&EPA
United States
EnviroimnU Protection
Agency
Review of the National Ambient Air Quality

Standards for Lead:


Policy Assessment of Scientific and Technical
Information


OAQPS Staff Paper - First Draft

-------
                                                 EPA-452/P-06-002
                                                    December 2006
Review of the National Ambient Air Quality Standards
for Lead:

Policy Assessment of Scientific
and Technical Information

OAQPS Staff Paper - First Draft
                  U.S. Environmental Protection Agency
                Office of Air Quality Planning and Standards
                  Research Triangle Park, North Carolina

-------
       DISCLAIMER


       This document has been reviewed by the Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency (EPA), and approved for publication. This first draft
document contains staff views and does not necessarily represent those of the EPA. Mention of
trade names or commercial products is not intended to constitute endorsement or
recommendation for use.

-------
PREFACE

       This document is part of the Environmental Protections Agency's (EPA's) review of the
National Ambient Air Quality Standards (NAAQS) for lead. Based on the information contained
in the Agency's Air Quality Criteria Document for Lead (October 2006; available at
http://www.epa.gOv/ttn/naaqs/standards/pb/s jb_cr_cd.html), this draft Staff Paper includes
assessments and preliminary analyses related to:
    1.  air quality characterization,
    2.  integration and evaluation of health information,
    3.  human exposure analysis and health risk assessment, and
    4.  evaluation and analysis of information on vegetation damage and other welfare effects.
This initial draft document does not include any conclusions or recommendations with regard to
potential retention or revision of the lead NAAQS.

       To date, the lead NAAQS review has followed our historic approach to reviewing
NAAQS, including issuance of a criteria document and a first draft staff paper.  The Agency is
now moving forward to implement a new, more efficient process for conducting NAAQS
reviews (http://www.epa.gov/ttn/naaqs/).  EPA intends to transition to that new process during
the course of the lead NAAQS review.

-------
                              Table of Contents

List of Tables	vii

List of Figures	ix

1   INTRODUCTION	1-1
    1.1    PURPOSE	1-1
    1.2    BACKGROUND	1-2
       1.2.1      Legislative Requirements	1-2
       1.2.2      History of Lead NAAQS Reviews	1-4
       1.2.3      Current Lead NAAQS Review	1-4
    1.3 GENERAL APPROACH AND ORGANIZATION OF THE DOCUMENT	1-6
    REFERENCES   	1-8

2   CHARACTERIZATION OF AMBIENT LEAD	2-1
    2.1    INTRODUCTION/BACKGROUND	2-1
    2.2    PROPERTIES OF AMBIENT LEAD	2-3
       2.2.1      Fate and Transport of Pb Particles	2-3
    2.3    SOURCES AND EMISSIONS TO THE ATMOSPHERE	2-5
       2.3.1      Data sources	2-5
       2.3.2      Confidence Level for Emission Estimates	2-6
       2.3.3      Trends in National Emissions: 1980 to 2002	2-7
       2.3.4      Source Categories with Largest National Total Pb Emissions	2-8
          23.4.1 Industrial/Commercial/Institutional/Process Heaters	2-10
          2.3.4.2 Utility Boilers	2-10
          2.3.4.3 Mobile Sources	2-10
          2.3.4.4 Iron and Steel Foundries	2-11
          2.3.4.5 Hazardous Waste Incineration/Combustion Facilities	2-12
          2.3.4.6 Primary Lead Smelting	2-12
          2.3.4.7 Secondary Lead Smelting	2-13
          2.3.4.8 Military Installations	2-13
          2.3.4.9 Mining	2-13
          2.3.4.10  Integrated Iron & Steel Manufacturing	2-13
          2.3.4.11  Municipal Waste Combustors:  Small & Large	2-14
          2.3.4.12  Pressed and Blown Glass and Glassware Manufacturing	2-14
          2.3.4.13  Electric Arc Furnace Steelmaking	2-15
          2.3.4.14  Lead Acid Battery Manufacturing	2-15
December 2006                         i             Draft - Do Not Quote or Cite

-------
          2.3.4.15 Primary Copper Smelting	2-15
          2.3.4.16 Portland Cement Manufacturing	2-16
       2.3.5      Geographic Distribution of Sources	2-16
          2.3.5.1 National Patterns in the Distribution of Lead Emissions	2-16
          2.3.5.2 Largest Pb Point Sources in the 2002 NEI	2-19
   2.4    AMBIENT AIR CONCENTRATIONS	2-20
       2.4.1      Ambient Pb Measurement Methods	2-21
          2.4.1.1 Sampling Frequency	2-21
          2.4.1.2 Inlet Design	2-21
          2.4.1.3 Sample Analysis	2-22
       2.4.2      Pb Monitoring Programs	2-22
          2.4.2.1 NAAQS Compliance Network	2-23
          2.4.2.2 PM2.5  Speciation Trends Network	2-25
          2.4.2.3 IMPROVE Network - PM2.5  Speciation	2-26
          2.4.2.4 National Air Toxics Trends Stations - PMio Speciation	2-27
       2.4.3      Ambient Pb Concentrations, Trends and Spatial Patterns	2-27
          2.4.3.1 PbinTSP	2-28
          2.4.3.2 PbinPM2.5	2-37
          2.4.3.3 PbinPMio	2-40
       2.4.4      Relationships among Different Particle-sized Pb Concentration	2-43
       2.4.5      Modeling Estimates (NATA- National  Scale Assessments)	2-45
          2.4.5.1 Methods	2-45
          2.4.5.2 Findings and Limitations	2-46
       2.4.6      Air Quality Summary	2-47
   2.5    ATMOSPHERIC DEPOSITION	2-48
       2.5.1      Temporal Trends	2-48
       2.5.2      Deposition Flux Estimates since the Last Review	2-49
   2.6    OUTDOOR DUST AND SOIL	2-49
       2.6.1      Outdoor Dust	2-49
       2.6.2      Soil	2-50
          2.6.2.1 Temporal Trends	2-50
          2.6.2.2 Current Surface Soil Concentrations	2-52
   2.7    SURFACE WATER AND SEDIMENT	2-53
       2.7.1      Temporal Trends	2-53
       2.7.2      Current Concentrations	2-55
   REFERENCES   	2-58
December 2006                         ii            Draft - Do Not Quote or Cite

-------
3  POLICY-RELEVANT ASSESSMENT OF HEALTH EFFECTS EVIDENCE	3-1
   3.1    INTRODUCTION	3-1
   3.2    INTERNAL DISPOSITION - BLOOD LEAD AS DOSE METRIC	3-2
   3.3    NATURE OF EFFECTS	3-6
      3.3.1      Developing Nervous System	3-10
          3.3.1.1 Endpoint for risk quantitation	3-11
          3.3.1.2 Metric and quantitative model for risk quantitation	3-12
      3.3.2      Cardiovascular System	3-15
      3.3.3      Heme Synthesis	3-16
      3.3.4      Renal System	3-16
      3.3.5      Immune System	3-17
      3.3.6      Adult Nervous System	3-17
   3.4    LEAD-RELATED IMPACTS ON PUBLIC HEALTH	3-19
      3.4.1      Potentially Susceptible or Vulnerable Subpopulations	3-19
      3.4.2      Potential Public Health Impact	3-21
   3.5    SUMMARY AND CONCLUSIONS	3-23
   REFERENCES   	3-26

4     CHARACTERIZATION OF HEALTH RISKS	4-1
   4.1    INTRODUCTION	4-1
      4.1.1      Overview of Risk Assessment from Last Review	4-2
   4.2    SCOPE OF PB EXPOSURE AND RISK ASSESSMENTS	4-4
      4.2.1      Conceptual Model for Human Health Risk Assessment	4-4
      4.2.2      Selection of Health Endpoint, Study Population, Dose-Metrics and
                Associated Concentration-Response Function	4-7
      4.2.3      Selection of Case Study Locations	4-7
          4.2.3.1 Primary Pb Smelter Case Study	4-8
          4.2.3.2 Additional Point Source (Secondary Smelter) Case Study	4-10
          4.2.3.3 Near Roadway (Urban) Location Case Study	4-10
      4.2.4      Air Quality Scenarios Covered in the Pilot Analysis	4-12
      4.2.5      Overview of the Exposure and Risk Modeling Approach Used in the
                Pilot Analysis	4-12
          4.2.5.1 CASAC Consultation Regarding Human Exposure and Health Risk
                Assessment	4-13
          4.2.5.2 Child Study Population	4-13
          4.2.5.3 Timeframe for Current Conditions	4-14
          4.2.5.4 Spatial Scale and Resolution	4-15
          4.2.5.5 Overview of Analytical Steps	4-16

December 2006                       in            Draft - Do Not Quote or Cite

-------
          4.2.5.6 Performance Evaluation	4-21
          4.2.5.7 Uncertainty/Sensitivity Analysis	4-21
   4.3    HUMAN EXPOSURE ASSESSMENT	4-23
       4.3.1      Spatial Templates	4-23
          4.3.1.1 Primary Pb Smelter Case Study	4-23
          4.3.1.2 Secondary Pb Smelter Case Study	4-24
          4.3.1.3 Near Roadway (Urban) Case Study	4-24
       4.3.2      Methods for Estimating Media Concentrations	4-25
          4.3.2.1 Ambient Air Concentrations	4-27
          4.3.2.2 Inhalation Exposure Concentrations	4-28
          4.3.2.3 Outdoor Soil Concentrations	4-29
          4.3.2.4 Indoor Dust Concentrations	4-33
       4.3.3      Methods for Estimating Blood Pb Levels	4-36
          4.3.3.1 Blood Pb Models	4-37
          4.3.3.2 Model Inputs	4-39
          4.3.3.3 Probabilistic Population Blood Pb Modeling Procedure	4-41
          4.3.3.4 GSD for Population Blood Pb Levels	4-45
       4.3.4      Projected Media Concentrations	4-46
       4.3.5      Projected Blood Pb Levels	4-48
       4.3.6      Performance Evaluation	4-53
          4.3.6.1 Media Concentrations	4-53
          4.3.6.2 Blood Pb Levels	4-55
   4.4    HEALTH RISK ASSESSMENT	4-58
       4.4.1      Method for Risk Characterization	4-58
          4.4.1.1 Concentration Response Function	4-58
          4.4.1.2 Derivation of Cutpoint	4-59
          4.4.1.3 Project! on of Population Risk	4-59
       4.4.2      Risk Estimates	4-60
          4.4.2.1 Primary Pb  Smelter Case Study	4-61
          4.4.2.2 Secondary Pb Smelter Case Study	4-64
          4.4.2.3 Near Roadway (Urban) Case Study	4-67
       4.4.3      Uncertainly Analysis (Sensitivity Analysis, Performance Evaluation
                 and Other Considerations)	4-69
          4.4.3.1 Sensitivity Analysis Methodology	4-69
          4.4.3.2 Sensitivity Analysis Results	4-73
          4.4.3.3 Additional Considerations	4-76
   4.5 SUMMARY OF FINDINGS AND CONSIDERATIONS FOR THE FULL-SCALE
       ASSESSMENT	4-78

December 2006                         iv            Draft - Do Not Quote or Cite

-------
      4.5.1      Summary of Findings in the Pilot Assessment	4-79
      4.5.2      Potential Areas for Enhancement in the Full-Scale Analysis	4-80
   REFERENCES   	4-82

5  THE PRIMARY LEAD NAAQS	5-1
   5.1    INTRODUCTION	5-1
   5.2    APPROACH	5-1
   5.3    PRIMARY LEAD STANDARD	5-3
      5.3.1      Level	5-3
          5.3.1.1 Sensitive Population	5-4
          5.3.1.2 Maximum Safe Blood Level	5-5
          5.3.1.3 Nonair Contribution	5-6
          5.3.1.4 Air Pb Level	5-6
          5.3.1.5 Margin of Safety	5-7
      5.3.2      Averaging Time, Form, and Indicator	5-7
   5.4    POLICY OPTIONS CONSIDERED IN THE LAST REVIEW	5-8
   REFERENCES   	5-11

6  POLICY RELEVANT ASSESSMENT OF WELFARE EFFECTS	6-1
   6.1    INTRODUCTION	6-1
   6.2    EFFECTS IN TERRESTRIAL ECOSYSTEMS	6-1
      6.2.1      Pathways of Exposure	6-3
      6.2.2      Effects of Lead on Energy Flow and Biogeocycling	6-3
      6.2.3      Tools for Identifying Ecotoxicity in Terrestrial Organisms	6-4
      6.2.4      Effects on Plants	6-5
      6.2.5      Effects on Birds and Mammals	6-6
      6.2.6      Effects on Decomposers and Soil Invertebrates	6-7
      6.2.7      Summary	6-7
   6.3    EFFECTS IN AQUATIC ECOSYSTEMS	6-8
      6.3.1      Tools for Identifying Ecotoxicity in Aquatic Organisms	6-8
      6.3.2      Effects in Marine/Estuarine Ecosystems	6-9
          6.3.2.1 Pathways of Exposure	6-9
          6.3.2.2 Effects on Organisms and Communities	6-9
      6.3.3      Effects in Freshwater Ecosystems	6-10
          6.3.3.1 Pathways of Exposure	6-10
          6.3.3.2 Effects at an Ecosystem Level	6-10
          6.3.3.3 Effects on Algae and Aquatic Plants	6-12
          6.3.3.4 Effects on Invertebrates	6-13

December 2006                        v             Draft - Do Not Quote or Cite

-------
          6.3.3.5 Effects on Fish and Waterfowl	6-14
       6.3.4     Summary	6-15
   6.4    SCREENING LEVEL RISK ASSESSMENT	6-15
       6.4.1     Overview of Analyses	6-15
       6.4.2     Measures of Exposure and Effect	6-19
       6.4.3     National-Scale Screen and Case Studies	6-20
          6.4.3.1 National Scale Screen	6-20
          6.4.3.2 Ecologically Vulnerable Location	6-21
          6.4.3.3 Primary Pb Smelter	6-22
          6.4.3.4 Secondary Pb  Smelter	6-23
          6.4.3.5 Near Roadway Non-Urban Case Study	6-23
       6.4.4     Screening Values	6-24
          6.4.4.1 Soil Screening Values	6-24
          6.4.4.2 Surface Water Screening Values	6-25
          6.4.4.3 Sediment Screening Values	6-25
       6.4.5     Results for Case Study Locations and Comparison to Screening Levels.... 6-25
          6.4.5.1 National-Scale Surface Water Screen	6-25
          6.4.5.2 National-Scale Sediment Screen	6-27
          6.4.5.3 Primary Pb Smelter	6-29
          6.4.5.4 Secondary Smelter	6-32
          6.4.5.5 Near Roadway Non-Urban Case Study	6-33
       6.4.6     Discussion	6-33
       6.4.7     Uncertainty and Variability	6-33
          6.4.7.1 Primary Pb Smelter Case Study	6-34
          6.4.7.2 Secondary Pb  Smelter Case Study	6-35
          6.4.7.3 Near Roadway Non-Urban Case Study	6-35
          6.4.7.4 National-Scale Surface Water Screen	6-35
          6.4.7.5 National-Scale Sediment Screen	6-36
   6.5    FUTURE ANALYSES	6-37
   6.6    THE SECONDARY LEAD NAAQS	6-37
       6.6.1     Introduction	6-37
       6.6.2     Approach	6-38
   REFERENCES   	6-40
December 2006                         vi             Draft - Do Not Quote or Cite

-------
                                    List of Tables

Table 2-1. Distribution of point sources within the 2002 NEI and associated estimated
           emissions	2-6
Table 2-2. Trend in Pb emissions (tpy) from 1980 to 2002	2-8
Table 2-3. Source categories emitting greater than 5 tpy of Pb in the 2002 NEI	2-9
Table 2-4. Point Sources with Pb emissions in 2002 NEI greater than or equal to 5 tpy	2-20
Table 2-5. Correlation among different TSP site-level statistics, 2003-2005	2-34
Table 2-6. FRM sites with Pb concentrations above the level of the current NAAQS,
          based on maximum quarterly average, 2003-2005	2-34
Table 2-7. Comparison of numbers of sites that exceed various TSP Pb levels using
          different averaging times or forms, 2003-2005	2-37
Table 2-8. Pb concentrations (|ig/m3), at four sites, in different PM size fractions:
          2003-2005	2-44

Table 3-1. Summary of Lowest Observed Effect Levels for Key  Lead-Induced Health Effects
           in Children (CD, Table 8-5)	3-8
Table 3-2.  Summary of Lowest Observed Effect Levels for Key Lead-Induced Health Effects
           in Adults (CD, Table 8-6) 	3-9
Table 3-3. Population subgroups with characteristics that may contribute to increased
          susceptibility or vulnerability to Pb health effects	3-22

Table 4-1. Timeframe (years) reflected in the characterizations of Pb media concentrations
           used in the pilot risk assessment	4-15
Table 4-2. Case study  approaches  for estimating media concentrations	4-26
Table 4-3. IEUBK input parameters and basis or derivation	4-40
Table 4-4. Projected ambient air and inhalation exposure concentrations, a	4-47
Table 4-5. Projected outdoor soil concentrations	4-47
Table 4-6. Projected indoor dust concentrations	4-48
Table 4-7. Projected blood Pb levels (ug/dL) for primary Pb smelter case study	4-51
Table 4-8. Projected blood Pb levels (ug/dL) for secondary Pb smelter case study	4-52
Table 4-9. Projected blood Pb levels (ug/dL) for near roadway (urban) case study	4-52
Table 4-10. Performance evaluation of approaches for ambient air concentrations	4-54
Table 4-11. Performance evaluation of approaches for outdoor soil concentrations	4-55
Table 4-12. Performance evaluation of approaches for blood Pb levels	4-56
Table 4-13. Projections of IQ loss for the primary Pb smelter case study - current conditions.4-62
Table 4-14. Projections of IQ loss for primary Pb smelter case study - NAAQS attainment. ..4-63
December 2006                         vii            Draft - Do Not Quote or Cite

-------
Table 4-15. Projections of IQ loss for secondary Pb smelter case study - modeled soil Pb
           approach	4-65
Table 4-16. Projections of IQ loss for secondary Pb smelter case study -hybrid soil Pb
           approach	4-66
Table 4-17. Projections of IQ loss for near roadway (urban) case study	4-68
Table 4-18. Modeling elements considered in the pilot sensitivity analysis (including
           summary of approaches used to derive alternate approaches/inputs)	4-71
Table 4-19. Summary of Sensitivity Analysis Results	4-74

Table 6-1. Models and Measurements Used for Ecological Risk Screening Assessment	6-18
Table 6-2. Soil Screening Values for Pb for Ecological Receptors	6-25
Table 6-3. Results of Aquatic Risk Screen - Locations at which Dissolved Pb
          Measurements Exceed AWQC, Excluding Mining Sites	6-27
Table 6-4. Concentrations of Total Pb in Sediments at Locations Near or Matching the
           15 Sites  at which Dissolved Pb Concentrations Exceeded the AWQC,
              Excluding Mining Sites	6-28
Table 6-5. HQs for Soils for Primary Pb Smelter Case Study	6-30
Table 6-6. HQs Calculated for Surface Waters for Primary Pb Smelter Case Study	6-31
Table 6-7. HQs Calculated for Sediments in Surface Waters for Primary Pb  Smelter
          Case Study	6-32
Table 6-8. HQs Calculated for Soils for Secondary Pb Smelter Case Study.a	6-33
Table 6-9. HQs Calculated for Ssoils Near Roadway Non-Urban Case Study	6-33
December 2006                         viii            Draft - Do Not Quote or Cite

-------
                                   List of Figures

Figure 2-1. Principal pathways of human and ecological exposure to Pb	2-2
Figure 2-2. Emission Density of All Pb Sources in the 2002 NEI	2-17
Figure 2-3. Emission Density of Mobile and Non Point Sources of Pb in 2002 NEI	2-18
Figure 2-4. Geographic distribution of point sources with >1 tpy Pb emissions in 2002 NEI.. 2-19
Figure 2-5. Change in the number of Pb TSP monitoring sites from  1980 to 2005	2-23
Figure 2-6. Pb TSP monitoring sites: 2003-2005	2-24
Figure 2-7. Pb PM2.5 (STN) monitoring sites	2-25
Figure 2-8. Pb PM2.5 (IMPROVE) monitoring sites	2-26
Figure 2-9. Pb PM10 (NATTS) monitoring  sites network	2-27
Figure 2-10. Airborne Pb (TSP) concentrations, averaged across continuously operating
           monitoring sites: 1980-2002	2-28
Figure 2-11. Distribution of TSP Pb concentrations (represented by 4 different statistics)
           at monitoring sites, 2003-2005	2-30
Figure 2-12. Distribution of monitor level TSP Pb annual mean concentrations for
          source-oriented and not sourced-oriented monitors, 2003-2005	2-31
Figure 2-13. Site level TSPPb, annual mean concentrations, 2003-2005	2-32
Figure 2-14. Site level TSP Pb, maximum quarterly mean concentrations, 2003-2005	2-33
Figure 2-15. Monthly average TSP Pb concentrations at 4 example monitor sites, 2003-2005.2-36
Figure 2-16. Distribution of PM2.5 Pb concentrations (represented by four different
          statistics) at  STN sites, 2003-2005	2-38
Figure 2-17. Site level 'urban' (STN) PM2.5 Pb annual means, 2003-2005	2-39
Figure 2-18. Distribution of PM10 Pb concentrations (represented by four different
          statistics), 2003-2005	2-41
Figure 2-19. Monitor level PM10 Pb annual means, 2002-2005	2-42
Figure 2-20. Comparison of national mean and median monitor level Pb, annual means
           for different size cut PM networks, 2002-2005	2-43
Figure 2-21. Modeled soil  concentrations of Pb in the South Coast Air Basin of
          California based on four re-suspension rates (A)	2-52
Figure 2-22. Pb concentrations in sediment samples in 12 Michigan lakes	2-54
Figure 2-23. Spatial distribution of dissolved lead in surface water (N = 3445).
           [CD, Figure AX7-2.2.7.]	2-56
Figure 2-24. Spatial distribution of total lead in bulk sediment <63 |im (N = 1466).
           [CD, Figure AX7-2.2.9]	2-57

Figure 4-1. Conceptual model for Pb human health risk assessment	4-5
Figure 4-2. Overview of analysis approach for the pilot analysis	4-17

December 2006                          ix             Draft - Do Not Quote or Cite

-------
Figure 4-3. Procedure for Generating Population Blood Pb Distributions	4-43





Figure 6-1. Overview of Ecological Screening Assessment	6-17
December 2006                         x             Draft - Do Not Quote or Cite

-------
 1                                     1   INTRODUCTION

 2    1.1   PURPOSE
 3          This first draft Staff Paper, prepared by staff in the U.S. Environmental Protection
 4    Agency's (EPA) Office of Air Quality Planning and Standards (OAQPS) (henceforth referred to
 5    as "staff), contains an initial evaluation of the policy implications of the key studies and
 6    scientific information contained in the final document, Air Quality Criteria for Lead (USEPA,
 7    2006a; henceforth referred to as the CD), prepared by EPA's National Center for Environmental
 8    Assessment (NCEA). This first draft also presents and interprets results from several initial or
 9    pilot analyses (e.g., human exposure analyses, human health risk assessments and environmental
10    assessments) that will inform EPA's current review of the national ambient air quality standards
11    (NAAQS) for lead (Pb). This draft Staff Paper, however, is an initial product intended to
12    provide an initial assessment of the issues for the Clean Air Scientific Advisory Committee
13    (CASAC) and public review and comment and, further, to solicit comment on the pilot exposure
14    and risk assessments, as well as on plans for the full-scale assessments.  This draft document
15    does not present any conclusions or recommendations with regard to potential retention or
16    revision of the primary (health-based) and secondary (welfare-based) Pb NAAQS.
17          The policy assessment to be presented in the final version of this document is intended to
18    help "bridge the gap" between the scientific assessment contained in the CD and the judgments
19    required of the EPA Administrator in determining whether it is appropriate to retain or revise the
20    NAAQS for Pb. In conducting this assessment, staff is aware of the dramatic alteration in the
21    basic patterns of air lead emissions in the U.S. since the listing of Pb as  a criteria pollutant and
22    the 1978 promulgation of the Pb NAAQS. The reduction of Pb in gasoline has resulted in
23    orders-of-magnitude reductions in airborne emissions of Pb, and a significant shift in the types of
24    sources with the greatest Pb emissions. An additional circumstance that has changed since 1978
25    is the enactment in 1990 of the Clean Air Act Amendments, in which Pb compounds were listed
26    as hazardous air pollutants under Section 112.  Section 112, as amended in 1990, requires EPA
27    to establish technology-based (or "MACT") emission standards for those listed source categories
28    emitting Pb compounds, and to establish risk-based standards,  as necessary, for those categories
29    of sources for which EPA has issued MACT standards. Given the significantly changed
30    circumstances since Pb was listed in 1976, we will evaluate the status of Pb as a criteria
            December 2006                      1 -1         Draft - Do Not Quote or Cite

-------
 1    pollutant1 in light of currently available information and assess whether revocation of the
 2    standard is an appropriate option for the Administrator to consider.
 3           In evaluating the adequacy of the current standard and policy alternatives, in the next
 4    draft of this document, emphasis will be placed on identifying those conclusions and
 5    uncertainties in the available scientific literature that are most pertinent to the indicator2,
 6    averaging times, forms3, and levels for primary (health-based) and secondary (welfare-based)
 7    standards, which must be considered collectively in evaluating the health and welfare protection
 8    afforded by Pb standards.  The final version of this document will evaluate the policy
 9    implications of the key studies and scientific information contained in the CD, identify the
10    critical elements to be considered in the  current review of the NAAQS for Pb, and present factors
11    relevant to the evaluation of current primary and secondary Pb NAAQS, as well as conclusions
12    and identification of options for the Administrator' consideration.
13           While this draft Staff Paper should be of use to all parties interested in the Pb NAAQS
14    review, it is written with an expectation that the reader has some familiarity with the technical
15    discussions contained in the CD.

16    1.2  BACKGROUND
17         1.2.1    Legislative Requirements
18           Two sections of the Clean Air Act (Act) govern the establishment and revision of the
19    NAAQS. Section 108 (42 U.S.C. 7408) directs the Administrator to identify and list each air
20    pollutant that "in his judgment, cause  or contribute to air pollution which may reasonably be
21    anticipated to endanger public health and welfare" and whose "presence ... in the ambient air
22    results  from numerous or diverse mobile or stationary sources" and to issue air quality criteria
23    for those that are listed.  Air quality criteria are to "accurately reflect the latest scientific
24    knowledge useful in indicating the kind  and extent of all identifiable effects on public health or
25    welfare which may be expected from the presence of [a] pollutant in ambient air ... "
26           Section  109 (42 U.S.C. 7409) directs the Administrator to propose and promulgate
27    "primary" and "secondary" NAAQS for pollutants listed under section 108. Section 109(b)(l)
28    defines a primary standard as one "the attainment and maintenance of which in the judgment of
29    the Administrator, based on [air quality] criteria and allowing an adequate margin  of safety, are
             1 Section 108 of the Clean Air Act states that the Administrator "shall, from time to time ... revise" the
      criteria pollutant list.
             2 The "indicator" of a standard designates the chemical associated with the standard.
             3 The "form" of a standard defines the air quality statistic that is to be compared to the level of the standard
      in determining whether an area attains the standard.
             December 2006                      1 -2         Draft - Do Not Quote or Cite

-------
 1    requisite to protect the public health."4 A secondary standard, as defined in Section 109(b)(2),
 2    must "specify a level of air quality the attainment and maintenance of which, in the judgment of
 3    the Administrator, based on criteria, is requisite to protect the public welfare from any known or
 4    anticipated adverse effects associated with the presence of [the] pollutant in the ambient air."5
 5           The requirement that primary standards include an adequate margin of safety was
 6    intended to address uncertainties associated with inconclusive scientific and technical
 7    information available at the time of standard setting. It was also intended to provide a reasonable
 8    degree of protection against hazards that research has not yet identified. Lead Industries
 9    Association v. EPA, 647 F.2d 1130,  1154 (D.C.  Cir 1980), cert, denied, 449 U.S. 1042 (1980);
10    American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 1981), cert, denied, 455
11    U.S. 1034 (1982). Both kinds of uncertainties are components of the risk associated with
12    pollution at levels below those  at which human health effects can be said to occur with
13    reasonable scientific certainty.  Thus, in selecting primary standards that include an adequate
14    margin of safety, the Administrator is seeking not only to prevent pollution levels that have been
15    demonstrated to be harmful but also to prevent lower pollutant levels that may pose an
16    unacceptable risk of harm, even if the risk is not precisely identified as to nature or degree.
17           In selecting a margin of safety, the EPA considers such factors as the nature and severity
18    of the health effects involved, the size of the sensitive population(s) at risk, and the kind and
19    degree of the uncertainties that must be addressed. The  selection of any particular approach to
20    providing an adequate margin of safety is a policy choice left specifically to the Administrator's
21    judgment. Lead Industries Association v. EPA, supra, 647 F.2d at 1161-62.
22           In setting standards that are "requisite" to protect public health and welfare, as provided
23    in section 109(b), EPA's task is to establish standards that are neither more nor less stringent
24    than necessary for these purposes. In so doing, EPA may  not consider the costs of implementing
25    the standards. See generally Whitman v. American Trucking Associations, 531 U.S. 457, 471,
26    475-76 (2001).
             4 The legislative history of section 109 indicates that a primary standard is to be set at "the maximum
      permissible ambient air level. .  . which will protect the health of any [sensitive] group of the population," and that
      for this purpose "reference should be made to a representative sample of persons comprising the sensitive group
      rather than to a single person in such a group." S. Rep. No. 91-1196, 91st Cong., 2d Sess. 10 (1970)
             5 Welfare effects as defined in section 302(h) (42 U.S.C. 7602(h)) include, but are not limited to, "effects
      on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility and climate, damage to
      and deterioration of property, and hazards to transportation, as well as effects on economic values and on personal
      comfort and well-being."
             December 2006                       1 -3        Draft - Do Not Quote or Cite

-------
 1          Section 109(d)(l) of the Act requires that "not later than December 31, 1980, and at 5-
 2   year intervals thereafter, the Administrator shall complete a thorough review of the criteria
 3   published under section 108 and the national ambient air quality standards . . . and shall make
 4   such revisions in such criteria and standards and promulgate such new standards as may be
 5   appropriate . .  . ." Section 109(d)(2) requires that an independent scientific review committee
 6   "shall complete a review of the criteria . . .  and the national primary and secondary ambient air
 7   quality standards . . . and shall recommend to the Administrator any new . . . standards and
 8   revisions of existing criteria and standards as may be appropriate . . . ." Since the early 1980's,
 9   this independent review function has been performed by the Clean Air Scientific Advisory
10   Committee (CASAC) of EPA's Science Advisory Board.

11         1.2.2  History of Lead NAAQS Reviews
12          On October 5, 1978 the EPA promulgated primary and secondary NAAQS for lead under
13   section 109 of the Act (43 FR 46246).  Both primary and secondary standards were set at a level
14   of 1.5 ug/m3 as a quarterly average (maximum arithmetic mean averaged over a calendar
15   quarter). This standard was based on the 1977 Air Quality Criteria for Lead (USEPA, 1977).
16          The most recent review was initiated in the mid 1980s. The scientific assessment for that
17   review is described in the 1986 Air Quality Criteria for Lead (USEPA, 1986a), the associated
18   Addendum (USEPA, 1986b) and the 1990 Supplement (USEPA, 1990a).  As part of the review,
19   the Agency designed and performed human exposure and health risk analyses (USEPA, 1989),
20   the results for which were presented in the  1990 Staff Paper (USEPA,  1990b).  Based on the
21   scientific assessment and the human exposure and health risk analyses, the 1990 Staff Paper
22   presented options for the Pb NAAQS level in the range of 0.5 to 1.5 |ig/m3, and suggested the
23   second highest monthly average in three years for the form (USEPA,  1990b). After
24   consideration of the documents developed during the review and the significantly  changed
25   circumstances since Pb was listed in 1976,  as noted above, the Agency did not propose any
26   revisions to the 1978 Pb NAAQS.  In a parallel effort, the Agency developed the broad, multi-
27   program, multimedia, integrated U.S. Strategy for Reducing Lead Exposure (USEPA, 1991).  As
28   part of implementing this strategy, the Agency focused efforts primarily on regulatory and
29   remedial clean-up actions aimed at reducing Pb exposures from a variety of non-air sources
30   judged to  pose more extensive public health risks to U.S. populations,  as well as on actions to
31   reduce Pb emissions to air.

32         1.2.3  Current Lead NAAQS Review
33          EPA initiated the review of the air quality criteria for Pb on November 9, 2004 with a
34   general call for information (69 FR 64926). A project work plan (USEPA, 2005a) for the
35   preparation of the CD was released in January 2005  for CASAC and public review.  EPA held a
            December 2006                     1 -4        Draft - Do Not Quote or Cite

-------
 1    series of workshops in August 2005, with invited recognized scientific experts, to discuss initial
 2    draft materials that dealt with various lead-related issues being addressed in the Pb air quality
 3    criteria document. These workshops helped to inform the preparation of the first draft CD
 4    (USEPA, 2005b), which was released for CASAC and public review in December, 2005 and
 5    discussed at the CASAC meeting held on February 28-March 1, 2006.
 6          A second draft CD (USEPA, 2006b) was released for CASAC and public review in May
 7    2006, and discussed at the CASAC meeting on June 28, 2006. A subsequent draft of Chapter 7 -
 8    Integrative Synthesis (Chapter 8 in the final CD), released on July 31, 2006, was discussed at an
 9    August 15, 2006 CASAC teleconference. The final CD was released on September 30, 2006
10    (USEPA, 2006a). While the CD focuses on new scientific information available since the last
11    review, it appropriately integrates that information with scientific criteria from previous reviews.
12          In February, 2006, EPA released the Plan for Review of the National Ambient Air Quality
13    Standards for Lead (USEPA 2006c) that described Agency plans and timeline for reviewing the
14    air quality criteria,  developing human exposure and risk assessments and an ecological risk
15    assessment, preparing a policy assessment, and developing the regulatory proposal and final
16    rulemaking.
17          In May, 2006, EPA released for CASAC and public review a draft Analysis Plan for
18    Human Health and Ecological Risk Assessment for the Review of the Lead National Ambient Air
19    Quality Standards (USEPA, 2006d) which was discussed at the June 29, 2006 CASAC meeting.
20    CASAC panel members' views were received at and subsequent to the meeting (Henderson,
21    2006), and considered in the implementation of the human health and ecological risk
22    assessments, the pilot phase of which is described in this first draft Staff Paper. As  described in
23    the May 2006 plan, the risk assessments are being performed in two phases:  1) pilot and 2) full-
24    scale.  With consideration of CASAC and public comments received on this document and the
25    analyses described within, staff plans to  develop and perform full-scale assessments.6 The full-
26    scale assessments will be presented  in the second draft of this document for public and CASAC
27    review. Based on the scientific and technical findings described therein, the second draft of this
28    document will present initial conclusions and alternative policy options regarding the Pb
29    NAAQS.  Comments received during CASAC and public review of the second draft will be
30    considered in preparation of the  final document.
31          The schedule for completion of this review is governed by a judicial order resolving a
32    lawsuit filed in May 2004, alleging that EPA had failed to complete the current review within the
            6 As discussed in Section 6.1, we do not at this time anticipate having funding to perform additional
      quantitative ecological risk assessment work for this review.
            December 2006                     1 -5        Draft - Do Not Quote or Cite

-------
 1    period provided by statute. Missouri Coalition for the Environment, v. EPA (No. 4:04CV00660
 2    ERW, Sept. 14, 2005). The order that now governs this review, entered by the court on
 3    September 14, 2005, provides that EPA will prepare the initial draft Staff Paper not later than
 4    January 1, 2007, and will finalize it no later than November 1, 2007. The order also specifies
 5    that EPA sign, for publication, notices of proposed and final rulemaking concerning its review of
 6    the Pb NAAQS no later than May 1, 2008 and September 1, 2008, respectively. EPA published
 7    a series of interim target dates in its Plan for Review of the Pb NAAQS (USEPA 2006c) that are
 8    designed to ensure that these deadlines will be met. With regard to the Staff Paper, this includes
 9    release of a second draft document in June 2007, followed by CASAC and public review, and a
10    final document in September 2007. There is also an interim target date for a proposed
11    rulemaking in February 2008.

12    1.3   GENERAL APPROACH AND ORGANIZATION OF THE DOCUMENT
13          The final version of this document will take into account the scientific evidence reviewed
14    in the CD and will include: 1) the results of air quality analyses, human exposure and health risk
15    assessments, and  environmental assessments; 2) an overall evaluation of the adequacy of the
16    current primary and secondary NAAQS; and 3) conclusions pertaining to a range of policy
17    choices available to address public health  and welfare effects associated with exposure to
18    ambient Pb resulting from emissions to the ambient air. This first draft Staff Paper includes
19    discussion of the  scientific evidence reviewed in the CD, as well as preliminary quantitative
20    analyses based  on available emissions and air quality information, information on deposition and
21    distribution of ambient Pb in other media, and estimated health and environmental risks related
22    to exposure to ambient Pb concentrations resulting from Pb emitted into the ambient air.
23          Following this introductory chapter, this draft Staff Paper  is organized into three main
24    parts: the characterization of ambient Pb; Pb-related health effects and primary Pb  NAAQS; and
25    Pb-related welfare effects and secondary Pb NAAQS. The content of these parts is discussed
26    more fully below.
27          The characterization of ambient Pb is presented in Chapter 2 and includes information on
28    Pb properties, current Pb  air quality patterns, historic trends, and background levels. In
29    recognition of the multimedia nature of Pb and the distribution into other media of Pb emitted
30    into the air, Chapter 2 also includes information on Pb in media other than air including outdoor
31    dust, soil, surface water and sediment. This chapter provides a frame of reference for exposure
32    and risk analyses  and subsequent discussion of the Pb NAAQS and alternative forms of Pb
33    standards.
34          Chapters 3 through 5 comprise the second main part of this document, dealing with
35    human health and primary standards.  Chapter 3 presents an overview of key policy-relevant

            December 2006                     1 -6        Draft - Do Not Quote  or Cite

-------
 1   health effects evidence, major health-related conclusions from the CD, and an examination of
 2   issues related to the quantitative assessment of health risks. Chapter 4 describes the scope and
 3   methods used in conducting human exposure and health risk assessments and presents initial
 4   results from those assessments.  Chapter 5 includes a preliminary discussion of the current
 5   primary standard. This first draft of the Staff paper begins the discussion of the current standard,
 6   but does not evaluate the standard in light of new information since the last review; that
 7   discussion will be included in the second draft.
 8          Chapter 6 comprises the third main part of this document. Chapter 6 presents a policy-
 9   relevant assessment of Pb welfare effects evidence and describes the scope and methods used in
10   conducting environmental risk assessments, as well as initial results from those assessments.
11   This chapter also includes a preliminary  discussion of the current secondary standard, but as with
12   the primary standard, an evaluation of the current secondary standard will be included in the
13   second draft of this document.
            December 2006                     1 -7         Draft - Do Not Quote or Cite

-------
  1    REFERENCES

  2
  3    Henderson, R. (2006) Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory Committee, to
  4            Administrator Stephen L. Johnson. Re: Clean Air Scientific Advisory Committee (CASAC) Lead Review
  5            Panel's Consultation on EPA's draft Analysis Plan for Human Health and Ecological Risk Assessment for
  6            the Review of the Lead National Ambient Air Quality Standards. July 26, 2006.

  7    U.S. Environmental Protection Agency. (1977) Air quality criteria for lead. Research Triangle Park, NC: Health
  8            Effects Research Laboratory, Criteria and Special Studies Office; EPA report no. EPA-600/8-77-017.
  9            Available from: NTIS, Springfield, VA; PB-280411.

10    U.S. Environmental Protection Agency. (1986a) Air quality criteria for lead. Research Triangle Park, NC: Office of
11            Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
12            EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.

13    U.S. Environmental Protection Agency. (1986b) Lead effects on cardiovascular function, early development, and
14            stature: an addendum to U.S. EPA Air Quality Criteria for Lead (1986). In:  Air quality criteria for lead, v.
15            1. Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria
16            and Assessment Office;  pp.  A1-A67; EPA report no. EPA-600/8-83/028aF. Available from: NTIS,
17            Springfield, VA; PB87-142378.

18    U.S. Environmental Protection Agency. (1989) Review of the national ambient air quality standards for lead:
19            Exposure analysis methodology and validation: OAQPS staff report. Research Triangle Park, NC: Office of
20            Air Quality Planning and Standards; report no. EPA-450/2-89/011. Available on the web:
21            http://www.epa.gov/ttn/naaqs/standards/pb/data/rnaaqsl  eamv.pdf

22    U.S. Environmental Protection Agency. (1990) Air quality criteria for lead: supplement to the 1986 addendum.
23            Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
24            Assessment Office; report no. EPA/600/8-89/049F. Available from: NTIS, Springfield, VA; PB91-138420.

25    U.S. Environmental Protection Agency. (1990b) Review of the  national ambient air quality standards for lead:
26            assessment of scientific and technical information: OAQPS staff paper. Research Triangle Park, NC: Office
27            of Air Quality Planning and Standards; report no. EPA-450/2-89/022. Available from: NTIS, Springfield,
28            VA; PB91-206185. Available on the web: http://www.epa.gov/ttn/naaqs/standards/pb/data/rnaaqsl  asti.pdf

29    U.S. Environmental Protection Agency. (1991) U.S. EPA Strategy for Reducing Lead Exposure.  Available from
30            U.S. EPA Headquarters  Library/Washington, D.C. (Library Code EJBD; Item Call Number:  EAP
31            100/1991.6; OCLC Number 2346675).

32    U.S. Environmental Protection Agency. (2005a) Project Work Plan for Revised Air Quality Criteria for Lead.
33            Research Triangle Park, NC: National Center for Environmental Assessment-RTF Report no. NCEA-R-
34            1465. CASAC Review Draft.

35    U.S. Environmental Protection Agency. (2005b) Air Quality Criteria for Lead (First External Review Draft).
36            Washington, DC, EPA/600/R-05/144aA-bA. Available online at: www.epa.gov/ncea/

37    U.S. Environmental Protection Agency. (2006a) Air Quality Criteria for Lead. Washington, DC, EPA/600/R-
38            5/144aF. Available online at: www.epa.gov/ncea/

39    U.S. Environmental Protection Agency. (2006b) Air Quality Criteria for Lead (Second External Review Draft).
40            Washington, DC, EPA/600/R-05/144aB-bB. Available online at: www.epa.gov/ncea/
              December 2006                        1 -8          Draft - Do Not Quote or Cite

-------
1    U.S. Environmental Protection Agency. (2006c) Plan for Review of the National Ambient Air Quality Standards for
2            Lead. Office of Air Quality Planning and Standards, Research Triangle Park, NC. Available online at:
3            http://www.epa.gov/ttn/naaqs/standards/pb/s_pb cr_pd.html

4    U.S. Environmental Protection Agency. (2006d) Analysis Plan for Human Health and Ecological Risk Assessment
5            for the Review of the Lead National Ambient Air Quality Standards. Office of Air Quality Planning and
6            Standards, Research Triangle Park, NC. Available online at:
7            http://www.epa.gOv/ttn/naaqs/standards/pb/s pb cr pd.html
             December 2006                         1 -9          Draft - Do Not Quote or Cite

-------
 1                      2   CHARACTERIZATION OF AMBIENT LEAD

 2         2.1   INTRODUCTION/BACKGROUND
 3          As described in the CD, "The multimedia aspects of Pb exposure can be seen in that Pb
 4    emissions to the air contribute to Pb concentrations in water, soil and dusts; Pb in soil and dust
 5    also can make important contributions to Pb concentrations in ambient air" (CD, p. 3-1).
 6    Exposures to Pb emitted into the air occur via multiple pathways.  As illustrated in Figure 2-1,
 7    pathways that are directly relevant to a review of the NAAQS include both newly emitted Pb
 8    from currently operating sources, and Pb emitted in the past, either from currently operating
 9    sources or historic sources, which are collectively referred to as "policy-relevant sources".
10    Consequently, this  document considers both airborne Pb as it contributes to human exposures
11    through direct inhalation of particles containing Pb, and also Pb deposited to dusts, soil and other
12    environmental media as it contributes to human exposures through ingestion, and to various
13    ecological exposures.  Further, Pb, once deposited, may also be re-entrained into the air.  In
14    addition, Figure 2-1 shows that people and the environment are also exposed to Pb that originates
15    from non-air sources, including Pb-paint or drinking water distribution systems. For purposes of
16    this review, the Pb  from these non-air sources is collectively referred to as "policy-relevant
17    background".1
            1 This categorization of policy-relevant and background sources is not intended to convey any particular
     policy decision at this stage regarding the Pb standard. Rather, it is simply intended to convey an area of interest to
     this review.

            December 2006                      2-1                Draft - Do Not Quote or Cite

-------
 2
 3
 4
 5
 6
 7
 8
 9
10
1 1
12
13
14
15
16
17
18
19
    Policy-relevant
     Background*
                                   Policy-relevant Sources
        Policy-relevant
        Background*
      Diet*
                       Newly emitted Pb
                                                    Historically emitted Pb
           i Paint
         I Drinking
         i water
                                  Outdoor Soil, Dusts
                                                                             Non-air Pb
                                                                              releases
                                                            Ecological Exposures
        Human Exposures
        -Inhalation & ingestion
      *Policy-relevant background sources and pathways are indicated by dashed lines.
      **Dietary exposure should not be considered to limited to policy-relevant background, as it reflects a combination
      of Pb introduced into food items during processing (policy-relevant background), as well as Pb associated with
      atmospheric deposition (policy-relevant sources).
 1   Figure 2-1.
             Principal pathways of human and ecological exposure to Pb. Among the
             policy-relevant pathways, heavy arrows indicate the predominant human
             exposures.
       This chapter generally characterizes airborne Pb and deposited atmospheric Pb in terms
of physical and chemical properties, measurement methods, recent concentrations and trends,
and relationships with human and ecological exposure.  The information provided here is
intended to provide context for information presented in subsequent chapters, including the
human exposure and risk assessments (Chapter 4) and the ecological risk assessment (Chapter 6).
Additionally, in the second draft of this document, the analyses presented in this chapter are
intended to be informative to policy considerations regarding the primary and secondary Pb
standards. The information presented in this chapter was drawn from the CD and additional
analyses of data from various Pb monitoring networks, emissions inventories, and modeling
activities.
       Section 2.2 presents information on the basic physical and chemical properties of
airborne Pb, including a discussion of environmental fate and transport.  Section 2.3 presents
information on the sources of atmospheric Pb, and their emissions, both on a cumulative national
basis, and on an individual source basis. Section 2.4 presents information on the methods used
to measure ambient Pb, and on Pb concentrations, trends, and spatial patterns in the U.S. Section
2.5 describes currently available information on atmospheric deposition of Pb. Sections 2.6 and
       December 2006
                                                 2-2
Draft - Do Not Quote or Cite

-------
 1    2.7 present information on Pb in other media pertinent to human and ecological exposure
 2    including outdoor dust, soil, surface waters and sediments.

 3         2.2  PROPERTIES OF AMBIENT LEAD
 4          Due to its physicochemical properties, Pb exists in the environment predominantly in
 5    solid form. Consequently upon emission into the air, Pb deposits onto surfaces or exists in the
 6    atmosphere as a component of atmospheric aerosol (CD, Section 2.1).  The various Pb
 7    compounds that are naturally occurring in the environment or are introduced by anthropogenic
 8    activities include oxides, chlorides (or other halides), sulfates, and sulfides (see CD, Table 2-5).
 9    A more complete discussion of the physical and chemical properties of Pb and Pb compounds is
10    provided in the CD (Section 2.1).  The Pb NAAQS pertains to the Pb content of all Pb
11    compounds that may be emitted to air (see Section 2.4.1 for discussion of collection and analysis
12    methods).
13          The relative presence of Pb among the various environmentally occurring compounds
14    influences its distribution within the environment, and the relative bioavailability of these
15    compounds has implications for human and ecological exposures and risks (CD, Sections 4.2.1,
16    8.1.3 and 8.2.3).  With regard to human exposures and risk, this is reflected in the exposure
17    modeling described in Chapter 4.  Lead speciation and bioavailability are discussed further with
18    regard to environmental effects in Chapter 6.

19         2.2.1   Fate and Transport of Pb Particles
20          The atmosphere is the major environmental transport pathway for anthropogenic Pb (CD,
21    p 2-52).  Lead can be transported in the atmosphere and undergo secondary dispersal via the
22    deposition and re-suspension of particles containing Pb. As described in the CD (Section 2.3.1)
23    and in Section 2.4.3 of this  document, airborne Pb particles generally have a bimodal distribution
24    with the greater mass of Pb found in the fine fraction (CD, p. 2-52).  Since small particles are
25    much slower to deposit than larger particles, Pb can be transported great distances in the
26    atmosphere.  Thus, Pb is widely dispersed, as evidenced by detection of Pb even in the most
27    remote places such as the arctic region (CD, pp.  2-52, 3-3).
28          Airborne concentrations of species emitted from a point source are frequently described
29    by a Gaussian distribution.  Gaussian models are, in general, reasonably accurate for small
30    geographic scales, e.g., within -50 km of the source (CD, p. 2-53).  The rate and direction of
31    dispersion are dependent both on pollutant characteristics and meteorological conditions.
32    Important meteorological factors influencing dispersion include wind speed, surface roughness,
33    inversion frequency, inversion duration, and temperature. Results are site specific. For long
34    range transport modeling, Lagrangian trajectory or Eulerian grid models are commonly
            December 2006                     2-3               Draft - Do Not Quote or Cite

-------
 1    employed. These models determine how a parcel of air moves relative to the moving fluid and a
 2    fixed coordinate system, respectively. Retrospective air mass trajectories based on hybrid
 3    models are also used. Results included a finding of airborne Pb in a less industrial country
 4    originating from emissions in several distant countries (CD, p. 2-54).
 5          Wet and dry deposition are the ultimate paths by which Pb particles are removed from the
 6    atmosphere.  Dry deposition is the process by which Pb particles are delivered from the
 7    atmosphere onto surfaces in the absence of precipitation. Factors that govern  dry deposition are
 8    the level of atmospheric turbulence, especially in the layer nearest the ground, particle size
 9    distributions and density, and the nature of the surface itself, such as smooth or rough. In the
10    commonly used model formulation for dry deposition, it is  assumed that the dry deposition flux
11    is directly proportional to the local concentration of the pollutant species, at some reference
12    height above the surface (e.g., 10 m or less), multiplied by the deposition velocity (CD, p. 2-55).
13    The concentration is computed by the dispersion models mentioned above, depending on local
14    versus regional or global applications. Estimates of dry deposition velocity constitute the
15    primary output of a large number of dry deposition models  that have been developed during the
16    past ten years and most of these rely on so-called "resistance schemes."  The  advantage of this
17    deposition velocity representation is that all the complexities of the dry deposition process are
18    bundled in a single parameter, but the disadvantage is that because this parameter addresses a
19    variety of processes, it is difficult to specify properly.  A large range of Pb deposition velocities
20    (0.05 to 1.3 cm/s) has been reported (CD, pp. 2-55 to 2-57 and Table 2-21).
21          Wet deposition, or the delivery of a pollutant to the  ground in precipitation, is the process
22    by which airborne pollutants are scavenged by precipitation and removed from the atmosphere.
23    The flux of a depositing species can be defined as the product of the rate of precipitation and  the
24    concentration of the chemical species in the precipitation (CD, pp. 2-57 to 2-59). Wet
25    deposition, is affected by: 1) nucleation scavenging (removal by direct incorporation into new
26    cloud droplets); 2) in-cloud scavenging (removal by incorporation into existing cloud droplets);
27    and 3) precipitation washout (removal by rain as it is falling to the ground). The size of particles
28    can influence wet deposition rates. Large particles are scavenged by  precipitation more
29    efficiently than smaller particles (CD, p. 2-59). Lead, beyond the influence of individual
30    sources, is found primarily in the submicron size range, and consequently does not undergo wet
31    deposition as easily as many of the crustal elements (CD, p. 2-59). Models vary in how they
32    treat wet deposition. Gaussian models focus solely on washout aspects, mainly because this
33    process is dominant within the 50 km limit of model applicability. Regional and global models
34    have more comprehensive treatment of wet  deposition.  Lead concentrations in precipitation
35    have shown a pronounced downward trend from the 1970s  into the 1990s, presumed primarily
36    due to the phase-out of leaded fuel (CD, pp. 2-60 to 2-61 and Table 2-22).

            December 2006                     2-4                Draft - Do Not Quote or  Cite

-------
 1          The re-suspension of soil-bound Pb particles and contaminated road dust can be a
 2    significant source of airborne Pb (CD, Section 2.3.3, and p. 2-62).  Studies of emissions in
 3    southern California indicate that Pb in re-suspended road dust may represent between 40% and
 4    90% of Pb emissions in some areas (CD, p. 2-65). Lead concentrations in suspended soil and
 5    dust, however, vary significantly (CD, p. 2-65).  In general, the main drivers of particle re-
 6    suspension are typically mechanical stressors such as vehicular traffic, construction and
 7    agricultural operations, and to a lesser extent, the wind.  Understanding the physics of re-
 8    suspension from natural winds requires analyzing the wind stresses on individual particles and
 9    although this analysis can be accurate on a small scale, predicting re-suspension on a large scale
10    generally focuses on empirical data for soil movement due to three processes: saltation, surface
11    creep, and suspension (CD, pp. 2-62 to 2-63). Further, rather than a continuous process, re-
12    suspension may occur as a series of events.  Short episodes of high wind speed, dry conditions,
13    and other factors  conducive to re-suspension may dominate annual averages of upward flux (CD,
14    p. 2-65).  These factors complicate emissions estimates.

15         2.3  SOURCES AND EMISSIONS  TO THE ATMOSPHERE
16          The purpose of this section is to describe the available information on sources of Pb into
17    the ambient air.  The section does not provide a  comprehensive list of all sources of Pb, nor does
18    it provide estimates of emission rates or emission factors for all source categories, since such
19    information is available for only a limited number of sources. Rather, the discussion here is
20    intended to identify the larger source categories, either on a national or local scale, and provide
21    some characterization of their emissions and distribution within the U.S. The data sources for
22    this information are described in Section 2.3.1.  Limitations of and confidence in the information
23    is discussed in Section 2.3.2. Section 2.3.3 summarizes temporal trends of major source sectors,
24    and Section 2.3.4 summarizes estimates of 2002 national emissions totals for the larger source
25    categories. And lastly, Section 2.3.5  describes the geographic distribution of emissions based on
26    the 2002 estimates.

27         2.3.1  Data sources
28          The Pb emissions information presented here is drawn largely from EPA's National
29    Emissions Inventory (NEI).  The NEI is based on inputs from State, Tribal and local air pollution
30    agencies and data obtained during the preparation of technical support information for the EPA's
31    hazardous air pollutant regulatory programs.  The Agency is currently developing version 2.1 of
32    the NEI for 2002 (http://www.epa.gov/ttn/chief/net/2002inventory.html).  The information
33    presented in this draft document is based on that version (USEPA, 2006), and comments
            December 2006                     2-5                Draft - Do Not Quote or Cite

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15

16
17
18

19
20
21
22
23
24
25
      received on the NEI and on this draft document will be considered in developing such
      information for the second draft of this document.
            There are some 13,000 industrial, commercial or institutional sources in the U.S. that
      contain one or more processes that emit Pb to the atmosphere and whose Pb emission estimates
      are included in the 2002 NEI (Table 2-1). Most of these sources emit < 0.1 tons per year (tpy).
      There are approximately 1300 point sources of Pb in the NEI with estimates of emissions greater
      than or equal to 0.1 tpy and these point sources, combined, emit 1037 tpy, or 94% of the Pb point
      source emissions. In other words, 94% of Pb point source emissions are emitted by the largest
      10% of the sources.
            In the NEI, emissions estimates for some of the point sources are in terms of mass of Pb
      compounds, whereas the non point source and mobile source emissions estimates are in terms of
      mass of the Pb only.  For the point sources, approximately 80% are reported as mass of Pb and
      most of the other 20% are reported as mass of Pb compounds.  The high molecular weight of Pb
      (as compared to elements with which it is associated in Pb compounds), however, reduces the
      impact of this reporting inconsistency.

      Table 2-1.  Distribution of point sources within the 2002 NEI and associated estimated
                 emissions.

Emissions
Range
(tpy)
<0.1
0.1 to 1.0
1.0 to 5
>5
Summary


Number
of Sources
11,822
1038
204
26
13,087

Total
Emissions
(tpy)
73
355
394
288
1114
Average
Emissions
per Source
(tpy)
<0.01
0.3
2
10

           2.3.2   Confidence Level for Emission Estimates
            The comprehensiveness of emission inventories depends upon what is known regarding
     which source types emit Pb, their locations and their operating characteristics, as well as the
     reporting of this information to the inventory. As described in Section 2.3.1, the NEI relies on
     information that is available from a variety of sources for this information. There are numerous
     steps, each with its own uncertainties, associated with the development of this information for
     use in the emissions inventory.  First, the categories emitting Pb must be identified. Second, the
            December 2006
                                               2-6
Draft - Do Not Quote or Cite

-------
 1    sources' processes and control devices must be known. Third, the activity throughputs and
 2    operating schedules of these sources must be known. Finally, we must have emission factors to
 3    relate emissions to the operating throughputs, process conditions and control devices. The
 4    process, control device, throughputs and operating schedules are generally available for each
 5    source. However, the emission factors represent average emissions for a source type and average
 6    emissions may differ significantly from source to source.  More information on emission factors
 7    and the estimation of emissions is found in the introduction to EPA's Compilation of Air
 8    Pollutant Emissions Factors (USEPA, 2006b). Further information on emission factors is
 9    available at:  http://www.epa.gov/ttn/chief/ap42/.
10          The process of identifying sources that emit Pb into the air has been ongoing since before
11    the Clean Air Act of 1970.  The NEI includes inventory estimates for Pb sources with some
12    exceptions, including re-suspended road dust, biomass burning and trace levels of Pb in motor
13    fuel and lubricating oil. For example, while Pb has not been added to automobile gasoline for
14    two decades, there are still deposits of Pb near roads. This Pb can be entrained into the air by
15    disturbance of soil near the roads or by burning of biomass materials that are near roads (CD,
16    Sections 2.3.3 and 2.2.1). We have not yet developed estimates for the NEI of Pb emissions
17    associated with re-suspension of Pb still residing in roadway dust and soil. Likewise, we have
18    not developed NEI estimates for Pb emitted from burning of biomass materials near roads or
19    how much of the Pb that accumulated away from roads due to transport and deposition is still
20    subject to emissions from forest fires or mechanical  disturbance of soils. As described in the CD
21    (Section 8.2.2), re-suspension of soil bound Pb and contaminated road dust may be a significant
22    source of airborne Pb, however, quantitative estimates for this process remains an area of
23    significant uncertainty.  Also, while Pb is no longer added to gasoline or diesel fuel, it is still
24    present as a trace contaminant in both fuels and there are trace amounts of Pb in lubricating oils.
25    These are not reflected in the emissions inventory

26         2.3.3  Trends in National Emissions:  1980 to 2002
27          Table 2-2 shows a downward trend in the fuel combustion, industrial process and solid
28    waste sectors from 1980 through 2002, as well as the dramatic reduction in Pb emissions in the
29    transportation sector due to the removal of Pb from gasoline. While the most dramatic
30    reductions occurred prior to 1990, Pb emissions were further reduced by 68% (from 5,000 to
31    1,600 tpy) between 1990 and 2002 (USEPA 1990; 2002 NEI). The greatest emission reductions
32    were from mobile sources, primary and secondary Pb and copper smelting, pulp and paper
33    manufacturing, inorganic paint pigment production and steel wire products. As discussed in the
34    CD (Section  2.2.4), reductions in mobile sources include some associated with the latter period
            December 2006                      2-7                Draft - Do Not Quote or Cite

-------
 1   of the "phase-out" of leaded gasoline. From a national inventory perspective, the categories that
 2   have the largest emissions in the 2002 NEI are discussed briefly in Section 2.3.4.

 3   Table 2-2.  Trend in Pb emissions (tpy) from 1980 to 2002.
                                  1980    1985    1990   1995
                                2002
          Transportation
          Fuel Combustion
          Industrial Processes
          Solid Waste
          Total
64,706
4,299
3,938
1,210
18,973
515
2,531
871
1,197
500
2,474
804
564
490
2,271
604
392
425
736
87
74,153  22,890   4,975   3,929   1,640
          Note: Estimates for 1980-1995 are from
          http://www.epa.qov/airtrends/econ-emissions.html.
          Estimates for 2002 are from Version 2.1 of the 2002 National Emissions
          Inventory, US EPA (USEPA, 2006a). The estimates for 2002 differ
          from those in Table 2-8 of the CD due to changes in the 2002 NEI
          subsequent to publication of the CD.	
 5         2.3.4   Source Categories with Largest National Total Pb Emissions
 6          Table 2-3 shows the sources of Pb emissions estimated to emit more than 5 tons per year
 7   of Pb in the 2002 NEI. The main sources of emissions in the 2002 NEI are comprised primarily
 8   of combustion-related emissions and industrial process-related emissions.  Point source
 9   emissions account for about 68% of the national Pb emissions in the 2002 NEI. The point source
10   emissions are roughly split between combustion and industrial processes; non road sources
11   (general aviation aircraft - leaded fuel) accounts for 24%.
12          Presence of a source category on this list does not necessarily provide an indication of the
13   significance of the emissions from individual sources within the source category.  A source
14   category, for example, may be composed of many small (i.e., low-emitting) sources, or of just a
15   few very large (high-emitting) sources.  Such aspects of a source category, which may influence
16   its potential for human and ecological impacts, are included in the following short descriptions of
17   the largest source categories identified in Table 2-3.
            December 2006
               2-8
Draft - Do Not Quote or Cite

-------
1    Table 2-3.  Source categories emitting greater than 5 tpy of Pb in the 2002 NEI.
      Source Category Name
      ALL CATEGORIES b
      Mobile sources (Leaded Aviation Fuel)
      Utility Boilers
      Industrial/Commercial/ Institutional Boilers & Process Heaters
      Iron and Steel Foundries
      Primary Lead Smelting
      Secondary Lead Smelting
      Mining
      Military Installations
      Municipal Waste Combustors
      Electric Arc Furnaces (EAF)
      Integrated Iron & Steel Manufacturing
      Pressed and Blown Glass and Glassware Manufacturing
      Lead Acid Battery Manufacturing
      Secondary Nonferrous Metals
      Hazardous Waste Incineration
      Solid Waste Incineration
      Primary Copper Smelting
      Portland Cement Manufacturing
      Primary Metal Products Manufacturing
      Pulp and Paper Production
      Industrial Inorganic Chemical Manufacturing
      Sewage Sludge Incineration
      Synthetic Rubber Manufacturing
      Secondary Aluminum Production
      Farm Machinery and Equipment Manufacturing
      Secondary Copper Smelting
      Stationary Reciprocating Internal Combustion Engines
      Industrial Machinery and Equipment:  Finishing Operations
      Nonferrous Foundries
      Ferroalloys Production
      Residential Heating
      Miscellaneous Metal Parts & Products (Surface Coating)
      Primary Nonferrous Metals—Zinc, Cadmium and Beryllium
      Engine Test Facilities
      Coke Ovens
      Surface Coating Operations (Auto Refinishing)
            2002 Total
            Emissions
            (TPY)a
                 1,640
                  392
                  221
                   191
                   110
                   59
                   46
                   38
                   33
                   33
                   32
                   32
                   31
                   25
                   23
                   22
                   22
                   22
                   22
                   20
                   10
                   10
                   10
                     9
                     9
                    7
                    7
                    7
                    6
                    6
                    5
                    5
                    5
                    5
      aSome values differ from those in the CD (Table 2-8) due to changes in the 2002
      NEI subsequent to CD publication.  Additionally, values just above 5 tpy have
      been rounded to 5.
      Includes emissions from 137 TPY emitted by 314 smaller categories (70 TPY in
      MACT categories and 67 TPY in non MACT).	
            December 2006
2-9
Draft - Do Not Quote or Cite

-------
 1         2.3.4.1  Industrial/Commercial/Institutional/Process Heaters
 2          Coal and/or other substances (e.g., oil, wood) are burned in boilers and process heaters to
 3    produce steam. With regard to boilers, the steam is used to produce electricity or provide heat,
 4    while process heaters are used in industrial processes. Given their use at a wide variety of
 5    facilities (e.g., refineries, chemical and manufacturing plants, etc), as well as in a "stand alone"
 6    mode to provide heat for large building complexes, there are thousands of these sources
 7    throughout the country, generally located in urban areas, and they range widely in size. Most
 8    coal-fired industrial boilers emit about 0.06 tpy while the larger ones emit about 0.07 tpy due to
 9    the use of high efficiency particulate matter (PM) control (ERG, 2002a).  U.S. EPA promulgated
10    a national emissions standard in 2004 for this category that will reduce Pb emissions (U.S. EPA,
11    2004).

12         2.3.4.2  Utility Boilers
13          This category includes boilers that burn coal, oil and natural gas (or, at times, other
14    substances such as wood and petroleum coke) to produce steam to produce electricity or to
15    provide co-generated heat for process operations. Lead is present as a natural trace metal in the
16    fuel and is emitted to air following combustion.  The extent of emissions depends on the
17    concentration of Pb in the fuel, the quantity of fuel burned, and PM control devices applied.
18    Most common PM control devices used in the U.S. (such as fabric filters and electrostatic
19    precipitators)  capture Pb  relatively well. However, some devices work somewhat better than
20    others.  Coal-fired utilities have the highest Pb emissions among these boilers. Oil-fired plants
21    emit somewhat lower amounts, and gas-fired plants emit very low levels of Pb (U.S. EPA,  1998).
22    There are approximately  1,300 coal-fired electric utility boilers in the U.S. ranging in  size from
23    25 to approximately 1,400 MWe.  Based on emission factor calculations,  a 325 MWe  coal-fired
24    boiler would be expected to emit approximately 0.021 tpy Pb, based on the use of an electrostatic
25    precipitator for PM control (USEP A, 1998).  Coal-fired utility boilers tend (there are exceptions)
26    to be located in non-urban areas and have stack heights that meet standards of good construction
27    practice.

28         2.3.4.3  Mobile Sources
29          Until 1995, when the phase-out of Pb in the nation's motor vehicle gasoline supply  was
30    complete, leaded gasoline was the dominant  source of Pb to the atmosphere.  Currently, Pb is
31    added to gasoline used in piston-engine aircraft and some types of race cars.  Depending  on the
32    grade of aviation fuel, or "avgas," the Pb content can range from 0.1 to 1.1 g of Pb per liter
33    (Chevron, 2000). About 33 percent of general aviation aircraft use avgas,  and the remainder use
34    jet fuel, which does not contain Pb additives  (CD, p. 2-51). Emissions of Pb from the use of
35    avgas is the largest  source of Pb to the air currently and is the only mobile source of Pb in the

            December 2006                     2-10               Draft - Do Not Quote or Cite

-------
 1    2002 inventory (Figure 2-2). The underlying data used in calculating the emissions of Pb from
 2    avgas are being reviewed which might lead to a changes in the emissions estimate for this
 3    source. Emissions from the combustion of leaded fuel include submicron inorganic Pb halides,
 4    as well as larger, coarse fraction Pb compounds (Habibi 1973).
 5          Vehicles used in racing are not regulated by the EPA under the Clean Air Act and can
 6    therefore use alkyl-lead additives to boost octane. EPA has formed a voluntary partnership with
 7    NASCAR with the goal of permanently removing alkyl-Pb from racing fuels used in the Busch,
 8    Nextel Cup (formerly known as Winston Cup), and Craftsman Truck Series (CD, p. 2-50). In
 9    January of 2006, NASCAR agreed to switch to unleaded fuel in its racecars and trucks beginning
10    in 2008.
11          Lead is also present as a trace contaminant in gasoline and diesel fuel and is a component
12    of lubricating oil. A range of Pb emission factors for motor vehicles is reported in the CD (pp. 2-
13    46, 2-47, 2-49). Mobile sources of roadside Pb contamination include deposition of Pb in
14    exhaust (largely originating from leaded gasoline), brake wear,  tire wear, and loss of Pb wheel
15    weights. Brake wear emissions are highly variable and depend on brake pad composition and
16    driving patterns. Wheel weights can become dislodged during quick stops and although
17    deposited pieces of wheel weights are quite large, Pb is very malleable and can be worn away
18    into respirable particles by being run over by vehicles (CD,  p. 2-50).
19          Lead measured in the vicinity of motor vehicle traffic is reported to have a bimodal size
20    distribution; a submicron mode that likely originates largely as  a product of combustion, and a
21    larger mode with an approximate size range of 1.0 to 18 um in diameter, which is likely a
22    product of physical processes such as road dust re-suspension and tire or brake wear, with some
23    contribution from exhaust (CD, p. 2-48).

24         2.3.4.4  Iron and Steel Foundries
25          Iron and steel foundries melt scrap, ingot, and other  forms of iron and steel and pour the
26    molten metal into molds for particular products.  The largest Pb emission sources at iron
27    foundries are large capacity furnaces, for which emissions release heights are on the order of 25-
28    30 feet.  Lead emissions from these furnaces generally range from about 0.3 to 3 tpy, depending
29    on the throughput of the furnace, the type and operating characteristics of the emission control
30    system, and the Pb content in the metal charged to the furnace.  In the U.S., there are about 650
31    existing foundries, all located in 44 of the lower 48 states, and most of which are iron foundries
32    operated by manufacturers of automobiles and large industrial equipment and their suppliers.
33    Foundries may be located in cities or in rural  areas.  There is a concentration of foundries in the
34    Midwest (IN, IL, OH, MI, WI, and MN) - roughly 40% of foundries with almost 60% of U.S.
35    production (USEPA, 2002b). Various regulations affecting Pb  emissions from this category


            December 2006                    2-11               Draft - Do Not Quote or Cite

-------
 1   were promulgated in 2004. Compliance with select work practices were required by April 2005,
 2   while compliance with other emission limitations are required by April 2007. The combined
 3   impact of these actions is projected to reduce Pb emissions from this category by approximately
 4   25 tpy (USEPA, 2004a).

