
Recreation
(Section 7.4)
Rocky Mountain National Park - No
area of park exceeds a W126
standard level of 15 ppm-hrs when
the W126 index values are adjusted
to just meeting the existing standard

Great Smoky Mountains National
Park and Sequoia/Kings National
Park — No area of parks exceeds a
W126 standard level of 15 ppm-hrs
when the W126 index values are
adjusted to just meeting the existing
standard

In screening-level assessment, of 214
parks, 3 parks remain above 7 ppm-
hrs after W126 index values are
adjusted to 15 ppm-hrs	
Rocky Mountain National Park - 59
percent of the park would meet
alternative W126 standard level of between
11 and 7 ppm-hrs when the W126 index
values are adjusted to just meeting the
existing standard

Great Smoky Mountains National Park and
Sequoia/Kings National Park — No area of
parks exceeds a W126 standard level of
1 Ippm-hrs when the W126 index values
are adjusted to just meeting the existing
standard

In screening-level assessment, of 214
parks, 2 parks remain above 7 ppm-hrs
after W126 index values are adjusted to 11
ppm-hrs	
Rocky Mountain National Park - 59
percent of the park would meet
alternative W126 standard level of
between 11 and 7 ppm-hrs when the
W126 index values are adjusted to just
meeting the existing standard

Great Smoky Mountains National Park
and Sequoia/Kings National Park — No
area of parks exceeds a W126
standard level of 7 ppm-hrs when the
W126 index values are adjusted to just
meeting the existing standard

In screening-level assessment, of 214
parks, no parks remain above 7 ppm-
hrs after W126 index values are
adjusted to 7 ppm-hrs	
                                                                                      8-26

-------
 1    8.4    Representativeness
 2          In conducting the national and case-study scale analyses of ecological effects and
 3    resulting impacts on ecosystem services, we worked to reflect appropriate representation of
 4    vegetation species, geographic regions, and timeframes. The following briefly discusses the
 5    representativeness across species, geography, and time in our analyses.
 6          8.4.1  Species Representativeness
 7          To estimate the effect of Os exposure on biomass loss, we used data on 12 tree species
 8    and 10 crop species.  The 12 species represent a range of sensitivities normally distributed
 9    around intermediately sensitive species. Several species are very non-sensitive, two species are
10    relatively more sensitive, and the remainder is between non-sensitive and moderately sensitive
11    species.  The data on the 12 species facilitate representation of species for which we do not have
12    data. For tree species, we used data for areas with at least one of the tree species present,
13    resulting in approximately 46.6 percent of the contiguous U.S. constituting the area being
14    assessed.  For 74 percent of the area being assessed, the species we know about made up 50
15    percent or less of total basal area cover. For another 12 percent of the area being assessed, the
16    species we know about made up between 50 and 75 percent of total basal area cover. For the
17    remaining 14 percent of the area being assessed, the species we know about made up over 75
18    percent of total basal area cover. Although we know that there are additional (Vsensitive
19    species, we do not have C-R functions for those species. We also used these C-R functions for
20    the tree and crop species in FASOMGHG, and to better employ the dynamic tradeoffs within the
21    model, FASOMGHG assigns proxy functions for O3 exposure C-R functions for additional
22    species.  For the iTree case-study  scale analysis  on carbon sequestration and pollution removal,
23    we chose the five urban areas based on data availability and presence of species with a W126 C-
24    R function. No urban areas with available vegetation data had more than three sensitive species
25    present.  Unlike FASOMGHG, the iTree model  does not provide tradeoffs between species, so
26    the species that do not have a C-R function were not assigned values, and thus were not part of
27    the carbon sequestration and pollution removal estimates.  Therefore, the majority of trees in
28    those urban areas were not  accounted for in the Os damages. For example, there are three tree
29    species present in these areas that we know are sensitive but for which no C-R function is
                                                8-27

-------
 1    available, excluding 80 - 90 percent of the total trees present in these two study areas. The
 2    species include northern red oak in Baltimore and southern red oak and tulip tree in Atlanta.
 3          We also qualitatively discuss many additional ecological effects and ecosystem services
 4    for which we do not have data to assess quantitatively; those ecological effects and related
 5    ecosystem services include supporting services such as net primary productivity; regulating
 6    services such as hydrologic cycle and pollination; provisioning services such as commercial non-
 7    timber forest products; and cultural services such as recreation, aesthetic and non-use values. In
 8    addition, other ecological effects that are causally or likely causally associated with Os exposure
 9    are not directly addressed in this risk and exposure assessment. These ecological effects include
10    terrestrial productivity, water cycle, biogeochemical cycle, and community composition.17
11            8.4.2  Geographic Representativeness
12          Nine of the 12 tree species used in the biomass analyses were in the eastern U.S. and
13    three were in the western U.S., with a few species such as Aspen and Cottonwood  in both the
14    eastern and western U.S. For the biomass loss analyses, by region we include the total basal area
15    covered by the 12 tree species assessed. In parts of the eastern U.S. - the Central,  East North
16    Central, and Northeast regions — from less than 1 percent to 4 percent of basal area assessed had
17    no data on percent cover of the 12 tree species.  In contrast, in parts of the western U.S. -
18    Southwest, West, West North Central regions — from 47 percent to 74 percent of basal area
19    assessed had no data on percent cover of the 12 tree species.
20          We applied C-R functions for 12 tree species and 10 crop species in FASOMGHG to
21    estimate nationwide effects on timber production, agricultural harvest, and carbon  sequestration.
22    While we used available C-R functions for tree and crop species, as well as the available models,
23    we had differential and inconsistent species coverage across the U.S., e.g., data were available
24    for more species in the eastern U.S. than in the western U.S., limiting our analyses. In addition,
25    to assess overall ecosystem-level effects from biomass loss, we combined the RBL values for
26    multiple tree species into a weighted RBL value and considered the weighted value in relation to
27    proportion of basal area  covered, both nationally and in Class I areas.
      17 For additional details on these other ecological effects, see Table 2-4 of the Integrated Science Assessment for
      Ozone and Related Photochemical Oxidants (U.S. EPA, 2013).
                                                 8-28

-------
 1           Also, in estimating the effect of 63 exposure on biomass loss and resulting changes in
 2    carbon sequestration and pollution removal capacity, for case-study scale analysis we used the
 3    iTree model and data from five urban areas. The urban areas represent diverse geography in the
 4    Northeast, Southeast, and Central regions, but we did not assess an urban area in the western part
 5    of the U.S.  Based on the monitored data from 2006 to 2008, Atlanta, Baltimore, and the urban
 6    areas in Tennessee have W126 index values over 20 ppm-hrs, with Atlanta having the highest
 7    index value. After adjusting to just meeting the existing standard, all of the urban areas show
 8    W126 index values between 5 and 7 ppm-hrs. Because there are more monitors in urban areas in
 9    the eastern U.S., we focused on urban areas in the eastern U.S. for the case-study analyses.
10           For the national-scale foliar injury analysis, we were limited by the available foliar injury
11    data. Biosite sampling was  discontinued in some states prior to our analysis.  Although we had
12    data for most regions of the  contiguous U.S., we did not have data for the Southwest and limited
13    data for the West and West North Central regions.  For example, over 2006 to 2010 there were
14    over 1,000 biosite index values each for the Northeast and Central regions and no biosite index
15    values for the Southwest.  In assessing foliar injury at parks, we conducted national scale
16    analyses, as well as a case-study scale analysis of national parks. In  assessing foliar injury at the
17    case-study scale, the three national parks represent diverse geographic areas — in the
18    Southeast/Central  (GRSM),  the Southwest (ROMO), and the West (SEKI). In the screening-
19    level assessment of foliar injury, we included 214 national parks in the contiguous U.S.
20           8.4.3   Temporal Representativeness
21           The biomass loss analysis relied upon the national-scale air quality surfaces described in
22    section 8.2.1.1 and in Chapter 4. A separate set of surfaces were created for the national-scale
23    analyses of presence/absence of and elevated foliar injury, for which national-scale surfaces were
24    generated by interpolating the unadjusted monitored annual W126 index values for the individual
25    years 2006 through 2010. Monitored Os index values in those years  vary considerably, and those
26    years represent a reasonable range of meteorological conditions that  affect Os formation. The
27    period also includes years with varying categories of soil moisture, which impacts the sensitivity
28    of plants to foliar injury.
29           Because the forestry and agriculture sectors are  interlinked and factors affecting one
30    sector can lead to changes in the other, we considered overall effects on producers and

