&EPA
EPA/600/R-11/048
   United States
   Environmental Protection
   Agency




    The DEARS Data Analysis Workshops:
    Summary of Findings and Discussions
  Office of Research and Development
  National Exposure Research Laboratory


-------

-------
 The DEARS Data Analysis Workshops

    Summary of Findings and Discussions
            Ron Williams
National Exposure Research Laboratory
  Office of Research and Development
 U.S. Environmental Protection Agency
   Research Triangle Park, NC 27711

-------
                                       Notice

The information in this document has been funded in whole by the U.S. Environmental
Protection Agency.  It has been subjected to the Agency's peer and administrative review and has
been approved for publication as an EPA document. Mention of trade names or commercial
products does not constitute endorsement or recommendation for us.

-------
                                  Acknowledgements

The Human Exposure and Atmospheric Sciences Division in the National Exposure Research
Laboratory thank the Steering Committee and volunteer workgroup participants for their time
and contributions to the workshop.  It thanks the NERL administrative support staff who
provided logistical support in conducting the workshop and in the production of this summary.
                                          in

-------
                              Table of Contents

Notice	ii
Acknowledgements	iii
1. Background	1
1.1 Overall Goal	2
1.2 Specific Objectives	2
1.3 Approach	2
2. Steering Committee and Workgroups Pre-Part 1 Workshop Discussion	4
2.1 Exposure Assessment and Uncertainty Workgroup	4
2.2 Human Exposure Modeling Workgroup	4
2.3 Atmospheric and Environmental Modeling Workgroup	4
2.4 Health and Epidemiology Workgroup	5
3. Part 1-Workshop Event	5
4. Part 2-Workshop Event	8
5. Common Needs	8
6. Workshop Conclusions  	9
7. Attachments	11
Attachment 1-Draft Research Questions	12
Attachment 2-Data Analysis Workshop Part 1 Workgroup Members	16
Attachment 3-Data Analysis Workshop Part 1 Agenda 	17
Attachment 4-Data Analysis Workshop Part 1 Invited Attendee List	18
Attachment 5-DEARS Study Objectives Progress Report Presentation	19
Attachment 6-Exposure Assessment and Uncertainty Workgroup Summary	39
Attachment 7-Human Exposure Modeling Workgroup Summary	44
Attachment 8-Environmental and Atmospheric Modeling Workgroup Summary 	48
Attachment 9-Health and Epidemiology Workgroup Summary	52
Attachment 10-Common Needs Summary	55
Attachment 11 -Data Analysis Workshop Part 2 Agenda	57
Attachment 12-Overall Workgroup Summaries	58
Attachment 13-DEARS Publication Plan	60
Attachment 14-Common Needs Priorities	62
                                        IV

-------
                      The DEARS Data Analysis Workshops

                          Summary of Findings and Discussions

                                Organizational Sponsor

U.S. Environmental Protection Agency, Office of Research and Development (ORD)

                               Chair: Ron Williams, (ORD)
  Steering Committee: Val Garcia (ORD), Karen Wesson (OAQPS), Jennifer-Richmond Bryant
                              (NCEA), Rich Cook (OTAQ)


1.     Background

The Detroit Exposure and Aerosol Research Study (DEARS) was a three-year study conducted
by the U.S. Environmental Protection Agency through the Office of Research and Development's
National Exposure Research  Laboratory (EPA/ORD/NERL).  Field  data collections were
completed  in 2007 and validation of the primary datasets was completed in late 2008. The
study's primary goal was to investigate the relationship of select air pollutant concentrations and
their sources measured at community air monitoring stations in comparison to those measured in
various neighborhoods in Wayne County,  Michigan. Six primary data analysis objectives were
components of the original study design  (www.epa.gov/dears).  These original  data  analysis
objectives were to: (1) characterize spatial and temporal relationships between pollutants, (2)
determine human exposure factors, (3) determine environmental exposure factors, (4) develop
enhanced human exposure models, (5) establish source contributions, and (6) investigate multi-
pollutant (particulate matter/gases/semi-volatiles) relationships.

The study was designed to significantly contribute to our understanding of how well air quality
information collected at community monitors accurately reflects  what neighborhoods and the
individuals living in these neighborhoods are exposed to  every day. It would provide needed
information on defining what factors affect an individual's exposure to various particulate matter
and air toxic sources. Exposure-related data from the  DEARS can be divided into five main
parts:  (1)  personal  monitoring,  (2) residential  indoor monitoring, (3) residential  outdoor
monitoring, (4) monitoring  performed at a central community  site, and (5)  survey information
related to environmental and human exposure factors.  The personal and residential monitoring
involved a  total of 145 participants over a three-year period of data collection. A total of nearly
36,000  individual  24-hr based exposure  measurements involving  particulate matter,  criteria
pollutant gases and other air pollutants of interest were obtained during the field measurements.

NERL and its collaborators have been actively analyzing  data to support the six original data
analysis objectives.   A detailed  summary  of the  six  objectives, progress to date,  ongoing
analyses, and current data summarization products was  recently developed and made available
on  the  DEARS website (www.epa.gov/dears/findings.html).   This summary was useful  in
assisting the NERL in determining the overall progress being made on the study as a whole as
well as gaining a perspective on study areas showing potential for unanticipated research benefits

-------
to ORD and its stakeholders. These stakeholders include among others, the National Center for
Exposure Assessment (NCEA), the Office of Air Quality, Planning and Standards (OAQPS), the
Office of Transportation and Air Quality (OTAQ), US EPA Region 5, and the National Health
and Environmental Effects Research Laboratory (NHEERL). NERL has been actively sharing
DEARS data with both internal and external  collaborators in  pursuit of the study objectives.
Significant progress had been made on data analyses during the 2008-2010 calendar periods and
information updating key collaborating institutions of these analyses would be beneficial to all
parties. A workshop setting was proposed where DEARS progress-to-date could be shared and
where potential new uses of exposure data to meet critical EPA needs could be explored.

1.1     Overall Goal

The goal of the effort was to develop a multi-institutional review of critical exposure-related data
gaps and determine the potential value of the DEARS in meeting those needs.

1.2     Specific Objectives

The DEARS Principal Investigator had a charge to:

   •   Develop a multi-institutional steering group interested in exposure-related research to
       guide overall workshop development and summarization,
   •   Establish workgroups responsible for  identifying  exposure-related data  gaps  in four
       primary research areas: (1) measurement assessment and uncertainty, (2) human exposure
       modeling,   (3)   atmospheric  and  environmental  modeling,  and  (4)  health  and
       epidemiology,
   •   Plan,  schedule  and  conduct a  series  of developmental meetings to accomplish the
       objectives above, ultimately resulting in one or more workshops each involving a report-
       out of findings-to-date, and
   •   Develop a summary (report) on the workshops and the ability of DEARS to meet the
       identified exposure-related data needs.
1.3    Approach

Senior scientists from the Office of Air Quality Planning and Standards (Karen Wesson/OAQPS),
the National Center for Exposure Assessment (Jennifer Richmond-Bryant/ NCEA), the Office of
Transportation and Air Quality (Rich Cook/OTAQ), and the Atmospheric Modeling and Analysis
Division (Val Garcia/NERL)  were  contacted by  the DEARS  Principal  Investigator (Ron
Williams/NERL) and invited to participate on the Steering Committee.  The Steering Committee
had the responsibility of examining the four draft exposure research areas of potential  mutual
interest proposed by NERL. These research areas were:

   (1) exposure assessment and uncertainty,
   (2) human exposure modeling,
   (3) atmospheric and environmental modeling, and
   (4) health and epidemiology.

-------
To facilitate this effort, the Steering Committee developed strawman exposure data gap questions
within each of the four areas (Attachment 1). The Steering Committee recruited other Agency
staff members who could provide expertise  on the data  gap issues  and act as Workgroup
members.   To the greatest extent possible, multi-institutional Workgroup member rosters were
developed for each of the four research areas being examined. A total of 23 Workgroup members
were invited to participate in discussions concerning one or more research areas.  A listing of the
Steering Committee and invited Workgroup members is provided (Attachment 2).

A series of meetings involving the Steering Committee and the various Workgroups were held
between August  19th and September 29th, 2010.  During this time, "champions", or individuals
who would summarize  individual Workgroup discussions, were named. These individuals had
the  responsibility of reporting back to the Steering Committee  what exposure-related data gaps
existed within each research area and then summarizing  these findings  in a common slide
presentation format.  A report-out by these  champions was  then presented to the Steering
Committee and the other Workgroups to gain input from all involved.

Slide sets meeting the Steering Committee's approval  were developed in anticipation of an
invitation-only workshop event to be sponsored by the NERL (Part 1). The Acting Director of the
NERL's Human Exposure and Atmospheric Sciences Division (Roy Fortmann) made personal
contact with all invitees who he believed would benefit from the workshop discussion as well as
provide valuable input to NERL  on the role exposure science must play  in advancing the
Agency's goals. This event was held on October 19th, 2010 on the EPA-RTP campus. The agenda
for  this event is provided (Attachment 3) as well  as the list of invited attendees (Attachment 4).
This event had the goal of providing a summary of the purpose of the DEARS Data Analysis
Workshops, defining the charge given to the Steering Committee and Workgroups, reporting on
the  DEARS overall study progress to date (Attachment 5) and providing a report out on each of
the  four research areas.  In particular, the champion of each research area had the responsibility
to report on exposure data gap issues of importance to the Agency. The individual presentations
associated with these reports are provided (Attachments 6-9). In addition,  a "Common Needs"
summary was reported (Attachment 10).  This summary reviewed information obtained from all
of the research area discussions. In doing so, it provided a simple tabular means of determining
what exposure data gaps were viewed as essential across the four research areas.

The last segment of the Part  1 workshop event focused on encouraging all  present to engage in
future discussions concerning how the DEARS data might have the potential to meet the agreed
upon  data gaps.  The ensuing November 2010 discussions  held between  the study's Principal
Investigator (Ron  Williams),  the Steering Committee,  and  all  Part  1 invitees  provided
information ultimately  reported out  during a November 30th  workshop event (Part  2).  The
agenda, summary presentations associated with each of the four critical research areas, DEARS
publication plan, and evaluation of the usefulness of the DEARS  to potentially meet specific
exposure data gaps associated with this event are included as Attachments 11-14.

-------
2.     Steering Committee and Workgroups Pre-Part 1 Workshop Discussions

The draft exposure research areas and strawman data gaps (Attachment 1) provided by NERL to
facilitate the  discussions with  the  Steering Committee proved  to  be  invaluable. Further
discussions indicated that the strawman data gaps should be refined by each of the subsequently-
developed Workgroups  rather than the Steering  Committee  itself.   This decision ultimately
resulted in very focused and highly productive meetings  concerning each  of the four research
areas. Partial summaries of these discussions are reported below.

2.1    Exposure Assessment and Uncertainty Workgroup:

This Workgroup considered what air pollutants were of interest to their  respective organizations.
Particulate matter (PM) and its component species were  of high interest as well as the multi-
pollutant environment consisting of PM, air toxics, criteria pollutant gases and semi-volatiles.
Workgroup  members were interested not only in data gap questions pertaining to pollutant
concentrations  but  also  in how  pollutants  related to one  another  in  time  and  space
(spatial/temporal).   The degree  of error involved in  obtaining pollutant measurements was
discussed as well as issues related to understanding source impacts.  Concerning source impacts,
Workgroup members raised questions as to how one might summarize source characterizations
using surrogates or  select species as identifying tools.  The  value of non-ambient pollutant
measurements (e.g., personal, residential) were issues deemed worthy of examination.

2.2    Human Exposure Modeling Workgroup:

These  individuals  examined a series  of draft  questions  concerning what pollutant data was
viewed as critical for development or evaluation of select models (e.g., PMSHEDS). This group
was asked to provide feedback on needed or perceived data needs ranging from PM (various size
fraction), VOCs and air toxics.  Draft  questions included issues  involving human activity and
environmental exposure factors.  Of particular interest to the steering committee were the inputs
needed to help advance the Exposure Model for Individuals (EMI), which NERL plans to use in
ongoing near-road field  studies.  The modelers comprising this workgroup believed that a more
systematic approach was needed to determine the state of the  science for this given area rather
than examining the draft questions.  Therefore, the majority of the discussions this group held
focused on defining what they collectively felt was the current state-of-the-science pertaining to
advancing human exposure modeling.