 5         2.3.4.5  Hazardous Waste Incineration/ Combustion Facilities
 6          Hazardous waste combustors include hazardous waste incinerators, as well as boilers and
 7   industrial furnaces that burn hazardous waste for energy or material recovery (e.g., production of
 8   halogen acid from the combustion of chlorine-bearing materials).  Industrial furnaces burning
 9   hazardous waste include cement kilns, lightweight aggregate kilns, and hydrochloric acid
10   production furnaces. Lead is a trace contaminant in the hazardous waste, fossil fuels, and raw
11   materials used in the combustors. In 2005, there were nearly 270 hazardous waste combustor
12   sources in operation in the United States (70 FR at 59530).  Approximately 40 percent of
13   hazardous waste combustors are located in the states of Texas and Louisiana. The source
14   categories with the largest number of combustors were boilers and incinerators with 116 and 107
15   sources, respectively.  On October 12, 2005, EPA finalized  standards implementing section
16   112(d) of the Clean Air Act by requiring all existing and new hazardous waste combustors to
17   meet HAP  emission standards reflecting the performance of the maximum achievable control
18   technology (MACT) (70 FR 59402). EPA promulgated emission standards for Pb and sources
19   must be in  compliance with these standards by October 2008. Following compliance with the
20   standards, EPA estimates that cumulative Pb emissions from hazardous waste combustors will be
21   reduced approximately to 4.0 tons per year (USEPA, 2005).  This represents a 95% reduction in
22   Pb emissions from  1990 levels.
23         2.3.4.6  Primary Lead Smelting
24          At primary Pb smelters, Pb-bearing ore concentrates are smelted to produce Pb metalThe
25   processes at a primary Pb smelter include: ore concentrate storage and handling; sintering of ore
26   concentrates; sinter crushing and handling; smelting of sinter to Pb metal; dressing, refining, and
27   alloying of Pb metal; and smelting of drosses. Lead is emitted from primary Pb smelters as
28   process emissions, process fugitive emissions, and fugitive dust emissions (CD, p. 2-21). U.S.
29   EPA promulgated a national emissions standard in 1999 for this category which includes an
30   emissions limit for Pb (U.S. EPA 1999a). In the 1990s, there were three operating primary Pb
31   smelters in the U.S: one in Montana and two in Missouri. In 2002, there were two in operation
32   (estimated  emissions shown in Table 2-3); one of the two had less than 1 tpy Pb emissions. As
33   of 2005, there is only one operating primary Pb smelter in the U.S. which is located in Missouri.
34   The estimate of Pb emitted from this facility in 2005 is 25 tons (CD, p. 2-20).
            December 2006                     2-12              Draft - Do Not Quote or Cite

-------
 1         2.3.4.7  Secondary Lead Smelting
 2           Secondary Pb smelters are recycling facilities that use blast, rotary, reverberatory, and/or
 3    electric furnaces to recover Pb metal from Pb-bearing scrap materials, primarily Pb-acid
 4    batteries. This category does not include remelters and refiners or primary Pb smelters.  At
 5    secondary Pb smelters, Pb may be emitted from: (1) process emissions contained in the primary
 6    exhaust of smelting furnaces, (2) process fugitive emissions associated with charging and
 7    tapping of smelting furnaces and Pb refining kettles, and (3) fugitive dust emissions from wind
 8    or mechanically induced entrainment of dust from stockpile and plant yards and roadways. U.S.
 9    EPA promulgated a national emissions standard in  1997 for this category which includes an
10    emissions limit for Pb (USEPA, 1997).  In 2002, there were 15 secondary smelters operating in
11    11 states, most of which are in the eastern half of the U.S. The 2002 emissions estimates for the
12    individual facilities indicate most having total emissions (fugitive and process) of less than 4 tpy,
13    and one facility having total emissions on the order of 12 tpy (USEPA, 2006a; EC/R, 2006).
14         2.3.4.8  Military Installations
15           This source category includes sources that are military facilities. The types of sources
16    contributing to Pb emissions from this category include, among others, rocket and engine test
17    facilities, ammunition manufacturing, weapons testing, waste combustion and boilers. While
18    there are over 300 military facilities in the NEI, only 10% emit over 0.1 tpy of Pb and only 3%
19    emit over 1 ton per year.   The two largest facilities (listed in Table 2-3) are a missile
20    ammunition production plant and a weapons testing facility and these two facilities account for
21    over 75% of the category emissions.

22         2.3.4.9  Mining
23           This category includes various mining facilities that extract ore from the earth containing
24    Pb, zinc, copper and/or other non-ferrous metals (such as gold and silver), and/or non-metallic
25    minerals such as talc and coal.  This category does  not include the smelting or refining of the
26    metals and minerals.  These facilities produce ore concentrates (such as Pb, zinc, and copper
27    concentrates) that are transported to other facilities where further processes, such as smelting and
28    refining take place. The 2002 NEI indicates 3 mining facilities in the U.S. emitting greater than
29    0.5 tpy Pb, one of which emits more than 5  tpy.  This facility is in Missouri and produces Pb,
30    zinc, and copper concentrates that are shipped to customers for further processing.
31         2.3.4.10 Integrated Iron & Steel Manufacturing
32           Integrated iron and steel manufacturing includes facilities engaged in the production of
33    steel from iron ore. The processes involved include sinter plants, blast furnaces that produce
34    iron, and basic oxygen process furnaces that produce steel, as well as several  ancillary processes

            December 2006                     2-13                Draft - Do Not Quote or Cite

-------
 1    including hot metal transfer, desulfurization, slag skimming, and ladle metallurgy. There are
 2    currently 17 facilities. The range of Pb emissions is from 2 to 8 tpy per facility. Stack heights
 3    range from heights of 30 - 50 feet. The facilities are located in 9 states; mostly in the Midwest
 4    (USEPA, 2003a). U.S. EPA promulgated a national emissions standard in 2003 for this category
 5    which includes an emissions limit for PM (USEPA, 2003c).

 6         2.3.4.11 Municipal Waste Combustors:  Small & Large
 7          Municipal waste combustors (MWCs) are units that incinerate municipal or municipal-
 8    type solid waste.  Currently about 14 % of the municipal waste generated in the US is
 9    incinerated. The amount of municipal waste incinerated has remained stable over the past
10    decade.  As described in the CD, the amount of Pb emitted from municipal waste combustors
11    depends on the amount of Pb in the refuse, with typical sources including paper, inks, cans and
12    other metal scrap and plastics (CD, pp. 2-35 to 2-36).  The Clean Air Act of 1990 required
13    MACT be applied to all new municipal waste incineration units and retrofitted to all  existing
14    municipal waste incineration units. The MACT retrofits at existing MWCs were completed by
15    2005 and national Pb emissions from  municipal waste incineration are now less than 10 tons per
16    year. This represents greater than a 97% reduction in national Pb emissions from these
17    incinerators since 1990. There are currently 66 large MWC plants and 26 small MWC plants
18    operating nationally. Current information indicates that individual large MWC plants general
19    emit less than 0.1 tpy Pb, while small MWC plants generally emit less than 0.02 tpy Pb (ERG,
20    2002b,c; Stevenson, 2002).

21         2.3.4.12 Pressed and Blown Glass and Glassware Manufacturing
22          The pressed and blown glass and  glassware manufacturing category includes
23    manufacturers of flat glass, glass containers, and other pressed and blown glass and glassware.
24    Lead is emitted primarily from the pressed and blown glass industry sector. Some container
25    plants also make a leaded-glass product, but this is not typical  of container glass plants.  Lead
26    may also be added to flat glass for use in microwaves and flat-screen TVs. Emissions from
27    individual facilities may range from a few pounds per year up  to several tons per year depending
28    on Pb content of their glass and the level  of control. Furnace stacks for these facilities are
29    typically of the order of 35-60 feet high.  As of 2005, about 22 tons of Pb is emitted from glass
30    manufacturing annually. Glass plants  are located in 35 States (RTI, 2006). U.S. EPA is currently
31    developing a regulation for HAP emissions from this category, which is scheduled for
32    promulgation in December 2007.
            December 2006                    2-14               Draft - Do Not Quote or Cite

-------
 1         2.3.4.13 Electric Arc Furnace Steelmaking
 2          In the Steelmaking process that uses an electric arc furnace (EAF), the primary raw
 3    material is scrap metal, which is melted and refined using electric energy. Since scrap metal is
 4    used instead of molten iron, there are no cokemaking or ironmaking operations associated with
 5    steel production that use an EAF.  There are currently 141 EAFs at 93 facilities. The total
 6    nationwide Pb  and Pb compound emissions are approximately 80 tons, and the average per
 7    facility is approximately 0.75 tpy.  Stack heights range from heights of 30 - 50 feet.  The
 8    facilities are located in 32 states; mostly in the northeast and Midwest, with ninety percent of the
 9    facilities located in urban areas. This information is drawn from multiple sources (Lehigh,
10    1982; Calspan, 1977; RTI, 2005). U.S. EPA is developing a regulation for HAP emissions from
11    this category, which is scheduled for promulgation in December 2007.

12         2.3.4.14 Lead Acid Battery Manufacturing
13          The Pb acid battery  manufacturing category includes establishments primarily engaged in
14    manufacturing storage batteries. Lead acid storage batteries are produced from Pb alloy ingots
15    and Pb oxide.  The Pb oxide may be prepared by the battery manufacturer or may be purchased
16    from a supplier. There are currently  58 facilities operating (data obtained from the Battery
17    Council International (BCI)); there is a general slow decline in the number of facilities.  The total
18    Pb and Pb compound emissions in the 2002 NEI from approximately 50 of the facilities included
19    in the NEI were 25 tons.  The range  of facility specific Pb and Pb compound emissions is from 1
20    x 10"5 to just below 10 tpy,  with an average of 0.5 tpy. Facilities are located in both urban and
21    rural areas. The facilities are located in 23 states and Puerto Rico (2002 NEI).
22         2.3.4.15        Primary Copper Smelting
23          This source category includes all industries which refine copper concentrate from mined
24    ore to anode grade copper, using pyrometallic processes.  Smelting includes the handling and
25    blending of ore concentrate; the drying of copper concentrate; the smelting of concentrate to
26    matte grade copper; the conversion of matte grade copper to blister grade copper; the refining of
27    blister grade copper to anode grade copper; and the pouring of copper anodes.  Seven primary
28    copper smelters are currently operating in the U.S. Six of these seven smelters use conventional
29    smelter technology which includes batch converter furnaces for the conversion of matte grade
30    copper to blister copper, while the seventh uses a continuous flash furnace.  Two of the three
31    largest smelters are located in AZ, and the third is in Utah. The largest facility emitted an
32    estimated 12.8 tons Pb in 2002. The estimated emissions for the other two large facilities are
33    between 0.1 to 5 tpy. No other source  in this category emits more than 0.1 tpy.  In 2002, U.S.
34    EPA promulgated a national emissions standard, including limits for PM, for this category
35    (USEPA, 2002d).

            December 2006                     2-15              Draft - Do Not Quote or Cite

-------
 1         2.3.4.16 Portland Cement Manufacturing
 2          Portland cement manufacturing is an energy intensive process in which cement is made
 3    by grinding and heating a mixture of raw materials such as limestone, clay, sand, and iron ore in
 4    a rotary kiln, which is a large furnace that is fueled by coal, oil, gas, coke and/or various waste
 5    materials.  Lead is a trace contaminant both of the raw materials and some fuel materials (e.g.,
 6    coal).  Thus, it is emitted in particulate materials from the kiln stacks, which range in height from
 7    near 10 meters to more than 100 meters, with relatively smaller releases from grinding, cooling,
 8    and materials handling steps in the manufacturing process.  Portland cement facilities tend to be
 9    located in portions of the country with limestone deposits and in rural areas or near small towns.
10    The largest numbers of facilities are in Pennsylvania and California, although a significant
11    percentage of facilities are in the Midwest. Between 1990 and 2002 total industry capacity grew
12    by 22 percent, although the number of facilities decreased slightly from 112 to 108, and as of
13    2004, there were 107 Portland cement plants in the U.S. (O'Hare, 2006). All but three facilities
14    report less than 1 tpy of Pb emissions. The highest estimated Pb emissions for a facility in the
15    2002 NEI is 5.4 tpy. In 1999, U.S. EPA promulgated a national emissions standard, including a
16    limit for PM, for this category (USEPA, 1999b).

17         2.3.5  Geographic Distribution of Sources
18         2.3.5.1  National Patterns in the Distribution of Pb Emissions
19          Figure 2-2 shows the geographic distribution and magnitude of Pb emissions in the U.S.
20    from all sources identified in the 2002 NEI, in terms of emissions density (defined here as tons
21    per area per county). This presentation indicates a broad distribution of the Pb emissions across
22    the US, with perhaps a concentration of emissions in a broad swath from Indiana to southern
23    New England, as well as in other scattered locations.  Figure 2-3 shows the emission density
24    specifically for mobile and non-point sources. Lastly, Figure 2-4 presents the geographic
25    distribution of point sources in the 2002 NEI with Pb emissions estimates greater than 1 tpy.
26
            December 2006                    2-16               Draft - Do Not Quote or Cite

-------
1

2

                                                         County Lead Emissions
                                                          (rpy/1000 sq miles)
Figure 2-2.   Emission Density of All Pb Sources in the 2002 NEL
       December 2006
                                               2-17
Draft - Do Not Quote or Cite

-------
                                                            County Lead Emissions
                                                             (tpy/1000 sq miles)
2   Figure 2-3.  Emission Density of Mobile and Non Point Sources of Pb in 2002 NEL
           December 2006
2-18
Draft - Do Not Quote or Cite

-------
                                                         Facility Emissions (tpy)
 2   Figure 2-4.  Geographic distribution of point sources with >1 tpy Pb emissions in 2002
 3                NEI.
 5         2.3.5.2  Largest Pb Point Sources in the 2002 NEI
 6          While Sections 2.3.3 and 2.3.4 focus on source categories that rank highest due to
 7   cumulative national Pb emissions, this section is intended to consider Pb emissions on the
 8   individual source level. As mentioned in Section 2.3.1 (see Table 2-1), the 2002 NEI includes 26
 9   facilities with emissions estimated to be greater than or equal to 5 tons per year (Table 2-4).
10   Most of these sources are metallurgical industries, followed by waste disposal facilities and
11   manufacturing processes. The information presented in Table 2-4 is based on the current version
12   of the NEI (USEPA, 2006a).
            December 2006
2-19
Draft - Do Not Quote or Cite

-------
1   Table 2-4.  Point Sources with Pb emissions in 2002 NEI greater than or equal to 5 tpy.
Source Category Name
Primary Lead Smelting
Military Installation
Mining
Secondary Nonferrous Metals
Primary Copper Smelting
Electric Arc Furnaces
Secondary Lead Smelting
Integrated Iron & Steel Manufacturing
Pressed and Blown Glass and Glassware Manufacturing
Military Installation
Hazardous Waste Incineration
Lead Acid Battery Manufacturing
Industrial and Commercial Machinery Manufacturing
Synthetic Rubber Products Manufacturing - Fabric Coating Mills
Commercial and Industrial Solid Waste Incineration
Iron and Steel Foundries
Integrated Iron & Steel Manufacturing
Integrated Iron & Steel Manufacturing
Mineral Products Manufacturing
Commercial and Industrial Solid Waste Incineration
Ferroalloys Production
Nonferrous Foundries
Portland Cement Manufacturing
Coke Oven
Iron and Steel Foundries
Mining
2002 Point
Emissions
State County Name (TPY)a
MO
OK
MO
TX
AZ
IL
MO
IN
TN
PA
AR
KY
KS
IN
AR
OH
IN
IN
NM
CT
OH
NE
MD
VA
IA
MO
a (USEPA, 2006a)
''Following compliance with the MACT standards in 2008, Pb emissions are
Jefferson County
Pittsburg County
Reynolds County
Potter County
Gila County
Peoria County
Iron County
Lake County
Madison County
Franklin County
Union County
Madison County
Marshall County
Cass County
Clark County
Cuyahoga County
Porter County
Lake County
Socorro County
Windham County
Washington County
Nemaha County
Frederick County
Buchanan County
Jefferson County
Reynolds County
estimated to be 0.07 tpy.
59
17
15
14
13
13
12
11
11
10
10b
10
8
7
7
7
7
6
6
6
6
6
5
5
5
5

3         2.4   AMBIENT AIR CONCENTRATIONS
4          The EPA has been measuring Pb in the atmosphere since the 1970s.  For the most part,
5   Pb concentrations have decreased dramatically over that period.  This decrease is primarily
6   attributed to the removal of Pb from gasoline, however, some isolated locations still Pb
7   concentrations above the level of the NAAQS.  The following sections describe the ambient Pb
8   measurement methods, the sites and networks where these measurements are made, as well as
9   how the ambient Pb concentrations vary geographically and temporally.
           December 2006
2-20
Draft - Do Not Quote or Cite

-------
 1         2.4.1   Ambient Pb Measurement Methods
 2          A number of methods are used to collect Pb and measure Pb concentrations in the
 3    atmosphere, however, most methods use a similar sample collection approaches. Ambient air is
 4    drawn through an inlet for a predetermined amount of time (typically 24 hours) and the PM is
 5    collected on a suitable filter media. After the sample has been collected, the filter may be used to
 6    determine the mass of PM collected prior to then being used for determination of Pb. The filter is
 7    chemically extracted and analyzed to determine the Pb concentration in the paniculate. The
 8    concentration of Pb found in the atmosphere, in |ig/m3, is calculated based on the concentration
 9    of Pb in the volume extracted, the size of the collection filter, and the volume of air drawn
10    through the filter.
11          The primary factors affecting the measurements made are the sampling frequency,
12    duration of sampling, type of inlet used, and the method of analyzing the filter for Pb content.
13    The following paragraphs  describe how these factors affect the Pb measurements.
14         2.4.1.1  Sampling Frequency
15          The frequency of Pb sampling used in the U.S. varies between one sample every day (1 in
16    1  sampling) to the more  common frequency of one sample  every 6 days (1 in 6 sampling). Semi-
17    continuous methods for the measurement of ambient metals (including Pb) are currently being
18    explored which would allow for more frequent sampling (as frequent as 1 sample per hour), but
19    more work is needed on these methods before they can be deployed in a network setting.
20          More frequent sampling reduces the uncertainty in estimates of quarterly or annual
21    averages associated with temporal variations in ambient concentrations. However, the costs of
22    sampling and analysis are  directly tied to sample  frequency. As such, it is necessary to evaluate
23    the reduction in measurement error versus the increase in sampling and analysis costs when
24    selecting the required sampling frequency. A discussion of the observed temporal variation of
25    Pb measurements is given later in this section.

26         2.4.1.2  Inlet Design
27          In ambient air monitors, a number of inlet designs have been developed that allow certain
28    particle size ranges to be sampled. The inlets use either impaction or cyclone techniques to
29    remove particles larger than a certain size (the size cutpoint) from the sample stream. Three
30    particle size cutpoints are used in ambient Pb measurements including total suspended PM
31    (TSP), PM less than or equal to 2.5 um in diameter (PM2.s), and PM less than or equal to 10 um
32    in diameter (PMio). The TSP inlet is designed to allow as much suspended particulate into the
33    sampling device as possible while protecting against precipitation and direct deposition on to the
34    filter (nominally 25 to 45 micrometers) (USEPA, 2004c).
            December 2006                     2-21               Draft - Do Not Quote or Cite

-------
 1          Sampling systems employing inlets other than the TSP inlet will not collect Pb contained
 2    in the PM larger than the size cutpoint.  Therefore, they do not provide an estimate of the total Pb
 3    in the ambient air.  This is particularly important near sources which may emit Pb in the larger
 4    PM size fractions (e.g., fugitive dust from materials handling and storage).

 5         2.4.1.3  Sample Analysis
 6          After the samples have been collected on filters and the filters have been weighed, the
 7    filters are analyzed for Pb content.  A number of analytical methods can be used to analyze the
 8    filters for Pb content including x-ray fluorescence analysis (XRF), proton-induced x-ray
 9    emission (PIXE), neutron activation analysis (NAA), atomic absorption (AA), or inductively-
10    coupled plasma mass spectrometry (ICP-MS) (CD, pp. 2-80 to 2-81).  A detailed discussion of
11    these methods was given in the 1986 CD, and the reader is referred to that document for more
12    information on these analytical methods. A search conducted on the Air Quality System
13    database 2shows that the method detection limits for all of these analytical methods (coupled
14    with the sampling methods) are very low, ranging from 0.01 ug/m3 to as low as 0.00001 |ig/m3,
15    and are adequate for NAAQS compliance purposes.

16         2.4.2   Pb Monitoring Programs
17          Ambient air Pb concentrations are measured by four monitoring networks in the United
18    States, all funded in whole or in part by EPA. These networks provide Pb measurements for 3
19    different size classes of airborne PM:  TSP, PM2.5, and PMio. The networks include the Pb
20    NAAQS compliance network, the PM2.5 Speciation Trends Network (STN), the Interagency
21    Monitoring of Protected Visual Environments (IMPROVE) network, and the National Air Toxics
22    Trends Stations (NATTS) network. The subsections below describe each network and the Pb
23    measurements made at these sites.  Comparisons of the data from these monitoring networks will
24    be discussed in section 2.4.4. Each network provides different types information on airborne Pb,
25    with the NAAQS compliance network providing data (on TSP Pb) most pertinent to this review.
26          In addition to these four networks, various organizations  have operated other sampling
27    sites yielding data on ambient air concentrations of Pb, often for limited periods and/or for
28    primary purposes other than quantification of Pb itself. Most of these data are accessible via the
29    Air Quality System. In an effort to gather as much air toxics data, including Pb, into one
30    database, the EPA and STAPPA/ALAPCO created the Air Toxics Data Archive. The Air Toxics
31    Data Archive can be accessed at:  http://vista.cira.colostate.edu/atda/.
            2 EPA's Air Quality System can be accessed at http://www.epa.gov/ttn/airs/airsaqs/

            December 2006                     2-22               Draft - Do Not Quote or Cite

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
     2.4.2.1  NAAQS Compliance Network
       This network is comprised of official state/local Pb monitoring stations which measure
Pb in TSP, i.e., particles up to 25 to 45 microns, for the purpose of determining compliance with
the Pb NAAQS. These stations use samplers and laboratory analysis methods which have either
Federal Reference Method (FRM) or Federal Equivalence Method (FEM) status. The FRM and
FEM method descriptions can be found in the U.S. Code of Federal Regulations, Section 40 part
50, Appendix G.  Sampling is conducted for 24-hour periods, with a typical sampling schedule of
1 in 6 days.  Some monitoring agencies "composite" samples by analyzing several consecutive
samples together to save costs and/or increase detection limits.
       The number of sites in the Pb NAAQS compliance network has decreased significantly
since the 1980s (see Figure 2-5).  The number of sites in the network reached its highest point in
1981 (946 sites). About 250 sampling sites operated during 2005.  This decline in the number of
Pb NAAQS compliance sites is attributable to the dramatic decrease in Pb concentrations
observed since the 1980s and the need to fund new monitoring objectives (e.g., PM2 5 and ozone
monitoring). Lead NAAQS compliance sites in lower concentration areas were shut down to
free up resources needed for monitoring of other pollutants such as PM2.5 and ozone.
           1000
        i
        W
        o
        I
             a.
             f?
             £
             •5
                 800
            600
            500
            400
            300
            200
            100
              0
                          \
                           \
                                                     Year
18   Figure 2-5.   Change in the number of Pb TSP monitoring sites from 1980 to 2005.
            December 2006
                                         2-23
Draft - Do Not Quote or Cite

-------
 1          The locations of sites in operation between 2003 and 2005 are shown in Figure 2-6.  The
 2   state/local agencies which operate these sites report the data to EPA's Air Quality System where
 3   they are accessible via several web-based tools. EPA's series of annual air quality trends reports
 4   have used data from this network to quantify trends in ambient air Pb concentrations. The most
 5   recent Trends report for Pb can be found at http://www.epa.gov/airtrends/lead.html.
 7   Figure 2-6.  Pb TSP monitoring sites: 2003-2005.
 8          A preliminary review of the Pb NAAQS Compliance Monitoring network's coverage of
 9   the highest Pb emitting sources (as identified in the current version of the 2002 NEI) was
10   conducted as part of preparing this draft document. This review indicates that many of the
11   highest Pb emitting sources in the 2002 NEI may not have nearby Pb NAAQS compliance
12   monitors. This preliminary review suggests that only 2 of 26 facilities (both Pb smelters3)
13   identified as emitting greater than 5 tpy have a Pb NAAQS compliance monitor within 1 mile.
14   We are currently completing a full review on the Pb NAAQS compliance network, including
15   quality assurance checks on details associated with monitor locations and aspects of the NEI
            3 Primary and secondary smelters were the source types given particular priority at the time of the last Pb
     NAAQS review (USEPA, 1990; USEPA, 1991).
            December 2006
2-24
Draft - Do Not Quote or Cite

-------
 1   sources, to confirm locations where Pb NAAQS compliance monitors are and should be located
 2   to ensure adequate monitoring around significant Pb sources. The findings of the full review will
 3   be described in the second draft of this document.

 4         2.4.2.2  PM2.5 Speciation Trends Network
 5          This is a U.S. network of about 200 PM2.5 speciation sites. This network consists of 54
 6   long-term trends sites [commonly referred to as the Speciation Trends Network (STN)] and
 7   about 150 supplemental sites. Most STN sites operate on a 1 in 3 day sampling schedule, while
 8   most supplemental sites operate on a 1 in 6 day sampling schedule. Nearly all of these state or
 9   locally operated sites are in urban areas, often at the location of highest known PM2.5
10   concentrations. Sites in this network determine the Pb concentrations in PM2.5 samples and, as
11   such, do not measure Pb in the size fraction >2.5  jim in diameter. Lead is quantified via the XRF
12   method.  The standard operating procedure for metals by XRF is available at:
13   http://www.epa.gov/ttnamtil/files/ambient/pm25/spec/xrfsop.pdf Data are managed through the
14   Air Quality System. These sites generally began operation around 2000.  The locations of these
15   sites are shown in Figure 2-7.
                                                            «  Trends
                                                            •:•  Supplemental
                                                            «  Tribal
16

17   Figure 2-7.  Pb PM2.5 (STN) monitoring sites.
18
            December 2006
2-25
Draft - Do Not Quote or Cite

-------
 1         2.4.2.3  IMPROVE Network - PM2.5 Speciation
 2          In the Interagency Monitoring of Protected Visual Environments (IMPROVE) network,
 3   PM2.5 monitors are placed in "Class I" areas (including National Parks and wilderness areas) and
 4   are mostly in rural locations. This network is administered by the National Park Service, largely
 5   with funding by EPA, on behalf of federal land management agencies and state air agencies that
 6   use the data to track trends in rural visibility.  Lead in the PM2 5 is quantified via the XRF
 7   method, as in the STN.  Data are managed and made accessible mainly through the IMPROVE
 8   website (http://vista.cira.colostate.edu/IMPROVE/), but also are available via the Air Quality
 9   System.  The oldest of these sites began operation in 1988, while many others began in the mid
10   1990s. The locations of these sites are shown in Figure 2-8.  There are 110 formally designated
11   "IMPROVE" sites located in or near national parks and other Class I visibility areas, virtually all
12   of these being rural. Approximately 80 additional sites at various urban and rural  locations,
13   requested and funded by various parties, are also informally treated as  part of the network.
14   Samplers are operated by several  different federal, state, and tribal host agencies on the same 1 in
15   3 day schedule as the STN.
                                      »    e
16

17   Figure 2-8.  Pb PM2.5 (IMPROVE) monitoring sites.
18


            December 2006                     2-26                Draft - Do Not Quote or Cite

-------
 1         2.4.2.4  National Air Toxics Trends Stations - PMi0 speciation
 2          The National Air Toxics Trends Stations (NATTS) network of 23  sites in mostly urban,
 3   but some rural, areas.  These sites are also operated by 21 state or local host agencies. All collect
 4   particulate matter as PMio for toxic metals analysis, typically on a 1 in 6 day sampling schedule.
 5   Lead in the collected sample is quantified via the ICP/MS method. The standard operating
 6   procedure for metals by ICP/MS is available at: http://www.epa.gov/ttn/amtic/airtox.html. These
 7   NATTS sites are relatively new, with 2004 being the first year in which all were operating. The
 8   Air Quality System can be accessed at http://www.epa.gov/ttn/airs/airsaqs/ (see Figure 2-9 for
 9   the locations of the NATTS monitoring sites).
10

11   Figure 2-9.  Pb PMio (NATTS) monitoring sites network.
12

13         2.4.3  Ambient Pb Concentrations, Trends and Spatial Patterns
14          The assessment of the available air quality at the time of the last NAAQS review (in and
15   just prior to 1990), described the dramatic changes in airborne Pb concentrations, primarily
16   associated with the reductions in use of leaded gasoline (USEPA, 1990). Given that change in
17   mobile source-related air concentrations, the focus of the last review was on areas near stationary
18   sources of Pb emissions (USEPA, 1990). Since that time, as described in Section 2.4.2, there
19   have been changes to the Pb TSP network and additional networks that produce Pb data have
            December 2006
2-27
Draft - Do Not Quote or Cite

-------
 1   been created. This section describes the available data that inform our current understanding of
 2   airborne Pb concentrations in U.S.
 4
 5
 6
 7
 8
 9
10
1 1
12
13
14
15
16
17
18
19
           2.4.3.1  PbinTSP
           2.4.3.1.1  Historical Trend
            Airborne concentrations of Pb in the United States have fallen dramatically over the last
     30 years due largely to the phase-out of leaded gasoline additives. Figure 2-10 shows the trend
     in overall U.S. airborne TSP Pb concentrations for a subset of the NAAQS FRM monitoring sites
     from 1983 through 2002. The data are plotted in terms of average per year of the maximum
     arithmetic mean averaged over a calendar quarter (the form of the current NAAQS) per
     monitoring site and are shown in relation to the current NAAQS of 1.5 ug/m3 (maximum
     quarterly average). The monitors used in this analysis are typically population-oriented urban
     monitors that are not source-oriented. Since 1983, major declines over several orders of
     magnitude have been observed not only in urban areas, but also in rural regions and remote
     locations.  The sharp decline through the 1980s has also been observed in Pb associated with fine
     particles (less than or equal to 2.5 microns) at remote and rural sites throughout the United States
     and have been attributed to the phase out of leaded gasoline (Eldred and Cahill, 1994).
                                                                         42 Sites
                                                            NAAQS
              - 90% of sites have concentrations below this line
                                                           1983-02: 94% decrease
                                                           1993-02: 57% decrease
          1.6
          1.4
          1.2
          1.0
          0.8
          0,6
          0.4
          0.2
          0.0
             83  84 85 86  87  88 89  90  91 92 93 94 95 96  97 98  99 00  01 02
                                             Year
     Figure 2-10. Airborne Pb (TSP) concentrations, averaged across continuously operating
                  monitoring sites: 1980-2002.
      CO
c
o
       u
       c
       o
      o
                                 70% of sites have concentrations below this line
            December 2006
                                               2-28
                                                           Draft - Do Not Quote or Cite

-------
 1         2.4.3.1.2 Current Concentrations
 2          Lead concentrations at very few locations in the U.S. (see discussion later in this section)
 3    exceed the current maximum quarterly average NAAQS. National average concentrations of
 4    TSP Pb measured in the NAAQS compliance network (i.e., via the FRM) are well below the
 5    NAAQS level of 1.5|ig/m3. The national composite average of annual means for all monitoring
 6    sites with at least one valid year of data was 0.08 |ig/m3 for the 3-year period, 2003-2005; the
 7    corresponding national composite median of the monitor level annual means was 0.02 |ig/m3.
 8          The distribution of the monitor site annual means is shown in Figure 2-11. This figure
 9    also shows the national distributions of: monitor level maximum quarterly average Pb
10    concentrations (i.e., the NAAQS metric); monitor level second maximum monthly average Pb
11    concentration (i.e., a candidate NAAQS replacement metric discussed in the last NAAQS
12    review); and monitor level average annual 98th percentile 24-hour concentration values (i.e., the
13    average across the 3 years of each year's 98th percentile value). The first three box-plots utilize
14    the same set of TSP Pb (FRM) data for 224 monitors.  The 98th percentile plot (198 total sites)
15    excludes data reported in 'composite' form (26 sites).
16          To be included in these TSP Pb characterization analyses, a site needed at least one
17    "complete" year consisting of at least three quarters of 6 or more observations.  One hundred
18    sixty one (161) of the 224 sites had complete  data for all three years (2003-2005), 35 monitors
19    had only two years of complete data; and 28 monitors had only one usable year of data.
20    [Excluding the 'composite' data (198 total sites), 140 sites had three years of utilized data, 31
21    sites had two years of data, and 27 sites had only one year of data.]
22          For 2003-2005, the national composite average of maximum quarterly mean Pb
23    concentrations was 0.16 |ig/m3; the corresponding national  composite median was 0.03 |ig/m3.
24    This median value is about fifty times lower than the 1.5 |ig/m3 NAAQS level.  For 2003-2005,
25    the national composite average of second maximum monthly average Pb concentrations was 0.18
26    |ig/m3 and the corresponding composite median was 0.03 |ig/m3. The monitor average 98th
27    percentile distribution is plotted on a different scale since those summary levels are much higher
28    than the other three  statistics.  The national composite mean of average 98th percentile
29    concentrations was 0.37 |ig/m3 and the corresponding median was 0.04 |ig/m3.
            December 2006                     2-29               Draft - Do Not Quote or Cite

-------
      O)
                Annual
                 mean
Maximum
quarterly
 mean
                                                             tu,
                                                          monthly mean
                                                                         .5-
          Annual 98th
           percentile
Statistic
Annual Mean
Maximum Quarterly Mean
2nd Maximum Monthly Mean
Annual 98th Percentile
# Sites
224
224
224
198
Maximum
1.450
4.093
5.022
8.260
95th
Percentile
0.371
0.719
0.968
1.907
75th
Percentile
0.043
0.100
0.109
0.185
Mean
0.075
0.155
0.183
0.373
Median
0.015
0.027
0.028
0.038
25th
Percentile
0.006
0.009
0.010
0.017
5th
Percentile
0.002
0.004
0.004
0.006
Mimimum
0.001
0.001
0.001
0.001
 2    Figure 2-11. Distribution of TSP Pb concentrations (represented by 4 different statistics)
 3                 at monitoring sites, 2003-2005.4
 4           For all four metrics plotted in Figure 2-11, the national means are substantially higher
 5    than the national medians. This is due to a small number of monitors with significantly higher
 6    levels. These monitors with higher concentrations are almost exclusively associated with
 7    industrial point sources. If source-oriented monitors were eliminated from the national level
 8    statistics shown in Figure 2-11, all of the national level statistics would be significantly lower
 9    and the means would be more comparable to the medians.
10           Figure 2-12 re-plots the monitor level maximum quarterly means and distinguishes
11    between the source-oriented monitors and those not identified as  such5. This plot shows that 95
12    percent of all monitors not identified as being source-oriented had a maximum quarterly  average
13    of 0.15 |ig/m3 or less (which is one tenth of the NAAQS  level). Almost 25 percent of the sites
              Box depicts inter-quartile range and median; whiskers depict 5th and 95th percentiles; asterisks identify
      composite averages.
             5 Sites were classified as ' source-oriented' if they were within one mile of an facility emitting at least one
      ton of Pb per year and/or they were previously identified as such (using a 2003 reference file).
             December 2006
            2-30
Draft - Do Not Quote or Cite

-------
1    identified as being source-oriented had maximum quarterly averages of 0.75 |ig/m3 or more
2    (which is 50 percent of the current NAAQS level).
 7
 8
 9
10
1 1
12
                                 CO
                                           Source-
                                           oriented
                                                            Not source-
                                                             oriented

Source-oriented sites
Not source-oriented sites
# Sites
59
165
Maximum
4.093
0.447
95th
Percentile
1.923
0.147
75th
Percentile
0.695
0.039
Mean
0.484
0.037
Median
0.252
0.016
25th
Percentile
0.081
0.007
5th
Percentile
0.019
0.003
Mimimum
0.010
0.001
5    Figure 2-12.  Distribution of monitor level TSP Pb annual mean concentrations for source-
6                 oriented and not sourced-oriented monitors, 2003-2005.6

            The monitor level values for two of the four discussed statistical metrics (annual average,
     and maximum quarterly mean) are mapped in Figures 2-13 and 2-14.  As seen when comparing
     these figures, the locations of the high concentration levels for both metrics are generally the
     same. In fact, there is significant correlation among all four of the monitor level summary
     metrics discussed above; see  Table 2-5.
           6 Box depicts inter-quartile range and median; whiskers depict 5th and 95th percentiles; asterisks identify
     composite averages.
           December 2006
                                                2-31
Draft - Do Not Quote or Cite

-------
                                                           > 0.50 pg/m3 (8 sites)
                                                           0.10 - 0.50 |jg/m3 (30 sites)
                                                           0.03-0.10|jg/m3(31 sites)
                                                           < 0.03 |jg/m3 (155 sites)
2   Figure 2-13. Site level TSP Pb, annual mean concentrations, 2003-2005.
                          December 2006
2-32
Draft - Do Not Quote or Cite

-------
                                                          9 > 1.5 pg/m3 (3 sites)
                                                             0.7-1.5|jg/m3(15sites)
                                                           o 0.1 -0.7 |jg/m3 (28 sites)
                                                           • <0.1 |jg/m3( 178 sites)
2   Figure 2-14.  Site level TSP Pb, maximum quarterly mean concentrations, 2003-2005.
                          December 2006
2-33
Draft - Do Not Quote or Cite

-------
 1          In the past several years only about one or two FRM sites per year have had measured
 2    maximum quarterly average TSP Pb levels that exceeded the NAAQS level (1.5 |ig/m3, as a
 3    maximum quarterly average).  These sites are shown in Table 2-6. Two areas are officially
 4    designated as nonattainment for the Pb NAAQS: East Helena Area portion of Lewis and Clark
 5    Counties, Montana7; and the area within the city limits of Herculaneum in Jefferson County,
 6    Missouri (http://www.epa.gov/air/oaqps/greenbk/lnca.html).

 7    Table 2-5.  Correlation among different TSP site-level statistics, 2003-2005.

Annual Mean
Maximum Quarterly Mean
2nd Maximum Monthly Mean
Annual 98th Percentile
Annual
Mean
1.00



Maximum
Quarterly
Mean
0.85
1.00


2nd Max
Monthly
Mean
0.86
0.99
1.00

Annual 98th
Percentile
0.94
0.93
0.93
1.00
10
11
12
13
14
15
16
17
18
19
20
21
Table 2-6.  FRM sites with Pb concentrations above the level of the current NAAQS, based
            on maximum quarterly average, 2003-2005.
State
AL
IN
MO
Area
Pike County
Delaware County
Herculeneum City
nonattainment area (in
Jefferson County)
Source
type
Secondary Pb
smelter
Secondary Pb
smelter
Primary Pb
smelter
Site
011090003
180350009
180350009
290990015
290990015
290990015
year
2003
2004
2004
2005
2005
2005
quarter
4
2
3
1
2
3
quarterly
mean
1.92
4.09
2.64
1.93
1.61
1.73
     2.4.3.1.3  Variability
       Some seasonal variability is common for air Pb concentrations.  However, the extent to
which seasonal variability is present depends on precipitation trends, changes in wind direction,
and mixing height variability for a given area. For monitors situated near Pb point sources,
factors related to the facilities' operations also contribute to temporal variability.
       Figure 2-15 plots monthly TSP Pb averages for the 2003-2005 time period for four
example sites. The two sites on the left, both source-oriented, have some of the highest
            7 The source associated with this area closed in early 200land monitoring ceased in late 2001.
            December 2006
                                          2-34
Draft - Do Not Quote or Cite

-------
 1   concentrations in the nation. The two sites on the right are presumably not source-oriented; their
 2   annual average concentrations are much lower, in fact, close to the national median.  Each of the
 3   four sites has unique monthly patterns.  The two sites on the top (one source-oriented and the
 4   other not) appear to have recurring seasonal patterns.  The monthly variation for the source site is
 5   probably related to the nearby source's  operations. The variation in the other site may be due to
 6   similar factors albeit from a smaller and/or further away emission source but there probably also
 7   is more of a meteorological impact. The two sites plotted on the bottom have more random
 8   variation in their monthly averages than the two on top.  In general, source oriented sites (such as
 9   the two on the left) typically have significantly more variation in their monthly averages than do
10   monitors that are not source oriented  (such as the two on the right). Note the wide relative range
11   of scale for the two left plots compared to the tight range for the two right plots. This difference
12   in magnitude of variation is illustrated by the ratios of highest monthly average to lowest
13   monthly average over the 3-year period for the four sites, 43 (top left) and 26 (bottom left) for
14   the two source-oriented locations and 4 for the two non-source-oriented locations.
            December 2006                     2-3 5                Draft - Do Not Quote or Cite

-------
           State=FL County=Hillsborough  Source-oriented
   State=MI County=Wayne   Not source-oriented
            2.0-


            1.8-


            1.6-


            1.4-


            1.2-


            1.0-


            0.8-


            0.6-


            0.4-


            0.2-


            o.o-
                                            9  10   11   12
               State=IN County= Delaware  Source-oriented
                               n
                                                                        1234567
                                                                                                     10  11   12
                                                                   State=IL County=Cook  Not Source-oriented
                                                                  0.051
                 1234567
                                             9   10  11   12
                                                                        1234567
                                                                                                     10  11   12
2   Figure 2-15. Monthly average TSP Pb concentrations at 4 example monitor sites, 2003-2005.
                            December 2006
2-36
Draft - Do Not Quote or Cite

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

14
15
16
17
18
19
20
21
22
            Monthly variation at near-source locations is better characterized by short-term averaging
     times (e.g., monthly) than longer-term averaging times (e.g., yearly). This is demonstrated in
     Table 2-7.  This table shows the number of TSP monitors (in the 224 site database used above)
     that exceed average levels of 0.5 to 1.5 |ig/m3 with averaging times or forms of 3-year, one year,
     quarterly, and maximum monthly, and second maximum monthly. For example, with a stated
     level equal to the current standard of 1.5 |ig/m3, no sites in this database exceed with an
     averaging time of 3 years, 2 sites exceed with an averaging time of 1 year, 3 sites exceed with a
     quarterly averaging time,  4 sites exceed based on the 2nd maximum monthly average and 11  on
     the first maximum monthly average . Using the lowest level examined, 0.5 |ig/m3, however,  10
     sites would exceed that level with an averaging time of 3 year; 16 sites would exceed that level
     with an averaging time of one year;  19 sites would exceed that level with a quarterly averaging
     time; 21 monitors would exceed that level with their second highest monthly average, and 32
     monitors would exceed that level with a maximum monthly average form.

     Table 2-7.   Comparison of numbers of sites that exceed various TSP Pb levels using
                 different averaging times or forms, 2003-2005.
Level
(ljg/m3)
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
Number of monitors that exceed level
3-year
annual
avg.
10
7
5
3
2
2
2
1
1
1
0
Max
annual
average
16
11
8
5
4
3
3
3
3
2
2
Max
quarterly
average
19
18
17
11
11
8
7
7
7
4
3
2nd max
monthly
average
21
18
15
15
14
12
9
8
8
5
4
Max
monthly
average
32
28
23
21
19
17
16
14
13
12
11
           2.4.3.2   Pb in PM2.5
            As noted in section 2.4.2 above, there are two national monitoring programs that collect
     ambient PM2.5 Pb information.  The EPA STN focuses mainly on urban areas and the IMPROVE
     network focuses mostly on rural environments, specifically those classified as "Class 1" areas
     (including National Parks and wilderness areas).  Figure 2-16 shows the STN site level
     distributions of annual means, maximum quarterly means, second maximum monthly means and
            December 2006
                                              2-37
Draft - Do Not Quote or Cite

-------
 1    annual 98th percentiles for 2003-2005.8 For this 3-year period, the national composite average
 2    of annual means for all sites with at least one valid year of data was 0.005 |ig/m3; the
 3    corresponding national composite median of the monitor level annual means was 0.004 |ig/m3.
 4    Both levels are more than an order of magnitude less than the similar TSP data illustrated in
 5    Figures 2-11.
              Annual
               mean
                               Maximum
                               quarterly
                                 mean
  2nd max
monthly mean
Annual 98th
 percentile
Statistic
Annual Mean
Maximum Quarterly Mean
2nd Maximum Monthly Mean
Annual 98th Percentile
# Sites
272
272
272
272
Maximum
0.0453
0.1681
0.1980
0.4715
95th
Percentile
0.0106
0.0178
0.0217
0.0439
75th
Percentile
0.0051
0.0075
0.0091
0.0170
Mean
0.0049
0.0078
0.0095
0.0189
Median
0.0038
0.0054
0.0062
0.0120
25th
Percentile
0.0033
0.0041
0.0047
0.0086
5th
Percentile
0.0021
0.0030
0.0035
0.0056
Mimimum
0.0010
0.0016
0.0023
0.0029
 7
 8

 9
10
Figure 2-16.  Distribution of PM2.5 Pb concentrations (represented by four different
              statistics) at STN sites, 2003-2005. 9
Figure 2-13 maps the 2002-2005 STN site level annual averages.
               To be included in PM2.5 Pb characterization analyses, a site needed at least one "complete" year
      consisting of at least three quarters of 11 or more observations. One hundred eighty two (182) of the 272 STN sites
      had complete data for all three years (2003-2005), 47 sites had only two years of complete data; and 43 sites had
      only one usable year of data.
             9 Box depicts inter-quartile range and median; whiskers depict 5th and 95th percentiles; asterisks identify
      composite averages.
             December 2006
                                             2-38
             Draft - Do Not Quote or Cite

-------
1

2
                                                         >0.150|jg/m3(5sites)
                                                         0.010 - 0.150 |jg/m3 (9 sites)
                                                       o  0.005-0.010|jg/m3 (39 sites)
                                                       •  < 0.005 |jg/m3 (219 sites)
Figure 2-17. Site level 'urban' (STN) PM2.5 Pb annual means, 2003-2005.
                          December 2006
                                                      2-39
Draft - Do Not Quote or Cite

-------
 1          For PM2.5 Pb measured in the IMPROVE program, levels are even lower. The 2003-
 2   2004 monitor level median and mean annual average levels are both less than 0.001 |ig/m3.
 3   Levels measured in the IMPROVE program are considerably lower than those obtained in the
 4   PM2.5 STN network, reflecting the fact that speciation monitors are generally located in urban
 5   areas while the IMPROVE sites are in national parks and wilderness  areas.  Published studies
 6   have also reported that concentrations of airborne Pb are sometimes several orders of magnitude
 7   higher in urban areas compared to remote regions (Schroeder et al., 1987; Malm and Sisler,
 8   2000). Rural areas tend to have Pb concentrations falling  somewhere between those of urban
 9   and remote areas.  Thus, urban populations are typically exposed to distinctly higher levels of
10   airborne Pb than rural or remote residents.

11         2.4.3.3  PbinPM10
12          Figure 2-18 shows distributions of PMi0 Pb site level annual means, max quarterly
13   means, second max monthly means, and 98th percentile concentrations for 2003-2005.10 For this
14   3-year period, the national composite average of annual means for all monitors with at least one
15   valid year of data was 0.007 |ig/m3; the corresponding national composite median of the monitor
16   level annual means was 0.006 |ig/m3.
            10 To be included in these PM10 Pb characterization analyses, a site needed at least one "complete" year
     consisting of at least three quarters of 11 or more observations. Five (5) of the 29 sites had complete data for all
     three years (2003-2005), 9 sites had only two years of complete data; and 15 sites had only one usable year of data.

            December 2006                     2-40               Draft - Do Not Quote or Cite

-------
                      0,03-
                              Annual
                               mean
                                        Maximum
                                        quarterly
                                         mean
  2nd max
monthly mean
Annual 98th
 percentile
Statistic
Annual Mean
Maximum Quarterly Mean
2nd Maximum Monthly Mean
Annual 98th Percentile
# Sites
29
29
29
29
Maximum
0.0203
0.0547
0.0426
0.2602
95th
Percentile
0.0151
0.0390
0.0260
0.0662
75th
Percentile
0.0088
0.0153
0.0120
0.0351
Mean
0.0067
0.0126
0.0112
0.0299
Median
0.0059
0.0098
0.0100
0.0161
25th
Percentile
0.0035
0.0048
0.0056
0.0095
5th
Percentile
0.0022
0.0027
0.0032
0.0060
Mimimum
0.0016
0.0027
0.0029
0.0060
2
O


4

5
Figure 2-18.  Distribution of PMio Pb concentrations (represented by four different
              statistics), 2003-2005.11
       The PMio Pb site means are mapped in Figure 2-19.
            11 Box depicts inter-quartile range and median; whiskers depict 5th and 95th percentiles; asterisks identify
     composite averages.
            December 2006
                                             2-41
            Draft - Do Not Quote or Cite

-------
1

2
                                                       • >0.150|jg/m3(1 site)
                                                          0.010-0.150|jg/m3(4sites)
                                                        o  0.005 - 0.010 |jg/m3 (13 sites)
                                                        •  < 0.005 |jg/m3 (11 sites)
Figure 2-19. Monitor level PM10 Pb annual means, 2002-2005.
                          December 2006
                                                       2-42
Draft - Do Not Quote or Cite

-------
 2         2.4.4  Relationships among Different Particle-sized Pb Concentration
 3          As described in Sections 2.4.1.2 and 2.4.2., airborne Pb concentrations are measured in
 4   three PM size fractions - TSP, PMio and PM2 5 - by the various monitoring networks. Figure 2-
 5   20 summarizes the annual means for the various PM size fractions and networks.  The TSP
 6   monitor averages for "all sites" and the "source-oriented" subset dwarf the other averages. An
 7   inset figure re-plots on a different scale the Pb averages for the TSP non source oriented
 8   monitors, the PMio NATS monitors, the PM2.5 STN monitors, and the PM2.5 IMPROVE
 9   monitors.  The TSP non source oriented averages are about 5 times larger than the PMio
10   averages; the PMio averages are about 1.5 times the PM2.5 urban averages; and the PM2.5 urban
11   averages are about 4 times the PM2 5 rural averages.
12
                  0.60
                  0.50
                  0.40
                  0.30
                  0.20
                  0.10
                  0.00
13

14
15
Statistic
TSP - all sites
TSP - source-oriented
TSP not source-oriented
PMIO(NATS)
PM2.5 - STN (mostly urban)
PM2.5 - IMPROVE (Nat. park/rural)
Obs
224
59
165
29
272
167
Maximum
1 .4502
4.0931
0.4467
0.0203
0.0453
0.0058
95th
Percentile
0.3711
1.9233
0.1467
0.0151
0.0106
0.0031
75th
Percentile
0.0431
0.6953
0.0386
0.0088
0.0051
0.0015
Mean
0.0750
0.4840
0.0373
0.0067
0.0049
0.0012
Median
0.0148
0.2523
0.0157
0.0059
0.0038
0.0009
25th
Percentile
0.0058
0.0813
0.0070
0.0035
0.0033
0.0005
5th
Percentile
0.0021
0.0193
0.0031
0.0022
0.0021
0.0004
Mimimum
0.0010
0.0100
0.0010
0.0016
0.0010
0.0003
Figure 2-20.  Comparison of national mean and median monitor level Pb, annual means for
             different size cut PM networks, 2002-2005.
            December 2006
                                         2-43
Draft - Do Not Quote or Cite

-------
 1
 2
 3
 4
 5
 6
 7
10
11
12
       There are not many sites where Pb measurements are made in different PM size fractions
at the same location and the same day (and where Pb values exceed the minimum detection
limit). Lead is measured in all three PM size fractions in only a few locations in the United
States. Table 2-8 shows Pb concentrations for 2003-2005 from four such monitoring sites:  one
located in Wayne Co., MI (Detroit); one located in St. Louis City, MO, and two located in Davis
County, Utah (Ogden). None of these sites are known to be source-oriented. At all four sites,
the majority of the TSP Pb appears to be in the PMi0 size cut. The first two sites have a slight
majority of their PMio TSP in the fine-sized fraction, but the Ogden, UT sites have significantly
more PMio Pb in the coarse-sized fraction.

Table 2-8.  Pb concentrations (jig/m3), at four sites, in different PM size fractions: 2003-
            2005.
Area
Wayne, Co., Ml
St. Louis City, MO
Davis County, UT
Davis County, UT
Site
261630033
295100085
490110001
490110004
Number of
collocated obs
143
53
22
11
Average for common days
TSP - Pb
0.023
0.013
0.030
0.007
PM10-Pb
0.020
0.012
0.031
0.007
PM2.5-Pb
0.011
0.008
0.004
0.002
             Note: Although Pb was measured in TSP, PM10, and PM2 5 at the same site in each of the above three locations,
             different PM monitoring methods (collection and/or analysis) were used for the different PM size fractions,
13           contributing to apparent anomaly of PM-|0 Pb value being higher than TSP Pb value for one of the Utah sites.
14
15           In a combined analysis of data from all co-located monitoring sites, there is typically a
16    good correlation between Pb measurements in TSP and PMio (average site level r of 0.96 at four
17    sites with 10+ paired observations) and to a lesser extent between Pb measurements in PMio and
18    PM2.5 (average r = 0.62 for 18 sites with 10+ paired observations).  The correlation between Pb
19    measurements in TSP and PM2.5 is generally quite lower (average r of = 0.38 for 35 sites with
20    10+ paired observations).  There is substantial variability in the correlation between Pb
21    concentrations in TSP and PM2.5 samples at different sites.  For those sites with at least 10 paired
22    observations, the correlation coefficients range from 0.00  to >0.99.
23           As described in the CD, several studies have investigated Pb concentrations in different
24    PM size fractions (CD, p. 3-13). For example, average Pb concentrations reported in a rural area
25    in the southeastern U.S. were 6.11 ng/m3 in PM2.5 and 15.04 ng/m3 in TSP samples, with the
26    average total mass concentration of 9.5 |ig/m3 and 19.1 |ig/m3 for PM2.5 and TSP, respectively
27    (Goforth and Christoforou, 2006); thus, Pb constituted a similar very small proportion  of
28    particles in each size fraction. Another study included two areas in the Los Angeles basin (Singh
29    et al, 2002).  In Downey, a site where refineries and traffic contribute heavily to particle
            December 2006
                                           2-44
Draft - Do Not Quote or Cite

-------
 1    concentrations, Pb was proportionally greater in the fine and ultrafine fractions of PMi0. In
 2    Riverside, which is considered a receptor site for particles transported from the Los Angeles
 3    basin and also has agricultural sources, Pb was proportionally greater in the coarse fraction of
 4    PMio. In Boston, MA, Pb concentrations of 326 ng/m3 and 75.6 ng/m3 were reported from PM2.5
 5    and PMio-2.5 (Thurston and Spengler, 1985).  Overall, these findings indicate that for locations
 6    primarily impacted by combustion sources, Pb concentrations appear to be higher in the fine
 7    fraction of particles. However, at locations impacted by mining and material handling sources,
 8    Pb contained in the larger particles can be of significantly higher concentrations than those for
 9    the fine particles.

10         2.4.5   Modeling Estimates (NATA- National Scale Assessments)
11          As part of the Agency's national air toxics assessment (NATA) activities, a national scale
12    assessment of hazardous  air pollutants including Pb compounds has been performed twice over
13    the past few years (USEPA 2006c, 2002c, 200 la). These two assessments included the use of the
14    NEI for the years 1996 and 1999,  respectively, with atmospheric dispersion modeling to predict
15    associated annual average Pb air concentrations across the country. A national scale assessment
16    is not yet available based on the 2002 NEI. A number of limitations are associated with the 1996
17    and 1999 ambient concentration estimates (see Section 2.4.5.2) and the underlying emissions
18    estimates (e.g., see Section 2.3.3). While the associated limitations handicap a reliance on the
19    absolute magnitude of these estimates, they may prove informative with regard to relative
20    patterns of concentrations across the country, and are presented in that light.

21         2.4.5.1  Methods
22          To develop national-scale  estimates of annual average ambient Pb concentrations, EPA
23    used the Assessment System for Population Exposure Nationwide (ASPEN) model. ASPEN
24    uses a Gaussian model formulation and climatological data to estimate long-term average
25    pollutant concentrations.  The ASPEN model takes into account important determinants of
26    pollutant concentrations,  such as:  rate of release, location of release, the height from which the
27    pollutants are released, wind speeds and directions from the meteorological stations nearest to
28    the release, breakdown of the pollutants in the atmosphere after being released (i.e., reactive
29    decay), settling of pollutants out of the atmosphere (i.e., deposition), and transformation of one
30    pollutant into another (i.e., secondary formation). ASPEN concentration estimates do not
31    account for day-of-week  or seasonal variations in emissions (USEPA, 200la).
32          For each source, the model calculates ground-level concentrations as a function of radial
33    distance and direction from the source at a set of receptors laid out in a radial grid pattern.  For
34    each grid receptor, concentrations are calculated for each of a standard set of stability class/wind
35    speed/wind direction combinations. These concentrations are averaged together using the annual

            December  2006                    2-45                Draft - Do Not Quote or Cite

-------
 1    frequency of occurrence of each combination (i.e., the climatology) as weightings to obtain
 2    annual average concentrations (USEPA, 200la). For the 1999 NATA assessment,
 3    meteorological data for 1999 were used and the frequency distributions were also stratified by
 4    time of day into eight 3-hour time blocks. This along with similar emission rate stratification
 5    helps to preserve any characteristic diurnal patterns that might be important in subsequent
 6    estimation of population exposure. The resulting output of ASPEN is a grid of annual average
 7    concentration estimates for each source/pollutant combination by time block (USEPA, 200la).
 8          Annual average concentration estimates for grid receptors surrounding each emission
 9    source are spatially interpolated to the census tract centroids within the 50 kilometers impact
10    zone, and contributions from all modeled sources are summed to give cumulative ambient
11    concentrations in each census tract. By accounting for all identified source categories (including
12    background concentrations, which are added to the ASPEN-calculated concentrations), the sum
13    of the concentration increments yields an estimate  of the overall Pb concentration within each
14    census tract.  For many pollutants modeled, total concentrations include a "background"
15    component which includes concentrations due to natural sources, sources not in the emissions
16    inventory, and long-range transport (USEPA, 200la).  In the case of Pb, however, a background
17    concentration value of zero was used.

18         2.4.5.2  Findings and Limitations
19          Historical studies show that Gaussian dispersion models, such as ASPEN, typically agree
20    with monitoring data within  a factor of 2 most of the time. In the case of Pb in the NATA
21    assessment, model estimates at monitor locations were generally lower than the monitor averages
22    for Pb, suggesting that the modeling system (i.e., emissions estimates, spatial allocation
23    estimates, dispersion modeling) may be systematically underestimating ambient concentrations.
24    This may be particularly true for Pb as metals tend to deposit rapidly with distance from the
25    source according to their particle size and weight.  The model-to-monitor analysis is described in
26    detail at http://www.epa.gov/ttn/atw/nata 1999/99compare.html.  The modeling system
27    underestimation may also be due in part to  a lack of accounting for emissions re-entrainment
28    (these "re-entrained" particles may be observed by the monitors, but they are not accounted for in
29    the emissions inventory, and thus would not contribute to the model estimate). For more details
30    on the limitations of the 1999 NATA national scale assessment,  see
31    http://www.epa.gov/ttn/atw/natal999/limitations.html.
32          Because higher Pb concentrations are associated with localized sources,  which are not
33    well-characterized by this modeling approach, national scale assessments such as this can only
34    provide answers to questions about emissions, ambient air concentrations, exposures and risks
35    across broad geographic areas (such as counties, states and the country) for that period. They  are


            December 2006                    2-46              Draft - Do Not Quote or Cite

-------
 1    also based on assumptions and methods that limit the range of questions that can be answered
 2    reliably such as identifying Pb exposures and risks for specific individuals, or identifying
 3    exposures and risks in small geographic regions such as a specific census tract.
 4          Given the limitations of this analysis with regard to estimating Pb concentrations
 5    nationally (see above), specific absolute ambient concentration estimates for Pb compounds
 6    generated by this analysis are not presented here.  The general pattern of results, presented
 7    elsewhere (USEPA, 2006c), is consistent with the following conclusions: 1) there are Pb
 8    concentrations projected in remote areas; 2) there are distinct geographical variations in ambient
 9    Pb concentrations; concentrations in rural areas are generally much lower than in urban areas;
10    and, 3) there are areas with high Pb concentrations associated with localized sources with high
11    emissions. These results also support the general conclusion that more detailed source and site
12    specific analyses are needed when addressing Pb impacts.

13         2.4.6   Air Quality Summary
14          Ambient air Pb concentrations are measured by four monitoring networks in the United
15    States, all funded in whole or in part by EPA.  These networks - the Pb NAAQS compliance
16    network, the PM2.5 STN, the PM2.5 IMPROVE network, and the PMi0 network - provide Pb
17    measurements for 3 different sizes of PM, and the PM2.5 size is measured separately in urban and
18    remote locations.
19          Airborne concentrations of TSP Pb in the United States have fallen dramatically over the
20    last 30 years due largely to the phase out of leaded gasoline additives. Despite this decline, there
21    have still been a small number of areas that have  not met the current Pb NAAQS over the past
22    few years. The sources of Pb in these areas are stationary sources (e.g. primary and/or secondary
23    smelters).  Except for the monitors in a limited number of areas, TSP Pb averages are quite low
24    with respect to the NAAQS.  The median monitor level maximum quarterly average for 2003-
25    2005 is about fifty times lower than the 1.5 |ig/m3 NAAQS level. However, there appears to be
26    significant 'under-monitoring' near known Pb emission point sources.
27          Some monthly variability is common for ambient Pb concentrations.  The current form of
28    the standard (quarterly average) attempts to account for seasonal variability.  As suggested
29    during the last review, a shorter averaging period  (monthly) would better capture short-term
30    increases  in Pb concentrations (USEPA 1990). Although there have only been 3 sites that
31    violated the 15. |ig/m3 max quarterly average NAAQS during the 2003 - 2005 period, 11  sites
32    violated that level with respect to a maximum monthly  average.
33          There are not many sites that collect ambient Pb data in all three size ranges. Analyses of
34    co-located Pb size data indicate that TSP-sized Pb and PMio-sized Pb are well correlated.  If
            December 2006                    2-47               Draft - Do Not Quote or Cite

-------
 1    further analyses corroborate this finding, specifically for source-oriented sites, PMio Pb
 2    measurements may be useful as a TSP Pb surrogate.
 3          The NATA national scale assessment estimates based on 1999 NEI reflect the quantity
 4    and distribution of Pb emissions, with the highest estimates associated with point sources. For
 5    example, the census tract with the highest estimated Pb concentration is located in the county
 6    with the highest Pb emissions estimate in the 1999 NEI, and the second highest census tract is
 7    located in a county with a now-closed major Pb smelter.  Limitations of the assessment,
 8    however, seem to contribute to uncertainty and potential underestimation of Pb concentrations.

 9         2.5   ATMOSPHERIC DEPOSITION
10          As  described in Section 2.2.2, deposition is the path by which Pb particles are removed
11    from the atmosphere and transferred to other environmental media, and, as discussed further in
12    Chapters 4 and 6, deposited Pb, plays a major role in human and ecological exposures. There are
13    several approaches described in the literature for estimating atmospheric deposition, or transfer
14    of Pb from the atmosphere to soil or water bodies. These include measurements of Pb in rainfall
15    (wet deposition) and on collection surfaces during dry periods (dry deposition); dry deposition
16    has also been estimated via measurements of airborne Pb particles coupled with estimates of
17    deposition  velocity (see CD, Section 2.3.2).  Studies that measure Pb in sediment or soil cores,
18    coupled with isotope dating methods (see CD, Sections 2.2.1  and 8.1.2), provide observations
19    informative of atmospheric deposition rates and trends. As there are currently no nationwide Pb
20    atmospheric deposition monitoring programs, the information in this section is drawn from a
21    variety of sources as discussed in the CD.

22         2.5.1   Temporal Trends
23          The available atmospheric studies of dry, wet and bulk deposition of Pb indicate a
24    pronounced downward trend in Pb deposition in the U.S. during the 1980s to early 1990s, likely
25    reflecting the reduction in atmospheric levels during that time period (CD, Section 2.3.2). As an
26    example, Pirrone and others (1995) estimated an order of magnitude reduction in dry deposition
27    from 1982  to 1991 in Detroit, Michigan (CD, Section 2.3.2).  Measurements of Pb in rainfall in
28    Lewes, Delaware (small town at mouth of Delaware Bay) have fallen from approximately 3 ug/L
29    in the early 1980s to less than 1 ug/L by  1989 (CD, pp. 2-60 and AX7-35; Scudlark et al., 1994).
30    Sediment core studies provide evidence of the larger historical pattern (CD, Section 2.3.1).  For
31    example, Jackson and others (2004) reported that deposition to the Okefenokee Swamp, Georgia,
32    USA peaked during the period from 1940s through 1970s, followed by a period of steady decline
33    into the 1990s (CD, Section 2.3.1).
            December 2006                     2-48               Draft - Do Not Quote or Cite

-------
 1         2.5.2   Deposition Flux Estimates since the Last Review
 2          Contemporary rates of total Pb loadings to terrestrial ecosystems are estimated at
 3    approximately 1 to 2 mg/m2year (CD, p. AX7-36). In association with the Great Lakes Water
 4    Quality Agreement between the United States and Canada, a deposition monitoring network was
 5    established to estimate regional atmospheric inputs to the Great Lakes (Voldner and Eisenreich,
 6    1989). Based on measurements from that network, total Pb deposition to the Great Lakes (Lakes
 7    Superior, Michigan and Erie) in the early 1990s was estimated to be on the order of 1.5 -2
 8    mg/m2-year (CD, pp. 2-57 and 2-60; Sweet et al., 1998).
 9          For Lakes Superior and Michigan, dry deposition estimates were greater than those for
10    wet deposition by a factor of 1.5 to 2, while dry deposition to Lake Erie was estimated to be less
11    than 80% of wet deposition (CD, pp. 2-57 and 2-60; Sweet et al., 1998).  In the mid-Atlantic
12    region during the 1990s, dry deposition was estimated to be equal to or lower than wet
13    deposition, contributing <50% of total deposition (CD, Section 2.3.2; Scudlark et al., 2005).
14    Reports of wet deposition for this region during the 1990s range from nearly 400 to just over 600
15    ug/m2-year (CD, Section 2.3.2).

16         2.6  OUTDOOR DUST AND SOIL
17          Lead in outdoor dust and  soil may be derived from a range of sources including current
18    and historical air emissions sources, as well as miscellaneous non-air sources (e.g., land disposal
19    of wastes and subsequent weathering). Both media may play a substantial role in human and
20    ecological exposures. With regard to human exposures, contaminated soil can be a potential
21    source of Pb exposure, particularly for children (CD, Section 3.2).  Another source of children's
22    exposure, as discussed in the CD (Sections  3.2 and 4.4), is house dust, which may  be derived
23    from Pb  in outdoor dust and soil as well as  from ambient air Pb.
24         2.6.1   Outdoor Dust
25           Outdoor dust refers to particles deposited on outdoor surfaces. Lead in outdoor dust has
26    been associated with active point sources as well as well as older urban areas. For example, a
27    50% reduction in dust Pb levels accompanied a 75% reduction in airborne Pb concentrations
28    associated with replacement of a  smelting facility in Canada (CD, pp. 3-23 to 3-24).
29    Additionally, Caravanos and others (2006b) have described Pb in dust (particulate matter)
30    deposited on surfaces in New York City.  Lead levels have been found to be higher in dust on or
31    near roadways, or in  older urban  areas as compared to newer or rural areas (CD, Sections 3.2.3
32    and 3.2.4; Caravanos et al  2006a,b).  As with surface soil, contact with outdoor dust may
33    contribute to incidental ingestion of environmental contaminants including Pb. Additionally,  as
34    stated in the CD (Section 2.3.3), the "re-suspension of soil-bound Pb particles and contaminated


            December 2006                     2-49               Draft - Do Not Quote or Cite

-------
 1    road dust is a significant source of airborne Pb". Re-suspension, thus, provides a pathway for Pb
 2    transport into residences and its contribution to Pb in house dust. As mentioned in Section 2.2.1,
 3    particles containing Pb may be resuspended into the air by a range of processes including wind
 4    and vehicular traffic, as well as other mechanical processes including pedestrian traffic,
 5    agricultural operations, and construction.

 6         2.6.2  Soil
 7           A reservoir of 0.5 to 4 g/m2 gasoline additive-derived Pb is estimated to exist in U. S.
 8    soils (CD, p. AX7-36), with most contained in the upper soil horizons  (O + A horizons).  Studies
 9    have indicated that industrial Pb can be strongly sequestered by organic matter and by secondary
10    minerals such as clays and oxides of Al, Fe, and Mn, (CD, pp. AX7-24 to AX7-39).
11    Accordingly, migration (e.g, to groundwater) and biological uptake of Pb in ecosystems is
12    considered to be relatively low, with variability of Pb mobility in different systems influenced by
13    factors including elevation and climate, vegetation type, acidity, and soil composition (CD,
14    Sections 2.3.5 and AX7.1.2.3). Generally then, forest floors are considered to currently act as
15    net sinks for Pb, and burial or movement of Pb over time down into lower soil/sediment layers
16    also tends to sequester it away from more biologically active parts of the watershed, unless later
17    disturbed or redistributed (CD, p. AX7-36). In areas of exposed soil, however, there is potential
18    for interaction with airborne Pb (as discussed in Sections 2.6.1 and 2.2.1).
19           As discussed below (Section 2.6.2.1), findings to date indicate those  systems less
20    influenced by point sources still responding to reduced Pb  deposition rates associated with
21    reduced atmospheric emissions of Pb, including those associated with  the phase-out of leaded
22    gasoline (see Section 2.3.3).  Situations near point sources and those involving historically
23    deposited Pb near roadways are less well characterized.  Section 2.6.2.2 summarizes estimates of
24    soil Pb concentrations  since the time of the last review.