                                                8-29

-------
 1    consumers associated with just meeting alternative W126 standard levels over time and across
 2    sectors.  In estimating the effect of 63 exposure on biomass loss and ecosystem services, we used
 3    the C-R functions for 12 tree seedlings in FASOMGHG to calculate relative yield changes over
 4    the entire lifespan of the trees, including percentage changes in national timber product market
 5    prices through 2040.
 6          At the national scale, we also used FASOMGHG to calculate changes in carbon
 7    sequestration by forests and agriculture through 2040.   At the case-study scale, we used iTree to
 8    estimate tree growth and calculate changes in carbon sequestration and pollution removal
 9    capacity in the five urban areas over a 25-year period.
10    8.5    Overall Confidence in Welfare Exposure and Risk Results
11          There are several important factors to consider when evaluating the overall confidence
12    we can express about the estimates of exposures and risks associated with just meeting the
13    existing and potential alternative W126 secondary standards.  As with any complex analysis
14    using estimated parameters and inputs from numerous data sources and models, there are many
15    sources of uncertainty that may affect estimated results. These sources of uncertainty are
16    discussed in each of the chapters related to air quality, biomass loss, visible foliar  injury and
17    ecosystem services.
18          The overall  effect of the combined set of uncertainties on confidence in the interpretation
19    of the results of the analyses is difficult to quantify. Due to differences in available information,
20    the degree to which each analysis was able to incorporate quantitative assessments of uncertainty
21    differed.  In general, we followed the WHO tiered approach to uncertainty characterization,
22    which includes both quantitative and qualitative assessments.  Chapters 4, 5, 6, and 7 include
23    tables identifying and characterizing the potential impact of key uncertainties  on risk estimates,
24    including the degree to which we were able to quantitatively address those uncertainties.
25          Below we discuss several  key limitations and uncertainties, which may have a large
26    impact on both overall confidence and confidence in individual analyses.
27            8.5.1 Uncertainties in Air Quality Analyses
28          The national W126 surface was created using the VNA technique to interpolate recent air
29    quality measurements of Os. In general, spatial interpolation techniques perform better in areas

                                                8-30

-------
 1    where the Os monitoring network is denser.  Therefore, we have lower confidence in the W126
 2    estimated in the rural areas in the West, Northwest, Southwest, and West North Central with few
 3    or no monitors.
 4           An additional uncertainty comes from the adjustment methodology, which used U.S.-
 5    wide NOx emissions reductions to adjust air quality to just meet the existing and alternative
 6    standard levels.  Consequently, meeting the standard levels at the highest monitor in each region
 7    (which generally occurs in or near a major urban area) leads to substantial reductions below the
 8    targeted level through the rest of the region.  These across-the-board NOx cuts do not represent
 9    an actual  control strategy, and it should be noted that resulting air quality could look different if
10    we used different assumptions about emissions reductions strategies.  However, the assumption
11    of broad regional or national NOx reductions is not unreasonable given current EPA regulations
12    such as (i) the Clean Air Interstate Rule (CAIR), which requires NOx emissions reductions
13    across the Eastern U.S.  to reduce regional ozone transport, and (ii) the multitude of onroad and
14    offroad mobile source rules that will lead to reductions in NOx emissions from these sources
15    across the country in future years.
16           Because the W126 estimates generated in the air quality analyses are inputs to the
17    vegetation risk analyses for biomass loss and foliar injury, any uncertainties in the air quality
18    analyses are propagated into those analyses.
19           8.5.2 Uncertainties in Biomass Loss Analyses
20           Even though we are certain that there are additional species adversely affected by O3-
21     related biomass loss, we only have C-R functions available to  quantify this loss for  12 tree
22     species and 10  crop species. This absence of information only allows a partial characterization
23     of the Os-related biomass loss impacts in trees and crops associated with recent Os index values
24     and with just meeting the existing and potential  alternative secondary standards. In addition,
25     there are uncertainties  inherent in these C-R functions, including the extrapolation of relative
26     biomass loss rates from tree seedlings to adult trees and information regarding within-species
27     variability. The overall confidence in the C-R function varies by species based on the number of
28     studies available for that species. Some species have low within-species variability (e.g., many
29     agricultural crops) and high seedling/adult comparability (e.g., Aspen), while other species do
30     not (e.g., Black Cherry). The uncertainties in the C-R functions for biomass loss and in the air

                                                8-31

-------
 1     quality analyses are propagated into the analysis of the impact of biomass loss on ecosystem
 2     services, including provisioning and regulating services.
 3           In the national-scale analyses of timber production, agricultural harvesting, and carbon
 4     sequestration, we used the FASOMGHG model, which includes functions for carbon
 5     sequestration, assumptions regarding proxy species, and non-W126 C-R functions for three
 6     crops. However, FASOMGHG does not include agriculture and forestry on public lands,
 7     changes in exports due to O3 into international trade projections, or forest adaptation. Despite
 8     the inherent limitations and uncertainties, we believe that the FASOMGHG model reflects
 9     reasonable and appropriate assumptions  for a national-scale assessment of changes in the
10     agricultural and forestry sectors due to changes in vegetation biomass associated with Os
11     exposure.
12           In the case study analyses of five urban areas, we used the iTree model, which includes
13     an urban tree inventory for each area and species-specific pollution removal and carbon
14     sequestration functions. However, iTree  does not account for the potential additional VOC
15     emissions from tree growth, which could contribute to O?, formation.  Despite the inherent
16     limitations and uncertainties, we believe that the iTree model reflects reasonable and
17     appropriate assumptions for a case study assessment of pollution removal and carbon
18     sequestration for changes in biomass associated with Os exposure.
19           8.5.3  Uncertainties in Visible Foliar Injury Analyses
20           To develop benchmarks for evaluating visible foliar injury, we conducted a national-scale
21    analysis using biosite sampling data combined with Os exposure and soil moisture. Evaluating
22    soil moisture is more  subjective than evaluating Os exposure because of its high spatial and
23    temporal variability within the Os season, and there is considerable  subjectivity in the
24    categorization of relative drought.  On balance, we believe that the spatial and temporal
25    resolution for the soil moisture data is likely to underestimate the potential of foliar injury that
26    could occur in some areas.  In addition, we are unaware of a clear threshold for drought below
27    which visible foliar injury would not occur.  In general, low soil moisture reduces the potential
28    for foliar injury, but injury  could still occur, and the degree of drought necessary to reduce
29    potential injury is not clear. Due to the absence of biosite injury data in the Southwest region and
30    limited biosite data in the West and West North Central regions, the benchmarks applied may not

                                                8-32

-------
 1    be applicable to these regions. We applied the benchmarks from the national-scale analysis to a
 2    screening-level analysis of 214 national parks and case studies of three national parks. Therefore,
 3    uncertainties in the foliar injury benchmarks and in the air quality analyses are propagated into
 4    the national park analyses. Additional uncertainties in the national park analyses are primarily
 5    related to the mapping, including park boundaries, vegetation species cover, and park amenities,
 6    such as scenic  overlooks and trails. In general, we have high confidence in the park mapping.
 7    8.6    Conclusions
 8           This welfare risk and exposure assessment provides information to further inform the
 9    following policy-relevant questions18: (1) in considering alternative standards, to what extent do
10    alternative standard levels, averaging times, and forms reduce estimated exposures and welfare
11    risks attributable to 63; (2) what range of alternative standard levels should be considered based
12    on the scientific information evaluated in the ISA, air quality analyses, and the welfare risk and
13    exposure assessment; and (3) what are the important uncertainties and limitations in the evidence
14    and assessments and how might those uncertainties and limitations be taken into consideration in
15    identifying alternative secondary standards for consideration. To develop information to help
16    inform these questions, we quantified ecological effects based on the relationship with the W126
17    metric  and assessed the associated impacts on ecosystem services. For some ecosystem  services,
18    such as commercial non-timber forest products, recreation, and aesthetic and non-use values, we
19    qualitatively assessed potential impacts  to services. We assessed impacts on ecosystem services
20    at the national  and case-study scales, as  well as across species, U.S. geographic regions  and
21    future years. Throughout the assessment, we characterized the uncertainties inherent in the
22    analyses.
23           To assess the ecological effect of biomass loss, we used C-R functions for tree seedlings
24    and crops to determine the range  of biomass loss associated with just meeting alternative W126
25    standard levels relative to just meeting the existing 75 ppb 8-hour standard. To compare
26    different levels of biomass loss to different W126 index values, we plotted the C-R functions as a
27    function of the percent biomass loss against varying W126 index values. For a one percent
28    biomass loss for trees, the estimated W126 index values were between 4 and 10 ppm-hrs; for a
29    two percent biomass loss for trees the estimated W126 index values were between 7 and 14 ppm-
      18 The policy-relevant questions were identified in the Integrated Review Plan for the Ozone National Ambient Air
      Quality Standards (IRP, US EPA, 201 la).
                                                8-33