2.3    Atmospheric and Environmental Modeling Workgroup:

This Workgroup considered what exposure data was  needed  to  advance current models  (like
CMAQ). Specific  components of their discussion indicated  that multi-pollutant issues were
viewed as  critical  and that those  involved  in modeling  needed  extensive  spatial/temporal
exposure datasets.  This included draft questions involving meteorology and the need for micro-
scale  source information.  Two key discussions points included the need for  data collections
involving both high frequency and high quality data. In addition, this workgroup decided that a
potentially beneficial approach concerning the subject matter was not a direct examination of the
draft questions, but rather an extensive discussion on what exposure monitoring data was needed

-------
to support CMAQ and other model development.

2.4    Health and Epi demi ol ogy Workgroup:

Members were initially asked to consider draft questions pertaining to what critical needs exist
following the most recent NOx/SOx, PM, and O?, Integrated Science Assessments (ISAs). The
role of multi-pollutant exposure issues was raised as well as how important it is to understand
exposures related to non-ambient spatial settings.  Questions related to the PMcoarse (PMio-2.s) size
fraction were proposed, including whether there exists sufficient interest on the role of various
non-ambient  PMcoarse sources  (e.g.,  residential  indoor) to  warrant further discussion.  Draft
questions pertaining to the value of volatile organic compounds (VOCs) were examined as well
as the value of exposure data of time  durations significantly smaller than the normal 24-hr based
metrics often  encountered.
Information obtained during deliberations of each Workgroup was then presented to the Steering
Committee for feedback.  The goal of the Steering Committee during this review was not to
change the summary from the respective Workgroup, but help each group gain a perspective on
the findings of all  other  groups and areas of collective agreement. An  approach (style) of
presentation was formalized and each Workgroup was asked to develop a slide set to be used for
the October 19th meeting (Part 1 event).
3.     Part 1- Workshop Event

An invitation only Workshop reporting event (Part 1) was held on October 19, 2010 on the EPA-
RTP campus. Following introductions and other housekeeping activities, the DEARS Principal
Investigator (Ron Williams) and select DEARS team members  (Janet Burke and Gary Norris)
provided a detailed description of the current progress of the study (Attachment 5).  This was
viewed by the  Steering Committee as  an essential part of the  Workshop activities because it
would inform all participants about the extensive exposure data collected in the DEARS as well
as progress associated with the six primary (original) data analysis objectives.  The latter was
viewed as being extremely helpful  in  helping   Workgroup members  understand what data
findings were already available as  well as the focus of current  or near-future planned analysis
activities.  The DEARS Data Analysis Progress presentation was  divided into a series of sub-
sections.  Each  sub-section provided information on one of the  six data  analysis objectives,
including: (1) data analysis progress during  the 2008-2010 time periods, (2) select  results
associated with one or more components, and (3) tabular and/or graphical examples of findings
highlighting results of interest to Workshop invitees. Examples  of information provided during
this component of the Part 1 event  included a thorough discussion of the DEARS study design,
the  types  of  pollutants  measured   and  their  frequency,  and  the   types  of human
activity/environmental exposure factor survey materials obtained.

DEARS  Objective #1  highlights  provided   a  discussion  of current exposure  assessment
collaborations, the spatial and temporal variability of select exposure measures and the observed
heterogeneity  of many of the pollutants  with  respect to  their  relationship with  a central

-------
community monitor.   Progress  reported for DEARS objectives #2 and  #3  included findings
relating the impact of indoor sources on total personal exposures to PM and the role of exposure
factors on observed residential air exchange  rates in the DEARS homes. Proximity effects from
localized line sources (near-roadway and stationary  mobile source emissions) were described.
The impact of ambient versus non-ambient  source effects and their role in personal exposures
were discussed. The DEARS survey materials and real-time personal PM monitoring appears to
have value in examining these issues.

Human exposure  modeling progress  (DEARS  objective  #4) detailed work investigating
meteorological impact on PM spatial relationships and expected DEARS inputs for the planned
EMI-NEXUS model. The ongoing collaborative work involving NERL and OAQPS researchers
and the CMAQ model was reported.  In brief, this effort will  utilize validated DEARS outdoor
exposure data as a means to compare CMAQ modeling output associated with the 2005 Detroit
airshed.  The comparability of National Air Toxics Assessment (NATA) modeled VOC results
with actual DEARS measures have been performed  and provided to OAQPS.  These findings
indicate that modeled human exposures to benzene might be underestimating total air exposures
to this pollutant due to non-ambient source impacts.

Significant progress has been  made on obtaining detailed speciation data needed to support
source allocation distributions for the DEARS airshed (Objective #5). Findings associated with
source impacts on the central community monitoring site were  reported. A key component of the
ongoing work is trying to understand the impact of biomass-related sources on the airshed. Data
from various marker species (e.g., potassium, levoglucosan, polynuclear aromatic hydrocarbons)
have been obtained using extensive laboratory analyses.  These inorganic and organic molecular
markers are being examined for their spatial  and temporal variability with respect to the airshed.
The acceptability of various markers to be used in source assessment activities represents a key
component of  this effort as it has broad value to  not only the DEARS  but the  source
apportionment community as a whole.

DEARS objective #6 (multi-pollutant relationships),  represented an area of high interest across
all of the Workgroups.  Progress to date indicated  that the  survey information  and exposure
monitoring data obtained in the DEARS (e.g., PM, VOC, criteria pollutant gases, questionnaires)
had great potential for providing advances in this research area. Findings indicated that use of
canonical correlations as well as mixed modeling analyses had  already elucidated the role human
and environmental exposure factors  play on total personal  exposures to numerous pollutant
groups.  As such, data analyses indicated  that non-ambient  NO2  and VOC  sources severely
limited the agreement between total personal exposures to these pollutants and ambient-based
measurement systems  (i.e.,  central  community  monitoring).  It  was  reported  that analyses
examining the spatial relationships between  various multi-pollutant groupings with other select
measures (e.g., VOCs, outdoor elemental carbon, NO2) are ongoing and the initial results were
described.

An additional objective of the  DEARS data analyses reported on during the Workshop Part 1
event was the  epidemiological  investigation of select environmental  exposures  and health
outcomes associated with a companion study performed by the University  of Michigan (Dr. Rob
Brook). This collaborative effort has already provided unique  findings including the role of PM

-------
of non-ambient  origin on  select cardiovascular outcomes and the observance  of heart rate
changes following personal exposures to fine particulate matter.  Speciation data has provided
the means to examine the multi-pollutant environment associated with some personal exposure
scenarios.  Personal NO2 exposures would appear to be more highly  associated  with reduced
brachial artery diameters as compared to select elements (iron, potassium) and their influence on
blood pressure changes. The DEARS exposure datasets and the companion health  outcome data
have the potential to examine a wide range of epidemiological research issues.

Each of the four Workgroups provided short summaries of their consensus findings on data gaps
(Attachments  6-9). The  Exposure Assessment and  Uncertainty  Workgroup indicated that
monitoring data  on lead (Pb) was of the highest priority due to an upcoming  Pb  ISA (draft in
2011).  The second greatest exposure-related need was the need for  summarized and interpreted
information  pertaining to the multi-pollutant environment.  The Human Exposure Modeling
Workgroup indicated that a number of critical exposure data gaps existed. These included single
and  multi-pollutant  concentrations and relationships,  the  impact of source(s)  on  human
exposures,  detailed human and  environmental  exposure factors, data sets containing  widely
varying temporal  and spatial  scales,  and  obtaining the  determinants  for  PMSHEDS and
potentially other models  supporting the national air quality  standards.  In similar fashion, the
Environmental and Atmospheric Modeling  Workgroup identified  data gaps associated  with
spatially dense and high temporal frequency collections of PM and  other pollutants. In essence,
this group was identifying the need for saturation-style collections  in a given location and that
there was a need for this monitoring to be extended over long temporal periods.  High quality
meteorology data was identified as being highly  valued.  This included the need for micro-scale
data  collections  to capture  key topographical features.   Data gaps  associated with identifying
local emission sources was reported as well as technically challenging efforts to provide data on
nitrogen cycling, free radical production and biogenic contributions to local  and regional air
sheds.  The Health and Epidemiology Workgroup identified  several research areas of interest.
These  included the need for more integrated exposure and health outcome  research beyond
cardiovascular and respiratory effects,  and mortality.  This Workgroup indicated that focused
human exposure panel studies involving select subpopulations were needed and that the multi-
pollutant environment  should be incorporated in development of  future study designs. High
quality exposure data  were viewed as  being critical in helping to advance epidemiological
research

One of the key products the steering committee produced and reported  at the Part  1 event was a
"Common Needs" document (Attachment 10). It quickly became apparent as the individual slide
sets were being reviewed for the Part 1  event that a  pattern of  commonality  often  existed
between a given exposure data need (e.g., Os &  PM and multi-pollutant measures) and the four
independent Workgroups. To facilitate discussion, a collective total of twenty exposure data gaps
were identified and a tabular format used to list individual Workgroup recommendations for such
data. Results of the Common Needs table were shared at the  Part 1  event and indicated that the
greatest number of identified needs were in the exposure assessment and uncertainty area (needs
in 16/20 critical data gaps).

The Part 1  event was  closed out by inviting all  attendees to  participant in an evaluation of the
ability  of DEARS data to meet some of the exposure gaps being reported. This discussion would

-------
take place in scheduled meetings involving the Steering committee, the various Workgroups, and
all interested parties, in early November 2010.

4.     Part 2- Workshop Event

The final DEARS data analysis workshop event (Part 2) was conducted on November 30, 2010.
These deliberations examined the DEARS study design and its reported (ongoing and planned)
data analysis objectives with each of the data gaps identified during the Part 1 effort. In doing
so, we would determine the extent to which DEARS data: (1) had already provided some benefit,
(2)  was expected to provide benefit based upon its  planned data analysis scheme and available
data, (3) could provide benefit but not currently targeted due to resource management, or (4) did
not have the ability to be of benefit due to the lack of the specific exposure data requirements.
To facilitate this effort, a summary slide set (Attachment 12) was developed that reiterated the
critical data needs  of the various working groups.  In  additional,  the DEARS publication plan
(Attachment 13), a constantly  evolving strategy first developed in 2007, was revised to highlight
the  latest efforts of NERL researchers and collaborators with respect to peer review publication
status.  All attachments were either presented or made available at the Part 2 event.

5.     Common Needs

The primary document provided to Part 2 attendees was a revised "Common Needs" document
(Attachment 14). This document now contained new information pertaining to: (1) the degree of
agreement between the various  subcommittee  working groups on each individual data gap, (2)
specific outputs determined to be needed (publications, databases, models), (3) the date by which
the  needed outputs were required to make an impact  on the science,  (4) the extent to  which
DEARS data might be of value in meeting the specific outputs, (5) specific descriptions of the
outputs the DEARS might  be able to provide, and (6) the current progress pertaining  to the
original six DEARS data analysis  objectives and  specific publications, models or databases
identified  as being  needed. The  Steering Committee and all subcommittee working groups were
unable to agree upon a priority of critical data needs. However, a simple listing of the number of
groups identifying  a given data need did provide a pseudo-ranking of the collective thought. In
particular, data needs associated with the following areas were viewed as highly important by all
groups:

    1.  high frequency spatial and  temporal measurement data (publications and  a database
       useful for modeling),
    2.  multi-pollutant measures and establishment of spatial/temporal relationships between and
       among these co-pollutants (publications and a database for modeling),
    3.  regional and local emission source identification (including indoor  settings) and the
       determination of key species making up each  source (publications and a database for
       modeling),
    4.  measurement data of short time resolution (publications and a  database for  modeling),
       and
    5.  air toxic measurement and relationship data pertaining to hazardous air pollutants (HAPs)
       including but not limited to VOCs. The association of these HAPs with health effects
       represents an area of immediate need (publications and a database for modeling).