25         2.6.2.1   Temporal Trends
26           Variability among soil systems in characteristics influencing Pb mobility contributes to
27    differences in current and projected temporal  trends in soil concentrations (e.g., CD, pp. 3-18 to
28    3-19, Sections 3.2.1-3.2.2, and pp. AX7-33 to AX7-34).
29           Studies of forest soils have concluded that the time for soils to  respond to reduced Pb
30    deposition rates (e.g., associated with Pb gasoline phase-out) is shorter than previously believed.
31    For example, Miller and Friedland (1994) projected that a 37% reduction in Pb concentration in
32    northern hardwood and subalpine forest soils would occur within 17 years and 77 years,
33    respectively. Kaste and Friedland (2003) traced atmospherically deposited Pb within forest soils
34    in Vermont and found  similar response times of 60 and 150 years for the two forest soils,
35    respectively.  They also concluded that the penetration of atmospherically delivered Pb in  soils is

             December 2006                     2-50               Draft - Do Not Quote  or Cite

-------
 1    currently limited to the upper 20 cm and that the heterogeneous distribution of Pb in soils would
 2    seem to indicate that the release of Pb to groundwater will be dispersed thereby reducing the
 3    likelihood of a large pulse to groundwater.  This study and those of Wang and Benoit (1997),
 4    Johnson et al. (1995), and Zhang (2003) conclude that forest surface soils do not act as sinks
 5    under current deposition rates for Pb and that a gradual migration into mineral soils is occurring,
 6    making the possibility of a large pulse to groundwater in the future from past Pb pollution
 7    unlikely.  Studies of the role of acidification in Pb mobility in sandy soils  (e.g., NJ pine barrens),
 8    however, suggest a greater risk of mobilization of Pb and organic matter into these mineral soils,
 9    with subsequent inputs to associated stream waters (CD, p. AX7-91).
10           Studies in urban areas of southern California, where Pb has accumulated from past
11    sources, suggests an environment in which Pb may remain at the soil surface (and other
12    surfaces), contributing to air concentrations via re-suspension for the near-term (CD, pp. 2-65 to
13    2-67 and 3-18 to 3-19). Figure 2-21 illustrates how the temporal trend in surface soil
14    concentrations at a location is considered to be influenced by the rate of re-suspension. Harris
15    and Davidson (2005) suggested that typical long-term values for re-suspension rate fall in the
16    range of 10"11 to 10"7 per second, based on wind speeds, with the range of 10"11 to 10"10 proposed
17    as a range appropriate to California's south coast air basin. Under these assumptions, the model
18    illustrated that the occurrence of re-suspension at this rate, would lead to little to no reduction in
19    soil Pb concentration in southern California over the next few hundred years (CD, pp. 2-65 to 2-
20    67 and 3-18 to 3-20).
            December 2006                     2-51               Draft - Do Not Quote or Cite

-------
              E
             7i>
              c
              o
             ~
              £
             **
              c
              o
              0
              c
              o
             O
          A =
          ,\ = ixi(r10/s
          A = 1x10-%
          A = 1x10'8/s
                      1970
                    2000
 /•N                                           I \*Cll

 3     Source: Reprinted from Harris and Davidson (2005). Units for re-suspension rate (A) are per second (/s).
 4    Figure 2-21.  Modeled soil concentrations of Pb in the South Coast Air Basin of California
 5                 based on four re-suspension rates (A).
 6           Temporal trends in surface soils near established point sources are not well characterized.
 7    Information described in the CD for areas surrounding smelters after implementation of pollution
 8    controls, although showing declines in Pb concentrations in outdoor dustfall, street dust and
 9    indoor dustfall, has not indicated a noticeable decline in soil Pb concentrations (CD, pp. 3-23 to
10    3-24). Further, Pb concentrations in "clean" soil placed in areas influenced by current sources
11    have been demonstrated to exhibit increasing temporal trends (USEPA, 2006d).  Concentrations
12    of Pb in the very top layer of material (within the upper 1 inch of soil, analyzed using portable x-
13    ray fluorescence) at locations less than a mile from  a primary Pb smelter exhibited statistically
14    significant increasing concentration over a four year period, with the average monthly  change in
15    Pb concentration ranging from 1 to 8 mg/kg (USEPA, 200 Ib, 2006d). Estimates of associated
16    steady-state surface soil Pb concentrations or the expected longer-term temporal pattern for this
17    situation have not been made.

18         2.6.2.2  Current Surface Soil Concentrations
19           Present concentrations of Pb in forest surface soils range from 40 to 100 mg/kg while
20    natural background levels would be expected to be  <1 mg/kg (CD, Section AX7.1.2.3). Urban
21    and roadside soils and those in areas of long-term Pb emissions from point sources have much
            December 2006
2-52
Draft - Do Not Quote or Cite

-------
 1    higher concentrations of Pb, ranging up to hundreds to tens of thousands of mg/kg (CD, Section
 2    3.2.1). For example, Pb surface soil concentrations near smelters has been found to range from
 3    1000s of mg/kg (dry weight) within approximately 100-250 meters, dropping to 200 mg/kg and
 4    below by distances of approximately 3-5 km (CD, Table 3-4).  Soil Pb concentrations of 500-800
 5    mg/kg have been reported near U.S. mines that are no longer active (CD, Table 3-6).

 6         2.7  SURFACE WATER AND SEDIMENT
 7          The primary source of Pb in aquatic systems is atmospheric deposition. Lead is also
 8    carried into water bodies via wastewater effluent from municipalities and industry, stormwater
 9    runoff, erosion, and accidental discharges  (CD, p.  AX7-142).  Most Pb occurring in aquatic
10    systems is associated with particles, with the distribution between particle-bound and dissolved
11    form being influenced by water chemistry as well  as suspended sediment levels (CD, pp. AX7-
12    117 to AX7-118; CD, Section AX7.2.2). The ratio of Pb in suspended solids to Pb in filtrate has
13    been described to vary from 4:1 in rural streams to 27:1  in urban streams (CD, p. AX7-118).
14          Water columns have been described as "transient reservoirs" for pollutants (CD, p.  2-75).
15    Once deposited to sediments, whether Pb is available for re-suspension back into the water
16    column with potential transport further down a watershed versus being buried into deeper
17    sediments depends on the aquatic system.  In open ocean waters (generally characterized by
18    depth and distance from continental sources), re-suspension to surface waters is unlikely.  In
19    more shallow systems, and additionally those influenced by land sources (e.g., stormwater runoff
20    as well as point sources), re-suspension may play a significant role in water column
21    concentrations. For example, studies in San Francisco Bay, the southern arm of which as an
22    average depth of 2 m, have indicated that Pb particles may be remobilized from surface
23    sediments into the water column (CD, AX7-141).

24         2.7.1   Temporal Trends
25          As discussed in the CD, many studies have investigated trends in Pb concentration in
26    sediment and surface waters (CD, Section AX7.2.2), with declines documented in many systems
27    and usually attributed to the phasing out of leaded gasoline.
28          Using sediment cores, temporal changes in Pb deposition and associated sediment Pb
29    concentration have been documented. In sediment cores from the Okefenokee Swamp, Pb
30    concentrations were approximately 0.5 mg/kg prior to industrial development, reached a
31    maximum of approximately 31  mg/kg from about  1935 to 1965, and following passage of the
32    Clean Air Act in 1970 concentrations have declined to about 18 mg/kg in 1990 (CD, p. AX7-
33    141). Researchers investigating trends in metals concentrations (roughly from 1970-2001) in
34    sediment cores from 35 reservoirs and lakes in urban and reference settings found that number of
            December 2006                     2-53               Draft - Do Not Quote or Cite

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

15
16
17
18
     lakes exhibiting decreasing trends in Pb concentration outnumbered increasing trends (83%
     versus 6%).  Mass accumulation rates of Pb in cores, adjusted for background concentrations,
     decreased from the 1970s to the 1990s, with a median change of 246%. The largest decreases
     were found in lakes located in dense urban watersheds, although anthropogenic mass
     accumulation rates in dense urban lakes remained elevated over those in lakes in undeveloped
     watersheds, indicating that urban fluvial source signals can overwhelm those from regional
     atmospheric sources (CD, p. AX7-141; Mahler et al, 2006).
            Figure 2-22 presents data on Pb concentrations in lake sediments from 12 lakes in the
     Great Lakes area.  Consistent with other studies, this study showed a peak in Pb concentrations
     consistent with peak use of leaded gasoline in the U.S. in the mid 70's and declining
     concentrations in most lake sediments through the mid 1990's.
                1995
                                        Litttefield*
                                       •Elk
                                        Higgins
                                        Mullett
                                                       Crystal M
                                                       Gratiot
                                                       Cadillac
                                                       Paw Paw
     Cass
     •Gull
     • Crystal B
     Whitmore
                1815
                    0.00      0.20      0.40      0.60      0.80      1.00
                           Lead sediment concentrations normalized to peak
            Source: Yohn et al. (2004).
                                                                                1.20
     Figure 2-22.  Pb concentrations in sediment samples in 12 Michigan lakes. The
                  concentrations are normalized by the peak Pb concentration in each lake;
                  peak Pb concentrations ranged from approximately 50 to 300 mg/kg.
            December 2006
                                              2-54
Draft - Do Not Quote or Cite

-------
 1         2.7.2   Current Concentrations
 2          An analysis of data from the United States Geological Survey (USGS) National Water-
 3    Quality Assessment (NAWQA) program is described in the CD. The NAWQA data set
 4    encompasses data, collected over the past 15 years, on Pb concentrations in flowing surface
 5    waters, bed sediment, and animal tissue for more than 50 river basins and aquifers throughout the
 6    country (CD, Section AX7.2.2.3). Based on analysis of these data, the mean dissolved Pb
 7    concentration in ambient surface waters of the U.S. is estimated to be 0.66 ug/L (range 0.04 to 30
 8    (J^g/L), as compared to a mean of 0.52 ug/L (range 0.04 to 8.4 ug/L) for the "natural" locations.
 9    The term ambient was used by NAWQA to describe the combined contribution of natural and
10    anthropogenic sources, and a separate set of samples was identified for natural locations (e.g.,
11    "forest", "rangeland", and "reference" sites).  The mean concentration of Pb in ambient bulk
12    sediment (<63 microns, grain size) is  120 ug/g dry weight (range 0.5 to 12,000 ug/g), as
13    compared to a mean of 109 ug/g dry weight (range 0.5 to 12,000 ug/g).
14          Geographic distribution of Pb concentrations in surface waters and sediments in this data
15    set are presented in Figures 2-23 and 2-24 (CD, Figures AX7-2.2.7 and AX7-2.2.9). Areas with
16    high surface water Pb concentrations were observed in Washington, Idaho, Utah, Colorado,
17    Arkansas, and Missouri, with the maximum measured Pb concentration occurring at a site in
18    Idaho with a land use classified as mining (CD, p. AX7-131). As was seen with surface water Pb
19    concentrations, the highest measured sediment Pb concentrations were found in Idaho, Utah, and
20    Colorado.  And also similar to the surface water findings, of the top 10 sediment Pb
21    concentrations recorded, 7 were measured at sites classified as mining land use (CD, p. AX7-
22    133).
23
            December 2006                    2-55               Draft - Do Not Quote or Cite

-------
                              Gf
I
2
                                                      Legend
                                                         Surface Water Dissolved Pb (pg/L)
                                                        o  Non-detect
                                                        0  0.51 - 5.44 (<99th precentiie)
                                                        • 5.45 - 29.78 (>99th percentile)
                                                                    6
                                                                    h
3    Figure 2-23. Spatial distribution of dissolved lead in surface water (N = 3445). [CD, Figure
4                 AX7-2.2.7.]
            December 2006
2-56
Draft - Do Not Quote or Cite

-------
i
2
                               •    o
                               «• .  a
                                                    Legend
                                                       Bulk Sediment <63u Total Lead (M9/9)
                                                         0 0.50- 18,00 (1st Quartile)
                                                         0 18.01 - 28.00 (2nd Quartile)
                                                         • 28.01 - 49.00 (3rd Quartile)
                                                         • 49.01 - 12000.00 (4th Quartile)
                                                  *.
                                                   \
                                                   v
                                                   ?•
                                                'A^"2
3
4
5
6
7
Figure 2-24. Spatial distribution of total lead in bulk sediment <63 um (N = 1466). [CD,
             Figure AX7-2.2.9]
       As described in the CD, dissolved surface water concentrations reported for lakes have
been generally much lower than the NAWQA values for lotic waters (CD, AX7-138).
           December 2006
                                          2-57
Draft - Do Not Quote or Cite

-------
  1     REFERENCES
  2

  3     Calspan Corporation. Assessment of Industrial Hazardous Waste Practices in the Metal Smelting and Refining
  4            Industry. Volume III: Ferrous Smelting and Refining. Prepared for EPA's Office of Solid Waste. No. SW-
  5            145c.3 1977.

  6     Caravanos, I; Weiss, A.L.; Jaeger, RJ. (2006a) An exterior and interior leaded dust deposition survey in New York
  7            City: results of a 2-year study. Environ. Res. 100: 159-164.

  8     Caravanos, I; Weiss, A.L.; Blaise, M. I; Jaeger, R.J. (2006b) A survey of spatially distributed exterior dust lead
  9            loadings in New York City. Environ. Res. 100: 165-172.

10     Chevron Products Company (2000) Aviation Fuels Technical Review IFTR-3.
11            http://www.chevron.com/products/prodserv/fuels/bulletin/aviationfuel/pdfs/aviation fuels.pdf

12     Eastern Research Group. (2002a) Development of Average Emission Factors and Baseline Emission Estimates for
13            the Industrial, Commercial, and Institutional Boilers and Process Heaters NESHAP. Memorandum to Jim
14            Eddinger, Office of Air Quality Planning and Standards, U.S. EPA. October, 2002. Docket number - OAR-
15            2002-0058-0022.

16     Eastern Research Group. (2002b) National Emission Trends for Large Municipal Waste Combustion Units, (years
17            1090 to 2005). Memorandum to Walt Stevenson. June 17, 2002, EPA Docket A-90-45 / Item VIII-B-7;

18     Eastern Research Group. (2002c) National Emission Trends for Small Municipal Waste Combustion Units. Memo
19            to Walt Stevenson. June 12, 2002, EPA Docket A-98-18 / Item VI-B-2

20     EC/R Incorporated. (2006) Secondary Lead Smelter Industry - Source Characterization for Residual Risk
21            Assessment. Prepared for USEPA Office of Air and Radiation, Office of Air Quality Planning and
22            Standards, Research Triangle Park, NC. November.

23     Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the United
24            States of America. Atmos. Environ. 28: 1009-1019.

25     Goforth, M. R.; Christoforou, C. S. (2006) Particle size distribution and atmospheric metals measurements in a rural
26            area in the South Eastern USA. Sci. Total Environ. 356: 217-227.

27     Habibi K. (1973) Characterization of Paniculate Matter in Vehicle Exhaust. Environ Sci & Technol 7(3):223-234.

28     Harrison, R. M.; Tilling, R.; Callen Romero, M. S.; Harrad, S.; Jarvis,  K. (2003) A study of trace metals and
29            polycyclic aromatic hydrocarbons in the roadside environment. Atmos. Environ. 37: 2391-2402.

30     Jackson, B.P., P.V. Winger, P.J. Lasier (2004) Atmospheric lead deposition to Okefenokee Swamp, Georgia, USA.
31            Environ Poll. 130: 445-451. Abu-Allaban, M.; Gillies, J. A.;  Gertler, A. W.; Clayton, R.; Proffitt, D.
32            (2003) Tailpipe, resuspended road dust, and brake-wear emission factors from on-road vehicles. Atmos.
33            Environ. 37: 5283-5293.

34     Johnson, C. E.; Siccama, T.  G.; Driscoll, C. T.; Likens, G. E.; Moeller, R. E. (1995) Changes in lead
35            biogeochemistry in response to decreasing atmospheric inputs. Ecol. Appl. 5: 813-822.

36     Kaste, J.; Friedland, A.; Stirrup, S. (2003) Using stable and radioactive isotopes to trace atmospherically deposited
37            Pb in montane forest soils. Environ. Sci. Technol. 37: 3560-3567.

38     Lehigh University. 1982. Characterization, Recovery, and Recycling of Electric Arc Furnace Dust. Final report
39            prepared for the U.S. Department of Commerce. February 1982.


              December 2006                        2-58                  Draft - Do Not Quote or Cite

-------
  1     Lough, G. C.; Schauer, J. I; Park, J.-S.; Shafer, M. M; Deminter, J. T.; Weinstein, J. P. (2005) Emissions of metals
  2            associated with motor vehicle roadways. Environ. Sci. Technol. 39: 826-836.

  3     Mahler, B. J.; Van Metre, P. C.; Callender, E. (2006) Trends in metals in urban and reference lake sediments across
  4            the United States, 1970 to 2001. Environ. Toxicol. Chem. 25: 1698-1709.

  5     Malm, W. C.; Sisler, J. F. (2000) Spatial patterns of major aerosol species and selected heavy metals in the United
  6            States. Fuel Process. Technol. 65: 473-501.

  7     Miller, E.K. and Friedland, A.J. (1994) Lead migration in forest soils: Response to changing atmospheric inputs.

  8     O'Hare, A. 2006. Email to Michele Price from Andy O'Hare, Portland Cement Association. February 28.

  9     Pirrone, N.,  GJ. Keeler, P.O. Warner. (1995) Trends of ambient concentrations and deposition fluxes of paniculate
10            trace metals in Detroit from 1982 to 1992. Sci Total Environ 162: 43-61.

11     RTI International. (2005) Summary of EPA's 2004 Survey of Minimills. June.

12     RTI International. (2006) Characterization of the Glass Manufacturing Industry, Glass Manufacturing Area Source
13            NESHAP. Memorandum to Susan Fairchild, Office of Air Quality Planning and Standards. May 5

14     Schell, W. R.; Barnes, R. S. (1986) Environmental isotope and anthropogenic tracers of recent lake sedimentation.
15            In:  Fritz, P.; Fontes, J. C., eds. Handbook of environmental isotope geochemistry; V. 2, the terrestrial
16            environment, B. New York, NY: Elsevier Science Publishers; pp. 169-206.

17     Schroeder, W. H.; Dobson, M.; Kane, D. M.; Johnson, N. D. (1987) Toxic trace elements associated with airborne
18            paniculate matter: a review. JAPCA 37: 1267-1285.

19     Scudlark, J.; Church, T.; Conko, K.; Ondov, J.; Han, M. (1994) The wet deposition of trace elements onDelmarva
20            and their utility as emission source indicators. Annapolis, MD: Maryland Department of Natural Resources,
21            Chesapeake Bay Research and Monitoring Program; report no. CBRM AD 94 3. Available from: NTIS,
22            Springfield, VA; PB94 178373. Available: http://esm.versar.com/pprp/bibliography/CBRM AD 94
23            3/CBRM AD 94 3 .pdf [24 August, 2006].

24     Scudlark, J.R.; Rice, K.C.; Conko, K.M.; Bricker, O.P.; Church, T.M. (2005) Transmission of atmospherically
25            derived trace  elements through an undeveloped, forested Maryland watershed. Water Air Soil Poll 163:53-
26            79.

27     Stevenson, W. (2002) Emissions from Large MWCs at MACT Compliance. Memo to Docket from Walt Stevenson.
28            EPA Docket a-90-45 / Item VIII-B-11.

29     Sweet, C. W.; Weiss, A.; Vermette, S. J. (1998) Atmospheric deposition of trace metals at three sites near the Great
30            Lakes. Water Air Soil Pollut. 103: 423 439.

31     Singh, M.; Jaques, P. A.; Sioutas, C. (2002) Size distribution and diurnal characteristics of particle-bound metals in
32            source and receptor sites of the Los Angeles Basin. Atmos. Environ. 36:  1675-1689.

33     Thurston, G. D.; Spengler, J. D. (1985) A quantitative assessment of source contributions to inhalable paniculate
34            matter pollution in metropolitan Boston. Atmos. Environ. 19: 9-25.

35     U.S. Environmental Protection Agency. (1990) Review of the national ambient air quality standards for lead:
36            assessment of scientific and technical information: OAQPS staff paper. Research Triangle Park, NC: Office
37            of Air Quality Planning and Standards;  report no. EPA-450/2-89/022. Available from: NTIS, Springfield,
38            VA; PB91-206185 and http://www.epa.gov/ttn/naaqs/standards/pb/data/rnaaqsl asti.pdf
              December 2006                         2-59                  Draft - Do Not Quote or Cite

-------
  1    U.S. Environmental Protection Agency. (1991) U.S. EPA Strategy for Reducing Lead Exposure. Available from
  2            U.S. EPA Headquarters Library/Washington, D.C. (Library Code EJBD; Item Call Number: EAP
  3            100/1991.6; OCLC Number 2346675).

  4    U.S. Environmental Protection Agency. (1997) National Emission Standards for Hazardous Air Pollutants for
  5            Secondary Lead Smelters: Direct Final Rule. Amendments to Rule. 13 June 1997.  Federal Register
  6            Volume 62, No. 114, page 32209.  Available at: http://www.epa.gov/ttn/atw/lead2nd/lead2pg.html

  7    U.S. Environmental Protection Agency. (1998) Study of Hazardous Air Pollutant Emissions from Electric Utility
  8            Steam Generating Units - Final Report to Congress. Office of Air Quality Planning and Standards. EPA
  9            453/R-98-004a. February.

10    U.S. Environmental Protection Agency. (1999a) National Emission Standards for Hazardous Air Pollutants for
11            Primary Lead Smelters: Final Rule.  4 June 1999.  Federal Register, Volume 64, No. 107, page 30194.
12            Available at:  http://www.epa.gov/ttn/atw/mactfnlalph.html

13    U.S. Environmental Protection Agency. (1999b) National Emission Standards for Hazardous Air Pollutants for
14            Portland Cement Manufacturing: Final Rule. 14 June 1999. Federal Register, Volume 64, No. 113.
15            Available at:  http://www.epa.gov/ttn/atw/pcem/pcempg.html

16    U.S. Environmental Protection Agency. (200la) National-Scale Air Toxics Assessment for 1996. EPA-453/R-01-
17            003. Research Triangle Park, NC.

18    U.S. Environmental Protection Agency. (200Ib) Quality Assurance Project Plan for a Site Characterization at the
19            Herculaneum Lead Smelter. Herculaneum, Missouri, CERCLIS ID No.: MOD 006266373. Prepared for
20            USEPA, Region VII, Superfund Division by USEPA Region VII Superfund Technical Assessment and
21            Response Team 2. September 10.

22    U.S. Environmental Protection Agency. (2002a) PBT national action plan for alkyl-Pb. Washington, DC: Persistent,
23            Bioaccumulative, and Toxic Pollutants (Pbt) Program. [13 October, 2005]  Available:
24            http://www.epa.gov/opptintr/pbt/cheminfo.htm

25    U.S. Environmental Protection Agency. (2002b) National Emission Standards for Hazardous Air Pollutants
26            (NESHAP) for Iron and Steel Foundries-Background Information for Proposed Standards. EPA-453/R-02-
27            013. Office of Air Quality Planning and Standards, Research Triangle Park, NC. December.

28    U.S. Environmental Protection Agency. (2002c) 1996 National Scale Air Toxics Assessment. Office of Air Quality
29            Planning and Standards,  http://www.epa.gov/ttn/atw/nata/

30    U.S. Environmental Protection Agency. (2002d) National Emission Standards for Hazardous Air Pollutants for
31            Primary Copper Smelters: Final Rule. 12 June 2002. Federal Register, Volume 67, No. 113, page 40478.
32            Available at:  http://www.epa.gov/ttn/atw/mactfnlalph.html

33    U.S. Environmental Protection Agency. (2003a) Emission estimates for integrated iron and steel plants.
34            Memorandum to Docket, February 3, 2003. Document no. IV-B-4 in Docket No. OAR-2002-0083

35    U.S. Environmental Protection Agency. (2003b) National air quality and emissions trends report. 2003 special
36            studies edition. Research Triangle Park, NC: Office of Air Quality  Standards; Emissions Monitoring and
37            Analysis Division; report no. EPA 454/R-03-005. (27 August, 2004). Available:
38            http://www.epa.gov/air/airtrends/aqtrnd03/toc.html

39    U.S. Environmental Protection Agency. (2003c) National Emission Standards for Hazardous Air Pollutants for
40            Integrated Iron and Steel Manufacturing: Final Rule.  20 May 2003. Federal Register, Volume 68, No. 97.
41            Available at: http://www.epa.gov/ttn/atw/iisteel/iisteelpg.html
              December 2006                        2-60                  Draft - Do Not Quote or Cite

-------
  1     U.S. Environmental Protection Agency. (2004a) National Emission Standards for Hazardous Air Pollutants for
  2            Industrial/Commercial/Institutional Boilers and Process Heaters: Final Rule.  13 September 2004. Federal
  3            Register, Volume 69, No. 176. Available at: http://www.epa.gov/ttn/atw/boiler/boilerpg.html

  4     U.S. Environmental Protection Agency. (2004b) National Emission Standards for Hazardous Air Pollutants for Iron
  5            and Steel Foundries; Final Rule. Federal Register 69(78): 21906-21940.  April 22.

  6     U.S. Environmental Protection Agency. (2004c) Air Quality Criteria for Particulate Matter. Volume I. EPA 600/P-
  7            99/002aF-bF, Washington, DC. Pages 1-4.

  8     U.S. Environmental Protection Agency. (2005) "Technical Support Document for HWC MACT Replacement
  9            Standards, Volume V: Emission Estimates and Engineering Costs," September 2005, Appendix C.

10     U.S. Environmental Protection Agency. (2006a) National Emissions Inventory for 2002, version 2.1, draft. Office of
11            Air Quality Planning and Standards, Research Triangle Park, NC. November 28.

12     U.S. Environmental Protection Agency. (2006b) Compilation of Air Pollutant Emission Factors, Volume 1:
13            Stationary Point and Area Sources. AP 42, Fifth Edition. Office of Air Quality Planning and Standards.
14            Current version available:  http://www.epa.gov/ttn/chief/ap42/index.html

15     U.S. Environmental Protection Agency. (2006c) 1999 National Scale Air Toxics Assessment. Office of Air Quality
16            Planning and Standards, http://www.epa.gov/ttn/atw/nata 1999/

17     U.S. Environmental Protection Agency. (2006d) Lead soil trend analysis through May, 2006. Evaluation by
18            individual quadrant. Herculaneum lead smelter site, Herculaneum, Missouri. Prepared by TetraTech for
19            U.S. EPA, Region 7.  Available on the web, at:
20            http://www.epa.gov/region7/cleanup/superfund/herculaneum_pbtrend thru may2006.pdf

21     Voldner, E.G. and Eisenreich, S.J. (1989)^4 Plan for Assessing Atmospheric Deposition to the Great Lakes, Water
22            Quality Board, International Joint Commission, Windsor, Ontario.

23     Wang, E. X.; Benoit, G. (1997) Fate and transport of contaminant lead in spodosols: a simple box model analysis.
24            Water Air Soil Pollut. 95: 381-397.

25     Zhang, Y.-H. (2003) 100 years of Pb deposition and transport in soils in Champaign, Illinois, U.S.A. Water Air Soil
26            Pollut. 146: 197-210.
              December 2006                         2-61                  Draft - Do Not Quote or Cite

-------
 1            3   POLICY-RELEVANT ASSESSMENT OF HEALTH EFFECTS
 2                                              EVIDENCE

 3         3.1   INTRODUCTION
 4          This chapter assesses key policy-relevant information on the known and potential health
 5    effects associated with exposure to ambient lead (Pb).  The presentation here summarizes the
 6    qualitative assessment of health evidence contained in the CD, as a basis for the evidence-based
 7    assessment of primary standards for Pb that will be presented in Chapter 5 of the second draft of
 8    this document. The focus is on health endpoints associated with the range of exposures
 9    considered to be most relevant to current exposure levels. This presentation also gives particular
10    attention to those endpoints for which there is quantitative health evidence available in this
11    review that provides a foundation for the quantitative health risk assessment discussed in Chapter
12    4 and used in the risk-based assessment of primary standards for Pb that will be presented in
13    Chapter 5 of the second draft of this document.
14          The presentation in this chapter recognizes several key aspects of the health evidence for
15    Pb.  First, because exposure to atmospheric Pb particles occurs not only via direct inhalation of
16    airborne particles, but also via ingestion of deposited particles (e.g., associated with soil and
17    dust), the exposure being assessed is multimedia and multi-pathway in nature, occurring via both
18    the inhalation and ingestion routes.  Second, the exposure index or dose metric most commonly
19    used and associated with health effects information is an internal biomarker (i.e., blood Pb).
20    Additionally, the exposure duration of interest (i.e., that influencing internal dose pertinent to
21    health effects of interest) may span months to potentially years, as does the time scale of the
22    environmental processes influencing Pb deposition and fate. Lastly, the nature of the evidence
23    for the health effects of greatest interest for this review is epidemiological data strongly
24    supported by toxicological data that provide biological plausibility and insights on mechanisms
25    of action.
26          At the time of the last review, Pb was recognized to produce multiple effects in a variety
27    of tissues and organ systems across a range of exposure levels, with blood Pb levels of 10-15
28    ug/dL being recognized as levels of concern for impaired neurobehavioral development in
29    infants and children (USEPA,  1990).  The current CD recognizes the existence of a wide array of
30    Pb-induced deleterious effects, including several in children and/or adults that are induced by
31    blood Pb levels extending well below 10 ug/dL, to below 5 ug/dL and possibly lower (CD,
32    Section 8.4).
33          In recognition of the multi-pathway aspects of Pb, and use of an internal exposure metric
34    in health risk assessment,  Section 3.2 describes our understanding of the internal disposition or
35    distribution of Pb, and the use  of blood Pb as an internal exposure or dose metric.  Section 3.3

            December  2006                     3-1                Draft - Do Not Quote or Cite

-------
 1    discusses the nature of Pb-induced health effects, giving emphasis to those with the strongest
 2    evidence, particularly those associated with the range of current exposure levels. Potential
 3    impacts of Pb exposures on public health, including recognition of potentially susceptible or
 4    vulnerable subpopulations, is discussed in Section 3.4.  Finally, Section 3.5 summarizes key
 5    policy-relevant conclusions about Pb-related health effects.

 6         3.2   INTERNAL DISPOSITION - BLOOD LEAD AS DOSE METRIC
 7           The health effects of Pb (discussed in the CD and summarized in Section 3.3 below) are
 8    remote from the portals of entry to the body (i.e., the respiratory system and gastrointestinal
 9    tract).  Consequently, the internal disposition and distribution of Pb is an integral aspect of the
10    relationship between exposure and effect. This section summarizes the current state of
11    knowledge of Pb disposition pertaining to both inhalation and ingestion routes of exposure (as
12    described in the CD).
13           Inhaled Pb particles deposit in the different regions of the respiratory tract as a function
14    of particle size (CD, pp. 4-3 to 4-4). Lead associated with smaller particles, which are
15    predominantly deposited in the pulmonary region, may, depending on solubility, be absorbed
16    into the general circulation or transported via phagocytic cells to the gastrointestinal tract (CD,
17    pp. 4-3). Lead associated with larger particles, that are predominantly deposited in the upper
18    respiratory tract (e.g., nasal pharyngeal and tracheobronchial regions), may be transported by
19    mucociliary transport into the esophagus and swallowed, thus making its way to the
20    gastrointestinal tract (CD, pp. 4-3 to 4-4), where it may be absorbed into the blood stream.
21           The absorption efficiency of Pb from the gastrointestinal (GI) tract varies with particle
22    size, as well as with the chemical form or matrix in which it is contained (CD, pp. 4-8 to 4-9).
23    One line of evidence for this comes from research using animal models to estimate relative
24    bioavailability (KB A) by comparing the absorbed fraction of ingested Pb for different test
25    materials relative to that for a highly water-soluble form of Pb. Relative bioavailability of Pb
26    from contaminated soils from different industrial sites (e.g., near Pb smelters, mines, etc), as
27    assessed in such models, have been found to differ markedly, with RBA values ranging from 6  to
28    100% (CD, pp. 4-8 to 4-10; Casteel et al., 2006).  As stated in the  CD, "variations in size and
29    mineral content of the Pb-bearing grains are the suspected cause of variations in the rate and
30    extent of GI absorption of Pb" occurring in soil from different contaminated locations (CD, p. 4-
31    9).
32           In addition to characteristics associated with the ingested Pb, GI absorption of Pb also
33    varies with an individual's physiology (e.g., maturity of the GI tract), and nutritional  status (e.g.,
34    iron and calcium deficiency increases absorption), as well as the presence of food in the GI tract
35    (CD, Section 4.2.1, pp. 4-5 to 4-8).   With regard to GI tract maturity, estimates  of Pb GI

             December 2006                      3-2                Draft - Do Not Quote or Cite

-------
 1    absorption reported in the past for young children (-40-50%) are higher than those reported for
 2    adults (CD, pp. 4-5 to 4-6). Several studies have reported that the presence of food in the GI
 3    tract reduces the absorption of water-soluble Pb (CD, p. 4-6).  A contributing factor to this
 4    phenomenon is the presence of calcium, iron, and phosphate in the food, which depresses Pb
 5    absorption (CD, pp. 4-6 to 4-7). Animal studies have also indicated that Vitamin D, which
 6    regulates calcium absorption,  enhances Pb absorption from the GI tract (CD, p. 4-7).
 7          Once in the blood stream, where approximately 99% of the Pb associates with red blood
 8    cells, the Pb is distributed throughout the body, with the bone serving as a large, long-term
 9    storage compartment, and soft tissues (e.g., kidney, liver, brain, etc) serving as smaller
10    compartments, in which Pb may be more mobile (CD, Sections 4.3.1.4 and 8.3.1.). Lead
11    accumulates in the bone during childhood development, and this accumulation continues through
12    adulthood. For example, more than 90% of the total Pb body burden in adults is stored in the
13    bone, while the storage in bone accounts for approximately 70% of a child's body burden (CD,
14    Section 4.2.2).
15          As described in the CD, Pb is exchanged between blood and bone and blood and soft
16    tissues (CD, Section 4.3.2). The exchanges between the blood and bone vary with "duration and
17    intensity of the exposure, age  and various physiological variables" (CD, p. 4-1). For example,
18    resorption of bone (e.g., in pregnant or nursing women, or associated with  osteoporosis in
19    postmenopausal women), results in a mobilization of Pb from bone into circulation (CD,
20    Sections 4.3.2.4 and 4.3.2.5).  Past exposures that contribute Pb to the bone, consequently, may
21    influence current levels of Pb  in blood.  Where past exposures were elevated in comparison to
22    recent exposures, this influence may complicate interpretations with regard to recent exposure
23    (CD, Sections 4.3.1.4 to 4.3.1.6). That is, higher blood Pb concentrations are not always
24    indicative of higher body burdens or cumulative exposure, but they are generally indicative of
25    higher exposures or Pb uptake over a somewhat recent past (CD, pp. 4-34 and 4-133). Bone
26    measurements, as a result of the generally slower Pb turnover in bone, are recognized as
27    providing a better measure of cumulative Pb exposure (CD, Section 8.3.2).
28          The bone pool of Pb is thought to be much more labile in children than in adults due to
29    the more rapid turnover of bone mineral as a result of growth (CD, p. 4-27).  As a result,
30    "changes in blood Pb concentration in children are thought to more closely parallel changes in
31    total body burden" (CD, p. 4-27). This is in contrast to adults, whose bone has accumulated
32    decades  of Pb exposures (with past exposures often greater than current ones), and for whom the
33    bone may be a significant source long after exposure has ended (CD, Section 4.3.2.5).
34          In several recent studies investigating the relationship between Pb exposure and blood Pb
35    in children (e.g., Lanphear and Roghmann 1997; Lanphear et al., 1998), blood Pb levels have
36    been shown to reflect Pb exposures, with particular influence associated with exposures to Pb in

            December 2006                      3-3                Draft - Do Not Quote or Cite

-------
 1    surface dust. Further, as stated in the CD "these and other studies of populations near active
 2    sources of air emissions (e.g., smelters, etc.), substantiate the effect of airborne Pb and
 3    resuspended soil Pb on interior dust and blood Pb" (CD, p. 8-22).
 4          As mentioned earlier blood Pb is generally described as reflecting recent exposures (CD,
 5    Section 4.3.1.4). Inhaled or ingested Pb quickly enters the blood, and Pb in the blood is available
 6    for exchange with the soft and skeletal tissues, conceptually viewed as the fast (half-life of-28
 7    days) and slow (half-life may be decades in adults) turnover pools, respectively (CD, Section
 8    4.3.1.4). Simulations using biokinetic models indicate that blood Pb levels in adults achieve a
 9    new quasi-steady state within 75-100 days (approximately 3-4 times the blood elimination half-
10    life) subsequent to abrupt increases in Pb uptake (CD, pp. 4-25 to 4-26).  Similar models indicate
11    a quicker response of blood Pb levels in children (CD, p. 4-27 and Figure 4-5).  Additionally,
12    response of the blood to reduction of a relatively brief Pb exposure  appears to be faster than for
13    an exposure of several years, with estimated half-lives of approximately 9 months as compared
14    to 30 months for the longer exposure response (CD, pp. 4-25 to 4-26).
15          Blood Pb levels are extensively used as an index or biomarker of exposure by national
16    and international health agencies, as well as in epidemiological (CD, Sections 4.3.1.3 and 8.3.2)
17    and toxicological studies of Pb health effects and dose-response relationships (CD, Chapter 5).
18    The prevalence of the use of blood Pb as an exposure index or biomarker is related to both the
19    ease of blood sample collection (CD, p. 4-19; Section 4.3.1) and by findings of association with a
20    variety of health effects (CD, Section 8.3.2). Accordingly, the U.S. Centers for Disease Control
21    and Prevention (CDC), and its predecessor agencies, have for many years used blood Pb level as
22    a metric for identifying children at risk of adverse health effects and for specifying particular
23    public health recommendations (CDC, 1991; CDC, 2005). In 1978, when the current Pb
24    NAAQS was established, the CDC recognized a blood Pb level of 30 ug/dL as a level warranting
25    individual  intervention (CDC, 1991).  In 1985, the CDC recognized a level of 25 ug/dL for
26    individual  child intervention, and in 1991, they recognized a level of 15 ug/dL for individual
27    intervention and a level of  10 ug/dL for implementing community-wide prevention activities
28    (CDC, 1991; CDC, 2005).  In 2005, with consideration of a review  of the evidence by their
29    advisory committee, CDC revised their statement on Preventing Lead Poisoning in Young
30    Children, specifically recognizing the evidence of adverse health effects in children with blood
31    Pb levels below 10 ug/dL and the data demonstrating that no "safe" threshold for blood Pb had
32    been identified,  and emphasizing the importance of preventative measures (CDC, 2005). *
              With the 2005 statement, CDC identified a variety of reasons, reflecting both scientific and practical
      considerations, for not lowering the 1991 level of concern, including a lack of effective clinical or public health

            December 2006                      3-4               Draft - Do Not Quote or Cite

-------
 1           Since 1976, the CDC has been monitoring blood Pb levels nationally through the
 2    National Health and Nutrition Examination Survey (NHANES).  This survey has documented
 3    the dramatic decline in mean blood Pb levels in the U.S. population that has occurred since the
 4    1970s and that coincides with  regulations regarding leaded fuels, leaded paint, and Pb-containing
 5    plumbing materials that have reduced Pb exposure among the general population  (CD, Sections
 6    4.3.1.3 and 8.3.3; Schwemberger et al., 2005).  Although levels in the U.S. general population,
 7    including geometric mean levels in children aged 1-5, have declined, mean levels have been
 8    found to differ among children of different socioeconomic status (SES) and other demographic
 9    characteristics  (CD, p. 4-21).  The health effects associated with blood Pb levels are extensively
10    discussed in the CD, while those of particular policy relevance for this review are summarized in
11    subsequent subsections of this chapter of this document.
12           Blood Pb levels are used as the index of exposure (or exposure metric) for prediction of
13    Pb associated health risk in the human exposure and health risk assessments performed for this
14    review (described in Chapter 4). This use of exposure-response functions that rely on blood Pb
15    (e.g., rather than ambient Pb concentration, if that were feasible) as the exposure metric in risk
16    assessments provides a reduced uncertainty as to causality aspects of Pb risk estimates, yet
17    imposes additional  effort on identifying specific risk contributions associated with specific Pb
18    exposure sources or pathways. For example, the blood Pb-response relationships developed in
19    epidemiological (or lexicological) studies  do not distinguish among different sources of Pb (e.g.,
20    inhalation, ingestion of dust, ingestion of dust containing paint, etc.) to the blood Pb
21    concentration.  In the exposure and risk assessments described in Chapter 4, exposure, dosimetry
22    and empirical models are used to inform estimates of the contributions of Pb to blood Pb levels
23    arising from ambient air related Pb versus other Pb sources.
24           The CD extensively discusses models in the peer reviewed literature that describe blood
25    Pb levels associated with Pb exposure, including summaries regarding the two pharmacokinetic
26    models that have been used in the human exposure assessment described in Chapter 4:  the
27    Integrated Exposure Uptake BioKinetic (IEUBK) model for Pb in children developed by EPA
28    (1994a,b; White et al., 1998; CD, Section 4.4.5); and, the model developed by Leggett (1985,
29    1992a, 1992b,  1993) for the International Commission on Radiological Protection, which
30    simulates Pb kinetics from birth through adulthood (CD, Section 4.4.6). The  performance of
31    these models has been evaluated with empirical data sets (CD, Sections 4.4.5.3, 4.4.6.2, 4.4.7.2).
      interventions to reliably and consistently reduce blood Pb levels that are already below 10 ug/dL, the lack of a
      demonstrated threshold for adverse effects, and concerns for deflecting resources from children with higher blood
      Pb levels (CDC, 2005).
            December 2006                      3-5                Draft - Do Not Quote or Cite

-------
 1    The IEUBK model, which unlike the others has an exposure pathway interface "has gained
 2    widespread use for risk assessment purposes in the United States" in evaluating multimedia Pb
 3    exposure impacts on blood Pb levels and distribution of Pb to bone and other tissues in young
 4    children <7 years old (CD, p. 8-23 and Sections 8.3.4 and 4.4.5.3). Aspects of the Leggett model
 5    have been used in an  'All Ages Lead Model', currently being developed by EPA (CD, pp. 4-118
 6    and 8-23).

 7         3.3   NATURE OF EFFECTS
 8           As described in the CD (Section 8.4.1), Pb has been demonstrated to exert "a broad array
 9    of deleterious effects  on multiple organ systems via widely diverse mechanisms  of action" (CD,
10    p. 8-24).  This array of health effects and the evidence associated with each effect is
11    comprehensively described in the CD.  This draft Staff Paper, however, is limited in focus to
12    those effects associated with the lowest Pb levels of exposure or blood Pb (i.e., those most
13    pertinent to ambient exposures). At the time of the last Staff Paper, the health effects of primary
14    interest included the following (USEPA 1990):
15         •   Heme biosynthesis and related functions;
16         •   Neurological development and function;
17         •   Reproduction and physical development;
18         •   Kidney function; and
19         •   Cardiovascular function.
20           As illustrated  by extensive discussion in the CD, the evidence for these effects remains,
21    and in most cases has been strengthened. Further, there has been substantial investigation of Pb
22    immunotoxicity. There is also evidence of Pb  carcinogenicity, primarily from animal studies,
23    with limited human evidence of suggestive associations (CD, Sections 5.6.2, 6.7, and 8.4.10).2
24           As stated in the CD, neurotoxic effects in children and cardiovascular effects in adults are
25    "currently clearly of greatest public health concern" (CD, p.  8-60). Further, the toxicological and
26    epidemiological information available  since the time of the last review "includes assessment of
27    new evidence substantiating risks of deleterious effects on certain health endpoints being induced
28    by distinctly lower than previously demonstrated Pb exposures indexed by blood-Pb levels
29    extending well below 10 ug/dL in children and/or adults" (CD, p.  8-25). For example, the
              Lead has been classified as a probable human carcinogen by the International Agency for Research on
      Cancer, based mainly on sufficient animal evidence, and as reasonably anticipated to be a human carcinogen by the
      U.S. National Toxicology Program (CD, Section 6.7.2). U.S. EPA classified it in the past as a probable carcinogen
      (http://www.epa.gov/iris/subst/0277.htmX

            December 2006                      3-6               Draft - Do Not Quote or Cite

-------
 1    overall weight of the available evidence, described in the CD, provides clear substantiation of
 2    neurocognitive decrements being associated in young children with blood Pb levels in the range
 3    of 5 to 10 ug/dL, and some analyses appear to show Pb effects on intellectual attainment of
 4    young children ranging from 2 to 8 ug/dL (CD, Sections 6.2, 8.4.2 and 8.4.2.6).  Table 3-1
 5    summarizes those Pb induced health effects for children, that given their occurrence in the range
 6    of current blood Pb levels, are most pertinent to the current review. Similar information for
 7    adults is presented in Table 3-2 (CD, Tables 8-5 and 8-6). These tables indicate  some health
 8    effects associated with blood Pb levels that extend below 5 ug/dL, and use the notation "(???)" to
 9    indicate that some studies have observed these effects at the lowest blood levels  considered (i.e.,
10    threshold levels for these effects cannot be discerned from the currently available studies).
            December 2006                      3-7                Draft - Do Not Quote or Cite

-------
1    Table 3-1.   Summary of Lowest Observed Effect Levels for Key Lead-Induced Health Effects in Children (CD, Table 8-5)
                Lowest Observed Effect
                   Blood Lead Level
    Neurological Effects
   Hematological Effects
        Immune Effects
                       30 (ig/dL
                                        Increased urinary 5-
                                        aminolevulinic acid
                        15 (ig/dL
                        10 (ig/dL
                        5 p.g/dL


                        0 (ig/dL
   Behavioral disturbances
(e.g., inattention, delinquency)

 Altered electrophysiological
          responses


Effects on neuromotor function

    CNS cognitive effects
      (e.g., IQ deficits)
 Erythrocyte protoporphyrin
       (EP) elevation
Inhibition of 5-aminolevulinic
  acid dehydratase (ALAD)
              I
  Pyrimidine -5' -nuclotidase
  (Py5N) activity inhibition
Effects on humoral (| serum IgE)
   and cell-mediated (j T-cell
     abundance) immunity
2
3
               Note: Arrows depict cases where weight of overall evidence strongly substantiates likely occurrence of type of effect in association with blood-Pb
               concentrations in range of 5-10 ng/dL, or possibly lower, as implied by (???). Although no evident threshold has yet been clearly established for those
               effects, the existence of such effects at still lower blood-Pb levels cannot be ruled out based on available data.

               Source: Adapted/updated from Table 1-17 of U.S. Environmental Protection Agency (1986a).
                               December 2006
                                              Draft - Do Not Quote or Cite

-------
2

3
Table 3-2.  Summary of Lowest Observed Effect Levels for Key Lead-Induced Health Effects in Adults (CD, Table 8-6)
                Lowest Observed Effect
                   Blood Lead Level
                       30 ug/dL
                                      Neurological Effects
Hematological Effects     Cardiovascular Effects
    Renal Effects
                                     Peripheral sensory nerve
                                          impairment
     Erythrocyte
 protoporphyrin (EP)
  elevation in males
Impaired Renal Tubular
      Function
                       20 ug/dL

                       15 ug/dL
                                      Cognitive impairment
                                         Postural sway
     Erythrocyte
 protoporphyrin (EP)
 elevation in females

  Increased urinary
5-aminolevulinic acid
                       10 ug/dL



                        5 ug/dL


                        0 ug/dL
                                                                    Inhibition of
                                                                5-aminolevulinic acid
                                                                dehydratase (ALAD)
                         Elevated blood pressure
                                                                                                                 Elevated serum creatine
                                                                                                                   (J, creatine clearance)
               Note:  Arrows depict cases where weight of overall evidence strongly substantiates likely occurrence of type of effect in association with blood-Pb
               concentrations in range of 5-10 ug/dL, or possibly lower, as implied by (???). Although no evident threshold has yet been clearly established for those
               effects, the existence of such effects at still lower blood-Pb levels cannot be ruled out based on available data.

               Source: Adapted/updated from Table 1-16 of U.S. Environmental Protection Agency (1986a).
                                December 2006
                                                                    3-9
                          Draft - Do Not Quote or Cite

-------
 1           The evidence for the health effects of greatest interest for this review (e.g., neurotoxic
 2    effects in children and cardiovascular effects in adults) is a combination of epidemiological and
 3    toxicological evidence. The epidemiological evidence is strongly supported by animal studies
 4    that substantiate the biological plausibility of the associations, in addition to providing an
 5    understanding of mechanisms of action for the effects (CD, Section 8.4.2).

 6         3.3.1   Developing Nervous System
 7           The nervous system has long been recognized as a target of Pb toxicity, with the
 8    developing  nervous system affected at lower exposures than the mature system (CD, Sections
 9    6.2.1, 6.2.2, and 8.4).  While blood Pb levels in U.S. children ages one to five years have
10    decreased notable since the late 1970s, newer studies have investigated and reported associations
11    of effects on the neurodevelopment of children with these more recent blood Pb levels (CD,
12    Chapter 6).  Effects on the  mature nervous system are discussed in a subsequent subsection of
13    this chapter (Section 3.3.6).
14           Functional manifestations of Pb neurotoxicity include sensory, motor, cognitive and
15    behavioral impacts. As stated in the CD, "extensive experimental laboratory animal evidence
16    has been generated that (a) substantiates well the plausibility of the epidemiologic findings
17    observed in human children and adults and (b) expands our understanding of likely mechanisms
18    underlying the neurotoxic effects" (CD, p. 8-25). Numerous epidemiological studies have
19    reported neurocognitive, neurobehavioral, sensory, and motor function effects in children at
20    blood Pb  levels below 10 ug/dL (CD, Section 6.2).  Studies with laboratory animals (discussed
21    in Section 5.3 of the CD) provide strong evidence with regard to the role of Pb in producing
22    these effects.
23           Effects on  cognition observed in epidemiological studies have included decrements in
24    intelligence test results, such as the widely used intelligence quotient (IQ)  score,  and in academic
25    achievement as assessed by various standardardized tests as well as by  class ranking and
26    graduation rates (CD, Section 6.2.16). As noted in the CD with regard to the latter,
27    "Associations between Pb exposure and academic achievement observed in the above-noted
28    studies were significant even after adjusting for IQ, suggesting that Pb-sensitive
29    neuropsychological processing and learning factors not reflected by global intelligence indices
30    might contribute to reduced performance on academic tasks" (CD, pp 8-29 to 8-30).
31           Other cognitive effects in children have been associated with Pb exposures including
32    effects on attention, executive functions, language, memory, learning and visuospatial
33    processing.  Attention and executive function effects have been associated with Pb exposures
34    indexed by  blood Pb levels below 10 ug/dL (CD, Section 6.2.5 and pp. 8-30 to 8-31). The
35    evidence  for the role of Pb  in this suite of effects includes experimental animal findings

             December 2006                     3-10        Draft - Do Not Quote or Cite

-------
 1    (discussed in CD, Section 8.4.2.1) of Pb effects on learning ability, memory and attention (CD,
 2    Section 5.3.5), as well as associated mechanistic findings. As stated in the CD, the "animal
 3    toxicology findings provide strong biological plausibility in support of the concept that Pb may
 4    impact one or more of these specific cognitive functions in humans" (CD, p. 8-31).
 5          The evidence described in the CD provides strong support for the association of
 6    neurocognitive decrements in young children with blood Pb levels in the range of 5 to 10 ug/dL,
 7    with some analyses indicating Pb effects on intellectual attainment of young children at blood Pb
 8    levels ranging from approximately 2 to 8 ug/dL (CD, Sections 6.2, 8.4.2 and 8.4.2.6; Lanphear et
 9    al., 2005; Lanphear et al., 2000; Al-Saleh et al 2001). These studies have observed these effects
10    at the lowest blood levels considered (i.e., threshold levels for these effects is not evident in the
11    populations studied).  Further, Pb-induced deficits observed in animal  and epidemiological
12    studies, for the most part, have been found to be persistent (CD, Sections 5.4 and 6.2.11).
13          Behavioral effects, including incidence of delinquent behavior, have been associated with
14    bone Pb and with blood Pb levels above 10 ug/dL (CD, Sections 6.2.6 and 8.4.2.2). For
15    example, the CD, based on results of several epidemiological investigations of the relationship
16    between Pb exposure and delinquent and criminal behavior in large cities,  concluded that "Pb
17    may play a role in the epigenesis of behavioral problems in inner-city  children independent of
18    other social and biomedical cofactors," although "the particular biological mechanisms that may
19    underlie Pb's effects on aggression,  impulsivity, and poor self-regulation are not yet well
20    understood" (CD, p. 8-32).
21          Sensory effects associated with Pb exposures during development have included those
22    related to hearing and vision.  The evidence has included findings from investigations with
23    animal models, as well as a limited number of epidemiological studies assessing hearing
24    thresholds and auditory processing (CD, Sections 6.2.7 and 7.4.2.3). In studies of children with
25    median blood Pb levels of 7 or 8 ug/dL, significant associations were found for increased hearing
26    thresholds with blood Pb levels extending below  10 ug/dL (CD, Sections 6.2.7 and 8.4.2.3, p.
27    AX6-23).
28          In the few epidemiological studies that have examined neuromotor function, early Pb
29    exposures, even those related to blood Pb levels below 10 ug/dL, have been associated with
30    deficits in neuromotor function.  Although from the animal studies "no clear pattern of Pb-
31    induced effects on motor activity has yet emerged", "many studies do  point to an increase in
32    activity, as seen with epidemiological findings" (CD, p. 8-36).

33         3.3.1.1  Endpoint for risk quantitation
34          Neurocognitive impact, specifically decrement in IQ in young  children, is a focus of the
35    quantitative risk assessment in this review (see Chapter 4) due to the strength of evidence for

            December 2006                     3-11         Draft - Do Not Quote or Cite

-------
 1    association with blood Pb levels below 10 ug/dL, and the strength of the dose-response
 2    information at these exposure levels. As discussed in the CD (Section 8.4.2) and by Rice (1996),
 3    while there is no direct animal test parallel to human IQ tests, "in animals a wide variety of tests
 4    that assess attention, learning, and memory suggest that Pb exposure results in a global deficit in
 5    functioning, just as it is indicated by decrements in IQ scores in children" (CD, p. 8-27). The
 6    following statements from the CD (p. 8-44) summarize the consistency and complementary
 7    nature of the animal and epidemiological evidence for this endpoint:
 8          "Findings from numerous  experimental studies of rats and of nonhuman primates, as
 9          discussed in Chapter 5, parallel the observed human neurocognitive deficits and the
10          processes responsible for them. Learning and other higher order cognitive processes show
11          the greatest similarities in Pb-induced deficits between humans and experimental
12          animals. Deficits in cognition are due to the combined and overlapping effects of Pb-
13          induced perseveration, inability to inhibit responding, inability to adapt to changing
14          behavioral requirements, aversion to delays, and distractibility. Higher level
15          neurocognitive functions are affected in both animals and humans at very low exposure
16          levels (<10 ug/dL),  more so than simple cognitive functions."
17
18          As stated in the CD, "epidemiologic studies of Pb and child development have
19    demonstrated inverse associations between blood Pb concentrations and children's IQ and other
20    outcomes at successively lower Pb exposure levels" over the past 30 years (CD, p. 6-64).  This is
21    supported by multiple studies performed over the past 15 years (see CD, Section 6.2.13), with
22    particularly compelling evidence for decrements in IQ at blood Pb levels below 10 ug/dL is
23    provided by a recent international pooled analysis of seven prospective studies (Lanphear et al.,
24    2005; CD, Section 6.2.13).  For example, this pooled analysis estimated a decline of 6.2 points
25    (with a 95% confidence interval bounded by 3.8 and 8.6) in full scale IQ occurring with a change
26    in blood Pb level across the entire pooled data set (measured concurrent with the IQ test), from
27    <1 ug/dL to 10 ug/dL (CD, p. 6-76).  This analysis (Lanphear et al., 2005) is relied upon in the
28    quantitative risk assessment for this  endpoint discussed in  Chapter 4.

29         3.3.1.2  Metric and  quantitative model for risk quantitation
30          The epidemiological studies  that have investigated blood Pb effects on IQ (see CD,
31    Section 6.2.3) have considered a variety of specific blood Pb metrics, including: 1) blood
32    concentration "concurrent" with the response assessment (e.g., with IQ testing), 2) average blood
33    concentration over the "lifetime" of the child at the time of response assessment (e.g., 6 or 7
34    years), 3) peak blood concentration during a particular age range and 4) early childhood blood
35    concentration (e.g., the mean of measurements  between 6 and 24 months age).   All four specific

            December 2006                     3-12        Draft - Do Not Quote or Cite

-------
 1    blood Pb metrics have been correlated with IQ (see CD, p. 6-62; Lanphear et al., 2005). In the
 2    international pooled analysis by Lanphear and others (2005), however, the concurrent and
 3    lifetime averaged measurements were considered "stronger predictors of lead-associated
 4    intellectual deficits than was maximal measured (peak) or early childhood blood lead
 5    concentrations," with the concurrent blood Pb level exhibiting the strongest relationship (CD, p.
 6    6-29).
 7           Using concurrent blood Pb level as the dose or exposure metric and IQ as the response
 8    from the pooled dataset of seven international studies, Lanphear and others (2005) employed
 9    mathematical models of various forms, including linear, cubic spline, the log-linear, and piece-
10    wise linear, in their investigation of the blood Pb concentration-response relationship (CD, p. 6-
11    29; Lanphear et al., 2005). They observed that the shape of the dose-response relationship is
12    nonlinear and the log-linear model provides a better fit for the data than  a linear one (CD, p. 6-29
13    and pp. 6-67 to 6-70; Lanphear et al., 2005). In addition, they found that no individual study
14    among the seven drove the results (CD  p. 6-30). Others have also analyzed the same dataset and
15    similarly concluded that, within the  ranges of the dataset's blood Pb levels, a log-linear
16    relationship was a significantly better fit than the linear relationship (p=0.009) with little
17    evidence of residual confounding from  included model variables (CD, Section 6.2.13;
18    Rothenberg and Rothenberg, 2005).
19           A nonlinear exposure-response relationship is also suggested by  several other studies that
20    have indicated a dose-response relationship, in terms of estimated IQ decline per ug/dL increase
21    in blood Pb, that  may be steeper at blood Pb levels below 10 ug/dL than at higher levels (CD, pp.
22    8-63 to 8-64). While, as discussed in the CD, this may at first seem at odds with certain
23    fundamental toxicological concepts, a number of examples of non- or supra-linear dose-response
24    relationships exist in toxicology, and this non-linear dose-effect relationship also occurs for
25    several Pb effects (CD, pp. 6-76 and 8-83 to 8-39).  With regard to this particular endpoint (IQ),
26    the CD states that it "is conceivable that the initial neurodevelopmental lesions at lower Pb levels
27    may be disrupting very different biological  mechanisms (e.g., early developmental processes in
28    the central nervous system) than the more severe effects of high exposures that result in
29    symptomatic Pb poisoning and frank mental retardation" (CD, p. 6-76).  In comparing  across the
30    individual studies and the pooled analysis, it is observed that at higher blood Pb levels, the slopes
31    derived for log-linear and linear models are almost identical, and for studies with lower blood Pb
32    levels, the slopes appear to be  steeper than those observed at higher blood Pb levels (CD, p.  8-78,
33    Figure 8-7).
34           Given the evidence summarized here and described in detail in the CD (Chapters 6 and
35    8), and consistent with recommendations from CAS AC on the risk assessment plan (Henderson,
36    2006), the assessment of children's risk described in Chapter 4 relies on the log-linear functions
             December 2006                     3-13         Draft - Do Not Quote or Cite

-------
 1    presented by Lanphear and others (2005) that relate absolute IQ as a function of the log of
 2    concurrent blood Pb, and lifetime average blood Pb, respectively. As discussed above, the slope
 3    of the exposure-response relationship described by these functions is greater at the lower blood
 4    Pb levels (e.g., less than 10 ug/dL). The impact of the nonlinear slope is illustrated by the
 5    estimated IQ decrements associated with increases in blood IQ for different ranges of blood Pb
 6    level. The IQ changes were 3.9 (with 95% confidence interval, CI, of 2.4-5.3), 1.9 (95% CI, 1.2-
 7    2.6) and 1.1 (95% CI, 0.7-1.5), for increases in concurrent blood Pb from 2.4 to 10 ug/dL, 10 to
 8    20 ug/dL, and 20 to 30 ug/dL, respectively (Lanphear et al., 2005).
 9           As discussed in the CD, threshold blood Pb levels for these  effects cannot be discerned
10    from the currently available epidemiological studies, and the evidence in the animal Pb
11    neurotoxicity literature does not identify a well-defined thresholds for any of the toxic
12    mechanisms of Pb (CD,  Sections 5.3.7 and 6.2). However, in recognition of a reduced
13    confidence in the characterization of the quantitative blood Pb concentration-response
14    relationship at the lowest blood Pb levels included in the current studies, as well as the
15    possibility of a threshold at or below these levels, the staff has employed a hypothetical threshold
16    or cutpoint in the pilot quantitative risk assessment described in chapter 4, below which it is
17    assumed that there is no individual response.  In this context, this cutpoint is not intended as  a
18    true biological  threshold. Rather it is intended simply to reflect a potential or hypothetical
19    inflection point at the lower end of the  concentration-response relationship.
20           In selecting the cutpoint for pilot risk assessment, we considered particularly two studies
21    (Lanphear et al., 2000 and Lanphear et al., 2005). In the study by Lanphear and others (2000)
22    that found associations of cognitive deficits in children aged 6-16 years of age using NHANES
23    III, the authors stratified their analyses into four blood Pb categories: <10 ug/dL (n=4,681); <7.5
24    ug/dL (n=4526); <5.0 ug/dL (n=4,043) and <2.5 ug/dL (n=2,467).  The lowest blood Pb group
25    was substantially smaller in size than the other groups, and additionally, although coefficients for
26    that category are fairly similar to the ones in higher level categories (sometimes slightly larger),
27    none of the coefficients were statistically significant, indicating a reduction in statistical power.
28    Additionally, in the pooled analysis by Lanphear and others (2005), from which the exposure-
29    response functions for the health risk assessment is drawn, the proportion of the pooled data  set
30    below 2.5 ug/dL (concurrent blood Pb) is quite  small. For example, the level of 2.4 ug/dL, is the
31    concurrent blood Pb level for the 5th percentile of the pooled data set, while 33.1 ug/dL is the
32    95th percentile. The 5th and 95th percentile values for lifetime average blood Pb are 6.1 and 47
33    ug/dL, respectively (Lanphear et al., 2005).
            December 2006                     3-14        Draft - Do Not Quote or Cite

-------
 1         3.3.2  Cardiovascular System
 2          Epidemiologic and experimental toxicology studies support the relationship between Pb
 3    exposure and increased adverse cardiovascular outcome, including increased blood pressure,
 4    increased incidence of hypertension, and cardiovascular morbidity and mortality (CD, Sections
 5    5.5, 6.5 and 8.4.3).
 6          The cardiovascular effect most frequently examined in epidemiological studies is
 7    increased blood pressure in adults, which has been repeatedly associated with Pb exposure (CD,
 8    Sections 8.4.3 and 6.5.7). The association has been observed with Pb levels in both bone and
 9    blood (including blood Pb levels below 10 ug/dL). This epidemiological evidence is supported
10    by evidence in numerous animal studies of arterial hypertension (HTN) with low Pb exposures,
11    an effect that persists in animals long after cessation of exposure (CD, Sections 5.5 and 8.4.3). A
12    recent meta-analysis by Nawrot and others (2005), that included a range of blood Pb levels from
13    2.3 to 63.8 ug/dL, reported an association of increased systolic blood pressure and decreased
14    diastolic pressure with increased blood Pb level, including levels below 10 ug/dL. The changes
15    observed, on the order of 1 millimeter mercury increase in systolic pressure per doubling of
16    blood Pb, have considerable significance at the population level  (CD, p. 8-45, Section 8.6.3).
17    Systolic blood pressure exerts a strong influence on more serious cardiovascular events by its
18    role in hypertension and its adverse cardiovascular sequelae (CD, p.  8-83).
19          Multiple studies of blood pressure and hypertension have reported positive associations
20    with bone Pb levels, highlighting the important role for cumulative past Pb exposure in
21    development of cardiovascular health effects (Sections  6.5.2.3 and 6.5.7).  Further, a study of
22    young adults who lived as children in an area of high Pb exposures indicates the potential for
23    childhood exposure to contribute to such effects later in life. In this study, higher bone Pb levels
24    were associated with higher systolic and diastolic blood pressure (CD, p. 6-138), while current
25    blood Pb  levels (mean of 2.2 ug/dL) were not associated with blood pressure effects (CD, p. 6-
26    124).
27          Several analyses of National Health and Nutrition Examination Survey (NHANES)
28    cohorts, including some recently released, have collectively suggested a "significant effect of Pb
29    on cardiovascular mortality in the general U.S. population" (CD, p. 8-88, Sections 6.5.3.2 and
30    8.6.3).  For example recent analyses of NHANES blood Pb data from 1976 to 1980 and 1988 to
31    1994 provide supportive evidence  for an increased risk of cardiovascular mortality, consistent
32    with projected likely increases in serious cardiovascular events (stroke, heart attack)  resulting
33    from Pb-induced increases in blood pressure (CD, Section 8.6.3).
            December 2006                     3-15        Draft - Do Not Quote or Cite

-------
 1         3.3.3   Heme Synthesis
 2          It has long been recognized that Pb exposure is associated with disruption of heme
 3    synthesis in both children and adults. At blood Pb levels above 30 ug/dL, such disruption leads
 4    to notable reductions in hemoglobin synthesis, and, at blood Pb levels above 40 ug/dL, to frank
 5    anemia, a clinical sign of severe Pb poisoning (CD, p. 8-47).  The evidence regarding effects on
 6    heme synthesis and other hematological parameters in animal and humans is strong, and includes
 7    documented quantitative relationships between exposure and effects in children and adults.
 8    Interference with heme synthesis was identified as one of the targets of low-level Pb toxicity in
 9    children during the time of the last NAAQS review (USEPA,  1990), and was the primary basis
10    for the initial setting of the Pb NAAQS in 1978 (USEPA, 1978).
11          Mechanisms associated with Pb interference with heme synthesis include inhibition of
12    the enzymes 5-aminolevulinic acid dehydratase (ALAD) and ferrochelatase (CD Sections 5.2.1,
13    6.9.1, 6.9.2). Inhibition of ALAD has been associated with increased blood Pb concentrations
14    across the range of 5 to 150 ug/dL. This information and evidence regarding associated
15    mechanisms is presented and discussed in detail in the CD (Sections 8.4.4, 5.2.1, 6.9.1  and
16    6.9.2).

17         3.3.4   Renal System
18          As described in the CD (Sections 5.7.3 and 8.4.5), Pb nephrotoxicity is mediated by
19    alterations in the glomerular filtration rate. The animal literature has described the occurrences
20    and mechanisms of Pb uptake by and accumulation in the kidney, and associated cellular
21    alterations (CD, Section 5.7). A set of screening tests involving markers of nephrotoxic effects
22    have been established for screening individuals exposed to Pb occupationally or environmentally
23    (CD, Section 5.7.1). In the epidemiological literature, associations between blood Pb and
24    indicators of renal function impairment (e.g., measures of glomerular integrity, such as creatinine
25    levels in urine) have been found at blood Pb levels extending below 10 ug/dL, to as low as ~2 to
26    4 ug/dL (CD,  Sections 6.4.4.1.5 and 8.4.5).  Associations are also observed  with cumulative Pb
27    dose, assessed via bone Pb, and longitudinal renal function decline (CD, p. 6-94).
28          Although previous observations from occupational studies have indicated much higher Pb
29    blood levels (e.g., >30-40 ug/dL) as affecting renal tubular function, the CD describes the recent
30    findings in non-occupational populations as providing "strong evidence that renal effects occur at
31    much lower blood Pb levels than previously recognized" (CD, p. 6-113). Exposure history  may
32    play a role in the differences in the two sets of evidence.  For  example, the CD recognizes that
33    "the data available to date are not sufficient to determine whether nephrotoxicity is related more
34    to such current blood-Pb levels, higher levels from past exposure, or both" (CD, p. 8-49).
35    Additionally, the CD suggests that the studies in the general population likely had larger

            December 2006                    3-16        Draft - Do Not  Quote or Cite

-------
 1    proportions of susceptible individuals than occupational cohorts, which may play a role in the
 2    findings of lower Pb dose thresholds for Pb renal effects in environmental compared to
 3    occupational research (CD, p. 6-107).
 4          The findings regarding Pb exposures and renal effects are of particular concern with
 5    regard to certain susceptible subpopulations as described in the CD (p. 6-113).
 6          "At levels of exposure in the general U.S. population overall, Pb combined with other
 7          risk factors, such as diabetes, hypertension, or chronic renal insufficiency from non-Pb
 8          related causes, can result in clinically relevant effects. Notably, the size of such
 9          susceptible populations is increasing in the United  States due to obesity."
10    That is, Pb is recognized as acting cumulatively with other renal risk factors to cause early onset
11    of renal insufficiency and/or a steeper rate of renal function decline in individuals already at risk
12    for renal disease (CD, p. 6-107).

13         3.3.5  Immune System
14          Since the time of the last review, there has been substantial research on the
15    immunotoxicity of Pb. As summarized in the CD, "studies across humans and a variety of
16    animal models are in general agreement concerning both the nature of the immunotoxicity
17    induced by Pb as well as the exposure conditions that are required to produce
18    immunomodulation" (CD, p. 5-244, Section 5.9). Lead is distinguished from other
19    immunotoxicants, however, by the fact that the most sensitive biomarkers of its immunotoxicity
20    are associated with specific functional  capacities that influence risk of disease, as opposed to
21    being associated with changes in immune cell numbers or  pathological changes of lymphatic
22    system organs (CD, Section 5.9.1).  The main immune system targets of Pb are macrophages
23    and T lymphocytes, leading to a potential for increased tissue inflammation, reduced cell-
24    mediated immunity, and increased risk of autoimmunity (See CD, Figure 5-18, Section 5.9.11).
25    Additionally, Pb exposures in both animal and human studies are associated with increased
26    production of IgE, an immunoglobulin involved in allergic responses and asthma (CD, Section
27    5.9.3.2). These effects have been reported in epidemiologic studies of children, and supported
28    by evidence in neonatal/juvenile animals, at blood-Pb levels extending below 10 ug/dL (CD, p.
29    6-197 and  Sections 5.9.10 and 8.4.6).