-------
 1    hrs; and for a five percent biomass loss for crops the estimated W126 index values were between
 2    12 and 17 ppm-hrs. We also conducted an analysis to estimate the ecosystem-level impacts of
 3    biomass loss from 12 tree species using weighted average RBL functions. These results indicate
 4    that of the area being assessed, the portion exceeding benchmarks of one to two percent biomass
 5    loss decreases as W126 index values decrease.  For example, after just meeting the existing
 6    standard, 20.8 percent and 12.4 percent of the total area being assessed exceeds benchmarks of
 7    one percent and two percent biomass loss in trees, respectively.  After just meeting an alternative
 8    standard level of 7 ppm-hrs, 11.5 percent and 7.7 percent of the total area being assessed still
 9    exceeds benchmarks of one and two percent biomass loss in trees, respectively.
10           We also used the FASOMGHG model to conduct national-scale analyses to quantify the
11    effects of biomass loss on timber production  and agricultural harvesting, as well as on carbon
12    sequestration.  Because the forestry and agriculture sectors are related, and trade-offs occur
13    between the sectors, we also calculated the resulting market-based welfare effects of 63 exposure
14    in the forestry and agriculture sectors.  For commercial timber production, because most areas
15    have W126 index values lower than 15 ppm-hrs when simulating meeting the existing standard,
16    the RYLs are below one percent, with  the exception of the Southwest, Southeast, Central, and
17    South regions.  For some sensitive crops, such as soybeans, RYLs above 1 percent remain at 15
18    ppm-hrs for the Southwest and Central regions; RYLs  are reduced to below one percent at
19    alternative W126 standard levels of 11 and 7 ppm-hrs.  Economic welfare impacts resulting from
20    just meeting the existing and alternative standards were largely similar between the forestry and
21    agricultural sectors - consumer surplus, or consumer gains, generally increased in both sectors
22    because higher productivity under lower W126 index values increased total production and
23    reduced market prices. For producer surplus, there were many examples for which producer
24    surplus declines. For example, in 2040, in the forestry sector when adjusting to meeting
25    alternative W126 standards of 15, 11, and 7 ppm-hrs, consumer surplus increases $597 million,
26    $712 million, and $779 million (i.e., 0.06, 0.08, and 0.08 percent), respectively, while producer
27    surplus decreases $839 million, $858 million, and $766 million, (i.e., about 0.6 percent),
28    respectively.
29           To assess the ecological effect  of foliar injury at a national scale, we conducted several
30    analyses based on a national data set on foliar injury from the  USFS's Forest Health Monitoring
31    Network.  We conducted the analyses using presence/absence of foliar injury, as well as using a
                                               8-34

-------
 1    cutoff for elevated injury; we conducted analyses across years in the data set, according to
 2    different moisture categories and across different geographic regions.  Generally, the results of
 3    these foliar injury analyses demonstrate a similar pattern - the proportion of biosites showing
 4    foliar injury increases steeply with W126 index values up to approximately 10 ppm-hrs and is
 5    relatively constant at W126 index levels above 10 ppm-hrs.  Using benchmarks derived from the
 6    national analysis, we conducted a screening-level assessment of foliar injury at national parks for
 7    214 national parks in the contiguous U.S. Generally, as the percentage of biosites showing foliar
 8    injury increases, the percentage of parks exceeding that benchmark decreases; similarly as the
 9    degree of foliar injury is elevated, the percentage of parks exceeding that benchmark decreases.
10    During 2006 to 2010, 58 percent of parks exceeded the benchmark criteria corresponding to the
11    base scenario (W126>10.46 ppm-hrs, 17.7 percent of biosites, without consideration of soil
12    moisture, any injury) for at least three years, and 34 percent of parks would exceed the
13    benchmark criteria for the elevated injury scenario (five percent of biosites, multiple moisture
14    categories, elevated foliar injury) for at least three years. Because the screening-level assessment
15    relies on annual estimates of W126 index values and soil moisture, we cannot fully evaluate just
16    meeting the existing and alternative standards because they are based on the 3-year average air
17    quality surfaces.  However, we can observe that after adjusting the W126 surfaces to just meet
18    the existing standard (3-year average), all of the 214 parks are below 10.46 ppm-hrs, which
19    corresponds to the annual benchmark criteria for the base  scenario.
20          In conclusion, we estimated that some exposures and risks remain after just meeting the
21    existing standard and that in many cases, just meeting alternative standard  levels results in
22    reductions in those remaining exposures and risks. Overall, the largest reduction in Os exposure-
23    related welfare risk occurs when moving from recent ambient conditions to meeting the existing
24    secondary standard of 75 ppb (equal to the existing primary standard).  When using monitored
25    W126 index values and adjusting for meeting the existing 63 standard of 75 ppb, only two of the
26    nine U.S. regions have 3-year index values remaining above 15  ppm-hrs (West — 18.9 ppm-hrs
27    and Southwest - 17.7 ppm-hrs). Four regions (East North Central, Northeast, Northwest, and
28    South) would meet 7 ppm-hrs, and two regions (Southeast and West North Central) have 3-year
29    index values between 9 and 12 ppm-hrs (Southeast -11.9 ppm-hrs and West North Central - 9.3
30    ppm-hrs).  When adjusting to just meeting the existing standard, the Central region would meet
31    an alternative W126 of 15 ppm-hrs, but further air quality adjustment would be needed for the

                                                8-35

-------
1
2
3
4
5
6
7
Central region to meet alternative standards of 11 and 7 ppm-hrs - alternate standard levels that
would protect against the recommended one to two percent biomass loss for trees and five
percent for crops. At an alternative W126 standard level of 15 ppm-hrs, ambient conditions and
related risk are not appreciably different than they are after just meeting the existing standard of
75 ppb, and the risk decreases at alternative W126 standard levels of 11 ppm-hrs and 7 ppm-hrs,
just not as much as the decrease in risk from recent conditions to the existing standard of 75 ppb.
                                               8-36

-------
 1                                    9      REFERENCES

 2    Acharya, G., and L.L. Bennett. (2001). Valuing Open Space and Land-Use Patterns in Urban
 3          Watersheds. Journal of Real Estate Finance and Economics 22(2/3):221-23 7.

 4    Adams, D.; Alig, R.; McCarl, B.A.; Murray, B.C. (2005). FASOMGHG Conceptual Structure,
 5          and Specification: Documentation. Available at
 6          http://agecon2.tamu.edu/people/faculty/mccarl-bruce/FASOM.html

 7    Alexander, S.J.;  Oswalt, S.N.; Emery, M.R. (2011). Nontimber forest products in the United
 8          States: Montreal Process indicators as measures of current conditions and sustainability.
 9          Gen. Tech. Rep. PNW-GTR-852. Portland, OR: U.S. Department of Agriculture, Forest
10          Service, pacific Northwest Research Station. 36 p. Available at
11          http://www.fs.fed.us/pnw/pubs/pnw gtr851 .pdf

12    Alexander, S.J.;  Weigand, J.; Blatner, K.A. (2002). Nontimber forest product commerce in
13          Nontimber forest products of the United States, Jones, E.T.; McLain, R.J.; Weigand, J.,
14          eds. University of Kansas Press.