-------
While the remaining 15 data gaps did not receive unanimous recommendations as key research
areas, this  should not be taken as a downplaying of their importance to  specific Workgroup
members or the needs  of the Agency as a whole.  For example, only two groups of the four
groups reported the need for Pb data (publications and/or a database) involving human exposures
to be available by the  summer of 2011.  This time  frame was needed to meet the  Pb ISA
publication inclusion criteria.  Therefore, each of the research  data gaps  needs to be viewed
accordingly. In this light, the overall workshop summary document  (Attachment 13) might best
serve as a review of data needs as they relate to important timelines rather than just specific data
needs.

A review of the current DEARS data analysis progress (pertaining  to its original six exposure
data analysis areas and  the additional collaborative epidemiological efforts) indicates progress
supporting fourteen of the twenty identified data gap needs.  Much of this progress relates to the
development of databases which have  been used  to support peer  review journal  article
development.  Examples of DEARS benefits to  the  state-of-the-science include publications
pertaining to the multi-pollutant environment, PMcoarse exposure assessment, the  impact of
various source settings on personal exposure assessment, and the association of PM and gaseous
co-pollutant species on observed human health outcomes.  There are, however,  areas where the
DEARS will provide little or  no benefit to  the reported data gaps.  These include needs
associated  with extensive  meteorological monitoring, high  frequency (short-time duration)
pollutant monitoring,  nitrogen cycling, and data needs from long duration  monitoring  at  a
consistent location.

No study,  including the DEARS can be expected to meet all needs, especially when many of
those needs were not a part of its original study design. Even so, this workshop pointed out the
value of conducting intensive, high impact exposure monitoring efforts like  the DEARS to meet
a variety of unforeseen exposure-related needs.  Part 2 workshop  attendees were thanked for
their overall contributions in developing all of the  materials for both events. They were invited
to continue independent discussions following the conclusion of the workshop event with the
DEARS Principal Investigator on  opportunities for collaborations associated with targeted data
gap needs.   While NERL  has been consistently releasing DEARS  data to all collaborators as
needs have been established, it is adjusting its database management priorities to target public
release of the DEARS data within the 2011-2012 time frame.

6.     Workshop Conclusions

The information obtained  by NERL during  the Workshop events  has provided the means to
examine planned data analyses  and the development  of other key  research products for their
overall value to the Agency. This will  allow for the  establishment of a more refined DEARS
publication plan. It has  also provided keen insight as to specific data analysis questions which
should be incorporated into such efforts. This has the potential for enabling  future data analyses
to target not only a given research area, but specifically focus on  exposure science-related issues
of high relevance to those involved in rule making or supporting risk assessments. In addition,
the Workshops have provided the means to effectively describe the original purpose and potential

-------
value of DEARS to the Agency.  This has resulted in enhanced communication between the
NERL and those interested in  developing new  collaborative research  efforts  involving the
DEARS. It is anticipated that a number of additional collaborations involving data analyses and
targeted  peer-review  manuscript  development  will  be  established  as  a result  of  this
communication.  One additional  aspect  of this  communication will be the development of a
summary report (this document) detailing the Workshop events and summary findings. As such,
NERL and non-NERL participants alike will have documentation pertaining to the identified
exposure  data gaps  and the common  needs  identified  by the various  Workgroups. This
information will be of value to each of them as they conduct their own future research planning
activities.
                                          10

-------
                          Attachment 1-Draft Research Questions
    Draft Research Questions


           DEARS Workgroup
        meeting August 31st, 2010
  Assessment and Uncertainty

       What pollutants are of greatest interest to
your organization?
       Is there a need to establish pollutant
relationships (temporal and spatial)? If so, what
are your interests?
       Are there methodological considerations
regarding uncertainty (data collection/analysis)
that your organization is concerned about?
  Assessment and Uncertainty

       Are concerned about the impact of non-ambient
sources of pollutants and the uncertainty of using just
ambient-based measures for risk assessment? If so, what
pollutants and their non-ambient sources concern you?
       What data do you feel is currently lacking in the
published literature concerning pollutant concentrations,
pollutant sources and their impact upon human
exposures. Where are you having to use assumptions in
your research?
  Assessment and Uncertainty

       What multi-pollutant uncertainties confront
your organization? What combinations are of
greatest interest?
       Concerning pollutant sources, what
information is needed to confront science areas of
greatest interest? Are you interested in non-
traditional source categories (residential indoor
sources, local,  non-NEI sources)?
                                            11

-------
  Assessment and Uncertainty

• Are there specific source categories that
you feel deserve special Agency attention at
this time (near-roadway, airports, power
plants)? If so, within this category, what
represents the greatest unknown (e.g.,
general pollutant concentrations and
gradients, impact on surrounding
neighborhoods,  spatial/temporal variability)?
  Assessment and Uncertainty

       Is there a need within your organization for
information on environmental and human
exposure factors? If so, what are they and how
would such information be useful to you?
       How would access to actual measurement
data be useful to you in moving specific research
areas forward in your organization?
     Health and Epidemiology

       What critical data needs exist following the most
recent ISAs for NOx/SOx, O3, and review for PM. Is there
a real need to understand epi implications of PM size
fractions, PM constituents, PM sources? Are there other
pollutant classes that need to be addressed (e.g., VOCs,
SVOCs.carbonyls, carbon)?
       Is it important to understand the impact of non-
ambient pollutants (or sources) or just the ambient upon
human health?
     Health and Epidemiology

       What multi-pollutant mixtures are of
concern? Are there still underlying issues of
surrogacy or confounding with certain pollutants
and PM and resulting health effects?
       Are there location (urban,suburban,rural)
specific effects that need to be explored?
                                                 12

-------
      Health and Epidemiology

        PMcoarse would appear to be a
 relativelyheterogeneous mass in some airsheds. Ambient
 monitoring would appear to be a very poorsurrogate for
 total PMcoarse exposures. Is there an interest in trying to
 establish PMcoarse health effects at this time?
        Is there an interest in trying to establish some
 marker of ambient PMcoarse (fresh) versus PMcoarse
 mass being retained within residences (resuspended)?
 Does it really matter?
      Health and Epidemiology

       Would you expect to see any impact of VOC or
SVOC (PAH) exposures upon certain cardiovascular
endpoints?
       Do source categorized epi findings (regional vs
mobile vs industrial, etc) provide value to issues you are
dealing with?
       Are epi outcomes associated with daily measures
all that we need to be concerned about or should finer
resolution (hourly) be investigated? If so, how could
findings related to shorter time domains support existing
24-hr based standards?
    Environmental and Atmospheric
                Modeling

       What measurement data inputs are needed
to support new or improved CMAQ modules? Is
there any particular time domain or spatial scale
that is particularly needed?
       Are there specific CMAQ modules or output
in need of evaluation versus actual physical
measurements?
   Environmental and Atmospheric
   Modeling

       What further integration of CMAQ and the
SHEDS (Stochastic Human Exposure and Dose
Simulation) is needed?
       Are there other (non-CMAQ) models of
interest needing evaluation?
       If you could design a field study meeting
your most critical data measurement need, what
would that study involve (location, duration, time
resolution, grid resolution, pollutant selection)
                                                  13

-------
   Environmental and Atmospheric
               Modeling

      What multipollutant mixtures are of the
highest interest?
      Is there potential to use micro-scale source
information (eg., presence of local gas station,
freeway distances) relative to neighborhood-based
pollutant measurements in existing or future
models? If so, what type of data would be
needed?
    Environmental and Atmospheric
               Modeling

       Is there environmental factor information
(meteorology, seasonality, etc) on a neighborhood
scale that would benefit current or future modeling
efforts? If so, what inputs would be needed?
       How small do we go relative to grid size? Is
there information available that would indicate
grids below 1 km might be needed with respect to
risk assessments in certain localities? If so, what
pollutants?
    Human Exposure Modeling

      Where do we stand relative to air
toxics modeling? What inputs are needed
and at what spatial and temporal domains?
      What pollutant information is needed
to develop a PMcoarse model?
    Human Exposure Modeling

      What would appear to be the most critical
human exposure factors needed as inputs to
updating the PM-SHEDS? Is there the potential
for developing SHEDS-VOC or some other
variant?
      Are there other human exposure models
that need evaluation using physical data
(PNEM?). Who would be the stakeholder?
                                              14

-------
    Human Exposure Modeling
        Is there a need to collect actual GPS-based
time/location information as an input into human exposure
models? Would coincidental exposure measurements also
be needed (at whattimedomain)? If so, which pollutants
are the most critical at this time?
        How do we enhance the integration of human
exposure models and atmospheric models? Whydo such
attempts provide value to risk assessment? What type of
data is needed to move some epidemiologists away from
using only ambient-based monitoring data instead of
integrated human exposure/air quality models in their
studies.
    Human Exposure Modeling

        Indoor air often contains concentrations of certain
pollutants well above ambient levels (e.g.,VOCs,
carbonyls). Are there critical data analyses needed to
support indoor air quality models in such instances? If so,
for what pollutants?
       What inputs are needed to further
develop/evaluate models like the EMI (exposure model for
individuals)?  What utility do such models provide to end
users? How can they be applied to risk assessment or
basic epi research?
                                                15

-------
             Attachment 2-Data Analysis Workshop Part 1 Workgroup Members

        Detroit Exposure and Aerosol Research Study (DEARS)

                          Data Analysis Workshop-Part 1
Workshop Responsibility
       Workgroup
Members
Workshop Steering Committee
Ron Williams (NERL), Val Garcia (NERL), Karen Wesson
(OAQPS), Jennifer Richmond-Bryant (NCEA), Rich Cook
(OTAQ)
Exposure Assessment and Uncertainty
Steve  McDow*  (NCEA);  Jennifer  Richmond-Bryant
(NCEA); Quingyu Meng (NCEA), Gary Norris (NERL),
Alan Vette (NERL)
Human Exposure Modeling
Janet Burke*  (NERL), Michael Breen (NERL),  Stephen
Graham (NCEA), Tom Long (NCEA),
Mark Morris (OAQPS), Lindsay Stanek (NCEA)
Environmental and Atmospheric Modeling
Health and Epidemiology
Karen Wesson*(OAQPS),  Deborah  Luecken  (NERL),
Wyatt Appel (NERL), Val Garcia (NERL-AMAD), Brian
Eder (NERL)

Lisa Baxter* (NERL), Ron Williams  (NERL), Joe Pinto
(NCEA), George Bollweg (Region 5),  Motria Caudill
(Region 5), Tom Luben (NCEA), Rich  Cook  (OTAQ),
Morta Fuoco (Region 5)
* denotes discussion champion
                                          16

-------
                  Attachment 3-Data Analysis Workshop Part 1 Agenda

        Detroit Exposure and Aerosol Research Study (DEARS)

                          Data Analysis Workshop-Part 1

                          EPA- RTP Campus-room C112

                                 October 19, 2010

                                   1:00-4:30 pm
Agenda
Welcome and introductions

Workshop overview

DEARS background
and results to date

Break (10 min)

Workgroup charge
and member introductions

Common research
data gaps

Uncertainty and assessment

Human exposure modeling

Break (10 min)

Atmospheric modeling

Health and epidemiology

Part 2 and next steps

Adjournment
Tim Watkins (NHEERL Acting Division Director)

Tim Watkins (NHEERL)

Ron Williams/Janet Burke/Gary Norris (NERL)




Ron Williams (NERL)


Ron Williams (NERL)


Steve McDow (NCEA)

Janet Burke (NERL)



Karen Wesson (OAQPS)

Lisa Baxter (NERL)

Tim Watkins (NHEERL)

Tim Watkins (NHEERL)
                                        17

-------
             Attachment 4-Data Analysis Workshop Part 1 Invited Attendee List
NERL Participant
Wyatt Appel
Tim Barzyk
Lisa Baxter
Sarah Bereznicki
Michael Breen
Janet Burke
Fred Dimmick
Rachelle Duval
Brian Eder
Roy Fortmann
Val Garcia
Andrew Geller
BJ George
Davy da Hammond
Kristin Isaccs
Kasey Kolvacik
David Kryak
Deborah Luecken
Gary Norris
John Offenberg
David Olson
Linda Sheldon
Alan Vette
Jon Sobus
Don Whitaker
Ron Williams
NCEA Participant
Jennifer Richmond-Bryant
Stephen Graham
Tom Long
Tom Luben
Steve McDow
Quingyu Meng
Mark Morris
Joe Pinto
Lindsay Stanek