30         3.3.6  Adult Nervous System
31          As  discussed in Section 3.3.1, the nervous system has long been recognized as a target of
32    Pb toxicity (CD Sections 5.3.1, 8.4.2).  As described in the CD, a blood Pb concentration of >14
33    ug/dL in those chronically exposed in the workplace is a possible threshold for various
34    neurological effects including peripheral sensory nerve impairment, visuomotor and memory
35    impairment, and postural sway abnormalities (CD, p. 6-87).  Past occupational exposure also
            December 2006                    3-17        Draft - Do Not Quote or Cite

-------
 1    increases the risk of developing amyotrophic lateral sclerosis (ALS) and motor neuron disease
 2    (CD, Section 6.3.5 and p. 6-87).  Essential tremor is also associated with Pb exposures, with
 3    studies indicating that the subpopulation of individuals with the ALAD2 allele have a 30-fold
 4    greater increased risk than those with only ALAD1 gene (CD, Sections 6.3.5 and 6.3.6 and p. 6-
 5    86).
 6          Epidemiological studies of elderly populations have also investigated associations
 7    between Pb exposures and impaired cognitive performance (CD, Section 6.3.3). While
 8    significant associations have not been consistently found in studies employing blood Pb as the
 9    exposure metric, significant associations have been reported with bone Pb levels as the exposure
10    metric (CD, Section 6.3.3.1), perhaps indicating a role of cumulative and/or past Pb exposures
11    (CD, p. 6-83).  The studies involving bone Pb have utilized several large cohorts of older adults
12    and reported associations with cognitive dysfunction in aging populations (CD, Section 6.3.3.1).
13    These include the Normative Aging Study (Rhodes et al., 2003; Payton et al., 1998; Wright et al.,
14    2003; Weisskopf et al., 2004), and the Study of Osteoporotic Fractures (Muldoon et al.,  1996).
15    The general finding of these studies of significant associations with bone Pb, but not blood Pb,
16    suggests that long-term cumulative exposure, more than current exposure, may contribute to
17    these neurotoxic effects in adults (CD, p. 6-83).
18          As discussed in the CD (Section 5.3.7), there is animal evidence supporting an increased
19    vulnerability among the elderly to Pb effects on cognitive function.  During the demineralization
20    of bone that occurs during aging, Pb may be released into the blood, thus augmenting blood Pb
21    associated with current ambient exposures (CD, Section 4.3.2.4). Research involving lifetime
22    exposure has found that senescent animals exhibit an increased vulnerability to Pb due to this
23    increased exposure from bone resorption and an apparently greater sensitivity to the biochemical
24    effects of Pb (CD, Section 5.3.7).  Additionally, animal studies indicate that cognitive function
25    effects in the elderly may also be related to physiological effects of Pb exposures in early
26    childhood (CD, p. 5-67). Laboratory animal research in rats and monkeys has demonstrated an
27    effect of early life exposure to Pb on latent upregulation of the gene associated with the
28    production of beta-amyloid precursor protein (APP). Increased expression of APP, which is
29    thought to have a role in Alzheimer's disease, was also observed. Upregulation of APP mRNA
30    and of APP was not demonstrated with Pb exposures in later life (CD, p. 5-67; Basha et al 2005).
31    Thus, early life exposure to Pb may contribute to neurocognitive effects later in life due to the
32    redistribution of Pb body burden from bone to brain and by enhanced vulnerability caused by
33    age-related degenerative changes in various organs, including brain (CD, p. 8-40).
            December 2006                     3-18        Draft - Do Not Quote or Cite

-------
 1         3.4   LEAD-RELATED IMPACTS ON PUBLIC HEALTH
 2           In addition to the advances in our knowledge and understanding of Pb health effects at
 3    lower exposures (e.g., using blood Pb as the index), there has been some change with regard to
 4    the U.S. population Pb burden since the time of the last Pb NAAQS review. For example, the
 5    geometric mean blood Pb level for U.S. children aged 1-5, as estimated by the U.S. Centers for
 6    Disease Control, declined from 2.7 ug/dL (95% CI: 2.5-3.0) in the 1991-1994 survey period to
 7    1.7 ug/dL (95% CI:  1.55-1.87) in the 2001-2002 survey period (CD, Section 4.3.1.3).3 Blood Pb
 8    levels have also declined in the U.S. adult population over this time period (CD, Section 4.3.1.3).
 9    These observation however, should not be interpreted to mean that blood Pb levels declined in all
10    communities, and by this amount. As noted in the CD, "blood-Pb levels have been declining at
11    differential rates for various general subpopulations, as a function of income, race, and certain
12    other demographic indicators such as age of housing" (CD, p.  8-21).
13           The following discussion draws from the CD to characterize subpopulations potentially at
14    risk for Pb-related effects and potential public health impacts associated with exposure to
15    ambient Pb.

16         3.4.1   Potentially Susceptible or Vulnerable Subpopulations
17           The CD summarizes information on factors affecting susceptibility to Pb toxicity, and
18    also recognizes associated factors of susceptibility within the individual discussions of specific
19    Pb effects.  Such factors include both physiological conditions contributing to a subgroup's
20    increased risk of effects at a given blood Pb level, and physiological conditions that contribute to
21    blood Pb levels higher than those otherwise associated with a given Pb exposure (CD, Section
22    8.5.3).  The term vulnerability additionally encompasses situations of elevated exposure (e.g.,
23    residing in old housing with Pb-containing paint or near sources of ambient Pb), as well as
24    socioeconomic factors (e.g., reduced access to health care or low SES) (USEPA, 2003) that can
25    contribute to increased risk of adverse health effects from Pb.
26           Three particular physiological factors contributing to increased risk of Pb effects at a
27    given blood Pb level are recognized in the CD (e.g., CD, Section 8.5.3). The first factor is age.
28    As summarized in Section 3.5.1 of this document, and described in detail in the CD (e.g.,
29    Sections 6.2, 8.4, 8.5, 8.6.2), the susceptibility of young children to the neurodevelopmental
30    effects  of Pb is well  recognized. A difficulty in identifying a discrete period of susceptibility has
31    been that the period  of peak exposure is around 18-27 months when hand-to-mouth activity is at
             3 These levels are in contrast to the geometric mean blood Pb level of 14.9 ug/dL reported for U.S. children
      (aged 6 months to 5 years) in 1976-1980 (CD, Section 4.3.1.3).
             December 2006                     3-19        Draft - Do Not Quote or Cite

-------
 1    its maximal (CD, p. 6-60). Earlier Pb literature described the first 3 years of life as a critical
 2    window of vulnerability to the neurodevelopmental impacts of Pb (CD, p. 6-60), however, recent
 3    epidemiologic studies have indicated a potential for susceptibility of children to concurrent Pb
 4    exposure to extend to the reaching of school age (CD, pp. 6-60 to 6-64).  It may be that the
 5    influence of concurrent blood Pb (and exposures contributing to it) remains important until
 6    school age with regard to the potential to affect cognitive development (CD, pp. 6-63 to 6-64;
 7    Chen et al., 2005). Additionally, the existing evidence regarding Pb immunotoxicity, and in
 8    particular, impacts on the  immunoglobulin, IgE, also indicates an increased susceptibility of
 9    children (CD, Sections 5.9.10, 6.8.3 and 8.4.6). Early childhood Pb exposures have also been
10    associated with increased  risk of cardiovascular and neurodegenerative effects in adulthood (CD,
11    p. 8-74).
12           A second physiological factor contributing to increased risk of Pb associated effects is
13    health status.  For example, subpopulations with pre-existing health conditions may also be at
14    increased susceptibility (as compared to the general population)  for particular Pb-associated
15    effects, which is most clear for renal and cardiovascular outcomes.  Those with higher baseline
16    blood pressure or hypertension may face a greater risk of adverse health impact from Pb-
17    associated cardiovascular effects, e.g., African Americans, as a group, have higher frequency of
18    hypertension than the general population or other ethnic groups (NCHS, 2005). As discussed in
19    the CD (Sections 6.4.7 and 8.4.5), those with diabetes, hypertension, and chronic renal
20    insufficiency have been shown to be at increased risk of Pb-associated reductions in renal
21    function. Such reductions have been reported at blood Pb levels ranging down to just below 5
22    ug/dL (CD, Section 6.4.4.1 and p. 8-72). Additionally, older age may be a risk factor for effects
23    on renal function (CD, p. 6-107).  Consequently, particularly vulnerable subpopulations may be
24    those that are  Pb exposed  and also at increased risk for obesity, diabetes, and hypertension; as
25    stated in the CD, frequently exposures to Pb occur in the same lower SES groups as these other
26    risk factors (CD, p. 8-89).
27           A third physiological  factor relates to genetic  polymorphisms. Subpopulations defined
28    by particular genetic polymorphisms have also been recognized with regard to susceptibility to
29    Pb toxicity. For example, presence of the ALAD allele appears to increase the magnitude of Pb-
30    associated renal dysfunction (CD, p. 8-71, Section 6.4.7.3), and also may play a role in the risks
31    of developing ALS or motor neuron disease that have been associated with past occupational
32    exposures to Pb (CD, p. 8-71, Sections 6.3.5 and 6.3.6).
33           Several physiological factors pertain to susceptibility or sensitivity by contributing to
34    increased blood Pb levels over those otherwise associated with a given Pb exposure (CD, Section
35    8.5.3).  These include nutritional status, which as recognized in Section 3.2, plays a role in Pb
36    absorption from the GI tract;  polymorphisms such as  those for the vitamin D receptor, which
            December 2006                     3 -20        Draft - Do Not Quote or Cite

-------
 1    studies suggest may contribute to increased Pb absorption from the GI tract; and bone
 2    demineralization, such as that occurring during pregnancy, lactation, and aging, which appear to
 3    influence the release of Pb from bone storage into the blood (CD, Sections 5.10.2.5, 4.3.2 and
 4    8.5.3).  An increased prevalence  of certain polymorphisms contributing to increased blood Pb
 5    levels may occur in particular subpopulations, increasing the sensitivity or vulnerability of that
 6    group to Pb associated effects. One example of this occurs with the vitamin D receptor or VDR
 7    gene, which is involved in calcium absorption through the gut. A study on blood Pb levels
 8    related to distribution of the FF genotype for this gene indicated that children with this genotype
 9    had the highest adjusted mean blood Pb concentrations at 2 years of age compared to children
10    with alternate genotypes for this  gene (CD, p.  6-56; p. 8-41).  As described by the CD, the "high
11    prevalence of FF genotypes in African-American children, compared to non-African American
12    children, may partially explain higher blood Pb concentrations often observed in African-
13    American children" (CD, p. 8-41).  Additionally, the last two NHANES surveys support findings
14    of significantly higher blood Pb levels in African-American children than whites, even after an
15    adjustment for urban residential status and family income, indicating that African-American
16    children are at increased risk for  elevated blood Pb levels compared to white children  (CD, p. 6-
17    54).
18           Differences in blood Pb levels among subpopulations living in  the same area have also
19    been identified that indicate an increased vulnerability to Pb  exposure  among some subgroups,
20    perhaps related to SES. For example, a study  of populations residing in a mining area found
21    highest blood Pb levels among African-American, Mexican-American, and poor children (CD,
22    pp. 3-26 and 8-13).

23         3.4.2   Potential Public Health Impact
24           There are several potential public health impacts associated with the current range of
25    population blood Pb levels, including potential impacts on population IQ, heart disease, and
26    chronic kidney disease (CD, Section 8.6).  The quantitative implications of potential Pb-related
27    population impacts related to these health impacts are discussed in the  CD (Sections 8.6.2, 8.6.3
28    and 8.6.4).  With regard to IQ, it  is noted that, given a somewhat uniform manifestation of Pb-
29    related decrements across the range of IQ scores in a population, "a downward shift in the mean
30    IQ value is not associated only with a substantial increase in the percentage of individuals
31    achieving very low scores, but also with substantial decreases in percentages achieving very high
32    scores" (CD, p. 8-81).  For example, for a population mean IQ of 100 (and standard deviation of
33    15), 2.3% of the population would score above 130, but a shift of the population to a mean of 95
34    results in only 0.99% of the population scoring above 130 (CD, pp. 8-81 to 8-82).
            December 2006                     3-21         Draft - Do Not Quote or Cite

-------
 1          In emphasizing the need to recognize distinctions between population and individual risk,
 2    the CD notes that a "point estimate indicating a modest mean change on a health index at the
 3    individual level can have substantial implications at the population level" (CD, p. 8-77).  For
 4    example, "the import of a decline for an individual's well-being is likely to vary depending on
 5    the portion of the IQ distribution" such that "for an individual functioning in the low range due
 6    to the influence of developmental risk factors other than Pb", a Pb-associated IQ decline of
 7    several points, might be sufficient to drop that individual into the range associated with increased
 8    risk of educational, vocational, and social handicap (CD, p. 8-77). Similarly, "although an
 9    increase of a few mmHg in blood pressure might not be of concern for an individual's well-
10    being, the same increase in the population mean might be associated with substantial increases in
11    the percentages of individuals with values that are sufficiently extreme that they exceed the
12    criteria used to diagnose hypertension" (CD, p. 8-77).
13          The magnitude of a public health impact is dependent upon the size of population
14    affected and type or severity of the effect. As summarized in Section  3.4.1, there are several
15    population groups that may be susceptible or vulnerable to effects associated with exposure to
16    Pb.  They include, young children, particularly those in families of low SES, as well as
17    individuals with hypertension, diabetes, and chronic renal insufficiency. Although
18    comprehensive estimates of the size of these groups residing in proximity to policy relevant
19    sources of ambient Pb have not been developed, total estimates of these population
20    subpopulations within the U.S. are substantial (Table 3-3).

21    Table 3-3.  Population subgroups with characteristics that may contribute to increased
22               susceptibility or vulnerability to Pb health effects.

Estimated # in
U.S. population*1
Year for estimate
Reference
Children"
Living in poverty
4.8 million
(20%)e
2005
DeNavas-Walt et al., 2006
Adultsb w.
hypertension0
~50 million
(25.6%)e
1999-2002
NCHS, 2005
Adultsb w.
Diabetes
1 8 million
(8.7%)e
2002
CDC, 2003
Adultsb w.
chronic kidney
disease
19.2 million
(ll%)e
1988-1994
Coresh et al., 2005
aChildren less than 6 years of age.
blndividuals greater than 20 year of age.
°Hypertension, defined as blood pressure of 140/90 millimeters of mercury (mm Hg) or higher, using blood-
pressure lowering medications, or having been told at least twice by a physician or other health professional
that they had high blood pressure (medical history).
dNote that there may be overlap among some groups (i.e., individuals may be counted in more than one
subgroup).
ePercent of age group.
            December 2006
3-22
Draft - Do Not Quote or Cite

-------
 1
 2           As described in the CD, and investigated in the pilot exposure and health risk assessment
 3    presented in Chapter 4 of this document, subpopulations residing near some sources of Pb
 4    emissions may be at increased risk of Pb exposures and associated effects.  The limited
 5    information available on air and soil concentrations of Pb indicates elevated concentrations near
 6    some stationary  sources (as compared with remote from such sources), including primary and
 7    secondary Pb smelters (see Chapters 2 and 4).  Using information from the  2000 U.S. Census
 8    and locations of currently operating primary and secondary Pb smelters, it is estimated that some
 9    76,000 persons,  including some 8600 children less than 7 years of age,4 reside within 2
10    kilometers of these sources.5  Emissions estimates described in Section 2.3 for individual
11    sources (e.g., Table 2-4) suggest a variety of other source types that may emit Pb in the range of
12    secondary Pb smelters.   Population size estimates near other Pb stationary point sources,
13    however, have not been developed for this first draft Staff Paper.  Additionally, the potential for
14    historically deposited Pb near roadways to contribute to increased risks of Pb exposure and
15    associated risk to populations residing nearby is suggested in the CD and also investigated in
16    Chapter 4 of this document.  Although estimates of the number of individuals, including
17    children, living within close proximity to roadways specifically recognized for this potential
18    have not been developed, these numbers may be substantial.6

19          3.5    SUMMARY AND CONCLUSIONS
20           Based  on the available health effects evidence and the evaluation and interpretation of
21    that evidence in the CD, summarized briefly above, the following conclusions have been drawn:
22          •   Lead exposures occur both by inhalation and by ingestion.  Ingestion of Pb-
23             contaminated dust has a strong influence on blood Pb levels in children.
             4 Total population counts are based on 2000 U.S. Census, derived for census blocks falling within 2 km of
      facility. In lieu of block-level age-specific counts, subgroup counts were derived from total population counts based
      on block counts and subgroup representation in block groups within 2 km of a facility, and rounded to nearest 100.
             5 The distance of 2 kilometers is consistent with estimates of distances associated with significant
      deposition or soil deposition for these types of sources in the pilot exposure and risk assessment described in
      Chapter 4.
             6 For example, the 2005 American Housing Survey, conducted by the U.S. Census Bureau indicates that
      some 14 million (or approximately 13% of) housing units are "within 300 feet of a 4-or-more-lane roadway, railroad
      or airport" (U.S. Census Bureau, 2006). Additionally, estimates developed for Colorado, Georgia and New York
      indicate that approximately 15-30% of the populations in those states reside within 75 meters of a major roadway
      (i.e., a "Limited Access Highway", "Highway", "Major Road" or "Ramp", as defined by the U.S. Census Feature
      Class Codes) (ICF, 2005).
             December 2006                      3 -23        Draft - Do Not Quote  or Cite

-------
 1         •   Children, in general and especially low SES children, are at increased risk for Pb
 2             exposure and Pb-induced adverse health effects.  This is due to several factors,
 3             including enhanced exposure to Pb via ingestion of soil-Pb and/or dust-Pb due to
 4             childhood hand-to-mouth activity.

 5         •   Once inhaled or ingested, Pb is distributed by the blood, with long-term storage
 6             accumulation in the bone.  Bone Pb levels provide a strong measure of cumulative
 7             exposure which has been associated with many of the effects summarized below,
 8             although difficulty of sample collection has precluded widespread use in
 9             epidemiological studies to date.

10         •   Blood levels of Pb are well accepted as an index of exposure (or exposure metric) for
11             which associations with the key effects (see below) have been observed.  In general,
12             associations with blood Pb are most robust for those effects for which past exposure
13             history poses less of a complicating factor, i.e., for effects during childhood.

14         •   Epidemiological studies have observed significant associations between Pb  exposures
15             and a broad range of health effects. Many of these associations have been found at
16             levels of blood Pb that are currently relevant for the U.S. population, with children
17             having blood Pb levels of 5-10 ug/dL or, perhaps somewhat lower, being at notable
18             risk.

19         •   Pb exposure is associated with a a variety of neurological effects in children, notably
20             intellectual attainment and school performance. Both qualitative and quantitative
21             evidence, with further support from animal research, indicates a robust effect of Pb
22             exposure on neurocognitive ability at blood Pb levels levels in the range of 5 to 10
23             M£/dL, and some analyses appear  to show Pb effects on intellectual attainment of
24             young children ranging from 2 to  8 ug/dL

25         •   The staff concludes that that it is appropriate to use log-linear concentration-response
26             models for the quantitative risk assessment (described in Chapter 4) for neurocognitive
27             ability in young children.

28         •   For children, the evidence is also  robust for Pb-induced disruption of heme synthesis at
29             blood Pb levels of 20-30 ug/dL. At blood Pb levels on the order of 10 ug/dL, and
30             slightly lower,  associations have been found with effects to the immune system,
31             resulting in altered macrophage function, increased IgE levels and associated increased
32             risk for autoimmunity and asthma.

33         •   In adults, epidemiological studies have consistently demonstrated associations between
34             Pb exposure and increased risk of adverse cardiovascular outcomes, including
35             increased blood pressure and incidence of hypertension. These associations have been
36             observed with bone Pb and, for some studies with blood Pb levels below  10 ug/dL.
37             Animal evidence provides confirmation of Pb effects on cardiovascular functions. For
38             these Pb effects, particularly susceptible subpopulations include those with a higher
39             baseline blood pressure. For example, African Americans, as a group, have greater
40             incidence of elevated  blood pressure than other ethnic groups.

41         •   Renal effects in adults, evidenced by reduced renal  filtration, have also been associated
42             with Pb exposures indexed by bone Pb levels and also with blood Pb below 10 ug/dL,

            December  2006                     3 -24        Draft - Do Not Quote or Cite

-------
1            with the potential adverse impact of such effects being enhanced for susceptible
2            subpopulations including those with diabetes, hypertension, and chronic renal
3            insufficiency.

4         •  Other Pb associated effects in adults occurring at or just above 10 ug/dL include
5            hematological (e.g., impact on heme synthesis pathway) and neurological effects, with
6            animal evidence providing support of Pb effects on these systems and evidence
7            regarding mechanism of action.
           December 2006                    3-25        Draft - Do Not Quote or Cite

-------
  1     REFERENCES

  2     Al-Saleh, I.; Nester, M; DeVol, E.; Shinwari, N.; Munchari, L.; Al-Shahria, S. (2001) Relationships between blood
  3             lead concentrations, intelligence, and academic achievement of Saudi Arabian schoolgirls. Int. J. Hyg.
  4             Environ. Health 204:  165-174.

  5     Centers for Disease Control (1991) Preventing lead poisoning in young children: a statement by the Centres for
  6             Disease Control. Atlanta, GA: U.S. Department of health and Human Services, Public Health Service;
  7             October 1. http://wonder.cdc.gov/wonder/prevguid/p0000029/p0000029.asp

  8     Centers for Disease Control and Prevention (2003) National Diabetes Fact Sheet.
  9             http://www.cdc.gov/diabetes/pubs/factsheet.htm

10     Centers for Disease Control and Prevention (2005) Preventing lead poisoning in young children: a statement by the
11             Centers for Disease Control and Prevention. Atlanta, GA: U.S. Department of Health and Human Services,
12             Public Health Service. August.

13     Coresh, J.; Astor, B.C.; Greene, T.; Eknoyan, G.; Levey, A.S. (2005) Prevalence of chronic kidney disease and
14             decreased kidney function in the adult US population:  Third National Health and Nutrition Examination
15             Survey. Am J Kidney Dis 41(1):  1-12.

16     DeNavas-Walt, C.; Proctor, B.D.; Lee, C.H. (2006) Income, Poverty, and Health Insurance Coverage in the United
17             States: 2005. U.S. Census Bureau, Current Population Reports, P60-231. U.S. Government Printing Office,
18             Washington, DC.

19     Henderson, R. (2006) Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory Committee, to
20             Administrator Stephen L. Johnson. Re: Clean Air Scientific Advisory Committee (CASAC) Lead Review
21             Panel's Consultation on EPA's draft Analysis Plan for Human Health and Ecological Risk Assessment for
22             the Review of the Lead National Ambient Air Quality Standards. July 26, 2006.

23     ICF, 2005. Estimating near roadway populations and areas for HAPEM6. Memorandum to Chad Bailey,  Office of
24             Transportation and Air Quality, U.S. EPA. December 28.  Docket EPA-HQ-OAR-2005-0036.

25     Lanphear, B. P.; Roghmann, K. J. (1997) Pathways of lead exposure in urban children. Environ. Res. 74: 67-73.

26     Lanphear, B. P.; Burgoon, D. A.; Rust, S. W.; Eberly, S.; Galke, W. (1998) Environmental exposures to lead and
27             urban children's blood lead levels. Environ. Res. 76: 120-130.

28     Lanphear, B. P.; Dietrich, K. N.;Auinger, P.; Cox, C. (2000) Cognitive deficits associated with blood lead
29             concentrations <10 ug/dL in US children and adolescents. Public Health Reports.  115: 521-529.

30     Lanphear, B. P.; Hornung, R.; Khoury, J.; Yolton, K.; Baghurst, P.; Bellinger, D. C.; Canfield, R. L.; Dietrich, K.
31             N.; Bornschein, R.; Greene, T.; Rothenberg, S. J.; Needleman, H. L.; Schnaas, L.; Wasserman, G.;
32             Graziano, J.; Roberts, R. (2005) Low-level environmental lead exposure and children's intellectual
33             function: an international pooled analysis. Environ. HealthPerspect. 113: 894-899.

34     Leggett, R. W. (1985) A model of the retention, translocation and excretion of systemic Pu. Health Phys. 49: 1115-
35             1137.

36     Leggett, R. W. (1992a) A retention-excretion model for americium in humans. Health Phys. 62: 288-310.

37     Leggett, R. W. (1992b) A generic age-specific biokinetic model for calcium-like  elements.  Radiat. Prot. Dosim.41:
38             183-198.
              December 2006                        3-26         Draft - Do Not Quote or Cite

-------
  1    Leggett, R. W. (1993) An age-specific kinetic model of lead metabolism in humans. Environ. Health Perspect. 101:
  2            598-616.

  3    Muldoon, S. B.; Cauley, J. A.; Kuller, L. H. ; Morrow, L.; Needleman, H. L. ; Scott, I; Hooper, F. J. (1996) Effects
  4            of blood lead levels on cognitive function of older women. Neuroepidemiology 15: 62 72.

  5    National Center for Health Statistics. (2005) Health, United States, 2005.  With Chartbook on Trends in the Health
  6            of Americans. Hyattsville, Maryland.

  7    Payton, M.; Riggs, K. M.; Spiro, A., Ill; Weiss, S. T.; Hu, H. (1998) Relations of bone and blood lead to cognitive
  8            function: the VA Normative Aging Study.  Neurotoxicol. Teratol. 20: 19 27.

  9    Rhodes, D.; Spiro, A., Ill; Aro, A.; Hu, H. (2003) Relationship of bone and blood lead levels to psychiatric
10            symptoms: The Normative Aging Study. J. Occup. Environ. Med. 45: 1144 1151.

11    Rice, D.C. (1996) Behavioral effects of lead: commonalities between experimental and epidemiologic data. Environ
12            Health Persp 104 (Suppl 2): 337-351.

13    Rothenberg, S.J.; Rothenberg, J.C. (2005)  Testing the dose-response specification in epidemiology: public health
14            aand policy consequences for lead. Environ. Health Perspect. 113: 1190-1195.

15    Schwemberger, MS, JE Mosby, MJ Doa, DE Jacobs, PJ Ashley, DJ Brody, MJ Brown, RL Jones, D Homa. May 27,
16            2005 Mortality and Morbidity Weekly Report 54(20):513-516.

17    U.S. Census Bureau. 2006. American Housing Survey for the United States: 2005. Current Housing Reports, Series
18            H150/05. U.S. Government Printing Office, Washington DC.

19    U.S. Environmental Protection Agency. (1978) National Primary and Secondary Ambient Air Quality Standards for
20            Lead. Federal Register 43(194): 46246-46263. Oct 5, 1978.

21    U.S. Environmental Protection Agency. (1986a) Air quality criteria for lead. Research Triangle Park, NC: Office of
22            Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
23            EPA-600/8-83/028aF-dF. 4v. Available from: NTIS,  Springfield, VA; PB87-142378.

24    U.S. Environmental Protection Agency. (1990) Review of the National Ambient Air Quality Standards for Lead:
25            Assessment of Scientific And Technical Information: OAQPS Staff Paper. Research Triangle Park, NC:
26            Office Of Air Quality Planning and Standards; report no. EPA-450/2-89/022. Available from: NTIS,
27            Springfield, VA; PB91-206185. http://www.epa.gov/ttn/naaqs/standards/pb/s_pb_pr sp.html.

28    U.S. Environmental Protection Agency. (1994a) Guidance manual for the integrated exposure uptake biokinetic
29            model for lead in children. Washington, DC: Office of Emergency and Remedial Response; report no.
30            EPA/540/R-93/081. Available from: NTIS, Springfield, VA; PB93-963510.

31    U.S. Environmental Protection Agency. (1994b) Technical support document: parameters and equations used in
32            integrated exposure uptake biokinetic model for lead in children (v 0.99d). Washington, DC: Office of
33            Solid Waste and Emergency Response; report no. EPA/540/R-94/040. Available from: NTIS, Springfield,
34            VA; PB94-963505.

35    U.S. Environmental Protection Agency. (2003) Framework for Cumulative Risk Assessment. Risk Assessment
36            Forum, Washington, DC, EPA/630/P-02/00IF. May

37    U.S. Environmental Protection Agency. (2006) Air  Quality Criteria for Lead. Washington, DC, EPA/600/R-
38            5/144aF. Available online at: www.epa.gov/ncea/
              December 2006                        3-27          Draft - Do Not Quote or Cite

-------
 1    Weisskopf, M. G.; Wright, R. O.; Schwartz, I; Spiro, A., Ill; Sparrow, D.; Aro, A.; Hu, H. (2004a) Cumulative lead
 2            exposure and prospective change in cognition among elderly men. The VA Normative Aging Study. Am. J.
 3            Epidemiol. 160: 1184 1193.

 4    White, P. D.; Van Leeuwan, P.; Davis, B. D.; Maddaloni, M.; Hogan, K. A.; Marcus, A. H.; Elias, R. W. (1998) The
 5            conceptual structure of the integrated exposure uptake biokinetic model for lead in children. Environ.
 6            Health Perspect. 106(suppl. 6): 1513-1530.

 7    Wright, R. O.; Hu, H.; Silverman, E. K.; Tsaih, S. W.; Schwartz, J.; Bellinger, D.; Palazuelos, E.; Weiss, S. T.;
 8            Hernandez Avila, M. (2003) Apolipoprotein E genotype predicts 24 month bayley scales infant
 9            development score. Pediatr. Res. 54: 819 825.

10
              December 2006                        3-28         Draft - Do Not Quote or Cite

-------
 1                       4   CHARACTERIZATION OF HEALTH RISKS

 2          4.1   INTRODUCTION
 3           This chapter presents the human exposure and health risk assessments conducted in
 4    support of the current review (throughout the remainder of this chapter, the term "risk
 5    assessment" will be used to refer to both the human exposure and health risk assessments
 6    collectively, unless specific reference to either the human exposure or health risk assessments is
 7    required). This risk assessment is being completed in two phases. The first phase (the pilot) is
 8    reflected in this first draft document.  The second phase (the full-scale assessment) will be
 9    reflected in the second draft of this document. The pilot assessment is intended primarily as a
10    demonstration of the risk assessment methodology being developed for the current review.
11    Consequently, exposure and risk results presented here are considered preliminary. The  pilot
12    assessment presents exposure and risk assessments for two air quality scenarios (current
13    conditions and attainment of the current NAAQS).
14           The risk assessment characterizes exposure and risk resulting from exposure to policy-
15    relevant sources of Pb (see Section 2.1).  Accordingly, the risk assessment is multi-pathway in
16    nature, including consideration both for direct inhalation and incidental ingestion of Pb in soil
17    and indoor dust that was originally released into outdoor air. In addition, because of the  need to
18    consider total Pb exposure in predicting blood Pb levels, the analysis also includes consideration
19    of policy-relevant background Pb exposure. As described in Section 2.1, exposure pathways
20    comprising background include diet, drinking water and Pb  paint (with Pb paint potentially
21    impacting both outdoor soil and indoor dust).1
22           As with the last review (see Section 4.1.1), this risk assessment utilizes  a case-study
23    approach wherein a set of specific locations associated with policy-relevant Pb  exposures are
24    evaluated in detail.  For the pilot analysis, three case studies have been selected including (a) a
25    primary Pb smelter (in Herculaneum Missouri), (b) a secondary Pb  smelter (in Troy, Alabama)
26    and (c) a near roadway (urban) location in Houston, Texas.2 Additional case studies may be
             1  In the pilot assessment, the contribution to food from air pathways is not explicitly included, such that
      dietary Pb exposure is treated as policy-relevant background. Further, although paint is a policy-relevant
      background_source, for this analysis, it may be reflected somewhat in estimates developed for policy-relevant
      sources, due to modeling constraints (see Section 4.4.3.3.2 and 4.5.2).
             2 Note, that the near roadway (urban) case study comprises a 1.5 mile road segment and the residents living
      within 200m of that road segment. Consequently, this case study provides perspective on the near roadway exposure
      scenario but is not intended to estimate total population risk for a larger urban or metropolitan area.  Such an area
      would likely include a large number of such road segments with buffered residential populations.
               December 2006                  4-1             Draft - Do Not Quote or Cite

-------
 1    evaluated for the full-scale analysis. The case studies modeled for the pilot have been selected to
 2    provide a preliminary perspective on the nature and magnitude of air-sourced Pb exposures and
 3    risk.  In addition, they provide a range of exposure scenarios in which to test the risk assessment
 4    methodology developed for the current review. Because of differences in the exposure scenarios
 5    and available data at each of the case study locations, the approach used for modeling exposure
 6    and risk differs among the case studies. Results from this pilot assessment, as well as comments
 7    received from the public and CASAC will inform staff decisions on the number and type of case
 8    studies to be included in the full-scale assessment.
 9          The remainder of this chapter is organized as follows. Section 4.1.1 provides an
10    overview of the human health risk assessment completed in the last review of the Pb NAAQS in
11    1990 (USEPA, 1990a).  Section 4.2 provides an overview of the  scope of the pilot exposure and
12    risk assessment, covering such topics as the conceptual model used in designing the analysis
13    (4.2.1), the selection of health endpoints and case study locations (4.2.2 and 4.2.3 respectively), a
14    description of the air quality scenarios covered in the pilot (4.2.4), and an overview of the key
15    components of the pilot exposure and risk assessment (4.2.5). Following the discussion of the
16    scope of the analysis, separate sections are dedicated to the exposure  analysis (4.3) and risk
17    assessment (4.4). Specifically,  Section 4.3 documents the methods and results of the human
18    exposure assessment completed for the analysis, which follows the analysis through the
19    estimation of blood Pb levels for child populations at the three case study locations. Section 4.4
20    presents the methods and results of the health risk assessment which characterizes the
21    distribution of IQ loss resulting from Pb exposure in the modeled child populations.  Separate
22    sections are dedicated to performance evaluation (4.3.6) and sensitivity analysis and uncertainty
23    (4.4.3). Finally, a summary of the risk results generated for the pilot, including discussion of
24    uncertainty and the identification of areas for potential enhancement for the full-scale analysis is
25    presented in Section 4.5.
26          Additional technical detail regarding both the exposure analysis and risk assessment
27    completed for the pilot analysis (beyond that presented in this chapter) can be found in Lead
28    Human Exposure and Health Risk Assessments and Ecological Risk Assessment for Selected
29    Areas (ICF, 2006; henceforth referred to as the Risk Assessment Report).

30         4.1.1  Overview of Risk Assessment from Last Review
31          In the risk assessment conducted in support of the last review, air quality scenarios were
32    compared in terms of their impact on the percentage of modeled populations that exceeded
33    specific blood Pb levels chosen with consideration of the health effects evidence at that time
34    (USEPA, 1990). The  1990 analysis focused on both children (birth through 7 years of age) and
35    middle-aged men residing in three case study locations (two near secondary Pb smelters and one

              December 2006                 4-2            Draft - Do Not Quote or Cite

-------
 1    near a primary Pb smelter). The analysis also introduced the use of pharmacokinetic blood Pb
 2    modeling for children, although it used empirically-derived slope models for adult men to relate
 3    changes in air Pb to changes in blood Pb.
 4           The following discussion presents a more detailed overview of the approach used in the
 5    1990 risk assessment. For children, the staff used blood Pb levels of 10 and 15 ug/dL to evaluate
 6    effects of alternate NAAQS.  These values were chosen with consideration of the health effects
 7    evidence at that time. Staff then used dispersion modeling (the Industrial Source Complex (ISC)
 8    model) combined with source characterization data for each point source to generate Pb air
 9    concentrations for each case study area. Statistically-derived relationships based on data from
10    other industrial locations, including Pb smelters, that linked concentrations of Pb in air to Pb in
11    indoor dust and outdoor soil were then used to predict Pb in these media for the three case study
12    locations, based on the modeled air Pb concentrations. An uptake/biokinetic model was also
13    developed to predict child blood Pb levels. This model was used in place of a statistically-based
14    regression slope model to allow consideration for the dynamic nature of Pb exposure in children.
15    EPA combined model-derived central tendency blood Pb levels with an estimated geometric
16    standard deviation (GSD) reflecting inter-individual variability in blood Pb levels, to generate
17    population distributions of blood Pb levels. These distributions were then used to estimate the
18    percentage of children at each case study location that exceeded the specified blood Pb levels of
19    interest for children (i.e., 10 and 15 ug/dL).
20           For adult men, the 1990 assessment used blood Pb levels of 10 and 12 ug/dL to compare
21    relative effects of alternate NAAQS. The same approach was used for generating media
22    concentrations for the adult analysis as was used for the child assessment. However, for adults as
23    noted above, the 1990 analysis used statistically-derived slope models to relate air Pb to blood Pb
24    levels with two versions of the slope models being employed: (a) the aggregate model which
25    predicts blood Pb in adults based solely on air Pb levels (here a single slope factor captures both
26    the direct inhalation pathway as well as the more complex pathway of Pb deposition to soil and
27    dust followed by incidental ingestion) and (b) the disaggregate model which uses media-specific
28    slopes to predict blood Pb based on Pb concentrations in soil, dust and air.  Since the projected
29    blood Pb levels were mean population levels, a GSD term was included to develop population-
30    level blood Pb distributions. The GSD estimates for adults and children were derived from
31    information on observed blood Pb levels in these subgroups. These population-level distributions
32    were then queried to identify the percentage of adult men at each case study location with
33    modeled blood Pb levels exceeding the levels of interest for adults (10 and 12 ug/dL).
34           The primary difference between the risk assessment approach used in the current pilot
35    analysis and the assessment completed in 1990 involves the risk metric employed. Rather than
36    estimating the percentage of study populations with exposures above blood Pb levels of interest
              December 2006                  4-3             Draft - Do Not Quote or Cite

-------
 1    as was done in the last review (i.e., 10, 12 and 15 ug/dL), the current pilot analysis estimates the
 2    degree of health decrement in study populations exposed to Pb. Specifically, the pilot analysis
 3    estimates the distribution of IQ loss associated with Pb exposure for child populations at each of
 4    the case study locations with that IQ loss further differentiated between background Pb exposure
 5    and policy-relevant exposures.

 6         4.2  SCOPE OF PB EXPOSURE AND RISK ASSESSMENTS
 7          This section describes the scope of the pilot analysis, including key elements in the
 8    design of both the exposure and risk assessments.

 9         4.2.1   Conceptual Model for Human Health Risk Assessment
10          This section presents the conceptual model (Figure 4-1) for the exposure and risk
11    assessment intended to illustrate the elements pertinent to evaluating public health risks
12    associated with environmental Pb exposures.
13          For the purposes of this risk assessment, as noted earlier, "background" refers to sources
14    of, and exposures to Pb associated with pathways that do not involve the release of Pb to ambient
15    outdoor air. Included among these would be pathways associated with indoor Pb paint, Pb in
16    drinking water, Pb introduced to food during processing, etc. Those elements considered
17    "background" for this analysis are shown in Figure 4-1 in non-bold (regular) type. As shown in
18    Figure 4-1, the pilot assessment will include contributions from all background sources. Those
19    sources considered policy-relevant (i.e., associated with the release of Pb to outdoor air) are
20    shown in Figure 4-1 in bold type. As noted earlier, the pilot assessment results for both blood Pb
21    levels and risk are differentiated as to contributions associated with background sources and
22    policy-relevant sources.
23          Based on recommendations from CAS AC and consideration of information presented in
24    the CD, the pilot assessment was designed to focus on children as the study population, blood Pb
25    as the dose metric, central nervous  system as the target for effects, and IQ decrement as the
              December 2006                 4-4            Draft - Do Not Quote or Cite

-------
1     Figure 4-1.   Conceptual model for Pb human health risk assessment.
      LLJ
      W)
      o
      Q.
      X
      LU
      V)
      O
      LU
      LU
      EE

Stationary and
Mobile Sources
i

Soil/Dust -
Outdoors3
REENTRAINM


Oth
Miscelte
I
ENT
                                            Ambient Air
1
DEPOSITION DEPOSITION
4 1
I Air * Indoor 1
^ r i
4
'l
. r
Soil/Dust -
Outdoors
i
'
DEPOSITION
4
Soil/Dust •
Outdoors
i

Soil/Dust -
indoors
i

i

Air - Indoor
DEPOSITION
J
Soil/Dust -
Indoors
i

Inhalation
                                                                                         Diel
                                                           Drinking
                                                            Water
                                                          Lead Paint -
                                                           Indoors
                                                                                                    SoiVDust • indoors
                          Young
                         Children
Adult
Men
 Adult
Women
                                                Blood
       Cardiovascular
          System

         {for adults!
                       Kidney

                     (for adults)
                                      Blood Pressure
                                         Change
                Hypertension
                 Incidence
                      Croalininc
                      Clearance
                      Decrement
          NOTES

          1 Components with gray background and solid borders will be included in this assessment.  Components depicted in gray
          but without borders were not included in the Pilot analysis, but are being considered for inclusion in the Full Scale analysis.
          Further, a distinction is made between components linked to exposure pathways involving ambient air (shown in bold) and
          components involving  other pathways (i.e., background).
          2 Includes contributions of historical sources, including (but not limited to) emissions from the use of leaded gasoline,
          historical emissions from stationary sources, and exterior leaded paints.
                 December 2006
                    4-5
                            Draft - Do Not Quote or Cite

-------
 1    metric associated with risk quantitation. Consequently, some of the elements presented in Figure
 2    4-1 are not reflected in the pilot analysis (e.g., adult populations, internal dose metric of bone Pb,
 3    potential cardiovascular effects in adults and blood pressure changes). Those elements in the
 4    conceptual model which are reflected in the pilot analysis have solid bolded borders and are
 5    shaded. Those excluded from the pilot analysis have light borders and are not shaded.

 6         •  Sources:  The focus of the pilot analysis is on policy-relevant sources. This includes
 7            both current sources of new Pb emissions  (e.g., ongoing industrial point source and
 8            fugitive emissions) and re-emission of historically deposited Pb (e.g., re-entrainment of
 9            Pb deposited historically near roadways).  However, because of the importance of
10            characterizing total Pb exposures in modeling blood Pb levels, other ("background")
11            sources of environmental Pb (e.g., diet and drinking water) are also considered to
12            varying degrees in the analysis.

13         •  Pathways:  Figure 4-1 is intended to generally illustrate the many pathways by which
14            Pb emitted into the environment becomes  available for human exposure. Those not
15            passing through ambient air are considered "background" for the purposes of this
16            assessment.

17         •  Routes:  The ingestion and inhalation routes are considered the primary routes of
18            human exposures to environmental Pb. The ingestion route (including incidental dust
19            and soil ingestion) is expected to have a greater contribution to the risk estimates. Both
20            routes are included in this assessment.

21         •  Exposed Populations: The Pb exposed populations can be characterized and stratified
22            based on  a variety of characteristics. Figure 4-1 identifies groups based primarily on
23            age or lifestage, which has an influence on behaviours that can influence exposure or
24            susceptibility (see Section 3.2 and 3.4.1 for additional detail on susceptible
25            populations). It is recognized that more specific factors (e.g., calcium deficiency) also
26            influence susceptibility.  However, currently available data do not generally support
27            quantitative modeling that differentiates between subpopulations on the basis of
28            enhanced or reduced susceptibility to Pb effects (e.g., concentration response functions
29            for IQ loss that differentiate between populations that are calcium  deficient and those
30            that are not).

31         •  Internal Disposition:  While Pb is distributed throughout the body, bone is an
32            established site of internal accumulation of Pb, while blood is an established internal
33            dose metric for purposes of both exposure and risk assessment.  The  pilot analysis
34            relies on blood Pb with corresponding dose-response functions.  However, the tools
35            employed in modeling blood Pb  levels in study populations recognize the role of bone
36            as a reservoir with the potential to act as both a source and storage site.

37         •  Endpoints:  Figure 4-1 generally identifies the wide variety of health endpoints
38            recognized in the CD (CD, p. E-8) as associated with Pb exposures.  As mentioned
39            previously, the endpoint of interest for this assessment is neurological effects in
40            children and more specifically, IQ decrement.
              December 2006                  4-6            Draft - Do Not Quote or Cite

-------
 1         •  Metrics: Figure 4-1 generally recognizes that there are many metrics that might be
 2            considered for risk assessment.  Recognizing the need for the metrics used in this
 3            assessment to have sufficient support for use in quantifying population health risk, the
 4            pilot analysis uses IQ decrement in children as its primary risk metric.
 6         4.2.2   Selection of Health Endpoint, Study Population, Dose-Metrics and Associated
 7                Concentration-Response Function
 8          As recognized in Chapter 3 (e.g., Section 3.3.1), the CD notes that recent epidemiological
 9    studies have strengthened the consensus that the developing nervous system is the most sensitive
10    endpoint in young children and that neurobehavioral deficits, including IQ impacts, appear to
11    occur at lower levels than previously believed (i.e. at levels < 10 ug/dL and possibly <5 ug/dL).
12    Consequently, for the pilot risk assessment, we selected the child neuro-developmental endpoint,
13    specifically focusing on the IQ loss metric (see Section 3.3.1.1). This assignment of priority to
14    children's IQ  as the endpoint assessed reflects consideration for evidence presented in the CD as
15    well as advice received during our consultation with CAS AC on the Analysis Plan for Human
16    Health and Ecological Risk Assessment For the Review of the Pb National Ambient Air Quality
17    Standards (USEPA, 2006a) (hereafter referred to as the Analysis Plan)  (Henderson, 2006).
18          As discussed in Section 3.3.1.2, a pooled analysis (Lanphear et al., 2005) was selected as
19    the basis for defining the relationship between Pb exposure and IQ loss. Furthermore, the pilot
20    analysis will use log-linear functions based on concurrent and lifetime average blood Pb metrics,
21    as discussed in Section 3.3.1.2. This decision was made after considering a range of
22    concentration-response functions provided in the pooled analysis (e.g.,  cubic-spline and log-
23    linear models).

24         4.2.3   Selection of Case Study Locations
25          In identifying the three case study locations modeled for the pilot analysis, the staff
26    followed the approach presented in the Analysis Plan.  The Analysis Plan included examples of
27    ambient Pb emissions sources and exposure scenarios that should be considered in selecting case
28    studies. These include a study area near a primary Pb smelter, one near another (or multiple)
29    significant stationary Pb sources, and a study area near an urban roadway. During our
30    consultation on the plan, CAS AC  generally expressed support for the approach, emphasizing the
31    need for inclusion of a near roadway location to provide coverage for the potential impact of
32    historically deposited (auto-sourced) Pb on public health (Henderson, 2006). The case studies
33    we have included in the pilot reflect these three situations and also reflect consideration for three
34    additional factors described in the Analysis Plan: (a) availability of site-specific monitoring data
35    for ambient air Pb, (b) availability of measurement data for other environmental media (soil and

              December 2006                 4-7             Draft - Do Not Quote or Cite

-------
 1    indoor dust) and biomonitoring of Pb exposure (i.e., blood Pb levels), and (c) consideration for
 2    demographics and socioeconomic factors related to Pb exposure and risk.
 3          The three case study locations modeled for the pilot assessment are (a) a primary Pb
 4    smelter (in Herculaneum Missouri), (b) a secondary Pb smelter (in Troy, Alabama) and (c) a near
 5    roadway location in Houston, Texas. The Herculaneum primary Pb smelter represents a
 6    relatively large point source that has been active for over a century and for which there exist a
 7    large amount of site-specific data characterizing both  media concentrations (soil, indoor dust,
 8    outdoor air) and population blood Pb levels. The secondary Pb smelter represents a somewhat
 9    smaller point source (compared to Herculaneum) with relatively less site-specific data
10    characterizing media concentrations and exposure levels. Finally, the near roadway (urban)
11    location meets the criteria that we identified for a near roadway exposure scenario (e.g., little
12    influence of point source emissions in the immediate vicinity, conditions which could contribute
13    to population exposure to near roadway-deposited Pb, and residential populations located near
14    the modeled road segment). We had also intended to identify a multiple-source case study
15    location (i.e., a location with multiple point and/or area sources with none being dominant).
16    However,  identification of a satisfactory location proved challenging preventing inclusion of this
17    type of case study in the pilot. Efforts may continue, during the full-scale analysis, to identify a
18    suitable multiple-source location.
19          Each of the three case study locations modeled for the pilot is briefly discussed below
20    including: (a) details related to each location which may be relevant to exposure and risk, (b) (for
21    the two point source locations) the magnitude of reported emissions for the facility, (c) the
22    magnitude of ambient air Pb levels at monitors associated with the case study location in the
23    context of overall monitored levels in the U. S. and (d) the availability of site-specific data
24    characterizing key media and Pb exposures (e.g., soil, blood Pb level data). The approach used
25    to identify the near roadway (urban) case study location, which is fairly complex, is also
26    discussed in some detail below.

27         4.2.3.1  Primary Pb Smelter Case Study
28          The facility in Herculaneum Missouri is the only remaining primary Pb smelter operating
29    within the U.S. It has been in operation for over a century contributing to Pb contamination of
30    the area surrounding the facility.3 However, over the past decade or more, remediation of yard
             3 Portions of this study area comprise an active Superfund site and are subject to ongoing evaluation under
      the Superfund program administered by the Office of Solid Waste and Emergency Response. Methods used in
      conducting the human health exposure and risk assessment for the pilot analysis have been selected to address policy
              December 2006                  4-8             Draft - Do Not Quote or Cite

-------
 1    soils have been completed for a significant number of the residences near the facility and in
 2    addition, the facility has extended its site boundaries to encompass many of the more heavily
 3    impacted houses and lots. The remediation activity introduced a complication to the risk
 4    modeling, especially aspects involving characterization of the relationship of ambient air Pb and
 5    residential soil Pb to indoor dust Pb (see Section 4.3.2.4).
 6          The U.S. Census estimates that, as of 2000, a total of 37,562 people live within  10 km of
 7    the facility (U.S. Census Bureau, 2005). Of these, 3,880 are children (0-7 years of age).
 8    Specifically, 171 children are within 2 km of the facility, 1,545 children are between 2 and 5 km
 9    and 2,164 children are between 5 and 10 km from the facility.
10          Total annual Pb emissions to ambient air for the Herculaneum facility are estimated at 25
11    tons/year for 2005 (see Section 2.3.4.6).
12          As of 2005, there are nine TSP air monitors located within 10 km of the Herculaneum
13    facility (as identified in EPA's Air Quality System database). Annual average concentrations of
14    Pb recorded at these monitors for 2005 ranged from (0.057 to 1.56 ug/m3) (USEPA, 2006b).  Of
15    the nine TSP monitors located within 10 km of the facility, all fall within the top 30% of the
16    2005 annual average levels for monitors in the database, with four of the nine monitors falling in
17    the top 10% (see Risk Assessment Report, Section 3.1.5 for additional detail).
18          The Herculaneum facility has more site-specific monitoring data available to support risk
19    assessment then the other two case study locations, including residential yard soil, indoor dust
20    and road dust Pb measurements collected in areas potentially impacted by the facility. In
21    addition, the Agency for Toxic Substances and Disease Registry (ATSDR) has conducted a
22    number of health consultations which involved the collection of blood Pb measurements for
23    children (ATSDR, 2003). The area within the city limits of Herculaneum is designated non-
24    attainment for the Pb NAAQS and a State Implementation Plan (SIP) was approved in 2002 (67
25    FR 18497). EPA determined the existing SIP to be inadequate to attain the current NAAQS in
26    2006 (71 FR 19432), and consequently a revised SIP is under development for the area. Air
27    dispersion modeling will play a role in development of this plan (Note, we intend to consider that
28    updated air quality modeling in conducting the full scale analysis for this case study).
29          The significant amount of site-specific data available for Herculaneum, paired with air
30    dispersion modeling for the facility conducted in support of SIP development for Pb, provides a
31    strong data set for this study area which enhances the modeling of exposure and risk. However,
32    the Herculaneum case study location also has a number of attributes that add complexity to the
33    modeling of Pb exposure and risk including (a) complex terrain and meteorology which

      questions relevant to the Pb NAAQS review and consequently, may differ from those used by the Superfund
      program.
              December 2006                 4-9            Draft - Do Not Quote or Cite

-------
 1    complicates the modeling of Pb transport in ambient air, (b) a large and complex facility with
 2    significant opportunity for fugitive emissions which makes source characterization challenging,
 3    and (c) a history of remediation activities which has contributed to widely varying residential soil
 4    Pb concentrations across the town.

 5         4.2.3.2  Additional Point Source (Secondary Smelter) Case Study
 6          The secondary Pb facility in Troy, Alabama, is one of 15 secondary Pb smelters operating
 7    within the U.S. as of 2002 (see Section 2.3.4.7) Secondary Pb smelters produce Pb from scrap
 8    and provide the primary means for recycling Pb-acid automotive batteries.
 9           According to the U.S. Census (US Census, 2005), as of 2000, a total of 17,901 people
10    live within 10 km of the facility. Of these, 1,672 are children (0-7 years of age). Specifically, 187
11    children are within 2 km of the facility, 896 children are between 2 and 5 km and 589 children
12    are between 5 and 10 km from the facility.
13          Total annual Pb emissions to ambient air for the Troy facility are estimated for 2002 at
14    4.56 tons/year (EC/R, 2006), which places the facility as the third highest emitter within its
15    source category.
16          There are two TSP air monitors located within  1 km of the  Troy facility, specifically
17    located between 300 and 800 m from the facility (as identified in EPA's Air Quality System
18    database). Annual average concentrations of Pb recorded at these monitors for 2000 range from
19    0.198 to 0.383 ug/m3 (USEPA, 2006b). These TSP values fall within the top 15% of TSP annual
20    average values for 2000 (see Risk Assessment Report, Section 3.2.5 for additional details).
21          In contrast to the Herculaneum facility, we have not identified soil or indoor dust Pb
22    measurements for this case study location and we  did not identified systematic blood Pb
23    assessments for children in the area.  This means that the exposure assessment conducted for the
24    Troy facility is more dependent on modeling  and consideration of measurements available for
25    similar locations and there is less opportunity for rigorous performance evaluation of modeling
26    due to the lack of site-specific measurement data.
27         4.2.3.3  Near Roadway (Urban) Location Case Study
28          In choosing a location for this case study, focus was placed on identifying a location with
29    ambient monitors that would provide reasonable coverage for near roadway Pb entrainment. The
30    team reviewed available air monitoring data from  two sources (a) the EPA's Air Quality System
31    database,  focusing on  TSP monitoring (USEPA 2006b) and (b) a series of air monitors placed by
32    Desert Research Institute (DRI) in support of an urban particulate matter speciation study (ICF,
33    2005). Ultimately, the DRI monitors were favored for the near roadway (urban) case study
34    because of their particulate matter speciation  data  which may be used in the full-scale analysis to
35    support more detailed source-apportionment of the Pb concentrations. In addition, in selecting
              December 2006                 4-10            Draft - Do Not Quote or Cite

-------
 1    the location for this case study, we also considered modeling conducted for the Agency's 1999
 2    NATA national-scale assessment (USEPA, 2006c). As described in Section 2.4.5.1, air
 3    dispersion modeling is conducted in this assessment at the census tract-level and includes
 4    modeling of all sources in the 1999 NEI.  Because the NEI does not include re-suspension as an
 5    emissions source, instances where measured (monitored) air Pb levels exceed predicted ambient
 6    air Pb concentrations may indicate potential contributions from re-entrainment.  Therefore, in
 7    choosing the near roadway (urban) case study location, we also favored those locations with a
 8    positive difference between monitored air Pb levels and the national-scale assessment modeled
 9    ambient air  concentrations. And finally, we also, favored those DRI monitor locations that have
10    relatively low area and point Pb emissions within 20 km, thus reducing the chance of Pb
11    emissions besides re-entrainment having a significant impact on Pb air concentrations in the
12    study area (i.e., increasing the chances that re-entrainment is the dominant Pb air source).
13          Ultimately, a case study location in Houston TX, comprising a 1.5 mile road segment and
14    the adjacent residential population, was selected to represent the near roadway exposure  scenario
15    based on the above criteria. This location has the following  characteristics, which contributed to
16    its inclusion in this analysis: (a) it is located in a relatively dry location where resuspension is
17    likely to occur, (b) DRI measured Pb concentrations in air are higher than the NATA modeled air
18    concentration estimates, and (c) there are relatively few point  sources of Pb emissions within 20
19    km of the site (no point sources within 5 km and only two within 20 km). Note,  however that one
20    of the two emissions sources is a large airport (airports are a potentially significant source of Pb
21    emissions due to the continued use of Pb in some aviation fuels) (CD Section 2.2.4).
22          According to the U.S. Census (U.S. Census Bureau,  2005), as of 2000, a total of  1,950
23    people live within the 200m buffer area surrounding the  modeled road segment. Of these, 320
24    are children (0-7 years of age). It is important to note, that the near roadway (urban) case study
25    was developed primarily to demonstration test the exposure and risk characterization approach
26    developed for this type of case study. It is likely that a far larger number of children would be
27    associated with near roadway exposures if a study area similar to that used for the two point-
28    source case  studies were considered (i.e., a larger number of road segments similar to that
29    evaluated here would be located within a 10 km-radius urban or suburban area).
30          The  DRI monitors measured PMi0 and collected  samples over three days in February,
31    2001. These resulted in an average air Pb concentration over this sampling period of 0.0030
32    ug/m3. Compared with annual average Pb concentrations obtained from the 36 PMio monitors in
33    the Air Quality System database (USEPA 2006b) the DRI monitor average falls around the lower
34    20th% (see Risk Assessment Report, Sections 3.3.4 and  4.3.2.1 for additional detail on
35    monitored ambient air data used in this assessment).

              December 2006                 4-11             Draft - Do Not Quote or Cite

-------
 1          As with the Troy case study location, the Houston near roadway location does not have
 2    any site-specific soil Pb data or monitored Pb exposure data in the form of blood Pb levels. This
 3    means that the exposure assessment conducted for the near roadway (urban) case study is more
 4    dependent on surrogate data collected from other industrial locations and there is less
 5    opportunity for rigorous performance evaluation of modeling due to the lack of site-specific
 6    measurement data.

 7         4.2.4   Air Quality Scenarios Covered in the Pilot Analysis
 8          The pilot analysis involved exposure and risk modeling for the current conditions and
 9    current NAAQS attainment air quality scenarios. Consideration for alternate NAAQS scenarios
10    will be covered as part of the full-scale analysis.
11          The current NAAQS attainment air quality scenario only applies to the primary Pb
12    smelter case study since monitors located within that study area have shown exceedances of the
13    Pb NAAQS (USEPA, 2006b). Monitors associated with the other two case study locations have
14    not had recent exceedances of the Pb NAAQS and consequently the current conditions and
15    current NAAQS attainment air quality scenarios are identical for these two locations.
16          Consideration of the current NAAQS attainment scenario for the primary Pb smelter case
17    study involved adjustment of modeled air quality results to reflect attainment of the current
18    NAAQS (see Section 4.3.2.1.1 for discussion of air quality modeling for this case  study).
19    Specifically, any quarterly-average modeled air concentrations at receptor points within the study
20    area estimated to exceed the NAAQS were reduced to the NAAQS level of 1.5 ppm. Annual
21    average air concentrations were then recalculated for those receptor points based on these
22    adjusted quarterly averages (as noted in Section 4.3.2, exposure modeling for the pilot analysis is
23    based on annual average media concentrations).

24         4.2.5   Overview of the Exposure and Risk Modeling Approach Used in the  Pilot
25                Analysis
26          This section provides an overview of the modeling approach implemented  for the pilot
27    analysis including: (a) consideration for recommendations provided by CASAC regarding
28    exposure and risk modeling, (b) description of the child study population evaluated for the pilot
29    analysis, (c) spatial scale of the analysis and the type of spatial template used in modeling, (d)
30    overview of key modeling steps (predicting media concentrations, modeling exposure, modeling
31    risk), (e) performance evaluation completed in support of the analysis and (f) the approach used
32    to address uncertainty.
              December 2006                  4-12           Draft - Do Not Quote or Cite

-------
 1          4.2.5.1  CASAC Consultation Regarding Human Exposure and Health Risk
 2                  Assessment
 3           The staff consulted with the CASAC on the draft Analysis Plan in June, 2006
 4    (Henderson, 2006). Some key comments provided by CASAC members on the Analysis Plan
 5    included: (a) placing a higher priority on modeling the child IQ metric than the adult endpoints
 6    (e.g., cardiovascular effects), (b) recognizing the importance of indoor dust loading by Pb
 7    contained in outdoor air as a factor in Pb-related exposure and risk for sources considered in this
 8    analysis, and (c) concurring with use of the IEUBK biokinetic model, suggesting that the linking
 9    of a probabilistic exposure module with IEUBK run in batch-mode to generate blood Pb
10    estimates for individual modeled children (as described  in the Analysis Plan) should be given
11    lower priority by staff in the pilot phase.4
12           The above comments, together with information presented in the CD, are reflected in the
13    design of the pilot analysis.  Important ramifications of this decision include: (a) a focus on
14    modeling exposure and risk (IQ loss) for the child cohort in the pilot analysis, (b) emphasis
15    placed on explicitly considering the linkage between ambient air-borne Pb  and Pb in indoor dust
16    and (c) development of a GIS-based blood Pb modeling approach that combines IEUBK with
17    spatially-differentiated exposure (dose) modeling, but still relies on the GSD-based approach for
18    ultimately characterizing variability in blood Pb levels related to behavior and biokinetics.

19          4.2.5.2  Child Study Population
20           As mentioned earlier, the pilot analysis focuses on estimating IQ loss in children
21    associated with exposure to  Pb  derived from policy-relevant sources within each study area. The
22    risk assessment conducted for each case study uses a simulated child population that begins
23    exposure in the same year and continues that exposure for 7 years (i.e., the study population is
24    assumed to be a single group, which begins exposure  in the study area all at the same age and
             4 Note, as discussed in Section 4.3.3.3, for the pilot we did implement a probabilistic population exposure
      model, however this model is based on coverage for inter-individual variability in behavior related to Pb exposure
      using the GSD approach described in the IEUBK Technical Support Document (USEPA 1994) and does not involve
      deterministic modeling of individual Pb exposure levels outside of IEUBK (i.e., the population-exposure modeling
      used in the pilot does not utilize the approach deemphasized by CASAC). Note also, that in relation to blood Pb
      modeling, CASAC has expressed support for the use of empirically-based regression models in addition to
      biokinetic models. As discussed in Section 4.2.5.5.2, the pilot analysis utilizes both categories of models with
      biokinetic blood Pb modeling being used to generate exposure and risk estimates in the analysis proper, and
      regression-based blood Pb modeling being considered as part of the sensitivity analysis completed for the pilot (see
      Section 4.4.3).
               December 2006                  4-13            Draft - Do Not Quote or Cite

-------
 1    continues that exposure until all modeled children are 7 years of age).5 Furthermore, it is
 2    assumed that no migration or immigration of these children occurs during this simulation period
 3    (i.e., none of the children move out of the study area and no children move in). With regard to
 4    characterizing Pb concentrations in media as part of exposure analysis, it is assumed that the
 5    media concentrations, after being defined for the start period of the simulation (when the
 6    simulated children are 1  day of age), hold steady and do not change6.  Note, however, that a
 7    variety of exposure factors and physiological parameters used in blood Pb modeling are allowed
 8    to change as each simulated child ages. We are considering potential refinements to this
 9    somewhat simplified modeling approach to address the needs of the full-scale assessment.

10         4.2.5.3 Timeframe for Current Conditions
11           The current conditions scenario modeled for the pilot is generally described as reflecting
12    conditions for the timeframe 2000 to 2005. As summarized in the last row of Table 4-1, 2000-
13    2005 data were used to characterize Pb media concentrations for the primary  Pb smelter.
14    Information used in the air and soil modeling for the secondary Pb smelter was collected
15    between 1997 and 2000. And finally, the information used to characterize Pb concentrations for
16    the near roadway (urban) case study reflects conditions in 2001.
             5 Modeling of blood Pb levels for the child population includes contributions representative of prenatal Pb
      exposure.
             6 While air concentrations of Pb exhibit notable temporal variation, particularly near point sources (see
      Section 2.4.3.1.3), less temporal variation is expected for other media. For the purposes of this pilot analysis,
      however, temporal variation (within or across years) is not simulated for any of the media.
              December 2006                  4-14            Draft - Do Not Quote or Cite

-------
 1    Table 4-1.   Timeframe (years) reflected in the characterizations of Pb media
 2                concentrations used in the pilot risk assessment.
Modeling input
Ambient outdoor air
Pb levels
Soil Pb levels
Indoor dust levels
Cumulative period
associated with data
Primary Pb smelter
2001-2005
(post-2001 SIP
emissions data)
2000-2005
(timeframe for soil
sampling at site)
2000-2005
(given that indoor dust
modeling relies on both
outdoor air data and
outdoor soil data, the
timeframe for indoor
dust modeling reflects
data used in
characterizing the other
two media)
2000-2005
Secondary Pb smelter
1997-2000
(stack test data used in
source characterization)
1997-2000
(modeled based on air
deposition)
1997-2000
(estimated based on
ambient air
concentrations and
outdoor soil
concentrations)
1997-2000
Near roadway (Urban)
2001
(DRI monitor data)
1998
(near roadway soil data
collected in Corpus
Christi, TX)
2001
(given that indoor dust
modeling relies on both
outdoor air data and
outdoor soil data, the
timeframe for indoor
dust modeling reflects
data used in
characterizing the other
two media)
-2001
 4         4.2.5.4  Spatial Scale and Resolution
 5           Exposure modeling conducted for the pilot is based on spatial templates customized for
 6    each case study location. These spatial templates divide the study area into smaller exposure
 7    zones which form the basis for exposure and risk modeling. For the point source-related case
 8    study locations (primary and secondary Pb smelters), these exposure zones are a combination of
 9    U.S. Census blocks and/or block groups.7 By contrast, the near roadway (urban) case study uses
10    exposure zones running parallel to the selected road segment, reflecting the focus placed on re-
11    entrainment of near roadway dust and the reduction in Pb concentrations with distance from the
12    road segment.
              US Census block groups vary in size from several city blocks in densely populated urban areas to many
      square miles in less populated rural areas. Their population count varies from 600 to 3000 people per block group
      with the typical block group in the U.S. containing 1,500 people. US Census blocks are more refined than block
      groups and typically contain several hundred people or less. Their size can vary from a single city block in urban
      areas to multiple square miles in less populated rural locations.
               December 2006                 4-15             Draft - Do Not Quote or Cite

-------
 1           At all case study locations, fate and transport modeling and/or empirical data are used to
 2    characterize Pb media concentrations (e.g., outdoor air, soil and indoor dust) for each exposure
 3    zone.  A central tendency estimate of concurrent and lifetime average blood Pb levels is derived
 4    for the children within each exposure zone. Inter-individual variability in blood Pb levels
 5    between children within a given zone is considered through the use of a statistically-derived
 6    GSD reflecting variability in blood Pb levels for children living in areas with similar levels of Pb
 7    contamination. Additional detail on the spatial templates used for each case study location can be
 8    found in Section 4.3.1.

 9         4.2.5.5   Overview of Analytical Steps
10           As illustrated in Figure 4-2, the risk assessment generally includes four analytical steps:
11    (a) characterization of the fate and transport of Pb released into outdoor air, including the
12    dispersion of Pb away from the point of release and the deposition of Pb onto surfaces, (b)
13    prediction of the resulting concentration of Pb  in media of concern including outdoor air, outdoor
14    soil and indoor dust, (c) use of these Pb concentrations together with estimates of Pb in
15    background media including diet to model blood Pb levels using biokinetic modeling and (d)
16    relating modeled blood Pb levels in children to degrees of IQ loss using concentration-response
17    functions derived from  epidemiology studies.  Figure 4-2 identifies the key input data sets,
18    modeling steps and intermediate model output in each of the four analytical steps.  The first two
19    steps are employed in estimating media concentrations, while the third step completes the
20    exposure assessment and the fourth is the risk assessment step. Each of the key elements of each
21    step is briefly described below. Details of the exposure and risk assessments, along with
22    exposure and risk  estimates, are described in Section 4.3 and 4.4, respectively.
23           Note, that the modeling approach discussed below is based on the analysis completed for
24    the current  conditions air quality scenario for the pilot analysis. The method used for the current
25    NAAQSattainment scenario is virtually identical, except that several U.S.  Census blocks within
26    the primary Pb smelter  case study area with air Pb levels exceeding the current NAAQS (1.5
27    |^g/m3) had  those levels reduced to values just meeting the NAAQS. Those adjusted ambient air
28    values were then used in exposure and risk modeling.
              December 2006                 4-16            Draft - Do Not Quote or Cite

-------
1
2
Figure 4-2.     Overview of analysis approach for the pilot analysis.
             1X1
             §
O
§
O

-------
 1    4.1.2.2 for additional detail on this procedure). The adjustment factors (or ratios) for the 0 to 4
 2    age group (the closest age group for which outputs are available to the age group of interest for
 3    this assessment) between NATA national-scale assessment Pb exposure concentrations and
 4    ambient air concentrations ranged from 0.37 to 0.42 for the Census tracts within the three case
 5    study areas. Use of these ratios for the 0 to 4 year old age group to represent the 0 to 7 year old
 6    age group modeled in the pilot contributes some uncertainty in the estimate of inhalation
 7    exposure concentrations.

 8         4.3.2.3  Outdoor Soil Concentrations
 9         Pb concentrations in outdoor soil where characterized for the three case studies using a
10    combination of modeling and empirical data. For the primary Pb smelter, empirical data were
11    used to estimate concentrations in the remediation zone closer to the facility, while statistical
12    extrapolation based on measurement data was used for the remainder of the study area. For the
13    secondary Pb  smelter case study, fate and transport modeling was used to  predict soil
14    concentrations. In addition, as mentioned earlier, a second hybrid - model plus  empirical -
15    approach was used to estimate soil concentrations based on measurement data from a surrogate
16    location. Soil concentrations for the near roadway (urban) case study are based on a combination
17    of empirical data obtained from a surrogate location and a spatial gradient decay function
18    obtained from the literature.
19         4.3.2.3.1  Primary Pb Smelter Case Study
20          Over the past 10 years, many of the residential yards closer to the primary Pb  smelter
21    have undergone remediation involving the removal of contaminated  soil and replacement using
22    "clean" soil. Extensive soil sampling has been conducted to support this remediation effort
23    including the collection of pre- and post-remediation Pb measurements. In addition, soil
24    measurements have been collected for locations outside of the remediation zone, which extends
25    out to about 1.5 km from the facility.
26          Characterization of soil Pb concentrations for this case study uses a combination of
27    measurement data (for blocks within the remediation zone) and statistically-based predictions
28    beyond the remediation zone. Soil levels within the remediation zone are based on the most
29    recent post-remediation measurements available for a given block (i.e., pre-remediation soil
30    levels are not used in estimating soil levels within the remediation zone).  This reflects the fact
31    that extensive remediation has occurred within the remediation zone and therefore, the latest
32    measurements from remediated yards are assumed representative of current conditions. Note,
33    that studies have shown recontamination of remediated yards (USEPA, 2006e) and consequently,
34    the remediation zone should be viewed as a dynamic zone with changing soil Pb levels.