15    Arbaugh, M; Bytnerowicz, A; Grulke, N; Fenn, M; Poth, M; Temple, P; Miller, P. (2003).
16          Photochemical smog effects in mixed conifer forests along a natural gradient of ozone
17          and nitrogen deposition in the  San Bernardino Mountains. Environ Int 29: 401-406.

18    Arbaugh, MJ;  Miller, PR; Carroll, JJ;  Takemoto, BL; Proctor, T. (1998). Relationships of ozone
19          exposure to pine injury in the Sierra Nevada and San Bernardino Mountains of
20          California, USA. Environ Pollut 101: 291-301.

21    Baumflek, M.J.;  Emery, M.R.; Ginger, C. (2010). Culturally and economically important
22          nontimber forest products of northern Maine. Gen.Tech. Rep. NRS-68. Newton Square,
23          PA: Y.s.  Department of Agriculture, Forest Service, Northern Research Station. 74p.
24          Available at
25          http://nrs.fs.fed.us/sustaining_forests/conserve_enhance/special_products/maine_ntfp/

26    Benoit, LF; Skelly, JM; Moore, LD; Dochinger, LS. (1982). Radial growth reductions of Pinus
27          strobus L correlated with foliar ozone sensitivity as an indicator of ozone-induced losses
28          in eastern forests.  Can J For Res 12: 673-678.

29    Berrocal, V.J., Gelfand, A.E., Holland, D.M. (2012). Space-Time Data Fusion Under Error in
30          Computer Model Output:  An Application to Modeling Air Quality. Biometrics, 68(3),
31          837-848.

32    Black, VJ; Black, CR; Roberts, JA; Stewart, CA. (2000). Impact of ozone on the reproductive
33          development of plants. New Phytol 147: 421-447.
                                                    9-1

-------
 1   Brown, R.B.; Xu, X.; Toth, J.F. Jr. (1998). Lifestyle options and economic strategies:
 2          susbsistence activities in the Mississippi Delta. Rural Sociology 63-4: p599-623.

 3   Brown, T. C., and T. C. Daniel. (1984). Modeling forest scenic beauty: Concepts and application
 4          to ponderosa pine. USDA Forest Service Research Paper RM-256. Rocky Mountain
 5          Forest and Range Experiment Station, Fort Collins, Colorado. Brown, T.C., Daniel, T.C.,
 6          1986. Predicting scenic beauty of forest timber stands. Forest Sci. 32, 471-487.

 7   Brown, T.C., and T.C. Daniel. (1986). Predicting scenic beauty of timber stands. Forest Science
 8          32:471-487.

 9   Burns, Russell M., and Barbara H. Honkala, tech. coords. (1990). Silvics of North America: 1.
10          Conifers; 2. Hardwoods. Agriculture Handbook 654. U.S. Department of Agriculture,
11          Forest Service, Washington, DC. vol.2, 877 p.

12   Butler, B.J. (2008). Family forest owners of the United States. (2006). Gen. Tech. Rep. NRS-
13          GTR-27. Newton Square, PA: U.S. Department of Agriculture, Forest Service, Northern
14          Research Station. 72 p.

15   Buyhoff, G.J.; Wellman, J.D.; Daniel, T.C.  (1982). Predicting scenic quality for mountain pine
16          beetle and western  spruce budworm damaged forest vistas. Forest Science 28:827-838.

17   Byun, D.W. and J.K.S. Ching. (1999). Science algorithms of the EPA Models-3 Community
18          Multiscale Air Qualty (CMAQ) Modeling System, (EPA/600/R-99/030 0). U.S. EPA.

19   Byun, D.W. and K.L. Schere. (2006). Review of the governing equations, computational
20          algorithms, and other components of the Models-3 Community Multiscale Air Quality
21          (CMAQ) modeling system. Applied Mechanics Reviews, 59. 51-77'.

22   CAL FIRE (California Department of Forestry and Fire Protection). (1996). California Fire Plan.
23          Available at http://cdfdata.fire.ca.gov/fire_er/fpp_planning_cafireplan

24   CAL FIRE (California Department of Forestry and Fire Protection). (2008). CAL FIRE 2007
25          Wildland Fire Summary. Available at
26          www.fire.ca.gov/communications/downloads/fact  sheets/2007Summary.pdf.

27   Carlton, A.G., R.W. Finder, P.V. Bhave and G.A. Pouliot. (2010). To What Extent Can Biogenic
28          SOA be Controlled, Environmental  Science & Technology 44:3376-3380.

29   Chappelka, A; Skelly, J; Somers, G; Renfro, J; Hildebrand, E. (1999a). Mature black cherry used
30          as a bioindicator of ozone injury. Water Air Soil Pollut 116: 261-266.
                                                    9-2

-------
 1    Chappelka, A; Somers, G; Renfro, J. (1999b). Visible ozone injury on forest trees in Great
 2          Smoky Mountains National Park, USA. Water Air Soil Pollut 116: 255-260.

 3    Chappelka, AH. (2002). Reproductive development of blackberry (Rubus cuneifolius) as
 4          influenced by ozone. New Phytol 155: 249-255.

 5    Chen, J., Zhao, R., Li, Z.  (2004). Voronoi-based k-order neighbor relations for spatial analysis.
 6          ISPRS J Photogrammetry Remote Sensing, 59(1-2), 60-72.

 7    Constable, J. V. H. and G. E. Taylor. (1997). Modeling the effects of elevated tropospheric Os on
 8          two varieties of Pinus ponderosa. Canadian Journal of Forest Research 27:527-53

 9    Coulson, R.N.; Klepzig, K.D. (2011). Southern Pine Beetle II. Gen. Tech. Rep.  SRS-140.
10          Asheville, NC: U.S. Department of Agriculture Forest Service, Southern Research
11          Station, 512p. Available at http://www.srs.fs.usda.gov/pubs/gtr/gtr_srs 140/gtr_srs 140.pdf

12    Daniel, T.C.; Brown, T.C.; King, D.A.; Richards, M.T.;  Stewart, W.P. (1989). Perceived scenic
13          beauty and contingent valuation of forest campgrounds. Forest Science 35:76-90.

14    Darbah, JNT; Kubiske, ME; Neilson, N; Oksanen, E; Vaapavuori, E; Karnosky, DF. (2007).
15          Impacts of elevated atmospheric CC>2 and 63 on  paper birch (Betula papyrifera):
16          Reproductive fitness. ScientificWorldJournal 7: 240-246.

17    Darbah, JNT; Kubiske, ME; Nelson, N;  Oksanen, E; Vapaavuori, E; Karnosky,  DF. (2008).
18          Effects of decadal exposure to interacting elevated CC>2 and/or 63 on paper birch (Betula
19          papyrifera) reproduction. Environ Pollut 155: 446-452.

20    Dickson, RE; Lewin, KF; Isebrands, JG; Coleman, MD; Heilman, WE; Riemenschneider, DE;
21          Sober, J; Host, GE; Zak, DR; Hendrey, GR; Pregitzer, KS; Karnosky, DF. (2000). Forest
22          Atmosphere Carbon Transfer and Storage (FACTS-II) the Aspen Free-Air CO2 and O3
23          Enrichment (FACE) project: An overview. (General Technical Report NC-214). St. Paul,
24          MN: U.S. Dept. of Agriculture, Forest Service.

25    Emery, M.R.; Ginger, C.; Newman, S.; Giammusso, M.R.B. (2003). Special forest products in
26          context: gatherers and gathering in the eastern United States., USDA Forest Service,
27          Northeastern Research Station, Newton Square, PA.  Available at
28          http://www.fs.fed.us/ne/newtown square/publications/technical reports/pdfs/2003/gtrne3
29          06.pdf

30    Emery, M.R.; Pierce, A.R. (2005). Interrupting the telos: locating subsistence in contemporary
31          U.S. forests. Environment and Planning, 37: p. 981-993.
                                                    9-3

-------
 1   ESRI. (2011). ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research
 2          Institute.

 3   Felzer, B; Kicklighter, D; Melillo, J; Wang, C; Xhuang, Q; Prinn, R. (2004). Effects of ozone on
 4          net primary production and carbon sequestration in the conterminous United States using
 5          a biogeochemistry model. Tellus B Chem Phys Meteorol 56: 230-248.