NHEERL Participant
Lucas Neas
Tim Watkins

OAQPS Participant
Karen Wesson

OTAQ Participant
Rich Cook

Region 5 Participant
George Bollweg
Motria Caudill
Marta Fuoco

OAR/ORIA Invitee
Laura Kolb
                                            ACE Interim Program Director Invitee
                                            Dan Costa
                                        18

-------
Attachment 5-DEARS Study Objectives Progress Report Presentation
         2010 DEARS Research
                   Update
 EPA's Exposure Research
                                                             Provide the link
                                                             between
                                                             source/regulatory
                                                             monitoring and
                                                             health outcomes
                                                             Goal:
                                                             Evaluate/model
                                                             the relationship
                                                             between PM at
                                                             ambient sites and
                                                             personal exposure
        Exposure Research Focus
     Determine the relationships between PM,  and select air
     toxics measured at the community, residential outdoor,
     residential indoor, andpersonal scale,
     Determine the chemical and physical factors that
     influence these relationships,
     Determine the human factors that influences these
     relationships,
     Develop andimprove human exposure models,
     Perform source apportionment using pollutant and
     exposure factor data inputs, and
     Determine the relationships between PM and criteria
     pollutant gases as potential confounders or surrogates of
     exposure
         The DEARS

Detroit Exposure and Aerosol Research
Study
 - Multi-pollutant design
 - Multi-seasonal in scope
 - Fully powered (1200 participant days)
 - Randomized participant recruitment
 - Integrated exposure and survey data collection
 - Varied housing stock
 - Varied monitoring settings
 - Varied spatial monitoring (census areas)
                                               19

-------
           DEARS Monitoring Design
          •  3 year field study initiated in July 2004 and
             completed on February 25, 2007
          •  Randomized household purposeful study design
          •  Non-smoking with no health or vocational
             exclusion
          •  Each year (winter/summer) had - 40 enrollees
          •  Households were monitored for 5 days in winter
             and 5 days in summer (-1200 total sampling
             days)
          •  - 36,000 individual 24-hr environmental records
          •  Concurrent monitoring at:
             - Central community site
             - Residential outdoors and indoors
             - Personal
                                                Detroit Was  Selected Because...
                                                •  Was an non-attainment area for PM? 5
                                                •  Projected non-attainment status after sulfur
                                                  reductions in 2010
                                                •  Large number of industrial point sources
                                                •  Heavy mobile source impact including diesel
                                                •  Potential for pollutant spatial variability
                                                •  Possibility of summer and winter season
                                                  variability
                                                •  Historic Speciation Trends Network site and
                                                  National Air Toxics Network Site data
                                                •  State and local interest
                                                •  Existing community partnerships
  vvEPA     Was Detroit a Good Choice?
      We believed...
• PM2.5 concentrations > 15 ug/m3-
• Local impact of industrial sources-
• Mobile source impact of certain  •
 areas
• Spatial variability of pollutants   •
 might exist
- Seasonal effects might exist for  •
 some pollutants
• Historical data would be useful   •
• MDEQ and others would be     •
 interested
• Local community groups might   .
 participate
       Data is showing...
> Some means above annual standard
> CRUISER data shows this potential
> Southfield freeway has obvious impacts
  from nearby roadway
* VOC, NO2, PAH, elemental variability
  observed
»• Acrolein, nitrate, sulfate, VOC, ozone,
  PAH seasonality
* Provided means to cross-check
> Multiple collaborators and partners
  have been established
•• Local groups formally involved in
  recruiting
                                                      DEARS Study Sites
                                                  (Emphasis on Proximity to Sources)
I
I-Industrial
 -Industrial (™t™.ni,
 -Diesel
 -Traffic/ Industrial
5-Industrial
 -Highway
.'-Regional
                                                            20

-------
      DEARS Measurements
Air Exchange Rate
   DEARS Data Attempted to Capture Ongoing Personal
   Exposure, Local and Regional Air Quality
                                             Methods Development
 Use of Novel Passive or Active
 Samplers
                    •25VOCs (aromatics/HCs
                    (9) + halogenated HCs (16))
                    • 3 carbonyls
                    •Continuous PM2,
                    •O3, NO2, SO2
Indoor & Outdoor Monitoring
                     Matched to
                     personal and
                     ambient
                     instrumentation
                                                                                   &
                                        21

-------
    Central Community Site
             Monitoring
               •Community-based
               monitoring at Allen
               Park, an MDEQ air site
               central to the study area
 Environmental and Human Factor Surveys

  .   t
  Jfc*
     Data Calendar of Events

Completion of field data collection (March 2007)
Recovery of final season of raw data from
contractor (July 2007)
Primary DEARS datasets released to
collaborators in late 2008 (ongoing).
Completion of season 6 XRF laboratory analyses
(Fall 2009)
Organic markers laboratory analyses (select
samples from select seasons)  anticipated to be
completed in 2010
APM 159 completed (September 2010)
  DEARS Data Analysis Objectives

 Six primary objectives:
 1.   Determine associations between concentrations at
     central site with those from outdoor and personal
     monitoring
 2.   Determine the factors affecting the mass
     concentration relationships described above
 3.   Identify human activity factors influencing personal
     exposures
 4.   Improve and evaluate models used to estimate
     residential and personal exposures
 5.   Investigate and apply source apportionment models
     to determine contributions from specific sources
I 6.   Determine the associations between ambient
     measurement of criteria pollutant gases (like ozone)
     with personal exposures to these gases, PM
     constituents and other air toxics.
                                            22

-------
       Data Analysis Phases
The DEARS data analysis plan had a four tiered
   analysis structure. These are:
1.  Performing descriptive statistics, searching for
   anomalies as a means to validate individual
   datasets, and establishing the relationships
   between various spatial measurements
   Use of various modeling approaches to
   integrate factors and ancillary data influencing
   the relationships established above
   Data from 1 and 2 above will be integrated into
   PM and air toxics human exposure modeling
   development.
   Data from 1 and 2 above, along with original
   data will be used to perform source
   apportionment modeling	
 Progress/Results Format
   1. Progress of overall effort
   2. Results
   3. Examples of findings
     Objective 1 Progress
 Validation and examination of raw data for
 outliers
 Data release (ongoing) to all
 stakeholders/collaborators since 2008
 (State of Michigan, SEMCOG, U of
 Michigan, Region 5, NERL-HEASD,
 OAQPS, OTAQ, NERL-AMAD, NCEA,
 EOHSI, NHEERL. Specific datasets
 released to meet needs.
 Public release scheduled for FYl 2
    Objective  1 Progress
Williams et al. Study design publication
with univariate analyses of PM, gases,
demographics and recruiting components
Thomburg et al. Publication of coarse PM
spatial and temporal relationships
Rodes et al. Publication of PM2.5 Spatial
and temporal relationships
George et al. Publication of community
versus outdoor PM2.5 relationships
manuscript
Wheeler et al. Publication of DEARS and
WOEAS PM relationships
                                         23

-------
    Objective  1 Progress
Vette et al. Pending draft of VOC spatial
and temporal relationships (Winter 2010).
Philips et al. Publication of DEARS
recruitment/retention strategy.
Stevens et al. Community and outdoor
elemental components relationships.
Pending journal review.
Niu et al. Publication of XRF vs ICP-MS
sample analysis considerations.
Johnson et al. Publication of article
describing indoor/outdoor relationships in
the MICA with DEARS input
      Objective 1 Results
Incredibly rich database, high quality and
quantity
Multiple users
Meets the Integrated, multidistiplinary
model
Both spatial and temporal variability
observed. A primary component of the
study design and the reason for selecting
Detroit
PM2 5 mass spatially homogenous outdoors
with respect to mass concentration
      Objective 1 Results
PM coarse mass heterogeneous across
study neighborhood. Noted seasonality in
relationships between monitoring sites
Community N03 monitors poor surrogate
for personal exposures. Many indoor
sources
Indoor VOC sources significantly impact
community versus personal relationship
With exceptions, PM2 5 outdoor elemental
and organic component mass highly
variable by season and location
      Objective 1 Results
Coarse spatial and temporal variability of
PM composition being determined using
highly sensitive ICP-MS methods
Development of methods needed to detect
organic markers in low volume samples
have been developed.
Biomass markers needed for source
apportionment have been quantified.
                                       24

-------
ftgft
1

Variability of PM and Exposure Factors
Metric Days Season

Personal PM2S 874
Indoor PMM 973
Outdoor PMIS 1347
Ambient PMI5 139
exposure factor
PFT air exchange 916 1-6
(h1)
Min Max Mean Std Dev Median 99%
(ug/m3) (ug/m3) (ug/m3) (ug/m3) (ug/m3) (ug/m3)
-0.5 255.3 20.3 20.9 15.6 113.6
-1.3 297.3 13.7 20.1 14.0 91.2
-1.1 85.6 16.3 9.7 14.5 43.4
2.3 66.4 16.9 10.E 14.0 63.0

0.1 17.0 1.5 1.5 1.0 7.5

H -.— -sas»^-- „__
                                                     oEPA     Spatial and Temporal PM Coarse Variability
Spatial and Temporal Elemental Variability
                      ,,  T
,ii!
                                                     sew
                                                                   Distribution of Personal PM2.5 Exposures
                                               25

-------
«>FFyV Impact of Housing Stock and Age on Air Exchange Rate
Home ABC
36-70yrs
71-lUOyrs
>100jrs
AERfhr-1)
N
178
100
9
Mean
1.534
i.syi
1 .701
Sid P5 P25
1.6!M 0.257 0.582
1.119 0.366 0.5SJO
1.67S 0 155 0547

Exposure Monitoring Area
3
66 2.227 1.41/7
42 1.583 1.413
4 69 1 629 2261
5 44 1.281 0.8SW
(, 83 1 199 0871
7 19 O.SW8 1.014
AF.R
0.575
0.374
0 ?"S
-., ;/;.
0.257
PSO P7S


1.063 1.918
1.226 2.41/7
1.219 1.72.1
(lir1)
P2S PSO
1.026 1.844
0.695 1.146
0.568 n.979
0.568 0.954
0.471 0986
0.366 0.552
P75
2.266
1 84V
1.726
1.789
0,92t>
pys

3. 683
3.592
5221
pys
5.221
4.119
4244
3.355
a, ass
4.422
™=L-a==SB5,,-._-~-.-._
                                                              VOC Results - Summer '04 -> Winter '06

 Personal and Ambient NO2 Mass
Concentration Relationships (ppb)
»Hft PM2! Spatial Composition Comparison (P-Values)
Comparison of outdoor sites with ambient monitor (summer/winter)
Llcmcn
<•„_
Potassium
M3tigaii*se
,„
Z,nc
Ud
liilA 1
M «
. 001 ''007

„„, „„
• 001 '001
M „„
liMA 3

049 015


002 • 001

EMA 4

039 048

«,„ ™
,,„•„,
*,«
EMA 5
», M
•001 .'001
««,,«0,
„,„ «,
. 001 .•'• 001

EMA 6
uuw
Oil 0 46

„„,„„.
004'0?S
uuu,
LMA 7

me 038


• 001 005

•Maamb^^M^
                                         26

-------
I Progress/Results-Objectives #2 and 3
  (factors impacting relationships)
      1. Progress of overall effort
      2. Results
      3. Examples of findings
  Objective 2 &3  Progress

Williams et al. Draft manuscript describing
the impact of human and environmental
factors on SO2, NO2 and O3 relationships
Meng et al. Draft manuscript exploring NO2
of ambient origin impact on exposure
Brook, Williams et al. Publication of effects
of short term PM exposures and observable
human health effects (0-4 hrs significant)
Brook, Williams et al. Publication of PM
measures of ambient and non-ambient
exposures versus observable human health
effects. Non-ambient sources impacted
outcomes
      Objective 2 &3  Progress

     Williams et al. Draft manuscript describing
     PM components and gases and their impact
     upon health effect linkages
     Brook, Williams et al. Submission linking
     personal temperatures to alterations in blood
     pressure
     Hammond, et al. Analyses defining short-
     term PM2 5 exposure sources. Will be linked
     to health effects
     Lawless et al. Draft manuscript on personal
     monitoring compliance and why knowledge
     of ETS exposures are critical in reducing
     measurement uncertainty
  Objective 2 & 3 Progress
Isaacs,  Burke et al. Completed report defining
factors influencing residential air exchange.
Being converted to manuscript.