              December 2006                4-29           Draft - Do Not Quote or Cite

-------
 1          Characterization of soil levels for blocks and block groups beyond the remediation zone
 2    are based on a regression model predicting soil Pb as a function of distance from the facility,
 3    which was fitted to pre-remediation soil measurement data (available closer to the facility) (Risk
 4    Assessment Report Section 4.1.3). The use of pre-remediation soil data in deriving the regression
 5    equation reflects the fact that little remediation has occurred in these more distant locations and
 6    consequently, spatial trends seen in the pre-remediation soil levels are more likely to be
 7    representative for these outer portions of the study area. The regression model  used in these
 8    estimates has an R2 of 0.92 which suggests a good fit and increases overall confidence in these
 9    statistical estimates. However, it should be noted that this increased confidence holds for areas of
10    interpolation (i.e., areas with sampling data used to fit the model - out to about 2.3 km from the
11    facility) more than for areas of extrapolation (areas without sampling data - beyond 2.3 km from
12    the facility).
13          Because sampling data were used either in establishing soil concentrations (within the
14    remediation zone) or in fitting the regression model (beyond the remediation zone), no
15    performance evaluation was conducted for this step in the analysis. Note, however, that future
16    efforts may consider performance evaluation based on such approaches as split-set validation
17    (i.e., fitting the regression model with half of the measurement data and evaluating the
18    performance of that model using the other half of the data).
19         4.3.2.3.2 Secondary Pb Smelter Case Study
20          As noted in Section 4.2.3.2, there are no soil sampling data for this case study,
21    necessitating the use of either (a) surrogate soil data (from a similar type of facility and study
22    area) or (b) fate and transport modeling to predict soil Pb levels. Surrogate soil data sufficient to
23    provide coverage for the entire study  area were not identified and consequently, fate and
24    transport modeling was employed.  Specifically, outdoor soil concentrations were calculated at
25    each block analyzed for the secondary Pb smelter case study using the AERMOD deposition
26    estimates and EPA's Multiple Pathways of Exposure (MPE) methodology (USEPA, 1998). The
27    MPE methodology represents the update of the Indirect Exposure Methodology,  or IEM
28    (USEPA 1990b; USEPA 1993).  MPE consists of a set of multimedia fate and  transport
29    algorithms developed by EPA's Office of Research and Development,  including a soil mixing
30    model which was used in this assessment to calculate the soil concentrations resulting from
31    deposition at the Troy site. The Human Health Risk Assessment Protocol (HHRAP), which
32    utilizes the same soil mixing algorithm, includes a database of input parameters which was used
33    to parameterize the equation for this assessment (USEPA, 2006d).
34          In the MPE/HHRAP algorithms, cumulative soil concentration  is calculated as a function
35    of particle deposition, soil mixing depth, bulk density, and a soil loss constant. The soil loss

              December 2006                 4-30           Draft - Do Not Quote or Cite

-------
 1    constant (in this case) was set up to be a function of loss due to leaching, erosion, and runoff.
 2    Concentration in the soil was calculated assuming constant deposition of Pb for the entire
 3    operating period of the facility (37 years).
 4           A background soil Pb concentration of 15 mg/kg (based on Gustavsson et al., 2001) was
 5    added to all modeled concentrations to produce a "total" Pb soil estimate.12  In this context,
 6    "background" refers to Pb in soil resulting from sources other than this particular secondary Pb
 7    smelter. Note, in presenting both exposure and risk results for this case study, the background
 8    soil Pb concentration described here was treated as part of background source exposure with the
 9    other contribution to soil Pb (from the facility itself) contributing to policy-relevant exposures.13
10           Modeled soil concentrations for the  secondary Pb smelter were compared to empirical
11    data obtained from a surrogate location (see Table 4-11).  Based on this comparison, which
12    suggested that modeled soil Pb concentrations for this case study might be significantly
13    underestimated, we included a second characterization of soil concentrations besides the purely
14    modeled approach described above. Specifically, measurement data from a surrogate secondary
15    Pb smelter location were used to "scale" up the modeled surface generated for this case study
16    location to more closely match the empirical data obtained from the surrogate location (at
17    specified distances from the facility). This second hybrid approach to estimate soil
18    concentrations was included to address uncertainty in estimating soil Pb concentrations for this
19    case study (see Tables 4-10 and 4-18, respectively, for additional details on the comparison of
20    modeled results for this case study against the empirical data from the surrogate location and the
21    hybrid approach that resulted from that comparison).
             12 Note, this background value of 15 mg/kg represents natural soil Pb concentrations and general
      anthropogenic activity. An argument could be made for considering a background value that might more closely
      reflect an urbanized scenario with greater coverage for auto-emitted lead and other sources (e.g., Pb paint). We will
      consider this issue of background in relation to the secondary Pb smelter as we refine the risk assessment for the full
      scale analysis (see Section 4.5.2).
             13 Explicit consideration for background soil Pb levels for the secondary Pb smelter case study reflect the
      fact that soil Pb levels for this case study area modeled. Consequently, they need to include (a) contributions from
      the facility of interest (captured in fate and transport modeling described here) and (b) background (non-facility
      related) Pb in soil (covered by the background value of 15 mg/kg discussed here). Because the other two case
      studies considered in the pilot use empirical data to characterize soil Pb levels, the contribution of background soil
      Pb to total soil Pb is implicitly reflected in the measured values and there is no need to add an additional background
      value as is done for the secondary Pb smelter.
               December 2006                  4-31             Draft - Do Not Quote or Cite

-------
 1         4.3.2.3.3 Near Roadway (Urban) Case Study
 2          Soil measurement data were not available for the DRI monitoring location in Houston
 3    and consequently, surrogate near roadway soil data were identified and used (together with an
 4    empirically-based decay function from the literature) to characterize potential soil Pb gradients
 5    for the six bands evaluated in the analysis.
 6          Near roadway soil data were identified for a number of locations around the country (see
 7    Risk Assessment Report, Sections 3.3.4 for a discussion of these data). Ultimately, sampling data
 8    from Corpus Christi, TX (Turner and Maynard, 2003) were selected for use a surrogates for this
 9    case study, primarily because of their proximity and geographic similarity to the case study
10    location. Turner and Maynard collected samples near the entrance ramp to Interstate 37, which
11    has higher traffic flow compared with the road segment in Houston used for this case study. The
12    Corpus Christi data included measurements at 2, 3, and 12m from the road with values of 731,
13    766 and 214 mg/kg, respectively (all samples were taken between 0 and 5 centimeters from the
14    surface).
15          The relatively sharp decrease in soil Pb levels with distance from the road reflected in
16    these data (i.e., the drop from 700+ mg/kg adjacent to the road to 214 mg/kg at 12 m), is
17    supported by other studies. Southerland and Tolosa (2000) reported that there is a linear
18    relationship between the log of soil Pb concentrations and  the log of distance from the road,
19    suggesting a sharply decreasing soil concentration gradient. Similarly, Filipelli et al., (2005) and
20    Hafen and Brinkmann (1996) have reported an exponential decay in soil Pb concentrations with
21    increasing distance from the roadway. In addition, review of available near roadway data
22    suggests that concentrations drop off to near urban background within a distance of about 50 m
23    from the road. A reasonable background urban soil Pb  level (given available data in the
24    literature) is 100 mg/kg.  This reflects a number of studies (Lejano and Ericson, 2005, and
25    Chinereje et al., 2004) (see Risk Assessment Report, Section 4.3.3 for additional discussion).
26          The data described above were used to develop  representative (surrogate) soil Pb levels
27    for each of the three bands used in this case study.  Specifically, for the 0-12 m band, an
28    assumption of a log-log (linear) relationship between soil Pb and distance was assumed and a
29    consequently, a log-log model was fitted using two data points from the Corpus Christi  data (766
30    mg/kg at  3 m and the 214 mg/kg at 12 m). The mean value predicted across this distance using
31    the fitted  regression line (388 mg/kg) was used for the 0-12 m band.  For the 12-75 m band,  an
32    assumption of a linear decay (between the 214 mg/kg datapoint at 12 m and a background
33    concentration of 100 ug/kg at 75 m) was used. This reflects and assumption that the exponential
34    decay in soil Pb levels has been largely realized by the  12 m distance, with decay becoming
35    linear at that point. This produced an average value of 157 mg/kg for the  12-75 m zone. And

               December 2006                4-32           Draft - Do Not Quote or  Cite

-------
 1    finally, urban background (100 mg/kg) was used for the 75-200 m band (see Risk Assessment
 2    Report Section 4.3.3 for additional details on the derivation of these soil concentrations).

 3         4.3.2.4  Indoor Dust Concentrations
 4           Pb in indoor dust can originate from a variety of sources including (a) outdoor soil which
 5    is tracked into the house, (b) Pb in outdoor soil which is entrained and subsequently transported
 6    indoors (c) Pb released directly into outdoor air through ongoing anthropogenic activity (e.g.,
 7    industrial point emissions) which is transported indoors and (d) interior sources of Pb (e.g., paint,
 8    hobbies) (Adgate et al., 1998, Von Lindern, 2003). In the exposure assessment conducted for the
 9    last review, indoor dust Pb concentrations were predicted based on Pb concentrations in outdoor
10    soil and ambient air (USEPA,  1989). This is also the case for the default approach in the
11    exposure component of the IEUBK model  (USEPA, 1994).
12           The importance of outdoor soil relative to outdoor air in influencing indoor dust Pb levels
13    appears to depend on the nature of the Pb sources involved.  Investigations in urban areas and
14    contaminated waste sites with  elevated soil Pb levels without active air point source emissions of
15    Pb have indicated a greater dependency of dust Pb concentrations on soil Pb concentrations than
16    on ambient  air concentrations (e.g.,  Abgate, 1998 and Von Lindern, 2003).  By contrast,
17    investigations in areas with current point sources of Pb (e.g., active Pb smelters) have implicated
18    ambient air Pb as an important source of Pb to indoor dust (Hilts, 2003). Contributions of
19    ambient air Pb to indoor dust Pb levels have also been illustrated by a deposition study
20    conducted in New York City (Caravanos et al., 2006).  Caravanos and others described Pb
21    deposition indoors resulting primarily from exterior environmental sources and not from interior
22    Pb sources.
23           In light of these differences between areas with and without active Pb point sources, we
24    have relied  on different air, soil and dust Pb relationships for estimating Pb levels in indoor dust
25    at the three  case study locations.
26         4.3.2.4.1  Primary Pb Smelter Case Study
27           We used different regression models for predicting Pb concentrations in indoor dust in
28    areas where soil has been remediated (see description in Section 4.2.3.1 for details on soil
29    remediation) and where it has not.  For the remediation zone near the facility, a regression
30    equation was developed using  dust Pb measurement data which had been collected from some of
31    the houses within this area,  (these data, while sufficient for supporting development of a site-
32    specific regression model, did  not have sufficient coverage to be used alone to represent indoor
33    dust Pb levels for that portion of the study area). For the remainder of the study area, we
34    employed the regression equation developed for the last review. We decided not to use the site-
35    specific dust Pb model for the  entire study  area because the soil Pb concentrations in the
              December 2006                 4-33           Draft - Do Not Quote or Cite

-------
 1    remediation zone are significantly different from those in the remainder of the study area as a
 2    result of the remediation activity.
 3          The dataset used to develop the model for the remediation zone was based on indoor dust
 4    samples collected in 17 houses within the remediation zone. Independent variables included in
 5    the analysis were: (a) estimated annual average Pb concentrations in air at census block centroids
 6    located within 200 meters of each of the 17 houses, (b) road dust Pb measurements for locations
 7    within 300 meters of each house and (c) post-remediation residential soil Pb measurements for
 8    the yard of each house.  Pre-remediation soil Pb concentrations were not included in the
 9    regression analysis since they were not expected to represent current conditions at the site.
10    Multiple samples for each medium associated with a specific house within the dataset (e.g.,
11    reflecting multiple samples collected over time) were averaged to produce a "temporally-
12    averaged" value. A number of regression models were evaluated, (see Risk Assessment Report,
13    Section 4.1.4), and the H6 model was ultimately selected based on goodness of fit and other
14    considerations. This model relates the natural log of indoor house dust to the natural log of
15    ambient air Pb (r2=0.701):
16
17          ln(house dust, mg/kg or ppm) = 8.3884 + 0.73639*ln(air Pb, ug/m3)
18
19          Several points regarding the other variables considered for the remediation zone
20    regression are noted here.  For  example, road dust Pb concentration was not found to have
21    significant predictive power for indoor dust Pb.  This may reflect the fact that the road dust Pb
22    dataset does not provide significant coverage for homes located near to the truck haul routes.
23    Additionally, yard soil Pb concentration was found to be slightly, and statistically significantly,
24    negatively correlated with indoor dust Pb levels. This counter-intuitive finding may be a result of
25    the existence within the remediation zone of a patchwork of remediated yards, such that the
26    remediation activity may have interfered with any correlation between yard soil Pb levels,
27    ambient air Pb levels and indoor dust Pb levels that might have existed previously.  The resulting
28    slight negative correlation of dust Pb levels with soil Pb levels led us to exclude soil Pb from the
29    model.  The y-intercept for the selected model may reflect a number of factors not correlated
30    with ambient air or distance from the facility, such as a general level of soil Pb contamination in
31    the area and indoor Pb paint.
32          For areas beyond the remediation zone, a regression equation developed during the last
33    review from data collected during the 1970s  and 1980s at a number of operational primary Pb
34    smelters, including the smelter at Herculaneum (i.e., this case study location) was used (USEPA,
35    1989, Appendix B).  This model (referred to  as the "AGG"  or "aggregate" model) predicts indoor
36    dust Pb concentration from both outdoor soil and ambient air Pb concentrations.  We have
              December 2006                  4-34            Draft - Do Not Quote or Cite

-------
 1    selected the AGG model for the non-remediation portion of the primary Pb smelter case study
 2    area since this area has not been subjected to extensive remediation and is therefore likely to
 3    resemble the locations included in the pooled dataset used in deriving this model. The AGG
 4    model, selected for areas beyond the remediation zone is the following:
 5
 6                  GG pooled analysis model (air+soil version):
 7                  House dust (mg/kg or ppm) = 31.3 +  638*air Pb (ug/m3) + 0.364*soil Pb (mg/kg)
 8
 9         4.3.2.4.2 Secondary Pb Smelter Case Study
10           A version of the same "AGG" model (USEPA, 1989) used for the primary Pb smelter
11    was also used for the secondary Pb smelter case study. However, in the case of the secondary Pb
12    smelter, an "air-only" version of the model (USEPA, 1989) was employed reflecting the reduced
13    overall confidence associated with soil characterization at this case study (as noted above, soil
14    concentrations at the secondary Pb smelter case study were modeled, while empirical data were
15    available for characterizing soil at the primary Pb smelter).  The "AGG" model for estimating
16    indoor dust (USEPA, 1989) was derived in two forms  including an air-only model that based
17    indoor dust concentrations on outdoor ambient air Pb (without explicitly considering outdoor soil
18    Pb levels) and an air+soil model which based estimates on both outdoor soil and ambient air Pb
19    data. It is important to note, however, that the air-only model does reflect (implicitly) some
20    consideration for the air-to-soil-to-indoor dust mechanism in the air signal.  Specifically, the
21    larger air factor for the air-only model (relative to the air+plus dust model's air factor) reflects
22    contribution of air Pb both directly to dust through penetration indoors and subsequent
23    deposition to surfaces and indirectly to dust through deposition to outdoor soil which impacts
24    indoor dust. (USEPA, 1989).14
25
26                  AGG pooled analysis model (air-only version):
27                  House dust (mg/kg or ppm) = 60 + 844*air Pb (ug/m3)
28
             14 Note, that for the sensitivity analysis ran focusing on the characterization of soil Pb concentrations at this
      case study, the alternate AGG soil+air dust model was used, rather than the AGG air-only dust model used in the ran
      described above. The decision to use the AGG soil+air model for the sensitivity analysis reflects the desire to make
      sure that the sensitivity analysis considered the full impact of higher soil Pb concentrations around the facility,
      including their impact on indoor dust Pb levels (use of the AGG air-only model, would have meant that the
      increased soil Pb concentrations considered in this sensitivity analysis ran would have only impacted exposure
      through soil ingestion and not through their impact on indoor dust).
              December 2006                  4-3 5            Draft - Do Not Quote or Cite

-------
 1          The AGG model used for the secondary smelter was based on a number of studies
 2    focusing mainly on primary Pb smelters (a number of primary Pb smelters were operational at
 3    the time of model development). This does introduce additional uncertainty into indoor dust
 4    predictions generated for the secondary Pb smelter using this model since factors related to
 5    indoor dust loading (particle size profiles and nature of the entrained Pb compounds) might differ
 6    for primary versus secondary Pb smelters resulting in differing degrees of indoor dust loading
 7    from stack-emitted Pb.
 8         4.3.2.4.3  Near Roadway (Urban) Case Study
 9          The same version of the "AGG" model (soil+air regression model) (USEPA, 1989) used
10    for the primary Pb smelter was also used for the near roadway (urban) case study.
11
12                 AGG pooled analysis model (air+soil version):
13                 House dust (mg/kg or ppm) = 31.3 + 638*air Pb (ug/m3) + 0.364*soil Pb (mg/kg)
14
15          A number of considerations went into the decision to use a soil+air version of the AGG
16    regression model for the near roadway (urban) case study. First, measurement data for a
17    surrogate near roadway location is used to characterize soil Pb levels for bands within the study
18    area which increases the overall confidence in soil characterization relative to the use of fate and
19    transport modeling. Having increased confidence in the soil Pb levels supports use of a model
20    that explicitly considers soil in predicting indoor dust (i.e., includes a soil factor in estimating
21    indoor dust Pb). Second, long-term  historical loading of near roadway soils has produced
22    relatively elevated levels of soil Pb (certainly within the 50m zone adjacent to the road) which
23    might contribute significantly to indoor dust levels.  Even if entrainment is relatively low,
24    resulting in a smaller contribution of soil Pb to indoor dust loading through this mechanism,
25    other mechanisms (tracking of soil indoors) could provide a means for soil to impact indoor dust.
26    Use of an AGG model which explicitly considers  soil in predicting indoor dust will allow these
27    mechanism (related to  soil loading dust) to be considered as part of exposure and risk analysis.  It
28    is important to note, however, that the soil+air AGG model was developed primarily based on
29    data collected near primary Pb smelters. Therefore, its use in predicting indoor dust levels for
30    houses near roadways (in areas with little current industrial Pb emissions) does introduce
31    uncertainty into the analysis.

32         4.3.3  Methods for Estimating Blood Pb  Levels
33          This section presents the methodology used to estimate blood Pb levels in the child study
34    populations. The section begins with an overview of the two biokinetic models used in this
35    analysis (IEUBK and Leggett). Input parameters used in running both models are then described,
              December 2006                4-36            Draft - Do Not Quote or Cite

-------
 1    with emphasis on those parameters expected to either introduce significant uncertainty into
 2    modeled blood Pb levels and/or those parameters which required more complex methods to
 3    develop input values for. The probabilistic approach used to generate population-level
 4    distributions of blood Pb levels for each study population is then described. The section ends
 5    with a discussion of the GSD used to reflect inter-individual variability in behavior related to Pb
 6    exposure and Pb biokinetics (a key component in modeling population-level blood Pb).

 7         4.3.3.1  Blood Pb Models
 8           The modeling of blood Pb levels is required for the pilot analysis for a number of
 9    reasons: (a) measured blood Pb levels are only available for a small fraction of the study
10    population associated with the primary Pb smelter and are not available for either of the other
11    two case studies (necessitating the need to model blood Pb levels), (b) exposure, characterized
12    using blood Pb, needs to be apportioned between policy-relevant and background Pb exposures,
13    which necessitates modeling capable of parsing blood Pb resulting from different exposure
14    pathways and (c) potential changes in existing blood Pb level distributions need to be predicted
15    given reductions in ambient air Pb levels. As discussed in Section 4.4.1 of the CD, there are two
16    broad categories of blood Pb models available to support exposure and risk assessment:
17         •  Statistical (regression) models., which attempt to apportion variance in measured blood
18            Pb levels for a study population to a range of determinants or control variables (e.g.,
19            surface dust Pb concentrations, air Pb concentrations). The development of these
20            models requires paired predictor-outcome data which restricts these empirical models
21            to the domain of their observations (i.e., to applications involving the study
22            population(s) and exposure scenarios used in their derivation or at least to scenarios
23            very similar to the original study conditions) (Section 4.4.1, CD).
24         •  Mechanistic models, which attempt to model the process of transfer of Pb from the
25            environment to human tissues. While these models are considerably more complex
26            compared with the regression models (in terms of both the number of variables and
27            their computational structure), by incorporating variables that vary temporally and
28            spatially, or across individuals or populations, mechanistic models can be extrapolated
29            to a wide range of scenarios, including those outside of the original populations and
30            exposure scenarios used to develop/parameterize the models (Section 4.4.1, CD)
31
32           Given concerns over applying regression models to populations and exposure  scenarios
33    other than those used in their derivation, we decided to place emphasis on mechanistic models in
34    conducting the exposure analysis for the pilot, given their greater flexibility in application. Note,
35    however, that regression models have been included as part of the sensitivity analysis addressing
36    uncertainty in blood Pb modeling - see  Section 4.4.3.1.
              December 2006                 4-37            Draft - Do Not Quote or Cite

-------
 1          The CD (Section 4.4.1) highlights three mechanistic (biokinetic) models developed over
 2    the past several decades including IEUBK for modeling child Pb exposure and two models for
 3    simulating Pb biokinetics from birth through adulthood (Leggett, 1993 and O'Flaherty, 1993,
 4    1995, 1998). All three models have the potential for application in Pb risk assessment and have
 5    been evaluated to varying degrees using empirical datasets (CD, Section 8.3.4).
 6          For the pilot analysis, we used the IEUBK and Leggett models to generate child blood Pb
 7    distributions for all three case studies. Inclusion of the Leggett and IEUBK models (together
 8    with the regression-based blood Pb model mentioned above) represents an effort to consider
 9    uncertainty in modeling blood Pb levels and is reflected in the sensitivity analysis completed for
10    the pilot (see Section 4.4.3.1).  A brief overview of the IEUBK and Leggett models is presented
11    below (for discussion of the regression-based model used in the sensitivity analysis, see Section
12    4.3.3.1).
13         4.3.3.1.1  IEUBK
14          A multi-compartmental pharmacokinetics model for children 0-7 years of age, which
15    predicts average quasi-steady state blood Pb concentrations corresponding to daily average
16    exposures, averaged over periods of a year of more. The exposure submodel provides average
17    daily intakes of Pb (averaged over a 1 year time increment) for inhalation (air, including
18    consideration for both outdoor and indoor) and ingestion (soil, indoor dust, diet and water)
19    (Section 4.4.5.1  of the CD). The model is intended to be applied to groups of children
20    experiencing similar levels of Pb exposure and will generate a representative central tendency
21    blood Pb  estimate for that group. Consideration for inter-individual variability in biokinetics and
22    behavior (e.g., varying rates of dietary Pb ingestion) is typically accomplished through the
23    incorporation of a GSD which, together with the lEUBK-generated average blood Pb level, can
24    be used to characterize the distribution of blood Pb levels for a group of modeled children.
25    Additional detail on the IEUBK model can be found in Section 4.4.5 of the CD.
26         4.3.3.1.2  Leggett
27           Originally developed from a model designed to simulate radiation doses  for bone-
28    seeking radionuclides, this biokinetic model has a temporal resolution of one day and can model
29    exposure  from infancy through adulthood. Note, that the day-level resolution in  Leggett does
30    allow more comprehensive treatment of the temporal pattern of exposure and its  shorter-term
31    impact on blood Pb levels than IEUBK, although for this analysis, which focuses on longer-term
32    trends in Pb exposure, this functionality is not that relevant.  The model does not include a
33    detailed pathway-level exposure submodel as does IEUBK, instead taking (as inputs) total
34    ingestion  and inhalation exposure. However, it is possible to link the Leggett model to a more
35    detailed pathway-level exposure model, thereby allowing a more detailed treatment of Pb
              December 2006                 4-3 8           Draft - Do Not Quote or Cite

-------
 1    exposure pathways and their impact on blood Pb. The use of this type of external exposure
 2    model including pathway-specific modeling of exposure levels was implemented for the pilot.
 3    As with IEUBK, Leggett is used to derive central tendency blood Pb levels for groups of
 4    similarly exposed children. The same GSD used for IEUBK is then used to produce estimates of
 5    the distribution of blood Pb levels within study populations. For additional details on the Leggett
 6    model see Section 4.4.6 of the CD.
 7           As noted above, both models (IEUBK and Leggett) were used in the pilot analysis
 8    essentially in unmodified form except for inclusion of an external exposure model for Leggett as
 9    mentioned above. Note, however, that a number of the input parameters for both models have
10    been adjusted to reflect the latest data on behavior, biokinetics and Pb exposure (key input
11    parameters for both models are discussed in the next section) (see Risk Assessment Report
12    Section 5.1 for addition detail on blood Pb modeling completed for the pilot).

13         4.3.3.2  Model Inputs
14           Both the IEUBK and Leggett models require the specification of a range of input
15    parameters addressing such factors as inhalation rates, inhalation exposure concentrations,
16    dietary consumption rates, incidental ingestion rates for soil and dust and route-specific
17    absorption factors.  In addition, as noted above, characterization of blood Pb levels using both
18    models include the application of a GSD reflecting inter-individual variability in both exposure
19    levels and biokinetic factors.
20           This section highlights a subset of the factors used in biokinetic blood Pb modeling for
21    the pilot, focusing on those factors (a) whose derivation involved a relatively complex analytical
22    process, thereby warranting discussion and/or (b) are subject to potentially significant
23    uncertainty resulting in their inclusion in the sensitivity analysis.
24         4.3.3.2.1  IEUBK Input Parameters
25           Exposure modeling completed for the three case study locations has generally identified
26    background exposure (diet and drinking water), incidental soil ingestion and incidental dust
27    ingestion as the pathways contributing most to total blood Pb in children (see Section 4.4.2 for
28    detailed pathway-specific results). Table 4-3 presents input parameters for IEUBK related to the
29    modeling of these key pathways (including their values and their basis for derivation).  Full
30    documentation of input parameters used for IEUBK modeling in the pilot are presented in
31    Section 5.1.4 of the Risk Assessment Report.
              December 2006                 4-39            Draft - Do Not Quote or Cite

-------
1    Table 4-3.  IEUBK input parameters and basis or derivation.
2
3
4
Parameter
Parameter
value*
Basis/Derivation
INGESTION - DRINKING WATER
Water
consumption
(L/day)
Water Pb
concentration
(ug/L)
Absolute
absorption
(unitless)
0.34,0.31,
0.31,0.33,
0.36,0.39,
0.42
4.61
0.5 (single
value used
across all age
ranges)
(USEPA, 2002a) Based on value for infants, 1-3 yr olds and value for 1-
10 yr olds (with trend lines used to interpolate intermediate age ranges).
Geometric mean of values reported in studies of U.S. and Canadian
populations (residential water) (CD, Section 3.3 Table 3-10).
Assumed similar to dietary absorption (see "Total percent accessible"
under Ingestion-Diet below)
INGESTION - DIET
Dietary Pb
intake
(ug/day)
Total fraction
accessible
(unitless)
3.16,2.60,
2.87, 2.74,
2.61,2.74,
2.99
0.5
Estimates based on (a) Pb food residue data from U.S. Food and Drug
Administration Total Diet Study (USFDA, 2001), (b) food consumption
data from NHANES III (USCDC, 1997), and (c) dietary consumption
rates (defaults) used in the IEUBK model (USEPA, 1994). See Website
for details on derivation and data used (Superfund recommendations -
web site : http ://www . epa.gov/superfund/lead/ieubkfaq . htm#input) .
Alexander et al., 1973 and Ziegler et al., 1978, as cited in the CD
(Section 4.2. 1). These two dietary balance studies suggest that 40-50%
of ingested Pb is absorbed by children (2wks to Syears of age).
INCIDENTAL INGESTION - SOIL and DUST
Soil/dust
weighting
factor
(unitless)
Total dust +
soil ingestion
(mg/day)
Total fraction
accessible
(soil and dust)
(unitless)
45
85, 135, 135,
135, 100, 90,
85
- primary Pb
smelter case
study: 0.48 for
soil and 0.26
for dust
- secondary Pb
smelter and
near roadway:
0.30 (for both
soil and dust)
This is the percent of total ingestion that is soil. Value reflects best
judgment and consideration for Clausing (Clausing, et al., 1987, as cited
in USEPA, 1989). The Clausing study looked at tracer studies of
ingestion rates for rainy days and non-rainy days, and assumed rainy
was all soil ingestion and non-rainy days was a combination of soil and
dust with the delta representing soil.
USEPA 1989, based on multiple studies focusing on children
- Site specific absorption factors for soil and indoor dust were derived
for the primary smelter case study using relative bioavailability (RBA)
estimates generated based on swine studies involving soil and dust
samples collected in the study area (Casteel, 2005). These RBAs were
converted to absolute bioavailability factors (i.e., total percent accessible
values) by applying the absolute bioavailability factor for the control
material (Pb acetate water solution also fed to the animals).
- secondary Pb smelter and near roadway values: (USEPA, 1989)
reflects evidence that Pb in dust and soil is as accessible as dietary Pb
and that dust/soil ingestion may occur away from mealtimes (resulting
in enhanced absorption relative to exposure during meal events).
OTHER
Maternal
blood Pb
(ng/dL)
1.94
NHANES IV (national geometric mean for adult women - all
nationalities) (Madeloni, 2005)
* When appropriate (i.e., when age-differentiation is required to capture variability and/or when sufficient data exist
to support age differentiation), values are presented for 0-1, 1-2, 2-3, 3-4, 4-5, 5-6 and 6-7 year olds. Otherwise,
when a single value is provided, it was not age-differentiated and was used for all age groups in IEUBK modeling.
               December 2006
                                             4-40
Draft - Do Not Quote or Cite

-------
 1         4.3.3.2.2  Leggett Input Parameters
 2           As noted earlier, the Leggett model does not include a detailed exposure module and
 3    instead accepts daily intake rates for the inhalation and ingestion routes. For the pilot analysis,
 4    the Leggett model has been linked to an external exposure module that allows us to model the
 5    contribution of specific pathways (e.g., dietary ingestion and indoor dust ingestion) to total
 6    ingestion and inhalation intake. Input parameters used both in the external exposure module and
 7    in the Leggett model proper, have been selected to match those used in the IEUBK model to the
 8    extent possible. Specifically, input parameters have been specified to insure as close a match as
 9    possible between the route-specific Pb uptake rates used in Leggett and those used in IEUBK.
10    This reflects a desire that any differences in the performance between IEUBK and Leggett stem
11    from fundamental differences in the way the two models treat the distribution and disposition of
12    Pb within the body and not from  differences in the Pb uptake rates provided to the two models.
13    Because of the similarities in the input parameters used in IEUBK and Leggett, the parameters
14    for Leggett will not be presented here. The reader is referred to Section 5.1.3.3 of the Risk
15    Assessment Report for additional details on the input parameters used in the Leggett modeling.

16         4.3.3.3  Probabilistic Population Blood Pb Modeling Procedure
17           This section provides an overview of the probabilistic modeling used to generate
18    distributions of blood Pb levels for children associated with each of the case study locations. As
19    discussed in Section 3.3.1, recent epidemiological studies have identified the concurrent and
20    lifetime average blood Pb metrics as most strongly correlated with IQ loss in children.  Therefore,
21    these two metrics have been used in generating IQ loss estimates for the pilot analysis and
22    consequently, exposure modeling conducted for the pilot is designed to characterize the
23    distribution of both concurrent and lifetime average blood Pb levels for study populations.
24           The goal  of this probabilistic exposure modeling is to generate population-level
25    distributions of blood Pb levels that allow (a) specific percentiles of exposure (e.g., 50th, 90th,
26    99th and  mean) within a study population to be identified and (b) allow the total blood Pb levels
27    associated with a given percentile to be further differentiated by exposure pathway  (e.g.,
28    background versus policy-relevant with the latter further differentiated as to ambient air
29    inhalation, indoor dust ingestion  and outdoor soil ingestion).15 Therefore, for example, we might
             15 As noted earlier, the modeling approach used for the pilot does not allow exposures resulting from the
      ingestion of soil and indoor dust to be further differentiated reflecting the contribution from older historically
      deposited lead and lead that has been released to the air more recently. Consequently, policy-relevant exposures
      (including the ingestion of outdoor soil and indoor dust) reflect the combined impact of older historical lead
      emissions and more recent emissions.
               December 2006                  4-41            Draft - Do Not Quote or Cite

-------
 1    have an estimate of exposure for the 99th percentile child at the primary Pb smelter, with that
 2    blood Pb level further differentiated as to the fraction coming from total background (diet and
 3    drinking water) and policy-relevant pathways including ambient air, indoor dust, and outdoor
 4    soil.
 5           The probabilistic exposure modeling relied on information in three areas as summarized
 6    below:

 7          •   Central-tendency blood Pb levels for each exposure zone: biokinetic blood Pb
 8             modeling described in the last section produces central-tendency blood Pb levels
 9             (concurrent and lifetime average) for each exposure zone in each case study area.
10             These blood Pb levels represent the "average" blood Pb levels projected for children
11             residing in each exposure zone. However, in reality, blood Pb levels are distributed
12             across the children within each zone, with that distribution being based around this
13             central-tendency estimate.  The variability in blood Pb levels is addressed through the
14             use of the GSD discussed in the last section (see third bullet below).

15          •   Demographics (childdistribution within study areas): The distribution of 0-7 year old
16             children within each case study area (represented as child counts within each exposure
17             zone) is used to insure that the generation of population-level blood Pb level
18             distributions for each case study reflects where children are  located.
19          •   GSD reflecting inter-individual variability in blood Pb levels: As discussed earlier, a
20             GSD is used to reflect inter-individual variability for blood Pb levels in groups of
21             similarly-exposed children. For the pilot, the GSD is combined with the central -
22             tendency blood Pb level  estimates described in the first bullet to generate a distribution
23             of blood Pb levels for the group of children located in each exposure zone.  The same
24             GSD (i.e.,  1.6) is used for all three case studies (see section  4.3.3.4 for additional
25             discussion of the  GSD).
26           The step-wise procedure used to generate  population-level blood Pb distributions for each
27    case study is illustrated in Figure 4-3 (Note, that the information described in the bullets above is
28    referenced in Figure 4-3  as input data used in population-level exposure modeling).
29           Several points related to the modeling approach  presented in Figure 4-3 need to be
30    highlighted. For the pilot analysis, 10,000 simulated individuals were generated for each case
31    study location, in order to insure that the population-level blood Pb distributions generated met
32    target stability goals (see Risk Assessment Report, Section 5.2.2).16 For all three case studies,
             16 An analysis of stability in the probabilistic modeling used for the pilot, showed that simulation runs with
      10,000 realizations achieved reasonable stability goals (i.e., average run-to-run variability of <10% for percentiles
      up to the 99.5th, with the 99.9th having average run-to-run variability <15%) (see Risk assessment Report Section
      5.2.2 for additional detail).

               December 2006                  4-42            Draft - Do Not Quote or Cite

-------
1     Figure 4-3.   Procedure for Generating Population Blood Pb Distributions.
              Input distributions and datasets used in the Monte Carlo-based population-exposure modeling:

                       • Central-tendency blood Pb level dataset: Blood Pb modeling (described in Section
                       4.3.3.1) generates a central-tendency blood Pb level for each exposure zone (e.g., block, or
                       blockgroup) within each study area (results dimensioned on blood metric and blood Pb
                       model).

                       • Demographic dataset: the child count (US Census 2000 children <7 yrs of age) for each
                       exposure zone is used to support population-weighted sampling of exposure zones.

                       • Adjustment factor distribution representing inter-individual variability in behavior
                       and biokinetics related to Pb exposure: The GSD of 1.6 has been centered on a median
                       value of 1.0 to produce a distribution that can be used to obtain adjustment factors reflecting
                       the behavior and biokinetics (related to Pb exposure) for a single simulated child.


              Monte Carlo-Based Population Blood Pb Level Modeling Procedure:
                   Step 1 - select a central-tendency blood Pb level for a specific
                   exposure zone based on population-weighted random sampling:
                   The central-tendency blood Pb level dataset is combined with the
                   demographic dataset to conduct population-weighted sampling of a
                   central-tendency blood Pb level for a specific exposure zone (i.e.,
                   sampling proportional to child count within each exposure zone is
                   used to select a specific zone and the central-tendency blood Pb level
                   for that zone is chosen as the output form Step 1).
                   Step 2 - select an adjustment factor reflecting inter-individual
                   variability in behavior and biokinetics related to Pb exposure for
                   a single simulated child: A value is sampled randomly from the
                   adjustment factor distribution described above. This single value
                   represents the behavior and biokinetics (related to Pb exposure) for a
                   single simulated child.
                   Step 3 - generate a blood Pb level for a single simulated
                   individual: The central-tendency blood Pb level selected in Step 1 is
                   multiplied by the adjustment factor in Step 2 to produce a blood Pb
                   level for a single simulated child within the exposure zone selected in
                   Step 1.
                   Step 4 - place simulated child blood Pb level (output of Step 3) in
                   pool of modeled blood Pb levels for that study area: This pool of
                   simulated child blood Pb levels represents the distribution of blood Pb
                   levels across the study area and reflects (a) the demographic
                   distribution of children across that study area (and their relation to Pb
                   levels in contact media) and (b) inter-individual variability in behavior
                   and biokinetics related to Pb exposure.
                                                I
                                Repeat procedure
                                10,000 times to
                                generate set of 10,000
                                simulated individuals
                                (Note, this set can be
                                weighted down to
                                reflect the actual
                                population of children
                                (0-7 years) within
                                each study area in
                                generating
                                population-count
                                related risk metrics).
                 December 2006
4-43
Draft - Do Not Quote or Cite

-------
 1    this simulation count obviously represents a higher total child count than actually is associated
 2    with the study area. Using a higher number of simulated individuals was necessary to generate
 3    blood Pb distributions with "stable" higher-end exposure estimates (if simulations matching the
 4    actual population count at each case study had been conducted, the distributions that would have
 5    resulted would have been "unstable" at higher percentiles). It is important to note, however, that
 6    in presenting population counts associated with individual percentiles (e.g., the number of
 7    children associated with a given population percentile), the population counts have been scaled
 8    to reflect the actual child counts associated with the study areas.
 9           A further point needs to be clarified regarding the differentiation of specific percentile
10    total blood Pb levels into pathway-specific fractions.  All simulated individuals associated with a
11    given exposure zone, were assigned the same pathway-specific apportionment, reflecting
12    biokinetic modeling conducted for each zone. The simulation described in Figure 4-3 involves
13    generating a set of simulated individual for each exposure zone by considering (a) the central
14    tendency blood Pb level generated using biokinetic modeling for that zone and (b) the GSD of
15    1.6 reflecting blood Pb variability. While this approach will produce a set of simulated
16    individuals with a range of total blood Pb levels, it is assumed that all of them have the  same
17    pathway-specific apportionment of those blood Pb levels (i.e., the same apportionment generated
18    for the  central-tendency blood Pb level modeled for that exposure zone). In reality, it is likely
19    that pathway apportionment would vary across children with different blood Pb levels located in
20    the same exposure zone (e.g., the  contribution of indoor dust exposure to total blood Pb would
21    differ for kids living near each other who demonstrate different total blood Pb levels). However,
22    the modeling approach used in the pilot does not account for this level of resolution in pathway
23    apportionment. Note, however, that pathway apportionment does differ across exposure zones
24    (i.e., each exposure zone has a different pattern of pathway apportionment for its simulated
25    children).
26           The modeling approach presented in Figure 4-3 and described above, generates
27    population-level distribution of total blood Pb levels with further pathway apportionment as
28    caveated above. These distributions can be used to generate several types of exposure metrics
29    including:
30         •  Population-weighted exposure percentiles: total blood Pb levels (with pathway
31            apportionment) for simulated individuals  representing specific points along the
32            population blood Pb distribution (e.g., 50, 90, 95, 99 and 99.5th percentile).
33         •  Incidence counts: number of children within a given study area projected to experience
34            a specific degree of Pb exposure (total blood Pb level).
35
              December 2006                 4-44            Draft - Do Not Quote or Cite

-------
 1         4.3.3.4  GSD for Population Blood Pb Levels
 2          Regarding the GSD used to reflect inter-individual variability in Pb exposure and
 3    biokinetics (i.e., the basis for the adjustment factor distribution described above and in Figure 4-
 4    3), a value of 1.6 has been used for the pilot analysis. This value reflects the distribution of blood
 5    Pb levels measured in children exposed to smelter emissions at the Midvale UT smelter (White
 6    et al., 1998). This value represents a reasonable central-tendency GSD for populations of
 7    children living in relatively small, defined areas where the sources and relative importance of
 8    different exposure pathways and media are similar across the exposed population. All three case
 9    study locations considered for the pilot include exposed child populations where the source of
10    interest (i.e., primary Pb smelter, secondary Pb smelter, or near roadway deposited Pb) is
11    expected to contribute significantly  to overall exposure, relative  to background (at least for areas
12    closer to the policy-relevant source  of interest). In this context, using a GSD reflecting the
13    distribution of blood Pb levels for children residing near a primary Pb smelter (i.e., a dominant
14    Pb source such as the Midvale UT smelter) seems appropriate in modeling the three case study
15    locations. It is also worth noting that the pre-remediation blood Pb GSD for children
16    participating in the Baltimore Urban Pb Soil Abatement Project was estimated at 1.5 (White et
17    al., 1998), which adds support to using the GSD of 1.6 for the near roadway (urban) case study,
18    which can be interpreted as being somewhat similar to a general urban scenario in terms of Pb
19    exposure.
20          Recent surveys of blood Pb levels in children at the national-level (NHANES IV data for
21    years 1999-2002), have found GSD values in the range of 2.03 to 2.23 (Hattis, 2005). These
22    GSDs, which are considerably larger than the values used in the  pilot, likely reflect the fact that,
23    while blood Pb levels for the majority of children in the U.S. have decreased significantly over
24    the last 1-2 decades, a small fraction of children still retains relatively elevated blood Pb levels
25    due to continued exposure to Pb paint and other artifact sources. Consequently,  as the median
26    and mean blood Pb levels have dropped, the extreme upper tail of the distribution is still
27    somewhat anchored by these high-exposure children, resulting in an increased GSD for the
28    overall population. However, because the three case studies modeled for the pilot reflect
29    exposure scenarios with dominant Pb sources (e.g., smelters or the more heavily contaminated
30    near roadway bands), it is believed that a smaller GSD (i.e., 1.6) is more appropriate for these
31    case studies (again, at least for portions of the study areas nearer to the sources of interest). Note,
32    however, that alternate GSDs have been considered in the sensitivity analysis (see Section
33    4.4.3.1).
              December 2006                 4-45            Draft - Do Not Quote or Cite

-------
 1         4.3.4  Projected Media Concentrations
 2           This section presents summaries of the media concentrations generated by the methods
 3    described in Section 4.3.2 for the three case study locations.  The complete set of media
 4    concentration estimates for each of the case studies is included in the Risk Assessment Report
 5    (Section 4.0). Table 4-4 presents summaries of projected annual average air concentration results
 6    along with inhalation exposure concentration results for the three case study locations. Table 4-5
 7    and Table 4-6 summarize the projected outdoor Pb  soil concentrations and indoor dust Pb
 8    concentrations, respectively, for the three study areas.
 9           Several factors should be noted in reviewing these summarized results. Both the ambient
10    air concentration and indoor dust concentration results are differentiated as to air quality scenario
11    (i.e., "current conditions" and "current NAAQS attainment"), since these differ for the primary
12    Pb smelter (note, that the indoor dust Pb concentrations, because they are based on air
13    concentrations, as well  as soil, will show a reduction under the "current NAAQS attainment"
14    scenario).  By contrast, summarized results for outdoor soil Pb are not differentiated by air
15    quality scenario, since for the pilot, potential changes in outdoor soil associated with attaining
16    the current NAAQS at the primary Pb smelter case study were not modeled.17
             17 Given that the reduction in air concentration associated with the current NAAQS attainment scenario is
      small, little reduction in soil Pb concentration is anticipated.
              December 2006                  4-46            Draft - Do Not Quote or Cite

-------
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
14
15
16
      Table 4-4.   Projected ambient air and inhalation exposure concentrations.

Statistic b
Current Conditions
Average
annual Pb air
concentration
(ug/m3)
Inhalation exposure
concentrations
(ug/m3)
Current NAAQS Attainment b'°
Average annual
Pbair
concentration
(ug/m3)
Inhalation exposure
concentrations
(ug/m3)
Primary Pb Smelter
Maximum
95th percentile
Median
5th percentile
Minimum
2.73
0.662
0.0221
0.00845
0.00541
1.14
0.277
0.00895
0.00329
0.00210
1.50
0.662
0.0221
0.00845
0.00541
0.628
0.277
0.00895
0.00329
0.00210
Secondary Pb Smelter
Maximum
95th percentile
Median
5th percentile
Minimum
0.536
0.0178
0.0047
0.0009
0.0005
0.238
0.008
0.0021
0.0004
0.0002
NA (study area projected to be in
attainment)
Near Roadway (Urban)
Maximum
95th percentile
Median
5th percentile
Minimum
0.008
0.008
0.005
0.005
0.005
0.0033
0.0033
0.0022
0.0022
0.0022
NA (study area projected to be in
attainment)
        a The 223 blocks and block groups with non-zero population selected for analysis were used to create this
summary. Note that in some of these blocks the 2000 U.S. Census indicates there are no children.
        b The 5th and 95th percentile values for the current conditions and current NAAQS attainment scenario (for
the primary Pb smelter case study) are identical because only two of 223 U.S. Census blocks within the study area
had modeled outdoor air Pb levels exceeding the NAAQS. This means that differences in the two air quality
scenarios are only evident at the extreme high-end of modeled media concentration distributions and associated
exposure levels.
        c Note that the "Average annual Pb air concentration" values presented here are the values used in
modeling exposure for the pilot analysis. As discussed in Section 4.3.2.1, consideration for the current NAAQS
attainment scenario required clipping of modeled quarterly average air concentrations values at the NAAQS level of
1.5 ug/m3, with subsequent recalculation of the annual-average air concentration values used in exposure modeling.
13    Table 4-5.   Projected outdoor soil concentrations.
Statistic
Maximum
95th percentile
Median
5th percentile
Minimum
Projected average soil concentration (mg/kg)
Primary Pb smelter
976
426
84.0
23.6
15.9
Secondary Pb smelter*
383-1,150
21.2-63.5
16.2-48.7
15.2-45.5
15-45.2
Near roadway (Urban)
388
157
100
100
100
        * Range reflects the two approaches used to characterize soil Pb levels for the secondary case study
including a purely modeling approach and a hybrid (model-empirical) approach, in which the modeled surface is
scaled up to match trends seen in soil Pb levels at a surrogate location (see Table 4-18).
                December 2006
                                              4-47
Draft - Do Not Quote or Cite

-------
 1    Table 4-6.  Projected indoor dust concentrations.
Statistic
Maximum
95th percentile
Median
5th percentile
Minimum
Projected average indoor dust concentration (mg/kg)
Primary Pb smelter
Current
Conditions
5,263
2,191
47
37
35
Current NAAQS
Attainment
3,335
2,191
47
37
35
Secondary Pb
Smelter*
133-142
74-75
63-64
61
60
Near Roadway
(Urban)
178
178
71
71
71
 2           * Range reflects dust Pb levels estimated as part of the sensitivity analysis examining uncertainty in
 3    outdoor soil Pb prediction for this case study location (see Table 4-18). That sensitivity analysis involved two
 4    different approaches to characterizing soil Pb levels. These different soil Pb estimates for the study area translated
 5    into different indoor dust Pb estimates reflected in this table. Note also, that the sensitivity analysis used two
 6    different approaches for predicting indoor dust Pb (i.e., the AGG air-only approach and the AGG air+soil approach -
 7    see Section 4.3.2.4). The fact that the lower percentiles do not have a range reflects the fact that outdoor soil has a
 8    very small effect on indoor dust Pb at greater distances from the facility (with indoor dust Pb at those distances
 9    driving primarily by the intercept term in the dust Pb models - see Section 4.3.2.4).
10

11          4.3.5  Projected Blood Pb Levels
12           This section presents summaries of blood Pb modeling completed for the three case study
13    locations in the form of population percentiles. These results are dimensioned as follows:

14          •  Air quality scenario: The primary Pb smelter had projected exceedances of the Pb
15             NAAQS, so projected air concentrations (and inhalation air concentrations) differ for
16             the two air quality scenarios, resulting in different modeled blood Pb levels.
17          •  Characterizing Pb concentrations in soil: Performance evaluation conducted for the
18             secondary Pb smelter focusing on soil Pb suggested that modeled soil Pb
19             concentrations for this case  study might be under-estimated by a factor of 3. This
20             resulted in the decision to include two scenarios for this case study in both the exposure
21             and risk assessment including:  (a) a model-only scenario where soil concentrations are
22             modeled and (b) a hybrid (model/empirical) approach where modeled estimates for this
23             study area are scaled up based on comparison to empirical soil Pb estimates obtained
24             for a surrogate secondary Pb smelter (see Tables 4-11 and 4-18). Consequently, two
25             sets of exposure estimates are presented for the secondary Pb  smelter.

26          •  Blood Pb model: Both the IEUBK and Leggett models were used to project blood Pb
27             levels for the case studies.

28          •  Blood Pb metric: Both concurrent and lifetime average blood  Pb levels were modeled
29             for the three case study locations.

30          •  Pathway apportionment. Each of the simulated individuals has had their projected total
31             blood Pb levels apportioned between contributing pathways. For the summary tables
32             presented in this section, pathway apportionment information is presented at the more
               December 2006
4-48
Draft - Do Not Quote or Cite

-------
 1            generalized level of (a) policy-relevant background (drinking water and diet) versus (b)
 2            policy-relevant exposures (air-inhalation, soil and dust ingestion). For more detailed
 3            pathway-specific breakdown of exposure results, refer to the risk results tables
 4            presented in Section 4.4.2.
 5
 6          Given the number of dimensions involved, results have been separated by case study,
 7    with Table 4-7 summarizing results for the primary Pb smelter,  Table 4-8 the secondary Pb
 8    smelter and Table 4-9 the near roadway (urban) case study.
 9          A number of observations can be made by reviewing the blood Pb results presented in
10    Tables 4-7 through 4-9. Generally, the IEUBK model generates higher blood Pb levels than the
11    Leggett model  (as seen when comparing population percentile blood Pb estimates across the two
12    models). Furthermore, the concurrent blood Pb levels are typically lower than the lifetime
13    average blood Pb levels; this is expected given that the lifetime  average values reflect
14    contributions from earlier years of exposure when Pb exposure is typically higher, while the
15    concurrent estimates represent modeled blood Pb at 7 years of age. When combined, these two
16    factors mean that the highest blood Pb levels are typically seen for the combination of the
17    IEUBK model  and the lifetime average blood Pb metric, while the lowest levels are seen with the
18    combination of the Leggett model and the concurrent blood Pb metric.
19          Speaking specifically to exposure estimates generated for individual case studies,
20    modeled blood Pb levels for the primary Pb smelter range from  0.2 to 28.6 ug/dL (1st percentile
21    to the  99.9th percentile simulated individual) for the current conditions exposure scenario and
22    from 0.2 to 24.9 ug/dL for the current NAAQS attainment scenario.  Central tendency blood Pb
23    levels  for this case study range from 0.7 to 1.9  ug/dL (50th percentile for both the current
24    conditions and current NAAQS attainment scenarios). The results for this case study demonstrate
25    a clear trend regarding pathway apportionment, with higher blood Pb levels reflecting a larger
26    contribution from policy-relevant sources relative to policy-relevant background. This is
27    expected, since simulated individual with higher blood Pb levels are generally located closer to
28    the facility where Pb concentrations in exposure media are higher, resulting in a higher
29    proportion of overall Pb exposure coming from facility-related Pb.
30          Speaking now to results generated for the secondary Pb  smelter case study, as noted
31    earlier, two sets of exposure estimates are presented for this case study, one reflecting a model-
32    only approach to estimating soil Pb concentrations and the other reflecting a hybrid (model+
33    empirical) approach. Projected blood Pb estimates for this case study range from 0.2 to 6.3
34    ug/dL (1st percentile to the 99.9th percentile simulated individual for the model-only approach
35    and the hybrid  approach). Unlike the primary Pb smelter case study, pathway apportionment is
36    fairly constant  across all of the population percentiles (i.e., there is no clear trend with higher

              December 2006                 4-49            Draft - Do Not Quote or Cite

-------
 1   percentiles being dominated by policy-relevant sources). This likely reflects the fact (discussed
 2   below in Section 4.4.3) that the highest impact areas near to the facility do not have any children
 3   (according to US Census data for 2000) and therefore, these portions of the study area that would
 4   have shown the highest gradients in exposure are not reflected in the population-weighted blood
 5   Pb distributions. Consequently, the blood Pb distribution for this case study is dominated by
 6   simulated individuals with fairly consistent patterns of exposure, in terms of pathway-
 7   apportionment (with policy-relevant sources generally contributing about 25-40% of total Pb
 8   exposure).
 9          The near roadway (urban) case study has modeled blood Pb levels ranging from 0.3 to
10   9.1 ug/dL (1st percentile to the 99.9th percentile simulated individual). As with the secondary Pb
11   smelter, pathway apportionment estimates  for this case study are fairly constant across the
12   population percentiles and show that typically, 60-65% of total Pb exposure comes from policy -
13   relevant sources. The absence of a clear trend in pathway apportionment likely reflects the fact
14   that this case study was modeled with a relatively small number of bands (3) extending out from
15   the modeled road segment which significantly reduces the opportunity for refined gradients in
16   exposure (i.e., exposure will be "clustered" into three subpopulations reflecting individuals
17   located within each of the three bands extending out from the road segment).
              December 2006                 4-50            Draft - Do Not Quote or Cite

-------
1    Table 4-7.  Projected blood Pb levels (ug/dL) for primary Pb smelter case study.
Statistic
IEUBK
(concurrent)
Blood
Pb
level
%from
Policy-
Relevant
Pathways*
IEUBK Leggett
(lifetime average) (concurrent)
Blood
Pb
level
%from
Policy- Blood
Relevant Pb
Pathways* level
% from
Policy-
Relevant
Pathways*
Leggett
(lifetime average)
Blood
Pb
level
% from
Policy-
Relevant
Pathways*
Current conditions exposure scenario
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
21.9
12.4
7.4
3.7
2.9
2.0
1.3
0.9
0.4
98%
96%
96%
78%
66%
55%
44%
44%
46%
28.6
16.9
10.6
5.3
4.1
2.7
1.8
1.2
0.5
98%
95%
95%
74%
71%
38%
75%
44%
46%
13.9
6.7
4.2
2.0
1.5
1.0
0.7
0.5
0.2
98%
95%
89%
44%
65%
72%
45%
65%
46%
22.9
11.1
6.8
3.1
2.3
1.6
1.1
0.7
0.3
98%
95%
83%
71%
78%
71%
52%
44%
46%
Current NAAQS attainment exposure scenario
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
18.4
11.2
7.9
3.7
2.9
2.0
1.4
0.9
0.4
97%
96%
87%
72%
60%
61%
71%
55%
38%
24.9
15.5
11.3
5.4
4.1
2.8
1.9
1.2
0.5
96%
95%
95%
91%
46%
71%
48%
44%
56%
12.7
6.4
4.3
2.0
1.5
1.0
0.7
0.5
0.2
96%
90%
95%
75%
78%
44%
46%
43%
41%
20.7
10.6
7.1
3.1
2.3
1.6
1.1
0.7
0.3
96%
90%
95%
65%
89%
55%
44%
38%
41%
2
O
4
5
6
7
* Policy-relevant pathways include inhalation, soil-ingestion and indoor dust ingestion (and exclude background
sources e.g., diet, drinking water).
              December 2006
                                          4-51
Draft - Do Not Quote or Cite

-------
1    Table 4-8.   Projected blood Pb levels (ug/dL) for secondary Pb smelter case study.
Statistic
IEUBK
(concurrent)
Blood
Pb
level
%from
Policy-
Relevant
Pathways*
IEUBK
(lifetime average)
Blood
Pb
level
%from
Policy-
Relevant
Pathways*
Leggett
(concurrent)
Blood
Pb
level
%from
Policy-
Relevant
Pathways
*
Leggett
(lifetime average)
Blood
Pb
level
% from
Policy-
Relevant
Pathways*
Model only approach for characterizing soil concentrations
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
Hybrid a
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
3.7
3.0
2.7
1.9
1.7
1.2
0.9
0.7
0.3
30%
30%
22%
30%
22%
22%
26%
23%
24%
4.7
3.9
3.5
2.5
2.1
1.6
1.1
0.8
0.4
22%
24%
27%
24%
24%
24%
22%
25%
26%
2.1
1.6
1.4
1.0
0.9
0.6
0.5
0.3
0.2
26%
30%
22%
31%
25%
24%
29%
26%
24%
3.0
2.3
1.9
1.4
1.2
0.9
0.7
0.5
0.2
22%
24%
26%
26%
23%
24%
23%
22%
23%
pproach (model + surrogate data) for characterizing soil concentrations
4.7
3.7
3.3
2.4
2.0
1.5
1.1
0.8
0.4
40%
38%
40%
38%
36%
38%
37%
39%
36%
6.3
4.9
4.4
3.2
2.7
2.0
1.5
1.1
0.5
38%
40%
37%
43%
37%
41%
38%
40%
36%
2.3
1.9
1.7
1.2
1.0
0.8
0.6
0.4
0.2
38%
39%
42%
38%
38%
38%
40%
36%
38%
3.3
2.7
2.5
1.8
1.5
1.1
0.8
0.6
0.3
37%
38%
40%
40%
39%
38%
39%
37%
40%
2
O
4
5
* Policy-relevant pathways include inhalation, soil-ingestion and indoor dust ingestion (and exclude background
sources e.g., diet, drinking water). Note, that for this case study, background sources also included a fraction of soil
Pb identified as background (see Section 4.3.2.3.2).
6    Table 4-9.   Projected blood Pb levels (ug/dL) for near roadway (urban) case study.
Statistic
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
IEUBK
(concurrent)
Blood
Pb
level
6.5
5.0
4.4
3.1
2.6
1.9
1.4
1.0
0.4
%from
Policy-
Relevant
Pathways
*
56%
56%
56%
56%
56%
56%
65%
65%
56%
IEUBK
(lifetime average)
Blood
Pb
level
9.1
7.0
6.2
4.4
3.6
2.7
1.9
1.4
0.6
%from
Policy-
Relevant
Pathways*
56%
56%
56%
56%
56%
56%
65%
65%
56%
Leggett
(concurrent)
Blood
Pb
level
4.2
3.1
2.7
1.9
1.6
1.2
0.8
0.6
0.3
% from
Policy-
Relevant
Pathways*
56%
56%
65%
56%
56%
56%
56%
56%
56%
Leggett
(lifetime average)
Blood
Pb
level
6.9
4.9
4.1
2.9
2.4
1.7
1.2
0.9
0.4
% from
Policy-
Relevant
Pathways*
80%
56%
56%
56%
65%
56%
65%
56%
56%
     * Policy-relevant pathways include inhalation, soil-ingestion and indoor dust ingestion (and exclude background
     sources e.g., diet, drinking water).
              December 2006
                                            4-52
Draft - Do Not Quote or Cite

-------
 1         4.3.6   Performance Evaluation
 2          This section describes performance evaluation completed in support of the pilot analysis
 3   (i.e., the comparison of modeled results to empirical data for purposes of assessing the
 4   representativeness of a particular modeling step). Performance evaluation for the exposure
 5   assessment focused on projections of Pb in ambient air and outdoor soil (discussed in Section
 6   4.3.2.1 and 4.3.2.3, respectively) and projections of Pb in blood (covered in Section 4.3.3). Those
 7   case studies for which media concentrations were estimated using empirical data as the basis,
 8   were not subjected to performance evaluation; only those estimates based directly on modeling
 9   were included.
10          Performance evaluation can provide insights into the degree of representativeness
11   associated with specific elements of exposure modeling by identifying systematic trends in either
12   over- or underestimation of modeled results relative to empirical data.

13         4.3.6.1  Media Concentrations
14          Table 4-10 describes the performance evaluation completed for modeled ambient air
15   concentrations and presents a summary of the results of that assessment for each of the case
16   studies.
              December 2006                  4-53            Draft - Do Not Quote or Cite

-------
1    Table 4-10. Performance evaluation of approaches for ambient air concentrations.
      Case study
               Description of
               performance
                evaluation
Results of the
performance
 evaluation
Implications for overall analysis
     Primary Pb
     smelter
                  Comparison of 2
                  yr-averaged
                  modeled air
                  concentrations
                  with annual
                  averaged Pb
                  measurements
                  (from 2001
                  through 2005)
                  from TSP
                  monitors located
                  within study area.
                               - Closest 2 monitors to
                               the facility (~300m) are
                               under-predicted by
                               modeling in 2001, but
                               over-predicted by
                               modeling (factor of 1.3
                               to  1.6). Overall trend is
                               over-prediction by
                               modeling for these
                               monitors.

                               - Remaining 7 TSP
                               monitors (800m to 3km
                               from facility) are under-
                               predicted by modeling
                               for all five years (0.05 to
                               0.6).	
                   Low/moderate overestimation for points
                   close to facility (for years 2002-2005) may
                   suggest similar degree of overestimation
                   for ambient air concentrations used in
                   inhalation modeling and indoor dust
                   prediction. Similarly, underestimations for
                   points further from facility (but still only
                   out to ~3km) may suggest moderate
                   underestimation of inhalation exposure.
                   Note, performance evaluation for this case
                   study is jeopardized somewhat by the fact
                   that meteorological data used in air
                   modeling is from 1997-1999, while
                   monitored data used in performance
                   evaluation are from later years (2001-
                   2005).
     Secondary
     Pb smelter
                  Comparison of 2
                  yr-averaged
                  modeled air
                  concentrations
                  with annual
                  averaged Pb
                  concentrations
                  from TSP
                  monitors (1999
                  and 2000) located
                  within study area.
                               2 monitors (400m and
                               650m from the facility)
                               are located in study area.
                               Modeled results
                               (geographically matched
                               to monitors) are
                               approximately three
                               times lower than the
                               monitored values. Note,
                               however that the
                               monitors are not down-
                               wind from the facility
                               and that the highest
                               modeled values (not
                               matched to the monitors)
                               are in the range of the
                               monitored values.
                   Although the results suggested a
                   significant underestimation of actual
                   measured ambient air Pb levels (based on
                   comparison at the monitors), because
                   monitored values fall within the range of
                   the highest modeled values predicted
                   across the study area (and because the
                   monitors are not located downwind from
                   the facility), concern over a significant
                   underestimation is reduced somewhat (see
                   Risk Assessment Report, Section 4.2.2.4
                   for additional discuss). Note, that the age
                   of the monitor data used in this
                   performance evaluation (1997-2000)
                   reduces the overall utility of the evaluation.
     Near
     Roadway
     (Urban)
             Performance evaluation was not completed for the near roadway study area since the
             characterization of ambient air Pb levels for this case study is based on empirical data
             (augmented with modeling characterizing spatial gradients in air pollution near roadways).
2
3
4
5
6
7
        Table 4-11 presents the results of the performance evaluation completed for modeled
outdoor soil concentrations. Note, that because the secondary Pb smelter is the only case study
to have its soil concentrations generated directly using modeling (the other two case studies
relying on direct,  or surrogate empirical data), only the secondary smelter was subjected to
performance evaluation of its outdoor soil modeling.
               December 2006
                                              4-54
                            Draft - Do Not Quote or Cite

-------
1   Table 4-11. Performance evaluation of approaches for outdoor soil concentrations.

Case
study
Primary
Pb
smelter
Secondary
Pb
smelter



Near
Roadway
(Urban)

Description of performance
evaluation
Results of the
performance
evaluation


Implications for overall analysis
Performance evaluation was not completed for the primary Pb smelter study area since soil Pb
concentrations were based directly on measured data (for the remediation zone closest to the
facility) and on statistically -extrapolated values for the remainder of the study area.
Modeled results for this case
study were compared to soil
concentration data collected near
a secondary Pb smelter
(Kimbrough et al., 1995). The
Kimbrough study presents soil
data collected around a secondary
Pb smelter located in an urban
area where there is the potential
for Pb impacts from multiple
sources. Note, Small et al., (1995)
also presents data collected near a
secondary Pb smelter in
Pennsylvania, but that study area
was subjected to significant
remediation of residential areas
near the source and consequently,
measurements may be
significantly biased down.

Modeled soil
concentrations are
approximately
three times lower
than measured
soil
concentrations
suggesting the
potential for a
significant
underestimation
of soil Pb levels
at this case study.

The potential for a significant
underestimation of soil Pb levels
translates into a potentially
significant impact (downward bias)
on exposure and risk results. This
finding led to an investigation of
this issue of soil modeling at the
secondary Pb smelter as part of the
sensitivity analysis. Specifically, we
developed an alternate soil
concentration coverage for the study
area by combining the modeled
results (used to characterize relative
spatial variability in soil levels
across the study area) with
surrogate data from Kimbrough (to
adjust the absolute soil Pb levels for
the study area), (see Table 4-16).
Performance evaluation was not completed for the near roadway study area since the
characterization of ambient air Pb levels for this case study is based on empirical data
(augmented with modeling characterizing spatial gradients in air pollution near roadways).
3         4.3.6.2  Blood Pb Levels
4          Performance evaluation for the blood Pb modeling involved comparing modeled blood
5   Pb levels for children (0-7 years of age) generated for the three case study locations against two
6   empirical data sets: (a) national-level central tendency blood Pb levels for children (0-5 years)
7   obtained through the NHANES IV survey (completed for all three locations) (CD, Table 4-1)
8   and (b) site-specific monitored blood Pb levels (available only for Herculaneum). Table 4-12
9   presents the results of performance evaluation completed for blood Pb modeling.
             December 2006
4-55
Draft - Do Not Quote or Cite

-------
1    Table 4-12. Performance evaluation of approaches for blood Pb levels.
     Case study
     Description of performance
             evaluation
   Results of the performance
          evaluation
             Implications for overall analysis
     All Case
     Study
     Populations
     (Part 1)
Comparison of modeled median blood
Pb levels for the three case study
populations against GM values
obtained from NHANES IV.

The purpose of this evaluation is to
determine whether there appears to be a
significant error in the characterization
of central tendency blood Pb levels
(e.g., whether modeled central tendency
levels are shown to be significantly
lower than corresponding national-
levels, which would suggest potential
under-estimate in modeling since we
would expect central tendency levels
within the study area to be equal to or
higher than national central-tendency
levels, given the presence of the
primary Pb  smelter and its impact on
exposure).
Modeled median lifetime blood
Pb modeled with IEUBK ranged
from 1.2 to 1.9 ug/dL (average
1.7); Median concurrent blood Pb
modeled with IEUBK ranged
from 0.9 to 1.4 ug/dL (average
1.2).

Leggett median lifetime blood Pb
ranged from 0.8 to 1.2 ug/dL
(average 1.0); Median concurrent
blood Pb ranged from 0.5 to 0.8
ug/dL (mean 0.7).

These compare with a GM value
from NHANES IV (for 0-5 yr
olds) of 1.7 (2001-2002) (CD,
Table 4-1).
The relatively close match between the modeled median
lifetime levels from IEUBK and the national GM value from
NHANES IV (for 0-5 yr olds) suggests that the IEUBK model
with the given set of exposure, intake, and uptake factors is
neither significantly over- or underestimating exposures for the
study population. Because "concurrent" blood Pb in this
analysis is defined as the  average blood Pb from age 6 to 7
years, when blood Pb levels are known to decline from values
seen in younger children, the median concurrent IEUBK values
appear to also generally consistent with population data. The
median blood Pb  estimates are insensitive to the relatively
small number of high-exposure block groups in the primary and
secondary Pb smelter case studies; thus the lower exposure
experienced by the large majority of the exposed populations
are dominating this metric

The Leggett model, however, with the selected exposure,
intake, and uptake factors, appears to be underestimating the
GM blood Pb statistics somewhat, compared to the national
population.  The reason for this is not clear, although it is
possible that the Pb exposure levels of the NHANES
population are actually lower than the combined background
and air-related levels used in the case studies.
                                  December 2006
                                                      4-56
                                  Draft - Do Not Quote or Cite

-------
Case study
Description of performance
        evaluation
Results of the performance
        evaluation
                                                                                      Implications for overall analysis
Primary Pb
smelter
(Part 2)
Comparison of upper-bound (extreme
high-end) range of modeled blood Pb
levels against the set of site-specific
measured blood Pb levels collected for
children <6 years of age in 2002
(U.S.DHHS, 2003).

The purpose of this comparison is to
compare the high-end of the modeled
blood Pb level distribution to the range
of empirical values obtained for this
segment of more highly exposed
children living close to the facility.
Ideally, the extreme high-end of the
modeled distribution (perhaps >99*
percentile) should be similar to the
high-end of the sampled population
(perhaps >90th percentile), reflecting the
fact that the measured data cover
highly-exposed children close to the
facility, while the modeled population
includes children further from the
facility who are less exposed.	
                                  58 site-specific blood Pb level
                                  measurements from 2001 yield
                                  the following percentiles:

                                  > 95th percentile is 20-29 ug/dL
                                  > 90th percentile is 10-19 ug/dL
                                  > 50th percentile is 0-9 ug/dL

                                  Here are percentile results
                                  generated from our site-specific
                                  modeling:

                                  > 99.9th percentile is 13-29 ug/dL
                                  > 99.5th percentile is 6-17 ug/dL
                                  > 99th percentile is 4-11 ug/dL
                                  > 95th percentile approaches
                                  5 ug/dL
                               These results show that the 58 screened blood Pb levels from
                               the DHHS study correspond to the extreme high-end of our
                               modeled distribution. This provides support for our modeled
                               results generating reasonable high-end estimates. It should be
                               noted that the top 5% of the screened children have higher
                               blood Pb levels (20-29 ug/dL) than the top 5% of our modeled
                               children (>5 ug/dL), but this is expected since the screening
                               analysis focuses on children located relatively close to the
                               facility,  while the children we modeled include children out to
                               10km from the facility (inclusion of children further out will
                               dilute the overall blood Pb distribution with children that live
                               further from the facility and are less exposed).
                              December 2006
                                                  4-57
                                 Draft - Do Not Quote or Cite

-------
 1         4.4   HEALTH RISK ASSESSMENT
 2           This section describes the approach used to characterize risk for the pilot assessment,
 3    including discussion of the modeling approach (4.4.1) and presentation of results (4.4.2). This
 4    section also includes the results of the sensitivity analysis (4.4.3).