 6   Freeman, Milton R. (1993). "The International Whaling Commission, Small-type Whaling, and
 7          Coming to Terms with Subsistence." Human Organization 52:243-51.

 8   Frey, H.C. and Samet, J.M. (2012a). CASAC Review of the EPA's Health Risk and Exposure
 9          Assessment for Ozone (First External Review Draft - Updated August 2012) and Welfare
10          Risk and Exposure Assessment for Ozone (First External Review Draft - Updated August
11          2012). EPA-CASAC-13-002. Available at
12          http://vosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/38CO
13          lBlEDD14C59E85257ABC004AOE18/$File/EPA-CASAC-13-002+unsigned.pdf

14   Frey, H.C. and Samet, J.M. (2012b). CASAC Review of the EPA's Policy Assessment for the
15          Review of the Ozone National Ambient Air Quality Standards (First External Review
16          Draft - August 2012) . ((PDF, 73 pp., 355,203 bytes). EPA-CASAC-13-003. Available at
17          http://vosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/E670
18          94C7FBBECD8685257AC200727082/$File/EPA-CASAC-13-003+unsigned.pdf

19   Gallai, N; Salles, JM; Settele, J.;Vaissiere, B.E. (2009). Economic Valuation of the vulnerability
20          of world agriculture confronted with pollinator decline. Environmental Economics
21          68:810-821.

22   Geoghegan, J., L.A. Wainger, and N.E. Bockstael. (1997). "Spatial Landscape Indices in a
23          Hedonic Framework: An Ecological Economics Analysis Using GIS." Ecological
24          Economics 23:251-264.

25   Gold, C. (1997). Voronoi methods in GIS. In: Algorithmic Foundation of Geographic
26          Information Systems (va Kereveld M., Nievergelt, J., Roos, T., Widmayer, P., eds).
27          Lecture Notes in Computer Sicnece, Vol 1340. Berlin: Springer-Verlag, 21-35.

28   Gregg, JW; Jones, CG; Dawson, TE. (2006). Physiological and developmental effects of O3 on
29          cottonwood growth in urban and rural sites. Ecol Appl 16: 2368-2381.

30   Grulke, N. E.; Minnich, R.A.; Paine, T.D.; Seybold, S.J.; Chavez, D.J.; Fenn, M.E.; Riggan, P.J.;
31          Dunn, A. (2009). Air pollution increases forest susceptibility to wildfires: a case study in
32          the San Bernardino Mountains in southern California in Developments in Environmental
33          Science, Volume 8. A. Bytnerowicz, M. Arbaugh, A. Riebau, and C. Anderson (eds.)
                                                   9-4

-------
 1   Guenther, A.; Karl, T.; Harley, P.; Wiedinmyer, C.; Palmer, P. I; Geron, C. (2006). Estimates of
 2          global terrestrial isoprene emissions using MEGAN (model of emissions of gases and
 3          aerosols from nature). Atmos. Chem. Phys. 6, 3181-3210.

 4   Gumpertz, ML; Rawlings, JO. (1992). Nonlinear regression with variance components:
 5          Modeling effects of ozone on crop yield. Crop Sci 32: 219-224.

 6   Haefele, M., R.A. Kramer, and T.P. Holmes. (1991). Estimating the Total Value of a Forest
 7          Quality in High-Elevation Spruce-Fir Forests. The Economic Value of Wilderness:
 8          Proceedings of the Conference. Gen. Tech. Rep. SE-78 (pp. 91-96).  Southeastern For
 9          Exper. Station. Asheville, NC: USD A Forest Service.

10   Hall, Karen R. and Richard R. Braham. (1998). Native Pines of Eastern North America. North
11          Carolina State University Department of Forestry. Available at
12          http://www.ncsu.edu/proj ect/dendrology/

13   Heck, WW; Cure, WW; Rawlings, JO; Zaragoza, LJ; Heagle, AS; Heggestad, HE; Kohut, RJ;
14          Kress, LW; Temple, PJ. (1984). Assessing impacts of ozone on agricultural crops: II.
15          Crop yield functions and alternative exposure statistics. J Air Pollut Control Assoc 34:
16          810-817.

17   Henderson, R. (2006a). Letter from CAS AC Chairman Rogene Henderson to EPA Administrator
18          Stephen Johnson, February 16, 2006, EPA-CASAC-06-003. Available at
19          http://vosemite.epa.gov/sab/sabproduct.nsf/69FBBlE21FBlE4428525712D004BA05D/$
20          File/casac con  06 003.pdf

21   Henderson, R. (2006b). Letter from CAS AC Chairman Rogene Henderson to EPA Administrator
22          Stephen Johnson, June 5, 2006, EPA-CASAC-06-007. Available at
23          http://vosemite.epa.gov/sab/sabproduct.nsf/0202D7053AC6E2AC852571870075ClD2/$
24          File/casac-06-007.pdf

25   Henderson, R. (2007). Clean Air Scientific Advisory Committee's (CAS AC) Review of the
26          Agency's Final Ozone Staff Paper. EPA-CASAC-07-002. March 26. Available at
27          http://vosemite.epa.gov/sab/sabproduct.nsf/FE915E916333D776852572AC007397B5/$F
28          ile/casac-07-002.pdf.

29   Hildebrand, E; Skelly, JM; Fredericksen, TS. (1996). Foliar response of ozone-sensitive
30          hardwood tree species from 1991 to 1993 in the Shenandoah National Park, Virginia. Can
31          J For Res 26: 658-669.

32   Hogsett, WE; Weber, JE; Tingey, D; Herstrom, A; Lee, EH; Laurence, JA. (1997).
33          Environmental auditing: An approach for characterizing tropospheric ozone risk to
34          forests. J Environ Manage 21: 105-120.

                                                   9-5

-------
 1   Holmes, T., and R. Kramer. (1995). "An Independent Sample Test of Yea-Saying and
 2          StartingPoint Bias in Dichotomous-Choice Contingent Valuation." Journal of
 3          Environmental Economics and Management 28:121-132.

 4   Hufford, M. (2000). Holding up the mountains: talk as a historical discourse. Smithsonian
 5          Folklife Center, Library of Congress, Washington D.C.

 6   IPCC (Intergovernmental Panel on Climate Change). (2007). Summary for policymakers. In: The
 7          Physical Science Basis. Contribution of Working Group I to the Fourth Assessment
 8          Report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom
 9          and New York, NY, USA: Cambridge University Press. Available at
10          http://www.ipcc.ch/publications  and data/ar4/wg2/en/spm.html

11   Irwin, E.G. (2002). "The Effects of Open Space on Residential Property Values." Land
12          Economics 78(4):465-480.

13   Jaegle, L., DJ. Jacob,  W.H. Brune, and P.O. Wenberg. (2001). Chemistry of HOX radicals in the
14          upper troposphere. Atmos. Environ., 35, 469-489.

15   Karl, T.R. (1986). The sensitivity of the Palmer Drought Severity Index and Palmer's Z-Index to
16          their calibration coefficients including potential evapotranspiration. J. Climate Appl.
17          Meteor., 25, 77-86.

18   Karl, T.R. and Koss, W.J. (1984). Regional and National Monthly, Seasonal, and Annual
19          Temperature Weighted by Area,  1895-1983. Historical Climatology Series 4-3, National
20          Climatic Data Center, Asheville, NC, 38 pp.

21   Karnosky, DF; Gagnon, ZE; Dickson, RE; Coleman, MD; Lee, EH; Isebrands, JG. (1996).
22          Changes in growth, leaf abscission, biomass associated with seasonal tropospheric ozone
23          exposures of Populus tremuloides clones and seedlings. Can J For Res 26: 23-37.

24   Kaval, P. and J. Loomis. (2003). Updated outdoor recreation use values with emphasis on
25          National Park recreation. Final Report, Cooperative Agreement 1200-99-009, Project
26          number IMDE-02-0070. Fort Collins, CO: Colorado State University, Department of
27          Agricultural and Resource Economics.

28   Kohut, R. (2007). Assessing the risk of foliar injury from ozone on vegetation in parks in the US
29          National Park Service's Vital Signs Network. Environ Pollut 149: 348-357.