Alion, Burke et al. WA to produce repeat on
residential infiltration parameters for PM
species and air toxics (2010). Subsequently will
be converted to manuscript.

Baxter et al. Publication of diesel impacting
residences near the Ambassador Bridge

Barzyk et al. Publication of methods to assess
near road impacts and preliminary findings.
                                              27

-------
     Objective'.
Current progress:
& 3 Progress
    George, Palma et al. Submission of DEARS
    VOCs andNATA modeling comparison
    George, Whitaker et al. Publication of
    comparison of outdoor PM2 5 to ambient
    monitor and impact of meteorological factors
    Thoraburg, Williams et al. Draft of participant
    compliance and impact on data quality for
    multiple pollutants
    Whitaker et al. Completion of secondary field
    and laboratory work needed to further evaluate
    carbonyl methodological considerations
    impacting DEARS data
  Objective 2 & 3- Results

Diesel emission marker impactedhomes close to
the Ambassador Bridge
 - Barzyk et al. Publication of methods to assess near
   road impacts and preliminary findings.
 - Baxter et al. Publication of diesel impacting residence:
   near the Ambassador Bridge
DEARS provided good agreement for benzene
spatiality versus NATA models.
 - George, Palma et al. Submission comparing NATA
   model predictions for benzene to DEARS
   m easurem ents Meteorologic al factors shown to
   impact both I'M-,.- and I'M coarse spatiality across the
   DEARS areas.
Residential air exchange rate distributions
strongly influenced by temperature and air
conditioning presence
    Objective 2 & 3- Results

  ETS impacted PM, NO, and VOC measures
  Indoor appliances impacted NO2 measures
  resulting in poor relationship between community
  and personal levels. Community levels of S02 and
  O3 reasonable surrogates for personal exposures.
  Indoor VOCs impacted VOC measures.
  Community monitors often poor surrogate for
  personal
  Both PM of ambient origin andPM of non-
  ambient origin had significant impacts on health
  Chronic impact (short duration) impact of PM
  exposures determined. Different health effects
  versus use of 24-hr based data
                                                                                                                 10
                                               28

-------
Effective Distance to Roadway (Southfield Fwy.)

i
1
i
M
•

. —
y


:;
•r~~>-L_. • . ' Concentrations
*^~-~r. (24 -hour)
w«™» "rban ,
background
levels at -250 m
; ;*
tWdk**[M«.M «»K,,,,0,,UnI,W,
                                                 Homes Near the Ambassador Bridge-Source
                                                              Impact
                                                                          Concentrations
                                                                          elevated and more
                                                                          variable when
                                                                          winds are from the
                                                                          Bridge
                                                        Wind 0-r.t! ro ii
 Progress/Results-Objective 4
| (human exposure modeling)

     1. Progress of overall effort
     2. Results
     3. Examples of findings
      Objective 4-Progress


Hammond, Burke et al. DEARS PM
coarse mass and components data being
targeted as planned inputs to SHEDS.
Laboratory data currently being pursued
Hammond et al. ME and human exposure
data analyses nearing completion. Planned
input for SHEDS and article preparation on
indoor and personal sources
                                                                                          11
                                      29

-------
                  Objective 4-Progress
           Burke et al. DEARS residential indoor/outdoor
           data being used to evaluate SHEDS modeled
           exposures to multiple air pollutants
           Breen et al. Using DEARS data to further
           evaluate residential air exchange model in EMI
           Breen et al. Using DEARS data to further
           evaluate EMI before developing EMI-based
           exposure metrics for NEXUS
                                                              Objective 4-Progress
                                                       George, et al. DEARS data used in evaluation of
                                                       NATA modeled results. Article in review.
                                                       Wesson et al (OAQPS). 2005 DEARS PM and
                                                       select air toxics data used to evaluate 2005
                                                       CMAQ output at 4 1cm and 1 km grid size
                                                       resolution for Detroit area.  All data sets
                                                       developed.
                                                       Appel & Leucken et al (AMAD). Direct input of
                                                       select 2005 air toxics data into CMAQ model
                                                       evaluation tool
L
      Objective 4- Results
Analyses for multiple modeling areas are
currently ongoing
DEARS providing data for improving
SHEDS inputs and evaluating model
output
DEARS data being used for evaluating
EMI for use in NEXUS modeling
DEARS being used as important evaluation
dataset for OAQPS and AMAD air quality
modeling for Detroit
                                                                     Conceptual EMI
                                                                                                              12
                                                    30

-------
        Air Exchange Rate (AER) Distributions
                    for SHEDS

        • Factors influencing DEARS AER
        • Compari son with other recent studi es

                              LMEACH   HT-TECH
               DetoftDEARS   fiJ.RloPA

    Comparison of DEARS and NATA
           Results Tor Detroit
          Ratio of DE4RS to NATA for be
     g  Comparison of DEARS with 2005 CMAQ
DE ARS Census TruMs       ReSUltS fOr Detroit

           .,-     Census Tract PM,j using CMAQ
| Progress/Results-Objective 5
       (source impacts)

     1. Progress of overall effort
     2. Results
     3. Examples of findings
                                                                                               13
                                            31

-------
       Objective 5 Progress
 Current progress:
| •   Completion of initial Allen Park source
     apportionment. Determination of appropriate
     data and models for use in the DEARS.
     Completion of laboratory analyses associated
     with select residential samples (2010).
     Organic markers analyzed and applied in source
     apportionment modeling, improving
     understanding of motor vehicle andbiomass
     sources (on-going).
     Ongoing determination of concentration impact
     on data quality, determination of marker
     selection for inclusion in future (integrated)
     source apportionment associated with EMA
     spatial comparison.
       Objective 5 Progress
I Current progress:
    Baxter et al. published impact of near road
    sources from the Ambassador Bridge.
    Duval et al. Submission of PM sources
    spatiality variability.
    Stockburger et al. Implementation of
    levoglucosan method. Was needed to validate
    biomass contribution to Detroit airshed.
    McDow et al. Selection and observations
    associated with organic markers for source
    apportionment in the DEARS. Source factor
    analyses have been completed
    Vette et al. Determination of the impact of
    sources and environmental factors upon
    multipollutant spatial and temporal variability
          Objective 5- Results
   High biomass source contribution (K
   signature) that agrees with other
   researcher's findings from Detroit.
   Multiple externals have now found similar
   impact. This is a very complicated issue.
   Substantial spatial variability observed for
   organic species (especially PAHs) on some
   days, other days are fairly uniform.  Spatial
   variability appears to be related to patterns
   of abundance (i.e.. diagnostic species
   ratios)
          Objective 5- Results
   Integration of both elemental and organic
   markers needed to improve source
   characterization in Detroit. Multiple
   sources having similar profiles complicate
   the source apportionment process.
   Multiple indoor sources exist.  Cooking a
   major contribution to total PM exposures.
                                                                                                            14
                                              32

-------
Spatial Variability of PM2.S Sources in the DEARS
                               '.Mb 5\erage sou-re
                               contributions (In %)for
                               DEARS Season 3 (Surniner
                               2005) ana 4 (Winter 2006)
                               Values in ng'm3 represent
                               sum of mixed Industrial and
                               steel manufacturing
                               contributions

                                                                            P-values for the Non-parametric One-way Wile oxon Score
                                                                            Test Comparing Variability of Sources in DEARS Ambient
                                                                                      and Residential Outdoor Samples
j.^sBiirniria

rt Wot Vehicles
;te el Manufacturing
YeV
017

-------
SPATIAL VARIABILITY FOR SEASON 1 OUTDOOR SAMPLES
             BENZO [ghi]PERYLENE
July 15, 2004*
                                   ick stripes indicate
                                 Allen Park Central Site
Fraction of Samples with n 25% of ' Most 25%
(n=22)
• Most (but <90%) samples agree with



m
/


|— I

/if1 J9 *$
&• .«p j«*
-? *$* Jr
^F A* 
-------
  Progress/Results-Objective 6
  (nuiltipollutant  relationships)

     1. Progress of overall effort
     2. Results
     3. Examples  of findings
     Objective 6 Progress
Current progress:
•   Canonical cm-relation analyses of select VOC
   data and human/environmental factors
   completed by EOHSI. Used as input in George
   et al manuscript and input into planned Vette et
   al. manuscript
•   Vette etal.,  completion of VOC and
   copollutant APM report. Basis for planned
   manuscript
•   Williams et  al. Completion of mixedmodel
   evaluation of factors influencing NO2, S02 and
   O3 relationships. Draft manuscript developed
      Objective 6 Progress
| Current progress:
 •   Meng, Pinto et al (NCEA). Modeling of
    PM and criteria gas copollutants and
    human/environmental factors. AAAR
    presentation developed into draft
    manuscript.
    Multipollutant impact upon health
    outcomes. Draft manuscript completed by
    Williams, Brook, et al.
        Objective 6-Results
  Significant indoor and personal exposure
  factors influencing NO2 spatial
  relationships.
  Significant NO2 spatiality observed in
  various neighborhoods-suspected near road
  influence
  ETS a major confounder of personal NO3
  and VOC'data.
  NO2 was detected in a high percentage.
  Detection limits impacted some personal
                                                                                                 17
                                         35

-------
s=,Ef*
' 	 Variability of Criteria Gas Concentrations
Metric
Personal
°3
Ambient
°3
Personal
NO2
Outdoor
NOj
Ambient
NOj
Personal
SO2
Ambient
SO.,
Days
936
199
891
989
198
928
195
Seasons
1-6
1-6
1-6
3-6
1-6
1-6
1-6
Win
(PPb>
-5.4

-3.0
-1.0
0.0
-4.0

Max

43.2

474.0
65.0
48.0
12.4

Mean

2.1

29.9
20.3
23.9
0.5

STD
(PPb)
3.3

37.4
9.7
8.2
1.7

Median

1.0

23.1
20.0
23.1
0.0

99%
(PPb)
15.0

188.0
48.0
47.5
6.0

— =«=«—
                                                                       v>EFV\   Examples of Exposure Survey Variables
                                                                            • season
                                                                            • air exchange
                                                                            • time spent in activities
                                                                            • outdoor temperatures
                                                                            • ets influence
                                                                            • use of fans
                                                                            • use of air conditioning
                                                                            • presence of open windows
                                                                            • cleaning episodes
• cooking episodes
• age of home
• proximity to roadways
• attached garages
• cooking fuel type
• use of clothes dryers
• volume of home
• use of candles
• use of incense
I
 Impact of Human and Environmental Factors on Personal NO2 Exposures
Indoor VOC Sources Impact Personal Exposures
V'
4'-
•-••
\,,
'-,'
• - Um

• -

1 ••
!•.-
00
*-.
• .
=£=
• ^


* li^

• ,**
0
1

1

*
• .i*

• ..

_ 4{-.}-.r . 	
".»*

                                                                                                                                         18
                                                                36

-------
     Multipollutant Comparisons
Correlation with Personal NO
                        Correlation with Indoor EC
  Indoor EC and personal Hu weaMy correlated
  Personal MSATs poorly correlated wrth personal NO..; slightly better
  with indoor EC
  Little seasonal difference observed
                                                    Progress/Results-Source to Effects
                                                    1. Integration of Cardiovascular Health Study
                                                    2. Results
                                                    3. Examples of findings
 In Home Health Data Collections
                                                                    Health Measures
                                                    4 hour fast prior to measures and reduced physical
                                                    activities. Resting state obtained prior to starting.