 5         4.4.1   Method for Risk Characterization
 6           Risk characterization for the pilot analysis focuses on modeling IQ loss in children using
 7    a log-linear concentration-response function obtained from a pooled analysis of epidemiology
 8    studies (Lanphear et al., 2005). This concentration-response function is combined with the
 9    population-level blood Pb distributions generated for each case study  (see Section 4.3.3 above)
10    to produce a distribution of IQ loss estimates for each study population. It is also possible to
11    apportion IQ loss between different exposure pathways using the pathway-apportionment
12    information generated as part of the exposure analysis (see Section 4.3.3.3 above).
13           Three key  elements of the risk methodology used for the pilot are described in greater
14    detail below, including: (a) the IQ-loss concentration response function used in the analysis, (b)
15    the cut-points or policy-thresholds (representing specific exposure levels below which IQ loss
16    will not be estimated) and (c) the step-wise analytical procedure used to generate the IQ loss
17    (risk) distributions.

18         4.4.1.1  Concentration Response Function
19           As discussed in Section 3.3.1.2, log-linear concentration response functions for IQ  loss
20    (for the concurrent and lifetime average blood Pb metrics) obtained from a large pooled study
21    (Lanphear et al., 2005) were used in this analysis. Specifically, these functions were used to
22    estimate IQ decrements associated with a specific increment of blood Pb exposure above
23    cutpoints established for the pilot analysis (see 4.4.1.2 below). The specific functions used in the
24    pilot include (see Risk Assessment Report Section 6.1.2 for additional detail on the concentration
25    response functions used and the application of the cutpoint):
26
27           Concurrent blood Pb metric log-linear IQ loss model:
28           IQ loss = -2.7 * In (concurrent blood Pb/concurrent cutpoint)
29
30           Lifetime averaged Pb metric log-linear IQ loss model:
31           IQ loss = -3.04 * In (lifetime average blood Pb/lifetime average cutpoint)
            December 2006                     4-58        Draft - Do Not Quote or Cite

-------
 1          4.4.1.2  Derivation of Cutpoint
 2           For the purposes of this analysis, we identified a blood Pb level below which risks would
 3    not be projected. In this context, the term "cutpoint" will be used for the lower-bound blood Pb
 4    level, below which IQ loss will not be estimated. Specifically, we chose the lower 5th percentile
 5    of blood Pb measurements from Lanphear et al. (2005) as the cutpoint. This reflects our
 6    recognition of the small sample size below this blood Pb level and the associated decreasing
 7    confidence in characterization of the concentration-response function in this blood Pb range (see
 8    Section 3.3.1.2).18 This lower 5th percentile of the sample blood Pb levels in the study,
 9    translates into two separate cutpoints for the two concentration-response functions: 2.4 ug/dL for
10    the prediction of IQ loss using concurrent blood Pb and 6.1 ug/dL for predictions using the
11    lifetime average metric. Because the  cutpoints used in the pilot analysis are not based on an
12    established biological threshold, these cutpoints  are considered policy- or hypothetical
13    thresholds.
14          4.4.1.3  Projection of Population Risk
15           Risk characterization completed for the pilot essentially involves converting the
16    population-level blood Pb distributions into population-level distributions of IQ loss, given
17    consideration for the cutpoints discussed in the last subsection. Specifically, each of the 10,000
18    simulated blood Pb levels generated for a given case study, is compared against the cutpoint. If
19    the simulated total blood Pb level is above the cutpoint, then an IQ loss estimate is generated
20    using the appropriate log-linear concentration response function described in Section 4.4.1.1
21    (i.e., using  either concurrent or lifetime average based on the blood Pb metric). If the simulated
22    blood Pb level is less than the cutpoint, then an IQ loss estimate  is not generated. Note, that the
23    application of this cutpoint approach results in a large fraction of the simulated individuals
24    modeled for each case study, not being assigned an IQ loss estimate because their Pb exposure
25    results in a projected blood Pb level less than the relevant cutpoint.
26           The pathway-apportioned IQ loss estimates generated using this approach are pooled to
27    form a population-level distribution of IQ loss for a given study  area. The point raised in Section
28    4.3.3.3 regarding pathway apportionment at the exposure zone-level (i.e., all simulated
             18 Note, however, that as discussed in Section 3.3.1.2 of this document and in the CD, a threshold blood Pb
      level for neurocognitive effects, including childhood IQ, has not been established and effects have been associated
      with the lowest Pb levels investigated (CD, Sections 6.2, 8.5.1 and 8.6.2). Consequently, this threshold reflects
      concerns over being able to characterize the nature of the concentration-response function for IQ loss and does not
      reflect evidence of a true biological threshold.
             December 2006                     4-59        Draft - Do Not Quote or Cite

-------
 1    individuals from a given zone having the same pathway apportionment) holds for the risk
 2    estimates as well.
 3           Just as with the population-level exposure estimates discussed in Section 4.3.3.3, risk
 4    estimates generated using the approach outlined here can be used to generate several types of
 5    risk metrics including:
 6         •   Population-weighted risk (IQ loss) percentiles: Total IQ loss (with pathway
 7             apportionment) for simulated individuals representing specific points along the
 8             population risk distribution (e.g., 50, 90, 95, 99 and 99.5th percentile simulated
 9             individuals).
10         •   Incidence counts: Number of children within a given study area projected to experience
11             a specific degree of risk (i.e., total IQ loss).
12
13         4.4.2  Risk Estimates
14           In this section we present risk results (IQ loss) generated for each of the three case
15    studies, presented as population percentiles, as well as the numbers of children associated with
16    each percentile (i.e., "Pop" in Tables 4-13 through 4-17). For example,  a Pop value of 5 for the
17    99th percentile risk estimate indicates a projection of 5 children with IQ loss at or above the 99th
18    percentile for that study area.  In addition to presenting results reflecting the aggregate Pb
19    exposure from all pathways, the fraction of the aggregate values associated with policy-relevant
20    background versus policy-relevant exposure pathways are also presented, with the latter category
21    further differentiated among three specific pathways (i.e., inhalation of ambient Pb, incidental
22    ingestion of outdoor soil and incidental ingestion of indoor dust).
23           Each of the risk results tables is also dimensioned on blood Pb model and blood Pb
24    metric, with separates  sets of results in each table being presented for permutations of these two
25    modeling options including: (a)  IEUBK (concurrent blood Pb level metric), (b) IEUBK (lifetime
26    average blood Pb level metric), (c) Leggett (concurrent blood Pb level metric) and (d) Leggett
27    (lifetime average blood Pb level metric). Inclusion of these four parallel sets of risk results
28    reflects the fact that both issues  - blood Pb modeling and the blood Pb metric used in estimating
29    IQ loss - represent potentially important sources of uncertainty in the pilot analysis. The decision
30    to include full sets of risk results dimensioned on these two key issues reflects the fact that no
31    clear "favored" approach (for either blood Pb model, or blood Pb metric) has been identified and
32    therefore, all four permutations of these modeling elements are given equal weight in presenting
33    risk results. Additional dimension are also included for two of the three case studies, as  described
34    below (e.g., inclusion of results  for two air quality scenarios for the primary Pb smelter  case
35    study).

             December 2006                      4-60         Draft - Do Not Quote or Cite

-------
 1         4.4.2.1  Primary Pb Smelter Case Study
 2          Risk results for the primary Pb smelter are further dimensioned on air quality scenario,
 3    with results for both the current conditions and current NAAQS attainment scenarios being
 4    presented. Results for the primary Pb smelter range from 0 IQ points lost to 6 IQ points lost (1st
 5    percentile to the 99.9th percentile simulated individual) for the current conditions air quality
 6    scenario (see Table 4-13). It is important to note that IQ point losses are  only projected for
 7    between 1 and 10% of the modeled population at this case study (39 to 388 children), depending
 8    on the concentration-response function used  (and associated cutpoint), with the remainder having
 9    projected blood Pb levels below the cutpoints used in the analysis. Risk results for the current
10    NAAQS scenario are similar to those for the current conditions scenario, with IQ loss estimates
11    ranging from 0 points to  6 points (1st percentile to the maximum simulated individual) (see
12    Table 4-14).
13          As with the exposure estimates discussed in Section 4.3.5, risk results for this  case study
14    exhibit a trend in terms of pathway apportionment, with higher risk  estimates reflecting a higher
15    proportion of total Pb exposure coming from policy-relevant sources.  For example, the current
16    conditions scenario (IEUBK + concurrent blood Pb metric) (Table 4-13) has a 99.9th  percentile
17    IQ loss of 6 points, with  approximately 98% of the Pb exposure coming  from policy-relevant
18    sources and the vast majority of that originating from the incidental  ingestion of indoor dust
19    containing Pb. It is worth noting that incidental indoor dust ingestion dominates policy-relevant
20    exposures for higher risk percentiles, but that incidental soil ingestion becomes increasingly
21    important (in terms of policy-relevant exposure) as risk and exposure decreases. This  is
22    expected, since indoor dust (driven by the air-to-dust pathway) likely dominates exposure close
23    to the facility  since many of the yards  have been remediated, decreasing the importance of
24    incidental soil-ingestion. However, as you move away from the facility towards lower-risk
25    zones, soils have not been remediated  and consequently, they become more important in
26    determining overall exposure and risk (with incidental ingestion of indoor dust decreasing
27    somewhat in terms of its relative contribution to overall Pb exposure).
            December 2006                     4-61        Draft - Do Not Quote or Cite

-------
1
2
Table 4-13.
conditions.
Projections of IQ loss for the primary Pb smelter case study - current
Percentile
Pop
Total
blood
Pb level
Oig/dL)
IQ
loss
Pathway Contribution*
Diet
Air-inh
Soil-ing
Dust-
ing
Total
BCK
Total
Policy-
Relevant
IEUBK (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
21.9
12.4
7.4
3.7
2.9
2.0
1.3
0.9
0.4
6
4
3
1
1
-
-
-
-
2%
4%
4%
22%
34%
45%
56%
56%
54%
4%
3%
3%
1%
0%
0%
1%
1%
0%
7%
6%
8%
68%
54%
41%
24%
26%
30%
86%
87%
85%
9%
11%
13%
19%
18%
15%
2%
4%
4%
22%
34%
45%
56%
56%
54%
98%
96%
96%
78%
66%
55%
44%
44%
46%
IEUBK (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
28.6
16.9
10.6
5.3
4.1
2.7
1.8
1.2
0.5
5
3
2
-
-
-
-
-
-
2%
5%
5%
26%
29%
62%
25%
56%
54%
4%
3%
3%
1%
1%
1%
1%
1%
0%
3%
12%
12%
63%
59%
20%
63%
26%
30%
90%
80%
80%
11%
11%
18%
11%
18%
15%
2%
5%
5%
26%
29%
62%
25%
56%
54%
98%
95%
95%
74%
71%
38%
75%
44%
46%
Leggett (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
13.9
6.7
4.2
2.0
1.5
1.0
0.7
0.5
0.2
5
3
2
-
-
-
-
-
-
2%
5%
11%
56%
35%
28%
55%
35%
54%
4%
3%
2%
1%
0%
1%
0%
0%
0%
7%
12%
13%
24%
54%
60%
31%
54%
30%
86%
80%
74%
19%
11%
11%
14%
11%
15%
2%
5%
11%
56%
35%
28%
55%
35%
54%
98%
95%
89%
44%
65%
72%
45%
65%
46%
Leggett (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
22.9
11.1
6.8
3.1
2.3
1.6
1.1
0.7
0.3
4
2
<1
-
-
-
-
-
-
2%
5%
17%
29%
22%
29%
48%
56%
54%
4%
3%
1%
1%
1%
1%
0%
1%
0%
3%
12%
32%
59%
68%
59%
38%
24%
30%
90%
80%
50%
11%
9%
11%
13%
19%
15%
2%
5%
17%
29%
22%
29%
48%
56%
54%
98%
95%
83%
71%
78%
71%
52%
44%
46%
            * inh (inhalation), ing (ingestion), diet (includes drinking water) and BCK (total background = diet
            drinking water).
           December 2006
                                           4-62
                                         Draft - Do Not Quote or Cite

-------
1    Table 4-14. Projections of IQ loss for primary Pb smelter case study - NAAQS attainment.
Percentile
Pop
Total
blood
Pb level
Oig/dL)
IQ
loss
Pathway Contribution*
Diet
Air-inh
Soil-ing
Dust-
ing
Total
BCK
Total
Policy-
Relevant
IEUBK (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
18.4
11.2
7.9
3.7
2.9
2.0
1.4
0.9
0.4
6
4
o
J
1
1
-
-
-
-
3%
4%
13%
28%
40%
39%
29%
45%
62%
3%
4%
0%
1%
0%
0%
1%
0%
1%
11%
6%
81%
60%
48%
48%
59%
41%
20%
82%
86%
6%
11%
12%
12%
11%
14%
18%
3%
4%
13%
28%
40%
39%
29%
45%
62%
97%
96%
87%
72%
60%
61%
71%
55%
38%
IEUBK (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
24.9
15.5
11.3
5.4
4.1
2.8
1.9
1.2
0.5
4
o
J
2
-
-
-
-
-
-
4%
5%
5%
9%
54%
29%
52%
56%
44%
4%
3%
3%
2%
0%
1%
1%
1%
0%
5%
12%
12%
15%
30%
59%
31%
24%
43%
88%
80%
80%
74%
15%
11%
16%
19%
12%
4%
5%
5%
9%
54%
29%
52%
56%
44%
96%
95%
95%
91%
46%
71%
48%
44%
56%
Leggett (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
12.7
6.4
4.3
2.0
1.5
1.0
0.7
0.5
0.2
4
o
J
2
-
-
-
-
-
-
4%
10%
5%
25%
22%
56%
54%
57%
59%
4%
1%
3%
1%
1%
1%
1%
1%
0%
5%
77%
12%
63%
68%
24%
26%
26%
24%
88%
12%
80%
11%
9%
19%
19%
17%
17%
4%
10%
5%
25%
22%
56%
54%
57%
59%
96%
90%
95%
75%
78%
44%
46%
43%
41%
Leggett (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
4
19
39
194
388
970
1,940
2,910
3,841
20.7
10.6
7.1
3.1
2.3
1.6
1.1
0.7
0.3
4
2
<1
-
-
-
-
-
-
4%
10%
5%
35%
11%
45%
56%
62%
59%
3%
1%
3%
0%
2%
0%
1%
1%
0%
6%
77%
12%
54%
14%
41%
24%
20%
24%
87%
12%
80%
11%
73%
14%
19%
18%
17%
4%
10%
5%
35%
11%
45%
56%
62%
59%
96%
90%
95%
65%
89%
55%
44%
38%
41%
2
3
* inh (inhalation), ing (ingestion), diet (includes drinking water) and BCK (total background = diet +
drinking water).
           December 2006
                                     4-63
Draft - Do Not Quote or Cite

-------
 1         4.4.2.2  Secondary Pb Smelter Case Study
 2           As mentioned earlier, two sets of risk results were generated for the secondary Pb smelter
 3    case study: (a) risk estimates for the current conditions air quality scenario using a model-only
 4    approach for characterizing soil Pb impacts (Table 4-15) and (b) risk estimates for the current
 5    conditions scenario using a hybrid (model+empirical data) approach for characterizing soil Pb
 6    (Table 4-16). Because there were no projected exceedances of the Pb NAAQS for this case study
 7    location, the current NAAQS attainment scenario is the same as the current condition scenario
 8    and is not presented separately here. Risk results for the model-only approach range from no
 9    adverse impact to 1 IQ points lost (1st percentile to the 99.9th percentile simulated individual).
10    Risk results for the hybrid approach range from 0 IQ points to 2 IQ points lost. It is important to
11    point out that whole IQ point losses are only projected for the top 1% of the modeled child
12    population (17 children) at this case study.
13           The differences between exposure and risk estimates for the two scenarios reflect the
14    higher soil concentrations associated with the hybrid (model + empirical data) approach (see
15    Section 4.3.2.3.2). This can be seen in the higher percentage of total blood Pb attributed to soil
16    ingestion for the hybrid scenario (-20%) versus model-only scenario (<10%). Interestingly, the
17    fraction of total blood Pb associated with dust ingestion is fairly similar (across percentile
18    results) for the two scenarios, suggesting that soil has little impact on indoor dust Pb
19    concentrations (i.e., even though we have increased the soil concentration in the hybrid scenario
20    relative to the model-only scenario and used a dust model in the hybrid scenario based in part of
21    soil concentrations, the overall impact on indoor dust seems to be small, given that the fraction of
22    total blood Pb coming from dust ingestion remains largely unchanged between the two
23    scenarios).
24           Risk results for the secondary Pb smelter should be considered in light of the fact that the
25    highest impact U.S. Census block within the study area (i.e., the block having the highest
26    modeled air concentration and deposition values and associated soil Pb levels), while having
27    adult residents did not have any children 0-7 years of age, according to the U.S. Census for 2000.
28    It is likely that, had children been located within this block, the upper tail of the exposure and
29    risk distributions would have been significantly elevated in  terms of blood Pb levels and IQ loss,
30    respectively. This issue is addressed in the sensitivity analysis conducted for the pilot (see Tables
31    4-18 and  4-19).
            December 2006                     4-64        Draft - Do Not Quote or Cite

-------
1
2
Table 4-15. Projections of IQ loss for secondary Pb smelter case study - modeled soil Pb
            approach.
Percentile
Pop
Total
blood
Pb level
Oig/dL)
IQ
loss
Pathway Contribution*
Diet
Air-inh
Soil-ing
Dust-
ing
Total
BCK
Total
Policy-
Relevant
IEUBK (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
3.7
3.0
2.7
1.9
1.7
1.2
0.9
0.7
0.3
1
1
<1
-
-
-
-
-
-
63%
63%
71%
63%
71%
71%
68%
70%
69%
0.8%
0.8%
0.1%
0.8%
0.0%
0.0%
0.4%
0.2%
0.2%
9.3%
9.2%
7.3%
9.2%
7.3%
7.3%
8.0%
7.5%
7.6%
27%
27%
22%
27%
22%
22%
24%
23%
23%
70%
70%
78%
70%
78%
78%
74%
77%
76%
30%
30%
22%
30%
22%
22%
26%
23%
24%
IEUBK (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
4.7
3.9
3.5
2.5
2.1
1.6
1.1
0.8
0.4
-
-
-
-
-
-
-
-
-
71%
69%
66%
69%
69%
69%
71%
68%
67%
0.0%
0.3%
0.5%
0.3%
0.2%
0.2%
0.1%
0.3%
0.5%
7.3%
7.7%
8.8%
7.8%
7.8%
7.6%
7.3%
8.1%
8.1%
22%
23%
25%
23%
23%
23%
22%
24%
25%
78%
76%
73%
76%
76%
76%
78%
75%
74%
22%
24%
27%
24%
24%
24%
22%
25%
26%
Leggett (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
2.1
1.6
1.4
1.0
0.9
0.6
0.5
0.3
0.2
-
-
-
-
-
-
-
-
-
67%
63%
70%
63%
68%
69%
65%
67%
69%
0.5%
0.8%
0.1%
0.8%
0.4%
0.3%
0.6%
0.5%
0.3%
8.2%
9.2%
7.4%
9.4%
7.9%
7.7%
9.4%
8.0%
7.6%
24%
27%
22%
27%
24%
23%
25%
25%
23%
74%
70%
78%
69%
75%
76%
71%
74%
76%
26%
30%
22%
31%
25%
24%
29%
26%
24%
Leggett (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
3.0
2.3
1.9
1.4
1.2
0.9
0.7
0.5
0.2
-
-
-
-
-
-
-
-
-
71%
69%
67%
67%
70%
69%
70%
71%
70%
0.0%
0.2%
0.4%
0.4%
0.1%
0.3%
0.1%
0.0%
0.1%
7.3%
7.6%
7.9%
8.2%
7.4%
7.7%
7.4%
7.3%
7.4%
22%
23%
24%
24%
22%
23%
22%
22%
22%
78%
76%
74%
74%
77%
76%
77%
78%
77%
22%
24%
26%
26%
23%
24%
23%
22%
23%
3
4
       * inh (inhalation), ing (ingestion), diet (includes drinking water) and BCK (total background = diet +
drinking water + fraction of soil Pb attributable to background, see Section 4.3.2.2).
            December 2006
                                            4-65
Draft - Do Not Quote or Cite

-------
1
2
Table 4-16. Projections of IQ loss for secondary Pb smelter case study -hybrid soil Pb
            approach.
Percentile
Pop
Total
blood
Pb level
Oig/dL)
IQ
loss
Pathway Contribution*
Diet
Air-inh
Soil-ing
Dust-
ing
Total
BCK
Total
Policy-
Relevant
IEUBK (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
4.7
3.7
3.3
2.4
2.0
1.5
1.1
0.8
0.4
2
1
1
<1
-
-
-
-
-
54%
56%
54%
56%
58%
56%
57%
56%
58%
0.4%
0.2%
0.3%
0.2%
0.1%
0.2%
0.1%
0.2%
0.0%
20%
19%
20%
19%
18%
19%
18%
19%
18%
26%
25%
25%
25%
24%
25%
24%
25%
24%
60%
62%
60%
62%
64%
62%
63%
61%
64%
40%
38%
40%
38%
36%
38%
37%
39%
36%
IEUBK (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
6.3
4.9
4.4
3.2
2.7
2.0
1.5
1.1
0.5
<1
-
-
-
-
-
-
-
-
56%
54%
57%
52%
57%
54%
57%
55%
58%
0.3%
0.3%
0.1%
0.5%
0.1%
0.4%
0.2%
0.4%
0.0%
19%
20%
19%
21%
18%
20%
19%
19%
18%
25%
25%
24%
27%
24%
26%
24%
26%
24%
62%
60%
63%
57%
63%
59%
62%
60%
64%
38%
40%
37%
43%
37%
41%
38%
40%
36%
Leggett (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
2.3
1.9
1.7
1.2
1.0
0.8
0.6
0.4
0.2
-
-
-
-
-
-
-
-
-
57%
55%
53%
57%
57%
56%
55%
58%
56%
0.2%
0.3%
0.4%
0.2%
0.2%
0.2%
0.3%
0.1%
0.2%
19%
19%
20%
19%
19%
19%
20%
18%
19%
24%
25%
26%
24%
24%
25%
25%
24%
25%
62%
61%
58%
62%
62%
62%
60%
64%
62%
38%
39%
42%
38%
38%
38%
40%
36%
38%
Leggett (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
2
8
17
84
167
418
836
1,254
1,655
3.3
2.7
2.5
1.8
1.5
1.1
0.8
0.6
0.3
-
-
-
-
-
-
-
-
-
57%
56%
55%
55%
56%
57%
55%
57%
55%
0.1%
0.2%
0.3%
0.3%
0.3%
0.2%
0.2%
0.1%
0.3%
19%
19%
20%
20%
19%
19%
19%
19%
20%
24%
25%
25%
25%
25%
24%
25%
24%
25%
63%
62%
60%
60%
61%
62%
61%
63%
60%
37%
38%
40%
40%
39%
38%
39%
37%
40%
4
5
       * inh (inhalation), ing (ingestion), diet (includes drinking water) and BCK (total background = diet
       drinking water + fraction of soil Pb attributable to background, see Section 4.3.2.2).
            December 2006
                                            4-66
Draft - Do Not Quote or Cite

-------
 1         4.4.2.3  Near Roadway (Urban) Case Study
 2          Risk results for the near roadway (urban) case study range from no adverse impacts to 3
 3    IQ points lost (1st percentile to the 99.9th percentile simulated individual) (see Table 4-17). It is
 4    important to note that whole IQ point losses are projected for the top 1-5% of the modeled
 5    population at this case study (3 to 16 children), depending on the concentration-response
 6    function (and  associated cutpoint), with the remainder having projected blood Pb levels that do
 7    not exceed the cutpoints used in the analysis.  In considering the risk results generated for this
 8    case study, it is important to reiterate the fact that the study area includes a 1.5 mile long urban
 9    road segment  and the residents living within 200m of that road  segment. Actual near roadway
10    exposures across an urban or metropolitan area would be comprised of many such road segments
11    and associated residential buffer areas and consequently would involve far larger study
12    populations.
13          As with the exposure estimates discussed in Section 4.3.5, risk results for this case study
14    exhibit a fairly consistent pattern regarding pathway apportionment, with policy-relevant sources
15    typically contributing about 55-65% of total Pb  exposure and policy-relevant background
16    contributing the rest (as noted  in Section 4.3.5, this likely reflects the fact that this case study was
17    modeled using a small number of exposure bands which reduces the specificity in capturing
18    gradients in population-level Pb exposure near the roadway). Unlike the other two case studies,
19    for the near roadway (urban) case study, policy-relevant source exposure is split almost evenly
20    between the incidental ingestion of soil and indoor dust. This likely reflects the relatively high
21    historical soil  concentrations associated with this case study, which increase the importance of
22    the incidental  soil ingestion pathway. Furthermore, the relatively low ambient air concentrations
23    (relative to the other case studies) mean that indoor dust loading will be reduced which will de-
24    emphasize that pathway in driving overall Pb exposure.
            December 2006                     4-67        Draft - Do Not Quote or Cite

-------
1    Table 4-17. Projections of IQ loss for near roadway (urban) case study.
Percentile
Pop
Total
blood
Pb level
(jig/dL)
IQ
loss
Pathway Contribution*
Diet
Air-inh
Soil-ing
Dust-
ing
Total
BCK
Total
Policy-
Relevant
IEUBK (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
<1
2
3
16
32
80
159
239
316
6.5
5.0
4.4
3.1
2.6
1.9
1.4
1.0
0.4
o
J
2
2
<1
<1
-
-
-
-
44%
44%
44%
44%
44%
44%
35%
35%
44%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
30%
30%
30%
30%
30%
30%
37%
37%
30%
26%
26%
26%
26%
26%
26%
27%
27%
26%
44%
44%
44%
44%
44%
44%
35%
35%
44%
56%
56%
56%
56%
56%
56%
65%
65%
56%
IEUBK (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
<1
2
3
16
32
80
159
239
316
9.1
7.0
6.2
4.4
3.6
2.7
1.9
1.4
0.6
1
<1
<1
-
-
-
-
-
-
44%
44%
44%
44%
44%
44%
35%
35%
44%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
30%
30%
30%
30%
30%
30%
37%
37%
30%
26%
26%
26%
26%
26%
26%
27%
27%
26%
44%
44%
44%
44%
44%
44%
35%
35%
44%
56%
56%
56%
56%
56%
56%
65%
65%
56%
Leggett (concurrent blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
<1
2
3
16
32
80
159
239
316
4.2
3.1
2.7
1.9
1.6
1.2
0.8
0.6
0.3
2
<1
<1
-
-
-
-
-
-
44%
44%
35%
44%
44%
44%
44%
44%
44%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
30%
30%
37%
30%
30%
30%
30%
30%
30%
26%
26%
27%
26%
26%
26%
26%
26%
26%
44%
44%
35%
44%
44%
44%
44%
44%
44%
56%
56%
65%
56%
56%
56%
56%
56%
56%
Leggett (lifetime average blood Pb metric)
99.9th
99.5th
99th
95th
90th
75th
Median
25th
1st
<1
2
3
16
32
80
159
239
316
6.9
4.9
4.1
2.9
2.4
1.7
1.2
0.9
0.4
<1
-
-
-
-
-
-
-
-
20%
44%
44%
44%
35%
44%
35%
44%
44%
0.1%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
52%
30%
30%
30%
37%
30%
37%
30%
30%
29%
26%
26%
26%
27%
26%
27%
26%
26%
20%
44%
44%
44%
35%
44%
35%
44%
44%
80%
56%
56%
56%
65%
56%
65%
56%
56%
2
3
* inh (inhalation), ing (ingestion), diet (includes drinking water) and BCK (total background = diet +
drinking water).
           December 2006
                                      4-68
Draft - Do Not Quote or Cite

-------
 1          4.4.3   Uncertainty Analysis (Sensitivity Analysis, Performance Evaluation and Other
 2                 Considerations)
 3           This section discusses uncertainty associated with the pilot risk analysis. As mentioned in
 4    Section 4.2.5.7, for the pilot analysis, we completed a sensitivity analysis focusing on the impact
 5    of uncertainty in individual modeling elements on risk results.  In addition to the sensitivity
 6    analysis results, the results of the performance evaluation conducted for the pilot analysis
 7    (section 4.3.6) can also be used to gain insights into elements of the analysis that might be
 8    subject to significant uncertainty. Finally, there are a range of potential sources of uncertainty,
 9    that, while not formally included in the sensitivity analysis or performance evaluation (due to a
10    lack of data), can still be discussed qualitatively. These sources are also addressed in this section.

11          4.4.3.1   Sensitivity Analysis Methodology
12           As mentioned in Section 4.2.5.7, the sensitivity  analysis completed for the pilot involved
13    a "one element at a time elasticity analysis" in which the full model was run with one of the
14    selected modeling elements adjusted to reflect an alternate (bounding if possible) input value or
15    modeling choice.19 The results of that run with the modified modeling element were then
16    compared to the "baseline risk run" to determine the magnitude of the impact on risk results
17    generated by modifying that one modeling element.20 This procedure was repeated for all of the
18    modeling elements selected for coverage in the sensitivity analysis.
19           The determination of the degree of impact from a given modeling element on risk results
20    was based on a comparison of specific risk percentiles between the baseline and sensitivity
21    analysis runs, including results generated for the 50th, 90th, 95th, 99th, and 99.5th simulated
22    individuals. For example, we might compare the 90th percentile risk for the baseline run with
23    that from a sensitivity analysis run reflecting a different option for a specific modeling element
24    (e.g., different input dataset or different modeling approach). The difference between these two
             19 Alternate models or input datasets for use in the sensitivity analysis were selected to reflect the range of
      options identified for a particular modeling element (i.e., given several values for a particular input parameter, the
      value representing the high or low bound on the range for that input would be selected). This approach reflects our
      desire that the sensitivity analysis capture, to the extent possible, the full impact of uncertainty in a particular
      modeling element on risk results.
             20 For purposes of the sensitivity analysis, the "baseline run" was defined as a full risk run for the primary
      Pb smelter case study involving the following modeling choices: IEUBK biokinetic model and the concurrent blood
      Pb metric. It is important to emphasize that the defining of a baseline run does not place greater confidence in this
      particular combination of modeling elements, but rather reflects the need to have a set of risk results for use in
      gauging the magnitude of the impact of alternative modeling elements on risk results.
             December 2006                     4-69         Draft - Do Not Quote or Cite

-------
 1    estimates of the 90th percentile risk would then be assigned to that particular modeling element
 2    to represent its impact on risk (i.e., the sensitivity of risk results to that modeling element).
 3          Results of the sensitivity analysis for a particular modeling element are presented both in
 4    terms of (a) their absolute impact on IQ  loss estimates (associated with the 99.9th percentile
 5    population percentile - see below) and (b) in terms of the percent difference between the
 6    baseline risk estimate and the estimates generated with modification of the element under
 7    consideration. Inclusion of a percentile impact metric in the sensitivity analysis makes it easier to
 8    rank modeling elements in terms of their overall impact on risk results. Note, that for some
 9    modeling elements, rather than having two data points (the baseline risk estimate and a single
10    alternative from the sensitivity analysis run), we actually have three alternatives (the baseline
11    risk estimate and a higher- and lower-end risk estimate generated by considering options
12    producing both lower risk and higher risk estimates relative to baseline). In this case, we present
13    all three risk estimates as well as the percent difference between the low- and high-end risk
14    estimate, with the baseline risk estimate encompassed within this percent range.
15          The majority of the sensitivity analysis is based the primary Pb smelter study area.
16    Specifically, both the baseline run as well as the sensitivity analysis runs examining alternate
17    options for specific modeling elements were completed using the primary Pb smelter. There are
18    two exceptions to this. In the first, an analysis was completed focused on the secondary Pb
19    smelter case study which considered the impact on risk estimates of locating children in  the
20    census block possessing the highest Pb impact on modeled media including ambient air, soil  and
21    indoor dust. Secondly, a  sensitivity analysis was completed (also for the secondary Pb smelter),
22    focusing on the characterization of Pb concentrations in outdoor soil. The results of the
23    sensitivity analysis based on the primary Pb smelter case study are generally applicable to the
24    other two case study locations. However, the results of the two analyses focused on the
25    secondary Pb smelter are relevant only for this case study since they are considering site-specific
26    factors related to exposure and risk modeling at this specific location.
27          Table 4-18 lists the modeling elements included in the sensitivity analysis and presents a
28    brief summary of the alternative modeling approaches/inputs used to represent each within the
29    sensitivity analysis. As mentioned earlier, consideration for which modeling elements would be
30    included in the sensitivity analysis was based on professional judgment by the  staff as to which
31    modeling elements were  likely to have significant impacts on risk and on consideration for data
32    availability. Those modeling elements expected to have a significant impact on risk results but
33    for which available data did not support inclusion in the sensitivity analysis are discussed
34    qualitatively.
            December 2006                     4-70        Draft - Do Not Quote or Cite

-------
1    Table 4-18. Modeling elements considered in the pilot sensitivity analysis (including
2                summary of approaches used to derive alternate approaches/inputs).
Modeling
Element

Description

Baseline Run

Sensitivity Analysis Run(s)
Media modeling (indoor dust)




Indoor dust

modeling







Statistical model used to
predict indoor dust
concentrations based on
outdoor soil and/or
outdoor ambient air



Two models used (see
Section 4.3.2.4.1):
A) site-specific
statistical (air-only)
model used for
remediation zone near
facility.
B) pooled analysis
model (air + soil) used
for areas further from
facility.



Two options considered:
A) use pooled analysis model (soil +
air) across entire study area.
B) use pooled analysis model (air
only) across entire study area.



Blood Pb modeling
Combined
assessment
focusing on oral
absorption for
background
sources:
- absolute
absorption factor
(water)
- absolute
absorption factor
(diet)
Combined
assessment of
factors related to
intake and uptake
modeling of soil
and indoor dust:
- soil/dust
weighting factor
- total fraction
accessible (soil
and dust)

Blood Pb
modeling





Used to estimate uptake
of Pb following dietary
consumption and drinking
water ingestion
(background sources).




Factors used to (a)
determine the fraction of
total soil+dust ingestion
that is for each media
(soil/dust weighting
factor) and then (b)
model uptake of Pb from
ingestion soil and dust
(total fraction accessible)

Biokinetic and statistical
(empirical) models used
to predict blood Pb in
children




Single value used for
both pathways (50
percentile)






- soil/dust weighting
factor: 45%
- total fraction
accessible: 48% for
soil and 26% for dust




IEUBK (see Section
4.3.3.1.1)





Alternate values based on study data
not identified. Therefore, conducted
simple test of mathematical elasticity
using factors that are 10% higher and
10% lower (i.e., 40% and 60% against
the baseline value of 50%).






Alternate values obtained from Von
Lindern et al., 2003.
- soil-dust weighting factor: 58%
- total fraction accessible: soil and dust
are both 18%



Two alternate models considered:
- an alternate biokinetic model:
Leggett (see Section 4. 3. 2.1)
- a statistical (empirical) model:
Lanphear et al., 1998 (see Risk
Assessment Report Section 6.3. 1)21
            21 The Lanphear empirical model was developed by relating blood Pb measurements in young children to a
     number of factors including: (a) Pb exposures in air, soil, house Pb loading, drinking water, (b) the presence/absence

            December 2006                      4-71         Draft - Do Not Quote or Cite

-------
Modeling
Element
Geometric
standard deviation
(GSD)
Description
Used to represent inter-
individual variability in
Pb biokinetics and
behavior related to Pb
exposure.
Baseline Run
1.6 (see Section
4.3.3.4)
Sensitivity Analysis Run(s)
Two alternate values considered:
- 1.3 (lower bound of GSDs provided
for children living near Pb smelters,
USEPA, 1989)
- 1.9 (reflects range of GSDs found in
NHEXAS study, USEPA, 2004).
IQ loss modeling
Statistical fit of
log-linear
concentration-
response function
model
Form of
concentration-
response function
Blood Pb metric
Cutpoint
Confidence intervals
associated with fit of the
Lanphpear pooled
analysis log-linear model
Actual form of the
concentration response
function (e.g., log-linear,
linear)
Type of blood Pb
measurement used to
represent exposure (i.e.,
concurrent, lifetime
average, peak)
Exposure level below
which there is insufficient
confidence (in the form of
the concentration-
response function) to
predict IQ loss
Best fit of log-linear
model used in baseline
run
Log-linear (Lanphear
etal.,2005)
Concurrent
Concurrent cutpoint:
2.4 ug/dL
Upper and lower 95th % confidence
intervals on the model used in
sensitivity analysis
Linear model with breakpoint at 10
ug/dL (i.e., linear model fit to subset
of sample data with blood Pb levels
<10 ug/dL) (Lanphear et al., 2005)
Two metrics considered:
- lifetime average
- peak annual averaged
Given that data applicable in
establishing lower cutpoints are
limited, alternate cutpoint was set as
1/2 baseline cutpoints (i.e., 1.2 ug/dL)
Sensitivity analyses focused on the secondary Pb smelter
Location of
children close to
the facility in the
U.S. Census block
with the greatest
Pb media impacts
Secondary Pb smelter has
a U.S. Census block
located close to the
facility with high air and
soil Pb impacts and with
adult residents, but with
no child residents (this
means it is not included
in projections of child
exposure and risk
estimates for the study
area)
Baseline run does not
consider this block in
calculating risk since
the U.S. Census data
used in the analysis
identify no child
residents within that
block
Calculate child exposure and risk
estimates for this high-impact block as
if there were children living there.
Specifically, calculate mean and high-
end percentile exposure and risk
estimates specifically for that block to
gain perspective for the magnitude of
risk which could exist if children did
live in that block.
of Pb paint and (c) a range of socioeconomic variables. The best fitting regression model included Pb concentrations
in soil, and Pb loadings in house dust In order to apply the Lanphear model in the pilot, it was necessary to convert
the estimated house dust exposure concentrations derived for each study area into dust Pb loadings. This was done
using a regression relationship based on the same underlying 1997 national Pb housing survey data used in
developing the Lanphear empirical blood Pb model (see Section 6.3.1 of the Risk Assessment Report for additional
details).
        December 2006
4-72
Draft - Do Not Quote or Cite

-------
          Modeling
          Element
      Description
    Baseline Run
     Sensitivity Analysis Run(s)
      Characterization
      ofPb
      concentrations in
      outdoor soil
Fate and transport
modeling used to predict
soil Pb levels based on
modeled air concentration
and deposition over the
study area. Performance
evaluation for soil
modeling suggests under-
prediction using MPE
fate and transport
modeling (compared with
measured soil data at
surrogate location - see
Table 4-11).
Model-only approach'.
use of MPE fate and
transport modeling to
predict soil
concentrations.
Hybrid model-empirical data
approach: use of MPE modeling to
generate a soil concentration surface
characterizing the relative spatial
profile of soil levels combined with
surrogate measurement data for soil Pb
near secondary smelters to "scale up"
the modeled surface to match the
surrogate data.  (Note, inclusion of this
hybrid model reflects consideration for
surrogate data (Kimbrough et al.,
1995) which suggests underestimation
by model-only approach - see
performance evaluation discussion in
Table 4-10 ). This hybrid modeling
approach resulted in soil
concentrations approximately 3X
higher than the model-only approach.22
 2          4.4.3.2   Sensitivity Analysis Results

 3            The results of the sensitivity analysis are summarized in Table 4-19 (see Table 4-18 for
 4    detail on each modeling element included in the sensitivity analysis). Because of the potential
 5    importance of higher-end risk percentiles in decision making, the presentation of sensitivity
 6    analysis results here includes only the range of predicted  99.9th percentile risk estimates.
 7    However, the percentile range associated with each modeling element (i.e., the percent difference
 8    between the baseline and alternative option runs) reflects the results for the 90th-99.9th
 9    percentiles. For a more complete presentation of the sensitivity analysis results please refer to the
10    Risk Assessment Report, Section 6.3.
                The hybrid (model-empirical data) approach considered in the sensitivity analysis for the secondary Pb
      smelter case study also includes a modification to the approach used in modeling indoor dust, compared with the
      baseline approach. While indoor dust for the baseline run (using soil data generated with the model-only approach)
      used the AGG air-only dust model, the sensitivity analysis run used the AGG air+soil dust model. As discussed in
      Section 4.3.2.4.2, the decision to switch to the AGG air+soil dust model reflects the fact that the increased soil Pb
      levels considered in the sensitivity analysis run required a dust model that factors those increased soil concentrations
      in predicting indoor dust in order to assess their full impact on modeled blood Pb levels (i.e., use of the AGG air-
      only dust model would have produced lower blood Pb levels since the impact of increased soil Pb concentrations on
      indoor blood Pb levels would not have been considered). Inclusion of the AGG air+soil dust model in the sensitivity
      analysis run does complicate interpretation of the results since they reflect the combined impact of the higher soil Pb
      levels as well as the effect of switching  from the AGG air-only dust model to the AGG air+soil dust model.

              December 2006                        4-73         Draft - Do Not Quote or Cite

-------
1   Table 4-19. Summary of Sensitivity Analysis Results.
Modeling Element
Risk (IQ loss) estimates for the 99.9th
simulated individual (baseline run and
sensitivity analysis run)*
Percentile difference
(baseline vs alternative
option) across population
percentiles: 90th to
99.9th **
Media modeling (indoor dust)
Indoor dust modeling
Baseline run: -5
Sensitivity analysis runs:
AGO (air-only): -5
AGO (air+soil): -4
-35% to +30%
Blood Pb modeling
Combined assessment:
- absolute absorption factor (water)
- absolute absorption factor (diet)
Combined assessment:
- soil/dust weighting factor
- total fraction accessible (soil and
dust)
Blood Pb models
GSD
Baseline run: -5
Sensitivity analysis runs:
AF diet, water (40%): -4
AF diet, water (60%): -6
Baseline run: -5
Sensitivity analysis run:
Van Lindern values : -4
Baseline run: -5
Sensitivity analysis runs:
Leggett: -5
Lanphear: -5
Baseline run: -5
Sensitivity analysis runs:
GSD (1.3): -5
GSD (1.9): -6
-88% to +52%
-5 to -88%
-100 to +55%
-51% to +124%
IQ loss modeling
Statistical fit of log-linear model
Form of concentration-response
function
Blood Pb metric
Baseline run: -5
Sensitivity analysis runs:
95th %LCL: -1
95th % UCL: -8
Baseline run: -5
Sensitivity analysis run:
Linear ( 10 ug/dL breakpoint): -9
Baseline run: -5
Sensitivity analysis runs:
Lifetime averaged: -4
Highest annual (peak): -8
-53% to +80%
-12% to +102%
-15% to +395%
          December 2006
4-74
Draft - Do Not Quote or Cite

-------
              Modeling Element
                                      Risk (IQ loss) estimates for the 99.9th
                                      simulated individual (baseline run and
                                     	sensitivity analysis run)*	
                Percentile difference
               (baseline vs alternative
              option) across population
                 percentiles: 90th to
                      99.9th **
      Cutpoint
                                     Baseline ran:

                                     Sensitivity analysis ran:
                                     1/2 baseline outpoint:
                                                                   -5
                   +43% to+412%
      Sensitivity analyses focused on the secondary Pb smelter
      Spatial distribution of children within
      the study area, particularly focusing
      on the portion of the study area
      closest to the emissions source
      (implemented using the secondary Pb
      smelter case study)
                                     This sensitivity analysis departs from the
                                     others in focusing only on risk for the
                                     single block of interest and not for the
                                     entire study area.

                                     Baseline ran:
                                     Note, the baseline ran has no children in
                                     this block, so there would be no IQ loss
                                     projected for the block with the highest
                                     Pb media concentrations.  However, the
                                     99.9th% risk level modeled for the entire
                                     study area ranges from -1 to -2 IQ points.

                                     Sensitivity analysis ran:
                                     Risk levels for max-impact block
                                     (assuming children are there):
                                     ~ mean:    -2 to -5
                                     - 95th%:   -4 to -7
                                     - 99th%:   -4 to -8
            Percentile difference is not
            applicable here because the
            baseline results is across the
            entire study area and the
            sensitivity results are for the
            high-impact block not
            included in the baseline ran.
            But the results clearly
            demonstrate that the extreme
            tail of this case study's risk
            distribution would be pushed
            higher if children were
            located in this high-impact
            block.
1
2
3
4
5
6
7
     Outdoor soil Pb modeling
     (implemented using the secondary Pb
     smelter case study)
                                     Baseline ran:   -1

                                     Sensitivity analysis ran:
                                     (X3) soil concentrations:  -2
            +100% to +300%
            It is interesting to note that,
            while differences in blood Pb
            levels are only on the order of
            30 to 50%, between the
            baseline and sensitivity
            analysis ran (across
            percentiles) given the non-
            linearities in the
            concentration-response
            function near this lower range
            of blood Pb levels, the
            differences in IQ can be far
            greater, as seen here.	
        * As noted in text, the 99.9th% risk estimates was selected as the basis for presenting results in this table
given the potential importance of higher-end risk results in supporting decision making. The sensitivity analysis did
include results for a range of percentile risk metrics (see Risk Assessment Report Section 6.3 for additional detail).
        ** Percent difference reflects sensitivity analysis results seen across the 90th to 99.9th percentile simulated
individuals (Note, that many of the highest sensitivity analysis results - those exceeding 100% - were seen for
impacts on the 90th-95th percentile simulated individual results and not for higher percentiles).
             December 2006
                                                   4-75
Draft - Do Not Quote or Cite

-------
 1           The sensitivity analysis results presented in Table 4-19 result in the following
 2    conclusions regarding the sensitivity of risk results generated for the pilot to specific sources of
 3    uncertainty:

 4         •   Modeling elements with the greatest impact: Modeling elements with high impacts
 5             (relative to the entire set considered in the sensitivity analysis) include: (a)
 6             characterization of soil Pb levels (for the secondary Pb smelter case study), (b) the
 7             blood Pb model, (c) the blood Pb metric, (d) the GSD, (e) the concentration-response
 8             function and (f) the cutpoint.
 9         •   Overall perspective on magnitude of uncertainty in high-end risk results:
10             Consideration of the range of impacts across modeling elements presented in Table 4-
11             19 suggests that overall uncertainty (resulting from these factors) would likely be under
12             one order of magnitude. Note, as stated earlier, this is only a qualitative assessment -
13             true quantitative uncertainty estimates would require a formal probabilistic uncertainty
14             analysis.

15         4.4.3.3  Additional Considerations
16           As the performance evaluation (described in Section 4.3.6) involved comparison of
17    modeled results against available empirical data, we are using this in addition to the sensitivity
18    analysis (described in  Section 4.4.3.1) to characterize potential uncertainty associated with
19    specific steps of the pilot analysis. This section also discusses potential sources of uncertainty
20    that have not been quantitatively investigated, but still deserve qualitative discussion.

21         4.4.3.3.1  Performance Evaluation
22           From the performance evaluation (presented in Section 4.3.6), we have drawn the
23    following conclusions regarding potential uncertainty in projected media concentrations and
24    blood Pb levels for the three  case studies:

25         •   Outdoor air Pb concentrations for the primary Pb smelter: Performance evaluation
26             completed for air modeling at this case study suggests an overestimation of air Pb
27             levels closer to the facility (in the range of perhaps 50-60%) with air concentrations
28             further from the facility potentially underestimated. The potential overestimation of air
29             concentrations near the facility is a potentially important factor since it could result in
30             overestimations of exposure (through both direct inhalation and indirectly through
31             indoor dust loading) for that portion of the study area likely to experience the highest
32             risks.
33         •   Outdoor air  concentrations for the secondary Pb smelter: Performance evaluation
34             completed for air modeling at this case study suggests the possibility of a potential
35             underestimation of air concentrations. This is an important observation since it could
36             results in under-predicted exposures and risks for this case study.
37         •   Outdoor soil concentrations at the secondary Pb smelter: Performance evaluation of
38             modeled outdoor Pb soil concentrations at the secondary Pb smelter suggests a
39             potentially significant underestimation (up to a factor of 3.0). This finding from the

             December 2006                     4-76        Draft - Do Not Quote or Cite

-------
 1             performance evaluation led to this issue being specifically addressed in the sensitivity
 2             analysis (see Tables 4-18 and 4-19) and this issue will not be discussed further here.

 3         •   Blood Pb levels for the primary Pb smelter. Two performance evaluations were
 4             conducted for blood Pb levels generated for this case study: (a) a comparison of the
 5             median modeled blood Pb level for the case study against the median blood Pb level
 6             for 1-5 yr olds in the U.S. (NHANES IV) and (b) a comparison of the high-end
 7             modeled blood Pb levels (95^-99.9th percentile) against the high-end percentiles from
 8             a blood screening study completed for children living near the smelter (US DHHS,
 9             2003). The results of the fist assessment suggest that our modeling is not generating
10             unreasonable central-tendency blood Pb levels for the study area.  The second
11             assessment (focusing on high-end exposure percentiles) suggests that our modeling is
12             also generating reasonable high-end exposures for that subpopulation, given the
13             measurement data that are available.

14         •   Blood Pb levels for the secondary Pb smelter and near roadway (urban) case study: As
15             with the primary Pb smelter, performance evaluation focusing on central-tendency
16             blood Pb levels for this case study suggest that our modeling is not generating
17             unreasonable central-tendency blood Pb levels.
18
19         4.4.3.3.2 Qualitative Discussion of Uncertainty
20           The staff has identified a number of modeling elements with the potential to introduce
21    uncertainty into the risk results, that were not addressed quantitatively in the sensitivity analysis
22    and performance  evaluation, the results of which are discussed above. These elements and the
23    potential uncertainty associated with them in the pilot analysis are  discussed below, with the
24    discussion organized by modeling step.
25
26           Characterizing media concentrations
27           Sources of potential uncertainty in characterizing media concentrations which are not
28    explicitly considered in sensitivity analysis or performance evaluation are described here.

29         •   Source characterization: The estimates of Pb emissions rates and particle size profiles
30             for the two point source case studies remains an area of uncertainty. The uncertainty is
31             greatest with regard to fugitive emissions, both on the facility property and any
32             associated re-suspension outside the property.
33         •   Paint Pb as a component of indoor dust: The models used in the pilot analysis to
34             estimate indoor dust Pb (based on Pb in air and soil), may include some contribution
35             from paint Pb in their estimates. Specifically, all of the statistical models used in the
36             pilot to  predict indoor dust Pb include intercept terms which represent that fraction of
37             indoor dust Pb not related to variations in outdoor soil Pb or air Pb. Given the data sets
38             from which regressions were derived, it is likely that these intercept terms reflect, to
39             some extent, paint Pb.  For the pilot, the total value predicted by these models for
40             indoor dust Pb was treated as policy-relevant and no effort was made to identify the

             December 2006                     4-77        Draft - Do Not Quote or Cite

-------
 1             fraction of predicted indoor dust Pb levels resulting from Pb paint. There is the
 2             potential, then, that the predictions of indoor dust Pb have over-estimated the policy -
 3             relevant component by not excluding the paint Pb signal.
 4         •   Characterizing the spatial gradient of outdoor air concentrations for the near roadway
 5             (urban) case study: The use of dispersion model results for highway-related diesel (and
 6             other PM component) emissions to derive spatial gradients for re-entrained near
 7             roadway air-bound Pb is subject to uncertainty.
 9          Exposure Analysis and Blood Pb modeling
10          Sources of potential uncertainty in characterizing exposure and estimating blood Pb
11    levels which are not explicitly considered in sensitivity analysis or performance evaluation are
12    described here.

13         •  Modeling of a static child population (in terms of media concentrations and residence
14            time within the study area): The risk results generated for the pilot assume that
15            modeled child populations  at all three case studies (a) come in contact with fixed media
16            concentrations (i.e., they do not change over time) and (b) reside for their entire
17            exposure period within the case study.  There is uncertainty associated with these
18            simplifying assumptions. For example, media concentrations may change with time;
19            this has been observed with recontamination of remediated soil near the primary Pb
20            smelter (USEPA, 2006e). With regard to residence time, it is likely that children
21            residing in the study area reflect a range of residence times, some having lived for
22            most, if not all of their lifetimes, while others may only spend only a fraction of their
23            time living near the facility. The assumption that all of the modeled children live in the
24            study area for their entire exposure period is conservative and will contribute to an
25            overestimation of Pb exposure for the case study.
26

27         4.5   SUMMARY OF FINDINGS AND CONSIDERATIONS FOR THE FULL-
28               SCALE ASSESSMENT
29          This section summarizes the risk results and analyses of uncertainty for the pilot
30    assessment (Section 4.5.1) and discusses plans for the full-scale assessment (Section 4.5.2) that
31    will be presented in the second draft  of this document. The detailed risk results for individual
32    case studies are presented in Sections 4.4.2.1, 4.4.2.2  and 4.4.2.3.
33          We note that the primary purpose of the pilot assessment was to test out methodologies
34    and design features of the assessment, as well as to assess the availability of different types of
35    information pertinent to this assessment.  Consequently, the risk results associated with the pilot
36    are not intended to reflect our best estimates of risk associated with these case studies. Rather,
37    they reflect preliminary estimates, limited by the initial application of our modeling tools and
38    information. We intend to build on our experience and findings associated with the pilot
39    assessment in designing and implementing the full-scale risk assessment, the purpose of which is
            December 2006                     4-78        Draft - Do Not Quote or Cite

-------
1    to inform the Agency's development of and consideration of NAAQS policy options with regard
2    to policy relevant sources of ambient Pb.
                                                                                             ,th
 3          4.5.1   Summary of Findings in the Pilot Assessment
 4           Risk results generated for the three case studies suggest that individuals in the upper 101
 5    to 5th percentile of exposure (depending on case study) have the potential for quantifiable IQ loss
 6    associated with projected Pb exposure with IQ decrements ranging from less than an IQ point to
 7    greater than six IQ points.  Among the three case studies, the greatest IQ loss was projected for
 8    children living in the vicinity of the primary Pb smelter case study, followed by the near roadway
 9    (urban) case study and the secondary Pb smelter case study. For the primary Pb smelter,
10    individuals in the upper 10th percentile (with IQ losses ranging from <1 to 6 points) had
11    exposures dominated by contributions from policy-relevant sources (-98%), with incidental
12    ingestion of indoor dust being by far the dominant pathway. For the secondary Pb smelter,
13    individuals in the upper 5th percentile  (with IQ losses ranging from <1 to 2 points) have between
14    20% and 40% of their exposure associated with policy-relevant sources (this range reflecting the
15    model-only approach and the hybrid approach,  respectively). Contributions to IQ loss from
16    policy-relevant sources are estimated to split evenly between soil  and dust ingestion for the
17    model-only scenario, but becomes dominated by soil ingestion, for the hybrid scenario (as would
18    be expected given the higher soil Pb concentrations associated with the hybrid scenario). For the
19    near roadway (urban) case  study, individuals in the upper 10th percentile (with IQ losses ranging
20    from <1 to 3 points) have about half of their exposure coming from policy-relevant sources, with
21    the majority of this coming from indoor dust ingestion.23
22           Risk results generated for the secondary Pb smelter should be carefully considered in
23    light of the fact that areas near the facility with  the highest projected Pb media concentrations,
24    while having adult residents, do not have any child residents. This resulted in these areas of
25    potentially higher exposures not contributing any risk for this case study.  A sensitivity analysis
26    examining this issue showed that IQ loss estimates for the closest (adult-only) U.S. Census block
27    would range from 2 to 8 points if child residents were included in that block. This finding
            23 Note, that the pilot analysis did not explicitly model paint Pb exposure. However, it is likely that
     modeling of Pb exposure resulting from indoor dust Pb ingestion, which is treated as policy-relevant in the pilot,
     does reflect some degree of paint Pb impact (see Section 4.2.6.5.1). Depending on the degree to which indoor dust
     Pb ingestion does reflect paint Pb, there is the potential that policy-relevant exposures and risk presented in the pilot
     may be over-stated. The issue of paint Pb impacts to indoor dust, and efforts to separate it out and treat it as
     background, is an area that will continue to be researched as part of the full scale analysis.
            December 2006                      4-79         Draft - Do Not Quote or Cite

-------
 1    illustrates the potential impact of population demographics in this assessment.  This factor will
 2    be considered in selecting additional case studies for the full-scale analysis.
 3           Performance evaluation completed for the pilot analysis (described in Section 4.3.6) has
 4    indicated the following with regard to model predictions for Pb concentrations in environmental
 5    media: (a) air modeling results for the primary Pb smelter suggest a potential for moderate
 6    overestimation of levels near the facility and underestimation of values further out, and (b) air
 7    modeling and soil modeling results for the secondary Pb smelter suggest a potentially significant
 8    underestimation of Pb concentrations in both media. Of these, the more significant conclusion
 9    appears to involve the secondary Pb smelter case study. These findings may contribute to an
10    overall low bias in exposure and risk results generated for this case study.  This, combined with
11    the finding regarding child residents in the census block with highest projected media
12    concentrations,  indicates the need to make improvements in our characterization of potential risk
13    associated with intermediate scale point source facilities in the full-scale assessment.
14           The sensitivity analysis indicated that risk estimates for the 90th to 99.9th percentiles
15    could vary by up to several  hundred percent depending on the approach or parameters used for
16    certain aspects of the analysis. Modeling elements identified as having potentially significant
17    impacts on risk results include: (a) characterization of soil Pb levels (for the secondary Pb
18    smelter case study), (b) the blood Pb  model, (c) the blood Pb metric, (d) the GSD for blood Pb
19    distributions (e) the concentration-response function, and (f) the cutpoint for the concentration-
20    response  function. All of these areas, and others discussed previously, will be considered in
21    finalizing our plans for the full-scale assessment.

22          4.5.2  Potential Areas for Enhancement in the Full-Scale Analysis
23           The staff intends that the full-scale risk assessment will provide a quantitative risk
24    characterization to inform consideration of policy options pertaining to policy relevant exposures
25    to ambient Pb.  Our plans for the full-scale analysis will reflect our critical assessment of the
26    pilot and  its results as well as  consideration of comments provided by the public and CAS AC.
27    For the full-scale analysis, we are planning to  include the following:
28          •  Additional case studies: We are considering including additional case studies, e.g.,
29             representative of additional  ambient Pb exposure situations.
30          •  Additional air quality scenarios: We will be considering additional air quality
31             scenarios to inform evaluation of alternate policy options.
             December 2006                      4-80        Draft - Do Not Quote or Cite

-------
 1           In addition, based on experience in the pilot assessment, we are considering the following
 2    potential enhancements for the full-scale assessment:

 3         •  Consideration for alternate exposure periods: We intend to consider the potential
 4            impact on risk results from modeling shorter exposure periods (i.e., periods shorter
 5            than the annual  average values used in the pilot).  Factors to consider in relation to this
 6            issue include response time for key risk-driving media (e.g., soil and indoor dust),
 7            given changes in ambient air Pb levels and the capability of blood Pb models for
 8            tracking shorter-term fluctuations in Pb exposure.

 9         •  Aspects of air dispersion modeling: We intend to review and refine,  as feasible, aspects
10            of the air dispersion modeling step. This will include source characterization, as well as
11            the choice of air dispersion model.
12         •  Characterization ofPb re-suspension: For the near roadway (urban) case study, we are
13            considering the  feasibility of using source apportionment analyses of available PM
14            speciation data to estimate the fraction of airborne Pb associated with re-entrainment of
15            previously deposited Pb.

16         •  Soil Pb modeling: We are considering the use of a compartmental mass balance model
17            (i.e., Total Risk Integrated Methodology Fate, Transport and Ecological exposure
18            model, TRIM.FaTE) (USEPA, 2002b, 2002c) to characterize temporal changes in  soil
19            Pb associated with different Pb emissions and deposition situations.  An additional
20            related consideration is characterization of background soil Pb levels (e.g., alternatives
21            to the 15 mg/kg background soil Pb level used in characterizing the secondary Pb
22            smelter case study).

23         •  Indoor dust Pb modeling: We intend to review and refine,  as feasible, aspects of the
24            indoor dust modeling step. This includes refining  our prediction of policy-relevant
25            source contributions, as differentiated from policy-relevant background sources (e.g.,
26            paint Pb).
27         •  Inhalation absorption estimates: We are considering ways to update the Pb inhalation
28            absorption factors used in the biokinetic blood Pb models with consideration of current
29            information both with regard to Pb particle size distribution associated with exposures
30            for the various case studies, and with regard to respiratory tract Pb particle deposition
31            and absorption.

32         •  Stability in probabilistic population- level exposure modeling: We will reexamine the
33            stability criteria used to establish the number of realizations used in probabilistic
34            exposure modeling for the pilot (10,000) to insure that we  have sufficiently stable high-
35            end results.
            December 2006                     4-81         Draft - Do Not Quote or Cite

-------
  1    REFERENCES

  2
  3    Adgate, J. L.; Willis, R. D.; Buckley, T. I; Chow, J. C; Watson, J. G.; Rhoads, G. G.; Lioy, P. J. (1998) Chemical
  4            mass balance source apportionment of Pb in house dust. Environ. Sci. Technol. 32: 108-114.EC/R
  5            Incorporated. (2006) Secondary Pb Smelter Industry ~ Residual Risk Assessment (draft). Praredred for
  6            USEPA Office of Air and Radiation, Office of Air Quality Planning and Standards, Research Triangle
  7            Park, NC. May.

  8    Agency for Toxic Substances and Disease Registry. (2003) Health Consultation; Blood Lead Results for 2002
  9            Calendar Year; Herculaneum Lead Smelter Site; Herculaneum, Jefferson County, Missouri, EPA facility
10            ID MOD006266373. August. Clay ton et al., 1999

11    Alexander, F.W., Delves, H.T., Clayton, B.E.; (1973).  The Uptake and Excretion by Children of Lead and Other
12            Contaminants. In: Earth, D., Berlin, A., Engel, R., Recht, P., Smeets, J. ets. Environmental Health Aspects
13            of Lead Proceedings, International Symposium: October, 1972, Amsterdam, The Netherlands. Luxemburg:
14            Commission of the European Communities, Center for Information and Documentation. Pp. 319-331.

15    Caravanos, J.; Weiss, A. L.; Jaeger, R. J. (2006) An exterior and interior Pbed dust deposition survey in New York
16            City: results of a 2-year study. Environ. Res. 100: 159-164.

17    Castelle, S.W.,and Tessman, R., (2005) Relative Bioavailability of Lead in House Dust and Soil from the
18            Herculaneum Lead Smelter Site in Herculaneum, Missouri. Prepared for Black and Veatch Special Projects
19            Corporation, May.

20    Chinereje, T., Ma, L.Q., Reeves, M., and Szulczewski, M. (2004) Pb distribution in near-surface soils of two
21            Florida cities: Gainesville and Miami. Geoderma. 119(2): 113-120.

22    Clausing, P., Brunekreef, B., van Wijnen, J.H. (1987). A method for estimating soil ingestion by children.
23            International Archive of Occupational and Environmental Health 59: 73-82.

24    EC/R Incorporated. (2006) Secondary Pb Smelter Industry - Source Characterization for Residual Risk Assessment.
25            Prepared for USEPA Office of Air and Radiation, Office of Air Quality Planning and Standards, Research
26            Triangle Park, NC. November.

27    Filippelli,  G.M., Laidlaw, M.A.S., Latimer, J.C., and Raftis, R. (2005) Urban Pb poisoning and medical geology: An
28            unfinished story.  GSA Today. 15(1):4-11.

29    Gustavsson, N., Bolviken, B., Smith, D.B., and Severson, R.C. (2001) Geochemical Landscapes of the
30            Conterminous United States - New Map Presentations of 22 Elements. U.S. Department of the Interior,
31            U.S. Geological Survey. Professional Paper 1648. Available at http://pubs.usgs.gov/pp/pl648.

32    Hafen, M.R. and Brinkmann, R. (1996) Analysis of Pb  in soils adjacent to an interstate highway in Tampa, Florida.
33            Environmental Geochemistry and Health. 18(4): 171-179.

34    Hattis, D. (2005) Preliminary Analysis of Blood Pb Distributional Data from Various Years of the National Health
3 5            and Nutrition Surveys (NHANES). Prepared for ICF International. October.

36    Henderson, R. (2006) Clean Air Scientific Advisory Committee (CASAC) Lead Review Panel's Consultation on
3 7            EPA's Draft Analysis Plan for Human Health and Ecological Risk Assessment for the Review of the Lead
3 8            National Ambient Air Quality Standards. Memorandum to Stephan Johnson, EPA Administrator, from Dr.
39            Rogene Henderson. July. Available at http://www.epa.gov/sab/pdf/casac-con-06-006.pdf. Singh and
40            Manivic, 1991.
              December 2006                        4-82          Draft - Do Not Quote or Cite

-------
  1    Hilts, S. R. (2003) Effect of smelter emission reductions on children's blood Pb levels. Sci. Total Environ. 303: 51-
  2            58.

  3    ICF International. (2005) Task 2: Near Roadway Concentrations (Revised): A Memorandum to Chad Bailey and
  4            Rich Cook, EPA from Jonathan Cohen and Arlene Rosenbaum. September.

  5    ICF International. (2006). Lead Human Exposure and Health Risk Assessments and Ecological Risk Assessment for
  6            Selected Areas. Pilot Phase. Draft Technical Report. Prepared for the U. S. EPA's Office of Air Quality
  7            Planning and Standards, Research Triangle Park, NC. December.

  8    Kimbrough, D.E. and Suffet, I.H. (1995) Off-site forensic determination of airborne elemental emissions by multi-
  9            media analysis: A case study at two secondary  lead smelters. Environmental Science & Technology.
10            29:2217-2221.

11    Lanphear, B. P., Burgoon, D.A., Rust, S.W. Eberly, S., Galke, AW. (1998) Environmental Exposures to Lead and
12            Urban Children's Blood Lead Levels. Environmental Research, 76: 120-130.

13    Lanphear, B.P., Hornung, R., Khoury, I, Yolton, K., Baghurst, P., Bellinger, D.C., Canfield, R.L., Dietrich, K.N.,
14            Bornschein, R., Greene, T., Rothenberg, S.J., Needleman, H.L., Schnaas, L., Wasserman, G., Graziano, I,
15            and Robe, R. (2005) Low-level environmental  Pb exposure and children's intellectual function: An
16            international pooled analysis. Environmental Health Perspectives. 113(7):894-899.

17    Leggett, R. W. (1993) An age-specific kinetic model of Pb metabolism in humans. Environ. Health Perspect.  101:
18            598-616.

19    Lejano, R.P. and Ericson, J.E. (2005) Tragedy of the temporal commons: Soil-bound Pb and the anachronicity of
20            risk. Journal of Environmental Planning and Management. 48(2):301-320.

21    Maddaloni,M., Bellew,M., Diamond,G., Follansbee,M., Gefell,D., Goodrum,?., Johnson,M., Koporec,K.,
22            Khoury,G., Luey,J., Odin,M., Troast,R., Van,L.P., and Zaragoza,L. (2005) Assessing Lead Risks at Non-
23            Residential Hazardous Waste Sites. Human and Ecological Risk Assessment 11:967-1005.

24    Manton, W.I., Angle, C.R., Stanek, K.L., Reese, Y.R., Kuehnemann, T.J. (2000). Acquisition and Retention of
25            Lead by Young Children. Environmental Research, 82: 60-80.

26    Missouri Department of Natural Resources (MDNR). (2000a) 2000 Revision of the State Implementation Plan for
27            Lead for the Doe Run Resources Corporation Primary Lead Smelter Herculaneum, Missouri, Public
28            Hearing - October 26, 2000. Jefferson City, MO: Air Pollution Control Program, Division of
29            Environmental Quality. October.

30    O'Flaherty, E. J. (1993) Physiologically based models for bone-seeking  elements. IV. Kinetics of Pb disposition in
31            humans. Toxicol. Appl. Pharmacol. 118: 16-29.

32    O'Flaherty, E. J. (1995) Physiologically based models for bone-seeking  elements: V. Pb absorption and disposition
33            in childhood. Toxicol. Appl. Pharmacol. 131: 297-308.

34    O'Flaherty, E. J. (1998) A physiologically based kinetic model for Pb in children and adults. Environ. Health
35            Perspect. 106(supp. 6): 1495-1503.

36    Oliver, D. P.; McLaughlin, M. J.; Naidu, R.; Smith, L. H.; Maynard, E.  J.; Calder, I. C. (1999) Measuring Pb
37            bioavailability from household dusts using an in vitro model. Environ. Sci. Technol.  33: 4434-4439.

38    Small, J.  S., Nunn, A. B., Forslund, B. L., and Daily, D. A. (1995), Source Attribution of Elevated Residential Soil
39            Lead near a Battery Recycling Site. Environmental Science & Technology. 29: 883-895.


              December 2006                        4-83          Draft - Do Not Quote or Cite

-------
  1     Sutherland, R.A., Tolosa, C.A., Tack, F.M.G., and Verloo, M.G. (2000) Characterization of selected element
  2            concentrations and enrichment ratios in background and anthropogenically impacted roadside areas.
  3            Archives of Environmental Contamination and Toxicology. 38:428-438.

  4     Turer, D.G. and Maynard, J.B. (2003) Heavy metal contamination in highway soils. Comparison of Corpus Christi,
  5            Texas and Cincinnati, Ohio shows organic matter is key to mobility. Clean Technologies and
  6            Environmental Policy. 4(4):235-245.

  7     U.S. Agency for Toxic Substances and Disease Registry. (2003) Health Consultation; Blood Lead Results for 2002
  8            Calendar Year; Herculaneum Lead Smelter Site; Herculaneum, Jefferson County, Missouri, EPA facility
  9            ID MOD006266373. August.