30   Krupa, S.; Muntifering, R.B.; Chappelka, A.H. (2004). Effects of ozone on plant nutritive
31          quality characteristics for ruminant animals. Botanica 54,  129-140.
                                                    9-6

-------
 1    Kubiske, ME; Quinn, VS; Heilman, WE; McDonald, EP; Marquardt, PE; Teclaw, RM; Friend,
 2          AL; Karnoskey, DF. (2006). Interannual climatic variation mediates elevated CC>2 and O?,
 3          effects on forest growth. Global Change Biol 12: 1054-1068.
 4
 5    Kubiske, ME; Quinn, VS; Marquardt, PE; Karnosky, DF. (2007). Effects of elevated
 6          atmospheric CC>2 and/or 63 on intra- and interspecific competitive ability of aspen. Plant
 7          Biol (Stuttg) 9: 342-355.

 8    Lee, EH; Hogsett, WE. (1999). Role of concentrations and time of day in developing ozone
 9          exposure indices for a secondary standard. J Air Waste Manag Assoc 49: 669-681.

10    Lee, EH; Hogsett, WE; Tingey, DT. (1994). Attainment and effects issues regarding alternative
11          secondary ozone air quality standards. J Environ Qual 23: 1129-1140.

12    Lee, EH; Tingey, DT; Hogsett, WE. (1987). Selection of the best exposure-response model using
13          various 7-hour ozone exposure statistics. Research Triangle Park, NC: U.S.
14          Environmental Protection Agency.

15    Lee, EH; Tingey, DT; Hogsett, WE. (1988b). Evaluation of ozone exposure indices in exposure-
16          response modeling. Environ Pollut 53: 43-62.

17    Lee, EH; Tingey, DT; Hogsett, WE. (1989). Interrelation of experimental exposure and ambient
18          air quality data for comparison of ozone exposure indices and estimating agricultural
19          losses. (EPA/600/3-89/047). Corvallis,  OR: U.S. Environmental Protection Agency.

20    Lefohn, A. S., Laurence, J. A.,  & Kohut, R. J. (1988). A comparison of indices that describe the
21          relationship between exposure to ozone and reduction in the yield of agricultural crops.
22          Atmospheric Environment (1967), 22(6), 1229-1240.

23    Lehrer, J.A., M. Bacou, B. Blankespoor, D. McCubbin, J. Sacks, C.R. Taylor, and D.A.
24          Weinstein.  (2007). Technical Report on Ozone Exposure, Risk, and Impact Assessments
25          for  Vegetation. EPA 452/R-07-002.

26    Lesser, VM; Rawlings, JO; Spruill, SE; Somerville, MC. (1990). Ozone effects on agricultural
27          crops:  Statistical methodologies and estimated dose-response relationships. Crop Sci 30:
28          148-155.

29    Little, E.L., Jr. (1971). Atlas of United States trees, volume 1, conifers and important hardwoods:
30          Misc. Pub.  1146. Washington, D.C.: U.S. Department of Agriculture. 9 p., 200 maps.

31    Little, E.L., Jr. (1976). Atlas of United States Trees, Volume 3, Minor Western Hardwoods. U.S.
32          Department of Agriculture Miscellaneous Publication 1314, 13 p., 290 maps.
                                                    9-7

-------
 1   Little, E.L., Jr. (1977). Atlas of United States Trees, Volume 4, Minor Eastern Hardwoods. U.S.
 2          Department of Agriculture Miscellaneous Publication 1342, 17 p., 230 maps.

 3   Little, E.L., Jr. (1978). Atlas of United States Trees, Volume 5, Florida. U.S. Department of
 4          Agriculture Miscellaneous Publication 1361, 262 maps.

 5   Mansfield, C.A., S.K. Pattanayak, W. McDow, R. MacDonald, and P. Halpin. (2005). Shades of
 6          Green: Measuring the Value of Urban Forests in the Housing Market. Journal of Forest
 7          Economics 11(3): 177-199.

 8   McLaughlin, SB; Nosal, M; Wullschleger, SD; Sun, G. (2007a). Interactive effects of ozone and
 9          climate on tree growth and water use in a southern Appalachian forest in the USA. New
10          Phytol 174: 109-124.

11   MEA (Millennium Ecosystem Assessment Board). (2005). Ecosystems and Human Well-being:
12          Current State and Trends, Volume 1. Edited by R. Hassan, R. Scholes, and N. Ash.
13          Washington: Island Press. Available at http://www.millenniumassessment.Org/8
14          documents/document. 766. aspx. pdf

15   Minnesota EVIPLAN Group (MIG, Inc). (1999). EVIPLAN System (data and software), 502 2nd
16          Street, Suite 301, Hudson, WI 54016. Available at www.implan.com

17   Morgan, PB; Bernacchi, CJ; Ort, DR; Long, SP. (2004). An in vivo analysis of the effect of
18          season-long open-air elevation of ozone to anticipated 2050 levels on photosynthesis in
19          soybean.  J Plant Physiol 135: 2348-2357.

20   Morgan, PB; Mies, TA; Bollero, GA; Nelson, RL; Long, SP. (2006). Season-long elevation of
21          ozone concentration to projected 2050 levels under fully open-air conditions substantially
22          decreases the growth and production of soybean. New Phytol 170: 333-343.

23   Mueller, J., J. Loomis, and A. Gonzalez-Caban. (2007). Do Repeated Wildfires Change
24          Homebuyers' Demand for Homes in High-Risk Areas? A Hedonic Analysis of the Short
25          and Long-Term Effects of Repeated Wildfires on House Prices in Southern California.
26          Journal of Real Estate Finance and Economics,  1-18.

27   National Association of State Foresters (NASF). (2009). Quadrennial Fire Review 2009.
28          Washington, DC: NASF. Quadrennial Fire and Fuel Review Final Report 2009.National
29          Wildfire Coordinating Group Executive Board January 2009.

30   National Climatic Data Center (NCDC). (2012a).U.S. Climatological Divisions. National
31          Oceanic and Atmospheric Administration. Available at http://www.ncdc.noaa.gov/temp-
32          and-precip/us-climate-divisions.php
                                                    9-8

-------
 1   National Climatic Data Center (NCDC). (2012b). Historical Palmer Drought Indices. National
 2          Oceanic and Atmospheric Administration. Available at
 3          http://wwwl.ncdc.noaa.gov/pub/data/cirs/

 4   National Climatic Data Center (NCDC). (2012c). Drought. National Oceanic and Atmospheric
 5          Administration. Available at http://www.ncdc.noaa.gov/sotc/drought/

 6   National Interagency Fire Center (NIFC). (2012). Statistics. Available at
 7          http://www.nifc. gov/firelnfo/firelnfo_stati sties, html

 8   National Park Service (NFS). (2002a). Comprehensive Survey of the American Public, Southeast
 9          Region Technical Report. Available at
10          http://www.nature.nps.gov/socialscience/docs/archive/SER  Tech Rep.pdf

11    National Park Service (NPS). (2002b). Comprehensive Survey of the American Public,
12          Intermountain Region Technical Report. Available at
13          http://www.nature.nps.gov/socialscience/docs/archive/IMR_Tech_Rep.pdf

14   National Park Service (NPS). (2002c). Comprehensive Survey of the American Public Pacific
15          West Region Technical Report. Available at
16          http://www.nature.nps.gOv/socialscience/docs/archive/P WR_Tech_Rep.pdf

17   National Park Service (NPS). (2003). Ozone Sensitive Plant Species on National Park Service
18          and U.S. Fish and Wildlife Service Lands: Results of a June 24-25, 2003 Workshop.
19          Baltimore, MD. Available at http://www.nature.nps.gov/air/pubs/pdf/baltfmalreportl.pdf

20   National Park Service (NPS). (2006b). Ozone Sensitive Plant  Species, by Park, November 2006.
21          Available at
22          http://www.nature.nps.gov/air/Permits/ARIS/docs/Ozone_Sensitive_ByPark_3600.pdf

23   National Park Service (NPS). (2010). Visitor Use Statistics: Annual Recreation Visitation by
24          Park Type or Region for: 2010. Available at https://irma.nps.gov

25   National Park Service (NPS). (2011). Economic benefits to Local Communities from National
26          Park Visitation and Payroll, 2010. Natural Resources Report NPS/NRSS/EGD/NRR-
27          2011/481. Available at
28          http://www.nature.nps.gov/socialscience/docs/NPSSvstemEstimates2009.pdf

29   National Park Service. (NPS). (2006). Management Policies 2006. Available at
30          http://www.nps.gov/policy/MP2006.pdf

31   Oregon State University, College of Forestry. Recreation Use  Values Database. Available at
32          http: //recvaluati on. fore stry. oregonstate. edu/
                                                     9-9

-------
 1    Outdoor Industry Foundation (OIF). (2012). The Economic Contribution of Active Outdoor
 2          Recreation - Technical Report on Methods and Findings. Available at
 3          http://www.outdoorindustry.org/images/ore reports/outdoorrecreationeconomy-
 4          technicalreport2012-oia.pdf

 5    Palmer, W.C. (1965). Meteorological drought. Research Paper No. 45. U.S. Weather Bureau.
 6          NOAA Library and Information Services Division. Washington, D.C.  Available at http://
 7          ncdc.noaa.gov/temp-and-precip/drought/docs/palmer.pdf.