                                                    Systolic blood pressure (SBP)- Omron 780 monitor
                                                    Diastolic blood pressure (DBF)- Omron 780 monitor
                                                    Heart rate (HR> Omron 780 monitor
                                                    Brachial  artery diameter (BAD)- lOmHzTerason 2000
                                                    ultrasound
                                                    Flow-mediated dialation (FMD)- lOrnHz Terason 2000
                                                    ultrasound
                                                    Nitroglycerin-mediated dilation (NMD)@0.4 mg
                                                    sublingual
                                                                                                           19
                                              37

-------
Health Effects Versus PM of Non-Ambient Origin
                   0.23 (0.12) fl)=O.Q73)
                   -0.24(Q.14Hp=0.096)
                                   Vest compliance (=-60%)
                                                                     Exposure and Heart Rate
                                                                  i

                                                                T
Multipollutant Effects on Health Outcomes
Model
PFe
PK
PFe
PK
PK
AFe
AZn
PNOj
PNO,
PN02
PK
AK
PNO;
Outcome
SBP
SBP
DBP
DBF
DBP
HR
HR
BAD
BAD
BAD
BAD
BAD
FMD
L>g
0
2
0
0
0
0
0
2
1
2
0
1
0
Gnif
All subjwts

Allsyhjwts
All subjects
V«t
Ailsuhjwts
AUsubjwls
AU$ubjert$
Vest
Vest
Attsubjects
AUsuhjwts
AUmltjefts
Risk
0.0128

0.0084
-0.0 124
-0.0447
-0.0163
0.0296
-0.0049
0.0041
-0.0067
-0.8007
0.0037
0.0468
SE
0.0054

0.0036
0.0056
0.0132
0.0065
0.0145
0.0023
0.0019
0.0029
0.0003
0.0017
0.0174
Rvalue
0.018?

0.0 188
0.0305
0.0016
0.0126
0.0429
0.0366
0.0353
0.0256
0.0442
0.0 3 SO
0.0079
                                                                        Data Analysis Perspective

                                                                    DEARS data analyses are primarily being
                                                                    performed between other maj or tasks by a small
                                                                    number of researchers. Competing time demands
                                                                    restricting progress in some areas
                                                                    Even so, progress across a wide number of
                                                                    objectives has being maintained.
                                                                    A large number of manuscripts are under journal
                                                                    review or draft development stage
                                                                    Objectives 1.2, and 6 have yielded majority of
                                                                    publications to date. Significant laboratory work
                                                                    in support of objective 5 has been performed.
                                                                    Modeling obj ectives ramping up.
                                                                    Large number of collaborations ongoing (internal
                                                                    and external)
                                                                                                                   20
                                                    38

-------
Attachment 6- Exposure Assessment and Uncertainty Workgroup  Summary
                                     ASSESSMENT & UNCERTAINTY
                                          Overall Summary

                           1) We need data on lead sooner than anything else

                           2) Next we need data and interpretation that will help us
                           to develop the multipollutant science assessment
                     What pollutants are of greatest interest to your organization?

                   In order of priority:

                   1) Lead is most urgent as the next ISA.
                   2) More generally, PM and ozone the most important risk drivers.  For
                      PM, specific components, size ranges, and sources are important,
                      especially elemental carbon,  metals, organics. coarse PM, fien PM;
                      ultrafine PM, traffic, secondary (ozone + SOA), industrial sources in
                      non-attainment areas (e.g. steel manufacturing).
                   3) Remaining criteria pollutants for which ISA's will be written.  (For
                      example, DEARS data on NO2 outdoor and personal exposure and
                      ozone  concentrations at Allen Park and from personal exposure
                      monitors would be useful.)
                   4) Surrogates of important multipollutant mixtures (e.g. EC or NOx for
                      traffic)
                   5) HAPs.
                                               39

-------
Is there a need to establish pollutant relationships? If so, what are your
   interests?

Yes, especially for multi-pollutant AQM. The key questions are:

1) what pollutants are can be grouped in reduced for epidemiological and
   exposure approaches?

2) what pollutants or groups of pollutants are the best surrogates for regional or
   near source exposure environments? 3) What pollutants or groups of
   pollutants have similar origin and are likely to benefit from the same control
   strategy?
   Are there methodological considerations regarding uncertainty
   data that your organization is concerned about?

   Yes.  Using network data for exposure assessment and
   epidemiology with increased reliance on using models instead of
   measurements to achieve more complete spatial and temporal data
   makes high quality measurements even more important.  Better
   quality measurements would be especially useful for elemental
   carbon.

  Are you concerned about the impact of non-ambient sources of
  pollutants and the uncertainty of using just ambient-based
  measures of risk assessment? If so, what pollutants and their
  non-abmient sources concern you?

  Only to diffrentiate ambient from indoor generated pollutants. Where
  data has been obtained, this is useful, but it is not as high a priority as
  understanding indoor penetration or spatial and temporal variability of
  pollutants from outdoor sources. The outdoor sources are of most
  importance to us.
                                  40

-------
   What data do you feel is currently lacking In the published literature
   concerning pollutant concentrations, pollutant sources and their
   impact upon human exposures. Where are you having to use
   assumptions in your research?

   PM component variability, PM size class variability, PM source
   composition, associations between pollutants,  indoor penetration, human
   activity patterns, intercity differences in both exposure factors and source
   apportionment, and the relationship of each of these with meteorology and
   land use; better tools for exposure modeling, air quality modeling, source
   apportionment, and hybrid and mixed models.
What multi-pollutant uncertainties confront your organization? What
   combinations are of greatest interest?

1) what pollutants can be grouped in reduced form epidemiological and
   exposure approaches
2) what pollutants or groups of pollutants are the best surrogates for regional
   or near source exposure environments?
3) What pollutants or groups of pollutants have similar origin and are likely to
   benefit from the same control strategy? The combinations of greatest
   interest are: traffic (PM, ultrafines, EC,  NOx, CO, BTEX, butadiene,
   aldehydes), secondary (SOA, sulfate, ozone), Eastern regional background
   (PM, SO2, NOx), urban industrial (PM,  EC, metals)
4) Understanding the relative importance of "exposure measurement error" in
   a multi-pollutant context for the broad classes of "exposure measurement
   error": 1)  instrument measurement error, 2) detection limits, 3) spatial
   misalignment, and 4) other discrepancies between monitor and exposure
   (e.g. activity patterns, micro-environment). Which uncertainty dominates
   and how does their relative importance vary between species?
                                  41

-------
   What multi-pollutant uncertainties confront your organization?
   What combinations are of greatest interest? - Continued
 5) Providing data that will help to understand interaction terms between pairs of
 pollutants relevant for epidemiological studies.
 6) Providing data that will help with data reduction for multipollutant
 epidemiological studies, development of optimal pollutant groupings by source,
 health outcome, and mode of action; and selection of optimal surrogates to be
 measured.
 7) Providing information for testing spatial and temporal interpolation
 approaches, land  use regression, air quality modeling inputs, human exposure
 models.
 8) Evaluate exposure model performance across pollutants to determine which
 are done well, and which are more challenging.
 9) Data that will help in the evaluation of synergistic effects in combined
 exposure to multiple pollutants.
Concerning pollutant sources, what information is needed to confront
science areas of greatest interest? Are you interested in non-traditional
source categories (residential indoor sources, local, non-NEI sources?)

Continued development of improvement of source apportionment models,
including better characterizing and reducing uncertainty, incorporating
meteorological data and compositional restraints into models, development of
routine practical evaluation procedures for model results, including uncertainty
analysis.

Are there specific source categories that you feel deserve special Agency
attention at this time (near-roadway, airports, power plants)? If so, within
this category, what represents the greatest unknown (e.g., general
pollutant concentrations and gradients, impact on surrounding
neighborhoods, spatial/temporal variability)?

Near-roadway, power plants, industrial sources of PM high in metals and EC,
urban activity such as demolition and construction.  Airports are a specific issue
for lead, but from DEARS are not designed for that purpose.
                                  42

-------
   Is there a need within your organization for information on
   environmental and human exposure factors?  If so, what are they and
   how would such information be useful to you?

Yes, to reduce exposure misclassification. Better information is needed on
   spatial and temporal variability, indoor penetration, and exposures of Pb of
   ambient origin, all criteria pollutants, ultrafine PM, coarse PM, EC,  metals,
   sulfate, nitrate, SOA, as well as human activity patterns and activity levels in
   different exposure environments, including indoor-home, indoor-work, indoor-
   school, commuting (in vehicle and otherwise).  More reliable ventilation and
   building related data would be useful.  Results of importance for
   environmental justice evaluation are especially needed.

How would access to actual measurement data be useful to you in moving
   specific research areas forward in your organization?

It would allow us to characterize concentration ranges and spatial patterns
   relevant to exposure, and 2) it would allow us to pose and investigate
   hypotheses related to exposure and  receptor modeling relevant for
   understanding  exposure to sources.
                  Suggested Questions for Workshop:
  1) What are the patterns and causes of spatial and temporal variability in
  outdoor concentrations of lead, PM25, Ozone, NOX, ultrafine PM, coarse PM,
  NOX, EC, and metals?

  2) What are the relationships between these species, what factors
  influence them, and what are their common sources?

  3) What are the most important exposure factors for these species and how
    do they differ among species? (e.g., spatial and temporal variability, indoor
    penetration, human activity, breathing rate, meteorology, built
    environment, others?)

  4) What urban and regional sources are the most important contributors of
    exposure to these species?

  S) How can DEARS data help to develop a better multi-pollutant air quality
  management strategy?
                                   43

-------
Attachment 7- Human Exposure Modeling Workgroup Summary
           Human  Exposure Modeling
         Data needs for:
         - Single pollutants
         - Multi-pollutant
           relationships
         - Individual sources
           or relative impacts
           of sources
         - Other (specific to
           human exposure)
Scale
Purpose
- NAAQS ISAs
- NAAQS
  Exposure & Risk
  Assessment
  (APEX model)
- NATA(HAPEM
  model)
                           Spatial scale:
                    national/regional/state/county/intra-urban
                           Temporal scale:
                       annual/seasonal/daily/hourly















Data Need
IN.IIll




NAAQS ISA
& Esp/Risk
Assess.






* for Single
ants?

03

PM^ mass
and species
nitrate, sulfate,
EC, OC, arid


PH0«« mass



TYPE
Amb lent air cone. /Indoorvs
outdoor/ Personal exp / Specific
micro or activity?
Ambient air
Personal exposures for
•4 H i H • "i i (ME & total)
Ambient ar
Indoor/outdoor & personal exp
For evaluation of exposure model
predicticns
NOT -residential micros, vehicles
for inputs
Hi^i exposure activities —near
roadways?
tat**,*
Moorfooldoorfan^oa..^
Personal exp for evaluation


SCALE
Spatial
(national / regional / state-
county / intraurban)
Intra-urbanvaria lity
Intra-urban varia lity
Intra-urbanvana lity
Variability acros regions;
within urban area


Variability across regions;
Variability across regions.