10     U.S. Census Bureau. (2005). United States Census 2000: Summary File 1. Public Information Office. Available at
11            http://www.census.gov/Press-Release/www/2001/sumfilel.html.

12     U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention
13            (CDC). 1997. National Health and Nutrition Examination Survey, III 1988-1994.  CD-ROM Series  11, No.
14            1 (July 1997).

15     U.S. Environmental Protection Agency. (1989) Review of National Ambient Air Quality Standard for Pb: Exposure
16            Analysis Methodology and Validation. Research Triangle Park, NC: Office of Air Quality Planning and
17            Standards. EPA-450/2-89-011. June.

18     U.S. Environmental Protection Agency. (1990a) Review of the national ambient air quality standards for Pb:
19            assessment of scientific and technical information. OAQPS staff report. Research Triangle Park, NC:
20            Office of Air Quality Planning and Standards; report no. EPA-450/2-89-022. Available from: NTIS,
21            Springfield, VA; PB89-207914.

22     U.S. Environmental Protection Agency. (1990b) Methodology for Assessing Health Risks  Associated  with Multiple
23            Pathways of Exposure to Combustor Emissions. Office of Health and Environmental Assessment. EPA
24            600/6-90/003.

25     U.S. Environmental Protection Agency. (1993) Addendum to Methodology for Assessing Health Risks Associated
26            with Indirect Exposure to Combustor Emissions. EPA 600/AP-93/003.

27     U.S. Environmental Protection Agency. (1994) Technical Support Document: Parameters and Equations Used in the
28            Integrated Exposure Uptake Biokinetic Model for Pb  in Children (v.099d). EPA 540/R-94/040.

29     U.S. Environmental Protection Agency. (1998) Methodology for Assessing Health Risks Associated with Multiple
30            Pathways of Exposure to Combustor Emissions. Cincinnati, OH: Update to USEPA/600/6-90/003,
31            EPA/NCEA (EPA 600/R-98/137). Available at
32            oaspub.epa.gov/eims/eimscomm.getfile ?p_download_id=427339.

33     U.S. Environmental Protection Agency. (2002a) Children-Specific Exposure Factors Handbook- Interim Draft.
34            Washington, DC: National Center for Environmental  Assessment; Office of Research and Development.
35            EPA-600-P-00-002B.

36     U.S. Environmental Protection Agency. (2002b). Total Risk Integrated Methodology: TRIM.FaTE Technical
37            Support Document, Volume I: Description of Module. EPA-453/R-02-01 la. September 2002. Emissions
3 8            Standards  and Air Quality Strategies and Standards Divisions. Office of Air Quality Planning and
3 9            Standards  - Research Triangle Park, NC.

40     U.S. Environmental Protection Agency. (2002c). Total Risk Integrated Methodology: TRIM.FaTE Technical
41            Support Document, Volume II: Description of Chemical Transport and Transformation Algorithms. EPA-


              December 2006                        4-84         Draft - Do Not Quote or  Cite

-------
  1            453/R-02-01 Ib. September 2002. Emissions Standards and Air Quality Strategies and Standards Divisions.
  2            Office of Air Quality Planning and Standards - Research Triangle Park, NC.

  3    U.S. Environmental Protection Agency. (2004) Human Exposure Measurements: The National Human Exposure
  4            Assessment Survey. U S Environmental Protection Agency. Available from:
  5            http://www.epa.gov/nerl/research/2001/g3-4.html.

  6    U.S. Environmental Protection Agency. (2006a). Analysis Plan for Human Health and Ecological Risk Assessment
  7            For the Review of the Pb National Ambient Air Quality Standards. OAQPS. Research Triangle Park, NC:
  8            Office of Air Quality Planning and Standards. (Plan developed as part of NAAQS review process). June
  9            27th, 2006.

10    U.S. Environmental Protection Agency. (2006b). AirData: Access to Air Pollution Data contained in the Air Quality
11            System database. Available at http://www.epa.gov/air/data.

12    U.S. Environmental Protection Agency. (2006c) 1999 National-Scale Air Toxics Assessment. Available at
13            http://www.epa.gov/ttn/atw/natal999/nsata99.html.

14    U.S. Environmental Protection Agency. (2006d) Human Health Risk Assessment Protocol for Hazardous Waste
15            Combustion Facilities. Office of Solid Waste and Emergency Response. EPA530-R-05-006. September.
16            Available at http://www.epa.gov/epaoswer/hazwaste/combust/risk.htm.

17    U.S. Environmental Protection Agency. (2006e) Lead soil trend analysis through May,  2006. Evaluation by
18            individual quadrant. Herculaneum lead smelter site, Herculaneum, Missouri. Prepared by TetraTech for
19            U.S. EPA, Region 7.

20    U.S. Food and Drug Administration (FDA). 2001. Total Diet Study. U. S. Food and Drug Administration Center for
21            Food Safety and Applied Nutrition Office of Plant and Dairy Foods and Beverages (June 2001). Available
22            online at The U.S. Food and Drug Administration Center for Food Safety and  Applied Nutrition Web site
23            (http://www.cfsan.fda.gov/~comm/tds-toc.html)

24    Von Lindern, I. H.; Spalinger, S. M.; Bero, B. N.; Petrosyan, V.; Von Braun, M. C. (2003a) The influence of soil
25            remediation on Pb in house dust. Sci. Total Environ. 303: 59-78.

26    White, P.O., Van Leeuwen, P., Davis, B.D., Maddaloni, M., Hogan, K.A., and Marcus, A.H. (1998) The conceptual
27            structure of the integrated exposure uptake biokinetic model for Pb in children. Environmental Health
28            Perspectives. 106(86): 1513-1530.

29    Ziegler, E.E., Edwardds, B.B., Jensen, R.L., Mahaffey, R.R., Fomon, J.S. (1978). Absorption and Retention of Lead
30            by Infants. Pediatric Research 12:  29-34.
              December 2006                        4-85          Draft - Do Not Quote or Cite

-------
 1                              5   THE PRIMARY LEAD NAAQS

 2         5.1   INTRODUCTION
 3           This first draft chapter discusses the general approach (Section 5.2) that is intended to be
 4    used in considering the adequacy of the current standard and in identifying policy alternatives for
 5    the next draft of this document.  The current Pb NAAQS and its derivation are summarized in
 6    Section 5.3, and conclusions from the Staff Paper prepared in the last review are presented in
 7    Section 5.4. Key uncertainties and research recommendations related to setting a primary lead
 8    standard will be identified in the next draft of this document.
 9           The current standard is 1.5 ug Pb/m3, as a maximum arithmetic mean averaged over a
10    calendar quarter, set to provide protection to the public, especially children as the particularly
11    sensitive population subgroup, against Pb-induced adverse health effects (43 FR 46246). In
12    identifying options for the Administrator's consideration in this review, we note that the final
13    decision on retaining or revising the current Pb standard is largely a public health policy
14    judgment.  A final decision should draw upon scientific information and analyses about health
15    effects, population exposure and risks, as well as judgments about the appropriate response to the
16    range of uncertainties that are inherent in the scientific evidence and analyses.  Our approach to
17    informing these judgments, discussed more fully below, is based on a recognition that the
18    available health effects evidence generally reflects a continuum consisting of ambient levels at
19    which scientists generally agree that health effects are likely to occur, through lower levels at
20    which the likelihood and magnitude of the response become increasingly uncertain.
21           This approach is consistent with the requirements of the NAAQS provisions of the Act
22    and with how EPA and the courts have historically interpreted the Act. These provisions require
23    the Administrator to establish primary standards that, in the Administrator's judgment, are
24    requisite to protect public health with an adequate margin of safety. In so doing, the
25    Administrator seeks to establish standards that are neither more nor less stringent than necessary
26    for this purpose. The Act does not require that primary standards be set at a zero-risk level  but
27    rather at a level that avoids unacceptable risks to public health, including the health of sensitive
28    groups.

29         5.2   APPROACH
30           As indicated in Chapter 1, the policy assessment to be presented in the final version of
31    this document is intended to inform judgments required by the EPA Administrator in
32    determining whether it is appropriate to retain or revise the NAAQS for Pb. In evaluating
33    whether it is appropriate to consider retaining the current primary Pb standard, or whether
34    consideration of revisions is appropriate, we intend to adopt an approach in this review that
             December 2006                      5-1         Draft - Do Not Quote or Cite

-------
 1    builds upon the general approach used in the initial setting of the standard, as well as in the last
 2    review, and reflects the broader body of evidence now available. As summarized in section 5.3,
 3    the 1978 notice of final rulemaking (43 FR 46246) outlined key factors considered in selecting
 4    the elements of a standard for Pb: the Pb  concentration (i.e., level); the averaging time; and the
 5    form (i.e., the air quality statistic to be used as a basis for determining compliance with the
 6    standard).  Decisions on these elements were based on an integration of information on health
 7    effects associated with exposure to ambient Pb; expert judgment on the adversity of such effects
 8    on individuals; and policy judgments as to when the standard is requisite to protect public health
 9    with an adequate margin of safety, which were informed by air quality and related analyses,
10    quantitative exposure and risk assessments when possible,  and qualitative assessment of impacts
11    that could not be quantified.
12          In developing conclusions and identifying options for the Pb standard in this review, staff
13    intends to take into account both evidence-based and quantitative exposure- and risk-based
14    considerations. A series of general  questions will frame our approach to reaching conclusions
15    and identifying options for consideration by the Administrator in deciding whether to retain or
16    revise the current primary Pb standard. Examples of questions that we intend to address in our
17    review include the following:
18         •  To what extent does newly available information reinforce or call into question
19            evidence of associations with effects identified in the last review?
20         •  To what extent has evidence of new effects and/or sensitive populations become
21            available since the last review?
22         •  To what extent have important  uncertainties identified in the  last review been reduced
23            and have new uncertainties emerged?
24         •  To what extent does newly available information reinforce or call into question any of
25            the basic elements of the current standard?
26    To the extent that the available information suggests that revision of the current standard may be
27    appropriate to consider, we intend to also address whether the currently available information
28    supports consideration of a standard that  is either more or less protective by addressing questions
29    such as the following:
30         •   Is there evidence that associations, especially likely causal associations, extend to air
31            quality levels that are as low as or lower than had previously  been observed, and what
32            are the important uncertainties  associated with that evidence?
33         •   Are exposures of concern and health risks estimated to occur in areas that meet the
34            current standard; are they important from a public health perspective; and what are the
35            important uncertainties associated with the estimated risks?
            December 2006                      5-2        Draft - Do Not Quote or Cite

-------
 1    To the extent that there is support for consideration of a revised standard, we will then identify
 2    ranges of standards (in terms of an indicator, averaging time, level, and form) that would reflect
 3    a range of alternative public health policy judgments, based on the currently available
 4    information, as to the degree of protection that is requisite to protect public health with an
 5    adequate margin of safety. In so doing, we would address the following questions:
 6         •  Does the evidence provide support for considering a different Pb indicator?
 7         •  Does the evidence provide support for considering different averaging times?
 8         •  What ranges  of levels and forms of alternative standards are supported by the evidence,
 9            and what are the uncertainties and limitations in that evidence?
10         •  To what extent do specific levels and forms of alternative standards reduce the
11            estimated exposures of concern and risks attributable to Pb, and what are the
12            uncertainties associated with the estimated exposure and risk reductions?
13
14          As noted in Chapter 1, staff will also evaluate removing Pb from the criteria pollutant list
15    and assess whether revocation of the Pb NAAQS is an option appropriate for the Administrator
16    to consider.  Section 108 of the Clean Air Act states that the Administrator "shall, from time to
17    time ... revise a list which includes each pollutant -
18          (A) Emissions of which, in his judgment, cause or contribute to air pollution which may
19          reasonably be anticipated to endanger public health or welfare;
20          (B) The presence of which in the ambient air results from numerous or diverse mobile or
21          stationary sources; and
22          (C) For which air quality criteria had not been issued before December 31, 1970, but for
23          which he plans to issue air quality criteria under this section."
24    In evaluating such an option, staff expects to consider, among other things, many of the same
25    issues identified earlier in the section. Information about the kinds and types of sources of Pb
26    emissions, as well as the quantities of emissions from those sources will also be important for
27    consideration.

28         5.3   PRIMARY LEAD STANDARD
29          As mentioned earlier, the current primary Pb NAAQS was promulgated in 1978.  The
30    basis for its establishment is described below.

31         5.3.1   Level
32          The level of the current NAAQS is 1.5 ug/m3. EPA's objective in setting the level of the
33    current standard was "to estimate the concentration of lead in the air to which all groups within
34    the general population  can be exposed for protracted periods without an unacceptable risk to

            December 2006                     5-3         Draft - Do Not Quote or Cite

-------
 1    health" (43 FR 46252).  Consistent with section 109 of the Clean Air Act, the Agency identified
 2    a level for the current standard that was not considered to be at the threshold for adverse health
 3    effects, but was at a lower level in order to provide a margin of safety (see Section 5.3.1.5).  As
 4    stated in the notice of final rulemaking (and further described in the following subsections),
 5          "This estimate was based on EPA's judgment in four key areas:
 6          (1) Determining the 'sensitive population' as that group within the general population
 7              which has the lowest threshold for adverse effects or greatest potential for exposure.
 8              EPA concludes that young children, aged 1 to 5, are the sensitive population.
 9          (2) Determining the safe level of total lead exposure for the sensitive population,
10              indicated by the concentration of lead in the blood. EPA concludes that the
11              maximum safe level of blood lead for an individual child is 30 ug Pb/dl and that
12              population blood lead, measured as the geometric mean, must be 15 ug Pb/dl in order
13              to place 99.5 percent of children in the United States below 30 ug Pb/dl.
14          (3) Attributing the contribution to blood lead from nonair pollution sources. EPA
15              concludes that 12 ug Pb/dl of population blood lead for children should be attributed
16              to nonair exposure.
17          (4) Determining the air lead level which is consistent with maintaining the mean
18              population blood lead level at 15 ug Pb/dl [the safe level]. Taking into account
19              exposure from other sources  (12 ug Pb/dl), EPA has designed the standard to limit air
20              contribution after achieving the standard to 3  ug Pb/dl. On the basis of an estimated
21              relationship of air lead to blood lead of 1 to 2, EPA concludes that the ambient air
22              standard should be 1.5 ug Pb/m3." (43 FR 46252)

23         5.3.1.1  Sensitive Population
24          The assessment of the science that was presented in the 1977 CD (USEPA,  1977),
25    indicated young children, aged 1 to 5, as the population group at particular risk from Pb
26    exposure.  Children were recognized to have a greater physiological sensitivity than adults to the
27    effects of Pb and a greater exposure. In  identifying young children as the sensitive population,
28    EPA also recognized the occurrence of subgroups with enhanced risk due to genetic factors,
29    dietary deficiencies or residence in urban areas.  Yet information was not available to estimate a
30    threshold for adverse effects for these subgroups separate from that of all young children.
31    Additionally, EPA recognized both a concern regarding potential risk to pregnant women and
32    fetuses, and a lack of information to establish that these subgroups are more at risk than young
33    children.  Accordingly, young children, aged 1 to 5, were identified as the group which has the
34    lowest threshold for adverse effects of greatest potential for exposure (i.e., the sensitive
3 5    population) (43 FR 46252).

            December  2006                     5-4         Draft - Do Not Quote or Cite

-------
 1         5.3.1.2  Maximum Safe Blood Level
 2          In identifying the maximum safe exposure, EPA relied upon the measurement of Pb in
 3    blood (43 FR 46252-46253).  The physiological effect of Pb that had been identified as occurring
 4    at the lowest blood Pb level, was inhibition of an enzyme integral to the pathway by which heme
 5    (the oxygen carrying protein of human blood) is synthesized, i.e., delta-aminolevulinic acid
 6    dehydratase (5-ALAD). The  1977 CD reported a threshold for inhibition of this enzyme in
 7    children at 10 ug Pb/dL.  The 1977 CD also reported a threshold of 15-20 ug/dL for elevation of
 8    protoporphyrin (EP), which is an indication of some disruption of the heme synthesis pathway.
 9    EPA concluded that this effect on the heme synthesis pathway (indicated by EP) was potentially
10    adverse. EPA further described a range of blood levels associated with a progression in
11    detrimental impact on the heme synthesis pathway. At the low end of the range (15-20 ug/dL),
12    the initial detection of EP associated with blood Pb, was not concluded to be associated with a
13    significant risk to health. The upper end of the range (40 ug/dL), the threshold associated with
14    clear evidence of heme synthesis impairment and other effects contributing to clinical symptoms
15    of anemia, was regarded as clearly adverse to health.  EPA also recognized the existence of
16    thresholds for additional adverse effects (e.g., nervous system deficits) occurring for some
17    children at just slightly higher blood Pb levels (e.g., 50 ug/dL). Additionally, EPA stated  that the
18    maximum safe blood level should not be higher than the blood  Pb level recognized by the CDC
19    as "elevated" (and indicative of the need for intervention).  In 1978, that level was 30 ug/dL1.
20          Once identifying the maximum safe blood level in individual children, EPA next made
21    the policy-based judgment regarding the target  mean blood level for the U. S. population of
22    young children (43 FR 46252-46253). With this judgment, EPA identified a target of 99.5
23    percent of this population to be brought below the maximum safe blood Pb level. This judgment
24    was based on consideration of the size of the sensitive subpopulation, and the recognition  that
25    there are special high risk groups of children within the general population. The population
26    statistics available at the time (the 1970 U.S. Census) indicated a total of 20 million children
27    younger than 5 years of age, with 15 million residing in urban areas and 5 million in center cities
28    where Pb exposure was thought likely to be "high". Concern about these high risk groups
29    influenced EPA's determination of 99.5%, deterring EPA from selecting a population percentage
30    lower than 99.5 (43 FR 46253). EPA then used standard statistical techniques to calculate the
31    population mean that would place 99.5 percent of the population below the maximum safe level.
32    Based on the then available data, EPA concluded that the blood Pb levels in the population of
            1 The CDC subsequently revised their advisory level for children's blood Pb to 25 ug/dL in 1985, and to 10
      ug/dL 1991. More details on this level are provided in Section 3.2.
            December 2006                     5-5        Draft - Do Not Quote or Cite

-------
 1    U.S. children were normally distributed with a geometric standard deviation of 1.3.  Based on
 2    standard statistical techniques, a thus described population in which 99.5 percent of the
 3    population has blood Pb levels below 30 ug/dL has a geometric mean blood level of 15 ug/dL.

 4         5.3.1.3  Nonair Contribution
 5          When setting the current NAAQS, EPA recognized that the air standard needed to take
 6    into account the contribution to blood Pb levels from Pb sources unrelated to air pollution.
 7    Consequently, the calculation of the current NAAQS included the subtraction of Pb contributed
 8    to blood Pb from nonair sources from the estimate of a safe mean population blood Pb level.
 9    Without this subtraction, EPA recognized that the combined exposure to Pb from air and nonair
10    sources would result in a blood Pb concentration exceeding the safe level (43 FR 46253).
11          In developing an estimate of this nonair contribution, EPA recognized the lack of detailed
12    or widespread information about the relative contribution of various sources to children's blood
13    levels, such that an estimate could only be made by inference from other empirical or theoretical
14    studies, often involving adults.  Additionally, EPA recognized the expectation that the
15    contribution to blood Pb levels from nonair sources would vary widely, was probably not in
16    constant proportion to air Pb contribution, and in some cases may alone exceed the target mean
17    population blood Pb level (43 FR 46253-46254).
18          The amount of blood Pb attributed to nonair sources was selected based primarily on
19    findings in studies of blood Pb levels in areas where air levels were low relative to other
20    locations in U.S. The air levels in these areas ranged from 0.1 to 0.7 ug/m3. The average of the
21    reported blood levels for children of various ages in these areas was on the order of 12 ug/dL.
22    So 12 ug/dL was identified as the nonair contribution, and subtracted from the population mean
23    target level of 15 ug/dL to yield a value of 3 ug/dL as the limit on the air contribution to blood
24    Pb.

25         5.3.1.4  Air Pb Level
26          In determining the air Pb level consistent with an air contribution of 3 ug Pb/dL, EPA
27    reviewed studies assessed in the 1977 CD that reported changes in blood Pb with different air Pb
28    levels. These studies included a study of children exposed to Pb from a primary Pb smelter,
29    controlled exposures of adult men to Pb in fine particulate matter, and a personal exposure study
30    involving several male cohorts exposed to Pb in a large urban area in the early 1970s (43 FR
31    46254).  Using all three studies, EPA calculated an average slope or ratio over the entire range of
32    data. That value was  1.95 (rounded to 2 ug ML blood Pb concentration to 1 ug /m3 air
33    concentration), and is recognized to fall within the range of values reported in the 1977 CD.  On
34    the basis of this  2 to 1 relationship, EPA concluded that the ambient air standard should be 1.5
35    ug Pb/m3 (43 FR 46254).
            December 2006                     5-6         Draft - Do Not Quote or Cite

-------
 1         5.3.1.5  Margin of Safety
 2          In consideration of the appropriate margin of safety during the development of the
 3    current NAAQS, EPA identified the following factors: (1) the 1977 CD reported multiple
 4    biological effects of Pb in practically all cell types, tissues and organ systems, of which the
 5    significance for health had not yet been fully studied; (2) no beneficial effects of Pb at then
 6    current environmental levels were recognized; (3) data were incomplete as to the extent to which
 7    children are indirectly exposed to air Pb that has moved to other environmental media, such as
 8    water, soil and dirt, and food; (4) Pb is chemically persistent and with continued uncontrolled
 9    emissions would continue to accumulate in human tissue and the environment; and (5) the
10    possibility that exposure associated with blood Pb levels previously considered safe might
11    influence neurological development and learning abilities of the young child (43 FR 46255).
12    Recognizing that estimating an appropriate margin of safety for the air Pb standard was
13    complicated by the multiple sources and media involved in Pb exposure, EPA chose to use
14    margin of safety considerations principally in establishing a maximum safe blood Pb level for
15    individual children (30 ug Pb/dL) and in determining the percentage of children to be placed
16    below this maximum  level (about 99.5). Additionally, in establishing other factors used in
17    calculating the standard, EPA used margin of safety in the sense of making careful judgment
18    based on available data, but these judgments were not considered to be at the precautionary
19    extreme of the range of data available at the time (43 FR 46251).
20          EPA further recognized that because of the variability between individuals in a
21    population experiencing a given level of Pb exposure, it was considered impossible  to provide
22    the same size margin  of safety for all members in the sensitive population or to define the margin
23    of safety in the standard as a simple percentage.  EPA believed that the factors it used in
24    designing the standards provided an adequate margin of safety for a large proportion of the
25    sensitive population.  The Agency did not believe that the margin was excessively large or on the
26    other hand that the air standard could protect everyone from elevated blood Pb levels (43 FR
27    46251).

28         5.3.2   Averaging Time, Form, and Indicator
29          The averaging time for the current standard is a calendar quarter. In the decision for this
30    aspect of the standard, the Agency also considered a monthly averaging period, but  concluded
31    that "a requirement for the averaging of air quality data over calendar quarter will improve the
32    validity of air quality  data gathered without a significant reduction  in the protectiveness of the
33    standards." As described in the notice for this decision (43 FR 46250),  this conclusion was
34    based on several points, including the following:


            December 2006                     5-7         Draft - Do Not Quote or Cite

-------
 1         •  An analysis of ambient measurements available at the time indicated that the
 2            distribution of air Pb levels was such that there was little possibility that there could be
 3            sustained periods greatly above the average value in situations where the quarterly
 4            standard was achieved.
 5         •  A recognition that the monitoring network may not actually represent the exposure
 6            situation for young children, such that it seemed likely that elevated air Pb levels when
 7            occurring would be close to Pb air pollution sources where young children would
 8            typically not encounter them for the full 24-hour period reported by the monitor.
 9         •  Medical evidence available at the time indicated that blood Pb levels re-equilibrate
10            slowly to changes in air exposure,  a finding that would serve to dampen the impact of
11            short-term period of exposure to elevated air Pb.
12         •  Direct exposure to air is only one of several routes of total exposure, thus lessening the
13            impact of a change in air Pb on blood Pb levels.
14
15          The statistical form of the current standard is as a not-to-be-exceeded or maximum value.
16    EPA set the standard as a ceiling value with the conclusion that this air level would be safe for
17    indefinite exposure for young children (43 FR 46250).
18          The indicator is total airborne Pb collected by high volume sampler (43 FR 46258).
19    EPA's selection of total suspended particulate Pb  as the indicator for the standard was based on
20    explicit recognition both  of the significance  of ingestion as an exposure pathway for Pb that had
21    deposited from the air and of the potential for Pb deposited from the air to become re-suspended
22    in respirable size particles in the air and available  for human inhalation exposure.  As stated in
23    the final rule, "a significant component of exposure can be ingestion of materials contaminated
24    by deposition of lead from the air", and that, "in addition to the indirect route  of ingestion and
25    absorption from the gastrointestinal tract, non-respirable Pb in the environment may, at some
26    point become respirable through weathering or mechanical action" (43 FR 46251).

27         5.4    POLICY OPTIONS CONSIDERED IN THE LAST REVIEW
28          During the 1980s, EPA initiated a review of the air quality criteria and NAAQS for Pb.
29    CAS AC and the public were fully involved in this review, which led to the publication of a
30    criteria document with associated addendum and a supplement (USEPA,  1986a, 1986b, 1990a),
31    an exposure analysis methods document (USEPA, 1989) and a staff paper (USEPA,  1990b).
32          Total emissions to air were estimated to have dropped by 94 percent between 1978 and
33    1987, with the vast majority of it attributed to the  reduction of Pb in gasoline. Accordingly, the
34    focus of this review was on areas near stationary sources of Pb emissions. Although such
35    sources were not considered to have made a significant contribution (as compared to Pb in
36    gasoline) to the overall Pb pollution across large, urban or regional areas, Pb emissions from
37    such sources were considered to have the potential for a significant impact on a local scale. Air,
            December 2006                     5-8        Draft - Do Not Quote or Cite

-------
 1    and especially soil and dust Pb concentrations had been associated with elevated levels of Pb
 2    absorption in children and adults in numerous Pb point source community studies. Exceedances
 3    of the current NAAQS were found at that time only in the vicinity of nonferrous smelters or
 4    other point sources of Pb.
 5           In summarizing and interpreting the health evidence presented in the 1986 CD and
 6    associated documents, the 1990  Staff Paper described the collective impact on children of the
 7    effects at blood Pb levels above  15 ug/dL as representing a clear pattern of adverse effects
 8    worthy of avoiding.  This is in contrast to EPA's identification of 30 ug/dL as a safe blood Pb
 9    level for individual children when the NAAQS was set in 1978.  The staff paper further stated
10    that at levels of 10-15 ug/dL, there was a convergence of evidence of Pb-induced interference
11    with a diverse set of physiological functions and processes, particularly evident in several
12    independent studies showing impaired neurobehavioral function and development. Further, the
13    available data did not indicate a  clear threshold in this blood Pb range.  Rather, it suggested a
14    continuum of health risks down  to the lowest levels measured.2
15           For the purposes  of comparing the relative protectiveness of alternative Pb NAAQS, the
16    staff conducted analyses to estimate the percentages of children with blood Pb levels above 10
17    ug/dL and above 15 ug/dL for several air quality scenarios developed for a small  set of
18    stationary source exposure case  studies.  These analyses omitted young children, whom it was
19    considered could not be substantially affected by any changes in atmospheric Pb emissions under
20    different standards (e.g.,  those with excessive pica3 and/or living in overtly deteriorated Pb-paint
21    homes). The results of the analyses of children populations living near two Pb smelters indicated
22    that substantial reductions in  Pb  exposure could be achieved through attainment of the current Pb
23    NAAQS. According to the best  estimate analyses,  over 99.5% of children living in areas
24    significantly affected by  the smelters would have blood Pb levels below 15 ug/dL if the current
25    standard was achieved. Progressive changes  in this number were estimated for the alternative
26    monthly Pb NAAQS levels evaluated, ranging from 1.5 ug/m3 to 0.5 ug/m3.
27           The staff paper, in light of the health effects evidence available at the time, in addition to
28    air quality, exposure and risk analyses, and other policy considerations, presented the following
29    staff conclusions with regard to the primary Pb NAAQS (USEPA, 1990b, pp. xii to xiv):
30           1)  "The range of standards ... should be from 0.5 to 1.5 ug/m3."
              In 1991, the CDC reduced their advisory level for children's blood Pb from 25 ug/dL to 10 ug/dL.
             3 Pica is an eating disorder typically defined by persistent cravings to eat non-food items.
             December 2006                      5-9        Draft - Do Not Quote or Cite

-------
 1          2)  "A monthly averaging period would better capture short-term increases in lead
 2              exposure and would more fully protect children's health than the current quarterly
 3              average."
 4          3)  "The most appropriate form of the standard appears to be the second highest monthly
 5              averages {sic} in a 3-year span. This form would be nearly as stringent as a form that
 6              does not permit any exceedances and allows for discounting of one "bad" month in 3
 7              years which may be caused, for example, by unusual meteorology."
 8          4)  "With a revision to a monthly averaging time more frequent sampling is needed,
 9              except in areas, like roadways  remote from lead point sources, where the standard is
10              not expected to be violated. In those situations, the current l-in-6 day sampling
11              schedule would sufficiently reflect air quality and trends."
12          5)  "Because exposure to atmospheric lead particles occurs not only via  direct inhalation,
13              but via ingestion of deposited particles as well, especially among young children, the
14              hi-volume  sampler provides a reasonable indicator for determining compliance with a
15              monthly standard and should be retained as the instrument to monitor compliance
16              with the lead NAAQS until more refined instruments can be developed."
17
18          After consideration of the documents developed during the review, EPA chose not to
19    propose revision of the NAAQS for Pb. During the same time period, the Agency published and
20    embarked on the implementation of a broad, multi-program, multi-media, integrated national
21    strategy to reduce Pb exposures (USEPA,  1991). As part of implementing this strategy, the
22    Agency focused efforts primarily on regulatory and remedial clean-up actions aimed at reducing
23    Pb exposures from a variety of non-air sources judged to pose more extensive public health risks
24    to U.S. populations, as well as on actions to reduce Pb emissions to air, particularly near
25    stationary sources. EPA established standards for Pb-based paint hazards and Pb dust cleanup
26    levels in most pre-1978 housing and child-occupied facilities.  Additionally, EPA has developed
27    standards for the management of Pb in solid and hazardous waste, oversees the cleanup  of Pb
28    contamination at  Superfund sites, and has  issued regulations to reduce Pb in drinking water
29    (http://www.epa.gov/lead/regulation.htm). Beyond these specific regulatory actions, the
30    Agency's Lead Awareness Program has continued to work to protect human health and  the
31    environment against the dangers  of Pb by  conducting research and designing educational
32    outreach activities and materials (http://www.epa.gov/lead/).  Actions to reduce Pb emissions to
33    air during the 1990s included enforcement of the NAAQS, as well as the promulgation of
34    regulations under Section  112 of the Clean Air Act, including national emissions standards for
35    hazardous air pollutants at primary and secondary Pb smelters, as well as other Pb sources.

            December 2006                     5-10        Draft - Do Not Quote or Cite

-------
  1    REFERENCES

  2
  3    U.S. Environmental Protection Agency. (1977) Air quality criteria for lead. Office of Research and Development.
  4            Washington, D.C. 20460. EPA-450/8-77-017. December.

  5    U.S. Environmental Protection Agency. (1986a) Air quality criteria for lead. Research Triangle Park, NC: Office of
  6            Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
  7            EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.  Available on the
  8            web: http://cfpub2.epa.gov/ncea/cfm/recordisplav.cfm?deid=32647

  9    U.S. Environmental Protection Agency. (1986b) Lead effects on cardiovascular function, early development, and
10            stature: an addendum to U.S. EPA Air Quality Criteria for Lead (1986). In: Air quality criteria for lead, v.
11            1. Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria
12            and Assessment Office; pp. A1-A67; EPA report no. EPA-600/8-83/028aF. Available from: NTIS,
13            Springfield, VA; PB87-142378.

14    U.S. Environmental Protection Agency. (1989) Review of the national ambient air quality standards for lead:
15            Exposure analysis methodology and validation: OAQPS staff report. Research Triangle Park, NC: Office of
16            Air Quality Planning and Standards; report no. EPA-450/2-89/011. Available on the web:
17            http://www.epa.gov/ttn/naaqs/standards/pb/sjbjr  td.html

18    U.S. Environmental Protection Agency. (1990a) Air quality criteria for lead: supplement to the 1986 addendum.
19            Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
20            Assessment Office; report no. EPA/600/8-89/049F. Available from: NTIS, Springfield, VA; PB91-138420.
21            Available on the web: http://cfpub2.epa.gov/ncea/cfm/recordisplav.cfm?deid=45189.

22    U.S. Environmental Protection Agency. (1990b) Review of the national ambient air quality standards for lead:
23            assessment of scientific and technical information: OAQPS staff paper. Research Triangle Park, NC: Office
24            of Air Quality Planning and Standards; report no. EPA-450/2-89/022. Available from: NTIS, Springfield,
25            VA; PB91-206185.  Available on the web: http://www.epa.gov/ttn/naaqs/standards/pb/data/rnaaqsl asti.pdf

26    U.S. Environmental Protection Agency. (1991) U.S. EPA Strategy for Reducing Lead Exposure.  Available from
27            U.S. EPA Headquarters Library/Washington, D.C. (Library Code EJBD; Item Call Number:  EAP
28            100/1991.6; OCLC Number 2346675).
              December 2006                        5-11         Draft - Do Not Quote or Cite

-------
 1           6   POLICY RELEVANT ASSESSMENT OF WELFARE EFFECTS

 2         6.1   INTRODUCTION
 3          This chapter presents information in support of the review of the secondary NAAQS for
 4    lead (Pb). Welfare effects addressed by the secondary NAAQS include, but are not limited to,
 5    effects on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather,
 6    visibility and climate, damage to and deterioration of property, and hazards to transportation, as
 7    well as effects on economic values and on personal comfort and well-being. Lead is persistent in
 8    the environment and accumulates in soils and sediments thereby providing long term exposures
 9    to organisms and ecosystems. Past emissions of Pb from the use of Pb additives in automobile
10    fuel significantly contributed to the widespread increase of Pb concentrations in the environment,
11    a portion of which remains today.
12          This chapter includes a summary of policy relevant information presented in the CD, with
13    effects of Pb in terrestrial ecosystems discussed in Section 6.2 and aquatic ecosystems discussed
14    in Section 6.3.  For some criteria pollutants, key effects and concentration responses are much
15    more fully understood.  For Pb, it is difficult to generalize effects due to the nature of the data
16    and the general lack of community or population level information on the effects of Pb.
17    Therefore, this chapter attempts to describe the effects of Pb on ecosystems by grouping known
18    effects into categories of organisms and summarizing the limited anecdotal information that is
19    available for broader ecosystem effects of Pb. Sections 6.4 and 6.5 describe the screening level
20    analyses that were conducted in this assessment in support of the current NAAQS review.  These
21    analyses are intended to identify areas of exposure for which there is the potential for adverse
22    effects from Pb and could be used to focus further analyses on those areas. At this time, we do
23    not anticipate having funding to perform additional ecological risk assessment work for this
24    review. That is, the focus for this review with regard to the secondary standard will be on what
25    we have learned from this pilot phase, in addition to the science assessment in the criteria
26    document.

27         6.2   EFFECTS IN TERRESTRIAL ECOSYSTEMS
28          Ecosystems near smelters, mines and other industrial sources of Pb have demonstrated a
29    wide variety of adverse effects including decreases in species diversity, loss of vegetation,
30    changes to community composition, decreased growth of vegetation, and increased number of
31    invasive species. Apportioning these effects between Pb and other stressors is problematic  since
32    these point sources also emit a wide variety of other heavy metals as well as 862 which may
33    cause toxic effects.  There are no field studies which have investigated effects of Pb additions
34    alone but some studies near large point sources of Pb have found significantly reduced species
35    composition and altered community structures. While these effects are significant, they are
            December 2006                      6-1         Draft - Do Not Quote or Cite

-------
 1    spatially limited: the majority of contamination occurs within 20 to 50 km of the emission source
 2    (CD, AX7.1.4.2).
 3          By far, the majority of Pb found in terrestrial ecosystems was deposited from the long
 4    range transport of Pb including Pb additives used in gasoline in the past few decades. There is
 5    little evidence that sites exposed to long range transport of Pb have experienced significant
 6    effects on ecosystem structure or function (CD, AX7.1.4.2). Studies have shown decreasing
 7    levels of Pb in vegetation which seems to correlate with decreases in atmospheric deposition of
 8    Pb resulting from the removal of Pb additives to gasoline (CD, AX 7.1.4.2). Little work,
 9    however, has been done on the effect of residual long term, low-level metal concentration on
10    species diversity.
11          As stated in the CD (Section 7.1), terrestrial ecosystems remain primarily sinks for Pb but
12    amounts retained in various soil layers vary based on forest type, climate, and litter cycling.
13    Once in the soil, the migration and distribution of Pb is controlled by a multitude of factors
14    including pH, precipitation, litter composition, and other factors which govern the rate at which
15    Pb is bound to organic materials in the soil (CD, Section 2.3.5).
16          Like most metals the solubility of Pb is increased at lower pH. However, the reduction of
17    pH may in turn decrease the solubility of dissolved organic material (DOM). Given the close
18    association between Pb mobility and complexation with DOM, a reduced pH does not
19    necessarily lead to increased movement of Pb through terrestrial systems and into surface waters.
20    Studies have shown that in areas with  moderately acidic soil (i.e., pH of 4.5 to 5.5) and abundant
21    DOM, there is no appreciable increase in the movement of Pb into surface waters compared to
22    those areas with neutral soils (i.e., pH  of approximately 7.0).  This appears to support the theory
23    that the movement of Pb in soils is limited by the solubilization and transport of DOM.  In sandy
24    soils without abundant DOM, moderate acidification appears likely to increase outputs of Pb to
25    surface waters (CD, AX 7.1.4.1).
26          Forest harvesting and management practices have significant and lasting effects on
27    organic matter cycling in forest ecosystems. Clear cutting, as well as other methods of tree
28    removal, leads to decreased organic matter for several  years after harvesting and organic matter
29    remaining in soils is exposed to higher temperatures and moisture which tend to increase rates of
30    decomposition.  Despite these effects,  studies have shown very little to no mobilization of Pb
31    from soils to surface waters following clear cutting. On possible explanation for this is that
32    mineral soils (those below the biologically active, organic layer of soil) are efficient in capturing
33    and retaining mobilized Pb. Loss of Pb in particulate form due to runoff and erosion in clear cut
34    areas remains a potential source of Pb  to surface waters.
35          As described in Chapter 2 (Sections 2.5 to 2.7) and in the CD (Chapter 7 and the Chapter
36    7 Annex), Pb  emitted anthropogenically into the atmosphere accumulates in surface soils and
37    vegetation throughout the Unites States as a result of wet and dry deposition. The following
            December 2006                      6-2         Draft - Do Not Quote or Cite

-------
 1    discussion relies heavily on information presented in Chapters 2, 7, 8 of the CD and the Chapter
 2    7 Annex of the CD.

 3         6.2.1  Pathways of Exposure
 4          The main pathways of exposure to Pb for animals are inhalation and ingestion.
 5    Inhalation exposures, which would be limited to areas immediately surrounding point sources,
 6    are not thought to be common and little information is available about inhalation in wildlife.
 7    Ingestion constitutes the main pathway of exposure for most organisms whether by incidental
 8    ingestion or prey contamination. For higher organisms which may ingest either contaminated
 9    plants or soils/sediments, the form and species of Pb ingested influences uptake and toxicity as
10    does the presence of other heavy metals. The relative toxicity of metal mixtures and their affects
11    on Pb toxicity is complex and varies greatly between species and metal.
12          For plants, direct deposition onto surfaces and uptake of dissolved Pb by roots is the main
13    exposure route (CD, Section 7.1.3). While the migration and biological uptake of Pb in
14    ecosystems is relatively low compared to other metals, there are many factors which may affect
15    the mobility of Pb, including elevation and climate, vegetation type, acidity, and soil
16    composition.  The bioavailability and accessibility of Pb to plants is determined largely by the
17    soil  pH, chemical form of Pb, presence of other metals, and source of the Pb in the ecosystem.
18    Low pH soils enhance bioavailability to plants and Pb chlorides and acetates are more
19    bioavailable than Pb oxides.  These factors directly relate to the ability of Pb complexes to enter
20    pore water in soils and sediments and thereby enter root tissues.

21         6.2.2  Effects of Lead on Energy Flow and Biogeocycling
22          Lead in soils and leaf litter can have a significant adverse effect on energy flow in
23    terrestrial ecosystems through reducing the rate of litter decomposition and by decreasing
24    photosynthetic rates in plants, both of which alter the ecosystem carbon cycling and may reduce
25    the ability of trees and other plants to obtain nutrients from the soil (CD, AX7.1.4.3). Recent
26    studies have associated high Pb concentrations  in soils, such as those found near point sources,
27    with reduced fungal and bacterial activity. This can lead to interruptions in various metabolic
28    pathways by either reducing symbiotic relationships between the roots of some types of plants
29    and  fungi and/or bacteria or by tying up nutrients needed for plant growth (CD, AX7.1.4.3).
30          In less contaminated areas removed from point sources, there is little evidence that Pb
31    represents a threat to energy flow or carbon cycling or that large pulses of Pb are likely to enter
32    surface waters. Recent studies have shown that atmospheric deposition of Pb has decreased
33    dramatically (>95%) over the last three decades and residence times in soils (the time for Pb to
34    move out of the biologically active layers of soil) range from about 60 years in deciduous forests
35    to 150 years in coniferous stands (CD, AX7.1.2.2).
36
            December 2006                     6-3         Draft - Do Not Quote or Cite

-------
 1         6.2.3   Tools for Identifying Ecotoxicity in Terrestrial Organisms
 2          In recognition of a need by EPA's Superfund Program to identify the potential for
 3    adverse effect from various pollutants in soils to ecosystems, a multi-stakeholder group,
 4    consisting of federal, state, consulting, industry, and academic participants developed Ecological
 5    Soil Screening Levels (Eco-SSLs) for various pollutants including Pb. Eco-SSLs describe the
 6    concentrations of contaminants in soils that would result in little or no measurable effect on
 7    ecological receptors (USEPA, 2005a). They are intentionally conservative in order to provide
 8    confidence that contaminants, which could present an unacceptable risk, are not screened out
 9    early in the evaluation process (intended to be a specific site under consideration of the
10    Superfund Program). That is, at or below these levels, adverse effects are  considered unlikely.
11    These values are defined in the Ecological Soil Screening Levels for Lead (USEPA, 2005a) as
12    "concentrations of contaminants in soil that are protective of ecological receptors that commonly
13    come into contact with soil or ingest biota that live in or on soil." They were derived separately
14    for four general categories of ecological receptors: plants, soil invertebrates, birds, and
15    mammals.
16           In the case of plants and soil invertebrates, Eco-SSLs are expressed as concentration of
17    Pb in soil (mg Pb /kg soil) and were developed with consideration of characteristics affecting
18    bioavailability (e.g., pH, organic content, etc).  The development of Eco-SSLs for avian and
19    mammalian wildlife involved a two step process:  1) derivation of a toxicity reference value
20    (TRV) in mg contaminant per kg body weight per day from available literature, and 2)
21    application of the TRV with information on soil intake, foraging habits, diet, contaminant uptake
22    by prey for a single species to derive an Eco-SSL in mg Pb per kg soil. In general for avian and
23    mammalian wildlife categories, a single TRV was developed (e.g., the reference dose for the
24    most sensitive of the adverse ecological effects on birds) for all species in each category.
25    However, default assumptions regarding incidental soil ingestion, foraging techniques,
26    contaminant intake by prey,  and overall diet composition generally resulted in different Eco-SSL
27    values, expressed as soil concentrations, for the different species in each receptor category. The
28    receptor category Eco-SSL was then  set equal to the lowest species-specific Eco-SSL (USEPA,
29    2005a; ICF, 2006).  The Eco-SSLs for Pb, as developed by EPA Superfund Program, for
30    terrestrial plants, birds, mammals, and soil invertebrates are 120 mg/kg, 11 mg/kg, 56 mg/kg and
31    1700 mg/kg, respectively. Section 2.6.2.2 discusses current concentrations of Pb in soils.
32    Values range from 40 to 100 mg Pb/kg soil in remote forests where historic deposition of Pb
33    from gasoline would be presumed to  be the major source to hundreds to tens of thousands of
34    mg/kg near point sources.
35          By comparing known or modeled soil concentrations of Pb to the Eco-SSL value derived
36    for each receptor group, Eco-SSL values can be used to identify locations for which further
37    analyses are warranted to determine adverse effects from Pb. Soil screening values, including
            December 2006                     6-4         Draft - Do Not Quote or Cite

-------
 1    Eco-SSLs, were used in this way in the ecological screening analyses conducted for this
 2    assessment and are discussed more fully in Section 6.4 of this document.

 3         6.2.4  Effects on Plants
 4          As discussed in Section 7.3.1  of the CD, atmospheric deposition of Pb onto vegetation is
 5    the primary route of exposure to plants from atmospheric Pb. Lead enters plant tissues primarily
 6    through direct transport, whether by surface deposition or through the soil. There is some uptake
 7    through root cell walls via pore water but little Pb is translocated to other parts of the plant by
 8    this mechanism. Most Pb that does enter plant tissues is deposited in the roots.
 9          Toxicity to plants occurs over a broad range of soil Pb concentrations (tens to thousands
10    of mg/kg) due in part, to the interaction between various soil processes and the bioavailability of
11    Pb to plants (CD, Section 7.1.4). Laboratory studies have shown great variation in toxicity to
12    plants based on the route of exposure and the form of Pb to which the plants are exposed. Two
13    main factors make it very difficult to determine concentration responses for plants in the field:  1)
14    the large number of confounding factors that need to be controlled for, and 2) the lack of good
15    field sites without multiple metal exposures. The 1986 CD (USEPA, 1986) indicated that most
16    plants experience reduced growth when Pb concentrations in pore water exceed 2 to 10 mg/kg
17    and when soil concentrations exceed  10,000 mg/kg under conditions of low bioavailability (e.g.,
18    high pH, oxide rather than acetate forms, etc.) Under increased bioavailability, Pb would cause
19    reduced growth  at much lower levels (e.g. <100 mg/kg). More recent studies have indeed
20    indicated effects at much lower levels than 10,000 mg/kg in the laboratory. For example, at
21    2,800 mg Pb/kg dry weight of soil, adverse effects on growth were found for radish shoots when
22    exposed to Pb-chloride in mildly acidic sandy loams and at 12,000 mg/kg for shoots under
23    similar exposures to Pb-oxide (CD, Section AX7.1.4).  Root cell elongation, another indicator of
24    growth, was inhibited in ryegrass at <2.5 mg/kg Pb-chloride and absence of root growth was
25    observed at 5 mg/kg. Elevated toxicity  was also found for red spruce and ryegrass when exposed
26    to Pb under low pH conditions (CD, Section AX7.1.3.1). There is a wide breadth of studies
27    discussed in the CD for various plants in the laboratory which indicate that Pb in concentrations
28    found in soils near point sources could  reduce plant growth. Despite this information, there are
29    very few reports of phytotoxicity from Pb exposure under field conditions. Indeed two studies
30    cited in Section  AX7.1.3.2 of the CD found no indication of toxicity in plants exposed to high
31    soil concentrations of Pb and other heavy metals near mining sites despite relatively high
32    concentrations of Pb in the vegetation (4000 |ig/g in Leita et al., 1989).  Overall, the
33    phytotoxicity of Pb is considered relatively low because little Pb enters plants from soil and what
34    Pb does enter into plant tissue is deposited in roots where it is either detoxified or sequestered.
            December 2006                     6-5         Draft - Do Not Quote or Cite

-------
 1         6.2.5   Effects on Birds and Mammals
 2          The primary source of Pb exposure to birds and mammals is through dietary intake of
 3    both contaminated food items and incidental ingestion of soils/sediments.  Direct inhalation of
 4    Pb rarely accounts for more than 10 to 15% of daily exposures and drinking water exposure is
 5    not a significant source of Pb for most organisms (CD, AX7.1.3.1).
 6          Physiological effects from Pb exposure in birds and mammals include increased lipid
 7    peroxidation (fat breakdown) and effects on blood component production (CD, Section
 8    AX7.1.2.5). Lipid peroxidation and fatty acid changes have been linked to changes in immune
 9    system response and bone formation. Other adverse effects may include changes in juvenile
10    growth rates; delay of reproductive maturity; behavioral effects, such as decreased predator
11    avoidance or lack of balance and coordination; and mortality.  This cascade of effects has the
12    potential to influence populations by reducing the number of organisms and the rate at which
13    they are replaced, as well as altering food web composition.
14          Toxic effects to birds from Pb exposure have been observed over a wide range of doses in
15    laboratory studies, usually measuring reproductive  success, but little to no data are available on
16    field populations.  Studies have found few significant effects in birds below doses of lOOmg/kg
17    in the diet and there is evidence that wide ranges of effects levels may be expected. Even in
18    studies focused on reproductive effects in the same species, effects from doses ranging from <1
19    to >100 mg Pb/kg bw/day have been observed (CD, AX7.1.3.5). This variation is also true for
20    other effects  (e.g. behavioral and physiological effects) which have been observed at lower
21    doses. As described in Section AX7.1.3.3 of the CD, no data are available on inhalation
22    exposures of birds and very little research has been done since the 1986 CD on toxicity from Pb
23    to birds not exposed to sediment (waterfowl). A discussion of effects to waterfowl can be found
24    in Section 6.3.2.4 of this document.
25          Soil Pb concentrations and potential toxicity to birds has been considered in the
26    development of Eco-SSLs by EPA's Office of Solid Waste and Emergency Response (USEPA,
27    2005b).  As discussed in Section 6.2.3, a soil Pb  concentration of 11  mg/kg dry weight of soil
28    was derived as the Eco-SSL for birds (woodcock) (CD, Section AX7.1.4). This concentration is
29    commonly exceeded in many areas including those not influenced by point sources (CD,
30    Sections 3.2 and AX7.1.2.3).
31          Toxic effects to mammals from Pb exposure have also been observed over a wide range
32    of doses in laboratory studies with little information available  for field populations or exposures.
33    Recent studies indicate that effects on wildlife survival would likely  occur at higher doses than
34    the 2 to 8 mg/kg-day reported in the 1986 CD. Several studies have recently reported no
35    observed adverse effect levels (NOAELs) for survival ranging from 3.5 to as high as 3200
36    mg/kg-day (CD, AX7.1.3.3).  No inhalation studies were found to evaluate endpoints in
37    mammals and in those studies used to develop toxicity endpoints, organisms were dosed using
            December 2006                     6-6         Draft - Do Not Quote or Cite

-------
 1    either ingestion or gavage (tube feeding) which may not necessarily simulate exposure levels in
 2    the field.
 3           A Pb Eco-SSL has been derived for mammals (shrews) at 56 mg/kg dry weight of soil
 4    based in part on toxicity reference values established for reproductive and growth effects
 5    (USEPA, 2005b). Soil concentrations exceeding 56 mg Pb/kg are not uncommon in
 6    urban/industrial locations or near major roadways and may indeed also occur in areas influenced
 7    by deposition of gasoline derived Pb without current Pb emission sources (CD, Section 3.2 and
 8    AX7.1.2.3).
 9          Several behavioral and physiological processes seem to alter the toxicity of Pb in birds
10    and mammals.  Nutritionally deficient diets, especially those low in calcium, lead to increased
11    uptake of Pb from the diet.  Studies have also shown that younger animals and females are
12    generally more sensitive to Pb, insectivorous animals may be more highly exposed than
13    herbivores, and higher trophic level organism are less exposed than lower trophic level
14    organisms.

15         6.2.6   Effects on Decomposers and Soil Invertebrates
16          Elevated concentrations of Pb in soils can lead to decreased decomposition rates  either by
17    direct toxicity to specific groups of decomposers, by deactivating enzymes excreted by
18    decomposers to break down organic material or by binding with organic matter and making it
19    resistant to the action of decomposers.  Direct adverse effects to invertebrates, such  as
20    earthworms and nematodes, include decreased survival, growth and reproduction. Toxicity has
21    been observed in soil invertebrates and microorganisms at concentrations of hundreds to
22    thousands of mg Pb/kg soil with significant variation due to soil parameters such as pH and
23    amount of organic matter (CD, Section AX 7.1.2).
24          As discussed in Section 6.4.2 and CD Section 7.1.4, an Eco-SSL of 1700 mg/kg dry
25    weight of soil has been derived for soil invertebrates (USEPA, 2005). This concentration does
26    not appear to be commonly exceeded in areas not directly influenced by point sources (CD,
27    Sections 3.2 and AX7.1.2.3).
28          Several physiological mechanisms for reducing Pb toxicity have been found among
29    invertebrates and microorganisms. These include enzyme mediated detoxification in two species
30    of spider, Pb storage in waste nodules in earthworms and storage as an inert compound,
31    pyromorphite, in nematodes.  Avoidance of contaminated substrates and reduced feeding has
32    also been observed in invertebrates.

33         6.2.7   Summary
34          Lead exists in the environment in various forms which vary widely in their ability to
35    cause adverse effects on ecosystems and organisms. Current levels of Pb in soil also vary widely
36    depending on the  source of Pb but in all ecosystems Pb concentrations exceed what is thought to
            December 2006                      6-7         Draft - Do Not Quote or Cite

-------
 1    be natural background levels. The deposition of gasoline-derived Pb into forest soils has
 2    produced a legacy of slow moving Pb that remains bound to organic materials despite the
 3    removal of Pb from most fuels and the resulting dramatic reductions in overall deposition rates.
 4    For areas influenced by point sources of air Pb, concentrations of Pb in soil may exceed by many
 5    orders of magnitude the concentrations which are considered harmful to laboratory organisms.
 6    Adverse effects associated with Pb include neurological, physiological and behavioral effects
 7    which may influence ecosystem structure and functioning.  Eco-SSLs have been developed for
 8    Superfund site characterizations to indicate concentrations of Pb in soils below which no adverse
 9    effects are expected to plants, soil invertebrates, birds and mammals. Values like these may be
10    used to identify areas in which there is the potential for adverse effects to any or all of these
11    receptors based on current concentrations of Pb in soils.

12         6.3  EFFECTS IN AQUATIC ECOSYSTEMS
13          Atmospheric Pb enters aquatic ecosystems primarily through the erosion and runoff of
14    soils containing Pb and deposition (wet and dry).  While overall deposition rates of atmospheric
15    Pb have decreased dramatically since the removal of Pb additives from gasoline, Pb continues to
16    accumulate and may be re-exposed in sediments and water bodies throughout the  U.S (CD,
17    Section 2.3.6).
18          Several physical and chemical factors govern the fate and bioavailability of Pb in aquatic
19    systems. A significant portion of Pb remains bound to suspended particulate matter in the water
20    column and eventually settles into the substrate. Species, pH, salinity, temperature, turbulence
21    and other factors govern the bioavailability of Pb in surface waters  (CD,  Section 7.2.2).

22         6.3.1   Tools for Identifying Ecotoxicity in Aquatic Organisms
23          Ambient Water Quality  Criteria (AWQC) were developed by U.S. EPA to provide
24    guidance to states and tribes to use in adopting water quality standards.  AWQC values are
25    available for freshwater and marine environments and for chronic and acute exposures. These
26    values vary with water hardness and are based on the amount of dissolved Pb in the water
27    column. They are derived from  toxicity testing on aquatic organisms, including fish,
28    invertebrates and algae and are  considered to be values below which no adverse effect is
29    anticipated (USEPA, 1993). Therefore these values are useful in identifying locations for which
30    there is the potential for adverse effect from Pb.  Section 6.4 describes how these  criteria were
31    used in the risk characterization for the ecological analyses that accompany this review.
32           A number of sediment ecotoxicity screening values have been developed  to identify the
33    concentration of Pb in sediment at which the potential for adverse effects occur. EPA has
34    recently published an equilibrium partitioning method for sediment which incorporates the
35    bioavailability of Pb and allows for mixtures of metals but may not account for ingestion of

            December 2006                     6-8         Draft - Do Not Quote or Cite

-------
 1    sediment by sediment dwelling organisms. There are other alternative approaches for deriving
 2    sediment criteria which are based more directly upon comparisons between concentrations of Pb
 3    in sediment and associated effects from toxicity tests. These methods do not account for
 4    bioavailability or metal mixtures but are compatible with data available from current water
 5    quality databases.  One of these methods developed by MacDonald et al (2000) is used in the
 6    analyses described in Section 6.4.

 7         6.3.2   Effects in Marine/Estuarine Ecosystems
 8          This section gives a brief overview of the information available for Pb in marine and
 9    estuarine systems.  Most Pb in marine systems is in the inorganic form, complexed with chloride
10    and carbonate ions. Increasing salinity increases the amount of Cl" and COj2' complexation and
11    reduces concentration of free Pb2+ thereby producing compounds with lower bioavailability.
12    There is less data available for the effects of Pb on saltwater organisms and ecosystems but
13    studies indicate lower concentrations of Pb in oceans and large lakes. Toxicity data as  expressed
14    in both the AQWC guidelines (USEPA, 1993) and CD, AX7.2.2, indicate a much higher
15    threshold for effects in saltwater environments.

16         6.3.2.1  Pathways of Exposure
17          Sources of Pb to marine and estuarine ecosystems include runoff from contaminated
18    watersheds, direct atmospheric deposition and turnover of contaminated sediment in areas of
19    high turbulence. Lead is primarily found in the open ocean in the dissolved form and is available
20    in sediment in a variety of complexed forms.  Lead concentrations in oceans were found to be
21    much lower than those measured in freshwater lotic environments and studies with estuarine
22    organisms have also shown reduced toxicity with increasing salinity, most likely due to increased
23    complexation with Cl" ions thereby reducing bioavailability. Studies in the Pacific Ocean near
24    Hawaii have found concentrations of total Pb between 5-11 ng/kg (CD, Section 7.2.2).

25         6.3.2.2  Effects on Organisms and Communities
26          Hematological and neurological responses, including red blood cell destruction, enzyme
27    inhibition and spinal curvature, were the most commonly reported effects in aquatic vertebrates.
28    Demonstrated effects in invertebrates include alteration of reproduction rates and reduced
29    growth.
30          Studies with marine protozoa indicate that at water column concentrations of 0.02 to 1.0
31    mg Pb/L, abundance, biomass and diversity are reduced.  In an estuarine community, Pb was
32    found to affect species abundance when sediment concentrations reached  1343 mg/kg dry
33    weight.  Inhibition of embryo development in commercial shellfish has been documented at
34    water concentrations of 50|ig/L (CD, AX 7.2.4.3).


            December 2006                      6-9        Draft - Do Not Quote or Cite

-------
 1          The toxicity of Pb in the marine or estuarine environment is highly dependent on salinity.
 2    A study of mysid shrimp reported a lethal concentration for 50% of the test organisms (LCso) of
 3    1140 |ig/L at a salinity of 5% and an LC50 of 4274 |ig/L at 25 % salinity.  There is also some
 4    evidence of gender sensitivity in that male copepods were more sensitive to Pb in sediment than
 5    females. Smaller fish have been shown to be more sensitive than larger fish of the same species.
 6    Studies on invertebrates have also shown that deposit feeders were most affected by elevated
 7    substrate concentrations.

 8         6.3.3   Effects in Freshwater Ecosystems
 9          This section gives a brief overview of information available for Pb in freshwater systems.
10    Most Pb in freshwater systems is in the inorganic form. Speciation is important in bioavailability
11    and is dependent upon factors such as pH, temperature and water hardness. In freshwater, Pb
12    typically forms strong inorganic complexes with OH" and CO32" and weak complexes with Cl".
13    Organic Pb compounds in freshwater, which may increase bioavailability,  arise from both natural
14    and anthropogenic sources. Concentrations of various forms of organic Pb complexes are largely
15    dependent on pH and water hardness.

16         6.3.3.1  Pathways of Exposure
17          The bioavailability and accessibility of Pb to aquatic organisms is determined largely by
18    the species of Pb that forms in the ecosystem. In an acidic environment (pH<4) the ionic form,
19    which is the more toxic form, of most metals generally predominates. As pH increases,
20    carbonate, oxide, hydroxide, and sulfide complexes usually predominate and tend to be less
21    toxic. Water hardness also influences toxicity by providing competition in the form of calcium
22    and magnesium to Pb binding sites on biological membranes.  Therefore, Pb  is least toxic in
23    neutral to basic pH levels and at increased water hardness.  A further discussion of speciation
24    and toxicity can be found in Section AX7.2.2.1 of the CD.
25          The US Geologic Service (USGS) has developed the National Water  Quality Assessment
26    (NAWQA) program which is a nation wide monitoring program that contains data on Pb
27    concentrations in surface water, bulk sediment, and biological tissues for samples collected in
28    many watersheds throughout the U.S.  While the data are not representative of the entire U.S.
29    and the analytical method employed for Pb was not as sophisticated as current methods, it is the
30    most comprehensive national database available. The mean concentration of Pb in U.S. surface
31    waters was 0.66|ig/L (ranging from 0.04 to 30) and in bulk sediment was 120.11 |ig/g (ranging
32    from 0.5 to 12,000) for data collected between 1991 and  2003 (CD, AX7.2.2.2 and Section 2.2).

33         6.3.3.2  Effects at an Ecosystem Level
34          The effects of Pb in aquatic ecosystems have not  been well studied for areas affected by
35    atmospheric deposition rather than point source pollution. Aquatic ecosystems near smelters,

            December 2006                     6-10        Draft - Do Not Quote or Cite

-------
 1    mines and other industrial sources of Pb have demonstrated a wide variety of effects including
 2    reduced species diversity, abundance and richness; decreased primary productivity, and
 3    alteration of nutrient cycling. Apportioning these effects between Pb and other stressors is
 4    problematic since these point sources also emit a wide variety of other heavy metals which may
 5    cause toxic effects in aquatic systems.
 6           Lead exposure may adversely affect organisms at different levels of organization, i.e.,
 7    individual organisms, populations, communities, or ecosystems. Generally,  however, there is
 8    insufficient information available for single contaminants in controlled studies to permit
 9    evaluation of specific impacts on higher levels of organization (beyond the individual organism).
10    Potential effects at the population level or higher are, of necessity, extrapolated from individual
11    level studies.  Available population, community, or ecosystem level studies are typically
12    conducted at sites that have been contaminated or adversely affected by multiple stressors
13    (several chemicals alone or combined with physical or biological stressors).  Therefore, the best
14    documented links between Pb and effects on the environment are with effects on individual
15    organisms.
16           However, several recent studies have attributed the presence of Pb to reduced primary
17    productivity, increased respiration, and alterations of community structure.   Specifically,
18    dissolved Pb at concentrations from 6 to 80 mg/L (concentrations higher than those found in the
19    NAWQA database) was found to reduce primary productivity and increase respiration in an algal
20    community. Laboratory microcosm studies have indicated reduced species abundance and
21    diversity in protozoan communities exposed to 0.02 to 1 mg Pb/L (CD, Section AX 7.2.5). Field
22    studies have associated the presence or bioaccumulation of Pb with reductions in species
23    abundance, richness, or diversity, particularly in sediment-dwelling communities (CD, Section
24    AX7.2.5). Most of the available data for Pb effects in aquatic ecosystems comes from either
25    laboratory studies which focused on only a few aspects of the natural system thereby neglecting
26    some of the factors known to influence bioavailability of Pb or from  complex natural systems
27    with many stressors and various sources of anthropogenic Pb, particularly direct mining waste
28    inputs (CD, AX7.2.5.2).  Thus, the effects of atmospheric Pb on aquatic ecological condition
29    remain to be defined.
30           There is a paucity of data in the general literature that explores the effects of Pb in
31    conjunction with all or several of the various components of ecological condition as defined by
32    the EPA (Young and Sanzone, 2002). Recent studies have attributed the presence of Pb to
33    adverse effects on biotic conditions such as abundance, diversity, reduced primary productivity,
34    and alteration of community structure (CD, Section 7.2.5). It is difficult to apportion effects
35    between Pb ands other stressors, however, and these studies did not generally account for
36    modifying factors that may mediate or exacerbate Pb effects.

             December 2006                     6-11        Draft - Do Not Quote or Cite

-------
 1          Lead concentrations in sediment vary with depth and are attributable to increased
 2    anthropogenic inputs over the last few decades.  Several studies have been undertaken to identify
 3    regional sources of Pb in eastern North America and the Great Lakes and have found positive
 4    correlations between Pb isotope ratios in the Great Lakes and known aerosol emissions from
 5    current and historic industrial sources in Canada and the U.S. These studies seem to indicate that
 6    current emissions are contributing somewhat to Pb in sediments (CD, AX7.2.2.3). Resuspension
 7    of historically deposited Pb in  sediments may also constitute a source of Pb in some systems for
 8    the foreseeable future (CD, AX 7.2.2.3).

 9         6.3.3.3  Effects on Algae  and Aquatic Plants
10          As primary producers in aquatic systems, algae and aquatic plants are vital to ecosystem
11    function and provide the foundation upon which the food web depends. Therefore impacts to
12    these organisms can create a chain of effects that impacts the entire ecosystem. Algae and
13    aquatic plants are exposed to Pb by either uptake from the water column or sediment.  Pb is most
14    bioavailable in the divalent form (Pb2+) and as such is adsorbed onto cell walls and accumulates
15    in the cell wall or surface of the plasma membrane of aquatic plants and algae (CD, AX7.2.3.1).
16    Bioconcentration of Pb, the accumulation of Pb inside an organism, may be quite high for both
17    algae and aquatic plants and have made them effective in the remediation of contaminated areas.
18    In aquatic plants as in terrestrial plants, Pb tends to be sequestered (bound and stored) in roots
19    much more than in shoots although some wetland plants have been found to accumulate high
20    levels of Pb in shoots as well.  Within the plants the sequestered Pb tends to be metabolically
21    unavailable until a certain concentration is reached which appears to be species specific.
22          Growth inhibition is exhibited by algae and aquatic plants over a broad range of Pb
23    concentrations in water (1000 to  > 100,000 |ig/L) due in part, to the interaction between various
24    biochemical factors and the bioavailability of Pb to these organisms (CD, AX7.2.3.1). Clinical
25    signs of Pb toxicity in algae include deformation and disintegration of cells, shortened
26    exponential growth phase, and inhibition of pigment synthesis which may ultimately lead to cell
27    death.  As reported in the CD  (Section AX7.2.3.1), studies have shown growth inhibition of
28    Closterium acerosum, a freshwater algae, at concentrations of 1,000 |ig/L Pb nitrate exposure
29    and an effects concentration for 50% of the test population (ECso) for growth inhibition of
30    Scenedesmus quadricauda has been reported at 13,180|ig/L. Other species of algae such as
31    Synechococcus aeruginosus were much more tolerant and required concentrations in excess of
32    82,000 |ig/L to elicit significant growth inhibition. In aquatic plants, toxicity studies have
33    focused on the effects of Pb on plant growth, chlorophyll concentration and protein content. An
34    ECso of 1,100  |ig/L was reported for growth inhibition for Azollapinnata, an aquatic fern, when
35    exposed to Pb-nitrate for 4 days.  Studies with duckweed, Lemna gibba, have reported an ECso of
36    3,750jig/L under the same conditions. These studies indicate the possibility of adverse impacts

            December 2006                      6-12         Draft - Do Not Quote or Cite

-------
 1    to algae and aquatic plants at concentrations which may be found in the vicinity of direct
 2    discharges from point sources but which would not be expected from ambient deposition.
 3           There are two main mechanisms by which algae and plants may moderate Pb toxicity:
 4    sequestration in roots or cell walls, and production of enzymes which complex Pb to make it
 5    metabolically inactive. Studies have shown phytochelatins, polypeptides which chelate heavy
 6    metal ions and make them biologically unavailable to the organism, may be synthesized in
 7    response to exposure to heavy metals (CD, AX7.1.2.4).

 8         6.3.3.4  Effects on Invertebrates
 9           Aquatic invertebrates serve an important role in aquatic ecosystems as both consumers of
10    detrital material and as a prey source for many other organisms. Therefore, adverse impacts to
11    invertebrates can dramatically alter or reduce ecosystem  function.  Invertebrates may accumulate
12    Pb in tissue through ingestion of food and water and adsorption from water. Dietary Pb may
13    contribute significantly to the chronic toxicity of Pb through ingestion of food which has
14    accumulated Pb or by incidental ingestion of sediments.  Studies which relate the effects of
15    dietary exposure and toxic effects in aquatic systems are rare; however, it may be assumed that
16    both dietary and waterborne exposures are important to overall Pb toxicity (CD, AX7.2.4.3).
17           Exposure to Pb can result in reduction of growth  rates and reproductive rates as well as
18    cause increased mortality.  As discussed in Section 6.3.2.1  of this document, both acute and
19    chronic toxicity of Pb can be significantly influenced by  water hardness and pH. A study by
20    Borgmann et.al (2005) with Hyalella azteca, a freshwater amphipod, showed a 23-fold increase
21    in acute toxicity in soft water (18 mg CaCOs/L) compared to hard water (124 mg CaCOs/L).
22    The influence of pH on Pb toxicity varies between invertebrate species. Studies have reported
23    increasing mortality with decreasing pH in some bivalves, cladocerans, amphipods, gastropods
24    and mayflies while some crustaceans and gastropods have shown no relationship between pH
25    and mortality under identical conditions. For the amphipod H. azteca, the low-observed effect
26    concentration (LOEC) for  survival in hard water at pH 8.27 was 192|ig/L as dissolved Pb and
27    466jig/L as total Pb leading to the conclusion that both waterborne and dietary Pb  contributed to
28    this reduced survival (CD, AX7.2.4.3).  Overall, adverse effects for the most sensitive
29    invertebrates studied, amphipods and waterfleas, occurred at concentrations ranging from 0.45 to
30    8,000|ig/L. Exposures to Pb in sediment can also produce toxic effects in sediment dwelling
31    invertebrates.  Acute effects in the water flea, Daphnia magna, included reduced mobility after
32    exposure to 7,000 mg Pb/kg dw for 48 hours while chronic exposure of midges to sediments
33    containing 31,900 mg Pb/kg dw resulted in 100% mortality over 14 days (CD,  AX7.2.4.3).
34    Overall, based on recorded Pb concentrations in the NAWQA database, there are some surface
35    waters and sediments in the U.S where effects on sensitive invertebrates would be expected but
36    apportioning these concentrations between air and non-air sources has not been done.