 8    Payton, R. (2007). Effects of Wildfire Smoke on UV Ozone Instruments, Regional Applied
 9          Research Effort, U.S. EPA Region 8, October 2007.
10
11    Pena, D.G. (1999). Cultural landscapes and biodiversity: the ethnoecology of an Upper Rio
12          Grandewatershed commons in Ethnoecology: Situated Knowledge/Located Lives Ed.
13          V.D. Nazarea, University of Arizona Press, Tucson, AZ, pp 107-132.

14    Peterson, DL; Arbaugh, MJ; Wakefield, VA; Miller, PR. (1987). Evidence of growth reduction
15          in ozone-injured Jeffrey pine (Pinus jeffreyi Grev and Balf) in Sequoia and Kings
16          Canyon National Parks. J Air Waste Manag Assoc 37: 906-912.

17    Pregitzer, KS; Burton, AJ; King, JS; Zak, DR. (2008). Soil respiration, root biomass, and root
18          turnover following long-term exposure of northern forests to elevated atmospheric Co-2
19          and tropospheric O-3. New Phytol 180: 153-161.

20    Rawlings, JO; Cure, WW. (1985). The Weibull function as  a dose-response model to describe
21          ozone effects on crop yields. Crop Sci 25: 807-814.

22    Rea et al,  (2012). Using Ecosystem Services To Inform Decisions  on U.S. Air Quality Standards.
23          Environ.  Sci. Technol.46 (12), 6481-6488.

24    Ribe, R. G. (1989). The Aesthetics of Forestry: What has empirical preference research taught
25          us? Environmental Management 13: 55-74.

26    Ribe, R.G. (1994). Scenic Beauty along the ROS. Journal of Environmental Research 42:199-
27          221.

28    Richardson, L.; Loomis, J. (2009). The total economic value of threatened, endangered and rare
29          specis: and updated meta-analysis. Ecological Economics,  68:1535-1548.

30    Rizzo, M. (2005). A Comparison of Different Rollback Methodologies Applied to Ozone
31          Concentrations. November 7, 2005. Available at
32          http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3  cr td.html
                                                    9-10

-------
 1   Rizzo, M. (2006). A Distributional Comparison between Different Rollback Methodologies
 2          Applied to Ambient Ozone Concentrations. May 31, 2006. Available at
 3          http://www.epa.gOv/ttn/naaqs/standards/ozone/s  o3crtd.html

 4   SAB CVPESS (Science Advisory Board Committee on Valuing the Protection of Ecological
 5          Systems and Services). (2009). Valuing the Protection of Ecological Systems and
 6          Services. EPA-SAB-09-012. May. Available at
 7          http://vosemite.epa.gov/sab/sabproduct.nsf/WebBOARD/SAB-09-
 8          012/$File/SAB%20Advisorv%20Report%20full%20web.pdf

 9   Samet. JM. (2009).  Consultation on Ambient Air Monitoring Issues Related to the Ozone
10          NAAQS.  EPA-CASAC-09-005. March. Available at
11          http://vosemite.epa.gov/sab/sabproduct.nsf/64B88B99C37A68CF852575710072D8CO/$
12          File/EP A-CASAC-09-005-unsigned.pdf

13   Samet. JM. (2011).  Consultation on EPA's Ozone National Ambient Air Quality Standards:
14          Scope and Methods Plan for Health Risk and Exposure Assessment (April 2011) and
15          Ozone National Ambient Air Quality Standards: Scope and Methods Plan for Welfare
16          Risk and Exposure Assessment (April 2011). EPA-CASAC-11-008. June.  Available at
17          yosemite.epa.gov/sab/sabproduct.nsf/0594FCC1374FCC5D852578B60069CDB2/$File/E
18          P A-C AS AC-11 -008-unsigned.pdf

19   Samuelson, L. J. &  Edwards, G. S. (1993). A comparison of sensitivity to ozone in seedlings and
20          trees of Quercus rubra L. New Phytol, 125, 373-9.

21   Sanz, J.; Muntifering, R.B.; Bermejo,  V.; Gimeno, B.S.; Elvira, S. (2005). Ozone and increased
22          nitrogen supply effects on the yield and nutritive quality of Trifolium subterraneum.
23          Atmospheric Environment 39: 5899-5907.

24   Schaub, M; Skelly,  JM;  Zhang, JW; Ferdinand, JA; Savage, JE; Stevenson, RE; Davis, DD;
25          Steiner, KC. (2005). Physiological and foliar symptom response in the  crowns of Prunus
26          serotina, Fraxinus americana and Acer rubrum canopy trees to ambient ozone under
27          forest conditions. Environ Pollut 133: 553-567.

28   Sitch, S; Cox, PM;  Collins, WJ; Huntingford, C. (2007). Indirect radiative forcing of climate
29          change through ozone effects on the land-carbon sink. Nature 448: 791-794.

30   Smith, G., J. Coulston, and B. O'Connell. (2008). Ozone bioindicators and forest health: A guide
31          to the evaluation, analysis, and interpretation of the ozone injury data in the Forest
32          Inventory and Analysis program. Gen. Tech. Rep. NRS-34. Newtown Square, PA: U.S.
33          Department of Agriculture, Forest Service, Northern Research Station.  Available at
34          http://www.treesearch.fs.fed.us/pubs/19036.
                                                   9-11

-------
 1    Smith, V.K., C. Poulos, and H. Kim. (2002). Treating Open Space as an Urban Amenity.
 2          Resource and Energy Economics 24:107-129.

 3    Somers, GL; Chappelka, AH; Rosseau, P; Renfro, JR. (1998). Empirical evidence of growth
 4          decline related to visible ozone injury. For Ecol Manage 104: 129-137.

 5    Southwick Associates. (2012). The Economic Contribution of Active Outdoor Recreation -
 6          Technical Report on Methods and Findings For: Outdoor Industry Foundation. Available
 7          at
 8          http://www.outdoorfoundation.org/pdf/ResearchRecreationEconomyTechnicalReport.pdf

 9    Taylor, C.R., K.H. Reichelderfer and S.R. Johnson. (1993). Agricultural Sector Models for the
10          United States: Descriptions and Selected Policy Applications. Iowa State University
11          Press, Ames, IA, 1993.

12    Timin B, Wesson K, Thurman J. Application of Model and Ambient Data Fusion Techniques to
13          Predict Current and Future Year PM2.5 Concentrations in Unmonitored Areas. (2010). Pp.
14          175-179 in Steyn DG, Rao St (eds). Air Pollution Modeling and Its Application XX.
15          Netherlands: Springer.

16    Tyrvainen, L., and A. Miettinen. (2000). "Property Prices and Urban Forest Amenities." Journal
17          of Economics and Environmental Management 39:205-223.

18    U.S. Census Bureau. 2006. County Business Patterns. Available at
19          http ://www. census. gov/econ/cbp/

20    U.S. Department of Agriculture (USD A). (2004). National Report on Sustainable Forests-2003
21          Documentation for the 2003 National Report: Indicators 35, 36, 37, 42, & 43 Available at
22          http ://warnell .forestry.uga. edu/nrrt/NSRE/MontreallndDoc.PDF

23    U.S. Department of Agriculture (USD A). (2010). National Survey on Recreation and the
24          Environment (NSRE): 2000-2002. The Interagency National Survey Consortium,
25          Coordinated by the USDA Forest Service, Recreation, Wilderness, and Demographics
26          Trends Research Group, Athens,  GA and the Human Dimensions Research Laboratory,
27          University of Tennessee, Knoxville, TN.