Temporal
(annual / seasonal /
daily/hourly)

Hourly to 8-hour daily
maximum
Daily -hourly
Daily -hourly,
Seasonal
Daily

Daily
Daily , Seasonal
Daily , Seasonal














                             44

-------
Data Needs for Single
Pollutants?
NAAQSISA
&Exp/Risk
Assess.
CO &, NO2
SO2
Pb
TYPE
Ambient air cone /Ihdoorvs
outdoor /Personal exp / Specific
micro or activity'
Ambia'i: air
High concentration MEs on-
road and near roadways
Indoor/outdoor ME factors for
model inputs
Personal exposures for
evaluation itaE&totaD
Indoor sources for model inputs
(as resources p^Tiiit)
Source emission profiles
Source concentration profiles
Ambient air
Personal exposures for
evaluation (ME AtotaH
Possible data needs forlUBK
combined exposure-dose model
used for children and All-ages
Pb model under development
[IUBK model considers
ingestion and inhalation]
SCALE
tfyatial,
(jiaticnal/ regional /state-
county / intraurban)
Intra-urban v ar lability
Intra-urbanvariability, near-
road cone gradient
In tra -urban variability
Intra-urban v ariability
Regional variability
SES variability
By source type (EGUs,
refineries, etc)
Near source gradient by
source type CEGUs,, etc )
In tra -urban variability
Intra-urban v ariability
Intra-urban v ariability
Temporal
(annual / seasonal /
daily / hourly)
Hourly
Hourly
Hourly to Daily
(concentration linked)
Hourly to 8-hour daily
Hourly to Daily
5-minute to hourly
Longer time scales-
aggregate exposure
o,ernu,liple years
Data Needs for Single
Pollutants?
NATA/
HAPEM
(BTEX)
Aldehydes
Other notional
risk drivers?
Other regional
risk drivers?
TYPE
Ambient air cone /Indoorvs
outdoor / Personal exp /
Specific micro or activity?
Ambient air from monitor or
model
Indoa-Mdoorfc,™,,^
Attached garages, other sources
for model inouts
Ambient air
MoaVo^rf™,^
Other residential sources'




SCALE
Spatial
(national / regional / state-
county /intraurban)
Nsh'nsl coverage at census
tract/block resolution
Regional
Regional
tract/block resolution
Regional
Regional




Temporal
(annual /seasonal/
daily /hourly)
monthly, diurnal
Annual, seasonal
monthly
Annual, seasonal,
monthly
Annual, seasonal,
monthly, diurnal
Annual, seasonal,
monthly
Annual, seasonal,
monthly




45

-------
Data Needs for Multip ollutant
Relationships?
NAAQS ISA
&Esp/Risk
NATA/
HAPEM
Ozone and PM
PM^ species
hpoT-mcally nitrate,
sulfate, EC, OC,
and crustal]
PMj 5 and air toxics
CO/NG2/OTWEC
(traffic pollutants)
PM^/NCysO,
Criteria pollutants
and toxics'?1?
TYPE
Ambient air cone. /Indoor
vs outdo or / Personal exp /
Specific micro or activity'
Ambient sir relationships
Ambient air relation ships
Indoor/outdoor relationships
Ambient air relationships
Indoor/outdoor & personal
Ambient air relation ships
Indoor/outdoor relationships
Microenv ironments (e g ,
office, school, vehicles)


SCALE
Spatial
(national / regional / state-
county / intraurban)
'•} anability across regions,
within urban areas
V ariabiltty across regions,
within urban areas

Variability across regions,
within urban areas

Variability across regions,
within urban areas


National1?1?

Temporal
(annual / seasonal /
daily /hourly)
Daily
Datly
"
Daily •"

Hourly to daily
' |. •:•:;. i.'.y seasonal)




Data Needs for Individual
Sources/Relative Impact or
Sources?
NAAQS IS A
& Exp /Risk
NATA/
HAPEM
Point Scwces(e.g.,
power plants,
industry)
t/fobile Sources
Regional vs. local
sources
Indoor/activity
Others for toxics'"
TYPE
Ambient air cone /Indoor
vs outdoor /Personal exp /
Specific micro or activity?
Ambient air
Indoor/outdoor & personal
Ambient air
Indoor/outdoor & personal
Ambient air
r..,,r- 	 1, r.*:,Hrt-ona:
Source strength (linked to
-Ch" :M-)


SCALE
(national / regional / state-
county/intraurban)
Variability across regions;

wSmiStia^as

Variability across regions,

within urban areas


Temporal
daily/hourly)
Daily1??

Daily"

Daily??

(sub-)Hourly, Daily


46

-------
Other Data Needs for Human
Exposure Modeling?
NAAQS ISA
& Esp/Eisk
NATA/
HAPEM
Air exchange rates
Sure ey questions
Human activity data
Air quality
modeling output
Others??
TYPE
'!
Direct measures of
residential air exchange
Non -residential (e g , office.
school, vehicles)
Residential window/door
opening and AC use
Non -residential HVAC (e g
office, school)
Vehicle window and
ventilation use
Source activities
Time spent outdoors at
moderate or greater exertion
for asthmatics (O3, SO2) &
outdoor workers (03)


SCALE
Spatial
(national /regional/
state-county / intraurban)
Variability at all spatial
scales




National to regional
Regional to intra -urban
variability
Regional to local-scale

Temporal
(annual / seasonal / daily /
hourly)
Daily -hourly for model
.np..h'. All :emf"Tral scales
for evaluation
Daily
Daily, monthly
Daily


Daily


47

-------
Attachment 8-Environmental and Atmospheric Modeling Workgroup  Summary
                              Environmental  and
                           Atmospheric Modeling
                  Question: What information from measurement
                  studies are useful in supporting CMAQ model

                  performance evaluation?
                  Response:
                     Ideally, measurements made for the purpose of model evaluation are spatially dense
                     and have a high temporal frequency (e.g. hourly). The spatial density is dependent
                     on the species of interest being measured, but generally air quality model
                     simulations are performed using horizontal grid spacing of 12km or iess.
                     Additionally, measurements should be made in one location for an extended period
                     of time (e.g. a year) in order to assess whether the model is able to capture the
                     seasonal variations in species concentrations that often occur, which is difficult to
                     impossible to do when monitoring equipment is frequently moved to different
                     locations.
                     Measurements of ozone, PM (total and speciated), and other criteria pollutants as
                     well as CMAQ modeled toxics would be useful.
                     In additional to air quality measurements, collocated measurements of
                     meteorological variables (e.g. temperature, wind speed and direction) are also
                     critical to assessing the air quality model performance.
                     Information on any local emissions sources. For example, if the monitor will be
                     impacted by a local emissions source, for which detailed information will probably
                     not be available in the emissions inventory (e.g. parked car, gas station), knowing
                     about this source would be useful.
                                              48

-------
Question: What information would be useful to
help make improvements CMAQ to further allow it
to support exposure modeling?


Response: Information that would help to improve
   CMAQ predicted concentrations on finer
   temporal and spatial scales. In these terms,
   CMAQ would benefit from:

   - continuous/hourly measurements,
   - measurements for extensive periods of time,

   - placement of measurement sites on the
     neighborhood-scale throughout an urban area.
Question: Is there potential use of micro-scale

source information (e.g. presence of local gas

station, freeway distance, etc) relative to

neighborhood-based pollutant measurements?

Response: Yes! When evaluating a model, it is important to
  understand the quality of the model inputs.  Knowing source
  information would allow one to understand the quality of the
  emissions being used in the model. Though it may be difficult to
  report actual source emissions during the measurement period, just
  knowing that the source is there and  it's location relative to the
  measurement site would be useful. The same could be said for
  meteorological inputs. If any meteorological measurements are
  taken, then they could be compared to those being used by the
  model. It is important to determine the difference between poor
  model performance because of inadequate model inputs and poor
  performance because a change should be made to the algorithms in
  the model.
                         49

-------
Question: If you could design a field study meeting

your most critical data measurement needs, what

would that study involve (location, duration, time

resolution, spatial resolution, pollutant selection)?

Response; While CM AQ performance comparisons with ozone and PM tefl us whether CM AQ is
  getting the "right answer", we need more complete set of chemical measurements to determine if
  we are properly characterizing the pathways for ozone and PM formation, i.e. the "right reason".
  Unless we can ensure the latter, we cannot have confidence that emissions reductions will really
  reduce ozone and PM2.5. In addition to a large scale sampling of criteria and toxic pollutants
  focused on evaluating CMAQ predictions, it is critical to have a few supersites that sample a more
  complete set of precursor and product concentrations. In particular:
• Organic nitrates
{including
speciatkm)
• Peroxyacyl nitrates
* NO2 and NO
• Nitric acid

•OH
•HO2
- Organic peroxy
radicals



• Isoprene and other
alkenes
• Oxygenated VOCs
• Biogenic SOA
tracers
• Anthropogenic SOA
tracers
 Question: Are there other models that could
 benefit from the measurement data for
 model validation?

 Response:  Yes, other photochemical
   models, such as CAMx, would also benefit
   from the measurement studies discussed
   here.  In addition,  dispersion models,  such
   as AERMOD, would also find this data
   useful.
                         50

-------
Question: What multipollutant mixtures are
of the highest interest?

Response: In general, ozone and PM (PM25&
  PM10_25) components are of interest. In addition,
  toxic pollutants which have high risk and are
  also prevalent in most urban areas are of
  interest. Until more information is known about
  additivities or synergies of toxic pollutants, a
  wide range of toxic compounds over consistent
  sampling times and locations should be
  collected.
                     51

-------
Attachment 9- Health and Epidemiology Workgroup Summary
          Health and  Epidemiology


                      Overall Summary

       1) We need more research on health outcomes beyond
       cardiovascular and respiratory effects, and mortality.

       2) More studies on susceptible populations is needed.

       3) Improvements in multi-pollutant modeling are needed.

       4) Better characterization of exposure is necessary.
       What links between air pollutants and
      disease have not been well established?

     • Developmental and reproductive
       outcomes
     • Cancer
     • Neurological/CNS outcomes
     • Inflammatory/Autoimmune diseases (e.g.,
       diabetes, rheumatoid arthritis)
     • Development/Progression of disease (e.g.
       asthma) (as opposed to prevalence)
                          52

-------
What potentially susceptible populations
            need further study?
•  Genetics/Epigenetics
•  Preexisting disease (e.g., Diabetes)
•  Obesity/BMI
•  Socioeconomic status
•  Race/Ethnicity
•  Children/Elderly (mechanistic evidence)
  What improvements in multi-pollutant
           modeling are needed?

  Multi-pollutant modeling
  - How can we disentangle the effects of individual
    pollutants from the ambient mixture?
     • Example: Health effects with CO are observed at very
      low levels.  Is the effect due to CO or is CO a marker
      for traffic?
  - How do we model multi-pollutant exposure in
    epidemiological models?
     • Are all pollutants put into the model as individual
      variables?
     • Can pollutants be combined into some type of indicator
      variable?
     • Can pollutants be combined into source categories?
                      53

-------
What improvements in characterizing
        exposures are needed?

Windows of exposure
- Establishing consistent exposure periods
   across studies
- Determining the appropriate windows of
   exposure depending on the health outcome
   and pollutant chemistry and dispersion
   characteristics
Coarse and Ultrafine PM
- Need a better understanding of the spatial
   and temporal patterns
What improvements in characterizing
        exposures are needed?

PM composition (All size fractions)
- Are certain constituents considered more toxic than
   others?
- Is the current monitoring adequate for health studies?
- Characterization of geographic heterogeneity in
   composition
Ambient vs. Non-ambient
- Non-ambient concentrations can be greater than ambient
   levels
- Assessments of non-ambient exposures are needed in the
   context of large epidemiological studies
- Are there differences in ambient vs. non-ambient
   exposures with respect to composition and size distribution
   (for PM)?
                     54

-------
Attachment 10-Common Needs Summary
    Common Needs
Data need
High frequency spatial and
temporal measurement data of
criteria pollutants
03 & PM and multipollutant
measures and establishment of
relationships, including health
effects
High quality meteorology
Macro and micro emission
source identification and
composition
Short time interval pollutant
measures
Nitrogen
cycling/radicals/biogenics
Application of other models
(CAMx; AERMOD)
Lead findings
Multipollutant groupings
PM component findings
Establishment of source
surrogates
HAPs-VOCs, andnon-HAP-
VOCs, PAHs, findings
including health effects
Improved models using
ambient data
Impact of personal & indoor
air quality on ambient
assessment, including health
effects
Human and environmental
exposure factors
Exposure &
uncertainty
/
/
/
/
/


/
/
/
/
/
/
/
/
Atmospheric
modeling
/
/
/
/
/
/
/
/



/



Human
exposure
modeling
/
/

/
/




/

/

/
/
Health &
epidemiology
/
/

/
/



/
/

/

/
/

            55

-------
Common Needs (continued)
Data need
exposure misclassification
Measurement uncertainty
error & exposure
misclassification
PMcoarse data findings
Effect of specific
multipollutant sources upon
susceptible subpopulations
Stationary monitoring of long
duration (consistent location)
Exposure &
uncertainty
/
/
/


Atmospheric
modeling




/
Human
exposure
modeling


/


Health &
epidemiology

/
/
/

56

-------
           Attachment 11-Data Analysis Workshop Part 2 Agenda

  Detroit Exposure and Aerosol Research Study
                     (DEARS)

            Data Analysis Workshop-Part 2

            EPA- RTP Campus-room C112

                November 3 Oth, 2010

                    1:00-3:00pm

                      Agenda


Welcome and introductions      Tim   Watkins   (NHEERL
                            Acting Division Director)

Workshop overview            Tim Watkins (NHEERE)

Part 1 Review                 Ron Williams (NERL)

Research data gaps             Ron Williams (NERL)
and the DEARS

Invitation to participants         Roy Fortmann (NERL)

Adjournment                 Roy Fortmann (NERL)
                         57

-------
  Attachment 12-Overall Workgroup Summaries
        ASSESSMENT & UNCERTAINTY
              Overall Summary

  1) We need data on lead sooner than anything
  else

  2) Next we need data and interpretation that
  will help us to develop the multipollutant
  science assessment
Environmental and Atmospheric Modeling


         Overall Summary Data Needs

1.  Spatially dense and high temporal frequency of
   PM, ozone, air toxics, etc
2.  Speciation of PM associated with saturation
   monitoring
3.  High quantity/high quality meteorological data
   of a given area
4.  Local emission sources
5.  Nitrogen cycling/radicals/biogenics
   measurements
                     58

-------
     Human Exposure Modeling

         Overall Summary Data Needs
1.  Single pollutant concentrations
2.  Multi-pollutant relationships
3.  Individual sources and impacts of sources
4.  Exposure factors (human/environmental)
5.  Wide range of temporal and spatial scales
   measurements
6.  Determinants for PMSHEDS and potentially
   other models in support of the NAAQS
      Health and Epidemiology

                Overall Summary

 1) We need more research on health outcomes beyond
 cardiovascular and respiratory effects, and mortality.