             December 2006                    6-13        Draft - Do Not Quote or Cite

-------
 1          There are several mechanisms by which invertebrates detoxify Pb.  Lead may be
 2    concentrated in some invertebrates by formation of granules which may be eventually excreted,
 3    sequestered within the exoskeleton and glandular cells, or bound to membranes in gills and other
 4    tissues. Avoidance behaviors have been documented for the aquatic snail, Physella columbiana,
 5    but few studies were found that reported avoidance behaviors in invertebrates. As neurological
 6    and behavioral effects may be important in determining the adverse effects of Pb, further
 7    research is needed in this area.

 8         6.3.3.5  Effects on Fish and Waterfowl
 9          Both the ingestion of contaminated sediment and prey items as well as direct absorption
10    from water contributes to fish exposures to Pb.  Dietary effects of Pb are not well studied in fish
11    but evidence supports that higher tissue concentrations have been found in fish with direct
12    contact with sediment.  Gale et al. (2002) found a good correlation between sediment
13    concentration and tissue concentrations in suckers and small sunfish, which feed directly from
14    the sediment, but not in smallmouth bass, which feed at a higher trophic level. Bioconcentration
15    does occur in freshwater fish and bioconcentration factors (BCFs) for brook trout and bluegill of
16    42 and 45, respectively, have been reported (CD, AX7.2.3.1).  Studies have also shown that fish
17    accumulate Pb more rapidly in low pH environments and when diets are calcium deficient.
18          Lead has been observed to have adverse effects on the production of some enzymes
19    which affect locomotor function as well as adverse blood chemistry effects in some fish.
20    Symptoms of Pb toxicity in fish include the production of excess mucous, spinal deformity,
21    anemia, darkening of the dorsal region, degeneration of the caudal fin, destruction of spinal
22    neurons, enzyme inhibition, growth inhibition, renal pathology, reproductive effects, and
23    mortality (CD, AX7.2.4.3). As in other organisms, Pb speciation, water pH and water hardness
24    play an important role in the  toxicity of Pb. Spinal deformities were found to occur at much
25    lower Pb concentrations in soft water than in hard water. Maximum acceptable threshold
26    concentrations (MATC), the  maximum concentrations at which no adverse effects were seen,
27    have been reported (CD, AX7.2.4.3) for rainbow and brook trout in soft water as 4.1 to 7.6 |ig/L
28    Pb and 58 to 118|ig/L Pb respectively. A LCso of 810 |ig/L was found using fathead minnows at
29    a pH of 6-6.5 while at the same water hardness the LCso was >5,400 |ig/L at a pH range of 7 -
30    8.5. Other studies have shown alterations in blood chemistry in fish from chronic and acute
31    exposures ranging from 100 to 10,000|ig/L Pb (CD, Section AX8.2.3.3). Therefore, given the
32    concentrations of Pb found in surface waters in the NAWQA database, there  are likely adverse
33    effects to fish populations in  some locations of the U.S. It is not clear what the ambient air
34    contributions of Pb are at these locations.
35          There are several physiological and behavioral mechanisms by which fish reduce
36    exposure and absorption of Pb. While the avoidance response to Pb in fish has not been well

            December 2006                     6-14        Draft - Do Not Quote or Cite

-------
 1    studied, it is known for other metals and is thought likely for Pb (CD, AX7.2.3.2). As in other
 2    organisms, gender and age are important variables in determining the adverse effects of Pb with
 3    females and young fish being more sensitive to Pb.
 4          Incidental ingestion of contaminated sediment is the primary route of exposure for
 5    waterfowl. Studies by Beyer et al. (2000) in the Coeur d'Alene watershed near mining and
 6    smelting activity have shown a range of effects for waterfowl based on sediment concentrations
 7    and corresponding blood Pb levels. This study suggested that a NOAEL blood concentration of
 8    0.20 mg/kg wet weight Pb corresponded to a sediment concentration of 24 mg/kg Pb. Sub-
 9    clinical poisoning (LOEL) occurred in swans when sediment concentration was 530 mg/kg Pb
10    which corresponded to a 0.68 mg/kg blood Pb  level.  Some mortality may occur with sediment
11    concentrations as low as 1800 mg/kg Pb and an LCso was found in swans at 3,600 mg/kg Pb in
12    sediment. While these values are somewhat site specific and are dependent on the bioavailability
13    of the Pb as well as the overall health and diet  of the animals, the correlation between blood Pb
14    levels and effects should be applicable irrespective of location-specific variables. Given  current
15    concentrations of Pb in sediment, it is likely that some adverse effects are occurring in waterfowl
16    exposed to point sources of Pb, whether through deposition or direct discharge.

17         6.3.4   Summary
18          Lead exists in the aquatic environment in various forms and under various chemical and
19    physical parameters which determine the ability of Pb to cause adverse effects either from
20    dissolved Pb in the water column or Pb in sediment. Current levels of Pb in water and sediment
21    also vary widely depending on the source of Pb. Conditions exist in which adverse effects to
22    organisms and thereby ecosystems may be anticipated given experimental results.  It is unlikely
23    that dissolved Pb in surface water constitutes a threat to ecosystems that are not directly
24    influenced by point sources. For Pb in sediment, the evidence is less clear.  It is likely that some
25    areas with long term historical deposition  of Pb to sediment from a variety of sources as well as
26    areas influenced by point sources have the potential for adverse effects to aquatic communities.
27    The long residence time of Pb in sediment and its  ability to be resuspended by turbulence make
28    Pb likely to be a factor for the foreseeable future.  Criteria have been developed to indicate
29    concentrations of Pb in water and sediment below which no adverse effects are expected  to
30    aquatic organisms. These values may be used to identify areas in which there is the potential for
31    adverse effects to receptors based on current concentrations of Pb in water and sediment.

32         6.4   SCREENING LEVEL RISK ASSESSMENT
33         6.4.1   Overview of Analyses
34          The ecological risk assessment consisted of case studies and a national scale surface
35    water and sediment screening assessment. In this analysis, activities for an additional case study

            December 2006                     6-15        Draft - Do Not Quote or Cite

-------
 1   (ecologically vulnerable location) focused on identification and description and did not include
 2   risk analyses. The case study analyses were designed to estimate the potential for ecological risks
 3   associated with exposures to Pb emitted into ambient air for three case studies: primary Pb
 4   smelter, secondary Pb smelter, and near roadway non-urban location.  Soil, surface water, and/or
 5   sediment concentrations, as  available, were compared to ecological screening benchmarks to
 6   assess the potential for ecological impacts from NAAQS-relevant sources.  Figure 6-1 provides
 7   an overview of the analyses  undertaken for this assessment.  These results are not definitive
 8   estimates of risk, but rather intended to focus attention on those locations at which there is the
 9   greatest likelihood for adverse effect.  The national-level screening assessment evaluated the
10   potential for ecological risks associated with the atmospheric deposition of Pb released into
11   ambient air at surface water  and sediment monitoring locations across the United States.
12           This overview shows the key types of information and models involved in each phase of
13   the analysis and how they are related to each other. Table 6-1 summarizes the use of these
14   information types and models for each case study. As indicated in these exhibits, the specific
15   approach for each case study differed based on the nature of the case study (e.g., type of source,
16   land use) and the site-specific measurements available.  In cases where the available
17   measurements were not sufficient to characterize the study area (e.g., due to insufficient spatial
18   coverage), these data were used for performance  evaluation.
            December 2006                     6-16        Draft - Do Not Quote or Cite

-------
Figure 6-1.   Overview of Ecological Screening Assessment.
Exposure Assessment: Estimating Media Concentrations

                                                          National-Scale
                                     Secondary Pb        Surface Water &
                                   Smelter Case Study     Sediment Screening
                                                           Assessment
       Primary Pb
        Smelter
       Case Study
Near Roadway
  N on-urban
  Case Study
     Site-specific soil,
     surface water, and
        sediment
      monitoring data
Site-specific soil
monitoring data
Stack and fugitive
   emissions
   estimates
                                               Air dispersion
                                                  model
                                              Lead deposition
                                             rates across study
                                                   area
                                                   Soil
                                                  mixing
                                                  model
                    Outdoor soil
                   concentrations
                                      Water column
                                      concentrations
                                                            Sediment
                                                          concentrations
         Risk
   Char a ctei izatio n
                                           Ecological
                                             Effects
                                           Assessment
                                                                                             Soil
                                                                                        screening values
                                                                 Freshwater
                                                               screening values
                                                                                           Sediment
                                                                                        screening values
                Media-specific Hazard Quotients
 December 2006
                 6-17
            Draft - Do Not Quote or Cite

-------
1   Table 6-1. Models and Measurements Used for Ecological Risk Screening Assessment.
Primary Pb Smelter
Location Missouri
Approximately 6km diameter,
Spatial extent and resolution
centered on point source
Secondary Pb
Smelter
Alabama
U.S. Census
blocks
Near Roadway Non-Urban
Corpus Christi, Texas
Atlee, Virginia
• Corpus Christi: single
transect perpendicular to
road; 0.5 - 4 m from
road
• Atlee: 140-m section of
road; 2 - 30 m from road
Exposure Assessment: Estimating Media Concentrations
Deposition
to soil
Soil cone.
Surface
water cone.
Sediment
cone.
Models
Measure-
ments
Models
Measure-
ments
Models
Measure-
ments
Models
Measure-
ments
n/a
n/a
n/a
Site-specific cone, of total Pb in
soil samples
(26 locations)
n/a
Site-specific cone, of dissolved
Pb in water column samples from
eight water bodies/drainage
areas (30 locations)
n/a
Site-specific cone, of total Pb in
sediment samples from five water
bodies /drainage areas (69 sites)
AERMOD3
n/a
MPEb
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Site-specific cone, of
total Pb in soil samples
(Corpus Christi: 2
locations; Atlee: 26
locations)
n/a
n/a
n/a
n/a
National-Scale Aquatic
Screen
Surface water bodies in
the U.S.
47 basin study units from
all regions of U.S.,
covering approx. 50
percent of U.S. land base

n/a
n/a
n/a
n/a
n/a
Site-specific cone, of dissolved
Pb in surface water samples
(430 samples)
n/a
Site-specific or nearby water
body cone, of total Pb in
sediment samples (15 sites)
Ecological Risk Assessment
Screening
ecotoxicity
benchmarks
Soil
Freshwater
- water
column
Freshwater
- sediment
Soil screening values
U.S. EPA Pb freshwater AWQC
for aquatic life derived based on
site-specific measured water
hardness (cone. Of CaCO3)°
Sediment screening values
based on MacDonald et al.
(2000) sediment quality
assessment guidelines
n/a
n/a
n/a
n/a
n/a
U.S. EPA Pb freshwater
AWQC for aquatic life derived
based on site-specific
measured water hardness
(cone, of CaCO3)°
Sediment screening values
based on MacDonald et al.
(2000) sediment quality
assessment guidelines
2 a American Meteorological Society/EPA Regulatory Model (AERMOD) (USEPA, 2004) b Multiple Pathways of Exposure (MPE) (USEPA,
J 1998) c These screening values are based on measured ecotoxicity data
          December 2006
6-18
Draft - Do Not Quote or Cite

-------
 1         6.4.2   Measures of Exposure and Effect
 2          The measures of exposure for these analyses are total Pb concentrations in soil, dissolved
 3    Pb concentrations in freshwater surface waters (water column), and total Pb concentrations in
 4    freshwater sediments.  These exposure concentrations were estimated for the three case studies
 5    and the national-scale screening analyses as described below:
 6
 7       •  For the primary Pb smelter case study, measured concentrations of total Pb in soil,
 8          dissolved Pb in surface waters, and total Pb in sediment were used to develop point
 9          estimates for sampling clusters thought to be associated with atmospheric Pb deposition,
10          rather than Pb associated with non-air sources, such as runoff from waste storage piles.
11
12       •  For the secondary Pb smelter case study, concentrations of Pb in soil were estimated
13          using fate and transport modeling based on EPA's MPE methodology (USEPA, 1998)
14          and data available  from similar locations.
15
16       •  For the near roadway non-urban case study, measured soil concentration data collected
17          from two interstate sampling locations, one with fairly high-density development (Corpus
18          Christi, Texas)  and another with medium-density  development (Atlee, Virginia), were
19          used to develop point estimates of Pb contaminated soils from historically-deposited Pb
20          for each location.
21
22       •  The national-level  surface water and sediment screening analyses compiled
23          measurements of dissolved Pb concentrations in surface water and total Pb in sediment
24          for locations across to the United States from available databases. Emissions, point
25          discharge, and land use data for the areas surrounding these locations were assessed to
26          identify locations where atmospheric Pb  deposition may be expected to contribute to
27          potential ecological impacts. The exposure assessment focused on these locations.
28
29          The Hazard Quotient (HQ) approach was used to  compare estimated media
30    concentrations with ecotoxicity screening values in soils, surface waters, and sediments around
31    the primary Pb smelter, for soils only around the secondary Pb smelter case study location and
32    the near roadway non-urban case study locations, and for surface water and sediment in the
33    national-level screen. The HQ is calculated as the ratio of the media concentration to the
34    ecotoxicity screening value. The HQ is represented by the following equation:
35
36          HQ = (estimated media concentration) / (ecotoxicity screening value)
37
38          For each case study, HQ values were calculated for each location where either modeled
39    or measured media concentrations were available. Separate soil HQ values were calculated for
40    each ecological receptor group for which an ecotoxicity screening value has been developed (i.e.,


            December 2006                     6-19         Draft - Do Not Quote or Cite

-------
 1    birds, mammals, soil invertebrates, and plants). HQ values less than 1.0 suggest that Pb
 2    concentrations in a specific medium are unlikely to pose significant risks to ecological receptors
 3    whereas HQ values greater than 1.0 indicate that the expected exposure exceeds the ecotoxicity
 4    screening value.

 5         6.4.3   National-Scale Screen and Case Studies
 6          This section provides an overview of the study locations selected for ecological screening
 7    risk assessment performed in support of the NAAQS review.  A national scale screen was
 8    conducted to look at current Pb concentrations in freshwaters and sediments throughout the U.S.
 9    and three case study locations were selected:  1) primary Pb smelter, 2) secondary Pb smelter and
10    3) near roadway non-urban. The primary and secondary smelter case studies represent an
11    extreme and moderate point source scenario while the near roadway non-urban location
12    represents a more ubiquitous exposure from historic gasoline Pb emissions along major
13    roadways.

14         6.4.3.1  National Scale Screen
15          A national scale assessment was performed using the NAWQA database to identify
16    locations in the U.S. in which concentrations of Pb in surface water and/or sediment exceed
17    established screening values and for which ambient air Pb  is likely to be a major factor. These
18    locations were identified using the methodology described below and in the risk assessment
19    report (ICF, 2006).
20         6.4.3.1.1 Fresh Surface Waters
21          A screening-level ecological risk assessment for aquatic ecosystems was conducted for
22    Pb concentrations in fresh surface waters of the United States to identify areas in which there are
23    concentrations in excess of EPA recommended national ambient water quality criteria (AWQC),
24    both chronic and acute, for the protection of freshwater aquatic life. In this assessment, we
25    identified locations at which Pb concentrations exceeded the EPA AWQC and for which air
26    sources are likely to be the major contributor to the Pb concentrations in the water (i.e., there are
27    no other obvious sources of Pb to the water).
28          As the geographic coverage achieved in this surface water screen is based entirely on the
29    geographic coverage of available measurements of dissolved Pb in the selected database, it was
30    important that the most appropriate dataset be used. It was concluded that of the commonly
31    available databases, including NWIS, STORET and NAWQA, the NAWQA data set is most
32    appropriate for a nationwide aquatic risk screen for several reasons. The inclusion of dissolved
33    Pb as an analyte is limited in all of the  databases (total Pb is measured more often).  None of the
34    databases provide the co-located measurements of water hardness in the same records as the
35    measurements of dissolved Pb.  STORET and NWIS include samples from more locations in the

            December 2006                    6-20       Draft - Do Not Quote or Cite

-------
 1    United States than does the NAWQA data set, but the sampling and reporting protocols
 2    represented in STORET and NWIS are less consistent from site to site. Data for dissolved Pb in
 3    NWIS are predominantly from the 1980s, and therefore do not represent current conditions.  The
 4    NAWQA data set, on the other hand, provides representative (though not complete) coverage of
 5    the United States, with samples through 2004 included in the database. The NAWQA data set
 6    also provides a consistent approach to  sampling and analysis of the elements using consistent
 7    quantitation limits across the country.  Given the sampling design for NAWQA and the
 8    consistency of the data across the country, it is considered to be more appropriate for a national-
 9    level aquatic risk screen than the other two data sets and was therefore used for this screen.
10         6.4.3.1.2  Lead in Sediments
11          Possible risks to sediment dwelling organisms were also examined at locations identified
12    in the surface water screen by comparing total Pb concentrations in sediments to ecotoxicity
13    benchmarks for sediments, generally referred to as sediment quality criteria. The preferred
14    approach for sediment data was to obtain it from surface water sampling locations in the
15    NAWQA database. It was not always  possible to sample sediments at locations where surface
16    water samples are desired.  Therefore,  some of the sites of interest do not have sediment samples
17    available from the same location. Where an exact match was not found, a nearby sampling
18    location was identified on the basis of latitude, and longitude, and name of the site location.

19         6.4.3.2  Ecologically Vulnerable Location
20          A literature search was conducted to identify an acidified forest or non-urban acidified
21    watershed ecosystem to which the following criteria could be applied:
22
23       •  Potential for increased bioavailability of Pb due to soil and water acidification;
24       •  Relatively distant from point sources of Pb emissions;
25       •  Relatively high elevation which may be subject to comparatively higher deposition of Pb
26          due to wind speed and precipitation as well as longer residence time; and
27       •  Availability of data on trends (temporal, elevation, etc.) of Pb concentration in various
28          environmental media.

29         Based on these criteria, we selected the Hubbard Brook Experimental Forest (HBEF) in the
30    White Mountains of New Hampshire for the ecologically vulnerable case study. The HBEF was
31    established by the USDA Forest Service in 1955 as a major center for hydrologic research in
32    New England.  The USDA proposed to use the small watershed approach at Hubbard Brook to
33    study linkages between hydrologic and nutrient flux and cycling in response to natural and
34    human disturbances, such as air pollution, forest cutting, land-use changes, increases in insect
35    populations and climatic factors.  The first grant was awarded by  the National  Science
36    Foundation (NSF) to Bormann and Likens in 1963 to support the research at the HBEF. Since

            December 2006                     6-21        Draft - Do Not Quote or Cite

-------
 1    that time there has been continuous support from the NSF and the U.S. Department of
 2    Agriculture (USDA) Forest Service. In 1988 the HBEF was designated as a Long-Term
 3    Ecological Research (LTER) site by the NSF.  On-going cooperative efforts among diverse
 4    educational institutions, private institutions, government agencies, foundations and corporations
 5    have resulted in one of the most extensive and longest continuous data bases on the hydrology,
 6    biology, geology and chemistry of natural ecosystems in the United States. This historical record
 7    makes FffiEF uniquely suited to the purpose of this review. As discussed earlier, an assessment
 8    of Pb related ecological risks for this case study location is not presented in the first draft of this
 9    document. Full development of this case study location including discussion of ecological
10    exposures and risk would be a useful enhancement for any future analyses.

11         6.4.3.3  Primary Pb Smelter
12          The primary Pb smelter case study location is one of the largest primary Pb smelters in
13    the world, is the only remaining operating Pb smelter in the United States, and is also the longest
14    operating smelter in the world, sustaining nearly continuous operation since 1892. Further
15    information on the surroundings and demographics in the vicinity of the primary smelter can be
16    found in the risk report (ICF, 2006). Portions  of this study area  comprise an active Superfund
17    site and are subject to ongoing evaluation under the Superfund program administered by the
18    Office of Solid Waste and Emergency Response. Methods used  in conducting ecological risk
19    assessment for the analysis have been  selected to address policy questions relevant to the Pb
20    NAAQS review and consequently, may differ from those used by the Superfund program.
21          The primary Pb smelter property is bordered on the east by the Mississippi River, on the
22    southeast by Joachim Creek, on the west and north-northwest by residential areas, and on the
23    south-southwest by a slag pile.  A large part of the slag pile is located in the floodplain wetlands
24    of Joachim Creek and the Mississippi River.
25          Ecological features near the facility include the Mississippi River, streams,  emergent and
26    scrub-shrub wetlands, and successional and mature bottomland hardwood forest tracts (ELM,
27    2005). Bottomland hardwood forests and agricultural fields are present to the west, south, and
28    east of the characterization area between the smelter's slag storage area and Joachim Creek. The
29    most mature bottomland hardwood forest is adjacent to Joachim Creek. Immediately south of
30    the facility is a mixture of floodplain forest, emergent marsh, and scrub-shrub wetland habitat
31    that is populated by willow trees.  Throughout much of the year, migratory birds such as the red-
32    tailed hawk, belted kingfisher,  and great blue herons utilize the habitat near the facility.  The
33    federally threatened bald eagle has been spotted on-site at the facility, which is known to be
34    within the habitat for the bird.  The facility is also within the habitat of the Indiana bat, which is
35    on the federal  and state endangered species lists. In addition, the state and federally endangered
36    pallid sturgeon has been identified in the Mississippi River adjacent to and downstream of the
            December 2006                     6-22        Draft - Do Not Quote or Cite

-------
 1    facility. The pink mucket, scaleshell, and gray bat also occur in Jefferson County and are on both
 2    the state and federal endangered lists.
 3         6.4.3.4  Secondary Pb Smelter
 4          The secondary Pb smelter location falls within the Alabama Coastal Plain in Pike County,
 5    Alabama. It is located in an area of disturbed forests, and is less than 2 km from Big Creek,
 6    which is part of the Pea River watershed. Big Creek is located approximately 0.5 m south
 7    southeast from the center of the facility. The surrounding area includes emergent and scrub shrub
 8    wetlands, forests, freshwater creeks, ponds, rivers, croplands, pastureland, and developed urban
 9    areas. The Pea River watershed drains into the Gulf of Mexico. The watershed is underlain by
10    coastal plain sediments, including sand, clay, and limestone; and the topography can be
11    characterized as gentle to moderate rolling hills (CPYRWMA 2006).  Diversity of terrestrial and
12    aquatic animal species is relatively high. The Choctawhatee and Pea River basins, in which the
13    secondary Pb smelter is located, contain 43  species of marine, estuarine, and freshwater fish
14    species (Cook and Kopaska-Merkel, 1996). Anadromous fish species  (i.e., saltwater fish that
15    must spawn in freshwater) found in the Pea River basin include the following: the threatened
16    Gulf sturgeon, Alabama shad, striped bass, and skipjack herring. The  Pea River basin also
17    provides habitat for 20 species of freshwater mussels (Cook and Kopaska-Merkel, 1996), as well
18    as numerous species of snails, snakes, and other invertebrates. Terrestrial species supported in
19    this region include a variety of birds, mammals, invertebrates, and vascular plants.  Other
20    terrestrial fauna found in the region include migratory birds, small mammals and invertebrate
21    species. A total of 34 vascular flora from Pike County are listed by the Alabama Natural
22    Heritage Inventory Program as state endangered, threatened, or of special concern (Alabama
23    Natural Heritage Inventory 2001).  According to NatureServe and the U.S. Fish and Wildlife
24    Service (USFWS), no species in Pike County are on the federal endangered species list (Outdoor
25    Alabama, 2003). A few species, however, are candidates for the federal list.

26         6.4.3.5  Near Roadway Non-Urban Case Study
27          The Houston, Texas, near roadway urban case study for the human health risk assessment
28    used  surrogate soil Pb concentration data measured at a sampling location in downtown Corpus
29    Christi, Texas (Turer and Maynard, 2003).  Air concentration data are not needed for this
30    assessment of ecological risks therefore a search for a more ecologically-relevant case study
31    location was conducted. Non-urban case study locations that provide soil concentration  data
32    were sought with the expectation that ecological receptors would be more likely to occur along
33    roads in less developed areas compared to the downtown location evaluated in the human health
34    risk assessment. Terrestrial wildlife could forage in Pb contaminated  soils alongside highways,
35    particular on roads traversing undeveloped areas.

            December 2006                     6-23         Draft - Do Not Quote or Cite

-------
 1          From the literature search for studies of Pb in near roadway soils, two non-urban sites for
 2    which soil Pb levels are available were identified for use in the ecological risk assessment. These
 3    locations are: (1) Interstate 37 near oil refineries in Corpus Christi, Texas (Turer and Maynard,
 4    2003) and (2) Interstate 95 in Atlee, Virginia, which connects to a moderately traveled, two-lane
 5    road (Speiran, 1998).
 6          Land cover data from 1992 within 1 m of the Corpus Christi, Texas study location
 7    showed 59 percent industrial, 10 percent low intensity residential, and 25 percent high intensity
 8    residential (Vogelmann et al., 2001). The remaining 5 percent of the surrounding area includes
 9    shrubland, row crops, pasture, grasses, and forested upland, including evergreen forest and
10    deciduous forest.
11          The 1992 land cover data within 1 m of the Atlee, Virginia, study locations showed 26
12    percent developed: 2 percent low-intensity residential and commercial and 24 percent industrial
13    and transportation. The remaining 74 percent included 25 percent deciduous forest, 14 percent
14    woody wetlands, and 12 percent pasture (Vogelmann et al., 2001). Smaller proportions of mixed
15    forest, evergreen forest, row crops, and transitional (barren) areas were also found.

16         6.4.4   Screening Values
17          The following is a discussion of specific  ecological screening values selected for use in
18    the risk assessment. The main discussion of the  development and derivation of these tools can
19    be found in Section 6.2.3 of this document and in the risk report (ICF, 2006). This discussion
20    outlines the ways in which the tools were used for this assessment to identify potential effect
21    from Pb  exposure to specific ecological endpoints in either localized case studies or in the
22    NAWQA monitoring database.

23         6.4.4.1  Soil Screening Values
24          In developing soil screening values for use in this assessment, assumptions inherent in the
25    derivation of the Superfund Eco-SSLs were examined, and as appropriate, augmented or
26    replaced with current species-specific information.  For example, the assumptions employed for
27    deriving the Eco-SSLs for avian and mammalian wildlife from the corresponding TRVs were
28    examined (ICF, 2006).  Soil screening values were derived for this assessment using the Eco-
29    SSL methodology (described in Section 6.2.3) with the TRVs for Pb (USEPA, 2005b) and
30    consideration of the inputs on diet composition, food intake rates, incidental soil ingestion, and
31    contaminant uptake by prey. The  soil screening  values shown in Table 6-2 for plants and soil
32    invertebrates are the Eco-SSL values (USEPA, 2005a) while the screening values for birds and
33    mammals are based on the Eco-SSL methodology but with modified inputs specific to this
34    assessment  (ICF, 2006).
            December 2006                    6-24        Draft - Do Not Quote or Cite

-------
 1    Table 6-2.  Soil Screening Values for Pb for Ecological Receptors
Ecological Receptor
Plants l
Soil Invertebrates :
Birds 2
Mammals 2
Soil Concentration
(mg Pb/kg soil dry
weight)
120
1700
38
112
 2         : Values obtained from Ecological Soil Screening Levels for Lead, Interim Final (USEPA, 2005 a).
 3         2 Values obtained by refinement described in risk report (ICF, 2006).
 4
 5         6.4.4.2  Surface Water Screening Values
 6          Specific screening values were calculated using the AWQC developed by EPA (1984) for
 7    the primary smelter case study location and the national scale screen based on site-specific water
 8    hardness data.  AWQC values for chronic exposures are called the criterion continuous
 9    concentration (CCC) and for acute exposures are called the criterion maximum concentration
10    (CMC), and they are available for freshwater and marine environments. For a CCC to be
11    exceeded, a 4-day average water concentration must exceed the CCC more than once every three
12    years (USEPA, 1984).

13         6.4.4.3  Sediment Screening Values
14          This risk screen uses sediment criteria developed by MacDonald et al  (2000) which
15    focuses on total Pb concentrations in sediment and identifies a threshold-effect concentration
16    (TEC) and a probable-effect concentration (PEC) as 35.8 mg/kg and 128 mg/kg respectively.
17    This methodology is described more fully in the risk assessment report (ICF,  2006).

18         6.4.5   Results for Case Study Locations and Comparison to Screening Value
19          To identify locations in which Pb concentrations in soil, water and/or  sediment may
20    potentially be harmful, each case study location was assessed using either empirical data or
21    model results for each media. These concentrations were then compared to screening values as
22    described in Section 6.4.2.  The HQ  approach was then used to compare estimated exposures for
23    geographic  areas around the case study sites with ecotoxicity benchmark values in each of three
24    media; soil, surface water, and sediment. HQs less than or equal to one suggest that ecological
25    risks are negligible. HQs greater than one indicate a potential for adverse effects and a more
26    refined analysis of sensitive receptors may be needed.

27         6.4.5.1  National-Scale Surface Water Screen
28          Based on EPA's re-evaluation of AWQC for metals (USEPA,  1993), the CCC for
29    relatively soft water (50 mg/L CaCOs) is 1.2|ig/L while increasing hardness resulted in higher
            December 2006                     6-25        Draft - Do Not Quote or Cite

-------
 1    CCC values. Therefore, the initial screen of dissolved Pb concentrations in surface water looked
 2    for measurements equal to or greater than 1.2|ig/L. This resulted in 42 sampling locations for
 3    which one or more measurements exceeded that screening value. Data for each measurement of
 4    dissolved Pb at these stations are provided in the risk assessment report (ICF, 2006b). For a
 5    definitive risk assessment, representing a given location with a single sampling measurement of
 6    dissolved Pb would not be considered.  However, for purposes of this risk screen, given the
 7    limited analyses for dissolved Pb, all 42 sampling locations were retained for analysis. Next, the
 8    location-specific CCC and CMC values were determined based on water hardness for those
 9    locations.  A review of the data on water hardness in the NAWQA data set for 1994 to 2004
10    indicated that the initial screening value of 1.2|ig/L was too high to identify all locations for
11    which dissolved Pb concentrations exceeded the CCC for the protection of aquatic life.  Many
12    waters in the United States are softer than anticipated (i.e., measured CaCOs concentrations
13    down to 1 mg/L).
14           A second screen was therefore conducted in which dissolved Pb measurements greater
15    than the quantitation limit (QL) but less than 1.2|ig/L were reviewed.  In the second screen, for
16    each  sampling location with one or more measured dissolved Pb concentrations above the QL
17    but less than 1.2|ig/L, collocated measurements of CaCOs were used to calculate a site-specific
18    CCC as described above. To attempt to isolate those locations where air derived Pb is the major
19    source of Pb to water; land use data was obtained from NAWQA for each location in which the
20    derived HQ was greater than 1.0.  The available categories of land use in the dataset  separated
21    mining sites but did not separate other activities which are likely to produce Pb (e.g.  smelting
22    sites were included in the  industrial category).  While it is likely that mining activities do
23    produce air emissions of Pb (Section 2.3.4), the data is lacking to apportion air and non-air Pb at
24    mining sites. Therefore, results for locations with mining as the land use category were
25    separated from the other land use types. Table 6-2 summarizes the HQs for the 15 non-mining
26    sites for which the chronic HQs exceed 1.0 in order of increasing HQ.  These locations are in
27    areas classified in the NAWQA database as urban and mixed, but also include forest, rangeland,
28    and a "reference" site in Alaska. The highest HQ is for the Alaska reference site and is based on
29    one measurement of dissolved Pb and one measurement of calcium carbonate. Thus, the
30    uncertainty associated with this HQ is high (ICF, 2006).
            December 2006                     6-26        Draft - Do Not Quote or Cite

-------
 1    Table 6-3.  Results of Aquatic Risk Screen - Locations at which Dissolved Pb
 2               Measurements Exceed AWQC, Excluding Mining Sites. a
Basin
ID
45 RIOG
44 UCOL
2 CONN
46 NROK
46 NROK
46 NROK
47 GRSL
2 CONN
31 OZRK
58 OAHU
1 NECB
2 CONN
46 NROK
31 OZRK
31 COOK
State
NM
CO
CT
WA
WA
ID
UT
CT
MO
HA
Rl
CT
ID
AR
AK
Station
ID
8331000
3.85E+14
1127000
12422500
12422000
12392155
4.05E+14
1119375
7018100
16212700
1112900
1124000
12419000
7050500
6.01E+14
Land
Use
Mixed
Other/Mixed
Mixed
Urban
Urban
Forest
Rangeland
Mixed
Forest
Mixed
Mixed
Mixed
Mixed
Mixed
Reference
Lead
CCC
(ug/L)
2.9
0.89
0.36
0.99
0.99
0.17
5.8
0.18
3.7
0.17
0.44
0.30
0.37
2.6
0.11
Pb Measurements
No.
>CCC
/ Total
N
1/12
1/4
3/22
4/28
2/20
4/17
1/2
5/20
1/2
1/2
3/3
11/23
2/26
1/8
1/1
No. < QL,
which
is > CCC
0
2
14
24
18
10
0
13
0
1
0
9
16
0
0
Hazard Quotient
Mean
[Pb]/
CCC
1.03
1.09
1.13
1.14
1.17
1.32
1.45
1.68
1.89
1.98
2.51
2.53
2.69
3.46
14.91
Max
[Pb]/
CCC
1.03
1.09
1.31
1.25
1.17
1.54
1.45
2.09
1.89
1.98
3.53
3.33
4.27
3.46
14.91
Max
[Pb]/
CMC
0.04
0.04
0.05
0.05
0.05
0.06
0.06
0.08
0.07
0.08
0.14
0.13
0.17
0.14
0.61
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

18
19
20
21
       a In order of increasing Hazard Quotient for the CCC aquatic toxicity benchmark. Additional information
characterizing these locations is provided in the risk report (ICF, 2006).

     When the 15 sampling locations in Table 6-2 are compared to NEI data, only three appear
to be near facilities emitting relatively large quantities of Pb to the atmosphere (i.e., more than 1
ton per year): one is in Oahu, Hawaii, one in Jewett City, Connecticut, and one in Manville,
Rhode Island.  An additional two sampling locations appear to be within 50 km of facilities
emitting relatively large quantities of Pb, both in Connecticut; however, whether these facilities
are close enough  to influence the Pb concentrations in the water column at these sampling sites is
unknown. Of the three sampling locations within 20 km of facilities emitting more than 1 ton of
Pb per year, the location in Rhode Island might also be receiving a large part of its Pb from
upstream discharges from metal ore processing facilities (i.e., there are six such discharges out of
14 National Pollutant Discharge Elimination System (NPDES) permitted facilities upstream of
this sampling location). More information on  emissions for these 15 locations can be found in
the risk report (ICF, 2006).

     6.4.5.2   National-Scale Sediment Screen
       Sediment characterization for the 15 sites identified in the AWQC screen was performed
using the hazard quotient method, where measures of total Pb concentrations in sediments were
compared with the sediment TEC and PEC values for the protection of sediment dwelling
            December 2006
                                         6-27
Draft - Do Not Quote or Cite

-------
 1    organisms.  The first step involved attempting to find matching sediment sampling locations in
 2    the NAWQA database. It was not always possible to find collocated sediment and surface water
 3    samples.  It was expected, therefore, that some of the 15 sites of interest would not have
 4    sediment samples available from the same location.  Where an exact match was not found, a
 5    nearby sampling location on the same water body was identified.
 6           Table 6-4 shows the HQs for measured total Pb concentrations in sediments at 12 of the
 7    15 surface water locations for which data were available. The HQs are calculated by dividing
 8    the TEC and PEC for sediment dwelling organisms from the consensus-based approach to
 9    sediment quality criteria (MacDonald et al., 2000).

10    Table 6-4.  Concentrations of Total Pb in Sediments at Locations Near or Matching the 15
11               Sites at which Dissolved Pb Concentrations Exceeded the AWQC, Excluding
12               Mining Sites."
Basin ID
45
2
46
46
47
31
58
1
2
46
31
31
RIOG
CONN
NROK
NROK
GRSL
OZRK
OAHU
NECB
CONN
NROK
OZRK
COOK
State
NM
CT
WA
ID
UT
MO
HA
RI
CT
ID
AR
AK
Land
Use
Mixed
Mixed
Urban
Forest
Rangeland
Forest
Mixed
Mixed
Mixed
Mixed
Mixed
Reference
Match
Yes
Near
Near
Yes
Yes
Yes
Yes
Yes
Near
Yes
Yes
Yes
Total [Pb]
(mg/kg dry
sediment)
23
68
47.3
24.9
2900
2300
59
240
68
1620
28
239
SW
HQb:
max
[Pb]/CCC
1.03
1.13
1.14
1.32
1.45
1.89
1.98
2.51
2.53
2.69
3.46
14.91
Pb Emissions
(tons/year) (b)
Fac<
20km
0.068
6.1
0.39
0.0
0.0
0.0
4.9
4.1
0.081
0.34
0.0
0.0
Fac<
50km
0.095
7.0
0.43
0.0
0.36
0.34
4.9
11.7
11.3
0.43
0.01
0.0
No.
Upstream
NPDES
permits for
metals
0
0
0
1
1
ND
ND
6
0
4
0
0
Sediment
Hazard
Quotients
[Pb]/
TEC
0.64
1.9
1.3
0.70
81
64
1.6
6.7
1.9
45
0.78
6.7
[Pb]/P
EC
0.18
0.53
0.37
0.19
23
18
0.46
1.9
0.53
13
0.22
1.9
13    aExhibit in increasing order of the surface water (SW) column risk hazard quotient (HQ). HQs exceeding 1.0 are
14    highlighted in bold type.
15    b Data collected for corresponding surface water locations
16    Abbreviations: [Pb] = total Pb concentration in sediments (mg/kg dry sediment). CCC = Criterion Continuous
17    Concentration (or chronic AWQC). TEC = threshold-effect concentration, and PEC = probable-effect concentration,
18    both from the consensus-based sediment quality criteria approach published by MacDonald et al. (2000; 2003).
19

20           Table 6-4 presents the HQs for risks to benthic organisms at the 9 matching and 3 near-
21    match locations at which dissolved Pb concentrations in the water column  exceeded the CCC
22    (i.e., chronic AWQC) for the protection of aquatic organisms in surface waters.  Nine of the
23    TEC-based HQs exceeded 1.0, and three were less than 1.0.  The three sites with HQs less than
24    1.0 are unlikely to pose risks to benthic aquatic communities based on the available data. None
            December 2006
6-28
Draft - Do Not Quote or Cite

-------
 1    of these three sites were those likely to be affected by air emissions of Pb from point sources
 2    (i.e., Pb emissions were less than 0.07 tons per year at all three locations).
 3           Five of the PEC-based HQs exceeded 1.0, indicating probable adverse effects.  Three of
 4    these exceeded a PEC-based HQ of 10, indicating a very high probability of adverse effects, and
 5    possibly higher severity of effects than at the locations with lower HQ values. None of these
 6    three locations were likely to be affected, however, by air emissions.  One in Idaho was
 7    downstream from several NPDES-permitted discharges of metals to surface waters (10th entry).
 8    The other two locations were found in Utah and Montana and it is possible that these
 9    concentrations reflect historical sediment contamination from mining operations.
10           Of the three locations for which air emissions of Pb from point sources appear to be more
11    likely to be contributing to ongoing Pb contamination of surface water and sediments (i.e.
12    locations in Connecticut, Hawaii, and Rhode Island, respectively), only one, the Blackstone
13    River in Manville, Rhode Island, is also likely to receive significant current Pb inputs from
14    upstream NPDES-permitted sites. In addition to Pb contamination of sediment through
15    deposition of current air emissions to surface waters, sediment at these three locations may be
16    contaminated by current and historic erosion of soils containing current and historic deposits of
17    Pb, particularly from leaded gasoline. The Quienebaug River in Connecticut (a near match
18    between the Jewett City and Clayville locations) and the water body at Waikakalaua Street near
19    Wahiawa, Oahu, Hawaii, had no other obvious inputs of Pb in our assessment than the point
20    sources within 20 km.  Both of those locations, however, are in "mixed" urbanized areas, and
21    therefore may also have historic Pb deposition from leaded gasoline and ongoing inputs of Pb to
22    sediments from erosion of soils contaminated by leaded  gasoline. A further discussion of
23    methodology for the sediment screen can be found in the risk  assessment report (ICF, 2006).

24         6.4.5.3  Primary Pb Smelter
25           A Characterization Area Investigation (CAI) was performed at the primary smelter
26    facility by ELM Consulting in 2005.  The investigation area included the smelter,  slag areas,  and
27    several haul roads within a 2.1  km radius from the facility as well as two "reference areas",
28    presumed to be outside the area of influence of the smelter,  6 to 7 km south of the facility.  The
29    area was evaluated for the potential for ecological impacts to soil, sediment, and surface water
30    from Pb originating from the facility. Data collected as part of the CAI were used.
31           To develop soil concentrations for this assessment, surface soil data were grouped into 3
32    geographic clusters: the west bank of Joachim Creek and two  "reference areas": Crystal City and
33    Festus Memorial Airport.  Surface water and sediment samples were taken from backwater and
34    low flow areas along Joachim Creek both upstream and downstream of the facility 800 m,  1.6
35    km and 3.2 km from the smelter. Additional samples were taken from the Mississippi River and
            December 2006                    6-29        Draft - Do Not Quote or Cite

-------
 1   a nearby pond. Details on the sampling methods used by ELM can be found in the risk
 2   assessment report (ICF, 2006).
 3          HQs calculated for each of the sampling clusters developed for this case study are
 4   provided here: soil results are listed in Table 6-5, surface water results are presented for Table 6-
 5   6, and sediment results are presented in Table 6-7.  HQs equal to or greater than 1.0 are bolded.
 6   All three of the soil sampling clusters (including the "reference areas") had HQs that exceeded
 7   1.0 for birds. The west bank of the Joachim Creek samples had HQs greater than 1 for plants
 8   and mammals also. The surface water sampling clusters all had HQs less than 1.0 as results were
 9   all below the detection limit of 3.0|ig/L.  However, three sediment sample clusters in Joachim
10   Creek (1,2, and 3) had HQs ranging from 1.0 to 2.2 and the U-shaped pond and one drainage
11   area had HQs greater than 3 but less than 5.

12   Table 6-5.  HQs for Soils  for Primary Pb Smelter Case Study.


Location of Sample Cluster

1 - West Bank of Joachim Creek
2 - Crystal City n
3 - Near Festus Airport n

HQ
for Plants

3.55
0.54
0.40
HQ

for Soil
Invertebrates
0.25
0.04
0.03

HQ
for Birds

11.19
1.70
1.28
HQ

for
Mammals
3.80
0.58
0.43
13   : Control samples taken outside perceived influence of the smelter.
14
            December 2006
6-30
Draft - Do Not Quote or Cite

-------
1   Table 6-6.  HQs Calculated for Surface Waters for Primary Pb Smelter Case Study.
Sample Location
and Cluster ID
HQ using CCC
(Chronic)
HQ using CMC
(Acute)
Joachim Creek
Cluster 1
Cluster 2
Cluster 3
Cluster 4
Cluster 5
0.39
0.40
0.39
0.41
0.41
0.02
0.02
0.02
0.02
0.02
Mississippi River
Upstream
Near Facility
Downstream
0.54
0.49
0.48
0.02
0.02
0.02
Emission Deposition
Cluster 1
CHRDDP
RRDP-02
DAMUP
0.69
0.24
0.47
0.40
0.03
0.01
0.02
0.02
3   Table 6-7.  HQs Calculated for Sediments in Surface Waters for Primary Pb Smelter Case
4              Study.
Location and
Cluster Sample ID
Hazard Quotient
Joachim Creek
Cluster 1
Cluster 2
Cluster 3
Cluster 4
Cluster 5
1.0
1.6
2.2
0.84
0.96
Mississippi River
Upstream
Near Facility
Downstream
0.41
0.84
0.34
Pond and Drainage Areas
U-shaped Pond
Cluster
ED1
ED2
4.8
3.1
0.41
          December 2006
6-31
Draft - Do Not Quote or Cite

-------
 1         6.4.5.4  Secondary Smelter
 2          For the secondary Pb smelter case study, as described in Section 4.2.3, two sets of
 3    modeled average Pb soil concentrations were used as exposure estimates for both the human
 4    health and the ecological risk assessment. The first set of concentrations was obtained by MPE
 5    modeling (Section 4.2.3.2). These modeled soil concentrations for the secondary Pb smelter were
 6    compared to empirical data obtained from a surrogate location. Based on this comparison, which
 7    suggested that modeled soil Pb concentrations for this case study might be significantly
 8    underestimated, we included a second characterization of soil  concentrations besides the purely
 9    modeled approach. Specifically, measurements from a surrogate secondary Pb smelter location
10    were used to "scale" up the modeled surface generated for this case study location to more
11    closely match the empirical  data obtained from the surrogate location (at specified distances
12    from the facility). The averages for 1-,  5-, or 10-km interval distances from the secondary Pb
13    smelter facility and the associated soil HQs calculated for each interval are presented in Table 6-
14    8.

15    Table 6-8.    HQs Calculated for Soils for Secondary Pb Smelter Case Study."
Distance Range
(m)

0
1000
2000
3000
4000
5000
10000
1000
2000
3000
4000
5000
10000
20000

0
1000
2000
3000
4000
5000
10000
1000
2000
3000
4000
5000
10000
20000
Modeled Soil Concentration
Datasets (mg/kg)
Total Pb Soil Cone, with
Background
86.6
20.7
17.3
16.3
15.8
15.4
15.1
Scaled 3x Total Pb Soil
Cone, with Background
260
62.3
51 .8
48.9
47,3
46.2
45.4
HQfor
Plants
HQ for Soil
Invertebrates
HQfor
Birds
HQfor
Mammals

0.72
0.17
0.14
0.14
0.13
0.13
0.13
0.05
0.012
0.010
0.010
0.009
0.009
0.009
2.28
0.54
0.46
0.43
0.42
0.41
0.40
0.77
0.18
0.15
0.15
0.14
0.14
0.13

2.17
0.52
0.43
0.41
0.39
0.39
0.38
0.15
0.037
0.030
0.029
0.028
0.027
0.027
6.84
1.64
1.36
1.29
1.24
1.22
1.19
2.32
0.56
0.46
0.44
0.42
0.41
0.41
16
17
18
19
20
21
      a HQ values greater than 1.0 are highlighted in bold type.

       The modeled soil concentrations within 1 km of the facility showed HQs of greater than
1.0 for avian wildlife. All soil concentrations for locations greater that 1 km from the facility
were associated with HQs less than 1.0 for this dataset. The three-times-higher-scaled soil
concentration dataset, developed based on soil data from similar locations, resulted in avian HQs
            December 2006
                                          6-32
Draft - Do Not Quote or Cite

-------
 1    greater than 1.0 for all distance intervals evaluated, including the farthest interval modeled, 10 to
 2    20 km from the facility. The scaled soil concentrations within 1 km of the facility also showed
 3    HQs greater than 1.0 for plants, birds, and mammals.
 4
 5
 6
 7
 8
 9
10
      6.4.5.5  Near Roadway Non-Urban Case Study
       Table 6-9 presents the HQ calculated for the Corpus Christi, Texas, near roadside soil
concentration data. HQs for birds were greater than 1.0 at all but one of the distances from the
road.  Mammalian HQs also were greater than 1.0 at the 2 m sampling distance from the
roadway.  Finally, plants also had HQs ranging from 2.83 and 5.42 at the 2 m distance. However
at the further distance from the roadway (4 m), birds and mammals still had HQs greater than 1.
11    Table 6-9.  HQs Calculated for Soils Near Roadway Non-Urban Case Study.
Sample location -
distance from
roadway
2m
2m
2m
4 m
Sample
depth
2.5 cm
10 cm
20 cm
2.5 cm
Total Pb
concentratio
n (mg/kg)
340
650
15
140
HQfor
Plants
2.83
5.42
0.13
1.17
HQfor Soil
Invertebrates
0.20
0.38
0.019
0.082
HQfor
Birds
8.95
17.1
0.395
3.68
HQfor
Mammals
3.04
5.80
0.13
1.25
12

13
14
15
16
17
18

19
20
21
22
23
24
25
26
27
28
      6.4.6  Discussion
       The results presented in this section of the document represent initial screening results for
the three case study locations and the national-level screen. These results are only indicative of
the potential for effects to terrestrial and aquatic systems from ambient Pb. It seems clear,
however, from this initial screening assessment that more refined analyses would be necessary in
order to characterize risk to various receptors from ambient Pb.

      6.4.7  Uncertainty and Variability
       This section addresses uncertainties and limitations associated with the primary Pb
smelter case study, the secondary Pb smelter facility case study, the near roadway non-urban
case studies and associated with the national-level screening for risks to aquatic organisms from
Pb deposition from air to surface waters. Note that limitations for the ecotoxicity  screening
values are described where they are introduced in Section 6.2.3.
       Uncertainties that apply across case studies include, but are not limited to, the following:

   •   The ecological risk screen is limited to specific case study locations and other locations
       for which dissolved Pb data were available and evaluated in the national-level surface
            December 2006
                                          6-33
Draft - Do Not Quote or Cite

-------
 1          water and sediment screen. Efforts were made to ensure that the exposure estimates were
 2          attributable to background level and air emissions of Pb; however, it is uncertain whether
 3          other sources might have actually contributed to the Pb exposure estimates.
 4
 5       •  A limitation to using the selected ecotoxicity screening values is that they might not be
 6          sufficient to identify risks to some threatened or endangered species or unusually
 7          sensitive aquatic ecosystems.
 8
 9       •  The database supporting the current AWQC for Pb  is over 20 years old.  There are data to
10          indicate that Pb bioconcentrates to some extent in invertebrates (e.g., bioconcentration
11          factors, or BCFs, of 500 to 1,700), and, to a lesser extent, in fish (e.g., BCFs of 42 to 45
12          in two species) in freshwater ecosystems. However, in 1984, data were insufficient to
13          estimate Final Tissue Residue Levels associated with adverse effects in fish, and thus the
14          BCFs did not influence the CCC value.  Also, EPA is evaluating whether pH may be a
15          better indicator of bioavailability compared to water hardness.
16
17       •  No adjustments were made for sediment-specific characteristics that might affect the
18          bioavailability of Pb in sediments in the derivation of the sediment quality criteria used
19          for this ecological risk screen.  Similarly, characteristics of soils for the case study
20          locations were not evaluated for measures of bioavailability.
21
22       •  Although the screening value for birds used in this analysis substituted more realistic
23          parameters for diet composition and assimilation efficiency, it was based on a
24          conservative estimate of the relative bioavailability of Pb in soil and natural diets
25          compared with water soluble Pb  added to an experimental pellet diet.  A recent site-
26          specific determination of a soil concentration protective of soil-invertebrate-consuming
27          birds suggested that the values of 38 mg/kg or even 83 mg/kg are  still  overly
28          conservative. This is possibly because the assimilation efficiency  of Pb in soils and
29          natural foods compared with the assimilation  efficiency of Pb acetate added to pelleted
30          diets is much less than 50 percent.
31

32         6.4.7.1   Primary Pb Smelter Case Study
33          The ELM Sampling and Analysis Plan (ELM, 2003) was designed to investigate possible
34    ecological risks from  all sources of Pb (and other contaminants) attributable to the primary Pb
35    smelter without a need to attribute the source of Pb in ecologically sensitive areas (ELM, 2003;
36    ELM, 2005).  For purposes of the Pb NAAQS review, it is  important to distinguish areas
37    impacted primarily from current or historic air deposition of Pb from areas impacted primarily
38    from other non-air sources (e.g. erosion  of mining waste piles, surface runoff from exposed
39    mining ores, direct waste discharges to water). While those areas impacted from these other
40    non-air sources are likely to be impacted from air deposition as well, it is not usually possible to
41    source apportion Pb in these areas. Therefore, these analyses attempt to focus on those areas
42    where it may be possible to identify effects from policy relevant sources.


            December 2006                     6-34         Draft - Do Not Quote or Cite

-------
 1          The soil sampling locations within a 2.1-km radius were all in areas that might have been
 2    subject to Pb inputs from Joachim Creek during flooding events. As such, the stations might not
 3    represent the concentrations of Pb in soils that result from air emissions from the smelter.  This
 4    limitation may overstate the risks from deposition of Pb emitted from the facility.

 5         6.4.7.2  Secondary Pb Smelter Case Study
 6          The ecological risk screen used modeled rather than measured media concentration data
 7    because measured data were not available for the case study location. Data were available for
 8    similar locations and these data were compared to the modeled results.  These results appeared to
 9    vary 3-fold therefore, scaled modeled data is also reported in this assessment. A full discussion
10    of the modeling steps can be found in Section 4.3.2. Fate and transport modeling limitations and
11    uncertainties are described in the risk assessment report (ICF, 2006).
12         6.4.7.3  Near Roadway Non-Urban Case  Study
13          Few measured data were available to evaluate ecological impacts of contaminated soils
14    near roadways in less developed areas where ecological receptors may be anticipated to occur.
15    The measured soil data for the Corpus Christi, Texas location 2 m from the roadway ranged from
16    15 mg Pb/kg at 20-cm depths to 650 mg Pb/kg at 10-cm depths. The Pb concentrations selected
17    at the Atlee, Virginia location ranged  from 17 mg/kg 15m from the roadway to 540 mg/kg 2 m
18    from the roadway; both samples were collected from 7.5-to 15-cm soil depths.  It is uncertain
19    how representative of other roadways these data are.
20          The soil concentration data were measured at sampling locations between 2 to 30 meters
21    away from intensely traveled roads, and the analysis did not evaluate the suitability of avian and
22    mammalian wildlife habitat in close proximity to roadways. Without this evaluation, it is
23    uncertain whether the assessment overestimates the ecological risks of Pb in roadway soils.
24          The assessment did not address surface water ecosystem impacts of Pb from near
25    roadway runoff of Pb contaminated soils. This may underestimate risks to aquatic receptors via
26    this exposure pathway.
27

28         6.4.7.4  National-Scale Surface Water Screen
29          The analysis revealed only two or three NAWQA sampling locations nationwide where
30    there appear to be risks to the  aquatic community from Pb that may  have originated from
31    atmospheric deposition. However, this is likely to be a large underestimate of the true number of
32    such sites for several reasons:
33
34       •   The NAWQA Study Units cover less than  50 percent of the land area of the United
35          States.

            December 2006                    6-3 5        Draft - Do Not Quote or Cite

-------
 1       •  Dissolved Pb was an analyte at only 16 percent of all NAWQA sampling locations.
 2       •  Dissolved Pb was measured only once or twice at many locations.
 3       •  For waters with a hardness of less than 47 mg/L as CaCOs, the CCC for dissolved Pb is
 4          less than the quantitation limit for dissolved Pb that was used until the fall of 2000 (i.e.,
 5          1 |ig/L).
 6       •  Fewer than 15 percent of samples analyzed for dissolved Pb between 1994 and 2004 were
 7          assessed with the lower quantitation limit of 0.08 |ig/L, which is a value that is
 8          sufficiently low to match the CCC for waters with a hardness as low as 4.7 mg/L CaCOs.
 9
10          The first two bullet points alone suggest that the number of such sites nationwide might
11    easily be at least ten times higher than what was represented in the NAWQA database.  In
12    addition, where the land use around a sampling location was classified as "mining," no
13    investigation was conducted to determine whether air  emissions from a nearby smelter might
14    also be contributing to the Pb in the water.
15          There are many sources of uncertainty in the results presented for the sampling locations
16    for which there were some data, including the following:
17        •  Many sampling locations are represented by only one or two measurements of dissolved
18           Pb.
19        •  The water hardness for some sampling locations was not measured or is represented by
20           only one or two measurements.
21        •  Where there are multiple measures of both dissolved Pb and water hardness at a given
22           location, no attempt was made to match sampling dates and times to develop time-
23           specific CCC values.1
24        •  The water hardness measured at some locations was less than the lowest value of 20
25           mg/L of CaCOs used to develop the equation to calculate a CCC.  The CCC equation is
26           not necessarily valid at values less than 20 mg/L CaCOs.
27        •  It is not known how quickly dissolved Pb concentrations changed at any of the locations.
28        •  The database  supporting the current AWQC for Pb is over 20 years old; new AWQC for
29           Pb may  be available in 2007.
30

31         6.4.7.5  National-Scale Sediment Screen
32          Results of this analysis cannot conclusively link any of the locations with probable
33    adverse effects of Pb in sediments on benthic communities to ongoing air emissions of Pb. This
34    analysis was limited to those 15 locations from the NAWQA database at which dissolved
35    concentrations of Pb in surface waters exceeded the chronic AWQC for Pb. Those 15  locations
36    are believed to represent a small fraction of surface waters in the U.S. for reasons given above.
            1 The coefficient of variation for water hardness measurements was less than 10 or 20 percent for many
      stations; however, at some locations, the coefficient of variation exceeded 50 percent, indicating higher fluctuations
      in water hardness measurements.

            December 2006                    6-3 6        Draft - Do Not Quote or Cite

-------
 1          An additional limitation is that where the land use around a sampling location was
 2    classified as "mining", no investigation was conducted to determine whether air emissions from
 3    a nearby smelter might also be contributing to the Pb in the water and sediments.  It was assumed
 4    that direct runoff and erosion from the mining sites to the surface waters would have contributed
 5    to the bulk of the Pb in sediments.
 6          Further limitations accrue from the sediment sampling data.  There were only nine exact
 7    matches and three near matches between the 15 surface water sampling locations  of interest and
 8    locations at which sediment samples also were analyzed.  Furthermore, there was a single
 9    sediment sample at each of the locations of interest, some of which were taken in  the early
10    1990s.
11          Finally, no adjustments were made for sediment-specific characteristics that might affect
12    the bioavailability of Pb in sediments in the derivation of the sediment quality criteria used for
13    this risk screen.

14         6.5   FUTURE ANALYSES
15          There are several expansions and refinements to this initial screening analysis that could
16    be considered for any future analyses. Additional case study locations could be identified,
17    particularly for near roadway scenarios, and a case study could be developed around the
18    ecologically vulnerable location identified in this draft. Development of more refined exposure
19    estimates for several of the case studies using ecosystem and habitat suitability models would
20    allow for exposure estimates that result in body burdens for target organisms which could be
21    directly compared to available concentration effects data. Broadening of the national-level screen
22    to focus on locations with known large air emissions would allow for a better estimate of media
23    concentrations in areas which are likely to be directly  influenced by ambient air concentrations.
24    Lastly, a more detailed discussion of the effect of Pb on ecosystem services  and a discussion of
25    research needs could be included.

26         6.6   THE SECONDARY LEAD NAAQS
27         6.6.1   Introduction
28          This first draft document discusses the general approach that is intended to be used in
29    considering the adequacy of the current standard and in identifying policy alternatives in the next
30    draft of this document. In addition,  the next draft will include key uncertainties and research
31    recommendations related to setting a secondary Pb standard.
32          The current secondary Pb standard is 1.5 ug Pb/m3,  as a maximum arithmetic mean
33    averaged over a calendar quarter, set equal to the primary standard (43 FR 46246). A final
34    decision should draw upon scientific information and  analyses about welfare effects, exposure
35    and risks, as well as judgments about the appropriate response to the range of uncertainties that

            December 2006                     6-3 7        Draft - Do Not Quote or Cite

-------
 1    are inherent in the scientific evidence and analyses. Our approach to informing these judgments,
 2    discussed more fully below, is based on a recognition that the available ecological evidence
 3    generally reflects a continuum consisting of ambient levels at which scientists generally agree
 4    that adverse ecological effects are likely to occur through lower levels at which the likelihood
 5    and magnitude of the response become increasingly uncertain.
 6           This approach is consistent with the requirements of the NAAQS provisions of the Act
 7    and with how EPA and the courts have historically interpreted the Act. These provisions require
 8    the Administrator to establish secondary standards that, in the Administrator's judgment, are
 9    requisite to protect public welfare. In so doing, the Administrator seeks to establish standards
10    that are neither more nor less stringent than necessary for this purpose. The Act does not require
11    that secondary standards be set at a zero-risk level but rather at a level that avoids unacceptable
12    risks to public welfare.

13         6.6.2  Approach
14           As indicated in Chapter 1, the policy assessment to be presented in the final version of
15    this document is intended to inform judgments required by the EPA Administrator in
16    determining whether it is appropriate to retain or revise the NAAQS for Pb.  In evaluating
17    whether it is appropriate  to consider retaining the current secondary Pb standard, or whether
18    consideration of revisions is appropriate, we intend to focus on the extent to which a broader
19    body of scientific evidence is now available that would inform  such decisions. As summarized
20    in section 5.2, the 1978 notice of final rulemaking (43 FR 46246) outlined key factors considered
21    in selecting the elements of a standard for Pb: the Pb concentration  (i.e., level); the averaging
22    time; and the form  (i.e., the air quality statistic to be used as a basis for determining compliance
23    with the standard).  Decisions on these elements were made only so far as to indicate that due to
24    a lack of relevant data at that time, the secondary standard should be set to be identical to the
25    primary standard.
26           In developing conclusions and identifying options for the Pb standard in this review, staff
27    intends to take into account both evidence-based and quantitative exposure- and risk-based
28    considerations. A series  of general questions will frame our approach to reaching conclusions
29    and identifying options for consideration by the Administrator as to whether consideration
30    should be given to retaining or revising the  current secondary Pb standard. Examples  of
31    questions that we intend to address in our review include the following:
32         •  To what extent has evidence of new effects and/or sensitive ecosystems become
33            available since the last review and to what extent are we able to characterize these
34            effects?
35         •  To what extent does newly available information support  or call into question any of
36            the basic  elements of the current standard?

             December 2006                     6-3 8        Draft - Do Not Quote or Cite

-------
 1         •   Is there evidence of associations, especially likely causal associations, in areas that
 2             meet the current standard? What are the important uncertainties associated with that
 3             evidence?
 4
 5           To the extent that there is support for consideration of a revised standard, we will then
 6    identify ranges of standards (in terms of an indicator, averaging time, level, and form) that would
 7    reflect a range of alternative public welfare policy judgments, based on the currently available
 8    information, as to the degree of protection that is requisite to protect public welfare.
 9           As noted in Chapter 1, staff will also evaluate removing Pb from the criteria pollutant list
10    and assess whether revocation of the Pb NAAQS is an appropriate option for the Administrator
11    to consider.  Section 108 of the Clean Air Act states that the Administrator "shall, from time to
12    time ... revise a list which includes each pollutant -
13           (A) Emissions of which, in his judgment, cause or contribute to air pollution which may
14           reasonably be anticipated to endanger public health or welfare;
15           (B) The presence of which in the ambient air results from numerous or diverse mobile or
16           stationary sources; and
17           (C) For which air quality criteria had not been issued before December 31,  1970, but for
18           which he plans to issue air quality criteria under this section."
19           In evaluating such an option, staff expects to consider, among other things, many of the
20    same issues identified earlier in the section. Information about the kinds and types  of sources of
21    Pb emissions, as well as the quantities of emissions from those sources will also be important for
22    consideration.
            December 2006                     6-39        Draft - Do Not Quote or Cite

-------
  1    REFERENCES

  2    Alabama Natural Heritage Inventory. (2001) Alabama Inventory List: The Rare, Threatened and Endangered
  3            Species, Animals and Natural Communities of Alabama. Montgomery, AL: Privately printed by the
  4            Alabama Natural Heritage Program.

  5    Beyer, W. N.; Audet, D. I; Heinz, G. H.; Hoffman, D. I; Day, D. (2000) Relation of waterfowl poisoning to
  6            sediment lead concentrations in the Coeur d'Alene River basin. Ecotoxicology 9: 207-218.

  7    Borgmann, U.; Couillard, Y.; Doyle, P.; Dixon, D. G. (2005) Toxicity of sixty-three metals and metalloids to
  8            Hyalella azteca at two levels of water hardness. Environ. Toxicol. Chem. 24:  641-652.

  9    Choctawhatchee Pea and Yellow Rivers Watershed Management Authority. (2006) About the Watersheds: The
10            Rivers.

11    Cook, M.R. and Kopaska-Merkel, D.C. (1996) Hydrologic Characterization of the WaterResources of the
12            Choctawhatchee-Pea Rivers Watershed - Phase I. Tuscaloosa, AL: Geological Survey of Alabama. 137p.

13    ELM Consulting LLC. (2003) Sampling Analysis Plan: Characterization Area Investigation: The Doe Run Company
14            Lead Smelter, Herculaneum, Missouri (Version 1). Prepared for the Doe Run Company. March.

15    ELM Consulting LLC. (2005) Characterization Area Investigation Report. July.

16    ICF International. (2006) Lead Human Exposure and Health Risk Assessments and Ecological Risk Assessment for
17            Selected Areas. Draft Technical Report to EPA. December.

18    MacDonald, D.D., Ingersoll, C.G., and Berger, T. A. (2000) Development and evaluation of consensus-based
19            sediment quality guidelines for freshwater ecosystems. Archives of Environmental Contamination and
20            Toxicology. 39:20-31.

21    MacDonald, D.D., Ingersoll, C.G., Smorong, D.E., Lindskoog, R.A., Sloane, G., and Biernacki, T. (2003)
22            Development and Evaluation of Numerical Sediment Quality Assessment Guidelines for Florida Inland
23            Waters. British Columbia: MacDonald Environmental Sciences, Lt. Columbia, MO: U.S. Geological
24            Survey. Prepared for: Florida Department of Environmental Protection, Tallahassee, FL. January.

25    Nature Serve. (2006) Nature Serve Explorer: An Online Encyclopedia of Life [Web Application]. Version 5.0.
26            Arlington, VA and Pike County, AL: NatureServe. Available at http://www.natureserve.org/explorer.

27    Outdoor Alabama. (2003)  Listing of Alabama Species on the Federal List for Threatened and Endangered Species or
28            Whose Status is a Concern. Available at http://www.outdooralabama.com/watchable-
29            wildlife/regulations/endangered-species.cfm.

30    Speiran, O.K. (1998) Selected heavy metals and other constituents in soil and stormwater runoff at the Interstate 95
31            Interchange near Atlee, Virginia, April  1993-May 1997. U.S. Geological Survey. Water-Resources
32            Investigations Report 98-4115, 39p.

33    Turer, D.G. and Maynard, J.B. (2003) Heavy metal contamination in highway soils. Comparison of Corpus Christi,
34            Texas and Cincinnati, Ohio shows organic matter is key to mobility. Clean Technologies and
35            Environmental Policy. 4(4):235-245.

36    U.S. Environmental Protection Agency (USEPA). (1984) Ambient Water Quality Criteria forLead -1984.
3 7            Washington, DC: Office of Water Regulations and Standards, Criteria  and Standards Division. EPA 440/5-
38            B4-027. Available at
39            http://vosemite.epa.gov/water/owrccatalog.nsf/065ca07e299b464685256ce50075clla/05edf009d
40            cdeld7d85256b060072307e!OpenDocument.


              December 2006                        6-40         Draft - Do Not Quote or Cite

-------
  1    U.S. Environmental Protection Agency. (1990) Review of the national ambient air quality standards for lead:
  2            assessment of scientific and technical information: OAQPS staff paper. Research Triangle Park, NC: Office
  3            of Air Quality Planning and Standards; report no. EPA-450/2-89/022. Available from: NTIS, Springfield,
  4            VA;PB91-206185.

  5    U.S. Environmental Protection Agency (USEPA). (1993) Office of Water Policy and TechnicalGuidance on
  6            Interpretation and Implementation of Aquatic Life Metals Criteria: A Memoradum by Martha G. Prothro.
  7            Washington, DC: U.S. Environmental Protection Agency, Office of Water. October. Available at
  8            http://www.epa.gov/waterscience/library/wqcriteria/metalsinterpret.pdf.

  9    U.S. Environmental Protection Agency (USEPA). (\99%) Methodology for Assessing Health Risks Associatedwith
10            Multiple Pathways of Exposure to Combustor Emissions. Cincinnati, OH: Update to EPA/600/6-90/003,
11            EPA/NCEA (EPA 600/R-98/137). Available at
12            oaspub.epa.gov/eims/eimscomm.getfile ?p_download_id=427339.

13    U.S. Environmental Protection Agency (USEPA). (2005a) Guidance for Developing Ecological Soil Screening
14            Levels. Washington, DC: Office of Solid Waste and Emergency Response.OSWER Directive 9285.7-55.
15            November.

16    U.S. Environmental Protection Agency (USEPA). (2005b) Ecological Soil Screening Levels forLead, Interim Final.
17            Washington, DC: Office of Solid Waste and Emergency Response. OSWER Directive 9285.7-70.
18            Available at http://www.epa.gov/ecotox/ecossl/pdf/eco-ssl lead.pdf.

19    U.S. Environmental Protection Agency (USEPA). (2005c) Procedures for the Derivation of Equilibrium
20            Partitioning Sediment Benchmarks (EBSs) for the Protection ofBenthic Organisms:Metal Mixtures
21            (Cadmium, Copper, Lead, Nickel, Silver, and Zinc). Washington, DC: Office of Research and
22            Development. EPA-600-R-02-011.

23    U.S. Environmental Protection Agency (USEPA). (2006) Air Quality Criteria for Lead, Volume I and II ofII.
24            Research Triangle Park, NC: National Center for Environmental Assessment - RTF Office. EPA/600/R-
25            5/144aF. October. Available at: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid= 158823

26    U.S. Geological Survey. (2004) National Water Quality Assessment Program. Available at
27            http://water.usgs.gov/nawqa/data.

28    Vogelmann, J.E., Howard, S.M., Yang, L., Larson, C.R., Wylie, B.K., and Van Driel, J.N. (2001) Completion of the
29            1990's National Land Cover Data Set for the conterminous United States. Photogrammetric Engineering
30            and Remote Sensing. 67:650-662. Available at http://edc.usgs.gov/products/landcover/nlcd.html.

31    Young, T. F.; Sanzone, S., eds. (2002) A framework for assessing and reporting on ecological condition: an SAB
32            report. Washington, DC: U.S. Environmental Protection Agency, Science Advisory Board; report no.  EPA-
33            SAB-EPEC-02-009. Available: http://www.epa.gov/sab/pdf/epec02009.pdf [9 December, 2003]
              December 2006                        6-41          Draft - Do Not Quote or Cite

-------
United States                              Office of Air Quality Planning and Standards              Publication No. EPA 452/P-06-002
Environmental Protection                   Air Quality Strategies and Standards Division                               December 2006
Agency                                          Research Triangle Park, NC

-------