28    U.S. Department of Agriculture (USDA). (2011). National Report on Sustainable Forests-2010
29          Available at http://www.fs.fed.us/research/sustain/

30    U.S. Department of Agriculture National Resources Conservation Service (USDA-NRCS).
31          (2013).  Fact Sheets & Plant Guides. Available at http://plants.usda.gov/java/factSheet
                                                    9-12

-------
 1   U.S. Department of the Interior, U.S. Fish and Wildlife Service, and U.S. Department of
 2          Commerce, U.S. Census Bureau. (2012). 2011 National Survey of Fishing, Hunting, and
 3          Wildlife-Associated Recreation. Available at
 4          http ://www. census. gov/prod/2012pub s/fhw 11 -nat.pdf

 5   U.S. EPA (U.S. Environmental Protection Agency). (1978). Air quality criteria for ozone and
 6          other photochemical oxidants (pp. 373). (EPA/600/8-78/004). Washington, DC.

 7   U.S. EPA (U.S. Environmental Protection Agency). (1984). Air quality criteria for ozone and
 8          other photochemical oxidants, Vol. 3 (pp. 405). (EPA/600/8-84/020A). Research Triangle
 9          Park, NC.

10   U.S. EPA (U.S. Environmental Protection Agency). (1996). Review of national ambient air
11          quality standards for ozone: Assessment of scientific and technical information:  OAQPS
12          staff paper. (EPA/452/R-96/007). Research Triangle Park, NC.

13   U.S. EPA (U.S. Environmental Protection Agency). (2006). Air quality criteria for ozone and
14          related photochemical oxidants (pp.  2118). (EPA/600/R-05/004AF). Research Triangle
15          Park,NC.

16   U.S. EPA (U.S. Environmental Protection Agency). (2007). Review of the national ambient air
17          quality standards for ozone: Policy assessment of scientific and technical information:
18          OAQPS staff paper. (EPA/452/R-07/007). July. Research Triangle Park, NC. Available at
19          http ://www. epa. gov/ttn/naaqs/standards/ozone/data/200701 ozone staff_paper. pdf

20    U.S. EPA (U.S. Environmental Protection Agency). (2009). The Risk and Exposure Assessment
21          for Review of the  Secondary National Ambient Air Quality  Standards for Oxides of
22          Nitrogen and Oxides of Sulfur. (EPA-452/R-09-008a). September. Available at
23          http://www.epa.gov/ttn/naaqs/standards/no2so2sec/data/NOxSOxREASep2009MainCont
24          ent.pdf

25   U.S. EPA (U.S. Environmental Protection Agency). (2009a). Integrated Review Plan for the
26          Ozone National Ambient Air Quality Standards - External Review Draft.  (EPA 452/D-
27          09-001). September.  Available at
28          http ://www. epa. gov/ttn/naaqs/standards/ozone/data/externalreviewdraftO3 IRP093 009. pdf

29   U.S. EPA (U.S. Environmental Protection Agency). (201 la). Integrated Review Plan for the
30          Ozone National Ambient Air Quality Standards - Final. (EPA 452/R-l 1-006). April.
31          Available at http://www.epa.gov/ttn/naaqs/standards/ozone/data/201 l_04_OzoneIRP.pdf.
                                                    9-13

-------
 1   U.S. EPA (U.S. Environmental Protection Agency). (201 Ib). Ozone National Ambient Air
 2          Quality Standards: Scope and Methods Plan for Health Risk and Exposure. (EPA-452/P-
 3          11-001). Research Triangle Park, NC. April. Available at
 4          http://www.epa.gov/ttn/naaqs/standards/ozone/data/2011 04 HealthREA.pdf

 5   U.S. EPA (U.S. Environmental Protection Agency). (201 Ic). Ozone National Ambient Air
 6          Quality Standards: Scope and Methods Plan for Welfare Risk and Exposure. (EPA 452/P-
 7          11-002). April. Research Triangle Park, NC. Available at
 8          http://www.epa.gov/ttn/naaqs/standards/ozone/data/20 ll_04_WelfareREA.pdf

 9   U.S. EPA (U.S. Environmental Protection Agency). (201 le). Policy Assessment for the Review
10          of the Secondary National Ambient Ai Quality Standards for Oxides of Nitrogen and
11          Oxides of Sulfur. (EPA/452/R-1 l-005a). February. Research Triangle Park, NC.
12          Available at
13          http://vosemite.epa.gOv/sab/sabproduct.nsf/0/A56303ClB3FC092B852577EC004DA8B
14          F/$File/NOx+SOx+PA+FINAL-main.pdf

15   U.S. EPA. (U.S.  Environmental Protection Agency). (2013). Integrated Science Assessment of
16          Ozone and Related Photochemical Oxidants (Final Report). U.S. Environmental
17          Protection Agency, Washington, DC, EPA/600/R-10/076F. Available at
18          http://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=247492#Download

19   U.S. EPA. (U.S.  Environmental Protection Agency). (2013b). Greenhouse Gas Equivalencies
20          Calculator. Updated September 2013. Available at
21          http://www.epa.gov/cleanenergy/energy-resources/calculator.html

22   U.S. Geological  Survey (USGS). (2003). FEDLANP020 - Federal Lands and Indian
23          Reservations of the United States. Reston, VA. Available at
24          http://coastalmap.marine.usgs.gov/GISdata/basemaps/boundaries/fedlands/fedlanp020.zip

25   USFS (U.S. Forest Service). (2011). Forest Health Monitoring Network.

26   USFS (U.S. Forest Service). (2013). Personal email communication with John Coulston, USFS,
27          June 2013.

28   USFS (U.S. Forest Service). Forest Health Technology Enterprise Team. Available at
29          http://www.fs.fed.us/foresthealth/technology/

30   Vollenweider, P; Woodcock, H; Kelty, MJ; Hofer, R, -M. (2003). Reduction of stem growth and
31          site dependency of leaf injury in Massachusetts black cherries exhibiting ozone
32          symptoms. Environ Pollut 125: 467-480.
                                                    9-14

-------
 1   Wegman, L. (2012).  Updates to information presented in the Scope and Methods Plans for the
 2          O3 NAAQS Health and Welfare Risk and Exposure Assessments.  Memorandum from
 3          Lydia Wegman, Division Director, Health and Environmental Impacts Division, Office
 4          of Air Quality Planning and Standards, Office of Air and Radiation, US EPA to Holly
 5          Stallworth, Designated Federal Officer, Clean Air Scientific Advisory Committee, US
 6          EPA Science  Advisory Board Staff Office. May 2, 2012. Available at
 7          www.epa.gov/ttn/naaqs/standards/ozone/data/20120502rea.pdf

 8   Weinstein, D.A., R.M. Beloin and R.D. Yanai. (1991). Modeling changes in red spruce carbon
 9          balance and allocation in response to interacting ozone and nutrient stresses. Tree
10          Physiol. 9:127-146.

11   Weinstein, DA; Gollands, B; Retzlaff, WA. (2001). The effects of ozone on a lower slope forest
12          of the Great Smoky Mountain National Park: Simulations linking an individual tree
13          model to a stand model. Forest Sci 47: 29-42.

14   Wells, B., Wesson, K., Jenkins, S. (2012). Analysis of Recent U.S.  Ozone Air Quality Data to
15          Support the O3 NAAQS Review and Quadratic Rollback Simulations to Support the First
16          Draft of the Risk and Exposure Assessment. Available at
17          http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3 2008 rea.html

18   West, J. I, A. M.  Fiore, V. Naik, L. W. Horowitz, M. D. Schwarzkopf, D. L. Mauzerall (2007).
19          Ozone air  quality and radiative forcing consequences of changes in ozone precursor
20          emissions. Geophys Res Lett, 34, L06806.
                                                    9-15

-------
United States                          Office of Air Quality Planning and Standards       Publication No. EPA-452/P-14-003 a
Environmental Protection               Health and Environmental Impacts Division                          February 2014
Agency                                      Research Triangle Park, NC

-------