 2) More studies on susceptible populations is needed.

 3) Improvements in multi-pollutant modeling are needed.

 4) Better characterization of exposure is necessary.
                     59

-------
                                                         Attachment  13-DEARS  Publication Plan

                                                                                                             Draft under devetoprn
                                       linked health effects study with the DEARS    Rob Brook
                                                                                                             Pending development by Whitaker
                                       Application of models developed for outdoor
                                       spatsal gradients to evaluate residential

Impact of mobile sources on residential concentrations and role of housing characteristics and a;r exchange
   an exposures of air pollutants                 rate on indoor infiltration of MSAPs         Possbily NCEA
                            vith
                                                                        Thomburg, Rodes
                                                                                                             Published 2009
                                                                                                             Undergoing final EPA .
                                                                                          60

-------
                                         DEARS Publication Plan, continued
Integration of DEARS and Health Heart
Integration of Health and PM components
                                                             61

-------
Attachment 14-Common Needs Priorities
  Common Needs
Data need

High frequency
spatial and
temporal
measurement
data of criteria
pollutants





03 & PM and
multipollutant
measures and
establishment of
relationships,
including health
effects


Macro and
micro emission
source
identification
and
composition


Checks

4










4








4







Output

1 .publications
defining
concentration
variablity
2. database






1 .publications
reporting spatial and
temporal
correlations and
impact of multi-
pollutant scenarios
upon health
outcomes
2. database
1 . publications
defining regional
sources. Articles
defining impacts of
non-ambient
sources, including
ETS
2. database
Need date

Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14






Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14




Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM 714



DEARS
useful
1. limited
2. limited









l.Yes
2.Yes







l.Yes
2. Yes






Products

1 . Continuous PM
measures and
resulting health
associations
2. available






1. Drafts of
multipollutant
impact on health and
factors impacting
exposures to 03,
N02, S02
relationships to
ambient sources
2. available
1 . Articles on
regional and indoor
sources
2. Database of local
sources



Current
progress
1. Short-term exposures and
health impacts under journal
review. Impact of indoor
sources on total exposures to
be drafted in 2011.
Ultimately to be challenged
versus CV outcomes in 2012.
2. Continuous personal &
indoor & outdoor & ambient
PM2.5 database has been
assembled.
1. Three articles under EPA
review
2. available






1. Submitted or drafted
2. available







            62

-------
Data need

Short time
interval
pollutant
measures



HAPs-VOCs,
and non-HAP-
VOCs, PAHs,
findings
including health
effects







PM component
findings










Checks

4






4












3











Output

1 . Publications on
concentration
variability. Pubs on
hourly data from
PM and gases
needed
2. database
1 . Publications on
HAP pollutant levels
and spatial/temporal
relationships.
Factors (sources)
influencing
relationships
important. Impact of
HAPS on health.
2. database



1 . publications on
organic and
inorganic speciation
for all PM size
fractions.
2. database for
model evaluation





Need date

Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14


Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14








Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14







DEARS
useful
1. Limited
2. Limited





l.Yes
2.Yes











l.Yes
2. Yes










Products

1. Temperature
impact on health.
Short-term PM
impacts on health
2. available


1 . Spatial and
temporal
relationships and
factors influencing
these relationships.
Method
considerations in
selecting organic
markers. Use of
organic markers in
source
apportionment
2. in development
1. PM2.5 and
PMcoarse
speciation. Spatial
and temporal
variability and
relationships.
Factors influencing
component
relationships with
respect to proximity
could be performed
2. in development
Current
progress
1. Articles in final journal
review.
2.temp database being
prepared for NCEA



1 . Articles to be submitted in
F Y 20 11. CV health impacts
analyzed in 2011.
2. database fully assembled
inFY2011.








1 . Articles reporting spatial
and/or temporal N03, S04,
EC & PM2.5 elemental data
in process. Other topics
could be examined.
PMcoarse findings to be
summarized in 2011 and
reported in 2012.
2. PM2.5 data available,
PMcoarse available in 2011.



63

-------
Data need

Impact of
personal &
indoor air
quality on
ambient
assessment,
including health
effects
















Human and
environmental
exposure factors






Checks

3























3








Output

1 . Publications on
non-ambient source
impacts to personal
exposures and
resulting health
outcomes.
2. Model evaluation
3. Model
development
4. database














1 . Publications on
factors impacting
use of ambient
monitoring as a
surrogate of
exposure.
2. database


Need date

Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM 714



















Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM 714




DEARS
useful
l.Yes
2.Yes
3. Yes
4.Yes




















l.Yes
2.Yes







Products

1 . articles on
personal&indoor& a
mbient gas and
PM2.5 speciation as
they relate to
ambient monitoring.
Articles reporting
impact of PM of
ambient and non-
ambient origin.
Articles reporting
impact of PM and
gas source origins
on health
2. PMSHEDS
evaluation
3. EMI development
4. available for
PM2.5 and gases





1 . Articles needed
defining factors and
their impact on
personal exposure
for PM, PM species
and gases.
2. Development of
factors database

Current
progress
1 . Article defining personal,
and non-ambient sources on
total exposures published
(20 10). Article on PM
sources and CV health in
press. Article on PM species
and gas impact on health
submitted. Personal, indoor,
outdoor, ambient elemental
relationships under EPA
review. VOC comparisons
drafted. Two articles on
exposures to 03, N02, S02
under EPA review. Article to
be developed in 2011
defining indoor sources,
ultimately to be linked with
CV outcomes (2012).
2. PMSHEDS evaluation
underway
3. EMI to be developed in
2011
4. PM2.5 and gas data
available
Articles drafted on factors
impacting personal exposure
to ambient and non-ambient
sources of 03, N02, S02.
Similar work proposed for
PM2.5 and its components
(2012)
2. gas and PM data and
factor data fully available

64

-------
Data need

PMcoarse data
findings















High quality
meteorology



Nitrogen
cycling/radicals/
biogenics

Application of
other models
(CAMx;
AERMOD)

Checks

3
















2 or less




2 or less



2 or less



Output

1 . Need articles on
spatial temporal
relationships and use
of ambient as a
surrogate of
exposure. Need
articles defining
coarse speciation
and its variability
due to time/space.
Need articles on
personal exposures.
Need pubs on
impact of coarse PM
on health effects
2. database for
model development
1 . Met database of a
given location of
sufficient depth and
duration.

1 . Database needed
as modeling input


1. Database for
model evaluation



Need date

Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM
714











Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14
Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14
DEARS
useful
l.Yes
2. Yes















l.No




l.No



1. Limited



Products

1 . Article on spatial
and temporal
variability.
2. Development of
speciated database












1 . None anticipated




1 . None anticipated



1 . None currently
planned



Current
progress
1. Article published in 2010.
Laboratory work being
completed on speciation
during 2011.
2. Speciated database
developed in 2011.











1 . No action




1 . No action



1 . No action



65

-------

Data need

Lead findings















Multipollutant
groupings










Checks

2 or less















2 or less











Output

1 . Publications on
spatial/temporal
relationships. Impact
of lead
concentrations at
low levels of
interest.
Establishment of
low dose health
response.
2. panel study
incorporating
biomarker recovery
3. exposure database


1 . Publications on
what pollutants
might be grouped to
represent a source.








Need date

Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM 714











Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM 714







DEARS
useful
l.Yes
2. No
3. Yes













l.Yes











Products

2. Articles on
personal & indoor &
outdoor
concentrations and
spatial/temporal
relationships. Article
on variability
between and within
neighborhoods.
2. No panel study
planned
3. Database released
to OTAQ/NCEA.
More work proposed
relating to PMcoarse
speciation
1 . Examination of
pollutants
representative of a
near-road
environment.
Examination of
HAPS that might be
representative of
select source
categories including
organic and
inorganic profiles.
Current
progress
1 .Articles either under
journal review or being
developed. Anticipate 2011
submission.
2. None
3. All available data released.










1 . 2009 pub on near road
VOCs but did not address
primary issue. Article
describing HAPS as source
markers to be submitted in
2011.






66

-------

Data need

Establishment
of source
surrogates



Improved
models using
ambient data


exposure
misclassificatio
n











Checks

2 or less





2 or less




2 or less













Output

1 . Publications on
identification of
surrogates to
represent an
exposure source
category
1. Database




1 . Publications on
determinants
relating erroneous
source identification
or over/under
representing one or
more source
contributions to
health effect
associations.
Specific discussions
on various error
source terms would
be of value.
Need date

Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14

Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14
Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14









DEARS
useful
1. Maybe





l.Yes




l.Yes













Products

1 . Sufficient data is
probably available
to examine this issue
in an exploratory
way.

1. Evaluation of
PMSHEDS.
Development of
EMI for NEXUS
using DEARS inputs
1 . Articles reporting
impact of ETS on
total personal
exposures and
impact of ETS on
observed health
effects for PM of
ambient and non-
ambient origin.
Articles describing
non-ambient source
impacts on N02
exposures.

Current
progress
1 . No action





1. PM2.5 and factor data
needed for PMSHEDS.
PMcoarse component inputs
needed. EMI-NEXUS to be
developed in 2011.
1 . Articles published or in
press concerning ETS
impacts (2010). More work
could be performed in this
area if resources permitted.
Non-ambient N02 source
impactions defined in
multiple articles under EPA
review.






67

-------
Data need
Measurement
uncertainty
error &
exposure
misclassifi-
ication
Effect of
specific
multipollutant
sources upon
susceptible
subpopulations
Stationary
monitoring of
long duration
(consistent
location)
Checks
2 or less
2 or less
2 or less
Output
1. Publications
relating imprecision
of a given
measurement to
accurately represent
an exposure metric
1 . Need pubs
dealing with impact
of copollutants on
various
subpopulations
1 . Database needed
for model input and
involving extensive
monitoring in a
given location
Need date
Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14
Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM
/14
Pb 7/11
03 7/11
NOx/SOx 6/12
Mult Pol 6/12
PM /14
DEARS
useful
l.Yes
l.No.
DEARS
was a
general
population
l.No
Products
1. Ability of an
ambient monitor to
act as an appropriate
surrogate for
personal exposures
to PM, PM
components, gases,
VOCs are being
examined and
summarized.
1. Supported
NEXUS
development
1 . None
Current
progress
1. 2010 publication defining
necessary PM2.5 personal
monitoring to reflect
acceptable error. Impact of
PM2.5 personal monitoring
compliance error on CV
outcomes in press. Articles
onVOCandSVOCerrorto
be reported in 2011.
1 . NEXUS is ongoing study
1 . No action.
68

-------

-------

&EPA
     United States
     Environmental Protection
     Agency
PRESORTED STANDARD
 POSTAGE & FEES PAID
        EPA
   PERMIT NO. G-35
     Office of Research and Development (8101R)
     Washington, DC 20460

     Official Business
     Penalty for Private Use
     $300
                                Recycled/Recyclable Printed on paper that contains a minimum of
                                  50% postconsumer fiber content processed chlorine free

-------